diff options
Diffstat (limited to '3rdparty/lzma/C')
118 files changed, 24225 insertions, 6913 deletions
diff --git a/3rdparty/lzma/C/7z.h b/3rdparty/lzma/C/7z.h index 4768151948d..9e27c015218 100644 --- a/3rdparty/lzma/C/7z.h +++ b/3rdparty/lzma/C/7z.h @@ -1,8 +1,8 @@ /* 7z.h -- 7z interface -2015-11-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __7Z_H -#define __7Z_H +#ifndef ZIP7_INC_7Z_H +#define ZIP7_INC_7Z_H #include "7zTypes.h" @@ -91,14 +91,16 @@ typedef struct UInt64 *CoderUnpackSizes; // for all coders in all folders Byte *CodersData; + + UInt64 RangeLimit; } CSzAr; UInt64 SzAr_GetFolderUnpackSize(const CSzAr *p, UInt32 folderIndex); SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex, - ILookInStream *stream, UInt64 startPos, + ILookInStreamPtr stream, UInt64 startPos, Byte *outBuffer, size_t outSize, - ISzAlloc *allocMain); + ISzAllocPtr allocMain); typedef struct { @@ -131,7 +133,7 @@ typedef struct #define SzArEx_GetFileSize(p, i) ((p)->UnpackPositions[(i) + 1] - (p)->UnpackPositions[i]) void SzArEx_Init(CSzArEx *p); -void SzArEx_Free(CSzArEx *p, ISzAlloc *alloc); +void SzArEx_Free(CSzArEx *p, ISzAllocPtr alloc); UInt64 SzArEx_GetFolderStreamPos(const CSzArEx *p, UInt32 folderIndex, UInt32 indexInFolder); int SzArEx_GetFolderFullPackSize(const CSzArEx *p, UInt32 folderIndex, UInt64 *resSize); @@ -172,15 +174,15 @@ UInt16 *SzArEx_GetFullNameUtf16_Back(const CSzArEx *p, size_t fileIndex, UInt16 SRes SzArEx_Extract( const CSzArEx *db, - ILookInStream *inStream, + ILookInStreamPtr inStream, UInt32 fileIndex, /* index of file */ UInt32 *blockIndex, /* index of solid block */ Byte **outBuffer, /* pointer to pointer to output buffer (allocated with allocMain) */ size_t *outBufferSize, /* buffer size for output buffer */ size_t *offset, /* offset of stream for required file in *outBuffer */ size_t *outSizeProcessed, /* size of file in *outBuffer */ - ISzAlloc *allocMain, - ISzAlloc *allocTemp); + ISzAllocPtr allocMain, + ISzAllocPtr allocTemp); /* @@ -194,8 +196,8 @@ SZ_ERROR_INPUT_EOF SZ_ERROR_FAIL */ -SRes SzArEx_Open(CSzArEx *p, ILookInStream *inStream, - ISzAlloc *allocMain, ISzAlloc *allocTemp); +SRes SzArEx_Open(CSzArEx *p, ILookInStreamPtr inStream, + ISzAllocPtr allocMain, ISzAllocPtr allocTemp); EXTERN_C_END diff --git a/3rdparty/lzma/C/7zAlloc.c b/3rdparty/lzma/C/7zAlloc.c index 3e848c9800d..2f0659af6f7 100644 --- a/3rdparty/lzma/C/7zAlloc.c +++ b/3rdparty/lzma/C/7zAlloc.c @@ -1,78 +1,89 @@ -/* 7zAlloc.c -- Allocation functions -2015-11-09 : Igor Pavlov : Public domain */ +/* 7zAlloc.c -- Allocation functions for 7z processing +2023-03-04 : Igor Pavlov : Public domain */ #include "Precomp.h" +#include <stdlib.h> + #include "7zAlloc.h" -/* #define _SZ_ALLOC_DEBUG */ -/* use _SZ_ALLOC_DEBUG to debug alloc/free operations */ +/* #define SZ_ALLOC_DEBUG */ +/* use SZ_ALLOC_DEBUG to debug alloc/free operations */ -#ifdef _SZ_ALLOC_DEBUG +#ifdef SZ_ALLOC_DEBUG +/* #ifdef _WIN32 -#include <windows.h> +#include "7zWindows.h" #endif +*/ #include <stdio.h> -int g_allocCount = 0; -int g_allocCountTemp = 0; +static int g_allocCount = 0; +static int g_allocCountTemp = 0; +static void Print_Alloc(const char *s, size_t size, int *counter) +{ + const unsigned size2 = (unsigned)size; + fprintf(stderr, "\n%s count = %10d : %10u bytes; ", s, *counter, size2); + (*counter)++; +} +static void Print_Free(const char *s, int *counter) +{ + (*counter)--; + fprintf(stderr, "\n%s count = %10d", s, *counter); +} #endif -void *SzAlloc(void *p, size_t size) +void *SzAlloc(ISzAllocPtr p, size_t size) { - UNUSED_VAR(p); + UNUSED_VAR(p) if (size == 0) return 0; - #ifdef _SZ_ALLOC_DEBUG - fprintf(stderr, "\nAlloc %10u bytes; count = %10d", (unsigned)size, g_allocCount); - g_allocCount++; + #ifdef SZ_ALLOC_DEBUG + Print_Alloc("Alloc", size, &g_allocCount); #endif return malloc(size); } -void SzFree(void *p, void *address) +void SzFree(ISzAllocPtr p, void *address) { - UNUSED_VAR(p); - #ifdef _SZ_ALLOC_DEBUG - if (address != 0) - { - g_allocCount--; - fprintf(stderr, "\nFree; count = %10d", g_allocCount); - } + UNUSED_VAR(p) + #ifdef SZ_ALLOC_DEBUG + if (address) + Print_Free("Free ", &g_allocCount); #endif free(address); } -void *SzAllocTemp(void *p, size_t size) +void *SzAllocTemp(ISzAllocPtr p, size_t size) { - UNUSED_VAR(p); + UNUSED_VAR(p) if (size == 0) return 0; - #ifdef _SZ_ALLOC_DEBUG - fprintf(stderr, "\nAlloc_temp %10u bytes; count = %10d", (unsigned)size, g_allocCountTemp); - g_allocCountTemp++; + #ifdef SZ_ALLOC_DEBUG + Print_Alloc("Alloc_temp", size, &g_allocCountTemp); + /* #ifdef _WIN32 return HeapAlloc(GetProcessHeap(), 0, size); #endif + */ #endif return malloc(size); } -void SzFreeTemp(void *p, void *address) +void SzFreeTemp(ISzAllocPtr p, void *address) { - UNUSED_VAR(p); - #ifdef _SZ_ALLOC_DEBUG - if (address != 0) - { - g_allocCountTemp--; - fprintf(stderr, "\nFree_temp; count = %10d", g_allocCountTemp); - } + UNUSED_VAR(p) + #ifdef SZ_ALLOC_DEBUG + if (address) + Print_Free("Free_temp ", &g_allocCountTemp); + /* #ifdef _WIN32 HeapFree(GetProcessHeap(), 0, address); return; #endif + */ #endif free(address); } diff --git a/3rdparty/lzma/C/7zAlloc.h b/3rdparty/lzma/C/7zAlloc.h index 2fd5bdbc800..b2b8b0cdd93 100644 --- a/3rdparty/lzma/C/7zAlloc.h +++ b/3rdparty/lzma/C/7zAlloc.h @@ -1,23 +1,19 @@ /* 7zAlloc.h -- Allocation functions -2013-03-25 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __7Z_ALLOC_H -#define __7Z_ALLOC_H +#ifndef ZIP7_INC_7Z_ALLOC_H +#define ZIP7_INC_7Z_ALLOC_H -#include <stdlib.h> +#include "7zTypes.h" -#ifdef __cplusplus -extern "C" { -#endif +EXTERN_C_BEGIN -void *SzAlloc(void *p, size_t size); -void SzFree(void *p, void *address); +void *SzAlloc(ISzAllocPtr p, size_t size); +void SzFree(ISzAllocPtr p, void *address); -void *SzAllocTemp(void *p, size_t size); -void SzFreeTemp(void *p, void *address); +void *SzAllocTemp(ISzAllocPtr p, size_t size); +void SzFreeTemp(ISzAllocPtr p, void *address); -#ifdef __cplusplus -} -#endif +EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/7zArcIn.c b/3rdparty/lzma/C/7zArcIn.c index 2beed3d5ccf..43fa7c21454 100644 --- a/3rdparty/lzma/C/7zArcIn.c +++ b/3rdparty/lzma/C/7zArcIn.c @@ -1,5 +1,5 @@ /* 7zArcIn.c -- 7z Input functions -2016-05-16 : Igor Pavlov : Public domain */ +2023-05-11 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -10,16 +10,17 @@ #include "7zCrc.h" #include "CpuArch.h" -#define MY_ALLOC(T, p, size, alloc) { \ - if ((p = (T *)IAlloc_Alloc(alloc, (size) * sizeof(T))) == NULL) return SZ_ERROR_MEM; } +#define MY_ALLOC(T, p, size, alloc) \ + { if ((p = (T *)ISzAlloc_Alloc(alloc, (size) * sizeof(T))) == NULL) return SZ_ERROR_MEM; } -#define MY_ALLOC_ZE(T, p, size, alloc) { if ((size) == 0) p = NULL; else MY_ALLOC(T, p, size, alloc) } +#define MY_ALLOC_ZE(T, p, size, alloc) \ + { if ((size) == 0) p = NULL; else MY_ALLOC(T, p, size, alloc) } #define MY_ALLOC_AND_CPY(to, size, from, alloc) \ { MY_ALLOC(Byte, to, size, alloc); memcpy(to, from, size); } #define MY_ALLOC_ZE_AND_CPY(to, size, from, alloc) \ - { if ((size) == 0) p = NULL; else { MY_ALLOC_AND_CPY(to, size, from, alloc) } } + { if ((size) == 0) to = NULL; else { MY_ALLOC_AND_CPY(to, size, from, alloc) } } #define k7zMajorVersion 0 @@ -58,9 +59,9 @@ enum EIdEnum const Byte k7zSignature[k7zSignatureSize] = {'7', 'z', 0xBC, 0xAF, 0x27, 0x1C}; -#define SzBitUi32s_Init(p) { (p)->Defs = NULL; (p)->Vals = NULL; } +#define SzBitUi32s_INIT(p) { (p)->Defs = NULL; (p)->Vals = NULL; } -static SRes SzBitUi32s_Alloc(CSzBitUi32s *p, size_t num, ISzAlloc *alloc) +static SRes SzBitUi32s_Alloc(CSzBitUi32s *p, size_t num, ISzAllocPtr alloc) { if (num == 0) { @@ -69,24 +70,24 @@ static SRes SzBitUi32s_Alloc(CSzBitUi32s *p, size_t num, ISzAlloc *alloc) } else { - MY_ALLOC(Byte, p->Defs, (num + 7) >> 3, alloc); - MY_ALLOC(UInt32, p->Vals, num, alloc); + MY_ALLOC(Byte, p->Defs, (num + 7) >> 3, alloc) + MY_ALLOC(UInt32, p->Vals, num, alloc) } return SZ_OK; } -void SzBitUi32s_Free(CSzBitUi32s *p, ISzAlloc *alloc) +static void SzBitUi32s_Free(CSzBitUi32s *p, ISzAllocPtr alloc) { - IAlloc_Free(alloc, p->Defs); p->Defs = NULL; - IAlloc_Free(alloc, p->Vals); p->Vals = NULL; + ISzAlloc_Free(alloc, p->Defs); p->Defs = NULL; + ISzAlloc_Free(alloc, p->Vals); p->Vals = NULL; } -#define SzBitUi64s_Init(p) { (p)->Defs = NULL; (p)->Vals = NULL; } +#define SzBitUi64s_INIT(p) { (p)->Defs = NULL; (p)->Vals = NULL; } -void SzBitUi64s_Free(CSzBitUi64s *p, ISzAlloc *alloc) +static void SzBitUi64s_Free(CSzBitUi64s *p, ISzAllocPtr alloc) { - IAlloc_Free(alloc, p->Defs); p->Defs = NULL; - IAlloc_Free(alloc, p->Vals); p->Vals = NULL; + ISzAlloc_Free(alloc, p->Defs); p->Defs = NULL; + ISzAlloc_Free(alloc, p->Vals); p->Vals = NULL; } @@ -96,7 +97,7 @@ static void SzAr_Init(CSzAr *p) p->NumFolders = 0; p->PackPositions = NULL; - SzBitUi32s_Init(&p->FolderCRCs); + SzBitUi32s_INIT(&p->FolderCRCs) p->FoCodersOffsets = NULL; p->FoStartPackStreamIndex = NULL; @@ -105,20 +106,22 @@ static void SzAr_Init(CSzAr *p) p->CoderUnpackSizes = NULL; p->CodersData = NULL; + + p->RangeLimit = 0; } -static void SzAr_Free(CSzAr *p, ISzAlloc *alloc) +static void SzAr_Free(CSzAr *p, ISzAllocPtr alloc) { - IAlloc_Free(alloc, p->PackPositions); + ISzAlloc_Free(alloc, p->PackPositions); SzBitUi32s_Free(&p->FolderCRCs, alloc); - IAlloc_Free(alloc, p->FoCodersOffsets); - IAlloc_Free(alloc, p->FoStartPackStreamIndex); - IAlloc_Free(alloc, p->FoToCoderUnpackSizes); - IAlloc_Free(alloc, p->FoToMainUnpackSizeIndex); - IAlloc_Free(alloc, p->CoderUnpackSizes); + ISzAlloc_Free(alloc, p->FoCodersOffsets); + ISzAlloc_Free(alloc, p->FoStartPackStreamIndex); + ISzAlloc_Free(alloc, p->FoToCoderUnpackSizes); + ISzAlloc_Free(alloc, p->FoToMainUnpackSizeIndex); + ISzAlloc_Free(alloc, p->CoderUnpackSizes); - IAlloc_Free(alloc, p->CodersData); + ISzAlloc_Free(alloc, p->CodersData); SzAr_Init(p); } @@ -140,23 +143,23 @@ void SzArEx_Init(CSzArEx *p) p->FileNameOffsets = NULL; p->FileNames = NULL; - SzBitUi32s_Init(&p->CRCs); - SzBitUi32s_Init(&p->Attribs); - // SzBitUi32s_Init(&p->Parents); - SzBitUi64s_Init(&p->MTime); - SzBitUi64s_Init(&p->CTime); + SzBitUi32s_INIT(&p->CRCs) + SzBitUi32s_INIT(&p->Attribs) + // SzBitUi32s_INIT(&p->Parents) + SzBitUi64s_INIT(&p->MTime) + SzBitUi64s_INIT(&p->CTime) } -void SzArEx_Free(CSzArEx *p, ISzAlloc *alloc) +void SzArEx_Free(CSzArEx *p, ISzAllocPtr alloc) { - IAlloc_Free(alloc, p->UnpackPositions); - IAlloc_Free(alloc, p->IsDirs); + ISzAlloc_Free(alloc, p->UnpackPositions); + ISzAlloc_Free(alloc, p->IsDirs); - IAlloc_Free(alloc, p->FolderToFile); - IAlloc_Free(alloc, p->FileToFolder); + ISzAlloc_Free(alloc, p->FolderToFile); + ISzAlloc_Free(alloc, p->FileToFolder); - IAlloc_Free(alloc, p->FileNameOffsets); - IAlloc_Free(alloc, p->FileNames); + ISzAlloc_Free(alloc, p->FileNameOffsets); + ISzAlloc_Free(alloc, p->FileNames); SzBitUi32s_Free(&p->CRCs, alloc); SzBitUi32s_Free(&p->Attribs, alloc); @@ -178,11 +181,20 @@ static int TestSignatureCandidate(const Byte *testBytes) return 1; } -#define SzData_Clear(p) { (p)->Data = NULL; (p)->Size = 0; } +#define SzData_CLEAR(p) { (p)->Data = NULL; (p)->Size = 0; } + +#define SZ_READ_BYTE_SD_NOCHECK(_sd_, dest) \ + (_sd_)->Size--; dest = *(_sd_)->Data++; + +#define SZ_READ_BYTE_SD(_sd_, dest) \ + if ((_sd_)->Size == 0) return SZ_ERROR_ARCHIVE; \ + SZ_READ_BYTE_SD_NOCHECK(_sd_, dest) -#define SZ_READ_BYTE_SD(_sd_, dest) if ((_sd_)->Size == 0) return SZ_ERROR_ARCHIVE; (_sd_)->Size--; dest = *(_sd_)->Data++; #define SZ_READ_BYTE(dest) SZ_READ_BYTE_SD(sd, dest) -#define SZ_READ_BYTE_2(dest) if (sd.Size == 0) return SZ_ERROR_ARCHIVE; sd.Size--; dest = *sd.Data++; + +#define SZ_READ_BYTE_2(dest) \ + if (sd.Size == 0) return SZ_ERROR_ARCHIVE; \ + sd.Size--; dest = *sd.Data++; #define SKIP_DATA(sd, size) { sd->Size -= (size_t)(size); sd->Data += (size_t)(size); } #define SKIP_DATA2(sd, size) { sd.Size -= (size_t)(size); sd.Data += (size_t)(size); } @@ -190,25 +202,25 @@ static int TestSignatureCandidate(const Byte *testBytes) #define SZ_READ_32(dest) if (sd.Size < 4) return SZ_ERROR_ARCHIVE; \ dest = GetUi32(sd.Data); SKIP_DATA2(sd, 4); -static MY_NO_INLINE SRes ReadNumber(CSzData *sd, UInt64 *value) +static Z7_NO_INLINE SRes ReadNumber(CSzData *sd, UInt64 *value) { Byte firstByte, mask; unsigned i; UInt32 v; - SZ_READ_BYTE(firstByte); + SZ_READ_BYTE(firstByte) if ((firstByte & 0x80) == 0) { *value = firstByte; return SZ_OK; } - SZ_READ_BYTE(v); + SZ_READ_BYTE(v) if ((firstByte & 0x40) == 0) { *value = (((UInt32)firstByte & 0x3F) << 8) | v; return SZ_OK; } - SZ_READ_BYTE(mask); + SZ_READ_BYTE(mask) *value = v | ((UInt32)mask << 8); mask = 0x20; for (i = 2; i < 8; i++) @@ -216,11 +228,11 @@ static MY_NO_INLINE SRes ReadNumber(CSzData *sd, UInt64 *value) Byte b; if ((firstByte & mask) == 0) { - UInt64 highPart = (unsigned)firstByte & (unsigned)(mask - 1); + const UInt64 highPart = (unsigned)firstByte & (unsigned)(mask - 1); *value |= (highPart << (8 * i)); return SZ_OK; } - SZ_READ_BYTE(b); + SZ_READ_BYTE(b) *value |= ((UInt64)b << (8 * i)); mask >>= 1; } @@ -228,7 +240,7 @@ static MY_NO_INLINE SRes ReadNumber(CSzData *sd, UInt64 *value) } -static MY_NO_INLINE SRes SzReadNumber32(CSzData *sd, UInt32 *value) +static Z7_NO_INLINE SRes SzReadNumber32(CSzData *sd, UInt32 *value) { Byte firstByte; UInt64 value64; @@ -242,7 +254,7 @@ static MY_NO_INLINE SRes SzReadNumber32(CSzData *sd, UInt32 *value) sd->Size--; return SZ_OK; } - RINOK(ReadNumber(sd, &value64)); + RINOK(ReadNumber(sd, &value64)) if (value64 >= (UInt32)0x80000000 - 1) return SZ_ERROR_UNSUPPORTED; if (value64 >= ((UInt64)(1) << ((sizeof(size_t) - 1) * 8 + 4))) @@ -256,10 +268,10 @@ static MY_NO_INLINE SRes SzReadNumber32(CSzData *sd, UInt32 *value) static SRes SkipData(CSzData *sd) { UInt64 size; - RINOK(ReadNumber(sd, &size)); + RINOK(ReadNumber(sd, &size)) if (size > sd->Size) return SZ_ERROR_ARCHIVE; - SKIP_DATA(sd, size); + SKIP_DATA(sd, size) return SZ_OK; } @@ -268,22 +280,22 @@ static SRes WaitId(CSzData *sd, UInt32 id) for (;;) { UInt64 type; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == id) return SZ_OK; if (type == k7zIdEnd) return SZ_ERROR_ARCHIVE; - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } } static SRes RememberBitVector(CSzData *sd, UInt32 numItems, const Byte **v) { - UInt32 numBytes = (numItems + 7) >> 3; + const UInt32 numBytes = (numItems + 7) >> 3; if (numBytes > sd->Size) return SZ_ERROR_ARCHIVE; *v = sd->Data; - SKIP_DATA(sd, numBytes); + SKIP_DATA(sd, numBytes) return SZ_OK; } @@ -305,48 +317,48 @@ static UInt32 CountDefinedBits(const Byte *bits, UInt32 numItems) return sum; } -static MY_NO_INLINE SRes ReadBitVector(CSzData *sd, UInt32 numItems, Byte **v, ISzAlloc *alloc) +static Z7_NO_INLINE SRes ReadBitVector(CSzData *sd, UInt32 numItems, Byte **v, ISzAllocPtr alloc) { Byte allAreDefined; Byte *v2; - UInt32 numBytes = (numItems + 7) >> 3; + const UInt32 numBytes = (numItems + 7) >> 3; *v = NULL; - SZ_READ_BYTE(allAreDefined); + SZ_READ_BYTE(allAreDefined) if (numBytes == 0) return SZ_OK; if (allAreDefined == 0) { if (numBytes > sd->Size) return SZ_ERROR_ARCHIVE; - MY_ALLOC_AND_CPY(*v, numBytes, sd->Data, alloc); - SKIP_DATA(sd, numBytes); + MY_ALLOC_AND_CPY(*v, numBytes, sd->Data, alloc) + SKIP_DATA(sd, numBytes) return SZ_OK; } - MY_ALLOC(Byte, *v, numBytes, alloc); + MY_ALLOC(Byte, *v, numBytes, alloc) v2 = *v; memset(v2, 0xFF, (size_t)numBytes); { - unsigned numBits = (unsigned)numItems & 7; + const unsigned numBits = (unsigned)numItems & 7; if (numBits != 0) - v2[numBytes - 1] = (Byte)((((UInt32)1 << numBits) - 1) << (8 - numBits)); + v2[(size_t)numBytes - 1] = (Byte)((((UInt32)1 << numBits) - 1) << (8 - numBits)); } return SZ_OK; } -static MY_NO_INLINE SRes ReadUi32s(CSzData *sd2, UInt32 numItems, CSzBitUi32s *crcs, ISzAlloc *alloc) +static Z7_NO_INLINE SRes ReadUi32s(CSzData *sd2, UInt32 numItems, CSzBitUi32s *crcs, ISzAllocPtr alloc) { UInt32 i; CSzData sd; UInt32 *vals; const Byte *defs; - MY_ALLOC_ZE(UInt32, crcs->Vals, numItems, alloc); + MY_ALLOC_ZE(UInt32, crcs->Vals, numItems, alloc) sd = *sd2; defs = crcs->Defs; vals = crcs->Vals; for (i = 0; i < numItems; i++) if (SzBitArray_Check(defs, i)) { - SZ_READ_32(vals[i]); + SZ_READ_32(vals[i]) } else vals[i] = 0; @@ -354,10 +366,10 @@ static MY_NO_INLINE SRes ReadUi32s(CSzData *sd2, UInt32 numItems, CSzBitUi32s *c return SZ_OK; } -static SRes ReadBitUi32s(CSzData *sd, UInt32 numItems, CSzBitUi32s *crcs, ISzAlloc *alloc) +static SRes ReadBitUi32s(CSzData *sd, UInt32 numItems, CSzBitUi32s *crcs, ISzAllocPtr alloc) { SzBitUi32s_Free(crcs, alloc); - RINOK(ReadBitVector(sd, numItems, &crcs->Defs, alloc)); + RINOK(ReadBitVector(sd, numItems, &crcs->Defs, alloc)) return ReadUi32s(sd, numItems, crcs, alloc); } @@ -365,36 +377,36 @@ static SRes SkipBitUi32s(CSzData *sd, UInt32 numItems) { Byte allAreDefined; UInt32 numDefined = numItems; - SZ_READ_BYTE(allAreDefined); + SZ_READ_BYTE(allAreDefined) if (!allAreDefined) { - size_t numBytes = (numItems + 7) >> 3; + const size_t numBytes = (numItems + 7) >> 3; if (numBytes > sd->Size) return SZ_ERROR_ARCHIVE; numDefined = CountDefinedBits(sd->Data, numItems); - SKIP_DATA(sd, numBytes); + SKIP_DATA(sd, numBytes) } if (numDefined > (sd->Size >> 2)) return SZ_ERROR_ARCHIVE; - SKIP_DATA(sd, (size_t)numDefined * 4); + SKIP_DATA(sd, (size_t)numDefined * 4) return SZ_OK; } -static SRes ReadPackInfo(CSzAr *p, CSzData *sd, ISzAlloc *alloc) +static SRes ReadPackInfo(CSzAr *p, CSzData *sd, ISzAllocPtr alloc) { - RINOK(SzReadNumber32(sd, &p->NumPackStreams)); + RINOK(SzReadNumber32(sd, &p->NumPackStreams)) - RINOK(WaitId(sd, k7zIdSize)); - MY_ALLOC(UInt64, p->PackPositions, (size_t)p->NumPackStreams + 1, alloc); + RINOK(WaitId(sd, k7zIdSize)) + MY_ALLOC(UInt64, p->PackPositions, (size_t)p->NumPackStreams + 1, alloc) { UInt64 sum = 0; UInt32 i; - UInt32 numPackStreams = p->NumPackStreams; + const UInt32 numPackStreams = p->NumPackStreams; for (i = 0; i < numPackStreams; i++) { UInt64 packSize; p->PackPositions[i] = sum; - RINOK(ReadNumber(sd, &packSize)); + RINOK(ReadNumber(sd, &packSize)) sum += packSize; if (sum < packSize) return SZ_ERROR_ARCHIVE; @@ -405,16 +417,16 @@ static SRes ReadPackInfo(CSzAr *p, CSzData *sd, ISzAlloc *alloc) for (;;) { UInt64 type; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdEnd) return SZ_OK; if (type == k7zIdCRC) { /* CRC of packed streams is unused now */ - RINOK(SkipBitUi32s(sd, p->NumPackStreams)); + RINOK(SkipBitUi32s(sd, p->NumPackStreams)) continue; } - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } } @@ -440,7 +452,7 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) f->NumPackStreams = 0; f->UnpackStream = 0; - RINOK(SzReadNumber32(sd, &numCoders)); + RINOK(SzReadNumber32(sd, &numCoders)) if (numCoders == 0 || numCoders > SZ_NUM_CODERS_IN_FOLDER_MAX) return SZ_ERROR_UNSUPPORTED; @@ -451,7 +463,7 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) unsigned idSize, j; UInt64 id; - SZ_READ_BYTE(mainByte); + SZ_READ_BYTE(mainByte) if ((mainByte & 0xC0) != 0) return SZ_ERROR_UNSUPPORTED; @@ -479,12 +491,12 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) { UInt32 numStreams; - RINOK(SzReadNumber32(sd, &numStreams)); + RINOK(SzReadNumber32(sd, &numStreams)) if (numStreams > k_NumCodersStreams_in_Folder_MAX) return SZ_ERROR_UNSUPPORTED; coder->NumStreams = (Byte)numStreams; - RINOK(SzReadNumber32(sd, &numStreams)); + RINOK(SzReadNumber32(sd, &numStreams)) if (numStreams != 1) return SZ_ERROR_UNSUPPORTED; } @@ -497,12 +509,12 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) if ((mainByte & 0x20) != 0) { UInt32 propsSize = 0; - RINOK(SzReadNumber32(sd, &propsSize)); + RINOK(SzReadNumber32(sd, &propsSize)) if (propsSize > sd->Size) return SZ_ERROR_ARCHIVE; if (propsSize >= 0x80) return SZ_ERROR_UNSUPPORTED; - coder->PropsOffset = sd->Data - dataStart; + coder->PropsOffset = (size_t)(sd->Data - dataStart); coder->PropsSize = (Byte)propsSize; sd->Data += (size_t)propsSize; sd->Size -= (size_t)propsSize; @@ -547,12 +559,12 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) { CSzBond *bp = f->Bonds + i; - RINOK(SzReadNumber32(sd, &bp->InIndex)); + RINOK(SzReadNumber32(sd, &bp->InIndex)) if (bp->InIndex >= numInStreams || streamUsed[bp->InIndex]) return SZ_ERROR_ARCHIVE; streamUsed[bp->InIndex] = True; - RINOK(SzReadNumber32(sd, &bp->OutIndex)); + RINOK(SzReadNumber32(sd, &bp->OutIndex)) if (bp->OutIndex >= numCoders || coderUsed[bp->OutIndex]) return SZ_ERROR_ARCHIVE; coderUsed[bp->OutIndex] = True; @@ -582,7 +594,7 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) for (i = 0; i < numPackStreams; i++) { UInt32 index; - RINOK(SzReadNumber32(sd, &index)); + RINOK(SzReadNumber32(sd, &index)) if (index >= numInStreams || streamUsed[index]) return SZ_ERROR_ARCHIVE; streamUsed[index] = True; @@ -596,7 +608,7 @@ SRes SzGetNextFolderItem(CSzFolder *f, CSzData *sd) } -static MY_NO_INLINE SRes SkipNumbers(CSzData *sd2, UInt32 num) +static Z7_NO_INLINE SRes SkipNumbers(CSzData *sd2, UInt32 num) { CSzData sd; sd = *sd2; @@ -604,7 +616,7 @@ static MY_NO_INLINE SRes SkipNumbers(CSzData *sd2, UInt32 num) { Byte firstByte, mask; unsigned i; - SZ_READ_BYTE_2(firstByte); + SZ_READ_BYTE_2(firstByte) if ((firstByte & 0x80) == 0) continue; if ((firstByte & 0x40) == 0) @@ -620,7 +632,7 @@ static MY_NO_INLINE SRes SkipNumbers(CSzData *sd2, UInt32 num) mask >>= 1; if (i > sd.Size) return SZ_ERROR_ARCHIVE; - SKIP_DATA2(sd, i); + SKIP_DATA2(sd, i) } *sd2 = sd; return SZ_OK; @@ -635,7 +647,7 @@ static SRes ReadUnpackInfo(CSzAr *p, CSzData *sd2, UInt32 numFoldersMax, const CBuf *tempBufs, UInt32 numTempBufs, - ISzAlloc *alloc) + ISzAllocPtr alloc) { CSzData sd; @@ -643,30 +655,30 @@ static SRes ReadUnpackInfo(CSzAr *p, const Byte *startBufPtr; Byte external; - RINOK(WaitId(sd2, k7zIdFolder)); + RINOK(WaitId(sd2, k7zIdFolder)) - RINOK(SzReadNumber32(sd2, &numFolders)); + RINOK(SzReadNumber32(sd2, &numFolders)) if (numFolders > numFoldersMax) return SZ_ERROR_UNSUPPORTED; p->NumFolders = numFolders; - SZ_READ_BYTE_SD(sd2, external); + SZ_READ_BYTE_SD(sd2, external) if (external == 0) sd = *sd2; else { UInt32 index; - RINOK(SzReadNumber32(sd2, &index)); + RINOK(SzReadNumber32(sd2, &index)) if (index >= numTempBufs) return SZ_ERROR_ARCHIVE; sd.Data = tempBufs[index].data; sd.Size = tempBufs[index].size; } - MY_ALLOC(size_t, p->FoCodersOffsets, (size_t)numFolders + 1, alloc); - MY_ALLOC(UInt32, p->FoStartPackStreamIndex, (size_t)numFolders + 1, alloc); - MY_ALLOC(UInt32, p->FoToCoderUnpackSizes, (size_t)numFolders + 1, alloc); - MY_ALLOC(Byte, p->FoToMainUnpackSizeIndex, (size_t)numFolders, alloc); + MY_ALLOC(size_t, p->FoCodersOffsets, (size_t)numFolders + 1, alloc) + MY_ALLOC(UInt32, p->FoStartPackStreamIndex, (size_t)numFolders + 1, alloc) + MY_ALLOC(UInt32, p->FoToCoderUnpackSizes, (size_t)numFolders + 1, alloc) + MY_ALLOC_ZE(Byte, p->FoToMainUnpackSizeIndex, (size_t)numFolders, alloc) startBufPtr = sd.Data; @@ -677,9 +689,9 @@ static SRes ReadUnpackInfo(CSzAr *p, { UInt32 numCoders, ci, numInStreams = 0; - p->FoCodersOffsets[fo] = sd.Data - startBufPtr; + p->FoCodersOffsets[fo] = (size_t)(sd.Data - startBufPtr); - RINOK(SzReadNumber32(&sd, &numCoders)); + RINOK(SzReadNumber32(&sd, &numCoders)) if (numCoders == 0 || numCoders > k_Scan_NumCoders_MAX) return SZ_ERROR_UNSUPPORTED; @@ -689,7 +701,7 @@ static SRes ReadUnpackInfo(CSzAr *p, unsigned idSize; UInt32 coderInStreams; - SZ_READ_BYTE_2(mainByte); + SZ_READ_BYTE_2(mainByte) if ((mainByte & 0xC0) != 0) return SZ_ERROR_UNSUPPORTED; idSize = (mainByte & 0xF); @@ -697,15 +709,15 @@ static SRes ReadUnpackInfo(CSzAr *p, return SZ_ERROR_UNSUPPORTED; if (idSize > sd.Size) return SZ_ERROR_ARCHIVE; - SKIP_DATA2(sd, idSize); + SKIP_DATA2(sd, idSize) coderInStreams = 1; if ((mainByte & 0x10) != 0) { UInt32 coderOutStreams; - RINOK(SzReadNumber32(&sd, &coderInStreams)); - RINOK(SzReadNumber32(&sd, &coderOutStreams)); + RINOK(SzReadNumber32(&sd, &coderInStreams)) + RINOK(SzReadNumber32(&sd, &coderOutStreams)) if (coderInStreams > k_Scan_NumCodersStreams_in_Folder_MAX || coderOutStreams != 1) return SZ_ERROR_UNSUPPORTED; } @@ -715,10 +727,10 @@ static SRes ReadUnpackInfo(CSzAr *p, if ((mainByte & 0x20) != 0) { UInt32 propsSize; - RINOK(SzReadNumber32(&sd, &propsSize)); + RINOK(SzReadNumber32(&sd, &propsSize)) if (propsSize > sd.Size) return SZ_ERROR_ARCHIVE; - SKIP_DATA2(sd, propsSize); + SKIP_DATA2(sd, propsSize) } } @@ -732,7 +744,7 @@ static SRes ReadUnpackInfo(CSzAr *p, Byte coderUsed[k_Scan_NumCoders_MAX]; UInt32 i; - UInt32 numBonds = numCoders - 1; + const UInt32 numBonds = numCoders - 1; if (numInStreams < numBonds) return SZ_ERROR_ARCHIVE; @@ -748,12 +760,12 @@ static SRes ReadUnpackInfo(CSzAr *p, { UInt32 index; - RINOK(SzReadNumber32(&sd, &index)); + RINOK(SzReadNumber32(&sd, &index)) if (index >= numInStreams || streamUsed[index]) return SZ_ERROR_ARCHIVE; streamUsed[index] = True; - RINOK(SzReadNumber32(&sd, &index)); + RINOK(SzReadNumber32(&sd, &index)) if (index >= numCoders || coderUsed[index]) return SZ_ERROR_ARCHIVE; coderUsed[index] = True; @@ -765,7 +777,7 @@ static SRes ReadUnpackInfo(CSzAr *p, for (i = 0; i < numPackStreams; i++) { UInt32 index; - RINOK(SzReadNumber32(&sd, &index)); + RINOK(SzReadNumber32(&sd, &index)) if (index >= numInStreams || streamUsed[index]) return SZ_ERROR_ARCHIVE; streamUsed[index] = True; @@ -797,10 +809,10 @@ static SRes ReadUnpackInfo(CSzAr *p, p->FoToCoderUnpackSizes[fo] = numCodersOutStreams; { - size_t dataSize = sd.Data - startBufPtr; + const size_t dataSize = (size_t)(sd.Data - startBufPtr); p->FoStartPackStreamIndex[fo] = packStreamIndex; p->FoCodersOffsets[fo] = dataSize; - MY_ALLOC_ZE_AND_CPY(p->CodersData, dataSize, startBufPtr, alloc); + MY_ALLOC_ZE_AND_CPY(p->CodersData, dataSize, startBufPtr, alloc) } if (external != 0) @@ -810,21 +822,21 @@ static SRes ReadUnpackInfo(CSzAr *p, sd = *sd2; } - RINOK(WaitId(&sd, k7zIdCodersUnpackSize)); + RINOK(WaitId(&sd, k7zIdCodersUnpackSize)) - MY_ALLOC_ZE(UInt64, p->CoderUnpackSizes, (size_t)numCodersOutStreams, alloc); + MY_ALLOC_ZE(UInt64, p->CoderUnpackSizes, (size_t)numCodersOutStreams, alloc) { UInt32 i; for (i = 0; i < numCodersOutStreams; i++) { - RINOK(ReadNumber(&sd, p->CoderUnpackSizes + i)); + RINOK(ReadNumber(&sd, p->CoderUnpackSizes + i)) } } for (;;) { UInt64 type; - RINOK(ReadID(&sd, &type)); + RINOK(ReadID(&sd, &type)) if (type == k7zIdEnd) { *sd2 = sd; @@ -832,10 +844,10 @@ static SRes ReadUnpackInfo(CSzAr *p, } if (type == k7zIdCRC) { - RINOK(ReadBitUi32s(&sd, numFolders, &p->FolderCRCs, alloc)); + RINOK(ReadBitUi32s(&sd, numFolders, &p->FolderCRCs, alloc)) continue; } - RINOK(SkipData(&sd)); + RINOK(SkipData(&sd)) } } @@ -860,13 +872,13 @@ static SRes ReadSubStreamsInfo(CSzAr *p, CSzData *sd, CSubStreamInfo *ssi) { UInt64 type = 0; UInt32 numSubDigests = 0; - UInt32 numFolders = p->NumFolders; + const UInt32 numFolders = p->NumFolders; UInt32 numUnpackStreams = numFolders; UInt32 numUnpackSizesInData = 0; for (;;) { - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdNumUnpackStream) { UInt32 i; @@ -876,7 +888,7 @@ static SRes ReadSubStreamsInfo(CSzAr *p, CSzData *sd, CSubStreamInfo *ssi) for (i = 0; i < numFolders; i++) { UInt32 numStreams; - RINOK(SzReadNumber32(sd, &numStreams)); + RINOK(SzReadNumber32(sd, &numStreams)) if (numUnpackStreams > numUnpackStreams + numStreams) return SZ_ERROR_UNSUPPORTED; numUnpackStreams += numStreams; @@ -885,12 +897,12 @@ static SRes ReadSubStreamsInfo(CSzAr *p, CSzData *sd, CSubStreamInfo *ssi) if (numStreams != 1 || !SzBitWithVals_Check(&p->FolderCRCs, i)) numSubDigests += numStreams; } - ssi->sdNumSubStreams.Size = sd->Data - ssi->sdNumSubStreams.Data; + ssi->sdNumSubStreams.Size = (size_t)(sd->Data - ssi->sdNumSubStreams.Data); continue; } if (type == k7zIdCRC || type == k7zIdSize || type == k7zIdEnd) break; - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } if (!ssi->sdNumSubStreams.Data) @@ -906,9 +918,9 @@ static SRes ReadSubStreamsInfo(CSzAr *p, CSzData *sd, CSubStreamInfo *ssi) if (type == k7zIdSize) { ssi->sdSizes.Data = sd->Data; - RINOK(SkipNumbers(sd, numUnpackSizesInData)); - ssi->sdSizes.Size = sd->Data - ssi->sdSizes.Data; - RINOK(ReadID(sd, &type)); + RINOK(SkipNumbers(sd, numUnpackSizesInData)) + ssi->sdSizes.Size = (size_t)(sd->Data - ssi->sdSizes.Data); + RINOK(ReadID(sd, &type)) } for (;;) @@ -918,14 +930,14 @@ static SRes ReadSubStreamsInfo(CSzAr *p, CSzData *sd, CSubStreamInfo *ssi) if (type == k7zIdCRC) { ssi->sdCRCs.Data = sd->Data; - RINOK(SkipBitUi32s(sd, numSubDigests)); - ssi->sdCRCs.Size = sd->Data - ssi->sdCRCs.Data; + RINOK(SkipBitUi32s(sd, numSubDigests)) + ssi->sdCRCs.Size = (size_t)(sd->Data - ssi->sdCRCs.Data); } else { - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) } } @@ -934,31 +946,35 @@ static SRes SzReadStreamsInfo(CSzAr *p, UInt32 numFoldersMax, const CBuf *tempBufs, UInt32 numTempBufs, UInt64 *dataOffset, CSubStreamInfo *ssi, - ISzAlloc *alloc) + ISzAllocPtr alloc) { UInt64 type; - SzData_Clear(&ssi->sdSizes); - SzData_Clear(&ssi->sdCRCs); - SzData_Clear(&ssi->sdNumSubStreams); + SzData_CLEAR(&ssi->sdSizes) + SzData_CLEAR(&ssi->sdCRCs) + SzData_CLEAR(&ssi->sdNumSubStreams) *dataOffset = 0; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdPackInfo) { - RINOK(ReadNumber(sd, dataOffset)); - RINOK(ReadPackInfo(p, sd, alloc)); - RINOK(ReadID(sd, &type)); + RINOK(ReadNumber(sd, dataOffset)) + if (*dataOffset > p->RangeLimit) + return SZ_ERROR_ARCHIVE; + RINOK(ReadPackInfo(p, sd, alloc)) + if (p->PackPositions[p->NumPackStreams] > p->RangeLimit - *dataOffset) + return SZ_ERROR_ARCHIVE; + RINOK(ReadID(sd, &type)) } if (type == k7zIdUnpackInfo) { - RINOK(ReadUnpackInfo(p, sd, numFoldersMax, tempBufs, numTempBufs, alloc)); - RINOK(ReadID(sd, &type)); + RINOK(ReadUnpackInfo(p, sd, numFoldersMax, tempBufs, numTempBufs, alloc)) + RINOK(ReadID(sd, &type)) } if (type == k7zIdSubStreamsInfo) { - RINOK(ReadSubStreamsInfo(p, sd, ssi)); - RINOK(ReadID(sd, &type)); + RINOK(ReadSubStreamsInfo(p, sd, ssi)) + RINOK(ReadID(sd, &type)) } else { @@ -970,19 +986,19 @@ static SRes SzReadStreamsInfo(CSzAr *p, } static SRes SzReadAndDecodePackedStreams( - ILookInStream *inStream, + ILookInStreamPtr inStream, CSzData *sd, CBuf *tempBufs, UInt32 numFoldersMax, UInt64 baseOffset, CSzAr *p, - ISzAlloc *allocTemp) + ISzAllocPtr allocTemp) { UInt64 dataStartPos; UInt32 fo; CSubStreamInfo ssi; - RINOK(SzReadStreamsInfo(p, sd, numFoldersMax, NULL, 0, &dataStartPos, &ssi, allocTemp)); + RINOK(SzReadStreamsInfo(p, sd, numFoldersMax, NULL, 0, &dataStartPos, &ssi, allocTemp)) dataStartPos += baseOffset; if (p->NumFolders == 0) @@ -994,7 +1010,7 @@ static SRes SzReadAndDecodePackedStreams( for (fo = 0; fo < p->NumFolders; fo++) { CBuf *tempBuf = tempBufs + fo; - UInt64 unpackSize = SzAr_GetFolderUnpackSize(p, fo); + const UInt64 unpackSize = SzAr_GetFolderUnpackSize(p, fo); if ((size_t)unpackSize != unpackSize) return SZ_ERROR_MEM; if (!Buf_Create(tempBuf, (size_t)unpackSize, allocTemp)) @@ -1004,8 +1020,8 @@ static SRes SzReadAndDecodePackedStreams( for (fo = 0; fo < p->NumFolders; fo++) { const CBuf *tempBuf = tempBufs + fo; - RINOK(LookInStream_SeekTo(inStream, dataStartPos)); - RINOK(SzAr_DecodeFolder(p, fo, inStream, dataStartPos, tempBuf->data, tempBuf->size, allocTemp)); + RINOK(LookInStream_SeekTo(inStream, dataStartPos)) + RINOK(SzAr_DecodeFolder(p, fo, inStream, dataStartPos, tempBuf->data, tempBuf->size, allocTemp)) } return SZ_OK; @@ -1028,22 +1044,22 @@ static SRes SzReadFileNames(const Byte *data, size_t size, UInt32 numFiles, size return SZ_ERROR_ARCHIVE; for (p = data + pos; #ifdef _WIN32 - *(const UInt16 *)p != 0 + *(const UInt16 *)(const void *)p != 0 #else p[0] != 0 || p[1] != 0 #endif ; p += 2); - pos = p - data + 2; + pos = (size_t)(p - data) + 2; *offsets++ = (pos >> 1); } while (--numFiles); return (pos == size) ? SZ_OK : SZ_ERROR_ARCHIVE; } -static MY_NO_INLINE SRes ReadTime(CSzBitUi64s *p, UInt32 num, +static Z7_NO_INLINE SRes ReadTime(CSzBitUi64s *p, UInt32 num, CSzData *sd2, const CBuf *tempBufs, UInt32 numTempBufs, - ISzAlloc *alloc) + ISzAllocPtr alloc) { CSzData sd; UInt32 i; @@ -1051,22 +1067,22 @@ static MY_NO_INLINE SRes ReadTime(CSzBitUi64s *p, UInt32 num, Byte *defs; Byte external; - RINOK(ReadBitVector(sd2, num, &p->Defs, alloc)); + RINOK(ReadBitVector(sd2, num, &p->Defs, alloc)) - SZ_READ_BYTE_SD(sd2, external); + SZ_READ_BYTE_SD(sd2, external) if (external == 0) sd = *sd2; else { UInt32 index; - RINOK(SzReadNumber32(sd2, &index)); + RINOK(SzReadNumber32(sd2, &index)) if (index >= numTempBufs) return SZ_ERROR_ARCHIVE; sd.Data = tempBufs[index].data; sd.Size = tempBufs[index].size; } - MY_ALLOC_ZE(CNtfsFileTime, p->Vals, num, alloc); + MY_ALLOC_ZE(CNtfsFileTime, p->Vals, num, alloc) vals = p->Vals; defs = p->Defs; for (i = 0; i < num; i++) @@ -1076,7 +1092,7 @@ static MY_NO_INLINE SRes ReadTime(CSzBitUi64s *p, UInt32 num, return SZ_ERROR_ARCHIVE; vals[i].Low = GetUi32(sd.Data); vals[i].High = GetUi32(sd.Data + 4); - SKIP_DATA2(sd, 8); + SKIP_DATA2(sd, 8) } else vals[i].High = vals[i].Low = 0; @@ -1094,10 +1110,10 @@ static MY_NO_INLINE SRes ReadTime(CSzBitUi64s *p, UInt32 num, static SRes SzReadHeader2( CSzArEx *p, /* allocMain */ CSzData *sd, - ILookInStream *inStream, + ILookInStreamPtr inStream, CBuf *tempBufs, UInt32 *numTempBufs, - ISzAlloc *allocMain, - ISzAlloc *allocTemp + ISzAllocPtr allocMain, + ISzAllocPtr allocTemp ) { CSubStreamInfo ssi; @@ -1105,26 +1121,26 @@ static SRes SzReadHeader2( { UInt64 type; - SzData_Clear(&ssi.sdSizes); - SzData_Clear(&ssi.sdCRCs); - SzData_Clear(&ssi.sdNumSubStreams); + SzData_CLEAR(&ssi.sdSizes) + SzData_CLEAR(&ssi.sdCRCs) + SzData_CLEAR(&ssi.sdNumSubStreams) ssi.NumSubDigests = 0; ssi.NumTotalSubStreams = 0; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdArchiveProperties) { for (;;) { UInt64 type2; - RINOK(ReadID(sd, &type2)); + RINOK(ReadID(sd, &type2)) if (type2 == k7zIdEnd) break; - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) } if (type == k7zIdAdditionalStreamsInfo) @@ -1133,6 +1149,8 @@ static SRes SzReadHeader2( SRes res; SzAr_Init(&tempAr); + tempAr.RangeLimit = p->db.RangeLimit; + res = SzReadAndDecodePackedStreams(inStream, sd, tempBufs, NUM_ADDITIONAL_STREAMS_MAX, p->startPosAfterHeader, &tempAr, allocTemp); *numTempBufs = tempAr.NumFolders; @@ -1140,15 +1158,15 @@ static SRes SzReadHeader2( if (res != SZ_OK) return res; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) } if (type == k7zIdMainStreamsInfo) { RINOK(SzReadStreamsInfo(&p->db, sd, (UInt32)1 << 30, tempBufs, *numTempBufs, - &p->dataPos, &ssi, allocMain)); + &p->dataPos, &ssi, allocMain)) p->dataPos += p->startPosAfterHeader; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) } if (type == k7zIdEnd) @@ -1166,23 +1184,23 @@ static SRes SzReadHeader2( const Byte *emptyStreams = NULL; const Byte *emptyFiles = NULL; - RINOK(SzReadNumber32(sd, &numFiles)); + RINOK(SzReadNumber32(sd, &numFiles)) p->NumFiles = numFiles; for (;;) { UInt64 type; UInt64 size; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdEnd) break; - RINOK(ReadNumber(sd, &size)); + RINOK(ReadNumber(sd, &size)) if (size > sd->Size) return SZ_ERROR_ARCHIVE; if (type >= ((UInt32)1 << 8)) { - SKIP_DATA(sd, size); + SKIP_DATA(sd, size) } else switch ((unsigned)type) { @@ -1192,7 +1210,7 @@ static SRes SzReadHeader2( const Byte *namesData; Byte external; - SZ_READ_BYTE(external); + SZ_READ_BYTE(external) if (external == 0) { namesSize = (size_t)size - 1; @@ -1201,7 +1219,7 @@ static SRes SzReadHeader2( else { UInt32 index; - RINOK(SzReadNumber32(sd, &index)); + RINOK(SzReadNumber32(sd, &index)) if (index >= *numTempBufs) return SZ_ERROR_ARCHIVE; namesData = (tempBufs)[index].data; @@ -1210,25 +1228,25 @@ static SRes SzReadHeader2( if ((namesSize & 1) != 0) return SZ_ERROR_ARCHIVE; - MY_ALLOC(size_t, p->FileNameOffsets, numFiles + 1, allocMain); - MY_ALLOC_ZE_AND_CPY(p->FileNames, namesSize, namesData, allocMain); + MY_ALLOC(size_t, p->FileNameOffsets, numFiles + 1, allocMain) + MY_ALLOC_ZE_AND_CPY(p->FileNames, namesSize, namesData, allocMain) RINOK(SzReadFileNames(p->FileNames, namesSize, numFiles, p->FileNameOffsets)) if (external == 0) { - SKIP_DATA(sd, namesSize); + SKIP_DATA(sd, namesSize) } break; } case k7zIdEmptyStream: { - RINOK(RememberBitVector(sd, numFiles, &emptyStreams)); + RINOK(RememberBitVector(sd, numFiles, &emptyStreams)) numEmptyStreams = CountDefinedBits(emptyStreams, numFiles); emptyFiles = NULL; break; } case k7zIdEmptyFile: { - RINOK(RememberBitVector(sd, numEmptyStreams, &emptyFiles)); + RINOK(RememberBitVector(sd, numEmptyStreams, &emptyFiles)) break; } case k7zIdWinAttrib: @@ -1237,22 +1255,22 @@ static SRes SzReadHeader2( CSzData sdSwitch; CSzData *sdPtr; SzBitUi32s_Free(&p->Attribs, allocMain); - RINOK(ReadBitVector(sd, numFiles, &p->Attribs.Defs, allocMain)); + RINOK(ReadBitVector(sd, numFiles, &p->Attribs.Defs, allocMain)) - SZ_READ_BYTE(external); + SZ_READ_BYTE(external) if (external == 0) sdPtr = sd; else { UInt32 index; - RINOK(SzReadNumber32(sd, &index)); + RINOK(SzReadNumber32(sd, &index)) if (index >= *numTempBufs) return SZ_ERROR_ARCHIVE; sdSwitch.Data = (tempBufs)[index].data; sdSwitch.Size = (tempBufs)[index].size; sdPtr = &sdSwitch; } - RINOK(ReadUi32s(sdPtr, numFiles, &p->Attribs, allocMain)); + RINOK(ReadUi32s(sdPtr, numFiles, &p->Attribs, allocMain)) break; } /* @@ -1265,11 +1283,11 @@ static SRes SzReadHeader2( break; } */ - case k7zIdMTime: RINOK(ReadTime(&p->MTime, numFiles, sd, tempBufs, *numTempBufs, allocMain)); break; - case k7zIdCTime: RINOK(ReadTime(&p->CTime, numFiles, sd, tempBufs, *numTempBufs, allocMain)); break; + case k7zIdMTime: RINOK(ReadTime(&p->MTime, numFiles, sd, tempBufs, *numTempBufs, allocMain)) break; + case k7zIdCTime: RINOK(ReadTime(&p->CTime, numFiles, sd, tempBufs, *numTempBufs, allocMain)) break; default: { - SKIP_DATA(sd, size); + SKIP_DATA(sd, size) } } } @@ -1280,10 +1298,10 @@ static SRes SzReadHeader2( for (;;) { UInt64 type; - RINOK(ReadID(sd, &type)); + RINOK(ReadID(sd, &type)) if (type == k7zIdEnd) break; - RINOK(SkipData(sd)); + RINOK(SkipData(sd)) } { @@ -1295,40 +1313,37 @@ static SRes SzReadHeader2( UInt64 unpackPos = 0; const Byte *digestsDefs = NULL; const Byte *digestsVals = NULL; - UInt32 digestsValsIndex = 0; - UInt32 digestIndex; - Byte allDigestsDefined = 0; + UInt32 digestIndex = 0; Byte isDirMask = 0; Byte crcMask = 0; Byte mask = 0x80; - MY_ALLOC(UInt32, p->FolderToFile, p->db.NumFolders + 1, allocMain); - MY_ALLOC_ZE(UInt32, p->FileToFolder, p->NumFiles, allocMain); - MY_ALLOC(UInt64, p->UnpackPositions, p->NumFiles + 1, allocMain); - MY_ALLOC_ZE(Byte, p->IsDirs, (p->NumFiles + 7) >> 3, allocMain); + MY_ALLOC(UInt32, p->FolderToFile, p->db.NumFolders + 1, allocMain) + MY_ALLOC_ZE(UInt32, p->FileToFolder, p->NumFiles, allocMain) + MY_ALLOC(UInt64, p->UnpackPositions, p->NumFiles + 1, allocMain) + MY_ALLOC_ZE(Byte, p->IsDirs, (p->NumFiles + 7) >> 3, allocMain) - RINOK(SzBitUi32s_Alloc(&p->CRCs, p->NumFiles, allocMain)); + RINOK(SzBitUi32s_Alloc(&p->CRCs, p->NumFiles, allocMain)) if (ssi.sdCRCs.Size != 0) { - SZ_READ_BYTE_SD(&ssi.sdCRCs, allDigestsDefined); + Byte allDigestsDefined = 0; + SZ_READ_BYTE_SD_NOCHECK(&ssi.sdCRCs, allDigestsDefined) if (allDigestsDefined) digestsVals = ssi.sdCRCs.Data; else { - size_t numBytes = (ssi.NumSubDigests + 7) >> 3; + const size_t numBytes = (ssi.NumSubDigests + 7) >> 3; digestsDefs = ssi.sdCRCs.Data; digestsVals = digestsDefs + numBytes; } } - digestIndex = 0; - for (i = 0; i < numFiles; i++, mask >>= 1) { if (mask == 0) { - UInt32 byteIndex = (i - 1) >> 3; + const UInt32 byteIndex = (i - 1) >> 3; p->IsDirs[byteIndex] = isDirMask; p->CRCs.Defs[byteIndex] = crcMask; isDirMask = 0; @@ -1366,18 +1381,17 @@ static SRes SzReadHeader2( numSubStreams = 1; if (ssi.sdNumSubStreams.Data) { - RINOK(SzReadNumber32(&ssi.sdNumSubStreams, &numSubStreams)); + RINOK(SzReadNumber32(&ssi.sdNumSubStreams, &numSubStreams)) } remSubStreams = numSubStreams; if (numSubStreams != 0) break; { - UInt64 folderUnpackSize = SzAr_GetFolderUnpackSize(&p->db, folderIndex); + const UInt64 folderUnpackSize = SzAr_GetFolderUnpackSize(&p->db, folderIndex); unpackPos += folderUnpackSize; if (unpackPos < folderUnpackSize) return SZ_ERROR_ARCHIVE; } - folderIndex++; } } @@ -1389,47 +1403,44 @@ static SRes SzReadHeader2( if (--remSubStreams == 0) { - UInt64 folderUnpackSize = SzAr_GetFolderUnpackSize(&p->db, folderIndex); - UInt64 startFolderUnpackPos = p->UnpackPositions[p->FolderToFile[folderIndex]]; + const UInt64 folderUnpackSize = SzAr_GetFolderUnpackSize(&p->db, folderIndex); + const UInt64 startFolderUnpackPos = p->UnpackPositions[p->FolderToFile[folderIndex]]; if (folderUnpackSize < unpackPos - startFolderUnpackPos) return SZ_ERROR_ARCHIVE; unpackPos = startFolderUnpackPos + folderUnpackSize; if (unpackPos < folderUnpackSize) return SZ_ERROR_ARCHIVE; - if (numSubStreams == 1 && SzBitWithVals_Check(&p->db.FolderCRCs, i)) + if (numSubStreams == 1 && SzBitWithVals_Check(&p->db.FolderCRCs, folderIndex)) { p->CRCs.Vals[i] = p->db.FolderCRCs.Vals[folderIndex]; crcMask |= mask; } - else if (allDigestsDefined || (digestsDefs && SzBitArray_Check(digestsDefs, digestIndex))) - { - p->CRCs.Vals[i] = GetUi32(digestsVals + (size_t)digestsValsIndex * 4); - digestsValsIndex++; - crcMask |= mask; - } - folderIndex++; } else { UInt64 v; - RINOK(ReadNumber(&ssi.sdSizes, &v)); + RINOK(ReadNumber(&ssi.sdSizes, &v)) unpackPos += v; if (unpackPos < v) return SZ_ERROR_ARCHIVE; - if (allDigestsDefined || (digestsDefs && SzBitArray_Check(digestsDefs, digestIndex))) + } + if ((crcMask & mask) == 0 && digestsVals) + { + if (!digestsDefs || SzBitArray_Check(digestsDefs, digestIndex)) { - p->CRCs.Vals[i] = GetUi32(digestsVals + (size_t)digestsValsIndex * 4); - digestsValsIndex++; + p->CRCs.Vals[i] = GetUi32(digestsVals); + digestsVals += 4; crcMask |= mask; } + digestIndex++; } } if (mask != 0x80) { - UInt32 byteIndex = (i - 1) >> 3; + const UInt32 byteIndex = (i - 1) >> 3; p->IsDirs[byteIndex] = isDirMask; p->CRCs.Defs[byteIndex] = crcMask; } @@ -1446,7 +1457,7 @@ static SRes SzReadHeader2( break; if (!ssi.sdNumSubStreams.Data) return SZ_ERROR_ARCHIVE; - RINOK(SzReadNumber32(&ssi.sdNumSubStreams, &numSubStreams)); + RINOK(SzReadNumber32(&ssi.sdNumSubStreams, &numSubStreams)) if (numSubStreams != 0) return SZ_ERROR_ARCHIVE; /* @@ -1471,9 +1482,9 @@ static SRes SzReadHeader2( static SRes SzReadHeader( CSzArEx *p, CSzData *sd, - ILookInStream *inStream, - ISzAlloc *allocMain, - ISzAlloc *allocTemp) + ILookInStreamPtr inStream, + ISzAllocPtr allocMain, + ISzAllocPtr allocTemp) { UInt32 i; UInt32 numTempBufs = 0; @@ -1490,7 +1501,7 @@ static SRes SzReadHeader( for (i = 0; i < NUM_ADDITIONAL_STREAMS_MAX; i++) Buf_Free(tempBufs + i, allocTemp); - RINOK(res); + RINOK(res) if (sd->Size != 0) return SZ_ERROR_FAIL; @@ -1500,9 +1511,9 @@ static SRes SzReadHeader( static SRes SzArEx_Open2( CSzArEx *p, - ILookInStream *inStream, - ISzAlloc *allocMain, - ISzAlloc *allocTemp) + ILookInStreamPtr inStream, + ISzAllocPtr allocMain, + ISzAllocPtr allocTemp) { Byte header[k7zStartHeaderSize]; Int64 startArcPos; @@ -1513,9 +1524,9 @@ static SRes SzArEx_Open2( SRes res; startArcPos = 0; - RINOK(inStream->Seek(inStream, &startArcPos, SZ_SEEK_CUR)); + RINOK(ILookInStream_Seek(inStream, &startArcPos, SZ_SEEK_CUR)) - RINOK(LookInStream_Read2(inStream, header, k7zStartHeaderSize, SZ_ERROR_NO_ARCHIVE)); + RINOK(LookInStream_Read2(inStream, header, k7zStartHeaderSize, SZ_ERROR_NO_ARCHIVE)) if (!TestSignatureCandidate(header)) return SZ_ERROR_NO_ARCHIVE; @@ -1526,11 +1537,13 @@ static SRes SzArEx_Open2( nextHeaderSize = GetUi64(header + 20); nextHeaderCRC = GetUi32(header + 28); - p->startPosAfterHeader = startArcPos + k7zStartHeaderSize; + p->startPosAfterHeader = (UInt64)startArcPos + k7zStartHeaderSize; if (CrcCalc(header + 12, 20) != GetUi32(header + 8)) return SZ_ERROR_CRC; + p->db.RangeLimit = nextHeaderOffset; + nextHeaderSizeT = (size_t)nextHeaderSize; if (nextHeaderSizeT != nextHeaderSize) return SZ_ERROR_MEM; @@ -1542,14 +1555,14 @@ static SRes SzArEx_Open2( { Int64 pos = 0; - RINOK(inStream->Seek(inStream, &pos, SZ_SEEK_END)); - if ((UInt64)pos < startArcPos + nextHeaderOffset || - (UInt64)pos < startArcPos + k7zStartHeaderSize + nextHeaderOffset || - (UInt64)pos < startArcPos + k7zStartHeaderSize + nextHeaderOffset + nextHeaderSize) + RINOK(ILookInStream_Seek(inStream, &pos, SZ_SEEK_END)) + if ((UInt64)pos < (UInt64)startArcPos + nextHeaderOffset || + (UInt64)pos < (UInt64)startArcPos + k7zStartHeaderSize + nextHeaderOffset || + (UInt64)pos < (UInt64)startArcPos + k7zStartHeaderSize + nextHeaderOffset + nextHeaderSize) return SZ_ERROR_INPUT_EOF; } - RINOK(LookInStream_SeekTo(inStream, startArcPos + k7zStartHeaderSize + nextHeaderOffset)); + RINOK(LookInStream_SeekTo(inStream, (UInt64)startArcPos + k7zStartHeaderSize + nextHeaderOffset)) if (!Buf_Create(&buf, nextHeaderSizeT, allocTemp)) return SZ_ERROR_MEM; @@ -1575,6 +1588,8 @@ static SRes SzArEx_Open2( Buf_Init(&tempBuf); SzAr_Init(&tempAr); + tempAr.RangeLimit = p->db.RangeLimit; + res = SzReadAndDecodePackedStreams(inStream, &sd, &tempBuf, 1, p->startPosAfterHeader, &tempAr, allocTemp); SzAr_Free(&tempAr, allocTemp); @@ -1622,10 +1637,10 @@ static SRes SzArEx_Open2( } -SRes SzArEx_Open(CSzArEx *p, ILookInStream *inStream, - ISzAlloc *allocMain, ISzAlloc *allocTemp) +SRes SzArEx_Open(CSzArEx *p, ILookInStreamPtr inStream, + ISzAllocPtr allocMain, ISzAllocPtr allocTemp) { - SRes res = SzArEx_Open2(p, inStream, allocMain, allocTemp); + const SRes res = SzArEx_Open2(p, inStream, allocMain, allocTemp); if (res != SZ_OK) SzArEx_Free(p, allocMain); return res; @@ -1634,17 +1649,17 @@ SRes SzArEx_Open(CSzArEx *p, ILookInStream *inStream, SRes SzArEx_Extract( const CSzArEx *p, - ILookInStream *inStream, + ILookInStreamPtr inStream, UInt32 fileIndex, UInt32 *blockIndex, Byte **tempBuf, size_t *outBufferSize, size_t *offset, size_t *outSizeProcessed, - ISzAlloc *allocMain, - ISzAlloc *allocTemp) + ISzAllocPtr allocMain, + ISzAllocPtr allocTemp) { - UInt32 folderIndex = p->FileToFolder[fileIndex]; + const UInt32 folderIndex = p->FileToFolder[fileIndex]; SRes res = SZ_OK; *offset = 0; @@ -1652,7 +1667,7 @@ SRes SzArEx_Extract( if (folderIndex == (UInt32)-1) { - IAlloc_Free(allocMain, *tempBuf); + ISzAlloc_Free(allocMain, *tempBuf); *blockIndex = folderIndex; *tempBuf = NULL; *outBufferSize = 0; @@ -1661,18 +1676,18 @@ SRes SzArEx_Extract( if (*tempBuf == NULL || *blockIndex != folderIndex) { - UInt64 unpackSizeSpec = SzAr_GetFolderUnpackSize(&p->db, folderIndex); + const UInt64 unpackSizeSpec = SzAr_GetFolderUnpackSize(&p->db, folderIndex); /* UInt64 unpackSizeSpec = - p->UnpackPositions[p->FolderToFile[folderIndex + 1]] - + p->UnpackPositions[p->FolderToFile[(size_t)folderIndex + 1]] - p->UnpackPositions[p->FolderToFile[folderIndex]]; */ - size_t unpackSize = (size_t)unpackSizeSpec; + const size_t unpackSize = (size_t)unpackSizeSpec; if (unpackSize != unpackSizeSpec) return SZ_ERROR_MEM; *blockIndex = folderIndex; - IAlloc_Free(allocMain, *tempBuf); + ISzAlloc_Free(allocMain, *tempBuf); *tempBuf = NULL; if (res == SZ_OK) @@ -1680,7 +1695,7 @@ SRes SzArEx_Extract( *outBufferSize = unpackSize; if (unpackSize != 0) { - *tempBuf = (Byte *)IAlloc_Alloc(allocMain, unpackSize); + *tempBuf = (Byte *)ISzAlloc_Alloc(allocMain, unpackSize); if (*tempBuf == NULL) res = SZ_ERROR_MEM; } @@ -1695,9 +1710,9 @@ SRes SzArEx_Extract( if (res == SZ_OK) { - UInt64 unpackPos = p->UnpackPositions[fileIndex]; + const UInt64 unpackPos = p->UnpackPositions[fileIndex]; *offset = (size_t)(unpackPos - p->UnpackPositions[p->FolderToFile[folderIndex]]); - *outSizeProcessed = (size_t)(p->UnpackPositions[fileIndex + 1] - unpackPos); + *outSizeProcessed = (size_t)(p->UnpackPositions[(size_t)fileIndex + 1] - unpackPos); if (*offset + *outSizeProcessed > *outBufferSize) return SZ_ERROR_FAIL; if (SzBitWithVals_Check(&p->CRCs, fileIndex)) @@ -1711,8 +1726,8 @@ SRes SzArEx_Extract( size_t SzArEx_GetFileNameUtf16(const CSzArEx *p, size_t fileIndex, UInt16 *dest) { - size_t offs = p->FileNameOffsets[fileIndex]; - size_t len = p->FileNameOffsets[fileIndex + 1] - offs; + const size_t offs = p->FileNameOffsets[fileIndex]; + const size_t len = p->FileNameOffsets[fileIndex + 1] - offs; if (dest != 0) { size_t i; @@ -1744,7 +1759,7 @@ size_t SzArEx_GetFullNameLen(const CSzArEx *p, size_t fileIndex) UInt16 *SzArEx_GetFullNameUtf16_Back(const CSzArEx *p, size_t fileIndex, UInt16 *dest) { - Bool needSlash; + BoolInt needSlash; if (!p->FileNameOffsets) { *(--dest) = 0; diff --git a/3rdparty/lzma/C/7zBuf.c b/3rdparty/lzma/C/7zBuf.c index 089a5c4f21f..8865c32a81b 100644 --- a/3rdparty/lzma/C/7zBuf.c +++ b/3rdparty/lzma/C/7zBuf.c @@ -1,5 +1,5 @@ /* 7zBuf.c -- Byte Buffer -2013-01-21 : Igor Pavlov : Public domain */ +2017-04-03 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -11,7 +11,7 @@ void Buf_Init(CBuf *p) p->size = 0; } -int Buf_Create(CBuf *p, size_t size, ISzAlloc *alloc) +int Buf_Create(CBuf *p, size_t size, ISzAllocPtr alloc) { p->size = 0; if (size == 0) @@ -19,8 +19,8 @@ int Buf_Create(CBuf *p, size_t size, ISzAlloc *alloc) p->data = 0; return 1; } - p->data = (Byte *)alloc->Alloc(alloc, size); - if (p->data != 0) + p->data = (Byte *)ISzAlloc_Alloc(alloc, size); + if (p->data) { p->size = size; return 1; @@ -28,9 +28,9 @@ int Buf_Create(CBuf *p, size_t size, ISzAlloc *alloc) return 0; } -void Buf_Free(CBuf *p, ISzAlloc *alloc) +void Buf_Free(CBuf *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->data); + ISzAlloc_Free(alloc, p->data); p->data = 0; p->size = 0; } diff --git a/3rdparty/lzma/C/7zBuf.h b/3rdparty/lzma/C/7zBuf.h index 65f1d7a716e..c0ba8a7b607 100644 --- a/3rdparty/lzma/C/7zBuf.h +++ b/3rdparty/lzma/C/7zBuf.h @@ -1,8 +1,8 @@ /* 7zBuf.h -- Byte Buffer -2013-01-18 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __7Z_BUF_H -#define __7Z_BUF_H +#ifndef ZIP7_INC_7Z_BUF_H +#define ZIP7_INC_7Z_BUF_H #include "7zTypes.h" @@ -15,8 +15,8 @@ typedef struct } CBuf; void Buf_Init(CBuf *p); -int Buf_Create(CBuf *p, size_t size, ISzAlloc *alloc); -void Buf_Free(CBuf *p, ISzAlloc *alloc); +int Buf_Create(CBuf *p, size_t size, ISzAllocPtr alloc); +void Buf_Free(CBuf *p, ISzAllocPtr alloc); typedef struct { @@ -27,8 +27,8 @@ typedef struct void DynBuf_Construct(CDynBuf *p); void DynBuf_SeekToBeg(CDynBuf *p); -int DynBuf_Write(CDynBuf *p, const Byte *buf, size_t size, ISzAlloc *alloc); -void DynBuf_Free(CDynBuf *p, ISzAlloc *alloc); +int DynBuf_Write(CDynBuf *p, const Byte *buf, size_t size, ISzAllocPtr alloc); +void DynBuf_Free(CDynBuf *p, ISzAllocPtr alloc); EXTERN_C_END diff --git a/3rdparty/lzma/C/7zBuf2.c b/3rdparty/lzma/C/7zBuf2.c index 430f76bf830..20834741654 100644 --- a/3rdparty/lzma/C/7zBuf2.c +++ b/3rdparty/lzma/C/7zBuf2.c @@ -1,5 +1,5 @@ /* 7zBuf2.c -- Byte Buffer -2014-08-22 : Igor Pavlov : Public domain */ +2017-04-03 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -19,19 +19,20 @@ void DynBuf_SeekToBeg(CDynBuf *p) p->pos = 0; } -int DynBuf_Write(CDynBuf *p, const Byte *buf, size_t size, ISzAlloc *alloc) +int DynBuf_Write(CDynBuf *p, const Byte *buf, size_t size, ISzAllocPtr alloc) { if (size > p->size - p->pos) { size_t newSize = p->pos + size; Byte *data; newSize += newSize / 4; - data = (Byte *)alloc->Alloc(alloc, newSize); - if (data == 0) + data = (Byte *)ISzAlloc_Alloc(alloc, newSize); + if (!data) return 0; p->size = newSize; - memcpy(data, p->data, p->pos); - alloc->Free(alloc, p->data); + if (p->pos != 0) + memcpy(data, p->data, p->pos); + ISzAlloc_Free(alloc, p->data); p->data = data; } if (size != 0) @@ -42,9 +43,9 @@ int DynBuf_Write(CDynBuf *p, const Byte *buf, size_t size, ISzAlloc *alloc) return 1; } -void DynBuf_Free(CDynBuf *p, ISzAlloc *alloc) +void DynBuf_Free(CDynBuf *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->data); + ISzAlloc_Free(alloc, p->data); p->data = 0; p->size = 0; p->pos = 0; diff --git a/3rdparty/lzma/C/7zCrc.c b/3rdparty/lzma/C/7zCrc.c index dc6d6abdc0a..c995a8be40c 100644 --- a/3rdparty/lzma/C/7zCrc.c +++ b/3rdparty/lzma/C/7zCrc.c @@ -1,5 +1,5 @@ -/* 7zCrc.c -- CRC32 init -2015-03-10 : Igor Pavlov : Public domain */ +/* 7zCrc.c -- CRC32 calculation and init +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -13,38 +13,51 @@ #else #define CRC_NUM_TABLES 9 - #define CRC_UINT32_SWAP(v) ((v >> 24) | ((v >> 8) & 0xFF00) | ((v << 8) & 0xFF0000) | (v << 24)) - - UInt32 MY_FAST_CALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table); - UInt32 MY_FAST_CALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table); + UInt32 Z7_FASTCALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table); + UInt32 Z7_FASTCALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table); #endif #ifndef MY_CPU_BE - UInt32 MY_FAST_CALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table); - UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table); + UInt32 Z7_FASTCALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table); + UInt32 Z7_FASTCALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table); #endif -typedef UInt32 (MY_FAST_CALL *CRC_FUNC)(UInt32 v, const void *data, size_t size, const UInt32 *table); - +/* +extern +CRC_FUNC g_CrcUpdateT4; CRC_FUNC g_CrcUpdateT4; +*/ +extern CRC_FUNC g_CrcUpdateT8; +CRC_FUNC g_CrcUpdateT8; +extern +CRC_FUNC g_CrcUpdateT0_32; +CRC_FUNC g_CrcUpdateT0_32; +extern +CRC_FUNC g_CrcUpdateT0_64; +CRC_FUNC g_CrcUpdateT0_64; +extern +CRC_FUNC g_CrcUpdate; CRC_FUNC g_CrcUpdate; UInt32 g_CrcTable[256 * CRC_NUM_TABLES]; -UInt32 MY_FAST_CALL CrcUpdate(UInt32 v, const void *data, size_t size) +UInt32 Z7_FASTCALL CrcUpdate(UInt32 v, const void *data, size_t size) { return g_CrcUpdate(v, data, size, g_CrcTable); } -UInt32 MY_FAST_CALL CrcCalc(const void *data, size_t size) +UInt32 Z7_FASTCALL CrcCalc(const void *data, size_t size) { return g_CrcUpdate(CRC_INIT_VAL, data, size, g_CrcTable) ^ CRC_INIT_VAL; } +#if CRC_NUM_TABLES < 4 \ + || (CRC_NUM_TABLES == 4 && defined(MY_CPU_BE)) \ + || (!defined(MY_CPU_LE) && !defined(MY_CPU_BE)) #define CRC_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) - -UInt32 MY_FAST_CALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const UInt32 *table) +UInt32 Z7_FASTCALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const UInt32 *table); +UInt32 Z7_FASTCALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const UInt32 *table) { const Byte *p = (const Byte *)data; const Byte *pEnd = p + size; @@ -52,8 +65,184 @@ UInt32 MY_FAST_CALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const U v = CRC_UPDATE_BYTE_2(v, *p); return v; } +#endif + +/* ---------- hardware CRC ---------- */ + +#ifdef MY_CPU_LE + +#if defined(MY_CPU_ARM_OR_ARM64) + +// #pragma message("ARM*") + + #if defined(_MSC_VER) + #if defined(MY_CPU_ARM64) + #if (_MSC_VER >= 1910) + #ifndef __clang__ + #define USE_ARM64_CRC + #include <intrin.h> + #endif + #endif + #endif + #elif (defined(__clang__) && (__clang_major__ >= 3)) \ + || (defined(__GNUC__) && (__GNUC__ > 4)) + #if !defined(__ARM_FEATURE_CRC32) + #define __ARM_FEATURE_CRC32 1 + #if defined(__clang__) + #if defined(MY_CPU_ARM64) + #define ATTRIB_CRC __attribute__((__target__("crc"))) + #else + #define ATTRIB_CRC __attribute__((__target__("armv8-a,crc"))) + #endif + #else + #if defined(MY_CPU_ARM64) + #define ATTRIB_CRC __attribute__((__target__("+crc"))) + #else + #define ATTRIB_CRC __attribute__((__target__("arch=armv8-a+crc"))) + #endif + #endif + #endif + #if defined(__ARM_FEATURE_CRC32) + #define USE_ARM64_CRC + #include <arm_acle.h> + #endif + #endif + +#else + +// no hardware CRC + +// #define USE_CRC_EMU + +#ifdef USE_CRC_EMU + +#pragma message("ARM64 CRC emulation") + +Z7_FORCE_INLINE +UInt32 __crc32b(UInt32 v, UInt32 data) +{ + const UInt32 *table = g_CrcTable; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); + return v; +} + +Z7_FORCE_INLINE +UInt32 __crc32w(UInt32 v, UInt32 data) +{ + const UInt32 *table = g_CrcTable; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + return v; +} + +Z7_FORCE_INLINE +UInt32 __crc32d(UInt32 v, UInt64 data) +{ + const UInt32 *table = g_CrcTable; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8; + return v; +} + +#endif // USE_CRC_EMU + +#endif // defined(MY_CPU_ARM64) && defined(MY_CPU_LE) + + + +#if defined(USE_ARM64_CRC) || defined(USE_CRC_EMU) + +#define T0_32_UNROLL_BYTES (4 * 4) +#define T0_64_UNROLL_BYTES (4 * 8) + +#ifndef ATTRIB_CRC +#define ATTRIB_CRC +#endif +// #pragma message("USE ARM HW CRC") + +ATTRIB_CRC +UInt32 Z7_FASTCALL CrcUpdateT0_32(UInt32 v, const void *data, size_t size, const UInt32 *table); +ATTRIB_CRC +UInt32 Z7_FASTCALL CrcUpdateT0_32(UInt32 v, const void *data, size_t size, const UInt32 *table) +{ + const Byte *p = (const Byte *)data; + UNUSED_VAR(table); + + for (; size != 0 && ((unsigned)(ptrdiff_t)p & (T0_32_UNROLL_BYTES - 1)) != 0; size--) + v = __crc32b(v, *p++); + + if (size >= T0_32_UNROLL_BYTES) + { + const Byte *lim = p + size; + size &= (T0_32_UNROLL_BYTES - 1); + lim -= size; + do + { + v = __crc32w(v, *(const UInt32 *)(const void *)(p)); + v = __crc32w(v, *(const UInt32 *)(const void *)(p + 4)); p += 2 * 4; + v = __crc32w(v, *(const UInt32 *)(const void *)(p)); + v = __crc32w(v, *(const UInt32 *)(const void *)(p + 4)); p += 2 * 4; + } + while (p != lim); + } + + for (; size != 0; size--) + v = __crc32b(v, *p++); + + return v; +} + +ATTRIB_CRC +UInt32 Z7_FASTCALL CrcUpdateT0_64(UInt32 v, const void *data, size_t size, const UInt32 *table); +ATTRIB_CRC +UInt32 Z7_FASTCALL CrcUpdateT0_64(UInt32 v, const void *data, size_t size, const UInt32 *table) +{ + const Byte *p = (const Byte *)data; + UNUSED_VAR(table); + + for (; size != 0 && ((unsigned)(ptrdiff_t)p & (T0_64_UNROLL_BYTES - 1)) != 0; size--) + v = __crc32b(v, *p++); + + if (size >= T0_64_UNROLL_BYTES) + { + const Byte *lim = p + size; + size &= (T0_64_UNROLL_BYTES - 1); + lim -= size; + do + { + v = __crc32d(v, *(const UInt64 *)(const void *)(p)); + v = __crc32d(v, *(const UInt64 *)(const void *)(p + 8)); p += 2 * 8; + v = __crc32d(v, *(const UInt64 *)(const void *)(p)); + v = __crc32d(v, *(const UInt64 *)(const void *)(p + 8)); p += 2 * 8; + } + while (p != lim); + } + + for (; size != 0; size--) + v = __crc32b(v, *p++); + + return v; +} + +#undef T0_32_UNROLL_BYTES +#undef T0_64_UNROLL_BYTES -void MY_FAST_CALL CrcGenerateTable() +#endif // defined(USE_ARM64_CRC) || defined(USE_CRC_EMU) + +#endif // MY_CPU_LE + + + + +void Z7_FASTCALL CrcGenerateTable(void) { UInt32 i; for (i = 0; i < 256; i++) @@ -61,68 +250,91 @@ void MY_FAST_CALL CrcGenerateTable() UInt32 r = i; unsigned j; for (j = 0; j < 8; j++) - r = (r >> 1) ^ (kCrcPoly & ~((r & 1) - 1)); + r = (r >> 1) ^ (kCrcPoly & ((UInt32)0 - (r & 1))); g_CrcTable[i] = r; } - for (; i < 256 * CRC_NUM_TABLES; i++) + for (i = 256; i < 256 * CRC_NUM_TABLES; i++) { - UInt32 r = g_CrcTable[i - 256]; + const UInt32 r = g_CrcTable[(size_t)i - 256]; g_CrcTable[i] = g_CrcTable[r & 0xFF] ^ (r >> 8); } #if CRC_NUM_TABLES < 4 - - g_CrcUpdate = CrcUpdateT1; - - #else - - #ifdef MY_CPU_LE - - g_CrcUpdateT4 = CrcUpdateT4; - g_CrcUpdate = CrcUpdateT4; - - #if CRC_NUM_TABLES >= 8 + g_CrcUpdate = CrcUpdateT1; + #elif defined(MY_CPU_LE) + // g_CrcUpdateT4 = CrcUpdateT4; + #if CRC_NUM_TABLES < 8 + g_CrcUpdate = CrcUpdateT4; + #else // CRC_NUM_TABLES >= 8 g_CrcUpdateT8 = CrcUpdateT8; - + /* #ifdef MY_CPU_X86_OR_AMD64 if (!CPU_Is_InOrder()) - g_CrcUpdate = CrcUpdateT8; #endif + */ + g_CrcUpdate = CrcUpdateT8; #endif - #else { - #ifndef MY_CPU_BE + #ifndef MY_CPU_BE UInt32 k = 0x01020304; const Byte *p = (const Byte *)&k; if (p[0] == 4 && p[1] == 3) { - g_CrcUpdateT4 = CrcUpdateT4; - g_CrcUpdate = CrcUpdateT4; - #if CRC_NUM_TABLES >= 8 - g_CrcUpdateT8 = CrcUpdateT8; - // g_CrcUpdate = CrcUpdateT8; + #if CRC_NUM_TABLES < 8 + // g_CrcUpdateT4 = CrcUpdateT4; + g_CrcUpdate = CrcUpdateT4; + #else // CRC_NUM_TABLES >= 8 + g_CrcUpdateT8 = CrcUpdateT8; + g_CrcUpdate = CrcUpdateT8; #endif } else if (p[0] != 1 || p[1] != 2) g_CrcUpdate = CrcUpdateT1; else - #endif + #endif // MY_CPU_BE { for (i = 256 * CRC_NUM_TABLES - 1; i >= 256; i--) { - UInt32 x = g_CrcTable[i - 256]; - g_CrcTable[i] = CRC_UINT32_SWAP(x); + const UInt32 x = g_CrcTable[(size_t)i - 256]; + g_CrcTable[i] = Z7_BSWAP32(x); } - g_CrcUpdateT4 = CrcUpdateT1_BeT4; - g_CrcUpdate = CrcUpdateT1_BeT4; - #if CRC_NUM_TABLES >= 8 - g_CrcUpdateT8 = CrcUpdateT1_BeT8; - // g_CrcUpdate = CrcUpdateT1_BeT8; + #if CRC_NUM_TABLES <= 4 + g_CrcUpdate = CrcUpdateT1; + #elif CRC_NUM_TABLES <= 8 + // g_CrcUpdateT4 = CrcUpdateT1_BeT4; + g_CrcUpdate = CrcUpdateT1_BeT4; + #else // CRC_NUM_TABLES > 8 + g_CrcUpdateT8 = CrcUpdateT1_BeT8; + g_CrcUpdate = CrcUpdateT1_BeT8; #endif } } - #endif + #endif // CRC_NUM_TABLES < 4 + #ifdef MY_CPU_LE + #ifdef USE_ARM64_CRC + if (CPU_IsSupported_CRC32()) + { + g_CrcUpdateT0_32 = CrcUpdateT0_32; + g_CrcUpdateT0_64 = CrcUpdateT0_64; + g_CrcUpdate = + #if defined(MY_CPU_ARM) + CrcUpdateT0_32; + #else + CrcUpdateT0_64; + #endif + } + #endif + + #ifdef USE_CRC_EMU + g_CrcUpdateT0_32 = CrcUpdateT0_32; + g_CrcUpdateT0_64 = CrcUpdateT0_64; + g_CrcUpdate = CrcUpdateT0_64; + #endif #endif } + +#undef kCrcPoly +#undef CRC64_NUM_TABLES +#undef CRC_UPDATE_BYTE_2 diff --git a/3rdparty/lzma/C/7zCrc.h b/3rdparty/lzma/C/7zCrc.h index 8fd57958715..4afaeae4a84 100644 --- a/3rdparty/lzma/C/7zCrc.h +++ b/3rdparty/lzma/C/7zCrc.h @@ -1,8 +1,8 @@ /* 7zCrc.h -- CRC32 calculation -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __7Z_CRC_H -#define __7Z_CRC_H +#ifndef ZIP7_INC_7Z_CRC_H +#define ZIP7_INC_7Z_CRC_H #include "7zTypes.h" @@ -11,14 +11,16 @@ EXTERN_C_BEGIN extern UInt32 g_CrcTable[]; /* Call CrcGenerateTable one time before other CRC functions */ -void MY_FAST_CALL CrcGenerateTable(void); +void Z7_FASTCALL CrcGenerateTable(void); #define CRC_INIT_VAL 0xFFFFFFFF #define CRC_GET_DIGEST(crc) ((crc) ^ CRC_INIT_VAL) #define CRC_UPDATE_BYTE(crc, b) (g_CrcTable[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) -UInt32 MY_FAST_CALL CrcUpdate(UInt32 crc, const void *data, size_t size); -UInt32 MY_FAST_CALL CrcCalc(const void *data, size_t size); +UInt32 Z7_FASTCALL CrcUpdate(UInt32 crc, const void *data, size_t size); +UInt32 Z7_FASTCALL CrcCalc(const void *data, size_t size); + +typedef UInt32 (Z7_FASTCALL *CRC_FUNC)(UInt32 v, const void *data, size_t size, const UInt32 *table); EXTERN_C_END diff --git a/3rdparty/lzma/C/7zCrcOpt.c b/3rdparty/lzma/C/7zCrcOpt.c index d1e1cd7bae1..9c649290df7 100644 --- a/3rdparty/lzma/C/7zCrcOpt.c +++ b/3rdparty/lzma/C/7zCrcOpt.c @@ -1,5 +1,5 @@ /* 7zCrcOpt.c -- CRC32 calculation -2015-03-01 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -9,26 +9,28 @@ #define CRC_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) -UInt32 MY_FAST_CALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table) +UInt32 Z7_FASTCALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table); +UInt32 Z7_FASTCALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table) { const Byte *p = (const Byte *)data; for (; size > 0 && ((unsigned)(ptrdiff_t)p & 3) != 0; size--, p++) v = CRC_UPDATE_BYTE_2(v, *p); for (; size >= 4; size -= 4, p += 4) { - v ^= *(const UInt32 *)p; + v ^= *(const UInt32 *)(const void *)p; v = - table[0x300 + ((v ) & 0xFF)] - ^ table[0x200 + ((v >> 8) & 0xFF)] - ^ table[0x100 + ((v >> 16) & 0xFF)] - ^ table[0x000 + ((v >> 24))]; + (table + 0x300)[((v ) & 0xFF)] + ^ (table + 0x200)[((v >> 8) & 0xFF)] + ^ (table + 0x100)[((v >> 16) & 0xFF)] + ^ (table + 0x000)[((v >> 24))]; } for (; size > 0; size--, p++) v = CRC_UPDATE_BYTE_2(v, *p); return v; } -UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table) +UInt32 Z7_FASTCALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table); +UInt32 Z7_FASTCALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table) { const Byte *p = (const Byte *)data; for (; size > 0 && ((unsigned)(ptrdiff_t)p & 7) != 0; size--, p++) @@ -36,18 +38,18 @@ UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const U for (; size >= 8; size -= 8, p += 8) { UInt32 d; - v ^= *(const UInt32 *)p; + v ^= *(const UInt32 *)(const void *)p; v = - table[0x700 + ((v ) & 0xFF)] - ^ table[0x600 + ((v >> 8) & 0xFF)] - ^ table[0x500 + ((v >> 16) & 0xFF)] - ^ table[0x400 + ((v >> 24))]; - d = *((const UInt32 *)p + 1); + (table + 0x700)[((v ) & 0xFF)] + ^ (table + 0x600)[((v >> 8) & 0xFF)] + ^ (table + 0x500)[((v >> 16) & 0xFF)] + ^ (table + 0x400)[((v >> 24))]; + d = *((const UInt32 *)(const void *)p + 1); v ^= - table[0x300 + ((d ) & 0xFF)] - ^ table[0x200 + ((d >> 8) & 0xFF)] - ^ table[0x100 + ((d >> 16) & 0xFF)] - ^ table[0x000 + ((d >> 24))]; + (table + 0x300)[((d ) & 0xFF)] + ^ (table + 0x200)[((d >> 8) & 0xFF)] + ^ (table + 0x100)[((d >> 16) & 0xFF)] + ^ (table + 0x000)[((d >> 24))]; } for (; size > 0; size--, p++) v = CRC_UPDATE_BYTE_2(v, *p); @@ -59,11 +61,11 @@ UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const U #ifndef MY_CPU_LE -#define CRC_UINT32_SWAP(v) ((v >> 24) | ((v >> 8) & 0xFF00) | ((v << 8) & 0xFF0000) | (v << 24)) +#define CRC_UINT32_SWAP(v) Z7_BSWAP32(v) #define CRC_UPDATE_BYTE_2_BE(crc, b) (table[(((crc) >> 24) ^ (b))] ^ ((crc) << 8)) -UInt32 MY_FAST_CALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table) +UInt32 Z7_FASTCALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table) { const Byte *p = (const Byte *)data; table += 0x100; @@ -72,19 +74,19 @@ UInt32 MY_FAST_CALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, co v = CRC_UPDATE_BYTE_2_BE(v, *p); for (; size >= 4; size -= 4, p += 4) { - v ^= *(const UInt32 *)p; + v ^= *(const UInt32 *)(const void *)p; v = - table[0x000 + ((v ) & 0xFF)] - ^ table[0x100 + ((v >> 8) & 0xFF)] - ^ table[0x200 + ((v >> 16) & 0xFF)] - ^ table[0x300 + ((v >> 24))]; + (table + 0x000)[((v ) & 0xFF)] + ^ (table + 0x100)[((v >> 8) & 0xFF)] + ^ (table + 0x200)[((v >> 16) & 0xFF)] + ^ (table + 0x300)[((v >> 24))]; } for (; size > 0; size--, p++) v = CRC_UPDATE_BYTE_2_BE(v, *p); return CRC_UINT32_SWAP(v); } -UInt32 MY_FAST_CALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table) +UInt32 Z7_FASTCALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table) { const Byte *p = (const Byte *)data; table += 0x100; @@ -94,18 +96,18 @@ UInt32 MY_FAST_CALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, co for (; size >= 8; size -= 8, p += 8) { UInt32 d; - v ^= *(const UInt32 *)p; + v ^= *(const UInt32 *)(const void *)p; v = - table[0x400 + ((v ) & 0xFF)] - ^ table[0x500 + ((v >> 8) & 0xFF)] - ^ table[0x600 + ((v >> 16) & 0xFF)] - ^ table[0x700 + ((v >> 24))]; - d = *((const UInt32 *)p + 1); + (table + 0x400)[((v ) & 0xFF)] + ^ (table + 0x500)[((v >> 8) & 0xFF)] + ^ (table + 0x600)[((v >> 16) & 0xFF)] + ^ (table + 0x700)[((v >> 24))]; + d = *((const UInt32 *)(const void *)p + 1); v ^= - table[0x000 + ((d ) & 0xFF)] - ^ table[0x100 + ((d >> 8) & 0xFF)] - ^ table[0x200 + ((d >> 16) & 0xFF)] - ^ table[0x300 + ((d >> 24))]; + (table + 0x000)[((d ) & 0xFF)] + ^ (table + 0x100)[((d >> 8) & 0xFF)] + ^ (table + 0x200)[((d >> 16) & 0xFF)] + ^ (table + 0x300)[((d >> 24))]; } for (; size > 0; size--, p++) v = CRC_UPDATE_BYTE_2_BE(v, *p); diff --git a/3rdparty/lzma/C/7zDec.c b/3rdparty/lzma/C/7zDec.c index c45d6bf8d4b..96c60359a47 100644 --- a/3rdparty/lzma/C/7zDec.c +++ b/3rdparty/lzma/C/7zDec.c @@ -1,11 +1,11 @@ /* 7zDec.c -- Decoding from 7z folder -2015-11-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" #include <string.h> -/* #define _7ZIP_PPMD_SUPPPORT */ +/* #define Z7_PPMD_SUPPORT */ #include "7z.h" #include "7zCrc.h" @@ -16,68 +16,93 @@ #include "Delta.h" #include "LzmaDec.h" #include "Lzma2Dec.h" -#ifdef _7ZIP_PPMD_SUPPPORT +#ifdef Z7_PPMD_SUPPORT #include "Ppmd7.h" #endif #define k_Copy 0 -#define k_Delta 3 +#ifndef Z7_NO_METHOD_LZMA2 #define k_LZMA2 0x21 +#endif #define k_LZMA 0x30101 -#define k_BCJ 0x3030103 #define k_BCJ2 0x303011B + +#if !defined(Z7_NO_METHODS_FILTERS) +#define Z7_USE_BRANCH_FILTER +#endif + +#if !defined(Z7_NO_METHODS_FILTERS) || \ + defined(Z7_USE_NATIVE_BRANCH_FILTER) && defined(MY_CPU_ARM64) +#define Z7_USE_FILTER_ARM64 +#ifndef Z7_USE_BRANCH_FILTER +#define Z7_USE_BRANCH_FILTER +#endif +#define k_ARM64 0xa +#endif + +#if !defined(Z7_NO_METHODS_FILTERS) || \ + defined(Z7_USE_NATIVE_BRANCH_FILTER) && defined(MY_CPU_ARMT) +#define Z7_USE_FILTER_ARMT +#ifndef Z7_USE_BRANCH_FILTER +#define Z7_USE_BRANCH_FILTER +#endif +#define k_ARMT 0x3030701 +#endif + +#ifndef Z7_NO_METHODS_FILTERS +#define k_Delta 3 +#define k_BCJ 0x3030103 #define k_PPC 0x3030205 #define k_IA64 0x3030401 #define k_ARM 0x3030501 -#define k_ARMT 0x3030701 #define k_SPARC 0x3030805 +#endif - -#ifdef _7ZIP_PPMD_SUPPPORT +#ifdef Z7_PPMD_SUPPORT #define k_PPMD 0x30401 typedef struct { - IByteIn p; + IByteIn vt; const Byte *cur; const Byte *end; const Byte *begin; UInt64 processed; - Bool extra; + BoolInt extra; SRes res; - ILookInStream *inStream; + ILookInStreamPtr inStream; } CByteInToLook; -static Byte ReadByte(void *pp) +static Byte ReadByte(IByteInPtr pp) { - CByteInToLook *p = (CByteInToLook *)pp; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CByteInToLook) if (p->cur != p->end) return *p->cur++; if (p->res == SZ_OK) { - size_t size = p->cur - p->begin; + size_t size = (size_t)(p->cur - p->begin); p->processed += size; - p->res = p->inStream->Skip(p->inStream, size); + p->res = ILookInStream_Skip(p->inStream, size); size = (1 << 25); - p->res = p->inStream->Look(p->inStream, (const void **)&p->begin, &size); + p->res = ILookInStream_Look(p->inStream, (const void **)&p->begin, &size); p->cur = p->begin; p->end = p->begin + size; if (size != 0) - return *p->cur++;; + return *p->cur++; } p->extra = True; return 0; } -static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStream *inStream, - Byte *outBuffer, SizeT outSize, ISzAlloc *allocMain) +static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream, + Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain) { CPpmd7 ppmd; CByteInToLook s; SRes res = SZ_OK; - s.p.Read = ReadByte; + s.vt.Read = ReadByte; s.inStream = inStream; s.begin = s.end = s.cur = NULL; s.extra = False; @@ -101,28 +126,32 @@ static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, I Ppmd7_Init(&ppmd, order); } { - CPpmd7z_RangeDec rc; - Ppmd7z_RangeDec_CreateVTable(&rc); - rc.Stream = &s.p; - if (!Ppmd7z_RangeDec_Init(&rc)) + ppmd.rc.dec.Stream = &s.vt; + if (!Ppmd7z_RangeDec_Init(&ppmd.rc.dec)) res = SZ_ERROR_DATA; - else if (s.extra) - res = (s.res != SZ_OK ? s.res : SZ_ERROR_DATA); - else + else if (!s.extra) { - SizeT i; - for (i = 0; i < outSize; i++) + Byte *buf = outBuffer; + const Byte *lim = buf + outSize; + for (; buf != lim; buf++) { - int sym = Ppmd7_DecodeSymbol(&ppmd, &rc.p); + int sym = Ppmd7z_DecodeSymbol(&ppmd); if (s.extra || sym < 0) break; - outBuffer[i] = (Byte)sym; + *buf = (Byte)sym; } - if (i != outSize) - res = (s.res != SZ_OK ? s.res : SZ_ERROR_DATA); - else if (s.processed + (s.cur - s.begin) != inSize || !Ppmd7z_RangeDec_IsFinishedOK(&rc)) + if (buf != lim) + res = SZ_ERROR_DATA; + else if (!Ppmd7z_RangeDec_IsFinishedOK(&ppmd.rc.dec)) + { + /* if (Ppmd7z_DecodeSymbol(&ppmd) != PPMD7_SYM_END || !Ppmd7z_RangeDec_IsFinishedOK(&ppmd.rc.dec)) */ res = SZ_ERROR_DATA; + } } + if (s.extra) + res = (s.res != SZ_OK ? s.res : SZ_ERROR_DATA); + else if (s.processed + (size_t)(s.cur - s.begin) != inSize) + res = SZ_ERROR_DATA; } Ppmd7_Free(&ppmd, allocMain); return res; @@ -131,14 +160,14 @@ static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, I #endif -static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStream *inStream, - Byte *outBuffer, SizeT outSize, ISzAlloc *allocMain) +static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream, + Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain) { CLzmaDec state; SRes res = SZ_OK; - LzmaDec_Construct(&state); - RINOK(LzmaDec_AllocateProbs(&state, props, propsSize, allocMain)); + LzmaDec_CONSTRUCT(&state) + RINOK(LzmaDec_AllocateProbs(&state, props, propsSize, allocMain)) state.dic = outBuffer; state.dicBufSize = outSize; LzmaDec_Init(&state); @@ -149,14 +178,14 @@ static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, I size_t lookahead = (1 << 18); if (lookahead > inSize) lookahead = (size_t)inSize; - res = inStream->Look(inStream, &inBuf, &lookahead); + res = ILookInStream_Look(inStream, &inBuf, &lookahead); if (res != SZ_OK) break; { SizeT inProcessed = (SizeT)lookahead, dicPos = state.dicPos; ELzmaStatus status; - res = LzmaDec_DecodeToDic(&state, outSize, inBuf, &inProcessed, LZMA_FINISH_END, &status); + res = LzmaDec_DecodeToDic(&state, outSize, (const Byte *)inBuf, &inProcessed, LZMA_FINISH_END, &status); lookahead -= inProcessed; inSize -= inProcessed; if (res != SZ_OK) @@ -178,7 +207,7 @@ static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, I break; } - res = inStream->Skip((void *)inStream, inProcessed); + res = ILookInStream_Skip(inStream, inProcessed); if (res != SZ_OK) break; } @@ -189,18 +218,18 @@ static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, I } -#ifndef _7Z_NO_METHOD_LZMA2 +#ifndef Z7_NO_METHOD_LZMA2 -static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStream *inStream, - Byte *outBuffer, SizeT outSize, ISzAlloc *allocMain) +static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream, + Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain) { CLzma2Dec state; SRes res = SZ_OK; - Lzma2Dec_Construct(&state); + Lzma2Dec_CONSTRUCT(&state) if (propsSize != 1) return SZ_ERROR_DATA; - RINOK(Lzma2Dec_AllocateProbs(&state, props[0], allocMain)); + RINOK(Lzma2Dec_AllocateProbs(&state, props[0], allocMain)) state.decoder.dic = outBuffer; state.decoder.dicBufSize = outSize; Lzma2Dec_Init(&state); @@ -211,14 +240,14 @@ static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, size_t lookahead = (1 << 18); if (lookahead > inSize) lookahead = (size_t)inSize; - res = inStream->Look(inStream, &inBuf, &lookahead); + res = ILookInStream_Look(inStream, &inBuf, &lookahead); if (res != SZ_OK) break; { SizeT inProcessed = (SizeT)lookahead, dicPos = state.decoder.dicPos; ELzmaStatus status; - res = Lzma2Dec_DecodeToDic(&state, outSize, inBuf, &inProcessed, LZMA_FINISH_END, &status); + res = Lzma2Dec_DecodeToDic(&state, outSize, (const Byte *)inBuf, &inProcessed, LZMA_FINISH_END, &status); lookahead -= inProcessed; inSize -= inProcessed; if (res != SZ_OK) @@ -237,7 +266,7 @@ static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, break; } - res = inStream->Skip((void *)inStream, inProcessed); + res = ILookInStream_Skip(inStream, inProcessed); if (res != SZ_OK) break; } @@ -250,7 +279,7 @@ static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, #endif -static SRes SzDecodeCopy(UInt64 inSize, ILookInStream *inStream, Byte *outBuffer) +static SRes SzDecodeCopy(UInt64 inSize, ILookInStreamPtr inStream, Byte *outBuffer) { while (inSize > 0) { @@ -258,35 +287,35 @@ static SRes SzDecodeCopy(UInt64 inSize, ILookInStream *inStream, Byte *outBuffer size_t curSize = (1 << 18); if (curSize > inSize) curSize = (size_t)inSize; - RINOK(inStream->Look(inStream, &inBuf, &curSize)); + RINOK(ILookInStream_Look(inStream, &inBuf, &curSize)) if (curSize == 0) return SZ_ERROR_INPUT_EOF; memcpy(outBuffer, inBuf, curSize); outBuffer += curSize; inSize -= curSize; - RINOK(inStream->Skip((void *)inStream, curSize)); + RINOK(ILookInStream_Skip(inStream, curSize)) } return SZ_OK; } -static Bool IS_MAIN_METHOD(UInt32 m) +static BoolInt IS_MAIN_METHOD(UInt32 m) { switch (m) { case k_Copy: case k_LZMA: - #ifndef _7Z_NO_METHOD_LZMA2 + #ifndef Z7_NO_METHOD_LZMA2 case k_LZMA2: - #endif - #ifdef _7ZIP_PPMD_SUPPPORT + #endif + #ifdef Z7_PPMD_SUPPORT case k_PPMD: - #endif + #endif return True; } return False; } -static Bool IS_SUPPORTED_CODER(const CSzCoderInfo *c) +static BoolInt IS_SUPPORTED_CODER(const CSzCoderInfo *c) { return c->NumStreams == 1 @@ -310,7 +339,7 @@ static SRes CheckSupportedFolder(const CSzFolder *f) } - #ifndef _7Z_NO_METHODS_FILTERS + #if defined(Z7_USE_BRANCH_FILTER) if (f->NumCoders == 2) { @@ -326,13 +355,20 @@ static SRes CheckSupportedFolder(const CSzFolder *f) return SZ_ERROR_UNSUPPORTED; switch ((UInt32)c->MethodID) { + #if !defined(Z7_NO_METHODS_FILTERS) case k_Delta: case k_BCJ: case k_PPC: case k_IA64: case k_SPARC: case k_ARM: + #endif + #ifdef Z7_USE_FILTER_ARM64 + case k_ARM64: + #endif + #ifdef Z7_USE_FILTER_ARMT case k_ARMT: + #endif break; default: return SZ_ERROR_UNSUPPORTED; @@ -365,14 +401,17 @@ static SRes CheckSupportedFolder(const CSzFolder *f) return SZ_ERROR_UNSUPPORTED; } -#define CASE_BRA_CONV(isa) case k_ ## isa: isa ## _Convert(outBuffer, outSize, 0, 0); break; + + + + static SRes SzFolder_Decode2(const CSzFolder *folder, const Byte *propsData, const UInt64 *unpackSizes, const UInt64 *packPositions, - ILookInStream *inStream, UInt64 startPos, - Byte *outBuffer, SizeT outSize, ISzAlloc *allocMain, + ILookInStreamPtr inStream, UInt64 startPos, + Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain, Byte *tempBuf[]) { UInt32 ci; @@ -380,7 +419,7 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, SizeT tempSize3 = 0; Byte *tempBuf3 = 0; - RINOK(CheckSupportedFolder(folder)); + RINOK(CheckSupportedFolder(folder)) for (ci = 0; ci < folder->NumCoders; ci++) { @@ -395,8 +434,8 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, SizeT outSizeCur = outSize; if (folder->NumCoders == 4) { - UInt32 indices[] = { 3, 2, 0 }; - UInt64 unpackSize = unpackSizes[ci]; + const UInt32 indices[] = { 3, 2, 0 }; + const UInt64 unpackSize = unpackSizes[ci]; si = indices[ci]; if (ci < 2) { @@ -404,7 +443,7 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, outSizeCur = (SizeT)unpackSize; if (outSizeCur != unpackSize) return SZ_ERROR_MEM; - temp = (Byte *)IAlloc_Alloc(allocMain, outSizeCur); + temp = (Byte *)ISzAlloc_Alloc(allocMain, outSizeCur); if (!temp && outSizeCur != 0) return SZ_ERROR_MEM; outBufCur = tempBuf[1 - ci] = temp; @@ -421,38 +460,38 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, return SZ_ERROR_UNSUPPORTED; } offset = packPositions[si]; - inSize = packPositions[si + 1] - offset; - RINOK(LookInStream_SeekTo(inStream, startPos + offset)); + inSize = packPositions[(size_t)si + 1] - offset; + RINOK(LookInStream_SeekTo(inStream, startPos + offset)) if (coder->MethodID == k_Copy) { if (inSize != outSizeCur) /* check it */ return SZ_ERROR_DATA; - RINOK(SzDecodeCopy(inSize, inStream, outBufCur)); + RINOK(SzDecodeCopy(inSize, inStream, outBufCur)) } else if (coder->MethodID == k_LZMA) { - RINOK(SzDecodeLzma(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)); + RINOK(SzDecodeLzma(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)) } - #ifndef _7Z_NO_METHOD_LZMA2 + #ifndef Z7_NO_METHOD_LZMA2 else if (coder->MethodID == k_LZMA2) { - RINOK(SzDecodeLzma2(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)); + RINOK(SzDecodeLzma2(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)) } - #endif - #ifdef _7ZIP_PPMD_SUPPPORT + #endif + #ifdef Z7_PPMD_SUPPORT else if (coder->MethodID == k_PPMD) { - RINOK(SzDecodePpmd(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)); + RINOK(SzDecodePpmd(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain)) } - #endif + #endif else return SZ_ERROR_UNSUPPORTED; } else if (coder->MethodID == k_BCJ2) { - UInt64 offset = packPositions[1]; - UInt64 s3Size = packPositions[2] - offset; + const UInt64 offset = packPositions[1]; + const UInt64 s3Size = packPositions[2] - offset; if (ci != 3) return SZ_ERROR_UNSUPPORTED; @@ -460,12 +499,12 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, tempSizes[2] = (SizeT)s3Size; if (tempSizes[2] != s3Size) return SZ_ERROR_MEM; - tempBuf[2] = (Byte *)IAlloc_Alloc(allocMain, tempSizes[2]); + tempBuf[2] = (Byte *)ISzAlloc_Alloc(allocMain, tempSizes[2]); if (!tempBuf[2] && tempSizes[2] != 0) return SZ_ERROR_MEM; - RINOK(LookInStream_SeekTo(inStream, startPos + offset)); - RINOK(SzDecodeCopy(s3Size, inStream, tempBuf[2])); + RINOK(LookInStream_SeekTo(inStream, startPos + offset)) + RINOK(SzDecodeCopy(s3Size, inStream, tempBuf[2])) if ((tempSizes[0] & 3) != 0 || (tempSizes[1] & 3) != 0 || @@ -484,26 +523,22 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, p.destLim = outBuffer + outSize; Bcj2Dec_Init(&p); - RINOK(Bcj2Dec_Decode(&p)); + RINOK(Bcj2Dec_Decode(&p)) { unsigned i; for (i = 0; i < 4; i++) if (p.bufs[i] != p.lims[i]) return SZ_ERROR_DATA; - - if (!Bcj2Dec_IsFinished(&p)) - return SZ_ERROR_DATA; - - if (p.dest != p.destLim - || p.state != BCJ2_STREAM_MAIN) + if (p.dest != p.destLim || !Bcj2Dec_IsMaybeFinished(&p)) return SZ_ERROR_DATA; } } } - #ifndef _7Z_NO_METHODS_FILTERS + #if defined(Z7_USE_BRANCH_FILTER) else if (ci == 1) { + #if !defined(Z7_NO_METHODS_FILTERS) if (coder->MethodID == k_Delta) { if (coder->PropsSize != 1) @@ -513,31 +548,53 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, Delta_Init(state); Delta_Decode(state, (unsigned)(propsData[coder->PropsOffset]) + 1, outBuffer, outSize); } + continue; } - else + #endif + + #ifdef Z7_USE_FILTER_ARM64 + if (coder->MethodID == k_ARM64) + { + UInt32 pc = 0; + if (coder->PropsSize == 4) + pc = GetUi32(propsData + coder->PropsOffset); + else if (coder->PropsSize != 0) + return SZ_ERROR_UNSUPPORTED; + z7_BranchConv_ARM64_Dec(outBuffer, outSize, pc); + continue; + } + #endif + + #if !defined(Z7_NO_METHODS_FILTERS) || defined(Z7_USE_FILTER_ARMT) { if (coder->PropsSize != 0) return SZ_ERROR_UNSUPPORTED; + #define CASE_BRA_CONV(isa) case k_ ## isa: Z7_BRANCH_CONV_DEC(isa)(outBuffer, outSize, 0); break; // pc = 0; switch (coder->MethodID) { + #if !defined(Z7_NO_METHODS_FILTERS) case k_BCJ: { - UInt32 state; - x86_Convert_Init(state); - x86_Convert(outBuffer, outSize, 0, &state, 0); + UInt32 state = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL; + z7_BranchConvSt_X86_Dec(outBuffer, outSize, 0, &state); // pc = 0 break; } CASE_BRA_CONV(PPC) CASE_BRA_CONV(IA64) CASE_BRA_CONV(SPARC) CASE_BRA_CONV(ARM) + #endif + #if !defined(Z7_NO_METHODS_FILTERS) || defined(Z7_USE_FILTER_ARMT) CASE_BRA_CONV(ARMT) + #endif default: return SZ_ERROR_UNSUPPORTED; } + continue; } - } - #endif + #endif + } // (c == 1) + #endif else return SZ_ERROR_UNSUPPORTED; } @@ -547,9 +604,9 @@ static SRes SzFolder_Decode2(const CSzFolder *folder, SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex, - ILookInStream *inStream, UInt64 startPos, + ILookInStreamPtr inStream, UInt64 startPos, Byte *outBuffer, size_t outSize, - ISzAlloc *allocMain) + ISzAllocPtr allocMain) { SRes res; CSzFolder folder; @@ -557,7 +614,7 @@ SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex, const Byte *data = p->CodersData + p->FoCodersOffsets[folderIndex]; sd.Data = data; - sd.Size = p->FoCodersOffsets[folderIndex + 1] - p->FoCodersOffsets[folderIndex]; + sd.Size = p->FoCodersOffsets[(size_t)folderIndex + 1] - p->FoCodersOffsets[folderIndex]; res = SzGetNextFolderItem(&folder, &sd); @@ -579,7 +636,7 @@ SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex, outBuffer, (SizeT)outSize, allocMain, tempBuf); for (i = 0; i < 3; i++) - IAlloc_Free(allocMain, tempBuf[i]); + ISzAlloc_Free(allocMain, tempBuf[i]); if (res == SZ_OK) if (SzBitWithVals_Check(&p->FolderCRCs, folderIndex)) diff --git a/3rdparty/lzma/C/7zFile.c b/3rdparty/lzma/C/7zFile.c index 041e5b15465..ba5daa133b9 100644 --- a/3rdparty/lzma/C/7zFile.c +++ b/3rdparty/lzma/C/7zFile.c @@ -1,5 +1,5 @@ /* 7zFile.c -- File IO -2009-11-24 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -7,9 +7,19 @@ #ifndef USE_WINDOWS_FILE -#ifndef UNDER_CE -#include <errno.h> -#endif + #include <errno.h> + + #ifndef USE_FOPEN + #include <stdio.h> + #include <fcntl.h> + #ifdef _WIN32 + #include <io.h> + typedef int ssize_t; + typedef int off_t; + #else + #include <unistd.h> + #endif + #endif #else @@ -23,30 +33,36 @@ And message can be "Network connection was lost" */ -#define kChunkSizeMax (1 << 22) - #endif +#define kChunkSizeMax (1 << 22) + void File_Construct(CSzFile *p) { #ifdef USE_WINDOWS_FILE p->handle = INVALID_HANDLE_VALUE; - #else + #elif defined(USE_FOPEN) p->file = NULL; + #else + p->fd = -1; #endif } #if !defined(UNDER_CE) || !defined(USE_WINDOWS_FILE) + static WRes File_Open(CSzFile *p, const char *name, int writeMode) { #ifdef USE_WINDOWS_FILE + p->handle = CreateFileA(name, writeMode ? GENERIC_WRITE : GENERIC_READ, FILE_SHARE_READ, NULL, writeMode ? CREATE_ALWAYS : OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL); return (p->handle != INVALID_HANDLE_VALUE) ? 0 : GetLastError(); - #else + + #elif defined(USE_FOPEN) + p->file = fopen(name, writeMode ? "wb+" : "rb"); return (p->file != 0) ? 0 : #ifdef UNDER_CE @@ -54,13 +70,34 @@ static WRes File_Open(CSzFile *p, const char *name, int writeMode) #else errno; #endif + + #else + + int flags = (writeMode ? (O_CREAT | O_EXCL | O_WRONLY) : O_RDONLY); + #ifdef O_BINARY + flags |= O_BINARY; + #endif + p->fd = open(name, flags, 0666); + return (p->fd != -1) ? 0 : errno; + #endif } WRes InFile_Open(CSzFile *p, const char *name) { return File_Open(p, name, 0); } -WRes OutFile_Open(CSzFile *p, const char *name) { return File_Open(p, name, 1); } + +WRes OutFile_Open(CSzFile *p, const char *name) +{ + #if defined(USE_WINDOWS_FILE) || defined(USE_FOPEN) + return File_Open(p, name, 1); + #else + p->fd = creat(name, 0666); + return (p->fd != -1) ? 0 : errno; + #endif +} + #endif + #ifdef USE_WINDOWS_FILE static WRes File_OpenW(CSzFile *p, const WCHAR *name, int writeMode) { @@ -78,74 +115,124 @@ WRes OutFile_OpenW(CSzFile *p, const WCHAR *name) { return File_OpenW(p, name, 1 WRes File_Close(CSzFile *p) { #ifdef USE_WINDOWS_FILE + if (p->handle != INVALID_HANDLE_VALUE) { if (!CloseHandle(p->handle)) return GetLastError(); p->handle = INVALID_HANDLE_VALUE; } - #else + + #elif defined(USE_FOPEN) + if (p->file != NULL) { int res = fclose(p->file); if (res != 0) + { + if (res == EOF) + return errno; return res; + } p->file = NULL; } + + #else + + if (p->fd != -1) + { + if (close(p->fd) != 0) + return errno; + p->fd = -1; + } + #endif + return 0; } + WRes File_Read(CSzFile *p, void *data, size_t *size) { size_t originalSize = *size; + *size = 0; if (originalSize == 0) return 0; #ifdef USE_WINDOWS_FILE - *size = 0; do { - DWORD curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : (DWORD)originalSize; + const DWORD curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : (DWORD)originalSize; DWORD processed = 0; - BOOL res = ReadFile(p->handle, data, curSize, &processed, NULL); + const BOOL res = ReadFile(p->handle, data, curSize, &processed, NULL); data = (void *)((Byte *)data + processed); originalSize -= processed; *size += processed; if (!res) return GetLastError(); + // debug : we can break here for partial reading mode + if (processed == 0) + break; + } + while (originalSize > 0); + + #elif defined(USE_FOPEN) + + do + { + const size_t curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : originalSize; + const size_t processed = fread(data, 1, curSize, p->file); + data = (void *)((Byte *)data + (size_t)processed); + originalSize -= processed; + *size += processed; + if (processed != curSize) + return ferror(p->file); + // debug : we can break here for partial reading mode if (processed == 0) break; } while (originalSize > 0); - return 0; #else - - *size = fread(data, 1, originalSize, p->file); - if (*size == originalSize) - return 0; - return ferror(p->file); - + + do + { + const size_t curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : originalSize; + const ssize_t processed = read(p->fd, data, curSize); + if (processed == -1) + return errno; + if (processed == 0) + break; + data = (void *)((Byte *)data + (size_t)processed); + originalSize -= (size_t)processed; + *size += (size_t)processed; + // debug : we can break here for partial reading mode + // break; + } + while (originalSize > 0); + #endif + + return 0; } + WRes File_Write(CSzFile *p, const void *data, size_t *size) { size_t originalSize = *size; + *size = 0; if (originalSize == 0) return 0; #ifdef USE_WINDOWS_FILE - *size = 0; do { - DWORD curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : (DWORD)originalSize; + const DWORD curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : (DWORD)originalSize; DWORD processed = 0; - BOOL res = WriteFile(p->handle, data, curSize, &processed, NULL); - data = (void *)((Byte *)data + processed); + const BOOL res = WriteFile(p->handle, data, curSize, &processed, NULL); + data = (const void *)((const Byte *)data + processed); originalSize -= processed; *size += processed; if (!res) @@ -154,61 +241,106 @@ WRes File_Write(CSzFile *p, const void *data, size_t *size) break; } while (originalSize > 0); - return 0; + + #elif defined(USE_FOPEN) + + do + { + const size_t curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : originalSize; + const size_t processed = fwrite(data, 1, curSize, p->file); + data = (void *)((Byte *)data + (size_t)processed); + originalSize -= processed; + *size += processed; + if (processed != curSize) + return ferror(p->file); + if (processed == 0) + break; + } + while (originalSize > 0); #else - *size = fwrite(data, 1, originalSize, p->file); - if (*size == originalSize) - return 0; - return ferror(p->file); - + do + { + const size_t curSize = (originalSize > kChunkSizeMax) ? kChunkSizeMax : originalSize; + const ssize_t processed = write(p->fd, data, curSize); + if (processed == -1) + return errno; + if (processed == 0) + break; + data = (const void *)((const Byte *)data + (size_t)processed); + originalSize -= (size_t)processed; + *size += (size_t)processed; + } + while (originalSize > 0); + #endif + + return 0; } + WRes File_Seek(CSzFile *p, Int64 *pos, ESzSeek origin) { #ifdef USE_WINDOWS_FILE - LARGE_INTEGER value; DWORD moveMethod; - value.LowPart = (DWORD)*pos; - value.HighPart = (LONG)((UInt64)*pos >> 16 >> 16); /* for case when UInt64 is 32-bit only */ - switch (origin) + UInt32 low = (UInt32)*pos; + LONG high = (LONG)((UInt64)*pos >> 16 >> 16); /* for case when UInt64 is 32-bit only */ + // (int) to eliminate clang warning + switch ((int)origin) { case SZ_SEEK_SET: moveMethod = FILE_BEGIN; break; case SZ_SEEK_CUR: moveMethod = FILE_CURRENT; break; case SZ_SEEK_END: moveMethod = FILE_END; break; default: return ERROR_INVALID_PARAMETER; } - value.LowPart = SetFilePointer(p->handle, value.LowPart, &value.HighPart, moveMethod); - if (value.LowPart == 0xFFFFFFFF) + low = SetFilePointer(p->handle, (LONG)low, &high, moveMethod); + if (low == (UInt32)0xFFFFFFFF) { WRes res = GetLastError(); if (res != NO_ERROR) return res; } - *pos = ((Int64)value.HighPart << 32) | value.LowPart; + *pos = ((Int64)high << 32) | low; return 0; #else - int moveMethod; - int res; - switch (origin) + int moveMethod; // = origin; + + switch ((int)origin) { case SZ_SEEK_SET: moveMethod = SEEK_SET; break; case SZ_SEEK_CUR: moveMethod = SEEK_CUR; break; case SZ_SEEK_END: moveMethod = SEEK_END; break; - default: return 1; + default: return EINVAL; } - res = fseek(p->file, (long)*pos, moveMethod); - *pos = ftell(p->file); - return res; - #endif + #if defined(USE_FOPEN) + { + int res = fseek(p->file, (long)*pos, moveMethod); + if (res == -1) + return errno; + *pos = ftell(p->file); + if (*pos == -1) + return errno; + return 0; + } + #else + { + off_t res = lseek(p->fd, (off_t)*pos, moveMethod); + if (res == -1) + return errno; + *pos = res; + return 0; + } + + #endif // USE_FOPEN + #endif // USE_WINDOWS_FILE } + WRes File_GetLength(CSzFile *p, UInt64 *length) { #ifdef USE_WINDOWS_FILE @@ -224,13 +356,31 @@ WRes File_GetLength(CSzFile *p, UInt64 *length) *length = (((UInt64)sizeHigh) << 32) + sizeLow; return 0; - #else + #elif defined(USE_FOPEN) long pos = ftell(p->file); int res = fseek(p->file, 0, SEEK_END); *length = ftell(p->file); fseek(p->file, pos, SEEK_SET); return res; + + #else + + off_t pos; + *length = 0; + pos = lseek(p->fd, 0, SEEK_CUR); + if (pos != -1) + { + const off_t len2 = lseek(p->fd, 0, SEEK_END); + const off_t res2 = lseek(p->fd, pos, SEEK_SET); + if (len2 != -1) + { + *length = (UInt64)len2; + if (res2 != -1) + return 0; + } + } + return errno; #endif } @@ -238,49 +388,56 @@ WRes File_GetLength(CSzFile *p, UInt64 *length) /* ---------- FileSeqInStream ---------- */ -static SRes FileSeqInStream_Read(void *pp, void *buf, size_t *size) +static SRes FileSeqInStream_Read(ISeqInStreamPtr pp, void *buf, size_t *size) { - CFileSeqInStream *p = (CFileSeqInStream *)pp; - return File_Read(&p->file, buf, size) == 0 ? SZ_OK : SZ_ERROR_READ; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileSeqInStream) + const WRes wres = File_Read(&p->file, buf, size); + p->wres = wres; + return (wres == 0) ? SZ_OK : SZ_ERROR_READ; } void FileSeqInStream_CreateVTable(CFileSeqInStream *p) { - p->s.Read = FileSeqInStream_Read; + p->vt.Read = FileSeqInStream_Read; } /* ---------- FileInStream ---------- */ -static SRes FileInStream_Read(void *pp, void *buf, size_t *size) +static SRes FileInStream_Read(ISeekInStreamPtr pp, void *buf, size_t *size) { - CFileInStream *p = (CFileInStream *)pp; - return (File_Read(&p->file, buf, size) == 0) ? SZ_OK : SZ_ERROR_READ; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileInStream) + const WRes wres = File_Read(&p->file, buf, size); + p->wres = wres; + return (wres == 0) ? SZ_OK : SZ_ERROR_READ; } -static SRes FileInStream_Seek(void *pp, Int64 *pos, ESzSeek origin) +static SRes FileInStream_Seek(ISeekInStreamPtr pp, Int64 *pos, ESzSeek origin) { - CFileInStream *p = (CFileInStream *)pp; - return File_Seek(&p->file, pos, origin); + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileInStream) + const WRes wres = File_Seek(&p->file, pos, origin); + p->wres = wres; + return (wres == 0) ? SZ_OK : SZ_ERROR_READ; } void FileInStream_CreateVTable(CFileInStream *p) { - p->s.Read = FileInStream_Read; - p->s.Seek = FileInStream_Seek; + p->vt.Read = FileInStream_Read; + p->vt.Seek = FileInStream_Seek; } /* ---------- FileOutStream ---------- */ -static size_t FileOutStream_Write(void *pp, const void *data, size_t size) +static size_t FileOutStream_Write(ISeqOutStreamPtr pp, const void *data, size_t size) { - CFileOutStream *p = (CFileOutStream *)pp; - File_Write(&p->file, data, &size); + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileOutStream) + const WRes wres = File_Write(&p->file, data, &size); + p->wres = wres; return size; } void FileOutStream_CreateVTable(CFileOutStream *p) { - p->s.Write = FileOutStream_Write; + p->vt.Write = FileOutStream_Write; } diff --git a/3rdparty/lzma/C/7zFile.h b/3rdparty/lzma/C/7zFile.h index 658987ed51a..f5069cd9ee2 100644 --- a/3rdparty/lzma/C/7zFile.h +++ b/3rdparty/lzma/C/7zFile.h @@ -1,17 +1,21 @@ /* 7zFile.h -- File IO -2013-01-18 : Igor Pavlov : Public domain */ +2023-03-05 : Igor Pavlov : Public domain */ -#ifndef __7Z_FILE_H -#define __7Z_FILE_H +#ifndef ZIP7_INC_FILE_H +#define ZIP7_INC_FILE_H #ifdef _WIN32 #define USE_WINDOWS_FILE +// #include <windows.h> #endif #ifdef USE_WINDOWS_FILE -#include <windows.h> +#include "7zWindows.h" + #else -#include <stdio.h> +// note: USE_FOPEN mode is limited to 32-bit file size +// #define USE_FOPEN +// #include <stdio.h> #endif #include "7zTypes.h" @@ -24,8 +28,10 @@ typedef struct { #ifdef USE_WINDOWS_FILE HANDLE handle; - #else + #elif defined(USE_FOPEN) FILE *file; + #else + int fd; #endif } CSzFile; @@ -54,8 +60,9 @@ WRes File_GetLength(CSzFile *p, UInt64 *length); typedef struct { - ISeqInStream s; + ISeqInStream vt; CSzFile file; + WRes wres; } CFileSeqInStream; void FileSeqInStream_CreateVTable(CFileSeqInStream *p); @@ -63,8 +70,9 @@ void FileSeqInStream_CreateVTable(CFileSeqInStream *p); typedef struct { - ISeekInStream s; + ISeekInStream vt; CSzFile file; + WRes wres; } CFileInStream; void FileInStream_CreateVTable(CFileInStream *p); @@ -72,8 +80,9 @@ void FileInStream_CreateVTable(CFileInStream *p); typedef struct { - ISeqOutStream s; + ISeqOutStream vt; CSzFile file; + WRes wres; } CFileOutStream; void FileOutStream_CreateVTable(CFileOutStream *p); diff --git a/3rdparty/lzma/C/7zStream.c b/3rdparty/lzma/C/7zStream.c index 88f9c42b1dd..74e75b65aa2 100644 --- a/3rdparty/lzma/C/7zStream.c +++ b/3rdparty/lzma/C/7zStream.c @@ -1,5 +1,5 @@ /* 7zStream.c -- 7z Stream functions -2013-11-12 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -7,12 +7,33 @@ #include "7zTypes.h" -SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorType) + +SRes SeqInStream_ReadMax(ISeqInStreamPtr stream, void *buf, size_t *processedSize) +{ + size_t size = *processedSize; + *processedSize = 0; + while (size != 0) + { + size_t cur = size; + const SRes res = ISeqInStream_Read(stream, buf, &cur); + *processedSize += cur; + buf = (void *)((Byte *)buf + cur); + size -= cur; + if (res != SZ_OK) + return res; + if (cur == 0) + return SZ_OK; + } + return SZ_OK; +} + +/* +SRes SeqInStream_Read2(ISeqInStreamPtr stream, void *buf, size_t size, SRes errorType) { while (size != 0) { size_t processed = size; - RINOK(stream->Read(stream, buf, &processed)); + RINOK(ISeqInStream_Read(stream, buf, &processed)) if (processed == 0) return errorType; buf = (void *)((Byte *)buf + processed); @@ -21,40 +42,44 @@ SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorT return SZ_OK; } -SRes SeqInStream_Read(ISeqInStream *stream, void *buf, size_t size) +SRes SeqInStream_Read(ISeqInStreamPtr stream, void *buf, size_t size) { return SeqInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF); } +*/ + -SRes SeqInStream_ReadByte(ISeqInStream *stream, Byte *buf) +SRes SeqInStream_ReadByte(ISeqInStreamPtr stream, Byte *buf) { size_t processed = 1; - RINOK(stream->Read(stream, buf, &processed)); + RINOK(ISeqInStream_Read(stream, buf, &processed)) return (processed == 1) ? SZ_OK : SZ_ERROR_INPUT_EOF; } -SRes LookInStream_SeekTo(ILookInStream *stream, UInt64 offset) + + +SRes LookInStream_SeekTo(ILookInStreamPtr stream, UInt64 offset) { - Int64 t = offset; - return stream->Seek(stream, &t, SZ_SEEK_SET); + Int64 t = (Int64)offset; + return ILookInStream_Seek(stream, &t, SZ_SEEK_SET); } -SRes LookInStream_LookRead(ILookInStream *stream, void *buf, size_t *size) +SRes LookInStream_LookRead(ILookInStreamPtr stream, void *buf, size_t *size) { const void *lookBuf; if (*size == 0) return SZ_OK; - RINOK(stream->Look(stream, &lookBuf, size)); + RINOK(ILookInStream_Look(stream, &lookBuf, size)) memcpy(buf, lookBuf, *size); - return stream->Skip(stream, *size); + return ILookInStream_Skip(stream, *size); } -SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes errorType) +SRes LookInStream_Read2(ILookInStreamPtr stream, void *buf, size_t size, SRes errorType) { while (size != 0) { size_t processed = size; - RINOK(stream->Read(stream, buf, &processed)); + RINOK(ILookInStream_Read(stream, buf, &processed)) if (processed == 0) return errorType; buf = (void *)((Byte *)buf + processed); @@ -63,61 +88,67 @@ SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes erro return SZ_OK; } -SRes LookInStream_Read(ILookInStream *stream, void *buf, size_t size) +SRes LookInStream_Read(ILookInStreamPtr stream, void *buf, size_t size) { return LookInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF); } -static SRes LookToRead_Look_Lookahead(void *pp, const void **buf, size_t *size) + + +#define GET_LookToRead2 Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CLookToRead2) + +static SRes LookToRead2_Look_Lookahead(ILookInStreamPtr pp, const void **buf, size_t *size) { SRes res = SZ_OK; - CLookToRead *p = (CLookToRead *)pp; + GET_LookToRead2 size_t size2 = p->size - p->pos; - if (size2 == 0 && *size > 0) + if (size2 == 0 && *size != 0) { p->pos = 0; - size2 = LookToRead_BUF_SIZE; - res = p->realStream->Read(p->realStream, p->buf, &size2); + p->size = 0; + size2 = p->bufSize; + res = ISeekInStream_Read(p->realStream, p->buf, &size2); p->size = size2; } - if (size2 < *size) + if (*size > size2) *size = size2; *buf = p->buf + p->pos; return res; } -static SRes LookToRead_Look_Exact(void *pp, const void **buf, size_t *size) +static SRes LookToRead2_Look_Exact(ILookInStreamPtr pp, const void **buf, size_t *size) { SRes res = SZ_OK; - CLookToRead *p = (CLookToRead *)pp; + GET_LookToRead2 size_t size2 = p->size - p->pos; - if (size2 == 0 && *size > 0) + if (size2 == 0 && *size != 0) { p->pos = 0; - if (*size > LookToRead_BUF_SIZE) - *size = LookToRead_BUF_SIZE; - res = p->realStream->Read(p->realStream, p->buf, size); + p->size = 0; + if (*size > p->bufSize) + *size = p->bufSize; + res = ISeekInStream_Read(p->realStream, p->buf, size); size2 = p->size = *size; } - if (size2 < *size) + if (*size > size2) *size = size2; *buf = p->buf + p->pos; return res; } -static SRes LookToRead_Skip(void *pp, size_t offset) +static SRes LookToRead2_Skip(ILookInStreamPtr pp, size_t offset) { - CLookToRead *p = (CLookToRead *)pp; + GET_LookToRead2 p->pos += offset; return SZ_OK; } -static SRes LookToRead_Read(void *pp, void *buf, size_t *size) +static SRes LookToRead2_Read(ILookInStreamPtr pp, void *buf, size_t *size) { - CLookToRead *p = (CLookToRead *)pp; + GET_LookToRead2 size_t rem = p->size - p->pos; if (rem == 0) - return p->realStream->Read(p->realStream, buf, size); + return ISeekInStream_Read(p->realStream, buf, size); if (rem > *size) rem = *size; memcpy(buf, p->buf + p->pos, rem); @@ -126,46 +157,43 @@ static SRes LookToRead_Read(void *pp, void *buf, size_t *size) return SZ_OK; } -static SRes LookToRead_Seek(void *pp, Int64 *pos, ESzSeek origin) +static SRes LookToRead2_Seek(ILookInStreamPtr pp, Int64 *pos, ESzSeek origin) { - CLookToRead *p = (CLookToRead *)pp; + GET_LookToRead2 p->pos = p->size = 0; - return p->realStream->Seek(p->realStream, pos, origin); + return ISeekInStream_Seek(p->realStream, pos, origin); } -void LookToRead_CreateVTable(CLookToRead *p, int lookahead) +void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead) { - p->s.Look = lookahead ? - LookToRead_Look_Lookahead : - LookToRead_Look_Exact; - p->s.Skip = LookToRead_Skip; - p->s.Read = LookToRead_Read; - p->s.Seek = LookToRead_Seek; + p->vt.Look = lookahead ? + LookToRead2_Look_Lookahead : + LookToRead2_Look_Exact; + p->vt.Skip = LookToRead2_Skip; + p->vt.Read = LookToRead2_Read; + p->vt.Seek = LookToRead2_Seek; } -void LookToRead_Init(CLookToRead *p) -{ - p->pos = p->size = 0; -} -static SRes SecToLook_Read(void *pp, void *buf, size_t *size) + +static SRes SecToLook_Read(ISeqInStreamPtr pp, void *buf, size_t *size) { - CSecToLook *p = (CSecToLook *)pp; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSecToLook) return LookInStream_LookRead(p->realStream, buf, size); } void SecToLook_CreateVTable(CSecToLook *p) { - p->s.Read = SecToLook_Read; + p->vt.Read = SecToLook_Read; } -static SRes SecToRead_Read(void *pp, void *buf, size_t *size) +static SRes SecToRead_Read(ISeqInStreamPtr pp, void *buf, size_t *size) { - CSecToRead *p = (CSecToRead *)pp; - return p->realStream->Read(p->realStream, buf, size); + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSecToRead) + return ILookInStream_Read(p->realStream, buf, size); } void SecToRead_CreateVTable(CSecToRead *p) { - p->s.Read = SecToRead_Read; + p->vt.Read = SecToRead_Read; } diff --git a/3rdparty/lzma/C/7zTypes.h b/3rdparty/lzma/C/7zTypes.h index 778413ef424..1fcb2473e18 100644 --- a/3rdparty/lzma/C/7zTypes.h +++ b/3rdparty/lzma/C/7zTypes.h @@ -1,11 +1,13 @@ /* 7zTypes.h -- Basic types -2013-11-12 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __7Z_TYPES_H -#define __7Z_TYPES_H +#ifndef ZIP7_7Z_TYPES_H +#define ZIP7_7Z_TYPES_H #ifdef _WIN32 /* #include <windows.h> */ +#else +#include <errno.h> #endif #include <stddef.h> @@ -42,22 +44,135 @@ EXTERN_C_BEGIN typedef int SRes; + +#ifdef _MSC_VER + #if _MSC_VER > 1200 + #define MY_ALIGN(n) __declspec(align(n)) + #else + #define MY_ALIGN(n) + #endif +#else + /* + // C11/C++11: + #include <stdalign.h> + #define MY_ALIGN(n) alignas(n) + */ + #define MY_ALIGN(n) __attribute__ ((aligned(n))) +#endif + + #ifdef _WIN32 + /* typedef DWORD WRes; */ typedef unsigned WRes; -#else +#define MY_SRes_HRESULT_FROM_WRes(x) HRESULT_FROM_WIN32(x) + +// #define MY_HRES_ERROR_INTERNAL_ERROR MY_SRes_HRESULT_FROM_WRes(ERROR_INTERNAL_ERROR) + +#else // _WIN32 + +// #define ENV_HAVE_LSTAT typedef int WRes; + +// (FACILITY_ERRNO = 0x800) is 7zip's FACILITY constant to represent (errno) errors in HRESULT +#define MY_FACILITY_ERRNO 0x800 +#define MY_FACILITY_WIN32 7 +#define MY_FACILITY_WRes MY_FACILITY_ERRNO + +#define MY_HRESULT_FROM_errno_CONST_ERROR(x) ((HRESULT)( \ + ( (HRESULT)(x) & 0x0000FFFF) \ + | (MY_FACILITY_WRes << 16) \ + | (HRESULT)0x80000000 )) + +#define MY_SRes_HRESULT_FROM_WRes(x) \ + ((HRESULT)(x) <= 0 ? ((HRESULT)(x)) : MY_HRESULT_FROM_errno_CONST_ERROR(x)) + +// we call macro HRESULT_FROM_WIN32 for system errors (WRes) that are (errno) +#define HRESULT_FROM_WIN32(x) MY_SRes_HRESULT_FROM_WRes(x) + +/* +#define ERROR_FILE_NOT_FOUND 2L +#define ERROR_ACCESS_DENIED 5L +#define ERROR_NO_MORE_FILES 18L +#define ERROR_LOCK_VIOLATION 33L +#define ERROR_FILE_EXISTS 80L +#define ERROR_DISK_FULL 112L +#define ERROR_NEGATIVE_SEEK 131L +#define ERROR_ALREADY_EXISTS 183L +#define ERROR_DIRECTORY 267L +#define ERROR_TOO_MANY_POSTS 298L + +#define ERROR_INTERNAL_ERROR 1359L +#define ERROR_INVALID_REPARSE_DATA 4392L +#define ERROR_REPARSE_TAG_INVALID 4393L +#define ERROR_REPARSE_TAG_MISMATCH 4394L +*/ + +// we use errno equivalents for some WIN32 errors: + +#define ERROR_INVALID_PARAMETER EINVAL +#define ERROR_INVALID_FUNCTION EINVAL +#define ERROR_ALREADY_EXISTS EEXIST +#define ERROR_FILE_EXISTS EEXIST +#define ERROR_PATH_NOT_FOUND ENOENT +#define ERROR_FILE_NOT_FOUND ENOENT +#define ERROR_DISK_FULL ENOSPC +// #define ERROR_INVALID_HANDLE EBADF + +// we use FACILITY_WIN32 for errors that has no errno equivalent +// Too many posts were made to a semaphore. +#define ERROR_TOO_MANY_POSTS ((HRESULT)0x8007012AL) +#define ERROR_INVALID_REPARSE_DATA ((HRESULT)0x80071128L) +#define ERROR_REPARSE_TAG_INVALID ((HRESULT)0x80071129L) + +// if (MY_FACILITY_WRes != FACILITY_WIN32), +// we use FACILITY_WIN32 for COM errors: +#define E_OUTOFMEMORY ((HRESULT)0x8007000EL) +#define E_INVALIDARG ((HRESULT)0x80070057L) +#define MY_E_ERROR_NEGATIVE_SEEK ((HRESULT)0x80070083L) + +/* +// we can use FACILITY_ERRNO for some COM errors, that have errno equivalents: +#define E_OUTOFMEMORY MY_HRESULT_FROM_errno_CONST_ERROR(ENOMEM) +#define E_INVALIDARG MY_HRESULT_FROM_errno_CONST_ERROR(EINVAL) +#define MY_E_ERROR_NEGATIVE_SEEK MY_HRESULT_FROM_errno_CONST_ERROR(EINVAL) +*/ + +#define TEXT(quote) quote + +#define FILE_ATTRIBUTE_READONLY 0x0001 +#define FILE_ATTRIBUTE_HIDDEN 0x0002 +#define FILE_ATTRIBUTE_SYSTEM 0x0004 +#define FILE_ATTRIBUTE_DIRECTORY 0x0010 +#define FILE_ATTRIBUTE_ARCHIVE 0x0020 +#define FILE_ATTRIBUTE_DEVICE 0x0040 +#define FILE_ATTRIBUTE_NORMAL 0x0080 +#define FILE_ATTRIBUTE_TEMPORARY 0x0100 +#define FILE_ATTRIBUTE_SPARSE_FILE 0x0200 +#define FILE_ATTRIBUTE_REPARSE_POINT 0x0400 +#define FILE_ATTRIBUTE_COMPRESSED 0x0800 +#define FILE_ATTRIBUTE_OFFLINE 0x1000 +#define FILE_ATTRIBUTE_NOT_CONTENT_INDEXED 0x2000 +#define FILE_ATTRIBUTE_ENCRYPTED 0x4000 + +#define FILE_ATTRIBUTE_UNIX_EXTENSION 0x8000 /* trick for Unix */ + #endif + #ifndef RINOK -#define RINOK(x) { int __result__ = (x); if (__result__ != 0) return __result__; } +#define RINOK(x) { const int _result_ = (x); if (_result_ != 0) return _result_; } +#endif + +#ifndef RINOK_WRes +#define RINOK_WRes(x) { const WRes _result_ = (x); if (_result_ != 0) return _result_; } #endif typedef unsigned char Byte; typedef short Int16; typedef unsigned short UInt16; -#ifdef _LZMA_UINT32_IS_ULONG +#ifdef Z7_DECL_Int32_AS_long typedef long Int32; typedef unsigned long UInt32; #else @@ -65,95 +180,186 @@ typedef int Int32; typedef unsigned int UInt32; #endif -#ifdef _SZ_NO_INT_64 -/* define _SZ_NO_INT_64, if your compiler doesn't support 64-bit integers. - NOTES: Some code will work incorrectly in that case! */ +#ifndef _WIN32 + +typedef int INT; +typedef Int32 INT32; +typedef unsigned int UINT; +typedef UInt32 UINT32; +typedef INT32 LONG; // LONG, ULONG and DWORD must be 32-bit for _WIN32 compatibility +typedef UINT32 ULONG; + +#undef DWORD +typedef UINT32 DWORD; + +#define VOID void + +#define HRESULT LONG + +typedef void *LPVOID; +// typedef void VOID; +// typedef ULONG_PTR DWORD_PTR, *PDWORD_PTR; +// gcc / clang on Unix : sizeof(long==sizeof(void*) in 32 or 64 bits) +typedef long INT_PTR; +typedef unsigned long UINT_PTR; +typedef long LONG_PTR; +typedef unsigned long DWORD_PTR; + +typedef size_t SIZE_T; + +#endif // _WIN32 + + +#define MY_HRES_ERROR_INTERNAL_ERROR ((HRESULT)0x8007054FL) + + +#ifdef Z7_DECL_Int64_AS_long typedef long Int64; typedef unsigned long UInt64; #else -#if defined(_MSC_VER) || defined(__BORLANDC__) +#if (defined(_MSC_VER) || defined(__BORLANDC__)) && !defined(__clang__) typedef __int64 Int64; typedef unsigned __int64 UInt64; -#define UINT64_CONST(n) n +#else +#if defined(__clang__) || defined(__GNUC__) +#include <stdint.h> +typedef int64_t Int64; +typedef uint64_t UInt64; #else typedef long long int Int64; typedef unsigned long long int UInt64; -#define UINT64_CONST(n) n ## ULL +// #define UINT64_CONST(n) n ## ULL +#endif #endif #endif -#ifdef _LZMA_NO_SYSTEM_SIZE_T -typedef UInt32 SizeT; +#define UINT64_CONST(n) n + + +#ifdef Z7_DECL_SizeT_AS_unsigned_int +typedef unsigned int SizeT; #else typedef size_t SizeT; #endif -typedef int Bool; +/* +#if (defined(_MSC_VER) && _MSC_VER <= 1200) +typedef size_t MY_uintptr_t; +#else +#include <stdint.h> +typedef uintptr_t MY_uintptr_t; +#endif +*/ + +typedef int BoolInt; +/* typedef BoolInt Bool; */ #define True 1 #define False 0 #ifdef _WIN32 -#define MY_STD_CALL __stdcall +#define Z7_STDCALL __stdcall #else -#define MY_STD_CALL +#define Z7_STDCALL #endif #ifdef _MSC_VER #if _MSC_VER >= 1300 -#define MY_NO_INLINE __declspec(noinline) +#define Z7_NO_INLINE __declspec(noinline) #else -#define MY_NO_INLINE +#define Z7_NO_INLINE #endif -#define MY_CDECL __cdecl -#define MY_FAST_CALL __fastcall +#define Z7_FORCE_INLINE __forceinline +#define Z7_CDECL __cdecl +#define Z7_FASTCALL __fastcall + +#else // _MSC_VER + +#if (defined(__GNUC__) && (__GNUC__ >= 4)) \ + || (defined(__clang__) && (__clang_major__ >= 4)) \ + || defined(__INTEL_COMPILER) \ + || defined(__xlC__) +#define Z7_NO_INLINE __attribute__((noinline)) +#define Z7_FORCE_INLINE __attribute__((always_inline)) inline #else +#define Z7_NO_INLINE +#define Z7_FORCE_INLINE +#endif -#define MY_NO_INLINE -#define MY_CDECL -#define MY_FAST_CALL +#define Z7_CDECL +#if defined(_M_IX86) \ + || defined(__i386__) +// #define Z7_FASTCALL __attribute__((fastcall)) +// #define Z7_FASTCALL __attribute__((cdecl)) +#define Z7_FASTCALL +#elif defined(MY_CPU_AMD64) +// #define Z7_FASTCALL __attribute__((ms_abi)) +#define Z7_FASTCALL +#else +#define Z7_FASTCALL #endif +#endif // _MSC_VER + /* The following interfaces use first parameter as pointer to structure */ -typedef struct +// #define Z7_C_IFACE_CONST_QUAL +#define Z7_C_IFACE_CONST_QUAL const + +#define Z7_C_IFACE_DECL(a) \ + struct a ## _; \ + typedef Z7_C_IFACE_CONST_QUAL struct a ## _ * a ## Ptr; \ + typedef struct a ## _ a; \ + struct a ## _ + + +Z7_C_IFACE_DECL (IByteIn) { - Byte (*Read)(void *p); /* reads one byte, returns 0 in case of EOF or error */ -} IByteIn; + Byte (*Read)(IByteInPtr p); /* reads one byte, returns 0 in case of EOF or error */ +}; +#define IByteIn_Read(p) (p)->Read(p) -typedef struct + +Z7_C_IFACE_DECL (IByteOut) { - void (*Write)(void *p, Byte b); -} IByteOut; + void (*Write)(IByteOutPtr p, Byte b); +}; +#define IByteOut_Write(p, b) (p)->Write(p, b) -typedef struct + +Z7_C_IFACE_DECL (ISeqInStream) { - SRes (*Read)(void *p, void *buf, size_t *size); + SRes (*Read)(ISeqInStreamPtr p, void *buf, size_t *size); /* if (input(*size) != 0 && output(*size) == 0) means end_of_stream. (output(*size) < input(*size)) is allowed */ -} ISeqInStream; +}; +#define ISeqInStream_Read(p, buf, size) (p)->Read(p, buf, size) +/* try to read as much as avail in stream and limited by (*processedSize) */ +SRes SeqInStream_ReadMax(ISeqInStreamPtr stream, void *buf, size_t *processedSize); /* it can return SZ_ERROR_INPUT_EOF */ -SRes SeqInStream_Read(ISeqInStream *stream, void *buf, size_t size); -SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorType); -SRes SeqInStream_ReadByte(ISeqInStream *stream, Byte *buf); +// SRes SeqInStream_Read(ISeqInStreamPtr stream, void *buf, size_t size); +// SRes SeqInStream_Read2(ISeqInStreamPtr stream, void *buf, size_t size, SRes errorType); +SRes SeqInStream_ReadByte(ISeqInStreamPtr stream, Byte *buf); -typedef struct + +Z7_C_IFACE_DECL (ISeqOutStream) { - size_t (*Write)(void *p, const void *buf, size_t size); + size_t (*Write)(ISeqOutStreamPtr p, const void *buf, size_t size); /* Returns: result - the number of actually written bytes. (result < size) means error */ -} ISeqOutStream; +}; +#define ISeqOutStream_Write(p, buf, size) (p)->Write(p, buf, size) typedef enum { @@ -162,78 +368,197 @@ typedef enum SZ_SEEK_END = 2 } ESzSeek; -typedef struct + +Z7_C_IFACE_DECL (ISeekInStream) { - SRes (*Read)(void *p, void *buf, size_t *size); /* same as ISeqInStream::Read */ - SRes (*Seek)(void *p, Int64 *pos, ESzSeek origin); -} ISeekInStream; + SRes (*Read)(ISeekInStreamPtr p, void *buf, size_t *size); /* same as ISeqInStream::Read */ + SRes (*Seek)(ISeekInStreamPtr p, Int64 *pos, ESzSeek origin); +}; +#define ISeekInStream_Read(p, buf, size) (p)->Read(p, buf, size) +#define ISeekInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin) -typedef struct + +Z7_C_IFACE_DECL (ILookInStream) { - SRes (*Look)(void *p, const void **buf, size_t *size); + SRes (*Look)(ILookInStreamPtr p, const void **buf, size_t *size); /* if (input(*size) != 0 && output(*size) == 0) means end_of_stream. (output(*size) > input(*size)) is not allowed (output(*size) < input(*size)) is allowed */ - SRes (*Skip)(void *p, size_t offset); + SRes (*Skip)(ILookInStreamPtr p, size_t offset); /* offset must be <= output(*size) of Look */ - - SRes (*Read)(void *p, void *buf, size_t *size); + SRes (*Read)(ILookInStreamPtr p, void *buf, size_t *size); /* reads directly (without buffer). It's same as ISeqInStream::Read */ - SRes (*Seek)(void *p, Int64 *pos, ESzSeek origin); -} ILookInStream; + SRes (*Seek)(ILookInStreamPtr p, Int64 *pos, ESzSeek origin); +}; -SRes LookInStream_LookRead(ILookInStream *stream, void *buf, size_t *size); -SRes LookInStream_SeekTo(ILookInStream *stream, UInt64 offset); +#define ILookInStream_Look(p, buf, size) (p)->Look(p, buf, size) +#define ILookInStream_Skip(p, offset) (p)->Skip(p, offset) +#define ILookInStream_Read(p, buf, size) (p)->Read(p, buf, size) +#define ILookInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin) + + +SRes LookInStream_LookRead(ILookInStreamPtr stream, void *buf, size_t *size); +SRes LookInStream_SeekTo(ILookInStreamPtr stream, UInt64 offset); /* reads via ILookInStream::Read */ -SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes errorType); -SRes LookInStream_Read(ILookInStream *stream, void *buf, size_t size); +SRes LookInStream_Read2(ILookInStreamPtr stream, void *buf, size_t size, SRes errorType); +SRes LookInStream_Read(ILookInStreamPtr stream, void *buf, size_t size); -#define LookToRead_BUF_SIZE (1 << 14) typedef struct { - ILookInStream s; - ISeekInStream *realStream; + ILookInStream vt; + ISeekInStreamPtr realStream; + size_t pos; - size_t size; - Byte buf[LookToRead_BUF_SIZE]; -} CLookToRead; + size_t size; /* it's data size */ + + /* the following variables must be set outside */ + Byte *buf; + size_t bufSize; +} CLookToRead2; + +void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead); + +#define LookToRead2_INIT(p) { (p)->pos = (p)->size = 0; } -void LookToRead_CreateVTable(CLookToRead *p, int lookahead); -void LookToRead_Init(CLookToRead *p); typedef struct { - ISeqInStream s; - ILookInStream *realStream; + ISeqInStream vt; + ILookInStreamPtr realStream; } CSecToLook; void SecToLook_CreateVTable(CSecToLook *p); + + typedef struct { - ISeqInStream s; - ILookInStream *realStream; + ISeqInStream vt; + ILookInStreamPtr realStream; } CSecToRead; void SecToRead_CreateVTable(CSecToRead *p); -typedef struct + +Z7_C_IFACE_DECL (ICompressProgress) { - SRes (*Progress)(void *p, UInt64 inSize, UInt64 outSize); + SRes (*Progress)(ICompressProgressPtr p, UInt64 inSize, UInt64 outSize); /* Returns: result. (result != SZ_OK) means break. Value (UInt64)(Int64)-1 for size means unknown value. */ -} ICompressProgress; +}; -typedef struct +#define ICompressProgress_Progress(p, inSize, outSize) (p)->Progress(p, inSize, outSize) + + + +typedef struct ISzAlloc ISzAlloc; +typedef const ISzAlloc * ISzAllocPtr; + +struct ISzAlloc { - void *(*Alloc)(void *p, size_t size); - void (*Free)(void *p, void *address); /* address can be 0 */ -} ISzAlloc; + void *(*Alloc)(ISzAllocPtr p, size_t size); + void (*Free)(ISzAllocPtr p, void *address); /* address can be 0 */ +}; + +#define ISzAlloc_Alloc(p, size) (p)->Alloc(p, size) +#define ISzAlloc_Free(p, a) (p)->Free(p, a) + +/* deprecated */ +#define IAlloc_Alloc(p, size) ISzAlloc_Alloc(p, size) +#define IAlloc_Free(p, a) ISzAlloc_Free(p, a) + + + + + +#ifndef MY_offsetof + #ifdef offsetof + #define MY_offsetof(type, m) offsetof(type, m) + /* + #define MY_offsetof(type, m) FIELD_OFFSET(type, m) + */ + #else + #define MY_offsetof(type, m) ((size_t)&(((type *)0)->m)) + #endif +#endif + + + +#ifndef Z7_container_of + +/* +#define Z7_container_of(ptr, type, m) container_of(ptr, type, m) +#define Z7_container_of(ptr, type, m) CONTAINING_RECORD(ptr, type, m) +#define Z7_container_of(ptr, type, m) ((type *)((char *)(ptr) - offsetof(type, m))) +#define Z7_container_of(ptr, type, m) (&((type *)0)->m == (ptr), ((type *)(((char *)(ptr)) - MY_offsetof(type, m)))) +*/ + +/* + GCC shows warning: "perhaps the 'offsetof' macro was used incorrectly" + GCC 3.4.4 : classes with constructor + GCC 4.8.1 : classes with non-public variable members" +*/ + +#define Z7_container_of(ptr, type, m) \ + ((type *)(void *)((char *)(void *) \ + (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m))) + +#define Z7_container_of_CONST(ptr, type, m) \ + ((const type *)(const void *)((const char *)(const void *) \ + (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m))) + +/* +#define Z7_container_of_NON_CONST_FROM_CONST(ptr, type, m) \ + ((type *)(void *)(const void *)((const char *)(const void *) \ + (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m))) +*/ + +#endif + +#define Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) ((type *)(void *)(ptr)) + +// #define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) +#define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_container_of(ptr, type, m) +// #define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_container_of_NON_CONST_FROM_CONST(ptr, type, m) + +#define Z7_CONTAINER_FROM_VTBL_CONST(ptr, type, m) Z7_container_of_CONST(ptr, type, m) + +#define Z7_CONTAINER_FROM_VTBL_CLS(ptr, type, m) Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) +/* +#define Z7_CONTAINER_FROM_VTBL_CLS(ptr, type, m) Z7_CONTAINER_FROM_VTBL(ptr, type, m) +*/ +#if defined (__clang__) || defined(__GNUC__) +#define Z7_DIAGNOSCTIC_IGNORE_BEGIN_CAST_QUAL \ + _Pragma("GCC diagnostic push") \ + _Pragma("GCC diagnostic ignored \"-Wcast-qual\"") +#define Z7_DIAGNOSCTIC_IGNORE_END_CAST_QUAL \ + _Pragma("GCC diagnostic pop") +#else +#define Z7_DIAGNOSCTIC_IGNORE_BEGIN_CAST_QUAL +#define Z7_DIAGNOSCTIC_IGNORE_END_CAST_QUAL +#endif + +#define Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR(ptr, type, m, p) \ + Z7_DIAGNOSCTIC_IGNORE_BEGIN_CAST_QUAL \ + type *p = Z7_CONTAINER_FROM_VTBL(ptr, type, m); \ + Z7_DIAGNOSCTIC_IGNORE_END_CAST_QUAL + +#define Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(type) \ + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR(pp, type, vt, p) + + +// #define ZIP7_DECLARE_HANDLE(name) typedef void *name; +#define Z7_DECLARE_HANDLE(name) struct name##_dummy{int unused;}; typedef struct name##_dummy *name; + + +#define Z7_memset_0_ARRAY(a) memset((a), 0, sizeof(a)) + +#ifndef Z7_ARRAY_SIZE +#define Z7_ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0])) +#endif -#define IAlloc_Alloc(p, size) (p)->Alloc((p), size) -#define IAlloc_Free(p, a) (p)->Free((p), a) #ifdef _WIN32 @@ -251,6 +576,22 @@ typedef struct #endif +#define k_PropVar_TimePrec_0 0 +#define k_PropVar_TimePrec_Unix 1 +#define k_PropVar_TimePrec_DOS 2 +#define k_PropVar_TimePrec_HighPrec 3 +#define k_PropVar_TimePrec_Base 16 +#define k_PropVar_TimePrec_100ns (k_PropVar_TimePrec_Base + 7) +#define k_PropVar_TimePrec_1ns (k_PropVar_TimePrec_Base + 9) + EXTERN_C_END #endif + +/* +#ifndef Z7_ST +#ifdef _7ZIP_ST +#define Z7_ST +#endif +#endif +*/ diff --git a/3rdparty/lzma/C/7zVersion.h b/3rdparty/lzma/C/7zVersion.h index 4f15081de78..7549239614e 100644 --- a/3rdparty/lzma/C/7zVersion.h +++ b/3rdparty/lzma/C/7zVersion.h @@ -1,14 +1,21 @@ -#define MY_VER_MAJOR 16 -#define MY_VER_MINOR 04 +#define MY_VER_MAJOR 23 +#define MY_VER_MINOR 01 #define MY_VER_BUILD 0 -#define MY_VERSION_NUMBERS "16.04" -#define MY_VERSION "16.04" -#define MY_DATE "2016-10-04" +#define MY_VERSION_NUMBERS "23.01" +#define MY_VERSION MY_VERSION_NUMBERS + +#ifdef MY_CPU_NAME + #define MY_VERSION_CPU MY_VERSION " (" MY_CPU_NAME ")" +#else + #define MY_VERSION_CPU MY_VERSION +#endif + +#define MY_DATE "2023-06-20" #undef MY_COPYRIGHT #undef MY_VERSION_COPYRIGHT_DATE #define MY_AUTHOR_NAME "Igor Pavlov" #define MY_COPYRIGHT_PD "Igor Pavlov : Public domain" -#define MY_COPYRIGHT_CR "Copyright (c) 1999-2016 Igor Pavlov" +#define MY_COPYRIGHT_CR "Copyright (c) 1999-2023 Igor Pavlov" #ifdef USE_COPYRIGHT_CR #define MY_COPYRIGHT MY_COPYRIGHT_CR @@ -16,4 +23,5 @@ #define MY_COPYRIGHT MY_COPYRIGHT_PD #endif -#define MY_VERSION_COPYRIGHT_DATE MY_VERSION " : " MY_COPYRIGHT " : " MY_DATE +#define MY_COPYRIGHT_DATE MY_COPYRIGHT " : " MY_DATE +#define MY_VERSION_COPYRIGHT_DATE MY_VERSION_CPU " : " MY_COPYRIGHT " : " MY_DATE diff --git a/3rdparty/lzma/C/7zWindows.h b/3rdparty/lzma/C/7zWindows.h new file mode 100644 index 00000000000..42c6db8bfc2 --- /dev/null +++ b/3rdparty/lzma/C/7zWindows.h @@ -0,0 +1,101 @@ +/* 7zWindows.h -- StdAfx +2023-04-02 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_7Z_WINDOWS_H +#define ZIP7_INC_7Z_WINDOWS_H + +#ifdef _WIN32 + +#if defined(__clang__) +# pragma clang diagnostic push +#endif + +#if defined(_MSC_VER) + +#pragma warning(push) +#pragma warning(disable : 4668) // '_WIN32_WINNT' is not defined as a preprocessor macro, replacing with '0' for '#if/#elif' + +#if _MSC_VER == 1900 +// for old kit10 versions +// #pragma warning(disable : 4255) // winuser.h(13979): warning C4255: 'GetThreadDpiAwarenessContext': +#endif +// win10 Windows Kit: +#endif // _MSC_VER + +#if defined(_MSC_VER) && _MSC_VER <= 1200 && !defined(_WIN64) +// for msvc6 without sdk2003 +#define RPC_NO_WINDOWS_H +#endif + +#if defined(__MINGW32__) || defined(__MINGW64__) +// #if defined(__GNUC__) && !defined(__clang__) +#include <windows.h> +#else +#include <Windows.h> +#endif +// #include <basetsd.h> +// #include <wtypes.h> + +// but if precompiled with clang-cl then we need +// #include <windows.h> +#if defined(_MSC_VER) +#pragma warning(pop) +#endif + +#if defined(__clang__) +# pragma clang diagnostic pop +#endif + +#if defined(_MSC_VER) && _MSC_VER <= 1200 && !defined(_WIN64) +#ifndef _W64 + +typedef long LONG_PTR, *PLONG_PTR; +typedef unsigned long ULONG_PTR, *PULONG_PTR; +typedef ULONG_PTR DWORD_PTR, *PDWORD_PTR; + +#define Z7_OLD_WIN_SDK +#endif // _W64 +#endif // _MSC_VER == 1200 + +#ifdef Z7_OLD_WIN_SDK + +#ifndef INVALID_FILE_ATTRIBUTES +#define INVALID_FILE_ATTRIBUTES ((DWORD)-1) +#endif +#ifndef INVALID_SET_FILE_POINTER +#define INVALID_SET_FILE_POINTER ((DWORD)-1) +#endif +#ifndef FILE_SPECIAL_ACCESS +#define FILE_SPECIAL_ACCESS (FILE_ANY_ACCESS) +#endif + +// ShlObj.h: +// #define BIF_NEWDIALOGSTYLE 0x0040 + +#pragma warning(disable : 4201) +// #pragma warning(disable : 4115) + +#undef VARIANT_TRUE +#define VARIANT_TRUE ((VARIANT_BOOL)-1) +#endif + +#endif // Z7_OLD_WIN_SDK + +#ifdef UNDER_CE +#undef VARIANT_TRUE +#define VARIANT_TRUE ((VARIANT_BOOL)-1) +#endif + + +#if defined(_MSC_VER) +#if _MSC_VER >= 1400 && _MSC_VER <= 1600 + // BaseTsd.h(148) : 'HandleToULong' : unreferenced inline function has been removed + // string.h + // #pragma warning(disable : 4514) +#endif +#endif + + +/* #include "7zTypes.h" */ + +#endif diff --git a/3rdparty/lzma/C/7zip_gcc_c.mak b/3rdparty/lzma/C/7zip_gcc_c.mak new file mode 100644 index 00000000000..f19a99bad5c --- /dev/null +++ b/3rdparty/lzma/C/7zip_gcc_c.mak @@ -0,0 +1,360 @@ + +MY_ARCH_2 = $(MY_ARCH) + +MY_ASM = jwasm +MY_ASM = asmc + +ifndef RC +#RC=windres.exe --target=pe-x86-64 +#RC=windres.exe -F pe-i386 +RC=windres.exe +endif + +PROGPATH = $(O)/$(PROG) +PROGPATH_STATIC = $(O)/$(PROG)s + +ifneq ($(CC), xlc) +CFLAGS_WARN_WALL = -Wall -Werror -Wextra +endif + +# for object file +CFLAGS_BASE_LIST = -c +# for ASM file +# CFLAGS_BASE_LIST = -S + +FLAGS_FLTO = +FLAGS_FLTO = -flto + +CFLAGS_BASE = $(MY_ARCH_2) -O2 $(CFLAGS_BASE_LIST) $(CFLAGS_WARN_WALL) $(CFLAGS_WARN) \ + -DNDEBUG -D_REENTRANT -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE + + +ifdef SystemDrive +IS_MINGW = 1 +else +ifdef SYSTEMDRIVE +# ifdef OS +IS_MINGW = 1 +endif +endif + +ifdef IS_MINGW +LDFLAGS_STATIC_2 = -static +else +ifndef DEF_FILE +ifndef IS_NOT_STANDALONE +ifndef MY_DYNAMIC_LINK +ifneq ($(CC), clang) +LDFLAGS_STATIC_2 = +# -static +# -static-libstdc++ -static-libgcc +endif +endif +endif +endif +endif + +LDFLAGS_STATIC = -DNDEBUG $(LDFLAGS_STATIC_2) + +ifdef DEF_FILE + + +ifdef IS_MINGW +SHARED_EXT=.dll +LDFLAGS = -shared -DEF $(DEF_FILE) $(LDFLAGS_STATIC) +else +SHARED_EXT=.so +LDFLAGS = -shared -fPIC $(LDFLAGS_STATIC) +CC_SHARED=-fPIC +endif + + +else + +LDFLAGS = $(LDFLAGS_STATIC) +# -s is not required for clang, do we need it for GGC ??? +# -s + +#-static -static-libgcc -static-libstdc++ + +ifdef IS_MINGW +SHARED_EXT=.exe +else +SHARED_EXT= +endif + +endif + + +PROGPATH = $(O)/$(PROG)$(SHARED_EXT) +PROGPATH_STATIC = $(O)/$(PROG)s$(SHARED_EXT) + +ifndef O +O=_o +endif + +ifdef IS_MINGW + +ifdef MSYSTEM +RM = rm -f +MY_MKDIR=mkdir -p +DEL_OBJ_EXE = -$(RM) $(PROGPATH) $(PROGPATH_STATIC) $(OBJS) +else +RM = del +MY_MKDIR=mkdir +DEL_OBJ_EXE = -$(RM) $(O)\*.o $(O)\$(PROG).exe $(O)\$(PROG).dll +endif + + +LIB2 = -lOle32 -loleaut32 -luuid -ladvapi32 -lUser32 -lShell32 + +CFLAGS_EXTRA = -DUNICODE -D_UNICODE +# -Wno-delete-non-virtual-dtor + + +else + +RM = rm -f +MY_MKDIR=mkdir -p +# CFLAGS_BASE := $(CFLAGS_BASE) -DZ7_ST +# CFLAGS_EXTRA = -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE + +# LOCAL_LIBS=-lpthread +# LOCAL_LIBS_DLL=$(LOCAL_LIBS) -ldl +LIB2 = -lpthread -ldl + +DEL_OBJ_EXE = -$(RM) $(PROGPATH) $(PROGPATH_STATIC) $(OBJS) + +endif + + +ifdef IS_X64 +AFLAGS_ABI = -elf64 -DABI_LINUX +else +AFLAGS_ABI = -elf -DABI_LINUX -DABI_CDECL +# -DABI_CDECL +# -DABI_LINUX +# -DABI_CDECL +endif +AFLAGS = $(AFLAGS_ABI) -Fo$(O)/ + +C_WARN_FLAGS = + +CFLAGS = $(LOCAL_FLAGS) $(CFLAGS_BASE2) $(CFLAGS_BASE) $(CFLAGS_EXTRA) $(C_WARN_FLAGS) $(FLAGS_FLTO) $(CC_SHARED) -o $@ + +STATIC_TARGET= +ifdef COMPL_STATIC +STATIC_TARGET=$(PROGPATH_STATIC) +endif + + +all: $(O) $(PROGPATH) $(STATIC_TARGET) + +$(O): + $(MY_MKDIR) $(O) + +ifneq ($(CC), $(CROSS_COMPILE)clang) +LFLAGS_STRIP = -s +endif + +LFLAGS_ALL = $(LFLAGS_STRIP) $(MY_ARCH_2) $(LDFLAGS) $(FLAGS_FLTO) $(LD_arch) $(OBJS) $(MY_LIBS) $(LIB2) +$(PROGPATH): $(OBJS) + $(CC) -o $(PROGPATH) $(LFLAGS_ALL) + +$(PROGPATH_STATIC): $(OBJS) + $(CC) -static -o $(PROGPATH_STATIC) $(LFLAGS_ALL) + + +ifndef NO_DEFAULT_RES +# old mingw without -FO +# windres.exe $(RFLAGS) resource.rc $O/resource.o +$O/resource.o: resource.rc + $(RC) $(RFLAGS) resource.rc $(O)/resource.o +endif +# windres.exe $(RFLAGS) resource.rc $(O)\resource.o +# windres.exe $(RFLAGS) resource.rc -FO $(O)/resource.o +# $(RC) $(RFLAGS) resource.rc -FO $(O)/resource.o + + + +$O/7zAlloc.o: ../../../C/7zAlloc.c + $(CC) $(CFLAGS) $< +$O/7zArcIn.o: ../../../C/7zArcIn.c + $(CC) $(CFLAGS) $< +$O/7zBuf.o: ../../../C/7zBuf.c + $(CC) $(CFLAGS) $< +$O/7zBuf2.o: ../../../C/7zBuf2.c + $(CC) $(CFLAGS) $< +$O/7zCrc.o: ../../../C/7zCrc.c + $(CC) $(CFLAGS) $< +$O/7zDec.o: ../../../C/7zDec.c + $(CC) $(CFLAGS) $< +$O/7zFile.o: ../../../C/7zFile.c + $(CC) $(CFLAGS) $< +$O/7zStream.o: ../../../C/7zStream.c + $(CC) $(CFLAGS) $< +$O/Aes.o: ../../../C/Aes.c + $(CC) $(CFLAGS) $< +$O/Alloc.o: ../../../C/Alloc.c + $(CC) $(CFLAGS) $< +$O/Bcj2.o: ../../../C/Bcj2.c + $(CC) $(CFLAGS) $< +$O/Bcj2Enc.o: ../../../C/Bcj2Enc.c + $(CC) $(CFLAGS) $< +$O/Blake2s.o: ../../../C/Blake2s.c + $(CC) $(CFLAGS) $< +$O/Bra.o: ../../../C/Bra.c + $(CC) $(CFLAGS) $< +$O/Bra86.o: ../../../C/Bra86.c + $(CC) $(CFLAGS) $< +$O/BraIA64.o: ../../../C/BraIA64.c + $(CC) $(CFLAGS) $< +$O/BwtSort.o: ../../../C/BwtSort.c + $(CC) $(CFLAGS) $< + +$O/CpuArch.o: ../../../C/CpuArch.c + $(CC) $(CFLAGS) $< +$O/Delta.o: ../../../C/Delta.c + $(CC) $(CFLAGS) $< +$O/DllSecur.o: ../../../C/DllSecur.c + $(CC) $(CFLAGS) $< +$O/HuffEnc.o: ../../../C/HuffEnc.c + $(CC) $(CFLAGS) $< +$O/LzFind.o: ../../../C/LzFind.c + $(CC) $(CFLAGS) $< + +# ifdef MT_FILES +$O/LzFindMt.o: ../../../C/LzFindMt.c + $(CC) $(CFLAGS) $< +$O/LzFindOpt.o: ../../../C/LzFindOpt.c + $(CC) $(CFLAGS) $< + +$O/Threads.o: ../../../C/Threads.c + $(CC) $(CFLAGS) $< +# endif + +$O/LzmaEnc.o: ../../../C/LzmaEnc.c + $(CC) $(CFLAGS) $< +$O/Lzma86Dec.o: ../../../C/Lzma86Dec.c + $(CC) $(CFLAGS) $< +$O/Lzma86Enc.o: ../../../C/Lzma86Enc.c + $(CC) $(CFLAGS) $< +$O/Lzma2Dec.o: ../../../C/Lzma2Dec.c + $(CC) $(CFLAGS) $< +$O/Lzma2DecMt.o: ../../../C/Lzma2DecMt.c + $(CC) $(CFLAGS) $< +$O/Lzma2Enc.o: ../../../C/Lzma2Enc.c + $(CC) $(CFLAGS) $< +$O/LzmaLib.o: ../../../C/LzmaLib.c + $(CC) $(CFLAGS) $< +$O/MtCoder.o: ../../../C/MtCoder.c + $(CC) $(CFLAGS) $< +$O/MtDec.o: ../../../C/MtDec.c + $(CC) $(CFLAGS) $< +$O/Ppmd7.o: ../../../C/Ppmd7.c + $(CC) $(CFLAGS) $< +$O/Ppmd7aDec.o: ../../../C/Ppmd7aDec.c + $(CC) $(CFLAGS) $< +$O/Ppmd7Dec.o: ../../../C/Ppmd7Dec.c + $(CC) $(CFLAGS) $< +$O/Ppmd7Enc.o: ../../../C/Ppmd7Enc.c + $(CC) $(CFLAGS) $< +$O/Ppmd8.o: ../../../C/Ppmd8.c + $(CC) $(CFLAGS) $< +$O/Ppmd8Dec.o: ../../../C/Ppmd8Dec.c + $(CC) $(CFLAGS) $< +$O/Ppmd8Enc.o: ../../../C/Ppmd8Enc.c + $(CC) $(CFLAGS) $< +$O/Sha1.o: ../../../C/Sha1.c + $(CC) $(CFLAGS) $< +$O/Sha256.o: ../../../C/Sha256.c + $(CC) $(CFLAGS) $< +$O/Sort.o: ../../../C/Sort.c + $(CC) $(CFLAGS) $< +$O/SwapBytes.o: ../../../C/SwapBytes.c + $(CC) $(CFLAGS) $< +$O/Xz.o: ../../../C/Xz.c + $(CC) $(CFLAGS) $< +$O/XzCrc64.o: ../../../C/XzCrc64.c + $(CC) $(CFLAGS) $< +$O/XzDec.o: ../../../C/XzDec.c + $(CC) $(CFLAGS) $< +$O/XzEnc.o: ../../../C/XzEnc.c + $(CC) $(CFLAGS) $< +$O/XzIn.o: ../../../C/XzIn.c + $(CC) $(CFLAGS) $< + + +ifdef USE_ASM +ifdef IS_X64 +USE_X86_ASM=1 +else +ifdef IS_X86 +USE_X86_ASM=1 +endif +endif +endif + +ifdef USE_X86_ASM +$O/7zCrcOpt.o: ../../../Asm/x86/7zCrcOpt.asm + $(MY_ASM) $(AFLAGS) $< +$O/XzCrc64Opt.o: ../../../Asm/x86/XzCrc64Opt.asm + $(MY_ASM) $(AFLAGS) $< +$O/AesOpt.o: ../../../Asm/x86/AesOpt.asm + $(MY_ASM) $(AFLAGS) $< +$O/Sha1Opt.o: ../../../Asm/x86/Sha1Opt.asm + $(MY_ASM) $(AFLAGS) $< +$O/Sha256Opt.o: ../../../Asm/x86/Sha256Opt.asm + $(MY_ASM) $(AFLAGS) $< +else +$O/7zCrcOpt.o: ../../7zCrcOpt.c + $(CC) $(CFLAGS) $< +$O/XzCrc64Opt.o: ../../XzCrc64Opt.c + $(CC) $(CFLAGS) $< +$O/Sha1Opt.o: ../../Sha1Opt.c + $(CC) $(CFLAGS) $< +$O/Sha256Opt.o: ../../Sha256Opt.c + $(CC) $(CFLAGS) $< +$O/AesOpt.o: ../../AesOpt.c + $(CC) $(CFLAGS) $< +endif + + +ifdef USE_LZMA_DEC_ASM + +ifdef IS_X64 +$O/LzmaDecOpt.o: ../../../Asm/x86/LzmaDecOpt.asm + $(MY_ASM) $(AFLAGS) $< +endif + +ifdef IS_ARM64 +$O/LzmaDecOpt.o: ../../../Asm/arm64/LzmaDecOpt.S ../../../Asm/arm64/7zAsm.S + $(CC) $(CFLAGS) $< +endif + +$O/LzmaDec.o: ../../LzmaDec.c + $(CC) $(CFLAGS) -DZ7_LZMA_DEC_OPT $< + +else + +$O/LzmaDec.o: ../../LzmaDec.c + $(CC) $(CFLAGS) $< + +endif + + + +$O/7zMain.o: ../../../C/Util/7z/7zMain.c + $(CC) $(CFLAGS) $< +$O/7zipInstall.o: ../../../C/Util/7zipInstall/7zipInstall.c + $(CC) $(CFLAGS) $< +$O/7zipUninstall.o: ../../../C/Util/7zipUninstall/7zipUninstall.c + $(CC) $(CFLAGS) $< +$O/LzmaUtil.o: ../../../C/Util/Lzma/LzmaUtil.c + $(CC) $(CFLAGS) $< +$O/XzUtil.o: ../../../C/Util/Xz/XzUtil.c + $(CC) $(CFLAGS) $< + + +clean: + -$(DEL_OBJ_EXE) diff --git a/3rdparty/lzma/C/Aes.c b/3rdparty/lzma/C/Aes.c index b10b9e58a55..bcaafab1188 100644 --- a/3rdparty/lzma/C/Aes.c +++ b/3rdparty/lzma/C/Aes.c @@ -1,10 +1,17 @@ /* Aes.c -- AES encryption / decryption -2016-05-21 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" -#include "Aes.h" #include "CpuArch.h" +#include "Aes.h" + +AES_CODE_FUNC g_AesCbc_Decode; +#ifndef Z7_SFX +AES_CODE_FUNC g_AesCbc_Encode; +AES_CODE_FUNC g_AesCtr_Code; +UInt32 g_Aes_SupportedFunctions_Flags; +#endif static UInt32 T[256 * 4]; static const Byte Sbox[256] = { @@ -25,23 +32,10 @@ static const Byte Sbox[256] = { 0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf, 0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16}; -void MY_FAST_CALL AesCbc_Encode(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCbc_Decode(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCtr_Code(UInt32 *ivAes, Byte *data, size_t numBlocks); - -void MY_FAST_CALL AesCbc_Encode_Intel(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCbc_Decode_Intel(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCtr_Code_Intel(UInt32 *ivAes, Byte *data, size_t numBlocks); - -AES_CODE_FUNC g_AesCbc_Encode; -AES_CODE_FUNC g_AesCbc_Decode; -AES_CODE_FUNC g_AesCtr_Code; static UInt32 D[256 * 4]; static Byte InvS[256]; -static const Byte Rcon[11] = { 0x00, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1b, 0x36 }; - #define xtime(x) ((((x) << 1) ^ (((x) & 0x80) != 0 ? 0x1B : 0)) & 0xFF) #define Ui32(a0, a1, a2, a3) ((UInt32)(a0) | ((UInt32)(a1) << 8) | ((UInt32)(a2) << 16) | ((UInt32)(a3) << 24)) @@ -49,7 +43,43 @@ static const Byte Rcon[11] = { 0x00, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0 #define gb0(x) ( (x) & 0xFF) #define gb1(x) (((x) >> ( 8)) & 0xFF) #define gb2(x) (((x) >> (16)) & 0xFF) -#define gb3(x) (((x) >> (24)) & 0xFF) +#define gb3(x) (((x) >> (24))) + +#define gb(n, x) gb ## n(x) + +#define TT(x) (T + (x << 8)) +#define DD(x) (D + (x << 8)) + + +// #define Z7_SHOW_AES_STATUS + +#ifdef MY_CPU_X86_OR_AMD64 + #define USE_HW_AES +#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE) + #if defined(__clang__) + #if (__clang_major__ >= 8) // fix that check + #define USE_HW_AES + #endif + #elif defined(__GNUC__) + #if (__GNUC__ >= 6) // fix that check + #define USE_HW_AES + #endif + #elif defined(_MSC_VER) + #if _MSC_VER >= 1910 + #define USE_HW_AES + #endif + #endif +#endif + +#ifdef USE_HW_AES +#ifdef Z7_SHOW_AES_STATUS +#include <stdio.h> +#define PRF(x) x +#else +#define PRF(x) +#endif +#endif + void AesGenTables(void) { @@ -60,46 +90,76 @@ void AesGenTables(void) for (i = 0; i < 256; i++) { { - UInt32 a1 = Sbox[i]; - UInt32 a2 = xtime(a1); - UInt32 a3 = a2 ^ a1; - T[ i] = Ui32(a2, a1, a1, a3); - T[0x100 + i] = Ui32(a3, a2, a1, a1); - T[0x200 + i] = Ui32(a1, a3, a2, a1); - T[0x300 + i] = Ui32(a1, a1, a3, a2); + const UInt32 a1 = Sbox[i]; + const UInt32 a2 = xtime(a1); + const UInt32 a3 = a2 ^ a1; + TT(0)[i] = Ui32(a2, a1, a1, a3); + TT(1)[i] = Ui32(a3, a2, a1, a1); + TT(2)[i] = Ui32(a1, a3, a2, a1); + TT(3)[i] = Ui32(a1, a1, a3, a2); } { - UInt32 a1 = InvS[i]; - UInt32 a2 = xtime(a1); - UInt32 a4 = xtime(a2); - UInt32 a8 = xtime(a4); - UInt32 a9 = a8 ^ a1; - UInt32 aB = a8 ^ a2 ^ a1; - UInt32 aD = a8 ^ a4 ^ a1; - UInt32 aE = a8 ^ a4 ^ a2; - D[ i] = Ui32(aE, a9, aD, aB); - D[0x100 + i] = Ui32(aB, aE, a9, aD); - D[0x200 + i] = Ui32(aD, aB, aE, a9); - D[0x300 + i] = Ui32(a9, aD, aB, aE); + const UInt32 a1 = InvS[i]; + const UInt32 a2 = xtime(a1); + const UInt32 a4 = xtime(a2); + const UInt32 a8 = xtime(a4); + const UInt32 a9 = a8 ^ a1; + const UInt32 aB = a8 ^ a2 ^ a1; + const UInt32 aD = a8 ^ a4 ^ a1; + const UInt32 aE = a8 ^ a4 ^ a2; + DD(0)[i] = Ui32(aE, a9, aD, aB); + DD(1)[i] = Ui32(aB, aE, a9, aD); + DD(2)[i] = Ui32(aD, aB, aE, a9); + DD(3)[i] = Ui32(a9, aD, aB, aE); } } - g_AesCbc_Encode = AesCbc_Encode; - g_AesCbc_Decode = AesCbc_Decode; - g_AesCtr_Code = AesCtr_Code; + { + AES_CODE_FUNC d = AesCbc_Decode; + #ifndef Z7_SFX + AES_CODE_FUNC e = AesCbc_Encode; + AES_CODE_FUNC c = AesCtr_Code; + UInt32 flags = 0; + #endif - #ifdef MY_CPU_X86_OR_AMD64 - if (CPU_Is_Aes_Supported()) + #ifdef USE_HW_AES + if (CPU_IsSupported_AES()) { - g_AesCbc_Encode = AesCbc_Encode_Intel; - g_AesCbc_Decode = AesCbc_Decode_Intel; - g_AesCtr_Code = AesCtr_Code_Intel; + // #pragma message ("AES HW") + PRF(printf("\n===AES HW\n")); + d = AesCbc_Decode_HW; + + #ifndef Z7_SFX + e = AesCbc_Encode_HW; + c = AesCtr_Code_HW; + flags = k_Aes_SupportedFunctions_HW; + #endif + + #ifdef MY_CPU_X86_OR_AMD64 + if (CPU_IsSupported_VAES_AVX2()) + { + PRF(printf("\n===vaes avx2\n")); + d = AesCbc_Decode_HW_256; + #ifndef Z7_SFX + c = AesCtr_Code_HW_256; + flags |= k_Aes_SupportedFunctions_HW_256; + #endif + } + #endif } #endif + + g_AesCbc_Decode = d; + #ifndef Z7_SFX + g_AesCbc_Encode = e; + g_AesCtr_Code = c; + g_Aes_SupportedFunctions_Flags = flags; + #endif + } } -#define HT(i, x, s) (T + (x << 8))[gb ## x(s[(i + x) & 3])] +#define HT(i, x, s) TT(x)[gb(x, s[(i + x) & 3])] #define HT4(m, i, s, p) m[i] = \ HT(i, 0, s) ^ \ @@ -113,11 +173,11 @@ void AesGenTables(void) HT4(m, 2, s, p); \ HT4(m, 3, s, p); \ -#define FT(i, x) Sbox[gb ## x(m[(i + x) & 3])] +#define FT(i, x) Sbox[gb(x, m[(i + x) & 3])] #define FT4(i) dest[i] = Ui32(FT(i, 0), FT(i, 1), FT(i, 2), FT(i, 3)) ^ w[i]; -#define HD(i, x, s) (D + (x << 8))[gb ## x(s[(i - x) & 3])] +#define HD(i, x, s) DD(x)[gb(x, s[(i - x) & 3])] #define HD4(m, i, s, p) m[i] = \ HD(i, 0, s) ^ \ @@ -131,13 +191,16 @@ void AesGenTables(void) HD4(m, 2, s, p); \ HD4(m, 3, s, p); \ -#define FD(i, x) InvS[gb ## x(m[(i - x) & 3])] +#define FD(i, x) InvS[gb(x, m[(i - x) & 3])] #define FD4(i) dest[i] = Ui32(FD(i, 0), FD(i, 1), FD(i, 2), FD(i, 3)) ^ w[i]; -void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize) +void Z7_FASTCALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize) { - unsigned i, wSize; - wSize = keySize + 28; + unsigned i, m; + const UInt32 *wLim; + UInt32 t; + UInt32 rcon = 1; + keySize /= 4; w[0] = ((UInt32)keySize / 2) + 3; w += 4; @@ -145,19 +208,29 @@ void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize) for (i = 0; i < keySize; i++, key += 4) w[i] = GetUi32(key); - for (; i < wSize; i++) + t = w[(size_t)keySize - 1]; + wLim = w + (size_t)keySize * 3 + 28; + m = 0; + do { - UInt32 t = w[i - 1]; - unsigned rem = i % keySize; - if (rem == 0) - t = Ui32(Sbox[gb1(t)] ^ Rcon[i / keySize], Sbox[gb2(t)], Sbox[gb3(t)], Sbox[gb0(t)]); - else if (keySize > 6 && rem == 4) + if (m == 0) + { + t = Ui32(Sbox[gb1(t)] ^ rcon, Sbox[gb2(t)], Sbox[gb3(t)], Sbox[gb0(t)]); + rcon <<= 1; + if (rcon & 0x100) + rcon = 0x1b; + m = keySize; + } + else if (m == 4 && keySize > 6) t = Ui32(Sbox[gb0(t)], Sbox[gb1(t)], Sbox[gb2(t)], Sbox[gb3(t)]); - w[i] = w[i - keySize] ^ t; + m--; + t ^= w[0]; + w[keySize] = t; } + while (++w != wLim); } -void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize) +void Z7_FASTCALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize) { unsigned i, num; Aes_SetKey_Enc(w, key, keySize); @@ -167,10 +240,10 @@ void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize) { UInt32 r = w[i]; w[i] = - D[ (unsigned)Sbox[gb0(r)]] ^ - D[0x100 + (unsigned)Sbox[gb1(r)]] ^ - D[0x200 + (unsigned)Sbox[gb2(r)]] ^ - D[0x300 + (unsigned)Sbox[gb3(r)]]; + DD(0)[Sbox[gb0(r)]] ^ + DD(1)[Sbox[gb1(r)]] ^ + DD(2)[Sbox[gb2(r)]] ^ + DD(3)[Sbox[gb3(r)]]; } } @@ -178,6 +251,7 @@ void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize) src and dest are pointers to 4 UInt32 words. src and dest can point to same block */ +// Z7_FORCE_INLINE static void Aes_Encode(const UInt32 *w, UInt32 *dest, const UInt32 *src) { UInt32 s[4]; @@ -191,16 +265,20 @@ static void Aes_Encode(const UInt32 *w, UInt32 *dest, const UInt32 *src) w += 4; for (;;) { - HT16(m, s, 0); + HT16(m, s, 0) if (--numRounds2 == 0) break; - HT16(s, m, 4); + HT16(s, m, 4) w += 8; } w += 4; - FT4(0); FT4(1); FT4(2); FT4(3); + FT4(0) + FT4(1) + FT4(2) + FT4(3) } +Z7_FORCE_INLINE static void Aes_Decode(const UInt32 *w, UInt32 *dest, const UInt32 *src) { UInt32 s[4]; @@ -214,12 +292,15 @@ static void Aes_Decode(const UInt32 *w, UInt32 *dest, const UInt32 *src) for (;;) { w -= 8; - HD16(m, s, 4); + HD16(m, s, 4) if (--numRounds2 == 0) break; - HD16(s, m, 0); + HD16(s, m, 0) } - FD4(0); FD4(1); FD4(2); FD4(3); + FD4(0) + FD4(1) + FD4(2) + FD4(3) } void AesCbc_Init(UInt32 *p, const Byte *iv) @@ -229,7 +310,7 @@ void AesCbc_Init(UInt32 *p, const Byte *iv) p[i] = GetUi32(iv + i * 4); } -void MY_FAST_CALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks) +void Z7_FASTCALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks) { for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE) { @@ -240,14 +321,14 @@ void MY_FAST_CALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks) Aes_Encode(p + 4, p, p); - SetUi32(data, p[0]); - SetUi32(data + 4, p[1]); - SetUi32(data + 8, p[2]); - SetUi32(data + 12, p[3]); + SetUi32(data, p[0]) + SetUi32(data + 4, p[1]) + SetUi32(data + 8, p[2]) + SetUi32(data + 12, p[3]) } } -void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks) +void Z7_FASTCALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks) { UInt32 in[4], out[4]; for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE) @@ -259,10 +340,10 @@ void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks) Aes_Decode(p + 4, out, in); - SetUi32(data, p[0] ^ out[0]); - SetUi32(data + 4, p[1] ^ out[1]); - SetUi32(data + 8, p[2] ^ out[2]); - SetUi32(data + 12, p[3] ^ out[3]); + SetUi32(data, p[0] ^ out[0]) + SetUi32(data + 4, p[1] ^ out[1]) + SetUi32(data + 8, p[2] ^ out[2]) + SetUi32(data + 12, p[3] ^ out[3]) p[0] = in[0]; p[1] = in[1]; @@ -271,25 +352,42 @@ void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks) } } -void MY_FAST_CALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks) +void Z7_FASTCALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks) { for (; numBlocks != 0; numBlocks--) { UInt32 temp[4]; - Byte buf[16]; - int i; + unsigned i; if (++p[0] == 0) p[1]++; Aes_Encode(p + 4, temp, p); - SetUi32(buf, temp[0]); - SetUi32(buf + 4, temp[1]); - SetUi32(buf + 8, temp[2]); - SetUi32(buf + 12, temp[3]); - - for (i = 0; i < 16; i++) - *data++ ^= buf[i]; + for (i = 0; i < 4; i++, data += 4) + { + const UInt32 t = temp[i]; + + #ifdef MY_CPU_LE_UNALIGN + *((UInt32 *)(void *)data) ^= t; + #else + data[0] = (Byte)(data[0] ^ (t & 0xFF)); + data[1] = (Byte)(data[1] ^ ((t >> 8) & 0xFF)); + data[2] = (Byte)(data[2] ^ ((t >> 16) & 0xFF)); + data[3] = (Byte)(data[3] ^ ((t >> 24))); + #endif + } } } + +#undef xtime +#undef Ui32 +#undef gb0 +#undef gb1 +#undef gb2 +#undef gb3 +#undef gb +#undef TT +#undef DD +#undef USE_HW_AES +#undef PRF diff --git a/3rdparty/lzma/C/Aes.h b/3rdparty/lzma/C/Aes.h index 64979b5bcbf..7f0182ac0f5 100644 --- a/3rdparty/lzma/C/Aes.h +++ b/3rdparty/lzma/C/Aes.h @@ -1,8 +1,8 @@ /* Aes.h -- AES encryption / decryption -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __AES_H -#define __AES_H +#ifndef ZIP7_INC_AES_H +#define ZIP7_INC_AES_H #include "7zTypes.h" @@ -20,18 +20,40 @@ void AesGenTables(void); /* aes - 16-byte aligned pointer to keyMode+roundKeys sequence */ /* keySize = 16 or 24 or 32 (bytes) */ -typedef void (MY_FAST_CALL *AES_SET_KEY_FUNC)(UInt32 *aes, const Byte *key, unsigned keySize); -void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *aes, const Byte *key, unsigned keySize); -void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *aes, const Byte *key, unsigned keySize); +typedef void (Z7_FASTCALL *AES_SET_KEY_FUNC)(UInt32 *aes, const Byte *key, unsigned keySize); +void Z7_FASTCALL Aes_SetKey_Enc(UInt32 *aes, const Byte *key, unsigned keySize); +void Z7_FASTCALL Aes_SetKey_Dec(UInt32 *aes, const Byte *key, unsigned keySize); /* ivAes - 16-byte aligned pointer to iv+keyMode+roundKeys sequence: UInt32[AES_NUM_IVMRK_WORDS] */ void AesCbc_Init(UInt32 *ivAes, const Byte *iv); /* iv size is AES_BLOCK_SIZE */ + /* data - 16-byte aligned pointer to data */ /* numBlocks - the number of 16-byte blocks in data array */ -typedef void (MY_FAST_CALL *AES_CODE_FUNC)(UInt32 *ivAes, Byte *data, size_t numBlocks); -extern AES_CODE_FUNC g_AesCbc_Encode; +typedef void (Z7_FASTCALL *AES_CODE_FUNC)(UInt32 *ivAes, Byte *data, size_t numBlocks); + extern AES_CODE_FUNC g_AesCbc_Decode; +#ifndef Z7_SFX +extern AES_CODE_FUNC g_AesCbc_Encode; extern AES_CODE_FUNC g_AesCtr_Code; +#define k_Aes_SupportedFunctions_HW (1 << 2) +#define k_Aes_SupportedFunctions_HW_256 (1 << 3) +extern UInt32 g_Aes_SupportedFunctions_Flags; +#endif + + +#define Z7_DECLARE_AES_CODE_FUNC(funcName) \ + void Z7_FASTCALL funcName(UInt32 *ivAes, Byte *data, size_t numBlocks); + +Z7_DECLARE_AES_CODE_FUNC (AesCbc_Encode) +Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode) +Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code) + +Z7_DECLARE_AES_CODE_FUNC (AesCbc_Encode_HW) +Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode_HW) +Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code_HW) + +Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode_HW_256) +Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code_HW_256) EXTERN_C_END diff --git a/3rdparty/lzma/C/AesOpt.c b/3rdparty/lzma/C/AesOpt.c index 8234f0d1465..cfa6413f4b8 100644 --- a/3rdparty/lzma/C/AesOpt.c +++ b/3rdparty/lzma/C/AesOpt.c @@ -1,184 +1,840 @@ -/* AesOpt.c -- Intel's AES -2013-11-12 : Igor Pavlov : Public domain */ +/* AesOpt.c -- AES optimized code for x86 AES hardware instructions +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" +#include "Aes.h" #include "CpuArch.h" #ifdef MY_CPU_X86_OR_AMD64 -#if _MSC_VER >= 1500 && !defined(__clang__) -#define USE_INTEL_AES + + #if defined(__INTEL_COMPILER) + #if (__INTEL_COMPILER >= 1110) + #define USE_INTEL_AES + #if (__INTEL_COMPILER >= 1900) + #define USE_INTEL_VAES + #endif + #endif + #elif defined(__clang__) && (__clang_major__ > 3 || __clang_major__ == 3 && __clang_minor__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ > 4 || __GNUC__ == 4 && __GNUC_MINOR__ >= 4) + #define USE_INTEL_AES + #if !defined(__AES__) + #define ATTRIB_AES __attribute__((__target__("aes"))) + #endif + #if defined(__clang__) && (__clang_major__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ >= 8) + #define USE_INTEL_VAES + #if !defined(__AES__) || !defined(__VAES__) || !defined(__AVX__) || !defined(__AVX2__) + #define ATTRIB_VAES __attribute__((__target__("aes,vaes,avx,avx2"))) + #endif + #endif + #elif defined(_MSC_VER) + #if (_MSC_VER > 1500) || (_MSC_FULL_VER >= 150030729) + #define USE_INTEL_AES + #if (_MSC_VER >= 1910) + #define USE_INTEL_VAES + #endif + #endif + #endif + +#ifndef ATTRIB_AES + #define ATTRIB_AES #endif +#ifndef ATTRIB_VAES + #define ATTRIB_VAES #endif + #ifdef USE_INTEL_AES #include <wmmintrin.h> -void MY_FAST_CALL AesCbc_Encode_Intel(__m128i *p, __m128i *data, size_t numBlocks) +#ifndef USE_INTEL_VAES +#define AES_TYPE_keys UInt32 +#define AES_TYPE_data Byte +// #define AES_TYPE_keys __m128i +// #define AES_TYPE_data __m128i +#endif + +#define AES_FUNC_START(name) \ + void Z7_FASTCALL name(UInt32 *ivAes, Byte *data8, size_t numBlocks) + // void Z7_FASTCALL name(__m128i *p, __m128i *data, size_t numBlocks) + +#define AES_FUNC_START2(name) \ +AES_FUNC_START (name); \ +ATTRIB_AES \ +AES_FUNC_START (name) + +#define MM_OP(op, dest, src) dest = op(dest, src); +#define MM_OP_m(op, src) MM_OP(op, m, src) + +#define MM_XOR( dest, src) MM_OP(_mm_xor_si128, dest, src) +#define AVX_XOR(dest, src) MM_OP(_mm256_xor_si256, dest, src) + + +AES_FUNC_START2 (AesCbc_Encode_HW) { + __m128i *p = (__m128i *)(void *)ivAes; + __m128i *data = (__m128i *)(void *)data8; __m128i m = *p; + const __m128i k0 = p[2]; + const __m128i k1 = p[3]; + const UInt32 numRounds2 = *(const UInt32 *)(p + 1) - 1; for (; numBlocks != 0; numBlocks--, data++) { - UInt32 numRounds2 = *(const UInt32 *)(p + 1) - 1; - const __m128i *w = p + 3; - m = _mm_xor_si128(m, *data); - m = _mm_xor_si128(m, p[2]); + UInt32 r = numRounds2; + const __m128i *w = p + 4; + __m128i temp = *data; + MM_XOR (temp, k0) + MM_XOR (m, temp) + MM_OP_m (_mm_aesenc_si128, k1) do { - m = _mm_aesenc_si128(m, w[0]); - m = _mm_aesenc_si128(m, w[1]); + MM_OP_m (_mm_aesenc_si128, w[0]) + MM_OP_m (_mm_aesenc_si128, w[1]) w += 2; } - while (--numRounds2 != 0); - m = _mm_aesenc_si128(m, w[0]); - m = _mm_aesenclast_si128(m, w[1]); + while (--r); + MM_OP_m (_mm_aesenclast_si128, w[0]) *data = m; } *p = m; } -#define NUM_WAYS 3 -#define AES_OP_W(op, n) { \ - const __m128i t = w[n]; \ - m0 = op(m0, t); \ - m1 = op(m1, t); \ - m2 = op(m2, t); \ - } +#define WOP_1(op) +#define WOP_2(op) WOP_1 (op) op (m1, 1) +#define WOP_3(op) WOP_2 (op) op (m2, 2) +#define WOP_4(op) WOP_3 (op) op (m3, 3) +#ifdef MY_CPU_AMD64 +#define WOP_5(op) WOP_4 (op) op (m4, 4) +#define WOP_6(op) WOP_5 (op) op (m5, 5) +#define WOP_7(op) WOP_6 (op) op (m6, 6) +#define WOP_8(op) WOP_7 (op) op (m7, 7) +#endif +/* +#define WOP_9(op) WOP_8 (op) op (m8, 8); +#define WOP_10(op) WOP_9 (op) op (m9, 9); +#define WOP_11(op) WOP_10(op) op (m10, 10); +#define WOP_12(op) WOP_11(op) op (m11, 11); +#define WOP_13(op) WOP_12(op) op (m12, 12); +#define WOP_14(op) WOP_13(op) op (m13, 13); +*/ + +#ifdef MY_CPU_AMD64 + #define NUM_WAYS 8 + #define WOP_M1 WOP_8 +#else + #define NUM_WAYS 4 + #define WOP_M1 WOP_4 +#endif + +#define WOP(op) op (m0, 0) WOP_M1(op) + + +#define DECLARE_VAR(reg, ii) __m128i reg; +#define LOAD_data( reg, ii) reg = data[ii]; +#define STORE_data( reg, ii) data[ii] = reg; +#if (NUM_WAYS > 1) +#define XOR_data_M1(reg, ii) MM_XOR (reg, data[ii- 1]) +#endif + +#define AVX_DECLARE_VAR(reg, ii) __m256i reg; +#define AVX_LOAD_data( reg, ii) reg = ((const __m256i *)(const void *)data)[ii]; +#define AVX_STORE_data( reg, ii) ((__m256i *)(void *)data)[ii] = reg; +#define AVX_XOR_data_M1(reg, ii) AVX_XOR (reg, (((const __m256i *)(const void *)(data - 1))[ii])) + +#define MM_OP_key(op, reg) MM_OP(op, reg, key); + +#define AES_DEC( reg, ii) MM_OP_key (_mm_aesdec_si128, reg) +#define AES_DEC_LAST( reg, ii) MM_OP_key (_mm_aesdeclast_si128, reg) +#define AES_ENC( reg, ii) MM_OP_key (_mm_aesenc_si128, reg) +#define AES_ENC_LAST( reg, ii) MM_OP_key (_mm_aesenclast_si128, reg) +#define AES_XOR( reg, ii) MM_OP_key (_mm_xor_si128, reg) + + +#define AVX_AES_DEC( reg, ii) MM_OP_key (_mm256_aesdec_epi128, reg) +#define AVX_AES_DEC_LAST( reg, ii) MM_OP_key (_mm256_aesdeclast_epi128, reg) +#define AVX_AES_ENC( reg, ii) MM_OP_key (_mm256_aesenc_epi128, reg) +#define AVX_AES_ENC_LAST( reg, ii) MM_OP_key (_mm256_aesenclast_epi128, reg) +#define AVX_AES_XOR( reg, ii) MM_OP_key (_mm256_xor_si256, reg) -#define AES_DEC(n) AES_OP_W(_mm_aesdec_si128, n) -#define AES_DEC_LAST(n) AES_OP_W(_mm_aesdeclast_si128, n) -#define AES_ENC(n) AES_OP_W(_mm_aesenc_si128, n) -#define AES_ENC_LAST(n) AES_OP_W(_mm_aesenclast_si128, n) +#define CTR_START(reg, ii) MM_OP (_mm_add_epi64, ctr, one) reg = ctr; +#define CTR_END( reg, ii) MM_XOR (data[ii], reg) -void MY_FAST_CALL AesCbc_Decode_Intel(__m128i *p, __m128i *data, size_t numBlocks) +#define AVX_CTR_START(reg, ii) MM_OP (_mm256_add_epi64, ctr2, two) reg = _mm256_xor_si256(ctr2, key); +#define AVX_CTR_END( reg, ii) AVX_XOR (((__m256i *)(void *)data)[ii], reg) + +#define WOP_KEY(op, n) { \ + const __m128i key = w[n]; \ + WOP(op); } + +#define AVX_WOP_KEY(op, n) { \ + const __m256i key = w[n]; \ + WOP(op); } + + +#define WIDE_LOOP_START \ + dataEnd = data + numBlocks; \ + if (numBlocks >= NUM_WAYS) \ + { dataEnd -= NUM_WAYS; do { \ + + +#define WIDE_LOOP_END \ + data += NUM_WAYS; \ + } while (data <= dataEnd); \ + dataEnd += NUM_WAYS; } \ + + +#define SINGLE_LOOP \ + for (; data < dataEnd; data++) + + +#define NUM_AES_KEYS_MAX 15 + +#define WIDE_LOOP_START_AVX(OP) \ + dataEnd = data + numBlocks; \ + if (numBlocks >= NUM_WAYS * 2) \ + { __m256i keys[NUM_AES_KEYS_MAX]; \ + UInt32 ii; \ + OP \ + for (ii = 0; ii < numRounds; ii++) \ + keys[ii] = _mm256_broadcastsi128_si256(p[ii]); \ + dataEnd -= NUM_WAYS * 2; do { \ + + +#define WIDE_LOOP_END_AVX(OP) \ + data += NUM_WAYS * 2; \ + } while (data <= dataEnd); \ + dataEnd += NUM_WAYS * 2; \ + OP \ + _mm256_zeroupper(); \ + } \ + +/* MSVC for x86: If we don't call _mm256_zeroupper(), and -arch:IA32 is not specified, + MSVC still can insert vzeroupper instruction. */ + + +AES_FUNC_START2 (AesCbc_Decode_HW) { + __m128i *p = (__m128i *)(void *)ivAes; + __m128i *data = (__m128i *)(void *)data8; __m128i iv = *p; - for (; numBlocks >= NUM_WAYS; numBlocks -= NUM_WAYS, data += NUM_WAYS) + const __m128i *wStart = p + *(const UInt32 *)(p + 1) * 2 + 2 - 1; + const __m128i *dataEnd; + p += 2; + + WIDE_LOOP_START { - UInt32 numRounds2 = *(const UInt32 *)(p + 1); - const __m128i *w = p + numRounds2 * 2; - __m128i m0, m1, m2; + const __m128i *w = wStart; + + WOP (DECLARE_VAR) + WOP (LOAD_data) + WOP_KEY (AES_XOR, 1) + + do { - const __m128i t = w[2]; - m0 = _mm_xor_si128(t, data[0]); - m1 = _mm_xor_si128(t, data[1]); - m2 = _mm_xor_si128(t, data[2]); + WOP_KEY (AES_DEC, 0) + w--; } - numRounds2--; + while (w != p); + WOP_KEY (AES_DEC_LAST, 0) + + MM_XOR (m0, iv) + WOP_M1 (XOR_data_M1) + iv = data[NUM_WAYS - 1]; + WOP (STORE_data) + } + WIDE_LOOP_END + + SINGLE_LOOP + { + const __m128i *w = wStart - 1; + __m128i m = _mm_xor_si128 (w[2], *data); do { - AES_DEC(1) - AES_DEC(0) + MM_OP_m (_mm_aesdec_si128, w[1]) + MM_OP_m (_mm_aesdec_si128, w[0]) w -= 2; } - while (--numRounds2 != 0); - AES_DEC(1) - AES_DEC_LAST(0) + while (w != p); + MM_OP_m (_mm_aesdec_si128, w[1]) + MM_OP_m (_mm_aesdeclast_si128, w[0]) + + MM_XOR (m, iv) + iv = *data; + *data = m; + } + + p[-2] = iv; +} + + +AES_FUNC_START2 (AesCtr_Code_HW) +{ + __m128i *p = (__m128i *)(void *)ivAes; + __m128i *data = (__m128i *)(void *)data8; + __m128i ctr = *p; + UInt32 numRoundsMinus2 = *(const UInt32 *)(p + 1) * 2 - 1; + const __m128i *dataEnd; + __m128i one = _mm_cvtsi32_si128(1); + + p += 2; + + WIDE_LOOP_START + { + const __m128i *w = p; + UInt32 r = numRoundsMinus2; + WOP (DECLARE_VAR) + WOP (CTR_START) + WOP_KEY (AES_XOR, 0) + w += 1; + do + { + WOP_KEY (AES_ENC, 0) + w += 1; + } + while (--r); + WOP_KEY (AES_ENC_LAST, 0) + + WOP (CTR_END) + } + WIDE_LOOP_END + SINGLE_LOOP + { + UInt32 numRounds2 = *(const UInt32 *)(p - 2 + 1) - 1; + const __m128i *w = p; + __m128i m; + MM_OP (_mm_add_epi64, ctr, one) + m = _mm_xor_si128 (ctr, p[0]); + w += 1; + do { - __m128i t; - t = _mm_xor_si128(m0, iv); iv = data[0]; data[0] = t; - t = _mm_xor_si128(m1, iv); iv = data[1]; data[1] = t; - t = _mm_xor_si128(m2, iv); iv = data[2]; data[2] = t; + MM_OP_m (_mm_aesenc_si128, w[0]) + MM_OP_m (_mm_aesenc_si128, w[1]) + w += 2; } + while (--numRounds2); + MM_OP_m (_mm_aesenc_si128, w[0]) + MM_OP_m (_mm_aesenclast_si128, w[1]) + MM_XOR (*data, m) } - for (; numBlocks != 0; numBlocks--, data++) + + p[-2] = ctr; +} + + + +#ifdef USE_INTEL_VAES + +/* +GCC before 2013-Jun: + <immintrin.h>: + #ifdef __AVX__ + #include <avxintrin.h> + #endif +GCC after 2013-Jun: + <immintrin.h>: + #include <avxintrin.h> +CLANG 3.8+: +{ + <immintrin.h>: + #if !defined(_MSC_VER) || defined(__AVX__) + #include <avxintrin.h> + #endif + + if (the compiler is clang for Windows and if global arch is not set for __AVX__) + [ if (defined(_MSC_VER) && !defined(__AVX__)) ] + { + <immintrin.h> doesn't include <avxintrin.h> + and we have 2 ways to fix it: + 1) we can define required __AVX__ before <immintrin.h> + or + 2) we can include <avxintrin.h> after <immintrin.h> + } +} + +If we include <avxintrin.h> manually for GCC/CLANG, it's +required that <immintrin.h> must be included before <avxintrin.h>. +*/ + +/* +#if defined(__clang__) && defined(_MSC_VER) +#define __AVX__ +#define __AVX2__ +#define __VAES__ +#endif +*/ + +#include <immintrin.h> +#if defined(__clang__) && defined(_MSC_VER) + #if !defined(__AVX__) + #include <avxintrin.h> + #endif + #if !defined(__AVX2__) + #include <avx2intrin.h> + #endif + #if !defined(__VAES__) + #include <vaesintrin.h> + #endif +#endif // __clang__ && _MSC_VER + + +#define VAES_FUNC_START2(name) \ +AES_FUNC_START (name); \ +ATTRIB_VAES \ +AES_FUNC_START (name) + +VAES_FUNC_START2 (AesCbc_Decode_HW_256) +{ + __m128i *p = (__m128i *)(void *)ivAes; + __m128i *data = (__m128i *)(void *)data8; + __m128i iv = *p; + const __m128i *dataEnd; + UInt32 numRounds = *(const UInt32 *)(p + 1) * 2 + 1; + p += 2; + + WIDE_LOOP_START_AVX(;) + { + const __m256i *w = keys + numRounds - 2; + + WOP (AVX_DECLARE_VAR) + WOP (AVX_LOAD_data) + AVX_WOP_KEY (AVX_AES_XOR, 1) + + do + { + AVX_WOP_KEY (AVX_AES_DEC, 0) + w--; + } + while (w != keys); + AVX_WOP_KEY (AVX_AES_DEC_LAST, 0) + + AVX_XOR (m0, _mm256_setr_m128i(iv, data[0])) + WOP_M1 (AVX_XOR_data_M1) + iv = data[NUM_WAYS * 2 - 1]; + WOP (AVX_STORE_data) + } + WIDE_LOOP_END_AVX(;) + + SINGLE_LOOP { - UInt32 numRounds2 = *(const UInt32 *)(p + 1); - const __m128i *w = p + numRounds2 * 2; - __m128i m = _mm_xor_si128(w[2], *data); - numRounds2--; + const __m128i *w = p + *(const UInt32 *)(p + 1 - 2) * 2 + 1 - 3; + __m128i m = _mm_xor_si128 (w[2], *data); do { - m = _mm_aesdec_si128(m, w[1]); - m = _mm_aesdec_si128(m, w[0]); + MM_OP_m (_mm_aesdec_si128, w[1]) + MM_OP_m (_mm_aesdec_si128, w[0]) w -= 2; } - while (--numRounds2 != 0); - m = _mm_aesdec_si128(m, w[1]); - m = _mm_aesdeclast_si128(m, w[0]); + while (w != p); + MM_OP_m (_mm_aesdec_si128, w[1]) + MM_OP_m (_mm_aesdeclast_si128, w[0]) - m = _mm_xor_si128(m, iv); + MM_XOR (m, iv) iv = *data; *data = m; } - *p = iv; + + p[-2] = iv; } -void MY_FAST_CALL AesCtr_Code_Intel(__m128i *p, __m128i *data, size_t numBlocks) + +/* +SSE2: _mm_cvtsi32_si128 : movd +AVX: _mm256_setr_m128i : vinsertf128 +AVX2: _mm256_add_epi64 : vpaddq ymm, ymm, ymm + _mm256_extracti128_si256 : vextracti128 + _mm256_broadcastsi128_si256 : vbroadcasti128 +*/ + +#define AVX_CTR_LOOP_START \ + ctr2 = _mm256_setr_m128i(_mm_sub_epi64(ctr, one), ctr); \ + two = _mm256_setr_m128i(one, one); \ + two = _mm256_add_epi64(two, two); \ + +// two = _mm256_setr_epi64x(2, 0, 2, 0); + +#define AVX_CTR_LOOP_ENC \ + ctr = _mm256_extracti128_si256 (ctr2, 1); \ + +VAES_FUNC_START2 (AesCtr_Code_HW_256) { + __m128i *p = (__m128i *)(void *)ivAes; + __m128i *data = (__m128i *)(void *)data8; __m128i ctr = *p; - __m128i one; - one.m128i_u64[0] = 1; - one.m128i_u64[1] = 0; - for (; numBlocks >= NUM_WAYS; numBlocks -= NUM_WAYS, data += NUM_WAYS) + UInt32 numRounds = *(const UInt32 *)(p + 1) * 2 + 1; + const __m128i *dataEnd; + __m128i one = _mm_cvtsi32_si128(1); + __m256i ctr2, two; + p += 2; + + WIDE_LOOP_START_AVX (AVX_CTR_LOOP_START) { - UInt32 numRounds2 = *(const UInt32 *)(p + 1) - 1; - const __m128i *w = p; - __m128i m0, m1, m2; - { - const __m128i t = w[2]; - ctr = _mm_add_epi64(ctr, one); m0 = _mm_xor_si128(ctr, t); - ctr = _mm_add_epi64(ctr, one); m1 = _mm_xor_si128(ctr, t); - ctr = _mm_add_epi64(ctr, one); m2 = _mm_xor_si128(ctr, t); - } - w += 3; + const __m256i *w = keys; + UInt32 r = numRounds - 2; + WOP (AVX_DECLARE_VAR) + AVX_WOP_KEY (AVX_CTR_START, 0) + + w += 1; do { - AES_ENC(0) - AES_ENC(1) - w += 2; + AVX_WOP_KEY (AVX_AES_ENC, 0) + w += 1; } - while (--numRounds2 != 0); - AES_ENC(0) - AES_ENC_LAST(1) - data[0] = _mm_xor_si128(data[0], m0); - data[1] = _mm_xor_si128(data[1], m1); - data[2] = _mm_xor_si128(data[2], m2); + while (--r); + AVX_WOP_KEY (AVX_AES_ENC_LAST, 0) + + WOP (AVX_CTR_END) } - for (; numBlocks != 0; numBlocks--, data++) + WIDE_LOOP_END_AVX (AVX_CTR_LOOP_ENC) + + SINGLE_LOOP { - UInt32 numRounds2 = *(const UInt32 *)(p + 1) - 1; + UInt32 numRounds2 = *(const UInt32 *)(p - 2 + 1) - 1; const __m128i *w = p; __m128i m; - ctr = _mm_add_epi64(ctr, one); - m = _mm_xor_si128(ctr, p[2]); - w += 3; + MM_OP (_mm_add_epi64, ctr, one) + m = _mm_xor_si128 (ctr, p[0]); + w += 1; do { - m = _mm_aesenc_si128(m, w[0]); - m = _mm_aesenc_si128(m, w[1]); + MM_OP_m (_mm_aesenc_si128, w[0]) + MM_OP_m (_mm_aesenc_si128, w[1]) w += 2; } - while (--numRounds2 != 0); - m = _mm_aesenc_si128(m, w[0]); - m = _mm_aesenclast_si128(m, w[1]); - *data = _mm_xor_si128(*data, m); + while (--numRounds2); + MM_OP_m (_mm_aesenc_si128, w[0]) + MM_OP_m (_mm_aesenclast_si128, w[1]) + MM_XOR (*data, m) } - *p = ctr; + + p[-2] = ctr; } +#endif // USE_INTEL_VAES + +#else // USE_INTEL_AES + +/* no USE_INTEL_AES */ + +#pragma message("AES HW_SW stub was used") + +#define AES_TYPE_keys UInt32 +#define AES_TYPE_data Byte + +#define AES_FUNC_START(name) \ + void Z7_FASTCALL name(UInt32 *p, Byte *data, size_t numBlocks) \ + +#define AES_COMPAT_STUB(name) \ + AES_FUNC_START(name); \ + AES_FUNC_START(name ## _HW) \ + { name(p, data, numBlocks); } + +AES_COMPAT_STUB (AesCbc_Encode) +AES_COMPAT_STUB (AesCbc_Decode) +AES_COMPAT_STUB (AesCtr_Code) + +#endif // USE_INTEL_AES + + +#ifndef USE_INTEL_VAES + +#pragma message("VAES HW_SW stub was used") + +#define VAES_COMPAT_STUB(name) \ + void Z7_FASTCALL name ## _256(UInt32 *p, Byte *data, size_t numBlocks); \ + void Z7_FASTCALL name ## _256(UInt32 *p, Byte *data, size_t numBlocks) \ + { name((AES_TYPE_keys *)(void *)p, (AES_TYPE_data *)(void *)data, numBlocks); } + +VAES_COMPAT_STUB (AesCbc_Decode_HW) +VAES_COMPAT_STUB (AesCtr_Code_HW) + +#endif // ! USE_INTEL_VAES + + +#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE) + + #if defined(__clang__) + #if (__clang_major__ >= 8) // fix that check + #define USE_HW_AES + #endif + #elif defined(__GNUC__) + #if (__GNUC__ >= 6) // fix that check + #define USE_HW_AES + #endif + #elif defined(_MSC_VER) + #if _MSC_VER >= 1910 + #define USE_HW_AES + #endif + #endif + +#ifdef USE_HW_AES + +// #pragma message("=== AES HW === ") + +#if defined(__clang__) || defined(__GNUC__) + #ifdef MY_CPU_ARM64 + #define ATTRIB_AES __attribute__((__target__("+crypto"))) + #else + #define ATTRIB_AES __attribute__((__target__("fpu=crypto-neon-fp-armv8"))) + #endif +#else + // _MSC_VER + // for arm32 + #define _ARM_USE_NEW_NEON_INTRINSICS +#endif + +#ifndef ATTRIB_AES + #define ATTRIB_AES +#endif + +#if defined(_MSC_VER) && defined(MY_CPU_ARM64) +#include <arm64_neon.h> #else +#include <arm_neon.h> +#endif + +typedef uint8x16_t v128; -void MY_FAST_CALL AesCbc_Encode(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCbc_Decode(UInt32 *ivAes, Byte *data, size_t numBlocks); -void MY_FAST_CALL AesCtr_Code(UInt32 *ivAes, Byte *data, size_t numBlocks); +#define AES_FUNC_START(name) \ + void Z7_FASTCALL name(UInt32 *ivAes, Byte *data8, size_t numBlocks) + // void Z7_FASTCALL name(v128 *p, v128 *data, size_t numBlocks) -void MY_FAST_CALL AesCbc_Encode_Intel(UInt32 *p, Byte *data, size_t numBlocks) +#define AES_FUNC_START2(name) \ +AES_FUNC_START (name); \ +ATTRIB_AES \ +AES_FUNC_START (name) + +#define MM_OP(op, dest, src) dest = op(dest, src); +#define MM_OP_m(op, src) MM_OP(op, m, src) +#define MM_OP1_m(op) m = op(m); + +#define MM_XOR( dest, src) MM_OP(veorq_u8, dest, src) +#define MM_XOR_m( src) MM_XOR(m, src) + +#define AES_E_m(k) MM_OP_m (vaeseq_u8, k) +#define AES_E_MC_m(k) AES_E_m (k) MM_OP1_m(vaesmcq_u8) + + +AES_FUNC_START2 (AesCbc_Encode_HW) { - AesCbc_Encode(p, data, numBlocks); + v128 *p = (v128*)(void*)ivAes; + v128 *data = (v128*)(void*)data8; + v128 m = *p; + const v128 k0 = p[2]; + const v128 k1 = p[3]; + const v128 k2 = p[4]; + const v128 k3 = p[5]; + const v128 k4 = p[6]; + const v128 k5 = p[7]; + const v128 k6 = p[8]; + const v128 k7 = p[9]; + const v128 k8 = p[10]; + const v128 k9 = p[11]; + const UInt32 numRounds2 = *(const UInt32 *)(p + 1); + const v128 *w = p + ((size_t)numRounds2 * 2); + const v128 k_z1 = w[1]; + const v128 k_z0 = w[2]; + for (; numBlocks != 0; numBlocks--, data++) + { + MM_XOR_m (*data); + AES_E_MC_m (k0) + AES_E_MC_m (k1) + AES_E_MC_m (k2) + AES_E_MC_m (k3) + AES_E_MC_m (k4) + AES_E_MC_m (k5) + AES_E_MC_m (k6) + AES_E_MC_m (k7) + AES_E_MC_m (k8) + if (numRounds2 >= 6) + { + AES_E_MC_m (k9) + AES_E_MC_m (p[12]) + if (numRounds2 != 6) + { + AES_E_MC_m (p[13]) + AES_E_MC_m (p[14]) + } + } + AES_E_m (k_z1) + MM_XOR_m (k_z0); + *data = m; + } + *p = m; } -void MY_FAST_CALL AesCbc_Decode_Intel(UInt32 *p, Byte *data, size_t numBlocks) + +#define WOP_1(op) +#define WOP_2(op) WOP_1 (op) op (m1, 1) +#define WOP_3(op) WOP_2 (op) op (m2, 2) +#define WOP_4(op) WOP_3 (op) op (m3, 3) +#define WOP_5(op) WOP_4 (op) op (m4, 4) +#define WOP_6(op) WOP_5 (op) op (m5, 5) +#define WOP_7(op) WOP_6 (op) op (m6, 6) +#define WOP_8(op) WOP_7 (op) op (m7, 7) + + #define NUM_WAYS 8 + #define WOP_M1 WOP_8 + +#define WOP(op) op (m0, 0) WOP_M1(op) + +#define DECLARE_VAR(reg, ii) v128 reg; +#define LOAD_data( reg, ii) reg = data[ii]; +#define STORE_data( reg, ii) data[ii] = reg; +#if (NUM_WAYS > 1) +#define XOR_data_M1(reg, ii) MM_XOR (reg, data[ii- 1]) +#endif + +#define MM_OP_key(op, reg) MM_OP (op, reg, key) + +#define AES_D_m(k) MM_OP_m (vaesdq_u8, k) +#define AES_D_IMC_m(k) AES_D_m (k) MM_OP1_m (vaesimcq_u8) + +#define AES_XOR( reg, ii) MM_OP_key (veorq_u8, reg) +#define AES_D( reg, ii) MM_OP_key (vaesdq_u8, reg) +#define AES_E( reg, ii) MM_OP_key (vaeseq_u8, reg) + +#define AES_D_IMC( reg, ii) AES_D (reg, ii) reg = vaesimcq_u8(reg); +#define AES_E_MC( reg, ii) AES_E (reg, ii) reg = vaesmcq_u8(reg); + +#define CTR_START(reg, ii) MM_OP (vaddq_u64, ctr, one) reg = vreinterpretq_u8_u64(ctr); +#define CTR_END( reg, ii) MM_XOR (data[ii], reg) + +#define WOP_KEY(op, n) { \ + const v128 key = w[n]; \ + WOP(op) } + +#define WIDE_LOOP_START \ + dataEnd = data + numBlocks; \ + if (numBlocks >= NUM_WAYS) \ + { dataEnd -= NUM_WAYS; do { \ + +#define WIDE_LOOP_END \ + data += NUM_WAYS; \ + } while (data <= dataEnd); \ + dataEnd += NUM_WAYS; } \ + +#define SINGLE_LOOP \ + for (; data < dataEnd; data++) + + +AES_FUNC_START2 (AesCbc_Decode_HW) { - AesCbc_Decode(p, data, numBlocks); + v128 *p = (v128*)(void*)ivAes; + v128 *data = (v128*)(void*)data8; + v128 iv = *p; + const v128 *wStart = p + ((size_t)*(const UInt32 *)(p + 1)) * 2; + const v128 *dataEnd; + p += 2; + + WIDE_LOOP_START + { + const v128 *w = wStart; + WOP (DECLARE_VAR) + WOP (LOAD_data) + WOP_KEY (AES_D_IMC, 2) + do + { + WOP_KEY (AES_D_IMC, 1) + WOP_KEY (AES_D_IMC, 0) + w -= 2; + } + while (w != p); + WOP_KEY (AES_D, 1) + WOP_KEY (AES_XOR, 0) + MM_XOR (m0, iv); + WOP_M1 (XOR_data_M1) + iv = data[NUM_WAYS - 1]; + WOP (STORE_data) + } + WIDE_LOOP_END + + SINGLE_LOOP + { + const v128 *w = wStart; + v128 m = *data; + AES_D_IMC_m (w[2]) + do + { + AES_D_IMC_m (w[1]); + AES_D_IMC_m (w[0]); + w -= 2; + } + while (w != p); + AES_D_m (w[1]); + MM_XOR_m (w[0]); + MM_XOR_m (iv); + iv = *data; + *data = m; + } + + p[-2] = iv; } -void MY_FAST_CALL AesCtr_Code_Intel(UInt32 *p, Byte *data, size_t numBlocks) + +AES_FUNC_START2 (AesCtr_Code_HW) { - AesCtr_Code(p, data, numBlocks); + v128 *p = (v128*)(void*)ivAes; + v128 *data = (v128*)(void*)data8; + uint64x2_t ctr = vreinterpretq_u64_u8(*p); + const v128 *wEnd = p + ((size_t)*(const UInt32 *)(p + 1)) * 2; + const v128 *dataEnd; + uint64x2_t one = vdupq_n_u64(0); + one = vsetq_lane_u64(1, one, 0); + p += 2; + + WIDE_LOOP_START + { + const v128 *w = p; + WOP (DECLARE_VAR) + WOP (CTR_START) + do + { + WOP_KEY (AES_E_MC, 0) + WOP_KEY (AES_E_MC, 1) + w += 2; + } + while (w != wEnd); + WOP_KEY (AES_E_MC, 0) + WOP_KEY (AES_E, 1) + WOP_KEY (AES_XOR, 2) + WOP (CTR_END) + } + WIDE_LOOP_END + + SINGLE_LOOP + { + const v128 *w = p; + v128 m; + CTR_START (m, 0); + do + { + AES_E_MC_m (w[0]); + AES_E_MC_m (w[1]); + w += 2; + } + while (w != wEnd); + AES_E_MC_m (w[0]) + AES_E_m (w[1]) + MM_XOR_m (w[2]) + CTR_END (m, 0) + } + + p[-2] = vreinterpretq_u8_u64(ctr); } -#endif +#endif // USE_HW_AES + +#endif // MY_CPU_ARM_OR_ARM64 + +#undef NUM_WAYS +#undef WOP_M1 +#undef WOP +#undef DECLARE_VAR +#undef LOAD_data +#undef STORE_data +#undef USE_INTEL_AES +#undef USE_HW_AES diff --git a/3rdparty/lzma/C/Alloc.c b/3rdparty/lzma/C/Alloc.c index 3f6ac41918e..d841bf20a35 100644 --- a/3rdparty/lzma/C/Alloc.c +++ b/3rdparty/lzma/C/Alloc.c @@ -1,33 +1,182 @@ /* Alloc.c -- Memory allocation functions -2015-02-21 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" #ifdef _WIN32 -#include <windows.h> +#include "7zWindows.h" #endif #include <stdlib.h> #include "Alloc.h" -/* #define _SZ_ALLOC_DEBUG */ +#ifdef _WIN32 +#ifdef Z7_LARGE_PAGES +#if defined(__clang__) || defined(__GNUC__) +typedef void (*Z7_voidFunction)(void); +#define MY_CAST_FUNC (Z7_voidFunction) +#elif defined(_MSC_VER) && _MSC_VER > 1920 +#define MY_CAST_FUNC (void *) +// #pragma warning(disable : 4191) // 'type cast': unsafe conversion from 'FARPROC' to 'void (__cdecl *)()' +#else +#define MY_CAST_FUNC +#endif +#endif // Z7_LARGE_PAGES +#endif // _WIN32 + +// #define SZ_ALLOC_DEBUG +/* #define SZ_ALLOC_DEBUG */ -/* use _SZ_ALLOC_DEBUG to debug alloc/free operations */ -#ifdef _SZ_ALLOC_DEBUG +/* use SZ_ALLOC_DEBUG to debug alloc/free operations */ +#ifdef SZ_ALLOC_DEBUG + +#include <string.h> #include <stdio.h> -int g_allocCount = 0; -int g_allocCountMid = 0; -int g_allocCountBig = 0; +static int g_allocCount = 0; +#ifdef _WIN32 +static int g_allocCountMid = 0; +static int g_allocCountBig = 0; #endif + +#define CONVERT_INT_TO_STR(charType, tempSize) \ + char temp[tempSize]; unsigned i = 0; \ + while (val >= 10) { temp[i++] = (char)('0' + (unsigned)(val % 10)); val /= 10; } \ + *s++ = (charType)('0' + (unsigned)val); \ + while (i != 0) { i--; *s++ = temp[i]; } \ + *s = 0; + +static void ConvertUInt64ToString(UInt64 val, char *s) +{ + CONVERT_INT_TO_STR(char, 24) +} + +#define GET_HEX_CHAR(t) ((char)(((t < 10) ? ('0' + t) : ('A' + (t - 10))))) + +static void ConvertUInt64ToHex(UInt64 val, char *s) +{ + UInt64 v = val; + unsigned i; + for (i = 1;; i++) + { + v >>= 4; + if (v == 0) + break; + } + s[i] = 0; + do + { + unsigned t = (unsigned)(val & 0xF); + val >>= 4; + s[--i] = GET_HEX_CHAR(t); + } + while (i); +} + +#define DEBUG_OUT_STREAM stderr + +static void Print(const char *s) +{ + fputs(s, DEBUG_OUT_STREAM); +} + +static void PrintAligned(const char *s, size_t align) +{ + size_t len = strlen(s); + for(;;) + { + fputc(' ', DEBUG_OUT_STREAM); + if (len >= align) + break; + ++len; + } + Print(s); +} + +static void PrintLn(void) +{ + Print("\n"); +} + +static void PrintHex(UInt64 v, size_t align) +{ + char s[32]; + ConvertUInt64ToHex(v, s); + PrintAligned(s, align); +} + +static void PrintDec(int v, size_t align) +{ + char s[32]; + ConvertUInt64ToString((unsigned)v, s); + PrintAligned(s, align); +} + +static void PrintAddr(void *p) +{ + PrintHex((UInt64)(size_t)(ptrdiff_t)p, 12); +} + + +#define PRINT_REALLOC(name, cnt, size, ptr) { \ + Print(name " "); \ + if (!ptr) PrintDec(cnt++, 10); \ + PrintHex(size, 10); \ + PrintAddr(ptr); \ + PrintLn(); } + +#define PRINT_ALLOC(name, cnt, size, ptr) { \ + Print(name " "); \ + PrintDec(cnt++, 10); \ + PrintHex(size, 10); \ + PrintAddr(ptr); \ + PrintLn(); } + +#define PRINT_FREE(name, cnt, ptr) if (ptr) { \ + Print(name " "); \ + PrintDec(--cnt, 10); \ + PrintAddr(ptr); \ + PrintLn(); } + +#else + +#ifdef _WIN32 +#define PRINT_ALLOC(name, cnt, size, ptr) +#endif +#define PRINT_FREE(name, cnt, ptr) +#define Print(s) +#define PrintLn() +#define PrintHex(v, align) +#define PrintAddr(p) + +#endif + + +/* +by specification: + malloc(non_NULL, 0) : returns NULL or a unique pointer value that can later be successfully passed to free() + realloc(NULL, size) : the call is equivalent to malloc(size) + realloc(non_NULL, 0) : the call is equivalent to free(ptr) + +in main compilers: + malloc(0) : returns non_NULL + realloc(NULL, 0) : returns non_NULL + realloc(non_NULL, 0) : returns NULL +*/ + + void *MyAlloc(size_t size) { if (size == 0) - return 0; - #ifdef _SZ_ALLOC_DEBUG + return NULL; + // PRINT_ALLOC("Alloc ", g_allocCount, size, NULL) + #ifdef SZ_ALLOC_DEBUG { void *p = malloc(size); - fprintf(stderr, "\nAlloc %10d bytes, count = %10d, addr = %8X", size, g_allocCount++, (unsigned)p); + if (p) + { + PRINT_ALLOC("Alloc ", g_allocCount, size, p) + } return p; } #else @@ -37,100 +186,350 @@ void *MyAlloc(size_t size) void MyFree(void *address) { - #ifdef _SZ_ALLOC_DEBUG - if (address != 0) - fprintf(stderr, "\nFree; count = %10d, addr = %8X", --g_allocCount, (unsigned)address); - #endif + PRINT_FREE("Free ", g_allocCount, address) + free(address); } +void *MyRealloc(void *address, size_t size) +{ + if (size == 0) + { + MyFree(address); + return NULL; + } + // PRINT_REALLOC("Realloc ", g_allocCount, size, address) + #ifdef SZ_ALLOC_DEBUG + { + void *p = realloc(address, size); + if (p) + { + PRINT_REALLOC("Realloc ", g_allocCount, size, address) + } + return p; + } + #else + return realloc(address, size); + #endif +} + + #ifdef _WIN32 void *MidAlloc(size_t size) { if (size == 0) - return 0; - #ifdef _SZ_ALLOC_DEBUG - fprintf(stderr, "\nAlloc_Mid %10d bytes; count = %10d", size, g_allocCountMid++); + return NULL; + #ifdef SZ_ALLOC_DEBUG + { + void *p = VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE); + if (p) + { + PRINT_ALLOC("Alloc-Mid", g_allocCountMid, size, p) + } + return p; + } + #else + return VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE); #endif - return VirtualAlloc(0, size, MEM_COMMIT, PAGE_READWRITE); } void MidFree(void *address) { - #ifdef _SZ_ALLOC_DEBUG - if (address != 0) - fprintf(stderr, "\nFree_Mid; count = %10d", --g_allocCountMid); - #endif - if (address == 0) + PRINT_FREE("Free-Mid", g_allocCountMid, address) + + if (!address) return; VirtualFree(address, 0, MEM_RELEASE); } -#ifndef MEM_LARGE_PAGES -#undef _7ZIP_LARGE_PAGES +#ifdef Z7_LARGE_PAGES + +#ifdef MEM_LARGE_PAGES + #define MY__MEM_LARGE_PAGES MEM_LARGE_PAGES +#else + #define MY__MEM_LARGE_PAGES 0x20000000 #endif -#ifdef _7ZIP_LARGE_PAGES +extern +SIZE_T g_LargePageSize; SIZE_T g_LargePageSize = 0; -typedef SIZE_T (WINAPI *GetLargePageMinimumP)(); -#endif +typedef SIZE_T (WINAPI *Func_GetLargePageMinimum)(VOID); -void SetLargePageSize() +void SetLargePageSize(void) { - #ifdef _7ZIP_LARGE_PAGES - SIZE_T size = 0; - GetLargePageMinimumP largePageMinimum = (GetLargePageMinimumP) - GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "GetLargePageMinimum"); - if (largePageMinimum == 0) + #ifdef Z7_LARGE_PAGES + SIZE_T size; + const + Func_GetLargePageMinimum fn = + (Func_GetLargePageMinimum) MY_CAST_FUNC GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), + "GetLargePageMinimum"); + if (!fn) return; - size = largePageMinimum(); + size = fn(); if (size == 0 || (size & (size - 1)) != 0) return; g_LargePageSize = size; #endif } +#endif // Z7_LARGE_PAGES void *BigAlloc(size_t size) { if (size == 0) - return 0; - #ifdef _SZ_ALLOC_DEBUG - fprintf(stderr, "\nAlloc_Big %10d bytes; count = %10d", size, g_allocCountBig++); - #endif - - #ifdef _7ZIP_LARGE_PAGES - if (g_LargePageSize != 0 && g_LargePageSize <= (1 << 30) && size >= (1 << 18)) + return NULL; + + PRINT_ALLOC("Alloc-Big", g_allocCountBig, size, NULL) + + #ifdef Z7_LARGE_PAGES { - void *res = VirtualAlloc(0, (size + g_LargePageSize - 1) & (~(g_LargePageSize - 1)), - MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE); - if (res != 0) - return res; + SIZE_T ps = g_LargePageSize; + if (ps != 0 && ps <= (1 << 30) && size > (ps / 2)) + { + size_t size2; + ps--; + size2 = (size + ps) & ~ps; + if (size2 >= size) + { + void *p = VirtualAlloc(NULL, size2, MEM_COMMIT | MY__MEM_LARGE_PAGES, PAGE_READWRITE); + if (p) + { + PRINT_ALLOC("Alloc-BM ", g_allocCountMid, size2, p) + return p; + } + } + } } #endif - return VirtualAlloc(0, size, MEM_COMMIT, PAGE_READWRITE); + + return MidAlloc(size); } void BigFree(void *address) { - #ifdef _SZ_ALLOC_DEBUG - if (address != 0) - fprintf(stderr, "\nFree_Big; count = %10d", --g_allocCountBig); + PRINT_FREE("Free-Big", g_allocCountBig, address) + MidFree(address); +} + +#endif // _WIN32 + + +static void *SzAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p) return MyAlloc(size); } +static void SzFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p) MyFree(address); } +const ISzAlloc g_Alloc = { SzAlloc, SzFree }; + +#ifdef _WIN32 +static void *SzMidAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p) return MidAlloc(size); } +static void SzMidFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p) MidFree(address); } +static void *SzBigAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p) return BigAlloc(size); } +static void SzBigFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p) BigFree(address); } +const ISzAlloc g_MidAlloc = { SzMidAlloc, SzMidFree }; +const ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree }; +#endif + +/* + uintptr_t : <stdint.h> C99 (optional) + : unsupported in VS6 +*/ + +#ifdef _WIN32 + typedef UINT_PTR UIntPtr; +#else + /* + typedef uintptr_t UIntPtr; + */ + typedef ptrdiff_t UIntPtr; +#endif + + +#define ADJUST_ALLOC_SIZE 0 +/* +#define ADJUST_ALLOC_SIZE (sizeof(void *) - 1) +*/ +/* + Use (ADJUST_ALLOC_SIZE = (sizeof(void *) - 1)), if + MyAlloc() can return address that is NOT multiple of sizeof(void *). +*/ + + +/* +#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((char *)(p) - ((size_t)(UIntPtr)(p) & ((align) - 1)))) +*/ +#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((((UIntPtr)(p)) & ~((UIntPtr)(align) - 1)))) + + +#if !defined(_WIN32) && defined(_POSIX_C_SOURCE) && (_POSIX_C_SOURCE >= 200112L) + #define USE_posix_memalign +#endif + +#ifndef USE_posix_memalign +#define MY_ALIGN_PTR_UP_PLUS(p, align) MY_ALIGN_PTR_DOWN(((char *)(p) + (align) + ADJUST_ALLOC_SIZE), align) +#endif + +/* + This posix_memalign() is for test purposes only. + We also need special Free() function instead of free(), + if this posix_memalign() is used. +*/ + +/* +static int posix_memalign(void **ptr, size_t align, size_t size) +{ + size_t newSize = size + align; + void *p; + void *pAligned; + *ptr = NULL; + if (newSize < size) + return 12; // ENOMEM + p = MyAlloc(newSize); + if (!p) + return 12; // ENOMEM + pAligned = MY_ALIGN_PTR_UP_PLUS(p, align); + ((void **)pAligned)[-1] = p; + *ptr = pAligned; + return 0; +} +*/ + +/* + ALLOC_ALIGN_SIZE >= sizeof(void *) + ALLOC_ALIGN_SIZE >= cache_line_size +*/ + +#define ALLOC_ALIGN_SIZE ((size_t)1 << 7) + +static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size) +{ + #ifndef USE_posix_memalign + + void *p; + void *pAligned; + size_t newSize; + UNUSED_VAR(pp) + + /* also we can allocate additional dummy ALLOC_ALIGN_SIZE bytes after aligned + block to prevent cache line sharing with another allocated blocks */ + + newSize = size + ALLOC_ALIGN_SIZE * 1 + ADJUST_ALLOC_SIZE; + if (newSize < size) + return NULL; + + p = MyAlloc(newSize); + + if (!p) + return NULL; + pAligned = MY_ALIGN_PTR_UP_PLUS(p, ALLOC_ALIGN_SIZE); + + Print(" size="); PrintHex(size, 8); + Print(" a_size="); PrintHex(newSize, 8); + Print(" ptr="); PrintAddr(p); + Print(" a_ptr="); PrintAddr(pAligned); + PrintLn(); + + ((void **)pAligned)[-1] = p; + + return pAligned; + + #else + + void *p; + UNUSED_VAR(pp) + if (posix_memalign(&p, ALLOC_ALIGN_SIZE, size)) + return NULL; + + Print(" posix_memalign="); PrintAddr(p); + PrintLn(); + + return p; + + #endif +} + + +static void SzAlignedFree(ISzAllocPtr pp, void *address) +{ + UNUSED_VAR(pp) + #ifndef USE_posix_memalign + if (address) + MyFree(((void **)address)[-1]); + #else + free(address); #endif +} + + +const ISzAlloc g_AlignedAlloc = { SzAlignedAlloc, SzAlignedFree }; + + + +#define MY_ALIGN_PTR_DOWN_1(p) MY_ALIGN_PTR_DOWN(p, sizeof(void *)) + +/* we align ptr to support cases where CAlignOffsetAlloc::offset is not multiply of sizeof(void *) */ +#define REAL_BLOCK_PTR_VAR(p) ((void **)MY_ALIGN_PTR_DOWN_1(p))[-1] +/* +#define REAL_BLOCK_PTR_VAR(p) ((void **)(p))[-1] +*/ + +static void *AlignOffsetAlloc_Alloc(ISzAllocPtr pp, size_t size) +{ + const CAlignOffsetAlloc *p = Z7_CONTAINER_FROM_VTBL_CONST(pp, CAlignOffsetAlloc, vt); + void *adr; + void *pAligned; + size_t newSize; + size_t extra; + size_t alignSize = (size_t)1 << p->numAlignBits; + + if (alignSize < sizeof(void *)) + alignSize = sizeof(void *); - if (address == 0) - return; - VirtualFree(address, 0, MEM_RELEASE); + if (p->offset >= alignSize) + return NULL; + + /* also we can allocate additional dummy ALLOC_ALIGN_SIZE bytes after aligned + block to prevent cache line sharing with another allocated blocks */ + extra = p->offset & (sizeof(void *) - 1); + newSize = size + alignSize + extra + ADJUST_ALLOC_SIZE; + if (newSize < size) + return NULL; + + adr = ISzAlloc_Alloc(p->baseAlloc, newSize); + + if (!adr) + return NULL; + + pAligned = (char *)MY_ALIGN_PTR_DOWN((char *)adr + + alignSize - p->offset + extra + ADJUST_ALLOC_SIZE, alignSize) + p->offset; + + PrintLn(); + Print("- Aligned: "); + Print(" size="); PrintHex(size, 8); + Print(" a_size="); PrintHex(newSize, 8); + Print(" ptr="); PrintAddr(adr); + Print(" a_ptr="); PrintAddr(pAligned); + PrintLn(); + + REAL_BLOCK_PTR_VAR(pAligned) = adr; + + return pAligned; } -#endif +static void AlignOffsetAlloc_Free(ISzAllocPtr pp, void *address) +{ + if (address) + { + const CAlignOffsetAlloc *p = Z7_CONTAINER_FROM_VTBL_CONST(pp, CAlignOffsetAlloc, vt); + PrintLn(); + Print("- Aligned Free: "); + PrintLn(); + ISzAlloc_Free(p->baseAlloc, REAL_BLOCK_PTR_VAR(address)); + } +} -static void *SzAlloc(void *p, size_t size) { UNUSED_VAR(p); return MyAlloc(size); } -static void SzFree(void *p, void *address) { UNUSED_VAR(p); MyFree(address); } -ISzAlloc g_Alloc = { SzAlloc, SzFree }; -static void *SzBigAlloc(void *p, size_t size) { UNUSED_VAR(p); return BigAlloc(size); } -static void SzBigFree(void *p, void *address) { UNUSED_VAR(p); BigFree(address); } -ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree }; +void AlignOffsetAlloc_CreateVTable(CAlignOffsetAlloc *p) +{ + p->vt.Alloc = AlignOffsetAlloc_Alloc; + p->vt.Free = AlignOffsetAlloc_Free; +} diff --git a/3rdparty/lzma/C/Alloc.h b/3rdparty/lzma/C/Alloc.h index bac0d87e2a0..fac5b62fb9f 100644 --- a/3rdparty/lzma/C/Alloc.h +++ b/3rdparty/lzma/C/Alloc.h @@ -1,19 +1,32 @@ /* Alloc.h -- Memory allocation functions -2015-02-21 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __COMMON_ALLOC_H -#define __COMMON_ALLOC_H +#ifndef ZIP7_INC_ALLOC_H +#define ZIP7_INC_ALLOC_H #include "7zTypes.h" EXTERN_C_BEGIN +/* + MyFree(NULL) : is allowed, as free(NULL) + MyAlloc(0) : returns NULL : but malloc(0) is allowed to return NULL or non_NULL + MyRealloc(NULL, 0) : returns NULL : but realloc(NULL, 0) is allowed to return NULL or non_NULL +MyRealloc() is similar to realloc() for the following cases: + MyRealloc(non_NULL, 0) : returns NULL and always calls MyFree(ptr) + MyRealloc(NULL, non_ZERO) : returns NULL, if allocation failed + MyRealloc(non_NULL, non_ZERO) : returns NULL, if reallocation failed +*/ + void *MyAlloc(size_t size); void MyFree(void *address); +void *MyRealloc(void *address, size_t size); #ifdef _WIN32 -void SetLargePageSize(); +#ifdef Z7_LARGE_PAGES +void SetLargePageSize(void); +#endif void *MidAlloc(size_t size); void MidFree(void *address); @@ -29,8 +42,29 @@ void BigFree(void *address); #endif -extern ISzAlloc g_Alloc; -extern ISzAlloc g_BigAlloc; +extern const ISzAlloc g_Alloc; + +#ifdef _WIN32 +extern const ISzAlloc g_BigAlloc; +extern const ISzAlloc g_MidAlloc; +#else +#define g_BigAlloc g_AlignedAlloc +#define g_MidAlloc g_AlignedAlloc +#endif + +extern const ISzAlloc g_AlignedAlloc; + + +typedef struct +{ + ISzAlloc vt; + ISzAllocPtr baseAlloc; + unsigned numAlignBits; /* ((1 << numAlignBits) >= sizeof(void *)) */ + size_t offset; /* (offset == (k * sizeof(void *)) && offset < (1 << numAlignBits) */ +} CAlignOffsetAlloc; + +void AlignOffsetAlloc_CreateVTable(CAlignOffsetAlloc *p); + EXTERN_C_END diff --git a/3rdparty/lzma/C/Bcj2.c b/3rdparty/lzma/C/Bcj2.c index 3c88e44fc1f..7cb57ad62dc 100644 --- a/3rdparty/lzma/C/Bcj2.c +++ b/3rdparty/lzma/C/Bcj2.c @@ -1,29 +1,24 @@ /* Bcj2.c -- BCJ2 Decoder (Converter for x86 code) -2015-08-01 : Igor Pavlov : Public domain */ +2023-03-01 : Igor Pavlov : Public domain */ #include "Precomp.h" #include "Bcj2.h" #include "CpuArch.h" -#define CProb UInt16 - #define kTopValue ((UInt32)1 << 24) -#define kNumModelBits 11 -#define kBitModelTotal (1 << kNumModelBits) +#define kNumBitModelTotalBits 11 +#define kBitModelTotal (1 << kNumBitModelTotalBits) #define kNumMoveBits 5 -#define _IF_BIT_0 ttt = *prob; bound = (p->range >> kNumModelBits) * ttt; if (p->code < bound) -#define _UPDATE_0 p->range = bound; *prob = (CProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); -#define _UPDATE_1 p->range -= bound; p->code -= bound; *prob = (CProb)(ttt - (ttt >> kNumMoveBits)); +// UInt32 bcj2_stats[256 + 2][2]; void Bcj2Dec_Init(CBcj2Dec *p) { unsigned i; - - p->state = BCJ2_DEC_STATE_OK; + p->state = BCJ2_STREAM_RC; // BCJ2_DEC_STATE_OK; p->ip = 0; - p->temp[3] = 0; + p->temp = 0; p->range = 0; p->code = 0; for (i = 0; i < sizeof(p->probs) / sizeof(p->probs[0]); i++) @@ -32,216 +27,248 @@ void Bcj2Dec_Init(CBcj2Dec *p) SRes Bcj2Dec_Decode(CBcj2Dec *p) { + UInt32 v = p->temp; + // const Byte *src; if (p->range <= 5) { - p->state = BCJ2_DEC_STATE_OK; + UInt32 code = p->code; + p->state = BCJ2_DEC_STATE_ERROR; /* for case if we return SZ_ERROR_DATA; */ for (; p->range != 5; p->range++) { - if (p->range == 1 && p->code != 0) + if (p->range == 1 && code != 0) return SZ_ERROR_DATA; - if (p->bufs[BCJ2_STREAM_RC] == p->lims[BCJ2_STREAM_RC]) { p->state = BCJ2_STREAM_RC; return SZ_OK; } - - p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++; + code = (code << 8) | *(p->bufs[BCJ2_STREAM_RC])++; + p->code = code; } - - if (p->code == 0xFFFFFFFF) + if (code == 0xffffffff) return SZ_ERROR_DATA; - - p->range = 0xFFFFFFFF; + p->range = 0xffffffff; } - else if (p->state >= BCJ2_DEC_STATE_ORIG_0) + // else { - while (p->state <= BCJ2_DEC_STATE_ORIG_3) + unsigned state = p->state; + // we check BCJ2_IS_32BIT_STREAM() here instead of check in the main loop + if (BCJ2_IS_32BIT_STREAM(state)) { - Byte *dest = p->dest; - if (dest == p->destLim) + const Byte *cur = p->bufs[state]; + if (cur == p->lims[state]) return SZ_OK; - *dest = p->temp[p->state++ - BCJ2_DEC_STATE_ORIG_0]; - p->dest = dest + 1; + p->bufs[state] = cur + 4; + { + const UInt32 ip = p->ip + 4; + v = GetBe32a(cur) - ip; + p->ip = ip; + } + state = BCJ2_DEC_STATE_ORIG_0; } - } - - /* - if (BCJ2_IS_32BIT_STREAM(p->state)) - { - const Byte *cur = p->bufs[p->state]; - if (cur == p->lims[p->state]) - return SZ_OK; - p->bufs[p->state] = cur + 4; - + if ((unsigned)(state - BCJ2_DEC_STATE_ORIG_0) < 4) { - UInt32 val; - Byte *dest; - SizeT rem; - - p->ip += 4; - val = GetBe32(cur) - p->ip; - dest = p->dest; - rem = p->destLim - dest; - if (rem < 4) + Byte *dest = p->dest; + for (;;) { - SizeT i; - SetUi32(p->temp, val); - for (i = 0; i < rem; i++) - dest[i] = p->temp[i]; - p->dest = dest + rem; - p->state = BCJ2_DEC_STATE_ORIG_0 + (unsigned)rem; - return SZ_OK; + if (dest == p->destLim) + { + p->state = state; + p->temp = v; + return SZ_OK; + } + *dest++ = (Byte)v; + p->dest = dest; + if (++state == BCJ2_DEC_STATE_ORIG_3 + 1) + break; + v >>= 8; } - SetUi32(dest, val); - p->temp[3] = (Byte)(val >> 24); - p->dest = dest + 4; - p->state = BCJ2_DEC_STATE_OK; } } - */ + // src = p->bufs[BCJ2_STREAM_MAIN]; for (;;) { + /* if (BCJ2_IS_32BIT_STREAM(p->state)) p->state = BCJ2_DEC_STATE_OK; else + */ { if (p->range < kTopValue) { if (p->bufs[BCJ2_STREAM_RC] == p->lims[BCJ2_STREAM_RC]) { p->state = BCJ2_STREAM_RC; + p->temp = v; return SZ_OK; } p->range <<= 8; p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++; } - { const Byte *src = p->bufs[BCJ2_STREAM_MAIN]; const Byte *srcLim; - Byte *dest; - SizeT num = p->lims[BCJ2_STREAM_MAIN] - src; - - if (num == 0) + Byte *dest = p->dest; { - p->state = BCJ2_STREAM_MAIN; - return SZ_OK; + const SizeT rem = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - src); + SizeT num = (SizeT)(p->destLim - dest); + if (num >= rem) + num = rem; + #define NUM_ITERS 4 + #if (NUM_ITERS & (NUM_ITERS - 1)) == 0 + num &= ~((SizeT)NUM_ITERS - 1); // if (NUM_ITERS == (1 << x)) + #else + num -= num % NUM_ITERS; // if (NUM_ITERS != (1 << x)) + #endif + srcLim = src + num; } - - dest = p->dest; - if (num > (SizeT)(p->destLim - dest)) + + #define NUM_SHIFT_BITS 24 + #define ONE_ITER(indx) { \ + const unsigned b = src[indx]; \ + *dest++ = (Byte)b; \ + v = (v << NUM_SHIFT_BITS) | b; \ + if (((b + (0x100 - 0xe8)) & 0xfe) == 0) break; \ + if (((v - (((UInt32)0x0f << (NUM_SHIFT_BITS)) + 0x80)) & \ + ((((UInt32)1 << (4 + NUM_SHIFT_BITS)) - 0x1) << 4)) == 0) break; \ + /* ++dest */; /* v = b; */ } + + if (src != srcLim) + for (;;) { - num = p->destLim - dest; - if (num == 0) - { - p->state = BCJ2_DEC_STATE_ORIG; - return SZ_OK; - } + /* The dependency chain of 2-cycle for (v) calculation is not big problem here. + But we can remove dependency chain with v = b in the end of loop. */ + ONE_ITER(0) + #if (NUM_ITERS > 1) + ONE_ITER(1) + #if (NUM_ITERS > 2) + ONE_ITER(2) + #if (NUM_ITERS > 3) + ONE_ITER(3) + #if (NUM_ITERS > 4) + ONE_ITER(4) + #if (NUM_ITERS > 5) + ONE_ITER(5) + #if (NUM_ITERS > 6) + ONE_ITER(6) + #if (NUM_ITERS > 7) + ONE_ITER(7) + #endif + #endif + #endif + #endif + #endif + #endif + #endif + + src += NUM_ITERS; + if (src == srcLim) + break; } - - srcLim = src + num; - if (p->temp[3] == 0x0F && (src[0] & 0xF0) == 0x80) - *dest = src[0]; - else for (;;) + if (src == srcLim) + #if (NUM_ITERS > 1) + for (;;) + #endif { - Byte b = *src; - *dest = b; - if (b != 0x0F) + #if (NUM_ITERS > 1) + if (src == p->lims[BCJ2_STREAM_MAIN] || dest == p->destLim) + #endif { - if ((b & 0xFE) == 0xE8) - break; - dest++; - if (++src != srcLim) - continue; - break; + const SizeT num = (SizeT)(src - p->bufs[BCJ2_STREAM_MAIN]); + p->bufs[BCJ2_STREAM_MAIN] = src; + p->dest = dest; + p->ip += (UInt32)num; + /* state BCJ2_STREAM_MAIN has more priority than BCJ2_STATE_ORIG */ + p->state = + src == p->lims[BCJ2_STREAM_MAIN] ? + (unsigned)BCJ2_STREAM_MAIN : + (unsigned)BCJ2_DEC_STATE_ORIG; + p->temp = v; + return SZ_OK; } - dest++; - if (++src == srcLim) - break; - if ((*src & 0xF0) != 0x80) - continue; - *dest = *src; - break; + #if (NUM_ITERS > 1) + ONE_ITER(0) + src++; + #endif } - - num = src - p->bufs[BCJ2_STREAM_MAIN]; - - if (src == srcLim) + { - p->temp[3] = src[-1]; - p->bufs[BCJ2_STREAM_MAIN] = src; + const SizeT num = (SizeT)(dest - p->dest); + p->dest = dest; // p->dest += num; + p->bufs[BCJ2_STREAM_MAIN] += num; // = src; p->ip += (UInt32)num; - p->dest += num; - p->state = - p->bufs[BCJ2_STREAM_MAIN] == - p->lims[BCJ2_STREAM_MAIN] ? - (unsigned)BCJ2_STREAM_MAIN : - (unsigned)BCJ2_DEC_STATE_ORIG; - return SZ_OK; } - { UInt32 bound, ttt; - CProb *prob; - Byte b = src[0]; - Byte prev = (Byte)(num == 0 ? p->temp[3] : src[-1]); - - p->temp[3] = b; - p->bufs[BCJ2_STREAM_MAIN] = src + 1; - num++; - p->ip += (UInt32)num; - p->dest += num; - - prob = p->probs + (unsigned)(b == 0xE8 ? 2 + (unsigned)prev : (b == 0xE9 ? 1 : 0)); - - _IF_BIT_0 + CBcj2Prob *prob; // unsigned index; + /* + prob = p->probs + (unsigned)((Byte)v == 0xe8 ? + 2 + (Byte)(v >> 8) : + ((v >> 5) & 1)); // ((Byte)v < 0xe8 ? 0 : 1)); + */ { - _UPDATE_0 + const unsigned c = ((v + 0x17) >> 6) & 1; + prob = p->probs + (unsigned) + (((0 - c) & (Byte)(v >> NUM_SHIFT_BITS)) + c + ((v >> 5) & 1)); + // (Byte) + // 8x->0 : e9->1 : xxe8->xx+2 + // 8x->0x100 : e9->0x101 : xxe8->xx + // (((0x100 - (e & ~v)) & (0x100 | (v >> 8))) + (e & v)); + // (((0x101 + (~e | v)) & (0x100 | (v >> 8))) + (e & v)); + } + ttt = *prob; + bound = (p->range >> kNumBitModelTotalBits) * ttt; + if (p->code < bound) + { + // bcj2_stats[prob - p->probs][0]++; + p->range = bound; + *prob = (CBcj2Prob)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); continue; } - _UPDATE_1 - + { + // bcj2_stats[prob - p->probs][1]++; + p->range -= bound; + p->code -= bound; + *prob = (CBcj2Prob)(ttt - (ttt >> kNumMoveBits)); + } } } } - { - UInt32 val; - unsigned cj = (p->temp[3] == 0xE8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP; + /* (v == 0xe8 ? 0 : 1) uses setcc instruction with additional zero register usage in x64 MSVC. */ + // const unsigned cj = ((Byte)v == 0xe8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP; + const unsigned cj = (((v + 0x57) >> 6) & 1) + BCJ2_STREAM_CALL; const Byte *cur = p->bufs[cj]; Byte *dest; SizeT rem; - if (cur == p->lims[cj]) { p->state = cj; break; } - - val = GetBe32(cur); + v = GetBe32a(cur); p->bufs[cj] = cur + 4; - - p->ip += 4; - val -= p->ip; + { + const UInt32 ip = p->ip + 4; + v -= ip; + p->ip = ip; + } dest = p->dest; - rem = p->destLim - dest; - + rem = (SizeT)(p->destLim - dest); if (rem < 4) { - SizeT i; - SetUi32(p->temp, val); - for (i = 0; i < rem; i++) - dest[i] = p->temp[i]; + if ((unsigned)rem > 0) { dest[0] = (Byte)v; v >>= 8; + if ((unsigned)rem > 1) { dest[1] = (Byte)v; v >>= 8; + if ((unsigned)rem > 2) { dest[2] = (Byte)v; v >>= 8; }}} + p->temp = v; p->dest = dest + rem; p->state = BCJ2_DEC_STATE_ORIG_0 + (unsigned)rem; break; } - - SetUi32(dest, val); - p->temp[3] = (Byte)(val >> 24); + SetUi32(dest, v) + v >>= 24; p->dest = dest + 4; } } @@ -251,6 +278,13 @@ SRes Bcj2Dec_Decode(CBcj2Dec *p) p->range <<= 8; p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++; } - return SZ_OK; } + +#undef NUM_ITERS +#undef ONE_ITER +#undef NUM_SHIFT_BITS +#undef kTopValue +#undef kNumBitModelTotalBits +#undef kBitModelTotal +#undef kNumMoveBits diff --git a/3rdparty/lzma/C/Bcj2.h b/3rdparty/lzma/C/Bcj2.h index 8824080acfb..4575545b62c 100644 --- a/3rdparty/lzma/C/Bcj2.h +++ b/3rdparty/lzma/C/Bcj2.h @@ -1,8 +1,8 @@ -/* Bcj2.h -- BCJ2 Converter for x86 code -2014-11-10 : Igor Pavlov : Public domain */ +/* Bcj2.h -- BCJ2 converter for x86 code (Branch CALL/JUMP variant2) +2023-03-02 : Igor Pavlov : Public domain */ -#ifndef __BCJ2_H -#define __BCJ2_H +#ifndef ZIP7_INC_BCJ2_H +#define ZIP7_INC_BCJ2_H #include "7zTypes.h" @@ -26,37 +26,68 @@ enum BCJ2_DEC_STATE_ORIG_3, BCJ2_DEC_STATE_ORIG, - BCJ2_DEC_STATE_OK + BCJ2_DEC_STATE_ERROR /* after detected data error */ }; enum { BCJ2_ENC_STATE_ORIG = BCJ2_NUM_STREAMS, - BCJ2_ENC_STATE_OK + BCJ2_ENC_STATE_FINISHED /* it's state after fully encoded stream */ }; -#define BCJ2_IS_32BIT_STREAM(s) ((s) == BCJ2_STREAM_CALL || (s) == BCJ2_STREAM_JUMP) +/* #define BCJ2_IS_32BIT_STREAM(s) ((s) == BCJ2_STREAM_CALL || (s) == BCJ2_STREAM_JUMP) */ +#define BCJ2_IS_32BIT_STREAM(s) ((unsigned)((unsigned)(s) - (unsigned)BCJ2_STREAM_CALL) < 2) /* CBcj2Dec / CBcj2Enc bufs sizes: BUF_SIZE(n) = lims[n] - bufs[n] -bufs sizes for BCJ2_STREAM_CALL and BCJ2_STREAM_JUMP must be mutliply of 4: +bufs sizes for BCJ2_STREAM_CALL and BCJ2_STREAM_JUMP must be multiply of 4: (BUF_SIZE(BCJ2_STREAM_CALL) & 3) == 0 (BUF_SIZE(BCJ2_STREAM_JUMP) & 3) == 0 */ +// typedef UInt32 CBcj2Prob; +typedef UInt16 CBcj2Prob; + +/* +BCJ2 encoder / decoder internal requirements: + - If last bytes of stream contain marker (e8/e8/0f8x), then + there is also encoded symbol (0 : no conversion) in RC stream. + - One case of overlapped instructions is supported, + if last byte of converted instruction is (0f) and next byte is (8x): + marker [xx xx xx 0f] 8x + then the pair (0f 8x) is treated as marker. +*/ + +/* ---------- BCJ2 Decoder ---------- */ + /* CBcj2Dec: -dest is allowed to overlap with bufs[BCJ2_STREAM_MAIN], with the following conditions: +(dest) is allowed to overlap with bufs[BCJ2_STREAM_MAIN], with the following conditions: bufs[BCJ2_STREAM_MAIN] >= dest && - bufs[BCJ2_STREAM_MAIN] - dest >= tempReserv + + bufs[BCJ2_STREAM_MAIN] - dest >= BUF_SIZE(BCJ2_STREAM_CALL) + BUF_SIZE(BCJ2_STREAM_JUMP) - tempReserv = 0 : for first call of Bcj2Dec_Decode - tempReserv = 4 : for any other calls of Bcj2Dec_Decode - overlap with offset = 1 is not allowed + reserve = bufs[BCJ2_STREAM_MAIN] - dest - + ( BUF_SIZE(BCJ2_STREAM_CALL) + + BUF_SIZE(BCJ2_STREAM_JUMP) ) + and additional conditions: + if (it's first call of Bcj2Dec_Decode() after Bcj2Dec_Init()) + { + (reserve != 1) : if (ver < v23.00) + } + else // if there are more than one calls of Bcj2Dec_Decode() after Bcj2Dec_Init()) + { + (reserve >= 6) : if (ver < v23.00) + (reserve >= 4) : if (ver >= v23.00) + We need that (reserve) because after first call of Bcj2Dec_Decode(), + CBcj2Dec::temp can contain up to 4 bytes for writing to (dest). + } + (reserve == 0) is allowed, if we decode full stream via single call of Bcj2Dec_Decode(). + (reserve == 0) also is allowed in case of multi-call, if we use fixed buffers, + and (reserve) is calculated from full (final) sizes of all streams before first call. */ typedef struct @@ -68,21 +99,65 @@ typedef struct unsigned state; /* BCJ2_STREAM_MAIN has more priority than BCJ2_STATE_ORIG */ - UInt32 ip; - Byte temp[4]; + UInt32 ip; /* property of starting base for decoding */ + UInt32 temp; /* Byte temp[4]; */ UInt32 range; UInt32 code; - UInt16 probs[2 + 256]; + CBcj2Prob probs[2 + 256]; } CBcj2Dec; + +/* Note: + Bcj2Dec_Init() sets (CBcj2Dec::ip = 0) + if (ip != 0) property is required, the caller must set CBcj2Dec::ip after Bcj2Dec_Init() +*/ void Bcj2Dec_Init(CBcj2Dec *p); -/* Returns: SZ_OK or SZ_ERROR_DATA */ + +/* Bcj2Dec_Decode(): + returns: + SZ_OK + SZ_ERROR_DATA : if data in 5 starting bytes of BCJ2_STREAM_RC stream are not correct +*/ SRes Bcj2Dec_Decode(CBcj2Dec *p); -#define Bcj2Dec_IsFinished(_p_) ((_p_)->code == 0) +/* To check that decoding was finished you can compare + sizes of processed streams with sizes known from another sources. + You must do at least one mandatory check from the two following options: + - the check for size of processed output (ORIG) stream. + - the check for size of processed input (MAIN) stream. + additional optional checks: + - the checks for processed sizes of all input streams (MAIN, CALL, JUMP, RC) + - the checks Bcj2Dec_IsMaybeFinished*() + also before actual decoding you can check that the + following condition is met for stream sizes: + ( size(ORIG) == size(MAIN) + size(CALL) + size(JUMP) ) +*/ +/* (state == BCJ2_STREAM_MAIN) means that decoder is ready for + additional input data in BCJ2_STREAM_MAIN stream. + Note that (state == BCJ2_STREAM_MAIN) is allowed for non-finished decoding. +*/ +#define Bcj2Dec_IsMaybeFinished_state_MAIN(_p_) ((_p_)->state == BCJ2_STREAM_MAIN) +/* if the stream decoding was finished correctly, then range decoder + part of CBcj2Dec also was finished, and then (CBcj2Dec::code == 0). + Note that (CBcj2Dec::code == 0) is allowed for non-finished decoding. +*/ +#define Bcj2Dec_IsMaybeFinished_code(_p_) ((_p_)->code == 0) + +/* use Bcj2Dec_IsMaybeFinished() only as additional check + after at least one mandatory check from the two following options: + - the check for size of processed output (ORIG) stream. + - the check for size of processed input (MAIN) stream. +*/ +#define Bcj2Dec_IsMaybeFinished(_p_) ( \ + Bcj2Dec_IsMaybeFinished_state_MAIN(_p_) && \ + Bcj2Dec_IsMaybeFinished_code(_p_)) + + + +/* ---------- BCJ2 Encoder ---------- */ typedef enum { @@ -91,6 +166,91 @@ typedef enum BCJ2_ENC_FINISH_MODE_END_STREAM } EBcj2Enc_FinishMode; +/* + BCJ2_ENC_FINISH_MODE_CONTINUE: + process non finished encoding. + It notifies the encoder that additional further calls + can provide more input data (src) than provided by current call. + In that case the CBcj2Enc encoder still can move (src) pointer + up to (srcLim), but CBcj2Enc encoder can store some of the last + processed bytes (up to 4 bytes) from src to internal CBcj2Enc::temp[] buffer. + at return: + (CBcj2Enc::src will point to position that includes + processed data and data copied to (temp[]) buffer) + That data from (temp[]) buffer will be used in further calls. + + BCJ2_ENC_FINISH_MODE_END_BLOCK: + finish encoding of current block (ended at srcLim) without RC flushing. + at return: if (CBcj2Enc::state == BCJ2_ENC_STATE_ORIG) && + CBcj2Enc::src == CBcj2Enc::srcLim) + : it shows that block encoding was finished. And the encoder is + ready for new (src) data or for stream finish operation. + finished block means + { + CBcj2Enc has completed block encoding up to (srcLim). + (1 + 4 bytes) or (2 + 4 bytes) CALL/JUMP cortages will + not cross block boundary at (srcLim). + temporary CBcj2Enc buffer for (ORIG) src data is empty. + 3 output uncompressed streams (MAIN, CALL, JUMP) were flushed. + RC stream was not flushed. And RC stream will cross block boundary. + } + Note: some possible implementation of BCJ2 encoder could + write branch marker (e8/e8/0f8x) in one call of Bcj2Enc_Encode(), + and it could calculate symbol for RC in another call of Bcj2Enc_Encode(). + BCJ2 encoder uses ip/fileIp/fileSize/relatLimit values to calculate RC symbol. + And these CBcj2Enc variables can have different values in different Bcj2Enc_Encode() calls. + So caller must finish each block with BCJ2_ENC_FINISH_MODE_END_BLOCK + to ensure that RC symbol is calculated and written in proper block. + + BCJ2_ENC_FINISH_MODE_END_STREAM + finish encoding of stream (ended at srcLim) fully including RC flushing. + at return: if (CBcj2Enc::state == BCJ2_ENC_STATE_FINISHED) + : it shows that stream encoding was finished fully, + and all output streams were flushed fully. + also Bcj2Enc_IsFinished() can be called. +*/ + + +/* + 32-bit relative offset in JUMP/CALL commands is + - (mod 4 GiB) for 32-bit x86 code + - signed Int32 for 64-bit x86-64 code + BCJ2 encoder also does internal relative to absolute address conversions. + And there are 2 possible ways to do it: + before v23: we used 32-bit variables and (mod 4 GiB) conversion + since v23: we use 64-bit variables and (signed Int32 offset) conversion. + The absolute address condition for conversion in v23: + ((UInt64)((Int64)ip64 - (Int64)fileIp64 + 5 + (Int32)offset) < (UInt64)fileSize64) + note that if (fileSize64 > 2 GiB). there is difference between + old (mod 4 GiB) way (v22) and new (signed Int32 offset) way (v23). + And new (v23) way is more suitable to encode 64-bit x86-64 code for (fileSize64 > 2 GiB) cases. +*/ + +/* +// for old (v22) way for conversion: +typedef UInt32 CBcj2Enc_ip_unsigned; +typedef Int32 CBcj2Enc_ip_signed; +#define BCJ2_ENC_FileSize_MAX ((UInt32)1 << 31) +*/ +typedef UInt64 CBcj2Enc_ip_unsigned; +typedef Int64 CBcj2Enc_ip_signed; + +/* maximum size of file that can be used for conversion condition */ +#define BCJ2_ENC_FileSize_MAX ((CBcj2Enc_ip_unsigned)0 - 2) + +/* default value of fileSize64_minus1 variable that means + that absolute address limitation will not be used */ +#define BCJ2_ENC_FileSizeField_UNLIMITED ((CBcj2Enc_ip_unsigned)0 - 1) + +/* calculate value that later can be set to CBcj2Enc::fileSize64_minus1 */ +#define BCJ2_ENC_GET_FileSizeField_VAL_FROM_FileSize(fileSize) \ + ((CBcj2Enc_ip_unsigned)(fileSize) - 1) + +/* set CBcj2Enc::fileSize64_minus1 variable from size of file */ +#define Bcj2Enc_SET_FileSize(p, fileSize) \ + (p)->fileSize64_minus1 = BCJ2_ENC_GET_FileSizeField_VAL_FROM_FileSize(fileSize); + + typedef struct { Byte *bufs[BCJ2_NUM_STREAMS]; @@ -101,45 +261,71 @@ typedef struct unsigned state; EBcj2Enc_FinishMode finishMode; - Byte prevByte; + Byte context; + Byte flushRem; + Byte isFlushState; Byte cache; UInt32 range; UInt64 low; UInt64 cacheSize; + + // UInt32 context; // for marker version, it can include marker flag. - UInt32 ip; - - /* 32-bit ralative offset in JUMP/CALL commands is - - (mod 4 GB) in 32-bit mode - - signed Int32 in 64-bit mode - We use (mod 4 GB) check for fileSize. - Use fileSize up to 2 GB, if you want to support 32-bit and 64-bit code conversion. */ - UInt32 fileIp; - UInt32 fileSize; /* (fileSize <= ((UInt32)1 << 31)), 0 means no_limit */ - UInt32 relatLimit; /* (relatLimit <= ((UInt32)1 << 31)), 0 means desable_conversion */ + /* (ip64) and (fileIp64) correspond to virtual source stream position + that doesn't include data in temp[] */ + CBcj2Enc_ip_unsigned ip64; /* current (ip) position */ + CBcj2Enc_ip_unsigned fileIp64; /* start (ip) position of current file */ + CBcj2Enc_ip_unsigned fileSize64_minus1; /* size of current file (for conversion limitation) */ + UInt32 relatLimit; /* (relatLimit <= ((UInt32)1 << 31)) : 0 means disable_conversion */ + // UInt32 relatExcludeBits; UInt32 tempTarget; - unsigned tempPos; - Byte temp[4 * 2]; - - unsigned flushPos; - - UInt16 probs[2 + 256]; + unsigned tempPos; /* the number of bytes that were copied to temp[] buffer + (tempPos <= 4) outside of Bcj2Enc_Encode() */ + // Byte temp[4]; // for marker version + Byte temp[8]; + CBcj2Prob probs[2 + 256]; } CBcj2Enc; void Bcj2Enc_Init(CBcj2Enc *p); -void Bcj2Enc_Encode(CBcj2Enc *p); -#define Bcj2Enc_Get_InputData_Size(p) ((SizeT)((p)->srcLim - (p)->src) + (p)->tempPos) -#define Bcj2Enc_IsFinished(p) ((p)->flushPos == 5) +/* +Bcj2Enc_Encode(): at exit: + p->State < BCJ2_NUM_STREAMS : we need more buffer space for output stream + (bufs[p->State] == lims[p->State]) + p->State == BCJ2_ENC_STATE_ORIG : we need more data in input src stream + (src == srcLim) + p->State == BCJ2_ENC_STATE_FINISHED : after fully encoded stream +*/ +void Bcj2Enc_Encode(CBcj2Enc *p); -#define BCJ2_RELAT_LIMIT_NUM_BITS 26 -#define BCJ2_RELAT_LIMIT ((UInt32)1 << BCJ2_RELAT_LIMIT_NUM_BITS) +/* Bcj2Enc encoder can look ahead for up 4 bytes of source stream. + CBcj2Enc::tempPos : is the number of bytes that were copied from input stream to temp[] buffer. + (CBcj2Enc::src) after Bcj2Enc_Encode() is starting position after + fully processed data and after data copied to temp buffer. + So if the caller needs to get real number of fully processed input + bytes (without look ahead data in temp buffer), + the caller must subtruct (CBcj2Enc::tempPos) value from processed size + value that is calculated based on current (CBcj2Enc::src): + cur_processed_pos = Calc_Big_Processed_Pos(enc.src)) - + Bcj2Enc_Get_AvailInputSize_in_Temp(&enc); +*/ +/* get the size of input data that was stored in temp[] buffer: */ +#define Bcj2Enc_Get_AvailInputSize_in_Temp(p) ((p)->tempPos) -/* limit for CBcj2Enc::fileSize variable */ -#define BCJ2_FileSize_MAX ((UInt32)1 << 31) +#define Bcj2Enc_IsFinished(p) ((p)->flushRem == 0) + +/* Note : the decoder supports overlapping of marker (0f 80). + But we can eliminate such overlapping cases by setting + the limit for relative offset conversion as + CBcj2Enc::relatLimit <= (0x0f << 24) == (240 MiB) +*/ +/* default value for CBcj2Enc::relatLimit */ +#define BCJ2_ENC_RELAT_LIMIT_DEFAULT ((UInt32)0x0f << 24) +#define BCJ2_ENC_RELAT_LIMIT_MAX ((UInt32)1 << 31) +// #define BCJ2_RELAT_EXCLUDE_NUM_BITS 5 EXTERN_C_END diff --git a/3rdparty/lzma/C/Bcj2Enc.c b/3rdparty/lzma/C/Bcj2Enc.c index 0ea36462cbe..79460bb150a 100644 --- a/3rdparty/lzma/C/Bcj2Enc.c +++ b/3rdparty/lzma/C/Bcj2Enc.c @@ -1,61 +1,62 @@ -/* Bcj2Enc.c -- BCJ2 Encoder (Converter for x86 code) -2014-11-10 : Igor Pavlov : Public domain */ +/* Bcj2Enc.c -- BCJ2 Encoder converter for x86 code (Branch CALL/JUMP variant2) +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" /* #define SHOW_STAT */ - #ifdef SHOW_STAT #include <stdio.h> -#define PRF(x) x +#define PRF2(s) printf("%s ip=%8x tempPos=%d src= %8x\n", s, (unsigned)p->ip64, p->tempPos, (unsigned)(p->srcLim - p->src)); #else -#define PRF(x) +#define PRF2(s) #endif -#include <windows.h> -#include <string.h> - #include "Bcj2.h" #include "CpuArch.h" -#define CProb UInt16 - #define kTopValue ((UInt32)1 << 24) -#define kNumModelBits 11 -#define kBitModelTotal (1 << kNumModelBits) +#define kNumBitModelTotalBits 11 +#define kBitModelTotal (1 << kNumBitModelTotalBits) #define kNumMoveBits 5 void Bcj2Enc_Init(CBcj2Enc *p) { unsigned i; - - p->state = BCJ2_ENC_STATE_OK; + p->state = BCJ2_ENC_STATE_ORIG; p->finishMode = BCJ2_ENC_FINISH_MODE_CONTINUE; - - p->prevByte = 0; - + p->context = 0; + p->flushRem = 5; + p->isFlushState = 0; p->cache = 0; - p->range = 0xFFFFFFFF; + p->range = 0xffffffff; p->low = 0; p->cacheSize = 1; - - p->ip = 0; - - p->fileIp = 0; - p->fileSize = 0; - p->relatLimit = BCJ2_RELAT_LIMIT; - + p->ip64 = 0; + p->fileIp64 = 0; + p->fileSize64_minus1 = BCJ2_ENC_FileSizeField_UNLIMITED; + p->relatLimit = BCJ2_ENC_RELAT_LIMIT_DEFAULT; + // p->relatExcludeBits = 0; p->tempPos = 0; - - p->flushPos = 0; - for (i = 0; i < sizeof(p->probs) / sizeof(p->probs[0]); i++) p->probs[i] = kBitModelTotal >> 1; } -static Bool MY_FAST_CALL RangeEnc_ShiftLow(CBcj2Enc *p) +// Z7_NO_INLINE +Z7_FORCE_INLINE +static BoolInt Bcj2_RangeEnc_ShiftLow(CBcj2Enc *p) { - if ((UInt32)p->low < (UInt32)0xFF000000 || (UInt32)(p->low >> 32) != 0) + const UInt32 low = (UInt32)p->low; + const unsigned high = (unsigned) + #if defined(Z7_MSC_VER_ORIGINAL) \ + && defined(MY_CPU_X86) \ + && defined(MY_CPU_LE) \ + && !defined(MY_CPU_64BIT) + // we try to rid of __aullshr() call in MSVS-x86 + (((const UInt32 *)&p->low)[1]); // [1] : for little-endian only + #else + (p->low >> 32); + #endif + if (low < (UInt32)0xff000000 || high != 0) { Byte *buf = p->bufs[BCJ2_STREAM_RC]; do @@ -66,247 +67,440 @@ static Bool MY_FAST_CALL RangeEnc_ShiftLow(CBcj2Enc *p) p->bufs[BCJ2_STREAM_RC] = buf; return True; } - *buf++ = (Byte)(p->cache + (Byte)(p->low >> 32)); - p->cache = 0xFF; + *buf++ = (Byte)(p->cache + high); + p->cache = 0xff; } while (--p->cacheSize); p->bufs[BCJ2_STREAM_RC] = buf; - p->cache = (Byte)((UInt32)p->low >> 24); + p->cache = (Byte)(low >> 24); } p->cacheSize++; - p->low = (UInt32)p->low << 8; + p->low = low << 8; return False; } -static void Bcj2Enc_Encode_2(CBcj2Enc *p) -{ - if (BCJ2_IS_32BIT_STREAM(p->state)) + +/* +We can use 2 alternative versions of code: +1) non-marker version: + Byte CBcj2Enc::context + Byte temp[8]; + Last byte of marker (e8/e9/[0f]8x) can be written to temp[] buffer. + Encoder writes last byte of marker (e8/e9/[0f]8x) to dest, only in conjunction + with writing branch symbol to range coder in same Bcj2Enc_Encode_2() call. + +2) marker version: + UInt32 CBcj2Enc::context + Byte CBcj2Enc::temp[4]; + MARKER_FLAG in CBcj2Enc::context shows that CBcj2Enc::context contains finded marker. + it's allowed that + one call of Bcj2Enc_Encode_2() writes last byte of marker (e8/e9/[0f]8x) to dest, + and another call of Bcj2Enc_Encode_2() does offset conversion. + So different values of (fileIp) and (fileSize) are possible + in these different Bcj2Enc_Encode_2() calls. + +Also marker version requires additional if((v & MARKER_FLAG) == 0) check in main loop. +So we use non-marker version. +*/ + +/* + Corner cases with overlap in multi-block. + before v23: there was one corner case, where converted instruction + could start in one sub-stream and finish in next sub-stream. + If multi-block (solid) encoding is used, + and BCJ2_ENC_FINISH_MODE_END_BLOCK is used for each sub-stream. + and (0f) is last byte of previous sub-stream + and (8x) is first byte of current sub-stream + then (0f 8x) pair is treated as marker by BCJ2 encoder and decoder. + BCJ2 encoder can converts 32-bit offset for that (0f 8x) cortage, + if that offset meets limit requirements. + If encoder allows 32-bit offset conversion for such overlap case, + then the data in 3 uncompressed BCJ2 streams for some sub-stream + can depend from data of previous sub-stream. + That corner case is not big problem, and it's rare case. + Since v23.00 we do additional check to prevent conversions in such overlap cases. +*/ + +/* + Bcj2Enc_Encode_2() output variables at exit: { - Byte *cur = p->bufs[p->state]; - if (cur == p->lims[p->state]) - return; - SetBe32(cur, p->tempTarget); - p->bufs[p->state] = cur + 4; + if (Bcj2Enc_Encode_2() exits with (p->state == BCJ2_ENC_STATE_ORIG)) + { + it means that encoder needs more input data. + if (p->srcLim == p->src) at exit, then + { + (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM) + all input data were read and processed, and we are ready for + new input data. + } + else + { + (p->srcLim != p->src) + (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE) + The encoder have found e8/e9/0f_8x marker, + and p->src points to last byte of that marker, + Bcj2Enc_Encode_2() needs more input data to get totally + 5 bytes (last byte of marker and 32-bit branch offset) + as continuous array starting from p->src. + (p->srcLim - p->src < 5) requirement is met after exit. + So non-processed resedue from p->src to p->srcLim is always less than 5 bytes. + } + } } +*/ - p->state = BCJ2_ENC_STATE_ORIG; - - for (;;) +Z7_NO_INLINE +static void Bcj2Enc_Encode_2(CBcj2Enc *p) +{ + if (!p->isFlushState) { - if (p->range < kTopValue) + const Byte *src; + UInt32 v; { - if (RangeEnc_ShiftLow(p)) - return; - p->range <<= 8; + const unsigned state = p->state; + if (BCJ2_IS_32BIT_STREAM(state)) + { + Byte *cur = p->bufs[state]; + if (cur == p->lims[state]) + return; + SetBe32a(cur, p->tempTarget) + p->bufs[state] = cur + 4; + } } + p->state = BCJ2_ENC_STATE_ORIG; // for main reason of exit + src = p->src; + v = p->context; + + // #define WRITE_CONTEXT p->context = v; // for marker version + #define WRITE_CONTEXT p->context = (Byte)v; + #define WRITE_CONTEXT_AND_SRC p->src = src; WRITE_CONTEXT + for (;;) { + // const Byte *src; + // UInt32 v; + CBcj2Enc_ip_unsigned ip; + if (p->range < kTopValue) + { + // to reduce register pressure and code size: we save and restore local variables. + WRITE_CONTEXT_AND_SRC + if (Bcj2_RangeEnc_ShiftLow(p)) + return; + p->range <<= 8; + src = p->src; + v = p->context; + } + // src = p->src; + // #define MARKER_FLAG ((UInt32)1 << 17) + // if ((v & MARKER_FLAG) == 0) // for marker version { - const Byte *src = p->src; const Byte *srcLim; - Byte *dest; - SizeT num = p->srcLim - src; - - if (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE) + Byte *dest = p->bufs[BCJ2_STREAM_MAIN]; { - if (num <= 4) - return; - num -= 4; + const SizeT remSrc = (SizeT)(p->srcLim - src); + SizeT rem = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - dest); + if (rem >= remSrc) + rem = remSrc; + srcLim = src + rem; } - else if (num == 0) - break; - - dest = p->bufs[BCJ2_STREAM_MAIN]; - if (num > (SizeT)(p->lims[BCJ2_STREAM_MAIN] - dest)) + /* p->context contains context of previous byte: + bits [0 : 7] : src[-1], if (src) was changed in this call + bits [8 : 31] : are undefined for non-marker version + */ + // v = p->context; + #define NUM_SHIFT_BITS 24 + #define CONV_FLAG ((UInt32)1 << 16) + #define ONE_ITER { \ + b = src[0]; \ + *dest++ = (Byte)b; \ + v = (v << NUM_SHIFT_BITS) | b; \ + if (((b + (0x100 - 0xe8)) & 0xfe) == 0) break; \ + if (((v - (((UInt32)0x0f << (NUM_SHIFT_BITS)) + 0x80)) & \ + ((((UInt32)1 << (4 + NUM_SHIFT_BITS)) - 0x1) << 4)) == 0) break; \ + src++; if (src == srcLim) { break; } } + + if (src != srcLim) + for (;;) { - num = p->lims[BCJ2_STREAM_MAIN] - dest; - if (num == 0) - { - p->state = BCJ2_STREAM_MAIN; - return; - } + /* clang can generate ineffective code with setne instead of two jcc instructions. + we can use 2 iterations and external (unsigned b) to avoid that ineffective code genaration. */ + unsigned b; + ONE_ITER + ONE_ITER } - - srcLim = src + num; + + ip = p->ip64 + (CBcj2Enc_ip_unsigned)(SizeT)(dest - p->bufs[BCJ2_STREAM_MAIN]); + p->bufs[BCJ2_STREAM_MAIN] = dest; + p->ip64 = ip; - if (p->prevByte == 0x0F && (src[0] & 0xF0) == 0x80) - *dest = src[0]; - else for (;;) + if (src == srcLim) { - Byte b = *src; - *dest = b; - if (b != 0x0F) + WRITE_CONTEXT_AND_SRC + if (src != p->srcLim) { - if ((b & 0xFE) == 0xE8) - break; - dest++; - if (++src != srcLim) - continue; - break; + p->state = BCJ2_STREAM_MAIN; + return; } - dest++; - if (++src == srcLim) - break; - if ((*src & 0xF0) != 0x80) - continue; - *dest = *src; + /* (p->src == p->srcLim) + (p->state == BCJ2_ENC_STATE_ORIG) */ + if (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM) + return; + /* (p->finishMode == BCJ2_ENC_FINISH_MODE_END_STREAM */ + // (p->flushRem == 5); + p->isFlushState = 1; break; } - - num = src - p->src; - - if (src == srcLim) - { - p->prevByte = src[-1]; - p->bufs[BCJ2_STREAM_MAIN] = dest; - p->src = src; - p->ip += (UInt32)num; - continue; - } - + src++; + // p->src = src; + } + // ip = p->ip; // for marker version + /* marker was found */ + /* (v) contains marker that was found: + bits [NUM_SHIFT_BITS : NUM_SHIFT_BITS + 7] + : value of src[-2] : xx/xx/0f + bits [0 : 7] : value of src[-1] : e8/e9/8x + */ + { { - Byte context = (Byte)(num == 0 ? p->prevByte : src[-1]); - Bool needConvert; - - p->bufs[BCJ2_STREAM_MAIN] = dest + 1; - p->ip += (UInt32)num + 1; - src++; - - needConvert = False; - + #if NUM_SHIFT_BITS != 24 + v &= ~(UInt32)CONV_FLAG; + #endif + // UInt32 relat = 0; if ((SizeT)(p->srcLim - src) >= 4) { - UInt32 relatVal = GetUi32(src); - if ((p->fileSize == 0 || (UInt32)(p->ip + 4 + relatVal - p->fileIp) < p->fileSize) - && ((relatVal + p->relatLimit) >> 1) < p->relatLimit) - needConvert = True; + /* + if (relat != 0 || (Byte)v != 0xe8) + BoolInt isBigOffset = True; + */ + const UInt32 relat = GetUi32(src); + /* + #define EXCLUDE_FLAG ((UInt32)1 << 4) + #define NEED_CONVERT(rel) ((((rel) + EXCLUDE_FLAG) & (0 - EXCLUDE_FLAG * 2)) != 0) + if (p->relatExcludeBits != 0) + { + const UInt32 flag = (UInt32)1 << (p->relatExcludeBits - 1); + isBigOffset = (((relat + flag) & (0 - flag * 2)) != 0); + } + // isBigOffset = False; // for debug + */ + ip -= p->fileIp64; + // Use the following if check, if (ip) is 64-bit: + if (ip > (((v + 0x20) >> 5) & 1)) // 23.00 : we eliminate milti-block overlap for (Of 80) and (e8/e9) + if ((CBcj2Enc_ip_unsigned)((CBcj2Enc_ip_signed)ip + 4 + (Int32)relat) <= p->fileSize64_minus1) + if (((UInt32)(relat + p->relatLimit) >> 1) < p->relatLimit) + v |= CONV_FLAG; } - + else if (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE) { - UInt32 bound; - unsigned ttt; - Byte b = src[-1]; - CProb *prob = p->probs + (unsigned)(b == 0xE8 ? 2 + (unsigned)context : (b == 0xE9 ? 1 : 0)); - - ttt = *prob; - bound = (p->range >> kNumModelBits) * ttt; - - if (!needConvert) + // (p->srcLim - src < 4) + // /* + // for non-marker version + p->ip64--; // p->ip = ip - 1; + p->bufs[BCJ2_STREAM_MAIN]--; + src--; + v >>= NUM_SHIFT_BITS; + // (0 < p->srcLim - p->src <= 4) + // */ + // v |= MARKER_FLAG; // for marker version + /* (p->state == BCJ2_ENC_STATE_ORIG) */ + WRITE_CONTEXT_AND_SRC + return; + } + { + const unsigned c = ((v + 0x17) >> 6) & 1; + CBcj2Prob *prob = p->probs + (unsigned) + (((0 - c) & (Byte)(v >> NUM_SHIFT_BITS)) + c + ((v >> 5) & 1)); + /* + ((Byte)v == 0xe8 ? 2 + ((Byte)(v >> 8)) : + ((Byte)v < 0xe8 ? 0 : 1)); // ((v >> 5) & 1)); + */ + const unsigned ttt = *prob; + const UInt32 bound = (p->range >> kNumBitModelTotalBits) * ttt; + if ((v & CONV_FLAG) == 0) { + // static int yyy = 0; yyy++; printf("\n!needConvert = %d\n", yyy); + // v = (Byte)v; // for marker version p->range = bound; - *prob = (CProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); - p->src = src; - p->prevByte = b; + *prob = (CBcj2Prob)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); + // WRITE_CONTEXT_AND_SRC continue; } - p->low += bound; p->range -= bound; - *prob = (CProb)(ttt - (ttt >> kNumMoveBits)); - + *prob = (CBcj2Prob)(ttt - (ttt >> kNumMoveBits)); + } + // p->context = src[3]; + { + // const unsigned cj = ((Byte)v == 0xe8 ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP); + const unsigned cj = (((v + 0x57) >> 6) & 1) + BCJ2_STREAM_CALL; + ip = p->ip64; + v = GetUi32(src); // relat + ip += 4; + p->ip64 = ip; + src += 4; + // p->src = src; { - UInt32 relatVal = GetUi32(src); - UInt32 absVal; - p->ip += 4; - absVal = p->ip + relatVal; - p->prevByte = src[3]; - src += 4; - p->src = src; + const UInt32 absol = (UInt32)ip + v; + Byte *cur = p->bufs[cj]; + v >>= 24; + // WRITE_CONTEXT + if (cur == p->lims[cj]) { - unsigned cj = (b == 0xE8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP; - Byte *cur = p->bufs[cj]; - if (cur == p->lims[cj]) - { - p->state = cj; - p->tempTarget = absVal; - return; - } - SetBe32(cur, absVal); - p->bufs[cj] = cur + 4; + p->state = cj; + p->tempTarget = absol; + WRITE_CONTEXT_AND_SRC + return; } + SetBe32a(cur, absol) + p->bufs[cj] = cur + 4; } } } } - } + } // end of loop } - if (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM) - return; - - for (; p->flushPos < 5; p->flushPos++) - if (RangeEnc_ShiftLow(p)) + for (; p->flushRem != 0; p->flushRem--) + if (Bcj2_RangeEnc_ShiftLow(p)) return; - p->state = BCJ2_ENC_STATE_OK; + p->state = BCJ2_ENC_STATE_FINISHED; } +/* +BCJ2 encoder needs look ahead for up to 4 bytes in (src) buffer. +So base function Bcj2Enc_Encode_2() + in BCJ2_ENC_FINISH_MODE_CONTINUE mode can return with + (p->state == BCJ2_ENC_STATE_ORIG && p->src < p->srcLim) +Bcj2Enc_Encode() solves that look ahead problem by using p->temp[] buffer. + so if (p->state == BCJ2_ENC_STATE_ORIG) after Bcj2Enc_Encode(), + then (p->src == p->srcLim). + And the caller's code is simpler with Bcj2Enc_Encode(). +*/ + +Z7_NO_INLINE void Bcj2Enc_Encode(CBcj2Enc *p) { - PRF(printf("\n")); - PRF(printf("---- ip = %8d tempPos = %8d src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src)); - + PRF2("\n----") if (p->tempPos != 0) { + /* extra: number of bytes that were copied from (src) to (temp) buffer in this call */ unsigned extra = 0; - + /* We will touch only minimal required number of bytes in input (src) stream. + So we will add input bytes from (src) stream to temp[] with step of 1 byte. + We don't add new bytes to temp[] before Bcj2Enc_Encode_2() call + in first loop iteration because + - previous call of Bcj2Enc_Encode() could use another (finishMode), + - previous call could finish with (p->state != BCJ2_ENC_STATE_ORIG). + the case with full temp[] buffer (p->tempPos == 4) is possible here. + */ for (;;) { + // (0 < p->tempPos <= 5) // in non-marker version + /* p->src : the current src data position including extra bytes + that were copied to temp[] buffer in this call */ const Byte *src = p->src; const Byte *srcLim = p->srcLim; - unsigned finishMode = p->finishMode; - - p->src = p->temp; - p->srcLim = p->temp + p->tempPos; + const EBcj2Enc_FinishMode finishMode = p->finishMode; if (src != srcLim) + { + /* if there are some src data after the data copied to temp[], + then we use MODE_CONTINUE for temp data */ p->finishMode = BCJ2_ENC_FINISH_MODE_CONTINUE; - - PRF(printf(" ip = %8d tempPos = %8d src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src)); - + } + p->src = p->temp; + p->srcLim = p->temp + p->tempPos; + PRF2(" ") Bcj2Enc_Encode_2(p); - { - unsigned num = (unsigned)(p->src - p->temp); - unsigned tempPos = p->tempPos - num; + const unsigned num = (unsigned)(p->src - p->temp); + const unsigned tempPos = p->tempPos - num; unsigned i; p->tempPos = tempPos; for (i = 0; i < tempPos; i++) - p->temp[i] = p->temp[i + num]; - + p->temp[i] = p->temp[(SizeT)i + num]; + // tempPos : number of bytes in temp buffer p->src = src; p->srcLim = srcLim; p->finishMode = finishMode; - - if (p->state != BCJ2_ENC_STATE_ORIG || src == srcLim) + if (p->state != BCJ2_ENC_STATE_ORIG) + { + // (p->tempPos <= 4) // in non-marker version + /* if (the reason of exit from Bcj2Enc_Encode_2() + is not BCJ2_ENC_STATE_ORIG), + then we exit from Bcj2Enc_Encode() with same reason */ + // optional code begin : we rollback (src) and tempPos, if it's possible: + if (extra >= tempPos) + extra = tempPos; + p->src = src - extra; + p->tempPos = tempPos - extra; + // optional code end : rollback of (src) and tempPos return; - + } + /* (p->tempPos <= 4) + (p->state == BCJ2_ENC_STATE_ORIG) + so encoder needs more data than in temp[] */ + if (src == srcLim) + return; // src buffer has no more input data. + /* (src != srcLim) + so we can provide more input data from src for Bcj2Enc_Encode_2() */ if (extra >= tempPos) { - p->src = src - tempPos; + /* (extra >= tempPos) means that temp buffer contains + only data from src buffer of this call. + So now we can encode without temp buffer */ + p->src = src - tempPos; // rollback (src) p->tempPos = 0; break; } - - p->temp[tempPos] = src[0]; + // we append one additional extra byte from (src) to temp[] buffer: + p->temp[tempPos] = *src; p->tempPos = tempPos + 1; + // (0 < p->tempPos <= 5) // in non-marker version p->src = src + 1; extra++; } } } - PRF(printf("++++ ip = %8d tempPos = %8d src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src)); - + PRF2("++++") + // (p->tempPos == 0) Bcj2Enc_Encode_2(p); + PRF2("====") if (p->state == BCJ2_ENC_STATE_ORIG) { const Byte *src = p->src; - unsigned rem = (unsigned)(p->srcLim - src); - unsigned i; - for (i = 0; i < rem; i++) - p->temp[i] = src[i]; - p->tempPos = rem; - p->src = src + rem; + const Byte *srcLim = p->srcLim; + const unsigned rem = (unsigned)(srcLim - src); + /* (rem <= 4) here. + if (p->src != p->srcLim), then + - we copy non-processed bytes from (p->src) to temp[] buffer, + - we set p->src equal to p->srcLim. + */ + if (rem) + { + unsigned i = 0; + p->src = srcLim; + p->tempPos = rem; + // (0 < p->tempPos <= 4) + do + p->temp[i] = src[i]; + while (++i != rem); + } + // (p->tempPos <= 4) + // (p->src == p->srcLim) } } + +#undef PRF2 +#undef CONV_FLAG +#undef MARKER_FLAG +#undef WRITE_CONTEXT +#undef WRITE_CONTEXT_AND_SRC +#undef ONE_ITER +#undef NUM_SHIFT_BITS +#undef kTopValue +#undef kNumBitModelTotalBits +#undef kBitModelTotal +#undef kNumMoveBits diff --git a/3rdparty/lzma/C/Bra.c b/3rdparty/lzma/C/Bra.c index cdb94569202..22e0e478eca 100644 --- a/3rdparty/lzma/C/Bra.c +++ b/3rdparty/lzma/C/Bra.c @@ -1,135 +1,420 @@ -/* Bra.c -- Converters for RISC code -2010-04-16 : Igor Pavlov : Public domain */ +/* Bra.c -- Branch converters for RISC code +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" #include "Bra.h" +#include "CpuArch.h" +#include "RotateDefs.h" -SizeT ARM_Convert(Byte *data, SizeT size, UInt32 ip, int encoding) +#if defined(MY_CPU_SIZEOF_POINTER) \ + && ( MY_CPU_SIZEOF_POINTER == 4 \ + || MY_CPU_SIZEOF_POINTER == 8) + #define BR_CONV_USE_OPT_PC_PTR +#endif + +#ifdef BR_CONV_USE_OPT_PC_PTR +#define BR_PC_INIT pc -= (UInt32)(SizeT)p; +#define BR_PC_GET (pc + (UInt32)(SizeT)p) +#else +#define BR_PC_INIT pc += (UInt32)size; +#define BR_PC_GET (pc - (UInt32)(SizeT)(lim - p)) +// #define BR_PC_INIT +// #define BR_PC_GET (pc + (UInt32)(SizeT)(p - data)) +#endif + +#define BR_CONVERT_VAL(v, c) if (encoding) v += c; else v -= c; +// #define BR_CONVERT_VAL(v, c) if (!encoding) c = (UInt32)0 - c; v += c; + +#define Z7_BRANCH_CONV(name) z7_BranchConv_ ## name + +#define Z7_BRANCH_FUNC_MAIN(name) \ +static \ +Z7_FORCE_INLINE \ +Z7_ATTRIB_NO_VECTOR \ +Byte *Z7_BRANCH_CONV(name)(Byte *p, SizeT size, UInt32 pc, int encoding) + +#define Z7_BRANCH_FUNC_IMP(name, m, encoding) \ +Z7_NO_INLINE \ +Z7_ATTRIB_NO_VECTOR \ +Byte *m(name)(Byte *data, SizeT size, UInt32 pc) \ + { return Z7_BRANCH_CONV(name)(data, size, pc, encoding); } \ + +#ifdef Z7_EXTRACT_ONLY +#define Z7_BRANCH_FUNCS_IMP(name) \ + Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_DEC, 0) +#else +#define Z7_BRANCH_FUNCS_IMP(name) \ + Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_DEC, 0) \ + Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_ENC, 1) +#endif + +#if defined(__clang__) +#define BR_EXTERNAL_FOR +#define BR_NEXT_ITERATION continue; +#else +#define BR_EXTERNAL_FOR for (;;) +#define BR_NEXT_ITERATION break; +#endif + +#if defined(__clang__) && (__clang_major__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ >= 1000) \ + // GCC is not good for __builtin_expect() here + /* || defined(_MSC_VER) && (_MSC_VER >= 1920) */ + // #define Z7_unlikely [[unlikely]] + // #define Z7_LIKELY(x) (__builtin_expect((x), 1)) + #define Z7_UNLIKELY(x) (__builtin_expect((x), 0)) + // #define Z7_likely [[likely]] +#else + // #define Z7_LIKELY(x) (x) + #define Z7_UNLIKELY(x) (x) + // #define Z7_likely +#endif + + +Z7_BRANCH_FUNC_MAIN(ARM64) { - SizeT i; - if (size < 4) - return 0; - size -= 4; - ip += 8; - for (i = 0; i <= size; i += 4) + // Byte *p = data; + const Byte *lim; + const UInt32 flag = (UInt32)1 << (24 - 4); + const UInt32 mask = ((UInt32)1 << 24) - (flag << 1); + size &= ~(SizeT)3; + // if (size == 0) return p; + lim = p + size; + BR_PC_INIT + pc -= 4; // because (p) will point to next instruction + + BR_EXTERNAL_FOR { - if (data[i + 3] == 0xEB) + // Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (;;) { - UInt32 dest; - UInt32 src = ((UInt32)data[i + 2] << 16) | ((UInt32)data[i + 1] << 8) | (data[i + 0]); - src <<= 2; - if (encoding) - dest = ip + (UInt32)i + src; - else - dest = src - (ip + (UInt32)i); - dest >>= 2; - data[i + 2] = (Byte)(dest >> 16); - data[i + 1] = (Byte)(dest >> 8); - data[i + 0] = (Byte)dest; + UInt32 v; + if Z7_UNLIKELY(p == lim) + return p; + v = GetUi32a(p); + p += 4; + if Z7_UNLIKELY(((v - 0x94000000) & 0xfc000000) == 0) + { + UInt32 c = BR_PC_GET >> 2; + BR_CONVERT_VAL(v, c) + v &= 0x03ffffff; + v |= 0x94000000; + SetUi32a(p - 4, v) + BR_NEXT_ITERATION + } + // v = rotlFixed(v, 8); v += (flag << 8) - 0x90; if Z7_UNLIKELY((v & ((mask << 8) + 0x9f)) == 0) + v -= 0x90000000; if Z7_UNLIKELY((v & 0x9f000000) == 0) + { + UInt32 z, c; + // v = rotrFixed(v, 8); + v += flag; if Z7_UNLIKELY(v & mask) continue; + z = (v & 0xffffffe0) | (v >> 26); + c = (BR_PC_GET >> (12 - 3)) & ~(UInt32)7; + BR_CONVERT_VAL(z, c) + v &= 0x1f; + v |= 0x90000000; + v |= z << 26; + v |= 0x00ffffe0 & ((z & (((flag << 1) - 1))) - flag); + SetUi32a(p - 4, v) + } } } - return i; } +Z7_BRANCH_FUNCS_IMP(ARM64) + -SizeT ARMT_Convert(Byte *data, SizeT size, UInt32 ip, int encoding) +Z7_BRANCH_FUNC_MAIN(ARM) { - SizeT i; - if (size < 4) - return 0; - size -= 4; - ip += 4; - for (i = 0; i <= size; i += 2) + // Byte *p = data; + const Byte *lim; + size &= ~(SizeT)3; + lim = p + size; + BR_PC_INIT + /* in ARM: branch offset is relative to the +2 instructions from current instruction. + (p) will point to next instruction */ + pc += 8 - 4; + + for (;;) { - if ((data[i + 1] & 0xF8) == 0xF0 && - (data[i + 3] & 0xF8) == 0xF8) + for (;;) { - UInt32 dest; - UInt32 src = - (((UInt32)data[i + 1] & 0x7) << 19) | - ((UInt32)data[i + 0] << 11) | - (((UInt32)data[i + 3] & 0x7) << 8) | - (data[i + 2]); - - src <<= 1; - if (encoding) - dest = ip + (UInt32)i + src; - else - dest = src - (ip + (UInt32)i); - dest >>= 1; - - data[i + 1] = (Byte)(0xF0 | ((dest >> 19) & 0x7)); - data[i + 0] = (Byte)(dest >> 11); - data[i + 3] = (Byte)(0xF8 | ((dest >> 8) & 0x7)); - data[i + 2] = (Byte)dest; - i += 2; + if Z7_UNLIKELY(p >= lim) { return p; } p += 4; if Z7_UNLIKELY(p[-1] == 0xeb) break; + if Z7_UNLIKELY(p >= lim) { return p; } p += 4; if Z7_UNLIKELY(p[-1] == 0xeb) break; + } + { + UInt32 v = GetUi32a(p - 4); + UInt32 c = BR_PC_GET >> 2; + BR_CONVERT_VAL(v, c) + v &= 0x00ffffff; + v |= 0xeb000000; + SetUi32a(p - 4, v) } } - return i; } +Z7_BRANCH_FUNCS_IMP(ARM) + -SizeT PPC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding) +Z7_BRANCH_FUNC_MAIN(PPC) { - SizeT i; - if (size < 4) - return 0; - size -= 4; - for (i = 0; i <= size; i += 4) + // Byte *p = data; + const Byte *lim; + size &= ~(SizeT)3; + lim = p + size; + BR_PC_INIT + pc -= 4; // because (p) will point to next instruction + + for (;;) { - if ((data[i] >> 2) == 0x12 && (data[i + 3] & 3) == 1) + UInt32 v; + for (;;) + { + if Z7_UNLIKELY(p == lim) + return p; + // v = GetBe32a(p); + v = *(UInt32 *)(void *)p; + p += 4; + // if ((v & 0xfc000003) == 0x48000001) break; + // if ((p[-4] & 0xFC) == 0x48 && (p[-1] & 3) == 1) break; + if Z7_UNLIKELY( + ((v - Z7_CONV_BE_TO_NATIVE_CONST32(0x48000001)) + & Z7_CONV_BE_TO_NATIVE_CONST32(0xfc000003)) == 0) break; + } { - UInt32 src = ((UInt32)(data[i + 0] & 3) << 24) | - ((UInt32)data[i + 1] << 16) | - ((UInt32)data[i + 2] << 8) | - ((UInt32)data[i + 3] & (~3)); - - UInt32 dest; - if (encoding) - dest = ip + (UInt32)i + src; - else - dest = src - (ip + (UInt32)i); - data[i + 0] = (Byte)(0x48 | ((dest >> 24) & 0x3)); - data[i + 1] = (Byte)(dest >> 16); - data[i + 2] = (Byte)(dest >> 8); - data[i + 3] &= 0x3; - data[i + 3] |= dest; + v = Z7_CONV_NATIVE_TO_BE_32(v); + { + UInt32 c = BR_PC_GET; + BR_CONVERT_VAL(v, c) + } + v &= 0x03ffffff; + v |= 0x48000000; + SetBe32a(p - 4, v) } } - return i; } +Z7_BRANCH_FUNCS_IMP(PPC) -SizeT SPARC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding) + +#ifdef Z7_CPU_FAST_ROTATE_SUPPORTED +#define BR_SPARC_USE_ROTATE +#endif + +Z7_BRANCH_FUNC_MAIN(SPARC) { - UInt32 i; - if (size < 4) - return 0; - size -= 4; - for (i = 0; i <= size; i += 4) + // Byte *p = data; + const Byte *lim; + const UInt32 flag = (UInt32)1 << 22; + size &= ~(SizeT)3; + lim = p + size; + BR_PC_INIT + pc -= 4; // because (p) will point to next instruction + for (;;) { - if ((data[i] == 0x40 && (data[i + 1] & 0xC0) == 0x00) || - (data[i] == 0x7F && (data[i + 1] & 0xC0) == 0xC0)) + UInt32 v; + for (;;) + { + if Z7_UNLIKELY(p == lim) + return p; + /* // the code without GetBe32a(): + { const UInt32 v = GetUi16a(p) & 0xc0ff; p += 4; if (v == 0x40 || v == 0xc07f) break; } + */ + v = GetBe32a(p); + p += 4; + #ifdef BR_SPARC_USE_ROTATE + v = rotlFixed(v, 2); + v += (flag << 2) - 1; + if Z7_UNLIKELY((v & (3 - (flag << 3))) == 0) + #else + v += (UInt32)5 << 29; + v ^= (UInt32)7 << 29; + v += flag; + if Z7_UNLIKELY((v & (0 - (flag << 1))) == 0) + #endif + break; + } + { + // UInt32 v = GetBe32a(p - 4); + #ifndef BR_SPARC_USE_ROTATE + v <<= 2; + #endif + { + UInt32 c = BR_PC_GET; + BR_CONVERT_VAL(v, c) + } + v &= (flag << 3) - 1; + #ifdef BR_SPARC_USE_ROTATE + v -= (flag << 2) - 1; + v = rotrFixed(v, 2); + #else + v -= (flag << 2); + v >>= 2; + v |= (UInt32)1 << 30; + #endif + SetBe32a(p - 4, v) + } + } +} +Z7_BRANCH_FUNCS_IMP(SPARC) + + +Z7_BRANCH_FUNC_MAIN(ARMT) +{ + // Byte *p = data; + Byte *lim; + size &= ~(SizeT)1; + // if (size == 0) return p; + if (size <= 2) return p; + size -= 2; + lim = p + size; + BR_PC_INIT + /* in ARM: branch offset is relative to the +2 instructions from current instruction. + (p) will point to the +2 instructions from current instruction */ + // pc += 4 - 4; + // if (encoding) pc -= 0xf800 << 1; else pc += 0xf800 << 1; + // #define ARMT_TAIL_PROC { goto armt_tail; } + #define ARMT_TAIL_PROC { return p; } + + do + { + /* in MSVC 32-bit x86 compilers: + UInt32 version : it loads value from memory with movzx + Byte version : it loads value to 8-bit register (AL/CL) + movzx version is slightly faster in some cpus + */ + unsigned b1; + // Byte / unsigned + b1 = p[1]; + // optimized version to reduce one (p >= lim) check: + // unsigned a1 = p[1]; b1 = p[3]; p += 2; if Z7_LIKELY((b1 & (a1 ^ 8)) < 0xf8) + for (;;) + { + unsigned b3; // Byte / UInt32 + /* (Byte)(b3) normalization can use low byte computations in MSVC. + It gives smaller code, and no loss of speed in some compilers/cpus. + But new MSVC 32-bit x86 compilers use more slow load + from memory to low byte register in that case. + So we try to use full 32-bit computations for faster code. + */ + // if (p >= lim) { ARMT_TAIL_PROC } b3 = b1 + 8; b1 = p[3]; p += 2; if ((b3 & b1) >= 0xf8) break; + if Z7_UNLIKELY(p >= lim) { ARMT_TAIL_PROC } b3 = p[3]; p += 2; if Z7_UNLIKELY((b3 & (b1 ^ 8)) >= 0xf8) break; + if Z7_UNLIKELY(p >= lim) { ARMT_TAIL_PROC } b1 = p[3]; p += 2; if Z7_UNLIKELY((b1 & (b3 ^ 8)) >= 0xf8) break; + } + { + /* we can adjust pc for (0xf800) to rid of (& 0x7FF) operation. + But gcc/clang for arm64 can use bfi instruction for full code here */ + UInt32 v = + ((UInt32)GetUi16a(p - 2) << 11) | + ((UInt32)GetUi16a(p) & 0x7FF); + /* + UInt32 v = + ((UInt32)p[1 - 2] << 19) + + (((UInt32)p[1] & 0x7) << 8) + + (((UInt32)p[-2] << 11)) + + (p[0]); + */ + p += 2; + { + UInt32 c = BR_PC_GET >> 1; + BR_CONVERT_VAL(v, c) + } + SetUi16a(p - 4, (UInt16)(((v >> 11) & 0x7ff) | 0xf000)) + SetUi16a(p - 2, (UInt16)(v | 0xf800)) + /* + p[-4] = (Byte)(v >> 11); + p[-3] = (Byte)(0xf0 | ((v >> 19) & 0x7)); + p[-2] = (Byte)v; + p[-1] = (Byte)(0xf8 | (v >> 8)); + */ + } + } + while (p < lim); + return p; + // armt_tail: + // if ((Byte)((lim[1] & 0xf8)) != 0xf0) { lim += 2; } return lim; + // return (Byte *)(lim + ((Byte)((lim[1] ^ 0xf0) & 0xf8) == 0 ? 0 : 2)); + // return (Byte *)(lim + (((lim[1] ^ ~0xfu) & ~7u) == 0 ? 0 : 2)); + // return (Byte *)(lim + 2 - (((((unsigned)lim[1] ^ 8) + 8) >> 7) & 2)); +} +Z7_BRANCH_FUNCS_IMP(ARMT) + + +// #define BR_IA64_NO_INLINE + +Z7_BRANCH_FUNC_MAIN(IA64) +{ + // Byte *p = data; + const Byte *lim; + size &= ~(SizeT)15; + lim = p + size; + pc -= 1 << 4; + pc >>= 4 - 1; + // pc -= 1 << 1; + + for (;;) + { + unsigned m; + for (;;) + { + if Z7_UNLIKELY(p == lim) + return p; + m = (unsigned)((UInt32)0x334b0000 >> (*p & 0x1e)); + p += 16; + pc += 1 << 1; + if (m &= 3) + break; + } { - UInt32 src = - ((UInt32)data[i + 0] << 24) | - ((UInt32)data[i + 1] << 16) | - ((UInt32)data[i + 2] << 8) | - ((UInt32)data[i + 3]); - UInt32 dest; - - src <<= 2; - if (encoding) - dest = ip + i + src; - else - dest = src - (ip + i); - dest >>= 2; - - dest = (((0 - ((dest >> 22) & 1)) << 22) & 0x3FFFFFFF) | (dest & 0x3FFFFF) | 0x40000000; - - data[i + 0] = (Byte)(dest >> 24); - data[i + 1] = (Byte)(dest >> 16); - data[i + 2] = (Byte)(dest >> 8); - data[i + 3] = (Byte)dest; + p += (ptrdiff_t)m * 5 - 20; // negative value is expected here. + do + { + const UInt32 t = + #if defined(MY_CPU_X86_OR_AMD64) + // we use 32-bit load here to reduce code size on x86: + GetUi32(p); + #else + GetUi16(p); + #endif + UInt32 z = GetUi32(p + 1) >> m; + p += 5; + if (((t >> m) & (0x70 << 1)) == 0 + && ((z - (0x5000000 << 1)) & (0xf000000 << 1)) == 0) + { + UInt32 v = (UInt32)((0x8fffff << 1) | 1) & z; + z ^= v; + #ifdef BR_IA64_NO_INLINE + v |= (v & ((UInt32)1 << (23 + 1))) >> 3; + { + UInt32 c = pc; + BR_CONVERT_VAL(v, c) + } + v &= (0x1fffff << 1) | 1; + #else + { + if (encoding) + { + // pc &= ~(0xc00000 << 1); // we just need to clear at least 2 bits + pc &= (0x1fffff << 1) | 1; + v += pc; + } + else + { + // pc |= 0xc00000 << 1; // we need to set at least 2 bits + pc |= ~(UInt32)((0x1fffff << 1) | 1); + v -= pc; + } + } + v &= ~(UInt32)(0x600000 << 1); + #endif + v += (0x700000 << 1); + v &= (0x8fffff << 1) | 1; + z |= v; + z <<= m; + SetUi32(p + 1 - 5, z) + } + m++; + } + while (m &= 3); // while (m < 4); } } - return i; } +Z7_BRANCH_FUNCS_IMP(IA64) diff --git a/3rdparty/lzma/C/Bra.h b/3rdparty/lzma/C/Bra.h index 855e37a6b50..a4ee568e215 100644 --- a/3rdparty/lzma/C/Bra.h +++ b/3rdparty/lzma/C/Bra.h @@ -1,64 +1,99 @@ /* Bra.h -- Branch converters for executables -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __BRA_H -#define __BRA_H +#ifndef ZIP7_INC_BRA_H +#define ZIP7_INC_BRA_H #include "7zTypes.h" EXTERN_C_BEGIN +#define Z7_BRANCH_CONV_DEC(name) z7_BranchConv_ ## name ## _Dec +#define Z7_BRANCH_CONV_ENC(name) z7_BranchConv_ ## name ## _Enc +#define Z7_BRANCH_CONV_ST_DEC(name) z7_BranchConvSt_ ## name ## _Dec +#define Z7_BRANCH_CONV_ST_ENC(name) z7_BranchConvSt_ ## name ## _Enc + +#define Z7_BRANCH_CONV_DECL(name) Byte * name(Byte *data, SizeT size, UInt32 pc) +#define Z7_BRANCH_CONV_ST_DECL(name) Byte * name(Byte *data, SizeT size, UInt32 pc, UInt32 *state) + +typedef Z7_BRANCH_CONV_DECL( (*z7_Func_BranchConv)); +typedef Z7_BRANCH_CONV_ST_DECL((*z7_Func_BranchConvSt)); + +#define Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL 0 +Z7_BRANCH_CONV_ST_DECL(Z7_BRANCH_CONV_ST_DEC(X86)); +Z7_BRANCH_CONV_ST_DECL(Z7_BRANCH_CONV_ST_ENC(X86)); + +#define Z7_BRANCH_FUNCS_DECL(name) \ +Z7_BRANCH_CONV_DECL(Z7_BRANCH_CONV_DEC(name)); \ +Z7_BRANCH_CONV_DECL(Z7_BRANCH_CONV_ENC(name)); + +Z7_BRANCH_FUNCS_DECL(ARM64) +Z7_BRANCH_FUNCS_DECL(ARM) +Z7_BRANCH_FUNCS_DECL(ARMT) +Z7_BRANCH_FUNCS_DECL(PPC) +Z7_BRANCH_FUNCS_DECL(SPARC) +Z7_BRANCH_FUNCS_DECL(IA64) + /* -These functions convert relative addresses to absolute addresses -in CALL instructions to increase the compression ratio. - - In: - data - data buffer - size - size of data - ip - current virtual Instruction Pinter (IP) value - state - state variable for x86 converter - encoding - 0 (for decoding), 1 (for encoding) - - Out: - state - state variable for x86 converter +These functions convert data that contain CPU instructions. +Each such function converts relative addresses to absolute addresses in some +branch instructions: CALL (in all converters) and JUMP (X86 converter only). +Such conversion allows to increase compression ratio, if we compress that data. + +There are 2 types of converters: + Byte * Conv_RISC (Byte *data, SizeT size, UInt32 pc); + Byte * ConvSt_X86(Byte *data, SizeT size, UInt32 pc, UInt32 *state); +Each Converter supports 2 versions: one for encoding +and one for decoding (_Enc/_Dec postfixes in function name). - Returns: - The number of processed bytes. If you call these functions with multiple calls, - you must start next call with first byte after block of processed bytes. +In params: + data : data buffer + size : size of data + pc : current virtual Program Counter (Instruction Pinter) value +In/Out param: + state : pointer to state variable (for X86 converter only) + +Return: + The pointer to position in (data) buffer after last byte that was processed. + If the caller calls converter again, it must call it starting with that position. + But the caller is allowed to move data in buffer. so pointer to + current processed position also will be changed for next call. + Also the caller must increase internal (pc) value for next call. +Each converter has some characteristics: Endian, Alignment, LookAhead. Type Endian Alignment LookAhead - x86 little 1 4 + X86 little 1 4 ARMT little 2 2 ARM little 4 0 + ARM64 little 4 0 PPC big 4 0 SPARC big 4 0 IA64 little 16 0 - size must be >= Alignment + LookAhead, if it's not last block. - If (size < Alignment + LookAhead), converter returns 0. - - Example: + (data) must be aligned for (Alignment). + processed size can be calculated as: + SizeT processed = Conv(data, size, pc) - data; + if (processed == 0) + it means that converter needs more data for processing. + If (size < Alignment + LookAhead) + then (processed == 0) is allowed. - UInt32 ip = 0; - for () - { - ; size must be >= Alignment + LookAhead, if it's not last block - SizeT processed = Convert(data, size, ip, 1); - data += processed; - size -= processed; - ip += processed; - } +Example code for conversion in loop: + UInt32 pc = 0; + size = 0; + for (;;) + { + size += Load_more_input_data(data + size); + SizeT processed = Conv(data, size, pc) - data; + if (processed == 0 && no_more_input_data_after_size) + break; // we stop convert loop + data += processed; + size -= processed; + pc += processed; + } */ -#define x86_Convert_Init(state) { state = 0; } -SizeT x86_Convert(Byte *data, SizeT size, UInt32 ip, UInt32 *state, int encoding); -SizeT ARM_Convert(Byte *data, SizeT size, UInt32 ip, int encoding); -SizeT ARMT_Convert(Byte *data, SizeT size, UInt32 ip, int encoding); -SizeT PPC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding); -SizeT SPARC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding); -SizeT IA64_Convert(Byte *data, SizeT size, UInt32 ip, int encoding); - EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/Bra86.c b/3rdparty/lzma/C/Bra86.c index 6db15e7ecd3..d81f392ae04 100644 --- a/3rdparty/lzma/C/Bra86.c +++ b/3rdparty/lzma/C/Bra86.c @@ -1,82 +1,187 @@ -/* Bra86.c -- Converter for x86 code (BCJ) -2013-11-12 : Igor Pavlov : Public domain */ +/* Bra86.c -- Branch converter for X86 code (BCJ) +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" #include "Bra.h" +#include "CpuArch.h" -#define Test86MSByte(b) ((((b) + 1) & 0xFE) == 0) -SizeT x86_Convert(Byte *data, SizeT size, UInt32 ip, UInt32 *state, int encoding) +#if defined(MY_CPU_SIZEOF_POINTER) \ + && ( MY_CPU_SIZEOF_POINTER == 4 \ + || MY_CPU_SIZEOF_POINTER == 8) + #define BR_CONV_USE_OPT_PC_PTR +#endif + +#ifdef BR_CONV_USE_OPT_PC_PTR +#define BR_PC_INIT pc -= (UInt32)(SizeT)p; // (MY_uintptr_t) +#define BR_PC_GET (pc + (UInt32)(SizeT)p) +#else +#define BR_PC_INIT pc += (UInt32)size; +#define BR_PC_GET (pc - (UInt32)(SizeT)(lim - p)) +// #define BR_PC_INIT +// #define BR_PC_GET (pc + (UInt32)(SizeT)(p - data)) +#endif + +#define BR_CONVERT_VAL(v, c) if (encoding) v += c; else v -= c; +// #define BR_CONVERT_VAL(v, c) if (!encoding) c = (UInt32)0 - c; v += c; + +#define Z7_BRANCH_CONV_ST(name) z7_BranchConvSt_ ## name + +#define BR86_NEED_CONV_FOR_MS_BYTE(b) ((((b) + 1) & 0xfe) == 0) + +#ifdef MY_CPU_LE_UNALIGN + #define BR86_PREPARE_BCJ_SCAN const UInt32 v = GetUi32(p) ^ 0xe8e8e8e8; + #define BR86_IS_BCJ_BYTE(n) ((v & ((UInt32)0xfe << (n) * 8)) == 0) +#else + #define BR86_PREPARE_BCJ_SCAN + // bad for MSVC X86 (partial write to byte reg): + #define BR86_IS_BCJ_BYTE(n) ((p[n - 4] & 0xfe) == 0xe8) + // bad for old MSVC (partial write to byte reg): + // #define BR86_IS_BCJ_BYTE(n) (((*p ^ 0xe8) & 0xfe) == 0) +#endif + +static +Z7_FORCE_INLINE +Z7_ATTRIB_NO_VECTOR +Byte *Z7_BRANCH_CONV_ST(X86)(Byte *p, SizeT size, UInt32 pc, UInt32 *state, int encoding) { - SizeT pos = 0; - UInt32 mask = *state & 7; if (size < 5) - return 0; - size -= 4; - ip += 5; + return p; + { + // Byte *p = data; + const Byte *lim = p + size - 4; + unsigned mask = (unsigned)*state; // & 7; +#ifdef BR_CONV_USE_OPT_PC_PTR + /* if BR_CONV_USE_OPT_PC_PTR is defined: we need to adjust (pc) for (+4), + because call/jump offset is relative to the next instruction. + if BR_CONV_USE_OPT_PC_PTR is not defined : we don't need to adjust (pc) for (+4), + because BR_PC_GET uses (pc - (lim - p)), and lim was adjusted for (-4) before. + */ + pc += 4; +#endif + BR_PC_INIT + goto start; - for (;;) + for (;; mask |= 4) { - Byte *p = data + pos; - const Byte *limit = data + size; - for (; p < limit; p++) - if ((*p & 0xFE) == 0xE8) - break; - + // cont: mask |= 4; + start: + if (p >= lim) + goto fin; { - SizeT d = (SizeT)(p - data - pos); - pos = (SizeT)(p - data); - if (p >= limit) - { - *state = (d > 2 ? 0 : mask >> (unsigned)d); - return pos; - } - if (d > 2) - mask = 0; - else - { - mask >>= (unsigned)d; - if (mask != 0 && (mask > 4 || mask == 3 || Test86MSByte(p[(mask >> 1) + 1]))) - { - mask = (mask >> 1) | 4; - pos++; - continue; - } - } + BR86_PREPARE_BCJ_SCAN + p += 4; + if (BR86_IS_BCJ_BYTE(0)) { goto m0; } mask >>= 1; + if (BR86_IS_BCJ_BYTE(1)) { goto m1; } mask >>= 1; + if (BR86_IS_BCJ_BYTE(2)) { goto m2; } mask = 0; + if (BR86_IS_BCJ_BYTE(3)) { goto a3; } } + goto main_loop; - if (Test86MSByte(p[4])) + m0: p--; + m1: p--; + m2: p--; + if (mask == 0) + goto a3; + if (p > lim) + goto fin_p; + + // if (((0x17u >> mask) & 1) == 0) + if (mask > 4 || mask == 3) + { + mask >>= 1; + continue; // goto cont; + } + mask >>= 1; + if (BR86_NEED_CONV_FOR_MS_BYTE(p[mask])) + continue; // goto cont; + // if (!BR86_NEED_CONV_FOR_MS_BYTE(p[3])) continue; // goto cont; { - UInt32 v = ((UInt32)p[4] << 24) | ((UInt32)p[3] << 16) | ((UInt32)p[2] << 8) | ((UInt32)p[1]); - UInt32 cur = ip + (UInt32)pos; - pos += 5; - if (encoding) - v += cur; - else - v -= cur; - if (mask != 0) + UInt32 v = GetUi32(p); + UInt32 c; + v += (1 << 24); if (v & 0xfe000000) continue; // goto cont; + c = BR_PC_GET; + BR_CONVERT_VAL(v, c) { - unsigned sh = (mask & 6) << 2; - if (Test86MSByte((Byte)(v >> sh))) + mask <<= 3; + if (BR86_NEED_CONV_FOR_MS_BYTE(v >> mask)) { - v ^= (((UInt32)0x100 << sh) - 1); - if (encoding) - v += cur; - else - v -= cur; + v ^= (((UInt32)0x100 << mask) - 1); + #ifdef MY_CPU_X86 + // for X86 : we can recalculate (c) to reduce register pressure + c = BR_PC_GET; + #endif + BR_CONVERT_VAL(v, c) } mask = 0; } - p[1] = (Byte)v; - p[2] = (Byte)(v >> 8); - p[3] = (Byte)(v >> 16); - p[4] = (Byte)(0 - ((v >> 24) & 1)); + // v = (v & ((1 << 24) - 1)) - (v & (1 << 24)); + v &= (1 << 25) - 1; v -= (1 << 24); + SetUi32(p, v) + p += 4; + goto main_loop; } - else + + main_loop: + if (p >= lim) + goto fin; + for (;;) { - mask = (mask >> 1) | 4; - pos++; + BR86_PREPARE_BCJ_SCAN + p += 4; + if (BR86_IS_BCJ_BYTE(0)) { goto a0; } + if (BR86_IS_BCJ_BYTE(1)) { goto a1; } + if (BR86_IS_BCJ_BYTE(2)) { goto a2; } + if (BR86_IS_BCJ_BYTE(3)) { goto a3; } + if (p >= lim) + goto fin; + } + + a0: p--; + a1: p--; + a2: p--; + a3: + if (p > lim) + goto fin_p; + // if (!BR86_NEED_CONV_FOR_MS_BYTE(p[3])) continue; // goto cont; + { + UInt32 v = GetUi32(p); + UInt32 c; + v += (1 << 24); if (v & 0xfe000000) continue; // goto cont; + c = BR_PC_GET; + BR_CONVERT_VAL(v, c) + // v = (v & ((1 << 24) - 1)) - (v & (1 << 24)); + v &= (1 << 25) - 1; v -= (1 << 24); + SetUi32(p, v) + p += 4; + goto main_loop; } } + +fin_p: + p--; +fin: + // the following processing for tail is optional and can be commented + /* + lim += 4; + for (; p < lim; p++, mask >>= 1) + if ((*p & 0xfe) == 0xe8) + break; + */ + *state = (UInt32)mask; + return p; + } } + + +#define Z7_BRANCH_CONV_ST_FUNC_IMP(name, m, encoding) \ +Z7_NO_INLINE \ +Z7_ATTRIB_NO_VECTOR \ +Byte *m(name)(Byte *data, SizeT size, UInt32 pc, UInt32 *state) \ + { return Z7_BRANCH_CONV_ST(name)(data, size, pc, state, encoding); } + +Z7_BRANCH_CONV_ST_FUNC_IMP(X86, Z7_BRANCH_CONV_ST_DEC, 0) +#ifndef Z7_EXTRACT_ONLY +Z7_BRANCH_CONV_ST_FUNC_IMP(X86, Z7_BRANCH_CONV_ST_ENC, 1) +#endif diff --git a/3rdparty/lzma/C/BraIA64.c b/3rdparty/lzma/C/BraIA64.c index fa60356b288..9dfe3e28958 100644 --- a/3rdparty/lzma/C/BraIA64.c +++ b/3rdparty/lzma/C/BraIA64.c @@ -1,69 +1,14 @@ /* BraIA64.c -- Converter for IA-64 code -2013-11-12 : Igor Pavlov : Public domain */ +2023-02-20 : Igor Pavlov : Public domain */ #include "Precomp.h" -#include "Bra.h" +// the code was moved to Bra.c -static const Byte kBranchTable[32] = -{ - 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, - 4, 4, 6, 6, 0, 0, 7, 7, - 4, 4, 0, 0, 4, 4, 0, 0 -}; +#ifdef _MSC_VER +#pragma warning(disable : 4206) // nonstandard extension used : translation unit is empty +#endif -SizeT IA64_Convert(Byte *data, SizeT size, UInt32 ip, int encoding) -{ - SizeT i; - if (size < 16) - return 0; - size -= 16; - for (i = 0; i <= size; i += 16) - { - UInt32 instrTemplate = data[i] & 0x1F; - UInt32 mask = kBranchTable[instrTemplate]; - UInt32 bitPos = 5; - int slot; - for (slot = 0; slot < 3; slot++, bitPos += 41) - { - UInt32 bytePos, bitRes; - UInt64 instruction, instNorm; - int j; - if (((mask >> slot) & 1) == 0) - continue; - bytePos = (bitPos >> 3); - bitRes = bitPos & 0x7; - instruction = 0; - for (j = 0; j < 6; j++) - instruction += (UInt64)data[i + j + bytePos] << (8 * j); - - instNorm = instruction >> bitRes; - if (((instNorm >> 37) & 0xF) == 0x5 && ((instNorm >> 9) & 0x7) == 0) - { - UInt32 src = (UInt32)((instNorm >> 13) & 0xFFFFF); - UInt32 dest; - src |= ((UInt32)(instNorm >> 36) & 1) << 20; - - src <<= 4; - - if (encoding) - dest = ip + (UInt32)i + src; - else - dest = src - (ip + (UInt32)i); - - dest >>= 4; - - instNorm &= ~((UInt64)(0x8FFFFF) << 13); - instNorm |= ((UInt64)(dest & 0xFFFFF) << 13); - instNorm |= ((UInt64)(dest & 0x100000) << (36 - 20)); - - instruction &= (1 << bitRes) - 1; - instruction |= (instNorm << bitRes); - for (j = 0; j < 6; j++) - data[i + j + bytePos] = (Byte)(instruction >> (8 * j)); - } - } - } - return i; -} +#if defined(__clang__) +#pragma GCC diagnostic ignored "-Wempty-translation-unit" +#endif diff --git a/3rdparty/lzma/C/Compiler.h b/3rdparty/lzma/C/Compiler.h index 5bba7ee5618..185a52deb50 100644 --- a/3rdparty/lzma/C/Compiler.h +++ b/3rdparty/lzma/C/Compiler.h @@ -1,8 +1,37 @@ -/* Compiler.h -2015-08-02 : Igor Pavlov : Public domain */ +/* Compiler.h : Compiler specific defines and pragmas +2023-04-02 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_COMPILER_H +#define ZIP7_INC_COMPILER_H + +#if defined(__clang__) +# define Z7_CLANG_VERSION (__clang_major__ * 10000 + __clang_minor__ * 100 + __clang_patchlevel__) +#endif +#if defined(__clang__) && defined(__apple_build_version__) +# define Z7_APPLE_CLANG_VERSION Z7_CLANG_VERSION +#elif defined(__clang__) +# define Z7_LLVM_CLANG_VERSION Z7_CLANG_VERSION +#elif defined(__GNUC__) +# define Z7_GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) +#endif + +#ifdef _MSC_VER +#if !defined(__clang__) && !defined(__GNUC__) +#define Z7_MSC_VER_ORIGINAL _MSC_VER +#endif +#endif + +#if defined(__MINGW32__) || defined(__MINGW64__) +#define Z7_MINGW +#endif + +// #pragma GCC diagnostic ignored "-Wunknown-pragmas" + +#ifdef __clang__ +// padding size of '' with 4 bytes to alignment boundary +#pragma GCC diagnostic ignored "-Wpadded" +#endif -#ifndef __7Z_COMPILER_H -#define __7Z_COMPILER_H #ifdef _MSC_VER @@ -13,17 +42,115 @@ #pragma warning(disable : 4214) // nonstandard extension used : bit field types other than int #endif - #if _MSC_VER >= 1300 - #pragma warning(disable : 4996) // This function or variable may be unsafe - #else - #pragma warning(disable : 4511) // copy constructor could not be generated - #pragma warning(disable : 4512) // assignment operator could not be generated - #pragma warning(disable : 4514) // unreferenced inline function has been removed - #pragma warning(disable : 4702) // unreachable code - #pragma warning(disable : 4710) // not inlined - #pragma warning(disable : 4786) // identifier was truncated to '255' characters in the debug information - #endif +#if defined(_MSC_VER) && _MSC_VER >= 1800 +#pragma warning(disable : 4464) // relative include path contains '..' +#endif + +// == 1200 : -O1 : for __forceinline +// >= 1900 : -O1 : for printf +#pragma warning(disable : 4710) // function not inlined + +#if _MSC_VER < 1900 +// winnt.h: 'Int64ShllMod32' +#pragma warning(disable : 4514) // unreferenced inline function has been removed +#endif + +#if _MSC_VER < 1300 +// #pragma warning(disable : 4702) // unreachable code +// Bra.c : -O1: +#pragma warning(disable : 4714) // function marked as __forceinline not inlined +#endif + +/* +#if _MSC_VER > 1400 && _MSC_VER <= 1900 +// strcat: This function or variable may be unsafe +// sysinfoapi.h: kit10: GetVersion was declared deprecated +#pragma warning(disable : 4996) +#endif +*/ + +#if _MSC_VER > 1200 +// -Wall warnings + +#pragma warning(disable : 4711) // function selected for automatic inline expansion +#pragma warning(disable : 4820) // '2' bytes padding added after data member + +#if _MSC_VER >= 1400 && _MSC_VER < 1920 +// 1400: string.h: _DBG_MEMCPY_INLINE_ +// 1600 - 191x : smmintrin.h __cplusplus' +// is not defined as a preprocessor macro, replacing with '0' for '#if/#elif' +#pragma warning(disable : 4668) + +// 1400 - 1600 : WinDef.h : 'FARPROC' : +// 1900 - 191x : immintrin.h: _readfsbase_u32 +// no function prototype given : converting '()' to '(void)' +#pragma warning(disable : 4255) +#endif + +#if _MSC_VER >= 1914 +// Compiler will insert Spectre mitigation for memory load if /Qspectre switch specified +#pragma warning(disable : 5045) +#endif + +#endif // _MSC_VER > 1200 +#endif // _MSC_VER + + +#if defined(__clang__) && (__clang_major__ >= 4) + #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE \ + _Pragma("clang loop unroll(disable)") \ + _Pragma("clang loop vectorize(disable)") + #define Z7_ATTRIB_NO_VECTORIZE +#elif defined(__GNUC__) && (__GNUC__ >= 5) + #define Z7_ATTRIB_NO_VECTORIZE __attribute__((optimize("no-tree-vectorize"))) + // __attribute__((optimize("no-unroll-loops"))); + #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE +#elif defined(_MSC_VER) && (_MSC_VER >= 1920) + #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE \ + _Pragma("loop( no_vector )") + #define Z7_ATTRIB_NO_VECTORIZE +#else + #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + #define Z7_ATTRIB_NO_VECTORIZE +#endif + +#if defined(MY_CPU_X86_OR_AMD64) && ( \ + defined(__clang__) && (__clang_major__ >= 4) \ + || defined(__GNUC__) && (__GNUC__ >= 5)) + #define Z7_ATTRIB_NO_SSE __attribute__((__target__("no-sse"))) +#else + #define Z7_ATTRIB_NO_SSE +#endif + +#define Z7_ATTRIB_NO_VECTOR \ + Z7_ATTRIB_NO_VECTORIZE \ + Z7_ATTRIB_NO_SSE + + +#if defined(__clang__) && (__clang_major__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ >= 1000) \ + /* || defined(_MSC_VER) && (_MSC_VER >= 1920) */ + // GCC is not good for __builtin_expect() + #define Z7_LIKELY(x) (__builtin_expect((x), 1)) + #define Z7_UNLIKELY(x) (__builtin_expect((x), 0)) + // #define Z7_unlikely [[unlikely]] + // #define Z7_likely [[likely]] +#else + #define Z7_LIKELY(x) (x) + #define Z7_UNLIKELY(x) (x) + // #define Z7_likely +#endif + +#if (defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 36000)) +#define Z7_DIAGNOSCTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER \ + _Pragma("GCC diagnostic push") \ + _Pragma("GCC diagnostic ignored \"-Wreserved-macro-identifier\"") +#define Z7_DIAGNOSCTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER \ + _Pragma("GCC diagnostic pop") +#else +#define Z7_DIAGNOSCTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER +#define Z7_DIAGNOSCTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER #endif #define UNUSED_VAR(x) (void)x; diff --git a/3rdparty/lzma/C/CpuArch.c b/3rdparty/lzma/C/CpuArch.c index 554ffa4fc7b..33f8a3ab4c1 100644 --- a/3rdparty/lzma/C/CpuArch.c +++ b/3rdparty/lzma/C/CpuArch.c @@ -1,200 +1,823 @@ /* CpuArch.c -- CPU specific code -2016-02-25: Igor Pavlov : Public domain */ +2023-05-18 : Igor Pavlov : Public domain */ #include "Precomp.h" +// #include <stdio.h> + #include "CpuArch.h" #ifdef MY_CPU_X86_OR_AMD64 -#if (defined(_MSC_VER) && !defined(MY_CPU_AMD64)) || defined(__GNUC__) -#define USE_ASM +#undef NEED_CHECK_FOR_CPUID +#if !defined(MY_CPU_AMD64) +#define NEED_CHECK_FOR_CPUID #endif -#if !defined(USE_ASM) && _MSC_VER >= 1500 -#include <intrin.h> +/* + cpuid instruction supports (subFunction) parameter in ECX, + that is used only with some specific (function) parameter values. + But we always use only (subFunction==0). +*/ +/* + __cpuid(): MSVC and GCC/CLANG use same function/macro name + but parameters are different. + We use MSVC __cpuid() parameters style for our z7_x86_cpuid() function. +*/ + +#if defined(__GNUC__) /* && (__GNUC__ >= 10) */ \ + || defined(__clang__) /* && (__clang_major__ >= 10) */ + +/* there was some CLANG/GCC compilers that have issues with + rbx(ebx) handling in asm blocks in -fPIC mode (__PIC__ is defined). + compiler's <cpuid.h> contains the macro __cpuid() that is similar to our code. + The history of __cpuid() changes in CLANG/GCC: + GCC: + 2007: it preserved ebx for (__PIC__ && __i386__) + 2013: it preserved rbx and ebx for __PIC__ + 2014: it doesn't preserves rbx and ebx anymore + we suppose that (__GNUC__ >= 5) fixed that __PIC__ ebx/rbx problem. + CLANG: + 2014+: it preserves rbx, but only for 64-bit code. No __PIC__ check. + Why CLANG cares about 64-bit mode only, and doesn't care about ebx (in 32-bit)? + Do we need __PIC__ test for CLANG or we must care about rbx even if + __PIC__ is not defined? +*/ + +#define ASM_LN "\n" + +#if defined(MY_CPU_AMD64) && defined(__PIC__) \ + && ((defined (__GNUC__) && (__GNUC__ < 5)) || defined(__clang__)) + +#define x86_cpuid_MACRO(p, func) { \ + __asm__ __volatile__ ( \ + ASM_LN "mov %%rbx, %q1" \ + ASM_LN "cpuid" \ + ASM_LN "xchg %%rbx, %q1" \ + : "=a" ((p)[0]), "=&r" ((p)[1]), "=c" ((p)[2]), "=d" ((p)[3]) : "0" (func), "2"(0)); } + + /* "=&r" selects free register. It can select even rbx, if that register is free. + "=&D" for (RDI) also works, but the code can be larger with "=&D" + "2"(0) means (subFunction = 0), + 2 is (zero-based) index in the output constraint list "=c" (ECX). */ + +#elif defined(MY_CPU_X86) && defined(__PIC__) \ + && ((defined (__GNUC__) && (__GNUC__ < 5)) || defined(__clang__)) + +#define x86_cpuid_MACRO(p, func) { \ + __asm__ __volatile__ ( \ + ASM_LN "mov %%ebx, %k1" \ + ASM_LN "cpuid" \ + ASM_LN "xchg %%ebx, %k1" \ + : "=a" ((p)[0]), "=&r" ((p)[1]), "=c" ((p)[2]), "=d" ((p)[3]) : "0" (func), "2"(0)); } + +#else + +#define x86_cpuid_MACRO(p, func) { \ + __asm__ __volatile__ ( \ + ASM_LN "cpuid" \ + : "=a" ((p)[0]), "=b" ((p)[1]), "=c" ((p)[2]), "=d" ((p)[3]) : "0" (func), "2"(0)); } + #endif -#if defined(USE_ASM) && !defined(MY_CPU_AMD64) -static UInt32 CheckFlag(UInt32 flag) + +void Z7_FASTCALL z7_x86_cpuid(UInt32 p[4], UInt32 func) { - #ifdef _MSC_VER - __asm pushfd; - __asm pop EAX; - __asm mov EDX, EAX; - __asm xor EAX, flag; - __asm push EAX; - __asm popfd; - __asm pushfd; - __asm pop EAX; - __asm xor EAX, EDX; - __asm push EDX; - __asm popfd; - __asm and flag, EAX; - #else - __asm__ __volatile__ ( - "pushf\n\t" - "pop %%EAX\n\t" - "movl %%EAX,%%EDX\n\t" - "xorl %0,%%EAX\n\t" - "push %%EAX\n\t" - "popf\n\t" - "pushf\n\t" - "pop %%EAX\n\t" - "xorl %%EDX,%%EAX\n\t" - "push %%EDX\n\t" - "popf\n\t" - "andl %%EAX, %0\n\t": - "=c" (flag) : "c" (flag) : - "%eax", "%edx"); - #endif - return flag; + x86_cpuid_MACRO(p, func) } -#define CHECK_CPUID_IS_SUPPORTED if (CheckFlag(1 << 18) == 0 || CheckFlag(1 << 21) == 0) return False; -#else -#define CHECK_CPUID_IS_SUPPORTED -#endif -void MyCPUID(UInt32 function, UInt32 *a, UInt32 *b, UInt32 *c, UInt32 *d) + +Z7_NO_INLINE +UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void) { - #ifdef USE_ASM + #if defined(NEED_CHECK_FOR_CPUID) + #define EFALGS_CPUID_BIT 21 + UInt32 a; + __asm__ __volatile__ ( + ASM_LN "pushf" + ASM_LN "pushf" + ASM_LN "pop %0" + // ASM_LN "movl %0, %1" + // ASM_LN "xorl $0x200000, %0" + ASM_LN "btc %1, %0" + ASM_LN "push %0" + ASM_LN "popf" + ASM_LN "pushf" + ASM_LN "pop %0" + ASM_LN "xorl (%%esp), %0" - #ifdef _MSC_VER + ASM_LN "popf" + ASM_LN + : "=&r" (a) // "=a" + : "i" (EFALGS_CPUID_BIT) + ); + if ((a & (1 << EFALGS_CPUID_BIT)) == 0) + return 0; + #endif + { + UInt32 p[4]; + x86_cpuid_MACRO(p, 0) + return p[0]; + } +} - UInt32 a2, b2, c2, d2; - __asm xor EBX, EBX; - __asm xor ECX, ECX; - __asm xor EDX, EDX; - __asm mov EAX, function; - __asm cpuid; - __asm mov a2, EAX; - __asm mov b2, EBX; - __asm mov c2, ECX; - __asm mov d2, EDX; +#undef ASM_LN - *a = a2; - *b = b2; - *c = c2; - *d = d2; +#elif !defined(_MSC_VER) - #else +/* +// for gcc/clang and other: we can try to use __cpuid macro: +#include <cpuid.h> +void Z7_FASTCALL z7_x86_cpuid(UInt32 p[4], UInt32 func) +{ + __cpuid(func, p[0], p[1], p[2], p[3]); +} +UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void) +{ + return (UInt32)__get_cpuid_max(0, NULL); +} +*/ +// for unsupported cpuid: +void Z7_FASTCALL z7_x86_cpuid(UInt32 p[4], UInt32 func) +{ + UNUSED_VAR(func) + p[0] = p[1] = p[2] = p[3] = 0; +} +UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void) +{ + return 0; +} - __asm__ __volatile__ ( - #if defined(MY_CPU_AMD64) && defined(__PIC__) - "mov %%rbx, %%rdi;" - "cpuid;" - "xchg %%rbx, %%rdi;" - : "=a" (*a) , - "=D" (*b) , - #elif defined(MY_CPU_X86) && defined(__PIC__) - "mov %%ebx, %%edi;" - "cpuid;" - "xchgl %%ebx, %%edi;" - : "=a" (*a) , - "=D" (*b) , - #else - "cpuid" - : "=a" (*a) , - "=b" (*b) , - #endif - "=c" (*c) , - "=d" (*d) - : "0" (function)) ; +#else // _MSC_VER + +#if !defined(MY_CPU_AMD64) +UInt32 __declspec(naked) Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void) +{ + #if defined(NEED_CHECK_FOR_CPUID) + #define EFALGS_CPUID_BIT 21 + __asm pushfd + __asm pushfd + /* + __asm pop eax + // __asm mov edx, eax + __asm btc eax, EFALGS_CPUID_BIT + __asm push eax + */ + __asm btc dword ptr [esp], EFALGS_CPUID_BIT + __asm popfd + __asm pushfd + __asm pop eax + // __asm xor eax, edx + __asm xor eax, [esp] + // __asm push edx + __asm popfd + __asm and eax, (1 shl EFALGS_CPUID_BIT) + __asm jz end_func #endif - - #else + __asm push ebx + __asm xor eax, eax // func + __asm xor ecx, ecx // subFunction (optional) for (func == 0) + __asm cpuid + __asm pop ebx + #if defined(NEED_CHECK_FOR_CPUID) + end_func: + #endif + __asm ret 0 +} - int CPUInfo[4]; - __cpuid(CPUInfo, function); - *a = CPUInfo[0]; - *b = CPUInfo[1]; - *c = CPUInfo[2]; - *d = CPUInfo[3]; +void __declspec(naked) Z7_FASTCALL z7_x86_cpuid(UInt32 p[4], UInt32 func) +{ + UNUSED_VAR(p) + UNUSED_VAR(func) + __asm push ebx + __asm push edi + __asm mov edi, ecx // p + __asm mov eax, edx // func + __asm xor ecx, ecx // subfunction (optional) for (func == 0) + __asm cpuid + __asm mov [edi ], eax + __asm mov [edi + 4], ebx + __asm mov [edi + 8], ecx + __asm mov [edi + 12], edx + __asm pop edi + __asm pop ebx + __asm ret 0 +} - #endif +#else // MY_CPU_AMD64 + + #if _MSC_VER >= 1600 + #include <intrin.h> + #define MY_cpuidex __cpuidex + #else +/* + __cpuid (func == (0 or 7)) requires subfunction number in ECX. + MSDN: The __cpuid intrinsic clears the ECX register before calling the cpuid instruction. + __cpuid() in new MSVC clears ECX. + __cpuid() in old MSVC (14.00) x64 doesn't clear ECX + We still can use __cpuid for low (func) values that don't require ECX, + but __cpuid() in old MSVC will be incorrect for some func values: (func == 7). + So here we use the hack for old MSVC to send (subFunction) in ECX register to cpuid instruction, + where ECX value is first parameter for FASTCALL / NO_INLINE func, + So the caller of MY_cpuidex_HACK() sets ECX as subFunction, and + old MSVC for __cpuid() doesn't change ECX and cpuid instruction gets (subFunction) value. + +DON'T remove Z7_NO_INLINE and Z7_FASTCALL for MY_cpuidex_HACK(): !!! +*/ +static +Z7_NO_INLINE void Z7_FASTCALL MY_cpuidex_HACK(UInt32 subFunction, UInt32 func, int *CPUInfo) +{ + UNUSED_VAR(subFunction) + __cpuid(CPUInfo, func); } + #define MY_cpuidex(info, func, func2) MY_cpuidex_HACK(func2, func, info) + #pragma message("======== MY_cpuidex_HACK WAS USED ========") + #endif // _MSC_VER >= 1600 -Bool x86cpuid_CheckAndRead(Cx86cpuid *p) +#if !defined(MY_CPU_AMD64) +/* inlining for __cpuid() in MSVC x86 (32-bit) produces big ineffective code, + so we disable inlining here */ +Z7_NO_INLINE +#endif +void Z7_FASTCALL z7_x86_cpuid(UInt32 p[4], UInt32 func) +{ + MY_cpuidex((int *)p, (int)func, 0); +} + +Z7_NO_INLINE +UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void) +{ + int a[4]; + MY_cpuidex(a, 0, 0); + return a[0]; +} + +#endif // MY_CPU_AMD64 +#endif // _MSC_VER + +#if defined(NEED_CHECK_FOR_CPUID) +#define CHECK_CPUID_IS_SUPPORTED { if (z7_x86_cpuid_GetMaxFunc() == 0) return 0; } +#else +#define CHECK_CPUID_IS_SUPPORTED +#endif +#undef NEED_CHECK_FOR_CPUID + + +static +BoolInt x86cpuid_Func_1(UInt32 *p) { CHECK_CPUID_IS_SUPPORTED - MyCPUID(0, &p->maxFunc, &p->vendor[0], &p->vendor[2], &p->vendor[1]); - MyCPUID(1, &p->ver, &p->b, &p->c, &p->d); + z7_x86_cpuid(p, 1); return True; } -static const UInt32 kVendors[][3] = +/* +static const UInt32 kVendors[][1] = { - { 0x756E6547, 0x49656E69, 0x6C65746E}, - { 0x68747541, 0x69746E65, 0x444D4163}, - { 0x746E6543, 0x48727561, 0x736C7561} + { 0x756E6547 }, // , 0x49656E69, 0x6C65746E }, + { 0x68747541 }, // , 0x69746E65, 0x444D4163 }, + { 0x746E6543 } // , 0x48727561, 0x736C7561 } }; +*/ + +/* +typedef struct +{ + UInt32 maxFunc; + UInt32 vendor[3]; + UInt32 ver; + UInt32 b; + UInt32 c; + UInt32 d; +} Cx86cpuid; + +enum +{ + CPU_FIRM_INTEL, + CPU_FIRM_AMD, + CPU_FIRM_VIA +}; +int x86cpuid_GetFirm(const Cx86cpuid *p); +#define x86cpuid_ver_GetFamily(ver) (((ver >> 16) & 0xff0) | ((ver >> 8) & 0xf)) +#define x86cpuid_ver_GetModel(ver) (((ver >> 12) & 0xf0) | ((ver >> 4) & 0xf)) +#define x86cpuid_ver_GetStepping(ver) (ver & 0xf) int x86cpuid_GetFirm(const Cx86cpuid *p) { unsigned i; - for (i = 0; i < sizeof(kVendors) / sizeof(kVendors[i]); i++) + for (i = 0; i < sizeof(kVendors) / sizeof(kVendors[0]); i++) { const UInt32 *v = kVendors[i]; - if (v[0] == p->vendor[0] && - v[1] == p->vendor[1] && - v[2] == p->vendor[2]) + if (v[0] == p->vendor[0] + // && v[1] == p->vendor[1] + // && v[2] == p->vendor[2] + ) return (int)i; } return -1; } -Bool CPU_Is_InOrder() +BoolInt CPU_Is_InOrder() { Cx86cpuid p; - int firm; UInt32 family, model; if (!x86cpuid_CheckAndRead(&p)) return True; - family = x86cpuid_GetFamily(p.ver); - model = x86cpuid_GetModel(p.ver); - - firm = x86cpuid_GetFirm(&p); + family = x86cpuid_ver_GetFamily(p.ver); + model = x86cpuid_ver_GetModel(p.ver); - switch (firm) + switch (x86cpuid_GetFirm(&p)) { case CPU_FIRM_INTEL: return (family < 6 || (family == 6 && ( - /* In-Order Atom CPU */ - model == 0x1C /* 45 nm, N4xx, D4xx, N5xx, D5xx, 230, 330 */ - || model == 0x26 /* 45 nm, Z6xx */ - || model == 0x27 /* 32 nm, Z2460 */ - || model == 0x35 /* 32 nm, Z2760 */ - || model == 0x36 /* 32 nm, N2xxx, D2xxx */ + // In-Order Atom CPU + model == 0x1C // 45 nm, N4xx, D4xx, N5xx, D5xx, 230, 330 + || model == 0x26 // 45 nm, Z6xx + || model == 0x27 // 32 nm, Z2460 + || model == 0x35 // 32 nm, Z2760 + || model == 0x36 // 32 nm, N2xxx, D2xxx ))); case CPU_FIRM_AMD: return (family < 5 || (family == 5 && (model < 6 || model == 0xA))); case CPU_FIRM_VIA: return (family < 6 || (family == 6 && model < 0xF)); } - return True; + return False; // v23 : unknown processors are not In-Order } +*/ + +#ifdef _WIN32 +#include "7zWindows.h" +#endif #if !defined(MY_CPU_AMD64) && defined(_WIN32) -#include <windows.h> -static Bool CPU_Sys_Is_SSE_Supported() + +/* for legacy SSE ia32: there is no user-space cpu instruction to check + that OS supports SSE register storing/restoring on context switches. + So we need some OS-specific function to check that it's safe to use SSE registers. +*/ + +Z7_FORCE_INLINE +static BoolInt CPU_Sys_Is_SSE_Supported(void) { - OSVERSIONINFO vi; - vi.dwOSVersionInfoSize = sizeof(vi); - if (!GetVersionEx(&vi)) - return False; - return (vi.dwMajorVersion >= 5); +#ifdef _MSC_VER + #pragma warning(push) + #pragma warning(disable : 4996) // `GetVersion': was declared deprecated +#endif + /* low byte is major version of Windows + We suppose that any Windows version since + Windows2000 (major == 5) supports SSE registers */ + return (Byte)GetVersion() >= 5; +#if defined(_MSC_VER) + #pragma warning(pop) +#endif } #define CHECK_SYS_SSE_SUPPORT if (!CPU_Sys_Is_SSE_Supported()) return False; #else #define CHECK_SYS_SSE_SUPPORT #endif -Bool CPU_Is_Aes_Supported() + +#if !defined(MY_CPU_AMD64) + +BoolInt CPU_IsSupported_CMOV(void) { - Cx86cpuid p; + UInt32 a[4]; + if (!x86cpuid_Func_1(&a[0])) + return 0; + return (a[3] >> 15) & 1; +} + +BoolInt CPU_IsSupported_SSE(void) +{ + UInt32 a[4]; CHECK_SYS_SSE_SUPPORT - if (!x86cpuid_CheckAndRead(&p)) + if (!x86cpuid_Func_1(&a[0])) + return 0; + return (a[3] >> 25) & 1; +} + +BoolInt CPU_IsSupported_SSE2(void) +{ + UInt32 a[4]; + CHECK_SYS_SSE_SUPPORT + if (!x86cpuid_Func_1(&a[0])) + return 0; + return (a[3] >> 26) & 1; +} + +#endif + + +static UInt32 x86cpuid_Func_1_ECX(void) +{ + UInt32 a[4]; + CHECK_SYS_SSE_SUPPORT + if (!x86cpuid_Func_1(&a[0])) + return 0; + return a[2]; +} + +BoolInt CPU_IsSupported_AES(void) +{ + return (x86cpuid_Func_1_ECX() >> 25) & 1; +} + +BoolInt CPU_IsSupported_SSSE3(void) +{ + return (x86cpuid_Func_1_ECX() >> 9) & 1; +} + +BoolInt CPU_IsSupported_SSE41(void) +{ + return (x86cpuid_Func_1_ECX() >> 19) & 1; +} + +BoolInt CPU_IsSupported_SHA(void) +{ + CHECK_SYS_SSE_SUPPORT + + if (z7_x86_cpuid_GetMaxFunc() < 7) + return False; + { + UInt32 d[4]; + z7_x86_cpuid(d, 7); + return (d[1] >> 29) & 1; + } +} + +/* +MSVC: _xgetbv() intrinsic is available since VS2010SP1. + MSVC also defines (_XCR_XFEATURE_ENABLED_MASK) macro in + <immintrin.h> that we can use or check. + For any 32-bit x86 we can use asm code in MSVC, + but MSVC asm code is huge after compilation. + So _xgetbv() is better + +ICC: _xgetbv() intrinsic is available (in what version of ICC?) + ICC defines (__GNUC___) and it supports gnu assembler + also ICC supports MASM style code with -use-msasm switch. + but ICC doesn't support __attribute__((__target__)) + +GCC/CLANG 9: + _xgetbv() is macro that works via __builtin_ia32_xgetbv() + and we need __attribute__((__target__("xsave")). + But with __target__("xsave") the function will be not + inlined to function that has no __target__("xsave") attribute. + If we want _xgetbv() call inlining, then we should use asm version + instead of calling _xgetbv(). + Note:intrinsic is broke before GCC 8.2: + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85684 +*/ + +#if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1100) \ + || defined(_MSC_VER) && (_MSC_VER >= 1600) && (_MSC_FULL_VER >= 160040219) \ + || defined(__GNUC__) && (__GNUC__ >= 9) \ + || defined(__clang__) && (__clang_major__ >= 9) +// we define ATTRIB_XGETBV, if we want to use predefined _xgetbv() from compiler +#if defined(__INTEL_COMPILER) +#define ATTRIB_XGETBV +#elif defined(__GNUC__) || defined(__clang__) +// we don't define ATTRIB_XGETBV here, because asm version is better for inlining. +// #define ATTRIB_XGETBV __attribute__((__target__("xsave"))) +#else +#define ATTRIB_XGETBV +#endif +#endif + +#if defined(ATTRIB_XGETBV) +#include <immintrin.h> +#endif + + +// XFEATURE_ENABLED_MASK/XCR0 +#define MY_XCR_XFEATURE_ENABLED_MASK 0 + +#if defined(ATTRIB_XGETBV) +ATTRIB_XGETBV +#endif +static UInt64 x86_xgetbv_0(UInt32 num) +{ +#if defined(ATTRIB_XGETBV) + { + return + #if (defined(_MSC_VER)) + _xgetbv(num); + #else + __builtin_ia32_xgetbv( + #if !defined(__clang__) + (int) + #endif + num); + #endif + } + +#elif defined(__GNUC__) || defined(__clang__) || defined(__SUNPRO_CC) + + UInt32 a, d; + #if defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) + __asm__ + ( + "xgetbv" + : "=a"(a), "=d"(d) : "c"(num) : "cc" + ); + #else // is old gcc + __asm__ + ( + ".byte 0x0f, 0x01, 0xd0" "\n\t" + : "=a"(a), "=d"(d) : "c"(num) : "cc" + ); + #endif + return ((UInt64)d << 32) | a; + // return a; + +#elif defined(_MSC_VER) && !defined(MY_CPU_AMD64) + + UInt32 a, d; + __asm { + push eax + push edx + push ecx + mov ecx, num; + // xor ecx, ecx // = MY_XCR_XFEATURE_ENABLED_MASK + _emit 0x0f + _emit 0x01 + _emit 0xd0 + mov a, eax + mov d, edx + pop ecx + pop edx + pop eax + } + return ((UInt64)d << 32) | a; + // return a; + +#else // it's unknown compiler + // #error "Need xgetbv function" + UNUSED_VAR(num) + // for MSVC-X64 we could call external function from external file. + /* Actually we had checked OSXSAVE/AVX in cpuid before. + So it's expected that OS supports at least AVX and below. */ + // if (num != MY_XCR_XFEATURE_ENABLED_MASK) return 0; // if not XCR0 + return + // (1 << 0) | // x87 + (1 << 1) // SSE + | (1 << 2); // AVX + +#endif +} + +#ifdef _WIN32 +/* + Windows versions do not know about new ISA extensions that + can be introduced. But we still can use new extensions, + even if Windows doesn't report about supporting them, + But we can use new extensions, only if Windows knows about new ISA extension + that changes the number or size of registers: SSE, AVX/XSAVE, AVX512 + So it's enough to check + MY_PF_AVX_INSTRUCTIONS_AVAILABLE + instead of + MY_PF_AVX2_INSTRUCTIONS_AVAILABLE +*/ +#define MY_PF_XSAVE_ENABLED 17 +// #define MY_PF_SSSE3_INSTRUCTIONS_AVAILABLE 36 +// #define MY_PF_SSE4_1_INSTRUCTIONS_AVAILABLE 37 +// #define MY_PF_SSE4_2_INSTRUCTIONS_AVAILABLE 38 +// #define MY_PF_AVX_INSTRUCTIONS_AVAILABLE 39 +// #define MY_PF_AVX2_INSTRUCTIONS_AVAILABLE 40 +// #define MY_PF_AVX512F_INSTRUCTIONS_AVAILABLE 41 +#endif + +BoolInt CPU_IsSupported_AVX(void) +{ + #ifdef _WIN32 + if (!IsProcessorFeaturePresent(MY_PF_XSAVE_ENABLED)) + return False; + /* PF_AVX_INSTRUCTIONS_AVAILABLE probably is supported starting from + some latest Win10 revisions. But we need AVX in older Windows also. + So we don't use the following check: */ + /* + if (!IsProcessorFeaturePresent(MY_PF_AVX_INSTRUCTIONS_AVAILABLE)) return False; - return (p.c >> 25) & 1; + */ + #endif + + /* + OS must use new special XSAVE/XRSTOR instructions to save + AVX registers when it required for context switching. + At OS statring: + OS sets CR4.OSXSAVE flag to signal the processor that OS supports the XSAVE extensions. + Also OS sets bitmask in XCR0 register that defines what + registers will be processed by XSAVE instruction: + XCR0.SSE[bit 0] - x87 registers and state + XCR0.SSE[bit 1] - SSE registers and state + XCR0.AVX[bit 2] - AVX registers and state + CR4.OSXSAVE is reflected to CPUID.1:ECX.OSXSAVE[bit 27]. + So we can read that bit in user-space. + XCR0 is available for reading in user-space by new XGETBV instruction. + */ + { + const UInt32 c = x86cpuid_Func_1_ECX(); + if (0 == (1 + & (c >> 28) // AVX instructions are supported by hardware + & (c >> 27))) // OSXSAVE bit: XSAVE and related instructions are enabled by OS. + return False; + } + + /* also we can check + CPUID.1:ECX.XSAVE [bit 26] : that shows that + XSAVE, XRESTOR, XSETBV, XGETBV instructions are supported by hardware. + But that check is redundant, because if OSXSAVE bit is set, then XSAVE is also set */ + + /* If OS have enabled XSAVE extension instructions (OSXSAVE == 1), + in most cases we expect that OS also will support storing/restoring + for AVX and SSE states at least. + But to be ensure for that we call user-space instruction + XGETBV(0) to get XCR0 value that contains bitmask that defines + what exact states(registers) OS have enabled for storing/restoring. + */ + + { + const UInt32 bm = (UInt32)x86_xgetbv_0(MY_XCR_XFEATURE_ENABLED_MASK); + // printf("\n=== XGetBV=%d\n", bm); + return 1 + & (bm >> 1) // SSE state is supported (set by OS) for storing/restoring + & (bm >> 2); // AVX state is supported (set by OS) for storing/restoring + } + // since Win7SP1: we can use GetEnabledXStateFeatures(); +} + + +BoolInt CPU_IsSupported_AVX2(void) +{ + if (!CPU_IsSupported_AVX()) + return False; + if (z7_x86_cpuid_GetMaxFunc() < 7) + return False; + { + UInt32 d[4]; + z7_x86_cpuid(d, 7); + // printf("\ncpuid(7): ebx=%8x ecx=%8x\n", d[1], d[2]); + return 1 + & (d[1] >> 5); // avx2 + } +} + +BoolInt CPU_IsSupported_VAES_AVX2(void) +{ + if (!CPU_IsSupported_AVX()) + return False; + if (z7_x86_cpuid_GetMaxFunc() < 7) + return False; + { + UInt32 d[4]; + z7_x86_cpuid(d, 7); + // printf("\ncpuid(7): ebx=%8x ecx=%8x\n", d[1], d[2]); + return 1 + & (d[1] >> 5) // avx2 + // & (d[1] >> 31) // avx512vl + & (d[2] >> 9); // vaes // VEX-256/EVEX + } +} + +BoolInt CPU_IsSupported_PageGB(void) +{ + CHECK_CPUID_IS_SUPPORTED + { + UInt32 d[4]; + z7_x86_cpuid(d, 0x80000000); + if (d[0] < 0x80000001) + return False; + z7_x86_cpuid(d, 0x80000001); + return (d[3] >> 26) & 1; + } +} + + +#elif defined(MY_CPU_ARM_OR_ARM64) + +#ifdef _WIN32 + +#include "7zWindows.h" + +BoolInt CPU_IsSupported_CRC32(void) { return IsProcessorFeaturePresent(PF_ARM_V8_CRC32_INSTRUCTIONS_AVAILABLE) ? 1 : 0; } +BoolInt CPU_IsSupported_CRYPTO(void) { return IsProcessorFeaturePresent(PF_ARM_V8_CRYPTO_INSTRUCTIONS_AVAILABLE) ? 1 : 0; } +BoolInt CPU_IsSupported_NEON(void) { return IsProcessorFeaturePresent(PF_ARM_NEON_INSTRUCTIONS_AVAILABLE) ? 1 : 0; } + +#else + +#if defined(__APPLE__) + +/* +#include <stdio.h> +#include <string.h> +static void Print_sysctlbyname(const char *name) +{ + size_t bufSize = 256; + char buf[256]; + int res = sysctlbyname(name, &buf, &bufSize, NULL, 0); + { + int i; + printf("\nres = %d : %s : '%s' : bufSize = %d, numeric", res, name, buf, (unsigned)bufSize); + for (i = 0; i < 20; i++) + printf(" %2x", (unsigned)(Byte)buf[i]); + + } +} +*/ +/* + Print_sysctlbyname("hw.pagesize"); + Print_sysctlbyname("machdep.cpu.brand_string"); +*/ + +static BoolInt z7_sysctlbyname_Get_BoolInt(const char *name) +{ + UInt32 val = 0; + if (z7_sysctlbyname_Get_UInt32(name, &val) == 0 && val == 1) + return 1; + return 0; +} + +BoolInt CPU_IsSupported_CRC32(void) +{ + return z7_sysctlbyname_Get_BoolInt("hw.optional.armv8_crc32"); +} + +BoolInt CPU_IsSupported_NEON(void) +{ + return z7_sysctlbyname_Get_BoolInt("hw.optional.neon"); +} + +#ifdef MY_CPU_ARM64 +#define APPLE_CRYPTO_SUPPORT_VAL 1 +#else +#define APPLE_CRYPTO_SUPPORT_VAL 0 +#endif + +BoolInt CPU_IsSupported_SHA1(void) { return APPLE_CRYPTO_SUPPORT_VAL; } +BoolInt CPU_IsSupported_SHA2(void) { return APPLE_CRYPTO_SUPPORT_VAL; } +BoolInt CPU_IsSupported_AES (void) { return APPLE_CRYPTO_SUPPORT_VAL; } + + +#else // __APPLE__ + +#include <sys/auxv.h> + +#define USE_HWCAP + +#ifdef USE_HWCAP + +#include <asm/hwcap.h> + + #define MY_HWCAP_CHECK_FUNC_2(name1, name2) \ + BoolInt CPU_IsSupported_ ## name1() { return (getauxval(AT_HWCAP) & (HWCAP_ ## name2)) ? 1 : 0; } + +#ifdef MY_CPU_ARM64 + #define MY_HWCAP_CHECK_FUNC(name) \ + MY_HWCAP_CHECK_FUNC_2(name, name) + MY_HWCAP_CHECK_FUNC_2(NEON, ASIMD) +// MY_HWCAP_CHECK_FUNC (ASIMD) +#elif defined(MY_CPU_ARM) + #define MY_HWCAP_CHECK_FUNC(name) \ + BoolInt CPU_IsSupported_ ## name() { return (getauxval(AT_HWCAP2) & (HWCAP2_ ## name)) ? 1 : 0; } + MY_HWCAP_CHECK_FUNC_2(NEON, NEON) +#endif + +#else // USE_HWCAP + + #define MY_HWCAP_CHECK_FUNC(name) \ + BoolInt CPU_IsSupported_ ## name() { return 0; } + MY_HWCAP_CHECK_FUNC(NEON) + +#endif // USE_HWCAP + +MY_HWCAP_CHECK_FUNC (CRC32) +MY_HWCAP_CHECK_FUNC (SHA1) +MY_HWCAP_CHECK_FUNC (SHA2) +MY_HWCAP_CHECK_FUNC (AES) + +#endif // __APPLE__ +#endif // _WIN32 + +#endif // MY_CPU_ARM_OR_ARM64 + + + +#ifdef __APPLE__ + +#include <sys/sysctl.h> + +int z7_sysctlbyname_Get(const char *name, void *buf, size_t *bufSize) +{ + return sysctlbyname(name, buf, bufSize, NULL, 0); +} + +int z7_sysctlbyname_Get_UInt32(const char *name, UInt32 *val) +{ + size_t bufSize = sizeof(*val); + const int res = z7_sysctlbyname_Get(name, val, &bufSize); + if (res == 0 && bufSize != sizeof(*val)) + return EFAULT; + return res; } #endif diff --git a/3rdparty/lzma/C/CpuArch.h b/3rdparty/lzma/C/CpuArch.h index e42ce281424..8e5d8a543f8 100644 --- a/3rdparty/lzma/C/CpuArch.h +++ b/3rdparty/lzma/C/CpuArch.h @@ -1,8 +1,8 @@ /* CpuArch.h -- CPU specific code -2016-06-09: Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __CPU_ARCH_H -#define __CPU_ARCH_H +#ifndef ZIP7_INC_CPU_ARCH_H +#define ZIP7_INC_CPU_ARCH_H #include "7zTypes.h" @@ -14,50 +14,166 @@ MY_CPU_BE means that CPU is BIG ENDIAN. If MY_CPU_LE and MY_CPU_BE are not defined, we don't know about ENDIANNESS of platform. MY_CPU_LE_UNALIGN means that CPU is LITTLE ENDIAN and CPU supports unaligned memory accesses. + +MY_CPU_64BIT means that processor can work with 64-bit registers. + MY_CPU_64BIT can be used to select fast code branch + MY_CPU_64BIT doesn't mean that (sizeof(void *) == 8) */ -#if defined(_M_X64) \ - || defined(_M_AMD64) \ - || defined(__x86_64__) \ - || defined(__AMD64__) \ - || defined(__amd64__) +#if defined(_M_X64) \ + || defined(_M_AMD64) \ + || defined(__x86_64__) \ + || defined(__AMD64__) \ + || defined(__amd64__) #define MY_CPU_AMD64 + #ifdef __ILP32__ + #define MY_CPU_NAME "x32" + #define MY_CPU_SIZEOF_POINTER 4 + #else + #define MY_CPU_NAME "x64" + #define MY_CPU_SIZEOF_POINTER 8 + #endif + #define MY_CPU_64BIT #endif -#if defined(MY_CPU_AMD64) \ - || defined(_M_IA64) \ - || defined(__AARCH64EL__) \ - || defined(__AARCH64EB__) + +#if defined(_M_IX86) \ + || defined(__i386__) + #define MY_CPU_X86 + #define MY_CPU_NAME "x86" + /* #define MY_CPU_32BIT */ + #define MY_CPU_SIZEOF_POINTER 4 +#endif + + +#if defined(_M_ARM64) \ + || defined(__AARCH64EL__) \ + || defined(__AARCH64EB__) \ + || defined(__aarch64__) + #define MY_CPU_ARM64 + #ifdef __ILP32__ + #define MY_CPU_NAME "arm64-32" + #define MY_CPU_SIZEOF_POINTER 4 + #else + #define MY_CPU_NAME "arm64" + #define MY_CPU_SIZEOF_POINTER 8 + #endif #define MY_CPU_64BIT #endif -#if defined(_M_IX86) || defined(__i386__) -#define MY_CPU_X86 + +#if defined(_M_ARM) \ + || defined(_M_ARM_NT) \ + || defined(_M_ARMT) \ + || defined(__arm__) \ + || defined(__thumb__) \ + || defined(__ARMEL__) \ + || defined(__ARMEB__) \ + || defined(__THUMBEL__) \ + || defined(__THUMBEB__) + #define MY_CPU_ARM + + #if defined(__thumb__) || defined(__THUMBEL__) || defined(_M_ARMT) + #define MY_CPU_ARMT + #define MY_CPU_NAME "armt" + #else + #define MY_CPU_ARM32 + #define MY_CPU_NAME "arm" + #endif + /* #define MY_CPU_32BIT */ + #define MY_CPU_SIZEOF_POINTER 4 +#endif + + +#if defined(_M_IA64) \ + || defined(__ia64__) + #define MY_CPU_IA64 + #define MY_CPU_NAME "ia64" + #define MY_CPU_64BIT #endif + +#if defined(__mips64) \ + || defined(__mips64__) \ + || (defined(__mips) && (__mips == 64 || __mips == 4 || __mips == 3)) + #define MY_CPU_NAME "mips64" + #define MY_CPU_64BIT +#elif defined(__mips__) + #define MY_CPU_NAME "mips" + /* #define MY_CPU_32BIT */ +#endif + + +#if defined(__ppc64__) \ + || defined(__powerpc64__) \ + || defined(__ppc__) \ + || defined(__powerpc__) \ + || defined(__PPC__) \ + || defined(_POWER) + +#define MY_CPU_PPC_OR_PPC64 + +#if defined(__ppc64__) \ + || defined(__powerpc64__) \ + || defined(_LP64) \ + || defined(__64BIT__) + #ifdef __ILP32__ + #define MY_CPU_NAME "ppc64-32" + #define MY_CPU_SIZEOF_POINTER 4 + #else + #define MY_CPU_NAME "ppc64" + #define MY_CPU_SIZEOF_POINTER 8 + #endif + #define MY_CPU_64BIT +#else + #define MY_CPU_NAME "ppc" + #define MY_CPU_SIZEOF_POINTER 4 + /* #define MY_CPU_32BIT */ +#endif +#endif + + +#if defined(__riscv) \ + || defined(__riscv__) + #if __riscv_xlen == 32 + #define MY_CPU_NAME "riscv32" + #elif __riscv_xlen == 64 + #define MY_CPU_NAME "riscv64" + #else + #define MY_CPU_NAME "riscv" + #endif +#endif + + #if defined(MY_CPU_X86) || defined(MY_CPU_AMD64) #define MY_CPU_X86_OR_AMD64 #endif -#if defined(MY_CPU_X86) \ - || defined(_M_ARM) \ - || defined(__ARMEL__) \ - || defined(__THUMBEL__) \ - || defined(__ARMEB__) \ - || defined(__THUMBEB__) - #define MY_CPU_32BIT +#if defined(MY_CPU_ARM) || defined(MY_CPU_ARM64) +#define MY_CPU_ARM_OR_ARM64 #endif -#if defined(_WIN32) && defined(_M_ARM) -#define MY_CPU_ARM_LE -#endif -#if defined(_WIN32) && defined(_M_IA64) -#define MY_CPU_IA64_LE +#ifdef _WIN32 + + #ifdef MY_CPU_ARM + #define MY_CPU_ARM_LE + #endif + + #ifdef MY_CPU_ARM64 + #define MY_CPU_ARM64_LE + #endif + + #ifdef _M_IA64 + #define MY_CPU_IA64_LE + #endif + #endif + #if defined(MY_CPU_X86_OR_AMD64) \ || defined(MY_CPU_ARM_LE) \ + || defined(MY_CPU_ARM64_LE) \ || defined(MY_CPU_IA64_LE) \ || defined(__LITTLE_ENDIAN__) \ || defined(__ARMEL__) \ @@ -86,14 +202,139 @@ MY_CPU_LE_UNALIGN means that CPU is LITTLE ENDIAN and CPU supports unaligned mem #define MY_CPU_BE #endif + #if defined(MY_CPU_LE) && defined(MY_CPU_BE) -Stop_Compiling_Bad_Endian + #error Stop_Compiling_Bad_Endian +#endif + +#if !defined(MY_CPU_LE) && !defined(MY_CPU_BE) + #error Stop_Compiling_CPU_ENDIAN_must_be_detected_at_compile_time +#endif + +#if defined(MY_CPU_32BIT) && defined(MY_CPU_64BIT) + #error Stop_Compiling_Bad_32_64_BIT +#endif + +#ifdef __SIZEOF_POINTER__ + #ifdef MY_CPU_SIZEOF_POINTER + #if MY_CPU_SIZEOF_POINTER != __SIZEOF_POINTER__ + #error Stop_Compiling_Bad_MY_CPU_PTR_SIZE + #endif + #else + #define MY_CPU_SIZEOF_POINTER __SIZEOF_POINTER__ + #endif +#endif + +#if defined(MY_CPU_SIZEOF_POINTER) && (MY_CPU_SIZEOF_POINTER == 4) +#if defined (_LP64) + #error Stop_Compiling_Bad_MY_CPU_PTR_SIZE +#endif +#endif + +#ifdef _MSC_VER + #if _MSC_VER >= 1300 + #define MY_CPU_pragma_pack_push_1 __pragma(pack(push, 1)) + #define MY_CPU_pragma_pop __pragma(pack(pop)) + #else + #define MY_CPU_pragma_pack_push_1 + #define MY_CPU_pragma_pop + #endif +#else + #ifdef __xlC__ + #define MY_CPU_pragma_pack_push_1 _Pragma("pack(1)") + #define MY_CPU_pragma_pop _Pragma("pack()") + #else + #define MY_CPU_pragma_pack_push_1 _Pragma("pack(push, 1)") + #define MY_CPU_pragma_pop _Pragma("pack(pop)") + #endif +#endif + + +#ifndef MY_CPU_NAME + #ifdef MY_CPU_LE + #define MY_CPU_NAME "LE" + #elif defined(MY_CPU_BE) + #define MY_CPU_NAME "BE" + #else + /* + #define MY_CPU_NAME "" + */ + #endif +#endif + + + + + +#ifdef __has_builtin + #define Z7_has_builtin(x) __has_builtin(x) +#else + #define Z7_has_builtin(x) 0 #endif +#define Z7_BSWAP32_CONST(v) \ + ( (((UInt32)(v) << 24) ) \ + | (((UInt32)(v) << 8) & (UInt32)0xff0000) \ + | (((UInt32)(v) >> 8) & (UInt32)0xff00 ) \ + | (((UInt32)(v) >> 24) )) + + +#if defined(_MSC_VER) && (_MSC_VER >= 1300) + +#include <stdlib.h> + +/* Note: these macros will use bswap instruction (486), that is unsupported in 386 cpu */ + +#pragma intrinsic(_byteswap_ushort) +#pragma intrinsic(_byteswap_ulong) +#pragma intrinsic(_byteswap_uint64) + +#define Z7_BSWAP16(v) _byteswap_ushort(v) +#define Z7_BSWAP32(v) _byteswap_ulong (v) +#define Z7_BSWAP64(v) _byteswap_uint64(v) +#define Z7_CPU_FAST_BSWAP_SUPPORTED + +#elif (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))) \ + || (defined(__clang__) && Z7_has_builtin(__builtin_bswap16)) + +#define Z7_BSWAP16(v) __builtin_bswap16(v) +#define Z7_BSWAP32(v) __builtin_bswap32(v) +#define Z7_BSWAP64(v) __builtin_bswap64(v) +#define Z7_CPU_FAST_BSWAP_SUPPORTED + +#else + +#define Z7_BSWAP16(v) ((UInt16) \ + ( ((UInt32)(v) << 8) \ + | ((UInt32)(v) >> 8) \ + )) + +#define Z7_BSWAP32(v) Z7_BSWAP32_CONST(v) + +#define Z7_BSWAP64(v) \ + ( ( ( (UInt64)(v) ) << 8 * 7 ) \ + | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 1) ) << 8 * 5 ) \ + | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 2) ) << 8 * 3 ) \ + | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 3) ) << 8 * 1 ) \ + | ( ( (UInt64)(v) >> 8 * 1 ) & ((UInt32)0xff << 8 * 3) ) \ + | ( ( (UInt64)(v) >> 8 * 3 ) & ((UInt32)0xff << 8 * 2) ) \ + | ( ( (UInt64)(v) >> 8 * 5 ) & ((UInt32)0xff << 8 * 1) ) \ + | ( ( (UInt64)(v) >> 8 * 7 ) ) \ + ) + +#endif + + + #ifdef MY_CPU_LE #if defined(MY_CPU_X86_OR_AMD64) \ - /* || defined(__AARCH64EL__) */ + || defined(MY_CPU_ARM64) + #define MY_CPU_LE_UNALIGN + #define MY_CPU_LE_UNALIGN_64 + #elif defined(__ARM_FEATURE_UNALIGNED) + /* gcc9 for 32-bit arm can use LDRD instruction that requires 32-bit alignment. + So we can't use unaligned 64-bit operations. */ #define MY_CPU_LE_UNALIGN #endif #endif @@ -103,11 +344,13 @@ Stop_Compiling_Bad_Endian #define GetUi16(p) (*(const UInt16 *)(const void *)(p)) #define GetUi32(p) (*(const UInt32 *)(const void *)(p)) +#ifdef MY_CPU_LE_UNALIGN_64 #define GetUi64(p) (*(const UInt64 *)(const void *)(p)) +#define SetUi64(p, v) { *(UInt64 *)(void *)(p) = (v); } +#endif -#define SetUi16(p, v) { *(UInt16 *)(p) = (v); } -#define SetUi32(p, v) { *(UInt32 *)(p) = (v); } -#define SetUi64(p, v) { *(UInt64 *)(p) = (v); } +#define SetUi16(p, v) { *(UInt16 *)(void *)(p) = (v); } +#define SetUi32(p, v) { *(UInt32 *)(void *)(p) = (v); } #else @@ -121,8 +364,6 @@ Stop_Compiling_Bad_Endian ((UInt32)((const Byte *)(p))[2] << 16) | \ ((UInt32)((const Byte *)(p))[3] << 24)) -#define GetUi64(p) (GetUi32(p) | ((UInt64)GetUi32(((const Byte *)(p)) + 4) << 32)) - #define SetUi16(p, v) { Byte *_ppp_ = (Byte *)(p); UInt32 _vvv_ = (v); \ _ppp_[0] = (Byte)_vvv_; \ _ppp_[1] = (Byte)(_vvv_ >> 8); } @@ -133,32 +374,28 @@ Stop_Compiling_Bad_Endian _ppp_[2] = (Byte)(_vvv_ >> 16); \ _ppp_[3] = (Byte)(_vvv_ >> 24); } -#define SetUi64(p, v) { Byte *_ppp2_ = (Byte *)(p); UInt64 _vvv2_ = (v); \ - SetUi32(_ppp2_ , (UInt32)_vvv2_); \ - SetUi32(_ppp2_ + 4, (UInt32)(_vvv2_ >> 32)); } - #endif -#if defined(MY_CPU_LE_UNALIGN) && /* defined(_WIN64) && */ (_MSC_VER >= 1300) - -/* Note: we use bswap instruction, that is unsupported in 386 cpu */ - -#include <stdlib.h> +#ifndef GetUi64 +#define GetUi64(p) (GetUi32(p) | ((UInt64)GetUi32(((const Byte *)(p)) + 4) << 32)) +#endif -#pragma intrinsic(_byteswap_ulong) -#pragma intrinsic(_byteswap_uint64) -#define GetBe32(p) _byteswap_ulong(*(const UInt32 *)(const Byte *)(p)) -#define GetBe64(p) _byteswap_uint64(*(const UInt64 *)(const Byte *)(p)) +#ifndef SetUi64 +#define SetUi64(p, v) { Byte *_ppp2_ = (Byte *)(p); UInt64 _vvv2_ = (v); \ + SetUi32(_ppp2_ , (UInt32)_vvv2_) \ + SetUi32(_ppp2_ + 4, (UInt32)(_vvv2_ >> 32)) } +#endif -#define SetBe32(p, v) (*(UInt32 *)(void *)(p)) = _byteswap_ulong(v) -#elif defined(MY_CPU_LE_UNALIGN) && defined (__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)) +#if defined(MY_CPU_LE_UNALIGN) && defined(Z7_CPU_FAST_BSWAP_SUPPORTED) -#define GetBe32(p) __builtin_bswap32(*(const UInt32 *)(const Byte *)(p)) -#define GetBe64(p) __builtin_bswap64(*(const UInt64 *)(const Byte *)(p)) +#define GetBe32(p) Z7_BSWAP32 (*(const UInt32 *)(const void *)(p)) +#define SetBe32(p, v) { (*(UInt32 *)(void *)(p)) = Z7_BSWAP32(v); } -#define SetBe32(p, v) (*(UInt32 *)(void *)(p)) = __builtin_bswap32(v) +#if defined(MY_CPU_LE_UNALIGN_64) +#define GetBe64(p) Z7_BSWAP64 (*(const UInt64 *)(const void *)(p)) +#endif #else @@ -168,8 +405,6 @@ Stop_Compiling_Bad_Endian ((UInt32)((const Byte *)(p))[2] << 8) | \ ((const Byte *)(p))[3] ) -#define GetBe64(p) (((UInt64)GetBe32(p) << 32) | GetBe32(((const Byte *)(p)) + 4)) - #define SetBe32(p, v) { Byte *_ppp_ = (Byte *)(p); UInt32 _vvv_ = (v); \ _ppp_[0] = (Byte)(_vvv_ >> 24); \ _ppp_[1] = (Byte)(_vvv_ >> 16); \ @@ -178,44 +413,109 @@ Stop_Compiling_Bad_Endian #endif +#ifndef GetBe64 +#define GetBe64(p) (((UInt64)GetBe32(p) << 32) | GetBe32(((const Byte *)(p)) + 4)) +#endif +#ifndef GetBe16 #define GetBe16(p) ( (UInt16) ( \ ((UInt16)((const Byte *)(p))[0] << 8) | \ ((const Byte *)(p))[1] )) +#endif + +#if defined(MY_CPU_BE) +#define Z7_CONV_BE_TO_NATIVE_CONST32(v) (v) +#define Z7_CONV_LE_TO_NATIVE_CONST32(v) Z7_BSWAP32_CONST(v) +#define Z7_CONV_NATIVE_TO_BE_32(v) (v) +#elif defined(MY_CPU_LE) +#define Z7_CONV_BE_TO_NATIVE_CONST32(v) Z7_BSWAP32_CONST(v) +#define Z7_CONV_LE_TO_NATIVE_CONST32(v) (v) +#define Z7_CONV_NATIVE_TO_BE_32(v) Z7_BSWAP32(v) +#else +#error Stop_Compiling_Unknown_Endian_CONV +#endif -#ifdef MY_CPU_X86_OR_AMD64 +#if defined(MY_CPU_BE) -typedef struct -{ - UInt32 maxFunc; - UInt32 vendor[3]; - UInt32 ver; - UInt32 b; - UInt32 c; - UInt32 d; -} Cx86cpuid; +#define GetBe32a(p) (*(const UInt32 *)(const void *)(p)) +#define GetBe16a(p) (*(const UInt16 *)(const void *)(p)) +#define SetBe32a(p, v) { *(UInt32 *)(void *)(p) = (v); } +#define SetBe16a(p, v) { *(UInt16 *)(void *)(p) = (v); } -enum -{ - CPU_FIRM_INTEL, - CPU_FIRM_AMD, - CPU_FIRM_VIA -}; +#define GetUi32a(p) GetUi32(p) +#define GetUi16a(p) GetUi16(p) +#define SetUi32a(p, v) SetUi32(p, v) +#define SetUi16a(p, v) SetUi16(p, v) -void MyCPUID(UInt32 function, UInt32 *a, UInt32 *b, UInt32 *c, UInt32 *d); +#elif defined(MY_CPU_LE) -Bool x86cpuid_CheckAndRead(Cx86cpuid *p); -int x86cpuid_GetFirm(const Cx86cpuid *p); +#define GetUi32a(p) (*(const UInt32 *)(const void *)(p)) +#define GetUi16a(p) (*(const UInt16 *)(const void *)(p)) +#define SetUi32a(p, v) { *(UInt32 *)(void *)(p) = (v); } +#define SetUi16a(p, v) { *(UInt16 *)(void *)(p) = (v); } -#define x86cpuid_GetFamily(ver) (((ver >> 16) & 0xFF0) | ((ver >> 8) & 0xF)) -#define x86cpuid_GetModel(ver) (((ver >> 12) & 0xF0) | ((ver >> 4) & 0xF)) -#define x86cpuid_GetStepping(ver) (ver & 0xF) +#define GetBe32a(p) GetBe32(p) +#define GetBe16a(p) GetBe16(p) +#define SetBe32a(p, v) SetBe32(p, v) +#define SetBe16a(p, v) SetBe16(p, v) -Bool CPU_Is_InOrder(); -Bool CPU_Is_Aes_Supported(); +#else +#error Stop_Compiling_Unknown_Endian_CPU_a +#endif + + +#if defined(MY_CPU_X86_OR_AMD64) \ + || defined(MY_CPU_ARM_OR_ARM64) \ + || defined(MY_CPU_PPC_OR_PPC64) + #define Z7_CPU_FAST_ROTATE_SUPPORTED +#endif + + +#ifdef MY_CPU_X86_OR_AMD64 + +void Z7_FASTCALL z7_x86_cpuid(UInt32 a[4], UInt32 function); +UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void); +#if defined(MY_CPU_AMD64) +#define Z7_IF_X86_CPUID_SUPPORTED +#else +#define Z7_IF_X86_CPUID_SUPPORTED if (z7_x86_cpuid_GetMaxFunc()) +#endif + +BoolInt CPU_IsSupported_AES(void); +BoolInt CPU_IsSupported_AVX(void); +BoolInt CPU_IsSupported_AVX2(void); +BoolInt CPU_IsSupported_VAES_AVX2(void); +BoolInt CPU_IsSupported_CMOV(void); +BoolInt CPU_IsSupported_SSE(void); +BoolInt CPU_IsSupported_SSE2(void); +BoolInt CPU_IsSupported_SSSE3(void); +BoolInt CPU_IsSupported_SSE41(void); +BoolInt CPU_IsSupported_SHA(void); +BoolInt CPU_IsSupported_PageGB(void); + +#elif defined(MY_CPU_ARM_OR_ARM64) + +BoolInt CPU_IsSupported_CRC32(void); +BoolInt CPU_IsSupported_NEON(void); + +#if defined(_WIN32) +BoolInt CPU_IsSupported_CRYPTO(void); +#define CPU_IsSupported_SHA1 CPU_IsSupported_CRYPTO +#define CPU_IsSupported_SHA2 CPU_IsSupported_CRYPTO +#define CPU_IsSupported_AES CPU_IsSupported_CRYPTO +#else +BoolInt CPU_IsSupported_SHA1(void); +BoolInt CPU_IsSupported_SHA2(void); +BoolInt CPU_IsSupported_AES(void); +#endif + +#endif +#if defined(__APPLE__) +int z7_sysctlbyname_Get(const char *name, void *buf, size_t *bufSize); +int z7_sysctlbyname_Get_UInt32(const char *name, UInt32 *val); #endif EXTERN_C_END diff --git a/3rdparty/lzma/C/Delta.c b/3rdparty/lzma/C/Delta.c index e3edd21edbb..c4a4499fe2a 100644 --- a/3rdparty/lzma/C/Delta.c +++ b/3rdparty/lzma/C/Delta.c @@ -1,5 +1,5 @@ /* Delta.c -- Delta converter -2009-05-26 : Igor Pavlov : Public domain */ +2021-02-09 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -12,53 +12,158 @@ void Delta_Init(Byte *state) state[i] = 0; } -static void MyMemCpy(Byte *dest, const Byte *src, unsigned size) -{ - unsigned i; - for (i = 0; i < size; i++) - dest[i] = src[i]; -} void Delta_Encode(Byte *state, unsigned delta, Byte *data, SizeT size) { - Byte buf[DELTA_STATE_SIZE]; - unsigned j = 0; - MyMemCpy(buf, state, delta); + Byte temp[DELTA_STATE_SIZE]; + + if (size == 0) + return; + + { + unsigned i = 0; + do + temp[i] = state[i]; + while (++i != delta); + } + + if (size <= delta) + { + unsigned i = 0, k; + do + { + Byte b = *data; + *data++ = (Byte)(b - temp[i]); + temp[i] = b; + } + while (++i != size); + + k = 0; + + do + { + if (i == delta) + i = 0; + state[k] = temp[i++]; + } + while (++k != delta); + + return; + } + { - SizeT i; - for (i = 0; i < size;) + Byte *p = data + size - delta; + { + unsigned i = 0; + do + state[i] = *p++; + while (++i != delta); + } { - for (j = 0; j < delta && i < size; i++, j++) + const Byte *lim = data + delta; + ptrdiff_t dif = -(ptrdiff_t)delta; + + if (((ptrdiff_t)size + dif) & 1) { - Byte b = data[i]; - data[i] = (Byte)(b - buf[j]); - buf[j] = b; + --p; *p = (Byte)(*p - p[dif]); } + + while (p != lim) + { + --p; *p = (Byte)(*p - p[dif]); + --p; *p = (Byte)(*p - p[dif]); + } + + dif = -dif; + + do + { + --p; *p = (Byte)(*p - temp[--dif]); + } + while (dif != 0); } } - if (j == delta) - j = 0; - MyMemCpy(state, buf + j, delta - j); - MyMemCpy(state + delta - j, buf, j); } + void Delta_Decode(Byte *state, unsigned delta, Byte *data, SizeT size) { - Byte buf[DELTA_STATE_SIZE]; - unsigned j = 0; - MyMemCpy(buf, state, delta); + unsigned i; + const Byte *lim; + + if (size == 0) + return; + + i = 0; + lim = data + size; + + if (size <= delta) + { + do + *data = (Byte)(*data + state[i++]); + while (++data != lim); + + for (; delta != i; state++, delta--) + *state = state[i]; + data -= i; + } + else { - SizeT i; - for (i = 0; i < size;) + /* + #define B(n) b ## n + #define I(n) Byte B(n) = state[n]; + #define U(n) { B(n) = (Byte)((B(n)) + *data++); data[-1] = (B(n)); } + #define F(n) if (data != lim) { U(n) } + + if (delta == 1) + { + I(0) + if ((lim - data) & 1) { U(0) } + while (data != lim) { U(0) U(0) } + data -= 1; + } + else if (delta == 2) { - for (j = 0; j < delta && i < size; i++, j++) + I(0) I(1) + lim -= 1; while (data < lim) { U(0) U(1) } + lim += 1; F(0) + data -= 2; + } + else if (delta == 3) + { + I(0) I(1) I(2) + lim -= 2; while (data < lim) { U(0) U(1) U(2) } + lim += 2; F(0) F(1) + data -= 3; + } + else if (delta == 4) + { + I(0) I(1) I(2) I(3) + lim -= 3; while (data < lim) { U(0) U(1) U(2) U(3) } + lim += 3; F(0) F(1) F(2) + data -= 4; + } + else + */ + { + do + { + *data = (Byte)(*data + state[i++]); + data++; + } + while (i != delta); + { - buf[j] = data[i] = (Byte)(buf[j] + data[i]); + ptrdiff_t dif = -(ptrdiff_t)delta; + do + *data = (Byte)(*data + data[dif]); + while (++data != lim); + data += dif; } } } - if (j == delta) - j = 0; - MyMemCpy(state, buf + j, delta - j); - MyMemCpy(state + delta - j, buf, j); + + do + *state++ = *data; + while (++data != lim); } diff --git a/3rdparty/lzma/C/Delta.h b/3rdparty/lzma/C/Delta.h index 2fa54ad67b3..706095417aa 100644 --- a/3rdparty/lzma/C/Delta.h +++ b/3rdparty/lzma/C/Delta.h @@ -1,8 +1,8 @@ /* Delta.h -- Delta converter -2013-01-18 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ -#ifndef __DELTA_H -#define __DELTA_H +#ifndef ZIP7_INC_DELTA_H +#define ZIP7_INC_DELTA_H #include "7zTypes.h" diff --git a/3rdparty/lzma/C/DllSecur.c b/3rdparty/lzma/C/DllSecur.c index 90352ce10f7..02a0f977ead 100644 --- a/3rdparty/lzma/C/DllSecur.c +++ b/3rdparty/lzma/C/DllSecur.c @@ -1,87 +1,111 @@ /* DllSecur.c -- DLL loading security -2016-10-04 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" #ifdef _WIN32 -#include <windows.h> +#include "7zWindows.h" #include "DllSecur.h" #ifndef UNDER_CE +#if (defined(__GNUC__) && (__GNUC__ >= 8)) || defined(__clang__) + // #pragma GCC diagnostic ignored "-Wcast-function-type" +#endif + +#if defined(__clang__) || defined(__GNUC__) +typedef void (*Z7_voidFunction)(void); +#define MY_CAST_FUNC (Z7_voidFunction) +#elif defined(_MSC_VER) && _MSC_VER > 1920 +#define MY_CAST_FUNC (void *) +// #pragma warning(disable : 4191) // 'type cast': unsafe conversion from 'FARPROC' to 'void (__cdecl *)()' +#else +#define MY_CAST_FUNC +#endif + typedef BOOL (WINAPI *Func_SetDefaultDllDirectories)(DWORD DirectoryFlags); #define MY_LOAD_LIBRARY_SEARCH_USER_DIRS 0x400 #define MY_LOAD_LIBRARY_SEARCH_SYSTEM32 0x800 +#define DELIM "\0" + static const char * const g_Dlls = + "userenv" + DELIM "setupapi" + DELIM "apphelp" + DELIM "propsys" + DELIM "dwmapi" + DELIM "cryptbase" + DELIM "oleacc" + DELIM "clbcatq" + DELIM "version" #ifndef _CONSOLE - "UXTHEME\0" + DELIM "uxtheme" #endif - "USERENV\0" - "SETUPAPI\0" - "APPHELP\0" - "PROPSYS\0" - "DWMAPI\0" - "CRYPTBASE\0" - "OLEACC\0" - "CLBCATQ\0" - ; + DELIM; + +#endif +#ifdef __clang__ + #pragma GCC diagnostic ignored "-Wdeprecated-declarations" +#endif +#if defined (_MSC_VER) && _MSC_VER >= 1900 +// sysinfoapi.h: kit10: GetVersion was declared deprecated +#pragma warning(disable : 4996) #endif -void LoadSecurityDlls() +#define IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN \ + if ((UInt16)GetVersion() != 6) { \ + const \ + Func_SetDefaultDllDirectories setDllDirs = \ + (Func_SetDefaultDllDirectories) MY_CAST_FUNC GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), \ + "SetDefaultDllDirectories"); \ + if (setDllDirs) if (setDllDirs(MY_LOAD_LIBRARY_SEARCH_SYSTEM32 | MY_LOAD_LIBRARY_SEARCH_USER_DIRS)) return; } + +void My_SetDefaultDllDirectories(void) { #ifndef UNDER_CE - - wchar_t buf[MAX_PATH + 100]; + IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN + #endif +} - { - // at Vista (ver 6.0) : CoCreateInstance(CLSID_ShellLink, ...) doesn't work after SetDefaultDllDirectories() : Check it ??? - OSVERSIONINFO vi; - vi.dwOSVersionInfoSize = sizeof(vi); - if (!GetVersionEx(&vi) || vi.dwMajorVersion != 6 || vi.dwMinorVersion != 0) - { - Func_SetDefaultDllDirectories setDllDirs = (Func_SetDefaultDllDirectories) - GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "SetDefaultDllDirectories"); - if (setDllDirs) - if (setDllDirs(MY_LOAD_LIBRARY_SEARCH_SYSTEM32 | MY_LOAD_LIBRARY_SEARCH_USER_DIRS)) - return; - } - } +void LoadSecurityDlls(void) +{ + #ifndef UNDER_CE + // at Vista (ver 6.0) : CoCreateInstance(CLSID_ShellLink, ...) doesn't work after SetDefaultDllDirectories() : Check it ??? + IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN { - unsigned len = GetSystemDirectoryW(buf, MAX_PATH + 2); - if (len == 0 || len > MAX_PATH) - return; - } - { + wchar_t buf[MAX_PATH + 100]; const char *dll; - unsigned pos = (unsigned)lstrlenW(buf); - + unsigned pos = GetSystemDirectoryW(buf, MAX_PATH + 2); + if (pos == 0 || pos > MAX_PATH) + return; if (buf[pos - 1] != '\\') buf[pos++] = '\\'; - - for (dll = g_Dlls; dll[0] != 0;) + for (dll = g_Dlls; *dll != 0;) { - unsigned k = 0; + wchar_t *dest = &buf[pos]; for (;;) { - char c = *dll++; - buf[pos + k] = c; - k++; + const char c = *dll++; if (c == 0) break; + *dest++ = (Byte)c; } - - lstrcatW(buf, L".dll"); + dest[0] = '.'; + dest[1] = 'd'; + dest[2] = 'l'; + dest[3] = 'l'; + dest[4] = 0; + // lstrcatW(buf, L".dll"); LoadLibraryExW(buf, NULL, LOAD_WITH_ALTERED_SEARCH_PATH); } } - #endif } -#endif +#endif // _WIN32 diff --git a/3rdparty/lzma/C/DllSecur.h b/3rdparty/lzma/C/DllSecur.h index e01f2247dcc..9fa41538265 100644 --- a/3rdparty/lzma/C/DllSecur.h +++ b/3rdparty/lzma/C/DllSecur.h @@ -1,8 +1,8 @@ /* DllSecur.h -- DLL loading for security -2016-06-08 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ -#ifndef __DLL_SECUR_H -#define __DLL_SECUR_H +#ifndef ZIP7_INC_DLL_SECUR_H +#define ZIP7_INC_DLL_SECUR_H #include "7zTypes.h" @@ -10,7 +10,8 @@ EXTERN_C_BEGIN #ifdef _WIN32 -void LoadSecurityDlls(); +void My_SetDefaultDllDirectories(void); +void LoadSecurityDlls(void); #endif diff --git a/3rdparty/lzma/C/LzFind.c b/3rdparty/lzma/C/LzFind.c index 2d05fa39534..0fbd5aae563 100644 --- a/3rdparty/lzma/C/LzFind.c +++ b/3rdparty/lzma/C/LzFind.c @@ -1,74 +1,134 @@ /* LzFind.c -- Match finder for LZ algorithms -2015-10-15 : Igor Pavlov : Public domain */ +2023-03-14 : Igor Pavlov : Public domain */ #include "Precomp.h" #include <string.h> +// #include <stdio.h> +#include "CpuArch.h" #include "LzFind.h" #include "LzHash.h" +#define kBlockMoveAlign (1 << 7) // alignment for memmove() +#define kBlockSizeAlign (1 << 16) // alignment for block allocation +#define kBlockSizeReserveMin (1 << 24) // it's 1/256 from 4 GB dictinary + #define kEmptyHashValue 0 -#define kMaxValForNormalize ((UInt32)0xFFFFFFFF) -#define kNormalizeStepMin (1 << 10) /* it must be power of 2 */ -#define kNormalizeMask (~(UInt32)(kNormalizeStepMin - 1)) -#define kMaxHistorySize ((UInt32)7 << 29) -#define kStartMaxLen 3 +#define kMaxValForNormalize ((UInt32)0) +// #define kMaxValForNormalize ((UInt32)(1 << 20) + 0xfff) // for debug + +// #define kNormalizeAlign (1 << 7) // alignment for speculated accesses + +#define GET_AVAIL_BYTES(p) \ + Inline_MatchFinder_GetNumAvailableBytes(p) + + +// #define kFix5HashSize (kHash2Size + kHash3Size + kHash4Size) +#define kFix5HashSize kFix4HashSize + +/* + HASH2_CALC: + if (hv) match, then cur[0] and cur[1] also match +*/ +#define HASH2_CALC hv = GetUi16(cur); + +// (crc[0 ... 255] & 0xFF) provides one-to-one correspondence to [0 ... 255] -static void LzInWindow_Free(CMatchFinder *p, ISzAlloc *alloc) +/* + HASH3_CALC: + if (cur[0]) and (h2) match, then cur[1] also match + if (cur[0]) and (hv) match, then cur[1] and cur[2] also match +*/ +#define HASH3_CALC { \ + UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ + h2 = temp & (kHash2Size - 1); \ + hv = (temp ^ ((UInt32)cur[2] << 8)) & p->hashMask; } + +#define HASH4_CALC { \ + UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ + h2 = temp & (kHash2Size - 1); \ + temp ^= ((UInt32)cur[2] << 8); \ + h3 = temp & (kHash3Size - 1); \ + hv = (temp ^ (p->crc[cur[3]] << kLzHash_CrcShift_1)) & p->hashMask; } + +#define HASH5_CALC { \ + UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ + h2 = temp & (kHash2Size - 1); \ + temp ^= ((UInt32)cur[2] << 8); \ + h3 = temp & (kHash3Size - 1); \ + temp ^= (p->crc[cur[3]] << kLzHash_CrcShift_1); \ + /* h4 = temp & p->hash4Mask; */ /* (kHash4Size - 1); */ \ + hv = (temp ^ (p->crc[cur[4]] << kLzHash_CrcShift_2)) & p->hashMask; } + +#define HASH_ZIP_CALC hv = ((cur[2] | ((UInt32)cur[0] << 8)) ^ p->crc[cur[1]]) & 0xFFFF; + + +static void LzInWindow_Free(CMatchFinder *p, ISzAllocPtr alloc) { - if (!p->directInput) + // if (!p->directInput) { - alloc->Free(alloc, p->bufferBase); - p->bufferBase = NULL; + ISzAlloc_Free(alloc, p->bufBase); + p->bufBase = NULL; } } -/* keepSizeBefore + keepSizeAfter + keepSizeReserv must be < 4G) */ -static int LzInWindow_Create(CMatchFinder *p, UInt32 keepSizeReserv, ISzAlloc *alloc) +static int LzInWindow_Create2(CMatchFinder *p, UInt32 blockSize, ISzAllocPtr alloc) { - UInt32 blockSize = p->keepSizeBefore + p->keepSizeAfter + keepSizeReserv; - if (p->directInput) - { - p->blockSize = blockSize; - return 1; - } - if (!p->bufferBase || p->blockSize != blockSize) + if (blockSize == 0) + return 0; + if (!p->bufBase || p->blockSize != blockSize) { + // size_t blockSizeT; LzInWindow_Free(p, alloc); p->blockSize = blockSize; - p->bufferBase = (Byte *)alloc->Alloc(alloc, (size_t)blockSize); + // blockSizeT = blockSize; + + // printf("\nblockSize = 0x%x\n", blockSize); + /* + #if defined _WIN64 + // we can allocate 4GiB, but still use UInt32 for (p->blockSize) + // we use UInt32 type for (p->blockSize), because + // we don't want to wrap over 4 GiB, + // when we use (p->streamPos - p->pos) that is UInt32. + if (blockSize >= (UInt32)0 - (UInt32)kBlockSizeAlign) + { + blockSizeT = ((size_t)1 << 32); + printf("\nchanged to blockSizeT = 4GiB\n"); + } + #endif + */ + + p->bufBase = (Byte *)ISzAlloc_Alloc(alloc, blockSize); + // printf("\nbufferBase = %p\n", p->bufBase); + // return 0; // for debug } - return (p->bufferBase != NULL); + return (p->bufBase != NULL); } -Byte *MatchFinder_GetPointerToCurrentPos(CMatchFinder *p) { return p->buffer; } +static const Byte *MatchFinder_GetPointerToCurrentPos(CMatchFinder *p) { return p->buffer; } -UInt32 MatchFinder_GetNumAvailableBytes(CMatchFinder *p) { return p->streamPos - p->pos; } +static UInt32 MatchFinder_GetNumAvailableBytes(CMatchFinder *p) { return GET_AVAIL_BYTES(p); } -void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue) -{ - p->posLimit -= subValue; - p->pos -= subValue; - p->streamPos -= subValue; -} +Z7_NO_INLINE static void MatchFinder_ReadBlock(CMatchFinder *p) { if (p->streamEndWasReached || p->result != SZ_OK) return; - /* We use (p->streamPos - p->pos) value. (p->streamPos < p->pos) is allowed. */ + /* We use (p->streamPos - p->pos) value. + (p->streamPos < p->pos) is allowed. */ if (p->directInput) { - UInt32 curSize = 0xFFFFFFFF - (p->streamPos - p->pos); + UInt32 curSize = 0xFFFFFFFF - GET_AVAIL_BYTES(p); if (curSize > p->directInputRem) curSize = (UInt32)p->directInputRem; - p->directInputRem -= curSize; p->streamPos += curSize; + p->directInputRem -= curSize; if (p->directInputRem == 0) p->streamEndWasReached = 1; return; @@ -76,12 +136,31 @@ static void MatchFinder_ReadBlock(CMatchFinder *p) for (;;) { - Byte *dest = p->buffer + (p->streamPos - p->pos); - size_t size = (p->bufferBase + p->blockSize - dest); + const Byte *dest = p->buffer + GET_AVAIL_BYTES(p); + size_t size = (size_t)(p->bufBase + p->blockSize - dest); if (size == 0) + { + /* we call ReadBlock() after NeedMove() and MoveBlock(). + NeedMove() and MoveBlock() povide more than (keepSizeAfter) + to the end of (blockSize). + So we don't execute this branch in normal code flow. + We can go here, if we will call ReadBlock() before NeedMove(), MoveBlock(). + */ + // p->result = SZ_ERROR_FAIL; // we can show error here return; + } - p->result = p->stream->Read(p->stream, dest, &size); + // #define kRead 3 + // if (size > kRead) size = kRead; // for debug + + /* + // we need cast (Byte *)dest. + #ifdef __clang__ + #pragma GCC diagnostic ignored "-Wcast-qual" + #endif + */ + p->result = ISeqInStream_Read(p->stream, + p->bufBase + (dest - p->bufBase), &size); if (p->result != SZ_OK) return; if (size == 0) @@ -90,47 +169,60 @@ static void MatchFinder_ReadBlock(CMatchFinder *p) return; } p->streamPos += (UInt32)size; - if (p->streamPos - p->pos > p->keepSizeAfter) + if (GET_AVAIL_BYTES(p) > p->keepSizeAfter) return; + /* here and in another (p->keepSizeAfter) checks we keep on 1 byte more than was requested by Create() function + (GET_AVAIL_BYTES(p) >= p->keepSizeAfter) - minimal required size */ } + + // on exit: (p->result != SZ_OK || p->streamEndWasReached || GET_AVAIL_BYTES(p) > p->keepSizeAfter) } + + +Z7_NO_INLINE void MatchFinder_MoveBlock(CMatchFinder *p) { - memmove(p->bufferBase, - p->buffer - p->keepSizeBefore, - (size_t)(p->streamPos - p->pos) + p->keepSizeBefore); - p->buffer = p->bufferBase + p->keepSizeBefore; + const size_t offset = (size_t)(p->buffer - p->bufBase) - p->keepSizeBefore; + const size_t keepBefore = (offset & (kBlockMoveAlign - 1)) + p->keepSizeBefore; + p->buffer = p->bufBase + keepBefore; + memmove(p->bufBase, + p->bufBase + (offset & ~((size_t)kBlockMoveAlign - 1)), + keepBefore + (size_t)GET_AVAIL_BYTES(p)); } +/* We call MoveBlock() before ReadBlock(). + So MoveBlock() can be wasteful operation, if the whole input data + can fit in current block even without calling MoveBlock(). + in important case where (dataSize <= historySize) + condition (p->blockSize > dataSize + p->keepSizeAfter) is met + So there is no MoveBlock() in that case case. +*/ + int MatchFinder_NeedMove(CMatchFinder *p) { if (p->directInput) return 0; - /* if (p->streamEndWasReached) return 0; */ - return ((size_t)(p->bufferBase + p->blockSize - p->buffer) <= p->keepSizeAfter); + if (p->streamEndWasReached || p->result != SZ_OK) + return 0; + return ((size_t)(p->bufBase + p->blockSize - p->buffer) <= p->keepSizeAfter); } void MatchFinder_ReadIfRequired(CMatchFinder *p) { - if (p->streamEndWasReached) - return; - if (p->keepSizeAfter >= p->streamPos - p->pos) + if (p->keepSizeAfter >= GET_AVAIL_BYTES(p)) MatchFinder_ReadBlock(p); } -static void MatchFinder_CheckAndMoveAndRead(CMatchFinder *p) -{ - if (MatchFinder_NeedMove(p)) - MatchFinder_MoveBlock(p); - MatchFinder_ReadBlock(p); -} + static void MatchFinder_SetDefaultSettings(CMatchFinder *p) { p->cutValue = 32; p->btMode = 1; p->numHashBytes = 4; + p->numHashBytes_Min = 2; + p->numHashOutBits = 0; p->bigHash = 0; } @@ -138,113 +230,248 @@ static void MatchFinder_SetDefaultSettings(CMatchFinder *p) void MatchFinder_Construct(CMatchFinder *p) { - UInt32 i; - p->bufferBase = NULL; + unsigned i; + p->buffer = NULL; + p->bufBase = NULL; p->directInput = 0; + p->stream = NULL; p->hash = NULL; + p->expectedDataSize = (UInt64)(Int64)-1; MatchFinder_SetDefaultSettings(p); for (i = 0; i < 256; i++) { - UInt32 r = i; + UInt32 r = (UInt32)i; unsigned j; for (j = 0; j < 8; j++) - r = (r >> 1) ^ (kCrcPoly & ~((r & 1) - 1)); + r = (r >> 1) ^ (kCrcPoly & ((UInt32)0 - (r & 1))); p->crc[i] = r; } } -static void MatchFinder_FreeThisClassMemory(CMatchFinder *p, ISzAlloc *alloc) +#undef kCrcPoly + +static void MatchFinder_FreeThisClassMemory(CMatchFinder *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->hash); + ISzAlloc_Free(alloc, p->hash); p->hash = NULL; } -void MatchFinder_Free(CMatchFinder *p, ISzAlloc *alloc) +void MatchFinder_Free(CMatchFinder *p, ISzAllocPtr alloc) { MatchFinder_FreeThisClassMemory(p, alloc); LzInWindow_Free(p, alloc); } -static CLzRef* AllocRefs(size_t num, ISzAlloc *alloc) +static CLzRef* AllocRefs(size_t num, ISzAllocPtr alloc) { - size_t sizeInBytes = (size_t)num * sizeof(CLzRef); + const size_t sizeInBytes = (size_t)num * sizeof(CLzRef); if (sizeInBytes / sizeof(CLzRef) != num) return NULL; - return (CLzRef *)alloc->Alloc(alloc, sizeInBytes); + return (CLzRef *)ISzAlloc_Alloc(alloc, sizeInBytes); } -int MatchFinder_Create(CMatchFinder *p, UInt32 historySize, - UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter, - ISzAlloc *alloc) +#if (kBlockSizeReserveMin < kBlockSizeAlign * 2) + #error Stop_Compiling_Bad_Reserve +#endif + + + +static UInt32 GetBlockSize(CMatchFinder *p, UInt32 historySize) { - UInt32 sizeReserv; - + UInt32 blockSize = (p->keepSizeBefore + p->keepSizeAfter); + /* if (historySize > kMaxHistorySize) - { - MatchFinder_Free(p, alloc); return 0; - } + */ + // printf("\nhistorySize == 0x%x\n", historySize); - sizeReserv = historySize >> 1; - if (historySize >= ((UInt32)3 << 30)) sizeReserv = historySize >> 3; - else if (historySize >= ((UInt32)2 << 30)) sizeReserv = historySize >> 2; + if (p->keepSizeBefore < historySize || blockSize < p->keepSizeBefore) // if 32-bit overflow + return 0; - sizeReserv += (keepAddBufferBefore + matchMaxLen + keepAddBufferAfter) / 2 + (1 << 19); + { + const UInt32 kBlockSizeMax = (UInt32)0 - (UInt32)kBlockSizeAlign; + const UInt32 rem = kBlockSizeMax - blockSize; + const UInt32 reserve = (blockSize >> (blockSize < ((UInt32)1 << 30) ? 1 : 2)) + + (1 << 12) + kBlockMoveAlign + kBlockSizeAlign; // do not overflow 32-bit here + if (blockSize >= kBlockSizeMax + || rem < kBlockSizeReserveMin) // we reject settings that will be slow + return 0; + if (reserve >= rem) + blockSize = kBlockSizeMax; + else + { + blockSize += reserve; + blockSize &= ~(UInt32)(kBlockSizeAlign - 1); + } + } + // printf("\n LzFind_blockSize = %x\n", blockSize); + // printf("\n LzFind_blockSize = %d\n", blockSize >> 20); + return blockSize; +} + +// input is historySize +static UInt32 MatchFinder_GetHashMask2(CMatchFinder *p, UInt32 hs) +{ + if (p->numHashBytes == 2) + return (1 << 16) - 1; + if (hs != 0) + hs--; + hs |= (hs >> 1); + hs |= (hs >> 2); + hs |= (hs >> 4); + hs |= (hs >> 8); + // we propagated 16 bits in (hs). Low 16 bits must be set later + if (hs >= (1 << 24)) + { + if (p->numHashBytes == 3) + hs = (1 << 24) - 1; + /* if (bigHash) mode, GetHeads4b() in LzFindMt.c needs (hs >= ((1 << 24) - 1))) */ + } + // (hash_size >= (1 << 16)) : Required for (numHashBytes > 2) + hs |= (1 << 16) - 1; /* don't change it! */ + // bt5: we adjust the size with recommended minimum size + if (p->numHashBytes >= 5) + hs |= (256 << kLzHash_CrcShift_2) - 1; + return hs; +} + +// input is historySize +static UInt32 MatchFinder_GetHashMask(CMatchFinder *p, UInt32 hs) +{ + if (p->numHashBytes == 2) + return (1 << 16) - 1; + if (hs != 0) + hs--; + hs |= (hs >> 1); + hs |= (hs >> 2); + hs |= (hs >> 4); + hs |= (hs >> 8); + // we propagated 16 bits in (hs). Low 16 bits must be set later + hs >>= 1; + if (hs >= (1 << 24)) + { + if (p->numHashBytes == 3) + hs = (1 << 24) - 1; + else + hs >>= 1; + /* if (bigHash) mode, GetHeads4b() in LzFindMt.c needs (hs >= ((1 << 24) - 1))) */ + } + // (hash_size >= (1 << 16)) : Required for (numHashBytes > 2) + hs |= (1 << 16) - 1; /* don't change it! */ + // bt5: we adjust the size with recommended minimum size + if (p->numHashBytes >= 5) + hs |= (256 << kLzHash_CrcShift_2) - 1; + return hs; +} + + +int MatchFinder_Create(CMatchFinder *p, UInt32 historySize, + UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter, + ISzAllocPtr alloc) +{ + /* we need one additional byte in (p->keepSizeBefore), + since we use MoveBlock() after (p->pos++) and before dictionary using */ + // keepAddBufferBefore = (UInt32)0xFFFFFFFF - (1 << 22); // for debug p->keepSizeBefore = historySize + keepAddBufferBefore + 1; - p->keepSizeAfter = matchMaxLen + keepAddBufferAfter; - - /* we need one additional byte, since we use MoveBlock after pos++ and before dictionary using */ - - if (LzInWindow_Create(p, sizeReserv, alloc)) + + keepAddBufferAfter += matchMaxLen; + /* we need (p->keepSizeAfter >= p->numHashBytes) */ + if (keepAddBufferAfter < p->numHashBytes) + keepAddBufferAfter = p->numHashBytes; + // keepAddBufferAfter -= 2; // for debug + p->keepSizeAfter = keepAddBufferAfter; + + if (p->directInput) + p->blockSize = 0; + if (p->directInput || LzInWindow_Create2(p, GetBlockSize(p, historySize), alloc)) { - UInt32 newCyclicBufferSize = historySize + 1; - UInt32 hs; - p->matchMaxLen = matchMaxLen; + size_t hashSizeSum; { - p->fixedHashSize = 0; - if (p->numHashBytes == 2) - hs = (1 << 16) - 1; + UInt32 hs; + UInt32 hsCur; + + if (p->numHashOutBits != 0) + { + unsigned numBits = p->numHashOutBits; + const unsigned nbMax = + (p->numHashBytes == 2 ? 16 : + (p->numHashBytes == 3 ? 24 : 32)); + if (numBits > nbMax) + numBits = nbMax; + if (numBits >= 32) + hs = (UInt32)0 - 1; + else + hs = ((UInt32)1 << numBits) - 1; + // (hash_size >= (1 << 16)) : Required for (numHashBytes > 2) + hs |= (1 << 16) - 1; /* don't change it! */ + if (p->numHashBytes >= 5) + hs |= (256 << kLzHash_CrcShift_2) - 1; + { + const UInt32 hs2 = MatchFinder_GetHashMask2(p, historySize); + if (hs > hs2) + hs = hs2; + } + hsCur = hs; + if (p->expectedDataSize < historySize) + { + const UInt32 hs2 = MatchFinder_GetHashMask2(p, (UInt32)p->expectedDataSize); + if (hsCur > hs2) + hsCur = hs2; + } + } else { - hs = historySize - 1; - hs |= (hs >> 1); - hs |= (hs >> 2); - hs |= (hs >> 4); - hs |= (hs >> 8); - hs >>= 1; - hs |= 0xFFFF; /* don't change it! It's required for Deflate */ - if (hs > (1 << 24)) + hs = MatchFinder_GetHashMask(p, historySize); + hsCur = hs; + if (p->expectedDataSize < historySize) { - if (p->numHashBytes == 3) - hs = (1 << 24) - 1; - else - hs >>= 1; - /* if (bigHash) mode, GetHeads4b() in LzFindMt.c needs (hs >= ((1 << 24) - 1))) */ + hsCur = MatchFinder_GetHashMask(p, (UInt32)p->expectedDataSize); + if (hsCur > hs) // is it possible? + hsCur = hs; } } - p->hashMask = hs; - hs++; - if (p->numHashBytes > 2) p->fixedHashSize += kHash2Size; - if (p->numHashBytes > 3) p->fixedHashSize += kHash3Size; - if (p->numHashBytes > 4) p->fixedHashSize += kHash4Size; - hs += p->fixedHashSize; + + p->hashMask = hsCur; + + hashSizeSum = hs; + hashSizeSum++; + if (hashSizeSum < hs) + return 0; + { + UInt32 fixedHashSize = 0; + if (p->numHashBytes > 2 && p->numHashBytes_Min <= 2) fixedHashSize += kHash2Size; + if (p->numHashBytes > 3 && p->numHashBytes_Min <= 3) fixedHashSize += kHash3Size; + // if (p->numHashBytes > 4) p->fixedHashSize += hs4; // kHash4Size; + hashSizeSum += fixedHashSize; + p->fixedHashSize = fixedHashSize; + } } + p->matchMaxLen = matchMaxLen; + { size_t newSize; size_t numSons; + const UInt32 newCyclicBufferSize = historySize + 1; // do not change it p->historySize = historySize; - p->hashSizeSum = hs; - p->cyclicBufferSize = newCyclicBufferSize; + p->cyclicBufferSize = newCyclicBufferSize; // it must be = (historySize + 1) numSons = newCyclicBufferSize; if (p->btMode) numSons <<= 1; - newSize = hs + numSons; + newSize = hashSizeSum + numSons; - if (p->hash && p->numRefs == newSize) + if (numSons < newCyclicBufferSize || newSize < numSons) + return 0; + + // aligned size is not required here, but it can be better for some loops + #define NUM_REFS_ALIGN_MASK 0xF + newSize = (newSize + NUM_REFS_ALIGN_MASK) & ~(size_t)NUM_REFS_ALIGN_MASK; + + // 22.02: we don't reallocate buffer, if old size is enough + if (p->hash && p->numRefs >= newSize) return 1; MatchFinder_FreeThisClassMemory(p, alloc); @@ -253,7 +480,7 @@ int MatchFinder_Create(CMatchFinder *p, UInt32 historySize, if (p->hash) { - p->son = p->hash + p->hashSizeSum; + p->son = p->hash + hashSizeSum; return 1; } } @@ -263,107 +490,397 @@ int MatchFinder_Create(CMatchFinder *p, UInt32 historySize, return 0; } + static void MatchFinder_SetLimits(CMatchFinder *p) { - UInt32 limit = kMaxValForNormalize - p->pos; - UInt32 limit2 = p->cyclicBufferSize - p->cyclicBufferPos; - - if (limit2 < limit) - limit = limit2; - limit2 = p->streamPos - p->pos; + UInt32 k; + UInt32 n = kMaxValForNormalize - p->pos; + if (n == 0) + n = (UInt32)(Int32)-1; // we allow (pos == 0) at start even with (kMaxValForNormalize == 0) - if (limit2 <= p->keepSizeAfter) + k = p->cyclicBufferSize - p->cyclicBufferPos; + if (k < n) + n = k; + + k = GET_AVAIL_BYTES(p); { - if (limit2 > 0) - limit2 = 1; + const UInt32 ksa = p->keepSizeAfter; + UInt32 mm = p->matchMaxLen; + if (k > ksa) + k -= ksa; // we must limit exactly to keepSizeAfter for ReadBlock + else if (k >= mm) + { + // the limitation for (p->lenLimit) update + k -= mm; // optimization : to reduce the number of checks + k++; + // k = 1; // non-optimized version : for debug + } + else + { + mm = k; + if (k != 0) + k = 1; + } + p->lenLimit = mm; } - else - limit2 -= p->keepSizeAfter; - - if (limit2 < limit) - limit = limit2; + if (k < n) + n = k; - { - UInt32 lenLimit = p->streamPos - p->pos; - if (lenLimit > p->matchMaxLen) - lenLimit = p->matchMaxLen; - p->lenLimit = lenLimit; - } - p->posLimit = p->pos + limit; + p->posLimit = p->pos + n; } -void MatchFinder_Init_2(CMatchFinder *p, int readData) + +void MatchFinder_Init_LowHash(CMatchFinder *p) { - UInt32 i; - UInt32 *hash = p->hash; - UInt32 num = p->hashSizeSum; - for (i = 0; i < num; i++) - hash[i] = kEmptyHashValue; - - p->cyclicBufferPos = 0; - p->buffer = p->bufferBase; - p->pos = p->streamPos = p->cyclicBufferSize; + size_t i; + CLzRef *items = p->hash; + const size_t numItems = p->fixedHashSize; + for (i = 0; i < numItems; i++) + items[i] = kEmptyHashValue; +} + + +void MatchFinder_Init_HighHash(CMatchFinder *p) +{ + size_t i; + CLzRef *items = p->hash + p->fixedHashSize; + const size_t numItems = (size_t)p->hashMask + 1; + for (i = 0; i < numItems; i++) + items[i] = kEmptyHashValue; +} + + +void MatchFinder_Init_4(CMatchFinder *p) +{ + if (!p->directInput) + p->buffer = p->bufBase; + { + /* kEmptyHashValue = 0 (Zero) is used in hash tables as NO-VALUE marker. + the code in CMatchFinderMt expects (pos = 1) */ + p->pos = + p->streamPos = + 1; // it's smallest optimal value. do not change it + // 0; // for debug + } p->result = SZ_OK; p->streamEndWasReached = 0; - - if (readData) - MatchFinder_ReadBlock(p); - - MatchFinder_SetLimits(p); } + +// (CYC_TO_POS_OFFSET == 0) is expected by some optimized code +#define CYC_TO_POS_OFFSET 0 +// #define CYC_TO_POS_OFFSET 1 // for debug + void MatchFinder_Init(CMatchFinder *p) { - MatchFinder_Init_2(p, True); + MatchFinder_Init_HighHash(p); + MatchFinder_Init_LowHash(p); + MatchFinder_Init_4(p); + // if (readData) + MatchFinder_ReadBlock(p); + + /* if we init (cyclicBufferPos = pos), then we can use one variable + instead of both (cyclicBufferPos) and (pos) : only before (cyclicBufferPos) wrapping */ + p->cyclicBufferPos = (p->pos - CYC_TO_POS_OFFSET); // init with relation to (pos) + // p->cyclicBufferPos = 0; // smallest value + // p->son[0] = p->son[1] = 0; // unused: we can init skipped record for speculated accesses. + MatchFinder_SetLimits(p); } - -static UInt32 MatchFinder_GetSubValue(CMatchFinder *p) + + + +#ifdef MY_CPU_X86_OR_AMD64 + #if defined(__clang__) && (__clang_major__ >= 4) \ + || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40701) + // || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1900) + + #define USE_LZFIND_SATUR_SUB_128 + #define USE_LZFIND_SATUR_SUB_256 + #define LZFIND_ATTRIB_SSE41 __attribute__((__target__("sse4.1"))) + #define LZFIND_ATTRIB_AVX2 __attribute__((__target__("avx2"))) + #elif defined(_MSC_VER) + #if (_MSC_VER >= 1600) + #define USE_LZFIND_SATUR_SUB_128 + #endif + #if (_MSC_VER >= 1900) + #define USE_LZFIND_SATUR_SUB_256 + #endif + #endif + +// #elif defined(MY_CPU_ARM_OR_ARM64) +#elif defined(MY_CPU_ARM64) + + #if defined(__clang__) && (__clang_major__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ >= 8) + #define USE_LZFIND_SATUR_SUB_128 + #ifdef MY_CPU_ARM64 + // #define LZFIND_ATTRIB_SSE41 __attribute__((__target__(""))) + #else + // #define LZFIND_ATTRIB_SSE41 __attribute__((__target__("fpu=crypto-neon-fp-armv8"))) + #endif + + #elif defined(_MSC_VER) + #if (_MSC_VER >= 1910) + #define USE_LZFIND_SATUR_SUB_128 + #endif + #endif + + #if defined(_MSC_VER) && defined(MY_CPU_ARM64) + #include <arm64_neon.h> + #else + #include <arm_neon.h> + #endif + +#endif + + +#ifdef USE_LZFIND_SATUR_SUB_128 + +// #define Z7_SHOW_HW_STATUS + +#ifdef Z7_SHOW_HW_STATUS +#include <stdio.h> +#define PRF(x) x +PRF(;) +#else +#define PRF(x) +#endif + + +#ifdef MY_CPU_ARM_OR_ARM64 + +#ifdef MY_CPU_ARM64 +// #define FORCE_LZFIND_SATUR_SUB_128 +#endif +typedef uint32x4_t LzFind_v128; +#define SASUB_128_V(v, s) \ + vsubq_u32(vmaxq_u32(v, s), s) + +#else // MY_CPU_ARM_OR_ARM64 + +#include <smmintrin.h> // sse4.1 + +typedef __m128i LzFind_v128; +// SSE 4.1 +#define SASUB_128_V(v, s) \ + _mm_sub_epi32(_mm_max_epu32(v, s), s) + +#endif // MY_CPU_ARM_OR_ARM64 + + +#define SASUB_128(i) \ + *( LzFind_v128 *)( void *)(items + (i) * 4) = SASUB_128_V( \ + *(const LzFind_v128 *)(const void *)(items + (i) * 4), sub2); + + +Z7_NO_INLINE +static +#ifdef LZFIND_ATTRIB_SSE41 +LZFIND_ATTRIB_SSE41 +#endif +void +Z7_FASTCALL +LzFind_SaturSub_128(UInt32 subValue, CLzRef *items, const CLzRef *lim) { - return (p->pos - p->historySize - 1) & kNormalizeMask; + const LzFind_v128 sub2 = + #ifdef MY_CPU_ARM_OR_ARM64 + vdupq_n_u32(subValue); + #else + _mm_set_epi32((Int32)subValue, (Int32)subValue, (Int32)subValue, (Int32)subValue); + #endif + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SASUB_128(0) SASUB_128(1) items += 2 * 4; + SASUB_128(0) SASUB_128(1) items += 2 * 4; + } + while (items != lim); } -void MatchFinder_Normalize3(UInt32 subValue, CLzRef *items, size_t numItems) + + +#ifdef USE_LZFIND_SATUR_SUB_256 + +#include <immintrin.h> // avx +/* +clang :immintrin.h uses +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__AVX2__) +#include <avx2intrin.h> +#endif +so we need <avxintrin.h> for clang-cl */ + +#if defined(__clang__) +#include <avxintrin.h> +#include <avx2intrin.h> +#endif + +// AVX2: +#define SASUB_256(i) \ + *( __m256i *)( void *)(items + (i) * 8) = \ + _mm256_sub_epi32(_mm256_max_epu32( \ + *(const __m256i *)(const void *)(items + (i) * 8), sub2), sub2); + +Z7_NO_INLINE +static +#ifdef LZFIND_ATTRIB_AVX2 +LZFIND_ATTRIB_AVX2 +#endif +void +Z7_FASTCALL +LzFind_SaturSub_256(UInt32 subValue, CLzRef *items, const CLzRef *lim) { - size_t i; - for (i = 0; i < numItems; i++) + const __m256i sub2 = _mm256_set_epi32( + (Int32)subValue, (Int32)subValue, (Int32)subValue, (Int32)subValue, + (Int32)subValue, (Int32)subValue, (Int32)subValue, (Int32)subValue); + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do { - UInt32 value = items[i]; - if (value <= subValue) - value = kEmptyHashValue; - else - value -= subValue; - items[i] = value; + SASUB_256(0) SASUB_256(1) items += 2 * 8; + SASUB_256(0) SASUB_256(1) items += 2 * 8; } + while (items != lim); } +#endif // USE_LZFIND_SATUR_SUB_256 + +#ifndef FORCE_LZFIND_SATUR_SUB_128 +typedef void (Z7_FASTCALL *LZFIND_SATUR_SUB_CODE_FUNC)( + UInt32 subValue, CLzRef *items, const CLzRef *lim); +static LZFIND_SATUR_SUB_CODE_FUNC g_LzFind_SaturSub; +#endif // FORCE_LZFIND_SATUR_SUB_128 + +#endif // USE_LZFIND_SATUR_SUB_128 + + +// kEmptyHashValue must be zero +// #define SASUB_32(i) { UInt32 v = items[i]; UInt32 m = v - subValue; if (v < subValue) m = kEmptyHashValue; items[i] = m; } +#define SASUB_32(i) { UInt32 v = items[i]; if (v < subValue) v = subValue; items[i] = v - subValue; } -static void MatchFinder_Normalize(CMatchFinder *p) +#ifdef FORCE_LZFIND_SATUR_SUB_128 + +#define DEFAULT_SaturSub LzFind_SaturSub_128 + +#else + +#define DEFAULT_SaturSub LzFind_SaturSub_32 + +Z7_NO_INLINE +static +void +Z7_FASTCALL +LzFind_SaturSub_32(UInt32 subValue, CLzRef *items, const CLzRef *lim) { - UInt32 subValue = MatchFinder_GetSubValue(p); - MatchFinder_Normalize3(subValue, p->hash, p->numRefs); - MatchFinder_ReduceOffsets(p, subValue); + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SASUB_32(0) SASUB_32(1) items += 2; + SASUB_32(0) SASUB_32(1) items += 2; + SASUB_32(0) SASUB_32(1) items += 2; + SASUB_32(0) SASUB_32(1) items += 2; + } + while (items != lim); } +#endif + + +Z7_NO_INLINE +void MatchFinder_Normalize3(UInt32 subValue, CLzRef *items, size_t numItems) +{ + #define LZFIND_NORM_ALIGN_BLOCK_SIZE (1 << 7) + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0 && ((unsigned)(ptrdiff_t)items & (LZFIND_NORM_ALIGN_BLOCK_SIZE - 1)) != 0; numItems--) + { + SASUB_32(0) + items++; + } + { + const size_t k_Align_Mask = (LZFIND_NORM_ALIGN_BLOCK_SIZE / 4 - 1); + CLzRef *lim = items + (numItems & ~(size_t)k_Align_Mask); + numItems &= k_Align_Mask; + if (items != lim) + { + #if defined(USE_LZFIND_SATUR_SUB_128) && !defined(FORCE_LZFIND_SATUR_SUB_128) + if (g_LzFind_SaturSub) + g_LzFind_SaturSub(subValue, items, lim); + else + #endif + DEFAULT_SaturSub(subValue, items, lim); + } + items = lim; + } + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0; numItems--) + { + SASUB_32(0) + items++; + } +} + + + +// call MatchFinder_CheckLimits() only after (p->pos++) update + +Z7_NO_INLINE static void MatchFinder_CheckLimits(CMatchFinder *p) { + if (// !p->streamEndWasReached && p->result == SZ_OK && + p->keepSizeAfter == GET_AVAIL_BYTES(p)) + { + // we try to read only in exact state (p->keepSizeAfter == GET_AVAIL_BYTES(p)) + if (MatchFinder_NeedMove(p)) + MatchFinder_MoveBlock(p); + MatchFinder_ReadBlock(p); + } + if (p->pos == kMaxValForNormalize) - MatchFinder_Normalize(p); - if (!p->streamEndWasReached && p->keepSizeAfter == p->streamPos - p->pos) - MatchFinder_CheckAndMoveAndRead(p); + if (GET_AVAIL_BYTES(p) >= p->numHashBytes) // optional optimization for last bytes of data. + /* + if we disable normalization for last bytes of data, and + if (data_size == 4 GiB), we don't call wastfull normalization, + but (pos) will be wrapped over Zero (0) in that case. + And we cannot resume later to normal operation + */ + { + // MatchFinder_Normalize(p); + /* after normalization we need (p->pos >= p->historySize + 1); */ + /* we can reduce subValue to aligned value, if want to keep alignment + of (p->pos) and (p->buffer) for speculated accesses. */ + const UInt32 subValue = (p->pos - p->historySize - 1) /* & ~(UInt32)(kNormalizeAlign - 1) */; + // const UInt32 subValue = (1 << 15); // for debug + // printf("\nMatchFinder_Normalize() subValue == 0x%x\n", subValue); + MatchFinder_REDUCE_OFFSETS(p, subValue) + MatchFinder_Normalize3(subValue, p->hash, (size_t)p->hashMask + 1 + p->fixedHashSize); + { + size_t numSonRefs = p->cyclicBufferSize; + if (p->btMode) + numSonRefs <<= 1; + MatchFinder_Normalize3(subValue, p->son, numSonRefs); + } + } + if (p->cyclicBufferPos == p->cyclicBufferSize) p->cyclicBufferPos = 0; + MatchFinder_SetLimits(p); } -static UInt32 * Hc_GetMatchesSpec(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const Byte *cur, CLzRef *son, - UInt32 _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue, - UInt32 *distances, UInt32 maxLen) + +/* + (lenLimit > maxLen) +*/ +Z7_FORCE_INLINE +static UInt32 * Hc_GetMatchesSpec(size_t lenLimit, UInt32 curMatch, UInt32 pos, const Byte *cur, CLzRef *son, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue, + UInt32 *d, unsigned maxLen) { + /* son[_cyclicBufferPos] = curMatch; for (;;) { UInt32 delta = pos - curMatch; if (cutValue-- == 0 || delta >= _cyclicBufferSize) - return distances; + return d; { const Byte *pb = cur - delta; curMatch = son[_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)]; @@ -375,35 +892,91 @@ static UInt32 * Hc_GetMatchesSpec(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, break; if (maxLen < len) { - *distances++ = maxLen = len; - *distances++ = delta - 1; + maxLen = len; + *d++ = len; + *d++ = delta - 1; if (len == lenLimit) - return distances; + return d; } } } } + */ + + const Byte *lim = cur + lenLimit; + son[_cyclicBufferPos] = curMatch; + + do + { + UInt32 delta; + + if (curMatch == 0) + break; + // if (curMatch2 >= curMatch) return NULL; + delta = pos - curMatch; + if (delta >= _cyclicBufferSize) + break; + { + ptrdiff_t diff; + curMatch = son[_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)]; + diff = (ptrdiff_t)0 - (ptrdiff_t)delta; + if (cur[maxLen] == cur[(ptrdiff_t)maxLen + diff]) + { + const Byte *c = cur; + while (*c == c[diff]) + { + if (++c == lim) + { + d[0] = (UInt32)(lim - cur); + d[1] = delta - 1; + return d + 2; + } + } + { + const unsigned len = (unsigned)(c - cur); + if (maxLen < len) + { + maxLen = len; + d[0] = (UInt32)len; + d[1] = delta - 1; + d += 2; + } + } + } + } + } + while (--cutValue); + + return d; } + +Z7_FORCE_INLINE UInt32 * GetMatchesSpec1(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const Byte *cur, CLzRef *son, - UInt32 _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue, - UInt32 *distances, UInt32 maxLen) + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue, + UInt32 *d, UInt32 maxLen) { - CLzRef *ptr0 = son + (_cyclicBufferPos << 1) + 1; - CLzRef *ptr1 = son + (_cyclicBufferPos << 1); - UInt32 len0 = 0, len1 = 0; - for (;;) + CLzRef *ptr0 = son + ((size_t)_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + unsigned len0 = 0, len1 = 0; + + UInt32 cmCheck; + + // if (curMatch >= pos) { *ptr0 = *ptr1 = kEmptyHashValue; return NULL; } + + cmCheck = (UInt32)(pos - _cyclicBufferSize); + if ((UInt32)pos <= _cyclicBufferSize) + cmCheck = 0; + + if (cmCheck < curMatch) + do { - UInt32 delta = pos - curMatch; - if (cutValue-- == 0 || delta >= _cyclicBufferSize) - { - *ptr0 = *ptr1 = kEmptyHashValue; - return distances; - } + const UInt32 delta = pos - curMatch; { - CLzRef *pair = son + ((_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); + CLzRef *pair = son + ((size_t)(_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); const Byte *pb = cur - delta; - UInt32 len = (len0 < len1 ? len0 : len1); + unsigned len = (len0 < len1 ? len0 : len1); + const UInt32 pair0 = pair[0]; if (pb[len] == cur[len]) { if (++len != lenLimit && pb[len] == cur[len]) @@ -412,52 +985,65 @@ UInt32 * GetMatchesSpec1(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const Byt break; if (maxLen < len) { - *distances++ = maxLen = len; - *distances++ = delta - 1; + maxLen = (UInt32)len; + *d++ = (UInt32)len; + *d++ = delta - 1; if (len == lenLimit) { - *ptr1 = pair[0]; + *ptr1 = pair0; *ptr0 = pair[1]; - return distances; + return d; } } } if (pb[len] < cur[len]) { *ptr1 = curMatch; + // const UInt32 curMatch2 = pair[1]; + // if (curMatch2 >= curMatch) { *ptr0 = *ptr1 = kEmptyHashValue; return NULL; } + // curMatch = curMatch2; + curMatch = pair[1]; ptr1 = pair + 1; - curMatch = *ptr1; len1 = len; } else { *ptr0 = curMatch; + curMatch = pair[0]; ptr0 = pair; - curMatch = *ptr0; len0 = len; } } } + while(--cutValue && cmCheck < curMatch); + + *ptr0 = *ptr1 = kEmptyHashValue; + return d; } + static void SkipMatchesSpec(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const Byte *cur, CLzRef *son, - UInt32 _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue) + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 cutValue) { - CLzRef *ptr0 = son + (_cyclicBufferPos << 1) + 1; - CLzRef *ptr1 = son + (_cyclicBufferPos << 1); - UInt32 len0 = 0, len1 = 0; - for (;;) + CLzRef *ptr0 = son + ((size_t)_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + unsigned len0 = 0, len1 = 0; + + UInt32 cmCheck; + + cmCheck = (UInt32)(pos - _cyclicBufferSize); + if ((UInt32)pos <= _cyclicBufferSize) + cmCheck = 0; + + if (// curMatch >= pos || // failure + cmCheck < curMatch) + do { - UInt32 delta = pos - curMatch; - if (cutValue-- == 0 || delta >= _cyclicBufferSize) + const UInt32 delta = pos - curMatch; { - *ptr0 = *ptr1 = kEmptyHashValue; - return; - } - { - CLzRef *pair = son + ((_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); + CLzRef *pair = son + ((size_t)(_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); const Byte *pb = cur - delta; - UInt32 len = (len0 < len1 ? len0 : len1); + unsigned len = (len0 < len1 ? len0 : len1); if (pb[len] == cur[len]) { while (++len != lenLimit) @@ -475,549 +1061,594 @@ static void SkipMatchesSpec(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const if (pb[len] < cur[len]) { *ptr1 = curMatch; + curMatch = pair[1]; ptr1 = pair + 1; - curMatch = *ptr1; len1 = len; } else { *ptr0 = curMatch; + curMatch = pair[0]; ptr0 = pair; - curMatch = *ptr0; len0 = len; } } } + while(--cutValue && cmCheck < curMatch); + + *ptr0 = *ptr1 = kEmptyHashValue; + return; } + #define MOVE_POS \ ++p->cyclicBufferPos; \ p->buffer++; \ - if (++p->pos == p->posLimit) MatchFinder_CheckLimits(p); + { const UInt32 pos1 = p->pos + 1; p->pos = pos1; if (pos1 == p->posLimit) MatchFinder_CheckLimits(p); } -#define MOVE_POS_RET MOVE_POS return offset; +#define MOVE_POS_RET MOVE_POS return distances; -static void MatchFinder_MovePos(CMatchFinder *p) { MOVE_POS; } +Z7_NO_INLINE +static void MatchFinder_MovePos(CMatchFinder *p) +{ + /* we go here at the end of stream data, when (avail < num_hash_bytes) + We don't update sons[cyclicBufferPos << btMode]. + So (sons) record will contain junk. And we cannot resume match searching + to normal operation, even if we will provide more input data in buffer. + p->sons[p->cyclicBufferPos << p->btMode] = 0; // kEmptyHashValue + if (p->btMode) + p->sons[(p->cyclicBufferPos << p->btMode) + 1] = 0; // kEmptyHashValue + */ + MOVE_POS +} #define GET_MATCHES_HEADER2(minLen, ret_op) \ - UInt32 lenLimit; UInt32 hv; const Byte *cur; UInt32 curMatch; \ - lenLimit = p->lenLimit; { if (lenLimit < minLen) { MatchFinder_MovePos(p); ret_op; }} \ + unsigned lenLimit; UInt32 hv; const Byte *cur; UInt32 curMatch; \ + lenLimit = (unsigned)p->lenLimit; { if (lenLimit < minLen) { MatchFinder_MovePos(p); ret_op; }} \ cur = p->buffer; -#define GET_MATCHES_HEADER(minLen) GET_MATCHES_HEADER2(minLen, return 0) -#define SKIP_HEADER(minLen) GET_MATCHES_HEADER2(minLen, continue) +#define GET_MATCHES_HEADER(minLen) GET_MATCHES_HEADER2(minLen, return distances) +#define SKIP_HEADER(minLen) do { GET_MATCHES_HEADER2(minLen, continue) + +#define MF_PARAMS(p) lenLimit, curMatch, p->pos, p->buffer, p->son, p->cyclicBufferPos, p->cyclicBufferSize, p->cutValue + +#define SKIP_FOOTER SkipMatchesSpec(MF_PARAMS(p)); MOVE_POS } while (--num); -#define MF_PARAMS(p) p->pos, p->buffer, p->son, p->cyclicBufferPos, p->cyclicBufferSize, p->cutValue +#define GET_MATCHES_FOOTER_BASE(_maxLen_, func) \ + distances = func(MF_PARAMS(p), \ + distances, (UInt32)_maxLen_); MOVE_POS_RET + +#define GET_MATCHES_FOOTER_BT(_maxLen_) \ + GET_MATCHES_FOOTER_BASE(_maxLen_, GetMatchesSpec1) + +#define GET_MATCHES_FOOTER_HC(_maxLen_) \ + GET_MATCHES_FOOTER_BASE(_maxLen_, Hc_GetMatchesSpec) -#define GET_MATCHES_FOOTER(offset, maxLen) \ - offset = (UInt32)(GetMatchesSpec1(lenLimit, curMatch, MF_PARAMS(p), \ - distances + offset, maxLen) - distances); MOVE_POS_RET; -#define SKIP_FOOTER \ - SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); MOVE_POS; #define UPDATE_maxLen { \ - ptrdiff_t diff = (ptrdiff_t)0 - d2; \ + const ptrdiff_t diff = (ptrdiff_t)0 - (ptrdiff_t)d2; \ const Byte *c = cur + maxLen; \ const Byte *lim = cur + lenLimit; \ for (; c != lim; c++) if (*(c + diff) != *c) break; \ - maxLen = (UInt32)(c - cur); } + maxLen = (unsigned)(c - cur); } -static UInt32 Bt2_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) +static UInt32* Bt2_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 offset; GET_MATCHES_HEADER(2) - HASH2_CALC; + HASH2_CALC curMatch = p->hash[hv]; p->hash[hv] = p->pos; - offset = 0; - GET_MATCHES_FOOTER(offset, 1) + GET_MATCHES_FOOTER_BT(1) } -UInt32 Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) +UInt32* Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 offset; GET_MATCHES_HEADER(3) - HASH_ZIP_CALC; + HASH_ZIP_CALC curMatch = p->hash[hv]; p->hash[hv] = p->pos; - offset = 0; - GET_MATCHES_FOOTER(offset, 2) + GET_MATCHES_FOOTER_BT(2) } -static UInt32 Bt3_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +#define SET_mmm \ + mmm = p->cyclicBufferSize; \ + if (pos < mmm) \ + mmm = pos; + + +static UInt32* Bt3_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 h2, d2, maxLen, offset, pos; + UInt32 mmm; + UInt32 h2, d2, pos; + unsigned maxLen; UInt32 *hash; GET_MATCHES_HEADER(3) - HASH3_CALC; + HASH3_CALC hash = p->hash; pos = p->pos; d2 = pos - hash[h2]; - curMatch = hash[kFix3HashSize + hv]; + curMatch = (hash + kFix3HashSize)[hv]; hash[h2] = pos; - hash[kFix3HashSize + hv] = pos; + (hash + kFix3HashSize)[hv] = pos; + + SET_mmm maxLen = 2; - offset = 0; - if (d2 < p->cyclicBufferSize && *(cur - d2) == *cur) + if (d2 < mmm && *(cur - d2) == *cur) { UPDATE_maxLen - distances[0] = maxLen; + distances[0] = (UInt32)maxLen; distances[1] = d2 - 1; - offset = 2; + distances += 2; if (maxLen == lenLimit) { - SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); - MOVE_POS_RET; + SkipMatchesSpec(MF_PARAMS(p)); + MOVE_POS_RET } } - GET_MATCHES_FOOTER(offset, maxLen) + GET_MATCHES_FOOTER_BT(maxLen) } -static UInt32 Bt4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +static UInt32* Bt4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 h2, h3, d2, d3, maxLen, offset, pos; + UInt32 mmm; + UInt32 h2, h3, d2, d3, pos; + unsigned maxLen; UInt32 *hash; GET_MATCHES_HEADER(4) - HASH4_CALC; + HASH4_CALC hash = p->hash; pos = p->pos; - d2 = pos - hash[ h2]; - d3 = pos - hash[kFix3HashSize + h3]; + d2 = pos - hash [h2]; + d3 = pos - (hash + kFix3HashSize)[h3]; + curMatch = (hash + kFix4HashSize)[hv]; - curMatch = hash[kFix4HashSize + hv]; + hash [h2] = pos; + (hash + kFix3HashSize)[h3] = pos; + (hash + kFix4HashSize)[hv] = pos; - hash[ h2] = pos; - hash[kFix3HashSize + h3] = pos; - hash[kFix4HashSize + hv] = pos; + SET_mmm - maxLen = 0; - offset = 0; + maxLen = 3; - if (d2 < p->cyclicBufferSize && *(cur - d2) == *cur) - { - distances[0] = maxLen = 2; - distances[1] = d2 - 1; - offset = 2; - } - - if (d2 != d3 && d3 < p->cyclicBufferSize && *(cur - d3) == *cur) + for (;;) { - maxLen = 3; - distances[offset + 1] = d3 - 1; - offset += 2; - d2 = d3; - } + if (d2 < mmm && *(cur - d2) == *cur) + { + distances[0] = 2; + distances[1] = d2 - 1; + distances += 2; + if (*(cur - d2 + 2) == cur[2]) + { + // distances[-2] = 3; + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + d2 = d3; + distances[1] = d3 - 1; + distances += 2; + } + else + break; + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + d2 = d3; + distances[1] = d3 - 1; + distances += 2; + } + else + break; - if (offset != 0) - { UPDATE_maxLen - distances[offset - 2] = maxLen; + distances[-2] = (UInt32)maxLen; if (maxLen == lenLimit) { - SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); - MOVE_POS_RET; + SkipMatchesSpec(MF_PARAMS(p)); + MOVE_POS_RET } + break; } - if (maxLen < 3) - maxLen = 3; - - GET_MATCHES_FOOTER(offset, maxLen) + GET_MATCHES_FOOTER_BT(maxLen) } -/* -static UInt32 Bt5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +static UInt32* Bt5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 h2, h3, h4, d2, d3, d4, maxLen, offset, pos; + UInt32 mmm; + UInt32 h2, h3, d2, d3, maxLen, pos; UInt32 *hash; GET_MATCHES_HEADER(5) - HASH5_CALC; + HASH5_CALC hash = p->hash; pos = p->pos; - d2 = pos - hash[ h2]; - d3 = pos - hash[kFix3HashSize + h3]; - d4 = pos - hash[kFix4HashSize + h4]; + d2 = pos - hash [h2]; + d3 = pos - (hash + kFix3HashSize)[h3]; + // d4 = pos - (hash + kFix4HashSize)[h4]; + + curMatch = (hash + kFix5HashSize)[hv]; - curMatch = hash[kFix5HashSize + hv]; + hash [h2] = pos; + (hash + kFix3HashSize)[h3] = pos; + // (hash + kFix4HashSize)[h4] = pos; + (hash + kFix5HashSize)[hv] = pos; - hash[ h2] = pos; - hash[kFix3HashSize + h3] = pos; - hash[kFix4HashSize + h4] = pos; - hash[kFix5HashSize + hv] = pos; + SET_mmm - maxLen = 0; - offset = 0; + maxLen = 4; - if (d2 < p->cyclicBufferSize && *(cur - d2) == *cur) + for (;;) { - distances[0] = maxLen = 2; - distances[1] = d2 - 1; - offset = 2; - if (*(cur - d2 + 2) == cur[2]) - distances[0] = maxLen = 3; - else if (d3 < p->cyclicBufferSize && *(cur - d3) == *cur) + if (d2 < mmm && *(cur - d2) == *cur) { - distances[2] = maxLen = 3; - distances[3] = d3 - 1; - offset = 4; + distances[0] = 2; + distances[1] = d2 - 1; + distances += 2; + if (*(cur - d2 + 2) == cur[2]) + { + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + distances[1] = d3 - 1; + distances += 2; + d2 = d3; + } + else + break; + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + distances[1] = d3 - 1; + distances += 2; d2 = d3; } - } - else if (d3 < p->cyclicBufferSize && *(cur - d3) == *cur) - { - distances[0] = maxLen = 3; - distances[1] = d3 - 1; - offset = 2; - d2 = d3; - } - - if (d2 != d4 && d4 < p->cyclicBufferSize - && *(cur - d4) == *cur - && *(cur - d4 + 3) == *(cur + 3)) - { - maxLen = 4; - distances[offset + 1] = d4 - 1; - offset += 2; - d2 = d4; - } - - if (offset != 0) - { + else + break; + + distances[-2] = 3; + if (*(cur - d2 + 3) != cur[3]) + break; UPDATE_maxLen - distances[offset - 2] = maxLen; + distances[-2] = (UInt32)maxLen; if (maxLen == lenLimit) { - SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); - MOVE_POS_RET; + SkipMatchesSpec(MF_PARAMS(p)); + MOVE_POS_RET } + break; } - - if (maxLen < 4) - maxLen = 4; - GET_MATCHES_FOOTER(offset, maxLen) + GET_MATCHES_FOOTER_BT(maxLen) } -*/ -static UInt32 Hc4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +static UInt32* Hc4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 h2, h3, d2, d3, maxLen, offset, pos; + UInt32 mmm; + UInt32 h2, h3, d2, d3, pos; + unsigned maxLen; UInt32 *hash; GET_MATCHES_HEADER(4) - HASH4_CALC; + HASH4_CALC hash = p->hash; pos = p->pos; - d2 = pos - hash[ h2]; - d3 = pos - hash[kFix3HashSize + h3]; - - curMatch = hash[kFix4HashSize + hv]; + d2 = pos - hash [h2]; + d3 = pos - (hash + kFix3HashSize)[h3]; + curMatch = (hash + kFix4HashSize)[hv]; - hash[ h2] = pos; - hash[kFix3HashSize + h3] = pos; - hash[kFix4HashSize + hv] = pos; + hash [h2] = pos; + (hash + kFix3HashSize)[h3] = pos; + (hash + kFix4HashSize)[hv] = pos; - maxLen = 0; - offset = 0; + SET_mmm - if (d2 < p->cyclicBufferSize && *(cur - d2) == *cur) - { - distances[0] = maxLen = 2; - distances[1] = d2 - 1; - offset = 2; - } - - if (d2 != d3 && d3 < p->cyclicBufferSize && *(cur - d3) == *cur) - { - maxLen = 3; - distances[offset + 1] = d3 - 1; - offset += 2; - d2 = d3; - } - - if (offset != 0) + maxLen = 3; + + for (;;) { + if (d2 < mmm && *(cur - d2) == *cur) + { + distances[0] = 2; + distances[1] = d2 - 1; + distances += 2; + if (*(cur - d2 + 2) == cur[2]) + { + // distances[-2] = 3; + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + d2 = d3; + distances[1] = d3 - 1; + distances += 2; + } + else + break; + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + d2 = d3; + distances[1] = d3 - 1; + distances += 2; + } + else + break; + UPDATE_maxLen - distances[offset - 2] = maxLen; + distances[-2] = (UInt32)maxLen; if (maxLen == lenLimit) { p->son[p->cyclicBufferPos] = curMatch; - MOVE_POS_RET; + MOVE_POS_RET } + break; } - if (maxLen < 3) - maxLen = 3; - - offset = (UInt32)(Hc_GetMatchesSpec(lenLimit, curMatch, MF_PARAMS(p), - distances + offset, maxLen) - (distances)); - MOVE_POS_RET + GET_MATCHES_FOOTER_HC(maxLen) } -/* -static UInt32 Hc5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +static UInt32 * Hc5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 h2, h3, h4, d2, d3, d4, maxLen, offset, pos + UInt32 mmm; + UInt32 h2, h3, d2, d3, maxLen, pos; UInt32 *hash; GET_MATCHES_HEADER(5) - HASH5_CALC; + HASH5_CALC hash = p->hash; pos = p->pos; - - d2 = pos - hash[ h2]; - d3 = pos - hash[kFix3HashSize + h3]; - d4 = pos - hash[kFix4HashSize + h4]; - curMatch = hash[kFix5HashSize + hv]; + d2 = pos - hash [h2]; + d3 = pos - (hash + kFix3HashSize)[h3]; + // d4 = pos - (hash + kFix4HashSize)[h4]; - hash[ h2] = pos; - hash[kFix3HashSize + h3] = pos; - hash[kFix4HashSize + h4] = pos; - hash[kFix5HashSize + hv] = pos; + curMatch = (hash + kFix5HashSize)[hv]; - maxLen = 0; - offset = 0; + hash [h2] = pos; + (hash + kFix3HashSize)[h3] = pos; + // (hash + kFix4HashSize)[h4] = pos; + (hash + kFix5HashSize)[hv] = pos; - if (d2 < p->cyclicBufferSize && *(cur - d2) == *cur) + SET_mmm + + maxLen = 4; + + for (;;) { - distances[0] = maxLen = 2; - distances[1] = d2 - 1; - offset = 2; - if (*(cur - d2 + 2) == cur[2]) - distances[0] = maxLen = 3; - else if (d3 < p->cyclicBufferSize && *(cur - d3) == *cur) + if (d2 < mmm && *(cur - d2) == *cur) + { + distances[0] = 2; + distances[1] = d2 - 1; + distances += 2; + if (*(cur - d2 + 2) == cur[2]) + { + } + else if (d3 < mmm && *(cur - d3) == *cur) + { + distances[1] = d3 - 1; + distances += 2; + d2 = d3; + } + else + break; + } + else if (d3 < mmm && *(cur - d3) == *cur) { - distances[2] = maxLen = 3; - distances[3] = d3 - 1; - offset = 4; + distances[1] = d3 - 1; + distances += 2; d2 = d3; } - } - else if (d3 < p->cyclicBufferSize && *(cur - d3) == *cur) - { - distances[0] = maxLen = 3; - distances[1] = d3 - 1; - offset = 2; - d2 = d3; - } - - if (d2 != d4 && d4 < p->cyclicBufferSize - && *(cur - d4) == *cur - && *(cur - d4 + 3) == *(cur + 3)) - { - maxLen = 4; - distances[offset + 1] = d4 - 1; - offset += 2; - d2 = d4; - } - - if (offset != 0) - { + else + break; + + distances[-2] = 3; + if (*(cur - d2 + 3) != cur[3]) + break; UPDATE_maxLen - distances[offset - 2] = maxLen; + distances[-2] = maxLen; if (maxLen == lenLimit) { p->son[p->cyclicBufferPos] = curMatch; - MOVE_POS_RET; + MOVE_POS_RET } + break; } - if (maxLen < 4) - maxLen = 4; - - offset = (UInt32)(Hc_GetMatchesSpec(lenLimit, curMatch, MF_PARAMS(p), - distances + offset, maxLen) - (distances)); - MOVE_POS_RET + GET_MATCHES_FOOTER_HC(maxLen) } -*/ -UInt32 Hc3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) + +UInt32* Hc3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances) { - UInt32 offset; GET_MATCHES_HEADER(3) - HASH_ZIP_CALC; + HASH_ZIP_CALC curMatch = p->hash[hv]; p->hash[hv] = p->pos; - offset = (UInt32)(Hc_GetMatchesSpec(lenLimit, curMatch, MF_PARAMS(p), - distances, 2) - (distances)); - MOVE_POS_RET + GET_MATCHES_FOOTER_HC(2) } + static void Bt2_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do + SKIP_HEADER(2) { - SKIP_HEADER(2) - HASH2_CALC; + HASH2_CALC curMatch = p->hash[hv]; p->hash[hv] = p->pos; - SKIP_FOOTER } - while (--num != 0); + SKIP_FOOTER } void Bt3Zip_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do + SKIP_HEADER(3) { - SKIP_HEADER(3) - HASH_ZIP_CALC; + HASH_ZIP_CALC curMatch = p->hash[hv]; p->hash[hv] = p->pos; - SKIP_FOOTER } - while (--num != 0); + SKIP_FOOTER } static void Bt3_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do + SKIP_HEADER(3) { UInt32 h2; UInt32 *hash; - SKIP_HEADER(3) - HASH3_CALC; + HASH3_CALC hash = p->hash; - curMatch = hash[kFix3HashSize + hv]; + curMatch = (hash + kFix3HashSize)[hv]; hash[h2] = - hash[kFix3HashSize + hv] = p->pos; - SKIP_FOOTER + (hash + kFix3HashSize)[hv] = p->pos; } - while (--num != 0); + SKIP_FOOTER } static void Bt4_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do + SKIP_HEADER(4) { UInt32 h2, h3; UInt32 *hash; - SKIP_HEADER(4) - HASH4_CALC; + HASH4_CALC hash = p->hash; - curMatch = hash[kFix4HashSize + hv]; - hash[ h2] = - hash[kFix3HashSize + h3] = - hash[kFix4HashSize + hv] = p->pos; - SKIP_FOOTER + curMatch = (hash + kFix4HashSize)[hv]; + hash [h2] = + (hash + kFix3HashSize)[h3] = + (hash + kFix4HashSize)[hv] = p->pos; } - while (--num != 0); + SKIP_FOOTER } -/* static void Bt5_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do + SKIP_HEADER(5) { - UInt32 h2, h3, h4; + UInt32 h2, h3; UInt32 *hash; - SKIP_HEADER(5) - HASH5_CALC; + HASH5_CALC hash = p->hash; - curMatch = hash[kFix5HashSize + hv]; - hash[ h2] = - hash[kFix3HashSize + h3] = - hash[kFix4HashSize + h4] = - hash[kFix5HashSize + hv] = p->pos; - SKIP_FOOTER + curMatch = (hash + kFix5HashSize)[hv]; + hash [h2] = + (hash + kFix3HashSize)[h3] = + // (hash + kFix4HashSize)[h4] = + (hash + kFix5HashSize)[hv] = p->pos; } - while (--num != 0); + SKIP_FOOTER } -*/ + + +#define HC_SKIP_HEADER(minLen) \ + do { if (p->lenLimit < minLen) { MatchFinder_MovePos(p); num--; continue; } { \ + const Byte *cur; \ + UInt32 *hash; \ + UInt32 *son; \ + UInt32 pos = p->pos; \ + UInt32 num2 = num; \ + /* (p->pos == p->posLimit) is not allowed here !!! */ \ + { const UInt32 rem = p->posLimit - pos; if (num2 > rem) num2 = rem; } \ + num -= num2; \ + { const UInt32 cycPos = p->cyclicBufferPos; \ + son = p->son + cycPos; \ + p->cyclicBufferPos = cycPos + num2; } \ + cur = p->buffer; \ + hash = p->hash; \ + do { \ + UInt32 curMatch; \ + UInt32 hv; + + +#define HC_SKIP_FOOTER \ + cur++; pos++; *son++ = curMatch; \ + } while (--num2); \ + p->buffer = cur; \ + p->pos = pos; \ + if (pos == p->posLimit) MatchFinder_CheckLimits(p); \ + }} while(num); \ + static void Hc4_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do - { + HC_SKIP_HEADER(4) + UInt32 h2, h3; - UInt32 *hash; - SKIP_HEADER(4) - HASH4_CALC; - hash = p->hash; - curMatch = hash[kFix4HashSize + hv]; - hash[ h2] = - hash[kFix3HashSize + h3] = - hash[kFix4HashSize + hv] = p->pos; - p->son[p->cyclicBufferPos] = curMatch; - MOVE_POS - } - while (--num != 0); + HASH4_CALC + curMatch = (hash + kFix4HashSize)[hv]; + hash [h2] = + (hash + kFix3HashSize)[h3] = + (hash + kFix4HashSize)[hv] = pos; + + HC_SKIP_FOOTER } -/* + static void Hc5_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do - { - UInt32 h2, h3, h4; - UInt32 *hash; - SKIP_HEADER(5) - HASH5_CALC; - hash = p->hash; - curMatch = p->hash[kFix5HashSize + hv]; - hash[ h2] = - hash[kFix3HashSize + h3] = - hash[kFix4HashSize + h4] = - hash[kFix5HashSize + hv] = p->pos; - p->son[p->cyclicBufferPos] = curMatch; - MOVE_POS - } - while (--num != 0); + HC_SKIP_HEADER(5) + + UInt32 h2, h3; + HASH5_CALC + curMatch = (hash + kFix5HashSize)[hv]; + hash [h2] = + (hash + kFix3HashSize)[h3] = + // (hash + kFix4HashSize)[h4] = + (hash + kFix5HashSize)[hv] = pos; + + HC_SKIP_FOOTER } -*/ + void Hc3Zip_MatchFinder_Skip(CMatchFinder *p, UInt32 num) { - do - { - SKIP_HEADER(3) - HASH_ZIP_CALC; - curMatch = p->hash[hv]; - p->hash[hv] = p->pos; - p->son[p->cyclicBufferPos] = curMatch; - MOVE_POS - } - while (--num != 0); + HC_SKIP_HEADER(3) + + HASH_ZIP_CALC + curMatch = hash[hv]; + hash[hv] = pos; + + HC_SKIP_FOOTER } -void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder *vTable) + +void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder2 *vTable) { vTable->Init = (Mf_Init_Func)MatchFinder_Init; vTable->GetNumAvailableBytes = (Mf_GetNumAvailableBytes_Func)MatchFinder_GetNumAvailableBytes; vTable->GetPointerToCurrentPos = (Mf_GetPointerToCurrentPos_Func)MatchFinder_GetPointerToCurrentPos; if (!p->btMode) { - /* if (p->numHashBytes <= 4) */ + if (p->numHashBytes <= 4) { vTable->GetMatches = (Mf_GetMatches_Func)Hc4_MatchFinder_GetMatches; vTable->Skip = (Mf_Skip_Func)Hc4_MatchFinder_Skip; } - /* else { vTable->GetMatches = (Mf_GetMatches_Func)Hc5_MatchFinder_GetMatches; vTable->Skip = (Mf_Skip_Func)Hc5_MatchFinder_Skip; } - */ } else if (p->numHashBytes == 2) { @@ -1029,16 +1660,58 @@ void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder *vTable) vTable->GetMatches = (Mf_GetMatches_Func)Bt3_MatchFinder_GetMatches; vTable->Skip = (Mf_Skip_Func)Bt3_MatchFinder_Skip; } - else /* if (p->numHashBytes == 4) */ + else if (p->numHashBytes == 4) { vTable->GetMatches = (Mf_GetMatches_Func)Bt4_MatchFinder_GetMatches; vTable->Skip = (Mf_Skip_Func)Bt4_MatchFinder_Skip; } - /* else { vTable->GetMatches = (Mf_GetMatches_Func)Bt5_MatchFinder_GetMatches; vTable->Skip = (Mf_Skip_Func)Bt5_MatchFinder_Skip; } - */ } + + + +void LzFindPrepare(void) +{ + #ifndef FORCE_LZFIND_SATUR_SUB_128 + #ifdef USE_LZFIND_SATUR_SUB_128 + LZFIND_SATUR_SUB_CODE_FUNC f = NULL; + #ifdef MY_CPU_ARM_OR_ARM64 + { + if (CPU_IsSupported_NEON()) + { + // #pragma message ("=== LzFind NEON") + PRF(printf("\n=== LzFind NEON\n")); + f = LzFind_SaturSub_128; + } + // f = 0; // for debug + } + #else // MY_CPU_ARM_OR_ARM64 + if (CPU_IsSupported_SSE41()) + { + // #pragma message ("=== LzFind SSE41") + PRF(printf("\n=== LzFind SSE41\n")); + f = LzFind_SaturSub_128; + + #ifdef USE_LZFIND_SATUR_SUB_256 + if (CPU_IsSupported_AVX2()) + { + // #pragma message ("=== LzFind AVX2") + PRF(printf("\n=== LzFind AVX2\n")); + f = LzFind_SaturSub_256; + } + #endif + } + #endif // MY_CPU_ARM_OR_ARM64 + g_LzFind_SaturSub = f; + #endif // USE_LZFIND_SATUR_SUB_128 + #endif // FORCE_LZFIND_SATUR_SUB_128 +} + + +#undef MOVE_POS +#undef MOVE_POS_RET +#undef PRF diff --git a/3rdparty/lzma/C/LzFind.h b/3rdparty/lzma/C/LzFind.h index d119944f445..a3f72c98700 100644 --- a/3rdparty/lzma/C/LzFind.h +++ b/3rdparty/lzma/C/LzFind.h @@ -1,8 +1,8 @@ /* LzFind.h -- Match finder for LZ algorithms -2015-10-15 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __LZ_FIND_H -#define __LZ_FIND_H +#ifndef ZIP7_INC_LZ_FIND_H +#define ZIP7_INC_LZ_FIND_H #include "7zTypes.h" @@ -10,12 +10,12 @@ EXTERN_C_BEGIN typedef UInt32 CLzRef; -typedef struct _CMatchFinder +typedef struct { - Byte *buffer; + const Byte *buffer; UInt32 pos; UInt32 posLimit; - UInt32 streamPos; + UInt32 streamPos; /* wrap over Zero is allowed (streamPos < pos). Use (UInt32)(streamPos - pos) */ UInt32 lenLimit; UInt32 cyclicBufferPos; @@ -32,8 +32,8 @@ typedef struct _CMatchFinder UInt32 hashMask; UInt32 cutValue; - Byte *bufferBase; - ISeqInStream *stream; + Byte *bufBase; + ISeqInStreamPtr stream; UInt32 blockSize; UInt32 keepSizeBefore; @@ -43,41 +43,79 @@ typedef struct _CMatchFinder size_t directInputRem; UInt32 historySize; UInt32 fixedHashSize; - UInt32 hashSizeSum; + Byte numHashBytes_Min; + Byte numHashOutBits; + Byte _pad2_[2]; SRes result; UInt32 crc[256]; size_t numRefs; + + UInt64 expectedDataSize; } CMatchFinder; -#define Inline_MatchFinder_GetPointerToCurrentPos(p) ((p)->buffer) +#define Inline_MatchFinder_GetPointerToCurrentPos(p) ((const Byte *)(p)->buffer) -#define Inline_MatchFinder_GetNumAvailableBytes(p) ((p)->streamPos - (p)->pos) +#define Inline_MatchFinder_GetNumAvailableBytes(p) ((UInt32)((p)->streamPos - (p)->pos)) +/* #define Inline_MatchFinder_IsFinishedOK(p) \ ((p)->streamEndWasReached \ && (p)->streamPos == (p)->pos \ && (!(p)->directInput || (p)->directInputRem == 0)) +*/ int MatchFinder_NeedMove(CMatchFinder *p); -Byte *MatchFinder_GetPointerToCurrentPos(CMatchFinder *p); +/* Byte *MatchFinder_GetPointerToCurrentPos(CMatchFinder *p); */ void MatchFinder_MoveBlock(CMatchFinder *p); void MatchFinder_ReadIfRequired(CMatchFinder *p); void MatchFinder_Construct(CMatchFinder *p); -/* Conditions: - historySize <= 3 GB - keepAddBufferBefore + matchMaxLen + keepAddBufferAfter < 511MB +/* (directInput = 0) is default value. + It's required to provide correct (directInput) value + before calling MatchFinder_Create(). + You can set (directInput) by any of the following calls: + - MatchFinder_SET_DIRECT_INPUT_BUF() + - MatchFinder_SET_STREAM() + - MatchFinder_SET_STREAM_MODE() +*/ + +#define MatchFinder_SET_DIRECT_INPUT_BUF(p, _src_, _srcLen_) { \ + (p)->stream = NULL; \ + (p)->directInput = 1; \ + (p)->buffer = (_src_); \ + (p)->directInputRem = (_srcLen_); } + +/* +#define MatchFinder_SET_STREAM_MODE(p) { \ + (p)->directInput = 0; } */ + +#define MatchFinder_SET_STREAM(p, _stream_) { \ + (p)->stream = _stream_; \ + (p)->directInput = 0; } + + int MatchFinder_Create(CMatchFinder *p, UInt32 historySize, UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter, - ISzAlloc *alloc); -void MatchFinder_Free(CMatchFinder *p, ISzAlloc *alloc); + ISzAllocPtr alloc); +void MatchFinder_Free(CMatchFinder *p, ISzAllocPtr alloc); void MatchFinder_Normalize3(UInt32 subValue, CLzRef *items, size_t numItems); -void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue); + +/* +#define MatchFinder_INIT_POS(p, val) \ + (p)->pos = (val); \ + (p)->streamPos = (val); +*/ + +// void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue); +#define MatchFinder_REDUCE_OFFSETS(p, subValue) \ + (p)->pos -= (subValue); \ + (p)->streamPos -= (subValue); + UInt32 * GetMatchesSpec1(UInt32 lenLimit, UInt32 curMatch, UInt32 pos, const Byte *buffer, CLzRef *son, - UInt32 _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 _cutValue, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 _cutValue, UInt32 *distances, UInt32 maxLen); /* @@ -89,29 +127,33 @@ Conditions: typedef void (*Mf_Init_Func)(void *object); typedef UInt32 (*Mf_GetNumAvailableBytes_Func)(void *object); typedef const Byte * (*Mf_GetPointerToCurrentPos_Func)(void *object); -typedef UInt32 (*Mf_GetMatches_Func)(void *object, UInt32 *distances); +typedef UInt32 * (*Mf_GetMatches_Func)(void *object, UInt32 *distances); typedef void (*Mf_Skip_Func)(void *object, UInt32); -typedef struct _IMatchFinder +typedef struct { Mf_Init_Func Init; Mf_GetNumAvailableBytes_Func GetNumAvailableBytes; Mf_GetPointerToCurrentPos_Func GetPointerToCurrentPos; Mf_GetMatches_Func GetMatches; Mf_Skip_Func Skip; -} IMatchFinder; +} IMatchFinder2; -void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder *vTable); +void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder2 *vTable); -void MatchFinder_Init_2(CMatchFinder *p, int readData); +void MatchFinder_Init_LowHash(CMatchFinder *p); +void MatchFinder_Init_HighHash(CMatchFinder *p); +void MatchFinder_Init_4(CMatchFinder *p); void MatchFinder_Init(CMatchFinder *p); -UInt32 Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances); -UInt32 Hc3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances); +UInt32* Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances); +UInt32* Hc3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances); void Bt3Zip_MatchFinder_Skip(CMatchFinder *p, UInt32 num); void Hc3Zip_MatchFinder_Skip(CMatchFinder *p, UInt32 num); +void LzFindPrepare(void); + EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/LzFindMt.c b/3rdparty/lzma/C/LzFindMt.c index a4ffa5ef9c8..5253e6ebb35 100644 --- a/3rdparty/lzma/C/LzFindMt.c +++ b/3rdparty/lzma/C/LzFindMt.c @@ -1,95 +1,215 @@ /* LzFindMt.c -- multithreaded Match finder for LZ algorithms -2015-10-15 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" -#include "LzHash.h" +// #include <stdio.h> + +#include "CpuArch.h" +#include "LzHash.h" #include "LzFindMt.h" +// #define LOG_ITERS + +// #define LOG_THREAD + +#ifdef LOG_THREAD +#include <stdio.h> +#define PRF(x) x +#else +#define PRF(x) +#endif + +#ifdef LOG_ITERS +#include <stdio.h> +extern UInt64 g_NumIters_Tree; +extern UInt64 g_NumIters_Loop; +extern UInt64 g_NumIters_Bytes; +#define LOG_ITER(x) x +#else +#define LOG_ITER(x) +#endif + +#define kMtHashBlockSize ((UInt32)1 << 17) +#define kMtHashNumBlocks (1 << 1) + +#define GET_HASH_BLOCK_OFFSET(i) (((i) & (kMtHashNumBlocks - 1)) * kMtHashBlockSize) + +#define kMtBtBlockSize ((UInt32)1 << 16) +#define kMtBtNumBlocks (1 << 4) + +#define GET_BT_BLOCK_OFFSET(i) (((i) & (kMtBtNumBlocks - 1)) * (size_t)kMtBtBlockSize) + +/* + HASH functions: + We use raw 8/16 bits from a[1] and a[2], + xored with crc(a[0]) and crc(a[3]). + We check a[0], a[3] only. We don't need to compare a[1] and a[2] in matches. + our crc() function provides one-to-one correspondence for low 8-bit values: + (crc[0...0xFF] & 0xFF) <-> [0...0xFF] +*/ + +#define MF(mt) ((mt)->MatchFinder) +#define MF_CRC (p->crc) + +// #define MF(mt) (&(mt)->MatchFinder) +// #define MF_CRC (p->MatchFinder.crc) + +#define MT_HASH2_CALC \ + h2 = (MF_CRC[cur[0]] ^ cur[1]) & (kHash2Size - 1); + +#define MT_HASH3_CALC { \ + UInt32 temp = MF_CRC[cur[0]] ^ cur[1]; \ + h2 = temp & (kHash2Size - 1); \ + h3 = (temp ^ ((UInt32)cur[2] << 8)) & (kHash3Size - 1); } + +/* +#define MT_HASH3_CALC__NO_2 { \ + UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ + h3 = (temp ^ ((UInt32)cur[2] << 8)) & (kHash3Size - 1); } + +#define MT_HASH4_CALC { \ + UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ + h2 = temp & (kHash2Size - 1); \ + temp ^= ((UInt32)cur[2] << 8); \ + h3 = temp & (kHash3Size - 1); \ + h4 = (temp ^ (p->crc[cur[3]] << kLzHash_CrcShift_1)) & p->hash4Mask; } + // (kHash4Size - 1); +*/ + + +Z7_NO_INLINE static void MtSync_Construct(CMtSync *p) { + p->affinity = 0; p->wasCreated = False; p->csWasInitialized = False; p->csWasEntered = False; - Thread_Construct(&p->thread); + Thread_CONSTRUCT(&p->thread) Event_Construct(&p->canStart); - Event_Construct(&p->wasStarted); Event_Construct(&p->wasStopped); Semaphore_Construct(&p->freeSemaphore); Semaphore_Construct(&p->filledSemaphore); } -static void MtSync_GetNextBlock(CMtSync *p) + +#define DEBUG_BUFFER_LOCK // define it to debug lock state + +#ifdef DEBUG_BUFFER_LOCK +#include <stdlib.h> +#define BUFFER_MUST_BE_LOCKED(p) if (!(p)->csWasEntered) exit(1); +#define BUFFER_MUST_BE_UNLOCKED(p) if ( (p)->csWasEntered) exit(1); +#else +#define BUFFER_MUST_BE_LOCKED(p) +#define BUFFER_MUST_BE_UNLOCKED(p) +#endif + +#define LOCK_BUFFER(p) { \ + BUFFER_MUST_BE_UNLOCKED(p); \ + CriticalSection_Enter(&(p)->cs); \ + (p)->csWasEntered = True; } + +#define UNLOCK_BUFFER(p) { \ + BUFFER_MUST_BE_LOCKED(p); \ + CriticalSection_Leave(&(p)->cs); \ + (p)->csWasEntered = False; } + + +Z7_NO_INLINE +static UInt32 MtSync_GetNextBlock(CMtSync *p) { + UInt32 numBlocks = 0; if (p->needStart) { + BUFFER_MUST_BE_UNLOCKED(p) p->numProcessedBlocks = 1; p->needStart = False; p->stopWriting = False; p->exit = False; - Event_Reset(&p->wasStarted); Event_Reset(&p->wasStopped); - Event_Set(&p->canStart); - Event_Wait(&p->wasStarted); } else { - CriticalSection_Leave(&p->cs); - p->csWasEntered = False; - p->numProcessedBlocks++; + UNLOCK_BUFFER(p) + // we free current block + numBlocks = p->numProcessedBlocks++; Semaphore_Release1(&p->freeSemaphore); } + + // buffer is UNLOCKED here Semaphore_Wait(&p->filledSemaphore); - CriticalSection_Enter(&p->cs); - p->csWasEntered = True; + LOCK_BUFFER(p) + return numBlocks; } -/* MtSync_StopWriting must be called if Writing was started */ +/* if Writing (Processing) thread was started, we must call MtSync_StopWriting() */ + +Z7_NO_INLINE static void MtSync_StopWriting(CMtSync *p) { - UInt32 myNumBlocks = p->numProcessedBlocks; if (!Thread_WasCreated(&p->thread) || p->needStart) return; - p->stopWriting = True; + + PRF(printf("\nMtSync_StopWriting %p\n", p)); + if (p->csWasEntered) { - CriticalSection_Leave(&p->cs); - p->csWasEntered = False; + /* we don't use buffer in this thread after StopWriting(). + So we UNLOCK buffer. + And we restore default UNLOCKED state for stopped thread */ + UNLOCK_BUFFER(p) } - Semaphore_Release1(&p->freeSemaphore); - + + /* We send (p->stopWriting) message and release freeSemaphore + to free current block. + So the thread will see (p->stopWriting) at some + iteration after Wait(freeSemaphore). + The thread doesn't need to fill all avail free blocks, + so we can get fast thread stop. + */ + + p->stopWriting = True; + Semaphore_Release1(&p->freeSemaphore); // check semaphore count !!! + + PRF(printf("\nMtSync_StopWriting %p : Event_Wait(&p->wasStopped)\n", p)); Event_Wait(&p->wasStopped); + PRF(printf("\nMtSync_StopWriting %p : Event_Wait() finsihed\n", p)); + + /* 21.03 : we don't restore samaphore counters here. + We will recreate and reinit samaphores in next start */ - while (myNumBlocks++ != p->numProcessedBlocks) - { - Semaphore_Wait(&p->filledSemaphore); - Semaphore_Release1(&p->freeSemaphore); - } p->needStart = True; } + +Z7_NO_INLINE static void MtSync_Destruct(CMtSync *p) { + PRF(printf("\nMtSync_Destruct %p\n", p)); + if (Thread_WasCreated(&p->thread)) { + /* we want thread to be in Stopped state before sending EXIT command. + note: stop(btSync) will stop (htSync) also */ MtSync_StopWriting(p); + /* thread in Stopped state here : (p->needStart == true) */ p->exit = True; - if (p->needStart) - Event_Set(&p->canStart); - Thread_Wait(&p->thread); - Thread_Close(&p->thread); + // if (p->needStart) // it's (true) + Event_Set(&p->canStart); // we send EXIT command to thread + Thread_Wait_Close(&p->thread); // we wait thread finishing } + if (p->csWasInitialized) { CriticalSection_Delete(&p->cs); p->csWasInitialized = False; } + p->csWasEntered = False; Event_Close(&p->canStart); - Event_Close(&p->wasStarted); Event_Close(&p->wasStopped); Semaphore_Close(&p->freeSemaphore); Semaphore_Close(&p->filledSemaphore); @@ -97,77 +217,251 @@ static void MtSync_Destruct(CMtSync *p) p->wasCreated = False; } -#define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; } -static SRes MtSync_Create2(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *obj, UInt32 numBlocks) +// #define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; } +// we want to get real system error codes here instead of SZ_ERROR_THREAD +#define RINOK_THREAD(x) RINOK_WRes(x) + + +// call it before each new file (when new starting is required): +Z7_NO_INLINE +static SRes MtSync_Init(CMtSync *p, UInt32 numBlocks) +{ + WRes wres; + // BUFFER_MUST_BE_UNLOCKED(p) + if (!p->needStart || p->csWasEntered) + return SZ_ERROR_FAIL; + wres = Semaphore_OptCreateInit(&p->freeSemaphore, numBlocks, numBlocks); + if (wres == 0) + wres = Semaphore_OptCreateInit(&p->filledSemaphore, 0, numBlocks); + return MY_SRes_HRESULT_FROM_WRes(wres); +} + + +static WRes MtSync_Create_WRes(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *obj) { + WRes wres; + if (p->wasCreated) return SZ_OK; - RINOK_THREAD(CriticalSection_Init(&p->cs)); + RINOK_THREAD(CriticalSection_Init(&p->cs)) p->csWasInitialized = True; + p->csWasEntered = False; - RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canStart)); - RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->wasStarted)); - RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->wasStopped)); - - RINOK_THREAD(Semaphore_Create(&p->freeSemaphore, numBlocks, numBlocks)); - RINOK_THREAD(Semaphore_Create(&p->filledSemaphore, 0, numBlocks)); + RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canStart)) + RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->wasStopped)) p->needStart = True; - - RINOK_THREAD(Thread_Create(&p->thread, startAddress, obj)); + p->exit = True; /* p->exit is unused before (canStart) Event. + But in case of some unexpected code failure we will get fast exit from thread */ + + // return ERROR_TOO_MANY_POSTS; // for debug + // return EINVAL; // for debug + + if (p->affinity != 0) + wres = Thread_Create_With_Affinity(&p->thread, startAddress, obj, (CAffinityMask)p->affinity); + else + wres = Thread_Create(&p->thread, startAddress, obj); + + RINOK_THREAD(wres) p->wasCreated = True; return SZ_OK; } -static SRes MtSync_Create(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *obj, UInt32 numBlocks) + +Z7_NO_INLINE +static SRes MtSync_Create(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *obj) { - SRes res = MtSync_Create2(p, startAddress, obj, numBlocks); - if (res != SZ_OK) - MtSync_Destruct(p); - return res; + const WRes wres = MtSync_Create_WRes(p, startAddress, obj); + if (wres == 0) + return 0; + MtSync_Destruct(p); + return MY_SRes_HRESULT_FROM_WRes(wres); } -void MtSync_Init(CMtSync *p) { p->needStart = True; } + +// ---------- HASH THREAD ---------- #define kMtMaxValForNormalize 0xFFFFFFFF +// #define kMtMaxValForNormalize ((1 << 21)) // for debug +// #define kNormalizeAlign (1 << 7) // alignment for speculated accesses -#define DEF_GetHeads2(name, v, action) \ - static void GetHeads ## name(const Byte *p, UInt32 pos, \ - UInt32 *hash, UInt32 hashMask, UInt32 *heads, UInt32 numHeads, const UInt32 *crc) \ - { action; for (; numHeads != 0; numHeads--) { \ - const UInt32 value = (v); p++; *heads++ = pos - hash[value]; hash[value] = pos++; } } +#ifdef MY_CPU_LE_UNALIGN + #define GetUi24hi_from32(p) ((UInt32)GetUi32(p) >> 8) +#else + #define GetUi24hi_from32(p) ((p)[1] ^ ((UInt32)(p)[2] << 8) ^ ((UInt32)(p)[3] << 16)) +#endif + +#define GetHeads_DECL(name) \ + static void GetHeads ## name(const Byte *p, UInt32 pos, \ + UInt32 *hash, UInt32 hashMask, UInt32 *heads, UInt32 numHeads, const UInt32 *crc) + +#define GetHeads_LOOP(v) \ + for (; numHeads != 0; numHeads--) { \ + const UInt32 value = (v); \ + p++; \ + *heads++ = pos - hash[value]; \ + hash[value] = pos++; } +#define DEF_GetHeads2(name, v, action) \ + GetHeads_DECL(name) { action \ + GetHeads_LOOP(v) } + #define DEF_GetHeads(name, v) DEF_GetHeads2(name, v, ;) -DEF_GetHeads2(2, (p[0] | ((UInt32)p[1] << 8)), UNUSED_VAR(hashMask); UNUSED_VAR(crc); ) -DEF_GetHeads(3, (crc[p[0]] ^ p[1] ^ ((UInt32)p[2] << 8)) & hashMask) -DEF_GetHeads(4, (crc[p[0]] ^ p[1] ^ ((UInt32)p[2] << 8) ^ (crc[p[3]] << 5)) & hashMask) -DEF_GetHeads(4b, (crc[p[0]] ^ p[1] ^ ((UInt32)p[2] << 8) ^ ((UInt32)p[3] << 16)) & hashMask) -/* DEF_GetHeads(5, (crc[p[0]] ^ p[1] ^ ((UInt32)p[2] << 8) ^ (crc[p[3]] << 5) ^ (crc[p[4]] << 3)) & hashMask) */ +DEF_GetHeads2(2, GetUi16(p), UNUSED_VAR(hashMask); UNUSED_VAR(crc); ) +DEF_GetHeads(3, (crc[p[0]] ^ GetUi16(p + 1)) & hashMask) +DEF_GetHeads2(3b, GetUi16(p) ^ ((UInt32)(p)[2] << 16), UNUSED_VAR(hashMask); UNUSED_VAR(crc); ) +// BT3 is not good for crc collisions for big hashMask values. + +/* +GetHeads_DECL(3b) +{ + UNUSED_VAR(hashMask); + UNUSED_VAR(crc); + { + const Byte *pLim = p + numHeads; + if (numHeads == 0) + return; + pLim--; + while (p < pLim) + { + UInt32 v1 = GetUi32(p); + UInt32 v0 = v1 & 0xFFFFFF; + UInt32 h0, h1; + p += 2; + v1 >>= 8; + h0 = hash[v0]; hash[v0] = pos; heads[0] = pos - h0; pos++; + h1 = hash[v1]; hash[v1] = pos; heads[1] = pos - h1; pos++; + heads += 2; + } + if (p == pLim) + { + UInt32 v0 = GetUi16(p) ^ ((UInt32)(p)[2] << 16); + *heads = pos - hash[v0]; + hash[v0] = pos; + } + } +} +*/ + +/* +GetHeads_DECL(4) +{ + unsigned sh = 0; + UNUSED_VAR(crc) + while ((hashMask & 0x80000000) == 0) + { + hashMask <<= 1; + sh++; + } + GetHeads_LOOP((GetUi32(p) * 0xa54a1) >> sh) +} +#define GetHeads4b GetHeads4 +*/ + +#define USE_GetHeads_LOCAL_CRC + +#ifdef USE_GetHeads_LOCAL_CRC + +GetHeads_DECL(4) +{ + UInt32 crc0[256]; + UInt32 crc1[256]; + { + unsigned i; + for (i = 0; i < 256; i++) + { + UInt32 v = crc[i]; + crc0[i] = v & hashMask; + crc1[i] = (v << kLzHash_CrcShift_1) & hashMask; + // crc1[i] = rotlFixed(v, 8) & hashMask; + } + } + GetHeads_LOOP(crc0[p[0]] ^ crc1[p[3]] ^ (UInt32)GetUi16(p+1)) +} + +GetHeads_DECL(4b) +{ + UInt32 crc0[256]; + { + unsigned i; + for (i = 0; i < 256; i++) + crc0[i] = crc[i] & hashMask; + } + GetHeads_LOOP(crc0[p[0]] ^ GetUi24hi_from32(p)) +} + +GetHeads_DECL(5) +{ + UInt32 crc0[256]; + UInt32 crc1[256]; + UInt32 crc2[256]; + { + unsigned i; + for (i = 0; i < 256; i++) + { + UInt32 v = crc[i]; + crc0[i] = v & hashMask; + crc1[i] = (v << kLzHash_CrcShift_1) & hashMask; + crc2[i] = (v << kLzHash_CrcShift_2) & hashMask; + } + } + GetHeads_LOOP(crc0[p[0]] ^ crc1[p[3]] ^ crc2[p[4]] ^ (UInt32)GetUi16(p+1)) +} + +GetHeads_DECL(5b) +{ + UInt32 crc0[256]; + UInt32 crc1[256]; + { + unsigned i; + for (i = 0; i < 256; i++) + { + UInt32 v = crc[i]; + crc0[i] = v & hashMask; + crc1[i] = (v << kLzHash_CrcShift_1) & hashMask; + } + } + GetHeads_LOOP(crc0[p[0]] ^ crc1[p[4]] ^ GetUi24hi_from32(p)) +} + +#else + +DEF_GetHeads(4, (crc[p[0]] ^ (crc[p[3]] << kLzHash_CrcShift_1) ^ (UInt32)GetUi16(p+1)) & hashMask) +DEF_GetHeads(4b, (crc[p[0]] ^ GetUi24hi_from32(p)) & hashMask) +DEF_GetHeads(5, (crc[p[0]] ^ (crc[p[3]] << kLzHash_CrcShift_1) ^ (crc[p[4]] << kLzHash_CrcShift_2) ^ (UInt32)GetUi16(p + 1)) & hashMask) +DEF_GetHeads(5b, (crc[p[0]] ^ (crc[p[4]] << kLzHash_CrcShift_1) ^ GetUi24hi_from32(p)) & hashMask) + +#endif + static void HashThreadFunc(CMatchFinderMt *mt) { CMtSync *p = &mt->hashSync; + PRF(printf("\nHashThreadFunc\n")); + for (;;) { - UInt32 numProcessedBlocks = 0; + UInt32 blockIndex = 0; + PRF(printf("\nHashThreadFunc : Event_Wait(&p->canStart)\n")); Event_Wait(&p->canStart); - Event_Set(&p->wasStarted); + PRF(printf("\nHashThreadFunc : Event_Wait(&p->canStart) : after \n")); + if (p->exit) + { + PRF(printf("\nHashThreadFunc : exit \n")); + return; + } + + MatchFinder_Init_HighHash(MF(mt)); + for (;;) { - if (p->exit) - return; - if (p->stopWriting) - { - p->numProcessedBlocks = numProcessedBlocks; - Event_Set(&p->wasStopped); - break; - } + PRF(printf("Hash thread block = %d pos = %d\n", (unsigned)blockIndex, mt->MatchFinder->pos)); { - CMatchFinder *mf = mt->MatchFinder; + CMatchFinder *mf = MF(mt); if (MatchFinder_NeedMove(mf)) { CriticalSection_Enter(&mt->btSync.cs); @@ -180,163 +474,178 @@ static void HashThreadFunc(CMatchFinderMt *mt) mt->pointerToCurPos -= offset; mt->buffer -= offset; } - CriticalSection_Leave(&mt->btSync.cs); CriticalSection_Leave(&mt->hashSync.cs); + CriticalSection_Leave(&mt->btSync.cs); continue; } Semaphore_Wait(&p->freeSemaphore); + if (p->exit) // exit is unexpected here. But we check it here for some failure case + return; + + // for faster stop : we check (p->stopWriting) after Wait(freeSemaphore) + if (p->stopWriting) + break; + MatchFinder_ReadIfRequired(mf); - if (mf->pos > (kMtMaxValForNormalize - kMtHashBlockSize)) - { - UInt32 subValue = (mf->pos - mf->historySize - 1); - MatchFinder_ReduceOffsets(mf, subValue); - MatchFinder_Normalize3(subValue, mf->hash + mf->fixedHashSize, (size_t)mf->hashMask + 1); - } { - UInt32 *heads = mt->hashBuf + ((numProcessedBlocks++) & kMtHashNumBlocksMask) * kMtHashBlockSize; - UInt32 num = mf->streamPos - mf->pos; + UInt32 *heads = mt->hashBuf + GET_HASH_BLOCK_OFFSET(blockIndex++); + UInt32 num = Inline_MatchFinder_GetNumAvailableBytes(mf); heads[0] = 2; heads[1] = num; + + /* heads[1] contains the number of avail bytes: + if (avail < mf->numHashBytes) : + { + it means that stream was finished + HASH_THREAD and BT_TREAD must move position for heads[1] (avail) bytes. + HASH_THREAD doesn't stop, + HASH_THREAD fills only the header (2 numbers) for all next blocks: + {2, NumHashBytes - 1}, {2,0}, {2,0}, ... , {2,0} + } + else + { + HASH_THREAD and BT_TREAD must move position for (heads[0] - 2) bytes; + } + */ + if (num >= mf->numHashBytes) { num = num - mf->numHashBytes + 1; if (num > kMtHashBlockSize - 2) num = kMtHashBlockSize - 2; + + if (mf->pos > (UInt32)kMtMaxValForNormalize - num) + { + const UInt32 subValue = (mf->pos - mf->historySize - 1); // & ~(UInt32)(kNormalizeAlign - 1); + MatchFinder_REDUCE_OFFSETS(mf, subValue) + MatchFinder_Normalize3(subValue, mf->hash + mf->fixedHashSize, (size_t)mf->hashMask + 1); + } + + heads[0] = 2 + num; mt->GetHeadsFunc(mf->buffer, mf->pos, mf->hash + mf->fixedHashSize, mf->hashMask, heads + 2, num, mf->crc); - heads[0] += num; } - mf->pos += num; + + mf->pos += num; // wrap over zero is allowed at the end of stream mf->buffer += num; } } Semaphore_Release1(&p->filledSemaphore); - } - } -} + } // for() processing end -static void MatchFinderMt_GetNextBlock_Hash(CMatchFinderMt *p) -{ - MtSync_GetNextBlock(&p->hashSync); - p->hashBufPosLimit = p->hashBufPos = ((p->hashSync.numProcessedBlocks - 1) & kMtHashNumBlocksMask) * kMtHashBlockSize; - p->hashBufPosLimit += p->hashBuf[p->hashBufPos++]; - p->hashNumAvail = p->hashBuf[p->hashBufPos++]; + // p->numBlocks_Sent = blockIndex; + Event_Set(&p->wasStopped); + } // for() thread end } -#define kEmptyHashValue 0 -/* #define MFMT_GM_INLINE */ + + +// ---------- BT THREAD ---------- + +/* we use one variable instead of two (cyclicBufferPos == pos) before CyclicBuf wrap. + here we define fixed offset of (p->pos) from (p->cyclicBufferPos) */ +#define CYC_TO_POS_OFFSET 0 +// #define CYC_TO_POS_OFFSET 1 // for debug + +#define MFMT_GM_INLINE #ifdef MFMT_GM_INLINE -#define NO_INLINE MY_FAST_CALL +/* + we use size_t for (pos) instead of UInt32 + to eliminate "movsx" BUG in old MSVC x64 compiler. +*/ -static Int32 NO_INLINE GetMatchesSpecN(UInt32 lenLimit, UInt32 pos, const Byte *cur, CLzRef *son, - UInt32 _cyclicBufferPos, UInt32 _cyclicBufferSize, UInt32 _cutValue, - UInt32 *_distances, UInt32 _maxLen, const UInt32 *hash, Int32 limit, UInt32 size, UInt32 *posRes) -{ - do - { - UInt32 *distances = _distances + 1; - UInt32 curMatch = pos - *hash++; - CLzRef *ptr0 = son + (_cyclicBufferPos << 1) + 1; - CLzRef *ptr1 = son + (_cyclicBufferPos << 1); - UInt32 len0 = 0, len1 = 0; - UInt32 cutValue = _cutValue; - UInt32 maxLen = _maxLen; - for (;;) - { - UInt32 delta = pos - curMatch; - if (cutValue-- == 0 || delta >= _cyclicBufferSize) - { - *ptr0 = *ptr1 = kEmptyHashValue; - break; - } - { - CLzRef *pair = son + ((_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); - const Byte *pb = cur - delta; - UInt32 len = (len0 < len1 ? len0 : len1); - if (pb[len] == cur[len]) - { - if (++len != lenLimit && pb[len] == cur[len]) - while (++len != lenLimit) - if (pb[len] != cur[len]) - break; - if (maxLen < len) - { - *distances++ = maxLen = len; - *distances++ = delta - 1; - if (len == lenLimit) - { - *ptr1 = pair[0]; - *ptr0 = pair[1]; - break; - } - } - } - if (pb[len] < cur[len]) - { - *ptr1 = curMatch; - ptr1 = pair + 1; - curMatch = *ptr1; - len1 = len; - } - else - { - *ptr0 = curMatch; - ptr0 = pair; - curMatch = *ptr0; - len0 = len; - } - } - } - pos++; - _cyclicBufferPos++; - cur++; - { - UInt32 num = (UInt32)(distances - _distances); - *_distances = num - 1; - _distances += num; - limit -= num; - } - } - while (limit > 0 && --size != 0); - *posRes = pos; - return limit; -} +UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son, + UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, + UInt32 *posRes); #endif -static void BtGetMatches(CMatchFinderMt *p, UInt32 *distances) + +static void BtGetMatches(CMatchFinderMt *p, UInt32 *d) { UInt32 numProcessed = 0; UInt32 curPos = 2; - UInt32 limit = kMtBtBlockSize - (p->matchMaxLen * 2); - distances[1] = p->hashNumAvail; + /* GetMatchesSpec() functions don't create (len = 1) + in [len, dist] match pairs, if (p->numHashBytes >= 2) + Also we suppose here that (matchMaxLen >= 2). + So the following code for (reserve) is not required + UInt32 reserve = (p->matchMaxLen * 2); + const UInt32 kNumHashBytes_Max = 5; // BT_HASH_BYTES_MAX + if (reserve < kNumHashBytes_Max - 1) + reserve = kNumHashBytes_Max - 1; + const UInt32 limit = kMtBtBlockSize - (reserve); + */ + + const UInt32 limit = kMtBtBlockSize - (p->matchMaxLen * 2); + + d[1] = p->hashNumAvail; + + if (p->failure_BT) + { + // printf("\n == 1 BtGetMatches() p->failure_BT\n"); + d[0] = 0; + // d[1] = 0; + return; + } while (curPos < limit) { if (p->hashBufPos == p->hashBufPosLimit) { - MatchFinderMt_GetNextBlock_Hash(p); - distances[1] = numProcessed + p->hashNumAvail; - if (p->hashNumAvail >= p->numHashBytes) + // MatchFinderMt_GetNextBlock_Hash(p); + UInt32 avail; + { + const UInt32 bi = MtSync_GetNextBlock(&p->hashSync); + const UInt32 k = GET_HASH_BLOCK_OFFSET(bi); + const UInt32 *h = p->hashBuf + k; + avail = h[1]; + p->hashBufPosLimit = k + h[0]; + p->hashNumAvail = avail; + p->hashBufPos = k + 2; + } + + { + /* we must prevent UInt32 overflow for avail total value, + if avail was increased with new hash block */ + UInt32 availSum = numProcessed + avail; + if (availSum < numProcessed) + availSum = (UInt32)(Int32)-1; + d[1] = availSum; + } + + if (avail >= p->numHashBytes) continue; - distances[0] = curPos + p->hashNumAvail; - distances += curPos; - for (; p->hashNumAvail != 0; p->hashNumAvail--) - *distances++ = 0; + + // if (p->hashBufPos != p->hashBufPosLimit) exit(1); + + /* (avail < p->numHashBytes) + It means that stream was finished. + And (avail) - is a number of remaining bytes, + we fill (d) for (avail) bytes for LZ_THREAD (receiver). + but we don't update (p->pos) and (p->cyclicBufferPos) here in BT_THREAD */ + + /* here we suppose that we have space enough: + (kMtBtBlockSize - curPos >= p->hashNumAvail) */ + p->hashNumAvail = 0; + d[0] = curPos + avail; + d += curPos; + for (; avail != 0; avail--) + *d++ = 0; return; } { UInt32 size = p->hashBufPosLimit - p->hashBufPos; - UInt32 lenLimit = p->matchMaxLen; UInt32 pos = p->pos; UInt32 cyclicBufferPos = p->cyclicBufferPos; + UInt32 lenLimit = p->matchMaxLen; if (lenLimit >= p->hashNumAvail) lenLimit = p->hashNumAvail; { @@ -348,10 +657,18 @@ static void BtGetMatches(CMatchFinderMt *p, UInt32 *distances) size = size2; } + if (pos > (UInt32)kMtMaxValForNormalize - size) + { + const UInt32 subValue = (pos - p->cyclicBufferSize); // & ~(UInt32)(kNormalizeAlign - 1); + pos -= subValue; + p->pos = pos; + MatchFinder_Normalize3(subValue, p->son, (size_t)p->cyclicBufferSize * 2); + } + #ifndef MFMT_GM_INLINE while (curPos < limit && size-- != 0) { - UInt32 *startDistances = distances + curPos; + UInt32 *startDistances = d + curPos; UInt32 num = (UInt32)(GetMatchesSpec1(lenLimit, pos - p->hashBuf[p->hashBufPos++], pos, p->buffer, p->son, cyclicBufferPos, p->cyclicBufferSize, p->cutValue, startDistances + 1, p->numHashBytes - 1) - startDistances); @@ -363,79 +680,112 @@ static void BtGetMatches(CMatchFinderMt *p, UInt32 *distances) } #else { - UInt32 posRes; - curPos = limit - GetMatchesSpecN(lenLimit, pos, p->buffer, p->son, cyclicBufferPos, p->cyclicBufferSize, p->cutValue, - distances + curPos, p->numHashBytes - 1, p->hashBuf + p->hashBufPos, (Int32)(limit - curPos), size, &posRes); - p->hashBufPos += posRes - pos; - cyclicBufferPos += posRes - pos; - p->buffer += posRes - pos; - pos = posRes; + UInt32 posRes = pos; + const UInt32 *d_end; + { + d_end = GetMatchesSpecN_2( + p->buffer + lenLimit - 1, + pos, p->buffer, p->son, p->cutValue, d + curPos, + p->numHashBytes - 1, p->hashBuf + p->hashBufPos, + d + limit, p->hashBuf + p->hashBufPos + size, + cyclicBufferPos, p->cyclicBufferSize, + &posRes); + } + { + if (!d_end) + { + // printf("\n == 2 BtGetMatches() p->failure_BT\n"); + // internal data failure + p->failure_BT = True; + d[0] = 0; + // d[1] = 0; + return; + } + } + curPos = (UInt32)(d_end - d); + { + const UInt32 processed = posRes - pos; + pos = posRes; + p->hashBufPos += processed; + cyclicBufferPos += processed; + p->buffer += processed; + } } #endif - numProcessed += pos - p->pos; - p->hashNumAvail -= pos - p->pos; - p->pos = pos; + { + const UInt32 processed = pos - p->pos; + numProcessed += processed; + p->hashNumAvail -= processed; + p->pos = pos; + } if (cyclicBufferPos == p->cyclicBufferSize) cyclicBufferPos = 0; p->cyclicBufferPos = cyclicBufferPos; } } - distances[0] = curPos; + d[0] = curPos; } + static void BtFillBlock(CMatchFinderMt *p, UInt32 globalBlockIndex) { CMtSync *sync = &p->hashSync; + + BUFFER_MUST_BE_UNLOCKED(sync) + if (!sync->needStart) { - CriticalSection_Enter(&sync->cs); - sync->csWasEntered = True; + LOCK_BUFFER(sync) } - BtGetMatches(p, p->btBuf + (globalBlockIndex & kMtBtNumBlocksMask) * kMtBtBlockSize); - - if (p->pos > kMtMaxValForNormalize - kMtBtBlockSize) - { - UInt32 subValue = p->pos - p->cyclicBufferSize; - MatchFinder_Normalize3(subValue, p->son, (size_t)p->cyclicBufferSize * 2); - p->pos -= subValue; - } + BtGetMatches(p, p->btBuf + GET_BT_BLOCK_OFFSET(globalBlockIndex)); + + /* We suppose that we have called GetNextBlock() from start. + So buffer is LOCKED */ - if (!sync->needStart) - { - CriticalSection_Leave(&sync->cs); - sync->csWasEntered = False; - } + UNLOCK_BUFFER(sync) } -void BtThreadFunc(CMatchFinderMt *mt) + +Z7_NO_INLINE +static void BtThreadFunc(CMatchFinderMt *mt) { CMtSync *p = &mt->btSync; for (;;) { UInt32 blockIndex = 0; Event_Wait(&p->canStart); - Event_Set(&p->wasStarted); + for (;;) { + PRF(printf(" BT thread block = %d pos = %d\n", (unsigned)blockIndex, mt->pos)); + /* (p->exit == true) is possible after (p->canStart) at first loop iteration + and is unexpected after more Wait(freeSemaphore) iterations */ if (p->exit) return; + + Semaphore_Wait(&p->freeSemaphore); + + // for faster stop : we check (p->stopWriting) after Wait(freeSemaphore) if (p->stopWriting) - { - p->numProcessedBlocks = blockIndex; - MtSync_StopWriting(&mt->hashSync); - Event_Set(&p->wasStopped); break; - } - Semaphore_Wait(&p->freeSemaphore); + BtFillBlock(mt, blockIndex++); + Semaphore_Release1(&p->filledSemaphore); } + + // we stop HASH_THREAD here + MtSync_StopWriting(&mt->hashSync); + + // p->numBlocks_Sent = blockIndex; + Event_Set(&p->wasStopped); } } + void MatchFinderMt_Construct(CMatchFinderMt *p) { p->hashBuf = NULL; @@ -443,24 +793,47 @@ void MatchFinderMt_Construct(CMatchFinderMt *p) MtSync_Construct(&p->btSync); } -static void MatchFinderMt_FreeMem(CMatchFinderMt *p, ISzAlloc *alloc) +static void MatchFinderMt_FreeMem(CMatchFinderMt *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->hashBuf); + ISzAlloc_Free(alloc, p->hashBuf); p->hashBuf = NULL; } -void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAlloc *alloc) +void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAllocPtr alloc) { - MtSync_Destruct(&p->hashSync); + /* + HASH_THREAD can use CriticalSection(s) btSync.cs and hashSync.cs. + So we must be sure that HASH_THREAD will not use CriticalSection(s) + after deleting CriticalSection here. + + we call ReleaseStream(p) + that calls StopWriting(btSync) + that calls StopWriting(hashSync), if it's required to stop HASH_THREAD. + after StopWriting() it's safe to destruct MtSync(s) in any order */ + + MatchFinderMt_ReleaseStream(p); + MtSync_Destruct(&p->btSync); + MtSync_Destruct(&p->hashSync); + + LOG_ITER( + printf("\nTree %9d * %7d iter = %9d = sum : bytes = %9d\n", + (UInt32)(g_NumIters_Tree / 1000), + (UInt32)(((UInt64)g_NumIters_Loop * 1000) / (g_NumIters_Tree + 1)), + (UInt32)(g_NumIters_Loop / 1000), + (UInt32)(g_NumIters_Bytes / 1000) + )); + MatchFinderMt_FreeMem(p, alloc); } + #define kHashBufferSize (kMtHashBlockSize * kMtHashNumBlocks) #define kBtBufferSize (kMtBtBlockSize * kMtBtNumBlocks) -static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE HashThreadFunc2(void *p) { HashThreadFunc((CMatchFinderMt *)p); return 0; } -static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE BtThreadFunc2(void *p) + +static THREAD_FUNC_DECL HashThreadFunc2(void *p) { HashThreadFunc((CMatchFinderMt *)p); return 0; } +static THREAD_FUNC_DECL BtThreadFunc2(void *p) { Byte allocaDummy[0x180]; unsigned i = 0; @@ -471,16 +844,17 @@ static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE BtThreadFunc2(void *p) return 0; } + SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddBufferBefore, - UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAlloc *alloc) + UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAllocPtr alloc) { - CMatchFinder *mf = p->MatchFinder; + CMatchFinder *mf = MF(p); p->historySize = historySize; if (kMtBtBlockSize <= matchMaxLen * 4) return SZ_ERROR_PARAM; if (!p->hashBuf) { - p->hashBuf = (UInt32 *)alloc->Alloc(alloc, (kHashBufferSize + kBtBufferSize) * sizeof(UInt32)); + p->hashBuf = (UInt32 *)ISzAlloc_Alloc(alloc, ((size_t)kHashBufferSize + (size_t)kBtBufferSize) * sizeof(UInt32)); if (!p->hashBuf) return SZ_ERROR_MEM; p->btBuf = p->hashBuf + kHashBufferSize; @@ -490,241 +864,456 @@ SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddB if (!MatchFinder_Create(mf, historySize, keepAddBufferBefore, matchMaxLen, keepAddBufferAfter, alloc)) return SZ_ERROR_MEM; - RINOK(MtSync_Create(&p->hashSync, HashThreadFunc2, p, kMtHashNumBlocks)); - RINOK(MtSync_Create(&p->btSync, BtThreadFunc2, p, kMtBtNumBlocks)); + RINOK(MtSync_Create(&p->hashSync, HashThreadFunc2, p)) + RINOK(MtSync_Create(&p->btSync, BtThreadFunc2, p)) return SZ_OK; } -/* Call it after ReleaseStream / SetStream */ -void MatchFinderMt_Init(CMatchFinderMt *p) + +SRes MatchFinderMt_InitMt(CMatchFinderMt *p) +{ + RINOK(MtSync_Init(&p->hashSync, kMtHashNumBlocks)) + return MtSync_Init(&p->btSync, kMtBtNumBlocks); +} + + +static void MatchFinderMt_Init(CMatchFinderMt *p) { - CMatchFinder *mf = p->MatchFinder; - p->btBufPos = p->btBufPosLimit = 0; - p->hashBufPos = p->hashBufPosLimit = 0; + CMatchFinder *mf = MF(p); + + p->btBufPos = + p->btBufPosLimit = NULL; + p->hashBufPos = + p->hashBufPosLimit = 0; + p->hashNumAvail = 0; // 21.03 + + p->failure_BT = False; /* Init without data reading. We don't want to read data in this thread */ - MatchFinder_Init_2(mf, False); + MatchFinder_Init_4(mf); + + MatchFinder_Init_LowHash(mf); p->pointerToCurPos = Inline_MatchFinder_GetPointerToCurrentPos(mf); p->btNumAvailBytes = 0; - p->lzPos = p->historySize + 1; + p->failure_LZ_BT = False; + // p->failure_LZ_LZ = False; + + p->lzPos = + 1; // optimal smallest value + // 0; // for debug: ignores match to start + // kNormalizeAlign; // for debug p->hash = mf->hash; p->fixedHashSize = mf->fixedHashSize; + // p->hash4Mask = mf->hash4Mask; p->crc = mf->crc; + // memcpy(p->crc, mf->crc, sizeof(mf->crc)); p->son = mf->son; p->matchMaxLen = mf->matchMaxLen; p->numHashBytes = mf->numHashBytes; - p->pos = mf->pos; - p->buffer = mf->buffer; - p->cyclicBufferPos = mf->cyclicBufferPos; + + /* (mf->pos) and (mf->streamPos) were already initialized to 1 in MatchFinder_Init_4() */ + // mf->streamPos = mf->pos = 1; // optimal smallest value + // 0; // for debug: ignores match to start + // kNormalizeAlign; // for debug + + /* we must init (p->pos = mf->pos) for BT, because + BT code needs (p->pos == delta_value_for_empty_hash_record == mf->pos) */ + p->pos = mf->pos; // do not change it + + p->cyclicBufferPos = (p->pos - CYC_TO_POS_OFFSET); p->cyclicBufferSize = mf->cyclicBufferSize; + p->buffer = mf->buffer; p->cutValue = mf->cutValue; + // p->son[0] = p->son[1] = 0; // unused: to init skipped record for speculated accesses. } + /* ReleaseStream is required to finish multithreading */ void MatchFinderMt_ReleaseStream(CMatchFinderMt *p) { + // Sleep(1); // for debug MtSync_StopWriting(&p->btSync); + // Sleep(200); // for debug /* p->MatchFinder->ReleaseStream(); */ } -static void MatchFinderMt_Normalize(CMatchFinderMt *p) -{ - MatchFinder_Normalize3(p->lzPos - p->historySize - 1, p->hash, p->fixedHashSize); - p->lzPos = p->historySize + 1; -} -static void MatchFinderMt_GetNextBlock_Bt(CMatchFinderMt *p) +Z7_NO_INLINE +static UInt32 MatchFinderMt_GetNextBlock_Bt(CMatchFinderMt *p) { - UInt32 blockIndex; - MtSync_GetNextBlock(&p->btSync); - blockIndex = ((p->btSync.numProcessedBlocks - 1) & kMtBtNumBlocksMask); - p->btBufPosLimit = p->btBufPos = blockIndex * kMtBtBlockSize; - p->btBufPosLimit += p->btBuf[p->btBufPos++]; - p->btNumAvailBytes = p->btBuf[p->btBufPos++]; - if (p->lzPos >= kMtMaxValForNormalize - kMtBtBlockSize) - MatchFinderMt_Normalize(p); + if (p->failure_LZ_BT) + p->btBufPos = p->failureBuf; + else + { + const UInt32 bi = MtSync_GetNextBlock(&p->btSync); + const UInt32 *bt = p->btBuf + GET_BT_BLOCK_OFFSET(bi); + { + const UInt32 numItems = bt[0]; + p->btBufPosLimit = bt + numItems; + p->btNumAvailBytes = bt[1]; + p->btBufPos = bt + 2; + if (numItems < 2 || numItems > kMtBtBlockSize) + { + p->failureBuf[0] = 0; + p->btBufPos = p->failureBuf; + p->btBufPosLimit = p->failureBuf + 1; + p->failure_LZ_BT = True; + // p->btNumAvailBytes = 0; + /* we don't want to decrease AvailBytes, that was load before. + that can be unxepected for the code that have loaded anopther value before */ + } + } + + if (p->lzPos >= (UInt32)kMtMaxValForNormalize - (UInt32)kMtBtBlockSize) + { + /* we don't check (lzPos) over exact avail bytes in (btBuf). + (fixedHashSize) is small, so normalization is fast */ + const UInt32 subValue = (p->lzPos - p->historySize - 1); // & ~(UInt32)(kNormalizeAlign - 1); + p->lzPos -= subValue; + MatchFinder_Normalize3(subValue, p->hash, p->fixedHashSize); + } + } + return p->btNumAvailBytes; } + + static const Byte * MatchFinderMt_GetPointerToCurrentPos(CMatchFinderMt *p) { return p->pointerToCurPos; } + #define GET_NEXT_BLOCK_IF_REQUIRED if (p->btBufPos == p->btBufPosLimit) MatchFinderMt_GetNextBlock_Bt(p); + static UInt32 MatchFinderMt_GetNumAvailableBytes(CMatchFinderMt *p) { - GET_NEXT_BLOCK_IF_REQUIRED; - return p->btNumAvailBytes; + if (p->btBufPos != p->btBufPosLimit) + return p->btNumAvailBytes; + return MatchFinderMt_GetNextBlock_Bt(p); } -static UInt32 * MixMatches2(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *distances) + +// #define CHECK_FAILURE_LZ(_match_, _pos_) if (_match_ >= _pos_) { p->failure_LZ_LZ = True; return d; } +#define CHECK_FAILURE_LZ(_match_, _pos_) + +static UInt32 * MixMatches2(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d) { - UInt32 h2, curMatch2; + UInt32 h2, c2; UInt32 *hash = p->hash; const Byte *cur = p->pointerToCurPos; - UInt32 lzPos = p->lzPos; + const UInt32 m = p->lzPos; MT_HASH2_CALC - curMatch2 = hash[h2]; - hash[h2] = lzPos; + c2 = hash[h2]; + hash[h2] = m; - if (curMatch2 >= matchMinPos) - if (cur[(ptrdiff_t)curMatch2 - lzPos] == cur[0]) + if (c2 >= matchMinPos) + { + CHECK_FAILURE_LZ(c2, m) + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m] == cur[0]) { - *distances++ = 2; - *distances++ = lzPos - curMatch2 - 1; + *d++ = 2; + *d++ = m - c2 - 1; } + } - return distances; + return d; } -static UInt32 * MixMatches3(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *distances) +static UInt32 * MixMatches3(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d) { - UInt32 h2, h3, curMatch2, curMatch3; + UInt32 h2, h3, c2, c3; UInt32 *hash = p->hash; const Byte *cur = p->pointerToCurPos; - UInt32 lzPos = p->lzPos; + const UInt32 m = p->lzPos; MT_HASH3_CALC - curMatch2 = hash[ h2]; - curMatch3 = hash[kFix3HashSize + h3]; + c2 = hash[h2]; + c3 = (hash + kFix3HashSize)[h3]; - hash[ h2] = lzPos; - hash[kFix3HashSize + h3] = lzPos; + hash[h2] = m; + (hash + kFix3HashSize)[h3] = m; - if (curMatch2 >= matchMinPos && cur[(ptrdiff_t)curMatch2 - lzPos] == cur[0]) + if (c2 >= matchMinPos) { - distances[1] = lzPos - curMatch2 - 1; - if (cur[(ptrdiff_t)curMatch2 - lzPos + 2] == cur[2]) + CHECK_FAILURE_LZ(c2, m) + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m] == cur[0]) { - distances[0] = 3; - return distances + 2; + d[1] = m - c2 - 1; + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m + 2] == cur[2]) + { + d[0] = 3; + return d + 2; + } + d[0] = 2; + d += 2; } - distances[0] = 2; - distances += 2; } - if (curMatch3 >= matchMinPos && cur[(ptrdiff_t)curMatch3 - lzPos] == cur[0]) + if (c3 >= matchMinPos) { - *distances++ = 3; - *distances++ = lzPos - curMatch3 - 1; + CHECK_FAILURE_LZ(c3, m) + if (cur[(ptrdiff_t)c3 - (ptrdiff_t)m] == cur[0]) + { + *d++ = 3; + *d++ = m - c3 - 1; + } } - return distances; + return d; } + +#define INCREASE_LZ_POS p->lzPos++; p->pointerToCurPos++; + /* -static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *distances) +static +UInt32* MatchFinderMt_GetMatches_Bt4(CMatchFinderMt *p, UInt32 *d) { - UInt32 h2, h3, h4, curMatch2, curMatch3, curMatch4; + const UInt32 *bt = p->btBufPos; + const UInt32 len = *bt++; + const UInt32 *btLim = bt + len; + UInt32 matchMinPos; + UInt32 avail = p->btNumAvailBytes - 1; + p->btBufPos = btLim; + + { + p->btNumAvailBytes = avail; + + #define BT_HASH_BYTES_MAX 5 + + matchMinPos = p->lzPos; + + if (len != 0) + matchMinPos -= bt[1]; + else if (avail < (BT_HASH_BYTES_MAX - 1) - 1) + { + INCREASE_LZ_POS + return d; + } + else + { + const UInt32 hs = p->historySize; + if (matchMinPos > hs) + matchMinPos -= hs; + else + matchMinPos = 1; + } + } + + for (;;) + { + + UInt32 h2, h3, c2, c3; UInt32 *hash = p->hash; const Byte *cur = p->pointerToCurPos; - UInt32 lzPos = p->lzPos; - MT_HASH4_CALC - - curMatch2 = hash[ h2]; - curMatch3 = hash[kFix3HashSize + h3]; - curMatch4 = hash[kFix4HashSize + h4]; + UInt32 m = p->lzPos; + MT_HASH3_CALC + + c2 = hash[h2]; + c3 = (hash + kFix3HashSize)[h3]; + + hash[h2] = m; + (hash + kFix3HashSize)[h3] = m; + + if (c2 >= matchMinPos && cur[(ptrdiff_t)c2 - (ptrdiff_t)m] == cur[0]) + { + d[1] = m - c2 - 1; + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m + 2] == cur[2]) + { + d[0] = 3; + d += 2; + break; + } + // else + { + d[0] = 2; + d += 2; + } + } + if (c3 >= matchMinPos && cur[(ptrdiff_t)c3 - (ptrdiff_t)m] == cur[0]) + { + *d++ = 3; + *d++ = m - c3 - 1; + } + break; + } + + if (len != 0) + { + do + { + const UInt32 v0 = bt[0]; + const UInt32 v1 = bt[1]; + bt += 2; + d[0] = v0; + d[1] = v1; + d += 2; + } + while (bt != btLim); + } + INCREASE_LZ_POS + return d; +} +*/ + + +static UInt32 * MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d) +{ + UInt32 h2, h3, /* h4, */ c2, c3 /* , c4 */; + UInt32 *hash = p->hash; + const Byte *cur = p->pointerToCurPos; + const UInt32 m = p->lzPos; + MT_HASH3_CALC + // MT_HASH4_CALC + c2 = hash[h2]; + c3 = (hash + kFix3HashSize)[h3]; + // c4 = (hash + kFix4HashSize)[h4]; - hash[ h2] = lzPos; - hash[kFix3HashSize + h3] = lzPos; - hash[kFix4HashSize + h4] = lzPos; + hash[h2] = m; + (hash + kFix3HashSize)[h3] = m; + // (hash + kFix4HashSize)[h4] = m; - if (curMatch2 >= matchMinPos && cur[(ptrdiff_t)curMatch2 - lzPos] == cur[0]) + // #define BT5_USE_H2 + // #ifdef BT5_USE_H2 + if (c2 >= matchMinPos && cur[(ptrdiff_t)c2 - (ptrdiff_t)m] == cur[0]) { - distances[1] = lzPos - curMatch2 - 1; - if (cur[(ptrdiff_t)curMatch2 - lzPos + 2] == cur[2]) + d[1] = m - c2 - 1; + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m + 2] == cur[2]) { - distances[0] = (cur[(ptrdiff_t)curMatch2 - lzPos + 3] == cur[3]) ? 4 : 3; - return distances + 2; + // d[0] = (cur[(ptrdiff_t)c2 - (ptrdiff_t)m + 3] == cur[3]) ? 4 : 3; + // return d + 2; + + if (cur[(ptrdiff_t)c2 - (ptrdiff_t)m + 3] == cur[3]) + { + d[0] = 4; + return d + 2; + } + d[0] = 3; + d += 2; + + #ifdef BT5_USE_H4 + if (c4 >= matchMinPos) + if ( + cur[(ptrdiff_t)c4 - (ptrdiff_t)m] == cur[0] && + cur[(ptrdiff_t)c4 - (ptrdiff_t)m + 3] == cur[3] + ) + { + *d++ = 4; + *d++ = m - c4 - 1; + } + #endif + return d; } - distances[0] = 2; - distances += 2; + d[0] = 2; + d += 2; } + // #endif - if (curMatch3 >= matchMinPos && cur[(ptrdiff_t)curMatch3 - lzPos] == cur[0]) + if (c3 >= matchMinPos && cur[(ptrdiff_t)c3 - (ptrdiff_t)m] == cur[0]) { - distances[1] = lzPos - curMatch3 - 1; - if (cur[(ptrdiff_t)curMatch3 - lzPos + 3] == cur[3]) + d[1] = m - c3 - 1; + if (cur[(ptrdiff_t)c3 - (ptrdiff_t)m + 3] == cur[3]) { - distances[0] = 4; - return distances + 2; + d[0] = 4; + return d + 2; } - distances[0] = 3; - distances += 2; + d[0] = 3; + d += 2; } - if (curMatch4 >= matchMinPos) + #ifdef BT5_USE_H4 + if (c4 >= matchMinPos) if ( - cur[(ptrdiff_t)curMatch4 - lzPos] == cur[0] && - cur[(ptrdiff_t)curMatch4 - lzPos + 3] == cur[3] + cur[(ptrdiff_t)c4 - (ptrdiff_t)m] == cur[0] && + cur[(ptrdiff_t)c4 - (ptrdiff_t)m + 3] == cur[3] ) { - *distances++ = 4; - *distances++ = lzPos - curMatch4 - 1; + *d++ = 4; + *d++ = m - c4 - 1; } + #endif - return distances; + return d; } -*/ -#define INCREASE_LZ_POS p->lzPos++; p->pointerToCurPos++; -static UInt32 MatchFinderMt2_GetMatches(CMatchFinderMt *p, UInt32 *distances) +static UInt32 * MatchFinderMt2_GetMatches(CMatchFinderMt *p, UInt32 *d) { - const UInt32 *btBuf = p->btBuf + p->btBufPos; - UInt32 len = *btBuf++; - p->btBufPos += 1 + len; + const UInt32 *bt = p->btBufPos; + const UInt32 len = *bt++; + const UInt32 *btLim = bt + len; + p->btBufPos = btLim; p->btNumAvailBytes--; + INCREASE_LZ_POS { - UInt32 i; - for (i = 0; i < len; i += 2) + while (bt != btLim) { - *distances++ = *btBuf++; - *distances++ = *btBuf++; + const UInt32 v0 = bt[0]; + const UInt32 v1 = bt[1]; + bt += 2; + d[0] = v0; + d[1] = v1; + d += 2; } } - INCREASE_LZ_POS - return len; + return d; } -static UInt32 MatchFinderMt_GetMatches(CMatchFinderMt *p, UInt32 *distances) -{ - const UInt32 *btBuf = p->btBuf + p->btBufPos; - UInt32 len = *btBuf++; - p->btBufPos += 1 + len; + +static UInt32 * MatchFinderMt_GetMatches(CMatchFinderMt *p, UInt32 *d) +{ + const UInt32 *bt = p->btBufPos; + UInt32 len = *bt++; + const UInt32 avail = p->btNumAvailBytes - 1; + p->btNumAvailBytes = avail; + p->btBufPos = bt + len; if (len == 0) { - /* change for bt5 ! */ - if (p->btNumAvailBytes-- >= 4) - len = (UInt32)(p->MixMatchesFunc(p, p->lzPos - p->historySize, distances) - (distances)); + #define BT_HASH_BYTES_MAX 5 + if (avail >= (BT_HASH_BYTES_MAX - 1) - 1) + { + UInt32 m = p->lzPos; + if (m > p->historySize) + m -= p->historySize; + else + m = 1; + d = p->MixMatchesFunc(p, m, d); + } } else { - /* Condition: there are matches in btBuf with length < p->numHashBytes */ - UInt32 *distances2; - p->btNumAvailBytes--; - distances2 = p->MixMatchesFunc(p, p->lzPos - btBuf[1], distances); + /* + first match pair from BinTree: (match_len, match_dist), + (match_len >= numHashBytes). + MixMatchesFunc() inserts only hash matches that are nearer than (match_dist) + */ + d = p->MixMatchesFunc(p, p->lzPos - bt[1], d); + // if (d) // check for failure do { - *distances2++ = *btBuf++; - *distances2++ = *btBuf++; + const UInt32 v0 = bt[0]; + const UInt32 v1 = bt[1]; + bt += 2; + d[0] = v0; + d[1] = v1; + d += 2; } - while ((len -= 2) != 0); - len = (UInt32)(distances2 - (distances)); + while (len -= 2); } INCREASE_LZ_POS - return len; + return d; } #define SKIP_HEADER2_MT do { GET_NEXT_BLOCK_IF_REQUIRED #define SKIP_HEADER_MT(n) SKIP_HEADER2_MT if (p->btNumAvailBytes-- >= (n)) { const Byte *cur = p->pointerToCurPos; UInt32 *hash = p->hash; -#define SKIP_FOOTER_MT } INCREASE_LZ_POS p->btBufPos += p->btBuf[p->btBufPos] + 1; } while (--num != 0); +#define SKIP_FOOTER_MT } INCREASE_LZ_POS p->btBufPos += (size_t)*p->btBufPos + 1; } while (--num != 0); static void MatchFinderMt0_Skip(CMatchFinderMt *p, UInt32 num) { @@ -746,58 +1335,72 @@ static void MatchFinderMt3_Skip(CMatchFinderMt *p, UInt32 num) SKIP_HEADER_MT(3) UInt32 h2, h3; MT_HASH3_CALC - hash[kFix3HashSize + h3] = + (hash + kFix3HashSize)[h3] = hash[ h2] = p->lzPos; SKIP_FOOTER_MT } /* +// MatchFinderMt4_Skip() is similar to MatchFinderMt3_Skip(). +// The difference is that MatchFinderMt3_Skip() updates hash for last 3 bytes of stream. + static void MatchFinderMt4_Skip(CMatchFinderMt *p, UInt32 num) { SKIP_HEADER_MT(4) - UInt32 h2, h3, h4; - MT_HASH4_CALC - hash[kFix4HashSize + h4] = - hash[kFix3HashSize + h3] = + UInt32 h2, h3; // h4 + MT_HASH3_CALC + // MT_HASH4_CALC + // (hash + kFix4HashSize)[h4] = + (hash + kFix3HashSize)[h3] = hash[ h2] = p->lzPos; SKIP_FOOTER_MT } */ -void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder *vTable) +void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder2 *vTable) { vTable->Init = (Mf_Init_Func)MatchFinderMt_Init; vTable->GetNumAvailableBytes = (Mf_GetNumAvailableBytes_Func)MatchFinderMt_GetNumAvailableBytes; vTable->GetPointerToCurrentPos = (Mf_GetPointerToCurrentPos_Func)MatchFinderMt_GetPointerToCurrentPos; vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt_GetMatches; - switch (p->MatchFinder->numHashBytes) + switch (MF(p)->numHashBytes) { case 2: p->GetHeadsFunc = GetHeads2; - p->MixMatchesFunc = (Mf_Mix_Matches)0; + p->MixMatchesFunc = (Mf_Mix_Matches)NULL; vTable->Skip = (Mf_Skip_Func)MatchFinderMt0_Skip; vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt2_GetMatches; break; case 3: - p->GetHeadsFunc = GetHeads3; + p->GetHeadsFunc = MF(p)->bigHash ? GetHeads3b : GetHeads3; p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches2; vTable->Skip = (Mf_Skip_Func)MatchFinderMt2_Skip; break; - default: - /* case 4: */ - p->GetHeadsFunc = p->MatchFinder->bigHash ? GetHeads4b : GetHeads4; + case 4: + p->GetHeadsFunc = MF(p)->bigHash ? GetHeads4b : GetHeads4; + + // it's fast inline version of GetMatches() + // vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt_GetMatches_Bt4; + p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches3; vTable->Skip = (Mf_Skip_Func)MatchFinderMt3_Skip; break; - /* default: - p->GetHeadsFunc = GetHeads5; + p->GetHeadsFunc = MF(p)->bigHash ? GetHeads5b : GetHeads5; p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches4; - vTable->Skip = (Mf_Skip_Func)MatchFinderMt4_Skip; + vTable->Skip = + (Mf_Skip_Func)MatchFinderMt3_Skip; + // (Mf_Skip_Func)MatchFinderMt4_Skip; break; - */ } } + +#undef RINOK_THREAD +#undef PRF +#undef MF +#undef GetUi24hi_from32 +#undef LOCK_BUFFER +#undef UNLOCK_BUFFER diff --git a/3rdparty/lzma/C/LzFindMt.h b/3rdparty/lzma/C/LzFindMt.h index 89b91fefd6e..db5923ea05b 100644 --- a/3rdparty/lzma/C/LzFindMt.h +++ b/3rdparty/lzma/C/LzFindMt.h @@ -1,39 +1,34 @@ /* LzFindMt.h -- multithreaded Match finder for LZ algorithms -2015-05-03 : Igor Pavlov : Public domain */ +2023-03-05 : Igor Pavlov : Public domain */ -#ifndef __LZ_FIND_MT_H -#define __LZ_FIND_MT_H +#ifndef ZIP7_INC_LZ_FIND_MT_H +#define ZIP7_INC_LZ_FIND_MT_H #include "LzFind.h" #include "Threads.h" EXTERN_C_BEGIN -#define kMtHashBlockSize (1 << 13) -#define kMtHashNumBlocks (1 << 3) -#define kMtHashNumBlocksMask (kMtHashNumBlocks - 1) +typedef struct +{ + UInt32 numProcessedBlocks; + CThread thread; + UInt64 affinity; -#define kMtBtBlockSize (1 << 14) -#define kMtBtNumBlocks (1 << 6) -#define kMtBtNumBlocksMask (kMtBtNumBlocks - 1) + BoolInt wasCreated; + BoolInt needStart; + BoolInt csWasInitialized; + BoolInt csWasEntered; -typedef struct _CMtSync -{ - Bool wasCreated; - Bool needStart; - Bool exit; - Bool stopWriting; + BoolInt exit; + BoolInt stopWriting; - CThread thread; CAutoResetEvent canStart; - CAutoResetEvent wasStarted; CAutoResetEvent wasStopped; CSemaphore freeSemaphore; CSemaphore filledSemaphore; - Bool csWasInitialized; - Bool csWasEntered; CCriticalSection cs; - UInt32 numProcessedBlocks; + // UInt32 numBlocks_Sent; } CMtSync; typedef UInt32 * (*Mf_Mix_Matches)(void *p, UInt32 matchMinPos, UInt32 *distances); @@ -44,23 +39,28 @@ typedef UInt32 * (*Mf_Mix_Matches)(void *p, UInt32 matchMinPos, UInt32 *distance typedef void (*Mf_GetHeads)(const Byte *buffer, UInt32 pos, UInt32 *hash, UInt32 hashMask, UInt32 *heads, UInt32 numHeads, const UInt32 *crc); -typedef struct _CMatchFinderMt +typedef struct { /* LZ */ const Byte *pointerToCurPos; UInt32 *btBuf; - UInt32 btBufPos; - UInt32 btBufPosLimit; + const UInt32 *btBufPos; + const UInt32 *btBufPosLimit; UInt32 lzPos; UInt32 btNumAvailBytes; UInt32 *hash; UInt32 fixedHashSize; + // UInt32 hash4Mask; UInt32 historySize; const UInt32 *crc; Mf_Mix_Matches MixMatchesFunc; - + UInt32 failure_LZ_BT; // failure in BT transfered to LZ + // UInt32 failure_LZ_LZ; // failure in LZ tables + UInt32 failureBuf[1]; + // UInt32 crc[256]; + /* LZ + BT */ CMtSync btSync; Byte btDummy[kMtCacheLineDummy]; @@ -70,6 +70,8 @@ typedef struct _CMatchFinderMt UInt32 hashBufPos; UInt32 hashBufPosLimit; UInt32 hashNumAvail; + UInt32 failure_BT; + CLzRef *son; UInt32 matchMaxLen; @@ -77,7 +79,7 @@ typedef struct _CMatchFinderMt UInt32 pos; const Byte *buffer; UInt32 cyclicBufferPos; - UInt32 cyclicBufferSize; /* it must be historySize + 1 */ + UInt32 cyclicBufferSize; /* it must be = (historySize + 1) */ UInt32 cutValue; /* BT + Hash */ @@ -87,13 +89,19 @@ typedef struct _CMatchFinderMt /* Hash */ Mf_GetHeads GetHeadsFunc; CMatchFinder *MatchFinder; + // CMatchFinder MatchFinder; } CMatchFinderMt; +// only for Mt part void MatchFinderMt_Construct(CMatchFinderMt *p); -void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAlloc *alloc); +void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAllocPtr alloc); + SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddBufferBefore, - UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAlloc *alloc); -void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder *vTable); + UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAllocPtr alloc); +void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder2 *vTable); + +/* call MatchFinderMt_InitMt() before IMatchFinder::Init() */ +SRes MatchFinderMt_InitMt(CMatchFinderMt *p); void MatchFinderMt_ReleaseStream(CMatchFinderMt *p); EXTERN_C_END diff --git a/3rdparty/lzma/C/LzFindOpt.c b/3rdparty/lzma/C/LzFindOpt.c new file mode 100644 index 00000000000..85bdc136d67 --- /dev/null +++ b/3rdparty/lzma/C/LzFindOpt.c @@ -0,0 +1,578 @@ +/* LzFindOpt.c -- multithreaded Match finder for LZ algorithms +2023-04-02 : Igor Pavlov : Public domain */ + +#include "Precomp.h" + +#include "CpuArch.h" +#include "LzFind.h" + +// #include "LzFindMt.h" + +// #define LOG_ITERS + +// #define LOG_THREAD + +#ifdef LOG_THREAD +#include <stdio.h> +#define PRF(x) x +#else +// #define PRF(x) +#endif + +#ifdef LOG_ITERS +#include <stdio.h> +UInt64 g_NumIters_Tree; +UInt64 g_NumIters_Loop; +UInt64 g_NumIters_Bytes; +#define LOG_ITER(x) x +#else +#define LOG_ITER(x) +#endif + +// ---------- BT THREAD ---------- + +#define USE_SON_PREFETCH +#define USE_LONG_MATCH_OPT + +#define kEmptyHashValue 0 + +// #define CYC_TO_POS_OFFSET 0 + +// #define CYC_TO_POS_OFFSET 1 // for debug + +/* +Z7_NO_INLINE +UInt32 * Z7_FASTCALL GetMatchesSpecN_1(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son, + UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, UInt32 *posRes) +{ + do + { + UInt32 delta; + if (hash == size) + break; + delta = *hash++; + + if (delta == 0 || delta > (UInt32)pos) + return NULL; + + lenLimit++; + + if (delta == (UInt32)pos) + { + CLzRef *ptr1 = son + ((size_t)pos << 1) - CYC_TO_POS_OFFSET * 2; + *d++ = 0; + ptr1[0] = kEmptyHashValue; + ptr1[1] = kEmptyHashValue; + } +else +{ + UInt32 *_distances = ++d; + + CLzRef *ptr0 = son + ((size_t)(pos) << 1) - CYC_TO_POS_OFFSET * 2 + 1; + CLzRef *ptr1 = son + ((size_t)(pos) << 1) - CYC_TO_POS_OFFSET * 2; + + const Byte *len0 = cur, *len1 = cur; + UInt32 cutValue = _cutValue; + const Byte *maxLen = cur + _maxLen; + + for (LOG_ITER(g_NumIters_Tree++);;) + { + LOG_ITER(g_NumIters_Loop++); + { + const ptrdiff_t diff = (ptrdiff_t)0 - (ptrdiff_t)delta; + CLzRef *pair = son + ((size_t)(((ptrdiff_t)pos - CYC_TO_POS_OFFSET) + diff) << 1); + const Byte *len = (len0 < len1 ? len0 : len1); + + #ifdef USE_SON_PREFETCH + const UInt32 pair0 = *pair; + #endif + + if (len[diff] == len[0]) + { + if (++len != lenLimit && len[diff] == len[0]) + while (++len != lenLimit) + { + LOG_ITER(g_NumIters_Bytes++); + if (len[diff] != len[0]) + break; + } + if (maxLen < len) + { + maxLen = len; + *d++ = (UInt32)(len - cur); + *d++ = delta - 1; + + if (len == lenLimit) + { + const UInt32 pair1 = pair[1]; + *ptr1 = + #ifdef USE_SON_PREFETCH + pair0; + #else + pair[0]; + #endif + *ptr0 = pair1; + + _distances[-1] = (UInt32)(d - _distances); + + #ifdef USE_LONG_MATCH_OPT + + if (hash == size || *hash != delta || lenLimit[diff] != lenLimit[0] || d >= limit) + break; + + { + for (;;) + { + hash++; + pos++; + cur++; + lenLimit++; + { + CLzRef *ptr = son + ((size_t)(pos) << 1) - CYC_TO_POS_OFFSET * 2; + #if 0 + *(UInt64 *)(void *)ptr = ((const UInt64 *)(const void *)ptr)[diff]; + #else + const UInt32 p0 = ptr[0 + (diff * 2)]; + const UInt32 p1 = ptr[1 + (diff * 2)]; + ptr[0] = p0; + ptr[1] = p1; + // ptr[0] = ptr[0 + (diff * 2)]; + // ptr[1] = ptr[1 + (diff * 2)]; + #endif + } + // PrintSon(son + 2, pos - 1); + // printf("\npos = %x delta = %x\n", pos, delta); + len++; + *d++ = 2; + *d++ = (UInt32)(len - cur); + *d++ = delta - 1; + if (hash == size || *hash != delta || lenLimit[diff] != lenLimit[0] || d >= limit) + break; + } + } + #endif + + break; + } + } + } + + { + const UInt32 curMatch = (UInt32)pos - delta; // (UInt32)(pos + diff); + if (len[diff] < len[0]) + { + delta = pair[1]; + if (delta >= curMatch) + return NULL; + *ptr1 = curMatch; + ptr1 = pair + 1; + len1 = len; + } + else + { + delta = *pair; + if (delta >= curMatch) + return NULL; + *ptr0 = curMatch; + ptr0 = pair; + len0 = len; + } + + delta = (UInt32)pos - delta; + + if (--cutValue == 0 || delta >= pos) + { + *ptr0 = *ptr1 = kEmptyHashValue; + _distances[-1] = (UInt32)(d - _distances); + break; + } + } + } + } // for (tree iterations) +} + pos++; + cur++; + } + while (d < limit); + *posRes = (UInt32)pos; + return d; +} +*/ + +/* define cbs if you use 2 functions. + GetMatchesSpecN_1() : (pos < _cyclicBufferSize) + GetMatchesSpecN_2() : (pos >= _cyclicBufferSize) + + do not define cbs if you use 1 function: + GetMatchesSpecN_2() +*/ + +// #define cbs _cyclicBufferSize + +/* + we use size_t for (pos) and (_cyclicBufferPos_ instead of UInt32 + to eliminate "movsx" BUG in old MSVC x64 compiler. +*/ + +UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son, + UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, + UInt32 *posRes); + +Z7_NO_INLINE +UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son, + UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, + UInt32 *posRes) +{ + do // while (hash != size) + { + UInt32 delta; + + #ifndef cbs + UInt32 cbs; + #endif + + if (hash == size) + break; + + delta = *hash++; + + if (delta == 0) + return NULL; + + lenLimit++; + + #ifndef cbs + cbs = _cyclicBufferSize; + if ((UInt32)pos < cbs) + { + if (delta > (UInt32)pos) + return NULL; + cbs = (UInt32)pos; + } + #endif + + if (delta >= cbs) + { + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + *d++ = 0; + ptr1[0] = kEmptyHashValue; + ptr1[1] = kEmptyHashValue; + } +else +{ + UInt32 *_distances = ++d; + + CLzRef *ptr0 = son + ((size_t)_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + + UInt32 cutValue = _cutValue; + const Byte *len0 = cur, *len1 = cur; + const Byte *maxLen = cur + _maxLen; + + // if (cutValue == 0) { *ptr0 = *ptr1 = kEmptyHashValue; } else + for (LOG_ITER(g_NumIters_Tree++);;) + { + LOG_ITER(g_NumIters_Loop++); + { + // SPEC code + CLzRef *pair = son + ((size_t)((ptrdiff_t)_cyclicBufferPos - (ptrdiff_t)delta + + (ptrdiff_t)(UInt32)(_cyclicBufferPos < delta ? cbs : 0) + ) << 1); + + const ptrdiff_t diff = (ptrdiff_t)0 - (ptrdiff_t)delta; + const Byte *len = (len0 < len1 ? len0 : len1); + + #ifdef USE_SON_PREFETCH + const UInt32 pair0 = *pair; + #endif + + if (len[diff] == len[0]) + { + if (++len != lenLimit && len[diff] == len[0]) + while (++len != lenLimit) + { + LOG_ITER(g_NumIters_Bytes++); + if (len[diff] != len[0]) + break; + } + if (maxLen < len) + { + maxLen = len; + *d++ = (UInt32)(len - cur); + *d++ = delta - 1; + + if (len == lenLimit) + { + const UInt32 pair1 = pair[1]; + *ptr1 = + #ifdef USE_SON_PREFETCH + pair0; + #else + pair[0]; + #endif + *ptr0 = pair1; + + _distances[-1] = (UInt32)(d - _distances); + + #ifdef USE_LONG_MATCH_OPT + + if (hash == size || *hash != delta || lenLimit[diff] != lenLimit[0] || d >= limit) + break; + + { + for (;;) + { + *d++ = 2; + *d++ = (UInt32)(lenLimit - cur); + *d++ = delta - 1; + cur++; + lenLimit++; + // SPEC + _cyclicBufferPos++; + { + // SPEC code + CLzRef *dest = son + ((size_t)(_cyclicBufferPos) << 1); + const CLzRef *src = dest + ((diff + + (ptrdiff_t)(UInt32)((_cyclicBufferPos < delta) ? cbs : 0)) << 1); + // CLzRef *ptr = son + ((size_t)(pos) << 1) - CYC_TO_POS_OFFSET * 2; + #if 0 + *(UInt64 *)(void *)dest = *((const UInt64 *)(const void *)src); + #else + const UInt32 p0 = src[0]; + const UInt32 p1 = src[1]; + dest[0] = p0; + dest[1] = p1; + #endif + } + pos++; + hash++; + if (hash == size || *hash != delta || lenLimit[diff] != lenLimit[0] || d >= limit) + break; + } // for() end for long matches + } + #endif + + break; // break from TREE iterations + } + } + } + { + const UInt32 curMatch = (UInt32)pos - delta; // (UInt32)(pos + diff); + if (len[diff] < len[0]) + { + delta = pair[1]; + *ptr1 = curMatch; + ptr1 = pair + 1; + len1 = len; + if (delta >= curMatch) + return NULL; + } + else + { + delta = *pair; + *ptr0 = curMatch; + ptr0 = pair; + len0 = len; + if (delta >= curMatch) + return NULL; + } + delta = (UInt32)pos - delta; + + if (--cutValue == 0 || delta >= cbs) + { + *ptr0 = *ptr1 = kEmptyHashValue; + _distances[-1] = (UInt32)(d - _distances); + break; + } + } + } + } // for (tree iterations) +} + pos++; + _cyclicBufferPos++; + cur++; + } + while (d < limit); + *posRes = (UInt32)pos; + return d; +} + + + +/* +typedef UInt32 uint32plus; // size_t + +UInt32 * Z7_FASTCALL GetMatchesSpecN_3(uint32plus lenLimit, size_t pos, const Byte *cur, CLzRef *son, + UInt32 _cutValue, UInt32 *d, uint32plus _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, + size_t _cyclicBufferPos, UInt32 _cyclicBufferSize, + UInt32 *posRes) +{ + do // while (hash != size) + { + UInt32 delta; + + #ifndef cbs + UInt32 cbs; + #endif + + if (hash == size) + break; + + delta = *hash++; + + if (delta == 0) + return NULL; + + #ifndef cbs + cbs = _cyclicBufferSize; + if ((UInt32)pos < cbs) + { + if (delta > (UInt32)pos) + return NULL; + cbs = (UInt32)pos; + } + #endif + + if (delta >= cbs) + { + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + *d++ = 0; + ptr1[0] = kEmptyHashValue; + ptr1[1] = kEmptyHashValue; + } +else +{ + CLzRef *ptr0 = son + ((size_t)_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + ((size_t)_cyclicBufferPos << 1); + UInt32 *_distances = ++d; + uint32plus len0 = 0, len1 = 0; + UInt32 cutValue = _cutValue; + uint32plus maxLen = _maxLen; + // lenLimit++; // const Byte *lenLimit = cur + _lenLimit; + + for (LOG_ITER(g_NumIters_Tree++);;) + { + LOG_ITER(g_NumIters_Loop++); + { + // const ptrdiff_t diff = (ptrdiff_t)0 - (ptrdiff_t)delta; + CLzRef *pair = son + ((size_t)((ptrdiff_t)_cyclicBufferPos - delta + + (ptrdiff_t)(UInt32)(_cyclicBufferPos < delta ? cbs : 0) + ) << 1); + const Byte *pb = cur - delta; + uint32plus len = (len0 < len1 ? len0 : len1); + + #ifdef USE_SON_PREFETCH + const UInt32 pair0 = *pair; + #endif + + if (pb[len] == cur[len]) + { + if (++len != lenLimit && pb[len] == cur[len]) + while (++len != lenLimit) + if (pb[len] != cur[len]) + break; + if (maxLen < len) + { + maxLen = len; + *d++ = (UInt32)len; + *d++ = delta - 1; + if (len == lenLimit) + { + { + const UInt32 pair1 = pair[1]; + *ptr0 = pair1; + *ptr1 = + #ifdef USE_SON_PREFETCH + pair0; + #else + pair[0]; + #endif + } + + _distances[-1] = (UInt32)(d - _distances); + + #ifdef USE_LONG_MATCH_OPT + + if (hash == size || *hash != delta || pb[lenLimit] != cur[lenLimit] || d >= limit) + break; + + { + const ptrdiff_t diff = (ptrdiff_t)0 - (ptrdiff_t)delta; + for (;;) + { + *d++ = 2; + *d++ = (UInt32)lenLimit; + *d++ = delta - 1; + _cyclicBufferPos++; + { + CLzRef *dest = son + ((size_t)_cyclicBufferPos << 1); + const CLzRef *src = dest + ((diff + + (ptrdiff_t)(UInt32)(_cyclicBufferPos < delta ? cbs : 0)) << 1); + #if 0 + *(UInt64 *)(void *)dest = *((const UInt64 *)(const void *)src); + #else + const UInt32 p0 = src[0]; + const UInt32 p1 = src[1]; + dest[0] = p0; + dest[1] = p1; + #endif + } + hash++; + pos++; + cur++; + pb++; + if (hash == size || *hash != delta || pb[lenLimit] != cur[lenLimit] || d >= limit) + break; + } + } + #endif + + break; + } + } + } + { + const UInt32 curMatch = (UInt32)pos - delta; + if (pb[len] < cur[len]) + { + delta = pair[1]; + *ptr1 = curMatch; + ptr1 = pair + 1; + len1 = len; + } + else + { + delta = *pair; + *ptr0 = curMatch; + ptr0 = pair; + len0 = len; + } + + { + if (delta >= curMatch) + return NULL; + delta = (UInt32)pos - delta; + if (delta >= cbs + // delta >= _cyclicBufferSize || delta >= pos + || --cutValue == 0) + { + *ptr0 = *ptr1 = kEmptyHashValue; + _distances[-1] = (UInt32)(d - _distances); + break; + } + } + } + } + } // for (tree iterations) +} + pos++; + _cyclicBufferPos++; + cur++; + } + while (d < limit); + *posRes = (UInt32)pos; + return d; +} +*/ diff --git a/3rdparty/lzma/C/LzHash.h b/3rdparty/lzma/C/LzHash.h index e7c942303dd..2b6290b64c0 100644 --- a/3rdparty/lzma/C/LzHash.h +++ b/3rdparty/lzma/C/LzHash.h @@ -1,57 +1,34 @@ -/* LzHash.h -- HASH functions for LZ algorithms -2015-04-12 : Igor Pavlov : Public domain */ +/* LzHash.h -- HASH constants for LZ algorithms +2023-03-05 : Igor Pavlov : Public domain */ -#ifndef __LZ_HASH_H -#define __LZ_HASH_H +#ifndef ZIP7_INC_LZ_HASH_H +#define ZIP7_INC_LZ_HASH_H + +/* + (kHash2Size >= (1 << 8)) : Required + (kHash3Size >= (1 << 16)) : Required +*/ #define kHash2Size (1 << 10) #define kHash3Size (1 << 16) -#define kHash4Size (1 << 20) +// #define kHash4Size (1 << 20) #define kFix3HashSize (kHash2Size) #define kFix4HashSize (kHash2Size + kHash3Size) -#define kFix5HashSize (kHash2Size + kHash3Size + kHash4Size) - -#define HASH2_CALC hv = cur[0] | ((UInt32)cur[1] << 8); - -#define HASH3_CALC { \ - UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ - h2 = temp & (kHash2Size - 1); \ - hv = (temp ^ ((UInt32)cur[2] << 8)) & p->hashMask; } - -#define HASH4_CALC { \ - UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ - h2 = temp & (kHash2Size - 1); \ - temp ^= ((UInt32)cur[2] << 8); \ - h3 = temp & (kHash3Size - 1); \ - hv = (temp ^ (p->crc[cur[3]] << 5)) & p->hashMask; } - -#define HASH5_CALC { \ - UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ - h2 = temp & (kHash2Size - 1); \ - temp ^= ((UInt32)cur[2] << 8); \ - h3 = temp & (kHash3Size - 1); \ - temp ^= (p->crc[cur[3]] << 5); \ - h4 = temp & (kHash4Size - 1); \ - hv = (temp ^ (p->crc[cur[4]] << 3)) & p->hashMask; } - -/* #define HASH_ZIP_CALC hv = ((cur[0] | ((UInt32)cur[1] << 8)) ^ p->crc[cur[2]]) & 0xFFFF; */ -#define HASH_ZIP_CALC hv = ((cur[2] | ((UInt32)cur[0] << 8)) ^ p->crc[cur[1]]) & 0xFFFF; - - -#define MT_HASH2_CALC \ - h2 = (p->crc[cur[0]] ^ cur[1]) & (kHash2Size - 1); - -#define MT_HASH3_CALC { \ - UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ - h2 = temp & (kHash2Size - 1); \ - h3 = (temp ^ ((UInt32)cur[2] << 8)) & (kHash3Size - 1); } - -#define MT_HASH4_CALC { \ - UInt32 temp = p->crc[cur[0]] ^ cur[1]; \ - h2 = temp & (kHash2Size - 1); \ - temp ^= ((UInt32)cur[2] << 8); \ - h3 = temp & (kHash3Size - 1); \ - h4 = (temp ^ (p->crc[cur[3]] << 5)) & (kHash4Size - 1); } +// #define kFix5HashSize (kHash2Size + kHash3Size + kHash4Size) + +/* + We use up to 3 crc values for hash: + crc0 + crc1 << Shift_1 + crc2 << Shift_2 + (Shift_1 = 5) and (Shift_2 = 10) is good tradeoff. + Small values for Shift are not good for collision rate. + Big value for Shift_2 increases the minimum size + of hash table, that will be slow for small files. +*/ + +#define kLzHash_CrcShift_1 5 +#define kLzHash_CrcShift_2 10 #endif diff --git a/3rdparty/lzma/C/Lzma2Dec.c b/3rdparty/lzma/C/Lzma2Dec.c index b84f88a4813..388cbc71732 100644 --- a/3rdparty/lzma/C/Lzma2Dec.c +++ b/3rdparty/lzma/C/Lzma2Dec.c @@ -1,5 +1,5 @@ /* Lzma2Dec.c -- LZMA2 Decoder -2015-11-09 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ /* #define SHOW_DEBUG_INFO */ @@ -14,28 +14,22 @@ #include "Lzma2Dec.h" /* -00000000 - EOS -00000001 U U - Uncompressed Reset Dic -00000010 U U - Uncompressed No Reset -100uuuuu U U P P - LZMA no reset -101uuuuu U U P P - LZMA reset state -110uuuuu U U P P S - LZMA reset state + new prop -111uuuuu U U P P S - LZMA reset state + new prop + reset dic +00000000 - End of data +00000001 U U - Uncompressed, reset dic, need reset state and set new prop +00000010 U U - Uncompressed, no reset +100uuuuu U U P P - LZMA, no reset +101uuuuu U U P P - LZMA, reset state +110uuuuu U U P P S - LZMA, reset state + set new prop +111uuuuu U U P P S - LZMA, reset state + set new prop, reset dic u, U - Unpack Size P - Pack Size S - Props */ -#define LZMA2_CONTROL_LZMA (1 << 7) -#define LZMA2_CONTROL_COPY_NO_RESET 2 #define LZMA2_CONTROL_COPY_RESET_DIC 1 -#define LZMA2_CONTROL_EOF 0 -#define LZMA2_IS_UNCOMPRESSED_STATE(p) (((p)->control & LZMA2_CONTROL_LZMA) == 0) - -#define LZMA2_GET_LZMA_MODE(p) (((p)->control >> 5) & 3) -#define LZMA2_IS_THERE_PROP(mode) ((mode) >= 2) +#define LZMA2_IS_UNCOMPRESSED_STATE(p) (((p)->control & (1 << 7)) == 0) #define LZMA2_LCLP_MAX 4 #define LZMA2_DIC_SIZE_FROM_PROP(p) (((UInt32)2 | ((p) & 1)) << ((p) / 2 + 11)) @@ -74,47 +68,57 @@ static SRes Lzma2Dec_GetOldProps(Byte prop, Byte *props) return SZ_OK; } -SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAlloc *alloc) +SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc) { Byte props[LZMA_PROPS_SIZE]; - RINOK(Lzma2Dec_GetOldProps(prop, props)); + RINOK(Lzma2Dec_GetOldProps(prop, props)) return LzmaDec_AllocateProbs(&p->decoder, props, LZMA_PROPS_SIZE, alloc); } -SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAlloc *alloc) +SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc) { Byte props[LZMA_PROPS_SIZE]; - RINOK(Lzma2Dec_GetOldProps(prop, props)); + RINOK(Lzma2Dec_GetOldProps(prop, props)) return LzmaDec_Allocate(&p->decoder, props, LZMA_PROPS_SIZE, alloc); } void Lzma2Dec_Init(CLzma2Dec *p) { p->state = LZMA2_STATE_CONTROL; - p->needInitDic = True; - p->needInitState = True; - p->needInitProp = True; + p->needInitLevel = 0xE0; + p->isExtraMode = False; + p->unpackSize = 0; + + // p->decoder.dicPos = 0; // we can use it instead of full init LzmaDec_Init(&p->decoder); } -static ELzma2State Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b) +// ELzma2State +static unsigned Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b) { switch (p->state) { case LZMA2_STATE_CONTROL: + p->isExtraMode = False; p->control = b; - PRF(printf("\n %4X ", (unsigned)p->decoder.dicPos)); - PRF(printf(" %2X", (unsigned)b)); - if (p->control == 0) + PRF(printf("\n %8X", (unsigned)p->decoder.dicPos)); + PRF(printf(" %02X", (unsigned)b)); + if (b == 0) return LZMA2_STATE_FINISHED; if (LZMA2_IS_UNCOMPRESSED_STATE(p)) { - if ((p->control & 0x7F) > 2) + if (b == LZMA2_CONTROL_COPY_RESET_DIC) + p->needInitLevel = 0xC0; + else if (b > 2 || p->needInitLevel == 0xE0) return LZMA2_STATE_ERROR; - p->unpackSize = 0; } else - p->unpackSize = (UInt32)(p->control & 0x1F) << 16; + { + if (b < p->needInitLevel) + return LZMA2_STATE_ERROR; + p->needInitLevel = 0; + p->unpackSize = (UInt32)(b & 0x1F) << 16; + } return LZMA2_STATE_UNPACK0; case LZMA2_STATE_UNPACK0: @@ -124,8 +128,8 @@ static ELzma2State Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b) case LZMA2_STATE_UNPACK1: p->unpackSize |= (UInt32)b; p->unpackSize++; - PRF(printf(" %8u", (unsigned)p->unpackSize)); - return (LZMA2_IS_UNCOMPRESSED_STATE(p)) ? LZMA2_STATE_DATA : LZMA2_STATE_PACK0; + PRF(printf(" %7u", (unsigned)p->unpackSize)); + return LZMA2_IS_UNCOMPRESSED_STATE(p) ? LZMA2_STATE_DATA : LZMA2_STATE_PACK0; case LZMA2_STATE_PACK0: p->packSize = (UInt32)b << 8; @@ -134,9 +138,9 @@ static ELzma2State Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b) case LZMA2_STATE_PACK1: p->packSize |= (UInt32)b; p->packSize++; - PRF(printf(" %8u", (unsigned)p->packSize)); - return LZMA2_IS_THERE_PROP(LZMA2_GET_LZMA_MODE(p)) ? LZMA2_STATE_PROP: - (p->needInitProp ? LZMA2_STATE_ERROR : LZMA2_STATE_DATA); + // if (p->packSize < 5) return LZMA2_STATE_ERROR; + PRF(printf(" %5u", (unsigned)p->packSize)); + return (p->control & 0x40) ? LZMA2_STATE_PROP : LZMA2_STATE_DATA; case LZMA2_STATE_PROP: { @@ -145,13 +149,12 @@ static ELzma2State Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b) return LZMA2_STATE_ERROR; lc = b % 9; b /= 9; - p->decoder.prop.pb = b / 5; + p->decoder.prop.pb = (Byte)(b / 5); lp = b % 5; if (lc + lp > LZMA2_LCLP_MAX) return LZMA2_STATE_ERROR; - p->decoder.prop.lc = lc; - p->decoder.prop.lp = lp; - p->needInitProp = False; + p->decoder.prop.lc = (Byte)lc; + p->decoder.prop.lp = (Byte)lp; return LZMA2_STATE_DATA; } } @@ -167,7 +170,8 @@ static void LzmaDec_UpdateWithUncompressed(CLzmaDec *p, const Byte *src, SizeT s p->processedPos += (UInt32)size; } -void LzmaDec_InitDicAndState(CLzmaDec *p, Bool initDic, Bool initState); +void LzmaDec_InitDicAndState(CLzmaDec *p, BoolInt initDic, BoolInt initState); + SRes Lzma2Dec_DecodeToDic(CLzma2Dec *p, SizeT dicLimit, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status) @@ -176,12 +180,17 @@ SRes Lzma2Dec_DecodeToDic(CLzma2Dec *p, SizeT dicLimit, *srcLen = 0; *status = LZMA_STATUS_NOT_SPECIFIED; - while (p->state != LZMA2_STATE_FINISHED) + while (p->state != LZMA2_STATE_ERROR) { - SizeT dicPos = p->decoder.dicPos; + SizeT dicPos; + + if (p->state == LZMA2_STATE_FINISHED) + { + *status = LZMA_STATUS_FINISHED_WITH_MARK; + return SZ_OK; + } - if (p->state == LZMA2_STATE_ERROR) - return SZ_ERROR_DATA; + dicPos = p->decoder.dicPos; if (dicPos == dicLimit && finishMode == LZMA_FINISH_ANY) { @@ -198,29 +207,25 @@ SRes Lzma2Dec_DecodeToDic(CLzma2Dec *p, SizeT dicLimit, } (*srcLen)++; p->state = Lzma2Dec_UpdateState(p, *src++); - if (dicPos == dicLimit && p->state != LZMA2_STATE_FINISHED) - { - p->state = LZMA2_STATE_ERROR; - return SZ_ERROR_DATA; - } + break; continue; } { - SizeT destSizeCur = dicLimit - dicPos; - SizeT srcSizeCur = inSize - *srcLen; + SizeT inCur = inSize - *srcLen; + SizeT outCur = dicLimit - dicPos; ELzmaFinishMode curFinishMode = LZMA_FINISH_ANY; - if (p->unpackSize <= destSizeCur) + if (outCur >= p->unpackSize) { - destSizeCur = (SizeT)p->unpackSize; + outCur = (SizeT)p->unpackSize; curFinishMode = LZMA_FINISH_END; } if (LZMA2_IS_UNCOMPRESSED_STATE(p)) { - if (*srcLen == inSize) + if (inCur == 0) { *status = LZMA_STATUS_NEEDS_MORE_INPUT; return SZ_OK; @@ -228,143 +233,249 @@ SRes Lzma2Dec_DecodeToDic(CLzma2Dec *p, SizeT dicLimit, if (p->state == LZMA2_STATE_DATA) { - Bool initDic = (p->control == LZMA2_CONTROL_COPY_RESET_DIC); - if (initDic) - p->needInitProp = p->needInitState = True; - else if (p->needInitDic) - { - p->state = LZMA2_STATE_ERROR; - return SZ_ERROR_DATA; - } - p->needInitDic = False; + BoolInt initDic = (p->control == LZMA2_CONTROL_COPY_RESET_DIC); LzmaDec_InitDicAndState(&p->decoder, initDic, False); } - if (srcSizeCur > destSizeCur) - srcSizeCur = destSizeCur; - - if (srcSizeCur == 0) - { - p->state = LZMA2_STATE_ERROR; - return SZ_ERROR_DATA; - } + if (inCur > outCur) + inCur = outCur; + if (inCur == 0) + break; - LzmaDec_UpdateWithUncompressed(&p->decoder, src, srcSizeCur); + LzmaDec_UpdateWithUncompressed(&p->decoder, src, inCur); - src += srcSizeCur; - *srcLen += srcSizeCur; - p->unpackSize -= (UInt32)srcSizeCur; + src += inCur; + *srcLen += inCur; + p->unpackSize -= (UInt32)inCur; p->state = (p->unpackSize == 0) ? LZMA2_STATE_CONTROL : LZMA2_STATE_DATA_CONT; } else { - SizeT outSizeProcessed; SRes res; if (p->state == LZMA2_STATE_DATA) { - unsigned mode = LZMA2_GET_LZMA_MODE(p); - Bool initDic = (mode == 3); - Bool initState = (mode != 0); - if ((!initDic && p->needInitDic) || (!initState && p->needInitState)) - { - p->state = LZMA2_STATE_ERROR; - return SZ_ERROR_DATA; - } - + BoolInt initDic = (p->control >= 0xE0); + BoolInt initState = (p->control >= 0xA0); LzmaDec_InitDicAndState(&p->decoder, initDic, initState); - p->needInitDic = False; - p->needInitState = False; p->state = LZMA2_STATE_DATA_CONT; } - if (srcSizeCur > p->packSize) - srcSizeCur = (SizeT)p->packSize; - - res = LzmaDec_DecodeToDic(&p->decoder, dicPos + destSizeCur, src, &srcSizeCur, curFinishMode, status); + if (inCur > p->packSize) + inCur = (SizeT)p->packSize; - src += srcSizeCur; - *srcLen += srcSizeCur; - p->packSize -= (UInt32)srcSizeCur; + res = LzmaDec_DecodeToDic(&p->decoder, dicPos + outCur, src, &inCur, curFinishMode, status); - outSizeProcessed = p->decoder.dicPos - dicPos; - p->unpackSize -= (UInt32)outSizeProcessed; + src += inCur; + *srcLen += inCur; + p->packSize -= (UInt32)inCur; + outCur = p->decoder.dicPos - dicPos; + p->unpackSize -= (UInt32)outCur; - RINOK(res); + if (res != 0) + break; + if (*status == LZMA_STATUS_NEEDS_MORE_INPUT) - return res; + { + if (p->packSize == 0) + break; + return SZ_OK; + } - if (srcSizeCur == 0 && outSizeProcessed == 0) + if (inCur == 0 && outCur == 0) { if (*status != LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK || p->unpackSize != 0 || p->packSize != 0) - { - p->state = LZMA2_STATE_ERROR; - return SZ_ERROR_DATA; - } + break; p->state = LZMA2_STATE_CONTROL; } - if (*status == LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK) - *status = LZMA_STATUS_NOT_FINISHED; + *status = LZMA_STATUS_NOT_SPECIFIED; } } } - *status = LZMA_STATUS_FINISHED_WITH_MARK; - return SZ_OK; + *status = LZMA_STATUS_NOT_SPECIFIED; + p->state = LZMA2_STATE_ERROR; + return SZ_ERROR_DATA; +} + + + + +ELzma2ParseStatus Lzma2Dec_Parse(CLzma2Dec *p, + SizeT outSize, + const Byte *src, SizeT *srcLen, + int checkFinishBlock) +{ + SizeT inSize = *srcLen; + *srcLen = 0; + + while (p->state != LZMA2_STATE_ERROR) + { + if (p->state == LZMA2_STATE_FINISHED) + return (ELzma2ParseStatus)LZMA_STATUS_FINISHED_WITH_MARK; + + if (outSize == 0 && !checkFinishBlock) + return (ELzma2ParseStatus)LZMA_STATUS_NOT_FINISHED; + + if (p->state != LZMA2_STATE_DATA && p->state != LZMA2_STATE_DATA_CONT) + { + if (*srcLen == inSize) + return (ELzma2ParseStatus)LZMA_STATUS_NEEDS_MORE_INPUT; + (*srcLen)++; + + p->state = Lzma2Dec_UpdateState(p, *src++); + + if (p->state == LZMA2_STATE_UNPACK0) + { + // if (p->decoder.dicPos != 0) + if (p->control == LZMA2_CONTROL_COPY_RESET_DIC || p->control >= 0xE0) + return LZMA2_PARSE_STATUS_NEW_BLOCK; + // if (outSize == 0) return LZMA_STATUS_NOT_FINISHED; + } + + // The following code can be commented. + // It's not big problem, if we read additional input bytes. + // It will be stopped later in LZMA2_STATE_DATA / LZMA2_STATE_DATA_CONT state. + + if (outSize == 0 && p->state != LZMA2_STATE_FINISHED) + { + // checkFinishBlock is true. So we expect that block must be finished, + // We can return LZMA_STATUS_NOT_SPECIFIED or LZMA_STATUS_NOT_FINISHED here + // break; + return (ELzma2ParseStatus)LZMA_STATUS_NOT_FINISHED; + } + + if (p->state == LZMA2_STATE_DATA) + return LZMA2_PARSE_STATUS_NEW_CHUNK; + + continue; + } + + if (outSize == 0) + return (ELzma2ParseStatus)LZMA_STATUS_NOT_FINISHED; + + { + SizeT inCur = inSize - *srcLen; + + if (LZMA2_IS_UNCOMPRESSED_STATE(p)) + { + if (inCur == 0) + return (ELzma2ParseStatus)LZMA_STATUS_NEEDS_MORE_INPUT; + if (inCur > p->unpackSize) + inCur = p->unpackSize; + if (inCur > outSize) + inCur = outSize; + p->decoder.dicPos += inCur; + src += inCur; + *srcLen += inCur; + outSize -= inCur; + p->unpackSize -= (UInt32)inCur; + p->state = (p->unpackSize == 0) ? LZMA2_STATE_CONTROL : LZMA2_STATE_DATA_CONT; + } + else + { + p->isExtraMode = True; + + if (inCur == 0) + { + if (p->packSize != 0) + return (ELzma2ParseStatus)LZMA_STATUS_NEEDS_MORE_INPUT; + } + else if (p->state == LZMA2_STATE_DATA) + { + p->state = LZMA2_STATE_DATA_CONT; + if (*src != 0) + { + // first byte of lzma chunk must be Zero + *srcLen += 1; + p->packSize--; + break; + } + } + + if (inCur > p->packSize) + inCur = (SizeT)p->packSize; + + src += inCur; + *srcLen += inCur; + p->packSize -= (UInt32)inCur; + + if (p->packSize == 0) + { + SizeT rem = outSize; + if (rem > p->unpackSize) + rem = p->unpackSize; + p->decoder.dicPos += rem; + p->unpackSize -= (UInt32)rem; + outSize -= rem; + if (p->unpackSize == 0) + p->state = LZMA2_STATE_CONTROL; + } + } + } + } + + p->state = LZMA2_STATE_ERROR; + return (ELzma2ParseStatus)LZMA_STATUS_NOT_SPECIFIED; } + + + SRes Lzma2Dec_DecodeToBuf(CLzma2Dec *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status) { SizeT outSize = *destLen, inSize = *srcLen; *srcLen = *destLen = 0; + for (;;) { - SizeT srcSizeCur = inSize, outSizeCur, dicPos; + SizeT inCur = inSize, outCur, dicPos; ELzmaFinishMode curFinishMode; SRes res; + if (p->decoder.dicPos == p->decoder.dicBufSize) p->decoder.dicPos = 0; dicPos = p->decoder.dicPos; - if (outSize > p->decoder.dicBufSize - dicPos) - { - outSizeCur = p->decoder.dicBufSize; - curFinishMode = LZMA_FINISH_ANY; - } - else + curFinishMode = LZMA_FINISH_ANY; + outCur = p->decoder.dicBufSize - dicPos; + + if (outCur >= outSize) { - outSizeCur = dicPos + outSize; + outCur = outSize; curFinishMode = finishMode; } - res = Lzma2Dec_DecodeToDic(p, outSizeCur, src, &srcSizeCur, curFinishMode, status); - src += srcSizeCur; - inSize -= srcSizeCur; - *srcLen += srcSizeCur; - outSizeCur = p->decoder.dicPos - dicPos; - memcpy(dest, p->decoder.dic + dicPos, outSizeCur); - dest += outSizeCur; - outSize -= outSizeCur; - *destLen += outSizeCur; + res = Lzma2Dec_DecodeToDic(p, dicPos + outCur, src, &inCur, curFinishMode, status); + + src += inCur; + inSize -= inCur; + *srcLen += inCur; + outCur = p->decoder.dicPos - dicPos; + memcpy(dest, p->decoder.dic + dicPos, outCur); + dest += outCur; + outSize -= outCur; + *destLen += outCur; if (res != 0) return res; - if (outSizeCur == 0 || outSize == 0) + if (outCur == 0 || outSize == 0) return SZ_OK; } } + SRes Lzma2Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - Byte prop, ELzmaFinishMode finishMode, ELzmaStatus *status, ISzAlloc *alloc) + Byte prop, ELzmaFinishMode finishMode, ELzmaStatus *status, ISzAllocPtr alloc) { CLzma2Dec p; SRes res; SizeT outSize = *destLen, inSize = *srcLen; *destLen = *srcLen = 0; *status = LZMA_STATUS_NOT_SPECIFIED; - Lzma2Dec_Construct(&p); - RINOK(Lzma2Dec_AllocateProbs(&p, prop, alloc)); + Lzma2Dec_CONSTRUCT(&p) + RINOK(Lzma2Dec_AllocateProbs(&p, prop, alloc)) p.decoder.dic = dest; p.decoder.dicBufSize = outSize; Lzma2Dec_Init(&p); @@ -376,3 +487,5 @@ SRes Lzma2Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, Lzma2Dec_FreeProbs(&p, alloc); return res; } + +#undef PRF diff --git a/3rdparty/lzma/C/Lzma2Dec.h b/3rdparty/lzma/C/Lzma2Dec.h index e6a0f6ed6db..1f5233a728a 100644 --- a/3rdparty/lzma/C/Lzma2Dec.h +++ b/3rdparty/lzma/C/Lzma2Dec.h @@ -1,8 +1,8 @@ /* Lzma2Dec.h -- LZMA2 Decoder -2015-05-13 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ -#ifndef __LZMA2_DEC_H -#define __LZMA2_DEC_H +#ifndef ZIP7_INC_LZMA2_DEC_H +#define ZIP7_INC_LZMA2_DEC_H #include "LzmaDec.h" @@ -12,25 +12,25 @@ EXTERN_C_BEGIN typedef struct { - CLzmaDec decoder; - UInt32 packSize; - UInt32 unpackSize; unsigned state; Byte control; - Bool needInitDic; - Bool needInitState; - Bool needInitProp; + Byte needInitLevel; + Byte isExtraMode; + Byte _pad_; + UInt32 packSize; + UInt32 unpackSize; + CLzmaDec decoder; } CLzma2Dec; -#define Lzma2Dec_Construct(p) LzmaDec_Construct(&(p)->decoder) -#define Lzma2Dec_FreeProbs(p, alloc) LzmaDec_FreeProbs(&(p)->decoder, alloc); -#define Lzma2Dec_Free(p, alloc) LzmaDec_Free(&(p)->decoder, alloc); +#define Lzma2Dec_CONSTRUCT(p) LzmaDec_CONSTRUCT(&(p)->decoder) +#define Lzma2Dec_Construct(p) Lzma2Dec_CONSTRUCT(p) +#define Lzma2Dec_FreeProbs(p, alloc) LzmaDec_FreeProbs(&(p)->decoder, alloc) +#define Lzma2Dec_Free(p, alloc) LzmaDec_Free(&(p)->decoder, alloc) -SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAlloc *alloc); -SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAlloc *alloc); +SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc); +SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc); void Lzma2Dec_Init(CLzma2Dec *p); - /* finishMode: It has meaning only if the decoding reaches output limit (*destLen or dicLimit). @@ -53,6 +53,47 @@ SRes Lzma2Dec_DecodeToBuf(CLzma2Dec *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status); +/* ---------- LZMA2 block and chunk parsing ---------- */ + +/* +Lzma2Dec_Parse() parses compressed data stream up to next independent block or next chunk data. +It can return LZMA_STATUS_* code or LZMA2_PARSE_STATUS_* code: + - LZMA2_PARSE_STATUS_NEW_BLOCK - there is new block, and 1 additional byte (control byte of next block header) was read from input. + - LZMA2_PARSE_STATUS_NEW_CHUNK - there is new chunk, and only lzma2 header of new chunk was read. + CLzma2Dec::unpackSize contains unpack size of that chunk +*/ + +typedef enum +{ +/* + LZMA_STATUS_NOT_SPECIFIED // data error + LZMA_STATUS_FINISHED_WITH_MARK + LZMA_STATUS_NOT_FINISHED // + LZMA_STATUS_NEEDS_MORE_INPUT + LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK // unused +*/ + LZMA2_PARSE_STATUS_NEW_BLOCK = LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK + 1, + LZMA2_PARSE_STATUS_NEW_CHUNK +} ELzma2ParseStatus; + +ELzma2ParseStatus Lzma2Dec_Parse(CLzma2Dec *p, + SizeT outSize, // output size + const Byte *src, SizeT *srcLen, + int checkFinishBlock // set (checkFinishBlock = 1), if it must read full input data, if decoder.dicPos reaches blockMax position. + ); + +/* +LZMA2 parser doesn't decode LZMA chunks, so we must read + full input LZMA chunk to decode some part of LZMA chunk. + +Lzma2Dec_GetUnpackExtra() returns the value that shows + max possible number of output bytes that can be output by decoder + at current input positon. +*/ + +#define Lzma2Dec_GetUnpackExtra(p) ((p)->isExtraMode ? (p)->unpackSize : 0) + + /* ---------- One Call Interface ---------- */ /* @@ -73,7 +114,7 @@ Returns: */ SRes Lzma2Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - Byte prop, ELzmaFinishMode finishMode, ELzmaStatus *status, ISzAlloc *alloc); + Byte prop, ELzmaFinishMode finishMode, ELzmaStatus *status, ISzAllocPtr alloc); EXTERN_C_END diff --git a/3rdparty/lzma/C/Lzma2DecMt.c b/3rdparty/lzma/C/Lzma2DecMt.c new file mode 100644 index 00000000000..4bc4ddeb5c4 --- /dev/null +++ b/3rdparty/lzma/C/Lzma2DecMt.c @@ -0,0 +1,1095 @@ +/* Lzma2DecMt.c -- LZMA2 Decoder Multi-thread +2023-04-13 : Igor Pavlov : Public domain */ + +#include "Precomp.h" + +// #define SHOW_DEBUG_INFO +// #define Z7_ST + +#ifdef SHOW_DEBUG_INFO +#include <stdio.h> +#endif + +#include "Alloc.h" + +#include "Lzma2Dec.h" +#include "Lzma2DecMt.h" + +#ifndef Z7_ST +#include "MtDec.h" + +#define LZMA2DECMT_OUT_BLOCK_MAX_DEFAULT (1 << 28) +#endif + + +#ifndef Z7_ST +#ifdef SHOW_DEBUG_INFO +#define PRF(x) x +#else +#define PRF(x) +#endif +#define PRF_STR(s) PRF(printf("\n" s "\n");) +#define PRF_STR_INT_2(s, d1, d2) PRF(printf("\n" s " %d %d\n", (unsigned)d1, (unsigned)d2);) +#endif + + +void Lzma2DecMtProps_Init(CLzma2DecMtProps *p) +{ + p->inBufSize_ST = 1 << 20; + p->outStep_ST = 1 << 20; + + #ifndef Z7_ST + p->numThreads = 1; + p->inBufSize_MT = 1 << 18; + p->outBlockMax = LZMA2DECMT_OUT_BLOCK_MAX_DEFAULT; + p->inBlockMax = p->outBlockMax + p->outBlockMax / 16; + #endif +} + + + +#ifndef Z7_ST + +/* ---------- CLzma2DecMtThread ---------- */ + +typedef struct +{ + CLzma2Dec dec; + Byte dec_created; + Byte needInit; + + Byte *outBuf; + size_t outBufSize; + + EMtDecParseState state; + ELzma2ParseStatus parseStatus; + + size_t inPreSize; + size_t outPreSize; + + size_t inCodeSize; + size_t outCodeSize; + SRes codeRes; + + CAlignOffsetAlloc alloc; + + Byte mtPad[1 << 7]; +} CLzma2DecMtThread; + +#endif + + +/* ---------- CLzma2DecMt ---------- */ + +struct CLzma2DecMt +{ + // ISzAllocPtr alloc; + ISzAllocPtr allocMid; + + CAlignOffsetAlloc alignOffsetAlloc; + CLzma2DecMtProps props; + Byte prop; + + ISeqInStreamPtr inStream; + ISeqOutStreamPtr outStream; + ICompressProgressPtr progress; + + BoolInt finishMode; + BoolInt outSize_Defined; + UInt64 outSize; + + UInt64 outProcessed; + UInt64 inProcessed; + BoolInt readWasFinished; + SRes readRes; + + Byte *inBuf; + size_t inBufSize; + Byte dec_created; + CLzma2Dec dec; + + size_t inPos; + size_t inLim; + + #ifndef Z7_ST + UInt64 outProcessed_Parse; + BoolInt mtc_WasConstructed; + CMtDec mtc; + CLzma2DecMtThread coders[MTDEC_THREADS_MAX]; + #endif +}; + + + +CLzma2DecMtHandle Lzma2DecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid) +{ + CLzma2DecMt *p = (CLzma2DecMt *)ISzAlloc_Alloc(alloc, sizeof(CLzma2DecMt)); + if (!p) + return NULL; + + // p->alloc = alloc; + p->allocMid = allocMid; + + AlignOffsetAlloc_CreateVTable(&p->alignOffsetAlloc); + p->alignOffsetAlloc.numAlignBits = 7; + p->alignOffsetAlloc.offset = 0; + p->alignOffsetAlloc.baseAlloc = alloc; + + p->inBuf = NULL; + p->inBufSize = 0; + p->dec_created = False; + + // Lzma2DecMtProps_Init(&p->props); + + #ifndef Z7_ST + p->mtc_WasConstructed = False; + { + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CLzma2DecMtThread *t = &p->coders[i]; + t->dec_created = False; + t->outBuf = NULL; + t->outBufSize = 0; + } + } + #endif + + return (CLzma2DecMtHandle)(void *)p; +} + + +#ifndef Z7_ST + +static void Lzma2DecMt_FreeOutBufs(CLzma2DecMt *p) +{ + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CLzma2DecMtThread *t = &p->coders[i]; + if (t->outBuf) + { + ISzAlloc_Free(p->allocMid, t->outBuf); + t->outBuf = NULL; + t->outBufSize = 0; + } + } +} + +#endif + + +static void Lzma2DecMt_FreeSt(CLzma2DecMt *p) +{ + if (p->dec_created) + { + Lzma2Dec_Free(&p->dec, &p->alignOffsetAlloc.vt); + p->dec_created = False; + } + if (p->inBuf) + { + ISzAlloc_Free(p->allocMid, p->inBuf); + p->inBuf = NULL; + } + p->inBufSize = 0; +} + + +// #define GET_CLzma2DecMt_p CLzma2DecMt *p = (CLzma2DecMt *)(void *)pp; + +void Lzma2DecMt_Destroy(CLzma2DecMtHandle p) +{ + // GET_CLzma2DecMt_p + + Lzma2DecMt_FreeSt(p); + + #ifndef Z7_ST + + if (p->mtc_WasConstructed) + { + MtDec_Destruct(&p->mtc); + p->mtc_WasConstructed = False; + } + { + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CLzma2DecMtThread *t = &p->coders[i]; + if (t->dec_created) + { + // we don't need to free dict here + Lzma2Dec_FreeProbs(&t->dec, &t->alloc.vt); // p->alloc !!! + t->dec_created = False; + } + } + } + Lzma2DecMt_FreeOutBufs(p); + + #endif + + ISzAlloc_Free(p->alignOffsetAlloc.baseAlloc, p); +} + + + +#ifndef Z7_ST + +static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCallbackInfo *cc) +{ + CLzma2DecMt *me = (CLzma2DecMt *)obj; + CLzma2DecMtThread *t = &me->coders[coderIndex]; + + PRF_STR_INT_2("Parse", coderIndex, cc->srcSize) + + cc->state = MTDEC_PARSE_CONTINUE; + + if (cc->startCall) + { + if (!t->dec_created) + { + Lzma2Dec_CONSTRUCT(&t->dec) + t->dec_created = True; + AlignOffsetAlloc_CreateVTable(&t->alloc); + { + /* (1 << 12) is expected size of one way in data cache. + We optimize alignment for cache line size of 128 bytes and smaller */ + const unsigned kNumAlignBits = 12; + const unsigned kNumCacheLineBits = 7; /* <= kNumAlignBits */ + t->alloc.numAlignBits = kNumAlignBits; + t->alloc.offset = ((UInt32)coderIndex * (((unsigned)1 << 11) + (1 << 8) + (1 << 6))) & (((unsigned)1 << kNumAlignBits) - ((unsigned)1 << kNumCacheLineBits)); + t->alloc.baseAlloc = me->alignOffsetAlloc.baseAlloc; + } + } + Lzma2Dec_Init(&t->dec); + + t->inPreSize = 0; + t->outPreSize = 0; + // t->blockWasFinished = False; + // t->finishedWithMark = False; + t->parseStatus = (ELzma2ParseStatus)LZMA_STATUS_NOT_SPECIFIED; + t->state = MTDEC_PARSE_CONTINUE; + + t->inCodeSize = 0; + t->outCodeSize = 0; + t->codeRes = SZ_OK; + + // (cc->srcSize == 0) is allowed + } + + { + ELzma2ParseStatus status; + BoolInt overflow; + UInt32 unpackRem = 0; + + int checkFinishBlock = True; + size_t limit = me->props.outBlockMax; + if (me->outSize_Defined) + { + UInt64 rem = me->outSize - me->outProcessed_Parse; + if (limit >= rem) + { + limit = (size_t)rem; + if (!me->finishMode) + checkFinishBlock = False; + } + } + + // checkFinishBlock = False, if we want to decode partial data + // that must be finished at position <= outBlockMax. + + { + const size_t srcOrig = cc->srcSize; + SizeT srcSize_Point = 0; + SizeT dicPos_Point = 0; + + cc->srcSize = 0; + overflow = False; + + for (;;) + { + SizeT srcCur = (SizeT)(srcOrig - cc->srcSize); + + status = Lzma2Dec_Parse(&t->dec, + (SizeT)limit - t->dec.decoder.dicPos, + cc->src + cc->srcSize, &srcCur, + checkFinishBlock); + + cc->srcSize += srcCur; + + if (status == LZMA2_PARSE_STATUS_NEW_CHUNK) + { + if (t->dec.unpackSize > me->props.outBlockMax - t->dec.decoder.dicPos) + { + overflow = True; + break; + } + continue; + } + + if (status == LZMA2_PARSE_STATUS_NEW_BLOCK) + { + if (t->dec.decoder.dicPos == 0) + continue; + // we decode small blocks in one thread + if (t->dec.decoder.dicPos >= (1 << 14)) + break; + dicPos_Point = t->dec.decoder.dicPos; + srcSize_Point = (SizeT)cc->srcSize; + continue; + } + + if ((int)status == LZMA_STATUS_NOT_FINISHED && checkFinishBlock + // && limit == t->dec.decoder.dicPos + // && limit == me->props.outBlockMax + ) + { + overflow = True; + break; + } + + unpackRem = Lzma2Dec_GetUnpackExtra(&t->dec); + break; + } + + if (dicPos_Point != 0 + && (int)status != LZMA2_PARSE_STATUS_NEW_BLOCK + && (int)status != LZMA_STATUS_FINISHED_WITH_MARK + && (int)status != LZMA_STATUS_NOT_SPECIFIED) + { + // we revert to latest newBlock state + status = LZMA2_PARSE_STATUS_NEW_BLOCK; + unpackRem = 0; + t->dec.decoder.dicPos = dicPos_Point; + cc->srcSize = srcSize_Point; + overflow = False; + } + } + + t->inPreSize += cc->srcSize; + t->parseStatus = status; + + if (overflow) + cc->state = MTDEC_PARSE_OVERFLOW; + else + { + size_t dicPos = t->dec.decoder.dicPos; + + if ((int)status != LZMA_STATUS_NEEDS_MORE_INPUT) + { + if (status == LZMA2_PARSE_STATUS_NEW_BLOCK) + { + cc->state = MTDEC_PARSE_NEW; + cc->srcSize--; // we don't need control byte of next block + t->inPreSize--; + } + else + { + cc->state = MTDEC_PARSE_END; + if ((int)status != LZMA_STATUS_FINISHED_WITH_MARK) + { + // (status == LZMA_STATUS_NOT_SPECIFIED) + // (status == LZMA_STATUS_NOT_FINISHED) + if (unpackRem != 0) + { + /* we also reserve space for max possible number of output bytes of current LZMA chunk */ + size_t rem = limit - dicPos; + if (rem > unpackRem) + rem = unpackRem; + dicPos += rem; + } + } + } + + me->outProcessed_Parse += dicPos; + } + + cc->outPos = dicPos; + t->outPreSize = (size_t)dicPos; + } + + t->state = cc->state; + return; + } +} + + +static SRes Lzma2DecMt_MtCallback_PreCode(void *pp, unsigned coderIndex) +{ + CLzma2DecMt *me = (CLzma2DecMt *)pp; + CLzma2DecMtThread *t = &me->coders[coderIndex]; + Byte *dest = t->outBuf; + + if (t->inPreSize == 0) + { + t->codeRes = SZ_ERROR_DATA; + return t->codeRes; + } + + if (!dest || t->outBufSize < t->outPreSize) + { + if (dest) + { + ISzAlloc_Free(me->allocMid, dest); + t->outBuf = NULL; + t->outBufSize = 0; + } + + dest = (Byte *)ISzAlloc_Alloc(me->allocMid, t->outPreSize + // + (1 << 28) + ); + // Sleep(200); + if (!dest) + return SZ_ERROR_MEM; + t->outBuf = dest; + t->outBufSize = t->outPreSize; + } + + t->dec.decoder.dic = dest; + t->dec.decoder.dicBufSize = (SizeT)t->outPreSize; + + t->needInit = True; + + return Lzma2Dec_AllocateProbs(&t->dec, me->prop, &t->alloc.vt); // alloc.vt +} + + +static SRes Lzma2DecMt_MtCallback_Code(void *pp, unsigned coderIndex, + const Byte *src, size_t srcSize, int srcFinished, + // int finished, int blockFinished, + UInt64 *inCodePos, UInt64 *outCodePos, int *stop) +{ + CLzma2DecMt *me = (CLzma2DecMt *)pp; + CLzma2DecMtThread *t = &me->coders[coderIndex]; + + UNUSED_VAR(srcFinished) + + PRF_STR_INT_2("Code", coderIndex, srcSize) + + *inCodePos = t->inCodeSize; + *outCodePos = 0; + *stop = True; + + if (t->needInit) + { + Lzma2Dec_Init(&t->dec); + t->needInit = False; + } + + { + ELzmaStatus status; + SizeT srcProcessed = (SizeT)srcSize; + BoolInt blockWasFinished = + ((int)t->parseStatus == LZMA_STATUS_FINISHED_WITH_MARK + || t->parseStatus == LZMA2_PARSE_STATUS_NEW_BLOCK); + + SRes res = Lzma2Dec_DecodeToDic(&t->dec, + (SizeT)t->outPreSize, + src, &srcProcessed, + blockWasFinished ? LZMA_FINISH_END : LZMA_FINISH_ANY, + &status); + + t->codeRes = res; + + t->inCodeSize += srcProcessed; + *inCodePos = t->inCodeSize; + t->outCodeSize = t->dec.decoder.dicPos; + *outCodePos = t->dec.decoder.dicPos; + + if (res != SZ_OK) + return res; + + if (srcProcessed == srcSize) + *stop = False; + + if (blockWasFinished) + { + if (srcSize != srcProcessed) + return SZ_ERROR_FAIL; + + if (t->inPreSize == t->inCodeSize) + { + if (t->outPreSize != t->outCodeSize) + return SZ_ERROR_FAIL; + *stop = True; + } + } + else + { + if (t->outPreSize == t->outCodeSize) + *stop = True; + } + + return SZ_OK; + } +} + + +#define LZMA2DECMT_STREAM_WRITE_STEP (1 << 24) + +static SRes Lzma2DecMt_MtCallback_Write(void *pp, unsigned coderIndex, + BoolInt needWriteToStream, + const Byte *src, size_t srcSize, BoolInt isCross, + BoolInt *needContinue, BoolInt *canRecode) +{ + CLzma2DecMt *me = (CLzma2DecMt *)pp; + const CLzma2DecMtThread *t = &me->coders[coderIndex]; + size_t size = t->outCodeSize; + const Byte *data = t->outBuf; + BoolInt needContinue2 = True; + + UNUSED_VAR(src) + UNUSED_VAR(srcSize) + UNUSED_VAR(isCross) + + PRF_STR_INT_2("Write", coderIndex, srcSize) + + *needContinue = False; + *canRecode = True; + + if ( + // t->parseStatus == LZMA_STATUS_FINISHED_WITH_MARK + t->state == MTDEC_PARSE_OVERFLOW + || t->state == MTDEC_PARSE_END) + needContinue2 = False; + + + if (!needWriteToStream) + return SZ_OK; + + me->mtc.inProcessed += t->inCodeSize; + + if (t->codeRes == SZ_OK) + if ((int)t->parseStatus == LZMA_STATUS_FINISHED_WITH_MARK + || t->parseStatus == LZMA2_PARSE_STATUS_NEW_BLOCK) + if (t->outPreSize != t->outCodeSize + || t->inPreSize != t->inCodeSize) + return SZ_ERROR_FAIL; + + *canRecode = False; + + if (me->outStream) + { + for (;;) + { + size_t cur = size; + size_t written; + if (cur > LZMA2DECMT_STREAM_WRITE_STEP) + cur = LZMA2DECMT_STREAM_WRITE_STEP; + + written = ISeqOutStream_Write(me->outStream, data, cur); + + me->outProcessed += written; + // me->mtc.writtenTotal += written; + if (written != cur) + return SZ_ERROR_WRITE; + data += cur; + size -= cur; + if (size == 0) + { + *needContinue = needContinue2; + return SZ_OK; + } + RINOK(MtProgress_ProgressAdd(&me->mtc.mtProgress, 0, 0)) + } + } + + return SZ_ERROR_FAIL; + /* + if (size > me->outBufSize) + return SZ_ERROR_OUTPUT_EOF; + memcpy(me->outBuf, data, size); + me->outBufSize -= size; + me->outBuf += size; + *needContinue = needContinue2; + return SZ_OK; + */ +} + +#endif + + +static SRes Lzma2Dec_Prepare_ST(CLzma2DecMt *p) +{ + if (!p->dec_created) + { + Lzma2Dec_CONSTRUCT(&p->dec) + p->dec_created = True; + } + + RINOK(Lzma2Dec_Allocate(&p->dec, p->prop, &p->alignOffsetAlloc.vt)) + + if (!p->inBuf || p->inBufSize != p->props.inBufSize_ST) + { + ISzAlloc_Free(p->allocMid, p->inBuf); + p->inBufSize = 0; + p->inBuf = (Byte *)ISzAlloc_Alloc(p->allocMid, p->props.inBufSize_ST); + if (!p->inBuf) + return SZ_ERROR_MEM; + p->inBufSize = p->props.inBufSize_ST; + } + + Lzma2Dec_Init(&p->dec); + + return SZ_OK; +} + + +static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p + #ifndef Z7_ST + , BoolInt tMode + #endif + ) +{ + SizeT wrPos; + size_t inPos, inLim; + const Byte *inData; + UInt64 inPrev, outPrev; + + CLzma2Dec *dec; + + #ifndef Z7_ST + if (tMode) + { + Lzma2DecMt_FreeOutBufs(p); + tMode = MtDec_PrepareRead(&p->mtc); + } + #endif + + RINOK(Lzma2Dec_Prepare_ST(p)) + + dec = &p->dec; + + inPrev = p->inProcessed; + outPrev = p->outProcessed; + + inPos = 0; + inLim = 0; + inData = NULL; + wrPos = dec->decoder.dicPos; + + for (;;) + { + SizeT dicPos; + SizeT size; + ELzmaFinishMode finishMode; + SizeT inProcessed; + ELzmaStatus status; + SRes res; + + SizeT outProcessed; + BoolInt outFinished; + BoolInt needStop; + + if (inPos == inLim) + { + #ifndef Z7_ST + if (tMode) + { + inData = MtDec_Read(&p->mtc, &inLim); + inPos = 0; + if (inData) + continue; + tMode = False; + inLim = 0; + } + #endif + + if (!p->readWasFinished) + { + inPos = 0; + inLim = p->inBufSize; + inData = p->inBuf; + p->readRes = ISeqInStream_Read(p->inStream, (void *)(p->inBuf), &inLim); + // p->readProcessed += inLim; + // inLim -= 5; p->readWasFinished = True; // for test + if (inLim == 0 || p->readRes != SZ_OK) + p->readWasFinished = True; + } + } + + dicPos = dec->decoder.dicPos; + { + SizeT next = dec->decoder.dicBufSize; + if (next - wrPos > p->props.outStep_ST) + next = wrPos + (SizeT)p->props.outStep_ST; + size = next - dicPos; + } + + finishMode = LZMA_FINISH_ANY; + if (p->outSize_Defined) + { + const UInt64 rem = p->outSize - p->outProcessed; + if (size >= rem) + { + size = (SizeT)rem; + if (p->finishMode) + finishMode = LZMA_FINISH_END; + } + } + + inProcessed = (SizeT)(inLim - inPos); + + res = Lzma2Dec_DecodeToDic(dec, dicPos + size, inData + inPos, &inProcessed, finishMode, &status); + + inPos += inProcessed; + p->inProcessed += inProcessed; + outProcessed = dec->decoder.dicPos - dicPos; + p->outProcessed += outProcessed; + + outFinished = (p->outSize_Defined && p->outSize <= p->outProcessed); + + needStop = (res != SZ_OK + || (inProcessed == 0 && outProcessed == 0) + || status == LZMA_STATUS_FINISHED_WITH_MARK + || (!p->finishMode && outFinished)); + + if (needStop || outProcessed >= size) + { + SRes res2; + { + size_t writeSize = dec->decoder.dicPos - wrPos; + size_t written = ISeqOutStream_Write(p->outStream, dec->decoder.dic + wrPos, writeSize); + res2 = (written == writeSize) ? SZ_OK : SZ_ERROR_WRITE; + } + + if (dec->decoder.dicPos == dec->decoder.dicBufSize) + dec->decoder.dicPos = 0; + wrPos = dec->decoder.dicPos; + + RINOK(res2) + + if (needStop) + { + if (res != SZ_OK) + return res; + + if (status == LZMA_STATUS_FINISHED_WITH_MARK) + { + if (p->finishMode) + { + if (p->outSize_Defined && p->outSize != p->outProcessed) + return SZ_ERROR_DATA; + } + return SZ_OK; + } + + if (!p->finishMode && outFinished) + return SZ_OK; + + if (status == LZMA_STATUS_NEEDS_MORE_INPUT) + return SZ_ERROR_INPUT_EOF; + + return SZ_ERROR_DATA; + } + } + + if (p->progress) + { + UInt64 inDelta = p->inProcessed - inPrev; + UInt64 outDelta = p->outProcessed - outPrev; + if (inDelta >= (1 << 22) || outDelta >= (1 << 22)) + { + RINOK(ICompressProgress_Progress(p->progress, p->inProcessed, p->outProcessed)) + inPrev = p->inProcessed; + outPrev = p->outProcessed; + } + } + } +} + + + +SRes Lzma2DecMt_Decode(CLzma2DecMtHandle p, + Byte prop, + const CLzma2DecMtProps *props, + ISeqOutStreamPtr outStream, const UInt64 *outDataSize, int finishMode, + // Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + // const Byte *inData, size_t inDataSize, + UInt64 *inProcessed, + // UInt64 *outProcessed, + int *isMT, + ICompressProgressPtr progress) +{ + // GET_CLzma2DecMt_p + #ifndef Z7_ST + BoolInt tMode; + #endif + + *inProcessed = 0; + + if (prop > 40) + return SZ_ERROR_UNSUPPORTED; + + p->prop = prop; + p->props = *props; + + p->inStream = inStream; + p->outStream = outStream; + p->progress = progress; + + p->outSize = 0; + p->outSize_Defined = False; + if (outDataSize) + { + p->outSize_Defined = True; + p->outSize = *outDataSize; + } + p->finishMode = finishMode; + + p->outProcessed = 0; + p->inProcessed = 0; + + p->readWasFinished = False; + p->readRes = SZ_OK; + + *isMT = False; + + + #ifndef Z7_ST + + tMode = False; + + // p->mtc.parseRes = SZ_OK; + + // p->mtc.numFilledThreads = 0; + // p->mtc.crossStart = 0; + // p->mtc.crossEnd = 0; + // p->mtc.allocError_for_Read_BlockIndex = 0; + // p->mtc.isAllocError = False; + + if (p->props.numThreads > 1) + { + IMtDecCallback2 vt; + + Lzma2DecMt_FreeSt(p); + + p->outProcessed_Parse = 0; + + if (!p->mtc_WasConstructed) + { + p->mtc_WasConstructed = True; + MtDec_Construct(&p->mtc); + } + + p->mtc.progress = progress; + p->mtc.inStream = inStream; + + // p->outBuf = NULL; + // p->outBufSize = 0; + /* + if (!outStream) + { + // p->outBuf = outBuf; + // p->outBufSize = *outBufSize; + // *outBufSize = 0; + return SZ_ERROR_PARAM; + } + */ + + // p->mtc.inBlockMax = p->props.inBlockMax; + p->mtc.alloc = &p->alignOffsetAlloc.vt; + // p->alignOffsetAlloc.baseAlloc; + // p->mtc.inData = inData; + // p->mtc.inDataSize = inDataSize; + p->mtc.mtCallback = &vt; + p->mtc.mtCallbackObject = p; + + p->mtc.inBufSize = p->props.inBufSize_MT; + + p->mtc.numThreadsMax = p->props.numThreads; + + *isMT = True; + + vt.Parse = Lzma2DecMt_MtCallback_Parse; + vt.PreCode = Lzma2DecMt_MtCallback_PreCode; + vt.Code = Lzma2DecMt_MtCallback_Code; + vt.Write = Lzma2DecMt_MtCallback_Write; + + { + BoolInt needContinue = False; + + SRes res = MtDec_Code(&p->mtc); + + /* + if (!outStream) + *outBufSize = p->outBuf - outBuf; + */ + + *inProcessed = p->mtc.inProcessed; + + needContinue = False; + + if (res == SZ_OK) + { + if (p->mtc.mtProgress.res != SZ_OK) + res = p->mtc.mtProgress.res; + else + needContinue = p->mtc.needContinue; + } + + if (!needContinue) + { + if (res == SZ_OK) + return p->mtc.readRes; + return res; + } + + tMode = True; + p->readRes = p->mtc.readRes; + p->readWasFinished = p->mtc.readWasFinished; + p->inProcessed = p->mtc.inProcessed; + + PRF_STR("----- decoding ST -----") + } + } + + #endif + + + *isMT = False; + + { + SRes res = Lzma2Dec_Decode_ST(p + #ifndef Z7_ST + , tMode + #endif + ); + + *inProcessed = p->inProcessed; + + // res = SZ_OK; // for test + if (res == SZ_ERROR_INPUT_EOF) + { + if (p->readRes != SZ_OK) + res = p->readRes; + } + else if (res == SZ_OK && p->readRes != SZ_OK) + res = p->readRes; + + /* + #ifndef Z7_ST + if (res == SZ_OK && tMode && p->mtc.parseRes != SZ_OK) + res = p->mtc.parseRes; + #endif + */ + + return res; + } +} + + +/* ---------- Read from CLzma2DecMtHandle Interface ---------- */ + +SRes Lzma2DecMt_Init(CLzma2DecMtHandle p, + Byte prop, + const CLzma2DecMtProps *props, + const UInt64 *outDataSize, int finishMode, + ISeqInStreamPtr inStream) +{ + // GET_CLzma2DecMt_p + + if (prop > 40) + return SZ_ERROR_UNSUPPORTED; + + p->prop = prop; + p->props = *props; + + p->inStream = inStream; + + p->outSize = 0; + p->outSize_Defined = False; + if (outDataSize) + { + p->outSize_Defined = True; + p->outSize = *outDataSize; + } + p->finishMode = finishMode; + + p->outProcessed = 0; + p->inProcessed = 0; + + p->inPos = 0; + p->inLim = 0; + + return Lzma2Dec_Prepare_ST(p); +} + + +SRes Lzma2DecMt_Read(CLzma2DecMtHandle p, + Byte *data, size_t *outSize, + UInt64 *inStreamProcessed) +{ + // GET_CLzma2DecMt_p + ELzmaFinishMode finishMode; + SRes readRes; + size_t size = *outSize; + + *outSize = 0; + *inStreamProcessed = 0; + + finishMode = LZMA_FINISH_ANY; + if (p->outSize_Defined) + { + const UInt64 rem = p->outSize - p->outProcessed; + if (size >= rem) + { + size = (size_t)rem; + if (p->finishMode) + finishMode = LZMA_FINISH_END; + } + } + + readRes = SZ_OK; + + for (;;) + { + SizeT inCur; + SizeT outCur; + ELzmaStatus status; + SRes res; + + if (p->inPos == p->inLim && readRes == SZ_OK) + { + p->inPos = 0; + p->inLim = p->props.inBufSize_ST; + readRes = ISeqInStream_Read(p->inStream, p->inBuf, &p->inLim); + } + + inCur = (SizeT)(p->inLim - p->inPos); + outCur = (SizeT)size; + + res = Lzma2Dec_DecodeToBuf(&p->dec, data, &outCur, + p->inBuf + p->inPos, &inCur, finishMode, &status); + + p->inPos += inCur; + p->inProcessed += inCur; + *inStreamProcessed += inCur; + p->outProcessed += outCur; + *outSize += outCur; + size -= outCur; + data += outCur; + + if (res != 0) + return res; + + /* + if (status == LZMA_STATUS_FINISHED_WITH_MARK) + return readRes; + + if (size == 0 && status != LZMA_STATUS_NEEDS_MORE_INPUT) + { + if (p->finishMode && p->outSize_Defined && p->outProcessed >= p->outSize) + return SZ_ERROR_DATA; + return readRes; + } + */ + + if (inCur == 0 && outCur == 0) + return readRes; + } +} + +#undef PRF +#undef PRF_STR +#undef PRF_STR_INT_2 diff --git a/3rdparty/lzma/C/Lzma2DecMt.h b/3rdparty/lzma/C/Lzma2DecMt.h new file mode 100644 index 00000000000..93a5cd59e63 --- /dev/null +++ b/3rdparty/lzma/C/Lzma2DecMt.h @@ -0,0 +1,81 @@ +/* Lzma2DecMt.h -- LZMA2 Decoder Multi-thread +2023-04-13 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_LZMA2_DEC_MT_H +#define ZIP7_INC_LZMA2_DEC_MT_H + +#include "7zTypes.h" + +EXTERN_C_BEGIN + +typedef struct +{ + size_t inBufSize_ST; + size_t outStep_ST; + + #ifndef Z7_ST + unsigned numThreads; + size_t inBufSize_MT; + size_t outBlockMax; + size_t inBlockMax; + #endif +} CLzma2DecMtProps; + +/* init to single-thread mode */ +void Lzma2DecMtProps_Init(CLzma2DecMtProps *p); + + +/* ---------- CLzma2DecMtHandle Interface ---------- */ + +/* Lzma2DecMt_ * functions can return the following exit codes: +SRes: + SZ_OK - OK + SZ_ERROR_MEM - Memory allocation error + SZ_ERROR_PARAM - Incorrect paramater in props + SZ_ERROR_WRITE - ISeqOutStream write callback error + // SZ_ERROR_OUTPUT_EOF - output buffer overflow - version with (Byte *) output + SZ_ERROR_PROGRESS - some break from progress callback + SZ_ERROR_THREAD - error in multithreading functions (only for Mt version) +*/ + +typedef struct CLzma2DecMt CLzma2DecMt; +typedef CLzma2DecMt * CLzma2DecMtHandle; +// Z7_DECLARE_HANDLE(CLzma2DecMtHandle) + +CLzma2DecMtHandle Lzma2DecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid); +void Lzma2DecMt_Destroy(CLzma2DecMtHandle p); + +SRes Lzma2DecMt_Decode(CLzma2DecMtHandle p, + Byte prop, + const CLzma2DecMtProps *props, + ISeqOutStreamPtr outStream, + const UInt64 *outDataSize, // NULL means undefined + int finishMode, // 0 - partial unpacking is allowed, 1 - if lzma2 stream must be finished + // Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + // const Byte *inData, size_t inDataSize, + + // out variables: + UInt64 *inProcessed, + int *isMT, /* out: (*isMT == 0), if single thread decoding was used */ + + // UInt64 *outProcessed, + ICompressProgressPtr progress); + + +/* ---------- Read from CLzma2DecMtHandle Interface ---------- */ + +SRes Lzma2DecMt_Init(CLzma2DecMtHandle pp, + Byte prop, + const CLzma2DecMtProps *props, + const UInt64 *outDataSize, int finishMode, + ISeqInStreamPtr inStream); + +SRes Lzma2DecMt_Read(CLzma2DecMtHandle pp, + Byte *data, size_t *outSize, + UInt64 *inStreamProcessed); + + +EXTERN_C_END + +#endif diff --git a/3rdparty/lzma/C/Lzma2Enc.c b/3rdparty/lzma/C/Lzma2Enc.c index 1b77a2bbc1c..703e146b57e 100644 --- a/3rdparty/lzma/C/Lzma2Enc.c +++ b/3rdparty/lzma/C/Lzma2Enc.c @@ -1,19 +1,18 @@ /* Lzma2Enc.c -- LZMA2 Encoder -2015-10-04 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ #include "Precomp.h" -/* #include <stdio.h> */ #include <string.h> -/* #define _7ZIP_ST */ +/* #define Z7_ST */ #include "Lzma2Enc.h" -#ifndef _7ZIP_ST +#ifndef Z7_ST #include "MtCoder.h" #else -#define NUM_MT_CODER_THREADS_MAX 1 +#define MTCODER_THREADS_MAX 1 #endif #define LZMA2_CONTROL_LZMA (1 << 7) @@ -35,50 +34,106 @@ #define PRF(x) /* x */ + +/* ---------- CLimitedSeqInStream ---------- */ + +typedef struct +{ + ISeqInStream vt; + ISeqInStreamPtr realStream; + UInt64 limit; + UInt64 processed; + int finished; +} CLimitedSeqInStream; + +static void LimitedSeqInStream_Init(CLimitedSeqInStream *p) +{ + p->limit = (UInt64)(Int64)-1; + p->processed = 0; + p->finished = 0; +} + +static SRes LimitedSeqInStream_Read(ISeqInStreamPtr pp, void *data, size_t *size) +{ + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CLimitedSeqInStream) + size_t size2 = *size; + SRes res = SZ_OK; + + if (p->limit != (UInt64)(Int64)-1) + { + const UInt64 rem = p->limit - p->processed; + if (size2 > rem) + size2 = (size_t)rem; + } + if (size2 != 0) + { + res = ISeqInStream_Read(p->realStream, data, &size2); + p->finished = (size2 == 0 ? 1 : 0); + p->processed += size2; + } + *size = size2; + return res; +} + + /* ---------- CLzma2EncInt ---------- */ typedef struct { CLzmaEncHandle enc; + Byte propsAreSet; + Byte propsByte; + Byte needInitState; + Byte needInitProp; UInt64 srcPos; - Byte props; - Bool needInitState; - Bool needInitProp; } CLzma2EncInt; -static SRes Lzma2EncInt_Init(CLzma2EncInt *p, const CLzma2EncProps *props) + +static SRes Lzma2EncInt_InitStream(CLzma2EncInt *p, const CLzma2EncProps *props) +{ + if (!p->propsAreSet) + { + SizeT propsSize = LZMA_PROPS_SIZE; + Byte propsEncoded[LZMA_PROPS_SIZE]; + RINOK(LzmaEnc_SetProps(p->enc, &props->lzmaProps)) + RINOK(LzmaEnc_WriteProperties(p->enc, propsEncoded, &propsSize)) + p->propsByte = propsEncoded[0]; + p->propsAreSet = True; + } + return SZ_OK; +} + +static void Lzma2EncInt_InitBlock(CLzma2EncInt *p) { - Byte propsEncoded[LZMA_PROPS_SIZE]; - SizeT propsSize = LZMA_PROPS_SIZE; - RINOK(LzmaEnc_SetProps(p->enc, &props->lzmaProps)); - RINOK(LzmaEnc_WriteProperties(p->enc, propsEncoded, &propsSize)); p->srcPos = 0; - p->props = propsEncoded[0]; p->needInitState = True; p->needInitProp = True; - return SZ_OK; } -SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle pp, ISeqInStream *inStream, UInt32 keepWindowSize, - ISzAlloc *alloc, ISzAlloc *allocBig); -SRes LzmaEnc_MemPrepare(CLzmaEncHandle pp, const Byte *src, SizeT srcLen, - UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig); -SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit, + +SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle p, ISeqInStreamPtr inStream, UInt32 keepWindowSize, + ISzAllocPtr alloc, ISzAllocPtr allocBig); +SRes LzmaEnc_MemPrepare(CLzmaEncHandle p, const Byte *src, SizeT srcLen, + UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig); +SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle p, BoolInt reInit, Byte *dest, size_t *destLen, UInt32 desiredPackSize, UInt32 *unpackSize); -const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle pp); -void LzmaEnc_Finish(CLzmaEncHandle pp); -void LzmaEnc_SaveState(CLzmaEncHandle pp); -void LzmaEnc_RestoreState(CLzmaEncHandle pp); +const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle p); +void LzmaEnc_Finish(CLzmaEncHandle p); +void LzmaEnc_SaveState(CLzmaEncHandle p); +void LzmaEnc_RestoreState(CLzmaEncHandle p); +/* +UInt32 LzmaEnc_GetNumAvailableBytes(CLzmaEncHandle p); +*/ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, - size_t *packSizeRes, ISeqOutStream *outStream) + size_t *packSizeRes, ISeqOutStreamPtr outStream) { size_t packSizeLimit = *packSizeRes; size_t packSize = packSizeLimit; UInt32 unpackSize = LZMA2_UNPACK_SIZE_MAX; unsigned lzHeaderSize = 5 + (p->needInitProp ? 1 : 0); - Bool useCopyBlock; + BoolInt useCopyBlock; SRes res; *packSizeRes = 0; @@ -112,7 +167,7 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, while (unpackSize > 0) { - UInt32 u = (unpackSize < LZMA2_COPY_CHUNK_SIZE) ? unpackSize : LZMA2_COPY_CHUNK_SIZE; + const UInt32 u = (unpackSize < LZMA2_COPY_CHUNK_SIZE) ? unpackSize : LZMA2_COPY_CHUNK_SIZE; if (packSizeLimit - destPos < u + 3) return SZ_ERROR_OUTPUT_EOF; outBuf[destPos++] = (Byte)(p->srcPos == 0 ? LZMA2_CONTROL_COPY_RESET_DIC : LZMA2_CONTROL_COPY_NO_RESET); @@ -126,7 +181,7 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, if (outStream) { *packSizeRes += destPos; - if (outStream->Write(outStream, outBuf, destPos) != destPos) + if (ISeqOutStream_Write(outStream, outBuf, destPos) != destPos) return SZ_ERROR_WRITE; destPos = 0; } @@ -141,9 +196,9 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, { size_t destPos = 0; - UInt32 u = unpackSize - 1; - UInt32 pm = (UInt32)(packSize - 1); - unsigned mode = (p->srcPos == 0) ? 3 : (p->needInitState ? (p->needInitProp ? 2 : 1) : 0); + const UInt32 u = unpackSize - 1; + const UInt32 pm = (UInt32)(packSize - 1); + const unsigned mode = (p->srcPos == 0) ? 3 : (p->needInitState ? (p->needInitProp ? 2 : 1) : 0); PRF(printf(" ")); @@ -154,7 +209,7 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, outBuf[destPos++] = (Byte)pm; if (p->needInitProp) - outBuf[destPos++] = p->props; + outBuf[destPos++] = p->propsByte; p->needInitProp = False; p->needInitState = False; @@ -162,7 +217,7 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, p->srcPos += unpackSize; if (outStream) - if (outStream->Write(outStream, outBuf, destPos) != destPos) + if (ISeqOutStream_Write(outStream, outBuf, destPos) != destPos) return SZ_ERROR_WRITE; *packSizeRes = destPos; @@ -176,14 +231,16 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf, void Lzma2EncProps_Init(CLzma2EncProps *p) { LzmaEncProps_Init(&p->lzmaProps); + p->blockSize = LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO; + p->numBlockThreads_Reduced = -1; + p->numBlockThreads_Max = -1; p->numTotalThreads = -1; - p->numBlockThreads = -1; - p->blockSize = 0; } void Lzma2EncProps_Normalize(CLzma2EncProps *p) { - int t1, t1n, t2, t3; + UInt64 fileSize; + int t1, t1n, t2, t2r, t3; { CLzmaEncProps lzmaProps = p->lzmaProps; LzmaEncProps_Normalize(&lzmaProps); @@ -191,11 +248,11 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p) } t1 = p->lzmaProps.numThreads; - t2 = p->numBlockThreads; + t2 = p->numBlockThreads_Max; t3 = p->numTotalThreads; - if (t2 > NUM_MT_CODER_THREADS_MAX) - t2 = NUM_MT_CODER_THREADS_MAX; + if (t2 > MTCODER_THREADS_MAX) + t2 = MTCODER_THREADS_MAX; if (t3 <= 0) { @@ -211,8 +268,8 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p) t1 = 1; t2 = t3; } - if (t2 > NUM_MT_CODER_THREADS_MAX) - t2 = NUM_MT_CODER_THREADS_MAX; + if (t2 > MTCODER_THREADS_MAX) + t2 = MTCODER_THREADS_MAX; } else if (t1 <= 0) { @@ -225,232 +282,189 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p) p->lzmaProps.numThreads = t1; + t2r = t2; + + fileSize = p->lzmaProps.reduceSize; + + if ( p->blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID + && p->blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO + && (p->blockSize < fileSize || fileSize == (UInt64)(Int64)-1)) + p->lzmaProps.reduceSize = p->blockSize; + LzmaEncProps_Normalize(&p->lzmaProps); + p->lzmaProps.reduceSize = fileSize; + t1 = p->lzmaProps.numThreads; - if (p->blockSize == 0) + if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID) { - UInt32 dictSize = p->lzmaProps.dictSize; - UInt64 blockSize = (UInt64)dictSize << 2; - const UInt32 kMinSize = (UInt32)1 << 20; - const UInt32 kMaxSize = (UInt32)1 << 28; - if (blockSize < kMinSize) blockSize = kMinSize; - if (blockSize > kMaxSize) blockSize = kMaxSize; - if (blockSize < dictSize) blockSize = dictSize; - p->blockSize = (size_t)blockSize; + t2r = t2 = 1; + t3 = t1; } - - if (t2 > 1 && p->lzmaProps.reduceSize != (UInt64)(Int64)-1) + else if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO && t2 <= 1) { - UInt64 temp = p->lzmaProps.reduceSize + p->blockSize - 1; - if (temp > p->lzmaProps.reduceSize) + /* if there is no block multi-threading, we use SOLID block */ + p->blockSize = LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID; + } + else + { + if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO) + { + const UInt32 kMinSize = (UInt32)1 << 20; + const UInt32 kMaxSize = (UInt32)1 << 28; + const UInt32 dictSize = p->lzmaProps.dictSize; + UInt64 blockSize = (UInt64)dictSize << 2; + if (blockSize < kMinSize) blockSize = kMinSize; + if (blockSize > kMaxSize) blockSize = kMaxSize; + if (blockSize < dictSize) blockSize = dictSize; + blockSize += (kMinSize - 1); + blockSize &= ~(UInt64)(kMinSize - 1); + p->blockSize = blockSize; + } + + if (t2 > 1 && fileSize != (UInt64)(Int64)-1) { - UInt64 numBlocks = temp / p->blockSize; + UInt64 numBlocks = fileSize / p->blockSize; + if (numBlocks * p->blockSize != fileSize) + numBlocks++; if (numBlocks < (unsigned)t2) { - t2 = (unsigned)numBlocks; - if (t2 == 0) - t2 = 1; - t3 = t1 * t2; + t2r = (int)numBlocks; + if (t2r == 0) + t2r = 1; + t3 = t1 * t2r; } } } - p->numBlockThreads = t2; + p->numBlockThreads_Max = t2; + p->numBlockThreads_Reduced = t2r; p->numTotalThreads = t3; } -static SRes Progress(ICompressProgress *p, UInt64 inSize, UInt64 outSize) +static SRes Progress(ICompressProgressPtr p, UInt64 inSize, UInt64 outSize) { - return (p && p->Progress(p, inSize, outSize) != SZ_OK) ? SZ_ERROR_PROGRESS : SZ_OK; + return (p && ICompressProgress_Progress(p, inSize, outSize) != SZ_OK) ? SZ_ERROR_PROGRESS : SZ_OK; } /* ---------- Lzma2 ---------- */ -typedef struct +struct CLzma2Enc { Byte propEncoded; CLzma2EncProps props; + UInt64 expectedDataSize; - Byte *outBuf; + Byte *tempBufLzma; - ISzAlloc *alloc; - ISzAlloc *allocBig; + ISzAllocPtr alloc; + ISzAllocPtr allocBig; - CLzma2EncInt coders[NUM_MT_CODER_THREADS_MAX]; + CLzma2EncInt coders[MTCODER_THREADS_MAX]; + + #ifndef Z7_ST + + ISeqOutStreamPtr outStream; + Byte *outBuf; + size_t outBuf_Rem; /* remainder in outBuf */ - #ifndef _7ZIP_ST + size_t outBufSize; /* size of allocated outBufs[i] */ + size_t outBufsDataSizes[MTCODER_BLOCKS_MAX]; + BoolInt mtCoder_WasConstructed; CMtCoder mtCoder; - #endif + Byte *outBufs[MTCODER_BLOCKS_MAX]; -} CLzma2Enc; + #endif +}; -/* ---------- Lzma2EncThread ---------- */ -static SRes Lzma2Enc_EncodeMt1(CLzma2EncInt *p, CLzma2Enc *mainEncoder, - ISeqOutStream *outStream, ISeqInStream *inStream, ICompressProgress *progress) +CLzma2EncHandle Lzma2Enc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig) { - UInt64 packTotal = 0; - SRes res = SZ_OK; - - if (!mainEncoder->outBuf) + CLzma2Enc *p = (CLzma2Enc *)ISzAlloc_Alloc(alloc, sizeof(CLzma2Enc)); + if (!p) + return NULL; + Lzma2EncProps_Init(&p->props); + Lzma2EncProps_Normalize(&p->props); + p->expectedDataSize = (UInt64)(Int64)-1; + p->tempBufLzma = NULL; + p->alloc = alloc; + p->allocBig = allocBig; { - mainEncoder->outBuf = (Byte *)IAlloc_Alloc(mainEncoder->alloc, LZMA2_CHUNK_SIZE_COMPRESSED_MAX); - if (!mainEncoder->outBuf) - return SZ_ERROR_MEM; + unsigned i; + for (i = 0; i < MTCODER_THREADS_MAX; i++) + p->coders[i].enc = NULL; } - RINOK(Lzma2EncInt_Init(p, &mainEncoder->props)); - RINOK(LzmaEnc_PrepareForLzma2(p->enc, inStream, LZMA2_KEEP_WINDOW_SIZE, - mainEncoder->alloc, mainEncoder->allocBig)); - - for (;;) + #ifndef Z7_ST + p->mtCoder_WasConstructed = False; { - size_t packSize = LZMA2_CHUNK_SIZE_COMPRESSED_MAX; - res = Lzma2EncInt_EncodeSubblock(p, mainEncoder->outBuf, &packSize, outStream); - if (res != SZ_OK) - break; - packTotal += packSize; - res = Progress(progress, p->srcPos, packTotal); - if (res != SZ_OK) - break; - if (packSize == 0) - break; + unsigned i; + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + p->outBufs[i] = NULL; + p->outBufSize = 0; } - - LzmaEnc_Finish(p->enc); + #endif - if (res == SZ_OK) - { - Byte b = 0; - if (outStream->Write(outStream, &b, 1) != 1) - return SZ_ERROR_WRITE; - } - - return res; + return (CLzma2EncHandle)p; } -#ifndef _7ZIP_ST - -typedef struct -{ - IMtCoderCallback funcTable; - CLzma2Enc *lzma2Enc; -} CMtCallbackImp; +#ifndef Z7_ST -static SRes MtCallbackImp_Code(void *pp, unsigned index, Byte *dest, size_t *destSize, - const Byte *src, size_t srcSize, int finished) +static void Lzma2Enc_FreeOutBufs(CLzma2Enc *p) { - CMtCallbackImp *imp = (CMtCallbackImp *)pp; - CLzma2Enc *mainEncoder = imp->lzma2Enc; - CLzma2EncInt *p = &mainEncoder->coders[index]; - - SRes res = SZ_OK; - { - size_t destLim = *destSize; - *destSize = 0; - - if (srcSize != 0) - { - RINOK(Lzma2EncInt_Init(p, &mainEncoder->props)); - - RINOK(LzmaEnc_MemPrepare(p->enc, src, srcSize, LZMA2_KEEP_WINDOW_SIZE, - mainEncoder->alloc, mainEncoder->allocBig)); - - while (p->srcPos < srcSize) - { - size_t packSize = destLim - *destSize; - res = Lzma2EncInt_EncodeSubblock(p, dest + *destSize, &packSize, NULL); - if (res != SZ_OK) - break; - *destSize += packSize; - - if (packSize == 0) - { - res = SZ_ERROR_FAIL; - break; - } - - if (MtProgress_Set(&mainEncoder->mtCoder.mtProgress, index, p->srcPos, *destSize) != SZ_OK) - { - res = SZ_ERROR_PROGRESS; - break; - } - } - - LzmaEnc_Finish(p->enc); - if (res != SZ_OK) - return res; - } - - if (finished) + unsigned i; + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + if (p->outBufs[i]) { - if (*destSize == destLim) - return SZ_ERROR_OUTPUT_EOF; - dest[(*destSize)++] = 0; + ISzAlloc_Free(p->alloc, p->outBufs[i]); + p->outBufs[i] = NULL; } - } - return res; + p->outBufSize = 0; } #endif +// #define GET_CLzma2Enc_p CLzma2Enc *p = (CLzma2Enc *)(void *)p; -/* ---------- Lzma2Enc ---------- */ - -CLzma2EncHandle Lzma2Enc_Create(ISzAlloc *alloc, ISzAlloc *allocBig) -{ - CLzma2Enc *p = (CLzma2Enc *)alloc->Alloc(alloc, sizeof(CLzma2Enc)); - if (!p) - return NULL; - Lzma2EncProps_Init(&p->props); - Lzma2EncProps_Normalize(&p->props); - p->outBuf = 0; - p->alloc = alloc; - p->allocBig = allocBig; - { - unsigned i; - for (i = 0; i < NUM_MT_CODER_THREADS_MAX; i++) - p->coders[i].enc = 0; - } - - #ifndef _7ZIP_ST - MtCoder_Construct(&p->mtCoder); - #endif - - return p; -} - -void Lzma2Enc_Destroy(CLzma2EncHandle pp) +void Lzma2Enc_Destroy(CLzma2EncHandle p) { - CLzma2Enc *p = (CLzma2Enc *)pp; + // GET_CLzma2Enc_p unsigned i; - for (i = 0; i < NUM_MT_CODER_THREADS_MAX; i++) + for (i = 0; i < MTCODER_THREADS_MAX; i++) { CLzma2EncInt *t = &p->coders[i]; if (t->enc) { LzmaEnc_Destroy(t->enc, p->alloc, p->allocBig); - t->enc = 0; + t->enc = NULL; } } - #ifndef _7ZIP_ST - MtCoder_Destruct(&p->mtCoder); + + #ifndef Z7_ST + if (p->mtCoder_WasConstructed) + { + MtCoder_Destruct(&p->mtCoder); + p->mtCoder_WasConstructed = False; + } + Lzma2Enc_FreeOutBufs(p); #endif - IAlloc_Free(p->alloc, p->outBuf); - IAlloc_Free(p->alloc, pp); + ISzAlloc_Free(p->alloc, p->tempBufLzma); + p->tempBufLzma = NULL; + + ISzAlloc_Free(p->alloc, p); } -SRes Lzma2Enc_SetProps(CLzma2EncHandle pp, const CLzma2EncProps *props) + +SRes Lzma2Enc_SetProps(CLzma2EncHandle p, const CLzma2EncProps *props) { - CLzma2Enc *p = (CLzma2Enc *)pp; + // GET_CLzma2Enc_p CLzmaEncProps lzmaProps = props->lzmaProps; LzmaEncProps_Normalize(&lzmaProps); if (lzmaProps.lc + lzmaProps.lp > LZMA2_LCLP_MAX) @@ -460,9 +474,17 @@ SRes Lzma2Enc_SetProps(CLzma2EncHandle pp, const CLzma2EncProps *props) return SZ_OK; } -Byte Lzma2Enc_WriteProperties(CLzma2EncHandle pp) + +void Lzma2Enc_SetDataSize(CLzma2EncHandle p, UInt64 expectedDataSiize) { - CLzma2Enc *p = (CLzma2Enc *)pp; + // GET_CLzma2Enc_p + p->expectedDataSize = expectedDataSiize; +} + + +Byte Lzma2Enc_WriteProperties(CLzma2EncHandle p) +{ + // GET_CLzma2Enc_p unsigned i; UInt32 dicSize = LzmaEncProps_GetDictSize(&p->props.lzmaProps); for (i = 0; i < 40; i++) @@ -471,50 +493,313 @@ Byte Lzma2Enc_WriteProperties(CLzma2EncHandle pp) return (Byte)i; } -SRes Lzma2Enc_Encode(CLzma2EncHandle pp, - ISeqOutStream *outStream, ISeqInStream *inStream, ICompressProgress *progress) + +static SRes Lzma2Enc_EncodeMt1( + CLzma2Enc *me, + CLzma2EncInt *p, + ISeqOutStreamPtr outStream, + Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + const Byte *inData, size_t inDataSize, + int finished, + ICompressProgressPtr progress) { - CLzma2Enc *p = (CLzma2Enc *)pp; - int i; + UInt64 unpackTotal = 0; + UInt64 packTotal = 0; + size_t outLim = 0; + CLimitedSeqInStream limitedInStream; + + if (outBuf) + { + outLim = *outBufSize; + *outBufSize = 0; + } + + if (!p->enc) + { + p->propsAreSet = False; + p->enc = LzmaEnc_Create(me->alloc); + if (!p->enc) + return SZ_ERROR_MEM; + } - for (i = 0; i < p->props.numBlockThreads; i++) + limitedInStream.realStream = inStream; + if (inStream) + { + limitedInStream.vt.Read = LimitedSeqInStream_Read; + } + + if (!outBuf) { - CLzma2EncInt *t = &p->coders[(unsigned)i]; - if (!t->enc) + // outStream version works only in one thread. So we use CLzma2Enc::tempBufLzma + if (!me->tempBufLzma) { - t->enc = LzmaEnc_Create(p->alloc); - if (!t->enc) + me->tempBufLzma = (Byte *)ISzAlloc_Alloc(me->alloc, LZMA2_CHUNK_SIZE_COMPRESSED_MAX); + if (!me->tempBufLzma) return SZ_ERROR_MEM; } } - #ifndef _7ZIP_ST - if (p->props.numBlockThreads > 1) + RINOK(Lzma2EncInt_InitStream(p, &me->props)) + + for (;;) { - CMtCallbackImp mtCallback; + SRes res = SZ_OK; + SizeT inSizeCur = 0; + + Lzma2EncInt_InitBlock(p); + + LimitedSeqInStream_Init(&limitedInStream); + limitedInStream.limit = me->props.blockSize; + + if (inStream) + { + UInt64 expected = (UInt64)(Int64)-1; + // inStream version works only in one thread. So we use CLzma2Enc::expectedDataSize + if (me->expectedDataSize != (UInt64)(Int64)-1 + && me->expectedDataSize >= unpackTotal) + expected = me->expectedDataSize - unpackTotal; + if (me->props.blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID + && expected > me->props.blockSize) + expected = (size_t)me->props.blockSize; + + LzmaEnc_SetDataSize(p->enc, expected); + + RINOK(LzmaEnc_PrepareForLzma2(p->enc, + &limitedInStream.vt, + LZMA2_KEEP_WINDOW_SIZE, + me->alloc, + me->allocBig)) + } + else + { + inSizeCur = (SizeT)(inDataSize - (size_t)unpackTotal); + if (me->props.blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID + && inSizeCur > me->props.blockSize) + inSizeCur = (SizeT)(size_t)me->props.blockSize; + + // LzmaEnc_SetDataSize(p->enc, inSizeCur); + + RINOK(LzmaEnc_MemPrepare(p->enc, + inData + (size_t)unpackTotal, inSizeCur, + LZMA2_KEEP_WINDOW_SIZE, + me->alloc, + me->allocBig)) + } + + for (;;) + { + size_t packSize = LZMA2_CHUNK_SIZE_COMPRESSED_MAX; + if (outBuf) + packSize = outLim - (size_t)packTotal; + + res = Lzma2EncInt_EncodeSubblock(p, + outBuf ? outBuf + (size_t)packTotal : me->tempBufLzma, &packSize, + outBuf ? NULL : outStream); + + if (res != SZ_OK) + break; + + packTotal += packSize; + if (outBuf) + *outBufSize = (size_t)packTotal; + + res = Progress(progress, unpackTotal + p->srcPos, packTotal); + if (res != SZ_OK) + break; + + /* + if (LzmaEnc_GetNumAvailableBytes(p->enc) == 0) + break; + */ + + if (packSize == 0) + break; + } + + LzmaEnc_Finish(p->enc); + + unpackTotal += p->srcPos; + + RINOK(res) - mtCallback.funcTable.Code = MtCallbackImp_Code; - mtCallback.lzma2Enc = p; + if (p->srcPos != (inStream ? limitedInStream.processed : inSizeCur)) + return SZ_ERROR_FAIL; + if (inStream ? limitedInStream.finished : (unpackTotal == inDataSize)) + { + if (finished) + { + if (outBuf) + { + const size_t destPos = *outBufSize; + if (destPos >= outLim) + return SZ_ERROR_OUTPUT_EOF; + outBuf[destPos] = LZMA2_CONTROL_EOF; // 0 + *outBufSize = destPos + 1; + } + else + { + const Byte b = LZMA2_CONTROL_EOF; // 0; + if (ISeqOutStream_Write(outStream, &b, 1) != 1) + return SZ_ERROR_WRITE; + } + } + return SZ_OK; + } + } +} + + + +#ifndef Z7_ST + +static SRes Lzma2Enc_MtCallback_Code(void *p, unsigned coderIndex, unsigned outBufIndex, + const Byte *src, size_t srcSize, int finished) +{ + CLzma2Enc *me = (CLzma2Enc *)p; + size_t destSize = me->outBufSize; + SRes res; + CMtProgressThunk progressThunk; + + Byte *dest = me->outBufs[outBufIndex]; + + me->outBufsDataSizes[outBufIndex] = 0; + + if (!dest) + { + dest = (Byte *)ISzAlloc_Alloc(me->alloc, me->outBufSize); + if (!dest) + return SZ_ERROR_MEM; + me->outBufs[outBufIndex] = dest; + } + + MtProgressThunk_CreateVTable(&progressThunk); + progressThunk.mtProgress = &me->mtCoder.mtProgress; + progressThunk.inSize = 0; + progressThunk.outSize = 0; + + res = Lzma2Enc_EncodeMt1(me, + &me->coders[coderIndex], + NULL, dest, &destSize, + NULL, src, srcSize, + finished, + &progressThunk.vt); + + me->outBufsDataSizes[outBufIndex] = destSize; + + return res; +} + + +static SRes Lzma2Enc_MtCallback_Write(void *p, unsigned outBufIndex) +{ + CLzma2Enc *me = (CLzma2Enc *)p; + size_t size = me->outBufsDataSizes[outBufIndex]; + const Byte *data = me->outBufs[outBufIndex]; + + if (me->outStream) + return ISeqOutStream_Write(me->outStream, data, size) == size ? SZ_OK : SZ_ERROR_WRITE; + + if (size > me->outBuf_Rem) + return SZ_ERROR_OUTPUT_EOF; + memcpy(me->outBuf, data, size); + me->outBuf_Rem -= size; + me->outBuf += size; + return SZ_OK; +} + +#endif + + + +SRes Lzma2Enc_Encode2(CLzma2EncHandle p, + ISeqOutStreamPtr outStream, + Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + const Byte *inData, size_t inDataSize, + ICompressProgressPtr progress) +{ + // GET_CLzma2Enc_p + + if (inStream && inData) + return SZ_ERROR_PARAM; + + if (outStream && outBuf) + return SZ_ERROR_PARAM; + + { + unsigned i; + for (i = 0; i < MTCODER_THREADS_MAX; i++) + p->coders[i].propsAreSet = False; + } + + #ifndef Z7_ST + + if (p->props.numBlockThreads_Reduced > 1) + { + IMtCoderCallback2 vt; + + if (!p->mtCoder_WasConstructed) + { + p->mtCoder_WasConstructed = True; + MtCoder_Construct(&p->mtCoder); + } + + vt.Code = Lzma2Enc_MtCallback_Code; + vt.Write = Lzma2Enc_MtCallback_Write; + + p->outStream = outStream; + p->outBuf = NULL; + p->outBuf_Rem = 0; + if (!outStream) + { + p->outBuf = outBuf; + p->outBuf_Rem = *outBufSize; + *outBufSize = 0; + } + + p->mtCoder.allocBig = p->allocBig; p->mtCoder.progress = progress; p->mtCoder.inStream = inStream; - p->mtCoder.outStream = outStream; - p->mtCoder.alloc = p->alloc; - p->mtCoder.mtCallback = &mtCallback.funcTable; + p->mtCoder.inData = inData; + p->mtCoder.inDataSize = inDataSize; + p->mtCoder.mtCallback = &vt; + p->mtCoder.mtCallbackObject = p; + + p->mtCoder.blockSize = (size_t)p->props.blockSize; + if (p->mtCoder.blockSize != p->props.blockSize) + return SZ_ERROR_PARAM; /* SZ_ERROR_MEM */ - p->mtCoder.blockSize = p->props.blockSize; - p->mtCoder.destBlockSize = p->props.blockSize + (p->props.blockSize >> 10) + 16; - if (p->mtCoder.destBlockSize < p->props.blockSize) { - p->mtCoder.destBlockSize = (size_t)0 - 1; - if (p->mtCoder.destBlockSize < p->props.blockSize) - return SZ_ERROR_FAIL; + const size_t destBlockSize = p->mtCoder.blockSize + (p->mtCoder.blockSize >> 10) + 16; + if (destBlockSize < p->mtCoder.blockSize) + return SZ_ERROR_PARAM; + if (p->outBufSize != destBlockSize) + Lzma2Enc_FreeOutBufs(p); + p->outBufSize = destBlockSize; } - p->mtCoder.numThreads = p->props.numBlockThreads; + + p->mtCoder.numThreadsMax = (unsigned)p->props.numBlockThreads_Max; + p->mtCoder.expectedDataSize = p->expectedDataSize; - return MtCoder_Code(&p->mtCoder); + { + const SRes res = MtCoder_Code(&p->mtCoder); + if (!outStream) + *outBufSize = (size_t)(p->outBuf - outBuf); + return res; + } } + #endif - return Lzma2Enc_EncodeMt1(&p->coders[0], p, outStream, inStream, progress); + + return Lzma2Enc_EncodeMt1(p, + &p->coders[0], + outStream, outBuf, outBufSize, + inStream, inData, inDataSize, + True, /* finished */ + progress); } + +#undef PRF diff --git a/3rdparty/lzma/C/Lzma2Enc.h b/3rdparty/lzma/C/Lzma2Enc.h index f409f184cc3..cb25275c6b8 100644 --- a/3rdparty/lzma/C/Lzma2Enc.h +++ b/3rdparty/lzma/C/Lzma2Enc.h @@ -1,18 +1,22 @@ /* Lzma2Enc.h -- LZMA2 Encoder -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ -#ifndef __LZMA2_ENC_H -#define __LZMA2_ENC_H +#ifndef ZIP7_INC_LZMA2_ENC_H +#define ZIP7_INC_LZMA2_ENC_H #include "LzmaEnc.h" EXTERN_C_BEGIN +#define LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO 0 +#define LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID ((UInt64)(Int64)-1) + typedef struct { CLzmaEncProps lzmaProps; - size_t blockSize; - int numBlockThreads; + UInt64 blockSize; + int numBlockThreads_Reduced; + int numBlockThreads_Max; int numTotalThreads; } CLzma2EncProps; @@ -22,40 +26,31 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p); /* ---------- CLzmaEnc2Handle Interface ---------- */ /* Lzma2Enc_* functions can return the following exit codes: -Returns: +SRes: SZ_OK - OK SZ_ERROR_MEM - Memory allocation error SZ_ERROR_PARAM - Incorrect paramater in props - SZ_ERROR_WRITE - Write callback error + SZ_ERROR_WRITE - ISeqOutStream write callback error + SZ_ERROR_OUTPUT_EOF - output buffer overflow - version with (Byte *) output SZ_ERROR_PROGRESS - some break from progress callback - SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version) + SZ_ERROR_THREAD - error in multithreading functions (only for Mt version) */ -typedef void * CLzma2EncHandle; +typedef struct CLzma2Enc CLzma2Enc; +typedef CLzma2Enc * CLzma2EncHandle; +// Z7_DECLARE_HANDLE(CLzma2EncHandle) -CLzma2EncHandle Lzma2Enc_Create(ISzAlloc *alloc, ISzAlloc *allocBig); +CLzma2EncHandle Lzma2Enc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig); void Lzma2Enc_Destroy(CLzma2EncHandle p); SRes Lzma2Enc_SetProps(CLzma2EncHandle p, const CLzma2EncProps *props); +void Lzma2Enc_SetDataSize(CLzma2EncHandle p, UInt64 expectedDataSiize); Byte Lzma2Enc_WriteProperties(CLzma2EncHandle p); -SRes Lzma2Enc_Encode(CLzma2EncHandle p, - ISeqOutStream *outStream, ISeqInStream *inStream, ICompressProgress *progress); - -/* ---------- One Call Interface ---------- */ - -/* Lzma2Encode -Return code: - SZ_OK - OK - SZ_ERROR_MEM - Memory allocation error - SZ_ERROR_PARAM - Incorrect paramater - SZ_ERROR_OUTPUT_EOF - output buffer overflow - SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version) -*/ - -/* -SRes Lzma2Encode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, - const CLzmaEncProps *props, Byte *propsEncoded, int writeEndMark, - ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig); -*/ +SRes Lzma2Enc_Encode2(CLzma2EncHandle p, + ISeqOutStreamPtr outStream, + Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + const Byte *inData, size_t inDataSize, + ICompressProgressPtr progress); EXTERN_C_END diff --git a/3rdparty/lzma/C/Lzma86.h b/3rdparty/lzma/C/Lzma86.h index bebed5cb76a..e7707e2c4ae 100644 --- a/3rdparty/lzma/C/Lzma86.h +++ b/3rdparty/lzma/C/Lzma86.h @@ -1,8 +1,8 @@ /* Lzma86.h -- LZMA + x86 (BCJ) Filter -2013-01-18 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ -#ifndef __LZMA86_H -#define __LZMA86_H +#ifndef ZIP7_INC_LZMA86_H +#define ZIP7_INC_LZMA86_H #include "7zTypes.h" diff --git a/3rdparty/lzma/C/Lzma86Dec.c b/3rdparty/lzma/C/Lzma86Dec.c index 21031745cef..f094d4c359a 100644 --- a/3rdparty/lzma/C/Lzma86Dec.c +++ b/3rdparty/lzma/C/Lzma86Dec.c @@ -1,5 +1,5 @@ /* Lzma86Dec.c -- LZMA + x86 (BCJ) Filter Decoder -2016-05-16 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -46,9 +46,8 @@ SRes Lzma86_Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen) return res; if (useFilter == 1) { - UInt32 x86State; - x86_Convert_Init(x86State); - x86_Convert(dest, *destLen, 0, &x86State, 0); + UInt32 x86State = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL; + z7_BranchConvSt_X86_Dec(dest, *destLen, 0, &x86State); } return SZ_OK; } diff --git a/3rdparty/lzma/C/Lzma86Enc.c b/3rdparty/lzma/C/Lzma86Enc.c index a15b58d6cf6..0cdde1c9a21 100644 --- a/3rdparty/lzma/C/Lzma86Enc.c +++ b/3rdparty/lzma/C/Lzma86Enc.c @@ -1,5 +1,5 @@ /* Lzma86Enc.c -- LZMA + x86 (BCJ) Filter Encoder -2016-05-16 : Igor Pavlov : Public domain */ +2023-03-03 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -11,14 +11,12 @@ #include "Bra.h" #include "LzmaEnc.h" -#define SZE_OUT_OVERFLOW SZE_DATA_ERROR - int Lzma86_Encode(Byte *dest, size_t *destLen, const Byte *src, size_t srcLen, int level, UInt32 dictSize, int filterMode) { size_t outSize2 = *destLen; Byte *filteredStream; - Bool useFilter; + BoolInt useFilter; int mainResult = SZ_ERROR_OUTPUT_EOF; CLzmaEncProps props; LzmaEncProps_Init(&props); @@ -48,15 +46,14 @@ int Lzma86_Encode(Byte *dest, size_t *destLen, const Byte *src, size_t srcLen, memcpy(filteredStream, src, srcLen); } { - UInt32 x86State; - x86_Convert_Init(x86State); - x86_Convert(filteredStream, srcLen, 0, &x86State, 1); + UInt32 x86State = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL; + z7_BranchConvSt_X86_Enc(filteredStream, srcLen, 0, &x86State); } } { size_t minSize = 0; - Bool bestIsFiltered = False; + BoolInt bestIsFiltered = False; /* passes for SZ_FILTER_AUTO: 0 - BCJ + LZMA @@ -71,7 +68,7 @@ int Lzma86_Encode(Byte *dest, size_t *destLen, const Byte *src, size_t srcLen, size_t outSizeProcessed = outSize2 - LZMA86_HEADER_SIZE; size_t outPropsSize = 5; SRes curRes; - Bool curModeIsFiltered = (numPasses > 1 && i == numPasses - 1); + BoolInt curModeIsFiltered = (numPasses > 1 && i == numPasses - 1); if (curModeIsFiltered && !bestIsFiltered) break; if (useFilter && i == 0) diff --git a/3rdparty/lzma/C/LzmaDec.c b/3rdparty/lzma/C/LzmaDec.c index 12dce11fbd5..69bb8bba9d4 100644 --- a/3rdparty/lzma/C/LzmaDec.c +++ b/3rdparty/lzma/C/LzmaDec.c @@ -1,90 +1,107 @@ /* LzmaDec.c -- LZMA Decoder -2016-05-16 : Igor Pavlov : Public domain */ +2023-04-07 : Igor Pavlov : Public domain */ #include "Precomp.h" -#include "LzmaDec.h" - #include <string.h> -#define kNumTopBits 24 -#define kTopValue ((UInt32)1 << kNumTopBits) +/* #include "CpuArch.h" */ +#include "LzmaDec.h" + +// #define kNumTopBits 24 +#define kTopValue ((UInt32)1 << 24) #define kNumBitModelTotalBits 11 #define kBitModelTotal (1 << kNumBitModelTotalBits) -#define kNumMoveBits 5 #define RC_INIT_SIZE 5 +#ifndef Z7_LZMA_DEC_OPT + +#define kNumMoveBits 5 #define NORMALIZE if (range < kTopValue) { range <<= 8; code = (code << 8) | (*buf++); } -#define IF_BIT_0(p) ttt = *(p); NORMALIZE; bound = (range >> kNumBitModelTotalBits) * ttt; if (code < bound) +#define IF_BIT_0(p) ttt = *(p); NORMALIZE; bound = (range >> kNumBitModelTotalBits) * (UInt32)ttt; if (code < bound) #define UPDATE_0(p) range = bound; *(p) = (CLzmaProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); #define UPDATE_1(p) range -= bound; code -= bound; *(p) = (CLzmaProb)(ttt - (ttt >> kNumMoveBits)); #define GET_BIT2(p, i, A0, A1) IF_BIT_0(p) \ - { UPDATE_0(p); i = (i + i); A0; } else \ - { UPDATE_1(p); i = (i + i) + 1; A1; } -#define GET_BIT(p, i) GET_BIT2(p, i, ; , ;) + { UPDATE_0(p) i = (i + i); A0; } else \ + { UPDATE_1(p) i = (i + i) + 1; A1; } + +#define TREE_GET_BIT(probs, i) { GET_BIT2(probs + i, i, ;, ;); } + +#define REV_BIT(p, i, A0, A1) IF_BIT_0(p + i) \ + { UPDATE_0(p + i) A0; } else \ + { UPDATE_1(p + i) A1; } +#define REV_BIT_VAR( p, i, m) REV_BIT(p, i, i += m; m += m, m += m; i += m; ) +#define REV_BIT_CONST(p, i, m) REV_BIT(p, i, i += m; , i += m * 2; ) +#define REV_BIT_LAST( p, i, m) REV_BIT(p, i, i -= m , ; ) -#define TREE_GET_BIT(probs, i) { GET_BIT((probs + i), i); } #define TREE_DECODE(probs, limit, i) \ { i = 1; do { TREE_GET_BIT(probs, i); } while (i < limit); i -= limit; } -/* #define _LZMA_SIZE_OPT */ +/* #define Z7_LZMA_SIZE_OPT */ -#ifdef _LZMA_SIZE_OPT +#ifdef Z7_LZMA_SIZE_OPT #define TREE_6_DECODE(probs, i) TREE_DECODE(probs, (1 << 6), i) #else #define TREE_6_DECODE(probs, i) \ { i = 1; \ - TREE_GET_BIT(probs, i); \ - TREE_GET_BIT(probs, i); \ - TREE_GET_BIT(probs, i); \ - TREE_GET_BIT(probs, i); \ - TREE_GET_BIT(probs, i); \ - TREE_GET_BIT(probs, i); \ + TREE_GET_BIT(probs, i) \ + TREE_GET_BIT(probs, i) \ + TREE_GET_BIT(probs, i) \ + TREE_GET_BIT(probs, i) \ + TREE_GET_BIT(probs, i) \ + TREE_GET_BIT(probs, i) \ i -= 0x40; } #endif -#define NORMAL_LITER_DEC GET_BIT(prob + symbol, symbol) +#define NORMAL_LITER_DEC TREE_GET_BIT(prob, symbol) #define MATCHED_LITER_DEC \ - matchByte <<= 1; \ - bit = (matchByte & offs); \ - probLit = prob + offs + bit + symbol; \ - GET_BIT2(probLit, symbol, offs &= ~bit, offs &= bit) + matchByte += matchByte; \ + bit = offs; \ + offs &= matchByte; \ + probLit = prob + (offs + bit + symbol); \ + GET_BIT2(probLit, symbol, offs ^= bit; , ;) + +#endif // Z7_LZMA_DEC_OPT + -#define NORMALIZE_CHECK if (range < kTopValue) { if (buf >= bufLimit) return DUMMY_ERROR; range <<= 8; code = (code << 8) | (*buf++); } +#define NORMALIZE_CHECK if (range < kTopValue) { if (buf >= bufLimit) return DUMMY_INPUT_EOF; range <<= 8; code = (code << 8) | (*buf++); } -#define IF_BIT_0_CHECK(p) ttt = *(p); NORMALIZE_CHECK; bound = (range >> kNumBitModelTotalBits) * ttt; if (code < bound) +#define IF_BIT_0_CHECK(p) ttt = *(p); NORMALIZE_CHECK bound = (range >> kNumBitModelTotalBits) * (UInt32)ttt; if (code < bound) #define UPDATE_0_CHECK range = bound; #define UPDATE_1_CHECK range -= bound; code -= bound; #define GET_BIT2_CHECK(p, i, A0, A1) IF_BIT_0_CHECK(p) \ - { UPDATE_0_CHECK; i = (i + i); A0; } else \ - { UPDATE_1_CHECK; i = (i + i) + 1; A1; } + { UPDATE_0_CHECK i = (i + i); A0; } else \ + { UPDATE_1_CHECK i = (i + i) + 1; A1; } #define GET_BIT_CHECK(p, i) GET_BIT2_CHECK(p, i, ; , ;) #define TREE_DECODE_CHECK(probs, limit, i) \ { i = 1; do { GET_BIT_CHECK(probs + i, i) } while (i < limit); i -= limit; } +#define REV_BIT_CHECK(p, i, m) IF_BIT_0_CHECK(p + i) \ + { UPDATE_0_CHECK i += m; m += m; } else \ + { UPDATE_1_CHECK m += m; i += m; } + + #define kNumPosBitsMax 4 #define kNumPosStatesMax (1 << kNumPosBitsMax) #define kLenNumLowBits 3 #define kLenNumLowSymbols (1 << kLenNumLowBits) -#define kLenNumMidBits 3 -#define kLenNumMidSymbols (1 << kLenNumMidBits) #define kLenNumHighBits 8 #define kLenNumHighSymbols (1 << kLenNumHighBits) -#define LenChoice 0 -#define LenChoice2 (LenChoice + 1) -#define LenLow (LenChoice2 + 1) -#define LenMid (LenLow + (kNumPosStatesMax << kLenNumLowBits)) -#define LenHigh (LenMid + (kNumPosStatesMax << kLenNumMidBits)) +#define LenLow 0 +#define LenHigh (LenLow + 2 * (kNumPosStatesMax << kLenNumLowBits)) #define kNumLenProbs (LenHigh + kLenNumHighSymbols) +#define LenChoice LenLow +#define LenChoice2 (LenLow + (1 << kLenNumLowBits)) #define kNumStates 12 +#define kNumStates2 16 #define kNumLitStates 7 #define kStartPosModelIndex 4 @@ -98,54 +115,130 @@ #define kAlignTableSize (1 << kNumAlignBits) #define kMatchMinLen 2 -#define kMatchSpecLenStart (kMatchMinLen + kLenNumLowSymbols + kLenNumMidSymbols + kLenNumHighSymbols) +#define kMatchSpecLenStart (kMatchMinLen + kLenNumLowSymbols * 2 + kLenNumHighSymbols) + +#define kMatchSpecLen_Error_Data (1 << 9) +#define kMatchSpecLen_Error_Fail (kMatchSpecLen_Error_Data - 1) -#define IsMatch 0 -#define IsRep (IsMatch + (kNumStates << kNumPosBitsMax)) +/* External ASM code needs same CLzmaProb array layout. So don't change it. */ + +/* (probs_1664) is faster and better for code size at some platforms */ +/* +#ifdef MY_CPU_X86_OR_AMD64 +*/ +#define kStartOffset 1664 +#define GET_PROBS p->probs_1664 +/* +#define GET_PROBS p->probs + kStartOffset +#else +#define kStartOffset 0 +#define GET_PROBS p->probs +#endif +*/ + +#define SpecPos (-kStartOffset) +#define IsRep0Long (SpecPos + kNumFullDistances) +#define RepLenCoder (IsRep0Long + (kNumStates2 << kNumPosBitsMax)) +#define LenCoder (RepLenCoder + kNumLenProbs) +#define IsMatch (LenCoder + kNumLenProbs) +#define Align (IsMatch + (kNumStates2 << kNumPosBitsMax)) +#define IsRep (Align + kAlignTableSize) #define IsRepG0 (IsRep + kNumStates) #define IsRepG1 (IsRepG0 + kNumStates) #define IsRepG2 (IsRepG1 + kNumStates) -#define IsRep0Long (IsRepG2 + kNumStates) -#define PosSlot (IsRep0Long + (kNumStates << kNumPosBitsMax)) -#define SpecPos (PosSlot + (kNumLenToPosStates << kNumPosSlotBits)) -#define Align (SpecPos + kNumFullDistances - kEndPosModelIndex) -#define LenCoder (Align + kAlignTableSize) -#define RepLenCoder (LenCoder + kNumLenProbs) -#define Literal (RepLenCoder + kNumLenProbs) - -#define LZMA_BASE_SIZE 1846 -#define LZMA_LIT_SIZE 0x300 +#define PosSlot (IsRepG2 + kNumStates) +#define Literal (PosSlot + (kNumLenToPosStates << kNumPosSlotBits)) +#define NUM_BASE_PROBS (Literal + kStartOffset) -#if Literal != LZMA_BASE_SIZE -StopCompilingDueBUG +#if Align != 0 && kStartOffset != 0 + #error Stop_Compiling_Bad_LZMA_kAlign #endif -#define LzmaProps_GetNumProbs(p) (Literal + ((UInt32)LZMA_LIT_SIZE << ((p)->lc + (p)->lp))) +#if NUM_BASE_PROBS != 1984 + #error Stop_Compiling_Bad_LZMA_PROBS +#endif + + +#define LZMA_LIT_SIZE 0x300 + +#define LzmaProps_GetNumProbs(p) (NUM_BASE_PROBS + ((UInt32)LZMA_LIT_SIZE << ((p)->lc + (p)->lp))) + + +#define CALC_POS_STATE(processedPos, pbMask) (((processedPos) & (pbMask)) << 4) +#define COMBINED_PS_STATE (posState + state) +#define GET_LEN_STATE (posState) #define LZMA_DIC_MIN (1 << 12) -/* First LZMA-symbol is always decoded. -And it decodes new LZMA-symbols while (buf < bufLimit), but "buf" is without last normalization +/* +p->remainLen : shows status of LZMA decoder: + < kMatchSpecLenStart : the number of bytes to be copied with (p->rep0) offset + = kMatchSpecLenStart : the LZMA stream was finished with end mark + = kMatchSpecLenStart + 1 : need init range coder + = kMatchSpecLenStart + 2 : need init range coder and state + = kMatchSpecLen_Error_Fail : Internal Code Failure + = kMatchSpecLen_Error_Data + [0 ... 273] : LZMA Data Error +*/ + +/* ---------- LZMA_DECODE_REAL ---------- */ +/* +LzmaDec_DecodeReal_3() can be implemented in external ASM file. +3 - is the code compatibility version of that function for check at link time. +*/ + +#define LZMA_DECODE_REAL LzmaDec_DecodeReal_3 + +/* +LZMA_DECODE_REAL() +In: + RangeCoder is normalized + if (p->dicPos == limit) + { + LzmaDec_TryDummy() was called before to exclude LITERAL and MATCH-REP cases. + So first symbol can be only MATCH-NON-REP. And if that MATCH-NON-REP symbol + is not END_OF_PAYALOAD_MARKER, then the function doesn't write any byte to dictionary, + the function returns SZ_OK, and the caller can use (p->remainLen) and (p->reps[0]) later. + } + +Processing: + The first LZMA symbol will be decoded in any case. + All main checks for limits are at the end of main loop, + It decodes additional LZMA-symbols while (p->buf < bufLimit && dicPos < limit), + RangeCoder is still without last normalization when (p->buf < bufLimit) is being checked. + But if (p->buf < bufLimit), the caller provided at least (LZMA_REQUIRED_INPUT_MAX + 1) bytes for + next iteration before limit (bufLimit + LZMA_REQUIRED_INPUT_MAX), + that is enough for worst case LZMA symbol with one additional RangeCoder normalization for one bit. + So that function never reads bufLimit [LZMA_REQUIRED_INPUT_MAX] byte. + Out: + RangeCoder is normalized Result: SZ_OK - OK - SZ_ERROR_DATA - Error - p->remainLen: - < kMatchSpecLenStart : normal remain - = kMatchSpecLenStart : finished - = kMatchSpecLenStart + 1 : Flush marker (unused now) - = kMatchSpecLenStart + 2 : State Init Marker (unused now) + p->remainLen: + < kMatchSpecLenStart : the number of bytes to be copied with (p->reps[0]) offset + = kMatchSpecLenStart : the LZMA stream was finished with end mark + + SZ_ERROR_DATA - error, when the MATCH-Symbol refers out of dictionary + p->remainLen : undefined + p->reps[*] : undefined */ -static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte *bufLimit) -{ - CLzmaProb *probs = p->probs; - unsigned state = p->state; +#ifdef Z7_LZMA_DEC_OPT + +int Z7_FASTCALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit); + +#else + +static +int Z7_FASTCALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit) +{ + CLzmaProb *probs = GET_PROBS; + unsigned state = (unsigned)p->state; UInt32 rep0 = p->reps[0], rep1 = p->reps[1], rep2 = p->reps[2], rep3 = p->reps[3]; unsigned pbMask = ((unsigned)1 << (p->prop.pb)) - 1; - unsigned lpMask = ((unsigned)1 << (p->prop.lp)) - 1; unsigned lc = p->prop.lc; + unsigned lpMask = ((unsigned)0x100 << p->prop.lp) - ((unsigned)0x100 >> lc); Byte *dic = p->dic; SizeT dicBufSize = p->dicBufSize; @@ -164,24 +257,23 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte CLzmaProb *prob; UInt32 bound; unsigned ttt; - unsigned posState = processedPos & pbMask; + unsigned posState = CALC_POS_STATE(processedPos, pbMask); - prob = probs + IsMatch + (state << kNumPosBitsMax) + posState; + prob = probs + IsMatch + COMBINED_PS_STATE; IF_BIT_0(prob) { unsigned symbol; - UPDATE_0(prob); + UPDATE_0(prob) prob = probs + Literal; if (processedPos != 0 || checkDicSize != 0) - prob += ((UInt32)LZMA_LIT_SIZE * (((processedPos & lpMask) << lc) + - (dic[(dicPos == 0 ? dicBufSize : dicPos) - 1] >> (8 - lc)))); + prob += (UInt32)3 * ((((processedPos << 8) + dic[(dicPos == 0 ? dicBufSize : dicPos) - 1]) & lpMask) << lc); processedPos++; if (state < kNumLitStates) { state -= (state < 4) ? state : 3; symbol = 1; - #ifdef _LZMA_SIZE_OPT + #ifdef Z7_LZMA_SIZE_OPT do { NORMAL_LITER_DEC } while (symbol < 0x100); #else NORMAL_LITER_DEC @@ -200,7 +292,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte unsigned offs = 0x100; state -= (state < 10) ? 3 : 6; symbol = 1; - #ifdef _LZMA_SIZE_OPT + #ifdef Z7_LZMA_SIZE_OPT do { unsigned bit; @@ -229,57 +321,62 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte } { - UPDATE_1(prob); + UPDATE_1(prob) prob = probs + IsRep + state; IF_BIT_0(prob) { - UPDATE_0(prob); + UPDATE_0(prob) state += kNumStates; prob = probs + LenCoder; } else { - UPDATE_1(prob); - if (checkDicSize == 0 && processedPos == 0) - return SZ_ERROR_DATA; + UPDATE_1(prob) prob = probs + IsRepG0 + state; IF_BIT_0(prob) { - UPDATE_0(prob); - prob = probs + IsRep0Long + (state << kNumPosBitsMax) + posState; + UPDATE_0(prob) + prob = probs + IsRep0Long + COMBINED_PS_STATE; IF_BIT_0(prob) { - UPDATE_0(prob); + UPDATE_0(prob) + + // that case was checked before with kBadRepCode + // if (checkDicSize == 0 && processedPos == 0) { len = kMatchSpecLen_Error_Data + 1; break; } + // The caller doesn't allow (dicPos == limit) case here + // so we don't need the following check: + // if (dicPos == limit) { state = state < kNumLitStates ? 9 : 11; len = 1; break; } + dic[dicPos] = dic[dicPos - rep0 + (dicPos < rep0 ? dicBufSize : 0)]; dicPos++; processedPos++; state = state < kNumLitStates ? 9 : 11; continue; } - UPDATE_1(prob); + UPDATE_1(prob) } else { UInt32 distance; - UPDATE_1(prob); + UPDATE_1(prob) prob = probs + IsRepG1 + state; IF_BIT_0(prob) { - UPDATE_0(prob); + UPDATE_0(prob) distance = rep1; } else { - UPDATE_1(prob); + UPDATE_1(prob) prob = probs + IsRepG2 + state; IF_BIT_0(prob) { - UPDATE_0(prob); + UPDATE_0(prob) distance = rep2; } else { - UPDATE_1(prob); + UPDATE_1(prob) distance = rep3; rep3 = rep2; } @@ -292,37 +389,37 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte prob = probs + RepLenCoder; } - #ifdef _LZMA_SIZE_OPT + #ifdef Z7_LZMA_SIZE_OPT { unsigned lim, offset; CLzmaProb *probLen = prob + LenChoice; IF_BIT_0(probLen) { - UPDATE_0(probLen); - probLen = prob + LenLow + (posState << kLenNumLowBits); + UPDATE_0(probLen) + probLen = prob + LenLow + GET_LEN_STATE; offset = 0; lim = (1 << kLenNumLowBits); } else { - UPDATE_1(probLen); + UPDATE_1(probLen) probLen = prob + LenChoice2; IF_BIT_0(probLen) { - UPDATE_0(probLen); - probLen = prob + LenMid + (posState << kLenNumMidBits); + UPDATE_0(probLen) + probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits); offset = kLenNumLowSymbols; - lim = (1 << kLenNumMidBits); + lim = (1 << kLenNumLowBits); } else { - UPDATE_1(probLen); + UPDATE_1(probLen) probLen = prob + LenHigh; - offset = kLenNumLowSymbols + kLenNumMidSymbols; + offset = kLenNumLowSymbols * 2; lim = (1 << kLenNumHighBits); } } - TREE_DECODE(probLen, lim, len); + TREE_DECODE(probLen, lim, len) len += offset; } #else @@ -330,33 +427,33 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte CLzmaProb *probLen = prob + LenChoice; IF_BIT_0(probLen) { - UPDATE_0(probLen); - probLen = prob + LenLow + (posState << kLenNumLowBits); + UPDATE_0(probLen) + probLen = prob + LenLow + GET_LEN_STATE; len = 1; - TREE_GET_BIT(probLen, len); - TREE_GET_BIT(probLen, len); - TREE_GET_BIT(probLen, len); + TREE_GET_BIT(probLen, len) + TREE_GET_BIT(probLen, len) + TREE_GET_BIT(probLen, len) len -= 8; } else { - UPDATE_1(probLen); + UPDATE_1(probLen) probLen = prob + LenChoice2; IF_BIT_0(probLen) { - UPDATE_0(probLen); - probLen = prob + LenMid + (posState << kLenNumMidBits); + UPDATE_0(probLen) + probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits); len = 1; - TREE_GET_BIT(probLen, len); - TREE_GET_BIT(probLen, len); - TREE_GET_BIT(probLen, len); + TREE_GET_BIT(probLen, len) + TREE_GET_BIT(probLen, len) + TREE_GET_BIT(probLen, len) } else { - UPDATE_1(probLen); + UPDATE_1(probLen) probLen = prob + LenHigh; - TREE_DECODE(probLen, (1 << kLenNumHighBits), len); - len += kLenNumLowSymbols + kLenNumMidSymbols; + TREE_DECODE(probLen, (1 << kLenNumHighBits), len) + len += kLenNumLowSymbols * 2; } } } @@ -367,7 +464,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte UInt32 distance; prob = probs + PosSlot + ((len < kNumLenToPosStates ? len : kNumLenToPosStates - 1) << kNumPosSlotBits); - TREE_6_DECODE(prob, distance); + TREE_6_DECODE(prob, distance) if (distance >= kStartPosModelIndex) { unsigned posSlot = (unsigned)distance; @@ -376,16 +473,16 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte if (posSlot < kEndPosModelIndex) { distance <<= numDirectBits; - prob = probs + SpecPos + distance - posSlot - 1; + prob = probs + SpecPos; { - UInt32 mask = 1; - unsigned i = 1; + UInt32 m = 1; + distance++; do { - GET_BIT2(prob + i, i, ; , distance |= mask); - mask <<= 1; + REV_BIT_VAR(prob, distance, m) } - while (--numDirectBits != 0); + while (--numDirectBits); + distance -= m; } } else @@ -412,19 +509,20 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte } */ } - while (--numDirectBits != 0); + while (--numDirectBits); prob = probs + Align; distance <<= kNumAlignBits; { unsigned i = 1; - GET_BIT2(prob + i, i, ; , distance |= 1); - GET_BIT2(prob + i, i, ; , distance |= 2); - GET_BIT2(prob + i, i, ; , distance |= 4); - GET_BIT2(prob + i, i, ; , distance |= 8); + REV_BIT_CONST(prob, i, 1) + REV_BIT_CONST(prob, i, 2) + REV_BIT_CONST(prob, i, 4) + REV_BIT_LAST (prob, i, 8) + distance |= i; } if (distance == (UInt32)0xFFFFFFFF) { - len += kMatchSpecLenStart; + len = kMatchSpecLenStart; state -= kNumStates; break; } @@ -435,20 +533,14 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte rep2 = rep1; rep1 = rep0; rep0 = distance + 1; - if (checkDicSize == 0) - { - if (distance >= processedPos) - { - p->dicPos = dicPos; - return SZ_ERROR_DATA; - } - } - else if (distance >= checkDicSize) + state = (state < kNumStates + kNumLitStates) ? kNumLitStates : kNumLitStates + 3; + if (distance >= (checkDicSize == 0 ? processedPos: checkDicSize)) { - p->dicPos = dicPos; - return SZ_ERROR_DATA; + len += kMatchSpecLen_Error_Data + kMatchMinLen; + // len = kMatchSpecLen_Error_Data; + // len += kMatchMinLen; + break; } - state = (state < kNumStates + kNumLitStates) ? kNumLitStates : kNumLitStates + 3; } len += kMatchMinLen; @@ -460,14 +552,19 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte if ((rem = limit - dicPos) == 0) { - p->dicPos = dicPos; - return SZ_ERROR_DATA; + /* + We stop decoding and return SZ_OK, and we can resume decoding later. + Any error conditions can be tested later in caller code. + For more strict mode we can stop decoding with error + // len += kMatchSpecLen_Error_Data; + */ + break; } curLen = ((rem < len) ? (unsigned)rem : len); pos = dicPos - rep0 + (dicPos < rep0 ? dicBufSize : 0); - processedPos += curLen; + processedPos += (UInt32)curLen; len -= curLen; if (curLen <= dicBufSize - pos) @@ -475,7 +572,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte Byte *dest = dic + dicPos; ptrdiff_t src = (ptrdiff_t)pos - (ptrdiff_t)dicPos; const Byte *lim = dest + curLen; - dicPos += curLen; + dicPos += (SizeT)curLen; do *(dest) = (Byte)*(dest + src); while (++dest != lim); @@ -495,113 +592,152 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte } while (dicPos < limit && buf < bufLimit); - NORMALIZE; + NORMALIZE p->buf = buf; p->range = range; p->code = code; - p->remainLen = len; + p->remainLen = (UInt32)len; // & (kMatchSpecLen_Error_Data - 1); // we can write real length for error matches too. p->dicPos = dicPos; p->processedPos = processedPos; p->reps[0] = rep0; p->reps[1] = rep1; p->reps[2] = rep2; p->reps[3] = rep3; - p->state = state; - + p->state = (UInt32)state; + if (len >= kMatchSpecLen_Error_Data) + return SZ_ERROR_DATA; return SZ_OK; } +#endif + + -static void MY_FAST_CALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit) +static void Z7_FASTCALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit) { - if (p->remainLen != 0 && p->remainLen < kMatchSpecLenStart) + unsigned len = (unsigned)p->remainLen; + if (len == 0 /* || len >= kMatchSpecLenStart */) + return; { - Byte *dic = p->dic; SizeT dicPos = p->dicPos; - SizeT dicBufSize = p->dicBufSize; - unsigned len = p->remainLen; - SizeT rep0 = p->reps[0]; /* we use SizeT to avoid the BUG of VC14 for AMD64 */ - SizeT rem = limit - dicPos; - if (rem < len) - len = (unsigned)(rem); + Byte *dic; + SizeT dicBufSize; + SizeT rep0; /* we use SizeT to avoid the BUG of VC14 for AMD64 */ + { + SizeT rem = limit - dicPos; + if (rem < len) + { + len = (unsigned)(rem); + if (len == 0) + return; + } + } if (p->checkDicSize == 0 && p->prop.dicSize - p->processedPos <= len) p->checkDicSize = p->prop.dicSize; - p->processedPos += len; - p->remainLen -= len; - while (len != 0) + p->processedPos += (UInt32)len; + p->remainLen -= (UInt32)len; + dic = p->dic; + rep0 = p->reps[0]; + dicBufSize = p->dicBufSize; + do { - len--; dic[dicPos] = dic[dicPos - rep0 + (dicPos < rep0 ? dicBufSize : 0)]; dicPos++; } + while (--len); p->dicPos = dicPos; } } -static int MY_FAST_CALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte *bufLimit) + +/* +At staring of new stream we have one of the following symbols: + - Literal - is allowed + - Non-Rep-Match - is allowed only if it's end marker symbol + - Rep-Match - is not allowed +We use early check of (RangeCoder:Code) over kBadRepCode to simplify main decoding code +*/ + +#define kRange0 0xFFFFFFFF +#define kBound0 ((kRange0 >> kNumBitModelTotalBits) << (kNumBitModelTotalBits - 1)) +#define kBadRepCode (kBound0 + (((kRange0 - kBound0) >> kNumBitModelTotalBits) << (kNumBitModelTotalBits - 1))) +#if kBadRepCode != (0xC0000000 - 0x400) + #error Stop_Compiling_Bad_LZMA_Check +#endif + + +/* +LzmaDec_DecodeReal2(): + It calls LZMA_DECODE_REAL() and it adjusts limit according (p->checkDicSize). + +We correct (p->checkDicSize) after LZMA_DECODE_REAL() and in LzmaDec_WriteRem(), +and we support the following state of (p->checkDicSize): + if (total_processed < p->prop.dicSize) then + { + (total_processed == p->processedPos) + (p->checkDicSize == 0) + } + else + (p->checkDicSize == p->prop.dicSize) +*/ + +static int Z7_FASTCALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte *bufLimit) { - do + if (p->checkDicSize == 0) { - SizeT limit2 = limit; - if (p->checkDicSize == 0) - { - UInt32 rem = p->prop.dicSize - p->processedPos; - if (limit - p->dicPos > rem) - limit2 = p->dicPos + rem; - } - - RINOK(LzmaDec_DecodeReal(p, limit2, bufLimit)); - + UInt32 rem = p->prop.dicSize - p->processedPos; + if (limit - p->dicPos > rem) + limit = p->dicPos + rem; + } + { + int res = LZMA_DECODE_REAL(p, limit, bufLimit); if (p->checkDicSize == 0 && p->processedPos >= p->prop.dicSize) p->checkDicSize = p->prop.dicSize; - - LzmaDec_WriteRem(p, limit); + return res; } - while (p->dicPos < limit && p->buf < bufLimit && p->remainLen < kMatchSpecLenStart); +} - if (p->remainLen > kMatchSpecLenStart) - p->remainLen = kMatchSpecLenStart; - return 0; -} typedef enum { - DUMMY_ERROR, /* unexpected end of input stream */ + DUMMY_INPUT_EOF, /* need more input data */ DUMMY_LIT, DUMMY_MATCH, DUMMY_REP } ELzmaDummy; -static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inSize) + +#define IS_DUMMY_END_MARKER_POSSIBLE(dummyRes) ((dummyRes) == DUMMY_MATCH) + +static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byte **bufOut) { UInt32 range = p->range; UInt32 code = p->code; - const Byte *bufLimit = buf + inSize; - const CLzmaProb *probs = p->probs; - unsigned state = p->state; + const Byte *bufLimit = *bufOut; + const CLzmaProb *probs = GET_PROBS; + unsigned state = (unsigned)p->state; ELzmaDummy res; + for (;;) { const CLzmaProb *prob; UInt32 bound; unsigned ttt; - unsigned posState = (p->processedPos) & ((1 << p->prop.pb) - 1); + unsigned posState = CALC_POS_STATE(p->processedPos, ((unsigned)1 << p->prop.pb) - 1); - prob = probs + IsMatch + (state << kNumPosBitsMax) + posState; + prob = probs + IsMatch + COMBINED_PS_STATE; IF_BIT_0_CHECK(prob) { UPDATE_0_CHECK - /* if (bufLimit - buf >= 7) return DUMMY_LIT; */ - prob = probs + Literal; if (p->checkDicSize != 0 || p->processedPos != 0) prob += ((UInt32)LZMA_LIT_SIZE * - ((((p->processedPos) & ((1 << (p->prop.lp)) - 1)) << p->prop.lc) + - (p->dic[(p->dicPos == 0 ? p->dicBufSize : p->dicPos) - 1] >> (8 - p->prop.lc)))); + ((((p->processedPos) & (((unsigned)1 << (p->prop.lp)) - 1)) << p->prop.lc) + + ((unsigned)p->dic[(p->dicPos == 0 ? p->dicBufSize : p->dicPos) - 1] >> (8 - p->prop.lc)))); if (state < kNumLitStates) { @@ -618,10 +754,11 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS { unsigned bit; const CLzmaProb *probLit; - matchByte <<= 1; - bit = (matchByte & offs); - probLit = prob + offs + bit + symbol; - GET_BIT2_CHECK(probLit, symbol, offs &= ~bit, offs &= bit) + matchByte += matchByte; + bit = offs; + offs &= matchByte; + probLit = prob + (offs + bit + symbol); + GET_BIT2_CHECK(probLit, symbol, offs ^= bit; , ; ) } while (symbol < 0x100); } @@ -630,55 +767,54 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS else { unsigned len; - UPDATE_1_CHECK; + UPDATE_1_CHECK prob = probs + IsRep + state; IF_BIT_0_CHECK(prob) { - UPDATE_0_CHECK; + UPDATE_0_CHECK state = 0; prob = probs + LenCoder; res = DUMMY_MATCH; } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK res = DUMMY_REP; prob = probs + IsRepG0 + state; IF_BIT_0_CHECK(prob) { - UPDATE_0_CHECK; - prob = probs + IsRep0Long + (state << kNumPosBitsMax) + posState; + UPDATE_0_CHECK + prob = probs + IsRep0Long + COMBINED_PS_STATE; IF_BIT_0_CHECK(prob) { - UPDATE_0_CHECK; - NORMALIZE_CHECK; - return DUMMY_REP; + UPDATE_0_CHECK + break; } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK } } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK prob = probs + IsRepG1 + state; IF_BIT_0_CHECK(prob) { - UPDATE_0_CHECK; + UPDATE_0_CHECK } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK prob = probs + IsRepG2 + state; IF_BIT_0_CHECK(prob) { - UPDATE_0_CHECK; + UPDATE_0_CHECK } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK } } } @@ -690,31 +826,31 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS const CLzmaProb *probLen = prob + LenChoice; IF_BIT_0_CHECK(probLen) { - UPDATE_0_CHECK; - probLen = prob + LenLow + (posState << kLenNumLowBits); + UPDATE_0_CHECK + probLen = prob + LenLow + GET_LEN_STATE; offset = 0; limit = 1 << kLenNumLowBits; } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK probLen = prob + LenChoice2; IF_BIT_0_CHECK(probLen) { - UPDATE_0_CHECK; - probLen = prob + LenMid + (posState << kLenNumMidBits); + UPDATE_0_CHECK + probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits); offset = kLenNumLowSymbols; - limit = 1 << kLenNumMidBits; + limit = 1 << kLenNumLowBits; } else { - UPDATE_1_CHECK; + UPDATE_1_CHECK probLen = prob + LenHigh; - offset = kLenNumLowSymbols + kLenNumMidSymbols; + offset = kLenNumLowSymbols * 2; limit = 1 << kLenNumHighBits; } } - TREE_DECODE_CHECK(probLen, limit, len); + TREE_DECODE_CHECK(probLen, limit, len) len += offset; } @@ -722,18 +858,16 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS { unsigned posSlot; prob = probs + PosSlot + - ((len < kNumLenToPosStates ? len : kNumLenToPosStates - 1) << + ((len < kNumLenToPosStates - 1 ? len : kNumLenToPosStates - 1) << kNumPosSlotBits); - TREE_DECODE_CHECK(prob, 1 << kNumPosSlotBits, posSlot); + TREE_DECODE_CHECK(prob, 1 << kNumPosSlotBits, posSlot) if (posSlot >= kStartPosModelIndex) { unsigned numDirectBits = ((posSlot >> 1) - 1); - /* if (bufLimit - buf >= 8) return DUMMY_MATCH; */ - if (posSlot < kEndPosModelIndex) { - prob = probs + SpecPos + ((2 | (posSlot & 1)) << numDirectBits) - posSlot - 1; + prob = probs + SpecPos + ((2 | (posSlot & 1)) << numDirectBits); } else { @@ -745,41 +879,44 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS code -= range & (((code - range) >> 31) - 1); /* if (code >= range) code -= range; */ } - while (--numDirectBits != 0); + while (--numDirectBits); prob = probs + Align; numDirectBits = kNumAlignBits; } { unsigned i = 1; + unsigned m = 1; do { - GET_BIT_CHECK(prob + i, i); + REV_BIT_CHECK(prob, i, m) } - while (--numDirectBits != 0); + while (--numDirectBits); } } } } + break; } - NORMALIZE_CHECK; + NORMALIZE_CHECK + + *bufOut = buf; return res; } - -void LzmaDec_InitDicAndState(CLzmaDec *p, Bool initDic, Bool initState) +void LzmaDec_InitDicAndState(CLzmaDec *p, BoolInt initDic, BoolInt initState); +void LzmaDec_InitDicAndState(CLzmaDec *p, BoolInt initDic, BoolInt initState) { - p->needFlush = 1; - p->remainLen = 0; + p->remainLen = kMatchSpecLenStart + 1; p->tempBufSize = 0; if (initDic) { p->processedPos = 0; p->checkDicSize = 0; - p->needInitState = 1; + p->remainLen = kMatchSpecLenStart + 2; } if (initState) - p->needInitState = 1; + p->remainLen = kMatchSpecLenStart + 2; } void LzmaDec_Init(CLzmaDec *p) @@ -788,53 +925,96 @@ void LzmaDec_Init(CLzmaDec *p) LzmaDec_InitDicAndState(p, True, True); } -static void LzmaDec_InitStateReal(CLzmaDec *p) -{ - SizeT numProbs = LzmaProps_GetNumProbs(&p->prop); - SizeT i; - CLzmaProb *probs = p->probs; - for (i = 0; i < numProbs; i++) - probs[i] = kBitModelTotal >> 1; - p->reps[0] = p->reps[1] = p->reps[2] = p->reps[3] = 1; - p->state = 0; - p->needInitState = 0; -} + +/* +LZMA supports optional end_marker. +So the decoder can lookahead for one additional LZMA-Symbol to check end_marker. +That additional LZMA-Symbol can require up to LZMA_REQUIRED_INPUT_MAX bytes in input stream. +When the decoder reaches dicLimit, it looks (finishMode) parameter: + if (finishMode == LZMA_FINISH_ANY), the decoder doesn't lookahead + if (finishMode != LZMA_FINISH_ANY), the decoder lookahead, if end_marker is possible for current position + +When the decoder lookahead, and the lookahead symbol is not end_marker, we have two ways: + 1) Strict mode (default) : the decoder returns SZ_ERROR_DATA. + 2) The relaxed mode (alternative mode) : we could return SZ_OK, and the caller + must check (status) value. The caller can show the error, + if the end of stream is expected, and the (status) is noit + LZMA_STATUS_FINISHED_WITH_MARK or LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK. +*/ + + +#define RETURN_NOT_FINISHED_FOR_FINISH \ + *status = LZMA_STATUS_NOT_FINISHED; \ + return SZ_ERROR_DATA; // for strict mode + // return SZ_OK; // for relaxed mode + SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status) { SizeT inSize = *srcLen; (*srcLen) = 0; - LzmaDec_WriteRem(p, dicLimit); - *status = LZMA_STATUS_NOT_SPECIFIED; - while (p->remainLen != kMatchSpecLenStart) + if (p->remainLen > kMatchSpecLenStart) { - int checkEndMarkNow; + if (p->remainLen > kMatchSpecLenStart + 2) + return p->remainLen == kMatchSpecLen_Error_Fail ? SZ_ERROR_FAIL : SZ_ERROR_DATA; + + for (; inSize > 0 && p->tempBufSize < RC_INIT_SIZE; (*srcLen)++, inSize--) + p->tempBuf[p->tempBufSize++] = *src++; + if (p->tempBufSize != 0 && p->tempBuf[0] != 0) + return SZ_ERROR_DATA; + if (p->tempBufSize < RC_INIT_SIZE) + { + *status = LZMA_STATUS_NEEDS_MORE_INPUT; + return SZ_OK; + } + p->code = + ((UInt32)p->tempBuf[1] << 24) + | ((UInt32)p->tempBuf[2] << 16) + | ((UInt32)p->tempBuf[3] << 8) + | ((UInt32)p->tempBuf[4]); - if (p->needFlush) - { - for (; inSize > 0 && p->tempBufSize < RC_INIT_SIZE; (*srcLen)++, inSize--) - p->tempBuf[p->tempBufSize++] = *src++; - if (p->tempBufSize < RC_INIT_SIZE) - { - *status = LZMA_STATUS_NEEDS_MORE_INPUT; - return SZ_OK; - } - if (p->tempBuf[0] != 0) - return SZ_ERROR_DATA; - p->code = - ((UInt32)p->tempBuf[1] << 24) - | ((UInt32)p->tempBuf[2] << 16) - | ((UInt32)p->tempBuf[3] << 8) - | ((UInt32)p->tempBuf[4]); - p->range = 0xFFFFFFFF; - p->needFlush = 0; - p->tempBufSize = 0; - } + if (p->checkDicSize == 0 + && p->processedPos == 0 + && p->code >= kBadRepCode) + return SZ_ERROR_DATA; + + p->range = 0xFFFFFFFF; + p->tempBufSize = 0; + + if (p->remainLen > kMatchSpecLenStart + 1) + { + SizeT numProbs = LzmaProps_GetNumProbs(&p->prop); + SizeT i; + CLzmaProb *probs = p->probs; + for (i = 0; i < numProbs; i++) + probs[i] = kBitModelTotal >> 1; + p->reps[0] = p->reps[1] = p->reps[2] = p->reps[3] = 1; + p->state = 0; + } + + p->remainLen = 0; + } + + for (;;) + { + if (p->remainLen == kMatchSpecLenStart) + { + if (p->code != 0) + return SZ_ERROR_DATA; + *status = LZMA_STATUS_FINISHED_WITH_MARK; + return SZ_OK; + } + + LzmaDec_WriteRem(p, dicLimit); + + { + // (p->remainLen == 0 || p->dicPos == dicLimit) + + int checkEndMarkNow = 0; - checkEndMarkNow = 0; if (p->dicPos >= dicLimit) { if (p->remainLen == 0 && p->code == 0) @@ -849,92 +1029,174 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr } if (p->remainLen != 0) { - *status = LZMA_STATUS_NOT_FINISHED; - return SZ_ERROR_DATA; + RETURN_NOT_FINISHED_FOR_FINISH } checkEndMarkNow = 1; } - if (p->needInitState) - LzmaDec_InitStateReal(p); - + // (p->remainLen == 0) + if (p->tempBufSize == 0) { - SizeT processed; const Byte *bufLimit; + int dummyProcessed = -1; + if (inSize < LZMA_REQUIRED_INPUT_MAX || checkEndMarkNow) { - int dummyRes = LzmaDec_TryDummy(p, src, inSize); - if (dummyRes == DUMMY_ERROR) + const Byte *bufOut = src + inSize; + + ELzmaDummy dummyRes = LzmaDec_TryDummy(p, src, &bufOut); + + if (dummyRes == DUMMY_INPUT_EOF) { - memcpy(p->tempBuf, src, inSize); - p->tempBufSize = (unsigned)inSize; + size_t i; + if (inSize >= LZMA_REQUIRED_INPUT_MAX) + break; (*srcLen) += inSize; + p->tempBufSize = (unsigned)inSize; + for (i = 0; i < inSize; i++) + p->tempBuf[i] = src[i]; *status = LZMA_STATUS_NEEDS_MORE_INPUT; return SZ_OK; } - if (checkEndMarkNow && dummyRes != DUMMY_MATCH) + + dummyProcessed = (int)(bufOut - src); + if ((unsigned)dummyProcessed > LZMA_REQUIRED_INPUT_MAX) + break; + + if (checkEndMarkNow && !IS_DUMMY_END_MARKER_POSSIBLE(dummyRes)) { - *status = LZMA_STATUS_NOT_FINISHED; - return SZ_ERROR_DATA; + unsigned i; + (*srcLen) += (unsigned)dummyProcessed; + p->tempBufSize = (unsigned)dummyProcessed; + for (i = 0; i < (unsigned)dummyProcessed; i++) + p->tempBuf[i] = src[i]; + // p->remainLen = kMatchSpecLen_Error_Data; + RETURN_NOT_FINISHED_FOR_FINISH } + bufLimit = src; + // we will decode only one iteration } else bufLimit = src + inSize - LZMA_REQUIRED_INPUT_MAX; + p->buf = src; - if (LzmaDec_DecodeReal2(p, dicLimit, bufLimit) != 0) - return SZ_ERROR_DATA; - processed = (SizeT)(p->buf - src); - (*srcLen) += processed; - src += processed; - inSize -= processed; + + { + int res = LzmaDec_DecodeReal2(p, dicLimit, bufLimit); + + SizeT processed = (SizeT)(p->buf - src); + + if (dummyProcessed < 0) + { + if (processed > inSize) + break; + } + else if ((unsigned)dummyProcessed != processed) + break; + + src += processed; + inSize -= processed; + (*srcLen) += processed; + + if (res != SZ_OK) + { + p->remainLen = kMatchSpecLen_Error_Data; + return SZ_ERROR_DATA; + } + } + continue; } - else + { - unsigned rem = p->tempBufSize, lookAhead = 0; - while (rem < LZMA_REQUIRED_INPUT_MAX && lookAhead < inSize) - p->tempBuf[rem++] = src[lookAhead++]; - p->tempBufSize = rem; + // we have some data in (p->tempBuf) + // in strict mode: tempBufSize is not enough for one Symbol decoding. + // in relaxed mode: tempBufSize not larger than required for one Symbol decoding. + + unsigned rem = p->tempBufSize; + unsigned ahead = 0; + int dummyProcessed = -1; + + while (rem < LZMA_REQUIRED_INPUT_MAX && ahead < inSize) + p->tempBuf[rem++] = src[ahead++]; + + // ahead - the size of new data copied from (src) to (p->tempBuf) + // rem - the size of temp buffer including new data from (src) + if (rem < LZMA_REQUIRED_INPUT_MAX || checkEndMarkNow) { - int dummyRes = LzmaDec_TryDummy(p, p->tempBuf, rem); - if (dummyRes == DUMMY_ERROR) + const Byte *bufOut = p->tempBuf + rem; + + ELzmaDummy dummyRes = LzmaDec_TryDummy(p, p->tempBuf, &bufOut); + + if (dummyRes == DUMMY_INPUT_EOF) { - (*srcLen) += lookAhead; + if (rem >= LZMA_REQUIRED_INPUT_MAX) + break; + p->tempBufSize = rem; + (*srcLen) += (SizeT)ahead; *status = LZMA_STATUS_NEEDS_MORE_INPUT; return SZ_OK; } - if (checkEndMarkNow && dummyRes != DUMMY_MATCH) + + dummyProcessed = (int)(bufOut - p->tempBuf); + + if ((unsigned)dummyProcessed < p->tempBufSize) + break; + + if (checkEndMarkNow && !IS_DUMMY_END_MARKER_POSSIBLE(dummyRes)) { - *status = LZMA_STATUS_NOT_FINISHED; - return SZ_ERROR_DATA; + (*srcLen) += (unsigned)dummyProcessed - p->tempBufSize; + p->tempBufSize = (unsigned)dummyProcessed; + // p->remainLen = kMatchSpecLen_Error_Data; + RETURN_NOT_FINISHED_FOR_FINISH } } + p->buf = p->tempBuf; - if (LzmaDec_DecodeReal2(p, dicLimit, p->buf) != 0) - return SZ_ERROR_DATA; { - unsigned kkk = (unsigned)(p->buf - p->tempBuf); - if (rem < kkk) - return SZ_ERROR_FAIL; /* some internal error */ - rem -= kkk; - if (lookAhead < rem) - return SZ_ERROR_FAIL; /* some internal error */ - lookAhead -= rem; + // we decode one symbol from (p->tempBuf) here, so the (bufLimit) is equal to (p->buf) + int res = LzmaDec_DecodeReal2(p, dicLimit, p->buf); + + SizeT processed = (SizeT)(p->buf - p->tempBuf); + rem = p->tempBufSize; + + if (dummyProcessed < 0) + { + if (processed > LZMA_REQUIRED_INPUT_MAX) + break; + if (processed < rem) + break; + } + else if ((unsigned)dummyProcessed != processed) + break; + + processed -= rem; + + src += processed; + inSize -= processed; + (*srcLen) += processed; + p->tempBufSize = 0; + + if (res != SZ_OK) + { + p->remainLen = kMatchSpecLen_Error_Data; + return SZ_ERROR_DATA; + } } - (*srcLen) += lookAhead; - src += lookAhead; - inSize -= lookAhead; - p->tempBufSize = 0; } + } } - if (p->code == 0) - *status = LZMA_STATUS_FINISHED_WITH_MARK; - return (p->code == 0) ? SZ_OK : SZ_ERROR_DATA; + + /* Some unexpected error: internal error of code, memory corruption or hardware failure */ + p->remainLen = kMatchSpecLen_Error_Fail; + return SZ_ERROR_FAIL; } + + SRes LzmaDec_DecodeToBuf(CLzmaDec *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status) { SizeT outSize = *destLen; @@ -975,19 +1237,19 @@ SRes LzmaDec_DecodeToBuf(CLzmaDec *p, Byte *dest, SizeT *destLen, const Byte *sr } } -void LzmaDec_FreeProbs(CLzmaDec *p, ISzAlloc *alloc) +void LzmaDec_FreeProbs(CLzmaDec *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->probs); + ISzAlloc_Free(alloc, p->probs); p->probs = NULL; } -static void LzmaDec_FreeDict(CLzmaDec *p, ISzAlloc *alloc) +static void LzmaDec_FreeDict(CLzmaDec *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->dic); + ISzAlloc_Free(alloc, p->dic); p->dic = NULL; } -void LzmaDec_Free(CLzmaDec *p, ISzAlloc *alloc) +void LzmaDec_Free(CLzmaDec *p, ISzAllocPtr alloc) { LzmaDec_FreeProbs(p, alloc); LzmaDec_FreeDict(p, alloc); @@ -1011,49 +1273,50 @@ SRes LzmaProps_Decode(CLzmaProps *p, const Byte *data, unsigned size) if (d >= (9 * 5 * 5)) return SZ_ERROR_UNSUPPORTED; - p->lc = d % 9; + p->lc = (Byte)(d % 9); d /= 9; - p->pb = d / 5; - p->lp = d % 5; + p->pb = (Byte)(d / 5); + p->lp = (Byte)(d % 5); return SZ_OK; } -static SRes LzmaDec_AllocateProbs2(CLzmaDec *p, const CLzmaProps *propNew, ISzAlloc *alloc) +static SRes LzmaDec_AllocateProbs2(CLzmaDec *p, const CLzmaProps *propNew, ISzAllocPtr alloc) { UInt32 numProbs = LzmaProps_GetNumProbs(propNew); if (!p->probs || numProbs != p->numProbs) { LzmaDec_FreeProbs(p, alloc); - p->probs = (CLzmaProb *)alloc->Alloc(alloc, numProbs * sizeof(CLzmaProb)); - p->numProbs = numProbs; + p->probs = (CLzmaProb *)ISzAlloc_Alloc(alloc, numProbs * sizeof(CLzmaProb)); if (!p->probs) return SZ_ERROR_MEM; + p->probs_1664 = p->probs + 1664; + p->numProbs = numProbs; } return SZ_OK; } -SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc) +SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc) { CLzmaProps propNew; - RINOK(LzmaProps_Decode(&propNew, props, propsSize)); - RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc)); + RINOK(LzmaProps_Decode(&propNew, props, propsSize)) + RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc)) p->prop = propNew; return SZ_OK; } -SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc) +SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc) { CLzmaProps propNew; SizeT dicBufSize; - RINOK(LzmaProps_Decode(&propNew, props, propsSize)); - RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc)); + RINOK(LzmaProps_Decode(&propNew, props, propsSize)) + RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc)) { UInt32 dictSize = propNew.dicSize; SizeT mask = ((UInt32)1 << 12) - 1; if (dictSize >= ((UInt32)1 << 30)) mask = ((UInt32)1 << 22) - 1; - else if (dictSize >= ((UInt32)1 << 22)) mask = ((UInt32)1 << 20) - 1;; + else if (dictSize >= ((UInt32)1 << 22)) mask = ((UInt32)1 << 20) - 1; dicBufSize = ((SizeT)dictSize + mask) & ~mask; if (dicBufSize < dictSize) dicBufSize = dictSize; @@ -1062,7 +1325,7 @@ SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAll if (!p->dic || dicBufSize != p->dicBufSize) { LzmaDec_FreeDict(p, alloc); - p->dic = (Byte *)alloc->Alloc(alloc, dicBufSize); + p->dic = (Byte *)ISzAlloc_Alloc(alloc, dicBufSize); if (!p->dic) { LzmaDec_FreeProbs(p, alloc); @@ -1076,7 +1339,7 @@ SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAll SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, const Byte *propData, unsigned propSize, ELzmaFinishMode finishMode, - ELzmaStatus *status, ISzAlloc *alloc) + ELzmaStatus *status, ISzAllocPtr alloc) { CLzmaDec p; SRes res; @@ -1085,8 +1348,8 @@ SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, *status = LZMA_STATUS_NOT_SPECIFIED; if (inSize < RC_INIT_SIZE) return SZ_ERROR_INPUT_EOF; - LzmaDec_Construct(&p); - RINOK(LzmaDec_AllocateProbs(&p, propData, propSize, alloc)); + LzmaDec_CONSTRUCT(&p) + RINOK(LzmaDec_AllocateProbs(&p, propData, propSize, alloc)) p.dic = dest; p.dicBufSize = outSize; LzmaDec_Init(&p); diff --git a/3rdparty/lzma/C/LzmaDec.h b/3rdparty/lzma/C/LzmaDec.h index cc44daef214..b0ce28fa02a 100644 --- a/3rdparty/lzma/C/LzmaDec.h +++ b/3rdparty/lzma/C/LzmaDec.h @@ -1,31 +1,36 @@ /* LzmaDec.h -- LZMA Decoder -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __LZMA_DEC_H -#define __LZMA_DEC_H +#ifndef ZIP7_INC_LZMA_DEC_H +#define ZIP7_INC_LZMA_DEC_H #include "7zTypes.h" EXTERN_C_BEGIN -/* #define _LZMA_PROB32 */ -/* _LZMA_PROB32 can increase the speed on some CPUs, +/* #define Z7_LZMA_PROB32 */ +/* Z7_LZMA_PROB32 can increase the speed on some CPUs, but memory usage for CLzmaDec::probs will be doubled in that case */ -#ifdef _LZMA_PROB32 -#define CLzmaProb UInt32 +typedef +#ifdef Z7_LZMA_PROB32 + UInt32 #else -#define CLzmaProb UInt16 + UInt16 #endif + CLzmaProb; /* ---------- LZMA Properties ---------- */ #define LZMA_PROPS_SIZE 5 -typedef struct _CLzmaProps +typedef struct { - unsigned lc, lp, pb; + Byte lc; + Byte lp; + Byte pb; + Byte _pad_; UInt32 dicSize; } CLzmaProps; @@ -47,32 +52,35 @@ SRes LzmaProps_Decode(CLzmaProps *p, const Byte *data, unsigned size); typedef struct { + /* Don't change this structure. ASM code can use it. */ CLzmaProps prop; CLzmaProb *probs; + CLzmaProb *probs_1664; Byte *dic; - const Byte *buf; - UInt32 range, code; - SizeT dicPos; SizeT dicBufSize; + SizeT dicPos; + const Byte *buf; + UInt32 range; + UInt32 code; UInt32 processedPos; UInt32 checkDicSize; - unsigned state; UInt32 reps[4]; - unsigned remainLen; - int needFlush; - int needInitState; + UInt32 state; + UInt32 remainLen; + UInt32 numProbs; unsigned tempBufSize; Byte tempBuf[LZMA_REQUIRED_INPUT_MAX]; } CLzmaDec; -#define LzmaDec_Construct(p) { (p)->dic = 0; (p)->probs = 0; } +#define LzmaDec_CONSTRUCT(p) { (p)->dic = NULL; (p)->probs = NULL; } +#define LzmaDec_Construct(p) LzmaDec_CONSTRUCT(p) void LzmaDec_Init(CLzmaDec *p); /* There are two types of LZMA streams: - 0) Stream with end mark. That end mark adds about 6 bytes to compressed size. - 1) Stream without end mark. You must know exact uncompressed size to decompress such stream. */ + - Stream with end mark. That end mark adds about 6 bytes to compressed size. + - Stream without end mark. You must know exact uncompressed size to decompress such stream. */ typedef enum { @@ -129,11 +137,11 @@ LzmaDec_Allocate* can return: SZ_ERROR_UNSUPPORTED - Unsupported properties */ -SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc); -void LzmaDec_FreeProbs(CLzmaDec *p, ISzAlloc *alloc); +SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc); +void LzmaDec_FreeProbs(CLzmaDec *p, ISzAllocPtr alloc); -SRes LzmaDec_Allocate(CLzmaDec *state, const Byte *prop, unsigned propsSize, ISzAlloc *alloc); -void LzmaDec_Free(CLzmaDec *state, ISzAlloc *alloc); +SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc); +void LzmaDec_Free(CLzmaDec *p, ISzAllocPtr alloc); /* ---------- Dictionary Interface ---------- */ @@ -142,7 +150,7 @@ void LzmaDec_Free(CLzmaDec *state, ISzAlloc *alloc); You must work with CLzmaDec variables directly in this interface. STEPS: - LzmaDec_Constr() + LzmaDec_Construct() LzmaDec_Allocate() for (each new stream) { @@ -174,6 +182,7 @@ Returns: LZMA_STATUS_NEEDS_MORE_INPUT LZMA_STATUS_MAYBE_FINISHED_WITHOUT_MARK SZ_ERROR_DATA - Data error + SZ_ERROR_FAIL - Some unexpected error: internal error of code, memory corruption or hardware failure */ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, @@ -216,11 +225,12 @@ Returns: SZ_ERROR_MEM - Memory allocation error SZ_ERROR_UNSUPPORTED - Unsupported properties SZ_ERROR_INPUT_EOF - It needs more bytes in input buffer (src). + SZ_ERROR_FAIL - Some unexpected error: internal error of code, memory corruption or hardware failure */ SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, const Byte *propData, unsigned propSize, ELzmaFinishMode finishMode, - ELzmaStatus *status, ISzAlloc *alloc); + ELzmaStatus *status, ISzAllocPtr alloc); EXTERN_C_END diff --git a/3rdparty/lzma/C/LzmaEnc.c b/3rdparty/lzma/C/LzmaEnc.c index 70df4569b7d..6d13cac8ce2 100644 --- a/3rdparty/lzma/C/LzmaEnc.c +++ b/3rdparty/lzma/C/LzmaEnc.c @@ -1,5 +1,5 @@ /* LzmaEnc.c -- LZMA Encoder -2016-05-16 : Igor Pavlov : Public domain */ +2023-04-13: Igor Pavlov : Public domain */ #include "Precomp.h" @@ -12,31 +12,36 @@ #include <stdio.h> #endif +#include "CpuArch.h" #include "LzmaEnc.h" #include "LzFind.h" -#ifndef _7ZIP_ST +#ifndef Z7_ST #include "LzFindMt.h" #endif +/* the following LzmaEnc_* declarations is internal LZMA interface for LZMA2 encoder */ + +SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle p, ISeqInStreamPtr inStream, UInt32 keepWindowSize, + ISzAllocPtr alloc, ISzAllocPtr allocBig); +SRes LzmaEnc_MemPrepare(CLzmaEncHandle p, const Byte *src, SizeT srcLen, + UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig); +SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle p, BoolInt reInit, + Byte *dest, size_t *destLen, UInt32 desiredPackSize, UInt32 *unpackSize); +const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle p); +void LzmaEnc_Finish(CLzmaEncHandle p); +void LzmaEnc_SaveState(CLzmaEncHandle p); +void LzmaEnc_RestoreState(CLzmaEncHandle p); + #ifdef SHOW_STAT static unsigned g_STAT_OFFSET = 0; #endif -#define kMaxHistorySize ((UInt32)3 << 29) -/* #define kMaxHistorySize ((UInt32)7 << 29) */ - -#define kBlockSizeMax ((1 << LZMA_NUM_BLOCK_SIZE_BITS) - 1) - -#define kBlockSize (9 << 10) -#define kUnpackBlockSize (1 << 18) -#define kMatchArraySize (1 << 21) -#define kMatchRecordMaxSize ((LZMA_MATCH_LEN_MAX * 2 + 3) * LZMA_MATCH_LEN_MAX) +/* for good normalization speed we still reserve 256 MB before 4 GB range */ +#define kLzmaMaxHistorySize ((UInt32)15 << 28) -#define kNumMaxDirectBits (31) - -#define kNumTopBits 24 -#define kTopValue ((UInt32)1 << kNumTopBits) +// #define kNumTopBits 24 +#define kTopValue ((UInt32)1 << 24) #define kNumBitModelTotalBits 11 #define kBitModelTotal (1 << kNumBitModelTotalBits) @@ -45,7 +50,9 @@ static unsigned g_STAT_OFFSET = 0; #define kNumMoveReducingBits 4 #define kNumBitPriceShiftBits 4 -#define kBitPrice (1 << kNumBitPriceShiftBits) +// #define kBitPrice (1 << kNumBitPriceShiftBits) + +#define REP_LEN_COUNT 64 void LzmaEncProps_Init(CLzmaEncProps *p) { @@ -53,7 +60,9 @@ void LzmaEncProps_Init(CLzmaEncProps *p) p->dictSize = p->mc = 0; p->reduceSize = (UInt64)(Int64)-1; p->lc = p->lp = p->pb = p->algo = p->fb = p->btMode = p->numHashBytes = p->numThreads = -1; + p->numHashOutBits = 0; p->writeEndMark = 0; + p->affinity = 0; } void LzmaEncProps_Normalize(CLzmaEncProps *p) @@ -62,15 +71,21 @@ void LzmaEncProps_Normalize(CLzmaEncProps *p) if (level < 0) level = 5; p->level = level; - if (p->dictSize == 0) p->dictSize = (level <= 5 ? (1 << (level * 2 + 14)) : (level == 6 ? (1 << 25) : (1 << 26))); + if (p->dictSize == 0) + p->dictSize = + ( level <= 3 ? ((UInt32)1 << (level * 2 + 16)) : + ( level <= 6 ? ((UInt32)1 << (level + 19)) : + ( level <= 7 ? ((UInt32)1 << 25) : ((UInt32)1 << 26) + ))); + if (p->dictSize > p->reduceSize) { - unsigned i; - for (i = 11; i <= 30; i++) - { - if ((UInt32)p->reduceSize <= ((UInt32)2 << i)) { p->dictSize = ((UInt32)2 << i); break; } - if ((UInt32)p->reduceSize <= ((UInt32)3 << i)) { p->dictSize = ((UInt32)3 << i); break; } - } + UInt32 v = (UInt32)p->reduceSize; + const UInt32 kReduceMin = ((UInt32)1 << 12); + if (v < kReduceMin) + v = kReduceMin; + if (p->dictSize > v) + p->dictSize = v; } if (p->lc < 0) p->lc = 3; @@ -80,12 +95,12 @@ void LzmaEncProps_Normalize(CLzmaEncProps *p) if (p->algo < 0) p->algo = (level < 5 ? 0 : 1); if (p->fb < 0) p->fb = (level < 7 ? 32 : 64); if (p->btMode < 0) p->btMode = (p->algo == 0 ? 0 : 1); - if (p->numHashBytes < 0) p->numHashBytes = 4; - if (p->mc == 0) p->mc = (16 + (p->fb >> 1)) >> (p->btMode ? 0 : 1); + if (p->numHashBytes < 0) p->numHashBytes = (p->btMode ? 4 : 5); + if (p->mc == 0) p->mc = (16 + ((unsigned)p->fb >> 1)) >> (p->btMode ? 0 : 1); if (p->numThreads < 0) p->numThreads = - #ifndef _7ZIP_ST + #ifndef Z7_ST ((p->btMode && p->algo) ? 2 : 1); #else 1; @@ -99,30 +114,97 @@ UInt32 LzmaEncProps_GetDictSize(const CLzmaEncProps *props2) return props.dictSize; } -#if (_MSC_VER >= 1400) -/* BSR code is fast for some new CPUs */ -/* #define LZMA_LOG_BSR */ + +/* +x86/x64: + +BSR: + IF (SRC == 0) ZF = 1, DEST is undefined; + AMD : DEST is unchanged; + IF (SRC != 0) ZF = 0; DEST is index of top non-zero bit + BSR is slow in some processors + +LZCNT: + IF (SRC == 0) CF = 1, DEST is size_in_bits_of_register(src) (32 or 64) + IF (SRC != 0) CF = 0, DEST = num_lead_zero_bits + IF (DEST == 0) ZF = 1; + +LZCNT works only in new processors starting from Haswell. +if LZCNT is not supported by processor, then it's executed as BSR. +LZCNT can be faster than BSR, if supported. +*/ + +// #define LZMA_LOG_BSR + +#if defined(MY_CPU_ARM_OR_ARM64) /* || defined(MY_CPU_X86_OR_AMD64) */ + + #if (defined(__clang__) && (__clang_major__ >= 6)) \ + || (defined(__GNUC__) && (__GNUC__ >= 6)) + #define LZMA_LOG_BSR + #elif defined(_MSC_VER) && (_MSC_VER >= 1300) + // #if defined(MY_CPU_ARM_OR_ARM64) + #define LZMA_LOG_BSR + // #endif + #endif #endif +// #include <intrin.h> + #ifdef LZMA_LOG_BSR -#define kDicLogSizeMaxCompress 32 +#if defined(__clang__) \ + || defined(__GNUC__) + +/* + C code: : (30 - __builtin_clz(x)) + gcc9/gcc10 for x64 /x86 : 30 - (bsr(x) xor 31) + clang10 for x64 : 31 + (bsr(x) xor -32) +*/ + + #define MY_clz(x) ((unsigned)__builtin_clz(x)) + // __lzcnt32 + // __builtin_ia32_lzcnt_u32 + +#else // #if defined(_MSC_VER) + + #ifdef MY_CPU_ARM_OR_ARM64 + + #define MY_clz _CountLeadingZeros -#define BSR2_RET(pos, res) { unsigned long zz; _BitScanReverse(&zz, (pos)); res = (zz + zz) + ((pos >> (zz - 1)) & 1); } + #else // if defined(MY_CPU_X86_OR_AMD64) -static UInt32 GetPosSlot1(UInt32 pos) + // #define MY_clz __lzcnt // we can use lzcnt (unsupported by old CPU) + // _BitScanReverse code is not optimal for some MSVC compilers + #define BSR2_RET(pos, res) { unsigned long zz; _BitScanReverse(&zz, (pos)); zz--; \ + res = (zz + zz) + (pos >> zz); } + + #endif // MY_CPU_X86_OR_AMD64 + +#endif // _MSC_VER + + +#ifndef BSR2_RET + + #define BSR2_RET(pos, res) { unsigned zz = 30 - MY_clz(pos); \ + res = (zz + zz) + (pos >> zz); } + +#endif + + +unsigned GetPosSlot1(UInt32 pos); +unsigned GetPosSlot1(UInt32 pos) { - UInt32 res; + unsigned res; BSR2_RET(pos, res); return res; } #define GetPosSlot2(pos, res) { BSR2_RET(pos, res); } #define GetPosSlot(pos, res) { if (pos < 2) res = pos; else BSR2_RET(pos, res); } -#else -#define kNumLogBits (9 + sizeof(size_t) / 2) -/* #define kNumLogBits (11 + sizeof(size_t) / 8 * 3) */ +#else // ! LZMA_LOG_BSR + +#define kNumLogBits (11 + sizeof(size_t) / 8 * 3) #define kDicLogSizeMaxCompress ((kNumLogBits - 1) * 2 + 7) @@ -145,18 +227,18 @@ static void LzmaEnc_FastPosInit(Byte *g_FastPos) /* we can use ((limit - pos) >> 31) only if (pos < ((UInt32)1 << 31)) */ /* -#define BSR2_RET(pos, res) { UInt32 zz = 6 + ((kNumLogBits - 1) & \ +#define BSR2_RET(pos, res) { unsigned zz = 6 + ((kNumLogBits - 1) & \ (0 - (((((UInt32)1 << (kNumLogBits + 6)) - 1) - pos) >> 31))); \ res = p->g_FastPos[pos >> zz] + (zz * 2); } */ /* -#define BSR2_RET(pos, res) { UInt32 zz = 6 + ((kNumLogBits - 1) & \ +#define BSR2_RET(pos, res) { unsigned zz = 6 + ((kNumLogBits - 1) & \ (0 - (((((UInt32)1 << (kNumLogBits)) - 1) - (pos >> 6)) >> 31))); \ res = p->g_FastPos[pos >> zz] + (zz * 2); } */ -#define BSR2_RET(pos, res) { UInt32 zz = (pos < (1 << (kNumLogBits + 6))) ? 6 : 6 + kNumLogBits - 1; \ +#define BSR2_RET(pos, res) { unsigned zz = (pos < (1 << (kNumLogBits + 6))) ? 6 : 6 + kNumLogBits - 1; \ res = p->g_FastPos[pos >> zz] + (zz * 2); } /* @@ -167,55 +249,57 @@ static void LzmaEnc_FastPosInit(Byte *g_FastPos) #define GetPosSlot1(pos) p->g_FastPos[pos] #define GetPosSlot2(pos, res) { BSR2_RET(pos, res); } -#define GetPosSlot(pos, res) { if (pos < kNumFullDistances) res = p->g_FastPos[pos]; else BSR2_RET(pos, res); } +#define GetPosSlot(pos, res) { if (pos < kNumFullDistances) res = p->g_FastPos[pos & (kNumFullDistances - 1)]; else BSR2_RET(pos, res); } -#endif +#endif // LZMA_LOG_BSR #define LZMA_NUM_REPS 4 -typedef unsigned CState; +typedef UInt16 CState; +typedef UInt16 CExtra; typedef struct { UInt32 price; - CState state; - int prev1IsChar; - int prev2; - - UInt32 posPrev2; - UInt32 backPrev2; - - UInt32 posPrev; - UInt32 backPrev; - UInt32 backs[LZMA_NUM_REPS]; + CExtra extra; + // 0 : normal + // 1 : LIT : MATCH + // > 1 : MATCH (extra-1) : LIT : REP0 (len) + UInt32 len; + UInt32 dist; + UInt32 reps[LZMA_NUM_REPS]; } COptimal; -#define kNumOpts (1 << 12) + +// 18.06 +#define kNumOpts (1 << 11) +#define kPackReserve (kNumOpts * 8) +// #define kNumOpts (1 << 12) +// #define kPackReserve (1 + kNumOpts * 2) #define kNumLenToPosStates 4 #define kNumPosSlotBits 6 -#define kDicLogSizeMin 0 +// #define kDicLogSizeMin 0 #define kDicLogSizeMax 32 #define kDistTableSizeMax (kDicLogSizeMax * 2) - #define kNumAlignBits 4 #define kAlignTableSize (1 << kNumAlignBits) #define kAlignMask (kAlignTableSize - 1) #define kStartPosModelIndex 4 #define kEndPosModelIndex 14 -#define kNumPosModels (kEndPosModelIndex - kStartPosModelIndex) - #define kNumFullDistances (1 << (kEndPosModelIndex >> 1)) -#ifdef _LZMA_PROB32 -#define CLzmaProb UInt32 +typedef +#ifdef Z7_LZMA_PROB32 + UInt32 #else -#define CLzmaProb UInt16 + UInt16 #endif + CLzmaProb; #define LZMA_PB_MAX 4 #define LZMA_LC_MAX 8 @@ -223,15 +307,11 @@ typedef struct #define LZMA_NUM_PB_STATES_MAX (1 << LZMA_PB_MAX) - #define kLenNumLowBits 3 #define kLenNumLowSymbols (1 << kLenNumLowBits) -#define kLenNumMidBits 3 -#define kLenNumMidSymbols (1 << kLenNumMidBits) #define kLenNumHighBits 8 #define kLenNumHighSymbols (1 << kLenNumHighBits) - -#define kLenNumSymbolsTotal (kLenNumLowSymbols + kLenNumMidSymbols + kLenNumHighSymbols) +#define kLenNumSymbolsTotal (kLenNumLowSymbols * 2 + kLenNumHighSymbols) #define LZMA_MATCH_LEN_MIN 2 #define LZMA_MATCH_LEN_MAX (LZMA_MATCH_LEN_MIN + kLenNumSymbolsTotal - 1) @@ -241,33 +321,37 @@ typedef struct typedef struct { - CLzmaProb choice; - CLzmaProb choice2; - CLzmaProb low[LZMA_NUM_PB_STATES_MAX << kLenNumLowBits]; - CLzmaProb mid[LZMA_NUM_PB_STATES_MAX << kLenNumMidBits]; + CLzmaProb low[LZMA_NUM_PB_STATES_MAX << (kLenNumLowBits + 1)]; CLzmaProb high[kLenNumHighSymbols]; } CLenEnc; typedef struct { - CLenEnc p; - UInt32 tableSize; + unsigned tableSize; UInt32 prices[LZMA_NUM_PB_STATES_MAX][kLenNumSymbolsTotal]; - UInt32 counters[LZMA_NUM_PB_STATES_MAX]; + // UInt32 prices1[LZMA_NUM_PB_STATES_MAX][kLenNumLowSymbols * 2]; + // UInt32 prices2[kLenNumSymbolsTotal]; } CLenPriceEnc; +#define GET_PRICE_LEN(p, posState, len) \ + ((p)->prices[posState][(size_t)(len) - LZMA_MATCH_LEN_MIN]) + +/* +#define GET_PRICE_LEN(p, posState, len) \ + ((p)->prices2[(size_t)(len) - 2] + ((p)->prices1[posState][((len) - 2) & (kLenNumLowSymbols * 2 - 1)] & (((len) - 2 - kLenNumLowSymbols * 2) >> 9))) +*/ typedef struct { UInt32 range; - Byte cache; + unsigned cache; UInt64 low; UInt64 cacheSize; Byte *buf; Byte *bufLim; Byte *bufBase; - ISeqOutStream *outStream; + ISeqOutStreamPtr outStream; UInt64 processed; SRes res; } CRangeEnc; @@ -277,208 +361,233 @@ typedef struct { CLzmaProb *litProbs; - UInt32 state; + unsigned state; UInt32 reps[LZMA_NUM_REPS]; - CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX]; + CLzmaProb posAlignEncoder[1 << kNumAlignBits]; CLzmaProb isRep[kNumStates]; CLzmaProb isRepG0[kNumStates]; CLzmaProb isRepG1[kNumStates]; CLzmaProb isRepG2[kNumStates]; + CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX]; CLzmaProb isRep0Long[kNumStates][LZMA_NUM_PB_STATES_MAX]; CLzmaProb posSlotEncoder[kNumLenToPosStates][1 << kNumPosSlotBits]; - CLzmaProb posEncoders[kNumFullDistances - kEndPosModelIndex]; - CLzmaProb posAlignEncoder[1 << kNumAlignBits]; + CLzmaProb posEncoders[kNumFullDistances]; - CLenPriceEnc lenEnc; - CLenPriceEnc repLenEnc; + CLenEnc lenProbs; + CLenEnc repLenProbs; + } CSaveState; -typedef struct +typedef UInt32 CProbPrice; + + +struct CLzmaEnc { void *matchFinderObj; - IMatchFinder matchFinder; + IMatchFinder2 matchFinder; - UInt32 optimumEndIndex; - UInt32 optimumCurrentIndex; + unsigned optCur; + unsigned optEnd; - UInt32 longestMatchLength; - UInt32 numPairs; + unsigned longestMatchLen; + unsigned numPairs; UInt32 numAvail; - UInt32 numFastBytes; - UInt32 additionalOffset; + unsigned state; + unsigned numFastBytes; + unsigned additionalOffset; UInt32 reps[LZMA_NUM_REPS]; - UInt32 state; + unsigned lpMask, pbMask; + CLzmaProb *litProbs; + CRangeEnc rc; + + UInt32 backRes; unsigned lc, lp, pb; - unsigned lpMask, pbMask; unsigned lclp; - CLzmaProb *litProbs; - - Bool fastMode; - Bool writeEndMark; - Bool finished; - Bool multiThread; - Bool needInit; + BoolInt fastMode; + BoolInt writeEndMark; + BoolInt finished; + BoolInt multiThread; + BoolInt needInit; + // BoolInt _maxMode; UInt64 nowPos64; - UInt32 matchPriceCount; - UInt32 alignPriceCount; + unsigned matchPriceCount; + // unsigned alignPriceCount; + int repLenEncCounter; - UInt32 distTableSize; + unsigned distTableSize; UInt32 dictSize; SRes result; - CRangeEnc rc; - - #ifndef _7ZIP_ST - Bool mtMode; + #ifndef Z7_ST + BoolInt mtMode; + // begin of CMatchFinderMt is used in LZ thread CMatchFinderMt matchFinderMt; + // end of CMatchFinderMt is used in BT and HASH threads + // #else + // CMatchFinder matchFinderBase; #endif - CMatchFinder matchFinderBase; - #ifndef _7ZIP_ST + + // we suppose that we have 8-bytes alignment after CMatchFinder + + #ifndef Z7_ST Byte pad[128]; #endif - COptimal opt[kNumOpts]; - - #ifndef LZMA_LOG_BSR - Byte g_FastPos[1 << kNumLogBits]; - #endif + // LZ thread + CProbPrice ProbPrices[kBitModelTotal >> kNumMoveReducingBits]; - UInt32 ProbPrices[kBitModelTotal >> kNumMoveReducingBits]; - UInt32 matches[LZMA_MATCH_LEN_MAX * 2 + 2 + 1]; + // we want {len , dist} pairs to be 8-bytes aligned in matches array + UInt32 matches[LZMA_MATCH_LEN_MAX * 2 + 2]; + // we want 8-bytes alignment here + UInt32 alignPrices[kAlignTableSize]; UInt32 posSlotPrices[kNumLenToPosStates][kDistTableSizeMax]; UInt32 distancesPrices[kNumLenToPosStates][kNumFullDistances]; - UInt32 alignPrices[kAlignTableSize]; - CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX]; + CLzmaProb posAlignEncoder[1 << kNumAlignBits]; CLzmaProb isRep[kNumStates]; CLzmaProb isRepG0[kNumStates]; CLzmaProb isRepG1[kNumStates]; CLzmaProb isRepG2[kNumStates]; + CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX]; CLzmaProb isRep0Long[kNumStates][LZMA_NUM_PB_STATES_MAX]; - CLzmaProb posSlotEncoder[kNumLenToPosStates][1 << kNumPosSlotBits]; - CLzmaProb posEncoders[kNumFullDistances - kEndPosModelIndex]; - CLzmaProb posAlignEncoder[1 << kNumAlignBits]; + CLzmaProb posEncoders[kNumFullDistances]; + CLenEnc lenProbs; + CLenEnc repLenProbs; + + #ifndef LZMA_LOG_BSR + Byte g_FastPos[1 << kNumLogBits]; + #endif + CLenPriceEnc lenEnc; CLenPriceEnc repLenEnc; + COptimal opt[kNumOpts]; + CSaveState saveState; - #ifndef _7ZIP_ST + // BoolInt mf_Failure; + #ifndef Z7_ST Byte pad2[128]; #endif -} CLzmaEnc; +}; -void LzmaEnc_SaveState(CLzmaEncHandle pp) -{ - CLzmaEnc *p = (CLzmaEnc *)pp; - CSaveState *dest = &p->saveState; - int i; - dest->lenEnc = p->lenEnc; - dest->repLenEnc = p->repLenEnc; - dest->state = p->state; +#define MFB (p->matchFinderBase) +/* +#ifndef Z7_ST +#define MFB (p->matchFinderMt.MatchFinder) +#endif +*/ - for (i = 0; i < kNumStates; i++) - { - memcpy(dest->isMatch[i], p->isMatch[i], sizeof(p->isMatch[i])); - memcpy(dest->isRep0Long[i], p->isRep0Long[i], sizeof(p->isRep0Long[i])); - } - for (i = 0; i < kNumLenToPosStates; i++) - memcpy(dest->posSlotEncoder[i], p->posSlotEncoder[i], sizeof(p->posSlotEncoder[i])); - memcpy(dest->isRep, p->isRep, sizeof(p->isRep)); - memcpy(dest->isRepG0, p->isRepG0, sizeof(p->isRepG0)); - memcpy(dest->isRepG1, p->isRepG1, sizeof(p->isRepG1)); - memcpy(dest->isRepG2, p->isRepG2, sizeof(p->isRepG2)); - memcpy(dest->posEncoders, p->posEncoders, sizeof(p->posEncoders)); - memcpy(dest->posAlignEncoder, p->posAlignEncoder, sizeof(p->posAlignEncoder)); - memcpy(dest->reps, p->reps, sizeof(p->reps)); - memcpy(dest->litProbs, p->litProbs, ((UInt32)0x300 << p->lclp) * sizeof(CLzmaProb)); -} +// #define GET_CLzmaEnc_p CLzmaEnc *p = (CLzmaEnc*)(void *)p; +// #define GET_const_CLzmaEnc_p const CLzmaEnc *p = (const CLzmaEnc*)(const void *)p; + +#define COPY_ARR(dest, src, arr) memcpy((dest)->arr, (src)->arr, sizeof((src)->arr)); -void LzmaEnc_RestoreState(CLzmaEncHandle pp) +#define COPY_LZMA_ENC_STATE(d, s, p) \ + (d)->state = (s)->state; \ + COPY_ARR(d, s, reps) \ + COPY_ARR(d, s, posAlignEncoder) \ + COPY_ARR(d, s, isRep) \ + COPY_ARR(d, s, isRepG0) \ + COPY_ARR(d, s, isRepG1) \ + COPY_ARR(d, s, isRepG2) \ + COPY_ARR(d, s, isMatch) \ + COPY_ARR(d, s, isRep0Long) \ + COPY_ARR(d, s, posSlotEncoder) \ + COPY_ARR(d, s, posEncoders) \ + (d)->lenProbs = (s)->lenProbs; \ + (d)->repLenProbs = (s)->repLenProbs; \ + memcpy((d)->litProbs, (s)->litProbs, ((UInt32)0x300 << (p)->lclp) * sizeof(CLzmaProb)); + +void LzmaEnc_SaveState(CLzmaEncHandle p) { - CLzmaEnc *dest = (CLzmaEnc *)pp; - const CSaveState *p = &dest->saveState; - int i; - dest->lenEnc = p->lenEnc; - dest->repLenEnc = p->repLenEnc; - dest->state = p->state; + // GET_CLzmaEnc_p + CSaveState *v = &p->saveState; + COPY_LZMA_ENC_STATE(v, p, p) +} - for (i = 0; i < kNumStates; i++) - { - memcpy(dest->isMatch[i], p->isMatch[i], sizeof(p->isMatch[i])); - memcpy(dest->isRep0Long[i], p->isRep0Long[i], sizeof(p->isRep0Long[i])); - } - for (i = 0; i < kNumLenToPosStates; i++) - memcpy(dest->posSlotEncoder[i], p->posSlotEncoder[i], sizeof(p->posSlotEncoder[i])); - memcpy(dest->isRep, p->isRep, sizeof(p->isRep)); - memcpy(dest->isRepG0, p->isRepG0, sizeof(p->isRepG0)); - memcpy(dest->isRepG1, p->isRepG1, sizeof(p->isRepG1)); - memcpy(dest->isRepG2, p->isRepG2, sizeof(p->isRepG2)); - memcpy(dest->posEncoders, p->posEncoders, sizeof(p->posEncoders)); - memcpy(dest->posAlignEncoder, p->posAlignEncoder, sizeof(p->posAlignEncoder)); - memcpy(dest->reps, p->reps, sizeof(p->reps)); - memcpy(dest->litProbs, p->litProbs, ((UInt32)0x300 << dest->lclp) * sizeof(CLzmaProb)); +void LzmaEnc_RestoreState(CLzmaEncHandle p) +{ + // GET_CLzmaEnc_p + const CSaveState *v = &p->saveState; + COPY_LZMA_ENC_STATE(p, v, p) } -SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2) + +Z7_NO_INLINE +SRes LzmaEnc_SetProps(CLzmaEncHandle p, const CLzmaEncProps *props2) { - CLzmaEnc *p = (CLzmaEnc *)pp; + // GET_CLzmaEnc_p CLzmaEncProps props = *props2; LzmaEncProps_Normalize(&props); if (props.lc > LZMA_LC_MAX || props.lp > LZMA_LP_MAX - || props.pb > LZMA_PB_MAX - || props.dictSize > ((UInt64)1 << kDicLogSizeMaxCompress) - || props.dictSize > kMaxHistorySize) + || props.pb > LZMA_PB_MAX) return SZ_ERROR_PARAM; + + if (props.dictSize > kLzmaMaxHistorySize) + props.dictSize = kLzmaMaxHistorySize; + + #ifndef LZMA_LOG_BSR + { + const UInt64 dict64 = props.dictSize; + if (dict64 > ((UInt64)1 << kDicLogSizeMaxCompress)) + return SZ_ERROR_PARAM; + } + #endif + p->dictSize = props.dictSize; { - unsigned fb = props.fb; + unsigned fb = (unsigned)props.fb; if (fb < 5) fb = 5; if (fb > LZMA_MATCH_LEN_MAX) fb = LZMA_MATCH_LEN_MAX; p->numFastBytes = fb; } - p->lc = props.lc; - p->lp = props.lp; - p->pb = props.pb; + p->lc = (unsigned)props.lc; + p->lp = (unsigned)props.lp; + p->pb = (unsigned)props.pb; p->fastMode = (props.algo == 0); - p->matchFinderBase.btMode = (Byte)(props.btMode ? 1 : 0); + // p->_maxMode = True; + MFB.btMode = (Byte)(props.btMode ? 1 : 0); + // MFB.btMode = (Byte)(props.btMode); { - UInt32 numHashBytes = 4; + unsigned numHashBytes = 4; if (props.btMode) { - if (props.numHashBytes < 2) - numHashBytes = 2; - else if (props.numHashBytes < 4) - numHashBytes = props.numHashBytes; + if (props.numHashBytes < 2) numHashBytes = 2; + else if (props.numHashBytes < 4) numHashBytes = (unsigned)props.numHashBytes; } - p->matchFinderBase.numHashBytes = numHashBytes; + if (props.numHashBytes >= 5) numHashBytes = 5; + + MFB.numHashBytes = numHashBytes; + // MFB.numHashBytes_Min = 2; + MFB.numHashOutBits = (Byte)props.numHashOutBits; } - p->matchFinderBase.cutValue = props.mc; + MFB.cutValue = props.mc; - p->writeEndMark = props.writeEndMark; + p->writeEndMark = (BoolInt)props.writeEndMark; - #ifndef _7ZIP_ST + #ifndef Z7_ST /* if (newMultiThread != _multiThread) { @@ -487,18 +596,34 @@ SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2) } */ p->multiThread = (props.numThreads > 1); + p->matchFinderMt.btSync.affinity = + p->matchFinderMt.hashSync.affinity = props.affinity; #endif return SZ_OK; } -static const int kLiteralNextStates[kNumStates] = {0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5}; -static const int kMatchNextStates[kNumStates] = {7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10}; -static const int kRepNextStates[kNumStates] = {8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11}; -static const int kShortRepNextStates[kNumStates]= {9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11}; -#define IsCharState(s) ((s) < 7) +void LzmaEnc_SetDataSize(CLzmaEncHandle p, UInt64 expectedDataSiize) +{ + // GET_CLzmaEnc_p + MFB.expectedDataSize = expectedDataSiize; +} + + +#define kState_Start 0 +#define kState_LitAfterMatch 4 +#define kState_LitAfterRep 5 +#define kState_MatchAfterLit 7 +#define kState_RepAfterLit 8 +static const Byte kLiteralNextStates[kNumStates] = {0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5}; +static const Byte kMatchNextStates[kNumStates] = {7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10}; +static const Byte kRepNextStates[kNumStates] = {8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11}; +static const Byte kShortRepNextStates[kNumStates]= {9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11}; + +#define IsLitState(s) ((s) < 7) +#define GetLenToPosState2(len) (((len) < kNumLenToPosStates - 1) ? (len) : kNumLenToPosStates - 1) #define GetLenToPosState(len) (((len) < kNumLenToPosStates + 1) ? (len) - 2 : kNumLenToPosStates - 1) #define kInfinityPrice (1 << 30) @@ -509,14 +634,16 @@ static void RangeEnc_Construct(CRangeEnc *p) p->bufBase = NULL; } -#define RangeEnc_GetProcessed(p) ((p)->processed + ((p)->buf - (p)->bufBase) + (p)->cacheSize) +#define RangeEnc_GetProcessed(p) ( (p)->processed + (size_t)((p)->buf - (p)->bufBase) + (p)->cacheSize) +#define RangeEnc_GetProcessed_sizet(p) ((size_t)(p)->processed + (size_t)((p)->buf - (p)->bufBase) + (size_t)(p)->cacheSize) #define RC_BUF_SIZE (1 << 16) -static int RangeEnc_Alloc(CRangeEnc *p, ISzAlloc *alloc) + +static int RangeEnc_Alloc(CRangeEnc *p, ISzAllocPtr alloc) { if (!p->bufBase) { - p->bufBase = (Byte *)alloc->Alloc(alloc, RC_BUF_SIZE); + p->bufBase = (Byte *)ISzAlloc_Alloc(alloc, RC_BUF_SIZE); if (!p->bufBase) return 0; p->bufLim = p->bufBase + RC_BUF_SIZE; @@ -524,19 +651,18 @@ static int RangeEnc_Alloc(CRangeEnc *p, ISzAlloc *alloc) return 1; } -static void RangeEnc_Free(CRangeEnc *p, ISzAlloc *alloc) +static void RangeEnc_Free(CRangeEnc *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->bufBase); - p->bufBase = 0; + ISzAlloc_Free(alloc, p->bufBase); + p->bufBase = NULL; } static void RangeEnc_Init(CRangeEnc *p) { - /* Stream.Init(); */ - p->low = 0; p->range = 0xFFFFFFFF; - p->cacheSize = 1; p->cache = 0; + p->low = 0; + p->cacheSize = 0; p->buf = p->bufBase; @@ -544,37 +670,48 @@ static void RangeEnc_Init(CRangeEnc *p) p->res = SZ_OK; } -static void RangeEnc_FlushStream(CRangeEnc *p) +Z7_NO_INLINE static void RangeEnc_FlushStream(CRangeEnc *p) { - size_t num; - if (p->res != SZ_OK) - return; - num = p->buf - p->bufBase; - if (num != p->outStream->Write(p->outStream, p->bufBase, num)) - p->res = SZ_ERROR_WRITE; + const size_t num = (size_t)(p->buf - p->bufBase); + if (p->res == SZ_OK) + { + if (num != ISeqOutStream_Write(p->outStream, p->bufBase, num)) + p->res = SZ_ERROR_WRITE; + } p->processed += num; p->buf = p->bufBase; } -static void MY_FAST_CALL RangeEnc_ShiftLow(CRangeEnc *p) +Z7_NO_INLINE static void Z7_FASTCALL RangeEnc_ShiftLow(CRangeEnc *p) { - if ((UInt32)p->low < (UInt32)0xFF000000 || (unsigned)(p->low >> 32) != 0) + UInt32 low = (UInt32)p->low; + unsigned high = (unsigned)(p->low >> 32); + p->low = (UInt32)(low << 8); + if (low < (UInt32)0xFF000000 || high != 0) { - Byte temp = p->cache; - do { Byte *buf = p->buf; - *buf++ = (Byte)(temp + (Byte)(p->low >> 32)); + *buf++ = (Byte)(p->cache + high); + p->cache = (unsigned)(low >> 24); p->buf = buf; if (buf == p->bufLim) RangeEnc_FlushStream(p); - temp = 0xFF; + if (p->cacheSize == 0) + return; + } + high += 0xFF; + for (;;) + { + Byte *buf = p->buf; + *buf++ = (Byte)(high); + p->buf = buf; + if (buf == p->bufLim) + RangeEnc_FlushStream(p); + if (--p->cacheSize == 0) + return; } - while (--p->cacheSize != 0); - p->cache = (Byte)((UInt32)p->low >> 24); } p->cacheSize++; - p->low = (UInt32)p->low << 8; } static void RangeEnc_FlushData(CRangeEnc *p) @@ -584,78 +721,121 @@ static void RangeEnc_FlushData(CRangeEnc *p) RangeEnc_ShiftLow(p); } -static void RangeEnc_EncodeDirectBits(CRangeEnc *p, UInt32 value, unsigned numBits) -{ - do - { - p->range >>= 1; - p->low += p->range & (0 - ((value >> --numBits) & 1)); - if (p->range < kTopValue) - { - p->range <<= 8; - RangeEnc_ShiftLow(p); - } - } - while (numBits != 0); -} +#define RC_NORM(p) if (range < kTopValue) { range <<= 8; RangeEnc_ShiftLow(p); } -static void RangeEnc_EncodeBit(CRangeEnc *p, CLzmaProb *prob, UInt32 symbol) -{ - UInt32 ttt = *prob; - UInt32 newBound = (p->range >> kNumBitModelTotalBits) * ttt; - if (symbol == 0) - { - p->range = newBound; - ttt += (kBitModelTotal - ttt) >> kNumMoveBits; - } - else - { - p->low += newBound; - p->range -= newBound; - ttt -= ttt >> kNumMoveBits; +#define RC_BIT_PRE(p, prob) \ + ttt = *(prob); \ + newBound = (range >> kNumBitModelTotalBits) * ttt; + +// #define Z7_LZMA_ENC_USE_BRANCH + +#ifdef Z7_LZMA_ENC_USE_BRANCH + +#define RC_BIT(p, prob, bit) { \ + RC_BIT_PRE(p, prob) \ + if (bit == 0) { range = newBound; ttt += (kBitModelTotal - ttt) >> kNumMoveBits; } \ + else { (p)->low += newBound; range -= newBound; ttt -= ttt >> kNumMoveBits; } \ + *(prob) = (CLzmaProb)ttt; \ + RC_NORM(p) \ } - *prob = (CLzmaProb)ttt; - if (p->range < kTopValue) - { - p->range <<= 8; - RangeEnc_ShiftLow(p); + +#else + +#define RC_BIT(p, prob, bit) { \ + UInt32 mask; \ + RC_BIT_PRE(p, prob) \ + mask = 0 - (UInt32)bit; \ + range &= mask; \ + mask &= newBound; \ + range -= mask; \ + (p)->low += mask; \ + mask = (UInt32)bit - 1; \ + range += newBound & mask; \ + mask &= (kBitModelTotal - ((1 << kNumMoveBits) - 1)); \ + mask += ((1 << kNumMoveBits) - 1); \ + ttt += (UInt32)((Int32)(mask - ttt) >> kNumMoveBits); \ + *(prob) = (CLzmaProb)ttt; \ + RC_NORM(p) \ } + +#endif + + + + +#define RC_BIT_0_BASE(p, prob) \ + range = newBound; *(prob) = (CLzmaProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits)); + +#define RC_BIT_1_BASE(p, prob) \ + range -= newBound; (p)->low += newBound; *(prob) = (CLzmaProb)(ttt - (ttt >> kNumMoveBits)); \ + +#define RC_BIT_0(p, prob) \ + RC_BIT_0_BASE(p, prob) \ + RC_NORM(p) + +#define RC_BIT_1(p, prob) \ + RC_BIT_1_BASE(p, prob) \ + RC_NORM(p) + +static void RangeEnc_EncodeBit_0(CRangeEnc *p, CLzmaProb *prob) +{ + UInt32 range, ttt, newBound; + range = p->range; + RC_BIT_PRE(p, prob) + RC_BIT_0(p, prob) + p->range = range; } -static void LitEnc_Encode(CRangeEnc *p, CLzmaProb *probs, UInt32 symbol) +static void LitEnc_Encode(CRangeEnc *p, CLzmaProb *probs, UInt32 sym) { - symbol |= 0x100; + UInt32 range = p->range; + sym |= 0x100; do { - RangeEnc_EncodeBit(p, probs + (symbol >> 8), (symbol >> 7) & 1); - symbol <<= 1; + UInt32 ttt, newBound; + // RangeEnc_EncodeBit(p, probs + (sym >> 8), (sym >> 7) & 1); + CLzmaProb *prob = probs + (sym >> 8); + UInt32 bit = (sym >> 7) & 1; + sym <<= 1; + RC_BIT(p, prob, bit) } - while (symbol < 0x10000); + while (sym < 0x10000); + p->range = range; } -static void LitEnc_EncodeMatched(CRangeEnc *p, CLzmaProb *probs, UInt32 symbol, UInt32 matchByte) +static void LitEnc_EncodeMatched(CRangeEnc *p, CLzmaProb *probs, UInt32 sym, UInt32 matchByte) { + UInt32 range = p->range; UInt32 offs = 0x100; - symbol |= 0x100; + sym |= 0x100; do { + UInt32 ttt, newBound; + CLzmaProb *prob; + UInt32 bit; matchByte <<= 1; - RangeEnc_EncodeBit(p, probs + (offs + (matchByte & offs) + (symbol >> 8)), (symbol >> 7) & 1); - symbol <<= 1; - offs &= ~(matchByte ^ symbol); + // RangeEnc_EncodeBit(p, probs + (offs + (matchByte & offs) + (sym >> 8)), (sym >> 7) & 1); + prob = probs + (offs + (matchByte & offs) + (sym >> 8)); + bit = (sym >> 7) & 1; + sym <<= 1; + offs &= ~(matchByte ^ sym); + RC_BIT(p, prob, bit) } - while (symbol < 0x10000); + while (sym < 0x10000); + p->range = range; } -static void LzmaEnc_InitPriceTables(UInt32 *ProbPrices) + + +static void LzmaEnc_InitPriceTables(CProbPrice *ProbPrices) { UInt32 i; - for (i = (1 << kNumMoveReducingBits) / 2; i < kBitModelTotal; i += (1 << kNumMoveReducingBits)) + for (i = 0; i < (kBitModelTotal >> kNumMoveReducingBits); i++) { - const int kCyclesBits = kNumBitPriceShiftBits; - UInt32 w = i; - UInt32 bitCount = 0; - int j; + const unsigned kCyclesBits = kNumBitPriceShiftBits; + UInt32 w = (i << kNumMoveReducingBits) + (1 << (kNumMoveReducingBits - 1)); + unsigned bitCount = 0; + unsigned j; for (j = 0; j < kCyclesBits; j++) { w = w * w; @@ -666,567 +846,663 @@ static void LzmaEnc_InitPriceTables(UInt32 *ProbPrices) bitCount++; } } - ProbPrices[i >> kNumMoveReducingBits] = ((kNumBitModelTotalBits << kCyclesBits) - 15 - bitCount); + ProbPrices[i] = (CProbPrice)(((unsigned)kNumBitModelTotalBits << kCyclesBits) - 15 - bitCount); + // printf("\n%3d: %5d", i, ProbPrices[i]); } } -#define GET_PRICE(prob, symbol) \ - p->ProbPrices[((prob) ^ (((-(int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits]; +#define GET_PRICE(prob, bit) \ + p->ProbPrices[((prob) ^ (unsigned)(((-(int)(bit))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits] -#define GET_PRICEa(prob, symbol) \ - ProbPrices[((prob) ^ ((-((int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits]; +#define GET_PRICEa(prob, bit) \ + ProbPrices[((prob) ^ (unsigned)((-((int)(bit))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits] #define GET_PRICE_0(prob) p->ProbPrices[(prob) >> kNumMoveReducingBits] #define GET_PRICE_1(prob) p->ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits] -#define GET_PRICE_0a(prob) ProbPrices[(prob) >> kNumMoveReducingBits] -#define GET_PRICE_1a(prob) ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits] +#define GET_PRICEa_0(prob) ProbPrices[(prob) >> kNumMoveReducingBits] +#define GET_PRICEa_1(prob) ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits] + -static UInt32 LitEnc_GetPrice(const CLzmaProb *probs, UInt32 symbol, const UInt32 *ProbPrices) +static UInt32 LitEnc_GetPrice(const CLzmaProb *probs, UInt32 sym, const CProbPrice *ProbPrices) { UInt32 price = 0; - symbol |= 0x100; + sym |= 0x100; do { - price += GET_PRICEa(probs[symbol >> 8], (symbol >> 7) & 1); - symbol <<= 1; + unsigned bit = sym & 1; + sym >>= 1; + price += GET_PRICEa(probs[sym], bit); } - while (symbol < 0x10000); + while (sym >= 2); return price; } -static UInt32 LitEnc_GetPriceMatched(const CLzmaProb *probs, UInt32 symbol, UInt32 matchByte, const UInt32 *ProbPrices) + +static UInt32 LitEnc_Matched_GetPrice(const CLzmaProb *probs, UInt32 sym, UInt32 matchByte, const CProbPrice *ProbPrices) { UInt32 price = 0; UInt32 offs = 0x100; - symbol |= 0x100; + sym |= 0x100; do { matchByte <<= 1; - price += GET_PRICEa(probs[offs + (matchByte & offs) + (symbol >> 8)], (symbol >> 7) & 1); - symbol <<= 1; - offs &= ~(matchByte ^ symbol); + price += GET_PRICEa(probs[offs + (matchByte & offs) + (sym >> 8)], (sym >> 7) & 1); + sym <<= 1; + offs &= ~(matchByte ^ sym); } - while (symbol < 0x10000); + while (sym < 0x10000); return price; } -static void RcTree_Encode(CRangeEnc *rc, CLzmaProb *probs, int numBitLevels, UInt32 symbol) +static void RcTree_ReverseEncode(CRangeEnc *rc, CLzmaProb *probs, unsigned numBits, unsigned sym) { - UInt32 m = 1; - int i; - for (i = numBitLevels; i != 0;) - { - UInt32 bit; - i--; - bit = (symbol >> i) & 1; - RangeEnc_EncodeBit(rc, probs + m, bit); - m = (m << 1) | bit; - } -} - -static void RcTree_ReverseEncode(CRangeEnc *rc, CLzmaProb *probs, int numBitLevels, UInt32 symbol) -{ - UInt32 m = 1; - int i; - for (i = 0; i < numBitLevels; i++) + UInt32 range = rc->range; + unsigned m = 1; + do { - UInt32 bit = symbol & 1; - RangeEnc_EncodeBit(rc, probs + m, bit); + UInt32 ttt, newBound; + unsigned bit = sym & 1; + // RangeEnc_EncodeBit(rc, probs + m, bit); + sym >>= 1; + RC_BIT(rc, probs + m, bit) m = (m << 1) | bit; - symbol >>= 1; - } -} - -static UInt32 RcTree_GetPrice(const CLzmaProb *probs, int numBitLevels, UInt32 symbol, const UInt32 *ProbPrices) -{ - UInt32 price = 0; - symbol |= (1 << numBitLevels); - while (symbol != 1) - { - price += GET_PRICEa(probs[symbol >> 1], symbol & 1); - symbol >>= 1; } - return price; + while (--numBits); + rc->range = range; } -static UInt32 RcTree_ReverseGetPrice(const CLzmaProb *probs, int numBitLevels, UInt32 symbol, const UInt32 *ProbPrices) -{ - UInt32 price = 0; - UInt32 m = 1; - int i; - for (i = numBitLevels; i != 0; i--) - { - UInt32 bit = symbol & 1; - symbol >>= 1; - price += GET_PRICEa(probs[m], bit); - m = (m << 1) | bit; - } - return price; -} static void LenEnc_Init(CLenEnc *p) { unsigned i; - p->choice = p->choice2 = kProbInitValue; - for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << kLenNumLowBits); i++) + for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << (kLenNumLowBits + 1)); i++) p->low[i] = kProbInitValue; - for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << kLenNumMidBits); i++) - p->mid[i] = kProbInitValue; for (i = 0; i < kLenNumHighSymbols; i++) p->high[i] = kProbInitValue; } -static void LenEnc_Encode(CLenEnc *p, CRangeEnc *rc, UInt32 symbol, UInt32 posState) +static void LenEnc_Encode(CLenEnc *p, CRangeEnc *rc, unsigned sym, unsigned posState) { - if (symbol < kLenNumLowSymbols) - { - RangeEnc_EncodeBit(rc, &p->choice, 0); - RcTree_Encode(rc, p->low + (posState << kLenNumLowBits), kLenNumLowBits, symbol); - } - else + UInt32 range, ttt, newBound; + CLzmaProb *probs = p->low; + range = rc->range; + RC_BIT_PRE(rc, probs) + if (sym >= kLenNumLowSymbols) { - RangeEnc_EncodeBit(rc, &p->choice, 1); - if (symbol < kLenNumLowSymbols + kLenNumMidSymbols) + RC_BIT_1(rc, probs) + probs += kLenNumLowSymbols; + RC_BIT_PRE(rc, probs) + if (sym >= kLenNumLowSymbols * 2) { - RangeEnc_EncodeBit(rc, &p->choice2, 0); - RcTree_Encode(rc, p->mid + (posState << kLenNumMidBits), kLenNumMidBits, symbol - kLenNumLowSymbols); - } - else - { - RangeEnc_EncodeBit(rc, &p->choice2, 1); - RcTree_Encode(rc, p->high, kLenNumHighBits, symbol - kLenNumLowSymbols - kLenNumMidSymbols); + RC_BIT_1(rc, probs) + rc->range = range; + // RcTree_Encode(rc, p->high, kLenNumHighBits, sym - kLenNumLowSymbols * 2); + LitEnc_Encode(rc, p->high, sym - kLenNumLowSymbols * 2); + return; } + sym -= kLenNumLowSymbols; } -} -static void LenEnc_SetPrices(CLenEnc *p, UInt32 posState, UInt32 numSymbols, UInt32 *prices, const UInt32 *ProbPrices) -{ - UInt32 a0 = GET_PRICE_0a(p->choice); - UInt32 a1 = GET_PRICE_1a(p->choice); - UInt32 b0 = a1 + GET_PRICE_0a(p->choice2); - UInt32 b1 = a1 + GET_PRICE_1a(p->choice2); - UInt32 i = 0; - for (i = 0; i < kLenNumLowSymbols; i++) + // RcTree_Encode(rc, probs + (posState << kLenNumLowBits), kLenNumLowBits, sym); { - if (i >= numSymbols) - return; - prices[i] = a0 + RcTree_GetPrice(p->low + (posState << kLenNumLowBits), kLenNumLowBits, i, ProbPrices); + unsigned m; + unsigned bit; + RC_BIT_0(rc, probs) + probs += (posState << (1 + kLenNumLowBits)); + bit = (sym >> 2) ; RC_BIT(rc, probs + 1, bit) m = (1 << 1) + bit; + bit = (sym >> 1) & 1; RC_BIT(rc, probs + m, bit) m = (m << 1) + bit; + bit = sym & 1; RC_BIT(rc, probs + m, bit) + rc->range = range; } - for (; i < kLenNumLowSymbols + kLenNumMidSymbols; i++) - { - if (i >= numSymbols) - return; - prices[i] = b0 + RcTree_GetPrice(p->mid + (posState << kLenNumMidBits), kLenNumMidBits, i - kLenNumLowSymbols, ProbPrices); - } - for (; i < numSymbols; i++) - prices[i] = b1 + RcTree_GetPrice(p->high, kLenNumHighBits, i - kLenNumLowSymbols - kLenNumMidSymbols, ProbPrices); } -static void MY_FAST_CALL LenPriceEnc_UpdateTable(CLenPriceEnc *p, UInt32 posState, const UInt32 *ProbPrices) +static void SetPrices_3(const CLzmaProb *probs, UInt32 startPrice, UInt32 *prices, const CProbPrice *ProbPrices) { - LenEnc_SetPrices(&p->p, posState, p->tableSize, p->prices[posState], ProbPrices); - p->counters[posState] = p->tableSize; + unsigned i; + for (i = 0; i < 8; i += 2) + { + UInt32 price = startPrice; + UInt32 prob; + price += GET_PRICEa(probs[1 ], (i >> 2)); + price += GET_PRICEa(probs[2 + (i >> 2)], (i >> 1) & 1); + prob = probs[4 + (i >> 1)]; + prices[i ] = price + GET_PRICEa_0(prob); + prices[i + 1] = price + GET_PRICEa_1(prob); + } } -static void LenPriceEnc_UpdateTables(CLenPriceEnc *p, UInt32 numPosStates, const UInt32 *ProbPrices) -{ - UInt32 posState; - for (posState = 0; posState < numPosStates; posState++) - LenPriceEnc_UpdateTable(p, posState, ProbPrices); -} -static void LenEnc_Encode2(CLenPriceEnc *p, CRangeEnc *rc, UInt32 symbol, UInt32 posState, Bool updatePrice, const UInt32 *ProbPrices) +Z7_NO_INLINE static void Z7_FASTCALL LenPriceEnc_UpdateTables( + CLenPriceEnc *p, + unsigned numPosStates, + const CLenEnc *enc, + const CProbPrice *ProbPrices) { - LenEnc_Encode(&p->p, rc, symbol, posState); - if (updatePrice) - if (--p->counters[posState] == 0) - LenPriceEnc_UpdateTable(p, posState, ProbPrices); -} + UInt32 b; + + { + unsigned prob = enc->low[0]; + UInt32 a, c; + unsigned posState; + b = GET_PRICEa_1(prob); + a = GET_PRICEa_0(prob); + c = b + GET_PRICEa_0(enc->low[kLenNumLowSymbols]); + for (posState = 0; posState < numPosStates; posState++) + { + UInt32 *prices = p->prices[posState]; + const CLzmaProb *probs = enc->low + (posState << (1 + kLenNumLowBits)); + SetPrices_3(probs, a, prices, ProbPrices); + SetPrices_3(probs + kLenNumLowSymbols, c, prices + kLenNumLowSymbols, ProbPrices); + } + } + + /* + { + unsigned i; + UInt32 b; + a = GET_PRICEa_0(enc->low[0]); + for (i = 0; i < kLenNumLowSymbols; i++) + p->prices2[i] = a; + a = GET_PRICEa_1(enc->low[0]); + b = a + GET_PRICEa_0(enc->low[kLenNumLowSymbols]); + for (i = kLenNumLowSymbols; i < kLenNumLowSymbols * 2; i++) + p->prices2[i] = b; + a += GET_PRICEa_1(enc->low[kLenNumLowSymbols]); + } + */ + + // p->counter = numSymbols; + // p->counter = 64; + { + unsigned i = p->tableSize; + + if (i > kLenNumLowSymbols * 2) + { + const CLzmaProb *probs = enc->high; + UInt32 *prices = p->prices[0] + kLenNumLowSymbols * 2; + i -= kLenNumLowSymbols * 2 - 1; + i >>= 1; + b += GET_PRICEa_1(enc->low[kLenNumLowSymbols]); + do + { + /* + p->prices2[i] = a + + // RcTree_GetPrice(enc->high, kLenNumHighBits, i - kLenNumLowSymbols * 2, ProbPrices); + LitEnc_GetPrice(probs, i - kLenNumLowSymbols * 2, ProbPrices); + */ + // UInt32 price = a + RcTree_GetPrice(probs, kLenNumHighBits - 1, sym, ProbPrices); + unsigned sym = --i + (1 << (kLenNumHighBits - 1)); + UInt32 price = b; + do + { + unsigned bit = sym & 1; + sym >>= 1; + price += GET_PRICEa(probs[sym], bit); + } + while (sym >= 2); + { + unsigned prob = probs[(size_t)i + (1 << (kLenNumHighBits - 1))]; + prices[(size_t)i * 2 ] = price + GET_PRICEa_0(prob); + prices[(size_t)i * 2 + 1] = price + GET_PRICEa_1(prob); + } + } + while (i); + { + unsigned posState; + size_t num = (p->tableSize - kLenNumLowSymbols * 2) * sizeof(p->prices[0][0]); + for (posState = 1; posState < numPosStates; posState++) + memcpy(p->prices[posState] + kLenNumLowSymbols * 2, p->prices[0] + kLenNumLowSymbols * 2, num); + } + } + } +} -static void MovePos(CLzmaEnc *p, UInt32 num) -{ +/* #ifdef SHOW_STAT g_STAT_OFFSET += num; printf("\n MovePos %u", num); #endif +*/ - if (num != 0) - { - p->additionalOffset += num; - p->matchFinder.Skip(p->matchFinderObj, num); - } -} +#define MOVE_POS(p, num) { \ + p->additionalOffset += (num); \ + p->matchFinder.Skip(p->matchFinderObj, (UInt32)(num)); } + -static UInt32 ReadMatchDistances(CLzmaEnc *p, UInt32 *numDistancePairsRes) +static unsigned ReadMatchDistances(CLzmaEnc *p, unsigned *numPairsRes) { - UInt32 lenRes = 0, numPairs; + unsigned numPairs; + + p->additionalOffset++; p->numAvail = p->matchFinder.GetNumAvailableBytes(p->matchFinderObj); - numPairs = p->matchFinder.GetMatches(p->matchFinderObj, p->matches); + { + const UInt32 *d = p->matchFinder.GetMatches(p->matchFinderObj, p->matches); + // if (!d) { p->mf_Failure = True; *numPairsRes = 0; return 0; } + numPairs = (unsigned)(d - p->matches); + } + *numPairsRes = numPairs; #ifdef SHOW_STAT printf("\n i = %u numPairs = %u ", g_STAT_OFFSET, numPairs / 2); g_STAT_OFFSET++; { - UInt32 i; + unsigned i; for (i = 0; i < numPairs; i += 2) printf("%2u %6u | ", p->matches[i], p->matches[i + 1]); } #endif - if (numPairs > 0) + if (numPairs == 0) + return 0; { - lenRes = p->matches[numPairs - 2]; - if (lenRes == p->numFastBytes) + const unsigned len = p->matches[(size_t)numPairs - 2]; + if (len != p->numFastBytes) + return len; { UInt32 numAvail = p->numAvail; if (numAvail > LZMA_MATCH_LEN_MAX) numAvail = LZMA_MATCH_LEN_MAX; { - const Byte *pbyCur = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; - const Byte *pby = pbyCur + lenRes; - ptrdiff_t dif = (ptrdiff_t)-1 - p->matches[numPairs - 1]; - const Byte *pbyLim = pbyCur + numAvail; - for (; pby != pbyLim && *pby == pby[dif]; pby++); - lenRes = (UInt32)(pby - pbyCur); + const Byte *p1 = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; + const Byte *p2 = p1 + len; + const ptrdiff_t dif = (ptrdiff_t)-1 - (ptrdiff_t)p->matches[(size_t)numPairs - 1]; + const Byte *lim = p1 + numAvail; + for (; p2 != lim && *p2 == p2[dif]; p2++) + {} + return (unsigned)(p2 - p1); } } } - p->additionalOffset++; - *numDistancePairsRes = numPairs; - return lenRes; } +#define MARK_LIT ((UInt32)(Int32)-1) -#define MakeAsChar(p) (p)->backPrev = (UInt32)(-1); (p)->prev1IsChar = False; -#define MakeAsShortRep(p) (p)->backPrev = 0; (p)->prev1IsChar = False; -#define IsShortRep(p) ((p)->backPrev == 0) +#define MakeAs_Lit(p) { (p)->dist = MARK_LIT; (p)->extra = 0; } +#define MakeAs_ShortRep(p) { (p)->dist = 0; (p)->extra = 0; } +#define IsShortRep(p) ((p)->dist == 0) -static UInt32 GetRepLen1Price(CLzmaEnc *p, UInt32 state, UInt32 posState) -{ - return - GET_PRICE_0(p->isRepG0[state]) + - GET_PRICE_0(p->isRep0Long[state][posState]); -} -static UInt32 GetPureRepPrice(CLzmaEnc *p, UInt32 repIndex, UInt32 state, UInt32 posState) +#define GetPrice_ShortRep(p, state, posState) \ + ( GET_PRICE_0(p->isRepG0[state]) + GET_PRICE_0(p->isRep0Long[state][posState])) + +#define GetPrice_Rep_0(p, state, posState) ( \ + GET_PRICE_1(p->isMatch[state][posState]) \ + + GET_PRICE_1(p->isRep0Long[state][posState])) \ + + GET_PRICE_1(p->isRep[state]) \ + + GET_PRICE_0(p->isRepG0[state]) + +Z7_FORCE_INLINE +static UInt32 GetPrice_PureRep(const CLzmaEnc *p, unsigned repIndex, size_t state, size_t posState) { UInt32 price; + UInt32 prob = p->isRepG0[state]; if (repIndex == 0) { - price = GET_PRICE_0(p->isRepG0[state]); + price = GET_PRICE_0(prob); price += GET_PRICE_1(p->isRep0Long[state][posState]); } else { - price = GET_PRICE_1(p->isRepG0[state]); + price = GET_PRICE_1(prob); + prob = p->isRepG1[state]; if (repIndex == 1) - price += GET_PRICE_0(p->isRepG1[state]); + price += GET_PRICE_0(prob); else { - price += GET_PRICE_1(p->isRepG1[state]); + price += GET_PRICE_1(prob); price += GET_PRICE(p->isRepG2[state], repIndex - 2); } } return price; } -static UInt32 GetRepPrice(CLzmaEnc *p, UInt32 repIndex, UInt32 len, UInt32 state, UInt32 posState) -{ - return p->repLenEnc.prices[posState][len - LZMA_MATCH_LEN_MIN] + - GetPureRepPrice(p, repIndex, state, posState); -} -static UInt32 Backward(CLzmaEnc *p, UInt32 *backRes, UInt32 cur) +static unsigned Backward(CLzmaEnc *p, unsigned cur) { - UInt32 posMem = p->opt[cur].posPrev; - UInt32 backMem = p->opt[cur].backPrev; - p->optimumEndIndex = cur; - do + unsigned wr = cur + 1; + p->optEnd = wr; + + for (;;) { - if (p->opt[cur].prev1IsChar) + UInt32 dist = p->opt[cur].dist; + unsigned len = (unsigned)p->opt[cur].len; + unsigned extra = (unsigned)p->opt[cur].extra; + cur -= len; + + if (extra) { - MakeAsChar(&p->opt[posMem]) - p->opt[posMem].posPrev = posMem - 1; - if (p->opt[cur].prev2) + wr--; + p->opt[wr].len = (UInt32)len; + cur -= extra; + len = extra; + if (extra == 1) { - p->opt[posMem - 1].prev1IsChar = False; - p->opt[posMem - 1].posPrev = p->opt[cur].posPrev2; - p->opt[posMem - 1].backPrev = p->opt[cur].backPrev2; + p->opt[wr].dist = dist; + dist = MARK_LIT; + } + else + { + p->opt[wr].dist = 0; + len--; + wr--; + p->opt[wr].dist = MARK_LIT; + p->opt[wr].len = 1; } } + + if (cur == 0) { - UInt32 posPrev = posMem; - UInt32 backCur = backMem; - - backMem = p->opt[posPrev].backPrev; - posMem = p->opt[posPrev].posPrev; - - p->opt[posPrev].backPrev = backCur; - p->opt[posPrev].posPrev = cur; - cur = posPrev; + p->backRes = dist; + p->optCur = wr; + return len; } + + wr--; + p->opt[wr].dist = dist; + p->opt[wr].len = (UInt32)len; } - while (cur != 0); - *backRes = p->opt[0].backPrev; - p->optimumCurrentIndex = p->opt[0].posPrev; - return p->optimumCurrentIndex; } -#define LIT_PROBS(pos, prevByte) (p->litProbs + ((((pos) & p->lpMask) << p->lc) + ((prevByte) >> (8 - p->lc))) * (UInt32)0x300) -static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes) -{ - UInt32 lenEnd, cur; - UInt32 reps[LZMA_NUM_REPS], repLens[LZMA_NUM_REPS]; - UInt32 *matches; - { +#define LIT_PROBS(pos, prevByte) \ + (p->litProbs + (UInt32)3 * (((((pos) << 8) + (prevByte)) & p->lpMask) << p->lc)) - UInt32 numAvail, mainLen, numPairs, repMaxIndex, i, posState, len; - UInt32 matchPrice, repMatchPrice, normalMatchPrice; - const Byte *data; - Byte curByte, matchByte; - if (p->optimumEndIndex != p->optimumCurrentIndex) - { - const COptimal *opt = &p->opt[p->optimumCurrentIndex]; - UInt32 lenRes = opt->posPrev - p->optimumCurrentIndex; - *backRes = opt->backPrev; - p->optimumCurrentIndex = opt->posPrev; - return lenRes; - } - p->optimumCurrentIndex = p->optimumEndIndex = 0; - - if (p->additionalOffset == 0) - mainLen = ReadMatchDistances(p, &numPairs); - else - { - mainLen = p->longestMatchLength; - numPairs = p->numPairs; - } - - numAvail = p->numAvail; - if (numAvail < 2) - { - *backRes = (UInt32)(-1); - return 1; - } - if (numAvail > LZMA_MATCH_LEN_MAX) - numAvail = LZMA_MATCH_LEN_MAX; +static unsigned GetOptimum(CLzmaEnc *p, UInt32 position) +{ + unsigned last, cur; + UInt32 reps[LZMA_NUM_REPS]; + unsigned repLens[LZMA_NUM_REPS]; + UInt32 *matches; - data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; - repMaxIndex = 0; - for (i = 0; i < LZMA_NUM_REPS; i++) { - UInt32 lenTest; - const Byte *data2; - reps[i] = p->reps[i]; - data2 = data - reps[i] - 1; - if (data[0] != data2[0] || data[1] != data2[1]) + UInt32 numAvail; + unsigned numPairs, mainLen, repMaxIndex, i, posState; + UInt32 matchPrice, repMatchPrice; + const Byte *data; + Byte curByte, matchByte; + + p->optCur = p->optEnd = 0; + + if (p->additionalOffset == 0) + mainLen = ReadMatchDistances(p, &numPairs); + else { - repLens[i] = 0; - continue; + mainLen = p->longestMatchLen; + numPairs = p->numPairs; } - for (lenTest = 2; lenTest < numAvail && data[lenTest] == data2[lenTest]; lenTest++); - repLens[i] = lenTest; - if (lenTest > repLens[repMaxIndex]) - repMaxIndex = i; - } - if (repLens[repMaxIndex] >= p->numFastBytes) - { - UInt32 lenRes; - *backRes = repMaxIndex; - lenRes = repLens[repMaxIndex]; - MovePos(p, lenRes - 1); - return lenRes; - } - - matches = p->matches; - if (mainLen >= p->numFastBytes) - { - *backRes = matches[numPairs - 1] + LZMA_NUM_REPS; - MovePos(p, mainLen - 1); - return mainLen; - } - curByte = *data; - matchByte = *(data - (reps[0] + 1)); - - if (mainLen < 2 && curByte != matchByte && repLens[repMaxIndex] < 2) - { - *backRes = (UInt32)-1; - return 1; - } - - p->opt[0].state = (CState)p->state; - - posState = (position & p->pbMask); - - { - const CLzmaProb *probs = LIT_PROBS(position, *(data - 1)); - p->opt[1].price = GET_PRICE_0(p->isMatch[p->state][posState]) + - (!IsCharState(p->state) ? - LitEnc_GetPriceMatched(probs, curByte, matchByte, p->ProbPrices) : - LitEnc_GetPrice(probs, curByte, p->ProbPrices)); - } - - MakeAsChar(&p->opt[1]); - - matchPrice = GET_PRICE_1(p->isMatch[p->state][posState]); - repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[p->state]); - - if (matchByte == curByte) - { - UInt32 shortRepPrice = repMatchPrice + GetRepLen1Price(p, p->state, posState); - if (shortRepPrice < p->opt[1].price) + + numAvail = p->numAvail; + if (numAvail < 2) { - p->opt[1].price = shortRepPrice; - MakeAsShortRep(&p->opt[1]); + p->backRes = MARK_LIT; + return 1; } - } - lenEnd = ((mainLen >= repLens[repMaxIndex]) ? mainLen : repLens[repMaxIndex]); - - if (lenEnd < 2) - { - *backRes = p->opt[1].backPrev; - return 1; - } - - p->opt[1].posPrev = 0; - for (i = 0; i < LZMA_NUM_REPS; i++) - p->opt[0].backs[i] = reps[i]; - - len = lenEnd; - do - p->opt[len--].price = kInfinityPrice; - while (len >= 2); - - for (i = 0; i < LZMA_NUM_REPS; i++) - { - UInt32 repLen = repLens[i]; - UInt32 price; - if (repLen < 2) - continue; - price = repMatchPrice + GetPureRepPrice(p, i, p->state, posState); - do + if (numAvail > LZMA_MATCH_LEN_MAX) + numAvail = LZMA_MATCH_LEN_MAX; + + data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; + repMaxIndex = 0; + + for (i = 0; i < LZMA_NUM_REPS; i++) { - UInt32 curAndLenPrice = price + p->repLenEnc.prices[posState][repLen - 2]; - COptimal *opt = &p->opt[repLen]; - if (curAndLenPrice < opt->price) + unsigned len; + const Byte *data2; + reps[i] = p->reps[i]; + data2 = data - reps[i]; + if (data[0] != data2[0] || data[1] != data2[1]) { - opt->price = curAndLenPrice; - opt->posPrev = 0; - opt->backPrev = i; - opt->prev1IsChar = False; + repLens[i] = 0; + continue; } + for (len = 2; len < numAvail && data[len] == data2[len]; len++) + {} + repLens[i] = len; + if (len > repLens[repMaxIndex]) + repMaxIndex = i; + if (len == LZMA_MATCH_LEN_MAX) // 21.03 : optimization + break; } - while (--repLen >= 2); - } - - normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[p->state]); + + if (repLens[repMaxIndex] >= p->numFastBytes) + { + unsigned len; + p->backRes = (UInt32)repMaxIndex; + len = repLens[repMaxIndex]; + MOVE_POS(p, len - 1) + return len; + } + + matches = p->matches; + #define MATCHES matches + // #define MATCHES p->matches + + if (mainLen >= p->numFastBytes) + { + p->backRes = MATCHES[(size_t)numPairs - 1] + LZMA_NUM_REPS; + MOVE_POS(p, mainLen - 1) + return mainLen; + } + + curByte = *data; + matchByte = *(data - reps[0]); - len = ((repLens[0] >= 2) ? repLens[0] + 1 : 2); - if (len <= mainLen) - { - UInt32 offs = 0; - while (len > matches[offs]) - offs += 2; - for (; ; len++) + last = repLens[repMaxIndex]; + if (last <= mainLen) + last = mainLen; + + if (last < 2 && curByte != matchByte) + { + p->backRes = MARK_LIT; + return 1; + } + + p->opt[0].state = (CState)p->state; + + posState = (position & p->pbMask); + { - COptimal *opt; - UInt32 distance = matches[offs + 1]; + const CLzmaProb *probs = LIT_PROBS(position, *(data - 1)); + p->opt[1].price = GET_PRICE_0(p->isMatch[p->state][posState]) + + (!IsLitState(p->state) ? + LitEnc_Matched_GetPrice(probs, curByte, matchByte, p->ProbPrices) : + LitEnc_GetPrice(probs, curByte, p->ProbPrices)); + } - UInt32 curAndLenPrice = normalMatchPrice + p->lenEnc.prices[posState][len - LZMA_MATCH_LEN_MIN]; - UInt32 lenToPosState = GetLenToPosState(len); - if (distance < kNumFullDistances) - curAndLenPrice += p->distancesPrices[lenToPosState][distance]; - else + MakeAs_Lit(&p->opt[1]) + + matchPrice = GET_PRICE_1(p->isMatch[p->state][posState]); + repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[p->state]); + + // 18.06 + if (matchByte == curByte && repLens[0] == 0) + { + UInt32 shortRepPrice = repMatchPrice + GetPrice_ShortRep(p, p->state, posState); + if (shortRepPrice < p->opt[1].price) { - UInt32 slot; - GetPosSlot2(distance, slot); - curAndLenPrice += p->alignPrices[distance & kAlignMask] + p->posSlotPrices[lenToPosState][slot]; + p->opt[1].price = shortRepPrice; + MakeAs_ShortRep(&p->opt[1]) } - opt = &p->opt[len]; - if (curAndLenPrice < opt->price) + if (last < 2) { - opt->price = curAndLenPrice; - opt->posPrev = 0; - opt->backPrev = distance + LZMA_NUM_REPS; - opt->prev1IsChar = False; + p->backRes = p->opt[1].dist; + return 1; } - if (len == matches[offs]) + } + + p->opt[1].len = 1; + + p->opt[0].reps[0] = reps[0]; + p->opt[0].reps[1] = reps[1]; + p->opt[0].reps[2] = reps[2]; + p->opt[0].reps[3] = reps[3]; + + // ---------- REP ---------- + + for (i = 0; i < LZMA_NUM_REPS; i++) + { + unsigned repLen = repLens[i]; + UInt32 price; + if (repLen < 2) + continue; + price = repMatchPrice + GetPrice_PureRep(p, i, p->state, posState); + do { - offs += 2; - if (offs == numPairs) - break; + UInt32 price2 = price + GET_PRICE_LEN(&p->repLenEnc, posState, repLen); + COptimal *opt = &p->opt[repLen]; + if (price2 < opt->price) + { + opt->price = price2; + opt->len = (UInt32)repLen; + opt->dist = (UInt32)i; + opt->extra = 0; + } } + while (--repLen >= 2); } - } + + + // ---------- MATCH ---------- + { + unsigned len = repLens[0] + 1; + if (len <= mainLen) + { + unsigned offs = 0; + UInt32 normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[p->state]); - cur = 0; + if (len < 2) + len = 2; + else + while (len > MATCHES[offs]) + offs += 2; + + for (; ; len++) + { + COptimal *opt; + UInt32 dist = MATCHES[(size_t)offs + 1]; + UInt32 price = normalMatchPrice + GET_PRICE_LEN(&p->lenEnc, posState, len); + unsigned lenToPosState = GetLenToPosState(len); + + if (dist < kNumFullDistances) + price += p->distancesPrices[lenToPosState][dist & (kNumFullDistances - 1)]; + else + { + unsigned slot; + GetPosSlot2(dist, slot) + price += p->alignPrices[dist & kAlignMask]; + price += p->posSlotPrices[lenToPosState][slot]; + } + + opt = &p->opt[len]; + + if (price < opt->price) + { + opt->price = price; + opt->len = (UInt32)len; + opt->dist = dist + LZMA_NUM_REPS; + opt->extra = 0; + } + + if (len == MATCHES[offs]) + { + offs += 2; + if (offs == numPairs) + break; + } + } + } + } + + + cur = 0; #ifdef SHOW_STAT2 /* if (position >= 0) */ { unsigned i; printf("\n pos = %4X", position); - for (i = cur; i <= lenEnd; i++) + for (i = cur; i <= last; i++) printf("\nprice[%4X] = %u", position - cur + i, p->opt[i].price); } #endif - } + + + // ---------- Optimal Parsing ---------- + for (;;) { - UInt32 numAvail; - UInt32 numAvailFull, newLen, numPairs, posPrev, state, posState, startLen; - UInt32 curPrice, curAnd1Price, matchPrice, repMatchPrice; - Bool nextIsChar; + unsigned numAvail; + UInt32 numAvailFull; + unsigned newLen, numPairs, prev, state, posState, startLen; + UInt32 litPrice, matchPrice, repMatchPrice; + BoolInt nextIsLit; Byte curByte, matchByte; const Byte *data; - COptimal *curOpt; - COptimal *nextOpt; + COptimal *curOpt, *nextOpt; - cur++; - if (cur == lenEnd) - return Backward(p, backRes, cur); + if (++cur == last) + break; + + // 18.06 + if (cur >= kNumOpts - 64) + { + unsigned j, best; + UInt32 price = p->opt[cur].price; + best = cur; + for (j = cur + 1; j <= last; j++) + { + UInt32 price2 = p->opt[j].price; + if (price >= price2) + { + price = price2; + best = j; + } + } + { + unsigned delta = best - cur; + if (delta != 0) + { + MOVE_POS(p, delta) + } + } + cur = best; + break; + } newLen = ReadMatchDistances(p, &numPairs); + if (newLen >= p->numFastBytes) { p->numPairs = numPairs; - p->longestMatchLength = newLen; - return Backward(p, backRes, cur); + p->longestMatchLen = newLen; + break; } - position++; + curOpt = &p->opt[cur]; - posPrev = curOpt->posPrev; - if (curOpt->prev1IsChar) - { - posPrev--; - if (curOpt->prev2) - { - state = p->opt[curOpt->posPrev2].state; - if (curOpt->backPrev2 < LZMA_NUM_REPS) - state = kRepNextStates[state]; - else - state = kMatchNextStates[state]; - } - else - state = p->opt[posPrev].state; - state = kLiteralNextStates[state]; - } - else - state = p->opt[posPrev].state; - if (posPrev == cur - 1) + + position++; + + // we need that check here, if skip_items in p->opt are possible + /* + if (curOpt->price >= kInfinityPrice) + continue; + */ + + prev = cur - curOpt->len; + + if (curOpt->len == 1) { + state = (unsigned)p->opt[prev].state; if (IsShortRep(curOpt)) state = kShortRepNextStates[state]; else @@ -1234,357 +1510,499 @@ static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes) } else { - UInt32 pos; const COptimal *prevOpt; - if (curOpt->prev1IsChar && curOpt->prev2) + UInt32 b0; + UInt32 dist = curOpt->dist; + + if (curOpt->extra) { - posPrev = curOpt->posPrev2; - pos = curOpt->backPrev2; - state = kRepNextStates[state]; + prev -= (unsigned)curOpt->extra; + state = kState_RepAfterLit; + if (curOpt->extra == 1) + state = (dist < LZMA_NUM_REPS ? kState_RepAfterLit : kState_MatchAfterLit); } else { - pos = curOpt->backPrev; - if (pos < LZMA_NUM_REPS) + state = (unsigned)p->opt[prev].state; + if (dist < LZMA_NUM_REPS) state = kRepNextStates[state]; else state = kMatchNextStates[state]; } - prevOpt = &p->opt[posPrev]; - if (pos < LZMA_NUM_REPS) + + prevOpt = &p->opt[prev]; + b0 = prevOpt->reps[0]; + + if (dist < LZMA_NUM_REPS) { - UInt32 i; - reps[0] = prevOpt->backs[pos]; - for (i = 1; i <= pos; i++) - reps[i] = prevOpt->backs[i - 1]; - for (; i < LZMA_NUM_REPS; i++) - reps[i] = prevOpt->backs[i]; + if (dist == 0) + { + reps[0] = b0; + reps[1] = prevOpt->reps[1]; + reps[2] = prevOpt->reps[2]; + reps[3] = prevOpt->reps[3]; + } + else + { + reps[1] = b0; + b0 = prevOpt->reps[1]; + if (dist == 1) + { + reps[0] = b0; + reps[2] = prevOpt->reps[2]; + reps[3] = prevOpt->reps[3]; + } + else + { + reps[2] = b0; + reps[0] = prevOpt->reps[dist]; + reps[3] = prevOpt->reps[dist ^ 1]; + } + } } else { - UInt32 i; - reps[0] = (pos - LZMA_NUM_REPS); - for (i = 1; i < LZMA_NUM_REPS; i++) - reps[i] = prevOpt->backs[i - 1]; + reps[0] = (dist - LZMA_NUM_REPS + 1); + reps[1] = b0; + reps[2] = prevOpt->reps[1]; + reps[3] = prevOpt->reps[2]; } } + curOpt->state = (CState)state; + curOpt->reps[0] = reps[0]; + curOpt->reps[1] = reps[1]; + curOpt->reps[2] = reps[2]; + curOpt->reps[3] = reps[3]; - curOpt->backs[0] = reps[0]; - curOpt->backs[1] = reps[1]; - curOpt->backs[2] = reps[2]; - curOpt->backs[3] = reps[3]; - - curPrice = curOpt->price; - nextIsChar = False; data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; curByte = *data; - matchByte = *(data - (reps[0] + 1)); + matchByte = *(data - reps[0]); posState = (position & p->pbMask); - curAnd1Price = curPrice + GET_PRICE_0(p->isMatch[state][posState]); + /* + The order of Price checks: + < LIT + <= SHORT_REP + < LIT : REP_0 + < REP [ : LIT : REP_0 ] + < MATCH [ : LIT : REP_0 ] + */ + { - const CLzmaProb *probs = LIT_PROBS(position, *(data - 1)); - curAnd1Price += - (!IsCharState(state) ? - LitEnc_GetPriceMatched(probs, curByte, matchByte, p->ProbPrices) : - LitEnc_GetPrice(probs, curByte, p->ProbPrices)); + UInt32 curPrice = curOpt->price; + unsigned prob = p->isMatch[state][posState]; + matchPrice = curPrice + GET_PRICE_1(prob); + litPrice = curPrice + GET_PRICE_0(prob); } - nextOpt = &p->opt[cur + 1]; - - if (curAnd1Price < nextOpt->price) + nextOpt = &p->opt[(size_t)cur + 1]; + nextIsLit = False; + + // here we can allow skip_items in p->opt, if we don't check (nextOpt->price < kInfinityPrice) + // 18.new.06 + if ((nextOpt->price < kInfinityPrice + // && !IsLitState(state) + && matchByte == curByte) + || litPrice > nextOpt->price + ) + litPrice = 0; + else { - nextOpt->price = curAnd1Price; - nextOpt->posPrev = cur; - MakeAsChar(nextOpt); - nextIsChar = True; + const CLzmaProb *probs = LIT_PROBS(position, *(data - 1)); + litPrice += (!IsLitState(state) ? + LitEnc_Matched_GetPrice(probs, curByte, matchByte, p->ProbPrices) : + LitEnc_GetPrice(probs, curByte, p->ProbPrices)); + + if (litPrice < nextOpt->price) + { + nextOpt->price = litPrice; + nextOpt->len = 1; + MakeAs_Lit(nextOpt) + nextIsLit = True; + } } - matchPrice = curPrice + GET_PRICE_1(p->isMatch[state][posState]); repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[state]); - if (matchByte == curByte && !(nextOpt->posPrev < cur && nextOpt->backPrev == 0)) + numAvailFull = p->numAvail; + { + unsigned temp = kNumOpts - 1 - cur; + if (numAvailFull > temp) + numAvailFull = (UInt32)temp; + } + + // 18.06 + // ---------- SHORT_REP ---------- + if (IsLitState(state)) // 18.new + if (matchByte == curByte) + if (repMatchPrice < nextOpt->price) // 18.new + // if (numAvailFull < 2 || data[1] != *(data - reps[0] + 1)) + if ( + // nextOpt->price >= kInfinityPrice || + nextOpt->len < 2 // we can check nextOpt->len, if skip items are not allowed in p->opt + || (nextOpt->dist != 0 + // && nextOpt->extra <= 1 // 17.old + ) + ) { - UInt32 shortRepPrice = repMatchPrice + GetRepLen1Price(p, state, posState); - if (shortRepPrice <= nextOpt->price) + UInt32 shortRepPrice = repMatchPrice + GetPrice_ShortRep(p, state, posState); + // if (shortRepPrice <= nextOpt->price) // 17.old + if (shortRepPrice < nextOpt->price) // 18.new { nextOpt->price = shortRepPrice; - nextOpt->posPrev = cur; - MakeAsShortRep(nextOpt); - nextIsChar = True; + nextOpt->len = 1; + MakeAs_ShortRep(nextOpt) + nextIsLit = False; } } - numAvailFull = p->numAvail; - { - UInt32 temp = kNumOpts - 1 - cur; - if (temp < numAvailFull) - numAvailFull = temp; - } - + if (numAvailFull < 2) continue; numAvail = (numAvailFull <= p->numFastBytes ? numAvailFull : p->numFastBytes); - if (!nextIsChar && matchByte != curByte) /* speed optimization */ + // numAvail <= p->numFastBytes + + // ---------- LIT : REP_0 ---------- + + if (!nextIsLit + && litPrice != 0 // 18.new + && matchByte != curByte + && numAvailFull > 2) { - /* try Literal + rep0 */ - UInt32 temp; - UInt32 lenTest2; - const Byte *data2 = data - reps[0] - 1; - UInt32 limit = p->numFastBytes + 1; - if (limit > numAvailFull) - limit = numAvailFull; - - for (temp = 1; temp < limit && data[temp] == data2[temp]; temp++); - lenTest2 = temp - 1; - if (lenTest2 >= 2) + const Byte *data2 = data - reps[0]; + if (data[1] == data2[1] && data[2] == data2[2]) { - UInt32 state2 = kLiteralNextStates[state]; - UInt32 posStateNext = (position + 1) & p->pbMask; - UInt32 nextRepMatchPrice = curAnd1Price + - GET_PRICE_1(p->isMatch[state2][posStateNext]) + - GET_PRICE_1(p->isRep[state2]); - /* for (; lenTest2 >= 2; lenTest2--) */ + unsigned len; + unsigned limit = p->numFastBytes + 1; + if (limit > numAvailFull) + limit = numAvailFull; + for (len = 3; len < limit && data[len] == data2[len]; len++) + {} + { - UInt32 curAndLenPrice; - COptimal *opt; - UInt32 offset = cur + 1 + lenTest2; - while (lenEnd < offset) - p->opt[++lenEnd].price = kInfinityPrice; - curAndLenPrice = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext); - opt = &p->opt[offset]; - if (curAndLenPrice < opt->price) + unsigned state2 = kLiteralNextStates[state]; + unsigned posState2 = (position + 1) & p->pbMask; + UInt32 price = litPrice + GetPrice_Rep_0(p, state2, posState2); { - opt->price = curAndLenPrice; - opt->posPrev = cur + 1; - opt->backPrev = 0; - opt->prev1IsChar = True; - opt->prev2 = False; + unsigned offset = cur + len; + + if (last < offset) + last = offset; + + // do + { + UInt32 price2; + COptimal *opt; + len--; + // price2 = price + GetPrice_Len_Rep_0(p, len, state2, posState2); + price2 = price + GET_PRICE_LEN(&p->repLenEnc, posState2, len); + + opt = &p->opt[offset]; + // offset--; + if (price2 < opt->price) + { + opt->price = price2; + opt->len = (UInt32)len; + opt->dist = 0; + opt->extra = 1; + } + } + // while (len >= 3); } } } } startLen = 2; /* speed optimization */ + { - UInt32 repIndex; - for (repIndex = 0; repIndex < LZMA_NUM_REPS; repIndex++) - { - UInt32 lenTest; - UInt32 lenTestTemp; - UInt32 price; - const Byte *data2 = data - reps[repIndex] - 1; - if (data[0] != data2[0] || data[1] != data2[1]) - continue; - for (lenTest = 2; lenTest < numAvail && data[lenTest] == data2[lenTest]; lenTest++); - while (lenEnd < cur + lenTest) - p->opt[++lenEnd].price = kInfinityPrice; - lenTestTemp = lenTest; - price = repMatchPrice + GetPureRepPrice(p, repIndex, state, posState); - do + // ---------- REP ---------- + unsigned repIndex = 0; // 17.old + // unsigned repIndex = IsLitState(state) ? 0 : 1; // 18.notused + for (; repIndex < LZMA_NUM_REPS; repIndex++) { - UInt32 curAndLenPrice = price + p->repLenEnc.prices[posState][lenTest - 2]; - COptimal *opt = &p->opt[cur + lenTest]; - if (curAndLenPrice < opt->price) + unsigned len; + UInt32 price; + const Byte *data2 = data - reps[repIndex]; + if (data[0] != data2[0] || data[1] != data2[1]) + continue; + + for (len = 2; len < numAvail && data[len] == data2[len]; len++) + {} + + // if (len < startLen) continue; // 18.new: speed optimization + { - opt->price = curAndLenPrice; - opt->posPrev = cur; - opt->backPrev = repIndex; - opt->prev1IsChar = False; + unsigned offset = cur + len; + if (last < offset) + last = offset; + } + { + unsigned len2 = len; + price = repMatchPrice + GetPrice_PureRep(p, repIndex, state, posState); + do + { + UInt32 price2 = price + GET_PRICE_LEN(&p->repLenEnc, posState, len2); + COptimal *opt = &p->opt[cur + len2]; + if (price2 < opt->price) + { + opt->price = price2; + opt->len = (UInt32)len2; + opt->dist = (UInt32)repIndex; + opt->extra = 0; + } + } + while (--len2 >= 2); } - } - while (--lenTest >= 2); - lenTest = lenTestTemp; - - if (repIndex == 0) - startLen = lenTest + 1; - /* if (_maxMode) */ + if (repIndex == 0) startLen = len + 1; // 17.old + // startLen = len + 1; // 18.new + + /* if (_maxMode) */ { - UInt32 lenTest2 = lenTest + 1; - UInt32 limit = lenTest2 + p->numFastBytes; + // ---------- REP : LIT : REP_0 ---------- + // numFastBytes + 1 + numFastBytes + + unsigned len2 = len + 1; + unsigned limit = len2 + p->numFastBytes; if (limit > numAvailFull) limit = numAvailFull; - for (; lenTest2 < limit && data[lenTest2] == data2[lenTest2]; lenTest2++); - lenTest2 -= lenTest + 1; - if (lenTest2 >= 2) + + len2 += 2; + if (len2 <= limit) + if (data[len2 - 2] == data2[len2 - 2]) + if (data[len2 - 1] == data2[len2 - 1]) { - UInt32 nextRepMatchPrice; - UInt32 state2 = kRepNextStates[state]; - UInt32 posStateNext = (position + lenTest) & p->pbMask; - UInt32 curAndLenCharPrice = - price + p->repLenEnc.prices[posState][lenTest - 2] + - GET_PRICE_0(p->isMatch[state2][posStateNext]) + - LitEnc_GetPriceMatched(LIT_PROBS(position + lenTest, data[lenTest - 1]), - data[lenTest], data2[lenTest], p->ProbPrices); - state2 = kLiteralNextStates[state2]; - posStateNext = (position + lenTest + 1) & p->pbMask; - nextRepMatchPrice = curAndLenCharPrice + - GET_PRICE_1(p->isMatch[state2][posStateNext]) + - GET_PRICE_1(p->isRep[state2]); + unsigned state2 = kRepNextStates[state]; + unsigned posState2 = (position + len) & p->pbMask; + price += GET_PRICE_LEN(&p->repLenEnc, posState, len) + + GET_PRICE_0(p->isMatch[state2][posState2]) + + LitEnc_Matched_GetPrice(LIT_PROBS(position + len, data[(size_t)len - 1]), + data[len], data2[len], p->ProbPrices); - /* for (; lenTest2 >= 2; lenTest2--) */ + // state2 = kLiteralNextStates[state2]; + state2 = kState_LitAfterRep; + posState2 = (posState2 + 1) & p->pbMask; + + + price += GetPrice_Rep_0(p, state2, posState2); + + for (; len2 < limit && data[len2] == data2[len2]; len2++) + {} + + len2 -= len; + // if (len2 >= 3) + { { - UInt32 curAndLenPrice; - COptimal *opt; - UInt32 offset = cur + lenTest + 1 + lenTest2; - while (lenEnd < offset) - p->opt[++lenEnd].price = kInfinityPrice; - curAndLenPrice = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext); - opt = &p->opt[offset]; - if (curAndLenPrice < opt->price) + unsigned offset = cur + len + len2; + + if (last < offset) + last = offset; + // do { - opt->price = curAndLenPrice; - opt->posPrev = cur + lenTest + 1; - opt->backPrev = 0; - opt->prev1IsChar = True; - opt->prev2 = True; - opt->posPrev2 = cur; - opt->backPrev2 = repIndex; + UInt32 price2; + COptimal *opt; + len2--; + // price2 = price + GetPrice_Len_Rep_0(p, len2, state2, posState2); + price2 = price + GET_PRICE_LEN(&p->repLenEnc, posState2, len2); + + opt = &p->opt[offset]; + // offset--; + if (price2 < opt->price) + { + opt->price = price2; + opt->len = (UInt32)len2; + opt->extra = (CExtra)(len + 1); + opt->dist = (UInt32)repIndex; + } } + // while (len2 >= 3); } } + } } + } } - } - /* for (UInt32 lenTest = 2; lenTest <= newLen; lenTest++) */ + + + // ---------- MATCH ---------- + /* for (unsigned len = 2; len <= newLen; len++) */ if (newLen > numAvail) { newLen = numAvail; - for (numPairs = 0; newLen > matches[numPairs]; numPairs += 2); - matches[numPairs] = newLen; + for (numPairs = 0; newLen > MATCHES[numPairs]; numPairs += 2); + MATCHES[numPairs] = (UInt32)newLen; numPairs += 2; } + + // startLen = 2; /* speed optimization */ + if (newLen >= startLen) { UInt32 normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[state]); - UInt32 offs, curBack, posSlot; - UInt32 lenTest; - while (lenEnd < cur + newLen) - p->opt[++lenEnd].price = kInfinityPrice; + UInt32 dist; + unsigned offs, posSlot, len; + + { + unsigned offset = cur + newLen; + if (last < offset) + last = offset; + } offs = 0; - while (startLen > matches[offs]) + while (startLen > MATCHES[offs]) offs += 2; - curBack = matches[offs + 1]; - GetPosSlot2(curBack, posSlot); - for (lenTest = /*2*/ startLen; ; lenTest++) + dist = MATCHES[(size_t)offs + 1]; + + // if (dist >= kNumFullDistances) + GetPosSlot2(dist, posSlot) + + for (len = /*2*/ startLen; ; len++) { - UInt32 curAndLenPrice = normalMatchPrice + p->lenEnc.prices[posState][lenTest - LZMA_MATCH_LEN_MIN]; - { - UInt32 lenToPosState = GetLenToPosState(lenTest); - COptimal *opt; - if (curBack < kNumFullDistances) - curAndLenPrice += p->distancesPrices[lenToPosState][curBack]; - else - curAndLenPrice += p->posSlotPrices[lenToPosState][posSlot] + p->alignPrices[curBack & kAlignMask]; - - opt = &p->opt[cur + lenTest]; - if (curAndLenPrice < opt->price) + UInt32 price = normalMatchPrice + GET_PRICE_LEN(&p->lenEnc, posState, len); { - opt->price = curAndLenPrice; - opt->posPrev = cur; - opt->backPrev = curBack + LZMA_NUM_REPS; - opt->prev1IsChar = False; - } + COptimal *opt; + unsigned lenNorm = len - 2; + lenNorm = GetLenToPosState2(lenNorm); + if (dist < kNumFullDistances) + price += p->distancesPrices[lenNorm][dist & (kNumFullDistances - 1)]; + else + price += p->posSlotPrices[lenNorm][posSlot] + p->alignPrices[dist & kAlignMask]; + + opt = &p->opt[cur + len]; + if (price < opt->price) + { + opt->price = price; + opt->len = (UInt32)len; + opt->dist = dist + LZMA_NUM_REPS; + opt->extra = 0; + } } - if (/*_maxMode && */lenTest == matches[offs]) + if (len == MATCHES[offs]) { - /* Try Match + Literal + Rep0 */ - const Byte *data2 = data - curBack - 1; - UInt32 lenTest2 = lenTest + 1; - UInt32 limit = lenTest2 + p->numFastBytes; + // if (p->_maxMode) { + // MATCH : LIT : REP_0 + + const Byte *data2 = data - dist - 1; + unsigned len2 = len + 1; + unsigned limit = len2 + p->numFastBytes; if (limit > numAvailFull) limit = numAvailFull; - for (; lenTest2 < limit && data[lenTest2] == data2[lenTest2]; lenTest2++); - lenTest2 -= lenTest + 1; - if (lenTest2 >= 2) + + len2 += 2; + if (len2 <= limit) + if (data[len2 - 2] == data2[len2 - 2]) + if (data[len2 - 1] == data2[len2 - 1]) { - UInt32 nextRepMatchPrice; - UInt32 state2 = kMatchNextStates[state]; - UInt32 posStateNext = (position + lenTest) & p->pbMask; - UInt32 curAndLenCharPrice = curAndLenPrice + - GET_PRICE_0(p->isMatch[state2][posStateNext]) + - LitEnc_GetPriceMatched(LIT_PROBS(position + lenTest, data[lenTest - 1]), - data[lenTest], data2[lenTest], p->ProbPrices); - state2 = kLiteralNextStates[state2]; - posStateNext = (posStateNext + 1) & p->pbMask; - nextRepMatchPrice = curAndLenCharPrice + - GET_PRICE_1(p->isMatch[state2][posStateNext]) + - GET_PRICE_1(p->isRep[state2]); - - /* for (; lenTest2 >= 2; lenTest2--) */ + for (; len2 < limit && data[len2] == data2[len2]; len2++) + {} + + len2 -= len; + + // if (len2 >= 3) + { + unsigned state2 = kMatchNextStates[state]; + unsigned posState2 = (position + len) & p->pbMask; + unsigned offset; + price += GET_PRICE_0(p->isMatch[state2][posState2]); + price += LitEnc_Matched_GetPrice(LIT_PROBS(position + len, data[(size_t)len - 1]), + data[len], data2[len], p->ProbPrices); + + // state2 = kLiteralNextStates[state2]; + state2 = kState_LitAfterMatch; + + posState2 = (posState2 + 1) & p->pbMask; + price += GetPrice_Rep_0(p, state2, posState2); + + offset = cur + len + len2; + + if (last < offset) + last = offset; + // do { - UInt32 offset = cur + lenTest + 1 + lenTest2; - UInt32 curAndLenPrice2; + UInt32 price2; COptimal *opt; - while (lenEnd < offset) - p->opt[++lenEnd].price = kInfinityPrice; - curAndLenPrice2 = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext); + len2--; + // price2 = price + GetPrice_Len_Rep_0(p, len2, state2, posState2); + price2 = price + GET_PRICE_LEN(&p->repLenEnc, posState2, len2); opt = &p->opt[offset]; - if (curAndLenPrice2 < opt->price) + // offset--; + if (price2 < opt->price) { - opt->price = curAndLenPrice2; - opt->posPrev = cur + lenTest + 1; - opt->backPrev = 0; - opt->prev1IsChar = True; - opt->prev2 = True; - opt->posPrev2 = cur; - opt->backPrev2 = curBack + LZMA_NUM_REPS; + opt->price = price2; + opt->len = (UInt32)len2; + opt->extra = (CExtra)(len + 1); + opt->dist = dist + LZMA_NUM_REPS; } } + // while (len2 >= 3); + } + } + offs += 2; if (offs == numPairs) break; - curBack = matches[offs + 1]; - if (curBack >= kNumFullDistances) - GetPosSlot2(curBack, posSlot); + dist = MATCHES[(size_t)offs + 1]; + // if (dist >= kNumFullDistances) + GetPosSlot2(dist, posSlot) } } } } + + do + p->opt[last].price = kInfinityPrice; + while (--last); + + return Backward(p, cur); } + + #define ChangePair(smallDist, bigDist) (((bigDist) >> 7) > (smallDist)) -static UInt32 GetOptimumFast(CLzmaEnc *p, UInt32 *backRes) + + +static unsigned GetOptimumFast(CLzmaEnc *p) { - UInt32 numAvail, mainLen, mainDist, numPairs, repIndex, repLen, i; + UInt32 numAvail, mainDist; + unsigned mainLen, numPairs, repIndex, repLen, i; const Byte *data; - const UInt32 *matches; if (p->additionalOffset == 0) mainLen = ReadMatchDistances(p, &numPairs); else { - mainLen = p->longestMatchLength; + mainLen = p->longestMatchLen; numPairs = p->numPairs; } numAvail = p->numAvail; - *backRes = (UInt32)-1; + p->backRes = MARK_LIT; if (numAvail < 2) return 1; + // if (mainLen < 2 && p->state == 0) return 1; // 18.06.notused if (numAvail > LZMA_MATCH_LEN_MAX) numAvail = LZMA_MATCH_LEN_MAX; data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; - repLen = repIndex = 0; + for (i = 0; i < LZMA_NUM_REPS; i++) { - UInt32 len; - const Byte *data2 = data - p->reps[i] - 1; + unsigned len; + const Byte *data2 = data - p->reps[i]; if (data[0] != data2[0] || data[1] != data2[1]) continue; - for (len = 2; len < numAvail && data[len] == data2[len]; len++); + for (len = 2; len < numAvail && data[len] == data2[len]; len++) + {} if (len >= p->numFastBytes) { - *backRes = i; - MovePos(p, len - 1); + p->backRes = (UInt32)i; + MOVE_POS(p, len - 1) return len; } if (len > repLen) @@ -1594,98 +2012,182 @@ static UInt32 GetOptimumFast(CLzmaEnc *p, UInt32 *backRes) } } - matches = p->matches; if (mainLen >= p->numFastBytes) { - *backRes = matches[numPairs - 1] + LZMA_NUM_REPS; - MovePos(p, mainLen - 1); + p->backRes = p->matches[(size_t)numPairs - 1] + LZMA_NUM_REPS; + MOVE_POS(p, mainLen - 1) return mainLen; } mainDist = 0; /* for GCC */ + if (mainLen >= 2) { - mainDist = matches[numPairs - 1]; - while (numPairs > 2 && mainLen == matches[numPairs - 4] + 1) + mainDist = p->matches[(size_t)numPairs - 1]; + while (numPairs > 2) { - if (!ChangePair(matches[numPairs - 3], mainDist)) + UInt32 dist2; + if (mainLen != p->matches[(size_t)numPairs - 4] + 1) + break; + dist2 = p->matches[(size_t)numPairs - 3]; + if (!ChangePair(dist2, mainDist)) break; numPairs -= 2; - mainLen = matches[numPairs - 2]; - mainDist = matches[numPairs - 1]; + mainLen--; + mainDist = dist2; } if (mainLen == 2 && mainDist >= 0x80) mainLen = 1; } - if (repLen >= 2 && ( - (repLen + 1 >= mainLen) || - (repLen + 2 >= mainLen && mainDist >= (1 << 9)) || - (repLen + 3 >= mainLen && mainDist >= (1 << 15)))) + if (repLen >= 2) + if ( repLen + 1 >= mainLen + || (repLen + 2 >= mainLen && mainDist >= (1 << 9)) + || (repLen + 3 >= mainLen && mainDist >= (1 << 15))) { - *backRes = repIndex; - MovePos(p, repLen - 1); + p->backRes = (UInt32)repIndex; + MOVE_POS(p, repLen - 1) return repLen; } if (mainLen < 2 || numAvail <= 2) return 1; - p->longestMatchLength = ReadMatchDistances(p, &p->numPairs); - if (p->longestMatchLength >= 2) { - UInt32 newDistance = matches[p->numPairs - 1]; - if ((p->longestMatchLength >= mainLen && newDistance < mainDist) || - (p->longestMatchLength == mainLen + 1 && !ChangePair(mainDist, newDistance)) || - (p->longestMatchLength > mainLen + 1) || - (p->longestMatchLength + 1 >= mainLen && mainLen >= 3 && ChangePair(newDistance, mainDist))) - return 1; + unsigned len1 = ReadMatchDistances(p, &p->numPairs); + p->longestMatchLen = len1; + + if (len1 >= 2) + { + UInt32 newDist = p->matches[(size_t)p->numPairs - 1]; + if ( (len1 >= mainLen && newDist < mainDist) + || (len1 == mainLen + 1 && !ChangePair(mainDist, newDist)) + || (len1 > mainLen + 1) + || (len1 + 1 >= mainLen && mainLen >= 3 && ChangePair(newDist, mainDist))) + return 1; + } } data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1; + for (i = 0; i < LZMA_NUM_REPS; i++) { - UInt32 len, limit; - const Byte *data2 = data - p->reps[i] - 1; + unsigned len, limit; + const Byte *data2 = data - p->reps[i]; if (data[0] != data2[0] || data[1] != data2[1]) continue; limit = mainLen - 1; - for (len = 2; len < limit && data[len] == data2[len]; len++); - if (len >= limit) - return 1; + for (len = 2;; len++) + { + if (len >= limit) + return 1; + if (data[len] != data2[len]) + break; + } + } + + p->backRes = mainDist + LZMA_NUM_REPS; + if (mainLen != 2) + { + MOVE_POS(p, mainLen - 2) } - *backRes = mainDist + LZMA_NUM_REPS; - MovePos(p, mainLen - 2); return mainLen; } -static void WriteEndMarker(CLzmaEnc *p, UInt32 posState) + + + +static void WriteEndMarker(CLzmaEnc *p, unsigned posState) { - UInt32 len; - RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 1); - RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 0); + UInt32 range; + range = p->rc.range; + { + UInt32 ttt, newBound; + CLzmaProb *prob = &p->isMatch[p->state][posState]; + RC_BIT_PRE(&p->rc, prob) + RC_BIT_1(&p->rc, prob) + prob = &p->isRep[p->state]; + RC_BIT_PRE(&p->rc, prob) + RC_BIT_0(&p->rc, prob) + } p->state = kMatchNextStates[p->state]; - len = LZMA_MATCH_LEN_MIN; - LenEnc_Encode2(&p->lenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices); - RcTree_Encode(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], kNumPosSlotBits, (1 << kNumPosSlotBits) - 1); - RangeEnc_EncodeDirectBits(&p->rc, (((UInt32)1 << 30) - 1) >> kNumAlignBits, 30 - kNumAlignBits); - RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, kAlignMask); + + p->rc.range = range; + LenEnc_Encode(&p->lenProbs, &p->rc, 0, posState); + range = p->rc.range; + + { + // RcTree_Encode_PosSlot(&p->rc, p->posSlotEncoder[0], (1 << kNumPosSlotBits) - 1); + CLzmaProb *probs = p->posSlotEncoder[0]; + unsigned m = 1; + do + { + UInt32 ttt, newBound; + RC_BIT_PRE(p, probs + m) + RC_BIT_1(&p->rc, probs + m) + m = (m << 1) + 1; + } + while (m < (1 << kNumPosSlotBits)); + } + { + // RangeEnc_EncodeDirectBits(&p->rc, ((UInt32)1 << (30 - kNumAlignBits)) - 1, 30 - kNumAlignBits); UInt32 range = p->range; + unsigned numBits = 30 - kNumAlignBits; + do + { + range >>= 1; + p->rc.low += range; + RC_NORM(&p->rc) + } + while (--numBits); + } + + { + // RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, kAlignMask); + CLzmaProb *probs = p->posAlignEncoder; + unsigned m = 1; + do + { + UInt32 ttt, newBound; + RC_BIT_PRE(p, probs + m) + RC_BIT_1(&p->rc, probs + m) + m = (m << 1) + 1; + } + while (m < kAlignTableSize); + } + p->rc.range = range; } + static SRes CheckErrors(CLzmaEnc *p) { if (p->result != SZ_OK) return p->result; if (p->rc.res != SZ_OK) p->result = SZ_ERROR_WRITE; - if (p->matchFinderBase.result != SZ_OK) + + #ifndef Z7_ST + if ( + // p->mf_Failure || + (p->mtMode && + ( // p->matchFinderMt.failure_LZ_LZ || + p->matchFinderMt.failure_LZ_BT)) + ) + { + p->result = MY_HRES_ERROR_INTERNAL_ERROR; + // printf("\nCheckErrors p->matchFinderMt.failureLZ\n"); + } + #endif + + if (MFB.result != SZ_OK) p->result = SZ_ERROR_READ; + if (p->result != SZ_OK) p->finished = True; return p->result; } -static SRes Flush(CLzmaEnc *p, UInt32 nowPos) + +Z7_NO_INLINE static SRes Flush(CLzmaEnc *p, UInt32 nowPos) { /* ReleaseMFStream(); */ p->finished = True; @@ -1696,61 +2198,140 @@ static SRes Flush(CLzmaEnc *p, UInt32 nowPos) return CheckErrors(p); } -static void FillAlignPrices(CLzmaEnc *p) + +Z7_NO_INLINE static void FillAlignPrices(CLzmaEnc *p) { - UInt32 i; - for (i = 0; i < kAlignTableSize; i++) - p->alignPrices[i] = RcTree_ReverseGetPrice(p->posAlignEncoder, kNumAlignBits, i, p->ProbPrices); - p->alignPriceCount = 0; + unsigned i; + const CProbPrice *ProbPrices = p->ProbPrices; + const CLzmaProb *probs = p->posAlignEncoder; + // p->alignPriceCount = 0; + for (i = 0; i < kAlignTableSize / 2; i++) + { + UInt32 price = 0; + unsigned sym = i; + unsigned m = 1; + unsigned bit; + UInt32 prob; + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit; + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit; + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit; + prob = probs[m]; + p->alignPrices[i ] = price + GET_PRICEa_0(prob); + p->alignPrices[i + 8] = price + GET_PRICEa_1(prob); + // p->alignPrices[i] = RcTree_ReverseGetPrice(p->posAlignEncoder, kNumAlignBits, i, p->ProbPrices); + } } -static void FillDistancesPrices(CLzmaEnc *p) + +Z7_NO_INLINE static void FillDistancesPrices(CLzmaEnc *p) { + // int y; for (y = 0; y < 100; y++) { + UInt32 tempPrices[kNumFullDistances]; - UInt32 i, lenToPosState; - for (i = kStartPosModelIndex; i < kNumFullDistances; i++) - { - UInt32 posSlot = GetPosSlot1(i); - UInt32 footerBits = ((posSlot >> 1) - 1); - UInt32 base = ((2 | (posSlot & 1)) << footerBits); - tempPrices[i] = RcTree_ReverseGetPrice(p->posEncoders + base - posSlot - 1, footerBits, i - base, p->ProbPrices); + unsigned i, lps; + + const CProbPrice *ProbPrices = p->ProbPrices; + p->matchPriceCount = 0; + + for (i = kStartPosModelIndex / 2; i < kNumFullDistances / 2; i++) + { + unsigned posSlot = GetPosSlot1(i); + unsigned footerBits = (posSlot >> 1) - 1; + unsigned base = ((2 | (posSlot & 1)) << footerBits); + const CLzmaProb *probs = p->posEncoders + (size_t)base * 2; + // tempPrices[i] = RcTree_ReverseGetPrice(p->posEncoders + base, footerBits, i - base, p->ProbPrices); + UInt32 price = 0; + unsigned m = 1; + unsigned sym = i; + unsigned offset = (unsigned)1 << footerBits; + base += i; + + if (footerBits) + do + { + unsigned bit = sym & 1; + sym >>= 1; + price += GET_PRICEa(probs[m], bit); + m = (m << 1) + bit; + } + while (--footerBits); + + { + unsigned prob = probs[m]; + tempPrices[base ] = price + GET_PRICEa_0(prob); + tempPrices[base + offset] = price + GET_PRICEa_1(prob); + } } - for (lenToPosState = 0; lenToPosState < kNumLenToPosStates; lenToPosState++) + for (lps = 0; lps < kNumLenToPosStates; lps++) { - UInt32 posSlot; - const CLzmaProb *encoder = p->posSlotEncoder[lenToPosState]; - UInt32 *posSlotPrices = p->posSlotPrices[lenToPosState]; - for (posSlot = 0; posSlot < p->distTableSize; posSlot++) - posSlotPrices[posSlot] = RcTree_GetPrice(encoder, kNumPosSlotBits, posSlot, p->ProbPrices); - for (posSlot = kEndPosModelIndex; posSlot < p->distTableSize; posSlot++) - posSlotPrices[posSlot] += ((((posSlot >> 1) - 1) - kNumAlignBits) << kNumBitPriceShiftBits); + unsigned slot; + unsigned distTableSize2 = (p->distTableSize + 1) >> 1; + UInt32 *posSlotPrices = p->posSlotPrices[lps]; + const CLzmaProb *probs = p->posSlotEncoder[lps]; + + for (slot = 0; slot < distTableSize2; slot++) + { + // posSlotPrices[slot] = RcTree_GetPrice(encoder, kNumPosSlotBits, slot, p->ProbPrices); + UInt32 price; + unsigned bit; + unsigned sym = slot + (1 << (kNumPosSlotBits - 1)); + unsigned prob; + bit = sym & 1; sym >>= 1; price = GET_PRICEa(probs[sym], bit); + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[sym], bit); + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[sym], bit); + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[sym], bit); + bit = sym & 1; sym >>= 1; price += GET_PRICEa(probs[sym], bit); + prob = probs[(size_t)slot + (1 << (kNumPosSlotBits - 1))]; + posSlotPrices[(size_t)slot * 2 ] = price + GET_PRICEa_0(prob); + posSlotPrices[(size_t)slot * 2 + 1] = price + GET_PRICEa_1(prob); + } + + { + UInt32 delta = ((UInt32)((kEndPosModelIndex / 2 - 1) - kNumAlignBits) << kNumBitPriceShiftBits); + for (slot = kEndPosModelIndex / 2; slot < distTableSize2; slot++) + { + posSlotPrices[(size_t)slot * 2 ] += delta; + posSlotPrices[(size_t)slot * 2 + 1] += delta; + delta += ((UInt32)1 << kNumBitPriceShiftBits); + } + } { - UInt32 *distancesPrices = p->distancesPrices[lenToPosState]; - for (i = 0; i < kStartPosModelIndex; i++) - distancesPrices[i] = posSlotPrices[i]; - for (; i < kNumFullDistances; i++) - distancesPrices[i] = posSlotPrices[GetPosSlot1(i)] + tempPrices[i]; + UInt32 *dp = p->distancesPrices[lps]; + + dp[0] = posSlotPrices[0]; + dp[1] = posSlotPrices[1]; + dp[2] = posSlotPrices[2]; + dp[3] = posSlotPrices[3]; + + for (i = 4; i < kNumFullDistances; i += 2) + { + UInt32 slotPrice = posSlotPrices[GetPosSlot1(i)]; + dp[i ] = slotPrice + tempPrices[i]; + dp[i + 1] = slotPrice + tempPrices[i + 1]; + } } } - p->matchPriceCount = 0; + // } } -void LzmaEnc_Construct(CLzmaEnc *p) + + +static void LzmaEnc_Construct(CLzmaEnc *p) { RangeEnc_Construct(&p->rc); - MatchFinder_Construct(&p->matchFinderBase); + MatchFinder_Construct(&MFB); - #ifndef _7ZIP_ST + #ifndef Z7_ST + p->matchFinderMt.MatchFinder = &MFB; MatchFinderMt_Construct(&p->matchFinderMt); - p->matchFinderMt.MatchFinder = &p->matchFinderBase; #endif { CLzmaEncProps props; LzmaEncProps_Init(&props); - LzmaEnc_SetProps(p, &props); + LzmaEnc_SetProps((CLzmaEncHandle)(void *)p, &props); } #ifndef LZMA_LOG_BSR @@ -1762,65 +2343,74 @@ void LzmaEnc_Construct(CLzmaEnc *p) p->saveState.litProbs = NULL; } -CLzmaEncHandle LzmaEnc_Create(ISzAlloc *alloc) +CLzmaEncHandle LzmaEnc_Create(ISzAllocPtr alloc) { void *p; - p = alloc->Alloc(alloc, sizeof(CLzmaEnc)); + p = ISzAlloc_Alloc(alloc, sizeof(CLzmaEnc)); if (p) LzmaEnc_Construct((CLzmaEnc *)p); return p; } -void LzmaEnc_FreeLits(CLzmaEnc *p, ISzAlloc *alloc) +static void LzmaEnc_FreeLits(CLzmaEnc *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->litProbs); - alloc->Free(alloc, p->saveState.litProbs); + ISzAlloc_Free(alloc, p->litProbs); + ISzAlloc_Free(alloc, p->saveState.litProbs); p->litProbs = NULL; p->saveState.litProbs = NULL; } -void LzmaEnc_Destruct(CLzmaEnc *p, ISzAlloc *alloc, ISzAlloc *allocBig) +static void LzmaEnc_Destruct(CLzmaEnc *p, ISzAllocPtr alloc, ISzAllocPtr allocBig) { - #ifndef _7ZIP_ST + #ifndef Z7_ST MatchFinderMt_Destruct(&p->matchFinderMt, allocBig); #endif - MatchFinder_Free(&p->matchFinderBase, allocBig); + MatchFinder_Free(&MFB, allocBig); LzmaEnc_FreeLits(p, alloc); RangeEnc_Free(&p->rc, alloc); } -void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAlloc *alloc, ISzAlloc *allocBig) +void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAllocPtr alloc, ISzAllocPtr allocBig) { - LzmaEnc_Destruct((CLzmaEnc *)p, alloc, allocBig); - alloc->Free(alloc, p); + // GET_CLzmaEnc_p + LzmaEnc_Destruct(p, alloc, allocBig); + ISzAlloc_Free(alloc, p); } -static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize, UInt32 maxUnpackSize) + +Z7_NO_INLINE +static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, UInt32 maxPackSize, UInt32 maxUnpackSize) { UInt32 nowPos32, startPos32; if (p->needInit) { + #ifndef Z7_ST + if (p->mtMode) + { + RINOK(MatchFinderMt_InitMt(&p->matchFinderMt)) + } + #endif p->matchFinder.Init(p->matchFinderObj); p->needInit = 0; } if (p->finished) return p->result; - RINOK(CheckErrors(p)); + RINOK(CheckErrors(p)) nowPos32 = (UInt32)p->nowPos64; startPos32 = nowPos32; if (p->nowPos64 == 0) { - UInt32 numPairs; + unsigned numPairs; Byte curByte; if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) == 0) return Flush(p, nowPos32); ReadMatchDistances(p, &numPairs); - RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][0], 0); - p->state = kLiteralNextStates[p->state]; + RangeEnc_EncodeBit_0(&p->rc, &p->isMatch[kState_Start][0]); + // p->state = kLiteralNextStates[p->state]; curByte = *(p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - p->additionalOffset); LitEnc_Encode(&p->rc, p->litProbs, curByte); p->additionalOffset--; @@ -1828,123 +2418,253 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize } if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) != 0) + for (;;) { - UInt32 pos, len, posState; - + UInt32 dist; + unsigned len, posState; + UInt32 range, ttt, newBound; + CLzmaProb *probs; + if (p->fastMode) - len = GetOptimumFast(p, &pos); + len = GetOptimumFast(p); else - len = GetOptimum(p, nowPos32, &pos); + { + unsigned oci = p->optCur; + if (p->optEnd == oci) + len = GetOptimum(p, nowPos32); + else + { + const COptimal *opt = &p->opt[oci]; + len = opt->len; + p->backRes = opt->dist; + p->optCur = oci + 1; + } + } + + posState = (unsigned)nowPos32 & p->pbMask; + range = p->rc.range; + probs = &p->isMatch[p->state][posState]; + + RC_BIT_PRE(&p->rc, probs) + + dist = p->backRes; #ifdef SHOW_STAT2 - printf("\n pos = %4X, len = %u pos = %u", nowPos32, len, pos); + printf("\n pos = %6X, len = %3u pos = %6u", nowPos32, len, dist); #endif - posState = nowPos32 & p->pbMask; - if (len == 1 && pos == (UInt32)-1) + if (dist == MARK_LIT) { Byte curByte; - CLzmaProb *probs; const Byte *data; + unsigned state; - RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 0); + RC_BIT_0(&p->rc, probs) + p->rc.range = range; data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - p->additionalOffset; - curByte = *data; probs = LIT_PROBS(nowPos32, *(data - 1)); - if (IsCharState(p->state)) + curByte = *data; + state = p->state; + p->state = kLiteralNextStates[state]; + if (IsLitState(state)) LitEnc_Encode(&p->rc, probs, curByte); else - LitEnc_EncodeMatched(&p->rc, probs, curByte, *(data - p->reps[0] - 1)); - p->state = kLiteralNextStates[p->state]; + LitEnc_EncodeMatched(&p->rc, probs, curByte, *(data - p->reps[0])); } else { - RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 1); - if (pos < LZMA_NUM_REPS) + RC_BIT_1(&p->rc, probs) + probs = &p->isRep[p->state]; + RC_BIT_PRE(&p->rc, probs) + + if (dist < LZMA_NUM_REPS) { - RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 1); - if (pos == 0) + RC_BIT_1(&p->rc, probs) + probs = &p->isRepG0[p->state]; + RC_BIT_PRE(&p->rc, probs) + if (dist == 0) { - RangeEnc_EncodeBit(&p->rc, &p->isRepG0[p->state], 0); - RangeEnc_EncodeBit(&p->rc, &p->isRep0Long[p->state][posState], ((len == 1) ? 0 : 1)); + RC_BIT_0(&p->rc, probs) + probs = &p->isRep0Long[p->state][posState]; + RC_BIT_PRE(&p->rc, probs) + if (len != 1) + { + RC_BIT_1_BASE(&p->rc, probs) + } + else + { + RC_BIT_0_BASE(&p->rc, probs) + p->state = kShortRepNextStates[p->state]; + } } else { - UInt32 distance = p->reps[pos]; - RangeEnc_EncodeBit(&p->rc, &p->isRepG0[p->state], 1); - if (pos == 1) - RangeEnc_EncodeBit(&p->rc, &p->isRepG1[p->state], 0); + RC_BIT_1(&p->rc, probs) + probs = &p->isRepG1[p->state]; + RC_BIT_PRE(&p->rc, probs) + if (dist == 1) + { + RC_BIT_0_BASE(&p->rc, probs) + dist = p->reps[1]; + } else { - RangeEnc_EncodeBit(&p->rc, &p->isRepG1[p->state], 1); - RangeEnc_EncodeBit(&p->rc, &p->isRepG2[p->state], pos - 2); - if (pos == 3) + RC_BIT_1(&p->rc, probs) + probs = &p->isRepG2[p->state]; + RC_BIT_PRE(&p->rc, probs) + if (dist == 2) + { + RC_BIT_0_BASE(&p->rc, probs) + dist = p->reps[2]; + } + else + { + RC_BIT_1_BASE(&p->rc, probs) + dist = p->reps[3]; p->reps[3] = p->reps[2]; + } p->reps[2] = p->reps[1]; } p->reps[1] = p->reps[0]; - p->reps[0] = distance; + p->reps[0] = dist; } - if (len == 1) - p->state = kShortRepNextStates[p->state]; - else + + RC_NORM(&p->rc) + + p->rc.range = range; + + if (len != 1) { - LenEnc_Encode2(&p->repLenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices); + LenEnc_Encode(&p->repLenProbs, &p->rc, len - LZMA_MATCH_LEN_MIN, posState); + --p->repLenEncCounter; p->state = kRepNextStates[p->state]; } } else { - UInt32 posSlot; - RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 0); + unsigned posSlot; + RC_BIT_0(&p->rc, probs) + p->rc.range = range; p->state = kMatchNextStates[p->state]; - LenEnc_Encode2(&p->lenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices); - pos -= LZMA_NUM_REPS; - GetPosSlot(pos, posSlot); - RcTree_Encode(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], kNumPosSlotBits, posSlot); + + LenEnc_Encode(&p->lenProbs, &p->rc, len - LZMA_MATCH_LEN_MIN, posState); + // --p->lenEnc.counter; + + dist -= LZMA_NUM_REPS; + p->reps[3] = p->reps[2]; + p->reps[2] = p->reps[1]; + p->reps[1] = p->reps[0]; + p->reps[0] = dist + 1; + + p->matchPriceCount++; + GetPosSlot(dist, posSlot) + // RcTree_Encode_PosSlot(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], posSlot); + { + UInt32 sym = (UInt32)posSlot + (1 << kNumPosSlotBits); + range = p->rc.range; + probs = p->posSlotEncoder[GetLenToPosState(len)]; + do + { + CLzmaProb *prob = probs + (sym >> kNumPosSlotBits); + UInt32 bit = (sym >> (kNumPosSlotBits - 1)) & 1; + sym <<= 1; + RC_BIT(&p->rc, prob, bit) + } + while (sym < (1 << kNumPosSlotBits * 2)); + p->rc.range = range; + } - if (posSlot >= kStartPosModelIndex) + if (dist >= kStartPosModelIndex) { - UInt32 footerBits = ((posSlot >> 1) - 1); - UInt32 base = ((2 | (posSlot & 1)) << footerBits); - UInt32 posReduced = pos - base; + unsigned footerBits = ((posSlot >> 1) - 1); - if (posSlot < kEndPosModelIndex) - RcTree_ReverseEncode(&p->rc, p->posEncoders + base - posSlot - 1, footerBits, posReduced); + if (dist < kNumFullDistances) + { + unsigned base = ((2 | (posSlot & 1)) << footerBits); + RcTree_ReverseEncode(&p->rc, p->posEncoders + base, footerBits, (unsigned)(dist /* - base */)); + } else { - RangeEnc_EncodeDirectBits(&p->rc, posReduced >> kNumAlignBits, footerBits - kNumAlignBits); - RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, posReduced & kAlignMask); - p->alignPriceCount++; + UInt32 pos2 = (dist | 0xF) << (32 - footerBits); + range = p->rc.range; + // RangeEnc_EncodeDirectBits(&p->rc, posReduced >> kNumAlignBits, footerBits - kNumAlignBits); + /* + do + { + range >>= 1; + p->rc.low += range & (0 - ((dist >> --footerBits) & 1)); + RC_NORM(&p->rc) + } + while (footerBits > kNumAlignBits); + */ + do + { + range >>= 1; + p->rc.low += range & (0 - (pos2 >> 31)); + pos2 += pos2; + RC_NORM(&p->rc) + } + while (pos2 != 0xF0000000); + + + // RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, posReduced & kAlignMask); + + { + unsigned m = 1; + unsigned bit; + bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit) m = (m << 1) + bit; + bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit) m = (m << 1) + bit; + bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit) m = (m << 1) + bit; + bit = dist & 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit) + p->rc.range = range; + // p->alignPriceCount++; + } } } - p->reps[3] = p->reps[2]; - p->reps[2] = p->reps[1]; - p->reps[1] = p->reps[0]; - p->reps[0] = pos; - p->matchPriceCount++; } } + + nowPos32 += (UInt32)len; p->additionalOffset -= len; - nowPos32 += len; + if (p->additionalOffset == 0) { UInt32 processed; + if (!p->fastMode) { - if (p->matchPriceCount >= (1 << 7)) + /* + if (p->alignPriceCount >= 16) // kAlignTableSize + FillAlignPrices(p); + if (p->matchPriceCount >= 128) FillDistancesPrices(p); - if (p->alignPriceCount >= kAlignTableSize) + if (p->lenEnc.counter <= 0) + LenPriceEnc_UpdateTables(&p->lenEnc, 1 << p->pb, &p->lenProbs, p->ProbPrices); + */ + if (p->matchPriceCount >= 64) + { FillAlignPrices(p); + // { int y; for (y = 0; y < 100; y++) { + FillDistancesPrices(p); + // }} + LenPriceEnc_UpdateTables(&p->lenEnc, (unsigned)1 << p->pb, &p->lenProbs, p->ProbPrices); + } + if (p->repLenEncCounter <= 0) + { + p->repLenEncCounter = REP_LEN_COUNT; + LenPriceEnc_UpdateTables(&p->repLenEnc, (unsigned)1 << p->pb, &p->repLenProbs, p->ProbPrices); + } } + if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) == 0) break; processed = nowPos32 - startPos32; - if (useLimits) + + if (maxPackSize) { - if (processed + kNumOpts + 300 >= maxUnpackSize || - RangeEnc_GetProcessed(&p->rc) + kNumOpts * 2 >= maxPackSize) + if (processed + kNumOpts + 300 >= maxUnpackSize + || RangeEnc_GetProcessed_sizet(&p->rc) + kPackReserve >= maxPackSize) break; } else if (processed >= (1 << 17)) @@ -1954,20 +2674,25 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize } } } + p->nowPos64 += nowPos32 - startPos32; return Flush(p, nowPos32); } + + #define kBigHashDicLimit ((UInt32)1 << 24) -static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig) +static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig) { UInt32 beforeSize = kNumOpts; + UInt32 dictSize; + if (!RangeEnc_Alloc(&p->rc, alloc)) return SZ_ERROR_MEM; - #ifndef _7ZIP_ST - p->mtMode = (p->multiThread && !p->fastMode && (p->matchFinderBase.btMode != 0)); + #ifndef Z7_ST + p->mtMode = (p->multiThread && !p->fastMode && (MFB.btMode != 0)); #endif { @@ -1975,8 +2700,8 @@ static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, I if (!p->litProbs || !p->saveState.litProbs || p->lclp != lclp) { LzmaEnc_FreeLits(p, alloc); - p->litProbs = (CLzmaProb *)alloc->Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb)); - p->saveState.litProbs = (CLzmaProb *)alloc->Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb)); + p->litProbs = (CLzmaProb *)ISzAlloc_Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb)); + p->saveState.litProbs = (CLzmaProb *)ISzAlloc_Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb)); if (!p->litProbs || !p->saveState.litProbs) { LzmaEnc_FreeLits(p, alloc); @@ -1986,43 +2711,71 @@ static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, I } } - p->matchFinderBase.bigHash = (Byte)(p->dictSize > kBigHashDicLimit ? 1 : 0); + MFB.bigHash = (Byte)(p->dictSize > kBigHashDicLimit ? 1 : 0); + + + dictSize = p->dictSize; + if (dictSize == ((UInt32)2 << 30) || + dictSize == ((UInt32)3 << 30)) + { + /* 21.03 : here we reduce the dictionary for 2 reasons: + 1) we don't want 32-bit back_distance matches in decoder for 2 GB dictionary. + 2) we want to elimate useless last MatchFinder_Normalize3() for corner cases, + where data size is aligned for 1 GB: 5/6/8 GB. + That reducing must be >= 1 for such corner cases. */ + dictSize -= 1; + } - if (beforeSize + p->dictSize < keepWindowSize) - beforeSize = keepWindowSize - p->dictSize; + if (beforeSize + dictSize < keepWindowSize) + beforeSize = keepWindowSize - dictSize; - #ifndef _7ZIP_ST + /* in worst case we can look ahead for + max(LZMA_MATCH_LEN_MAX, numFastBytes + 1 + numFastBytes) bytes. + we send larger value for (keepAfter) to MantchFinder_Create(): + (numFastBytes + LZMA_MATCH_LEN_MAX + 1) + */ + + #ifndef Z7_ST if (p->mtMode) { - RINOK(MatchFinderMt_Create(&p->matchFinderMt, p->dictSize, beforeSize, p->numFastBytes, LZMA_MATCH_LEN_MAX, allocBig)); + RINOK(MatchFinderMt_Create(&p->matchFinderMt, dictSize, beforeSize, + p->numFastBytes, LZMA_MATCH_LEN_MAX + 1 /* 18.04 */ + , allocBig)) p->matchFinderObj = &p->matchFinderMt; + MFB.bigHash = (Byte)(MFB.hashMask >= 0xFFFFFF ? 1 : 0); MatchFinderMt_CreateVTable(&p->matchFinderMt, &p->matchFinder); } else #endif { - if (!MatchFinder_Create(&p->matchFinderBase, p->dictSize, beforeSize, p->numFastBytes, LZMA_MATCH_LEN_MAX, allocBig)) + if (!MatchFinder_Create(&MFB, dictSize, beforeSize, + p->numFastBytes, LZMA_MATCH_LEN_MAX + 1 /* 21.03 */ + , allocBig)) return SZ_ERROR_MEM; - p->matchFinderObj = &p->matchFinderBase; - MatchFinder_CreateVTable(&p->matchFinderBase, &p->matchFinder); + p->matchFinderObj = &MFB; + MatchFinder_CreateVTable(&MFB, &p->matchFinder); } return SZ_OK; } -void LzmaEnc_Init(CLzmaEnc *p) +static void LzmaEnc_Init(CLzmaEnc *p) { - UInt32 i; + unsigned i; p->state = 0; - for (i = 0 ; i < LZMA_NUM_REPS; i++) - p->reps[i] = 0; + p->reps[0] = + p->reps[1] = + p->reps[2] = + p->reps[3] = 1; RangeEnc_Init(&p->rc); + for (i = 0; i < (1 << kNumAlignBits); i++) + p->posAlignEncoder[i] = kProbInitValue; for (i = 0; i < kNumStates; i++) { - UInt32 j; + unsigned j; for (j = 0; j < LZMA_NUM_PB_STATES_MAX; j++) { p->isMatch[i][j] = kProbInitValue; @@ -2035,41 +2788,49 @@ void LzmaEnc_Init(CLzmaEnc *p) } { - UInt32 num = (UInt32)0x300 << (p->lp + p->lc); - CLzmaProb *probs = p->litProbs; - for (i = 0; i < num; i++) - probs[i] = kProbInitValue; - } - - { for (i = 0; i < kNumLenToPosStates; i++) { CLzmaProb *probs = p->posSlotEncoder[i]; - UInt32 j; + unsigned j; for (j = 0; j < (1 << kNumPosSlotBits); j++) probs[j] = kProbInitValue; } } { - for (i = 0; i < kNumFullDistances - kEndPosModelIndex; i++) + for (i = 0; i < kNumFullDistances; i++) p->posEncoders[i] = kProbInitValue; } - LenEnc_Init(&p->lenEnc.p); - LenEnc_Init(&p->repLenEnc.p); + { + UInt32 num = (UInt32)0x300 << (p->lp + p->lc); + UInt32 k; + CLzmaProb *probs = p->litProbs; + for (k = 0; k < num; k++) + probs[k] = kProbInitValue; + } + - for (i = 0; i < (1 << kNumAlignBits); i++) - p->posAlignEncoder[i] = kProbInitValue; + LenEnc_Init(&p->lenProbs); + LenEnc_Init(&p->repLenProbs); + + p->optEnd = 0; + p->optCur = 0; + + { + for (i = 0; i < kNumOpts; i++) + p->opt[i].price = kInfinityPrice; + } - p->optimumEndIndex = 0; - p->optimumCurrentIndex = 0; p->additionalOffset = 0; - p->pbMask = (1 << p->pb) - 1; - p->lpMask = (1 << p->lp) - 1; + p->pbMask = ((unsigned)1 << p->pb) - 1; + p->lpMask = ((UInt32)0x100 << p->lp) - ((unsigned)0x100 >> p->lc); + + // p->mf_Failure = False; } -void LzmaEnc_InitPrices(CLzmaEnc *p) + +static void LzmaEnc_InitPrices(CLzmaEnc *p) { if (!p->fastMode) { @@ -2080,122 +2841,125 @@ void LzmaEnc_InitPrices(CLzmaEnc *p) p->lenEnc.tableSize = p->repLenEnc.tableSize = p->numFastBytes + 1 - LZMA_MATCH_LEN_MIN; - LenPriceEnc_UpdateTables(&p->lenEnc, 1 << p->pb, p->ProbPrices); - LenPriceEnc_UpdateTables(&p->repLenEnc, 1 << p->pb, p->ProbPrices); + + p->repLenEncCounter = REP_LEN_COUNT; + + LenPriceEnc_UpdateTables(&p->lenEnc, (unsigned)1 << p->pb, &p->lenProbs, p->ProbPrices); + LenPriceEnc_UpdateTables(&p->repLenEnc, (unsigned)1 << p->pb, &p->repLenProbs, p->ProbPrices); } -static SRes LzmaEnc_AllocAndInit(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig) +static SRes LzmaEnc_AllocAndInit(CLzmaEnc *p, UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig) { - UInt32 i; - for (i = 0; i < (UInt32)kDicLogSizeMaxCompress; i++) + unsigned i; + for (i = kEndPosModelIndex / 2; i < kDicLogSizeMax; i++) if (p->dictSize <= ((UInt32)1 << i)) break; p->distTableSize = i * 2; p->finished = False; p->result = SZ_OK; - RINOK(LzmaEnc_Alloc(p, keepWindowSize, alloc, allocBig)); + p->nowPos64 = 0; + p->needInit = 1; + RINOK(LzmaEnc_Alloc(p, keepWindowSize, alloc, allocBig)) LzmaEnc_Init(p); LzmaEnc_InitPrices(p); - p->nowPos64 = 0; return SZ_OK; } -static SRes LzmaEnc_Prepare(CLzmaEncHandle pp, ISeqOutStream *outStream, ISeqInStream *inStream, - ISzAlloc *alloc, ISzAlloc *allocBig) +static SRes LzmaEnc_Prepare(CLzmaEncHandle p, + ISeqOutStreamPtr outStream, + ISeqInStreamPtr inStream, + ISzAllocPtr alloc, ISzAllocPtr allocBig) { - CLzmaEnc *p = (CLzmaEnc *)pp; - p->matchFinderBase.stream = inStream; - p->needInit = 1; + // GET_CLzmaEnc_p + MatchFinder_SET_STREAM(&MFB, inStream) p->rc.outStream = outStream; return LzmaEnc_AllocAndInit(p, 0, alloc, allocBig); } -SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle pp, - ISeqInStream *inStream, UInt32 keepWindowSize, - ISzAlloc *alloc, ISzAlloc *allocBig) +SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle p, + ISeqInStreamPtr inStream, UInt32 keepWindowSize, + ISzAllocPtr alloc, ISzAllocPtr allocBig) { - CLzmaEnc *p = (CLzmaEnc *)pp; - p->matchFinderBase.stream = inStream; - p->needInit = 1; + // GET_CLzmaEnc_p + MatchFinder_SET_STREAM(&MFB, inStream) return LzmaEnc_AllocAndInit(p, keepWindowSize, alloc, allocBig); } -static void LzmaEnc_SetInputBuf(CLzmaEnc *p, const Byte *src, SizeT srcLen) -{ - p->matchFinderBase.directInput = 1; - p->matchFinderBase.bufferBase = (Byte *)src; - p->matchFinderBase.directInputRem = srcLen; -} - -SRes LzmaEnc_MemPrepare(CLzmaEncHandle pp, const Byte *src, SizeT srcLen, - UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig) +SRes LzmaEnc_MemPrepare(CLzmaEncHandle p, + const Byte *src, SizeT srcLen, + UInt32 keepWindowSize, + ISzAllocPtr alloc, ISzAllocPtr allocBig) { - CLzmaEnc *p = (CLzmaEnc *)pp; - LzmaEnc_SetInputBuf(p, src, srcLen); - p->needInit = 1; - + // GET_CLzmaEnc_p + MatchFinder_SET_DIRECT_INPUT_BUF(&MFB, src, srcLen) + LzmaEnc_SetDataSize(p, srcLen); return LzmaEnc_AllocAndInit(p, keepWindowSize, alloc, allocBig); } -void LzmaEnc_Finish(CLzmaEncHandle pp) +void LzmaEnc_Finish(CLzmaEncHandle p) { - #ifndef _7ZIP_ST - CLzmaEnc *p = (CLzmaEnc *)pp; + #ifndef Z7_ST + // GET_CLzmaEnc_p if (p->mtMode) MatchFinderMt_ReleaseStream(&p->matchFinderMt); #else - UNUSED_VAR(pp); + UNUSED_VAR(p) #endif } typedef struct { - ISeqOutStream funcTable; + ISeqOutStream vt; Byte *data; - SizeT rem; - Bool overflow; -} CSeqOutStreamBuf; + size_t rem; + BoolInt overflow; +} CLzmaEnc_SeqOutStreamBuf; -static size_t MyWrite(void *pp, const void *data, size_t size) +static size_t SeqOutStreamBuf_Write(ISeqOutStreamPtr pp, const void *data, size_t size) { - CSeqOutStreamBuf *p = (CSeqOutStreamBuf *)pp; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CLzmaEnc_SeqOutStreamBuf) if (p->rem < size) { size = p->rem; p->overflow = True; } - memcpy(p->data, data, size); - p->rem -= size; - p->data += size; + if (size != 0) + { + memcpy(p->data, data, size); + p->rem -= size; + p->data += size; + } return size; } -UInt32 LzmaEnc_GetNumAvailableBytes(CLzmaEncHandle pp) +/* +UInt32 LzmaEnc_GetNumAvailableBytes(CLzmaEncHandle p) { - const CLzmaEnc *p = (CLzmaEnc *)pp; + GET_const_CLzmaEnc_p return p->matchFinder.GetNumAvailableBytes(p->matchFinderObj); } +*/ - -const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle pp) +const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle p) { - const CLzmaEnc *p = (CLzmaEnc *)pp; + // GET_const_CLzmaEnc_p return p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - p->additionalOffset; } -SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit, +// (desiredPackSize == 0) is not allowed +SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle p, BoolInt reInit, Byte *dest, size_t *destLen, UInt32 desiredPackSize, UInt32 *unpackSize) { - CLzmaEnc *p = (CLzmaEnc *)pp; + // GET_CLzmaEnc_p UInt64 nowPos64; SRes res; - CSeqOutStreamBuf outStream; + CLzmaEnc_SeqOutStreamBuf outStream; - outStream.funcTable.Write = MyWrite; + outStream.vt.Write = SeqOutStreamBuf_Write; outStream.data = dest; outStream.rem = *destLen; outStream.overflow = False; @@ -2207,11 +2971,11 @@ SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit, if (reInit) LzmaEnc_Init(p); LzmaEnc_InitPrices(p); - nowPos64 = p->nowPos64; RangeEnc_Init(&p->rc); - p->rc.outStream = &outStream.funcTable; - - res = LzmaEnc_CodeOneBlock(p, True, desiredPackSize, *unpackSize); + p->rc.outStream = &outStream.vt; + nowPos64 = p->nowPos64; + + res = LzmaEnc_CodeOneBlock(p, desiredPackSize, *unpackSize); *unpackSize = (UInt32)(p->nowPos64 - nowPos64); *destLen -= outStream.rem; @@ -2222,11 +2986,12 @@ SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit, } -static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress) +Z7_NO_INLINE +static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgressPtr progress) { SRes res = SZ_OK; - #ifndef _7ZIP_ST + #ifndef Z7_ST Byte allocaDummy[0x300]; allocaDummy[0] = 0; allocaDummy[1] = allocaDummy[0]; @@ -2234,12 +2999,12 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress) for (;;) { - res = LzmaEnc_CodeOneBlock(p, False, 0, 0); + res = LzmaEnc_CodeOneBlock(p, 0, 0); if (res != SZ_OK || p->finished) break; if (progress) { - res = progress->Progress(progress, p->nowPos64, RangeEnc_GetProcessed(&p->rc)); + res = ICompressProgress_Progress(progress, p->nowPos64, RangeEnc_GetProcessed(&p->rc)); if (res != SZ_OK) { res = SZ_ERROR_PROGRESS; @@ -2248,10 +3013,10 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress) } } - LzmaEnc_Finish(p); + LzmaEnc_Finish((CLzmaEncHandle)(void *)p); /* - if (res == S_OK && !Inline_MatchFinder_IsFinishedOK(&p->matchFinderBase)) + if (res == SZ_OK && !Inline_MatchFinder_IsFinishedOK(&MFB)) res = SZ_ERROR_FAIL; } */ @@ -2260,59 +3025,75 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress) } -SRes LzmaEnc_Encode(CLzmaEncHandle pp, ISeqOutStream *outStream, ISeqInStream *inStream, ICompressProgress *progress, - ISzAlloc *alloc, ISzAlloc *allocBig) +SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, ICompressProgressPtr progress, + ISzAllocPtr alloc, ISzAllocPtr allocBig) { - RINOK(LzmaEnc_Prepare(pp, outStream, inStream, alloc, allocBig)); - return LzmaEnc_Encode2((CLzmaEnc *)pp, progress); + // GET_CLzmaEnc_p + RINOK(LzmaEnc_Prepare(p, outStream, inStream, alloc, allocBig)) + return LzmaEnc_Encode2(p, progress); } -SRes LzmaEnc_WriteProperties(CLzmaEncHandle pp, Byte *props, SizeT *size) +SRes LzmaEnc_WriteProperties(CLzmaEncHandle p, Byte *props, SizeT *size) { - CLzmaEnc *p = (CLzmaEnc *)pp; - unsigned i; - UInt32 dictSize = p->dictSize; if (*size < LZMA_PROPS_SIZE) return SZ_ERROR_PARAM; *size = LZMA_PROPS_SIZE; - props[0] = (Byte)((p->pb * 5 + p->lp) * 9 + p->lc); - - if (dictSize >= ((UInt32)1 << 22)) { - UInt32 kDictMask = ((UInt32)1 << 20) - 1; - if (dictSize < (UInt32)0xFFFFFFFF - kDictMask) - dictSize = (dictSize + kDictMask) & ~kDictMask; - } - else for (i = 11; i <= 30; i++) - { - if (dictSize <= ((UInt32)2 << i)) { dictSize = (2 << i); break; } - if (dictSize <= ((UInt32)3 << i)) { dictSize = (3 << i); break; } + // GET_CLzmaEnc_p + const UInt32 dictSize = p->dictSize; + UInt32 v; + props[0] = (Byte)((p->pb * 5 + p->lp) * 9 + p->lc); + + // we write aligned dictionary value to properties for lzma decoder + if (dictSize >= ((UInt32)1 << 21)) + { + const UInt32 kDictMask = ((UInt32)1 << 20) - 1; + v = (dictSize + kDictMask) & ~kDictMask; + if (v < dictSize) + v = dictSize; + } + else + { + unsigned i = 11 * 2; + do + { + v = (UInt32)(2 + (i & 1)) << (i >> 1); + i++; + } + while (v < dictSize); + } + + SetUi32(props + 1, v) + return SZ_OK; } +} - for (i = 0; i < 4; i++) - props[1 + i] = (Byte)(dictSize >> (8 * i)); - return SZ_OK; + +unsigned LzmaEnc_IsWriteEndMark(CLzmaEncHandle p) +{ + // GET_CLzmaEnc_p + return (unsigned)p->writeEndMark; } -SRes LzmaEnc_MemEncode(CLzmaEncHandle pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, - int writeEndMark, ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig) +SRes LzmaEnc_MemEncode(CLzmaEncHandle p, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, + int writeEndMark, ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig) { SRes res; - CLzmaEnc *p = (CLzmaEnc *)pp; + // GET_CLzmaEnc_p - CSeqOutStreamBuf outStream; + CLzmaEnc_SeqOutStreamBuf outStream; - outStream.funcTable.Write = MyWrite; + outStream.vt.Write = SeqOutStreamBuf_Write; outStream.data = dest; outStream.rem = *destLen; outStream.overflow = False; p->writeEndMark = writeEndMark; - p->rc.outStream = &outStream.funcTable; + p->rc.outStream = &outStream.vt; - res = LzmaEnc_MemPrepare(pp, src, srcLen, 0, alloc, allocBig); + res = LzmaEnc_MemPrepare(p, src, srcLen, 0, alloc, allocBig); if (res == SZ_OK) { @@ -2321,7 +3102,7 @@ SRes LzmaEnc_MemEncode(CLzmaEncHandle pp, Byte *dest, SizeT *destLen, const Byte res = SZ_ERROR_FAIL; } - *destLen -= outStream.rem; + *destLen -= (SizeT)outStream.rem; if (outStream.overflow) return SZ_ERROR_OUTPUT_EOF; return res; @@ -2330,9 +3111,9 @@ SRes LzmaEnc_MemEncode(CLzmaEncHandle pp, Byte *dest, SizeT *destLen, const Byte SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, const CLzmaEncProps *props, Byte *propsEncoded, SizeT *propsSize, int writeEndMark, - ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig) + ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig) { - CLzmaEnc *p = (CLzmaEnc *)LzmaEnc_Create(alloc); + CLzmaEncHandle p = LzmaEnc_Create(alloc); SRes res; if (!p) return SZ_ERROR_MEM; @@ -2349,3 +3130,15 @@ SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, LzmaEnc_Destroy(p, alloc, allocBig); return res; } + + +/* +#ifndef Z7_ST +void LzmaEnc_GetLzThreads(CLzmaEncHandle p, HANDLE lz_threads[2]) +{ + GET_const_CLzmaEnc_p + lz_threads[0] = p->matchFinderMt.hashSync.thread; + lz_threads[1] = p->matchFinderMt.btSync.thread; +} +#endif +*/ diff --git a/3rdparty/lzma/C/LzmaEnc.h b/3rdparty/lzma/C/LzmaEnc.h index cffe220bbf6..9f8039a1032 100644 --- a/3rdparty/lzma/C/LzmaEnc.h +++ b/3rdparty/lzma/C/LzmaEnc.h @@ -1,8 +1,8 @@ /* LzmaEnc.h -- LZMA Encoder -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ -#ifndef __LZMA_ENC_H -#define __LZMA_ENC_H +#ifndef ZIP7_INC_LZMA_ENC_H +#define ZIP7_INC_LZMA_ENC_H #include "7zTypes.h" @@ -10,14 +10,12 @@ EXTERN_C_BEGIN #define LZMA_PROPS_SIZE 5 -typedef struct _CLzmaEncProps +typedef struct { - int level; /* 0 <= level <= 9 */ + int level; /* 0 <= level <= 9 */ UInt32 dictSize; /* (1 << 12) <= dictSize <= (1 << 27) for 32-bit version - (1 << 12) <= dictSize <= (1 << 30) for 64-bit version - default = (1 << 24) */ - UInt64 reduceSize; /* estimated size of data that will be compressed. default = 0xFFFFFFFF. - Encoder uses this value to reduce dictionary size */ + (1 << 12) <= dictSize <= (3 << 29) for 64-bit version + default = (1 << 24) */ int lc; /* 0 <= lc <= 8, default = 3 */ int lp; /* 0 <= lp <= 4, default = 0 */ int pb; /* 0 <= pb <= 4, default = 2 */ @@ -25,9 +23,17 @@ typedef struct _CLzmaEncProps int fb; /* 5 <= fb <= 273, default = 32 */ int btMode; /* 0 - hashChain Mode, 1 - binTree mode - normal, default = 1 */ int numHashBytes; /* 2, 3 or 4, default = 4 */ - UInt32 mc; /* 1 <= mc <= (1 << 30), default = 32 */ + unsigned numHashOutBits; /* default = ? */ + UInt32 mc; /* 1 <= mc <= (1 << 30), default = 32 */ unsigned writeEndMark; /* 0 - do not write EOPM, 1 - write EOPM, default = 0 */ int numThreads; /* 1 or 2, default = 2 */ + + // int _pad; + + UInt64 reduceSize; /* estimated size of data that will be compressed. default = (UInt64)(Int64)-1. + Encoder uses this value to reduce dictionary size */ + + UInt64 affinity; } CLzmaEncProps; void LzmaEncProps_Init(CLzmaEncProps *p); @@ -37,41 +43,40 @@ UInt32 LzmaEncProps_GetDictSize(const CLzmaEncProps *props2); /* ---------- CLzmaEncHandle Interface ---------- */ -/* LzmaEnc_* functions can return the following exit codes: -Returns: +/* LzmaEnc* functions can return the following exit codes: +SRes: SZ_OK - OK SZ_ERROR_MEM - Memory allocation error SZ_ERROR_PARAM - Incorrect paramater in props - SZ_ERROR_WRITE - Write callback error. + SZ_ERROR_WRITE - ISeqOutStream write callback error + SZ_ERROR_OUTPUT_EOF - output buffer overflow - version with (Byte *) output SZ_ERROR_PROGRESS - some break from progress callback - SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version) + SZ_ERROR_THREAD - error in multithreading functions (only for Mt version) */ -typedef void * CLzmaEncHandle; +typedef struct CLzmaEnc CLzmaEnc; +typedef CLzmaEnc * CLzmaEncHandle; +// Z7_DECLARE_HANDLE(CLzmaEncHandle) + +CLzmaEncHandle LzmaEnc_Create(ISzAllocPtr alloc); +void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAllocPtr alloc, ISzAllocPtr allocBig); -CLzmaEncHandle LzmaEnc_Create(ISzAlloc *alloc); -void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAlloc *alloc, ISzAlloc *allocBig); SRes LzmaEnc_SetProps(CLzmaEncHandle p, const CLzmaEncProps *props); +void LzmaEnc_SetDataSize(CLzmaEncHandle p, UInt64 expectedDataSiize); SRes LzmaEnc_WriteProperties(CLzmaEncHandle p, Byte *properties, SizeT *size); -SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStream *outStream, ISeqInStream *inStream, - ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig); +unsigned LzmaEnc_IsWriteEndMark(CLzmaEncHandle p); + +SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, + ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig); SRes LzmaEnc_MemEncode(CLzmaEncHandle p, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, - int writeEndMark, ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig); + int writeEndMark, ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig); -/* ---------- One Call Interface ---------- */ -/* LzmaEncode -Return code: - SZ_OK - OK - SZ_ERROR_MEM - Memory allocation error - SZ_ERROR_PARAM - Incorrect paramater - SZ_ERROR_OUTPUT_EOF - output buffer overflow - SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version) -*/ +/* ---------- One Call Interface ---------- */ SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen, const CLzmaEncProps *props, Byte *propsEncoded, SizeT *propsSize, int writeEndMark, - ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig); + ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig); EXTERN_C_END diff --git a/3rdparty/lzma/C/LzmaLib.c b/3rdparty/lzma/C/LzmaLib.c index 706e9e58cd6..785e8848744 100644 --- a/3rdparty/lzma/C/LzmaLib.c +++ b/3rdparty/lzma/C/LzmaLib.c @@ -1,12 +1,14 @@ /* LzmaLib.c -- LZMA library wrapper -2015-06-13 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ + +#include "Precomp.h" #include "Alloc.h" #include "LzmaDec.h" #include "LzmaEnc.h" #include "LzmaLib.h" -MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen, +Z7_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen, unsigned char *outProps, size_t *outPropsSize, int level, /* 0 <= level <= 9, default = 5 */ unsigned dictSize, /* use (1 << N) or (3 << N). 4 KB < dictSize <= 128 MB */ @@ -32,7 +34,7 @@ MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char } -MY_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t *srcLen, +Z7_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t *srcLen, const unsigned char *props, size_t propsSize) { ELzmaStatus status; diff --git a/3rdparty/lzma/C/LzmaLib.h b/3rdparty/lzma/C/LzmaLib.h index 88fa87d350c..d7c0724de5e 100644 --- a/3rdparty/lzma/C/LzmaLib.h +++ b/3rdparty/lzma/C/LzmaLib.h @@ -1,14 +1,14 @@ /* LzmaLib.h -- LZMA library interface -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __LZMA_LIB_H -#define __LZMA_LIB_H +#ifndef ZIP7_INC_LZMA_LIB_H +#define ZIP7_INC_LZMA_LIB_H #include "7zTypes.h" EXTERN_C_BEGIN -#define MY_STDAPI int MY_STD_CALL +#define Z7_STDAPI int Z7_STDCALL #define LZMA_PROPS_SIZE 5 @@ -40,14 +40,16 @@ outPropsSize - level - compression level: 0 <= level <= 9; level dictSize algo fb - 0: 16 KB 0 32 - 1: 64 KB 0 32 - 2: 256 KB 0 32 - 3: 1 MB 0 32 - 4: 4 MB 0 32 + 0: 64 KB 0 32 + 1: 256 KB 0 32 + 2: 1 MB 0 32 + 3: 4 MB 0 32 + 4: 16 MB 0 32 5: 16 MB 1 32 6: 32 MB 1 32 - 7+: 64 MB 1 64 + 7: 32 MB 1 64 + 8: 64 MB 1 64 + 9: 64 MB 1 64 The default value for "level" is 5. @@ -83,6 +85,11 @@ fb - Word size (the number of fast bytes). numThreads - The number of thereads. 1 or 2. The default value is 2. Fast mode (algo = 0) can use only 1 thread. +In: + dest - output data buffer + destLen - output data buffer size + src - input data + srcLen - input data size Out: destLen - processed output size Returns: @@ -93,7 +100,7 @@ Returns: SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version) */ -MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen, +Z7_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen, unsigned char *outProps, size_t *outPropsSize, /* *outPropsSize must be = 5 */ int level, /* 0 <= level <= 9, default = 5 */ unsigned dictSize, /* default = (1 << 24) */ @@ -108,8 +115,8 @@ MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char LzmaUncompress -------------- In: - dest - output data - destLen - output data size + dest - output data buffer + destLen - output data buffer size src - input data srcLen - input data size Out: @@ -123,7 +130,7 @@ Returns: SZ_ERROR_INPUT_EOF - it needs more bytes in input buffer (src) */ -MY_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, SizeT *srcLen, +Z7_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, SizeT *srcLen, const unsigned char *props, size_t propsSize); EXTERN_C_END diff --git a/3rdparty/lzma/C/MtCoder.c b/3rdparty/lzma/C/MtCoder.c index 8d9731d0a28..6f58abb2a39 100644 --- a/3rdparty/lzma/C/MtCoder.c +++ b/3rdparty/lzma/C/MtCoder.c @@ -1,327 +1,571 @@ /* MtCoder.c -- Multi-thread Coder -2015-10-13 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ #include "Precomp.h" #include "MtCoder.h" -void LoopThread_Construct(CLoopThread *p) -{ - Thread_Construct(&p->thread); - Event_Construct(&p->startEvent); - Event_Construct(&p->finishedEvent); -} - -void LoopThread_Close(CLoopThread *p) -{ - Thread_Close(&p->thread); - Event_Close(&p->startEvent); - Event_Close(&p->finishedEvent); -} +#ifndef Z7_ST -static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE LoopThreadFunc(void *pp) +static SRes MtProgressThunk_Progress(ICompressProgressPtr pp, UInt64 inSize, UInt64 outSize) { - CLoopThread *p = (CLoopThread *)pp; - for (;;) + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CMtProgressThunk) + UInt64 inSize2 = 0; + UInt64 outSize2 = 0; + if (inSize != (UInt64)(Int64)-1) { - if (Event_Wait(&p->startEvent) != 0) - return SZ_ERROR_THREAD; - if (p->stop) - return 0; - p->res = p->func(p->param); - if (Event_Set(&p->finishedEvent) != 0) - return SZ_ERROR_THREAD; + inSize2 = inSize - p->inSize; + p->inSize = inSize; } + if (outSize != (UInt64)(Int64)-1) + { + outSize2 = outSize - p->outSize; + p->outSize = outSize; + } + return MtProgress_ProgressAdd(p->mtProgress, inSize2, outSize2); } -WRes LoopThread_Create(CLoopThread *p) -{ - p->stop = 0; - RINOK(AutoResetEvent_CreateNotSignaled(&p->startEvent)); - RINOK(AutoResetEvent_CreateNotSignaled(&p->finishedEvent)); - return Thread_Create(&p->thread, LoopThreadFunc, p); -} -WRes LoopThread_StopAndWait(CLoopThread *p) +void MtProgressThunk_CreateVTable(CMtProgressThunk *p) { - p->stop = 1; - if (Event_Set(&p->startEvent) != 0) - return SZ_ERROR_THREAD; - return Thread_Wait(&p->thread); + p->vt.Progress = MtProgressThunk_Progress; } -WRes LoopThread_StartSubThread(CLoopThread *p) { return Event_Set(&p->startEvent); } -WRes LoopThread_WaitSubThread(CLoopThread *p) { return Event_Wait(&p->finishedEvent); } -static SRes Progress(ICompressProgress *p, UInt64 inSize, UInt64 outSize) -{ - return (p && p->Progress(p, inSize, outSize) != SZ_OK) ? SZ_ERROR_PROGRESS : SZ_OK; -} -static void MtProgress_Init(CMtProgress *p, ICompressProgress *progress) -{ - unsigned i; - for (i = 0; i < NUM_MT_CODER_THREADS_MAX; i++) - p->inSizes[i] = p->outSizes[i] = 0; - p->totalInSize = p->totalOutSize = 0; - p->progress = progress; - p->res = SZ_OK; -} +#define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; } -static void MtProgress_Reinit(CMtProgress *p, unsigned index) -{ - p->inSizes[index] = 0; - p->outSizes[index] = 0; -} -#define UPDATE_PROGRESS(size, prev, total) \ - if (size != (UInt64)(Int64)-1) { total += size - prev; prev = size; } +static THREAD_FUNC_DECL ThreadFunc(void *pp); -SRes MtProgress_Set(CMtProgress *p, unsigned index, UInt64 inSize, UInt64 outSize) -{ - SRes res; - CriticalSection_Enter(&p->cs); - UPDATE_PROGRESS(inSize, p->inSizes[index], p->totalInSize) - UPDATE_PROGRESS(outSize, p->outSizes[index], p->totalOutSize) - if (p->res == SZ_OK) - p->res = Progress(p->progress, p->totalInSize, p->totalOutSize); - res = p->res; - CriticalSection_Leave(&p->cs); - return res; -} -static void MtProgress_SetError(CMtProgress *p, SRes res) +static SRes MtCoderThread_CreateAndStart(CMtCoderThread *t) { - CriticalSection_Enter(&p->cs); - if (p->res == SZ_OK) - p->res = res; - CriticalSection_Leave(&p->cs); + WRes wres = AutoResetEvent_OptCreate_And_Reset(&t->startEvent); + if (wres == 0) + { + t->stop = False; + if (!Thread_WasCreated(&t->thread)) + wres = Thread_Create(&t->thread, ThreadFunc, t); + if (wres == 0) + wres = Event_Set(&t->startEvent); + } + if (wres == 0) + return SZ_OK; + return MY_SRes_HRESULT_FROM_WRes(wres); } -static void MtCoder_SetError(CMtCoder* p, SRes res) + +static void MtCoderThread_Destruct(CMtCoderThread *t) { - CriticalSection_Enter(&p->cs); - if (p->res == SZ_OK) - p->res = res; - CriticalSection_Leave(&p->cs); -} + if (Thread_WasCreated(&t->thread)) + { + t->stop = 1; + Event_Set(&t->startEvent); + Thread_Wait_Close(&t->thread); + } -/* ---------- MtThread ---------- */ + Event_Close(&t->startEvent); -void CMtThread_Construct(CMtThread *p, CMtCoder *mtCoder) -{ - p->mtCoder = mtCoder; - p->outBuf = 0; - p->inBuf = 0; - Event_Construct(&p->canRead); - Event_Construct(&p->canWrite); - LoopThread_Construct(&p->thread); + if (t->inBuf) + { + ISzAlloc_Free(t->mtCoder->allocBig, t->inBuf); + t->inBuf = NULL; + } } -#define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; } -static void CMtThread_CloseEvents(CMtThread *p) -{ - Event_Close(&p->canRead); - Event_Close(&p->canWrite); -} -static void CMtThread_Destruct(CMtThread *p) + +/* + ThreadFunc2() returns: + SZ_OK - in all normal cases (even for stream error or memory allocation error) + SZ_ERROR_THREAD - in case of failure in system synch function +*/ + +static SRes ThreadFunc2(CMtCoderThread *t) { - CMtThread_CloseEvents(p); + CMtCoder *mtc = t->mtCoder; - if (Thread_WasCreated(&p->thread.thread)) + for (;;) { - LoopThread_StopAndWait(&p->thread); - LoopThread_Close(&p->thread); - } + unsigned bi; + SRes res; + SRes res2; + BoolInt finished; + unsigned bufIndex; + size_t size; + const Byte *inData; + UInt64 readProcessed = 0; + + RINOK_THREAD(Event_Wait(&mtc->readEvent)) + + /* after Event_Wait(&mtc->readEvent) we must call Event_Set(&mtc->readEvent) in any case to unlock another threads */ + + if (mtc->stopReading) + { + return Event_Set(&mtc->readEvent) == 0 ? SZ_OK : SZ_ERROR_THREAD; + } - if (p->mtCoder->alloc) - IAlloc_Free(p->mtCoder->alloc, p->outBuf); - p->outBuf = 0; + res = MtProgress_GetError(&mtc->mtProgress); + + size = 0; + inData = NULL; + finished = True; - if (p->mtCoder->alloc) - IAlloc_Free(p->mtCoder->alloc, p->inBuf); - p->inBuf = 0; -} + if (res == SZ_OK) + { + size = mtc->blockSize; + if (mtc->inStream) + { + if (!t->inBuf) + { + t->inBuf = (Byte *)ISzAlloc_Alloc(mtc->allocBig, mtc->blockSize); + if (!t->inBuf) + res = SZ_ERROR_MEM; + } + if (res == SZ_OK) + { + res = SeqInStream_ReadMax(mtc->inStream, t->inBuf, &size); + readProcessed = mtc->readProcessed + size; + mtc->readProcessed = readProcessed; + } + if (res != SZ_OK) + { + mtc->readRes = res; + /* after reading error - we can stop encoding of previous blocks */ + MtProgress_SetError(&mtc->mtProgress, res); + } + else + finished = (size != mtc->blockSize); + } + else + { + size_t rem; + readProcessed = mtc->readProcessed; + rem = mtc->inDataSize - (size_t)readProcessed; + if (size > rem) + size = rem; + inData = mtc->inData + (size_t)readProcessed; + readProcessed += size; + mtc->readProcessed = readProcessed; + finished = (mtc->inDataSize == (size_t)readProcessed); + } + } -#define MY_BUF_ALLOC(buf, size, newSize) \ - if (buf == 0 || size != newSize) \ - { IAlloc_Free(p->mtCoder->alloc, buf); \ - size = newSize; buf = (Byte *)IAlloc_Alloc(p->mtCoder->alloc, size); \ - if (buf == 0) return SZ_ERROR_MEM; } + /* we must get some block from blocksSemaphore before Event_Set(&mtc->readEvent) */ -static SRes CMtThread_Prepare(CMtThread *p) -{ - MY_BUF_ALLOC(p->inBuf, p->inBufSize, p->mtCoder->blockSize) - MY_BUF_ALLOC(p->outBuf, p->outBufSize, p->mtCoder->destBlockSize) + res2 = SZ_OK; - p->stopReading = False; - p->stopWriting = False; - RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canRead)); - RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canWrite)); + if (Semaphore_Wait(&mtc->blocksSemaphore) != 0) + { + res2 = SZ_ERROR_THREAD; + if (res == SZ_OK) + { + res = res2; + // MtProgress_SetError(&mtc->mtProgress, res); + } + } - return SZ_OK; -} + bi = mtc->blockIndex; -static SRes FullRead(ISeqInStream *stream, Byte *data, size_t *processedSize) -{ - size_t size = *processedSize; - *processedSize = 0; - while (size != 0) - { - size_t curSize = size; - SRes res = stream->Read(stream, data, &curSize); - *processedSize += curSize; - data += curSize; - size -= curSize; - RINOK(res); - if (curSize == 0) - return SZ_OK; - } - return SZ_OK; -} + if (++mtc->blockIndex >= mtc->numBlocksMax) + mtc->blockIndex = 0; -#define GET_NEXT_THREAD(p) &p->mtCoder->threads[p->index == p->mtCoder->numThreads - 1 ? 0 : p->index + 1] + bufIndex = (unsigned)(int)-1; -static SRes MtThread_Process(CMtThread *p, Bool *stop) -{ - CMtThread *next; - *stop = True; - if (Event_Wait(&p->canRead) != 0) - return SZ_ERROR_THREAD; - - next = GET_NEXT_THREAD(p); - - if (p->stopReading) - { - next->stopReading = True; - return Event_Set(&next->canRead) == 0 ? SZ_OK : SZ_ERROR_THREAD; - } + if (res == SZ_OK) + res = MtProgress_GetError(&mtc->mtProgress); - { - size_t size = p->mtCoder->blockSize; - size_t destSize = p->outBufSize; - - RINOK(FullRead(p->mtCoder->inStream, p->inBuf, &size)); - next->stopReading = *stop = (size != p->mtCoder->blockSize); - if (Event_Set(&next->canRead) != 0) - return SZ_ERROR_THREAD; - - RINOK(p->mtCoder->mtCallback->Code(p->mtCoder->mtCallback, p->index, - p->outBuf, &destSize, p->inBuf, size, *stop)); - - MtProgress_Reinit(&p->mtCoder->mtProgress, p->index); - - if (Event_Wait(&p->canWrite) != 0) - return SZ_ERROR_THREAD; - if (p->stopWriting) - return SZ_ERROR_FAIL; - if (p->mtCoder->outStream->Write(p->mtCoder->outStream, p->outBuf, destSize) != destSize) - return SZ_ERROR_WRITE; - return Event_Set(&next->canWrite) == 0 ? SZ_OK : SZ_ERROR_THREAD; + if (res != SZ_OK) + finished = True; + + if (!finished) + { + if (mtc->numStartedThreads < mtc->numStartedThreadsLimit + && mtc->expectedDataSize != readProcessed) + { + res = MtCoderThread_CreateAndStart(&mtc->threads[mtc->numStartedThreads]); + if (res == SZ_OK) + mtc->numStartedThreads++; + else + { + MtProgress_SetError(&mtc->mtProgress, res); + finished = True; + } + } + } + + if (finished) + mtc->stopReading = True; + + RINOK_THREAD(Event_Set(&mtc->readEvent)) + + if (res2 != SZ_OK) + return res2; + + if (res == SZ_OK) + { + CriticalSection_Enter(&mtc->cs); + bufIndex = mtc->freeBlockHead; + mtc->freeBlockHead = mtc->freeBlockList[bufIndex]; + CriticalSection_Leave(&mtc->cs); + + res = mtc->mtCallback->Code(mtc->mtCallbackObject, t->index, bufIndex, + mtc->inStream ? t->inBuf : inData, size, finished); + + // MtProgress_Reinit(&mtc->mtProgress, t->index); + + if (res != SZ_OK) + MtProgress_SetError(&mtc->mtProgress, res); + } + + { + CMtCoderBlock *block = &mtc->blocks[bi]; + block->res = res; + block->bufIndex = bufIndex; + block->finished = finished; + } + + #ifdef MTCODER_USE_WRITE_THREAD + RINOK_THREAD(Event_Set(&mtc->writeEvents[bi])) + #else + { + unsigned wi; + { + CriticalSection_Enter(&mtc->cs); + wi = mtc->writeIndex; + if (wi == bi) + mtc->writeIndex = (unsigned)(int)-1; + else + mtc->ReadyBlocks[bi] = True; + CriticalSection_Leave(&mtc->cs); + } + + if (wi != bi) + { + if (res != SZ_OK || finished) + return 0; + continue; + } + + if (mtc->writeRes != SZ_OK) + res = mtc->writeRes; + + for (;;) + { + if (res == SZ_OK && bufIndex != (unsigned)(int)-1) + { + res = mtc->mtCallback->Write(mtc->mtCallbackObject, bufIndex); + if (res != SZ_OK) + { + mtc->writeRes = res; + MtProgress_SetError(&mtc->mtProgress, res); + } + } + + if (++wi >= mtc->numBlocksMax) + wi = 0; + { + BoolInt isReady; + + CriticalSection_Enter(&mtc->cs); + + if (bufIndex != (unsigned)(int)-1) + { + mtc->freeBlockList[bufIndex] = mtc->freeBlockHead; + mtc->freeBlockHead = bufIndex; + } + + isReady = mtc->ReadyBlocks[wi]; + + if (isReady) + mtc->ReadyBlocks[wi] = False; + else + mtc->writeIndex = wi; + + CriticalSection_Leave(&mtc->cs); + + RINOK_THREAD(Semaphore_Release1(&mtc->blocksSemaphore)) + + if (!isReady) + break; + } + + { + CMtCoderBlock *block = &mtc->blocks[wi]; + if (res == SZ_OK && block->res != SZ_OK) + res = block->res; + bufIndex = block->bufIndex; + finished = block->finished; + } + } + } + #endif + + if (finished || res != SZ_OK) + return 0; } } -static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE ThreadFunc(void *pp) + +static THREAD_FUNC_DECL ThreadFunc(void *pp) { - CMtThread *p = (CMtThread *)pp; + CMtCoderThread *t = (CMtCoderThread *)pp; for (;;) { - Bool stop; - CMtThread *next = GET_NEXT_THREAD(p); - SRes res = MtThread_Process(p, &stop); - if (res != SZ_OK) + if (Event_Wait(&t->startEvent) != 0) + return (THREAD_FUNC_RET_TYPE)SZ_ERROR_THREAD; + if (t->stop) + return 0; { - MtCoder_SetError(p->mtCoder, res); - MtProgress_SetError(&p->mtCoder->mtProgress, res); - next->stopReading = True; - next->stopWriting = True; - Event_Set(&next->canRead); - Event_Set(&next->canWrite); - return res; + SRes res = ThreadFunc2(t); + CMtCoder *mtc = t->mtCoder; + if (res != SZ_OK) + { + MtProgress_SetError(&mtc->mtProgress, res); + } + + #ifndef MTCODER_USE_WRITE_THREAD + { + unsigned numFinished = (unsigned)InterlockedIncrement(&mtc->numFinishedThreads); + if (numFinished == mtc->numStartedThreads) + if (Event_Set(&mtc->finishedEvent) != 0) + return (THREAD_FUNC_RET_TYPE)SZ_ERROR_THREAD; + } + #endif } - if (stop) - return 0; } } -void MtCoder_Construct(CMtCoder* p) + + +void MtCoder_Construct(CMtCoder *p) { unsigned i; - p->alloc = 0; - for (i = 0; i < NUM_MT_CODER_THREADS_MAX; i++) + + p->blockSize = 0; + p->numThreadsMax = 0; + p->expectedDataSize = (UInt64)(Int64)-1; + + p->inStream = NULL; + p->inData = NULL; + p->inDataSize = 0; + + p->progress = NULL; + p->allocBig = NULL; + + p->mtCallback = NULL; + p->mtCallbackObject = NULL; + + p->allocatedBufsSize = 0; + + Event_Construct(&p->readEvent); + Semaphore_Construct(&p->blocksSemaphore); + + for (i = 0; i < MTCODER_THREADS_MAX; i++) { - CMtThread *t = &p->threads[i]; + CMtCoderThread *t = &p->threads[i]; + t->mtCoder = p; t->index = i; - CMtThread_Construct(t, p); + t->inBuf = NULL; + t->stop = False; + Event_Construct(&t->startEvent); + Thread_CONSTRUCT(&t->thread) } + + #ifdef MTCODER_USE_WRITE_THREAD + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + Event_Construct(&p->writeEvents[i]); + #else + Event_Construct(&p->finishedEvent); + #endif + CriticalSection_Init(&p->cs); CriticalSection_Init(&p->mtProgress.cs); } -void MtCoder_Destruct(CMtCoder* p) + + + +static void MtCoder_Free(CMtCoder *p) { unsigned i; - for (i = 0; i < NUM_MT_CODER_THREADS_MAX; i++) - CMtThread_Destruct(&p->threads[i]); + + /* + p->stopReading = True; + if (Event_IsCreated(&p->readEvent)) + Event_Set(&p->readEvent); + */ + + for (i = 0; i < MTCODER_THREADS_MAX; i++) + MtCoderThread_Destruct(&p->threads[i]); + + Event_Close(&p->readEvent); + Semaphore_Close(&p->blocksSemaphore); + + #ifdef MTCODER_USE_WRITE_THREAD + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + Event_Close(&p->writeEvents[i]); + #else + Event_Close(&p->finishedEvent); + #endif +} + + +void MtCoder_Destruct(CMtCoder *p) +{ + MtCoder_Free(p); + CriticalSection_Delete(&p->cs); CriticalSection_Delete(&p->mtProgress.cs); } + SRes MtCoder_Code(CMtCoder *p) { - unsigned i, numThreads = p->numThreads; + unsigned numThreads = p->numThreadsMax; + unsigned numBlocksMax; + unsigned i; SRes res = SZ_OK; - p->res = SZ_OK; + + if (numThreads > MTCODER_THREADS_MAX) + numThreads = MTCODER_THREADS_MAX; + numBlocksMax = MTCODER_GET_NUM_BLOCKS_FROM_THREADS(numThreads); + + if (p->blockSize < ((UInt32)1 << 26)) numBlocksMax++; + if (p->blockSize < ((UInt32)1 << 24)) numBlocksMax++; + if (p->blockSize < ((UInt32)1 << 22)) numBlocksMax++; + + if (numBlocksMax > MTCODER_BLOCKS_MAX) + numBlocksMax = MTCODER_BLOCKS_MAX; + + if (p->blockSize != p->allocatedBufsSize) + { + for (i = 0; i < MTCODER_THREADS_MAX; i++) + { + CMtCoderThread *t = &p->threads[i]; + if (t->inBuf) + { + ISzAlloc_Free(p->allocBig, t->inBuf); + t->inBuf = NULL; + } + } + p->allocatedBufsSize = p->blockSize; + } + + p->readRes = SZ_OK; MtProgress_Init(&p->mtProgress, p->progress); - for (i = 0; i < numThreads; i++) + #ifdef MTCODER_USE_WRITE_THREAD + for (i = 0; i < numBlocksMax; i++) + { + RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->writeEvents[i])) + } + #else + RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->finishedEvent)) + #endif + + { + RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->readEvent)) + RINOK_THREAD(Semaphore_OptCreateInit(&p->blocksSemaphore, numBlocksMax, numBlocksMax)) + } + + for (i = 0; i < MTCODER_BLOCKS_MAX - 1; i++) + p->freeBlockList[i] = i + 1; + p->freeBlockList[MTCODER_BLOCKS_MAX - 1] = (unsigned)(int)-1; + p->freeBlockHead = 0; + + p->readProcessed = 0; + p->blockIndex = 0; + p->numBlocksMax = numBlocksMax; + p->stopReading = False; + + #ifndef MTCODER_USE_WRITE_THREAD + p->writeIndex = 0; + p->writeRes = SZ_OK; + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + p->ReadyBlocks[i] = False; + p->numFinishedThreads = 0; + #endif + + p->numStartedThreadsLimit = numThreads; + p->numStartedThreads = 0; + + // for (i = 0; i < numThreads; i++) { - RINOK(CMtThread_Prepare(&p->threads[i])); + CMtCoderThread *nextThread = &p->threads[p->numStartedThreads++]; + RINOK(MtCoderThread_CreateAndStart(nextThread)) } - for (i = 0; i < numThreads; i++) + RINOK_THREAD(Event_Set(&p->readEvent)) + + #ifdef MTCODER_USE_WRITE_THREAD { - CMtThread *t = &p->threads[i]; - CLoopThread *lt = &t->thread; + unsigned bi = 0; - if (!Thread_WasCreated(<->thread)) + for (;; bi++) { - lt->func = ThreadFunc; - lt->param = t; + if (bi >= numBlocksMax) + bi = 0; + + RINOK_THREAD(Event_Wait(&p->writeEvents[bi])) - if (LoopThread_Create(lt) != SZ_OK) { - res = SZ_ERROR_THREAD; - break; + const CMtCoderBlock *block = &p->blocks[bi]; + unsigned bufIndex = block->bufIndex; + BoolInt finished = block->finished; + if (res == SZ_OK && block->res != SZ_OK) + res = block->res; + + if (bufIndex != (unsigned)(int)-1) + { + if (res == SZ_OK) + { + res = p->mtCallback->Write(p->mtCallbackObject, bufIndex); + if (res != SZ_OK) + MtProgress_SetError(&p->mtProgress, res); + } + + CriticalSection_Enter(&p->cs); + { + p->freeBlockList[bufIndex] = p->freeBlockHead; + p->freeBlockHead = bufIndex; + } + CriticalSection_Leave(&p->cs); + } + + RINOK_THREAD(Semaphore_Release1(&p->blocksSemaphore)) + + if (finished) + break; } } } + #else + { + WRes wres = Event_Wait(&p->finishedEvent); + res = MY_SRes_HRESULT_FROM_WRes(wres); + } + #endif if (res == SZ_OK) - { - unsigned j; - for (i = 0; i < numThreads; i++) - { - CMtThread *t = &p->threads[i]; - if (LoopThread_StartSubThread(&t->thread) != SZ_OK) - { - res = SZ_ERROR_THREAD; - p->threads[0].stopReading = True; - break; - } - } + res = p->readRes; - Event_Set(&p->threads[0].canWrite); - Event_Set(&p->threads[0].canRead); + if (res == SZ_OK) + res = p->mtProgress.res; - for (j = 0; j < i; j++) - LoopThread_WaitSubThread(&p->threads[j].thread); - } + #ifndef MTCODER_USE_WRITE_THREAD + if (res == SZ_OK) + res = p->writeRes; + #endif - for (i = 0; i < numThreads; i++) - CMtThread_CloseEvents(&p->threads[i]); - return (res == SZ_OK) ? p->res : res; + if (res != SZ_OK) + MtCoder_Free(p); + return res; } + +#endif + +#undef RINOK_THREAD diff --git a/3rdparty/lzma/C/MtCoder.h b/3rdparty/lzma/C/MtCoder.h index f0f06da281f..1231d3c2a54 100644 --- a/3rdparty/lzma/C/MtCoder.h +++ b/3rdparty/lzma/C/MtCoder.h @@ -1,98 +1,141 @@ /* MtCoder.h -- Multi-thread Coder -2009-11-19 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ -#ifndef __MT_CODER_H -#define __MT_CODER_H +#ifndef ZIP7_INC_MT_CODER_H +#define ZIP7_INC_MT_CODER_H -#include "Threads.h" +#include "MtDec.h" EXTERN_C_BEGIN -typedef struct -{ - CThread thread; - CAutoResetEvent startEvent; - CAutoResetEvent finishedEvent; - int stop; - - THREAD_FUNC_TYPE func; - LPVOID param; - THREAD_FUNC_RET_TYPE res; -} CLoopThread; - -void LoopThread_Construct(CLoopThread *p); -void LoopThread_Close(CLoopThread *p); -WRes LoopThread_Create(CLoopThread *p); -WRes LoopThread_StopAndWait(CLoopThread *p); -WRes LoopThread_StartSubThread(CLoopThread *p); -WRes LoopThread_WaitSubThread(CLoopThread *p); - -#ifndef _7ZIP_ST -#define NUM_MT_CODER_THREADS_MAX 32 +/* + if ( defined MTCODER_USE_WRITE_THREAD) : main thread writes all data blocks to output stream + if (not defined MTCODER_USE_WRITE_THREAD) : any coder thread can write data blocks to output stream +*/ +/* #define MTCODER_USE_WRITE_THREAD */ + +#ifndef Z7_ST + #define MTCODER_GET_NUM_BLOCKS_FROM_THREADS(numThreads) ((numThreads) + (numThreads) / 8 + 1) + #define MTCODER_THREADS_MAX 64 + #define MTCODER_BLOCKS_MAX (MTCODER_GET_NUM_BLOCKS_FROM_THREADS(MTCODER_THREADS_MAX) + 3) #else -#define NUM_MT_CODER_THREADS_MAX 1 + #define MTCODER_THREADS_MAX 1 + #define MTCODER_BLOCKS_MAX 1 #endif + +#ifndef Z7_ST + + typedef struct { - UInt64 totalInSize; - UInt64 totalOutSize; - ICompressProgress *progress; - SRes res; - CCriticalSection cs; - UInt64 inSizes[NUM_MT_CODER_THREADS_MAX]; - UInt64 outSizes[NUM_MT_CODER_THREADS_MAX]; -} CMtProgress; + ICompressProgress vt; + CMtProgress *mtProgress; + UInt64 inSize; + UInt64 outSize; +} CMtProgressThunk; + +void MtProgressThunk_CreateVTable(CMtProgressThunk *p); + +#define MtProgressThunk_INIT(p) { (p)->inSize = 0; (p)->outSize = 0; } + -SRes MtProgress_Set(CMtProgress *p, unsigned index, UInt64 inSize, UInt64 outSize); +struct CMtCoder_; -struct _CMtCoder; typedef struct { - struct _CMtCoder *mtCoder; - Byte *outBuf; - size_t outBufSize; - Byte *inBuf; - size_t inBufSize; + struct CMtCoder_ *mtCoder; unsigned index; - CLoopThread thread; + int stop; + Byte *inBuf; + + CAutoResetEvent startEvent; + CThread thread; +} CMtCoderThread; - Bool stopReading; - Bool stopWriting; - CAutoResetEvent canRead; - CAutoResetEvent canWrite; -} CMtThread; typedef struct { - SRes (*Code)(void *p, unsigned index, Byte *dest, size_t *destSize, + SRes (*Code)(void *p, unsigned coderIndex, unsigned outBufIndex, const Byte *src, size_t srcSize, int finished); -} IMtCoderCallback; + SRes (*Write)(void *p, unsigned outBufIndex); +} IMtCoderCallback2; + + +typedef struct +{ + SRes res; + unsigned bufIndex; + BoolInt finished; +} CMtCoderBlock; -typedef struct _CMtCoder + +typedef struct CMtCoder_ { - size_t blockSize; - size_t destBlockSize; - unsigned numThreads; + /* input variables */ - ISeqInStream *inStream; - ISeqOutStream *outStream; - ICompressProgress *progress; - ISzAlloc *alloc; + size_t blockSize; /* size of input block */ + unsigned numThreadsMax; + UInt64 expectedDataSize; + + ISeqInStreamPtr inStream; + const Byte *inData; + size_t inDataSize; + + ICompressProgressPtr progress; + ISzAllocPtr allocBig; + + IMtCoderCallback2 *mtCallback; + void *mtCallbackObject; + + + /* internal variables */ + + size_t allocatedBufsSize; + + CAutoResetEvent readEvent; + CSemaphore blocksSemaphore; + + BoolInt stopReading; + SRes readRes; + + #ifdef MTCODER_USE_WRITE_THREAD + CAutoResetEvent writeEvents[MTCODER_BLOCKS_MAX]; + #else + CAutoResetEvent finishedEvent; + SRes writeRes; + unsigned writeIndex; + Byte ReadyBlocks[MTCODER_BLOCKS_MAX]; + LONG numFinishedThreads; + #endif + + unsigned numStartedThreadsLimit; + unsigned numStartedThreads; + + unsigned numBlocksMax; + unsigned blockIndex; + UInt64 readProcessed; - IMtCoderCallback *mtCallback; CCriticalSection cs; - SRes res; + + unsigned freeBlockHead; + unsigned freeBlockList[MTCODER_BLOCKS_MAX]; CMtProgress mtProgress; - CMtThread threads[NUM_MT_CODER_THREADS_MAX]; + CMtCoderBlock blocks[MTCODER_BLOCKS_MAX]; + CMtCoderThread threads[MTCODER_THREADS_MAX]; } CMtCoder; -void MtCoder_Construct(CMtCoder* p); -void MtCoder_Destruct(CMtCoder* p); + +void MtCoder_Construct(CMtCoder *p); +void MtCoder_Destruct(CMtCoder *p); SRes MtCoder_Code(CMtCoder *p); + +#endif + + EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/MtDec.c b/3rdparty/lzma/C/MtDec.c new file mode 100644 index 00000000000..782069926d5 --- /dev/null +++ b/3rdparty/lzma/C/MtDec.c @@ -0,0 +1,1114 @@ +/* MtDec.c -- Multi-thread Decoder +2023-04-02 : Igor Pavlov : Public domain */ + +#include "Precomp.h" + +// #define SHOW_DEBUG_INFO + +// #include <stdio.h> +#include <string.h> + +#ifdef SHOW_DEBUG_INFO +#include <stdio.h> +#endif + +#include "MtDec.h" + +#ifndef Z7_ST + +#ifdef SHOW_DEBUG_INFO +#define PRF(x) x +#else +#define PRF(x) +#endif + +#define PRF_STR_INT(s, d) PRF(printf("\n" s " %d\n", (unsigned)d)) + +void MtProgress_Init(CMtProgress *p, ICompressProgressPtr progress) +{ + p->progress = progress; + p->res = SZ_OK; + p->totalInSize = 0; + p->totalOutSize = 0; +} + + +SRes MtProgress_Progress_ST(CMtProgress *p) +{ + if (p->res == SZ_OK && p->progress) + if (ICompressProgress_Progress(p->progress, p->totalInSize, p->totalOutSize) != SZ_OK) + p->res = SZ_ERROR_PROGRESS; + return p->res; +} + + +SRes MtProgress_ProgressAdd(CMtProgress *p, UInt64 inSize, UInt64 outSize) +{ + SRes res; + CriticalSection_Enter(&p->cs); + + p->totalInSize += inSize; + p->totalOutSize += outSize; + if (p->res == SZ_OK && p->progress) + if (ICompressProgress_Progress(p->progress, p->totalInSize, p->totalOutSize) != SZ_OK) + p->res = SZ_ERROR_PROGRESS; + res = p->res; + + CriticalSection_Leave(&p->cs); + return res; +} + + +SRes MtProgress_GetError(CMtProgress *p) +{ + SRes res; + CriticalSection_Enter(&p->cs); + res = p->res; + CriticalSection_Leave(&p->cs); + return res; +} + + +void MtProgress_SetError(CMtProgress *p, SRes res) +{ + CriticalSection_Enter(&p->cs); + if (p->res == SZ_OK) + p->res = res; + CriticalSection_Leave(&p->cs); +} + + +#define RINOK_THREAD(x) RINOK_WRes(x) + + +struct CMtDecBufLink_ +{ + struct CMtDecBufLink_ *next; + void *pad[3]; +}; + +typedef struct CMtDecBufLink_ CMtDecBufLink; + +#define MTDEC__LINK_DATA_OFFSET sizeof(CMtDecBufLink) +#define MTDEC__DATA_PTR_FROM_LINK(link) ((Byte *)(link) + MTDEC__LINK_DATA_OFFSET) + + + +static THREAD_FUNC_DECL MtDec_ThreadFunc(void *pp); + + +static WRes MtDecThread_CreateEvents(CMtDecThread *t) +{ + WRes wres = AutoResetEvent_OptCreate_And_Reset(&t->canWrite); + if (wres == 0) + { + wres = AutoResetEvent_OptCreate_And_Reset(&t->canRead); + if (wres == 0) + return SZ_OK; + } + return wres; +} + + +static SRes MtDecThread_CreateAndStart(CMtDecThread *t) +{ + WRes wres = MtDecThread_CreateEvents(t); + // wres = 17; // for test + if (wres == 0) + { + if (Thread_WasCreated(&t->thread)) + return SZ_OK; + wres = Thread_Create(&t->thread, MtDec_ThreadFunc, t); + if (wres == 0) + return SZ_OK; + } + return MY_SRes_HRESULT_FROM_WRes(wres); +} + + +void MtDecThread_FreeInBufs(CMtDecThread *t) +{ + if (t->inBuf) + { + void *link = t->inBuf; + t->inBuf = NULL; + do + { + void *next = ((CMtDecBufLink *)link)->next; + ISzAlloc_Free(t->mtDec->alloc, link); + link = next; + } + while (link); + } +} + + +static void MtDecThread_CloseThread(CMtDecThread *t) +{ + if (Thread_WasCreated(&t->thread)) + { + Event_Set(&t->canWrite); /* we can disable it. There are no threads waiting canWrite in normal cases */ + Event_Set(&t->canRead); + Thread_Wait_Close(&t->thread); + } + + Event_Close(&t->canRead); + Event_Close(&t->canWrite); +} + +static void MtDec_CloseThreads(CMtDec *p) +{ + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + MtDecThread_CloseThread(&p->threads[i]); +} + +static void MtDecThread_Destruct(CMtDecThread *t) +{ + MtDecThread_CloseThread(t); + MtDecThread_FreeInBufs(t); +} + + + +static SRes MtDec_GetError_Spec(CMtDec *p, UInt64 interruptIndex, BoolInt *wasInterrupted) +{ + SRes res; + CriticalSection_Enter(&p->mtProgress.cs); + *wasInterrupted = (p->needInterrupt && interruptIndex > p->interruptIndex); + res = p->mtProgress.res; + CriticalSection_Leave(&p->mtProgress.cs); + return res; +} + +static SRes MtDec_Progress_GetError_Spec(CMtDec *p, UInt64 inSize, UInt64 outSize, UInt64 interruptIndex, BoolInt *wasInterrupted) +{ + SRes res; + CriticalSection_Enter(&p->mtProgress.cs); + + p->mtProgress.totalInSize += inSize; + p->mtProgress.totalOutSize += outSize; + if (p->mtProgress.res == SZ_OK && p->mtProgress.progress) + if (ICompressProgress_Progress(p->mtProgress.progress, p->mtProgress.totalInSize, p->mtProgress.totalOutSize) != SZ_OK) + p->mtProgress.res = SZ_ERROR_PROGRESS; + + *wasInterrupted = (p->needInterrupt && interruptIndex > p->interruptIndex); + res = p->mtProgress.res; + + CriticalSection_Leave(&p->mtProgress.cs); + + return res; +} + +static void MtDec_Interrupt(CMtDec *p, UInt64 interruptIndex) +{ + CriticalSection_Enter(&p->mtProgress.cs); + if (!p->needInterrupt || interruptIndex < p->interruptIndex) + { + p->interruptIndex = interruptIndex; + p->needInterrupt = True; + } + CriticalSection_Leave(&p->mtProgress.cs); +} + +Byte *MtDec_GetCrossBuff(CMtDec *p) +{ + Byte *cr = p->crossBlock; + if (!cr) + { + cr = (Byte *)ISzAlloc_Alloc(p->alloc, MTDEC__LINK_DATA_OFFSET + p->inBufSize); + if (!cr) + return NULL; + p->crossBlock = cr; + } + return MTDEC__DATA_PTR_FROM_LINK(cr); +} + + +/* + MtDec_ThreadFunc2() returns: + 0 - in all normal cases (even for stream error or memory allocation error) + (!= 0) - WRes error return by system threading function +*/ + +// #define MTDEC_ProgessStep (1 << 22) +#define MTDEC_ProgessStep (1 << 0) + +static WRes MtDec_ThreadFunc2(CMtDecThread *t) +{ + CMtDec *p = t->mtDec; + + PRF_STR_INT("MtDec_ThreadFunc2", t->index) + + // SetThreadAffinityMask(GetCurrentThread(), 1 << t->index); + + for (;;) + { + SRes res, codeRes; + BoolInt wasInterrupted, isAllocError, overflow, finish; + SRes threadingErrorSRes; + BoolInt needCode, needWrite, needContinue; + + size_t inDataSize_Start; + UInt64 inDataSize; + // UInt64 inDataSize_Full; + + UInt64 blockIndex; + + UInt64 inPrev = 0; + UInt64 outPrev = 0; + UInt64 inCodePos; + UInt64 outCodePos; + + Byte *afterEndData = NULL; + size_t afterEndData_Size = 0; + BoolInt afterEndData_IsCross = False; + + BoolInt canCreateNewThread = False; + // CMtDecCallbackInfo parse; + CMtDecThread *nextThread; + + PRF_STR_INT("=============== Event_Wait(&t->canRead)", t->index) + + RINOK_THREAD(Event_Wait(&t->canRead)) + if (p->exitThread) + return 0; + + PRF_STR_INT("after Event_Wait(&t->canRead)", t->index) + + // if (t->index == 3) return 19; // for test + + blockIndex = p->blockIndex++; + + // PRF(printf("\ncanRead\n")) + + res = MtDec_Progress_GetError_Spec(p, 0, 0, blockIndex, &wasInterrupted); + + finish = p->readWasFinished; + needCode = False; + needWrite = False; + isAllocError = False; + overflow = False; + + inDataSize_Start = 0; + inDataSize = 0; + // inDataSize_Full = 0; + + if (res == SZ_OK && !wasInterrupted) + { + // if (p->inStream) + { + CMtDecBufLink *prev = NULL; + CMtDecBufLink *link = (CMtDecBufLink *)t->inBuf; + size_t crossSize = p->crossEnd - p->crossStart; + + PRF(printf("\ncrossSize = %d\n", crossSize)); + + for (;;) + { + if (!link) + { + link = (CMtDecBufLink *)ISzAlloc_Alloc(p->alloc, MTDEC__LINK_DATA_OFFSET + p->inBufSize); + if (!link) + { + finish = True; + // p->allocError_for_Read_BlockIndex = blockIndex; + isAllocError = True; + break; + } + link->next = NULL; + if (prev) + { + // static unsigned g_num = 0; + // printf("\n%6d : %x", ++g_num, (unsigned)(size_t)((Byte *)link - (Byte *)prev)); + prev->next = link; + } + else + t->inBuf = (void *)link; + } + + { + Byte *data = MTDEC__DATA_PTR_FROM_LINK(link); + Byte *parseData = data; + size_t size; + + if (crossSize != 0) + { + inDataSize = crossSize; + // inDataSize_Full = inDataSize; + inDataSize_Start = crossSize; + size = crossSize; + parseData = MTDEC__DATA_PTR_FROM_LINK(p->crossBlock) + p->crossStart; + PRF(printf("\ncross : crossStart = %7d crossEnd = %7d finish = %1d", + (int)p->crossStart, (int)p->crossEnd, (int)finish)); + } + else + { + size = p->inBufSize; + + res = SeqInStream_ReadMax(p->inStream, data, &size); + + // size = 10; // test + + inDataSize += size; + // inDataSize_Full = inDataSize; + if (!prev) + inDataSize_Start = size; + + p->readProcessed += size; + finish = (size != p->inBufSize); + if (finish) + p->readWasFinished = True; + + // res = E_INVALIDARG; // test + + if (res != SZ_OK) + { + // PRF(printf("\nRead error = %d\n", res)) + // we want to decode all data before error + p->readRes = res; + // p->readError_BlockIndex = blockIndex; + p->readWasFinished = True; + finish = True; + res = SZ_OK; + // break; + } + + if (inDataSize - inPrev >= MTDEC_ProgessStep) + { + res = MtDec_Progress_GetError_Spec(p, 0, 0, blockIndex, &wasInterrupted); + if (res != SZ_OK || wasInterrupted) + break; + inPrev = inDataSize; + } + } + + { + CMtDecCallbackInfo parse; + + parse.startCall = (prev == NULL); + parse.src = parseData; + parse.srcSize = size; + parse.srcFinished = finish; + parse.canCreateNewThread = True; + + PRF(printf("\nParse size = %d\n", (unsigned)size)); + + p->mtCallback->Parse(p->mtCallbackObject, t->index, &parse); + + PRF(printf(" Parse processed = %d, state = %d \n", (unsigned)parse.srcSize, (unsigned)parse.state)); + + needWrite = True; + canCreateNewThread = parse.canCreateNewThread; + + // printf("\n\n%12I64u %12I64u", (UInt64)p->mtProgress.totalInSize, (UInt64)p->mtProgress.totalOutSize); + + if ( + // parseRes != SZ_OK || + // inDataSize - (size - parse.srcSize) > p->inBlockMax + // || + parse.state == MTDEC_PARSE_OVERFLOW + // || wasInterrupted + ) + { + // Overflow or Parse error - switch from MT decoding to ST decoding + finish = True; + overflow = True; + + { + PRF(printf("\n Overflow")); + // PRF(printf("\nisBlockFinished = %d", (unsigned)parse.blockWasFinished)); + PRF(printf("\n inDataSize = %d", (unsigned)inDataSize)); + } + + if (crossSize != 0) + memcpy(data, parseData, size); + p->crossStart = 0; + p->crossEnd = 0; + break; + } + + if (crossSize != 0) + { + memcpy(data, parseData, parse.srcSize); + p->crossStart += parse.srcSize; + } + + if (parse.state != MTDEC_PARSE_CONTINUE || finish) + { + // we don't need to parse in current thread anymore + + if (parse.state == MTDEC_PARSE_END) + finish = True; + + needCode = True; + // p->crossFinished = finish; + + if (parse.srcSize == size) + { + // full parsed - no cross transfer + p->crossStart = 0; + p->crossEnd = 0; + break; + } + + if (parse.state == MTDEC_PARSE_END) + { + afterEndData = parseData + parse.srcSize; + afterEndData_Size = size - parse.srcSize; + if (crossSize != 0) + afterEndData_IsCross = True; + // we reduce data size to required bytes (parsed only) + inDataSize -= afterEndData_Size; + if (!prev) + inDataSize_Start = parse.srcSize; + break; + } + + { + // partial parsed - need cross transfer + if (crossSize != 0) + inDataSize = parse.srcSize; // it's only parsed now + else + { + // partial parsed - is not in initial cross block - we need to copy new data to cross block + Byte *cr = MtDec_GetCrossBuff(p); + if (!cr) + { + { + PRF(printf("\ncross alloc error error\n")); + // res = SZ_ERROR_MEM; + finish = True; + // p->allocError_for_Read_BlockIndex = blockIndex; + isAllocError = True; + break; + } + } + + { + size_t crSize = size - parse.srcSize; + inDataSize -= crSize; + p->crossEnd = crSize; + p->crossStart = 0; + memcpy(cr, parseData + parse.srcSize, crSize); + } + } + + // inDataSize_Full = inDataSize; + if (!prev) + inDataSize_Start = parse.srcSize; // it's partial size (parsed only) + + finish = False; + break; + } + } + + if (parse.srcSize != size) + { + res = SZ_ERROR_FAIL; + PRF(printf("\nfinished error SZ_ERROR_FAIL = %d\n", res)); + break; + } + } + } + + prev = link; + link = link->next; + + if (crossSize != 0) + { + crossSize = 0; + p->crossStart = 0; + p->crossEnd = 0; + } + } + } + + if (res == SZ_OK) + res = MtDec_GetError_Spec(p, blockIndex, &wasInterrupted); + } + + codeRes = SZ_OK; + + if (res == SZ_OK && needCode && !wasInterrupted) + { + codeRes = p->mtCallback->PreCode(p->mtCallbackObject, t->index); + if (codeRes != SZ_OK) + { + needCode = False; + finish = True; + // SZ_ERROR_MEM is expected error here. + // if (codeRes == SZ_ERROR_MEM) - we will try single-thread decoding later. + // if (codeRes != SZ_ERROR_MEM) - we can stop decoding or try single-thread decoding. + } + } + + if (res != SZ_OK || wasInterrupted) + finish = True; + + nextThread = NULL; + threadingErrorSRes = SZ_OK; + + if (!finish) + { + if (p->numStartedThreads < p->numStartedThreads_Limit && canCreateNewThread) + { + SRes res2 = MtDecThread_CreateAndStart(&p->threads[p->numStartedThreads]); + if (res2 == SZ_OK) + { + // if (p->numStartedThreads % 1000 == 0) PRF(printf("\n numStartedThreads=%d\n", p->numStartedThreads)); + p->numStartedThreads++; + } + else + { + PRF(printf("\nERROR: numStartedThreads=%d\n", p->numStartedThreads)); + if (p->numStartedThreads == 1) + { + // if only one thread is possible, we leave muti-threading code + finish = True; + needCode = False; + threadingErrorSRes = res2; + } + else + p->numStartedThreads_Limit = p->numStartedThreads; + } + } + + if (!finish) + { + unsigned nextIndex = t->index + 1; + nextThread = &p->threads[nextIndex >= p->numStartedThreads ? 0 : nextIndex]; + RINOK_THREAD(Event_Set(&nextThread->canRead)) + // We have started executing for new iteration (with next thread) + // And that next thread now is responsible for possible exit from decoding (threading_code) + } + } + + // each call of Event_Set(&nextThread->canRead) must be followed by call of Event_Set(&nextThread->canWrite) + // if ( !finish ) we must call Event_Set(&nextThread->canWrite) in any case + // if ( finish ) we switch to single-thread mode and there are 2 ways at the end of current iteration (current block): + // - if (needContinue) after Write(&needContinue), we restore decoding with new iteration + // - otherwise we stop decoding and exit from MtDec_ThreadFunc2() + + // Don't change (finish) variable in the further code + + + // ---------- CODE ---------- + + inPrev = 0; + outPrev = 0; + inCodePos = 0; + outCodePos = 0; + + if (res == SZ_OK && needCode && codeRes == SZ_OK) + { + BoolInt isStartBlock = True; + CMtDecBufLink *link = (CMtDecBufLink *)t->inBuf; + + for (;;) + { + size_t inSize; + int stop; + + if (isStartBlock) + inSize = inDataSize_Start; + else + { + UInt64 rem = inDataSize - inCodePos; + inSize = p->inBufSize; + if (inSize > rem) + inSize = (size_t)rem; + } + + inCodePos += inSize; + stop = True; + + codeRes = p->mtCallback->Code(p->mtCallbackObject, t->index, + (const Byte *)MTDEC__DATA_PTR_FROM_LINK(link), inSize, + (inCodePos == inDataSize), // srcFinished + &inCodePos, &outCodePos, &stop); + + if (codeRes != SZ_OK) + { + PRF(printf("\nCode Interrupt error = %x\n", codeRes)); + // we interrupt only later blocks + MtDec_Interrupt(p, blockIndex); + break; + } + + if (stop || inCodePos == inDataSize) + break; + + { + const UInt64 inDelta = inCodePos - inPrev; + const UInt64 outDelta = outCodePos - outPrev; + if (inDelta >= MTDEC_ProgessStep || outDelta >= MTDEC_ProgessStep) + { + // Sleep(1); + res = MtDec_Progress_GetError_Spec(p, inDelta, outDelta, blockIndex, &wasInterrupted); + if (res != SZ_OK || wasInterrupted) + break; + inPrev = inCodePos; + outPrev = outCodePos; + } + } + + link = link->next; + isStartBlock = False; + } + } + + + // ---------- WRITE ---------- + + RINOK_THREAD(Event_Wait(&t->canWrite)) + + { + BoolInt isErrorMode = False; + BoolInt canRecode = True; + BoolInt needWriteToStream = needWrite; + + if (p->exitThread) return 0; // it's never executed in normal cases + + if (p->wasInterrupted) + wasInterrupted = True; + else + { + if (codeRes != SZ_OK) // || !needCode // check it !!! + { + p->wasInterrupted = True; + p->codeRes = codeRes; + if (codeRes == SZ_ERROR_MEM) + isAllocError = True; + } + + if (threadingErrorSRes) + { + p->wasInterrupted = True; + p->threadingErrorSRes = threadingErrorSRes; + needWriteToStream = False; + } + if (isAllocError) + { + p->wasInterrupted = True; + p->isAllocError = True; + needWriteToStream = False; + } + if (overflow) + { + p->wasInterrupted = True; + p->overflow = True; + needWriteToStream = False; + } + } + + if (needCode) + { + if (wasInterrupted) + { + inCodePos = 0; + outCodePos = 0; + } + { + const UInt64 inDelta = inCodePos - inPrev; + const UInt64 outDelta = outCodePos - outPrev; + // if (inDelta != 0 || outDelta != 0) + res = MtProgress_ProgressAdd(&p->mtProgress, inDelta, outDelta); + } + } + + needContinue = (!finish); + + // if (res == SZ_OK && needWrite && !wasInterrupted) + if (needWrite) + { + // p->inProcessed += inCodePos; + + PRF(printf("\n--Write afterSize = %d\n", (unsigned)afterEndData_Size)); + + res = p->mtCallback->Write(p->mtCallbackObject, t->index, + res == SZ_OK && needWriteToStream && !wasInterrupted, // needWrite + afterEndData, afterEndData_Size, afterEndData_IsCross, + &needContinue, + &canRecode); + + // res = SZ_ERROR_FAIL; // for test + + PRF(printf("\nAfter Write needContinue = %d\n", (unsigned)needContinue)); + PRF(printf("\nprocessed = %d\n", (unsigned)p->inProcessed)); + + if (res != SZ_OK) + { + PRF(printf("\nWrite error = %d\n", res)); + isErrorMode = True; + p->wasInterrupted = True; + } + if (res != SZ_OK + || (!needContinue && !finish)) + { + PRF(printf("\nWrite Interrupt error = %x\n", res)); + MtDec_Interrupt(p, blockIndex); + } + } + + if (canRecode) + if (!needCode + || res != SZ_OK + || p->wasInterrupted + || codeRes != SZ_OK + || wasInterrupted + || p->numFilledThreads != 0 + || isErrorMode) + { + if (p->numFilledThreads == 0) + p->filledThreadStart = t->index; + if (inDataSize != 0 || !finish) + { + t->inDataSize_Start = inDataSize_Start; + t->inDataSize = inDataSize; + p->numFilledThreads++; + } + PRF(printf("\np->numFilledThreads = %d\n", p->numFilledThreads)); + PRF(printf("p->filledThreadStart = %d\n", p->filledThreadStart)); + } + + if (!finish) + { + RINOK_THREAD(Event_Set(&nextThread->canWrite)) + } + else + { + if (needContinue) + { + // we restore decoding with new iteration + RINOK_THREAD(Event_Set(&p->threads[0].canWrite)) + } + else + { + // we exit from decoding + if (t->index == 0) + return SZ_OK; + p->exitThread = True; + } + RINOK_THREAD(Event_Set(&p->threads[0].canRead)) + } + } + } +} + +#ifdef _WIN32 +#define USE_ALLOCA +#endif + +#ifdef USE_ALLOCA +#ifdef _WIN32 +#include <malloc.h> +#else +#include <stdlib.h> +#endif +#endif + + +static THREAD_FUNC_DECL MtDec_ThreadFunc1(void *pp) +{ + WRes res; + + CMtDecThread *t = (CMtDecThread *)pp; + CMtDec *p; + + // fprintf(stdout, "\n%d = %p\n", t->index, &t); + + res = MtDec_ThreadFunc2(t); + p = t->mtDec; + if (res == 0) + return (THREAD_FUNC_RET_TYPE)(UINT_PTR)p->exitThreadWRes; + { + // it's unexpected situation for some threading function error + if (p->exitThreadWRes == 0) + p->exitThreadWRes = res; + PRF(printf("\nthread exit error = %d\n", res)); + p->exitThread = True; + Event_Set(&p->threads[0].canRead); + Event_Set(&p->threads[0].canWrite); + MtProgress_SetError(&p->mtProgress, MY_SRes_HRESULT_FROM_WRes(res)); + } + return (THREAD_FUNC_RET_TYPE)(UINT_PTR)res; +} + +static Z7_NO_INLINE THREAD_FUNC_DECL MtDec_ThreadFunc(void *pp) +{ + #ifdef USE_ALLOCA + CMtDecThread *t = (CMtDecThread *)pp; + // fprintf(stderr, "\n%d = %p - before", t->index, &t); + t->allocaPtr = alloca(t->index * 128); + #endif + return MtDec_ThreadFunc1(pp); +} + + +int MtDec_PrepareRead(CMtDec *p) +{ + if (p->crossBlock && p->crossStart == p->crossEnd) + { + ISzAlloc_Free(p->alloc, p->crossBlock); + p->crossBlock = NULL; + } + + { + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + if (i > p->numStartedThreads + || p->numFilledThreads <= + (i >= p->filledThreadStart ? + i - p->filledThreadStart : + i + p->numStartedThreads - p->filledThreadStart)) + MtDecThread_FreeInBufs(&p->threads[i]); + } + + return (p->numFilledThreads != 0) || (p->crossStart != p->crossEnd); +} + + +const Byte *MtDec_Read(CMtDec *p, size_t *inLim) +{ + while (p->numFilledThreads != 0) + { + CMtDecThread *t = &p->threads[p->filledThreadStart]; + + if (*inLim != 0) + { + { + void *link = t->inBuf; + void *next = ((CMtDecBufLink *)link)->next; + ISzAlloc_Free(p->alloc, link); + t->inBuf = next; + } + + if (t->inDataSize == 0) + { + MtDecThread_FreeInBufs(t); + if (--p->numFilledThreads == 0) + break; + if (++p->filledThreadStart == p->numStartedThreads) + p->filledThreadStart = 0; + t = &p->threads[p->filledThreadStart]; + } + } + + { + size_t lim = t->inDataSize_Start; + if (lim != 0) + t->inDataSize_Start = 0; + else + { + UInt64 rem = t->inDataSize; + lim = p->inBufSize; + if (lim > rem) + lim = (size_t)rem; + } + t->inDataSize -= lim; + *inLim = lim; + return (const Byte *)MTDEC__DATA_PTR_FROM_LINK(t->inBuf); + } + } + + { + size_t crossSize = p->crossEnd - p->crossStart; + if (crossSize != 0) + { + const Byte *data = MTDEC__DATA_PTR_FROM_LINK(p->crossBlock) + p->crossStart; + *inLim = crossSize; + p->crossStart = 0; + p->crossEnd = 0; + return data; + } + *inLim = 0; + if (p->crossBlock) + { + ISzAlloc_Free(p->alloc, p->crossBlock); + p->crossBlock = NULL; + } + return NULL; + } +} + + +void MtDec_Construct(CMtDec *p) +{ + unsigned i; + + p->inBufSize = (size_t)1 << 18; + + p->numThreadsMax = 0; + + p->inStream = NULL; + + // p->inData = NULL; + // p->inDataSize = 0; + + p->crossBlock = NULL; + p->crossStart = 0; + p->crossEnd = 0; + + p->numFilledThreads = 0; + + p->progress = NULL; + p->alloc = NULL; + + p->mtCallback = NULL; + p->mtCallbackObject = NULL; + + p->allocatedBufsSize = 0; + + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CMtDecThread *t = &p->threads[i]; + t->mtDec = p; + t->index = i; + t->inBuf = NULL; + Event_Construct(&t->canRead); + Event_Construct(&t->canWrite); + Thread_CONSTRUCT(&t->thread) + } + + // Event_Construct(&p->finishedEvent); + + CriticalSection_Init(&p->mtProgress.cs); +} + + +static void MtDec_Free(CMtDec *p) +{ + unsigned i; + + p->exitThread = True; + + for (i = 0; i < MTDEC_THREADS_MAX; i++) + MtDecThread_Destruct(&p->threads[i]); + + // Event_Close(&p->finishedEvent); + + if (p->crossBlock) + { + ISzAlloc_Free(p->alloc, p->crossBlock); + p->crossBlock = NULL; + } +} + + +void MtDec_Destruct(CMtDec *p) +{ + MtDec_Free(p); + + CriticalSection_Delete(&p->mtProgress.cs); +} + + +SRes MtDec_Code(CMtDec *p) +{ + unsigned i; + + p->inProcessed = 0; + + p->blockIndex = 1; // it must be larger than not_defined index (0) + p->isAllocError = False; + p->overflow = False; + p->threadingErrorSRes = SZ_OK; + + p->needContinue = True; + + p->readWasFinished = False; + p->needInterrupt = False; + p->interruptIndex = (UInt64)(Int64)-1; + + p->readProcessed = 0; + p->readRes = SZ_OK; + p->codeRes = SZ_OK; + p->wasInterrupted = False; + + p->crossStart = 0; + p->crossEnd = 0; + + p->filledThreadStart = 0; + p->numFilledThreads = 0; + + { + unsigned numThreads = p->numThreadsMax; + if (numThreads > MTDEC_THREADS_MAX) + numThreads = MTDEC_THREADS_MAX; + p->numStartedThreads_Limit = numThreads; + p->numStartedThreads = 0; + } + + if (p->inBufSize != p->allocatedBufsSize) + { + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CMtDecThread *t = &p->threads[i]; + if (t->inBuf) + MtDecThread_FreeInBufs(t); + } + if (p->crossBlock) + { + ISzAlloc_Free(p->alloc, p->crossBlock); + p->crossBlock = NULL; + } + + p->allocatedBufsSize = p->inBufSize; + } + + MtProgress_Init(&p->mtProgress, p->progress); + + // RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->finishedEvent)) + p->exitThread = False; + p->exitThreadWRes = 0; + + { + WRes wres; + SRes sres; + CMtDecThread *nextThread = &p->threads[p->numStartedThreads++]; + // wres = MtDecThread_CreateAndStart(nextThread); + wres = MtDecThread_CreateEvents(nextThread); + if (wres == 0) { wres = Event_Set(&nextThread->canWrite); + if (wres == 0) { wres = Event_Set(&nextThread->canRead); + if (wres == 0) { THREAD_FUNC_RET_TYPE res = MtDec_ThreadFunc(nextThread); + wres = (WRes)(UINT_PTR)res; + if (wres != 0) + { + p->needContinue = False; + MtDec_CloseThreads(p); + }}}} + + // wres = 17; // for test + // wres = Event_Wait(&p->finishedEvent); + + sres = MY_SRes_HRESULT_FROM_WRes(wres); + + if (sres != 0) + p->threadingErrorSRes = sres; + + if ( + // wres == 0 + // wres != 0 + // || p->mtc.codeRes == SZ_ERROR_MEM + p->isAllocError + || p->threadingErrorSRes != SZ_OK + || p->overflow) + { + // p->needContinue = True; + } + else + p->needContinue = False; + + if (p->needContinue) + return SZ_OK; + + // if (sres != SZ_OK) + return sres; + // return SZ_ERROR_FAIL; + } +} + +#endif + +#undef PRF diff --git a/3rdparty/lzma/C/MtDec.h b/3rdparty/lzma/C/MtDec.h new file mode 100644 index 00000000000..c28e8d9acfa --- /dev/null +++ b/3rdparty/lzma/C/MtDec.h @@ -0,0 +1,202 @@ +/* MtDec.h -- Multi-thread Decoder +2023-04-02 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_MT_DEC_H +#define ZIP7_INC_MT_DEC_H + +#include "7zTypes.h" + +#ifndef Z7_ST +#include "Threads.h" +#endif + +EXTERN_C_BEGIN + +#ifndef Z7_ST + +#ifndef Z7_ST + #define MTDEC_THREADS_MAX 32 +#else + #define MTDEC_THREADS_MAX 1 +#endif + + +typedef struct +{ + ICompressProgressPtr progress; + SRes res; + UInt64 totalInSize; + UInt64 totalOutSize; + CCriticalSection cs; +} CMtProgress; + +void MtProgress_Init(CMtProgress *p, ICompressProgressPtr progress); +SRes MtProgress_Progress_ST(CMtProgress *p); +SRes MtProgress_ProgressAdd(CMtProgress *p, UInt64 inSize, UInt64 outSize); +SRes MtProgress_GetError(CMtProgress *p); +void MtProgress_SetError(CMtProgress *p, SRes res); + +struct CMtDec; + +typedef struct +{ + struct CMtDec_ *mtDec; + unsigned index; + void *inBuf; + + size_t inDataSize_Start; // size of input data in start block + UInt64 inDataSize; // total size of input data in all blocks + + CThread thread; + CAutoResetEvent canRead; + CAutoResetEvent canWrite; + void *allocaPtr; +} CMtDecThread; + +void MtDecThread_FreeInBufs(CMtDecThread *t); + + +typedef enum +{ + MTDEC_PARSE_CONTINUE, // continue this block with more input data + MTDEC_PARSE_OVERFLOW, // MT buffers overflow, need switch to single-thread + MTDEC_PARSE_NEW, // new block + MTDEC_PARSE_END // end of block threading. But we still can return to threading after Write(&needContinue) +} EMtDecParseState; + +typedef struct +{ + // in + int startCall; + const Byte *src; + size_t srcSize; + // in : (srcSize == 0) is allowed + // out : it's allowed to return less that actually was used ? + int srcFinished; + + // out + EMtDecParseState state; + BoolInt canCreateNewThread; + UInt64 outPos; // check it (size_t) +} CMtDecCallbackInfo; + + +typedef struct +{ + void (*Parse)(void *p, unsigned coderIndex, CMtDecCallbackInfo *ci); + + // PreCode() and Code(): + // (SRes_return_result != SZ_OK) means stop decoding, no need another blocks + SRes (*PreCode)(void *p, unsigned coderIndex); + SRes (*Code)(void *p, unsigned coderIndex, + const Byte *src, size_t srcSize, int srcFinished, + UInt64 *inCodePos, UInt64 *outCodePos, int *stop); + // stop - means stop another Code calls + + + /* Write() must be called, if Parse() was called + set (needWrite) if + { + && (was not interrupted by progress) + && (was not interrupted in previous block) + } + + out: + if (*needContinue), decoder still need to continue decoding with new iteration, + even after MTDEC_PARSE_END + if (*canRecode), we didn't flush current block data, so we still can decode current block later. + */ + SRes (*Write)(void *p, unsigned coderIndex, + BoolInt needWriteToStream, + const Byte *src, size_t srcSize, BoolInt isCross, + // int srcFinished, + BoolInt *needContinue, + BoolInt *canRecode); + +} IMtDecCallback2; + + + +typedef struct CMtDec_ +{ + /* input variables */ + + size_t inBufSize; /* size of input block */ + unsigned numThreadsMax; + // size_t inBlockMax; + unsigned numThreadsMax_2; + + ISeqInStreamPtr inStream; + // const Byte *inData; + // size_t inDataSize; + + ICompressProgressPtr progress; + ISzAllocPtr alloc; + + IMtDecCallback2 *mtCallback; + void *mtCallbackObject; + + + /* internal variables */ + + size_t allocatedBufsSize; + + BoolInt exitThread; + WRes exitThreadWRes; + + UInt64 blockIndex; + BoolInt isAllocError; + BoolInt overflow; + SRes threadingErrorSRes; + + BoolInt needContinue; + + // CAutoResetEvent finishedEvent; + + SRes readRes; + SRes codeRes; + + BoolInt wasInterrupted; + + unsigned numStartedThreads_Limit; + unsigned numStartedThreads; + + Byte *crossBlock; + size_t crossStart; + size_t crossEnd; + UInt64 readProcessed; + BoolInt readWasFinished; + UInt64 inProcessed; + + unsigned filledThreadStart; + unsigned numFilledThreads; + + #ifndef Z7_ST + BoolInt needInterrupt; + UInt64 interruptIndex; + CMtProgress mtProgress; + CMtDecThread threads[MTDEC_THREADS_MAX]; + #endif +} CMtDec; + + +void MtDec_Construct(CMtDec *p); +void MtDec_Destruct(CMtDec *p); + +/* +MtDec_Code() returns: + SZ_OK - in most cases + MY_SRes_HRESULT_FROM_WRes(WRes_error) - in case of unexpected error in threading function +*/ + +SRes MtDec_Code(CMtDec *p); +Byte *MtDec_GetCrossBuff(CMtDec *p); + +int MtDec_PrepareRead(CMtDec *p); +const Byte *MtDec_Read(CMtDec *p, size_t *inLim); + +#endif + +EXTERN_C_END + +#endif diff --git a/3rdparty/lzma/C/Ppmd.h b/3rdparty/lzma/C/Ppmd.h index 5655b26d73c..66b26266ffd 100644 --- a/3rdparty/lzma/C/Ppmd.h +++ b/3rdparty/lzma/C/Ppmd.h @@ -1,15 +1,24 @@ /* Ppmd.h -- PPMD codec common code -2016-05-16 : Igor Pavlov : Public domain +2023-03-05 : Igor Pavlov : Public domain This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ -#ifndef __PPMD_H -#define __PPMD_H +#ifndef ZIP7_INC_PPMD_H +#define ZIP7_INC_PPMD_H #include "CpuArch.h" EXTERN_C_BEGIN -#ifdef MY_CPU_32BIT +#if defined(MY_CPU_SIZEOF_POINTER) && (MY_CPU_SIZEOF_POINTER == 4) +/* + PPMD code always uses 32-bit internal fields in PPMD structures to store internal references in main block. + if (PPMD_32BIT is defined), the PPMD code stores internal pointers to 32-bit reference fields. + if (PPMD_32BIT is NOT defined), the PPMD code stores internal UInt32 offsets to reference fields. + if (pointer size is 64-bit), then (PPMD_32BIT) mode is not allowed, + if (pointer size is 32-bit), then (PPMD_32BIT) mode is optional, + and it's allowed to disable PPMD_32BIT mode even if pointer is 32-bit. + PPMD code works slightly faster in (PPMD_32BIT) mode. +*/ #define PPMD_32BIT #endif @@ -28,7 +37,7 @@ EXTERN_C_BEGIN #define PPMD_N4 ((128 + 3 - 1 * PPMD_N1 - 2 * PPMD_N2 - 3 * PPMD_N3) / 4) #define PPMD_NUM_INDEXES (PPMD_N1 + PPMD_N2 + PPMD_N3 + PPMD_N4) -#pragma pack(push, 1) +MY_CPU_pragma_pack_push_1 /* Most compilers works OK here even without #pragma pack(push, 1), but some GCC compilers need it. */ /* SEE-contexts for PPM-contexts with masked symbols */ @@ -39,45 +48,120 @@ typedef struct Byte Count; /* Count to next change of Shift */ } CPpmd_See; -#define Ppmd_See_Update(p) if ((p)->Shift < PPMD_PERIOD_BITS && --(p)->Count == 0) \ - { (p)->Summ <<= 1; (p)->Count = (Byte)(3 << (p)->Shift++); } +#define Ppmd_See_UPDATE(p) \ + { if ((p)->Shift < PPMD_PERIOD_BITS && --(p)->Count == 0) \ + { (p)->Summ = (UInt16)((p)->Summ << 1); \ + (p)->Count = (Byte)(3 << (p)->Shift++); }} + typedef struct { Byte Symbol; Byte Freq; - UInt16 SuccessorLow; - UInt16 SuccessorHigh; + UInt16 Successor_0; + UInt16 Successor_1; } CPpmd_State; -#pragma pack(pop) - -typedef - #ifdef PPMD_32BIT - CPpmd_State * - #else - UInt32 - #endif - CPpmd_State_Ref; - -typedef - #ifdef PPMD_32BIT - void * - #else - UInt32 - #endif - CPpmd_Void_Ref; - -typedef - #ifdef PPMD_32BIT - Byte * - #else - UInt32 - #endif - CPpmd_Byte_Ref; +typedef struct CPpmd_State2_ +{ + Byte Symbol; + Byte Freq; +} CPpmd_State2; + +typedef struct CPpmd_State4_ +{ + UInt16 Successor_0; + UInt16 Successor_1; +} CPpmd_State4; + +MY_CPU_pragma_pop + +/* + PPMD code can write full CPpmd_State structure data to CPpmd*_Context + at (byte offset = 2) instead of some fields of original CPpmd*_Context structure. + + If we use pointers to different types, but that point to shared + memory space, we can have aliasing problem (strict aliasing). + + XLC compiler in -O2 mode can change the order of memory write instructions + in relation to read instructions, if we have use pointers to different types. + + To solve that aliasing problem we use combined CPpmd*_Context structure + with unions that contain the fields from both structures: + the original CPpmd*_Context and CPpmd_State. + So we can access the fields from both structures via one pointer, + and the compiler doesn't change the order of write instructions + in relation to read instructions. + + If we don't use memory write instructions to shared memory in + some local code, and we use only reading instructions (read only), + then probably it's safe to use pointers to different types for reading. +*/ + + + +#ifdef PPMD_32BIT + + #define Ppmd_Ref_Type(type) type * + #define Ppmd_GetRef(p, ptr) (ptr) + #define Ppmd_GetPtr(p, ptr) (ptr) + #define Ppmd_GetPtr_Type(p, ptr, note_type) (ptr) + +#else + + #define Ppmd_Ref_Type(type) UInt32 + #define Ppmd_GetRef(p, ptr) ((UInt32)((Byte *)(ptr) - (p)->Base)) + #define Ppmd_GetPtr(p, offs) ((void *)((p)->Base + (offs))) + #define Ppmd_GetPtr_Type(p, offs, type) ((type *)Ppmd_GetPtr(p, offs)) + +#endif // PPMD_32BIT + + +typedef Ppmd_Ref_Type(CPpmd_State) CPpmd_State_Ref; +typedef Ppmd_Ref_Type(void) CPpmd_Void_Ref; +typedef Ppmd_Ref_Type(Byte) CPpmd_Byte_Ref; + + +/* +#ifdef MY_CPU_LE_UNALIGN +// the unaligned 32-bit access latency can be too large, if the data is not in L1 cache. +#define Ppmd_GET_SUCCESSOR(p) ((CPpmd_Void_Ref)*(const UInt32 *)(const void *)&(p)->Successor_0) +#define Ppmd_SET_SUCCESSOR(p, v) *(UInt32 *)(void *)(void *)&(p)->Successor_0 = (UInt32)(v) + +#else +*/ + +/* + We can write 16-bit halves to 32-bit (Successor) field in any selected order. + But the native order is more consistent way. + So we use the native order, if LE/BE order can be detected here at compile time. +*/ + +#ifdef MY_CPU_BE + + #define Ppmd_GET_SUCCESSOR(p) \ + ( (CPpmd_Void_Ref) (((UInt32)(p)->Successor_0 << 16) | (p)->Successor_1) ) + + #define Ppmd_SET_SUCCESSOR(p, v) { \ + (p)->Successor_0 = (UInt16)(((UInt32)(v) >> 16) /* & 0xFFFF */); \ + (p)->Successor_1 = (UInt16)((UInt32)(v) /* & 0xFFFF */); } + +#else + + #define Ppmd_GET_SUCCESSOR(p) \ + ( (CPpmd_Void_Ref) ((p)->Successor_0 | ((UInt32)(p)->Successor_1 << 16)) ) + + #define Ppmd_SET_SUCCESSOR(p, v) { \ + (p)->Successor_0 = (UInt16)((UInt32)(v) /* & 0xFFFF */); \ + (p)->Successor_1 = (UInt16)(((UInt32)(v) >> 16) /* & 0xFFFF */); } + +#endif + +// #endif + #define PPMD_SetAllBitsIn256Bytes(p) \ - { unsigned z; for (z = 0; z < 256 / sizeof(p[0]); z += 8) { \ + { size_t z; for (z = 0; z < 256 / sizeof(p[0]); z += 8) { \ p[z+7] = p[z+6] = p[z+5] = p[z+4] = p[z+3] = p[z+2] = p[z+1] = p[z+0] = ~(size_t)0; }} EXTERN_C_END diff --git a/3rdparty/lzma/C/Ppmd7.c b/3rdparty/lzma/C/Ppmd7.c index eda8eb70a22..6e1307e2d39 100644 --- a/3rdparty/lzma/C/Ppmd7.c +++ b/3rdparty/lzma/C/Ppmd7.c @@ -1,5 +1,5 @@ /* Ppmd7.c -- PPMdH codec -2016-05-21 : Igor Pavlov : Public domain +2023-04-02 : Igor Pavlov : Public domain This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ #include "Precomp.h" @@ -8,21 +8,23 @@ This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ #include "Ppmd7.h" -const Byte PPMD7_kExpEscape[16] = { 25, 14, 9, 7, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2 }; -static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051}; +/* define PPMD7_ORDER_0_SUPPPORT to suport order-0 mode, unsupported by orignal PPMd var.H. code */ +// #define PPMD7_ORDER_0_SUPPPORT + +MY_ALIGN(16) +static const Byte PPMD7_kExpEscape[16] = { 25, 14, 9, 7, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2 }; +MY_ALIGN(16) +static const UInt16 PPMD7_kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051}; #define MAX_FREQ 124 #define UNIT_SIZE 12 #define U2B(nu) ((UInt32)(nu) * UNIT_SIZE) -#define U2I(nu) (p->Units2Indx[(nu) - 1]) -#define I2U(indx) (p->Indx2Units[indx]) +#define U2I(nu) (p->Units2Indx[(size_t)(nu) - 1]) +#define I2U(indx) ((unsigned)p->Indx2Units[indx]) +#define I2U_UInt16(indx) ((UInt16)p->Indx2Units[indx]) -#ifdef PPMD_32BIT - #define REF(ptr) (ptr) -#else - #define REF(ptr) ((UInt32)((Byte *)(ptr) - (p)->Base)) -#endif +#define REF(ptr) Ppmd_GetRef(p, ptr) #define STATS_REF(ptr) ((CPpmd_State_Ref)REF(ptr)) @@ -31,17 +33,11 @@ static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x #define ONE_STATE(ctx) Ppmd7Context_OneState(ctx) #define SUFFIX(ctx) CTX((ctx)->Suffix) -typedef CPpmd7_Context * CTX_PTR; +typedef CPpmd7_Context * PPMD7_CTX_PTR; struct CPpmd7_Node_; -typedef - #ifdef PPMD_32BIT - struct CPpmd7_Node_ * - #else - UInt32 - #endif - CPpmd7_Node_Ref; +typedef Ppmd_Ref_Type(struct CPpmd7_Node_) CPpmd7_Node_Ref; typedef struct CPpmd7_Node_ { @@ -51,17 +47,13 @@ typedef struct CPpmd7_Node_ CPpmd7_Node_Ref Prev; } CPpmd7_Node; -#ifdef PPMD_32BIT - #define NODE(ptr) (ptr) -#else - #define NODE(offs) ((CPpmd7_Node *)(p->Base + (offs))) -#endif +#define NODE(r) Ppmd_GetPtr_Type(p, r, CPpmd7_Node) void Ppmd7_Construct(CPpmd7 *p) { unsigned i, k, m; - p->Base = 0; + p->Base = NULL; for (i = 0, k = 0; i < PPMD_NUM_INDEXES; i++) { @@ -77,6 +69,7 @@ void Ppmd7_Construct(CPpmd7 *p) for (i = 0; i < 3; i++) p->NS2Indx[i] = (Byte)i; + for (m = i, k = 1; i < 256; i++) { p->NS2Indx[i] = (Byte)m; @@ -84,181 +77,236 @@ void Ppmd7_Construct(CPpmd7 *p) k = (++m) - 2; } - memset(p->HB2Flag, 0, 0x40); - memset(p->HB2Flag + 0x40, 8, 0x100 - 0x40); + memcpy(p->ExpEscape, PPMD7_kExpEscape, 16); } -void Ppmd7_Free(CPpmd7 *p, ISzAlloc *alloc) + +void Ppmd7_Free(CPpmd7 *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->Base); + ISzAlloc_Free(alloc, p->Base); p->Size = 0; - p->Base = 0; + p->Base = NULL; } -Bool Ppmd7_Alloc(CPpmd7 *p, UInt32 size, ISzAlloc *alloc) + +BoolInt Ppmd7_Alloc(CPpmd7 *p, UInt32 size, ISzAllocPtr alloc) { - if (p->Base == 0 || p->Size != size) + if (!p->Base || p->Size != size) { Ppmd7_Free(p, alloc); - p->AlignOffset = - #ifdef PPMD_32BIT - (4 - size) & 3; - #else - 4 - (size & 3); - #endif - if ((p->Base = (Byte *)alloc->Alloc(alloc, p->AlignOffset + size - #ifndef PPMD_32BIT - + UNIT_SIZE - #endif - )) == 0) + p->AlignOffset = (4 - size) & 3; + if ((p->Base = (Byte *)ISzAlloc_Alloc(alloc, p->AlignOffset + size)) == NULL) return False; p->Size = size; } return True; } -static void InsertNode(CPpmd7 *p, void *node, unsigned indx) + + +// ---------- Internal Memory Allocator ---------- + +/* We can use CPpmd7_Node in list of free units (as in Ppmd8) + But we still need one additional list walk pass in Ppmd7_GlueFreeBlocks(). + So we use simple CPpmd_Void_Ref instead of CPpmd7_Node in Ppmd7_InsertNode() / Ppmd7_RemoveNode() +*/ + +#define EMPTY_NODE 0 + + +static void Ppmd7_InsertNode(CPpmd7 *p, void *node, unsigned indx) { *((CPpmd_Void_Ref *)node) = p->FreeList[indx]; + // ((CPpmd7_Node *)node)->Next = (CPpmd7_Node_Ref)p->FreeList[indx]; + p->FreeList[indx] = REF(node); + } -static void *RemoveNode(CPpmd7 *p, unsigned indx) + +static void *Ppmd7_RemoveNode(CPpmd7 *p, unsigned indx) { CPpmd_Void_Ref *node = (CPpmd_Void_Ref *)Ppmd7_GetPtr(p, p->FreeList[indx]); p->FreeList[indx] = *node; + // CPpmd7_Node *node = NODE((CPpmd7_Node_Ref)p->FreeList[indx]); + // p->FreeList[indx] = node->Next; return node; } -static void SplitBlock(CPpmd7 *p, void *ptr, unsigned oldIndx, unsigned newIndx) + +static void Ppmd7_SplitBlock(CPpmd7 *p, void *ptr, unsigned oldIndx, unsigned newIndx) { unsigned i, nu = I2U(oldIndx) - I2U(newIndx); ptr = (Byte *)ptr + U2B(I2U(newIndx)); if (I2U(i = U2I(nu)) != nu) { unsigned k = I2U(--i); - InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1); + Ppmd7_InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1); } - InsertNode(p, ptr, i); + Ppmd7_InsertNode(p, ptr, i); } -static void GlueFreeBlocks(CPpmd7 *p) + +/* we use CPpmd7_Node_Union union to solve XLC -O2 strict pointer aliasing problem */ + +typedef union { - #ifdef PPMD_32BIT - CPpmd7_Node headItem; - CPpmd7_Node_Ref head = &headItem; - #else - CPpmd7_Node_Ref head = p->AlignOffset + p->Size; - #endif - - CPpmd7_Node_Ref n = head; - unsigned i; + CPpmd7_Node Node; + CPpmd7_Node_Ref NextRef; +} CPpmd7_Node_Union; + +/* Original PPmdH (Ppmd7) code uses doubly linked list in Ppmd7_GlueFreeBlocks() + we use single linked list similar to Ppmd8 code */ + +static void Ppmd7_GlueFreeBlocks(CPpmd7 *p) +{ + /* + we use first UInt16 field of 12-bytes UNITs as record type stamp + CPpmd_State { Byte Symbol; Byte Freq; : Freq != 0 + CPpmd7_Context { UInt16 NumStats; : NumStats != 0 + CPpmd7_Node { UInt16 Stamp : Stamp == 0 for free record + : Stamp == 1 for head record and guard + Last 12-bytes UNIT in array is always contains 12-bytes order-0 CPpmd7_Context record. + */ + CPpmd7_Node_Ref head, n = 0; + p->GlueCount = 255; - /* create doubly-linked list of free blocks */ - for (i = 0; i < PPMD_NUM_INDEXES; i++) + + /* we set guard NODE at LoUnit */ + if (p->LoUnit != p->HiUnit) + ((CPpmd7_Node *)(void *)p->LoUnit)->Stamp = 1; + { - UInt16 nu = I2U(i); - CPpmd7_Node_Ref next = (CPpmd7_Node_Ref)p->FreeList[i]; - p->FreeList[i] = 0; - while (next != 0) + /* Create list of free blocks. + We still need one additional list walk pass before Glue. */ + unsigned i; + for (i = 0; i < PPMD_NUM_INDEXES; i++) { - CPpmd7_Node *node = NODE(next); - node->Next = n; - n = NODE(n)->Prev = next; - next = *(const CPpmd7_Node_Ref *)node; - node->Stamp = 0; - node->NU = (UInt16)nu; + const UInt16 nu = I2U_UInt16(i); + CPpmd7_Node_Ref next = (CPpmd7_Node_Ref)p->FreeList[i]; + p->FreeList[i] = 0; + while (next != 0) + { + /* Don't change the order of the following commands: */ + CPpmd7_Node_Union *un = (CPpmd7_Node_Union *)NODE(next); + const CPpmd7_Node_Ref tmp = next; + next = un->NextRef; + un->Node.Stamp = EMPTY_NODE; + un->Node.NU = nu; + un->Node.Next = n; + n = tmp; + } } } - NODE(head)->Stamp = 1; - NODE(head)->Next = n; - NODE(n)->Prev = head; - if (p->LoUnit != p->HiUnit) - ((CPpmd7_Node *)p->LoUnit)->Stamp = 1; - - /* Glue free blocks */ - while (n != head) + + head = n; + /* Glue and Fill must walk the list in same direction */ { - CPpmd7_Node *node = NODE(n); - UInt32 nu = (UInt32)node->NU; - for (;;) + /* Glue free blocks */ + CPpmd7_Node_Ref *prev = &head; + while (n) { - CPpmd7_Node *node2 = NODE(n) + nu; - nu += node2->NU; - if (node2->Stamp != 0 || nu >= 0x10000) - break; - NODE(node2->Prev)->Next = node2->Next; - NODE(node2->Next)->Prev = node2->Prev; - node->NU = (UInt16)nu; + CPpmd7_Node *node = NODE(n); + UInt32 nu = node->NU; + n = node->Next; + if (nu == 0) + { + *prev = n; + continue; + } + prev = &node->Next; + for (;;) + { + CPpmd7_Node *node2 = node + nu; + nu += node2->NU; + if (node2->Stamp != EMPTY_NODE || nu >= 0x10000) + break; + node->NU = (UInt16)nu; + node2->NU = 0; + } } - n = node->Next; } - + /* Fill lists of free blocks */ - for (n = NODE(head)->Next; n != head;) + for (n = head; n != 0;) { CPpmd7_Node *node = NODE(n); - unsigned nu; - CPpmd7_Node_Ref next = node->Next; - for (nu = node->NU; nu > 128; nu -= 128, node += 128) - InsertNode(p, node, PPMD_NUM_INDEXES - 1); + UInt32 nu = node->NU; + unsigned i; + n = node->Next; + if (nu == 0) + continue; + for (; nu > 128; nu -= 128, node += 128) + Ppmd7_InsertNode(p, node, PPMD_NUM_INDEXES - 1); if (I2U(i = U2I(nu)) != nu) { unsigned k = I2U(--i); - InsertNode(p, node + k, nu - k - 1); + Ppmd7_InsertNode(p, node + k, (unsigned)nu - k - 1); } - InsertNode(p, node, i); - n = next; + Ppmd7_InsertNode(p, node, i); } } -static void *AllocUnitsRare(CPpmd7 *p, unsigned indx) + +Z7_NO_INLINE +static void *Ppmd7_AllocUnitsRare(CPpmd7 *p, unsigned indx) { unsigned i; - void *retVal; + if (p->GlueCount == 0) { - GlueFreeBlocks(p); + Ppmd7_GlueFreeBlocks(p); if (p->FreeList[indx] != 0) - return RemoveNode(p, indx); + return Ppmd7_RemoveNode(p, indx); } + i = indx; + do { if (++i == PPMD_NUM_INDEXES) { UInt32 numBytes = U2B(I2U(indx)); + Byte *us = p->UnitsStart; p->GlueCount--; - return ((UInt32)(p->UnitsStart - p->Text) > numBytes) ? (p->UnitsStart -= numBytes) : (NULL); + return ((UInt32)(us - p->Text) > numBytes) ? (p->UnitsStart = us - numBytes) : NULL; } } while (p->FreeList[i] == 0); - retVal = RemoveNode(p, i); - SplitBlock(p, retVal, i, indx); - return retVal; + + { + void *block = Ppmd7_RemoveNode(p, i); + Ppmd7_SplitBlock(p, block, i, indx); + return block; + } } -static void *AllocUnits(CPpmd7 *p, unsigned indx) + +static void *Ppmd7_AllocUnits(CPpmd7 *p, unsigned indx) { - UInt32 numBytes; if (p->FreeList[indx] != 0) - return RemoveNode(p, indx); - numBytes = U2B(I2U(indx)); - if (numBytes <= (UInt32)(p->HiUnit - p->LoUnit)) + return Ppmd7_RemoveNode(p, indx); { - void *retVal = p->LoUnit; - p->LoUnit += numBytes; - return retVal; + UInt32 numBytes = U2B(I2U(indx)); + Byte *lo = p->LoUnit; + if ((UInt32)(p->HiUnit - lo) >= numBytes) + { + p->LoUnit = lo + numBytes; + return lo; + } } - return AllocUnitsRare(p, indx); + return Ppmd7_AllocUnitsRare(p, indx); } -#define MyMem12Cpy(dest, src, num) \ - { UInt32 *d = (UInt32 *)dest; const UInt32 *s = (const UInt32 *)src; UInt32 n = num; \ - do { d[0] = s[0]; d[1] = s[1]; d[2] = s[2]; s += 3; d += 3; } while (--n); } +#define MEM_12_CPY(dest, src, num) \ + { UInt32 *d = (UInt32 *)dest; const UInt32 *z = (const UInt32 *)src; UInt32 n = num; \ + do { d[0] = z[0]; d[1] = z[1]; d[2] = z[2]; z += 3; d += 3; } while (--n); } + + +/* static void *ShrinkUnits(CPpmd7 *p, void *oldPtr, unsigned oldNU, unsigned newNU) { unsigned i0 = U2I(oldNU); @@ -267,28 +315,33 @@ static void *ShrinkUnits(CPpmd7 *p, void *oldPtr, unsigned oldNU, unsigned newNU return oldPtr; if (p->FreeList[i1] != 0) { - void *ptr = RemoveNode(p, i1); - MyMem12Cpy(ptr, oldPtr, newNU); - InsertNode(p, oldPtr, i0); + void *ptr = Ppmd7_RemoveNode(p, i1); + MEM_12_CPY(ptr, oldPtr, newNU) + Ppmd7_InsertNode(p, oldPtr, i0); return ptr; } - SplitBlock(p, oldPtr, i0, i1); + Ppmd7_SplitBlock(p, oldPtr, i0, i1); return oldPtr; } +*/ -#define SUCCESSOR(p) ((CPpmd_Void_Ref)((p)->SuccessorLow | ((UInt32)(p)->SuccessorHigh << 16))) +#define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p) static void SetSuccessor(CPpmd_State *p, CPpmd_Void_Ref v) { - (p)->SuccessorLow = (UInt16)((UInt32)(v) & 0xFFFF); - (p)->SuccessorHigh = (UInt16)(((UInt32)(v) >> 16) & 0xFFFF); + Ppmd_SET_SUCCESSOR(p, v) } -static void RestartModel(CPpmd7 *p) + + +Z7_NO_INLINE +static +void Ppmd7_RestartModel(CPpmd7 *p) { - unsigned i, k, m; + unsigned i, k; memset(p->FreeList, 0, sizeof(p->FreeList)); + p->Text = p->Base + p->AlignOffset; p->HiUnit = p->Text + p->Size; p->LoUnit = p->UnitsStart = p->HiUnit - p->Size / 8 / UNIT_SIZE * 7 * UNIT_SIZE; @@ -298,57 +351,110 @@ static void RestartModel(CPpmd7 *p) p->RunLength = p->InitRL = -(Int32)((p->MaxOrder < 12) ? p->MaxOrder : 12) - 1; p->PrevSuccess = 0; - p->MinContext = p->MaxContext = (CTX_PTR)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */ - p->MinContext->Suffix = 0; - p->MinContext->NumStats = 256; - p->MinContext->SummFreq = 256 + 1; - p->FoundState = (CPpmd_State *)p->LoUnit; /* AllocUnits(p, PPMD_NUM_INDEXES - 1); */ - p->LoUnit += U2B(256 / 2); - p->MinContext->Stats = REF(p->FoundState); - for (i = 0; i < 256; i++) { - CPpmd_State *s = &p->FoundState[i]; - s->Symbol = (Byte)i; - s->Freq = 1; - SetSuccessor(s, 0); + CPpmd7_Context *mc = (PPMD7_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */ + CPpmd_State *s = (CPpmd_State *)p->LoUnit; /* Ppmd7_AllocUnits(p, PPMD_NUM_INDEXES - 1); */ + + p->LoUnit += U2B(256 / 2); + p->MaxContext = p->MinContext = mc; + p->FoundState = s; + + mc->NumStats = 256; + mc->Union2.SummFreq = 256 + 1; + mc->Union4.Stats = REF(s); + mc->Suffix = 0; + + for (i = 0; i < 256; i++, s++) + { + s->Symbol = (Byte)i; + s->Freq = 1; + SetSuccessor(s, 0); + } + + #ifdef PPMD7_ORDER_0_SUPPPORT + if (p->MaxOrder == 0) + { + CPpmd_Void_Ref r = REF(mc); + s = p->FoundState; + for (i = 0; i < 256; i++, s++) + SetSuccessor(s, r); + return; + } + #endif } for (i = 0; i < 128; i++) + + + for (k = 0; k < 8; k++) { + unsigned m; UInt16 *dest = p->BinSumm[i] + k; - UInt16 val = (UInt16)(PPMD_BIN_SCALE - kInitBinEsc[k] / (i + 2)); + const UInt16 val = (UInt16)(PPMD_BIN_SCALE - PPMD7_kInitBinEsc[k] / (i + 2)); for (m = 0; m < 64; m += 8) dest[m] = val; } - + + for (i = 0; i < 25; i++) - for (k = 0; k < 16; k++) + { + + CPpmd_See *s = p->See[i]; + + + + unsigned summ = ((5 * i + 10) << (PPMD_PERIOD_BITS - 4)); + for (k = 0; k < 16; k++, s++) { - CPpmd_See *s = &p->See[i][k]; - s->Summ = (UInt16)((5 * i + 10) << (s->Shift = PPMD_PERIOD_BITS - 4)); + s->Summ = (UInt16)summ; + s->Shift = (PPMD_PERIOD_BITS - 4); s->Count = 4; } + } + + p->DummySee.Summ = 0; /* unused */ + p->DummySee.Shift = PPMD_PERIOD_BITS; + p->DummySee.Count = 64; /* unused */ } + void Ppmd7_Init(CPpmd7 *p, unsigned maxOrder) { p->MaxOrder = maxOrder; - RestartModel(p); - p->DummySee.Shift = PPMD_PERIOD_BITS; - p->DummySee.Summ = 0; /* unused */ - p->DummySee.Count = 64; /* unused */ + + Ppmd7_RestartModel(p); } -static CTX_PTR CreateSuccessors(CPpmd7 *p, Bool skip) + + +/* + Ppmd7_CreateSuccessors() + It's called when (FoundState->Successor) is RAW-Successor, + that is the link to position in Raw text. + So we create Context records and write the links to + FoundState->Successor and to identical RAW-Successors in suffix + contexts of MinContex. + + The function returns: + if (OrderFall == 0) then MinContext is already at MAX order, + { return pointer to new or existing context of same MAX order } + else + { return pointer to new real context that will be (Order+1) in comparison with MinContext + + also it can return pointer to real context of same order, +*/ + +Z7_NO_INLINE +static PPMD7_CTX_PTR Ppmd7_CreateSuccessors(CPpmd7 *p) { - CPpmd_State upState; - CTX_PTR c = p->MinContext; + PPMD7_CTX_PTR c = p->MinContext; CPpmd_Byte_Ref upBranch = (CPpmd_Byte_Ref)SUCCESSOR(p->FoundState); - CPpmd_State *ps[PPMD7_MAX_ORDER]; + Byte newSym, newFreq; unsigned numPs = 0; - - if (!skip) + CPpmd_State *ps[PPMD7_MAX_ORDER]; + + if (p->OrderFall != 0) ps[numPs++] = p->FoundState; while (c->Suffix) @@ -356,54 +462,83 @@ static CTX_PTR CreateSuccessors(CPpmd7 *p, Bool skip) CPpmd_Void_Ref successor; CPpmd_State *s; c = SUFFIX(c); + + if (c->NumStats != 1) { - for (s = STATS(c); s->Symbol != p->FoundState->Symbol; s++); + Byte sym = p->FoundState->Symbol; + for (s = STATS(c); s->Symbol != sym; s++); + } else + { s = ONE_STATE(c); + + } successor = SUCCESSOR(s); if (successor != upBranch) { + // (c) is real record Context here, c = CTX(successor); if (numPs == 0) + { + // (c) is real record MAX Order Context here, + // So we don't need to create any new contexts. return c; + } break; } ps[numPs++] = s; } - upState.Symbol = *(const Byte *)Ppmd7_GetPtr(p, upBranch); - SetSuccessor(&upState, upBranch + 1); + // All created contexts will have single-symbol with new RAW-Successor + // All new RAW-Successors will point to next position in RAW text + // after FoundState->Successor + + newSym = *(const Byte *)Ppmd7_GetPtr(p, upBranch); + upBranch++; + if (c->NumStats == 1) - upState.Freq = ONE_STATE(c)->Freq; + newFreq = ONE_STATE(c)->Freq; else { UInt32 cf, s0; CPpmd_State *s; - for (s = STATS(c); s->Symbol != upState.Symbol; s++); - cf = s->Freq - 1; - s0 = c->SummFreq - c->NumStats - cf; - upState.Freq = (Byte)(1 + ((2 * cf <= s0) ? (5 * cf > s0) : ((2 * cf + 3 * s0 - 1) / (2 * s0)))); + for (s = STATS(c); s->Symbol != newSym; s++); + cf = (UInt32)s->Freq - 1; + s0 = (UInt32)c->Union2.SummFreq - c->NumStats - cf; + /* + cf - is frequency of symbol that will be Successor in new context records. + s0 - is commulative frequency sum of another symbols from parent context. + max(newFreq)= (s->Freq + 1), when (s0 == 1) + we have requirement (Ppmd7Context_OneState()->Freq <= 128) in BinSumm[] + so (s->Freq < 128) - is requirement for multi-symbol contexts + */ + newFreq = (Byte)(1 + ((2 * cf <= s0) ? (5 * cf > s0) : (2 * cf + s0 - 1) / (2 * s0) + 1)); } + // Create new single-symbol contexts from low order to high order in loop + do { - /* Create Child */ - CTX_PTR c1; /* = AllocContext(p); */ + PPMD7_CTX_PTR c1; + /* = AllocContext(p); */ if (p->HiUnit != p->LoUnit) - c1 = (CTX_PTR)(p->HiUnit -= UNIT_SIZE); + c1 = (PPMD7_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); else if (p->FreeList[0] != 0) - c1 = (CTX_PTR)RemoveNode(p, 0); + c1 = (PPMD7_CTX_PTR)Ppmd7_RemoveNode(p, 0); else { - c1 = (CTX_PTR)AllocUnitsRare(p, 0); + c1 = (PPMD7_CTX_PTR)Ppmd7_AllocUnitsRare(p, 0); if (!c1) return NULL; } + c1->NumStats = 1; - *ONE_STATE(c1) = upState; + ONE_STATE(c1)->Symbol = newSym; + ONE_STATE(c1)->Freq = newFreq; + SetSuccessor(ONE_STATE(c1), upBranch); c1->Suffix = REF(c); SetSuccessor(ps[--numPs], REF(c1)); c = c1; @@ -413,21 +548,26 @@ static CTX_PTR CreateSuccessors(CPpmd7 *p, Bool skip) return c; } -static void SwapStates(CPpmd_State *t1, CPpmd_State *t2) -{ - CPpmd_State tmp = *t1; - *t1 = *t2; - *t2 = tmp; -} -static void UpdateModel(CPpmd7 *p) + +#define SWAP_STATES(s) \ + { CPpmd_State tmp = s[0]; s[0] = s[-1]; s[-1] = tmp; } + + +void Ppmd7_UpdateModel(CPpmd7 *p); +Z7_NO_INLINE +void Ppmd7_UpdateModel(CPpmd7 *p) { - CPpmd_Void_Ref successor, fSuccessor = SUCCESSOR(p->FoundState); - CTX_PTR c; + CPpmd_Void_Ref maxSuccessor, minSuccessor; + PPMD7_CTX_PTR c, mc; unsigned s0, ns; - + + + if (p->FoundState->Freq < MAX_FREQ / 4 && p->MinContext->Suffix != 0) { + /* Update Freqs in Suffix Context */ + c = SUFFIX(p->MinContext); if (c->NumStats == 1) @@ -439,73 +579,133 @@ static void UpdateModel(CPpmd7 *p) else { CPpmd_State *s = STATS(c); - if (s->Symbol != p->FoundState->Symbol) + Byte sym = p->FoundState->Symbol; + + if (s->Symbol != sym) { - do { s++; } while (s->Symbol != p->FoundState->Symbol); + do + { + // s++; if (s->Symbol == sym) break; + s++; + } + while (s->Symbol != sym); + if (s[0].Freq >= s[-1].Freq) { - SwapStates(&s[0], &s[-1]); + SWAP_STATES(s) s--; } } + if (s->Freq < MAX_FREQ - 9) { - s->Freq += 2; - c->SummFreq += 2; + s->Freq = (Byte)(s->Freq + 2); + c->Union2.SummFreq = (UInt16)(c->Union2.SummFreq + 2); } } } + if (p->OrderFall == 0) { - p->MinContext = p->MaxContext = CreateSuccessors(p, True); - if (p->MinContext == 0) + /* MAX ORDER context */ + /* (FoundState->Successor) is RAW-Successor. */ + p->MaxContext = p->MinContext = Ppmd7_CreateSuccessors(p); + if (!p->MinContext) { - RestartModel(p); + Ppmd7_RestartModel(p); return; } SetSuccessor(p->FoundState, REF(p->MinContext)); return; } + + + /* NON-MAX ORDER context */ - *p->Text++ = p->FoundState->Symbol; - successor = REF(p->Text); - if (p->Text >= p->UnitsStart) { - RestartModel(p); - return; + Byte *text = p->Text; + *text++ = p->FoundState->Symbol; + p->Text = text; + if (text >= p->UnitsStart) + { + Ppmd7_RestartModel(p); + return; + } + maxSuccessor = REF(text); } - if (fSuccessor) + minSuccessor = SUCCESSOR(p->FoundState); + + if (minSuccessor) { - if (fSuccessor <= successor) + // there is Successor for FoundState in MinContext. + // So the next context will be one order higher than MinContext. + + if (minSuccessor <= maxSuccessor) { - CTX_PTR cs = CreateSuccessors(p, False); - if (cs == NULL) + // minSuccessor is RAW-Successor. So we will create real contexts records: + PPMD7_CTX_PTR cs = Ppmd7_CreateSuccessors(p); + if (!cs) { - RestartModel(p); + Ppmd7_RestartModel(p); return; } - fSuccessor = REF(cs); + minSuccessor = REF(cs); } + + // minSuccessor now is real Context pointer that points to existing (Order+1) context + if (--p->OrderFall == 0) { - successor = fSuccessor; + /* + if we move to MaxOrder context, then minSuccessor will be common Succesor for both: + MinContext that is (MaxOrder - 1) + MaxContext that is (MaxOrder) + so we don't need new RAW-Successor, and we can use real minSuccessor + as succssors for both MinContext and MaxContext. + */ + maxSuccessor = minSuccessor; + + /* + if (MaxContext != MinContext) + { + there was order fall from MaxOrder and we don't need current symbol + to transfer some RAW-Succesors to real contexts. + So we roll back pointer in raw data for one position. + } + */ p->Text -= (p->MaxContext != p->MinContext); } } else { - SetSuccessor(p->FoundState, successor); - fSuccessor = REF(p->MinContext); + /* + FoundState has NULL-Successor here. + And only root 0-order context can contain NULL-Successors. + We change Successor in FoundState to RAW-Successor, + And next context will be same 0-order root Context. + */ + SetSuccessor(p->FoundState, maxSuccessor); + minSuccessor = REF(p->MinContext); } - - s0 = p->MinContext->SummFreq - (ns = p->MinContext->NumStats) - (p->FoundState->Freq - 1); - - for (c = p->MaxContext; c != p->MinContext; c = SUFFIX(c)) + + mc = p->MinContext; + c = p->MaxContext; + + p->MaxContext = p->MinContext = CTX(minSuccessor); + + if (c == mc) + return; + + // s0 : is pure Escape Freq + s0 = mc->Union2.SummFreq - (ns = mc->NumStats) - ((unsigned)p->FoundState->Freq - 1); + + do { unsigned ns1; - UInt32 cf, sf; + UInt32 sum; + if ((ns1 = c->NumStats) != 1) { if ((ns1 & 1) == 0) @@ -513,92 +713,139 @@ static void UpdateModel(CPpmd7 *p) /* Expand for one UNIT */ unsigned oldNU = ns1 >> 1; unsigned i = U2I(oldNU); - if (i != U2I(oldNU + 1)) + if (i != U2I((size_t)oldNU + 1)) { - void *ptr = AllocUnits(p, i + 1); + void *ptr = Ppmd7_AllocUnits(p, i + 1); void *oldPtr; if (!ptr) { - RestartModel(p); + Ppmd7_RestartModel(p); return; } oldPtr = STATS(c); - MyMem12Cpy(ptr, oldPtr, oldNU); - InsertNode(p, oldPtr, i); - c->Stats = STATS_REF(ptr); + MEM_12_CPY(ptr, oldPtr, oldNU) + Ppmd7_InsertNode(p, oldPtr, i); + c->Union4.Stats = STATS_REF(ptr); } } - c->SummFreq = (UInt16)(c->SummFreq + (2 * ns1 < ns) + 2 * ((4 * ns1 <= ns) & (c->SummFreq <= 8 * ns1))); + sum = c->Union2.SummFreq; + /* max increase of Escape_Freq is 3 here. + total increase of Union2.SummFreq for all symbols is less than 256 here */ + sum += (UInt32)(2 * ns1 < ns) + 2 * ((unsigned)(4 * ns1 <= ns) & (sum <= 8 * ns1)); + /* original PPMdH uses 16-bit variable for (sum) here. + But (sum < 0x9000). So we don't truncate (sum) to 16-bit */ + // sum = (UInt16)sum; } else { - CPpmd_State *s = (CPpmd_State*)AllocUnits(p, 0); + // instead of One-symbol context we create 2-symbol context + CPpmd_State *s = (CPpmd_State*)Ppmd7_AllocUnits(p, 0); if (!s) { - RestartModel(p); + Ppmd7_RestartModel(p); return; } - *s = *ONE_STATE(c); - c->Stats = REF(s); - if (s->Freq < MAX_FREQ / 4 - 1) - s->Freq <<= 1; - else - s->Freq = MAX_FREQ - 4; - c->SummFreq = (UInt16)(s->Freq + p->InitEsc + (ns > 3)); - } - cf = 2 * (UInt32)p->FoundState->Freq * (c->SummFreq + 6); - sf = (UInt32)s0 + c->SummFreq; - if (cf < 6 * sf) - { - cf = 1 + (cf > sf) + (cf >= 4 * sf); - c->SummFreq += 3; - } - else - { - cf = 4 + (cf >= 9 * sf) + (cf >= 12 * sf) + (cf >= 15 * sf); - c->SummFreq = (UInt16)(c->SummFreq + cf); + { + unsigned freq = c->Union2.State2.Freq; + // s = *ONE_STATE(c); + s->Symbol = c->Union2.State2.Symbol; + s->Successor_0 = c->Union4.State4.Successor_0; + s->Successor_1 = c->Union4.State4.Successor_1; + // SetSuccessor(s, c->Union4.Stats); // call it only for debug purposes to check the order of + // (Successor_0 and Successor_1) in LE/BE. + c->Union4.Stats = REF(s); + if (freq < MAX_FREQ / 4 - 1) + freq <<= 1; + else + freq = MAX_FREQ - 4; + // (max(s->freq) == 120), when we convert from 1-symbol into 2-symbol context + s->Freq = (Byte)freq; + // max(InitEsc = PPMD7_kExpEscape[*]) is 25. So the max(escapeFreq) is 26 here + sum = freq + p->InitEsc + (ns > 3); + } } + { CPpmd_State *s = STATS(c) + ns1; - SetSuccessor(s, successor); + UInt32 cf = 2 * (sum + 6) * (UInt32)p->FoundState->Freq; + UInt32 sf = (UInt32)s0 + sum; s->Symbol = p->FoundState->Symbol; - s->Freq = (Byte)cf; c->NumStats = (UInt16)(ns1 + 1); + SetSuccessor(s, maxSuccessor); + + if (cf < 6 * sf) + { + cf = (UInt32)1 + (cf > sf) + (cf >= 4 * sf); + sum += 3; + /* It can add (0, 1, 2) to Escape_Freq */ + } + else + { + cf = (UInt32)4 + (cf >= 9 * sf) + (cf >= 12 * sf) + (cf >= 15 * sf); + sum += cf; + } + + c->Union2.SummFreq = (UInt16)sum; + s->Freq = (Byte)cf; } + c = SUFFIX(c); } - p->MaxContext = p->MinContext = CTX(fSuccessor); + while (c != mc); } -static void Rescale(CPpmd7 *p) + + +Z7_NO_INLINE +static void Ppmd7_Rescale(CPpmd7 *p) { unsigned i, adder, sumFreq, escFreq; CPpmd_State *stats = STATS(p->MinContext); CPpmd_State *s = p->FoundState; + + /* Sort the list by Freq */ + if (s != stats) { CPpmd_State tmp = *s; - for (; s != stats; s--) + do s[0] = s[-1]; + while (--s != stats); *s = tmp; } - escFreq = p->MinContext->SummFreq - s->Freq; - s->Freq += 4; - adder = (p->OrderFall != 0); - s->Freq = (Byte)((s->Freq + adder) >> 1); + sumFreq = s->Freq; + escFreq = p->MinContext->Union2.SummFreq - sumFreq; + + /* + if (p->OrderFall == 0), adder = 0 : it's allowed to remove symbol from MAX Order context + if (p->OrderFall != 0), adder = 1 : it's NOT allowed to remove symbol from NON-MAX Order context + */ + + adder = (p->OrderFall != 0); + + #ifdef PPMD7_ORDER_0_SUPPPORT + adder |= (p->MaxOrder == 0); // we don't remove symbols from order-0 context + #endif + + sumFreq = (sumFreq + 4 + adder) >> 1; + i = (unsigned)p->MinContext->NumStats - 1; + s->Freq = (Byte)sumFreq; - i = p->MinContext->NumStats - 1; do { - escFreq -= (++s)->Freq; - s->Freq = (Byte)((s->Freq + adder) >> 1); - sumFreq += s->Freq; - if (s[0].Freq > s[-1].Freq) + unsigned freq = (++s)->Freq; + escFreq -= freq; + freq = (freq + adder) >> 1; + sumFreq += freq; + s->Freq = (Byte)freq; + if (freq > s[-1].Freq) { + CPpmd_State tmp = *s; CPpmd_State *s1 = s; - CPpmd_State tmp = *s1; do + { s1[0] = s1[-1]; - while (--s1 != stats && tmp.Freq > s1[-1].Freq); + } + while (--s1 != stats && freq > s1[-1].Freq); *s1 = tmp; } } @@ -606,47 +853,89 @@ static void Rescale(CPpmd7 *p) if (s->Freq == 0) { - unsigned numStats = p->MinContext->NumStats; - unsigned n0, n1; - do { i++; } while ((--s)->Freq == 0); + /* Remove all items with Freq == 0 */ + CPpmd7_Context *mc; + unsigned numStats, numStatsNew, n0, n1; + + i = 0; do { i++; } while ((--s)->Freq == 0); + + /* We increase (escFreq) for the number of removed symbols. + So we will have (0.5) increase for Escape_Freq in avarage per + removed symbol after Escape_Freq halving */ escFreq += i; - p->MinContext->NumStats = (UInt16)(p->MinContext->NumStats - i); - if (p->MinContext->NumStats == 1) + mc = p->MinContext; + numStats = mc->NumStats; + numStatsNew = numStats - i; + mc->NumStats = (UInt16)(numStatsNew); + n0 = (numStats + 1) >> 1; + + if (numStatsNew == 1) { - CPpmd_State tmp = *stats; + /* Create Single-Symbol context */ + unsigned freq = stats->Freq; + do { - tmp.Freq = (Byte)(tmp.Freq - (tmp.Freq >> 1)); escFreq >>= 1; + freq = (freq + 1) >> 1; } while (escFreq > 1); - InsertNode(p, stats, U2I(((numStats + 1) >> 1))); - *(p->FoundState = ONE_STATE(p->MinContext)) = tmp; + + s = ONE_STATE(mc); + *s = *stats; + s->Freq = (Byte)freq; // (freq <= 260 / 4) + p->FoundState = s; + Ppmd7_InsertNode(p, stats, U2I(n0)); return; } - n0 = (numStats + 1) >> 1; - n1 = (p->MinContext->NumStats + 1) >> 1; + + n1 = (numStatsNew + 1) >> 1; if (n0 != n1) - p->MinContext->Stats = STATS_REF(ShrinkUnits(p, stats, n0, n1)); + { + // p->MinContext->Union4.Stats = STATS_REF(ShrinkUnits(p, stats, n0, n1)); + unsigned i0 = U2I(n0); + unsigned i1 = U2I(n1); + if (i0 != i1) + { + if (p->FreeList[i1] != 0) + { + void *ptr = Ppmd7_RemoveNode(p, i1); + p->MinContext->Union4.Stats = STATS_REF(ptr); + MEM_12_CPY(ptr, (const void *)stats, n1) + Ppmd7_InsertNode(p, stats, i0); + } + else + Ppmd7_SplitBlock(p, stats, i0, i1); + } + } + } + { + CPpmd7_Context *mc = p->MinContext; + mc->Union2.SummFreq = (UInt16)(sumFreq + escFreq - (escFreq >> 1)); + // Escape_Freq halving here + p->FoundState = STATS(mc); } - p->MinContext->SummFreq = (UInt16)(sumFreq + escFreq - (escFreq >> 1)); - p->FoundState = STATS(p->MinContext); } + CPpmd_See *Ppmd7_MakeEscFreq(CPpmd7 *p, unsigned numMasked, UInt32 *escFreq) { CPpmd_See *see; - unsigned nonMasked = p->MinContext->NumStats - numMasked; - if (p->MinContext->NumStats != 256) + const CPpmd7_Context *mc = p->MinContext; + unsigned numStats = mc->NumStats; + if (numStats != 256) { - see = p->See[(unsigned)p->NS2Indx[nonMasked - 1]] + - (nonMasked < (unsigned)SUFFIX(p->MinContext)->NumStats - p->MinContext->NumStats) + - 2 * (unsigned)(p->MinContext->SummFreq < 11 * p->MinContext->NumStats) + - 4 * (unsigned)(numMasked > nonMasked) + + unsigned nonMasked = numStats - numMasked; + see = p->See[(unsigned)p->NS2Indx[(size_t)nonMasked - 1]] + + (nonMasked < (unsigned)SUFFIX(mc)->NumStats - numStats) + + 2 * (unsigned)(mc->Union2.SummFreq < 11 * numStats) + + 4 * (unsigned)(numMasked > nonMasked) + p->HiBitsFlag; { - unsigned r = (see->Summ >> see->Shift); - see->Summ = (UInt16)(see->Summ - r); + // if (see->Summ) field is larger than 16-bit, we need only low 16 bits of Summ + unsigned summ = (UInt16)see->Summ; // & 0xFFFF + unsigned r = (summ >> see->Shift); + see->Summ = (UInt16)(summ - r); *escFreq = r + (r == 0); } } @@ -658,53 +947,176 @@ CPpmd_See *Ppmd7_MakeEscFreq(CPpmd7 *p, unsigned numMasked, UInt32 *escFreq) return see; } -static void NextContext(CPpmd7 *p) + +static void Ppmd7_NextContext(CPpmd7 *p) { - CTX_PTR c = CTX(SUCCESSOR(p->FoundState)); - if (p->OrderFall == 0 && (Byte *)c > p->Text) - p->MinContext = p->MaxContext = c; + PPMD7_CTX_PTR c = CTX(SUCCESSOR(p->FoundState)); + if (p->OrderFall == 0 && (const Byte *)c > p->Text) + p->MaxContext = p->MinContext = c; else - UpdateModel(p); + Ppmd7_UpdateModel(p); } + void Ppmd7_Update1(CPpmd7 *p) { CPpmd_State *s = p->FoundState; - s->Freq += 4; - p->MinContext->SummFreq += 4; - if (s[0].Freq > s[-1].Freq) + unsigned freq = s->Freq; + freq += 4; + p->MinContext->Union2.SummFreq = (UInt16)(p->MinContext->Union2.SummFreq + 4); + s->Freq = (Byte)freq; + if (freq > s[-1].Freq) { - SwapStates(&s[0], &s[-1]); + SWAP_STATES(s) p->FoundState = --s; - if (s->Freq > MAX_FREQ) - Rescale(p); + if (freq > MAX_FREQ) + Ppmd7_Rescale(p); } - NextContext(p); + Ppmd7_NextContext(p); } + void Ppmd7_Update1_0(CPpmd7 *p) { - p->PrevSuccess = (2 * p->FoundState->Freq > p->MinContext->SummFreq); - p->RunLength += p->PrevSuccess; - p->MinContext->SummFreq += 4; - if ((p->FoundState->Freq += 4) > MAX_FREQ) - Rescale(p); - NextContext(p); + CPpmd_State *s = p->FoundState; + CPpmd7_Context *mc = p->MinContext; + unsigned freq = s->Freq; + unsigned summFreq = mc->Union2.SummFreq; + p->PrevSuccess = (2 * freq > summFreq); + p->RunLength += (int)p->PrevSuccess; + mc->Union2.SummFreq = (UInt16)(summFreq + 4); + freq += 4; + s->Freq = (Byte)freq; + if (freq > MAX_FREQ) + Ppmd7_Rescale(p); + Ppmd7_NextContext(p); } + +/* void Ppmd7_UpdateBin(CPpmd7 *p) { - p->FoundState->Freq = (Byte)(p->FoundState->Freq + (p->FoundState->Freq < 128 ? 1: 0)); + unsigned freq = p->FoundState->Freq; + p->FoundState->Freq = (Byte)(freq + (freq < 128)); p->PrevSuccess = 1; p->RunLength++; - NextContext(p); + Ppmd7_NextContext(p); } +*/ void Ppmd7_Update2(CPpmd7 *p) { - p->MinContext->SummFreq += 4; - if ((p->FoundState->Freq += 4) > MAX_FREQ) - Rescale(p); + CPpmd_State *s = p->FoundState; + unsigned freq = s->Freq; + freq += 4; p->RunLength = p->InitRL; - UpdateModel(p); + p->MinContext->Union2.SummFreq = (UInt16)(p->MinContext->Union2.SummFreq + 4); + s->Freq = (Byte)freq; + if (freq > MAX_FREQ) + Ppmd7_Rescale(p); + Ppmd7_UpdateModel(p); +} + + + +/* +PPMd Memory Map: +{ + [ 0 ] contains subset of original raw text, that is required to create context + records, Some symbols are not written, when max order context was reached + [ Text ] free area + [ UnitsStart ] CPpmd_State vectors and CPpmd7_Context records + [ LoUnit ] free area for CPpmd_State and CPpmd7_Context items +[ HiUnit ] CPpmd7_Context records + [ Size ] end of array } + +These addresses don't cross at any time. +And the following condtions is true for addresses: + (0 <= Text < UnitsStart <= LoUnit <= HiUnit <= Size) + +Raw text is BYTE--aligned. +the data in block [ UnitsStart ... Size ] contains 12-bytes aligned UNITs. + +Last UNIT of array at offset (Size - 12) is root order-0 CPpmd7_Context record. +The code can free UNITs memory blocks that were allocated to store CPpmd_State vectors. +The code doesn't free UNITs allocated for CPpmd7_Context records. + +The code calls Ppmd7_RestartModel(), when there is no free memory for allocation. +And Ppmd7_RestartModel() changes the state to orignal start state, with full free block. + + +The code allocates UNITs with the following order: + +Allocation of 1 UNIT for Context record + - from free space (HiUnit) down to (LoUnit) + - from FreeList[0] + - Ppmd7_AllocUnitsRare() + +Ppmd7_AllocUnits() for CPpmd_State vectors: + - from FreeList[i] + - from free space (LoUnit) up to (HiUnit) + - Ppmd7_AllocUnitsRare() + +Ppmd7_AllocUnitsRare() + - if (GlueCount == 0) + { Glue lists, GlueCount = 255, allocate from FreeList[i]] } + - loop for all higher sized FreeList[...] lists + - from (UnitsStart - Text), GlueCount-- + - ERROR + + +Each Record with Context contains the CPpmd_State vector, where each +CPpmd_State contains the link to Successor. +There are 3 types of Successor: + 1) NULL-Successor - NULL pointer. NULL-Successor links can be stored + only in 0-order Root Context Record. + We use 0 value as NULL-Successor + 2) RAW-Successor - the link to position in raw text, + that "RAW-Successor" is being created after first + occurrence of new symbol for some existing context record. + (RAW-Successor > 0). + 3) RECORD-Successor - the link to CPpmd7_Context record of (Order+1), + that record is being created when we go via RAW-Successor again. + +For any successors at any time: the following condtions are true for Successor links: +(NULL-Successor < RAW-Successor < UnitsStart <= RECORD-Successor) + + +---------- Symbol Frequency, SummFreq and Range in Range_Coder ---------- + +CPpmd7_Context::SummFreq = Sum(Stats[].Freq) + Escape_Freq + +The PPMd code tries to fulfill the condition: + (SummFreq <= (256 * 128 = RC::kBot)) + +We have (Sum(Stats[].Freq) <= 256 * 124), because of (MAX_FREQ = 124) +So (4 = 128 - 124) is average reserve for Escape_Freq for each symbol. +If (CPpmd_State::Freq) is not aligned for 4, the reserve can be 5, 6 or 7. +SummFreq and Escape_Freq can be changed in Ppmd7_Rescale() and *Update*() functions. +Ppmd7_Rescale() can remove symbols only from max-order contexts. So Escape_Freq can increase after multiple calls of Ppmd7_Rescale() for +max-order context. + +When the PPMd code still break (Total <= RC::Range) condition in range coder, +we have two ways to resolve that problem: + 1) we can report error, if we want to keep compatibility with original PPMd code that has no fix for such cases. + 2) we can reduce (Total) value to (RC::Range) by reducing (Escape_Freq) part of (Total) value. +*/ + +#undef MAX_FREQ +#undef UNIT_SIZE +#undef U2B +#undef U2I +#undef I2U +#undef I2U_UInt16 +#undef REF +#undef STATS_REF +#undef CTX +#undef STATS +#undef ONE_STATE +#undef SUFFIX +#undef NODE +#undef EMPTY_NODE +#undef MEM_12_CPY +#undef SUCCESSOR +#undef SWAP_STATES diff --git a/3rdparty/lzma/C/Ppmd7.h b/3rdparty/lzma/C/Ppmd7.h index 87eefde804a..d9eb326d60b 100644 --- a/3rdparty/lzma/C/Ppmd7.h +++ b/3rdparty/lzma/C/Ppmd7.h @@ -1,13 +1,11 @@ -/* Ppmd7.h -- PPMdH compression codec -2016-05-21 : Igor Pavlov : Public domain -This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ - -/* This code supports virtual RangeDecoder and includes the implementation -of RangeCoder from 7z, instead of RangeCoder from original PPMd var.H. -If you need the compatibility with original PPMd var.H, you can use external RangeDecoder */ +/* Ppmd7.h -- Ppmd7 (PPMdH) compression codec +2023-04-02 : Igor Pavlov : Public domain +This code is based on: + PPMd var.H (2001): Dmitry Shkarin : Public domain */ + -#ifndef __PPMD7_H -#define __PPMD7_H +#ifndef ZIP7_INC_PPMD7_H +#define ZIP7_INC_PPMD7_H #include "Ppmd.h" @@ -21,23 +19,56 @@ EXTERN_C_BEGIN struct CPpmd7_Context_; -typedef - #ifdef PPMD_32BIT - struct CPpmd7_Context_ * - #else - UInt32 - #endif - CPpmd7_Context_Ref; +typedef Ppmd_Ref_Type(struct CPpmd7_Context_) CPpmd7_Context_Ref; + +// MY_CPU_pragma_pack_push_1 typedef struct CPpmd7_Context_ { UInt16 NumStats; - UInt16 SummFreq; - CPpmd_State_Ref Stats; + + + union + { + UInt16 SummFreq; + CPpmd_State2 State2; + } Union2; + + union + { + CPpmd_State_Ref Stats; + CPpmd_State4 State4; + } Union4; + CPpmd7_Context_Ref Suffix; } CPpmd7_Context; -#define Ppmd7Context_OneState(p) ((CPpmd_State *)&(p)->SummFreq) +// MY_CPU_pragma_pop + +#define Ppmd7Context_OneState(p) ((CPpmd_State *)&(p)->Union2) + + + + +typedef struct +{ + UInt32 Range; + UInt32 Code; + UInt32 Low; + IByteInPtr Stream; +} CPpmd7_RangeDec; + + +typedef struct +{ + UInt32 Range; + Byte Cache; + // Byte _dummy_[3]; + UInt64 Low; + UInt64 CacheSize; + IByteOutPtr Stream; +} CPpmd7z_RangeEnc; + typedef struct { @@ -48,92 +79,102 @@ typedef struct UInt32 Size; UInt32 GlueCount; - Byte *Base, *LoUnit, *HiUnit, *Text, *UnitsStart; UInt32 AlignOffset; + Byte *Base, *LoUnit, *HiUnit, *Text, *UnitsStart; + - Byte Indx2Units[PPMD_NUM_INDEXES]; + + + union + { + CPpmd7_RangeDec dec; + CPpmd7z_RangeEnc enc; + } rc; + + Byte Indx2Units[PPMD_NUM_INDEXES + 2]; // +2 for alignment Byte Units2Indx[128]; CPpmd_Void_Ref FreeList[PPMD_NUM_INDEXES]; - Byte NS2Indx[256], NS2BSIndx[256], HB2Flag[256]; + + Byte NS2BSIndx[256], NS2Indx[256]; + Byte ExpEscape[16]; CPpmd_See DummySee, See[25][16]; UInt16 BinSumm[128][64]; + // int LastSymbol; } CPpmd7; + void Ppmd7_Construct(CPpmd7 *p); -Bool Ppmd7_Alloc(CPpmd7 *p, UInt32 size, ISzAlloc *alloc); -void Ppmd7_Free(CPpmd7 *p, ISzAlloc *alloc); +BoolInt Ppmd7_Alloc(CPpmd7 *p, UInt32 size, ISzAllocPtr alloc); +void Ppmd7_Free(CPpmd7 *p, ISzAllocPtr alloc); void Ppmd7_Init(CPpmd7 *p, unsigned maxOrder); #define Ppmd7_WasAllocated(p) ((p)->Base != NULL) /* ---------- Internal Functions ---------- */ -extern const Byte PPMD7_kExpEscape[16]; - -#ifdef PPMD_32BIT - #define Ppmd7_GetPtr(p, ptr) (ptr) - #define Ppmd7_GetContext(p, ptr) (ptr) - #define Ppmd7_GetStats(p, ctx) ((ctx)->Stats) -#else - #define Ppmd7_GetPtr(p, offs) ((void *)((p)->Base + (offs))) - #define Ppmd7_GetContext(p, offs) ((CPpmd7_Context *)Ppmd7_GetPtr((p), (offs))) - #define Ppmd7_GetStats(p, ctx) ((CPpmd_State *)Ppmd7_GetPtr((p), ((ctx)->Stats))) -#endif +#define Ppmd7_GetPtr(p, ptr) Ppmd_GetPtr(p, ptr) +#define Ppmd7_GetContext(p, ptr) Ppmd_GetPtr_Type(p, ptr, CPpmd7_Context) +#define Ppmd7_GetStats(p, ctx) Ppmd_GetPtr_Type(p, (ctx)->Union4.Stats, CPpmd_State) void Ppmd7_Update1(CPpmd7 *p); void Ppmd7_Update1_0(CPpmd7 *p); void Ppmd7_Update2(CPpmd7 *p); -void Ppmd7_UpdateBin(CPpmd7 *p); + +#define PPMD7_HiBitsFlag_3(sym) ((((unsigned)sym + 0xC0) >> (8 - 3)) & (1 << 3)) +#define PPMD7_HiBitsFlag_4(sym) ((((unsigned)sym + 0xC0) >> (8 - 4)) & (1 << 4)) +// #define PPMD7_HiBitsFlag_3(sym) ((sym) < 0x40 ? 0 : (1 << 3)) +// #define PPMD7_HiBitsFlag_4(sym) ((sym) < 0x40 ? 0 : (1 << 4)) #define Ppmd7_GetBinSumm(p) \ - &p->BinSumm[(unsigned)Ppmd7Context_OneState(p->MinContext)->Freq - 1][p->PrevSuccess + \ - p->NS2BSIndx[Ppmd7_GetContext(p, p->MinContext->Suffix)->NumStats - 1] + \ - (p->HiBitsFlag = p->HB2Flag[p->FoundState->Symbol]) + \ - 2 * p->HB2Flag[(unsigned)Ppmd7Context_OneState(p->MinContext)->Symbol] + \ - ((p->RunLength >> 26) & 0x20)] + &p->BinSumm[(size_t)(unsigned)Ppmd7Context_OneState(p->MinContext)->Freq - 1] \ + [ p->PrevSuccess + ((p->RunLength >> 26) & 0x20) \ + + p->NS2BSIndx[(size_t)Ppmd7_GetContext(p, p->MinContext->Suffix)->NumStats - 1] \ + + PPMD7_HiBitsFlag_4(Ppmd7Context_OneState(p->MinContext)->Symbol) \ + + (p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol)) ] CPpmd_See *Ppmd7_MakeEscFreq(CPpmd7 *p, unsigned numMasked, UInt32 *scale); -/* ---------- Decode ---------- */ +/* +We support two versions of Ppmd7 (PPMdH) methods that use same CPpmd7 structure: + 1) Ppmd7a_*: original PPMdH + 2) Ppmd7z_*: modified PPMdH with 7z Range Coder +Ppmd7_*: the structures and functions that are common for both versions of PPMd7 (PPMdH) +*/ -typedef struct -{ - UInt32 (*GetThreshold)(void *p, UInt32 total); - void (*Decode)(void *p, UInt32 start, UInt32 size); - UInt32 (*DecodeBit)(void *p, UInt32 size0); -} IPpmd7_RangeDec; +/* ---------- Decode ---------- */ -typedef struct -{ - IPpmd7_RangeDec p; - UInt32 Range; - UInt32 Code; - IByteIn *Stream; -} CPpmd7z_RangeDec; +#define PPMD7_SYM_END (-1) +#define PPMD7_SYM_ERROR (-2) -void Ppmd7z_RangeDec_CreateVTable(CPpmd7z_RangeDec *p); -Bool Ppmd7z_RangeDec_Init(CPpmd7z_RangeDec *p); -#define Ppmd7z_RangeDec_IsFinishedOK(p) ((p)->Code == 0) +/* +You must set (CPpmd7::rc.dec.Stream) before Ppmd7*_RangeDec_Init() -int Ppmd7_DecodeSymbol(CPpmd7 *p, IPpmd7_RangeDec *rc); +Ppmd7*_DecodeSymbol() +out: + >= 0 : decoded byte + -1 : PPMD7_SYM_END : End of payload marker + -2 : PPMD7_SYM_ERROR : Data error +*/ +/* Ppmd7a_* : original PPMdH */ +BoolInt Ppmd7a_RangeDec_Init(CPpmd7_RangeDec *p); +#define Ppmd7a_RangeDec_IsFinishedOK(p) ((p)->Code == 0) +int Ppmd7a_DecodeSymbol(CPpmd7 *p); -/* ---------- Encode ---------- */ +/* Ppmd7z_* : modified PPMdH with 7z Range Coder */ +BoolInt Ppmd7z_RangeDec_Init(CPpmd7_RangeDec *p); +#define Ppmd7z_RangeDec_IsFinishedOK(p) ((p)->Code == 0) +int Ppmd7z_DecodeSymbol(CPpmd7 *p); +// Byte *Ppmd7z_DecodeSymbols(CPpmd7 *p, Byte *buf, const Byte *lim); -typedef struct -{ - UInt64 Low; - UInt32 Range; - Byte Cache; - UInt64 CacheSize; - IByteOut *Stream; -} CPpmd7z_RangeEnc; -void Ppmd7z_RangeEnc_Init(CPpmd7z_RangeEnc *p); -void Ppmd7z_RangeEnc_FlushData(CPpmd7z_RangeEnc *p); +/* ---------- Encode ---------- */ -void Ppmd7_EncodeSymbol(CPpmd7 *p, CPpmd7z_RangeEnc *rc, int symbol); +void Ppmd7z_Init_RangeEnc(CPpmd7 *p); +void Ppmd7z_Flush_RangeEnc(CPpmd7 *p); +// void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol); +void Ppmd7z_EncodeSymbols(CPpmd7 *p, const Byte *buf, const Byte *lim); EXTERN_C_END diff --git a/3rdparty/lzma/C/Ppmd7Dec.c b/3rdparty/lzma/C/Ppmd7Dec.c index 04b4b09e330..832382828d1 100644 --- a/3rdparty/lzma/C/Ppmd7Dec.c +++ b/3rdparty/lzma/C/Ppmd7Dec.c @@ -1,189 +1,312 @@ -/* Ppmd7Dec.c -- PPMdH Decoder -2010-03-12 : Igor Pavlov : Public domain -This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ +/* Ppmd7Dec.c -- Ppmd7z (PPMdH with 7z Range Coder) Decoder +2023-04-02 : Igor Pavlov : Public domain +This code is based on: + PPMd var.H (2001): Dmitry Shkarin : Public domain */ + #include "Precomp.h" #include "Ppmd7.h" -#define kTopValue (1 << 24) +#define kTopValue ((UInt32)1 << 24) + + +#define READ_BYTE(p) IByteIn_Read((p)->Stream) -Bool Ppmd7z_RangeDec_Init(CPpmd7z_RangeDec *p) +BoolInt Ppmd7z_RangeDec_Init(CPpmd7_RangeDec *p) { unsigned i; p->Code = 0; p->Range = 0xFFFFFFFF; - if (p->Stream->Read((void *)p->Stream) != 0) + if (READ_BYTE(p) != 0) return False; for (i = 0; i < 4; i++) - p->Code = (p->Code << 8) | p->Stream->Read((void *)p->Stream); + p->Code = (p->Code << 8) | READ_BYTE(p); return (p->Code < 0xFFFFFFFF); } -static UInt32 Range_GetThreshold(void *pp, UInt32 total) -{ - CPpmd7z_RangeDec *p = (CPpmd7z_RangeDec *)pp; - return (p->Code) / (p->Range /= total); -} +#define RC_NORM_BASE(p) if ((p)->Range < kTopValue) \ + { (p)->Code = ((p)->Code << 8) | READ_BYTE(p); (p)->Range <<= 8; -static void Range_Normalize(CPpmd7z_RangeDec *p) -{ - if (p->Range < kTopValue) - { - p->Code = (p->Code << 8) | p->Stream->Read((void *)p->Stream); - p->Range <<= 8; - if (p->Range < kTopValue) - { - p->Code = (p->Code << 8) | p->Stream->Read((void *)p->Stream); - p->Range <<= 8; - } - } -} +#define RC_NORM_1(p) RC_NORM_BASE(p) } +#define RC_NORM(p) RC_NORM_BASE(p) RC_NORM_BASE(p) }} -static void Range_Decode(void *pp, UInt32 start, UInt32 size) -{ - CPpmd7z_RangeDec *p = (CPpmd7z_RangeDec *)pp; - p->Code -= start * p->Range; - p->Range *= size; - Range_Normalize(p); -} +// we must use only one type of Normalization from two: LOCAL or REMOTE +#define RC_NORM_LOCAL(p) // RC_NORM(p) +#define RC_NORM_REMOTE(p) RC_NORM(p) -static UInt32 Range_DecodeBit(void *pp, UInt32 size0) -{ - CPpmd7z_RangeDec *p = (CPpmd7z_RangeDec *)pp; - UInt32 newBound = (p->Range >> 14) * size0; - UInt32 symbol; - if (p->Code < newBound) - { - symbol = 0; - p->Range = newBound; - } - else - { - symbol = 1; - p->Code -= newBound; - p->Range -= newBound; - } - Range_Normalize(p); - return symbol; -} +#define R (&p->rc.dec) -void Ppmd7z_RangeDec_CreateVTable(CPpmd7z_RangeDec *p) +Z7_FORCE_INLINE +// Z7_NO_INLINE +static void Ppmd7z_RD_Decode(CPpmd7 *p, UInt32 start, UInt32 size) { - p->p.GetThreshold = Range_GetThreshold; - p->p.Decode = Range_Decode; - p->p.DecodeBit = Range_DecodeBit; + + + R->Code -= start * R->Range; + R->Range *= size; + RC_NORM_LOCAL(R) } +#define RC_Decode(start, size) Ppmd7z_RD_Decode(p, start, size); +#define RC_DecodeFinal(start, size) RC_Decode(start, size) RC_NORM_REMOTE(R) +#define RC_GetThreshold(total) (R->Code / (R->Range /= (total))) + -#define MASK(sym) ((signed char *)charMask)[sym] +#define CTX(ref) ((CPpmd7_Context *)Ppmd7_GetContext(p, ref)) +// typedef CPpmd7_Context * CTX_PTR; +#define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p) +void Ppmd7_UpdateModel(CPpmd7 *p); -int Ppmd7_DecodeSymbol(CPpmd7 *p, IPpmd7_RangeDec *rc) +#define MASK(sym) ((unsigned char *)charMask)[sym] +// Z7_FORCE_INLINE +// static +int Ppmd7z_DecodeSymbol(CPpmd7 *p) { size_t charMask[256 / sizeof(size_t)]; + if (p->MinContext->NumStats != 1) { CPpmd_State *s = Ppmd7_GetStats(p, p->MinContext); unsigned i; UInt32 count, hiCnt; - if ((count = rc->GetThreshold(rc, p->MinContext->SummFreq)) < (hiCnt = s->Freq)) + const UInt32 summFreq = p->MinContext->Union2.SummFreq; + + + + + count = RC_GetThreshold(summFreq); + hiCnt = count; + + if ((Int32)(count -= s->Freq) < 0) { - Byte symbol; - rc->Decode(rc, 0, s->Freq); + Byte sym; + RC_DecodeFinal(0, s->Freq) p->FoundState = s; - symbol = s->Symbol; + sym = s->Symbol; Ppmd7_Update1_0(p); - return symbol; + return sym; } + p->PrevSuccess = 0; - i = p->MinContext->NumStats - 1; + i = (unsigned)p->MinContext->NumStats - 1; + do { - if ((hiCnt += (++s)->Freq) > count) + if ((Int32)(count -= (++s)->Freq) < 0) { - Byte symbol; - rc->Decode(rc, hiCnt - s->Freq, s->Freq); + Byte sym; + RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq) p->FoundState = s; - symbol = s->Symbol; + sym = s->Symbol; Ppmd7_Update1(p); - return symbol; + return sym; } } while (--i); - if (count >= p->MinContext->SummFreq) - return -2; - p->HiBitsFlag = p->HB2Flag[p->FoundState->Symbol]; - rc->Decode(rc, hiCnt, p->MinContext->SummFreq - hiCnt); - PPMD_SetAllBitsIn256Bytes(charMask); - MASK(s->Symbol) = 0; - i = p->MinContext->NumStats - 1; - do { MASK((--s)->Symbol) = 0; } while (--i); + + if (hiCnt >= summFreq) + return PPMD7_SYM_ERROR; + + hiCnt -= count; + RC_Decode(hiCnt, summFreq - hiCnt) + + p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol); + PPMD_SetAllBitsIn256Bytes(charMask) + // i = p->MinContext->NumStats - 1; + // do { MASK((--s)->Symbol) = 0; } while (--i); + { + CPpmd_State *s2 = Ppmd7_GetStats(p, p->MinContext); + MASK(s->Symbol) = 0; + do + { + unsigned sym0 = s2[0].Symbol; + unsigned sym1 = s2[1].Symbol; + s2 += 2; + MASK(sym0) = 0; + MASK(sym1) = 0; + } + while (s2 < s); + } } else { + CPpmd_State *s = Ppmd7Context_OneState(p->MinContext); UInt16 *prob = Ppmd7_GetBinSumm(p); - if (rc->DecodeBit(rc, *prob) == 0) + UInt32 pr = *prob; + UInt32 size0 = (R->Range >> 14) * pr; + pr = PPMD_UPDATE_PROB_1(pr); + + if (R->Code < size0) { - Byte symbol; - *prob = (UInt16)PPMD_UPDATE_PROB_0(*prob); - symbol = (p->FoundState = Ppmd7Context_OneState(p->MinContext))->Symbol; - Ppmd7_UpdateBin(p); - return symbol; + Byte sym; + *prob = (UInt16)(pr + (1 << PPMD_INT_BITS)); + + // RangeDec_DecodeBit0(size0); + R->Range = size0; + RC_NORM_1(R) + /* we can use single byte normalization here because of + (min(BinSumm[][]) = 95) > (1 << (14 - 8)) */ + + // sym = (p->FoundState = Ppmd7Context_OneState(p->MinContext))->Symbol; + // Ppmd7_UpdateBin(p); + { + unsigned freq = s->Freq; + CPpmd7_Context *c = CTX(SUCCESSOR(s)); + sym = s->Symbol; + p->FoundState = s; + p->PrevSuccess = 1; + p->RunLength++; + s->Freq = (Byte)(freq + (freq < 128)); + // NextContext(p); + if (p->OrderFall == 0 && (const Byte *)c > p->Text) + p->MaxContext = p->MinContext = c; + else + Ppmd7_UpdateModel(p); + } + return sym; } - *prob = (UInt16)PPMD_UPDATE_PROB_1(*prob); - p->InitEsc = PPMD7_kExpEscape[*prob >> 10]; - PPMD_SetAllBitsIn256Bytes(charMask); + + *prob = (UInt16)pr; + p->InitEsc = p->ExpEscape[pr >> 10]; + + // RangeDec_DecodeBit1(size0); + + R->Code -= size0; + R->Range -= size0; + RC_NORM_LOCAL(R) + + PPMD_SetAllBitsIn256Bytes(charMask) MASK(Ppmd7Context_OneState(p->MinContext)->Symbol) = 0; p->PrevSuccess = 0; } + for (;;) { - CPpmd_State *ps[256], *s; + CPpmd_State *s, *s2; UInt32 freqSum, count, hiCnt; + CPpmd_See *see; - unsigned i, num, numMasked = p->MinContext->NumStats; + CPpmd7_Context *mc; + unsigned numMasked; + RC_NORM_REMOTE(R) + mc = p->MinContext; + numMasked = mc->NumStats; + do { p->OrderFall++; - if (!p->MinContext->Suffix) - return -1; - p->MinContext = Ppmd7_GetContext(p, p->MinContext->Suffix); + if (!mc->Suffix) + return PPMD7_SYM_END; + mc = Ppmd7_GetContext(p, mc->Suffix); } - while (p->MinContext->NumStats == numMasked); - hiCnt = 0; - s = Ppmd7_GetStats(p, p->MinContext); - i = 0; - num = p->MinContext->NumStats - numMasked; - do + while (mc->NumStats == numMasked); + + s = Ppmd7_GetStats(p, mc); + { - int k = (int)(MASK(s->Symbol)); - hiCnt += (s->Freq & k); - ps[i] = s++; - i -= k; + unsigned num = mc->NumStats; + unsigned num2 = num / 2; + + num &= 1; + hiCnt = (s->Freq & (unsigned)(MASK(s->Symbol))) & (0 - (UInt32)num); + s += num; + p->MinContext = mc; + + do + { + unsigned sym0 = s[0].Symbol; + unsigned sym1 = s[1].Symbol; + s += 2; + hiCnt += (s[-2].Freq & (unsigned)(MASK(sym0))); + hiCnt += (s[-1].Freq & (unsigned)(MASK(sym1))); + } + while (--num2); } - while (i != num); - + see = Ppmd7_MakeEscFreq(p, numMasked, &freqSum); freqSum += hiCnt; - count = rc->GetThreshold(rc, freqSum); + + + + + count = RC_GetThreshold(freqSum); if (count < hiCnt) { - Byte symbol; - CPpmd_State **pps = ps; - for (hiCnt = 0; (hiCnt += (*pps)->Freq) <= count; pps++); - s = *pps; - rc->Decode(rc, hiCnt - s->Freq, s->Freq); - Ppmd_See_Update(see); + Byte sym; + + s = Ppmd7_GetStats(p, p->MinContext); + hiCnt = count; + // count -= s->Freq & (unsigned)(MASK(s->Symbol)); + // if ((Int32)count >= 0) + { + for (;;) + { + count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break; + // count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break; + } + } + s--; + RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq) + + // new (see->Summ) value can overflow over 16-bits in some rare cases + Ppmd_See_UPDATE(see) p->FoundState = s; - symbol = s->Symbol; + sym = s->Symbol; Ppmd7_Update2(p); - return symbol; + return sym; } + if (count >= freqSum) - return -2; - rc->Decode(rc, hiCnt, freqSum - hiCnt); + return PPMD7_SYM_ERROR; + + RC_Decode(hiCnt, freqSum - hiCnt) + + // We increase (see->Summ) for sum of Freqs of all non_Masked symbols. + // new (see->Summ) value can overflow over 16-bits in some rare cases see->Summ = (UInt16)(see->Summ + freqSum); - do { MASK(ps[--i]->Symbol) = 0; } while (i != 0); + + s = Ppmd7_GetStats(p, p->MinContext); + s2 = s + p->MinContext->NumStats; + do + { + MASK(s->Symbol) = 0; + s++; + } + while (s != s2); } } + +/* +Byte *Ppmd7z_DecodeSymbols(CPpmd7 *p, Byte *buf, const Byte *lim) +{ + int sym = 0; + if (buf != lim) + do + { + sym = Ppmd7z_DecodeSymbol(p); + if (sym < 0) + break; + *buf = (Byte)sym; + } + while (++buf < lim); + p->LastSymbol = sym; + return buf; +} +*/ + +#undef kTopValue +#undef READ_BYTE +#undef RC_NORM_BASE +#undef RC_NORM_1 +#undef RC_NORM +#undef RC_NORM_LOCAL +#undef RC_NORM_REMOTE +#undef R +#undef RC_Decode +#undef RC_DecodeFinal +#undef RC_GetThreshold +#undef CTX +#undef SUCCESSOR +#undef MASK diff --git a/3rdparty/lzma/C/Ppmd7Enc.c b/3rdparty/lzma/C/Ppmd7Enc.c index e6dd3d2fc29..41106bab459 100644 --- a/3rdparty/lzma/C/Ppmd7Enc.c +++ b/3rdparty/lzma/C/Ppmd7Enc.c @@ -1,104 +1,123 @@ -/* Ppmd7Enc.c -- PPMdH Encoder -2015-09-28 : Igor Pavlov : Public domain -This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */ +/* Ppmd7Enc.c -- Ppmd7z (PPMdH with 7z Range Coder) Encoder +2023-04-02 : Igor Pavlov : Public domain +This code is based on: + PPMd var.H (2001): Dmitry Shkarin : Public domain */ + #include "Precomp.h" #include "Ppmd7.h" -#define kTopValue (1 << 24) +#define kTopValue ((UInt32)1 << 24) + +#define R (&p->rc.enc) -void Ppmd7z_RangeEnc_Init(CPpmd7z_RangeEnc *p) +void Ppmd7z_Init_RangeEnc(CPpmd7 *p) { - p->Low = 0; - p->Range = 0xFFFFFFFF; - p->Cache = 0; - p->CacheSize = 1; + R->Low = 0; + R->Range = 0xFFFFFFFF; + R->Cache = 0; + R->CacheSize = 1; } -static void RangeEnc_ShiftLow(CPpmd7z_RangeEnc *p) +Z7_NO_INLINE +static void Ppmd7z_RangeEnc_ShiftLow(CPpmd7 *p) { - if ((UInt32)p->Low < (UInt32)0xFF000000 || (unsigned)(p->Low >> 32) != 0) + if ((UInt32)R->Low < (UInt32)0xFF000000 || (unsigned)(R->Low >> 32) != 0) { - Byte temp = p->Cache; + Byte temp = R->Cache; do { - p->Stream->Write(p->Stream, (Byte)(temp + (Byte)(p->Low >> 32))); + IByteOut_Write(R->Stream, (Byte)(temp + (Byte)(R->Low >> 32))); temp = 0xFF; } - while (--p->CacheSize != 0); - p->Cache = (Byte)((UInt32)p->Low >> 24); + while (--R->CacheSize != 0); + R->Cache = (Byte)((UInt32)R->Low >> 24); } - p->CacheSize++; - p->Low = (UInt32)p->Low << 8; + R->CacheSize++; + R->Low = (UInt32)((UInt32)R->Low << 8); } -static void RangeEnc_Encode(CPpmd7z_RangeEnc *p, UInt32 start, UInt32 size, UInt32 total) -{ - p->Low += start * (p->Range /= total); - p->Range *= size; - while (p->Range < kTopValue) - { - p->Range <<= 8; - RangeEnc_ShiftLow(p); - } -} +#define RC_NORM_BASE(p) if (R->Range < kTopValue) { R->Range <<= 8; Ppmd7z_RangeEnc_ShiftLow(p); +#define RC_NORM_1(p) RC_NORM_BASE(p) } +#define RC_NORM(p) RC_NORM_BASE(p) RC_NORM_BASE(p) }} -static void RangeEnc_EncodeBit_0(CPpmd7z_RangeEnc *p, UInt32 size0) -{ - p->Range = (p->Range >> 14) * size0; - while (p->Range < kTopValue) - { - p->Range <<= 8; - RangeEnc_ShiftLow(p); - } -} +// we must use only one type of Normalization from two: LOCAL or REMOTE +#define RC_NORM_LOCAL(p) // RC_NORM(p) +#define RC_NORM_REMOTE(p) RC_NORM(p) -static void RangeEnc_EncodeBit_1(CPpmd7z_RangeEnc *p, UInt32 size0) +/* +#define Ppmd7z_RangeEnc_Encode(p, start, _size_) \ + { UInt32 size = _size_; \ + R->Low += start * R->Range; \ + R->Range *= size; \ + RC_NORM_LOCAL(p); } +*/ + +Z7_FORCE_INLINE +// Z7_NO_INLINE +static void Ppmd7z_RangeEnc_Encode(CPpmd7 *p, UInt32 start, UInt32 size) { - UInt32 newBound = (p->Range >> 14) * size0; - p->Low += newBound; - p->Range -= newBound; - while (p->Range < kTopValue) - { - p->Range <<= 8; - RangeEnc_ShiftLow(p); - } + R->Low += start * R->Range; + R->Range *= size; + RC_NORM_LOCAL(p) } -void Ppmd7z_RangeEnc_FlushData(CPpmd7z_RangeEnc *p) +void Ppmd7z_Flush_RangeEnc(CPpmd7 *p) { unsigned i; for (i = 0; i < 5; i++) - RangeEnc_ShiftLow(p); + Ppmd7z_RangeEnc_ShiftLow(p); } -#define MASK(sym) ((signed char *)charMask)[sym] -void Ppmd7_EncodeSymbol(CPpmd7 *p, CPpmd7z_RangeEnc *rc, int symbol) +#define RC_Encode(start, size) Ppmd7z_RangeEnc_Encode(p, start, size); +#define RC_EncodeFinal(start, size) RC_Encode(start, size) RC_NORM_REMOTE(p) + +#define CTX(ref) ((CPpmd7_Context *)Ppmd7_GetContext(p, ref)) +#define SUFFIX(ctx) CTX((ctx)->Suffix) +// typedef CPpmd7_Context * CTX_PTR; +#define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p) + +void Ppmd7_UpdateModel(CPpmd7 *p); + +#define MASK(sym) ((unsigned char *)charMask)[sym] + +Z7_FORCE_INLINE +static +void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol) { size_t charMask[256 / sizeof(size_t)]; + if (p->MinContext->NumStats != 1) { CPpmd_State *s = Ppmd7_GetStats(p, p->MinContext); UInt32 sum; unsigned i; + + + + + R->Range /= p->MinContext->Union2.SummFreq; + if (s->Symbol == symbol) { - RangeEnc_Encode(rc, 0, s->Freq, p->MinContext->SummFreq); + // R->Range /= p->MinContext->Union2.SummFreq; + RC_EncodeFinal(0, s->Freq) p->FoundState = s; Ppmd7_Update1_0(p); return; } p->PrevSuccess = 0; sum = s->Freq; - i = p->MinContext->NumStats - 1; + i = (unsigned)p->MinContext->NumStats - 1; do { if ((++s)->Symbol == symbol) { - RangeEnc_Encode(rc, sum, s->Freq, p->MinContext->SummFreq); + // R->Range /= p->MinContext->Union2.SummFreq; + RC_EncodeFinal(sum, s->Freq) p->FoundState = s; Ppmd7_Update1(p); return; @@ -106,82 +125,214 @@ void Ppmd7_EncodeSymbol(CPpmd7 *p, CPpmd7z_RangeEnc *rc, int symbol) sum += s->Freq; } while (--i); + + // R->Range /= p->MinContext->Union2.SummFreq; + RC_Encode(sum, p->MinContext->Union2.SummFreq - sum) - p->HiBitsFlag = p->HB2Flag[p->FoundState->Symbol]; - PPMD_SetAllBitsIn256Bytes(charMask); - MASK(s->Symbol) = 0; - i = p->MinContext->NumStats - 1; - do { MASK((--s)->Symbol) = 0; } while (--i); - RangeEnc_Encode(rc, sum, p->MinContext->SummFreq - sum, p->MinContext->SummFreq); + p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol); + PPMD_SetAllBitsIn256Bytes(charMask) + // MASK(s->Symbol) = 0; + // i = p->MinContext->NumStats - 1; + // do { MASK((--s)->Symbol) = 0; } while (--i); + { + CPpmd_State *s2 = Ppmd7_GetStats(p, p->MinContext); + MASK(s->Symbol) = 0; + do + { + unsigned sym0 = s2[0].Symbol; + unsigned sym1 = s2[1].Symbol; + s2 += 2; + MASK(sym0) = 0; + MASK(sym1) = 0; + } + while (s2 < s); + } } else { UInt16 *prob = Ppmd7_GetBinSumm(p); CPpmd_State *s = Ppmd7Context_OneState(p->MinContext); + UInt32 pr = *prob; + const UInt32 bound = (R->Range >> 14) * pr; + pr = PPMD_UPDATE_PROB_1(pr); if (s->Symbol == symbol) { - RangeEnc_EncodeBit_0(rc, *prob); - *prob = (UInt16)PPMD_UPDATE_PROB_0(*prob); - p->FoundState = s; - Ppmd7_UpdateBin(p); + *prob = (UInt16)(pr + (1 << PPMD_INT_BITS)); + // RangeEnc_EncodeBit_0(p, bound); + R->Range = bound; + RC_NORM_1(p) + + // p->FoundState = s; + // Ppmd7_UpdateBin(p); + { + const unsigned freq = s->Freq; + CPpmd7_Context *c = CTX(SUCCESSOR(s)); + p->FoundState = s; + p->PrevSuccess = 1; + p->RunLength++; + s->Freq = (Byte)(freq + (freq < 128)); + // NextContext(p); + if (p->OrderFall == 0 && (const Byte *)c > p->Text) + p->MaxContext = p->MinContext = c; + else + Ppmd7_UpdateModel(p); + } return; } - else - { - RangeEnc_EncodeBit_1(rc, *prob); - *prob = (UInt16)PPMD_UPDATE_PROB_1(*prob); - p->InitEsc = PPMD7_kExpEscape[*prob >> 10]; - PPMD_SetAllBitsIn256Bytes(charMask); - MASK(s->Symbol) = 0; - p->PrevSuccess = 0; - } + + *prob = (UInt16)pr; + p->InitEsc = p->ExpEscape[pr >> 10]; + // RangeEnc_EncodeBit_1(p, bound); + R->Low += bound; + R->Range -= bound; + RC_NORM_LOCAL(p) + + PPMD_SetAllBitsIn256Bytes(charMask) + MASK(s->Symbol) = 0; + p->PrevSuccess = 0; } + for (;;) { - UInt32 escFreq; CPpmd_See *see; CPpmd_State *s; - UInt32 sum; - unsigned i, numMasked = p->MinContext->NumStats; + UInt32 sum, escFreq; + CPpmd7_Context *mc; + unsigned i, numMasked; + + RC_NORM_REMOTE(p) + + mc = p->MinContext; + numMasked = mc->NumStats; + do { p->OrderFall++; - if (!p->MinContext->Suffix) + if (!mc->Suffix) return; /* EndMarker (symbol = -1) */ - p->MinContext = Ppmd7_GetContext(p, p->MinContext->Suffix); + mc = Ppmd7_GetContext(p, mc->Suffix); + i = mc->NumStats; } - while (p->MinContext->NumStats == numMasked); + while (i == numMasked); + + p->MinContext = mc; - see = Ppmd7_MakeEscFreq(p, numMasked, &escFreq); - s = Ppmd7_GetStats(p, p->MinContext); + // see = Ppmd7_MakeEscFreq(p, numMasked, &escFreq); + { + if (i != 256) + { + unsigned nonMasked = i - numMasked; + see = p->See[(unsigned)p->NS2Indx[(size_t)nonMasked - 1]] + + p->HiBitsFlag + + (nonMasked < (unsigned)SUFFIX(mc)->NumStats - i) + + 2 * (unsigned)(mc->Union2.SummFreq < 11 * i) + + 4 * (unsigned)(numMasked > nonMasked); + { + // if (see->Summ) field is larger than 16-bit, we need only low 16 bits of Summ + unsigned summ = (UInt16)see->Summ; // & 0xFFFF + unsigned r = (summ >> see->Shift); + see->Summ = (UInt16)(summ - r); + escFreq = r + (r == 0); + } + } + else + { + see = &p->DummySee; + escFreq = 1; + } + } + + s = Ppmd7_GetStats(p, mc); sum = 0; - i = p->MinContext->NumStats; + // i = mc->NumStats; + do { - int cur = s->Symbol; - if (cur == symbol) + const unsigned cur = s->Symbol; + if ((int)cur == symbol) { - UInt32 low = sum; - CPpmd_State *s1 = s; - do + const UInt32 low = sum; + const UInt32 freq = s->Freq; + unsigned num2; + + Ppmd_See_UPDATE(see) + p->FoundState = s; + sum += escFreq; + + num2 = i / 2; + i &= 1; + sum += freq & (0 - (UInt32)i); + if (num2 != 0) { - sum += (s->Freq & (int)(MASK(s->Symbol))); - s++; + s += i; + for (;;) + { + unsigned sym0 = s[0].Symbol; + unsigned sym1 = s[1].Symbol; + s += 2; + sum += (s[-2].Freq & (unsigned)(MASK(sym0))); + sum += (s[-1].Freq & (unsigned)(MASK(sym1))); + if (--num2 == 0) + break; + } } - while (--i); - RangeEnc_Encode(rc, low, s1->Freq, sum + escFreq); - Ppmd_See_Update(see); - p->FoundState = s1; + + + R->Range /= sum; + RC_EncodeFinal(low, freq) Ppmd7_Update2(p); return; } - sum += (s->Freq & (int)(MASK(cur))); - MASK(cur) = 0; + sum += (s->Freq & (unsigned)(MASK(cur))); s++; } while (--i); - RangeEnc_Encode(rc, sum, escFreq, sum + escFreq); - see->Summ = (UInt16)(see->Summ + sum + escFreq); + { + const UInt32 total = sum + escFreq; + see->Summ = (UInt16)(see->Summ + total); + + R->Range /= total; + RC_Encode(sum, escFreq) + } + + { + const CPpmd_State *s2 = Ppmd7_GetStats(p, p->MinContext); + s--; + MASK(s->Symbol) = 0; + do + { + const unsigned sym0 = s2[0].Symbol; + const unsigned sym1 = s2[1].Symbol; + s2 += 2; + MASK(sym0) = 0; + MASK(sym1) = 0; + } + while (s2 < s); + } } } + + +void Ppmd7z_EncodeSymbols(CPpmd7 *p, const Byte *buf, const Byte *lim) +{ + for (; buf < lim; buf++) + { + Ppmd7z_EncodeSymbol(p, *buf); + } +} + +#undef kTopValue +#undef WRITE_BYTE +#undef RC_NORM_BASE +#undef RC_NORM_1 +#undef RC_NORM +#undef RC_NORM_LOCAL +#undef RC_NORM_REMOTE +#undef R +#undef RC_Encode +#undef RC_EncodeFinal +#undef SUFFIX +#undef CTX +#undef SUCCESSOR +#undef MASK diff --git a/3rdparty/lzma/C/Precomp.h b/3rdparty/lzma/C/Precomp.h index e8ff8b40e81..69afb2ffd10 100644 --- a/3rdparty/lzma/C/Precomp.h +++ b/3rdparty/lzma/C/Precomp.h @@ -1,8 +1,8 @@ /* Precomp.h -- StdAfx -2013-11-12 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __7Z_PRECOMP_H -#define __7Z_PRECOMP_H +#ifndef ZIP7_INC_PRECOMP_H +#define ZIP7_INC_PRECOMP_H #include "Compiler.h" /* #include "7zTypes.h" */ diff --git a/3rdparty/lzma/C/RotateDefs.h b/3rdparty/lzma/C/RotateDefs.h index 8f01d1a6c59..c16b4f8e6a1 100644 --- a/3rdparty/lzma/C/RotateDefs.h +++ b/3rdparty/lzma/C/RotateDefs.h @@ -1,14 +1,14 @@ /* RotateDefs.h -- Rotate functions -2015-03-25 : Igor Pavlov : Public domain */ +2023-06-18 : Igor Pavlov : Public domain */ -#ifndef __ROTATE_DEFS_H -#define __ROTATE_DEFS_H +#ifndef ZIP7_INC_ROTATE_DEFS_H +#define ZIP7_INC_ROTATE_DEFS_H #ifdef _MSC_VER #include <stdlib.h> -/* don't use _rotl with MINGW. It can insert slow call to function. */ +/* don't use _rotl with old MINGW. It can insert slow call to function. */ /* #if (_MSC_VER >= 1200) */ #pragma intrinsic(_rotl) @@ -18,12 +18,32 @@ #define rotlFixed(x, n) _rotl((x), (n)) #define rotrFixed(x, n) _rotr((x), (n)) +#if (_MSC_VER >= 1300) +#define Z7_ROTL64(x, n) _rotl64((x), (n)) +#define Z7_ROTR64(x, n) _rotr64((x), (n)) +#else +#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (64 - (n)))) +#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (64 - (n)))) +#endif + #else /* new compilers can translate these macros to fast commands. */ +#if defined(__clang__) && (__clang_major__ >= 4) \ + || defined(__GNUC__) && (__GNUC__ >= 5) +/* GCC 4.9.0 and clang 3.5 can recognize more correct version: */ +#define rotlFixed(x, n) (((x) << (n)) | ((x) >> (-(n) & 31))) +#define rotrFixed(x, n) (((x) >> (n)) | ((x) << (-(n) & 31))) +#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (-(n) & 63))) +#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (-(n) & 63))) +#else +/* for old GCC / clang: */ #define rotlFixed(x, n) (((x) << (n)) | ((x) >> (32 - (n)))) #define rotrFixed(x, n) (((x) >> (n)) | ((x) << (32 - (n)))) +#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (64 - (n)))) +#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (64 - (n)))) +#endif #endif diff --git a/3rdparty/lzma/C/Sha256.c b/3rdparty/lzma/C/Sha256.c index b5bc19b9602..018cf6f4b4b 100644 --- a/3rdparty/lzma/C/Sha256.c +++ b/3rdparty/lzma/C/Sha256.c @@ -1,5 +1,5 @@ -/* Crypto/Sha256.c -- SHA-256 Hash -2015-11-14 : Igor Pavlov : Public domain +/* Sha256.c -- SHA-256 Hash +2023-04-02 : Igor Pavlov : Public domain This code is based on public domain code from Wei Dai's Crypto++ library. */ #include "Precomp.h" @@ -10,16 +10,108 @@ This code is based on public domain code from Wei Dai's Crypto++ library. */ #include "RotateDefs.h" #include "Sha256.h" +#if defined(_MSC_VER) && (_MSC_VER < 1900) +// #define USE_MY_MM +#endif + +#ifdef MY_CPU_X86_OR_AMD64 + #ifdef _MSC_VER + #if _MSC_VER >= 1200 + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #elif defined(__clang__) + #if (__clang_major__ >= 8) // fix that check + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #elif defined(__GNUC__) + #if (__GNUC__ >= 8) // fix that check + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #elif defined(__INTEL_COMPILER) + #if (__INTEL_COMPILER >= 1800) // fix that check + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #endif +#elif defined(MY_CPU_ARM_OR_ARM64) + #ifdef _MSC_VER + #if _MSC_VER >= 1910 + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #elif defined(__clang__) + #if (__clang_major__ >= 8) // fix that check + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #elif defined(__GNUC__) + #if (__GNUC__ >= 6) // fix that check + #define Z7_COMPILER_SHA256_SUPPORTED + #endif + #endif +#endif + +void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks); + +#ifdef Z7_COMPILER_SHA256_SUPPORTED + void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks); + + static SHA256_FUNC_UPDATE_BLOCKS g_SHA256_FUNC_UPDATE_BLOCKS = Sha256_UpdateBlocks; + static SHA256_FUNC_UPDATE_BLOCKS g_SHA256_FUNC_UPDATE_BLOCKS_HW; + + #define SHA256_UPDATE_BLOCKS(p) p->func_UpdateBlocks +#else + #define SHA256_UPDATE_BLOCKS(p) Sha256_UpdateBlocks +#endif + + +BoolInt Sha256_SetFunction(CSha256 *p, unsigned algo) +{ + SHA256_FUNC_UPDATE_BLOCKS func = Sha256_UpdateBlocks; + + #ifdef Z7_COMPILER_SHA256_SUPPORTED + if (algo != SHA256_ALGO_SW) + { + if (algo == SHA256_ALGO_DEFAULT) + func = g_SHA256_FUNC_UPDATE_BLOCKS; + else + { + if (algo != SHA256_ALGO_HW) + return False; + func = g_SHA256_FUNC_UPDATE_BLOCKS_HW; + if (!func) + return False; + } + } + #else + if (algo > 1) + return False; + #endif + + p->func_UpdateBlocks = func; + return True; +} + + /* define it for speed optimization */ -#ifndef _SFX -#define _SHA256_UNROLL -#define _SHA256_UNROLL2 + +#ifdef Z7_SFX + #define STEP_PRE 1 + #define STEP_MAIN 1 +#else + #define STEP_PRE 2 + #define STEP_MAIN 4 + // #define Z7_SHA256_UNROLL #endif -/* #define _SHA256_UNROLL2 */ +#undef Z7_SHA256_BIG_W +#if STEP_MAIN != 16 + #define Z7_SHA256_BIG_W +#endif -void Sha256_Init(CSha256 *p) + + + +void Sha256_InitState(CSha256 *p) { + p->count = 0; p->state[0] = 0x6a09e667; p->state[1] = 0xbb67ae85; p->state[2] = 0x3c6ef372; @@ -28,7 +120,17 @@ void Sha256_Init(CSha256 *p) p->state[5] = 0x9b05688c; p->state[6] = 0x1f83d9ab; p->state[7] = 0x5be0cd19; - p->count = 0; +} + +void Sha256_Init(CSha256 *p) +{ + p->func_UpdateBlocks = + #ifdef Z7_COMPILER_SHA256_SUPPORTED + g_SHA256_FUNC_UPDATE_BLOCKS; + #else + NULL; + #endif + Sha256_InitState(p); } #define S0(x) (rotrFixed(x, 2) ^ rotrFixed(x,13) ^ rotrFixed(x, 22)) @@ -36,61 +138,100 @@ void Sha256_Init(CSha256 *p) #define s0(x) (rotrFixed(x, 7) ^ rotrFixed(x,18) ^ (x >> 3)) #define s1(x) (rotrFixed(x,17) ^ rotrFixed(x,19) ^ (x >> 10)) -#define blk0(i) (W[i]) -#define blk2(i) (W[i] += s1(W[((i)-2)&15]) + W[((i)-7)&15] + s0(W[((i)-15)&15])) - #define Ch(x,y,z) (z^(x&(y^z))) #define Maj(x,y,z) ((x&y)|(z&(x|y))) -#ifdef _SHA256_UNROLL2 -#define R(a,b,c,d,e,f,g,h, i) \ - h += S1(e) + Ch(e,f,g) + K[(i)+(j)] + (j ? blk2(i) : blk0(i)); \ - d += h; \ - h += S0(a) + Maj(a, b, c) - -#define RX_8(i) \ - R(a,b,c,d,e,f,g,h, i); \ - R(h,a,b,c,d,e,f,g, i+1); \ - R(g,h,a,b,c,d,e,f, i+2); \ - R(f,g,h,a,b,c,d,e, i+3); \ - R(e,f,g,h,a,b,c,d, i+4); \ - R(d,e,f,g,h,a,b,c, i+5); \ - R(c,d,e,f,g,h,a,b, i+6); \ - R(b,c,d,e,f,g,h,a, i+7) +#define W_PRE(i) (W[(i) + (size_t)(j)] = GetBe32(data + ((size_t)(j) + i) * 4)) -#define RX_16 RX_8(0); RX_8(8); +#define blk2_main(j, i) s1(w(j, (i)-2)) + w(j, (i)-7) + s0(w(j, (i)-15)) +#ifdef Z7_SHA256_BIG_W + // we use +i instead of +(i) to change the order to solve CLANG compiler warning for signed/unsigned. + #define w(j, i) W[(size_t)(j) + i] + #define blk2(j, i) (w(j, i) = w(j, (i)-16) + blk2_main(j, i)) #else + #if STEP_MAIN == 16 + #define w(j, i) W[(i) & 15] + #else + #define w(j, i) W[((size_t)(j) + (i)) & 15] + #endif + #define blk2(j, i) (w(j, i) += blk2_main(j, i)) +#endif -#define a(i) T[(0-(i))&7] -#define b(i) T[(1-(i))&7] -#define c(i) T[(2-(i))&7] -#define d(i) T[(3-(i))&7] -#define e(i) T[(4-(i))&7] -#define f(i) T[(5-(i))&7] -#define g(i) T[(6-(i))&7] -#define h(i) T[(7-(i))&7] - -#define R(i) \ - h(i) += S1(e(i)) + Ch(e(i),f(i),g(i)) + K[(i)+(j)] + (j ? blk2(i) : blk0(i)); \ - d(i) += h(i); \ - h(i) += S0(a(i)) + Maj(a(i), b(i), c(i)) \ +#define W_MAIN(i) blk2(j, i) -#ifdef _SHA256_UNROLL -#define RX_8(i) R(i+0); R(i+1); R(i+2); R(i+3); R(i+4); R(i+5); R(i+6); R(i+7); -#define RX_16 RX_8(0); RX_8(8); +#define T1(wx, i) \ + tmp = h + S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \ + h = g; \ + g = f; \ + f = e; \ + e = d + tmp; \ + tmp += S0(a) + Maj(a, b, c); \ + d = c; \ + c = b; \ + b = a; \ + a = tmp; \ -#else +#define R1_PRE(i) T1( W_PRE, i) +#define R1_MAIN(i) T1( W_MAIN, i) -#define RX_16 unsigned i; for (i = 0; i < 16; i++) { R(i); } +#if (!defined(Z7_SHA256_UNROLL) || STEP_MAIN < 8) && (STEP_MAIN >= 4) +#define R2_MAIN(i) \ + R1_MAIN(i) \ + R1_MAIN(i + 1) \ #endif + + +#if defined(Z7_SHA256_UNROLL) && STEP_MAIN >= 8 + +#define T4( a,b,c,d,e,f,g,h, wx, i) \ + h += S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \ + tmp = h; \ + h += d; \ + d = tmp + S0(a) + Maj(a, b, c); \ + +#define R4( wx, i) \ + T4 ( a,b,c,d,e,f,g,h, wx, (i )); \ + T4 ( d,a,b,c,h,e,f,g, wx, (i+1)); \ + T4 ( c,d,a,b,g,h,e,f, wx, (i+2)); \ + T4 ( b,c,d,a,f,g,h,e, wx, (i+3)); \ + +#define R4_PRE(i) R4( W_PRE, i) +#define R4_MAIN(i) R4( W_MAIN, i) + + +#define T8( a,b,c,d,e,f,g,h, wx, i) \ + h += S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \ + d += h; \ + h += S0(a) + Maj(a, b, c); \ + +#define R8( wx, i) \ + T8 ( a,b,c,d,e,f,g,h, wx, i ); \ + T8 ( h,a,b,c,d,e,f,g, wx, i+1); \ + T8 ( g,h,a,b,c,d,e,f, wx, i+2); \ + T8 ( f,g,h,a,b,c,d,e, wx, i+3); \ + T8 ( e,f,g,h,a,b,c,d, wx, i+4); \ + T8 ( d,e,f,g,h,a,b,c, wx, i+5); \ + T8 ( c,d,e,f,g,h,a,b, wx, i+6); \ + T8 ( b,c,d,e,f,g,h,a, wx, i+7); \ + +#define R8_PRE(i) R8( W_PRE, i) +#define R8_MAIN(i) R8( W_MAIN, i) + #endif -static const UInt32 K[64] = { +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks); + +// static +extern MY_ALIGN(64) +const UInt32 SHA256_K_ARRAY[64]; + +MY_ALIGN(64) +const UInt32 SHA256_K_ARRAY[64] = { 0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5, 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, @@ -109,30 +250,27 @@ static const UInt32 K[64] = { 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2 }; -static void Sha256_WriteByteBlock(CSha256 *p) -{ - UInt32 W[16]; - unsigned j; - UInt32 *state; +#define K SHA256_K_ARRAY - #ifdef _SHA256_UNROLL2 - UInt32 a,b,c,d,e,f,g,h; + +Z7_NO_INLINE +void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks) +{ + UInt32 W + #ifdef Z7_SHA256_BIG_W + [64]; #else - UInt32 T[8]; + [16]; #endif - for (j = 0; j < 16; j += 4) - { - const Byte *ccc = p->buffer + j * 4; - W[j ] = GetBe32(ccc); - W[j + 1] = GetBe32(ccc + 4); - W[j + 2] = GetBe32(ccc + 8); - W[j + 3] = GetBe32(ccc + 12); - } + unsigned j; - state = p->state; + UInt32 a,b,c,d,e,f,g,h; - #ifdef _SHA256_UNROLL2 + #if !defined(Z7_SHA256_UNROLL) || (STEP_MAIN <= 4) || (STEP_PRE <= 4) + UInt32 tmp; + #endif + a = state[0]; b = state[1]; c = state[2]; @@ -141,39 +279,96 @@ static void Sha256_WriteByteBlock(CSha256 *p) f = state[5]; g = state[6]; h = state[7]; - #else - for (j = 0; j < 8; j++) - T[j] = state[j]; - #endif - for (j = 0; j < 64; j += 16) + while (numBlocks) { - RX_16 + + for (j = 0; j < 16; j += STEP_PRE) + { + #if STEP_PRE > 4 + + #if STEP_PRE < 8 + R4_PRE(0); + #else + R8_PRE(0); + #if STEP_PRE == 16 + R8_PRE(8); + #endif + #endif + + #else + + R1_PRE(0) + #if STEP_PRE >= 2 + R1_PRE(1) + #if STEP_PRE >= 4 + R1_PRE(2) + R1_PRE(3) + #endif + #endif + + #endif + } + + for (j = 16; j < 64; j += STEP_MAIN) + { + #if defined(Z7_SHA256_UNROLL) && STEP_MAIN >= 8 + + #if STEP_MAIN < 8 + R4_MAIN(0) + #else + R8_MAIN(0) + #if STEP_MAIN == 16 + R8_MAIN(8) + #endif + #endif + + #else + + R1_MAIN(0) + #if STEP_MAIN >= 2 + R1_MAIN(1) + #if STEP_MAIN >= 4 + R2_MAIN(2) + #if STEP_MAIN >= 8 + R2_MAIN(4) + R2_MAIN(6) + #if STEP_MAIN >= 16 + R2_MAIN(8) + R2_MAIN(10) + R2_MAIN(12) + R2_MAIN(14) + #endif + #endif + #endif + #endif + #endif + } + + a += state[0]; state[0] = a; + b += state[1]; state[1] = b; + c += state[2]; state[2] = c; + d += state[3]; state[3] = d; + e += state[4]; state[4] = e; + f += state[5]; state[5] = f; + g += state[6]; state[6] = g; + h += state[7]; state[7] = h; + + data += 64; + numBlocks--; } - #ifdef _SHA256_UNROLL2 - state[0] += a; - state[1] += b; - state[2] += c; - state[3] += d; - state[4] += e; - state[5] += f; - state[6] += g; - state[7] += h; - #else - for (j = 0; j < 8; j++) - state[j] += T[j]; - #endif - /* Wipe variables */ /* memset(W, 0, sizeof(W)); */ - /* memset(T, 0, sizeof(T)); */ } #undef S0 #undef S1 #undef s0 #undef s1 +#undef K + +#define Sha256_UpdateBlock(p) SHA256_UPDATE_BLOCKS(p)(p->state, p->buffer, 1) void Sha256_Update(CSha256 *p, const Byte *data, size_t size) { @@ -193,25 +388,26 @@ void Sha256_Update(CSha256 *p, const Byte *data, size_t size) return; } - size -= num; - memcpy(p->buffer + pos, data, num); - data += num; + if (pos != 0) + { + size -= num; + memcpy(p->buffer + pos, data, num); + data += num; + Sha256_UpdateBlock(p); + } } - - for (;;) { - Sha256_WriteByteBlock(p); - if (size < 64) - break; - size -= 64; - memcpy(p->buffer, data, 64); - data += 64; - } - - if (size != 0) + size_t numBlocks = size >> 6; + SHA256_UPDATE_BLOCKS(p)(p->state, data, numBlocks); + size &= 0x3F; + if (size == 0) + return; + data += (numBlocks << 6); memcpy(p->buffer, data, size); + } } + void Sha256_Final(CSha256 *p, Byte *digest) { unsigned pos = (unsigned)p->count & 0x3F; @@ -219,30 +415,102 @@ void Sha256_Final(CSha256 *p, Byte *digest) p->buffer[pos++] = 0x80; - while (pos != (64 - 8)) + if (pos > (64 - 8)) + { + while (pos != 64) { p->buffer[pos++] = 0; } + // memset(&p->buf.buffer[pos], 0, 64 - pos); + Sha256_UpdateBlock(p); + pos = 0; + } + + /* + if (pos & 3) + { + p->buffer[pos] = 0; + p->buffer[pos + 1] = 0; + p->buffer[pos + 2] = 0; + pos += 3; + pos &= ~3; + } { - pos &= 0x3F; - if (pos == 0) - Sha256_WriteByteBlock(p); - p->buffer[pos++] = 0; + for (; pos < 64 - 8; pos += 4) + *(UInt32 *)(&p->buffer[pos]) = 0; } + */ + + memset(&p->buffer[pos], 0, (64 - 8) - pos); { UInt64 numBits = (p->count << 3); - SetBe32(p->buffer + 64 - 8, (UInt32)(numBits >> 32)); - SetBe32(p->buffer + 64 - 4, (UInt32)(numBits)); + SetBe32(p->buffer + 64 - 8, (UInt32)(numBits >> 32)) + SetBe32(p->buffer + 64 - 4, (UInt32)(numBits)) } - Sha256_WriteByteBlock(p); + Sha256_UpdateBlock(p); for (i = 0; i < 8; i += 2) { UInt32 v0 = p->state[i]; - UInt32 v1 = p->state[i + 1]; - SetBe32(digest , v0); - SetBe32(digest + 4, v1); + UInt32 v1 = p->state[(size_t)i + 1]; + SetBe32(digest , v0) + SetBe32(digest + 4, v1) digest += 8; } - Sha256_Init(p); + Sha256_InitState(p); } + + +void Sha256Prepare(void) +{ + #ifdef Z7_COMPILER_SHA256_SUPPORTED + SHA256_FUNC_UPDATE_BLOCKS f, f_hw; + f = Sha256_UpdateBlocks; + f_hw = NULL; + #ifdef MY_CPU_X86_OR_AMD64 + #ifndef USE_MY_MM + if (CPU_IsSupported_SHA() + && CPU_IsSupported_SSSE3() + // && CPU_IsSupported_SSE41() + ) + #endif + #else + if (CPU_IsSupported_SHA2()) + #endif + { + // printf("\n========== HW SHA256 ======== \n"); + f = f_hw = Sha256_UpdateBlocks_HW; + } + g_SHA256_FUNC_UPDATE_BLOCKS = f; + g_SHA256_FUNC_UPDATE_BLOCKS_HW = f_hw; + #endif +} + +#undef S0 +#undef S1 +#undef s0 +#undef s1 +#undef Ch +#undef Maj +#undef W_MAIN +#undef W_PRE +#undef w +#undef blk2_main +#undef blk2 +#undef T1 +#undef T4 +#undef T8 +#undef R1_PRE +#undef R1_MAIN +#undef R2_MAIN +#undef R4 +#undef R4_PRE +#undef R4_MAIN +#undef R8 +#undef R8_PRE +#undef R8_MAIN +#undef STEP_PRE +#undef STEP_MAIN +#undef Z7_SHA256_BIG_W +#undef Z7_SHA256_UNROLL +#undef Z7_COMPILER_SHA256_SUPPORTED diff --git a/3rdparty/lzma/C/Sha256.h b/3rdparty/lzma/C/Sha256.h index 3f455dbc0d8..9e0422320c7 100644 --- a/3rdparty/lzma/C/Sha256.h +++ b/3rdparty/lzma/C/Sha256.h @@ -1,26 +1,76 @@ /* Sha256.h -- SHA-256 Hash -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __CRYPTO_SHA256_H -#define __CRYPTO_SHA256_H +#ifndef ZIP7_INC_SHA256_H +#define ZIP7_INC_SHA256_H #include "7zTypes.h" EXTERN_C_BEGIN -#define SHA256_DIGEST_SIZE 32 +#define SHA256_NUM_BLOCK_WORDS 16 +#define SHA256_NUM_DIGEST_WORDS 8 + +#define SHA256_BLOCK_SIZE (SHA256_NUM_BLOCK_WORDS * 4) +#define SHA256_DIGEST_SIZE (SHA256_NUM_DIGEST_WORDS * 4) + +typedef void (Z7_FASTCALL *SHA256_FUNC_UPDATE_BLOCKS)(UInt32 state[8], const Byte *data, size_t numBlocks); + +/* + if (the system supports different SHA256 code implementations) + { + (CSha256::func_UpdateBlocks) will be used + (CSha256::func_UpdateBlocks) can be set by + Sha256_Init() - to default (fastest) + Sha256_SetFunction() - to any algo + } + else + { + (CSha256::func_UpdateBlocks) is ignored. + } +*/ typedef struct { - UInt32 state[8]; + SHA256_FUNC_UPDATE_BLOCKS func_UpdateBlocks; UInt64 count; - Byte buffer[64]; + UInt64 _pad_2[2]; + UInt32 state[SHA256_NUM_DIGEST_WORDS]; + + Byte buffer[SHA256_BLOCK_SIZE]; } CSha256; + +#define SHA256_ALGO_DEFAULT 0 +#define SHA256_ALGO_SW 1 +#define SHA256_ALGO_HW 2 + +/* +Sha256_SetFunction() +return: + 0 - (algo) value is not supported, and func_UpdateBlocks was not changed + 1 - func_UpdateBlocks was set according (algo) value. +*/ + +BoolInt Sha256_SetFunction(CSha256 *p, unsigned algo); + +void Sha256_InitState(CSha256 *p); void Sha256_Init(CSha256 *p); void Sha256_Update(CSha256 *p, const Byte *data, size_t size); void Sha256_Final(CSha256 *p, Byte *digest); + + + +// void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks); + +/* +call Sha256Prepare() once at program start. +It prepares all supported implementations, and detects the fastest implementation. +*/ + +void Sha256Prepare(void); + EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/Sha256Opt.c b/3rdparty/lzma/C/Sha256Opt.c new file mode 100644 index 00000000000..e4465e3e7d7 --- /dev/null +++ b/3rdparty/lzma/C/Sha256Opt.c @@ -0,0 +1,386 @@ +/* Sha256Opt.c -- SHA-256 optimized code for SHA-256 hardware instructions +2023-04-02 : Igor Pavlov : Public domain */ + +#include "Precomp.h" +#include "Compiler.h" +#include "CpuArch.h" + +#if defined(_MSC_VER) +#if (_MSC_VER < 1900) && (_MSC_VER >= 1200) +// #define USE_MY_MM +#endif +#endif + +#ifdef MY_CPU_X86_OR_AMD64 + #if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1600) // fix that check + #define USE_HW_SHA + #elif defined(Z7_LLVM_CLANG_VERSION) && (Z7_LLVM_CLANG_VERSION >= 30800) \ + || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 50100) \ + || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40900) + #define USE_HW_SHA + #if !defined(_INTEL_COMPILER) + // icc defines __GNUC__, but icc doesn't support __attribute__(__target__) + #if !defined(__SHA__) || !defined(__SSSE3__) + #define ATTRIB_SHA __attribute__((__target__("sha,ssse3"))) + #endif + #endif + #elif defined(_MSC_VER) + #ifdef USE_MY_MM + #define USE_VER_MIN 1300 + #else + #define USE_VER_MIN 1900 + #endif + #if (_MSC_VER >= USE_VER_MIN) + #define USE_HW_SHA + #endif + #endif +// #endif // MY_CPU_X86_OR_AMD64 + +#ifdef USE_HW_SHA + +// #pragma message("Sha256 HW") + +// sse/sse2/ssse3: +#include <tmmintrin.h> +// sha*: +#include <immintrin.h> + +#if defined (__clang__) && defined(_MSC_VER) + // #if !defined(__SSSE3__) + // #endif + #if !defined(__SHA__) + #include <shaintrin.h> + #endif +#else + +#ifdef USE_MY_MM +#include "My_mm.h" +#endif + +#endif + +/* +SHA256 uses: +SSE2: + _mm_loadu_si128 + _mm_storeu_si128 + _mm_set_epi32 + _mm_add_epi32 + _mm_shuffle_epi32 / pshufd + + + +SSSE3: + _mm_shuffle_epi8 / pshufb + _mm_alignr_epi8 +SHA: + _mm_sha256* +*/ + +// K array must be aligned for 16-bytes at least. +// The compiler can look align attribute and selects +// movdqu - for code without align attribute +// movdqa - for code with align attribute +extern +MY_ALIGN(64) +const UInt32 SHA256_K_ARRAY[64]; + +#define K SHA256_K_ARRAY + + +#define ADD_EPI32(dest, src) dest = _mm_add_epi32(dest, src); +#define SHA256_MSG1(dest, src) dest = _mm_sha256msg1_epu32(dest, src); +#define SHA25G_MSG2(dest, src) dest = _mm_sha256msg2_epu32(dest, src); + + +#define LOAD_SHUFFLE(m, k) \ + m = _mm_loadu_si128((const __m128i *)(const void *)(data + (k) * 16)); \ + m = _mm_shuffle_epi8(m, mask); \ + +#define SM1(g0, g1, g2, g3) \ + SHA256_MSG1(g3, g0); \ + +#define SM2(g0, g1, g2, g3) \ + tmp = _mm_alignr_epi8(g1, g0, 4); \ + ADD_EPI32(g2, tmp) \ + SHA25G_MSG2(g2, g1); \ + +// #define LS0(k, g0, g1, g2, g3) LOAD_SHUFFLE(g0, k) +// #define LS1(k, g0, g1, g2, g3) LOAD_SHUFFLE(g1, k+1) + + +#define NNN(g0, g1, g2, g3) + + +#define RND2(t0, t1) \ + t0 = _mm_sha256rnds2_epu32(t0, t1, msg); + +#define RND2_0(m, k) \ + msg = _mm_add_epi32(m, *(const __m128i *) (const void *) &K[(k) * 4]); \ + RND2(state0, state1); \ + msg = _mm_shuffle_epi32(msg, 0x0E); \ + + +#define RND2_1 \ + RND2(state1, state0); \ + + +// We use scheme with 3 rounds ahead for SHA256_MSG1 / 2 rounds ahead for SHA256_MSG2 + +#define R4(k, g0, g1, g2, g3, OP0, OP1) \ + RND2_0(g0, k) \ + OP0(g0, g1, g2, g3) \ + RND2_1 \ + OP1(g0, g1, g2, g3) \ + +#define R16(k, OP0, OP1, OP2, OP3, OP4, OP5, OP6, OP7) \ + R4 ( (k)*4+0, m0,m1,m2,m3, OP0, OP1 ) \ + R4 ( (k)*4+1, m1,m2,m3,m0, OP2, OP3 ) \ + R4 ( (k)*4+2, m2,m3,m0,m1, OP4, OP5 ) \ + R4 ( (k)*4+3, m3,m0,m1,m2, OP6, OP7 ) \ + +#define PREPARE_STATE \ + tmp = _mm_shuffle_epi32(state0, 0x1B); /* abcd */ \ + state0 = _mm_shuffle_epi32(state1, 0x1B); /* efgh */ \ + state1 = state0; \ + state0 = _mm_unpacklo_epi64(state0, tmp); /* cdgh */ \ + state1 = _mm_unpackhi_epi64(state1, tmp); /* abef */ \ + + +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks); +#ifdef ATTRIB_SHA +ATTRIB_SHA +#endif +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks) +{ + const __m128i mask = _mm_set_epi32(0x0c0d0e0f, 0x08090a0b, 0x04050607, 0x00010203); + __m128i tmp; + __m128i state0, state1; + + if (numBlocks == 0) + return; + + state0 = _mm_loadu_si128((const __m128i *) (const void *) &state[0]); + state1 = _mm_loadu_si128((const __m128i *) (const void *) &state[4]); + + PREPARE_STATE + + do + { + __m128i state0_save, state1_save; + __m128i m0, m1, m2, m3; + __m128i msg; + // #define msg tmp + + state0_save = state0; + state1_save = state1; + + LOAD_SHUFFLE (m0, 0) + LOAD_SHUFFLE (m1, 1) + LOAD_SHUFFLE (m2, 2) + LOAD_SHUFFLE (m3, 3) + + + + R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 ) + R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 ) + R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 ) + R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN ) + + ADD_EPI32(state0, state0_save) + ADD_EPI32(state1, state1_save) + + data += 64; + } + while (--numBlocks); + + PREPARE_STATE + + _mm_storeu_si128((__m128i *) (void *) &state[0], state0); + _mm_storeu_si128((__m128i *) (void *) &state[4], state1); +} + +#endif // USE_HW_SHA + +#elif defined(MY_CPU_ARM_OR_ARM64) + + #if defined(__clang__) + #if (__clang_major__ >= 8) // fix that check + #define USE_HW_SHA + #endif + #elif defined(__GNUC__) + #if (__GNUC__ >= 6) // fix that check + #define USE_HW_SHA + #endif + #elif defined(_MSC_VER) + #if _MSC_VER >= 1910 + #define USE_HW_SHA + #endif + #endif + +#ifdef USE_HW_SHA + +// #pragma message("=== Sha256 HW === ") + +#if defined(__clang__) || defined(__GNUC__) + #ifdef MY_CPU_ARM64 + #define ATTRIB_SHA __attribute__((__target__("+crypto"))) + #else + #define ATTRIB_SHA __attribute__((__target__("fpu=crypto-neon-fp-armv8"))) + #endif +#else + // _MSC_VER + // for arm32 + #define _ARM_USE_NEW_NEON_INTRINSICS +#endif + +#if defined(_MSC_VER) && defined(MY_CPU_ARM64) +#include <arm64_neon.h> +#else +#include <arm_neon.h> +#endif + +typedef uint32x4_t v128; +// typedef __n128 v128; // MSVC + +#ifdef MY_CPU_BE + #define MY_rev32_for_LE(x) +#else + #define MY_rev32_for_LE(x) x = vreinterpretq_u32_u8(vrev32q_u8(vreinterpretq_u8_u32(x))) +#endif + +#define LOAD_128(_p) (*(const v128 *)(const void *)(_p)) +#define STORE_128(_p, _v) *(v128 *)(void *)(_p) = (_v) + +#define LOAD_SHUFFLE(m, k) \ + m = LOAD_128((data + (k) * 16)); \ + MY_rev32_for_LE(m); \ + +// K array must be aligned for 16-bytes at least. +extern +MY_ALIGN(64) +const UInt32 SHA256_K_ARRAY[64]; + +#define K SHA256_K_ARRAY + + +#define SHA256_SU0(dest, src) dest = vsha256su0q_u32(dest, src); +#define SHA25G_SU1(dest, src2, src3) dest = vsha256su1q_u32(dest, src2, src3); + +#define SM1(g0, g1, g2, g3) SHA256_SU0(g3, g0) +#define SM2(g0, g1, g2, g3) SHA25G_SU1(g2, g0, g1) +#define NNN(g0, g1, g2, g3) + + +#define R4(k, g0, g1, g2, g3, OP0, OP1) \ + msg = vaddq_u32(g0, *(const v128 *) (const void *) &K[(k) * 4]); \ + tmp = state0; \ + state0 = vsha256hq_u32( state0, state1, msg ); \ + state1 = vsha256h2q_u32( state1, tmp, msg ); \ + OP0(g0, g1, g2, g3); \ + OP1(g0, g1, g2, g3); \ + + +#define R16(k, OP0, OP1, OP2, OP3, OP4, OP5, OP6, OP7) \ + R4 ( (k)*4+0, m0, m1, m2, m3, OP0, OP1 ) \ + R4 ( (k)*4+1, m1, m2, m3, m0, OP2, OP3 ) \ + R4 ( (k)*4+2, m2, m3, m0, m1, OP4, OP5 ) \ + R4 ( (k)*4+3, m3, m0, m1, m2, OP6, OP7 ) \ + + +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks); +#ifdef ATTRIB_SHA +ATTRIB_SHA +#endif +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks) +{ + v128 state0, state1; + + if (numBlocks == 0) + return; + + state0 = LOAD_128(&state[0]); + state1 = LOAD_128(&state[4]); + + do + { + v128 state0_save, state1_save; + v128 m0, m1, m2, m3; + v128 msg, tmp; + + state0_save = state0; + state1_save = state1; + + LOAD_SHUFFLE (m0, 0) + LOAD_SHUFFLE (m1, 1) + LOAD_SHUFFLE (m2, 2) + LOAD_SHUFFLE (m3, 3) + + R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 ); + R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 ); + R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 ); + R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN ); + + state0 = vaddq_u32(state0, state0_save); + state1 = vaddq_u32(state1, state1_save); + + data += 64; + } + while (--numBlocks); + + STORE_128(&state[0], state0); + STORE_128(&state[4], state1); +} + +#endif // USE_HW_SHA + +#endif // MY_CPU_ARM_OR_ARM64 + + +#ifndef USE_HW_SHA + +// #error Stop_Compiling_UNSUPPORTED_SHA +// #include <stdlib.h> + +// #include "Sha256.h" +void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks); + +#pragma message("Sha256 HW-SW stub was used") + +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks); +void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks) +{ + Sha256_UpdateBlocks(state, data, numBlocks); + /* + UNUSED_VAR(state); + UNUSED_VAR(data); + UNUSED_VAR(numBlocks); + exit(1); + return; + */ +} + +#endif + + + +#undef K +#undef RND2 +#undef RND2_0 +#undef RND2_1 + +#undef MY_rev32_for_LE +#undef NNN +#undef LOAD_128 +#undef STORE_128 +#undef LOAD_SHUFFLE +#undef SM1 +#undef SM2 + +#undef NNN +#undef R4 +#undef R16 +#undef PREPARE_STATE +#undef USE_HW_SHA +#undef ATTRIB_SHA +#undef USE_VER_MIN diff --git a/3rdparty/lzma/C/Sort.h b/3rdparty/lzma/C/Sort.h index 2e2963a23da..1817b652f55 100644 --- a/3rdparty/lzma/C/Sort.h +++ b/3rdparty/lzma/C/Sort.h @@ -1,8 +1,8 @@ /* Sort.h -- Sort functions -2014-04-05 : Igor Pavlov : Public domain */ +2023-03-05 : Igor Pavlov : Public domain */ -#ifndef __7Z_SORT_H -#define __7Z_SORT_H +#ifndef ZIP7_INC_SORT_H +#define ZIP7_INC_SORT_H #include "7zTypes.h" diff --git a/3rdparty/lzma/C/SwapBytes.c b/3rdparty/lzma/C/SwapBytes.c new file mode 100644 index 00000000000..7901bbaa87e --- /dev/null +++ b/3rdparty/lzma/C/SwapBytes.c @@ -0,0 +1,800 @@ +/* SwapBytes.c -- Byte Swap conversion filter +2023-04-07 : Igor Pavlov : Public domain */ + +#include "Precomp.h" + +#include "Compiler.h" +#include "CpuArch.h" +#include "RotateDefs.h" +#include "SwapBytes.h" + +typedef UInt16 CSwapUInt16; +typedef UInt32 CSwapUInt32; + +// #define k_SwapBytes_Mode_BASE 0 + +#ifdef MY_CPU_X86_OR_AMD64 + +#define k_SwapBytes_Mode_SSE2 1 +#define k_SwapBytes_Mode_SSSE3 2 +#define k_SwapBytes_Mode_AVX2 3 + + // #if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1900) + #if defined(__clang__) && (__clang_major__ >= 4) \ + || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40701) + #define k_SwapBytes_Mode_MAX k_SwapBytes_Mode_AVX2 + #define SWAP_ATTRIB_SSE2 __attribute__((__target__("sse2"))) + #define SWAP_ATTRIB_SSSE3 __attribute__((__target__("ssse3"))) + #define SWAP_ATTRIB_AVX2 __attribute__((__target__("avx2"))) + #elif defined(_MSC_VER) + #if (_MSC_VER == 1900) + #pragma warning(disable : 4752) // found Intel(R) Advanced Vector Extensions; consider using /arch:AVX + #endif + #if (_MSC_VER >= 1900) + #define k_SwapBytes_Mode_MAX k_SwapBytes_Mode_AVX2 + #elif (_MSC_VER >= 1500) // (VS2008) + #define k_SwapBytes_Mode_MAX k_SwapBytes_Mode_SSSE3 + #elif (_MSC_VER >= 1310) // (VS2003) + #define k_SwapBytes_Mode_MAX k_SwapBytes_Mode_SSE2 + #endif + #endif // _MSC_VER + +/* +// for debug +#ifdef k_SwapBytes_Mode_MAX +#undef k_SwapBytes_Mode_MAX +#endif +*/ + +#ifndef k_SwapBytes_Mode_MAX +#define k_SwapBytes_Mode_MAX 0 +#endif + +#if (k_SwapBytes_Mode_MAX != 0) && defined(MY_CPU_AMD64) + #define k_SwapBytes_Mode_MIN k_SwapBytes_Mode_SSE2 +#else + #define k_SwapBytes_Mode_MIN 0 +#endif + +#if (k_SwapBytes_Mode_MAX >= k_SwapBytes_Mode_AVX2) + #define USE_SWAP_AVX2 +#endif +#if (k_SwapBytes_Mode_MAX >= k_SwapBytes_Mode_SSSE3) + #define USE_SWAP_SSSE3 +#endif +#if (k_SwapBytes_Mode_MAX >= k_SwapBytes_Mode_SSE2) + #define USE_SWAP_128 +#endif + +#if k_SwapBytes_Mode_MAX <= k_SwapBytes_Mode_MIN || !defined(USE_SWAP_128) +#define FORCE_SWAP_MODE +#endif + + +#ifdef USE_SWAP_128 +/* + <mmintrin.h> MMX +<xmmintrin.h> SSE +<emmintrin.h> SSE2 +<pmmintrin.h> SSE3 +<tmmintrin.h> SSSE3 +<smmintrin.h> SSE4.1 +<nmmintrin.h> SSE4.2 +<ammintrin.h> SSE4A +<wmmintrin.h> AES +<immintrin.h> AVX, AVX2, FMA +*/ + +#include <emmintrin.h> // sse2 +// typedef __m128i v128; + +#define SWAP2_128(i) { \ + const __m128i v = *(const __m128i *)(const void *)(items + (i) * 8); \ + *( __m128i *)( void *)(items + (i) * 8) = \ + _mm_or_si128( \ + _mm_slli_epi16(v, 8), \ + _mm_srli_epi16(v, 8)); } +// _mm_or_si128() has more ports to execute than _mm_add_epi16(). + +static +#ifdef SWAP_ATTRIB_SSE2 +SWAP_ATTRIB_SSE2 +#endif +void +Z7_FASTCALL +SwapBytes2_128(CSwapUInt16 *items, const CSwapUInt16 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP2_128(0) SWAP2_128(1) items += 2 * 8; + SWAP2_128(0) SWAP2_128(1) items += 2 * 8; + } + while (items != lim); +} + +/* +// sse2 +#define SWAP4_128_pack(i) { \ + __m128i v = *(const __m128i *)(const void *)(items + (i) * 4); \ + __m128i v0 = _mm_unpacklo_epi8(v, mask); \ + __m128i v1 = _mm_unpackhi_epi8(v, mask); \ + v0 = _mm_shufflelo_epi16(v0, 0x1b); \ + v1 = _mm_shufflelo_epi16(v1, 0x1b); \ + v0 = _mm_shufflehi_epi16(v0, 0x1b); \ + v1 = _mm_shufflehi_epi16(v1, 0x1b); \ + *(__m128i *)(void *)(items + (i) * 4) = _mm_packus_epi16(v0, v1); } + +static +#ifdef SWAP_ATTRIB_SSE2 +SWAP_ATTRIB_SSE2 +#endif +void +Z7_FASTCALL +SwapBytes4_128_pack(CSwapUInt32 *items, const CSwapUInt32 *lim) +{ + const __m128i mask = _mm_setzero_si128(); + // const __m128i mask = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP4_128_pack(0); items += 1 * 4; + // SWAP4_128_pack(0); SWAP4_128_pack(1); items += 2 * 4; + } + while (items != lim); +} + +// sse2 +#define SWAP4_128_shift(i) { \ + __m128i v = *(const __m128i *)(const void *)(items + (i) * 4); \ + __m128i v2; \ + v2 = _mm_or_si128( \ + _mm_slli_si128(_mm_and_si128(v, mask), 1), \ + _mm_and_si128(_mm_srli_si128(v, 1), mask)); \ + v = _mm_or_si128( \ + _mm_slli_epi32(v, 24), \ + _mm_srli_epi32(v, 24)); \ + *(__m128i *)(void *)(items + (i) * 4) = _mm_or_si128(v2, v); } + +static +#ifdef SWAP_ATTRIB_SSE2 +SWAP_ATTRIB_SSE2 +#endif +void +Z7_FASTCALL +SwapBytes4_128_shift(CSwapUInt32 *items, const CSwapUInt32 *lim) +{ + #define M1 0xff00 + const __m128i mask = _mm_set_epi32(M1, M1, M1, M1); + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + // SWAP4_128_shift(0) SWAP4_128_shift(1) items += 2 * 4; + // SWAP4_128_shift(0) SWAP4_128_shift(1) items += 2 * 4; + SWAP4_128_shift(0); items += 1 * 4; + } + while (items != lim); +} +*/ + + +#if defined(USE_SWAP_SSSE3) || defined(USE_SWAP_AVX2) + +#define SWAP_SHUF_REV_SEQ_2_VALS(v) (v)+1, (v) +#define SWAP_SHUF_REV_SEQ_4_VALS(v) (v)+3, (v)+2, (v)+1, (v) + +#define SWAP2_SHUF_MASK_16_BYTES \ + SWAP_SHUF_REV_SEQ_2_VALS (0 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (1 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (2 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (3 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (4 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (5 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (6 * 2), \ + SWAP_SHUF_REV_SEQ_2_VALS (7 * 2) + +#define SWAP4_SHUF_MASK_16_BYTES \ + SWAP_SHUF_REV_SEQ_4_VALS (0 * 4), \ + SWAP_SHUF_REV_SEQ_4_VALS (1 * 4), \ + SWAP_SHUF_REV_SEQ_4_VALS (2 * 4), \ + SWAP_SHUF_REV_SEQ_4_VALS (3 * 4) + +#if defined(USE_SWAP_AVX2) +/* if we use 256_BIT_INIT_MASK, each static array mask will be larger for 16 bytes */ +// #define SWAP_USE_256_BIT_INIT_MASK +#endif + +#if defined(SWAP_USE_256_BIT_INIT_MASK) && defined(USE_SWAP_AVX2) +#define SWAP_MASK_INIT_SIZE 32 +#else +#define SWAP_MASK_INIT_SIZE 16 +#endif + +MY_ALIGN(SWAP_MASK_INIT_SIZE) +static const Byte k_ShufMask_Swap2[] = +{ + SWAP2_SHUF_MASK_16_BYTES + #if SWAP_MASK_INIT_SIZE > 16 + , SWAP2_SHUF_MASK_16_BYTES + #endif +}; + +MY_ALIGN(SWAP_MASK_INIT_SIZE) +static const Byte k_ShufMask_Swap4[] = +{ + SWAP4_SHUF_MASK_16_BYTES + #if SWAP_MASK_INIT_SIZE > 16 + , SWAP4_SHUF_MASK_16_BYTES + #endif +}; + + +#ifdef USE_SWAP_SSSE3 + +#include <tmmintrin.h> // ssse3 + +#define SHUF_128(i) *(items + (i)) = \ + _mm_shuffle_epi8(*(items + (i)), mask); // SSSE3 + +// Z7_NO_INLINE +static +#ifdef SWAP_ATTRIB_SSSE3 +SWAP_ATTRIB_SSSE3 +#endif +Z7_ATTRIB_NO_VECTORIZE +void +Z7_FASTCALL +ShufBytes_128(void *items8, const void *lim8, const void *mask128_ptr) +{ + __m128i *items = (__m128i *)items8; + const __m128i *lim = (const __m128i *)lim8; + // const __m128i mask = _mm_set_epi8(SHUF_SWAP2_MASK_16_VALS); + // const __m128i mask = _mm_set_epi8(SHUF_SWAP4_MASK_16_VALS); + // const __m128i mask = _mm_load_si128((const __m128i *)(const void *)&(k_ShufMask_Swap4[0])); + // const __m128i mask = _mm_load_si128((const __m128i *)(const void *)&(k_ShufMask_Swap4[0])); + // const __m128i mask = *(const __m128i *)(const void *)&(k_ShufMask_Swap4[0]); + const __m128i mask = *(const __m128i *)mask128_ptr; + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SHUF_128(0) SHUF_128(1) items += 2; + SHUF_128(0) SHUF_128(1) items += 2; + } + while (items != lim); +} + +#endif // USE_SWAP_SSSE3 + + + +#ifdef USE_SWAP_AVX2 + +#include <immintrin.h> // avx, avx2 +#if defined(__clang__) +#include <avxintrin.h> +#include <avx2intrin.h> +#endif + +#define SHUF_256(i) *(items + (i)) = \ + _mm256_shuffle_epi8(*(items + (i)), mask); // AVX2 + +// Z7_NO_INLINE +static +#ifdef SWAP_ATTRIB_AVX2 +SWAP_ATTRIB_AVX2 +#endif +Z7_ATTRIB_NO_VECTORIZE +void +Z7_FASTCALL +ShufBytes_256(void *items8, const void *lim8, const void *mask128_ptr) +{ + __m256i *items = (__m256i *)items8; + const __m256i *lim = (const __m256i *)lim8; + /* + UNUSED_VAR(mask128_ptr) + __m256i mask = + for Swap4: _mm256_setr_epi8(SWAP4_SHUF_MASK_16_BYTES, SWAP4_SHUF_MASK_16_BYTES); + for Swap2: _mm256_setr_epi8(SWAP2_SHUF_MASK_16_BYTES, SWAP2_SHUF_MASK_16_BYTES); + */ + const __m256i mask = + #if SWAP_MASK_INIT_SIZE > 16 + *(const __m256i *)(const void *)mask128_ptr; + #else + /* msvc: broadcastsi128() version reserves the stack for no reason + msvc 19.29-: _mm256_insertf128_si256() / _mm256_set_m128i)) versions use non-avx movdqu xmm0,XMMWORD PTR [r8] + msvc 19.30+ (VS2022): replaces _mm256_set_m128i(m,m) to vbroadcastf128(m) as we want + */ + // _mm256_broadcastsi128_si256(*mask128_ptr); + /* + #define MY_mm256_set_m128i(hi, lo) _mm256_insertf128_si256(_mm256_castsi128_si256(lo), (hi), 1) + MY_mm256_set_m128i + */ + _mm256_set_m128i( + *(const __m128i *)mask128_ptr, + *(const __m128i *)mask128_ptr); + #endif + + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SHUF_256(0) SHUF_256(1) items += 2; + SHUF_256(0) SHUF_256(1) items += 2; + } + while (items != lim); +} + +#endif // USE_SWAP_AVX2 +#endif // USE_SWAP_SSSE3 || USE_SWAP_AVX2 +#endif // USE_SWAP_128 + + + +// compile message "NEON intrinsics not available with the soft-float ABI" +#elif defined(MY_CPU_ARM_OR_ARM64) || \ + (defined(__ARM_ARCH) && (__ARM_ARCH >= 7)) +// #elif defined(MY_CPU_ARM64) + + #if defined(__clang__) && (__clang_major__ >= 8) \ + || defined(__GNUC__) && (__GNUC__ >= 8) + #if (defined(__ARM_ARCH) && (__ARM_ARCH >= 7)) \ + || defined(MY_CPU_ARM64) + #define USE_SWAP_128 + #endif + #ifdef MY_CPU_ARM64 + // #define SWAP_ATTRIB_NEON __attribute__((__target__(""))) + #else + // #define SWAP_ATTRIB_NEON __attribute__((__target__("fpu=crypto-neon-fp-armv8"))) + #endif + #elif defined(_MSC_VER) + #if (_MSC_VER >= 1910) + #define USE_SWAP_128 + #endif + #endif + + #if defined(_MSC_VER) && defined(MY_CPU_ARM64) + #include <arm64_neon.h> + #else + #include <arm_neon.h> + #endif + +#ifndef USE_SWAP_128 + #define FORCE_SWAP_MODE +#else + +#ifdef MY_CPU_ARM64 + // for debug : comment it + #define FORCE_SWAP_MODE +#else + #define k_SwapBytes_Mode_NEON 1 +#endif +// typedef uint8x16_t v128; +#define SWAP2_128(i) *(uint8x16_t *) (void *)(items + (i) * 8) = \ + vrev16q_u8(*(const uint8x16_t *)(const void *)(items + (i) * 8)); +#define SWAP4_128(i) *(uint8x16_t *) (void *)(items + (i) * 4) = \ + vrev32q_u8(*(const uint8x16_t *)(const void *)(items + (i) * 4)); + +// Z7_NO_INLINE +static +#ifdef SWAP_ATTRIB_NEON +SWAP_ATTRIB_NEON +#endif +Z7_ATTRIB_NO_VECTORIZE +void +Z7_FASTCALL +SwapBytes2_128(CSwapUInt16 *items, const CSwapUInt16 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP2_128(0) SWAP2_128(1) items += 2 * 8; + SWAP2_128(0) SWAP2_128(1) items += 2 * 8; + } + while (items != lim); +} + +// Z7_NO_INLINE +static +#ifdef SWAP_ATTRIB_NEON +SWAP_ATTRIB_NEON +#endif +Z7_ATTRIB_NO_VECTORIZE +void +Z7_FASTCALL +SwapBytes4_128(CSwapUInt32 *items, const CSwapUInt32 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP4_128(0) SWAP4_128(1) items += 2 * 4; + SWAP4_128(0) SWAP4_128(1) items += 2 * 4; + } + while (items != lim); +} + +#endif // USE_SWAP_128 + +#else // MY_CPU_ARM_OR_ARM64 +#define FORCE_SWAP_MODE +#endif // MY_CPU_ARM_OR_ARM64 + + + + + + +#if defined(Z7_MSC_VER_ORIGINAL) && defined(MY_CPU_X86) + /* _byteswap_ushort() in MSVC x86 32-bit works via slow { mov dh, al; mov dl, ah } + So we use own versions of byteswap function */ + #if (_MSC_VER < 1400 ) // old MSVC-X86 without _rotr16() support + #define SWAP2_16(i) { UInt32 v = items[i]; v += (v << 16); v >>= 8; items[i] = (CSwapUInt16)v; } + #else // is new MSVC-X86 with fast _rotr16() + #include <intrin.h> + #define SWAP2_16(i) { items[i] = _rotr16(items[i], 8); } + #endif +#else // is not MSVC-X86 + #define SWAP2_16(i) { CSwapUInt16 v = items[i]; items[i] = Z7_BSWAP16(v); } +#endif // MSVC-X86 + +#if defined(Z7_CPU_FAST_BSWAP_SUPPORTED) + #define SWAP4_32(i) { CSwapUInt32 v = items[i]; items[i] = Z7_BSWAP32(v); } +#else + #define SWAP4_32(i) \ + { UInt32 v = items[i]; \ + v = ((v & 0xff00ff) << 8) + ((v >> 8) & 0xff00ff); \ + v = rotlFixed(v, 16); \ + items[i] = v; } +#endif + + + + +#if defined(FORCE_SWAP_MODE) && defined(USE_SWAP_128) + #define DEFAULT_Swap2 SwapBytes2_128 + #if !defined(MY_CPU_X86_OR_AMD64) + #define DEFAULT_Swap4 SwapBytes4_128 + #endif +#endif + +#if !defined(DEFAULT_Swap2) || !defined(DEFAULT_Swap4) + +#define SWAP_BASE_FUNCS_PREFIXES \ +Z7_FORCE_INLINE \ +static \ +Z7_ATTRIB_NO_VECTOR \ +void Z7_FASTCALL + + +#ifdef MY_CPU_64BIT + +#if defined(MY_CPU_ARM64) \ + && defined(__ARM_ARCH) && (__ARM_ARCH >= 8) \ + && ( (defined(__GNUC__) && (__GNUC__ >= 4)) \ + || (defined(__clang__) && (__clang_major__ >= 4))) + + #define SWAP2_64_VAR(v) asm ("rev16 %x0,%x0" : "+r" (v)); + #define SWAP4_64_VAR(v) asm ("rev32 %x0,%x0" : "+r" (v)); + +#else // is not ARM64-GNU + +#if !defined(MY_CPU_X86_OR_AMD64) || (k_SwapBytes_Mode_MIN == 0) || !defined(USE_SWAP_128) + #define SWAP2_64_VAR(v) \ + v = ( 0x00ff00ff00ff00ff & (v >> 8)) \ + + ((0x00ff00ff00ff00ff & v) << 8); + /* plus gives faster code in MSVC */ +#endif + +#ifdef Z7_CPU_FAST_BSWAP_SUPPORTED + #define SWAP4_64_VAR(v) \ + v = Z7_BSWAP64(v); \ + v = Z7_ROTL64(v, 32); +#else + #define SWAP4_64_VAR(v) \ + v = ( 0x000000ff000000ff & (v >> 24)) \ + + ((0x000000ff000000ff & v) << 24 ) \ + + ( 0x0000ff000000ff00 & (v >> 8)) \ + + ((0x0000ff000000ff00 & v) << 8 ) \ + ; +#endif + +#endif // ARM64-GNU + + +#ifdef SWAP2_64_VAR + +#define SWAP2_64(i) { \ + UInt64 v = *(const UInt64 *)(const void *)(items + (i) * 4); \ + SWAP2_64_VAR(v) \ + *(UInt64 *)(void *)(items + (i) * 4) = v; } + +SWAP_BASE_FUNCS_PREFIXES +SwapBytes2_64(CSwapUInt16 *items, const CSwapUInt16 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP2_64(0) SWAP2_64(1) items += 2 * 4; + SWAP2_64(0) SWAP2_64(1) items += 2 * 4; + } + while (items != lim); +} + + #define DEFAULT_Swap2 SwapBytes2_64 + #if !defined(FORCE_SWAP_MODE) + #define SWAP2_DEFAULT_MODE 0 + #endif +#else // !defined(SWAP2_64_VAR) + #define DEFAULT_Swap2 SwapBytes2_128 + #if !defined(FORCE_SWAP_MODE) + #define SWAP2_DEFAULT_MODE 1 + #endif +#endif // SWAP2_64_VAR + + +#define SWAP4_64(i) { \ + UInt64 v = *(const UInt64 *)(const void *)(items + (i) * 2); \ + SWAP4_64_VAR(v) \ + *(UInt64 *)(void *)(items + (i) * 2) = v; } + +SWAP_BASE_FUNCS_PREFIXES +SwapBytes4_64(CSwapUInt32 *items, const CSwapUInt32 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP4_64(0) SWAP4_64(1) items += 2 * 2; + SWAP4_64(0) SWAP4_64(1) items += 2 * 2; + } + while (items != lim); +} + +#define DEFAULT_Swap4 SwapBytes4_64 + +#else // is not 64BIT + + +#if defined(MY_CPU_ARM_OR_ARM64) \ + && defined(__ARM_ARCH) && (__ARM_ARCH >= 6) \ + && ( (defined(__GNUC__) && (__GNUC__ >= 4)) \ + || (defined(__clang__) && (__clang_major__ >= 4))) + +#ifdef MY_CPU_64BIT + #define SWAP2_32_VAR(v) asm ("rev16 %w0,%w0" : "+r" (v)); +#else + #define SWAP2_32_VAR(v) asm ("rev16 %0,%0" : "+r" (v)); // for clang/gcc + // asm ("rev16 %r0,%r0" : "+r" (a)); // for gcc +#endif + +#elif defined(_MSC_VER) && (_MSC_VER < 1300) && defined(MY_CPU_X86) \ + || !defined(Z7_CPU_FAST_BSWAP_SUPPORTED) \ + || !defined(Z7_CPU_FAST_ROTATE_SUPPORTED) + // old msvc doesn't support _byteswap_ulong() + #define SWAP2_32_VAR(v) \ + v = ((v & 0xff00ff) << 8) + ((v >> 8) & 0xff00ff); + +#else // is not ARM and is not old-MSVC-X86 and fast BSWAP/ROTATE are supported + #define SWAP2_32_VAR(v) \ + v = Z7_BSWAP32(v); \ + v = rotlFixed(v, 16); + +#endif // GNU-ARM* + +#define SWAP2_32(i) { \ + UInt32 v = *(const UInt32 *)(const void *)(items + (i) * 2); \ + SWAP2_32_VAR(v); \ + *(UInt32 *)(void *)(items + (i) * 2) = v; } + + +SWAP_BASE_FUNCS_PREFIXES +SwapBytes2_32(CSwapUInt16 *items, const CSwapUInt16 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP2_32(0) SWAP2_32(1) items += 2 * 2; + SWAP2_32(0) SWAP2_32(1) items += 2 * 2; + } + while (items != lim); +} + + +SWAP_BASE_FUNCS_PREFIXES +SwapBytes4_32(CSwapUInt32 *items, const CSwapUInt32 *lim) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + do + { + SWAP4_32(0) SWAP4_32(1) items += 2; + SWAP4_32(0) SWAP4_32(1) items += 2; + } + while (items != lim); +} + +#define DEFAULT_Swap2 SwapBytes2_32 +#define DEFAULT_Swap4 SwapBytes4_32 +#if !defined(FORCE_SWAP_MODE) + #define SWAP2_DEFAULT_MODE 0 +#endif + +#endif // MY_CPU_64BIT +#endif // if !defined(DEFAULT_Swap2) || !defined(DEFAULT_Swap4) + + + +#if !defined(FORCE_SWAP_MODE) +static unsigned g_SwapBytes_Mode; +#endif + +/* size of largest unrolled loop iteration: 128 bytes = 4 * 32 bytes (AVX). */ +#define SWAP_ITERATION_BLOCK_SIZE_MAX (1 << 7) + +// 32 bytes for (AVX) or 2 * 16-bytes for NEON. +#define SWAP_VECTOR_ALIGN_SIZE (1 << 5) + +Z7_NO_INLINE +void z7_SwapBytes2(CSwapUInt16 *items, size_t numItems) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0 && ((unsigned)(ptrdiff_t)items & (SWAP_VECTOR_ALIGN_SIZE - 1)) != 0; numItems--) + { + SWAP2_16(0) + items++; + } + { + const size_t k_Align_Mask = SWAP_ITERATION_BLOCK_SIZE_MAX / sizeof(CSwapUInt16) - 1; + size_t numItems2 = numItems; + CSwapUInt16 *lim; + numItems &= k_Align_Mask; + numItems2 &= ~(size_t)k_Align_Mask; + lim = items + numItems2; + if (numItems2 != 0) + { + #if !defined(FORCE_SWAP_MODE) + #ifdef MY_CPU_X86_OR_AMD64 + #ifdef USE_SWAP_AVX2 + if (g_SwapBytes_Mode > k_SwapBytes_Mode_SSSE3) + ShufBytes_256((__m256i *)(void *)items, + (const __m256i *)(const void *)lim, + (const __m128i *)(const void *)&(k_ShufMask_Swap2[0])); + else + #endif + #ifdef USE_SWAP_SSSE3 + if (g_SwapBytes_Mode >= k_SwapBytes_Mode_SSSE3) + ShufBytes_128((__m128i *)(void *)items, + (const __m128i *)(const void *)lim, + (const __m128i *)(const void *)&(k_ShufMask_Swap2[0])); + else + #endif + #endif // MY_CPU_X86_OR_AMD64 + #if SWAP2_DEFAULT_MODE == 0 + if (g_SwapBytes_Mode != 0) + SwapBytes2_128(items, lim); + else + #endif + #endif // FORCE_SWAP_MODE + DEFAULT_Swap2(items, lim); + } + items = lim; + } + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0; numItems--) + { + SWAP2_16(0) + items++; + } +} + + +Z7_NO_INLINE +void z7_SwapBytes4(CSwapUInt32 *items, size_t numItems) +{ + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0 && ((unsigned)(ptrdiff_t)items & (SWAP_VECTOR_ALIGN_SIZE - 1)) != 0; numItems--) + { + SWAP4_32(0) + items++; + } + { + const size_t k_Align_Mask = SWAP_ITERATION_BLOCK_SIZE_MAX / sizeof(CSwapUInt32) - 1; + size_t numItems2 = numItems; + CSwapUInt32 *lim; + numItems &= k_Align_Mask; + numItems2 &= ~(size_t)k_Align_Mask; + lim = items + numItems2; + if (numItems2 != 0) + { + #if !defined(FORCE_SWAP_MODE) + #ifdef MY_CPU_X86_OR_AMD64 + #ifdef USE_SWAP_AVX2 + if (g_SwapBytes_Mode > k_SwapBytes_Mode_SSSE3) + ShufBytes_256((__m256i *)(void *)items, + (const __m256i *)(const void *)lim, + (const __m128i *)(const void *)&(k_ShufMask_Swap4[0])); + else + #endif + #ifdef USE_SWAP_SSSE3 + if (g_SwapBytes_Mode >= k_SwapBytes_Mode_SSSE3) + ShufBytes_128((__m128i *)(void *)items, + (const __m128i *)(const void *)lim, + (const __m128i *)(const void *)&(k_ShufMask_Swap4[0])); + else + #endif + #else // MY_CPU_X86_OR_AMD64 + + if (g_SwapBytes_Mode != 0) + SwapBytes4_128(items, lim); + else + #endif // MY_CPU_X86_OR_AMD64 + #endif // FORCE_SWAP_MODE + DEFAULT_Swap4(items, lim); + } + items = lim; + } + Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE + for (; numItems != 0; numItems--) + { + SWAP4_32(0) + items++; + } +} + + +// #define SHOW_HW_STATUS + +#ifdef SHOW_HW_STATUS +#include <stdio.h> +#define PRF(x) x +#else +#define PRF(x) +#endif + +void z7_SwapBytesPrepare(void) +{ +#ifndef FORCE_SWAP_MODE + unsigned mode = 0; // k_SwapBytes_Mode_BASE; + +#ifdef MY_CPU_ARM_OR_ARM64 + { + if (CPU_IsSupported_NEON()) + { + // #pragma message ("=== SwapBytes NEON") + PRF(printf("\n=== SwapBytes NEON\n");) + mode = k_SwapBytes_Mode_NEON; + } + } +#else // MY_CPU_ARM_OR_ARM64 + { + #ifdef USE_SWAP_AVX2 + if (CPU_IsSupported_AVX2()) + { + // #pragma message ("=== SwapBytes AVX2") + PRF(printf("\n=== SwapBytes AVX2\n");) + mode = k_SwapBytes_Mode_AVX2; + } + else + #endif + #ifdef USE_SWAP_SSSE3 + if (CPU_IsSupported_SSSE3()) + { + // #pragma message ("=== SwapBytes SSSE3") + PRF(printf("\n=== SwapBytes SSSE3\n");) + mode = k_SwapBytes_Mode_SSSE3; + } + else + #endif + #if !defined(MY_CPU_AMD64) + if (CPU_IsSupported_SSE2()) + #endif + { + // #pragma message ("=== SwapBytes SSE2") + PRF(printf("\n=== SwapBytes SSE2\n");) + mode = k_SwapBytes_Mode_SSE2; + } + } +#endif // MY_CPU_ARM_OR_ARM64 + g_SwapBytes_Mode = mode; + // g_SwapBytes_Mode = 0; // for debug +#endif // FORCE_SWAP_MODE + PRF(printf("\n=== SwapBytesPrepare\n");) +} + +#undef PRF diff --git a/3rdparty/lzma/C/SwapBytes.h b/3rdparty/lzma/C/SwapBytes.h new file mode 100644 index 00000000000..d442467386e --- /dev/null +++ b/3rdparty/lzma/C/SwapBytes.h @@ -0,0 +1,17 @@ +/* SwapBytes.h -- Byte Swap conversion filter +2023-04-02 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_SWAP_BYTES_H +#define ZIP7_INC_SWAP_BYTES_H + +#include "7zTypes.h" + +EXTERN_C_BEGIN + +void z7_SwapBytes2(UInt16 *data, size_t numItems); +void z7_SwapBytes4(UInt32 *data, size_t numItems); +void z7_SwapBytesPrepare(void); + +EXTERN_C_END + +#endif diff --git a/3rdparty/lzma/C/Threads.c b/3rdparty/lzma/C/Threads.c index d3d0912da42..cf52bd30307 100644 --- a/3rdparty/lzma/C/Threads.c +++ b/3rdparty/lzma/C/Threads.c @@ -1,55 +1,134 @@ /* Threads.c -- multithreading library -2014-09-21 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ #include "Precomp.h" -#ifndef UNDER_CE +#ifdef _WIN32 + +#ifndef USE_THREADS_CreateThread #include <process.h> #endif #include "Threads.h" -static WRes GetError() +static WRes GetError(void) { - DWORD res = GetLastError(); - return (res) ? (WRes)(res) : 1; + const DWORD res = GetLastError(); + return res ? (WRes)res : 1; } -WRes HandleToWRes(HANDLE h) { return (h != 0) ? 0 : GetError(); } -WRes BOOLToWRes(BOOL v) { return v ? 0 : GetError(); } +static WRes HandleToWRes(HANDLE h) { return (h != NULL) ? 0 : GetError(); } +static WRes BOOLToWRes(BOOL v) { return v ? 0 : GetError(); } WRes HandlePtr_Close(HANDLE *p) { if (*p != NULL) + { if (!CloseHandle(*p)) return GetError(); - *p = NULL; + *p = NULL; + } return 0; } -WRes Handle_WaitObject(HANDLE h) { return (WRes)WaitForSingleObject(h, INFINITE); } +WRes Handle_WaitObject(HANDLE h) +{ + DWORD dw = WaitForSingleObject(h, INFINITE); + /* + (dw) result: + WAIT_OBJECT_0 // 0 + WAIT_ABANDONED // 0x00000080 : is not compatible with Win32 Error space + WAIT_TIMEOUT // 0x00000102 : is compatible with Win32 Error space + WAIT_FAILED // 0xFFFFFFFF + */ + if (dw == WAIT_FAILED) + { + dw = GetLastError(); + if (dw == 0) + return WAIT_FAILED; + } + return (WRes)dw; +} + +#define Thread_Wait(p) Handle_WaitObject(*(p)) + +WRes Thread_Wait_Close(CThread *p) +{ + WRes res = Thread_Wait(p); + WRes res2 = Thread_Close(p); + return (res != 0 ? res : res2); +} WRes Thread_Create(CThread *p, THREAD_FUNC_TYPE func, LPVOID param) { /* Windows Me/98/95: threadId parameter may not be NULL in _beginthreadex/CreateThread functions */ - - #ifdef UNDER_CE - - DWORD threadId; - *p = CreateThread(0, 0, func, param, 0, &threadId); - #else + #ifdef USE_THREADS_CreateThread + DWORD threadId; + *p = CreateThread(NULL, 0, func, param, 0, &threadId); + + #else + unsigned threadId; - *p = (HANDLE)_beginthreadex(NULL, 0, func, param, 0, &threadId); - + *p = (HANDLE)(_beginthreadex(NULL, 0, func, param, 0, &threadId)); + #endif /* maybe we must use errno here, but probably GetLastError() is also OK. */ return HandleToWRes(*p); } -WRes Event_Create(CEvent *p, BOOL manualReset, int signaled) + +WRes Thread_Create_With_Affinity(CThread *p, THREAD_FUNC_TYPE func, LPVOID param, CAffinityMask affinity) +{ + #ifdef USE_THREADS_CreateThread + + UNUSED_VAR(affinity) + return Thread_Create(p, func, param); + + #else + + /* Windows Me/98/95: threadId parameter may not be NULL in _beginthreadex/CreateThread functions */ + HANDLE h; + WRes wres; + unsigned threadId; + h = (HANDLE)(_beginthreadex(NULL, 0, func, param, CREATE_SUSPENDED, &threadId)); + *p = h; + wres = HandleToWRes(h); + if (h) + { + { + // DWORD_PTR prevMask = + SetThreadAffinityMask(h, (DWORD_PTR)affinity); + /* + if (prevMask == 0) + { + // affinity change is non-critical error, so we can ignore it + // wres = GetError(); + } + */ + } + { + DWORD prevSuspendCount = ResumeThread(h); + /* ResumeThread() returns: + 0 : was_not_suspended + 1 : was_resumed + -1 : error + */ + if (prevSuspendCount == (DWORD)-1) + wres = GetError(); + } + } + + /* maybe we must use errno here, but probably GetLastError() is also OK. */ + return wres; + + #endif +} + + +static WRes Event_Create(CEvent *p, BOOL manualReset, int signaled) { *p = CreateEvent(NULL, manualReset, (signaled ? TRUE : FALSE), NULL); return HandleToWRes(*p); @@ -66,10 +145,22 @@ WRes AutoResetEvent_CreateNotSignaled(CAutoResetEvent *p) { return AutoResetEven WRes Semaphore_Create(CSemaphore *p, UInt32 initCount, UInt32 maxCount) { + // negative ((LONG)maxCount) is not supported in WIN32::CreateSemaphore() *p = CreateSemaphore(NULL, (LONG)initCount, (LONG)maxCount, NULL); return HandleToWRes(*p); } +WRes Semaphore_OptCreateInit(CSemaphore *p, UInt32 initCount, UInt32 maxCount) +{ + // if (Semaphore_IsCreated(p)) + { + WRes wres = Semaphore_Close(p); + if (wres != 0) + return wres; + } + return Semaphore_Create(p, initCount, maxCount); +} + static WRes Semaphore_Release(CSemaphore *p, LONG releaseCount, LONG *previousCount) { return BOOLToWRes(ReleaseSemaphore(*p, releaseCount, previousCount)); } WRes Semaphore_ReleaseN(CSemaphore *p, UInt32 num) @@ -78,8 +169,13 @@ WRes Semaphore_Release1(CSemaphore *p) { return Semaphore_ReleaseN(p, 1); } WRes CriticalSection_Init(CCriticalSection *p) { - /* InitializeCriticalSection can raise only STATUS_NO_MEMORY exception */ + /* InitializeCriticalSection() can raise exception: + Windows XP, 2003 : can raise a STATUS_NO_MEMORY exception + Windows Vista+ : no exceptions */ #ifdef _MSC_VER + #ifdef __clang__ + #pragma GCC diagnostic ignored "-Wlanguage-extension-token" + #endif __try #endif { @@ -87,7 +183,380 @@ WRes CriticalSection_Init(CCriticalSection *p) /* InitializeCriticalSectionAndSpinCount(p, 0); */ } #ifdef _MSC_VER - __except (EXCEPTION_EXECUTE_HANDLER) { return 1; } + __except (EXCEPTION_EXECUTE_HANDLER) { return ERROR_NOT_ENOUGH_MEMORY; } + #endif + return 0; +} + + + + +#else // _WIN32 + +// ---------- POSIX ---------- + +#ifndef __APPLE__ +#ifndef Z7_AFFINITY_DISABLE +// _GNU_SOURCE can be required for pthread_setaffinity_np() / CPU_ZERO / CPU_SET +// clang < 3.6 : unknown warning group '-Wreserved-id-macro' +// clang 3.6 - 12.01 : gives warning "macro name is a reserved identifier" +// clang >= 13 : do not give warning +#if !defined(_GNU_SOURCE) + #if defined(__clang__) && (__clang_major__ >= 4) && (__clang_major__ <= 12) + #pragma GCC diagnostic ignored "-Wreserved-id-macro" #endif +#define _GNU_SOURCE +#endif // !defined(_GNU_SOURCE) +#endif // Z7_AFFINITY_DISABLE +#endif // __APPLE__ + +#include "Threads.h" + +#include <errno.h> +#include <stdlib.h> +#include <string.h> +#ifdef Z7_AFFINITY_SUPPORTED +// #include <sched.h> +#endif + + +// #include <stdio.h> +// #define PRF(p) p +#define PRF(p) +#define Print(s) PRF(printf("\n%s\n", s);) + +WRes Thread_Create_With_CpuSet(CThread *p, THREAD_FUNC_TYPE func, LPVOID param, const CCpuSet *cpuSet) +{ + // new thread in Posix probably inherits affinity from parrent thread + Print("Thread_Create_With_CpuSet") + + pthread_attr_t attr; + int ret; + // int ret2; + + p->_created = 0; + + RINOK(pthread_attr_init(&attr)) + + ret = pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_JOINABLE); + + if (!ret) + { + if (cpuSet) + { + #ifdef Z7_AFFINITY_SUPPORTED + + /* + printf("\n affinity :"); + unsigned i; + for (i = 0; i < sizeof(*cpuSet) && i < 8; i++) + { + Byte b = *((const Byte *)cpuSet + i); + char temp[32]; + #define GET_HEX_CHAR(t) ((char)(((t < 10) ? ('0' + t) : ('A' + (t - 10))))) + temp[0] = GET_HEX_CHAR((b & 0xF)); + temp[1] = GET_HEX_CHAR((b >> 4)); + // temp[0] = GET_HEX_CHAR((b >> 4)); // big-endian + // temp[1] = GET_HEX_CHAR((b & 0xF)); // big-endian + temp[2] = 0; + printf("%s", temp); + } + printf("\n"); + */ + + // ret2 = + pthread_attr_setaffinity_np(&attr, sizeof(*cpuSet), cpuSet); + // if (ret2) ret = ret2; + #endif + } + + ret = pthread_create(&p->_tid, &attr, func, param); + + if (!ret) + { + p->_created = 1; + /* + if (cpuSet) + { + // ret2 = + pthread_setaffinity_np(p->_tid, sizeof(*cpuSet), cpuSet); + // if (ret2) ret = ret2; + } + */ + } + } + // ret2 = + pthread_attr_destroy(&attr); + // if (ret2 != 0) ret = ret2; + return ret; +} + + +WRes Thread_Create(CThread *p, THREAD_FUNC_TYPE func, LPVOID param) +{ + return Thread_Create_With_CpuSet(p, func, param, NULL); +} + + +WRes Thread_Create_With_Affinity(CThread *p, THREAD_FUNC_TYPE func, LPVOID param, CAffinityMask affinity) +{ + Print("Thread_Create_WithAffinity") + CCpuSet cs; + unsigned i; + CpuSet_Zero(&cs); + for (i = 0; i < sizeof(affinity) * 8; i++) + { + if (affinity == 0) + break; + if (affinity & 1) + { + CpuSet_Set(&cs, i); + } + affinity >>= 1; + } + return Thread_Create_With_CpuSet(p, func, param, &cs); +} + + +WRes Thread_Close(CThread *p) +{ + // Print("Thread_Close") + int ret; + if (!p->_created) + return 0; + + ret = pthread_detach(p->_tid); + p->_tid = 0; + p->_created = 0; + return ret; +} + + +WRes Thread_Wait_Close(CThread *p) +{ + // Print("Thread_Wait_Close") + void *thread_return; + int ret; + if (!p->_created) + return EINVAL; + + ret = pthread_join(p->_tid, &thread_return); + // probably we can't use that (_tid) after pthread_join(), so we close thread here + p->_created = 0; + p->_tid = 0; + return ret; +} + + + +static WRes Event_Create(CEvent *p, int manualReset, int signaled) +{ + RINOK(pthread_mutex_init(&p->_mutex, NULL)) + RINOK(pthread_cond_init(&p->_cond, NULL)) + p->_manual_reset = manualReset; + p->_state = (signaled ? True : False); + p->_created = 1; + return 0; +} + +WRes ManualResetEvent_Create(CManualResetEvent *p, int signaled) + { return Event_Create(p, True, signaled); } +WRes ManualResetEvent_CreateNotSignaled(CManualResetEvent *p) + { return ManualResetEvent_Create(p, 0); } +WRes AutoResetEvent_Create(CAutoResetEvent *p, int signaled) + { return Event_Create(p, False, signaled); } +WRes AutoResetEvent_CreateNotSignaled(CAutoResetEvent *p) + { return AutoResetEvent_Create(p, 0); } + + +WRes Event_Set(CEvent *p) +{ + RINOK(pthread_mutex_lock(&p->_mutex)) + p->_state = True; + int res1 = pthread_cond_broadcast(&p->_cond); + int res2 = pthread_mutex_unlock(&p->_mutex); + return (res2 ? res2 : res1); +} + +WRes Event_Reset(CEvent *p) +{ + RINOK(pthread_mutex_lock(&p->_mutex)) + p->_state = False; + return pthread_mutex_unlock(&p->_mutex); +} + +WRes Event_Wait(CEvent *p) +{ + RINOK(pthread_mutex_lock(&p->_mutex)) + while (p->_state == False) + { + // ETIMEDOUT + // ret = + pthread_cond_wait(&p->_cond, &p->_mutex); + // if (ret != 0) break; + } + if (p->_manual_reset == False) + { + p->_state = False; + } + return pthread_mutex_unlock(&p->_mutex); +} + +WRes Event_Close(CEvent *p) +{ + if (!p->_created) + return 0; + p->_created = 0; + { + int res1 = pthread_mutex_destroy(&p->_mutex); + int res2 = pthread_cond_destroy(&p->_cond); + return (res1 ? res1 : res2); + } +} + + +WRes Semaphore_Create(CSemaphore *p, UInt32 initCount, UInt32 maxCount) +{ + if (initCount > maxCount || maxCount < 1) + return EINVAL; + RINOK(pthread_mutex_init(&p->_mutex, NULL)) + RINOK(pthread_cond_init(&p->_cond, NULL)) + p->_count = initCount; + p->_maxCount = maxCount; + p->_created = 1; return 0; } + + +WRes Semaphore_OptCreateInit(CSemaphore *p, UInt32 initCount, UInt32 maxCount) +{ + if (Semaphore_IsCreated(p)) + { + /* + WRes wres = Semaphore_Close(p); + if (wres != 0) + return wres; + */ + if (initCount > maxCount || maxCount < 1) + return EINVAL; + // return EINVAL; // for debug + p->_count = initCount; + p->_maxCount = maxCount; + return 0; + } + return Semaphore_Create(p, initCount, maxCount); +} + + +WRes Semaphore_ReleaseN(CSemaphore *p, UInt32 releaseCount) +{ + UInt32 newCount; + int ret; + + if (releaseCount < 1) + return EINVAL; + + RINOK(pthread_mutex_lock(&p->_mutex)) + + newCount = p->_count + releaseCount; + if (newCount > p->_maxCount) + ret = ERROR_TOO_MANY_POSTS; // EINVAL; + else + { + p->_count = newCount; + ret = pthread_cond_broadcast(&p->_cond); + } + RINOK(pthread_mutex_unlock(&p->_mutex)) + return ret; +} + +WRes Semaphore_Wait(CSemaphore *p) +{ + RINOK(pthread_mutex_lock(&p->_mutex)) + while (p->_count < 1) + { + pthread_cond_wait(&p->_cond, &p->_mutex); + } + p->_count--; + return pthread_mutex_unlock(&p->_mutex); +} + +WRes Semaphore_Close(CSemaphore *p) +{ + if (!p->_created) + return 0; + p->_created = 0; + { + int res1 = pthread_mutex_destroy(&p->_mutex); + int res2 = pthread_cond_destroy(&p->_cond); + return (res1 ? res1 : res2); + } +} + + + +WRes CriticalSection_Init(CCriticalSection *p) +{ + // Print("CriticalSection_Init") + if (!p) + return EINTR; + return pthread_mutex_init(&p->_mutex, NULL); +} + +void CriticalSection_Enter(CCriticalSection *p) +{ + // Print("CriticalSection_Enter") + if (p) + { + // int ret = + pthread_mutex_lock(&p->_mutex); + } +} + +void CriticalSection_Leave(CCriticalSection *p) +{ + // Print("CriticalSection_Leave") + if (p) + { + // int ret = + pthread_mutex_unlock(&p->_mutex); + } +} + +void CriticalSection_Delete(CCriticalSection *p) +{ + // Print("CriticalSection_Delete") + if (p) + { + // int ret = + pthread_mutex_destroy(&p->_mutex); + } +} + +LONG InterlockedIncrement(LONG volatile *addend) +{ + // Print("InterlockedIncrement") + #ifdef USE_HACK_UNSAFE_ATOMIC + LONG val = *addend + 1; + *addend = val; + return val; + #else + + #if defined(__clang__) && (__clang_major__ >= 8) + #pragma GCC diagnostic ignored "-Watomic-implicit-seq-cst" + #endif + return __sync_add_and_fetch(addend, 1); + #endif +} + +#endif // _WIN32 + +WRes AutoResetEvent_OptCreate_And_Reset(CAutoResetEvent *p) +{ + if (Event_IsCreated(p)) + return Event_Reset(p); + return AutoResetEvent_CreateNotSignaled(p); +} + +#undef PRF +#undef Print diff --git a/3rdparty/lzma/C/Threads.h b/3rdparty/lzma/C/Threads.h index 9b3e1c556d2..4028464a337 100644 --- a/3rdparty/lzma/C/Threads.h +++ b/3rdparty/lzma/C/Threads.h @@ -1,38 +1,144 @@ /* Threads.h -- multithreading library -2013-11-12 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __7Z_THREADS_H -#define __7Z_THREADS_H +#ifndef ZIP7_INC_THREADS_H +#define ZIP7_INC_THREADS_H #ifdef _WIN32 -#include <windows.h> +#include "7zWindows.h" + +#else + +#if defined(__linux__) +#if !defined(__APPLE__) && !defined(_AIX) && !defined(__ANDROID__) +#ifndef Z7_AFFINITY_DISABLE +#define Z7_AFFINITY_SUPPORTED +// #pragma message(" ==== Z7_AFFINITY_SUPPORTED") +// #define _GNU_SOURCE +#endif +#endif +#endif + +#include <pthread.h> + #endif #include "7zTypes.h" EXTERN_C_BEGIN +#ifdef _WIN32 + WRes HandlePtr_Close(HANDLE *h); WRes Handle_WaitObject(HANDLE h); typedef HANDLE CThread; -#define Thread_Construct(p) *(p) = NULL + +#define Thread_CONSTRUCT(p) { *(p) = NULL; } #define Thread_WasCreated(p) (*(p) != NULL) #define Thread_Close(p) HandlePtr_Close(p) -#define Thread_Wait(p) Handle_WaitObject(*(p)) +// #define Thread_Wait(p) Handle_WaitObject(*(p)) -typedef #ifdef UNDER_CE - DWORD + // if (USE_THREADS_CreateThread is defined), we use _beginthreadex() + // if (USE_THREADS_CreateThread is not definned), we use CreateThread() + #define USE_THREADS_CreateThread +#endif + +typedef + #ifdef USE_THREADS_CreateThread + DWORD + #else + unsigned + #endif + THREAD_FUNC_RET_TYPE; + +#define THREAD_FUNC_RET_ZERO 0 + +typedef DWORD_PTR CAffinityMask; +typedef DWORD_PTR CCpuSet; + +#define CpuSet_Zero(p) *(p) = (0) +#define CpuSet_Set(p, cpu) *(p) |= ((DWORD_PTR)1 << (cpu)) + +#else // _WIN32 + +typedef struct +{ + pthread_t _tid; + int _created; +} CThread; + +#define Thread_CONSTRUCT(p) { (p)->_tid = 0; (p)->_created = 0; } +#define Thread_WasCreated(p) ((p)->_created != 0) +WRes Thread_Close(CThread *p); +// #define Thread_Wait Thread_Wait_Close + +typedef void * THREAD_FUNC_RET_TYPE; +#define THREAD_FUNC_RET_ZERO NULL + + +typedef UInt64 CAffinityMask; + +#ifdef Z7_AFFINITY_SUPPORTED + +typedef cpu_set_t CCpuSet; +#define CpuSet_Zero(p) CPU_ZERO(p) +#define CpuSet_Set(p, cpu) CPU_SET(cpu, p) +#define CpuSet_IsSet(p, cpu) CPU_ISSET(cpu, p) + +#else + +typedef UInt64 CCpuSet; +#define CpuSet_Zero(p) *(p) = (0) +#define CpuSet_Set(p, cpu) *(p) |= ((UInt64)1 << (cpu)) +#define CpuSet_IsSet(p, cpu) ((*(p) & ((UInt64)1 << (cpu))) != 0) + +#endif + + +#endif // _WIN32 + + +#define THREAD_FUNC_CALL_TYPE Z7_STDCALL + +#if defined(_WIN32) && defined(__GNUC__) +/* GCC compiler for x86 32-bit uses the rule: + the stack is 16-byte aligned before CALL instruction for function calling. + But only root function main() contains instructions that + set 16-byte alignment for stack pointer. And another functions + just keep alignment, if it was set in some parent function. + + The problem: + if we create new thread in MinGW (GCC) 32-bit x86 via _beginthreadex() or CreateThread(), + the root function of thread doesn't set 16-byte alignment. + And stack frames in all child functions also will be unaligned in that case. + + Here we set (force_align_arg_pointer) attribute for root function of new thread. + Do we need (force_align_arg_pointer) also for another systems? */ + + #define THREAD_FUNC_ATTRIB_ALIGN_ARG __attribute__((force_align_arg_pointer)) + // #define THREAD_FUNC_ATTRIB_ALIGN_ARG // for debug : bad alignment in SSE functions #else - unsigned + #define THREAD_FUNC_ATTRIB_ALIGN_ARG #endif - THREAD_FUNC_RET_TYPE; -#define THREAD_FUNC_CALL_TYPE MY_STD_CALL -#define THREAD_FUNC_DECL THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE +#define THREAD_FUNC_DECL THREAD_FUNC_ATTRIB_ALIGN_ARG THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE + typedef THREAD_FUNC_RET_TYPE (THREAD_FUNC_CALL_TYPE * THREAD_FUNC_TYPE)(void *); WRes Thread_Create(CThread *p, THREAD_FUNC_TYPE func, LPVOID param); +WRes Thread_Create_With_Affinity(CThread *p, THREAD_FUNC_TYPE func, LPVOID param, CAffinityMask affinity); +WRes Thread_Wait_Close(CThread *p); + +#ifdef _WIN32 +#define Thread_Create_With_CpuSet(p, func, param, cs) \ + Thread_Create_With_Affinity(p, func, param, *cs) +#else +WRes Thread_Create_With_CpuSet(CThread *p, THREAD_FUNC_TYPE func, LPVOID param, const CCpuSet *cpuSet); +#endif + + +#ifdef _WIN32 typedef HANDLE CEvent; typedef CEvent CAutoResetEvent; @@ -49,10 +155,12 @@ WRes AutoResetEvent_Create(CAutoResetEvent *p, int signaled); WRes AutoResetEvent_CreateNotSignaled(CAutoResetEvent *p); typedef HANDLE CSemaphore; -#define Semaphore_Construct(p) (*p) = NULL +#define Semaphore_Construct(p) *(p) = NULL +#define Semaphore_IsCreated(p) (*(p) != NULL) #define Semaphore_Close(p) HandlePtr_Close(p) #define Semaphore_Wait(p) Handle_WaitObject(*(p)) WRes Semaphore_Create(CSemaphore *p, UInt32 initCount, UInt32 maxCount); +WRes Semaphore_OptCreateInit(CSemaphore *p, UInt32 initCount, UInt32 maxCount); WRes Semaphore_ReleaseN(CSemaphore *p, UInt32 num); WRes Semaphore_Release1(CSemaphore *p); @@ -62,6 +170,71 @@ WRes CriticalSection_Init(CCriticalSection *p); #define CriticalSection_Enter(p) EnterCriticalSection(p) #define CriticalSection_Leave(p) LeaveCriticalSection(p) + +#else // _WIN32 + +typedef struct _CEvent +{ + int _created; + int _manual_reset; + int _state; + pthread_mutex_t _mutex; + pthread_cond_t _cond; +} CEvent; + +typedef CEvent CAutoResetEvent; +typedef CEvent CManualResetEvent; + +#define Event_Construct(p) (p)->_created = 0 +#define Event_IsCreated(p) ((p)->_created) + +WRes ManualResetEvent_Create(CManualResetEvent *p, int signaled); +WRes ManualResetEvent_CreateNotSignaled(CManualResetEvent *p); +WRes AutoResetEvent_Create(CAutoResetEvent *p, int signaled); +WRes AutoResetEvent_CreateNotSignaled(CAutoResetEvent *p); + +WRes Event_Set(CEvent *p); +WRes Event_Reset(CEvent *p); +WRes Event_Wait(CEvent *p); +WRes Event_Close(CEvent *p); + + +typedef struct _CSemaphore +{ + int _created; + UInt32 _count; + UInt32 _maxCount; + pthread_mutex_t _mutex; + pthread_cond_t _cond; +} CSemaphore; + +#define Semaphore_Construct(p) (p)->_created = 0 +#define Semaphore_IsCreated(p) ((p)->_created) + +WRes Semaphore_Create(CSemaphore *p, UInt32 initCount, UInt32 maxCount); +WRes Semaphore_OptCreateInit(CSemaphore *p, UInt32 initCount, UInt32 maxCount); +WRes Semaphore_ReleaseN(CSemaphore *p, UInt32 num); +#define Semaphore_Release1(p) Semaphore_ReleaseN(p, 1) +WRes Semaphore_Wait(CSemaphore *p); +WRes Semaphore_Close(CSemaphore *p); + + +typedef struct _CCriticalSection +{ + pthread_mutex_t _mutex; +} CCriticalSection; + +WRes CriticalSection_Init(CCriticalSection *p); +void CriticalSection_Delete(CCriticalSection *cs); +void CriticalSection_Enter(CCriticalSection *cs); +void CriticalSection_Leave(CCriticalSection *cs); + +LONG InterlockedIncrement(LONG volatile *addend); + +#endif // _WIN32 + +WRes AutoResetEvent_OptCreate_And_Reset(CAutoResetEvent *p); + EXTERN_C_END #endif diff --git a/3rdparty/lzma/C/Util/7z/7z.dsp b/3rdparty/lzma/C/Util/7z/7z.dsp index be0f0a74376..11e1b03e504 100644 --- a/3rdparty/lzma/C/Util/7z/7z.dsp +++ b/3rdparty/lzma/C/Util/7z/7z.dsp @@ -42,7 +42,7 @@ RSC=rc.exe # PROP Ignore_Export_Lib 0 # PROP Target_Dir "" # ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c -# ADD CPP /nologo /MD /W4 /WX /GX /O2 /D "NDEBUG" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /FAcs /Yu"Precomp.h" /FD /c +# ADD CPP /nologo /MD /W4 /WX /GX /O2 /D "NDEBUG" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /D "Z7_PPMD_SUPPORT" /FAcs /Yu"Precomp.h" /FD /c # ADD BASE RSC /l 0x419 /d "NDEBUG" # ADD RSC /l 0x419 /d "NDEBUG" BSC32=bscmake.exe @@ -67,7 +67,7 @@ LINK32=link.exe # PROP Ignore_Export_Lib 0 # PROP Target_Dir "" # ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c -# ADD CPP /nologo /W4 /WX /Gm /GX /ZI /Od /D "_DEBUG" /D "_SZ_ALLOC_DEBUG2" /D "_SZ_NO_INT_64_A" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /Yu"Precomp.h" /FD /GZ /c +# ADD CPP /nologo /W4 /WX /Gm /GX /ZI /Od /D "_DEBUG" /D "_SZ_ALLOC_DEBUG2" /D "_SZ_NO_INT_64_A" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /D "Z7_PPMD_SUPPORT" /Yu"Precomp.h" /FD /GZ /c # ADD BASE RSC /l 0x419 /d "_DEBUG" # ADD RSC /l 0x419 /d "_DEBUG" BSC32=bscmake.exe @@ -145,6 +145,10 @@ SOURCE=..\..\7zTypes.h # End Source File # Begin Source File +SOURCE=..\..\7zWindows.h +# End Source File +# Begin Source File + SOURCE=..\..\Bcj2.c # End Source File # Begin Source File diff --git a/3rdparty/lzma/C/Util/7z/7zMain.c b/3rdparty/lzma/C/Util/7z/7zMain.c index 1837e9d606a..547920ac31f 100644 --- a/3rdparty/lzma/C/Util/7z/7zMain.c +++ b/3rdparty/lzma/C/Util/7z/7zMain.c @@ -1,11 +1,13 @@ /* 7zMain.c - Test application for 7z Decoder -2016-05-16 : Igor Pavlov : Public domain */ +2023-04-04 : Igor Pavlov : Public domain */ #include "Precomp.h" #include <stdio.h> #include <string.h> +#include "../../CpuArch.h" + #include "../../7z.h" #include "../../7zAlloc.h" #include "../../7zBuf.h" @@ -18,12 +20,29 @@ #ifdef _WIN32 #include <direct.h> #else +#include <stdlib.h> +#include <time.h> +#ifdef __GNUC__ +#include <sys/time.h> +#endif +#include <fcntl.h> +// #include <utime.h> #include <sys/stat.h> #include <errno.h> #endif #endif -static ISzAlloc g_Alloc = { SzAlloc, SzFree }; +#define kInputBufSize ((size_t)1 << 18) + +static const ISzAlloc g_Alloc = { SzAlloc, SzFree }; +// static const ISzAlloc g_Alloc_temp = { SzAllocTemp, SzFreeTemp }; + + +static void Print(const char *s) +{ + fputs(s, stdout); +} + static int Buf_EnsureSize(CBuf *dest, size_t size) { @@ -34,19 +53,19 @@ static int Buf_EnsureSize(CBuf *dest, size_t size) } #ifndef _WIN32 -#define _USE_UTF8 +#define MY_USE_UTF8 #endif -/* #define _USE_UTF8 */ +/* #define MY_USE_UTF8 */ -#ifdef _USE_UTF8 +#ifdef MY_USE_UTF8 -#define _UTF8_START(n) (0x100 - (1 << (7 - (n)))) +#define MY_UTF8_START(n) (0x100 - (1 << (7 - (n)))) -#define _UTF8_RANGE(n) (((UInt32)1) << ((n) * 5 + 6)) +#define MY_UTF8_RANGE(n) (((UInt32)1) << ((n) * 5 + 6)) -#define _UTF8_HEAD(n, val) ((Byte)(_UTF8_START(n) + (val >> (6 * (n))))) -#define _UTF8_CHAR(n, val) ((Byte)(0x80 + (((val) >> (6 * (n))) & 0x3F))) +#define MY_UTF8_HEAD(n, val) ((Byte)(MY_UTF8_START(n) + (val >> (6 * (n))))) +#define MY_UTF8_CHAR(n, val) ((Byte)(0x80 + (((val) >> (6 * (n))) & 0x3F))) static size_t Utf16_To_Utf8_Calc(const UInt16 *src, const UInt16 *srcLim) { @@ -63,7 +82,7 @@ static size_t Utf16_To_Utf8_Calc(const UInt16 *src, const UInt16 *srcLim) if (val < 0x80) continue; - if (val < _UTF8_RANGE(1)) + if (val < MY_UTF8_RANGE(1)) { size++; continue; @@ -71,7 +90,7 @@ static size_t Utf16_To_Utf8_Calc(const UInt16 *src, const UInt16 *srcLim) if (val >= 0xD800 && val < 0xDC00 && src != srcLim) { - UInt32 c2 = *src; + const UInt32 c2 = *src; if (c2 >= 0xDC00 && c2 < 0xE000) { src++; @@ -96,37 +115,37 @@ static Byte *Utf16_To_Utf8(Byte *dest, const UInt16 *src, const UInt16 *srcLim) if (val < 0x80) { - *dest++ = (char)val; + *dest++ = (Byte)val; continue; } - if (val < _UTF8_RANGE(1)) + if (val < MY_UTF8_RANGE(1)) { - dest[0] = _UTF8_HEAD(1, val); - dest[1] = _UTF8_CHAR(0, val); + dest[0] = MY_UTF8_HEAD(1, val); + dest[1] = MY_UTF8_CHAR(0, val); dest += 2; continue; } if (val >= 0xD800 && val < 0xDC00 && src != srcLim) { - UInt32 c2 = *src; + const UInt32 c2 = *src; if (c2 >= 0xDC00 && c2 < 0xE000) { src++; val = (((val - 0xD800) << 10) | (c2 - 0xDC00)) + 0x10000; - dest[0] = _UTF8_HEAD(3, val); - dest[1] = _UTF8_CHAR(2, val); - dest[2] = _UTF8_CHAR(1, val); - dest[3] = _UTF8_CHAR(0, val); + dest[0] = MY_UTF8_HEAD(3, val); + dest[1] = MY_UTF8_CHAR(2, val); + dest[2] = MY_UTF8_CHAR(1, val); + dest[3] = MY_UTF8_CHAR(0, val); dest += 4; continue; } } - dest[0] = _UTF8_HEAD(2, val); - dest[1] = _UTF8_CHAR(1, val); - dest[2] = _UTF8_CHAR(0, val); + dest[0] = MY_UTF8_HEAD(2, val); + dest[1] = MY_UTF8_CHAR(1, val); + dest[2] = MY_UTF8_CHAR(0, val); dest += 3; } } @@ -144,27 +163,27 @@ static SRes Utf16_To_Utf8Buf(CBuf *dest, const UInt16 *src, size_t srcLen) #endif static SRes Utf16_To_Char(CBuf *buf, const UInt16 *s - #ifndef _USE_UTF8 + #ifndef MY_USE_UTF8 , UINT codePage #endif ) { unsigned len = 0; - for (len = 0; s[len] != 0; len++); + for (len = 0; s[len] != 0; len++) {} - #ifndef _USE_UTF8 + #ifndef MY_USE_UTF8 { - unsigned size = len * 3 + 100; + const unsigned size = len * 3 + 100; if (!Buf_EnsureSize(buf, size)) return SZ_ERROR_MEM; { buf->data[0] = 0; if (len != 0) { - char defaultChar = '_'; + const char defaultChar = '_'; BOOL defUsed; - unsigned numChars = 0; - numChars = WideCharToMultiByte(codePage, 0, s, len, (char *)buf->data, size, &defaultChar, &defUsed); + const unsigned numChars = (unsigned)WideCharToMultiByte( + codePage, 0, (LPCWSTR)s, (int)len, (char *)buf->data, (int)size, &defaultChar, &defUsed); if (numChars == 0 || numChars >= size) return SZ_ERROR_FAIL; buf->data[numChars] = 0; @@ -180,8 +199,8 @@ static SRes Utf16_To_Char(CBuf *buf, const UInt16 *s #ifdef _WIN32 #ifndef USE_WINDOWS_FILE static UINT g_FileCodePage = CP_ACP; + #define MY_FILE_CODE_PAGE_PARAM ,g_FileCodePage #endif - #define MY_FILE_CODE_PAGE_PARAM ,g_FileCodePage #else #define MY_FILE_CODE_PAGE_PARAM #endif @@ -190,14 +209,14 @@ static WRes MyCreateDir(const UInt16 *name) { #ifdef USE_WINDOWS_FILE - return CreateDirectoryW(name, NULL) ? 0 : GetLastError(); + return CreateDirectoryW((LPCWSTR)name, NULL) ? 0 : GetLastError(); #else CBuf buf; WRes res; Buf_Init(&buf); - RINOK(Utf16_To_Char(&buf, name MY_FILE_CODE_PAGE_PARAM)); + RINOK(Utf16_To_Char(&buf, name MY_FILE_CODE_PAGE_PARAM)) res = #ifdef _WIN32 @@ -215,35 +234,36 @@ static WRes MyCreateDir(const UInt16 *name) static WRes OutFile_OpenUtf16(CSzFile *p, const UInt16 *name) { #ifdef USE_WINDOWS_FILE - return OutFile_OpenW(p, name); + return OutFile_OpenW(p, (LPCWSTR)name); #else CBuf buf; WRes res; Buf_Init(&buf); - RINOK(Utf16_To_Char(&buf, name MY_FILE_CODE_PAGE_PARAM)); + RINOK(Utf16_To_Char(&buf, name MY_FILE_CODE_PAGE_PARAM)) res = OutFile_Open(p, (const char *)buf.data); Buf_Free(&buf, &g_Alloc); return res; #endif } + static SRes PrintString(const UInt16 *s) { CBuf buf; SRes res; Buf_Init(&buf); res = Utf16_To_Char(&buf, s - #ifndef _USE_UTF8 + #ifndef MY_USE_UTF8 , CP_OEMCP #endif ); if (res == SZ_OK) - fputs((const char *)buf.data, stdout); + Print((const char *)buf.data); Buf_Free(&buf, &g_Alloc); return res; } -static void UInt64ToStr(UInt64 value, char *s) +static void UInt64ToStr(UInt64 value, char *s, int numDigits) { char temp[32]; int pos = 0; @@ -253,6 +273,10 @@ static void UInt64ToStr(UInt64 value, char *s) value /= 10; } while (value != 0); + + for (numDigits -= pos; numDigits > 0; numDigits--) + *s++ = ' '; + do *s++ = temp[--pos]; while (pos); @@ -266,8 +290,10 @@ static char *UIntToStr(char *s, unsigned value, int numDigits) do temp[pos++] = (char)('0' + (value % 10)); while (value /= 10); + for (numDigits -= pos; numDigits > 0; numDigits--) *s++ = '0'; + do *s++ = temp[--pos]; while (pos); @@ -281,17 +307,142 @@ static void UIntToStr_2(char *s, unsigned value) s[1] = (char)('0' + (value % 10)); } + #define PERIOD_4 (4 * 365 + 1) #define PERIOD_100 (PERIOD_4 * 25 - 1) #define PERIOD_400 (PERIOD_100 * 4 + 1) -static void ConvertFileTimeToString(const CNtfsFileTime *nt, char *s) + + +#ifndef _WIN32 + +// MS uses long for BOOL, but long is 32-bit in MS. So we use int. +// typedef long BOOL; +typedef int BOOL; + +typedef struct _FILETIME +{ + DWORD dwLowDateTime; + DWORD dwHighDateTime; +} FILETIME; + +static LONG TIME_GetBias() +{ + const time_t utc = time(NULL); + struct tm *ptm = localtime(&utc); + const int localdaylight = ptm->tm_isdst; /* daylight for local timezone */ + ptm = gmtime(&utc); + ptm->tm_isdst = localdaylight; /* use local daylight, not that of Greenwich */ + const LONG bias = (int)(mktime(ptm) - utc); + return bias; +} + +#define TICKS_PER_SEC 10000000 + +#define GET_TIME_64(pft) ((pft)->dwLowDateTime | ((UInt64)(pft)->dwHighDateTime << 32)) + +#define SET_FILETIME(ft, v64) \ + (ft)->dwLowDateTime = (DWORD)v64; \ + (ft)->dwHighDateTime = (DWORD)(v64 >> 32); + +#define WINAPI +#define TRUE 1 + +static BOOL WINAPI FileTimeToLocalFileTime(const FILETIME *fileTime, FILETIME *localFileTime) +{ + UInt64 v = GET_TIME_64(fileTime); + v = (UInt64)((Int64)v - (Int64)TIME_GetBias() * TICKS_PER_SEC); + SET_FILETIME(localFileTime, v) + return TRUE; +} + +static const UInt32 kNumTimeQuantumsInSecond = 10000000; +static const UInt32 kFileTimeStartYear = 1601; +static const UInt32 kUnixTimeStartYear = 1970; +static const UInt64 kUnixTimeOffset = + (UInt64)60 * 60 * 24 * (89 + 365 * (kUnixTimeStartYear - kFileTimeStartYear)); + +static Int64 Time_FileTimeToUnixTime64(const FILETIME *ft) +{ + const UInt64 winTime = GET_TIME_64(ft); + return (Int64)(winTime / kNumTimeQuantumsInSecond) - (Int64)kUnixTimeOffset; +} + +#if defined(_AIX) + #define MY_ST_TIMESPEC st_timespec +#else + #define MY_ST_TIMESPEC timespec +#endif + +static void FILETIME_To_timespec(const FILETIME *ft, struct MY_ST_TIMESPEC *ts) +{ + if (ft) + { + const Int64 sec = Time_FileTimeToUnixTime64(ft); + // time_t is long + const time_t sec2 = (time_t)sec; + if (sec2 == sec) + { + ts->tv_sec = sec2; + const UInt64 winTime = GET_TIME_64(ft); + ts->tv_nsec = (long)((winTime % 10000000) * 100); + return; + } + } + // else + { + ts->tv_sec = 0; + // ts.tv_nsec = UTIME_NOW; // set to the current time + ts->tv_nsec = UTIME_OMIT; // keep old timesptamp + } +} + +static WRes Set_File_FILETIME(const UInt16 *name, const FILETIME *mTime) +{ + struct timespec times[2]; + + const int flags = 0; // follow link + // = AT_SYMLINK_NOFOLLOW; // don't follow link + + CBuf buf; + int res; + Buf_Init(&buf); + RINOK(Utf16_To_Char(&buf, name MY_FILE_CODE_PAGE_PARAM)) + FILETIME_To_timespec(NULL, ×[0]); + FILETIME_To_timespec(mTime, ×[1]); + res = utimensat(AT_FDCWD, (const char *)buf.data, times, flags); + Buf_Free(&buf, &g_Alloc); + if (res == 0) + return 0; + return errno; +} + +#endif + +static void NtfsFileTime_to_FILETIME(const CNtfsFileTime *t, FILETIME *ft) +{ + ft->dwLowDateTime = (DWORD)(t->Low); + ft->dwHighDateTime = (DWORD)(t->High); +} + +static void ConvertFileTimeToString(const CNtfsFileTime *nTime, char *s) { unsigned year, mon, hour, min, sec; Byte ms[] = { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 }; unsigned t; UInt32 v; - UInt64 v64 = nt->Low | ((UInt64)nt->High << 32); + // UInt64 v64 = nt->Low | ((UInt64)nt->High << 32); + UInt64 v64; + { + FILETIME fileTime, locTime; + NtfsFileTime_to_FILETIME(nTime, &fileTime); + if (!FileTimeToLocalFileTime(&fileTime, &locTime)) + { + locTime.dwHighDateTime = + locTime.dwLowDateTime = 0; + } + v64 = locTime.dwLowDateTime | ((UInt64)locTime.dwHighDateTime << 32); + } v64 /= 10000000; sec = (unsigned)(v64 % 60); v64 /= 60; min = (unsigned)(v64 % 60); v64 /= 60; @@ -310,7 +461,7 @@ static void ConvertFileTimeToString(const CNtfsFileTime *nt, char *s) ms[1] = 29; for (mon = 0;; mon++) { - unsigned d = ms[mon]; + const unsigned d = ms[mon]; if (v < d) break; v -= d; @@ -323,12 +474,56 @@ static void ConvertFileTimeToString(const CNtfsFileTime *nt, char *s) UIntToStr_2(s, sec); s[2] = 0; } -void PrintError(char *sz) +static void PrintLF(void) +{ + Print("\n"); +} + +static void PrintError(char *s) +{ + Print("\nERROR: "); + Print(s); + PrintLF(); +} + +static void PrintError_WRes(const char *message, WRes wres) { - printf("\nERROR: %s\n", sz); + Print("\nERROR: "); + Print(message); + PrintLF(); + { + char s[32]; + UIntToStr(s, (unsigned)wres, 1); + Print("System error code: "); + Print(s); + } + // sprintf(buffer + strlen(buffer), "\nSystem error code: %d", (unsigned)wres); + #ifdef _WIN32 + { + char *s = NULL; + if (FormatMessageA(FORMAT_MESSAGE_ALLOCATE_BUFFER | + FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS, + NULL, wres, 0, (LPSTR) &s, 0, NULL) != 0 && s) + { + Print(" : "); + Print(s); + LocalFree(s); + } + } + #else + { + const char *s = strerror(wres); + if (s) + { + Print(" : "); + Print(s); + } + } + #endif + PrintLF(); } -static void GetAttribString(UInt32 wa, Bool isDir, char *s) +static void GetAttribString(UInt32 wa, BoolInt isDir, char *s) { #ifdef USE_WINDOWS_FILE s[0] = (char)(((wa & FILE_ATTRIBUTE_DIRECTORY) != 0 || isDir) ? 'D' : '.'); @@ -343,25 +538,27 @@ static void GetAttribString(UInt32 wa, Bool isDir, char *s) #endif } + // #define NUM_PARENTS_MAX 128 -int MY_CDECL main(int numargs, char *args[]) +int Z7_CDECL main(int numargs, char *args[]) { + ISzAlloc allocImp; + ISzAlloc allocTempImp; + CFileInStream archiveStream; - CLookToRead lookStream; + CLookToRead2 lookStream; CSzArEx db; SRes res; - ISzAlloc allocImp; - ISzAlloc allocTempImp; UInt16 *temp = NULL; size_t tempSize = 0; // UInt32 parents[NUM_PARENTS_MAX]; - printf("\n7z ANSI-C Decoder " MY_VERSION_COPYRIGHT_DATE "\n\n"); + Print("\n7z Decoder " MY_VERSION_CPU " : " MY_COPYRIGHT_DATE "\n\n"); if (numargs == 1) { - printf( + Print( "Usage: 7zDec <command> <archive_name>\n\n" "<Commands>\n" " e: Extract files from archive (without using directory names)\n" @@ -370,7 +567,7 @@ int MY_CDECL main(int numargs, char *args[]) " x: eXtract files with full paths\n"); return 0; } - + if (numargs < 3) { PrintError("incorrect command"); @@ -381,33 +578,52 @@ int MY_CDECL main(int numargs, char *args[]) g_FileCodePage = AreFileApisANSI() ? CP_ACP : CP_OEMCP; #endif - allocImp.Alloc = SzAlloc; - allocImp.Free = SzFree; - allocTempImp.Alloc = SzAllocTemp; - allocTempImp.Free = SzFreeTemp; + allocImp = g_Alloc; + allocTempImp = g_Alloc; + // allocTempImp = g_Alloc_temp; - #ifdef UNDER_CE - if (InFile_OpenW(&archiveStream.file, L"\test.7z")) - #else - if (InFile_Open(&archiveStream.file, args[2])) - #endif { - PrintError("can not open input file"); - return 1; + WRes wres = + #ifdef UNDER_CE + InFile_OpenW(&archiveStream.file, L"\test.7z"); // change it + #else + InFile_Open(&archiveStream.file, args[2]); + #endif + if (wres != 0) + { + PrintError_WRes("cannot open input file", wres); + return 1; + } } FileInStream_CreateVTable(&archiveStream); - LookToRead_CreateVTable(&lookStream, False); - - lookStream.realStream = &archiveStream.s; - LookToRead_Init(&lookStream); + archiveStream.wres = 0; + LookToRead2_CreateVTable(&lookStream, False); + lookStream.buf = NULL; - CrcGenerateTable(); + res = SZ_OK; + { + lookStream.buf = (Byte *)ISzAlloc_Alloc(&allocImp, kInputBufSize); + if (!lookStream.buf) + res = SZ_ERROR_MEM; + else + { + lookStream.bufSize = kInputBufSize; + lookStream.realStream = &archiveStream.vt; + LookToRead2_INIT(&lookStream) + } + } + + CrcGenerateTable(); + SzArEx_Init(&db); - - res = SzArEx_Open(&db, &lookStream.s, &allocImp, &allocTempImp); + + if (res == SZ_OK) + { + res = SzArEx_Open(&db, &lookStream.vt, &allocImp, &allocTempImp); + } if (res == SZ_OK) { @@ -442,7 +658,7 @@ int MY_CDECL main(int numargs, char *args[]) size_t outSizeProcessed = 0; // const CSzFileItem *f = db.Files + i; size_t len; - unsigned isDir = SzArEx_IsDir(&db, i); + const BoolInt isDir = SzArEx_IsDir(&db, i); if (listCommand == 0 && isDir && !fullPaths) continue; len = SzArEx_GetFileNameUtf16(&db, i, NULL); @@ -477,7 +693,7 @@ int MY_CDECL main(int numargs, char *args[]) GetAttribString(SzBitWithVals_Check(&db.Attribs, i) ? db.Attribs.Vals[i] : 0, isDir, attr); fileSize = SzArEx_GetFileSize(&db, i); - UInt64ToStr(fileSize, s); + UInt64ToStr(fileSize, s, 10); if (SzBitWithVals_Check(&db.MTime, i)) ConvertFileTimeToString(&db.MTime.Vals[i], t); @@ -489,29 +705,33 @@ int MY_CDECL main(int numargs, char *args[]) t[j] = '\0'; } - printf("%s %s %10s ", t, attr, s); + Print(t); + Print(" "); + Print(attr); + Print(" "); + Print(s); + Print(" "); res = PrintString(temp); if (res != SZ_OK) break; if (isDir) - printf("/"); - printf("\n"); + Print("/"); + PrintLF(); continue; } - fputs(testCommand ? - "Testing ": - "Extracting ", - stdout); + Print(testCommand ? + "T ": + "- "); res = PrintString(temp); if (res != SZ_OK) break; if (isDir) - printf("/"); + Print("/"); else { - res = SzArEx_Extract(&db, &lookStream.s, i, + res = SzArEx_Extract(&db, &lookStream.vt, i, &blockIndex, &outBuffer, &outBufferSize, &offset, &outSizeProcessed, &allocImp, &allocTempImp); @@ -543,62 +763,126 @@ int MY_CDECL main(int numargs, char *args[]) if (isDir) { MyCreateDir(destPath); - printf("\n"); + PrintLF(); continue; } - else if (OutFile_OpenUtf16(&outFile, destPath)) + else { - PrintError("can not open output file"); - res = SZ_ERROR_FAIL; - break; + const WRes wres = OutFile_OpenUtf16(&outFile, destPath); + if (wres != 0) + { + PrintError_WRes("cannot open output file", wres); + res = SZ_ERROR_FAIL; + break; + } } processedSize = outSizeProcessed; - if (File_Write(&outFile, outBuffer + offset, &processedSize) != 0 || processedSize != outSizeProcessed) { - PrintError("can not write output file"); - res = SZ_ERROR_FAIL; - break; + const WRes wres = File_Write(&outFile, outBuffer + offset, &processedSize); + if (wres != 0 || processedSize != outSizeProcessed) + { + PrintError_WRes("cannot write output file", wres); + res = SZ_ERROR_FAIL; + break; + } } - - if (File_Close(&outFile)) + { - PrintError("can not close output file"); - res = SZ_ERROR_FAIL; - break; + FILETIME mtime; + FILETIME *mtimePtr = NULL; + + #ifdef USE_WINDOWS_FILE + FILETIME ctime; + FILETIME *ctimePtr = NULL; + #endif + + if (SzBitWithVals_Check(&db.MTime, i)) + { + const CNtfsFileTime *t = &db.MTime.Vals[i]; + mtime.dwLowDateTime = (DWORD)(t->Low); + mtime.dwHighDateTime = (DWORD)(t->High); + mtimePtr = &mtime; + } + + #ifdef USE_WINDOWS_FILE + if (SzBitWithVals_Check(&db.CTime, i)) + { + const CNtfsFileTime *t = &db.CTime.Vals[i]; + ctime.dwLowDateTime = (DWORD)(t->Low); + ctime.dwHighDateTime = (DWORD)(t->High); + ctimePtr = &ctime; + } + + if (mtimePtr || ctimePtr) + SetFileTime(outFile.handle, ctimePtr, NULL, mtimePtr); + #endif + + { + const WRes wres = File_Close(&outFile); + if (wres != 0) + { + PrintError_WRes("cannot close output file", wres); + res = SZ_ERROR_FAIL; + break; + } + } + + #ifndef USE_WINDOWS_FILE + #ifdef _WIN32 + mtimePtr = mtimePtr; + #else + if (mtimePtr) + Set_File_FILETIME(destPath, mtimePtr); + #endif + #endif } #ifdef USE_WINDOWS_FILE if (SzBitWithVals_Check(&db.Attribs, i)) - SetFileAttributesW(destPath, db.Attribs.Vals[i]); + { + UInt32 attrib = db.Attribs.Vals[i]; + /* p7zip stores posix attributes in high 16 bits and adds 0x8000 as marker. + We remove posix bits, if we detect posix mode field */ + if ((attrib & 0xF0000000) != 0) + attrib &= 0x7FFF; + SetFileAttributesW((LPCWSTR)destPath, attrib); + } #endif } - printf("\n"); + PrintLF(); } - IAlloc_Free(&allocImp, outBuffer); + ISzAlloc_Free(&allocImp, outBuffer); } } - SzArEx_Free(&db, &allocImp); SzFree(NULL, temp); + SzArEx_Free(&db, &allocImp); + ISzAlloc_Free(&allocImp, lookStream.buf); File_Close(&archiveStream.file); if (res == SZ_OK) { - printf("\nEverything is Ok\n"); + Print("\nEverything is Ok\n"); return 0; } if (res == SZ_ERROR_UNSUPPORTED) PrintError("decoder doesn't support this archive"); else if (res == SZ_ERROR_MEM) - PrintError("can not allocate memory"); + PrintError("cannot allocate memory"); else if (res == SZ_ERROR_CRC) PrintError("CRC error"); + else if (res == SZ_ERROR_READ /* || archiveStream.Res != 0 */) + PrintError_WRes("Read Error", archiveStream.wres); else - printf("\nERROR #%d\n", res); + { + char s[32]; + UInt64ToStr((unsigned)res, s, 0); + PrintError(s); + } return 1; } diff --git a/3rdparty/lzma/C/Util/7z/Precomp.h b/3rdparty/lzma/C/Util/7z/Precomp.h index 588a66f7eb8..bc8fa219b58 100644 --- a/3rdparty/lzma/C/Util/7z/Precomp.h +++ b/3rdparty/lzma/C/Util/7z/Precomp.h @@ -1,8 +1,12 @@ /* Precomp.h -- StdAfx -2013-06-16 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __7Z_PRECOMP_H -#define __7Z_PRECOMP_H +#ifndef ZIP7_INC_PRECOMP_H +#define ZIP7_INC_PRECOMP_H + +#if defined(_MSC_VER) && _MSC_VER >= 1800 +#pragma warning(disable : 4464) // relative include path contains '..' +#endif #include "../../Compiler.h" #include "../../7zTypes.h" diff --git a/3rdparty/lzma/C/Util/7z/makefile b/3rdparty/lzma/C/Util/7z/makefile index 9a49fd515a6..dfc560e9255 100644 --- a/3rdparty/lzma/C/Util/7z/makefile +++ b/3rdparty/lzma/C/Util/7z/makefile @@ -1,4 +1,4 @@ -CFLAGS = $(CFLAGS) -D_7ZIP_PPMD_SUPPPORT +CFLAGS = $(CFLAGS) -DZ7_PPMD_SUPPORT -DZ7_EXTRACT_ONLY PROG = 7zDec.exe diff --git a/3rdparty/lzma/C/Util/7z/makefile.gcc b/3rdparty/lzma/C/Util/7z/makefile.gcc index 51053bad65c..f48d36211d0 100644 --- a/3rdparty/lzma/C/Util/7z/makefile.gcc +++ b/3rdparty/lzma/C/Util/7z/makefile.gcc @@ -1,75 +1,32 @@ -PROG = 7zDec -CXX = gcc -LIB = -RM = rm -f -CFLAGS = -c -O2 -Wall - -OBJS = 7zMain.o 7zAlloc.o 7zArcIn.o 7zBuf.o 7zBuf2.o 7zCrc.o 7zCrcOpt.o 7zDec.o CpuArch.o Delta.o LzmaDec.o Lzma2Dec.o Bra.o Bra86.o BraIA64.o Bcj2.o Ppmd7.o Ppmd7Dec.o 7zFile.o 7zStream.o - -all: $(PROG) - -$(PROG): $(OBJS) - $(CXX) -o $(PROG) $(LDFLAGS) $(OBJS) $(LIB) - -7zMain.o: 7zMain.c - $(CXX) $(CFLAGS) 7zMain.c - -7zAlloc.o: ../../7zAlloc.c - $(CXX) $(CFLAGS) ../../7zAlloc.c - -7zArcIn.o: ../../7zArcIn.c - $(CXX) $(CFLAGS) ../../7zArcIn.c - -7zBuf.o: ../../7zBuf.c - $(CXX) $(CFLAGS) ../../7zBuf.c - -7zBuf2.o: ../../7zBuf2.c - $(CXX) $(CFLAGS) ../../7zBuf2.c - -7zCrc.o: ../../7zCrc.c - $(CXX) $(CFLAGS) ../../7zCrc.c - -7zCrcOpt.o: ../../7zCrc.c - $(CXX) $(CFLAGS) ../../7zCrcOpt.c - -7zDec.o: ../../7zDec.c - $(CXX) $(CFLAGS) -D_7ZIP_PPMD_SUPPPORT ../../7zDec.c - -CpuArch.o: ../../CpuArch.c - $(CXX) $(CFLAGS) ../../CpuArch.c - -Delta.o: ../../Delta.c - $(CXX) $(CFLAGS) ../../Delta.c - -LzmaDec.o: ../../LzmaDec.c - $(CXX) $(CFLAGS) ../../LzmaDec.c - -Lzma2Dec.o: ../../Lzma2Dec.c - $(CXX) $(CFLAGS) ../../Lzma2Dec.c - -Bra.o: ../../Bra.c - $(CXX) $(CFLAGS) ../../Bra.c - -Bra86.o: ../../Bra86.c - $(CXX) $(CFLAGS) ../../Bra86.c - -BraIA64.o: ../../BraIA64.c - $(CXX) $(CFLAGS) ../../BraIA64.c - -Bcj2.o: ../../Bcj2.c - $(CXX) $(CFLAGS) ../../Bcj2.c - -Ppmd7.o: ../../Ppmd7.c - $(CXX) $(CFLAGS) ../../Ppmd7.c - -Ppmd7Dec.o: ../../Ppmd7Dec.c - $(CXX) $(CFLAGS) ../../Ppmd7Dec.c - -7zFile.o: ../../7zFile.c - $(CXX) $(CFLAGS) ../../7zFile.c - -7zStream.o: ../../7zStream.c - $(CXX) $(CFLAGS) ../../7zStream.c - -clean: - -$(RM) $(PROG) $(OBJS) +PROG = 7zdec + +LOCAL_FLAGS = -DZ7_PPMD_SUPPORT -DZ7_EXTRACT_ONLY + +include ../../../CPP/7zip/LzmaDec_gcc.mak + + +OBJS = \ + $(LZMA_DEC_OPT_OBJS) \ + $O/Bcj2.o \ + $O/Bra.o \ + $O/Bra86.o \ + $O/BraIA64.o \ + $O/CpuArch.o \ + $O/Delta.o \ + $O/Lzma2Dec.o \ + $O/LzmaDec.o \ + $O/Ppmd7.o \ + $O/Ppmd7Dec.o \ + $O/7zCrc.o \ + $O/7zCrcOpt.o \ + $O/7zAlloc.o \ + $O/7zArcIn.o \ + $O/7zBuf.o \ + $O/7zBuf2.o \ + $O/7zDec.o \ + $O/7zMain.o \ + $O/7zFile.o \ + $O/7zStream.o \ + + +include ../../7zip_gcc_c.mak diff --git a/3rdparty/lzma/C/Util/Lzma/LzmaUtil.c b/3rdparty/lzma/C/Util/Lzma/LzmaUtil.c index e5857ceff09..b9b974bd1bb 100644 --- a/3rdparty/lzma/C/Util/Lzma/LzmaUtil.c +++ b/3rdparty/lzma/C/Util/Lzma/LzmaUtil.c @@ -1,57 +1,111 @@ /* LzmaUtil.c -- Test application for LZMA compression -2015-11-08 : Igor Pavlov : Public domain */ +2023-03-07 : Igor Pavlov : Public domain */ -#include "../../Precomp.h" +#include "Precomp.h" #include <stdio.h> #include <stdlib.h> #include <string.h> +#include "../../CpuArch.h" + #include "../../Alloc.h" #include "../../7zFile.h" #include "../../7zVersion.h" +#include "../../LzFind.h" #include "../../LzmaDec.h" #include "../../LzmaEnc.h" -const char *kCantReadMessage = "Can not read input file"; -const char *kCantWriteMessage = "Can not write output file"; -const char *kCantAllocateMessage = "Can not allocate memory"; -const char *kDataErrorMessage = "Data error"; +static const char * const kCantReadMessage = "Cannot read input file"; +static const char * const kCantWriteMessage = "Cannot write output file"; +static const char * const kCantAllocateMessage = "Cannot allocate memory"; +static const char * const kDataErrorMessage = "Data error"; -void PrintHelp(char *buffer) +static void Print(const char *s) { - strcat(buffer, "\nLZMA Utility " MY_VERSION_COPYRIGHT_DATE "\n" - "\nUsage: lzma <e|d> inputFile outputFile\n" - " e: encode file\n" - " d: decode file\n"); + fputs(s, stdout); } -int PrintError(char *buffer, const char *message) +static void PrintHelp(void) +{ + Print( + "\n" "LZMA-C " MY_VERSION_CPU " : " MY_COPYRIGHT_DATE + "\n" + "\n" "Usage: lzma <e|d> inputFile outputFile" + "\n" " e: encode file" + "\n" " d: decode file" + "\n"); +} + +static int PrintError(const char *message) +{ + Print("\nError: "); + Print(message); + Print("\n"); + return 1; +} + +#define CONVERT_INT_TO_STR(charType, tempSize) \ + unsigned char temp[tempSize]; unsigned i = 0; \ + while (val >= 10) { temp[i++] = (unsigned char)('0' + (unsigned)(val % 10)); val /= 10; } \ + *s++ = (charType)('0' + (unsigned)val); \ + while (i != 0) { i--; *s++ = (charType)temp[i]; } \ + *s = 0; \ + return s; + +static char * Convert_unsigned_To_str(unsigned val, char *s) { - strcat(buffer, "\nError: "); - strcat(buffer, message); - strcat(buffer, "\n"); + CONVERT_INT_TO_STR(char, 32) +} + +static void Print_unsigned(unsigned code) +{ + char str[32]; + Convert_unsigned_To_str(code, str); + Print(str); +} + +static int PrintError_WRes(const char *message, WRes wres) +{ + PrintError(message); + Print("\nSystem error code: "); + Print_unsigned((unsigned)wres); + #ifndef _WIN32 + { + const char *s = strerror(wres); + if (s) + { + Print(" : "); + Print(s); + } + } + #endif + Print("\n"); return 1; } -int PrintErrorNumber(char *buffer, SRes val) +static int PrintErrorNumber(SRes val) { - sprintf(buffer + strlen(buffer), "\nError code: %x\n", (unsigned)val); + Print("\n7-Zip error code: "); + Print_unsigned((unsigned)val); + Print("\n"); return 1; } -int PrintUserError(char *buffer) +static int PrintUserError(void) { - return PrintError(buffer, "Incorrect command"); + return PrintError("Incorrect command"); } + #define IN_BUF_SIZE (1 << 16) #define OUT_BUF_SIZE (1 << 16) -static SRes Decode2(CLzmaDec *state, ISeqOutStream *outStream, ISeqInStream *inStream, + +static SRes Decode2(CLzmaDec *state, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, UInt64 unpackSize) { - int thereIsSize = (unpackSize != (UInt64)(Int64)-1); + const int thereIsSize = (unpackSize != (UInt64)(Int64)-1); Byte inBuf[IN_BUF_SIZE]; Byte outBuf[OUT_BUF_SIZE]; size_t inPos = 0, inSize = 0, outPos = 0; @@ -61,7 +115,7 @@ static SRes Decode2(CLzmaDec *state, ISeqOutStream *outStream, ISeqInStream *inS if (inPos == inSize) { inSize = IN_BUF_SIZE; - RINOK(inStream->Read(inStream, inBuf, &inSize)); + RINOK(inStream->Read(inStream, inBuf, &inSize)) inPos = 0; } { @@ -101,7 +155,8 @@ static SRes Decode2(CLzmaDec *state, ISeqOutStream *outStream, ISeqInStream *inS } } -static SRes Decode(ISeqOutStream *outStream, ISeqInStream *inStream) + +static SRes Decode(ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream) { UInt64 unpackSize; int i; @@ -114,27 +169,29 @@ static SRes Decode(ISeqOutStream *outStream, ISeqInStream *inStream) /* Read and parse header */ - RINOK(SeqInStream_Read(inStream, header, sizeof(header))); - + { + size_t size = sizeof(header); + RINOK(SeqInStream_ReadMax(inStream, header, &size)) + if (size != sizeof(header)) + return SZ_ERROR_INPUT_EOF; + } unpackSize = 0; for (i = 0; i < 8; i++) unpackSize += (UInt64)header[LZMA_PROPS_SIZE + i] << (i * 8); - LzmaDec_Construct(&state); - RINOK(LzmaDec_Allocate(&state, header, LZMA_PROPS_SIZE, &g_Alloc)); + LzmaDec_CONSTRUCT(&state) + RINOK(LzmaDec_Allocate(&state, header, LZMA_PROPS_SIZE, &g_Alloc)) res = Decode2(&state, outStream, inStream, unpackSize); LzmaDec_Free(&state, &g_Alloc); return res; } -static SRes Encode(ISeqOutStream *outStream, ISeqInStream *inStream, UInt64 fileSize, char *rs) +static SRes Encode(ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, UInt64 fileSize) { CLzmaEncHandle enc; SRes res; CLzmaEncProps props; - UNUSED_VAR(rs); - enc = LzmaEnc_Create(&g_Alloc); if (enc == 0) return SZ_ERROR_MEM; @@ -163,63 +220,77 @@ static SRes Encode(ISeqOutStream *outStream, ISeqInStream *inStream, UInt64 file return res; } -int main2(int numArgs, const char *args[], char *rs) + +int Z7_CDECL main(int numArgs, const char *args[]) { CFileSeqInStream inStream; CFileOutStream outStream; char c; int res; int encodeMode; - Bool useOutFile = False; + BoolInt useOutFile = False; + + LzFindPrepare(); FileSeqInStream_CreateVTable(&inStream); File_Construct(&inStream.file); + inStream.wres = 0; FileOutStream_CreateVTable(&outStream); File_Construct(&outStream.file); + outStream.wres = 0; if (numArgs == 1) { - PrintHelp(rs); + PrintHelp(); return 0; } if (numArgs < 3 || numArgs > 4 || strlen(args[1]) != 1) - return PrintUserError(rs); + return PrintUserError(); c = args[1][0]; encodeMode = (c == 'e' || c == 'E'); if (!encodeMode && c != 'd' && c != 'D') - return PrintUserError(rs); + return PrintUserError(); + /* { size_t t4 = sizeof(UInt32); size_t t8 = sizeof(UInt64); if (t4 != 4 || t8 != 8) - return PrintError(rs, "Incorrect UInt32 or UInt64"); + return PrintError("Incorrect UInt32 or UInt64"); } + */ - if (InFile_Open(&inStream.file, args[2]) != 0) - return PrintError(rs, "Can not open input file"); + { + const WRes wres = InFile_Open(&inStream.file, args[2]); + if (wres != 0) + return PrintError_WRes("Cannot open input file", wres); + } if (numArgs > 3) { + WRes wres; useOutFile = True; - if (OutFile_Open(&outStream.file, args[3]) != 0) - return PrintError(rs, "Can not open output file"); + wres = OutFile_Open(&outStream.file, args[3]); + if (wres != 0) + return PrintError_WRes("Cannot open output file", wres); } else if (encodeMode) - PrintUserError(rs); + PrintUserError(); if (encodeMode) { UInt64 fileSize; - File_GetLength(&inStream.file, &fileSize); - res = Encode(&outStream.s, &inStream.s, fileSize, rs); + const WRes wres = File_GetLength(&inStream.file, &fileSize); + if (wres != 0) + return PrintError_WRes("Cannot get file length", wres); + res = Encode(&outStream.vt, &inStream.vt, fileSize); } else { - res = Decode(&outStream.s, useOutFile ? &inStream.s : NULL); + res = Decode(&outStream.vt, useOutFile ? &inStream.vt : NULL); } if (useOutFile) @@ -229,22 +300,14 @@ int main2(int numArgs, const char *args[], char *rs) if (res != SZ_OK) { if (res == SZ_ERROR_MEM) - return PrintError(rs, kCantAllocateMessage); + return PrintError(kCantAllocateMessage); else if (res == SZ_ERROR_DATA) - return PrintError(rs, kDataErrorMessage); + return PrintError(kDataErrorMessage); else if (res == SZ_ERROR_WRITE) - return PrintError(rs, kCantWriteMessage); + return PrintError_WRes(kCantWriteMessage, outStream.wres); else if (res == SZ_ERROR_READ) - return PrintError(rs, kCantReadMessage); - return PrintErrorNumber(rs, res); + return PrintError_WRes(kCantReadMessage, inStream.wres); + return PrintErrorNumber(res); } return 0; } - -int MY_CDECL main(int numArgs, const char *args[]) -{ - char rs[800] = { 0 }; - int res = main2(numArgs, args, rs); - fputs(rs, stdout); - return res; -} diff --git a/3rdparty/lzma/C/Util/Lzma/LzmaUtil.dsp b/3rdparty/lzma/C/Util/Lzma/LzmaUtil.dsp index 43b8e176b22..e2e7d428296 100644 --- a/3rdparty/lzma/C/Util/Lzma/LzmaUtil.dsp +++ b/3rdparty/lzma/C/Util/Lzma/LzmaUtil.dsp @@ -98,10 +98,18 @@ SOURCE=..\..\7zStream.c # End Source File # Begin Source File +SOURCE=..\..\7zTypes.h +# End Source File +# Begin Source File + SOURCE=..\..\7zVersion.h # End Source File # Begin Source File +SOURCE=..\..\7zWindows.h +# End Source File +# Begin Source File + SOURCE=..\..\Alloc.c # End Source File # Begin Source File @@ -110,6 +118,10 @@ SOURCE=..\..\Alloc.h # End Source File # Begin Source File +SOURCE=..\..\Compiler.h +# End Source File +# Begin Source File + SOURCE=..\..\CpuArch.h # End Source File # Begin Source File @@ -130,6 +142,10 @@ SOURCE=..\..\LzFindMt.h # End Source File # Begin Source File +SOURCE=..\..\LzFindOpt.c +# End Source File +# Begin Source File + SOURCE=..\..\LzHash.h # End Source File # Begin Source File @@ -154,15 +170,19 @@ SOURCE=.\LzmaUtil.c # End Source File # Begin Source File -SOURCE=..\..\Threads.c +SOURCE=..\..\Precomp.h # End Source File # Begin Source File -SOURCE=..\..\Threads.h +SOURCE=.\Precomp.h # End Source File # Begin Source File -SOURCE=..\..\Types.h +SOURCE=..\..\Threads.c +# End Source File +# Begin Source File + +SOURCE=..\..\Threads.h # End Source File # End Target # End Project diff --git a/3rdparty/lzma/C/Util/Lzma/Precomp.h b/3rdparty/lzma/C/Util/Lzma/Precomp.h new file mode 100644 index 00000000000..bc8fa219b58 --- /dev/null +++ b/3rdparty/lzma/C/Util/Lzma/Precomp.h @@ -0,0 +1,14 @@ +/* Precomp.h -- StdAfx +2023-03-04 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_PRECOMP_H +#define ZIP7_INC_PRECOMP_H + +#if defined(_MSC_VER) && _MSC_VER >= 1800 +#pragma warning(disable : 4464) // relative include path contains '..' +#endif + +#include "../../Compiler.h" +#include "../../7zTypes.h" + +#endif diff --git a/3rdparty/lzma/C/Util/Lzma/makefile b/3rdparty/lzma/C/Util/Lzma/makefile index 4795322786b..7813bdb059d 100644 --- a/3rdparty/lzma/C/Util/Lzma/makefile +++ b/3rdparty/lzma/C/Util/Lzma/makefile @@ -8,8 +8,10 @@ LIB_OBJS = \ C_OBJS = \ $O\Alloc.obj \ + $O\CpuArch.obj \ $O\LzFind.obj \ $O\LzFindMt.obj \ + $O\LzFindOpt.obj \ $O\LzmaDec.obj \ $O\LzmaEnc.obj \ $O\7zFile.obj \ diff --git a/3rdparty/lzma/C/Util/Lzma/makefile.gcc b/3rdparty/lzma/C/Util/Lzma/makefile.gcc index 67aa8b17957..2acb0b808d2 100644 --- a/3rdparty/lzma/C/Util/Lzma/makefile.gcc +++ b/3rdparty/lzma/C/Util/Lzma/makefile.gcc @@ -1,44 +1,21 @@ -PROG = lzma -CXX = g++ -LIB = -RM = rm -f -CFLAGS = -c -O2 -Wall -D_7ZIP_ST +PROG = 7lzma -OBJS = \ - LzmaUtil.o \ - Alloc.o \ - LzFind.o \ - LzmaDec.o \ - LzmaEnc.o \ - 7zFile.o \ - 7zStream.o \ - - -all: $(PROG) - -$(PROG): $(OBJS) - $(CXX) -o $(PROG) $(LDFLAGS) $(OBJS) $(LIB) $(LIB2) - -LzmaUtil.o: LzmaUtil.c - $(CXX) $(CFLAGS) LzmaUtil.c - -Alloc.o: ../../Alloc.c - $(CXX) $(CFLAGS) ../../Alloc.c +include ../../../CPP/7zip/LzmaDec_gcc.mak -LzFind.o: ../../LzFind.c - $(CXX) $(CFLAGS) ../../LzFind.c -LzmaDec.o: ../../LzmaDec.c - $(CXX) $(CFLAGS) ../../LzmaDec.c - -LzmaEnc.o: ../../LzmaEnc.c - $(CXX) $(CFLAGS) ../../LzmaEnc.c - -7zFile.o: ../../7zFile.c - $(CXX) $(CFLAGS) ../../7zFile.c - -7zStream.o: ../../7zStream.c - $(CXX) $(CFLAGS) ../../7zStream.c - -clean: - -$(RM) $(PROG) $(OBJS) +OBJS = \ + $(LZMA_DEC_OPT_OBJS) \ + $O/7zFile.o \ + $O/7zStream.o \ + $O/Alloc.o \ + $O/CpuArch.o \ + $O/LzFind.o \ + $O/LzFindMt.o \ + $O/LzFindOpt.o \ + $O/LzmaDec.o \ + $O/LzmaEnc.o \ + $O/LzmaUtil.o \ + $O/Threads.o \ + + +include ../../7zip_gcc_c.mak diff --git a/3rdparty/lzma/C/Util/LzmaLib/LzmaLib.dsp b/3rdparty/lzma/C/Util/LzmaLib/LzmaLib.dsp index 3421de83a1d..bacd9679324 100644 --- a/3rdparty/lzma/C/Util/LzmaLib/LzmaLib.dsp +++ b/3rdparty/lzma/C/Util/LzmaLib/LzmaLib.dsp @@ -101,6 +101,10 @@ SOURCE=.\LzmaLib.def SOURCE=.\LzmaLibExports.c # End Source File +# Begin Source File + +SOURCE=.\Precomp.h +# End Source File # End Group # Begin Source File @@ -108,6 +112,10 @@ SOURCE=..\..\7zTypes.h # End Source File # Begin Source File +SOURCE=..\..\7zWindows.h +# End Source File +# Begin Source File + SOURCE=..\..\Alloc.c # End Source File # Begin Source File @@ -116,6 +124,14 @@ SOURCE=..\..\Alloc.h # End Source File # Begin Source File +SOURCE=..\..\Compiler.h +# End Source File +# Begin Source File + +SOURCE=..\..\CpuArch.h +# End Source File +# Begin Source File + SOURCE=..\..\IStream.h # End Source File # Begin Source File @@ -136,6 +152,10 @@ SOURCE=..\..\LzFindMt.h # End Source File # Begin Source File +SOURCE=..\..\LzFindOpt.c +# End Source File +# Begin Source File + SOURCE=..\..\LzHash.h # End Source File # Begin Source File @@ -164,6 +184,10 @@ SOURCE=..\..\LzmaLib.h # End Source File # Begin Source File +SOURCE=..\..\Precomp.h +# End Source File +# Begin Source File + SOURCE=.\resource.rc # End Source File # Begin Source File diff --git a/3rdparty/lzma/C/Util/LzmaLib/LzmaLibExports.c b/3rdparty/lzma/C/Util/LzmaLib/LzmaLibExports.c index 4a28a9a6b34..a46c9a80732 100644 --- a/3rdparty/lzma/C/Util/LzmaLib/LzmaLibExports.c +++ b/3rdparty/lzma/C/Util/LzmaLib/LzmaLibExports.c @@ -1,14 +1,15 @@ /* LzmaLibExports.c -- LZMA library DLL Entry point -2015-11-08 : Igor Pavlov : Public domain */ +2023-03-05 : Igor Pavlov : Public domain */ -#include "../../Precomp.h" +#include "Precomp.h" -#include <windows.h> +#include "../../7zWindows.h" +BOOL WINAPI DllMain(HINSTANCE hInstance, DWORD dwReason, LPVOID lpReserved); BOOL WINAPI DllMain(HINSTANCE hInstance, DWORD dwReason, LPVOID lpReserved) { - UNUSED_VAR(hInstance); - UNUSED_VAR(dwReason); - UNUSED_VAR(lpReserved); + UNUSED_VAR(hInstance) + UNUSED_VAR(dwReason) + UNUSED_VAR(lpReserved) return TRUE; } diff --git a/3rdparty/lzma/C/Util/LzmaLib/Precomp.c b/3rdparty/lzma/C/Util/LzmaLib/Precomp.c new file mode 100644 index 00000000000..01605e3c261 --- /dev/null +++ b/3rdparty/lzma/C/Util/LzmaLib/Precomp.c @@ -0,0 +1,4 @@ +/* Precomp.c -- StdAfx +2013-01-21 : Igor Pavlov : Public domain */ + +#include "Precomp.h" diff --git a/3rdparty/lzma/C/Util/LzmaLib/Precomp.h b/3rdparty/lzma/C/Util/LzmaLib/Precomp.h new file mode 100644 index 00000000000..bc8fa219b58 --- /dev/null +++ b/3rdparty/lzma/C/Util/LzmaLib/Precomp.h @@ -0,0 +1,14 @@ +/* Precomp.h -- StdAfx +2023-03-04 : Igor Pavlov : Public domain */ + +#ifndef ZIP7_INC_PRECOMP_H +#define ZIP7_INC_PRECOMP_H + +#if defined(_MSC_VER) && _MSC_VER >= 1800 +#pragma warning(disable : 4464) // relative include path contains '..' +#endif + +#include "../../Compiler.h" +#include "../../7zTypes.h" + +#endif diff --git a/3rdparty/lzma/C/Util/LzmaLib/makefile b/3rdparty/lzma/C/Util/LzmaLib/makefile index 74103bb0540..b8e054efca3 100644 --- a/3rdparty/lzma/C/Util/LzmaLib/makefile +++ b/3rdparty/lzma/C/Util/LzmaLib/makefile @@ -11,14 +11,17 @@ LIB_OBJS = \ C_OBJS = \ $O\Alloc.obj \ + $O\CpuArch.obj \ $O\LzFind.obj \ $O\LzFindMt.obj \ + $O\LzFindOpt.obj \ $O\LzmaDec.obj \ $O\LzmaEnc.obj \ $O\LzmaLib.obj \ $O\Threads.obj \ OBJS = \ + $O\Precomp.obj \ $(LIB_OBJS) \ $(C_OBJS) \ $O\resource.res @@ -28,7 +31,24 @@ OBJS = \ $(SLIBPATH): $O $(OBJS) lib -out:$(SLIBPATH) $(OBJS) $(LIBS) + +MAK_SINGLE_FILE = 1 + +$O\Precomp.obj: Precomp.c + $(CCOMPL_PCH) + +!IFDEF MAK_SINGLE_FILE + $(LIB_OBJS): $(*B).c - $(COMPL_O2) + $(CCOMPL_USE) $(C_OBJS): ../../$(*B).c - $(COMPL_O2) + $(CCOMPL_USE) + +!ELSE + +{.}.c{$O}.obj:: + $(CCOMPLB_USE) +{../../../C}.c{$O}.obj:: + $(CCOMPLB_USE) + +!ENDIF diff --git a/3rdparty/lzma/C/Util/SfxSetup/Precomp.h b/3rdparty/lzma/C/Util/SfxSetup/Precomp.h index 588a66f7eb8..bc8fa219b58 100644 --- a/3rdparty/lzma/C/Util/SfxSetup/Precomp.h +++ b/3rdparty/lzma/C/Util/SfxSetup/Precomp.h @@ -1,8 +1,12 @@ /* Precomp.h -- StdAfx -2013-06-16 : Igor Pavlov : Public domain */ +2023-03-04 : Igor Pavlov : Public domain */ -#ifndef __7Z_PRECOMP_H -#define __7Z_PRECOMP_H +#ifndef ZIP7_INC_PRECOMP_H +#define ZIP7_INC_PRECOMP_H + +#if defined(_MSC_VER) && _MSC_VER >= 1800 +#pragma warning(disable : 4464) // relative include path contains '..' +#endif #include "../../Compiler.h" #include "../../7zTypes.h" diff --git a/3rdparty/lzma/C/Util/SfxSetup/SfxSetup.c b/3rdparty/lzma/C/Util/SfxSetup/SfxSetup.c index 57e73d86528..7304a0be40d 100644 --- a/3rdparty/lzma/C/Util/SfxSetup/SfxSetup.c +++ b/3rdparty/lzma/C/Util/SfxSetup/SfxSetup.c @@ -1,5 +1,5 @@ /* SfxSetup.c - 7z SFX Setup -2016-05-16 : Igor Pavlov : Public domain */ +2019-02-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -24,6 +24,14 @@ #define k_EXE_ExtIndex 2 +#define kInputBufSize ((size_t)1 << 18) + + +#define wcscat lstrcatW +#define wcslen (size_t)lstrlenW +#define wcscpy lstrcpyW +// wcsncpy() and lstrcpynW() work differently. We don't use them. + static const char * const kExts[] = { "bat" @@ -62,7 +70,7 @@ static unsigned FindExt(const wchar_t *s, unsigned *extLen) return len; } -#define MAKE_CHAR_UPPER(c) ((((c) >= 'a' && (c) <= 'z') ? (c) -= 0x20 : (c))) +#define MAKE_CHAR_UPPER(c) ((((c) >= 'a' && (c) <= 'z') ? (c) - 0x20 : (c))) static unsigned FindItem(const char * const *items, unsigned num, const wchar_t *s, unsigned len) { @@ -70,13 +78,13 @@ static unsigned FindItem(const char * const *items, unsigned num, const wchar_t for (i = 0; i < num; i++) { const char *item = items[i]; - unsigned itemLen = (unsigned)strlen(item); + const unsigned itemLen = (unsigned)strlen(item); unsigned j; if (len != itemLen) continue; for (j = 0; j < len; j++) { - unsigned c = (Byte)item[j]; + const unsigned c = (Byte)item[j]; if (c != s[j] && MAKE_CHAR_UPPER(c) != s[j]) break; } @@ -94,10 +102,20 @@ static BOOL WINAPI HandlerRoutine(DWORD ctrlType) } #endif + +#ifdef _CONSOLE +static void PrintStr(const char *s) +{ + fputs(s, stdout); +} +#endif + static void PrintErrorMessage(const char *message) { #ifdef _CONSOLE - printf("\n7-Zip Error: %s\n", message); + PrintStr("\n7-Zip Error: "); + PrintStr(message); + PrintStr("\n"); #else #ifdef UNDER_CE WCHAR messageW[256 + 4]; @@ -125,7 +143,7 @@ static WRes MyCreateDir(const WCHAR *name) #define kSignatureSearchLimit (1 << 22) -static Bool FindSignature(CSzFile *stream, UInt64 *resPos) +static BoolInt FindSignature(CSzFile *stream, UInt64 *resPos) { Byte buf[kBufferSize]; size_t numPrevBytes = 0; @@ -161,7 +179,7 @@ static Bool FindSignature(CSzFile *stream, UInt64 *resPos) } } -static Bool DoesFileOrDirExist(const WCHAR *path) +static BoolInt DoesFileOrDirExist(const WCHAR *path) { WIN32_FIND_DATAW fd; HANDLE handle; @@ -177,7 +195,7 @@ static WRes RemoveDirWithSubItems(WCHAR *path) WIN32_FIND_DATAW fd; HANDLE handle; WRes res = 0; - size_t len = wcslen(path); + const size_t len = wcslen(path); wcscpy(path + len, L"*"); handle = FindFirstFileW(path, &fd); path[len] = L'\0'; @@ -226,7 +244,7 @@ static WRes RemoveDirWithSubItems(WCHAR *path) } #ifdef _CONSOLE -int MY_CDECL main() +int Z7_CDECL main(void) #else int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, #ifdef UNDER_CE @@ -238,7 +256,7 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, #endif { CFileInStream archiveStream; - CLookToRead lookStream; + CLookToRead2 lookStream; CSzArEx db; SRes res = SZ_OK; ISzAlloc allocImp; @@ -252,7 +270,7 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, DWORD winRes; const wchar_t *cmdLineParams; const char *errorMessage = NULL; - Bool useShellExecute = True; + BoolInt useShellExecute = True; DWORD exitCode = 0; LoadSecurityDlls(); @@ -275,7 +293,8 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, allocTempImp.Free = SzFreeTemp; FileInStream_CreateVTable(&archiveStream); - LookToRead_CreateVTable(&lookStream, False); + LookToRead2_CreateVTable(&lookStream, False); + lookStream.buf = NULL; winRes = GetModuleFileNameW(NULL, sfxPath, MAX_PATH); if (winRes == 0 || winRes > MAX_PATH) @@ -284,10 +303,10 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, cmdLineParams = GetCommandLineW(); #ifndef UNDER_CE { - Bool quoteMode = False; + BoolInt quoteMode = False; for (;; cmdLineParams++) { - wchar_t c = *cmdLineParams; + const wchar_t c = *cmdLineParams; if (c == L'\"') quoteMode = !quoteMode; else if (c == 0 || (c == L' ' && !quoteMode)) @@ -321,7 +340,7 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, unsigned k; for (k = 0; k < 8; k++) { - unsigned t = value & 0xF; + const unsigned t = value & 0xF; value >>= 4; s[7 - k] = (wchar_t)((t < 10) ? ('0' + t) : ('A' + (t - 10))); } @@ -376,14 +395,22 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, if (res == SZ_OK) { - lookStream.realStream = &archiveStream.s; - LookToRead_Init(&lookStream); + lookStream.buf = (Byte *)ISzAlloc_Alloc(&allocImp, kInputBufSize); + if (!lookStream.buf) + res = SZ_ERROR_MEM; + else + { + lookStream.bufSize = kInputBufSize; + lookStream.realStream = &archiveStream.vt; + LookToRead2_INIT(&lookStream) + } } SzArEx_Init(&db); + if (res == SZ_OK) { - res = SzArEx_Open(&db, &lookStream.s, &allocImp, &allocTempImp); + res = SzArEx_Open(&db, &lookStream.vt, &allocImp, &allocTempImp); } if (res == SZ_OK) @@ -409,9 +436,9 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, temp = path + pathLen; - SzArEx_GetFileNameUtf16(&db, i, temp); + SzArEx_GetFileNameUtf16(&db, i, (UInt16 *)temp); { - res = SzArEx_Extract(&db, &lookStream.s, i, + res = SzArEx_Extract(&db, &lookStream.vt, i, &blockIndex, &outBuffer, &outBufferSize, &offset, &outSizeProcessed, &allocImp, &allocTempImp); @@ -444,11 +471,11 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, unsigned extLen; const WCHAR *name = temp + nameStartPos; unsigned len = (unsigned)wcslen(name); - unsigned nameLen = FindExt(temp + nameStartPos, &extLen); - unsigned extPrice = FindItem(kExts, sizeof(kExts) / sizeof(kExts[0]), name + len - extLen, extLen); - unsigned namePrice = FindItem(kNames, sizeof(kNames) / sizeof(kNames[0]), name, nameLen); + const unsigned nameLen = FindExt(temp + nameStartPos, &extLen); + const unsigned extPrice = FindItem(kExts, sizeof(kExts) / sizeof(kExts[0]), name + len - extLen, extLen); + const unsigned namePrice = FindItem(kNames, sizeof(kNames) / sizeof(kNames[0]), name, nameLen); - unsigned price = namePrice + extPrice * 64 + (nameStartPos == 0 ? 0 : (1 << 12)); + const unsigned price = namePrice + extPrice * 64 + (nameStartPos == 0 ? 0 : (1 << 12)); if (minPrice > price) { minPrice = price; @@ -489,7 +516,7 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, #endif { - SRes res2 = File_Close(&outFile); + const SRes res2 = File_Close(&outFile); if (res != SZ_OK) break; if (res2 != SZ_OK) @@ -516,16 +543,19 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, { WCHAR *temp = path + pathLen; UInt32 j; - SzArEx_GetFileNameUtf16(&db, executeFileIndex, temp); + SzArEx_GetFileNameUtf16(&db, executeFileIndex, (UInt16 *)temp); for (j = 0; temp[j] != 0; j++) if (temp[j] == '/') temp[j] = CHAR_PATH_SEPARATOR; } } - IAlloc_Free(&allocImp, outBuffer); + ISzAlloc_Free(&allocImp, outBuffer); } + SzArEx_Free(&db, &allocImp); + ISzAlloc_Free(&allocImp, lookStream.buf); + File_Close(&archiveStream.file); if (res == SZ_OK) @@ -536,7 +566,7 @@ int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, WCHAR oldCurDir[MAX_PATH + 2]; oldCurDir[0] = 0; { - DWORD needLen = GetCurrentDirectory(MAX_PATH + 1, oldCurDir); + const DWORD needLen = GetCurrentDirectory(MAX_PATH + 1, oldCurDir); if (needLen == 0 || needLen > MAX_PATH) oldCurDir[0] = 0; SetCurrentDirectory(workCurDir); diff --git a/3rdparty/lzma/C/Util/SfxSetup/makefile b/3rdparty/lzma/C/Util/SfxSetup/makefile index 76cd9112464..bc0cf8bc71c 100644 --- a/3rdparty/lzma/C/Util/SfxSetup/makefile +++ b/3rdparty/lzma/C/Util/SfxSetup/makefile @@ -1,4 +1,8 @@ PROG = 7zS2.sfx +MY_FIXED = 1 + +CFLAGS = $(CFLAGS) \ + -DZ7_EXTRACT_ONLY \ C_OBJS = \ $O\7zAlloc.obj \ diff --git a/3rdparty/lzma/C/Util/SfxSetup/makefile_con b/3rdparty/lzma/C/Util/SfxSetup/makefile_con index ed9174d3b20..9f4b91669ca 100644 --- a/3rdparty/lzma/C/Util/SfxSetup/makefile_con +++ b/3rdparty/lzma/C/Util/SfxSetup/makefile_con @@ -1,5 +1,8 @@ PROG = 7zS2con.sfx -CFLAGS = $(CFLAGS) -D_CONSOLE +MY_FIXED = 1 + +CFLAGS = $(CFLAGS) -D_CONSOLE \ + -DZ7_EXTRACT_ONLY \ C_OBJS = \ $O\7zAlloc.obj \ diff --git a/3rdparty/lzma/C/Xz.c b/3rdparty/lzma/C/Xz.c index 1ef6878252c..4ad071060a7 100644 --- a/3rdparty/lzma/C/Xz.c +++ b/3rdparty/lzma/C/Xz.c @@ -1,5 +1,5 @@ /* Xz.c - Xz -2015-05-01 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -9,7 +9,7 @@ #include "XzCrc64.h" const Byte XZ_SIG[XZ_SIG_SIZE] = { 0xFD, '7', 'z', 'X', 'Z', 0 }; -const Byte XZ_FOOTER_SIG[XZ_FOOTER_SIG_SIZE] = { 'Y', 'Z' }; +/* const Byte XZ_FOOTER_SIG[XZ_FOOTER_SIG_SIZE] = { 'Y', 'Z' }; */ unsigned Xz_WriteVarInt(Byte *buf, UInt64 v) { @@ -20,28 +20,28 @@ unsigned Xz_WriteVarInt(Byte *buf, UInt64 v) v >>= 7; } while (v != 0); - buf[i - 1] &= 0x7F; + buf[(size_t)i - 1] &= 0x7F; return i; } void Xz_Construct(CXzStream *p) { - p->numBlocks = p->numBlocksAllocated = 0; - p->blocks = 0; + p->numBlocks = 0; + p->blocks = NULL; p->flags = 0; } -void Xz_Free(CXzStream *p, ISzAlloc *alloc) +void Xz_Free(CXzStream *p, ISzAllocPtr alloc) { - alloc->Free(alloc, p->blocks); - p->numBlocks = p->numBlocksAllocated = 0; - p->blocks = 0; + ISzAlloc_Free(alloc, p->blocks); + p->numBlocks = 0; + p->blocks = NULL; } unsigned XzFlags_GetCheckSize(CXzStreamFlags f) { unsigned t = XzFlags_GetCheckType(f); - return (t == 0) ? 0 : (4 << ((t - 1) / 3)); + return (t == 0) ? 0 : ((unsigned)4 << ((t - 1) / 3)); } void XzCheck_Init(CXzCheck *p, unsigned mode) @@ -70,7 +70,7 @@ int XzCheck_Final(CXzCheck *p, Byte *digest) switch (p->mode) { case XZ_CHECK_CRC32: - SetUi32(digest, CRC_GET_DIGEST(p->crc)); + SetUi32(digest, CRC_GET_DIGEST(p->crc)) break; case XZ_CHECK_CRC64: { diff --git a/3rdparty/lzma/C/Xz.h b/3rdparty/lzma/C/Xz.h index be3a1c36220..d5001f6cace 100644 --- a/3rdparty/lzma/C/Xz.h +++ b/3rdparty/lzma/C/Xz.h @@ -1,21 +1,23 @@ /* Xz.h - Xz interface -2015-05-01 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ -#ifndef __XZ_H -#define __XZ_H +#ifndef ZIP7_INC_XZ_H +#define ZIP7_INC_XZ_H #include "Sha256.h" +#include "Delta.h" EXTERN_C_BEGIN #define XZ_ID_Subblock 1 #define XZ_ID_Delta 3 -#define XZ_ID_X86 4 -#define XZ_ID_PPC 5 -#define XZ_ID_IA64 6 -#define XZ_ID_ARM 7 -#define XZ_ID_ARMT 8 +#define XZ_ID_X86 4 +#define XZ_ID_PPC 5 +#define XZ_ID_IA64 6 +#define XZ_ID_ARM 7 +#define XZ_ID_ARMT 8 #define XZ_ID_SPARC 9 +#define XZ_ID_ARM64 0xa #define XZ_ID_LZMA2 0x21 unsigned Xz_ReadVarInt(const Byte *p, size_t maxSize, UInt64 *value); @@ -47,12 +49,13 @@ typedef struct CXzFilter filters[XZ_NUM_FILTERS_MAX]; } CXzBlock; -#define XzBlock_GetNumFilters(p) (((p)->flags & XZ_BF_NUM_FILTERS_MASK) + 1) +#define XzBlock_GetNumFilters(p) (((unsigned)(p)->flags & XZ_BF_NUM_FILTERS_MASK) + 1) #define XzBlock_HasPackSize(p) (((p)->flags & XZ_BF_PACK_SIZE) != 0) #define XzBlock_HasUnpackSize(p) (((p)->flags & XZ_BF_UNPACK_SIZE) != 0) +#define XzBlock_HasUnsupportedFlags(p) (((p)->flags & ~(XZ_BF_NUM_FILTERS_MASK | XZ_BF_PACK_SIZE | XZ_BF_UNPACK_SIZE)) != 0) SRes XzBlock_Parse(CXzBlock *p, const Byte *header); -SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStream *inStream, Bool *isIndex, UInt32 *headerSizeRes); +SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStreamPtr inStream, BoolInt *isIndex, UInt32 *headerSizeRes); /* ---------- xz stream ---------- */ @@ -60,7 +63,13 @@ SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStream *inStream, Bool *isIndex, UInt #define XZ_FOOTER_SIG_SIZE 2 extern const Byte XZ_SIG[XZ_SIG_SIZE]; + +/* extern const Byte XZ_FOOTER_SIG[XZ_FOOTER_SIG_SIZE]; +*/ + +#define XZ_FOOTER_SIG_0 'Y' +#define XZ_FOOTER_SIG_1 'Z' #define XZ_STREAM_FLAGS_SIZE 2 #define XZ_STREAM_CRC_SIZE 4 @@ -94,7 +103,7 @@ typedef UInt16 CXzStreamFlags; unsigned XzFlags_GetCheckSize(CXzStreamFlags f); SRes Xz_ParseHeader(CXzStreamFlags *p, const Byte *buf); -SRes Xz_ReadHeader(CXzStreamFlags *p, ISeqInStream *inStream); +SRes Xz_ReadHeader(CXzStreamFlags *p, ISeqInStreamPtr inStream); typedef struct { @@ -105,14 +114,14 @@ typedef struct typedef struct { CXzStreamFlags flags; + // Byte _pad[6]; size_t numBlocks; - size_t numBlocksAllocated; CXzBlockSizes *blocks; UInt64 startOffset; } CXzStream; void Xz_Construct(CXzStream *p); -void Xz_Free(CXzStream *p, ISzAlloc *alloc); +void Xz_Free(CXzStream *p, ISzAllocPtr alloc); #define XZ_SIZE_OVERFLOW ((UInt64)(Int64)-1) @@ -127,12 +136,15 @@ typedef struct } CXzs; void Xzs_Construct(CXzs *p); -void Xzs_Free(CXzs *p, ISzAlloc *alloc); -SRes Xzs_ReadBackward(CXzs *p, ILookInStream *inStream, Int64 *startOffset, ICompressProgress *progress, ISzAlloc *alloc); +void Xzs_Free(CXzs *p, ISzAllocPtr alloc); +SRes Xzs_ReadBackward(CXzs *p, ILookInStreamPtr inStream, Int64 *startOffset, ICompressProgressPtr progress, ISzAllocPtr alloc); UInt64 Xzs_GetNumBlocks(const CXzs *p); UInt64 Xzs_GetUnpackSize(const CXzs *p); + +// ECoderStatus values are identical to ELzmaStatus values of LZMA2 decoder + typedef enum { CODER_STATUS_NOT_SPECIFIED, /* use main error code instead */ @@ -141,43 +153,69 @@ typedef enum CODER_STATUS_NEEDS_MORE_INPUT /* you must provide more input bytes */ } ECoderStatus; + +// ECoderFinishMode values are identical to ELzmaFinishMode + typedef enum { CODER_FINISH_ANY, /* finish at any point */ CODER_FINISH_END /* block must be finished at the end */ } ECoderFinishMode; -typedef struct _IStateCoder + +typedef struct { - void *p; - void (*Free)(void *p, ISzAlloc *alloc); - SRes (*SetProps)(void *p, const Byte *props, size_t propSize, ISzAlloc *alloc); + void *p; // state object; + void (*Free)(void *p, ISzAllocPtr alloc); + SRes (*SetProps)(void *p, const Byte *props, size_t propSize, ISzAllocPtr alloc); void (*Init)(void *p); - SRes (*Code)(void *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - int srcWasFinished, ECoderFinishMode finishMode, int *wasFinished); + SRes (*Code2)(void *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, + int srcWasFinished, ECoderFinishMode finishMode, + // int *wasFinished, + ECoderStatus *status); + SizeT (*Filter)(void *p, Byte *data, SizeT size); } IStateCoder; + +typedef struct +{ + UInt32 methodId; + UInt32 delta; + UInt32 ip; + UInt32 X86_State; + Byte delta_State[DELTA_STATE_SIZE]; +} CXzBcFilterStateBase; + +typedef SizeT (*Xz_Func_BcFilterStateBase_Filter)(CXzBcFilterStateBase *p, Byte *data, SizeT size); + +SRes Xz_StateCoder_Bc_SetFromMethod_Func(IStateCoder *p, UInt64 id, + Xz_Func_BcFilterStateBase_Filter func, ISzAllocPtr alloc); + + #define MIXCODER_NUM_FILTERS_MAX 4 typedef struct { - ISzAlloc *alloc; + ISzAllocPtr alloc; Byte *buf; unsigned numCoders; + + Byte *outBuf; + size_t outBufSize; + size_t outWritten; // is equal to lzmaDecoder.dicPos (in outBuf mode) + BoolInt wasFinished; + SRes res; + ECoderStatus status; + // BoolInt SingleBufMode; + int finished[MIXCODER_NUM_FILTERS_MAX - 1]; size_t pos[MIXCODER_NUM_FILTERS_MAX - 1]; size_t size[MIXCODER_NUM_FILTERS_MAX - 1]; UInt64 ids[MIXCODER_NUM_FILTERS_MAX]; + SRes results[MIXCODER_NUM_FILTERS_MAX]; IStateCoder coders[MIXCODER_NUM_FILTERS_MAX]; } CMixCoder; -void MixCoder_Construct(CMixCoder *p, ISzAlloc *alloc); -void MixCoder_Free(CMixCoder *p); -void MixCoder_Init(CMixCoder *p); -SRes MixCoder_SetFromMethod(CMixCoder *p, unsigned coderIndex, UInt64 methodId); -SRes MixCoder_Code(CMixCoder *p, Byte *dest, SizeT *destLen, - const Byte *src, SizeT *srcLen, int srcWasFinished, - ECoderFinishMode finishMode, ECoderStatus *status); typedef enum { @@ -191,6 +229,7 @@ typedef enum XZ_STATE_BLOCK_FOOTER } EXzState; + typedef struct { EXzState state; @@ -204,7 +243,7 @@ typedef struct UInt64 packSize; UInt64 unpackSize; - UInt64 numBlocks; + UInt64 numBlocks; // number of finished blocks in current stream UInt64 indexSize; UInt64 indexPos; UInt64 padSize; @@ -218,15 +257,73 @@ typedef struct CXzBlock block; CXzCheck check; CSha256 sha; + + BoolInt parseMode; + BoolInt headerParsedOk; + BoolInt decodeToStreamSignature; + unsigned decodeOnlyOneBlock; + + Byte *outBuf; + size_t outBufSize; + size_t outDataWritten; // the size of data in (outBuf) that were fully unpacked + Byte shaDigest[SHA256_DIGEST_SIZE]; Byte buf[XZ_BLOCK_HEADER_SIZE_MAX]; } CXzUnpacker; -void XzUnpacker_Construct(CXzUnpacker *p, ISzAlloc *alloc); +/* alloc : aligned for cache line allocation is better */ +void XzUnpacker_Construct(CXzUnpacker *p, ISzAllocPtr alloc); void XzUnpacker_Init(CXzUnpacker *p); +void XzUnpacker_SetOutBuf(CXzUnpacker *p, Byte *outBuf, size_t outBufSize); void XzUnpacker_Free(CXzUnpacker *p); /* + XzUnpacker + The sequence for decoding functions: + { + XzUnpacker_Construct() + [Decoding_Calls] + XzUnpacker_Free() + } + + [Decoding_Calls] + + There are 3 types of interfaces for [Decoding_Calls] calls: + + Interface-1 : Partial output buffers: + { + XzUnpacker_Init() + for() + { + XzUnpacker_Code(); + } + XzUnpacker_IsStreamWasFinished() + } + + Interface-2 : Direct output buffer: + Use it, if you know exact size of decoded data, and you need + whole xz unpacked data in one output buffer. + xz unpacker doesn't allocate additional buffer for lzma2 dictionary in that mode. + { + XzUnpacker_Init() + XzUnpacker_SetOutBufMode(); // to set output buffer and size + for() + { + XzUnpacker_Code(); // (dest = NULL) in XzUnpacker_Code() + } + XzUnpacker_IsStreamWasFinished() + } + + Interface-3 : Direct output buffer : One call full decoding + It unpacks whole input buffer to output buffer in one call. + It uses Interface-2 internally. + { + XzUnpacker_CodeFull() + XzUnpacker_IsStreamWasFinished() + } +*/ + +/* finishMode: It has meaning only if the decoding reaches output limit (*destLen). CODER_FINISH_ANY - use smallest number of input bytes @@ -236,8 +333,12 @@ Returns: SZ_OK status: CODER_STATUS_NOT_FINISHED, - CODER_STATUS_NEEDS_MORE_INPUT - maybe there are more xz streams, - call XzUnpacker_IsStreamWasFinished to check that current stream was finished + CODER_STATUS_NEEDS_MORE_INPUT - the decoder can return it in two cases: + 1) it needs more input data to finish current xz stream + 2) xz stream was finished successfully. But the decoder supports multiple + concatented xz streams. So it expects more input data for new xz streams. + Call XzUnpacker_IsStreamWasFinished() to check that latest xz stream was finished successfully. + SZ_ERROR_MEM - Memory allocation error SZ_ERROR_DATA - Data error SZ_ERROR_UNSUPPORTED - Unsupported method or method properties @@ -255,20 +356,179 @@ Returns: SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, - const Byte *src, SizeT *srcLen, ECoderFinishMode finishMode, - ECoderStatus *status); + const Byte *src, SizeT *srcLen, int srcFinished, + ECoderFinishMode finishMode, ECoderStatus *status); -Bool XzUnpacker_IsStreamWasFinished(CXzUnpacker *p); +SRes XzUnpacker_CodeFull(CXzUnpacker *p, Byte *dest, SizeT *destLen, + const Byte *src, SizeT *srcLen, + ECoderFinishMode finishMode, ECoderStatus *status); /* -Call XzUnpacker_GetExtraSize after XzUnpacker_Code function to detect real size of -xz stream in two cases: -XzUnpacker_Code() returns: +If you decode full xz stream(s), then you can call XzUnpacker_IsStreamWasFinished() +after successful XzUnpacker_CodeFull() or after last call of XzUnpacker_Code(). +*/ + +BoolInt XzUnpacker_IsStreamWasFinished(const CXzUnpacker *p); + +/* +XzUnpacker_GetExtraSize() returns then number of unconfirmed bytes, + if it's in (XZ_STATE_STREAM_HEADER) state or in (XZ_STATE_STREAM_PADDING) state. +These bytes can be some data after xz archive, or +it can be start of new xz stream. + +Call XzUnpacker_GetExtraSize() after XzUnpacker_Code() function to detect real size of +xz stream in two cases, if XzUnpacker_Code() returns: res == SZ_OK && status == CODER_STATUS_NEEDS_MORE_INPUT res == SZ_ERROR_NO_ARCHIVE */ -UInt64 XzUnpacker_GetExtraSize(CXzUnpacker *p); +UInt64 XzUnpacker_GetExtraSize(const CXzUnpacker *p); + + +/* + for random block decoding: + XzUnpacker_Init(); + set CXzUnpacker::streamFlags + XzUnpacker_PrepareToRandomBlockDecoding() + loop + { + XzUnpacker_Code() + XzUnpacker_IsBlockFinished() + } +*/ + +void XzUnpacker_PrepareToRandomBlockDecoding(CXzUnpacker *p); +BoolInt XzUnpacker_IsBlockFinished(const CXzUnpacker *p); + +#define XzUnpacker_GetPackSizeForIndex(p) ((p)->packSize + (p)->blockHeaderSize + XzFlags_GetCheckSize((p)->streamFlags)) + + + + + + +/* ---- Single-Thread and Multi-Thread xz Decoding with Input/Output Streams ---- */ + +/* + if (CXzDecMtProps::numThreads > 1), the decoder can try to use + Multi-Threading. The decoder analyses xz block header, and if + there are pack size and unpack size values stored in xz block header, + the decoder reads compressed data of block to internal buffers, + and then it can start parallel decoding, if there are another blocks. + The decoder can switch back to Single-Thread decoding after some conditions. + + The sequence of calls for xz decoding with in/out Streams: + { + XzDecMt_Create() + XzDecMtProps_Init(XzDecMtProps) to set default values of properties + // then you can change some XzDecMtProps parameters with required values + // here you can set the number of threads and (memUseMax) - the maximum + Memory usage for multithreading decoding. + for() + { + XzDecMt_Decode() // one call per one file + } + XzDecMt_Destroy() + } +*/ + + +typedef struct +{ + size_t inBufSize_ST; // size of input buffer for Single-Thread decoding + size_t outStep_ST; // size of output buffer for Single-Thread decoding + BoolInt ignoreErrors; // if set to 1, the decoder can ignore some errors and it skips broken parts of data. + + #ifndef Z7_ST + unsigned numThreads; // the number of threads for Multi-Thread decoding. if (umThreads == 1) it will use Single-thread decoding + size_t inBufSize_MT; // size of small input data buffers for Multi-Thread decoding. Big number of such small buffers can be created + size_t memUseMax; // the limit of total memory usage for Multi-Thread decoding. + // it's recommended to set (memUseMax) manually to value that is smaller of total size of RAM in computer. + #endif +} CXzDecMtProps; + +void XzDecMtProps_Init(CXzDecMtProps *p); + +typedef struct CXzDecMt CXzDecMt; +typedef CXzDecMt * CXzDecMtHandle; +// Z7_DECLARE_HANDLE(CXzDecMtHandle) + +/* + alloc : XzDecMt uses CAlignOffsetAlloc internally for addresses allocated by (alloc). + allocMid : for big allocations, aligned allocation is better +*/ + +CXzDecMtHandle XzDecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid); +void XzDecMt_Destroy(CXzDecMtHandle p); + + +typedef struct +{ + Byte UnpackSize_Defined; + Byte NumStreams_Defined; + Byte NumBlocks_Defined; + + Byte DataAfterEnd; // there are some additional data after good xz streams, and that data is not new xz stream. + Byte DecodingTruncated; // Decoding was Truncated, we need only partial output data + + UInt64 InSize; // pack size processed. That value doesn't include the data after + // end of xz stream, if that data was not correct + UInt64 OutSize; + + UInt64 NumStreams; + UInt64 NumBlocks; + + SRes DecodeRes; // the error code of xz streams data decoding + SRes ReadRes; // error code from ISeqInStream:Read() + SRes ProgressRes; // error code from ICompressProgress:Progress() + + SRes CombinedRes; // Combined result error code that shows main rusult + // = S_OK, if there is no error. + // but check also (DataAfterEnd) that can show additional minor errors. + + SRes CombinedRes_Type; // = SZ_ERROR_READ, if error from ISeqInStream + // = SZ_ERROR_PROGRESS, if error from ICompressProgress + // = SZ_ERROR_WRITE, if error from ISeqOutStream + // = SZ_ERROR_* codes for decoding +} CXzStatInfo; + +void XzStatInfo_Clear(CXzStatInfo *p); + +/* + +XzDecMt_Decode() +SRes: it's combined decoding result. It also is equal to stat->CombinedRes. + + SZ_OK - no error + check also output value in (stat->DataAfterEnd) + that can show additional possible error + + SZ_ERROR_MEM - Memory allocation error + SZ_ERROR_NO_ARCHIVE - is not xz archive + SZ_ERROR_ARCHIVE - Headers error + SZ_ERROR_DATA - Data Error + SZ_ERROR_UNSUPPORTED - Unsupported method or method properties + SZ_ERROR_CRC - CRC Error + SZ_ERROR_INPUT_EOF - it needs more input data + SZ_ERROR_WRITE - ISeqOutStream error + (SZ_ERROR_READ) - ISeqInStream errors + (SZ_ERROR_PROGRESS) - ICompressProgress errors + // SZ_ERROR_THREAD - error in multi-threading functions + MY_SRes_HRESULT_FROM_WRes(WRes_error) - error in multi-threading function +*/ + +SRes XzDecMt_Decode(CXzDecMtHandle p, + const CXzDecMtProps *props, + const UInt64 *outDataSize, // NULL means undefined + int finishMode, // 0 - partial unpacking is allowed, 1 - xz stream(s) must be finished + ISeqOutStreamPtr outStream, + // Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + // const Byte *inData, size_t inDataSize, + CXzStatInfo *stat, // out: decoding results and statistics + int *isMT, // out: 0 means that ST (Single-Thread) version was used + // 1 means that MT (Multi-Thread) version was used + ICompressProgressPtr progress); EXTERN_C_END diff --git a/3rdparty/lzma/C/XzCrc64.c b/3rdparty/lzma/C/XzCrc64.c index e2da63b6ebe..c2fad6cdaaa 100644 --- a/3rdparty/lzma/C/XzCrc64.c +++ b/3rdparty/lzma/C/XzCrc64.c @@ -1,5 +1,5 @@ /* XzCrc64.c -- CRC64 calculation -2015-03-01 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -9,42 +9,33 @@ #define kCrc64Poly UINT64_CONST(0xC96C5795D7870F42) #ifdef MY_CPU_LE - #define CRC_NUM_TABLES 4 + #define CRC64_NUM_TABLES 4 #else - #define CRC_NUM_TABLES 5 - #define CRC_UINT64_SWAP(v) \ - ((v >> 56) \ - | ((v >> 40) & ((UInt64)0xFF << 8)) \ - | ((v >> 24) & ((UInt64)0xFF << 16)) \ - | ((v >> 8) & ((UInt64)0xFF << 24)) \ - | ((v << 8) & ((UInt64)0xFF << 32)) \ - | ((v << 24) & ((UInt64)0xFF << 40)) \ - | ((v << 40) & ((UInt64)0xFF << 48)) \ - | ((v << 56))) + #define CRC64_NUM_TABLES 5 - UInt64 MY_FAST_CALL XzCrc64UpdateT1_BeT4(UInt64 v, const void *data, size_t size, const UInt64 *table); + UInt64 Z7_FASTCALL XzCrc64UpdateT1_BeT4(UInt64 v, const void *data, size_t size, const UInt64 *table); #endif #ifndef MY_CPU_BE - UInt64 MY_FAST_CALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, const UInt64 *table); + UInt64 Z7_FASTCALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, const UInt64 *table); #endif -typedef UInt64 (MY_FAST_CALL *CRC_FUNC)(UInt64 v, const void *data, size_t size, const UInt64 *table); +typedef UInt64 (Z7_FASTCALL *CRC64_FUNC)(UInt64 v, const void *data, size_t size, const UInt64 *table); -static CRC_FUNC g_Crc64Update; -UInt64 g_Crc64Table[256 * CRC_NUM_TABLES]; +static CRC64_FUNC g_Crc64Update; +UInt64 g_Crc64Table[256 * CRC64_NUM_TABLES]; -UInt64 MY_FAST_CALL Crc64Update(UInt64 v, const void *data, size_t size) +UInt64 Z7_FASTCALL Crc64Update(UInt64 v, const void *data, size_t size) { return g_Crc64Update(v, data, size, g_Crc64Table); } -UInt64 MY_FAST_CALL Crc64Calc(const void *data, size_t size) +UInt64 Z7_FASTCALL Crc64Calc(const void *data, size_t size) { return g_Crc64Update(CRC64_INIT_VAL, data, size, g_Crc64Table) ^ CRC64_INIT_VAL; } -void MY_FAST_CALL Crc64GenerateTable() +void Z7_FASTCALL Crc64GenerateTable(void) { UInt32 i; for (i = 0; i < 256; i++) @@ -52,12 +43,12 @@ void MY_FAST_CALL Crc64GenerateTable() UInt64 r = i; unsigned j; for (j = 0; j < 8; j++) - r = (r >> 1) ^ (kCrc64Poly & ~((r & 1) - 1)); + r = (r >> 1) ^ (kCrc64Poly & ((UInt64)0 - (r & 1))); g_Crc64Table[i] = r; } - for (; i < 256 * CRC_NUM_TABLES; i++) + for (i = 256; i < 256 * CRC64_NUM_TABLES; i++) { - UInt64 r = g_Crc64Table[i - 256]; + const UInt64 r = g_Crc64Table[(size_t)i - 256]; g_Crc64Table[i] = g_Crc64Table[r & 0xFF] ^ (r >> 8); } @@ -74,13 +65,16 @@ void MY_FAST_CALL Crc64GenerateTable() else #endif { - for (i = 256 * CRC_NUM_TABLES - 1; i >= 256; i--) + for (i = 256 * CRC64_NUM_TABLES - 1; i >= 256; i--) { - UInt64 x = g_Crc64Table[i - 256]; - g_Crc64Table[i] = CRC_UINT64_SWAP(x); + const UInt64 x = g_Crc64Table[(size_t)i - 256]; + g_Crc64Table[i] = Z7_BSWAP64(x); } g_Crc64Update = XzCrc64UpdateT1_BeT4; } } #endif } + +#undef kCrc64Poly +#undef CRC64_NUM_TABLES diff --git a/3rdparty/lzma/C/XzCrc64.h b/3rdparty/lzma/C/XzCrc64.h index 08dbc330c27..ca46869a644 100644 --- a/3rdparty/lzma/C/XzCrc64.h +++ b/3rdparty/lzma/C/XzCrc64.h @@ -1,8 +1,8 @@ /* XzCrc64.h -- CRC64 calculation -2013-01-18 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ -#ifndef __XZ_CRC64_H -#define __XZ_CRC64_H +#ifndef ZIP7_INC_XZ_CRC64_H +#define ZIP7_INC_XZ_CRC64_H #include <stddef.h> @@ -12,14 +12,14 @@ EXTERN_C_BEGIN extern UInt64 g_Crc64Table[]; -void MY_FAST_CALL Crc64GenerateTable(void); +void Z7_FASTCALL Crc64GenerateTable(void); #define CRC64_INIT_VAL UINT64_CONST(0xFFFFFFFFFFFFFFFF) #define CRC64_GET_DIGEST(crc) ((crc) ^ CRC64_INIT_VAL) #define CRC64_UPDATE_BYTE(crc, b) (g_Crc64Table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) -UInt64 MY_FAST_CALL Crc64Update(UInt64 crc, const void *data, size_t size); -UInt64 MY_FAST_CALL Crc64Calc(const void *data, size_t size); +UInt64 Z7_FASTCALL Crc64Update(UInt64 crc, const void *data, size_t size); +UInt64 Z7_FASTCALL Crc64Calc(const void *data, size_t size); EXTERN_C_END diff --git a/3rdparty/lzma/C/XzCrc64Opt.c b/3rdparty/lzma/C/XzCrc64Opt.c index c4fde65a10b..d03374c0063 100644 --- a/3rdparty/lzma/C/XzCrc64Opt.c +++ b/3rdparty/lzma/C/XzCrc64Opt.c @@ -1,5 +1,5 @@ /* XzCrc64Opt.c -- CRC64 calculation -2015-03-01 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -7,24 +7,25 @@ #ifndef MY_CPU_BE -#define CRC_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) +#define CRC64_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8)) -UInt64 MY_FAST_CALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, const UInt64 *table) +UInt64 Z7_FASTCALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, const UInt64 *table); +UInt64 Z7_FASTCALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, const UInt64 *table) { const Byte *p = (const Byte *)data; for (; size > 0 && ((unsigned)(ptrdiff_t)p & 3) != 0; size--, p++) - v = CRC_UPDATE_BYTE_2(v, *p); + v = CRC64_UPDATE_BYTE_2(v, *p); for (; size >= 4; size -= 4, p += 4) { - UInt32 d = (UInt32)v ^ *(const UInt32 *)p; + const UInt32 d = (UInt32)v ^ *(const UInt32 *)(const void *)p; v = (v >> 32) - ^ table[0x300 + ((d ) & 0xFF)] - ^ table[0x200 + ((d >> 8) & 0xFF)] - ^ table[0x100 + ((d >> 16) & 0xFF)] - ^ table[0x000 + ((d >> 24))]; + ^ (table + 0x300)[((d ) & 0xFF)] + ^ (table + 0x200)[((d >> 8) & 0xFF)] + ^ (table + 0x100)[((d >> 16) & 0xFF)] + ^ (table + 0x000)[((d >> 24))]; } for (; size > 0; size--, p++) - v = CRC_UPDATE_BYTE_2(v, *p); + v = CRC64_UPDATE_BYTE_2(v, *p); return v; } @@ -33,37 +34,28 @@ UInt64 MY_FAST_CALL XzCrc64UpdateT4(UInt64 v, const void *data, size_t size, con #ifndef MY_CPU_LE -#define CRC_UINT64_SWAP(v) \ - ((v >> 56) \ - | ((v >> 40) & ((UInt64)0xFF << 8)) \ - | ((v >> 24) & ((UInt64)0xFF << 16)) \ - | ((v >> 8) & ((UInt64)0xFF << 24)) \ - | ((v << 8) & ((UInt64)0xFF << 32)) \ - | ((v << 24) & ((UInt64)0xFF << 40)) \ - | ((v << 40) & ((UInt64)0xFF << 48)) \ - | ((v << 56))) +#define CRC64_UPDATE_BYTE_2_BE(crc, b) (table[(Byte)((crc) >> 56) ^ (b)] ^ ((crc) << 8)) -#define CRC_UPDATE_BYTE_2_BE(crc, b) (table[(Byte)((crc) >> 56) ^ (b)] ^ ((crc) << 8)) - -UInt64 MY_FAST_CALL XzCrc64UpdateT1_BeT4(UInt64 v, const void *data, size_t size, const UInt64 *table) +UInt64 Z7_FASTCALL XzCrc64UpdateT1_BeT4(UInt64 v, const void *data, size_t size, const UInt64 *table); +UInt64 Z7_FASTCALL XzCrc64UpdateT1_BeT4(UInt64 v, const void *data, size_t size, const UInt64 *table) { const Byte *p = (const Byte *)data; table += 0x100; - v = CRC_UINT64_SWAP(v); + v = Z7_BSWAP64(v); for (; size > 0 && ((unsigned)(ptrdiff_t)p & 3) != 0; size--, p++) - v = CRC_UPDATE_BYTE_2_BE(v, *p); + v = CRC64_UPDATE_BYTE_2_BE(v, *p); for (; size >= 4; size -= 4, p += 4) { - UInt32 d = (UInt32)(v >> 32) ^ *(const UInt32 *)p; + const UInt32 d = (UInt32)(v >> 32) ^ *(const UInt32 *)(const void *)p; v = (v << 32) - ^ table[0x000 + ((d ) & 0xFF)] - ^ table[0x100 + ((d >> 8) & 0xFF)] - ^ table[0x200 + ((d >> 16) & 0xFF)] - ^ table[0x300 + ((d >> 24))]; + ^ (table + 0x000)[((d ) & 0xFF)] + ^ (table + 0x100)[((d >> 8) & 0xFF)] + ^ (table + 0x200)[((d >> 16) & 0xFF)] + ^ (table + 0x300)[((d >> 24))]; } for (; size > 0; size--, p++) - v = CRC_UPDATE_BYTE_2_BE(v, *p); - return CRC_UINT64_SWAP(v); + v = CRC64_UPDATE_BYTE_2_BE(v, *p); + return Z7_BSWAP64(v); } #endif diff --git a/3rdparty/lzma/C/XzDec.c b/3rdparty/lzma/C/XzDec.c index 1edc0245637..a5f703966d6 100644 --- a/3rdparty/lzma/C/XzDec.c +++ b/3rdparty/lzma/C/XzDec.c @@ -1,14 +1,33 @@ /* XzDec.c -- Xz Decode -2015-11-09 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ #include "Precomp.h" +// #include <stdio.h> + +// #define XZ_DUMP + /* #define XZ_DUMP */ #ifdef XZ_DUMP #include <stdio.h> #endif +// #define SHOW_DEBUG_INFO + +#ifdef SHOW_DEBUG_INFO +#include <stdio.h> +#endif + +#ifdef SHOW_DEBUG_INFO +#define PRF(x) x +#else +#define PRF(x) +#endif + +#define PRF_STR(s) PRF(printf("\n" s "\n")) +#define PRF_STR_INT(s, d) PRF(printf("\n" s " %d\n", (unsigned)d)) + #include <stdlib.h> #include <string.h> @@ -19,16 +38,18 @@ #include "Delta.h" #include "Lzma2Dec.h" +// #define USE_SUBBLOCK + #ifdef USE_SUBBLOCK #include "Bcj3Dec.c" -#include "SbDec.c" +#include "SbDec.h" #endif #include "Xz.h" #define XZ_CHECK_SIZE_MAX 64 -#define CODER_BUF_SIZE (1 << 17) +#define CODER_BUF_SIZE ((size_t)1 << 17) unsigned Xz_ReadVarInt(const Byte *p, size_t maxSize, UInt64 *value) { @@ -46,7 +67,8 @@ unsigned Xz_ReadVarInt(const Byte *p, size_t maxSize, UInt64 *value) return 0; } -/* ---------- BraState ---------- */ + +/* ---------- XzBcFilterState ---------- */ #define BRA_BUF_SIZE (1 << 14) @@ -55,26 +77,29 @@ typedef struct size_t bufPos; size_t bufConv; size_t bufTotal; + Byte *buf; // must be aligned for 4 bytes + Xz_Func_BcFilterStateBase_Filter filter_func; + // int encodeMode; + CXzBcFilterStateBase base; + // Byte buf[BRA_BUF_SIZE]; +} CXzBcFilterState; - UInt32 methodId; - int encodeMode; - UInt32 delta; - UInt32 ip; - UInt32 x86State; - Byte deltaState[DELTA_STATE_SIZE]; - - Byte buf[BRA_BUF_SIZE]; -} CBraState; -static void BraState_Free(void *pp, ISzAlloc *alloc) +static void XzBcFilterState_Free(void *pp, ISzAllocPtr alloc) { - alloc->Free(alloc, pp); + if (pp) + { + CXzBcFilterState *p = ((CXzBcFilterState *)pp); + ISzAlloc_Free(alloc, p->buf); + ISzAlloc_Free(alloc, pp); + } } -static SRes BraState_SetProps(void *pp, const Byte *props, size_t propSize, ISzAlloc *alloc) + +static SRes XzBcFilterState_SetProps(void *pp, const Byte *props, size_t propSize, ISzAllocPtr alloc) { - CBraState *p = ((CBraState *)pp); - UNUSED_VAR(alloc); + CXzBcFilterStateBase *p = &((CXzBcFilterState *)pp)->base; + UNUSED_VAR(alloc) p->ip = 0; if (p->methodId == XZ_ID_Delta) { @@ -92,6 +117,7 @@ static SRes BraState_SetProps(void *pp, const Byte *props, size_t propSize, ISzA case XZ_ID_PPC: case XZ_ID_ARM: case XZ_ID_SPARC: + case XZ_ID_ARM64: if ((v & 3) != 0) return SZ_ERROR_UNSUPPORTED; break; @@ -112,78 +138,110 @@ static SRes BraState_SetProps(void *pp, const Byte *props, size_t propSize, ISzA return SZ_OK; } -static void BraState_Init(void *pp) + +static void XzBcFilterState_Init(void *pp) { - CBraState *p = ((CBraState *)pp); + CXzBcFilterState *p = ((CXzBcFilterState *)pp); p->bufPos = p->bufConv = p->bufTotal = 0; - x86_Convert_Init(p->x86State); - if (p->methodId == XZ_ID_Delta) - Delta_Init(p->deltaState); + p->base.X86_State = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL; + if (p->base.methodId == XZ_ID_Delta) + Delta_Init(p->base.delta_State); } -#define CASE_BRA_CONV(isa) case XZ_ID_ ## isa: p->bufConv = isa ## _Convert(p->buf, p->bufTotal, p->ip, p->encodeMode); break; -static SRes BraState_Code(void *pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - int srcWasFinished, ECoderFinishMode finishMode, int *wasFinished) +static const z7_Func_BranchConv g_Funcs_BranchConv_RISC_Dec[] = { - CBraState *p = ((CBraState *)pp); - SizeT destLenOrig = *destLen; - SizeT srcLenOrig = *srcLen; - UNUSED_VAR(finishMode); + Z7_BRANCH_CONV_DEC(PPC), + Z7_BRANCH_CONV_DEC(IA64), + Z7_BRANCH_CONV_DEC(ARM), + Z7_BRANCH_CONV_DEC(ARMT), + Z7_BRANCH_CONV_DEC(SPARC), + Z7_BRANCH_CONV_DEC(ARM64) +}; + +static SizeT XzBcFilterStateBase_Filter_Dec(CXzBcFilterStateBase *p, Byte *data, SizeT size) +{ + switch (p->methodId) + { + case XZ_ID_Delta: + Delta_Decode(p->delta_State, p->delta, data, size); + break; + case XZ_ID_X86: + size = (SizeT)(z7_BranchConvSt_X86_Dec(data, size, p->ip, &p->X86_State) - data); + break; + default: + if (p->methodId >= XZ_ID_PPC) + { + const UInt32 i = p->methodId - XZ_ID_PPC; + if (i < Z7_ARRAY_SIZE(g_Funcs_BranchConv_RISC_Dec)) + size = (SizeT)(g_Funcs_BranchConv_RISC_Dec[i](data, size, p->ip) - data); + } + break; + } + p->ip += (UInt32)size; + return size; +} + + +static SizeT XzBcFilterState_Filter(void *pp, Byte *data, SizeT size) +{ + CXzBcFilterState *p = ((CXzBcFilterState *)pp); + return p->filter_func(&p->base, data, size); +} + + +static SRes XzBcFilterState_Code2(void *pp, + Byte *dest, SizeT *destLen, + const Byte *src, SizeT *srcLen, int srcWasFinished, + ECoderFinishMode finishMode, + // int *wasFinished + ECoderStatus *status) +{ + CXzBcFilterState *p = ((CXzBcFilterState *)pp); + SizeT destRem = *destLen; + SizeT srcRem = *srcLen; + UNUSED_VAR(finishMode) + *destLen = 0; *srcLen = 0; - *wasFinished = 0; - while (destLenOrig > 0) + // *wasFinished = False; + *status = CODER_STATUS_NOT_FINISHED; + + while (destRem != 0) { - if (p->bufPos != p->bufConv) { - size_t curSize = p->bufConv - p->bufPos; - if (curSize > destLenOrig) - curSize = destLenOrig; - memcpy(dest, p->buf + p->bufPos, curSize); - p->bufPos += curSize; - *destLen += curSize; - dest += curSize; - destLenOrig -= curSize; - continue; + size_t size = p->bufConv - p->bufPos; + if (size) + { + if (size > destRem) + size = destRem; + memcpy(dest, p->buf + p->bufPos, size); + p->bufPos += size; + *destLen += size; + dest += size; + destRem -= size; + continue; + } } + p->bufTotal -= p->bufPos; memmove(p->buf, p->buf + p->bufPos, p->bufTotal); p->bufPos = 0; p->bufConv = 0; { - size_t curSize = BRA_BUF_SIZE - p->bufTotal; - if (curSize > srcLenOrig) - curSize = srcLenOrig; - memcpy(p->buf + p->bufTotal, src, curSize); - *srcLen += curSize; - src += curSize; - srcLenOrig -= curSize; - p->bufTotal += curSize; + size_t size = BRA_BUF_SIZE - p->bufTotal; + if (size > srcRem) + size = srcRem; + memcpy(p->buf + p->bufTotal, src, size); + *srcLen += size; + src += size; + srcRem -= size; + p->bufTotal += size; } if (p->bufTotal == 0) break; - switch (p->methodId) - { - case XZ_ID_Delta: - if (p->encodeMode) - Delta_Encode(p->deltaState, p->delta, p->buf, p->bufTotal); - else - Delta_Decode(p->deltaState, p->delta, p->buf, p->bufTotal); - p->bufConv = p->bufTotal; - break; - case XZ_ID_X86: - p->bufConv = x86_Convert(p->buf, p->bufTotal, p->ip, &p->x86State, p->encodeMode); - break; - CASE_BRA_CONV(PPC) - CASE_BRA_CONV(IA64) - CASE_BRA_CONV(ARM) - CASE_BRA_CONV(ARMT) - CASE_BRA_CONV(SPARC) - default: - return SZ_ERROR_UNSUPPORTED; - } - p->ip += (UInt32)p->bufConv; + + p->bufConv = p->filter_func(&p->base, p->buf, p->bufTotal); if (p->bufConv == 0) { @@ -192,52 +250,69 @@ static SRes BraState_Code(void *pp, Byte *dest, SizeT *destLen, const Byte *src, p->bufConv = p->bufTotal; } } - if (p->bufTotal == p->bufPos && srcLenOrig == 0 && srcWasFinished) - *wasFinished = 1; + + if (p->bufTotal == p->bufPos && srcRem == 0 && srcWasFinished) + { + *status = CODER_STATUS_FINISHED_WITH_MARK; + // *wasFinished = 1; + } + return SZ_OK; } -SRes BraState_SetFromMethod(IStateCoder *p, UInt64 id, int encodeMode, ISzAlloc *alloc) + +#define XZ_IS_SUPPORTED_FILTER_ID(id) \ + ((id) >= XZ_ID_Delta && (id) <= XZ_ID_ARM64) + +SRes Xz_StateCoder_Bc_SetFromMethod_Func(IStateCoder *p, UInt64 id, + Xz_Func_BcFilterStateBase_Filter func, ISzAllocPtr alloc) { - CBraState *decoder; - if (id != XZ_ID_Delta && - id != XZ_ID_X86 && - id != XZ_ID_PPC && - id != XZ_ID_IA64 && - id != XZ_ID_ARM && - id != XZ_ID_ARMT && - id != XZ_ID_SPARC) + CXzBcFilterState *decoder; + if (!XZ_IS_SUPPORTED_FILTER_ID(id)) return SZ_ERROR_UNSUPPORTED; - p->p = 0; - decoder = (CBraState *)alloc->Alloc(alloc, sizeof(CBraState)); - if (decoder == 0) - return SZ_ERROR_MEM; - decoder->methodId = (UInt32)id; - decoder->encodeMode = encodeMode; - p->p = decoder; - p->Free = BraState_Free; - p->SetProps = BraState_SetProps; - p->Init = BraState_Init; - p->Code = BraState_Code; + decoder = (CXzBcFilterState *)p->p; + if (!decoder) + { + decoder = (CXzBcFilterState *)ISzAlloc_Alloc(alloc, sizeof(CXzBcFilterState)); + if (!decoder) + return SZ_ERROR_MEM; + decoder->buf = ISzAlloc_Alloc(alloc, BRA_BUF_SIZE); + if (!decoder->buf) + { + ISzAlloc_Free(alloc, decoder); + return SZ_ERROR_MEM; + } + p->p = decoder; + p->Free = XzBcFilterState_Free; + p->SetProps = XzBcFilterState_SetProps; + p->Init = XzBcFilterState_Init; + p->Code2 = XzBcFilterState_Code2; + p->Filter = XzBcFilterState_Filter; + decoder->filter_func = func; + } + decoder->base.methodId = (UInt32)id; + // decoder->encodeMode = encodeMode; return SZ_OK; } + + /* ---------- SbState ---------- */ #ifdef USE_SUBBLOCK -static void SbState_Free(void *pp, ISzAlloc *alloc) +static void SbState_Free(void *pp, ISzAllocPtr alloc) { CSbDec *p = (CSbDec *)pp; SbDec_Free(p); - alloc->Free(alloc, pp); + ISzAlloc_Free(alloc, pp); } -static SRes SbState_SetProps(void *pp, const Byte *props, size_t propSize, ISzAlloc *alloc) +static SRes SbState_SetProps(void *pp, const Byte *props, size_t propSize, ISzAllocPtr alloc) { - UNUSED_VAR(pp); - UNUSED_VAR(props); - UNUSED_VAR(alloc); + UNUSED_VAR(pp) + UNUSED_VAR(props) + UNUSED_VAR(alloc) return (propSize == 0) ? SZ_OK : SZ_ERROR_UNSUPPORTED; } @@ -246,12 +321,14 @@ static void SbState_Init(void *pp) SbDec_Init((CSbDec *)pp); } -static SRes SbState_Code(void *pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - int srcWasFinished, ECoderFinishMode finishMode, int *wasFinished) +static SRes SbState_Code2(void *pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, + int srcWasFinished, ECoderFinishMode finishMode, + // int *wasFinished + ECoderStatus *status) { CSbDec *p = (CSbDec *)pp; SRes res; - UNUSED_VAR(srcWasFinished); + UNUSED_VAR(srcWasFinished) p->dest = dest; p->destLen = *destLen; p->src = src; @@ -260,102 +337,185 @@ static SRes SbState_Code(void *pp, Byte *dest, SizeT *destLen, const Byte *src, res = SbDec_Decode((CSbDec *)pp); *destLen -= p->destLen; *srcLen -= p->srcLen; - *wasFinished = (*destLen == 0 && *srcLen == 0); /* change it */ + // *wasFinished = (*destLen == 0 && *srcLen == 0); /* change it */ + *status = (*destLen == 0 && *srcLen == 0) ? + CODER_STATUS_FINISHED_WITH_MARK : + CODER_STATUS_NOT_FINISHED; return res; } -SRes SbState_SetFromMethod(IStateCoder *p, ISzAlloc *alloc) +static SRes SbState_SetFromMethod(IStateCoder *p, ISzAllocPtr alloc) { - CSbDec *decoder; - p->p = 0; - decoder = alloc->Alloc(alloc, sizeof(CSbDec)); - if (decoder == 0) - return SZ_ERROR_MEM; - p->p = decoder; - p->Free = SbState_Free; - p->SetProps = SbState_SetProps; - p->Init = SbState_Init; - p->Code = SbState_Code; + CSbDec *decoder = (CSbDec *)p->p; + if (!decoder) + { + decoder = (CSbDec *)ISzAlloc_Alloc(alloc, sizeof(CSbDec)); + if (!decoder) + return SZ_ERROR_MEM; + p->p = decoder; + p->Free = SbState_Free; + p->SetProps = SbState_SetProps; + p->Init = SbState_Init; + p->Code2 = SbState_Code2; + p->Filter = NULL; + } SbDec_Construct(decoder); SbDec_SetAlloc(decoder, alloc); return SZ_OK; } + #endif -/* ---------- Lzma2State ---------- */ -static void Lzma2State_Free(void *pp, ISzAlloc *alloc) + +/* ---------- Lzma2 ---------- */ + +typedef struct +{ + CLzma2Dec decoder; + BoolInt outBufMode; +} CLzma2Dec_Spec; + + +static void Lzma2State_Free(void *pp, ISzAllocPtr alloc) { - Lzma2Dec_Free((CLzma2Dec *)pp, alloc); - alloc->Free(alloc, pp); + CLzma2Dec_Spec *p = (CLzma2Dec_Spec *)pp; + if (p->outBufMode) + Lzma2Dec_FreeProbs(&p->decoder, alloc); + else + Lzma2Dec_Free(&p->decoder, alloc); + ISzAlloc_Free(alloc, pp); } -static SRes Lzma2State_SetProps(void *pp, const Byte *props, size_t propSize, ISzAlloc *alloc) +static SRes Lzma2State_SetProps(void *pp, const Byte *props, size_t propSize, ISzAllocPtr alloc) { if (propSize != 1) return SZ_ERROR_UNSUPPORTED; - return Lzma2Dec_Allocate((CLzma2Dec *)pp, props[0], alloc); + { + CLzma2Dec_Spec *p = (CLzma2Dec_Spec *)pp; + if (p->outBufMode) + return Lzma2Dec_AllocateProbs(&p->decoder, props[0], alloc); + else + return Lzma2Dec_Allocate(&p->decoder, props[0], alloc); + } } static void Lzma2State_Init(void *pp) { - Lzma2Dec_Init((CLzma2Dec *)pp); + Lzma2Dec_Init(&((CLzma2Dec_Spec *)pp)->decoder); } -static SRes Lzma2State_Code(void *pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, - int srcWasFinished, ECoderFinishMode finishMode, int *wasFinished) + +/* + if (outBufMode), then (dest) is not used. Use NULL. + Data is unpacked to (spec->decoder.decoder.dic) output buffer. +*/ + +static SRes Lzma2State_Code2(void *pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, + int srcWasFinished, ECoderFinishMode finishMode, + // int *wasFinished, + ECoderStatus *status) { - ELzmaStatus status; + CLzma2Dec_Spec *spec = (CLzma2Dec_Spec *)pp; + ELzmaStatus status2; /* ELzmaFinishMode fm = (finishMode == LZMA_FINISH_ANY) ? LZMA_FINISH_ANY : LZMA_FINISH_END; */ - SRes res = Lzma2Dec_DecodeToBuf((CLzma2Dec *)pp, dest, destLen, src, srcLen, (ELzmaFinishMode)finishMode, &status); - UNUSED_VAR(srcWasFinished); - *wasFinished = (status == LZMA_STATUS_FINISHED_WITH_MARK); + SRes res; + UNUSED_VAR(srcWasFinished) + if (spec->outBufMode) + { + SizeT dicPos = spec->decoder.decoder.dicPos; + SizeT dicLimit = dicPos + *destLen; + res = Lzma2Dec_DecodeToDic(&spec->decoder, dicLimit, src, srcLen, (ELzmaFinishMode)finishMode, &status2); + *destLen = spec->decoder.decoder.dicPos - dicPos; + } + else + res = Lzma2Dec_DecodeToBuf(&spec->decoder, dest, destLen, src, srcLen, (ELzmaFinishMode)finishMode, &status2); + // *wasFinished = (status2 == LZMA_STATUS_FINISHED_WITH_MARK); + // ECoderStatus values are identical to ELzmaStatus values of LZMA2 decoder + *status = (ECoderStatus)status2; return res; } -static SRes Lzma2State_SetFromMethod(IStateCoder *p, ISzAlloc *alloc) + +static SRes Lzma2State_SetFromMethod(IStateCoder *p, Byte *outBuf, size_t outBufSize, ISzAllocPtr alloc) +{ + CLzma2Dec_Spec *spec = (CLzma2Dec_Spec *)p->p; + if (!spec) + { + spec = (CLzma2Dec_Spec *)ISzAlloc_Alloc(alloc, sizeof(CLzma2Dec_Spec)); + if (!spec) + return SZ_ERROR_MEM; + p->p = spec; + p->Free = Lzma2State_Free; + p->SetProps = Lzma2State_SetProps; + p->Init = Lzma2State_Init; + p->Code2 = Lzma2State_Code2; + p->Filter = NULL; + Lzma2Dec_CONSTRUCT(&spec->decoder) + } + spec->outBufMode = False; + if (outBuf) + { + spec->outBufMode = True; + spec->decoder.decoder.dic = outBuf; + spec->decoder.decoder.dicBufSize = outBufSize; + } + return SZ_OK; +} + + +static SRes Lzma2State_ResetOutBuf(IStateCoder *p, Byte *outBuf, size_t outBufSize) { - CLzma2Dec *decoder = (CLzma2Dec *)alloc->Alloc(alloc, sizeof(CLzma2Dec)); - p->p = decoder; - if (decoder == 0) - return SZ_ERROR_MEM; - p->Free = Lzma2State_Free; - p->SetProps = Lzma2State_SetProps; - p->Init = Lzma2State_Init; - p->Code = Lzma2State_Code; - Lzma2Dec_Construct(decoder); + CLzma2Dec_Spec *spec = (CLzma2Dec_Spec *)p->p; + if ((spec->outBufMode && !outBuf) || (!spec->outBufMode && outBuf)) + return SZ_ERROR_FAIL; + if (outBuf) + { + spec->decoder.decoder.dic = outBuf; + spec->decoder.decoder.dicBufSize = outBufSize; + } return SZ_OK; } -void MixCoder_Construct(CMixCoder *p, ISzAlloc *alloc) + +static void MixCoder_Construct(CMixCoder *p, ISzAllocPtr alloc) { unsigned i; p->alloc = alloc; p->buf = NULL; p->numCoders = 0; + + p->outBufSize = 0; + p->outBuf = NULL; + // p->SingleBufMode = False; + for (i = 0; i < MIXCODER_NUM_FILTERS_MAX; i++) p->coders[i].p = NULL; } -void MixCoder_Free(CMixCoder *p) + +static void MixCoder_Free(CMixCoder *p) { unsigned i; - for (i = 0; i < p->numCoders; i++) + p->numCoders = 0; + for (i = 0; i < MIXCODER_NUM_FILTERS_MAX; i++) { IStateCoder *sc = &p->coders[i]; - if (p->alloc && sc->p) + if (sc->p) + { sc->Free(sc->p, p->alloc); + sc->p = NULL; + } } - p->numCoders = 0; if (p->buf) { - p->alloc->Free(p->alloc, p->buf); + ISzAlloc_Free(p->alloc, p->buf); p->buf = NULL; /* 9.31: the BUG was fixed */ } } -void MixCoder_Init(CMixCoder *p) +static void MixCoder_Init(CMixCoder *p) { unsigned i; for (i = 0; i < MIXCODER_NUM_FILTERS_MAX - 1; i++) @@ -368,50 +528,154 @@ void MixCoder_Init(CMixCoder *p) { IStateCoder *coder = &p->coders[i]; coder->Init(coder->p); + p->results[i] = SZ_OK; } + p->outWritten = 0; + p->wasFinished = False; + p->res = SZ_OK; + p->status = CODER_STATUS_NOT_SPECIFIED; } -SRes MixCoder_SetFromMethod(CMixCoder *p, unsigned coderIndex, UInt64 methodId) + +static SRes MixCoder_SetFromMethod(CMixCoder *p, unsigned coderIndex, UInt64 methodId, Byte *outBuf, size_t outBufSize) { IStateCoder *sc = &p->coders[coderIndex]; p->ids[coderIndex] = methodId; switch (methodId) { - case XZ_ID_LZMA2: return Lzma2State_SetFromMethod(sc, p->alloc); + case XZ_ID_LZMA2: return Lzma2State_SetFromMethod(sc, outBuf, outBufSize, p->alloc); #ifdef USE_SUBBLOCK case XZ_ID_Subblock: return SbState_SetFromMethod(sc, p->alloc); #endif } if (coderIndex == 0) return SZ_ERROR_UNSUPPORTED; - return BraState_SetFromMethod(sc, methodId, 0, p->alloc); + return Xz_StateCoder_Bc_SetFromMethod_Func(sc, methodId, + XzBcFilterStateBase_Filter_Dec, p->alloc); } -SRes MixCoder_Code(CMixCoder *p, Byte *dest, SizeT *destLen, + +static SRes MixCoder_ResetFromMethod(CMixCoder *p, unsigned coderIndex, UInt64 methodId, Byte *outBuf, size_t outBufSize) +{ + IStateCoder *sc = &p->coders[coderIndex]; + switch (methodId) + { + case XZ_ID_LZMA2: return Lzma2State_ResetOutBuf(sc, outBuf, outBufSize); + } + return SZ_ERROR_UNSUPPORTED; +} + + + +/* + if (destFinish) - then unpack data block is finished at (*destLen) position, + and we can return data that were not processed by filter + +output (status) can be : + CODER_STATUS_NOT_FINISHED + CODER_STATUS_FINISHED_WITH_MARK + CODER_STATUS_NEEDS_MORE_INPUT - not implemented still +*/ + +static SRes MixCoder_Code(CMixCoder *p, + Byte *dest, SizeT *destLen, int destFinish, const Byte *src, SizeT *srcLen, int srcWasFinished, - ECoderFinishMode finishMode, ECoderStatus *status) + ECoderFinishMode finishMode) { SizeT destLenOrig = *destLen; SizeT srcLenOrig = *srcLen; - Bool allFinished = True; + *destLen = 0; *srcLen = 0; - *status = CODER_STATUS_NOT_FINISHED; - if (!p->buf) + if (p->wasFinished) + return p->res; + + p->status = CODER_STATUS_NOT_FINISHED; + + // if (p->SingleBufMode) + if (p->outBuf) { - p->buf = (Byte *)p->alloc->Alloc(p->alloc, CODER_BUF_SIZE * (MIXCODER_NUM_FILTERS_MAX - 1)); - if (!p->buf) - return SZ_ERROR_MEM; + SRes res; + SizeT destLen2, srcLen2; + int wasFinished; + + PRF_STR("------- MixCoder Single ----------") + + srcLen2 = srcLenOrig; + destLen2 = destLenOrig; + + { + IStateCoder *coder = &p->coders[0]; + res = coder->Code2(coder->p, NULL, &destLen2, src, &srcLen2, srcWasFinished, finishMode, + // &wasFinished, + &p->status); + wasFinished = (p->status == CODER_STATUS_FINISHED_WITH_MARK); + } + + p->res = res; + + /* + if (wasFinished) + p->status = CODER_STATUS_FINISHED_WITH_MARK; + else + { + if (res == SZ_OK) + if (destLen2 != destLenOrig) + p->status = CODER_STATUS_NEEDS_MORE_INPUT; + } + */ + + + *srcLen = srcLen2; + src += srcLen2; + p->outWritten += destLen2; + + if (res != SZ_OK || srcWasFinished || wasFinished) + p->wasFinished = True; + + if (p->numCoders == 1) + *destLen = destLen2; + else if (p->wasFinished) + { + unsigned i; + size_t processed = p->outWritten; + + for (i = 1; i < p->numCoders; i++) + { + IStateCoder *coder = &p->coders[i]; + processed = coder->Filter(coder->p, p->outBuf, processed); + if (wasFinished || (destFinish && p->outWritten == destLenOrig)) + processed = p->outWritten; + PRF_STR_INT("filter", i) + } + *destLen = processed; + } + return res; } + PRF_STR("standard mix") + if (p->numCoders != 1) + { + if (!p->buf) + { + p->buf = (Byte *)ISzAlloc_Alloc(p->alloc, CODER_BUF_SIZE * (MIXCODER_NUM_FILTERS_MAX - 1)); + if (!p->buf) + return SZ_ERROR_MEM; + } + finishMode = CODER_FINISH_ANY; + } for (;;) { - Bool processed = False; + BoolInt processed = False; + BoolInt allFinished = True; + SRes resMain = SZ_OK; unsigned i; + + p->status = CODER_STATUS_NOT_FINISHED; /* if (p->numCoders == 1 && *destLen == destLenOrig && finishMode == LZMA_FINISH_ANY) break; @@ -421,79 +685,106 @@ SRes MixCoder_Code(CMixCoder *p, Byte *dest, SizeT *destLen, { SRes res; IStateCoder *coder = &p->coders[i]; - Byte *destCur; - SizeT destLenCur, srcLenCur; - const Byte *srcCur; - int srcFinishedCur; + Byte *dest2; + SizeT destLen2, srcLen2; // destLen2_Orig; + const Byte *src2; + int srcFinished2; int encodingWasFinished; + ECoderStatus status2; if (i == 0) { - srcCur = src; - srcLenCur = srcLenOrig - *srcLen; - srcFinishedCur = srcWasFinished; + src2 = src; + srcLen2 = srcLenOrig - *srcLen; + srcFinished2 = srcWasFinished; } else { - srcCur = p->buf + (CODER_BUF_SIZE * (i - 1)) + p->pos[i - 1]; - srcLenCur = p->size[i - 1] - p->pos[i - 1]; - srcFinishedCur = p->finished[i - 1]; + size_t k = i - 1; + src2 = p->buf + (CODER_BUF_SIZE * k) + p->pos[k]; + srcLen2 = p->size[k] - p->pos[k]; + srcFinished2 = p->finished[k]; } if (i == p->numCoders - 1) { - destCur = dest; - destLenCur = destLenOrig - *destLen; + dest2 = dest; + destLen2 = destLenOrig - *destLen; } else { if (p->pos[i] != p->size[i]) continue; - destCur = p->buf + (CODER_BUF_SIZE * i); - destLenCur = CODER_BUF_SIZE; + dest2 = p->buf + (CODER_BUF_SIZE * i); + destLen2 = CODER_BUF_SIZE; } - res = coder->Code(coder->p, destCur, &destLenCur, srcCur, &srcLenCur, srcFinishedCur, finishMode, &encodingWasFinished); + // destLen2_Orig = destLen2; + + if (p->results[i] != SZ_OK) + { + if (resMain == SZ_OK) + resMain = p->results[i]; + continue; + } + + res = coder->Code2(coder->p, + dest2, &destLen2, + src2, &srcLen2, srcFinished2, + finishMode, + // &encodingWasFinished, + &status2); + + if (res != SZ_OK) + { + p->results[i] = res; + if (resMain == SZ_OK) + resMain = res; + } + encodingWasFinished = (status2 == CODER_STATUS_FINISHED_WITH_MARK); + if (!encodingWasFinished) + { allFinished = False; + if (p->numCoders == 1 && res == SZ_OK) + p->status = status2; + } if (i == 0) { - *srcLen += srcLenCur; - src += srcLenCur; + *srcLen += srcLen2; + src += srcLen2; } else - { - p->pos[i - 1] += srcLenCur; - } + p->pos[(size_t)i - 1] += srcLen2; if (i == p->numCoders - 1) { - *destLen += destLenCur; - dest += destLenCur; + *destLen += destLen2; + dest += destLen2; } else { - p->size[i] = destLenCur; + p->size[i] = destLen2; p->pos[i] = 0; p->finished[i] = encodingWasFinished; } - if (res != SZ_OK) - return res; - - if (destLenCur != 0 || srcLenCur != 0) + if (destLen2 != 0 || srcLen2 != 0) processed = True; } + if (!processed) - break; + { + if (allFinished) + p->status = CODER_STATUS_FINISHED_WITH_MARK; + return resMain; + } } - if (allFinished) - *status = CODER_STATUS_FINISHED_WITH_MARK; - return SZ_OK; } + SRes Xz_ParseHeader(CXzStreamFlags *p, const Byte *buf) { *p = (CXzStreamFlags)GetBe16(buf + XZ_SIG_SIZE); @@ -503,18 +794,45 @@ SRes Xz_ParseHeader(CXzStreamFlags *p, const Byte *buf) return XzFlags_IsSupported(*p) ? SZ_OK : SZ_ERROR_UNSUPPORTED; } -static Bool Xz_CheckFooter(CXzStreamFlags flags, UInt64 indexSize, const Byte *buf) +static BoolInt Xz_CheckFooter(CXzStreamFlags flags, UInt64 indexSize, const Byte *buf) { - return - indexSize == (((UInt64)GetUi32(buf + 4) + 1) << 2) && - (GetUi32(buf) == CrcCalc(buf + 4, 6) && - flags == GetBe16(buf + 8) && - memcmp(buf + 10, XZ_FOOTER_SIG, XZ_FOOTER_SIG_SIZE) == 0); + return indexSize == (((UInt64)GetUi32(buf + 4) + 1) << 2) + && GetUi32(buf) == CrcCalc(buf + 4, 6) + && flags == GetBe16(buf + 8) + && buf[10] == XZ_FOOTER_SIG_0 + && buf[11] == XZ_FOOTER_SIG_1; } #define READ_VARINT_AND_CHECK(buf, pos, size, res) \ { unsigned s = Xz_ReadVarInt(buf + pos, size - pos, res); \ - if (s == 0) return SZ_ERROR_ARCHIVE; pos += s; } + if (s == 0) return SZ_ERROR_ARCHIVE; \ + pos += s; } + + +static BoolInt XzBlock_AreSupportedFilters(const CXzBlock *p) +{ + const unsigned numFilters = XzBlock_GetNumFilters(p) - 1; + unsigned i; + { + const CXzFilter *f = &p->filters[numFilters]; + if (f->id != XZ_ID_LZMA2 || f->propsSize != 1 || f->props[0] > 40) + return False; + } + + for (i = 0; i < numFilters; i++) + { + const CXzFilter *f = &p->filters[i]; + if (f->id == XZ_ID_Delta) + { + if (f->propsSize != 1) + return False; + } + else if (!XZ_IS_SUPPORTED_FILTER_ID(f->id) + || (f->propsSize != 0 && f->propsSize != 4)) + return False; + } + return True; +} SRes XzBlock_Parse(CXzBlock *p, const Byte *header) @@ -523,31 +841,35 @@ SRes XzBlock_Parse(CXzBlock *p, const Byte *header) unsigned numFilters, i; unsigned headerSize = (unsigned)header[0] << 2; + /* (headerSize != 0) : another code checks */ + if (CrcCalc(header, headerSize) != GetUi32(header + headerSize)) return SZ_ERROR_ARCHIVE; pos = 1; - if (pos == headerSize) - return SZ_ERROR_ARCHIVE; p->flags = header[pos++]; + p->packSize = (UInt64)(Int64)-1; if (XzBlock_HasPackSize(p)) { - READ_VARINT_AND_CHECK(header, pos, headerSize, &p->packSize); + READ_VARINT_AND_CHECK(header, pos, headerSize, &p->packSize) if (p->packSize == 0 || p->packSize + headerSize >= (UInt64)1 << 63) return SZ_ERROR_ARCHIVE; } + p->unpackSize = (UInt64)(Int64)-1; if (XzBlock_HasUnpackSize(p)) - READ_VARINT_AND_CHECK(header, pos, headerSize, &p->unpackSize); + { + READ_VARINT_AND_CHECK(header, pos, headerSize, &p->unpackSize) + } numFilters = XzBlock_GetNumFilters(p); for (i = 0; i < numFilters; i++) { CXzFilter *filter = p->filters + i; UInt64 size; - READ_VARINT_AND_CHECK(header, pos, headerSize, &filter->id); - READ_VARINT_AND_CHECK(header, pos, headerSize, &size); + READ_VARINT_AND_CHECK(header, pos, headerSize, &filter->id) + READ_VARINT_AND_CHECK(header, pos, headerSize, &size) if (size > headerSize - pos || size > XZ_FILTER_PROPS_SIZE_MAX) return SZ_ERROR_ARCHIVE; filter->propsSize = (UInt32)size; @@ -564,48 +886,69 @@ SRes XzBlock_Parse(CXzBlock *p, const Byte *header) #endif } + if (XzBlock_HasUnsupportedFlags(p)) + return SZ_ERROR_UNSUPPORTED; + while (pos < headerSize) if (header[pos++] != 0) return SZ_ERROR_ARCHIVE; return SZ_OK; } -SRes XzDec_Init(CMixCoder *p, const CXzBlock *block) + + + +static SRes XzDecMix_Init(CMixCoder *p, const CXzBlock *block, Byte *outBuf, size_t outBufSize) { unsigned i; - Bool needReInit = True; + BoolInt needReInit = True; unsigned numFilters = XzBlock_GetNumFilters(block); - - if (numFilters == p->numCoders) + + if (numFilters == p->numCoders && ((p->outBuf && outBuf) || (!p->outBuf && !outBuf))) { + needReInit = False; for (i = 0; i < numFilters; i++) if (p->ids[i] != block->filters[numFilters - 1 - i].id) + { + needReInit = True; break; - needReInit = (i != numFilters); + } } + + // p->SingleBufMode = (outBuf != NULL); + p->outBuf = outBuf; + p->outBufSize = outBufSize; + + // p->SingleBufMode = False; + // outBuf = NULL; if (needReInit) { MixCoder_Free(p); - p->numCoders = numFilters; for (i = 0; i < numFilters; i++) { - const CXzFilter *f = &block->filters[numFilters - 1 - i]; - RINOK(MixCoder_SetFromMethod(p, i, f->id)); + RINOK(MixCoder_SetFromMethod(p, i, block->filters[numFilters - 1 - i].id, outBuf, outBufSize)) } + p->numCoders = numFilters; } - + else + { + RINOK(MixCoder_ResetFromMethod(p, 0, block->filters[numFilters - 1].id, outBuf, outBufSize)) + } + for (i = 0; i < numFilters; i++) { const CXzFilter *f = &block->filters[numFilters - 1 - i]; IStateCoder *sc = &p->coders[i]; - RINOK(sc->SetProps(sc->p, f->props, f->propsSize, p->alloc)); + RINOK(sc->SetProps(sc->p, f->props, f->propsSize, p->alloc)) } MixCoder_Init(p); return SZ_OK; } + + void XzUnpacker_Init(CXzUnpacker *p) { p->state = XZ_STATE_STREAM_HEADER; @@ -614,75 +957,171 @@ void XzUnpacker_Init(CXzUnpacker *p) p->numFinishedStreams = 0; p->numTotalBlocks = 0; p->padSize = 0; + p->decodeOnlyOneBlock = 0; + + p->parseMode = False; + p->decodeToStreamSignature = False; + + // p->outBuf = NULL; + // p->outBufSize = 0; + p->outDataWritten = 0; } -void XzUnpacker_Construct(CXzUnpacker *p, ISzAlloc *alloc) + +void XzUnpacker_SetOutBuf(CXzUnpacker *p, Byte *outBuf, size_t outBufSize) +{ + p->outBuf = outBuf; + p->outBufSize = outBufSize; +} + + +void XzUnpacker_Construct(CXzUnpacker *p, ISzAllocPtr alloc) { MixCoder_Construct(&p->decoder, alloc); + p->outBuf = NULL; + p->outBufSize = 0; XzUnpacker_Init(p); } + void XzUnpacker_Free(CXzUnpacker *p) { MixCoder_Free(&p->decoder); } + +void XzUnpacker_PrepareToRandomBlockDecoding(CXzUnpacker *p) +{ + p->indexSize = 0; + p->numBlocks = 0; + Sha256_Init(&p->sha); + p->state = XZ_STATE_BLOCK_HEADER; + p->pos = 0; + p->decodeOnlyOneBlock = 1; +} + + +static void XzUnpacker_UpdateIndex(CXzUnpacker *p, UInt64 packSize, UInt64 unpackSize) +{ + Byte temp[32]; + unsigned num = Xz_WriteVarInt(temp, packSize); + num += Xz_WriteVarInt(temp + num, unpackSize); + Sha256_Update(&p->sha, temp, num); + p->indexSize += num; + p->numBlocks++; +} + + + SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, - const Byte *src, SizeT *srcLen, ECoderFinishMode finishMode, ECoderStatus *status) + const Byte *src, SizeT *srcLen, int srcFinished, + ECoderFinishMode finishMode, ECoderStatus *status) { SizeT destLenOrig = *destLen; SizeT srcLenOrig = *srcLen; *destLen = 0; *srcLen = 0; *status = CODER_STATUS_NOT_SPECIFIED; + for (;;) { - SizeT srcRem = srcLenOrig - *srcLen; + SizeT srcRem; if (p->state == XZ_STATE_BLOCK) { SizeT destLen2 = destLenOrig - *destLen; SizeT srcLen2 = srcLenOrig - *srcLen; SRes res; + + ECoderFinishMode finishMode2 = finishMode; + BoolInt srcFinished2 = srcFinished; + BoolInt destFinish = False; + + if (p->block.packSize != (UInt64)(Int64)-1) + { + UInt64 rem = p->block.packSize - p->packSize; + if (srcLen2 >= rem) + { + srcFinished2 = True; + srcLen2 = (SizeT)rem; + } + if (rem == 0 && p->block.unpackSize == p->unpackSize) + return SZ_ERROR_DATA; + } + + if (p->block.unpackSize != (UInt64)(Int64)-1) + { + UInt64 rem = p->block.unpackSize - p->unpackSize; + if (destLen2 >= rem) + { + destFinish = True; + finishMode2 = CODER_FINISH_END; + destLen2 = (SizeT)rem; + } + } + + /* if (srcLen2 == 0 && destLen2 == 0) { *status = CODER_STATUS_NOT_FINISHED; return SZ_OK; } + */ - res = MixCoder_Code(&p->decoder, dest, &destLen2, src, &srcLen2, False, finishMode, status); - XzCheck_Update(&p->check, dest, destLen2); + { + res = MixCoder_Code(&p->decoder, + (p->outBuf ? NULL : dest), &destLen2, destFinish, + src, &srcLen2, srcFinished2, + finishMode2); + + *status = p->decoder.status; + XzCheck_Update(&p->check, (p->outBuf ? p->outBuf + p->outDataWritten : dest), destLen2); + if (!p->outBuf) + dest += destLen2; + p->outDataWritten += destLen2; + } (*srcLen) += srcLen2; src += srcLen2; p->packSize += srcLen2; - (*destLen) += destLen2; - dest += destLen2; p->unpackSize += destLen2; - - RINOK(res); - - if (*status == CODER_STATUS_FINISHED_WITH_MARK) - { - Byte temp[32]; - unsigned num = Xz_WriteVarInt(temp, p->packSize + p->blockHeaderSize + XzFlags_GetCheckSize(p->streamFlags)); - num += Xz_WriteVarInt(temp + num, p->unpackSize); - Sha256_Update(&p->sha, temp, num); - p->indexSize += num; - p->numBlocks++; + + RINOK(res) + + if (*status != CODER_STATUS_FINISHED_WITH_MARK) + { + if (p->block.packSize == p->packSize + && *status == CODER_STATUS_NEEDS_MORE_INPUT) + { + PRF_STR("CODER_STATUS_NEEDS_MORE_INPUT") + *status = CODER_STATUS_NOT_SPECIFIED; + return SZ_ERROR_DATA; + } + return SZ_OK; + } + { + XzUnpacker_UpdateIndex(p, XzUnpacker_GetPackSizeForIndex(p), p->unpackSize); p->state = XZ_STATE_BLOCK_FOOTER; p->pos = 0; p->alignPos = 0; + *status = CODER_STATUS_NOT_SPECIFIED; + + if ((p->block.packSize != (UInt64)(Int64)-1 && p->block.packSize != p->packSize) + || (p->block.unpackSize != (UInt64)(Int64)-1 && p->block.unpackSize != p->unpackSize)) + { + PRF_STR("ERROR: block.size mismatch") + return SZ_ERROR_DATA; + } } - else if (srcLen2 == 0 && destLen2 == 0) - return SZ_OK; - - continue; + // continue; } - if (srcRem == 0) + srcRem = srcLenOrig - *srcLen; + + // XZ_STATE_BLOCK_FOOTER can transit to XZ_STATE_BLOCK_HEADER without input bytes + if (srcRem == 0 && p->state != XZ_STATE_BLOCK_FOOTER) { *status = CODER_STATUS_NEEDS_MORE_INPUT; return SZ_OK; @@ -696,17 +1135,19 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, { if (p->pos < XZ_SIG_SIZE && *src != XZ_SIG[p->pos]) return SZ_ERROR_NO_ARCHIVE; + if (p->decodeToStreamSignature) + return SZ_OK; p->buf[p->pos++] = *src++; (*srcLen)++; } else { - RINOK(Xz_ParseHeader(&p->streamFlags, p->buf)); + RINOK(Xz_ParseHeader(&p->streamFlags, p->buf)) p->numStartedStreams++; - p->state = XZ_STATE_BLOCK_HEADER; - Sha256_Init(&p->sha); p->indexSize = 0; p->numBlocks = 0; + Sha256_Init(&p->sha); + p->state = XZ_STATE_BLOCK_HEADER; p->pos = 0; } break; @@ -720,6 +1161,8 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, (*srcLen)++; if (p->buf[0] == 0) { + if (p->decodeOnlyOneBlock) + return SZ_ERROR_DATA; p->indexPreSize = 1 + Xz_WriteVarInt(p->buf + 1, p->numBlocks); p->indexPos = p->indexPreSize; p->indexSize += p->indexPreSize; @@ -727,10 +1170,13 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, Sha256_Init(&p->sha); p->crc = CrcUpdate(CRC_INIT_VAL, p->buf, p->indexPreSize); p->state = XZ_STATE_STREAM_INDEX; + break; } p->blockHeaderSize = ((UInt32)p->buf[0] << 2) + 4; + break; } - else if (p->pos != p->blockHeaderSize) + + if (p->pos != p->blockHeaderSize) { UInt32 cur = p->blockHeaderSize - p->pos; if (cur > srcRem) @@ -742,21 +1188,33 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, } else { - RINOK(XzBlock_Parse(&p->block, p->buf)); + RINOK(XzBlock_Parse(&p->block, p->buf)) + if (!XzBlock_AreSupportedFilters(&p->block)) + return SZ_ERROR_UNSUPPORTED; p->numTotalBlocks++; p->state = XZ_STATE_BLOCK; p->packSize = 0; p->unpackSize = 0; XzCheck_Init(&p->check, XzFlags_GetCheckType(p->streamFlags)); - RINOK(XzDec_Init(&p->decoder, &p->block)); + if (p->parseMode) + { + p->headerParsedOk = True; + return SZ_OK; + } + RINOK(XzDecMix_Init(&p->decoder, &p->block, p->outBuf, p->outBufSize)) } break; } case XZ_STATE_BLOCK_FOOTER: { - if (((p->packSize + p->alignPos) & 3) != 0) + if ((((unsigned)p->packSize + p->alignPos) & 3) != 0) { + if (srcRem == 0) + { + *status = CODER_STATUS_NEEDS_MORE_INPUT; + return SZ_OK; + } (*srcLen)++; p->alignPos++; if (*src++ != 0) @@ -768,20 +1226,31 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, UInt32 cur = checkSize - p->pos; if (cur != 0) { + if (srcRem == 0) + { + *status = CODER_STATUS_NEEDS_MORE_INPUT; + return SZ_OK; + } if (cur > srcRem) cur = (UInt32)srcRem; memcpy(p->buf + p->pos, src, cur); p->pos += cur; (*srcLen) += cur; src += cur; + if (checkSize != p->pos) + break; } - else { Byte digest[XZ_CHECK_SIZE_MAX]; p->state = XZ_STATE_BLOCK_HEADER; p->pos = 0; if (XzCheck_Final(&p->check, digest) && memcmp(digest, p->buf, checkSize) != 0) return SZ_ERROR_CRC; + if (p->decodeOnlyOneBlock) + { + *status = CODER_STATUS_FINISHED_WITH_MARK; + return SZ_OK; + } } } break; @@ -841,9 +1310,10 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, } else { + const Byte *ptr = p->buf; p->state = XZ_STATE_STREAM_FOOTER; p->pos = 0; - if (CRC_GET_DIGEST(p->crc) != GetUi32(p->buf)) + if (CRC_GET_DIGEST(p->crc) != GetUi32(ptr)) return SZ_ERROR_CRC; } break; @@ -897,17 +1367,1509 @@ SRes XzUnpacker_Code(CXzUnpacker *p, Byte *dest, SizeT *destLen, */ } -Bool XzUnpacker_IsStreamWasFinished(CXzUnpacker *p) + +SRes XzUnpacker_CodeFull(CXzUnpacker *p, Byte *dest, SizeT *destLen, + const Byte *src, SizeT *srcLen, + ECoderFinishMode finishMode, ECoderStatus *status) +{ + XzUnpacker_Init(p); + XzUnpacker_SetOutBuf(p, dest, *destLen); + + return XzUnpacker_Code(p, + NULL, destLen, + src, srcLen, True, + finishMode, status); +} + + +BoolInt XzUnpacker_IsBlockFinished(const CXzUnpacker *p) +{ + return (p->state == XZ_STATE_BLOCK_HEADER) && (p->pos == 0); +} + +BoolInt XzUnpacker_IsStreamWasFinished(const CXzUnpacker *p) { return (p->state == XZ_STATE_STREAM_PADDING) && (((UInt32)p->padSize & 3) == 0); } -UInt64 XzUnpacker_GetExtraSize(CXzUnpacker *p) +UInt64 XzUnpacker_GetExtraSize(const CXzUnpacker *p) { UInt64 num = 0; if (p->state == XZ_STATE_STREAM_PADDING) - num += p->padSize; + num = p->padSize; else if (p->state == XZ_STATE_STREAM_HEADER) - num += p->padSize + p->pos; + num = p->padSize + p->pos; return num; } + + + + + + + + + + + + + + + + + + + + + +#ifndef Z7_ST +#include "MtDec.h" +#endif + + +void XzDecMtProps_Init(CXzDecMtProps *p) +{ + p->inBufSize_ST = 1 << 18; + p->outStep_ST = 1 << 20; + p->ignoreErrors = False; + + #ifndef Z7_ST + p->numThreads = 1; + p->inBufSize_MT = 1 << 18; + p->memUseMax = sizeof(size_t) << 28; + #endif +} + + + +#ifndef Z7_ST + +/* ---------- CXzDecMtThread ---------- */ + +typedef struct +{ + Byte *outBuf; + size_t outBufSize; + size_t outPreSize; + size_t inPreSize; + size_t inPreHeaderSize; + size_t blockPackSize_for_Index; // including block header and checksum. + size_t blockPackTotal; // including stream header, block header and checksum. + size_t inCodeSize; + size_t outCodeSize; + ECoderStatus status; + SRes codeRes; + BoolInt skipMode; + // BoolInt finishedWithMark; + EMtDecParseState parseState; + BoolInt parsing_Truncated; + BoolInt atBlockHeader; + CXzStreamFlags streamFlags; + // UInt64 numFinishedStreams + UInt64 numStreams; + UInt64 numTotalBlocks; + UInt64 numBlocks; + + BoolInt dec_created; + CXzUnpacker dec; + + Byte mtPad[1 << 7]; +} CXzDecMtThread; + +#endif + + +/* ---------- CXzDecMt ---------- */ + +struct CXzDecMt +{ + CAlignOffsetAlloc alignOffsetAlloc; + ISzAllocPtr allocMid; + + CXzDecMtProps props; + size_t unpackBlockMaxSize; + + ISeqInStreamPtr inStream; + ISeqOutStreamPtr outStream; + ICompressProgressPtr progress; + + BoolInt finishMode; + BoolInt outSize_Defined; + UInt64 outSize; + + UInt64 outProcessed; + UInt64 inProcessed; + UInt64 readProcessed; + BoolInt readWasFinished; + SRes readRes; + SRes writeRes; + + Byte *outBuf; + size_t outBufSize; + Byte *inBuf; + size_t inBufSize; + + CXzUnpacker dec; + + ECoderStatus status; + SRes codeRes; + + #ifndef Z7_ST + BoolInt mainDecoderWasCalled; + // int statErrorDefined; + int finishedDecoderIndex; + + // global values that are used in Parse stage + CXzStreamFlags streamFlags; + // UInt64 numFinishedStreams + UInt64 numStreams; + UInt64 numTotalBlocks; + UInt64 numBlocks; + + // UInt64 numBadBlocks; + SRes mainErrorCode; // it's set to error code, if the size Code() output doesn't patch the size from Parsing stage + // it can be = SZ_ERROR_INPUT_EOF + // it can be = SZ_ERROR_DATA, in some another cases + BoolInt isBlockHeaderState_Parse; + BoolInt isBlockHeaderState_Write; + UInt64 outProcessed_Parse; + BoolInt parsing_Truncated; + + BoolInt mtc_WasConstructed; + CMtDec mtc; + CXzDecMtThread coders[MTDEC_THREADS_MAX]; + #endif +}; + + + +CXzDecMtHandle XzDecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid) +{ + CXzDecMt *p = (CXzDecMt *)ISzAlloc_Alloc(alloc, sizeof(CXzDecMt)); + if (!p) + return NULL; + + AlignOffsetAlloc_CreateVTable(&p->alignOffsetAlloc); + p->alignOffsetAlloc.baseAlloc = alloc; + p->alignOffsetAlloc.numAlignBits = 7; + p->alignOffsetAlloc.offset = 0; + + p->allocMid = allocMid; + + p->outBuf = NULL; + p->outBufSize = 0; + p->inBuf = NULL; + p->inBufSize = 0; + + XzUnpacker_Construct(&p->dec, &p->alignOffsetAlloc.vt); + + p->unpackBlockMaxSize = 0; + + XzDecMtProps_Init(&p->props); + + #ifndef Z7_ST + p->mtc_WasConstructed = False; + { + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CXzDecMtThread *coder = &p->coders[i]; + coder->dec_created = False; + coder->outBuf = NULL; + coder->outBufSize = 0; + } + } + #endif + + return (CXzDecMtHandle)p; +} + + +#ifndef Z7_ST + +static void XzDecMt_FreeOutBufs(CXzDecMt *p) +{ + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CXzDecMtThread *coder = &p->coders[i]; + if (coder->outBuf) + { + ISzAlloc_Free(p->allocMid, coder->outBuf); + coder->outBuf = NULL; + coder->outBufSize = 0; + } + } + p->unpackBlockMaxSize = 0; +} + +#endif + + + +static void XzDecMt_FreeSt(CXzDecMt *p) +{ + XzUnpacker_Free(&p->dec); + + if (p->outBuf) + { + ISzAlloc_Free(p->allocMid, p->outBuf); + p->outBuf = NULL; + } + p->outBufSize = 0; + + if (p->inBuf) + { + ISzAlloc_Free(p->allocMid, p->inBuf); + p->inBuf = NULL; + } + p->inBufSize = 0; +} + + +// #define GET_CXzDecMt_p CXzDecMt *p = pp; + +void XzDecMt_Destroy(CXzDecMtHandle p) +{ + // GET_CXzDecMt_p + + XzDecMt_FreeSt(p); + + #ifndef Z7_ST + + if (p->mtc_WasConstructed) + { + MtDec_Destruct(&p->mtc); + p->mtc_WasConstructed = False; + } + { + unsigned i; + for (i = 0; i < MTDEC_THREADS_MAX; i++) + { + CXzDecMtThread *t = &p->coders[i]; + if (t->dec_created) + { + // we don't need to free dict here + XzUnpacker_Free(&t->dec); + t->dec_created = False; + } + } + } + XzDecMt_FreeOutBufs(p); + + #endif + + ISzAlloc_Free(p->alignOffsetAlloc.baseAlloc, p); +} + + + +#ifndef Z7_ST + +static void XzDecMt_Callback_Parse(void *obj, unsigned coderIndex, CMtDecCallbackInfo *cc) +{ + CXzDecMt *me = (CXzDecMt *)obj; + CXzDecMtThread *coder = &me->coders[coderIndex]; + size_t srcSize = cc->srcSize; + + cc->srcSize = 0; + cc->outPos = 0; + cc->state = MTDEC_PARSE_CONTINUE; + + cc->canCreateNewThread = True; + + if (cc->startCall) + { + coder->outPreSize = 0; + coder->inPreSize = 0; + coder->inPreHeaderSize = 0; + coder->parseState = MTDEC_PARSE_CONTINUE; + coder->parsing_Truncated = False; + coder->skipMode = False; + coder->codeRes = SZ_OK; + coder->status = CODER_STATUS_NOT_SPECIFIED; + coder->inCodeSize = 0; + coder->outCodeSize = 0; + + coder->numStreams = me->numStreams; + coder->numTotalBlocks = me->numTotalBlocks; + coder->numBlocks = me->numBlocks; + + if (!coder->dec_created) + { + XzUnpacker_Construct(&coder->dec, &me->alignOffsetAlloc.vt); + coder->dec_created = True; + } + + XzUnpacker_Init(&coder->dec); + + if (me->isBlockHeaderState_Parse) + { + coder->dec.streamFlags = me->streamFlags; + coder->atBlockHeader = True; + XzUnpacker_PrepareToRandomBlockDecoding(&coder->dec); + } + else + { + coder->atBlockHeader = False; + me->isBlockHeaderState_Parse = True; + } + + coder->dec.numStartedStreams = me->numStreams; + coder->dec.numTotalBlocks = me->numTotalBlocks; + coder->dec.numBlocks = me->numBlocks; + } + + while (!coder->skipMode) + { + ECoderStatus status; + SRes res; + size_t srcSize2 = srcSize; + size_t destSize = (size_t)0 - 1; + + coder->dec.parseMode = True; + coder->dec.headerParsedOk = False; + + PRF_STR_INT("Parse", srcSize2) + + res = XzUnpacker_Code(&coder->dec, + NULL, &destSize, + cc->src, &srcSize2, cc->srcFinished, + CODER_FINISH_END, &status); + + // PRF(printf(" res = %d, srcSize2 = %d", res, (unsigned)srcSize2)); + + coder->codeRes = res; + coder->status = status; + cc->srcSize += srcSize2; + srcSize -= srcSize2; + coder->inPreHeaderSize += srcSize2; + coder->inPreSize = coder->inPreHeaderSize; + + if (res != SZ_OK) + { + cc->state = + coder->parseState = MTDEC_PARSE_END; + /* + if (res == SZ_ERROR_MEM) + return res; + return SZ_OK; + */ + return; // res; + } + + if (coder->dec.headerParsedOk) + { + const CXzBlock *block = &coder->dec.block; + if (XzBlock_HasUnpackSize(block) + // && block->unpackSize <= me->props.outBlockMax + && XzBlock_HasPackSize(block)) + { + { + if (block->unpackSize * 2 * me->mtc.numStartedThreads > me->props.memUseMax) + { + cc->state = MTDEC_PARSE_OVERFLOW; + return; // SZ_OK; + } + } + { + UInt64 packSize = block->packSize; + UInt64 packSizeAligned = packSize + ((0 - (unsigned)packSize) & 3); + UInt32 checkSize = XzFlags_GetCheckSize(coder->dec.streamFlags); + UInt64 blockPackSum = coder->inPreSize + packSizeAligned + checkSize; + // if (blockPackSum <= me->props.inBlockMax) + // unpackBlockMaxSize + { + coder->blockPackSize_for_Index = (size_t)(coder->dec.blockHeaderSize + packSize + checkSize); + coder->blockPackTotal = (size_t)blockPackSum; + coder->outPreSize = (size_t)block->unpackSize; + coder->streamFlags = coder->dec.streamFlags; + me->streamFlags = coder->dec.streamFlags; + coder->skipMode = True; + break; + } + } + } + } + else + // if (coder->inPreSize <= me->props.inBlockMax) + { + if (!cc->srcFinished) + return; // SZ_OK; + cc->state = + coder->parseState = MTDEC_PARSE_END; + return; // SZ_OK; + } + cc->state = MTDEC_PARSE_OVERFLOW; + return; // SZ_OK; + } + + // ---------- skipMode ---------- + { + UInt64 rem = coder->blockPackTotal - coder->inPreSize; + size_t cur = srcSize; + if (cur > rem) + cur = (size_t)rem; + cc->srcSize += cur; + coder->inPreSize += cur; + srcSize -= cur; + + if (coder->inPreSize == coder->blockPackTotal) + { + if (srcSize == 0) + { + if (!cc->srcFinished) + return; // SZ_OK; + cc->state = MTDEC_PARSE_END; + } + else if ((cc->src)[cc->srcSize] == 0) // we check control byte of next block + cc->state = MTDEC_PARSE_END; + else + { + cc->state = MTDEC_PARSE_NEW; + + { + size_t blockMax = me->unpackBlockMaxSize; + if (blockMax < coder->outPreSize) + blockMax = coder->outPreSize; + { + UInt64 required = (UInt64)blockMax * (me->mtc.numStartedThreads + 1) * 2; + if (me->props.memUseMax < required) + cc->canCreateNewThread = False; + } + } + + if (me->outSize_Defined) + { + // next block can be zero size + const UInt64 rem2 = me->outSize - me->outProcessed_Parse; + if (rem2 < coder->outPreSize) + { + coder->parsing_Truncated = True; + cc->state = MTDEC_PARSE_END; + } + me->outProcessed_Parse += coder->outPreSize; + } + } + } + else if (cc->srcFinished) + cc->state = MTDEC_PARSE_END; + else + return; // SZ_OK; + + coder->parseState = cc->state; + cc->outPos = coder->outPreSize; + + me->numStreams = coder->dec.numStartedStreams; + me->numTotalBlocks = coder->dec.numTotalBlocks; + me->numBlocks = coder->dec.numBlocks + 1; + return; // SZ_OK; + } +} + + +static SRes XzDecMt_Callback_PreCode(void *pp, unsigned coderIndex) +{ + CXzDecMt *me = (CXzDecMt *)pp; + CXzDecMtThread *coder = &me->coders[coderIndex]; + Byte *dest; + + if (!coder->dec.headerParsedOk) + return SZ_OK; + + dest = coder->outBuf; + + if (!dest || coder->outBufSize < coder->outPreSize) + { + if (dest) + { + ISzAlloc_Free(me->allocMid, dest); + coder->outBuf = NULL; + coder->outBufSize = 0; + } + { + size_t outPreSize = coder->outPreSize; + if (outPreSize == 0) + outPreSize = 1; + dest = (Byte *)ISzAlloc_Alloc(me->allocMid, outPreSize); + } + if (!dest) + return SZ_ERROR_MEM; + coder->outBuf = dest; + coder->outBufSize = coder->outPreSize; + + if (coder->outBufSize > me->unpackBlockMaxSize) + me->unpackBlockMaxSize = coder->outBufSize; + } + + // return SZ_ERROR_MEM; + + XzUnpacker_SetOutBuf(&coder->dec, coder->outBuf, coder->outBufSize); + + { + SRes res = XzDecMix_Init(&coder->dec.decoder, &coder->dec.block, coder->outBuf, coder->outBufSize); + // res = SZ_ERROR_UNSUPPORTED; // to test + coder->codeRes = res; + if (res != SZ_OK) + { + // if (res == SZ_ERROR_MEM) return res; + if (me->props.ignoreErrors && res != SZ_ERROR_MEM) + return SZ_OK; + return res; + } + } + + return SZ_OK; +} + + +static SRes XzDecMt_Callback_Code(void *pp, unsigned coderIndex, + const Byte *src, size_t srcSize, int srcFinished, + // int finished, int blockFinished, + UInt64 *inCodePos, UInt64 *outCodePos, int *stop) +{ + CXzDecMt *me = (CXzDecMt *)pp; + CXzDecMtThread *coder = &me->coders[coderIndex]; + + *inCodePos = coder->inCodeSize; + *outCodePos = coder->outCodeSize; + *stop = True; + + if (srcSize > coder->inPreSize - coder->inCodeSize) + return SZ_ERROR_FAIL; + + if (coder->inCodeSize < coder->inPreHeaderSize) + { + size_t step = coder->inPreHeaderSize - coder->inCodeSize; + if (step > srcSize) + step = srcSize; + src += step; + srcSize -= step; + coder->inCodeSize += step; + *inCodePos = coder->inCodeSize; + if (coder->inCodeSize < coder->inPreHeaderSize) + { + *stop = False; + return SZ_OK; + } + } + + if (!coder->dec.headerParsedOk) + return SZ_OK; + if (!coder->outBuf) + return SZ_OK; + + if (coder->codeRes == SZ_OK) + { + ECoderStatus status; + SRes res; + size_t srcProcessed = srcSize; + size_t outSizeCur = coder->outPreSize - coder->dec.outDataWritten; + + // PRF(printf("\nCallback_Code: Code %d %d\n", (unsigned)srcSize, (unsigned)outSizeCur)); + + res = XzUnpacker_Code(&coder->dec, + NULL, &outSizeCur, + src, &srcProcessed, srcFinished, + // coder->finishedWithMark ? CODER_FINISH_END : CODER_FINISH_ANY, + CODER_FINISH_END, + &status); + + // PRF(printf(" res = %d, srcSize2 = %d, outSizeCur = %d", res, (unsigned)srcProcessed, (unsigned)outSizeCur)); + + coder->codeRes = res; + coder->status = status; + coder->inCodeSize += srcProcessed; + coder->outCodeSize = coder->dec.outDataWritten; + *inCodePos = coder->inCodeSize; + *outCodePos = coder->outCodeSize; + + if (res == SZ_OK) + { + if (srcProcessed == srcSize) + *stop = False; + return SZ_OK; + } + } + + if (me->props.ignoreErrors && coder->codeRes != SZ_ERROR_MEM) + { + *inCodePos = coder->inPreSize; + *outCodePos = coder->outPreSize; + return SZ_OK; + } + return coder->codeRes; +} + + +#define XZDECMT_STREAM_WRITE_STEP (1 << 24) + +static SRes XzDecMt_Callback_Write(void *pp, unsigned coderIndex, + BoolInt needWriteToStream, + const Byte *src, size_t srcSize, BoolInt isCross, + // int srcFinished, + BoolInt *needContinue, + BoolInt *canRecode) +{ + CXzDecMt *me = (CXzDecMt *)pp; + const CXzDecMtThread *coder = &me->coders[coderIndex]; + + // PRF(printf("\nWrite processed = %d srcSize = %d\n", (unsigned)me->mtc.inProcessed, (unsigned)srcSize)); + + *needContinue = False; + *canRecode = True; + + if (!needWriteToStream) + return SZ_OK; + + if (!coder->dec.headerParsedOk || !coder->outBuf) + { + if (me->finishedDecoderIndex < 0) + me->finishedDecoderIndex = (int)coderIndex; + return SZ_OK; + } + + if (me->finishedDecoderIndex >= 0) + return SZ_OK; + + me->mtc.inProcessed += coder->inCodeSize; + + *canRecode = False; + + { + SRes res; + size_t size = coder->outCodeSize; + Byte *data = coder->outBuf; + + // we use in me->dec: sha, numBlocks, indexSize + + if (!me->isBlockHeaderState_Write) + { + XzUnpacker_PrepareToRandomBlockDecoding(&me->dec); + me->dec.decodeOnlyOneBlock = False; + me->dec.numStartedStreams = coder->dec.numStartedStreams; + me->dec.streamFlags = coder->streamFlags; + + me->isBlockHeaderState_Write = True; + } + + me->dec.numTotalBlocks = coder->dec.numTotalBlocks; + XzUnpacker_UpdateIndex(&me->dec, coder->blockPackSize_for_Index, coder->outPreSize); + + if (coder->outPreSize != size) + { + if (me->props.ignoreErrors) + { + memset(data + size, 0, coder->outPreSize - size); + size = coder->outPreSize; + } + // me->numBadBlocks++; + if (me->mainErrorCode == SZ_OK) + { + if ((int)coder->status == LZMA_STATUS_NEEDS_MORE_INPUT) + me->mainErrorCode = SZ_ERROR_INPUT_EOF; + else + me->mainErrorCode = SZ_ERROR_DATA; + } + } + + if (me->writeRes != SZ_OK) + return me->writeRes; + + res = SZ_OK; + { + if (me->outSize_Defined) + { + const UInt64 rem = me->outSize - me->outProcessed; + if (size > rem) + size = (SizeT)rem; + } + + for (;;) + { + size_t cur = size; + size_t written; + if (cur > XZDECMT_STREAM_WRITE_STEP) + cur = XZDECMT_STREAM_WRITE_STEP; + + written = ISeqOutStream_Write(me->outStream, data, cur); + + // PRF(printf("\nWritten ask = %d written = %d\n", (unsigned)cur, (unsigned)written)); + + me->outProcessed += written; + if (written != cur) + { + me->writeRes = SZ_ERROR_WRITE; + res = me->writeRes; + break; + } + data += cur; + size -= cur; + // PRF_STR_INT("Written size =", size) + if (size == 0) + break; + res = MtProgress_ProgressAdd(&me->mtc.mtProgress, 0, 0); + if (res != SZ_OK) + break; + } + } + + if (coder->codeRes != SZ_OK) + if (!me->props.ignoreErrors) + { + me->finishedDecoderIndex = (int)coderIndex; + return res; + } + + RINOK(res) + + if (coder->inPreSize != coder->inCodeSize + || coder->blockPackTotal != coder->inCodeSize) + { + me->finishedDecoderIndex = (int)coderIndex; + return SZ_OK; + } + + if (coder->parseState != MTDEC_PARSE_END) + { + *needContinue = True; + return SZ_OK; + } + } + + // (coder->state == MTDEC_PARSE_END) means that there are no other working threads + // so we can use mtc variables without lock + + PRF_STR_INT("Write MTDEC_PARSE_END", me->mtc.inProcessed) + + me->mtc.mtProgress.totalInSize = me->mtc.inProcessed; + { + CXzUnpacker *dec = &me->dec; + + PRF_STR_INT("PostSingle", srcSize) + + { + size_t srcProcessed = srcSize; + ECoderStatus status; + size_t outSizeCur = 0; + SRes res; + + // dec->decodeOnlyOneBlock = False; + dec->decodeToStreamSignature = True; + + me->mainDecoderWasCalled = True; + + if (coder->parsing_Truncated) + { + me->parsing_Truncated = True; + return SZ_OK; + } + + /* + We have processed all xz-blocks of stream, + And xz unpacker is at XZ_STATE_BLOCK_HEADER state, where + (src) is a pointer to xz-Index structure. + We finish reading of current xz-Stream, including Zero padding after xz-Stream. + We exit, if we reach extra byte (first byte of new-Stream or another data). + But we don't update input stream pointer for that new extra byte. + If extra byte is not correct first byte of xz-signature, + we have SZ_ERROR_NO_ARCHIVE error here. + */ + + res = XzUnpacker_Code(dec, + NULL, &outSizeCur, + src, &srcProcessed, + me->mtc.readWasFinished, // srcFinished + CODER_FINISH_END, // CODER_FINISH_ANY, + &status); + + // res = SZ_ERROR_ARCHIVE; // for failure test + + me->status = status; + me->codeRes = res; + + if (isCross) + me->mtc.crossStart += srcProcessed; + + me->mtc.inProcessed += srcProcessed; + me->mtc.mtProgress.totalInSize = me->mtc.inProcessed; + + srcSize -= srcProcessed; + src += srcProcessed; + + if (res != SZ_OK) + { + return SZ_OK; + // return res; + } + + if (dec->state == XZ_STATE_STREAM_HEADER) + { + *needContinue = True; + me->isBlockHeaderState_Parse = False; + me->isBlockHeaderState_Write = False; + + if (!isCross) + { + Byte *crossBuf = MtDec_GetCrossBuff(&me->mtc); + if (!crossBuf) + return SZ_ERROR_MEM; + if (srcSize != 0) + memcpy(crossBuf, src, srcSize); + me->mtc.crossStart = 0; + me->mtc.crossEnd = srcSize; + } + + PRF_STR_INT("XZ_STATE_STREAM_HEADER crossEnd = ", (unsigned)me->mtc.crossEnd) + + return SZ_OK; + } + + if (status != CODER_STATUS_NEEDS_MORE_INPUT || srcSize != 0) + { + return SZ_ERROR_FAIL; + } + + if (me->mtc.readWasFinished) + { + return SZ_OK; + } + } + + { + size_t inPos; + size_t inLim; + // const Byte *inData; + UInt64 inProgressPrev = me->mtc.inProcessed; + + // XzDecMt_Prepare_InBuf_ST(p); + Byte *crossBuf = MtDec_GetCrossBuff(&me->mtc); + if (!crossBuf) + return SZ_ERROR_MEM; + + inPos = 0; + inLim = 0; + + // inData = crossBuf; + + for (;;) + { + SizeT inProcessed; + SizeT outProcessed; + ECoderStatus status; + SRes res; + + if (inPos == inLim) + { + if (!me->mtc.readWasFinished) + { + inPos = 0; + inLim = me->mtc.inBufSize; + me->mtc.readRes = ISeqInStream_Read(me->inStream, (void *)crossBuf, &inLim); + me->mtc.readProcessed += inLim; + if (inLim == 0 || me->mtc.readRes != SZ_OK) + me->mtc.readWasFinished = True; + } + } + + inProcessed = inLim - inPos; + outProcessed = 0; + + res = XzUnpacker_Code(dec, + NULL, &outProcessed, + crossBuf + inPos, &inProcessed, + (inProcessed == 0), // srcFinished + CODER_FINISH_END, &status); + + me->codeRes = res; + me->status = status; + inPos += inProcessed; + me->mtc.inProcessed += inProcessed; + me->mtc.mtProgress.totalInSize = me->mtc.inProcessed; + + if (res != SZ_OK) + { + return SZ_OK; + // return res; + } + + if (dec->state == XZ_STATE_STREAM_HEADER) + { + *needContinue = True; + me->mtc.crossStart = inPos; + me->mtc.crossEnd = inLim; + me->isBlockHeaderState_Parse = False; + me->isBlockHeaderState_Write = False; + return SZ_OK; + } + + if (status != CODER_STATUS_NEEDS_MORE_INPUT) + return SZ_ERROR_FAIL; + + if (me->mtc.progress) + { + UInt64 inDelta = me->mtc.inProcessed - inProgressPrev; + if (inDelta >= (1 << 22)) + { + RINOK(MtProgress_Progress_ST(&me->mtc.mtProgress)) + inProgressPrev = me->mtc.inProcessed; + } + } + if (me->mtc.readWasFinished) + return SZ_OK; + } + } + } +} + + +#endif + + + +void XzStatInfo_Clear(CXzStatInfo *p) +{ + p->InSize = 0; + p->OutSize = 0; + + p->NumStreams = 0; + p->NumBlocks = 0; + + p->UnpackSize_Defined = False; + + p->NumStreams_Defined = False; + p->NumBlocks_Defined = False; + + p->DataAfterEnd = False; + p->DecodingTruncated = False; + + p->DecodeRes = SZ_OK; + p->ReadRes = SZ_OK; + p->ProgressRes = SZ_OK; + + p->CombinedRes = SZ_OK; + p->CombinedRes_Type = SZ_OK; +} + + + +/* + XzDecMt_Decode_ST() can return SZ_OK or the following errors + - SZ_ERROR_MEM for memory allocation error + - error from XzUnpacker_Code() function + - SZ_ERROR_WRITE for ISeqOutStream::Write(). stat->CombinedRes_Type = SZ_ERROR_WRITE in that case + - ICompressProgress::Progress() error, stat->CombinedRes_Type = SZ_ERROR_PROGRESS. + But XzDecMt_Decode_ST() doesn't return ISeqInStream::Read() errors. + ISeqInStream::Read() result is set to p->readRes. + also it can set stat->CombinedRes_Type to SZ_ERROR_WRITE or SZ_ERROR_PROGRESS. +*/ + +static SRes XzDecMt_Decode_ST(CXzDecMt *p + #ifndef Z7_ST + , BoolInt tMode + #endif + , CXzStatInfo *stat) +{ + size_t outPos; + size_t inPos, inLim; + const Byte *inData; + UInt64 inPrev, outPrev; + + CXzUnpacker *dec; + + #ifndef Z7_ST + if (tMode) + { + XzDecMt_FreeOutBufs(p); + tMode = MtDec_PrepareRead(&p->mtc); + } + #endif + + if (!p->outBuf || p->outBufSize != p->props.outStep_ST) + { + ISzAlloc_Free(p->allocMid, p->outBuf); + p->outBufSize = 0; + p->outBuf = (Byte *)ISzAlloc_Alloc(p->allocMid, p->props.outStep_ST); + if (!p->outBuf) + return SZ_ERROR_MEM; + p->outBufSize = p->props.outStep_ST; + } + + if (!p->inBuf || p->inBufSize != p->props.inBufSize_ST) + { + ISzAlloc_Free(p->allocMid, p->inBuf); + p->inBufSize = 0; + p->inBuf = (Byte *)ISzAlloc_Alloc(p->allocMid, p->props.inBufSize_ST); + if (!p->inBuf) + return SZ_ERROR_MEM; + p->inBufSize = p->props.inBufSize_ST; + } + + dec = &p->dec; + dec->decodeToStreamSignature = False; + // dec->decodeOnlyOneBlock = False; + + XzUnpacker_SetOutBuf(dec, NULL, 0); + + inPrev = p->inProcessed; + outPrev = p->outProcessed; + + inPos = 0; + inLim = 0; + inData = NULL; + outPos = 0; + + for (;;) + { + SizeT outSize; + BoolInt finished; + ECoderFinishMode finishMode; + SizeT inProcessed; + ECoderStatus status; + SRes res; + + SizeT outProcessed; + + + + if (inPos == inLim) + { + #ifndef Z7_ST + if (tMode) + { + inData = MtDec_Read(&p->mtc, &inLim); + inPos = 0; + if (inData) + continue; + tMode = False; + inLim = 0; + } + #endif + + if (!p->readWasFinished) + { + inPos = 0; + inLim = p->inBufSize; + inData = p->inBuf; + p->readRes = ISeqInStream_Read(p->inStream, (void *)p->inBuf, &inLim); + p->readProcessed += inLim; + if (inLim == 0 || p->readRes != SZ_OK) + p->readWasFinished = True; + } + } + + outSize = p->props.outStep_ST - outPos; + + finishMode = CODER_FINISH_ANY; + if (p->outSize_Defined) + { + const UInt64 rem = p->outSize - p->outProcessed; + if (outSize >= rem) + { + outSize = (SizeT)rem; + if (p->finishMode) + finishMode = CODER_FINISH_END; + } + } + + inProcessed = inLim - inPos; + outProcessed = outSize; + + res = XzUnpacker_Code(dec, p->outBuf + outPos, &outProcessed, + inData + inPos, &inProcessed, + (inPos == inLim), // srcFinished + finishMode, &status); + + p->codeRes = res; + p->status = status; + + inPos += inProcessed; + outPos += outProcessed; + p->inProcessed += inProcessed; + p->outProcessed += outProcessed; + + finished = ((inProcessed == 0 && outProcessed == 0) || res != SZ_OK); + + if (finished || outProcessed >= outSize) + if (outPos != 0) + { + const size_t written = ISeqOutStream_Write(p->outStream, p->outBuf, outPos); + // p->outProcessed += written; // 21.01: BUG fixed + if (written != outPos) + { + stat->CombinedRes_Type = SZ_ERROR_WRITE; + return SZ_ERROR_WRITE; + } + outPos = 0; + } + + if (p->progress && res == SZ_OK) + { + if (p->inProcessed - inPrev >= (1 << 22) || + p->outProcessed - outPrev >= (1 << 22)) + { + res = ICompressProgress_Progress(p->progress, p->inProcessed, p->outProcessed); + if (res != SZ_OK) + { + stat->CombinedRes_Type = SZ_ERROR_PROGRESS; + stat->ProgressRes = res; + return res; + } + inPrev = p->inProcessed; + outPrev = p->outProcessed; + } + } + + if (finished) + { + // p->codeRes is preliminary error from XzUnpacker_Code. + // and it can be corrected later as final result + // so we return SZ_OK here instead of (res); + return SZ_OK; + // return res; + } + } +} + + + +/* +XzStatInfo_SetStat() transforms + CXzUnpacker return code and status to combined CXzStatInfo results. + it can convert SZ_OK to SZ_ERROR_INPUT_EOF + it can convert SZ_ERROR_NO_ARCHIVE to SZ_OK and (DataAfterEnd = 1) +*/ + +static void XzStatInfo_SetStat(const CXzUnpacker *dec, + int finishMode, + // UInt64 readProcessed, + UInt64 inProcessed, + SRes res, // it's result from CXzUnpacker unpacker + ECoderStatus status, + BoolInt decodingTruncated, + CXzStatInfo *stat) +{ + UInt64 extraSize; + + stat->DecodingTruncated = (Byte)(decodingTruncated ? 1 : 0); + stat->InSize = inProcessed; + stat->NumStreams = dec->numStartedStreams; + stat->NumBlocks = dec->numTotalBlocks; + + stat->UnpackSize_Defined = True; + stat->NumStreams_Defined = True; + stat->NumBlocks_Defined = True; + + extraSize = XzUnpacker_GetExtraSize(dec); + + if (res == SZ_OK) + { + if (status == CODER_STATUS_NEEDS_MORE_INPUT) + { + // CODER_STATUS_NEEDS_MORE_INPUT is expected status for correct xz streams + // any extra data is part of correct data + extraSize = 0; + // if xz stream was not finished, then we need more data + if (!XzUnpacker_IsStreamWasFinished(dec)) + res = SZ_ERROR_INPUT_EOF; + } + else + { + // CODER_STATUS_FINISHED_WITH_MARK is not possible for multi stream xz decoding + // so he we have (status == CODER_STATUS_NOT_FINISHED) + // if (status != CODER_STATUS_FINISHED_WITH_MARK) + if (!decodingTruncated || finishMode) + res = SZ_ERROR_DATA; + } + } + else if (res == SZ_ERROR_NO_ARCHIVE) + { + /* + SZ_ERROR_NO_ARCHIVE is possible for 2 states: + XZ_STATE_STREAM_HEADER - if bad signature or bad CRC + XZ_STATE_STREAM_PADDING - if non-zero padding data + extraSize and inProcessed don't include "bad" byte + */ + // if (inProcessed == extraSize), there was no any good xz stream header, and we keep error + if (inProcessed != extraSize) // if there were good xz streams before error + { + // if (extraSize != 0 || readProcessed != inProcessed) + { + // he we suppose that all xz streams were finsihed OK, and we have + // some extra data after all streams + stat->DataAfterEnd = True; + res = SZ_OK; + } + } + } + + if (stat->DecodeRes == SZ_OK) + stat->DecodeRes = res; + + stat->InSize -= extraSize; +} + + + +SRes XzDecMt_Decode(CXzDecMtHandle p, + const CXzDecMtProps *props, + const UInt64 *outDataSize, int finishMode, + ISeqOutStreamPtr outStream, + // Byte *outBuf, size_t *outBufSize, + ISeqInStreamPtr inStream, + // const Byte *inData, size_t inDataSize, + CXzStatInfo *stat, + int *isMT, + ICompressProgressPtr progress) +{ + // GET_CXzDecMt_p + #ifndef Z7_ST + BoolInt tMode; + #endif + + XzStatInfo_Clear(stat); + + p->props = *props; + + p->inStream = inStream; + p->outStream = outStream; + p->progress = progress; + // p->stat = stat; + + p->outSize = 0; + p->outSize_Defined = False; + if (outDataSize) + { + p->outSize_Defined = True; + p->outSize = *outDataSize; + } + + p->finishMode = finishMode; + + // p->outSize = 457; p->outSize_Defined = True; p->finishMode = False; // for test + + p->writeRes = SZ_OK; + p->outProcessed = 0; + p->inProcessed = 0; + p->readProcessed = 0; + p->readWasFinished = False; + p->readRes = SZ_OK; + + p->codeRes = SZ_OK; + p->status = CODER_STATUS_NOT_SPECIFIED; + + XzUnpacker_Init(&p->dec); + + *isMT = False; + + /* + p->outBuf = NULL; + p->outBufSize = 0; + if (!outStream) + { + p->outBuf = outBuf; + p->outBufSize = *outBufSize; + *outBufSize = 0; + } + */ + + + #ifndef Z7_ST + + p->isBlockHeaderState_Parse = False; + p->isBlockHeaderState_Write = False; + // p->numBadBlocks = 0; + p->mainErrorCode = SZ_OK; + p->mainDecoderWasCalled = False; + + tMode = False; + + if (p->props.numThreads > 1) + { + IMtDecCallback2 vt; + BoolInt needContinue; + SRes res; + // we just free ST buffers here + // but we still keep state variables, that was set in XzUnpacker_Init() + XzDecMt_FreeSt(p); + + p->outProcessed_Parse = 0; + p->parsing_Truncated = False; + + p->numStreams = 0; + p->numTotalBlocks = 0; + p->numBlocks = 0; + p->finishedDecoderIndex = -1; + + if (!p->mtc_WasConstructed) + { + p->mtc_WasConstructed = True; + MtDec_Construct(&p->mtc); + } + + p->mtc.mtCallback = &vt; + p->mtc.mtCallbackObject = p; + + p->mtc.progress = progress; + p->mtc.inStream = inStream; + p->mtc.alloc = &p->alignOffsetAlloc.vt; + // p->mtc.inData = inData; + // p->mtc.inDataSize = inDataSize; + p->mtc.inBufSize = p->props.inBufSize_MT; + // p->mtc.inBlockMax = p->props.inBlockMax; + p->mtc.numThreadsMax = p->props.numThreads; + + *isMT = True; + + vt.Parse = XzDecMt_Callback_Parse; + vt.PreCode = XzDecMt_Callback_PreCode; + vt.Code = XzDecMt_Callback_Code; + vt.Write = XzDecMt_Callback_Write; + + + res = MtDec_Code(&p->mtc); + + + stat->InSize = p->mtc.inProcessed; + + p->inProcessed = p->mtc.inProcessed; + p->readRes = p->mtc.readRes; + p->readWasFinished = p->mtc.readWasFinished; + p->readProcessed = p->mtc.readProcessed; + + tMode = True; + needContinue = False; + + if (res == SZ_OK) + { + if (p->mtc.mtProgress.res != SZ_OK) + { + res = p->mtc.mtProgress.res; + stat->ProgressRes = res; + stat->CombinedRes_Type = SZ_ERROR_PROGRESS; + } + else + needContinue = p->mtc.needContinue; + } + + if (!needContinue) + { + { + SRes codeRes; + BoolInt truncated = False; + ECoderStatus status; + const CXzUnpacker *dec; + + stat->OutSize = p->outProcessed; + + if (p->finishedDecoderIndex >= 0) + { + const CXzDecMtThread *coder = &p->coders[(unsigned)p->finishedDecoderIndex]; + codeRes = coder->codeRes; + dec = &coder->dec; + status = coder->status; + } + else if (p->mainDecoderWasCalled) + { + codeRes = p->codeRes; + dec = &p->dec; + status = p->status; + truncated = p->parsing_Truncated; + } + else + return SZ_ERROR_FAIL; + + if (p->mainErrorCode != SZ_OK) + stat->DecodeRes = p->mainErrorCode; + + XzStatInfo_SetStat(dec, p->finishMode, + // p->mtc.readProcessed, + p->mtc.inProcessed, + codeRes, status, + truncated, + stat); + } + + if (res == SZ_OK) + { + stat->ReadRes = p->mtc.readRes; + + if (p->writeRes != SZ_OK) + { + res = p->writeRes; + stat->CombinedRes_Type = SZ_ERROR_WRITE; + } + else if (p->mtc.readRes != SZ_OK + // && p->mtc.inProcessed == p->mtc.readProcessed + && stat->DecodeRes == SZ_ERROR_INPUT_EOF) + { + res = p->mtc.readRes; + stat->CombinedRes_Type = SZ_ERROR_READ; + } + else if (stat->DecodeRes != SZ_OK) + res = stat->DecodeRes; + } + + stat->CombinedRes = res; + if (stat->CombinedRes_Type == SZ_OK) + stat->CombinedRes_Type = res; + return res; + } + + PRF_STR("----- decoding ST -----") + } + + #endif + + + *isMT = False; + + { + SRes res = XzDecMt_Decode_ST(p + #ifndef Z7_ST + , tMode + #endif + , stat + ); + + #ifndef Z7_ST + // we must set error code from MT decoding at first + if (p->mainErrorCode != SZ_OK) + stat->DecodeRes = p->mainErrorCode; + #endif + + XzStatInfo_SetStat(&p->dec, + p->finishMode, + // p->readProcessed, + p->inProcessed, + p->codeRes, p->status, + False, // truncated + stat); + + stat->ReadRes = p->readRes; + + if (res == SZ_OK) + { + if (p->readRes != SZ_OK + // && p->inProcessed == p->readProcessed + && stat->DecodeRes == SZ_ERROR_INPUT_EOF) + { + // we set read error as combined error, only if that error was the reason + // of decoding problem + res = p->readRes; + stat->CombinedRes_Type = SZ_ERROR_READ; + } + else if (stat->DecodeRes != SZ_OK) + res = stat->DecodeRes; + } + + stat->CombinedRes = res; + if (stat->CombinedRes_Type == SZ_OK) + stat->CombinedRes_Type = res; + return res; + } +} + +#undef PRF +#undef PRF_STR +#undef PRF_STR_INT_2 diff --git a/3rdparty/lzma/C/XzEnc.c b/3rdparty/lzma/C/XzEnc.c index ccac7a8ed4b..22408e26189 100644 --- a/3rdparty/lzma/C/XzEnc.c +++ b/3rdparty/lzma/C/XzEnc.c @@ -1,5 +1,5 @@ /* XzEnc.c -- Xz Encode -2015-09-16 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -7,7 +7,6 @@ #include <string.h> #include "7zCrc.h" -#include "Alloc.h" #include "Bra.h" #include "CpuArch.h" @@ -19,23 +18,42 @@ #include "XzEnc.h" -#define XzBlock_ClearFlags(p) (p)->flags = 0; -#define XzBlock_SetNumFilters(p, n) (p)->flags |= ((n) - 1); +// #define Z7_ST + +#ifndef Z7_ST +#include "MtCoder.h" +#else +#define MTCODER_THREADS_MAX 1 +#define MTCODER_BLOCKS_MAX 1 +#endif + +#define XZ_GET_PAD_SIZE(dataSize) ((4 - ((unsigned)(dataSize) & 3)) & 3) + +/* max pack size for LZMA2 block + check-64bytrs: */ +#define XZ_GET_MAX_BLOCK_PACK_SIZE(unpackSize) ((unpackSize) + ((unpackSize) >> 10) + 16 + 64) + +#define XZ_GET_ESTIMATED_BLOCK_TOTAL_PACK_SIZE(unpackSize) (XZ_BLOCK_HEADER_SIZE_MAX + XZ_GET_MAX_BLOCK_PACK_SIZE(unpackSize)) + + +// #define XzBlock_ClearFlags(p) (p)->flags = 0; +#define XzBlock_ClearFlags_SetNumFilters(p, n) (p)->flags = (Byte)((n) - 1); #define XzBlock_SetHasPackSize(p) (p)->flags |= XZ_BF_PACK_SIZE; #define XzBlock_SetHasUnpackSize(p) (p)->flags |= XZ_BF_UNPACK_SIZE; -static SRes WriteBytes(ISeqOutStream *s, const void *buf, UInt32 size) + +static SRes WriteBytes(ISeqOutStreamPtr s, const void *buf, size_t size) { - return (s->Write(s, buf, size) == size) ? SZ_OK : SZ_ERROR_WRITE; + return (ISeqOutStream_Write(s, buf, size) == size) ? SZ_OK : SZ_ERROR_WRITE; } -static SRes WriteBytesAndCrc(ISeqOutStream *s, const void *buf, UInt32 size, UInt32 *crc) +static SRes WriteBytes_UpdateCrc(ISeqOutStreamPtr s, const void *buf, size_t size, UInt32 *crc) { *crc = CrcUpdate(*crc, buf, size); return WriteBytes(s, buf, size); } -static SRes Xz_WriteHeader(CXzStreamFlags f, ISeqOutStream *s) + +static SRes Xz_WriteHeader(CXzStreamFlags f, ISeqOutStreamPtr s) { UInt32 crc; Byte header[XZ_STREAM_HEADER_SIZE]; @@ -43,12 +61,12 @@ static SRes Xz_WriteHeader(CXzStreamFlags f, ISeqOutStream *s) header[XZ_SIG_SIZE] = (Byte)(f >> 8); header[XZ_SIG_SIZE + 1] = (Byte)(f & 0xFF); crc = CrcCalc(header + XZ_SIG_SIZE, XZ_STREAM_FLAGS_SIZE); - SetUi32(header + XZ_SIG_SIZE + XZ_STREAM_FLAGS_SIZE, crc); + SetUi32(header + XZ_SIG_SIZE + XZ_STREAM_FLAGS_SIZE, crc) return WriteBytes(s, header, XZ_STREAM_HEADER_SIZE); } -static SRes XzBlock_WriteHeader(const CXzBlock *p, ISeqOutStream *s) +static SRes XzBlock_WriteHeader(const CXzBlock *p, ISeqOutStreamPtr s) { Byte header[XZ_BLOCK_HEADER_SIZE_MAX]; @@ -73,103 +91,159 @@ static SRes XzBlock_WriteHeader(const CXzBlock *p, ISeqOutStream *s) header[pos++] = 0; header[0] = (Byte)(pos >> 2); - SetUi32(header + pos, CrcCalc(header, pos)); + SetUi32(header + pos, CrcCalc(header, pos)) return WriteBytes(s, header, pos + 4); } -static SRes Xz_WriteFooter(CXzStream *p, ISeqOutStream *s) + + +typedef struct { - Byte buf[32]; - UInt64 globalPos; - { - UInt32 crc = CRC_INIT_VAL; - unsigned pos = 1 + Xz_WriteVarInt(buf + 1, p->numBlocks); - size_t i; + size_t numBlocks; + size_t size; + size_t allocated; + Byte *blocks; +} CXzEncIndex; - globalPos = pos; - buf[0] = 0; - RINOK(WriteBytesAndCrc(s, buf, pos, &crc)); - for (i = 0; i < p->numBlocks; i++) - { - const CXzBlockSizes *block = &p->blocks[i]; - pos = Xz_WriteVarInt(buf, block->totalSize); - pos += Xz_WriteVarInt(buf + pos, block->unpackSize); - globalPos += pos; - RINOK(WriteBytesAndCrc(s, buf, pos, &crc)); - } - - pos = ((unsigned)globalPos & 3); - - if (pos != 0) - { - buf[0] = buf[1] = buf[2] = 0; - RINOK(WriteBytesAndCrc(s, buf, 4 - pos, &crc)); - globalPos += 4 - pos; - } - { - SetUi32(buf, CRC_GET_DIGEST(crc)); - RINOK(WriteBytes(s, buf, 4)); - globalPos += 4; - } - } +static void XzEncIndex_Construct(CXzEncIndex *p) +{ + p->numBlocks = 0; + p->size = 0; + p->allocated = 0; + p->blocks = NULL; +} +static void XzEncIndex_Init(CXzEncIndex *p) +{ + p->numBlocks = 0; + p->size = 0; +} + +static void XzEncIndex_Free(CXzEncIndex *p, ISzAllocPtr alloc) +{ + if (p->blocks) { - UInt32 indexSize = (UInt32)((globalPos >> 2) - 1); - SetUi32(buf + 4, indexSize); - buf[8] = (Byte)(p->flags >> 8); - buf[9] = (Byte)(p->flags & 0xFF); - SetUi32(buf, CrcCalc(buf + 4, 6)); - memcpy(buf + 10, XZ_FOOTER_SIG, XZ_FOOTER_SIG_SIZE); - return WriteBytes(s, buf, 12); + ISzAlloc_Free(alloc, p->blocks); + p->blocks = NULL; } + p->numBlocks = 0; + p->size = 0; + p->allocated = 0; } -static SRes Xz_AddIndexRecord(CXzStream *p, UInt64 unpackSize, UInt64 totalSize, ISzAlloc *alloc) +static SRes XzEncIndex_ReAlloc(CXzEncIndex *p, size_t newSize, ISzAllocPtr alloc) { - if (!p->blocks || p->numBlocksAllocated == p->numBlocks) + Byte *blocks = (Byte *)ISzAlloc_Alloc(alloc, newSize); + if (!blocks) + return SZ_ERROR_MEM; + if (p->size != 0) + memcpy(blocks, p->blocks, p->size); + if (p->blocks) + ISzAlloc_Free(alloc, p->blocks); + p->blocks = blocks; + p->allocated = newSize; + return SZ_OK; +} + + +static SRes XzEncIndex_PreAlloc(CXzEncIndex *p, UInt64 numBlocks, UInt64 unpackSize, UInt64 totalSize, ISzAllocPtr alloc) +{ + UInt64 pos; { - size_t num = p->numBlocks * 2 + 1; - size_t newSize = sizeof(CXzBlockSizes) * num; - CXzBlockSizes *blocks; - if (newSize / sizeof(CXzBlockSizes) != num) - return SZ_ERROR_MEM; - blocks = (CXzBlockSizes *)alloc->Alloc(alloc, newSize); - if (!blocks) + Byte buf[32]; + unsigned pos2 = Xz_WriteVarInt(buf, totalSize); + pos2 += Xz_WriteVarInt(buf + pos2, unpackSize); + pos = numBlocks * pos2; + } + + if (pos <= p->allocated - p->size) + return SZ_OK; + { + UInt64 newSize64 = p->size + pos; + size_t newSize = (size_t)newSize64; + if (newSize != newSize64) return SZ_ERROR_MEM; - if (p->numBlocks != 0) - { - memcpy(blocks, p->blocks, p->numBlocks * sizeof(CXzBlockSizes)); - alloc->Free(alloc, p->blocks); - } - p->blocks = blocks; - p->numBlocksAllocated = num; + return XzEncIndex_ReAlloc(p, newSize, alloc); } +} + + +static SRes XzEncIndex_AddIndexRecord(CXzEncIndex *p, UInt64 unpackSize, UInt64 totalSize, ISzAllocPtr alloc) +{ + Byte buf[32]; + unsigned pos = Xz_WriteVarInt(buf, totalSize); + pos += Xz_WriteVarInt(buf + pos, unpackSize); + + if (pos > p->allocated - p->size) { - CXzBlockSizes *block = &p->blocks[p->numBlocks++]; - block->unpackSize = unpackSize; - block->totalSize = totalSize; + size_t newSize = p->allocated * 2 + 16 * 2; + if (newSize < p->size + pos) + return SZ_ERROR_MEM; + RINOK(XzEncIndex_ReAlloc(p, newSize, alloc)) } + memcpy(p->blocks + p->size, buf, pos); + p->size += pos; + p->numBlocks++; return SZ_OK; } +static SRes XzEncIndex_WriteFooter(const CXzEncIndex *p, CXzStreamFlags flags, ISeqOutStreamPtr s) +{ + Byte buf[32]; + UInt64 globalPos; + UInt32 crc = CRC_INIT_VAL; + unsigned pos = 1 + Xz_WriteVarInt(buf + 1, p->numBlocks); + + globalPos = pos; + buf[0] = 0; + RINOK(WriteBytes_UpdateCrc(s, buf, pos, &crc)) + RINOK(WriteBytes_UpdateCrc(s, p->blocks, p->size, &crc)) + globalPos += p->size; + + pos = XZ_GET_PAD_SIZE(globalPos); + buf[1] = 0; + buf[2] = 0; + buf[3] = 0; + globalPos += pos; + + crc = CrcUpdate(crc, buf + 4 - pos, pos); + SetUi32(buf + 4, CRC_GET_DIGEST(crc)) + + SetUi32(buf + 8 + 4, (UInt32)(globalPos >> 2)) + buf[8 + 8] = (Byte)(flags >> 8); + buf[8 + 9] = (Byte)(flags & 0xFF); + SetUi32(buf + 8, CrcCalc(buf + 8 + 4, 6)) + buf[8 + 10] = XZ_FOOTER_SIG_0; + buf[8 + 11] = XZ_FOOTER_SIG_1; + + return WriteBytes(s, buf + 4 - pos, pos + 4 + 12); +} + + + /* ---------- CSeqCheckInStream ---------- */ typedef struct { - ISeqInStream p; - ISeqInStream *realStream; + ISeqInStream vt; + ISeqInStreamPtr realStream; + const Byte *data; + UInt64 limit; UInt64 processed; + int realStreamFinished; CXzCheck check; } CSeqCheckInStream; -static void SeqCheckInStream_Init(CSeqCheckInStream *p, unsigned mode) +static void SeqCheckInStream_Init(CSeqCheckInStream *p, unsigned checkMode) { + p->limit = (UInt64)(Int64)-1; p->processed = 0; - XzCheck_Init(&p->check, mode); + p->realStreamFinished = 0; + XzCheck_Init(&p->check, checkMode); } static void SeqCheckInStream_GetDigest(CSeqCheckInStream *p, Byte *digest) @@ -177,12 +251,31 @@ static void SeqCheckInStream_GetDigest(CSeqCheckInStream *p, Byte *digest) XzCheck_Final(&p->check, digest); } -static SRes SeqCheckInStream_Read(void *pp, void *data, size_t *size) +static SRes SeqCheckInStream_Read(ISeqInStreamPtr pp, void *data, size_t *size) { - CSeqCheckInStream *p = (CSeqCheckInStream *)pp; - SRes res = p->realStream->Read(p->realStream, data, size); - XzCheck_Update(&p->check, data, *size); - p->processed += *size; + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSeqCheckInStream) + size_t size2 = *size; + SRes res = SZ_OK; + + if (p->limit != (UInt64)(Int64)-1) + { + UInt64 rem = p->limit - p->processed; + if (size2 > rem) + size2 = (size_t)rem; + } + if (size2 != 0) + { + if (p->realStream) + { + res = ISeqInStream_Read(p->realStream, data, &size2); + p->realStreamFinished = (size2 == 0) ? 1 : 0; + } + else + memcpy(data, p->data + (size_t)p->processed, size2); + XzCheck_Update(&p->check, data, size2); + p->processed += size2; + } + *size = size2; return res; } @@ -191,15 +284,24 @@ static SRes SeqCheckInStream_Read(void *pp, void *data, size_t *size) typedef struct { - ISeqOutStream p; - ISeqOutStream *realStream; + ISeqOutStream vt; + ISeqOutStreamPtr realStream; + Byte *outBuf; + size_t outBufLimit; UInt64 processed; } CSeqSizeOutStream; -static size_t MyWrite(void *pp, const void *data, size_t size) +static size_t SeqSizeOutStream_Write(ISeqOutStreamPtr pp, const void *data, size_t size) { - CSeqSizeOutStream *p = (CSeqSizeOutStream *)pp; - size = p->realStream->Write(p->realStream, data, size); + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSeqSizeOutStream) + if (p->realStream) + size = ISeqOutStream_Write(p->realStream, data, size); + else + { + if (size > p->outBufLimit - (size_t)p->processed) + return 0; + memcpy(p->outBuf + (size_t)p->processed, data, size); + } p->processed += size; return size; } @@ -211,8 +313,8 @@ static size_t MyWrite(void *pp, const void *data, size_t size) typedef struct { - ISeqInStream p; - ISeqInStream *realStream; + ISeqInStream vt; + ISeqInStreamPtr realStream; IStateCoder StateCoder; Byte *buf; size_t curPos; @@ -220,10 +322,62 @@ typedef struct int srcWasFinished; } CSeqInFilter; -static SRes SeqInFilter_Read(void *pp, void *data, size_t *size) + +static const z7_Func_BranchConv g_Funcs_BranchConv_RISC_Enc[] = { - CSeqInFilter *p = (CSeqInFilter *)pp; - size_t sizeOriginal = *size; + Z7_BRANCH_CONV_ENC(PPC), + Z7_BRANCH_CONV_ENC(IA64), + Z7_BRANCH_CONV_ENC(ARM), + Z7_BRANCH_CONV_ENC(ARMT), + Z7_BRANCH_CONV_ENC(SPARC), + Z7_BRANCH_CONV_ENC(ARM64) +}; + +static SizeT XzBcFilterStateBase_Filter_Enc(CXzBcFilterStateBase *p, Byte *data, SizeT size) +{ + switch (p->methodId) + { + case XZ_ID_Delta: + Delta_Encode(p->delta_State, p->delta, data, size); + break; + case XZ_ID_X86: + size = (SizeT)(z7_BranchConvSt_X86_Enc(data, size, p->ip, &p->X86_State) - data); + break; + default: + if (p->methodId >= XZ_ID_PPC) + { + const UInt32 i = p->methodId - XZ_ID_PPC; + if (i < Z7_ARRAY_SIZE(g_Funcs_BranchConv_RISC_Enc)) + size = (SizeT)(g_Funcs_BranchConv_RISC_Enc[i](data, size, p->ip) - data); + } + break; + } + p->ip += (UInt32)size; + return size; +} + + +static SRes SeqInFilter_Init(CSeqInFilter *p, const CXzFilter *props, ISzAllocPtr alloc) +{ + if (!p->buf) + { + p->buf = (Byte *)ISzAlloc_Alloc(alloc, FILTER_BUF_SIZE); + if (!p->buf) + return SZ_ERROR_MEM; + } + p->curPos = p->endPos = 0; + p->srcWasFinished = 0; + RINOK(Xz_StateCoder_Bc_SetFromMethod_Func(&p->StateCoder, props->id, XzBcFilterStateBase_Filter_Enc, alloc)) + RINOK(p->StateCoder.SetProps(p->StateCoder.p, props->props, props->propsSize, alloc)) + p->StateCoder.Init(p->StateCoder.p); + return SZ_OK; +} + + +static SRes SeqInFilter_Read(ISeqInStreamPtr pp, void *data, size_t *size) +{ + Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSeqInFilter) + const size_t sizeOriginal = *size; if (sizeOriginal == 0) return SZ_OK; *size = 0; @@ -234,19 +388,22 @@ static SRes SeqInFilter_Read(void *pp, void *data, size_t *size) { p->curPos = 0; p->endPos = FILTER_BUF_SIZE; - RINOK(p->realStream->Read(p->realStream, p->buf, &p->endPos)); + RINOK(ISeqInStream_Read(p->realStream, p->buf, &p->endPos)) if (p->endPos == 0) p->srcWasFinished = 1; } { SizeT srcLen = p->endPos - p->curPos; - int wasFinished; + ECoderStatus status; SRes res; *size = sizeOriginal; - res = p->StateCoder.Code(p->StateCoder.p, data, size, p->buf + p->curPos, &srcLen, - p->srcWasFinished, CODER_FINISH_ANY, &wasFinished); + res = p->StateCoder.Code2(p->StateCoder.p, + (Byte *)data, size, + p->buf + p->curPos, &srcLen, + p->srcWasFinished, CODER_FINISH_ANY, + &status); p->curPos += srcLen; - if (*size != 0 || srcLen == 0 || res != 0) + if (*size != 0 || srcLen == 0 || res != SZ_OK) return res; } } @@ -255,34 +412,22 @@ static SRes SeqInFilter_Read(void *pp, void *data, size_t *size) static void SeqInFilter_Construct(CSeqInFilter *p) { p->buf = NULL; - p->p.Read = SeqInFilter_Read; + p->StateCoder.p = NULL; + p->vt.Read = SeqInFilter_Read; } -static void SeqInFilter_Free(CSeqInFilter *p) +static void SeqInFilter_Free(CSeqInFilter *p, ISzAllocPtr alloc) { - if (p->buf) + if (p->StateCoder.p) { - g_Alloc.Free(&g_Alloc, p->buf); - p->buf = NULL; + p->StateCoder.Free(p->StateCoder.p, alloc); + p->StateCoder.p = NULL; } -} - -SRes BraState_SetFromMethod(IStateCoder *p, UInt64 id, int encodeMode, ISzAlloc *alloc); - -static SRes SeqInFilter_Init(CSeqInFilter *p, const CXzFilter *props) -{ - if (!p->buf) + if (p->buf) { - p->buf = g_Alloc.Alloc(&g_Alloc, FILTER_BUF_SIZE); - if (!p->buf) - return SZ_ERROR_MEM; + ISzAlloc_Free(alloc, p->buf); + p->buf = NULL; } - p->curPos = p->endPos = 0; - p->srcWasFinished = 0; - RINOK(BraState_SetFromMethod(&p->StateCoder, props->id, 1, &g_Alloc)); - RINOK(p->StateCoder.SetProps(p->StateCoder.p, props->props, props->propsSize, &g_Alloc)); - p->StateCoder.Init(p->StateCoder.p); - return SZ_OK; } @@ -292,24 +437,24 @@ static SRes SeqInFilter_Init(CSeqInFilter *p, const CXzFilter *props) typedef struct { - ISeqInStream p; - ISeqInStream *inStream; + ISeqInStream vt; + ISeqInStreamPtr inStream; CSbEnc enc; } CSbEncInStream; -static SRes SbEncInStream_Read(void *pp, void *data, size_t *size) +static SRes SbEncInStream_Read(ISeqInStreamPtr pp, void *data, size_t *size) { - CSbEncInStream *p = (CSbEncInStream *)pp; + CSbEncInStream *p = Z7_CONTAINER_FROM_VTBL(pp, CSbEncInStream, vt); size_t sizeOriginal = *size; if (sizeOriginal == 0) - return S_OK; + return SZ_OK; for (;;) { if (p->enc.needRead && !p->enc.readWasFinished) { size_t processed = p->enc.needReadSizeMax; - RINOK(p->inStream->Read(p->inStream, p->enc.buf + p->enc.readPos, &processed)); + RINOK(p->inStream->Read(p->inStream, p->enc.buf + p->enc.readPos, &processed)) p->enc.readPos += processed; if (processed == 0) { @@ -320,16 +465,16 @@ static SRes SbEncInStream_Read(void *pp, void *data, size_t *size) } *size = sizeOriginal; - RINOK(SbEnc_Read(&p->enc, data, size)); + RINOK(SbEnc_Read(&p->enc, data, size)) if (*size != 0 || !p->enc.needRead) - return S_OK; + return SZ_OK; } } -void SbEncInStream_Construct(CSbEncInStream *p, ISzAlloc *alloc) +void SbEncInStream_Construct(CSbEncInStream *p, ISzAllocPtr alloc) { SbEnc_Construct(&p->enc, alloc); - p->p.Read = SbEncInStream_Read; + p->vt.Read = SbEncInStream_Read; } SRes SbEncInStream_Init(CSbEncInStream *p) @@ -345,43 +490,233 @@ void SbEncInStream_Free(CSbEncInStream *p) #endif + +/* ---------- CXzProps ---------- */ + + +void XzFilterProps_Init(CXzFilterProps *p) +{ + p->id = 0; + p->delta = 0; + p->ip = 0; + p->ipDefined = False; +} + +void XzProps_Init(CXzProps *p) +{ + p->checkId = XZ_CHECK_CRC32; + p->blockSize = XZ_PROPS_BLOCK_SIZE_AUTO; + p->numBlockThreads_Reduced = -1; + p->numBlockThreads_Max = -1; + p->numTotalThreads = -1; + p->reduceSize = (UInt64)(Int64)-1; + p->forceWriteSizesInHeader = 0; + // p->forceWriteSizesInHeader = 1; + + XzFilterProps_Init(&p->filterProps); + Lzma2EncProps_Init(&p->lzma2Props); +} + + +static void XzEncProps_Normalize_Fixed(CXzProps *p) +{ + UInt64 fileSize; + int t1, t1n, t2, t2r, t3; + { + CLzma2EncProps tp = p->lzma2Props; + if (tp.numTotalThreads <= 0) + tp.numTotalThreads = p->numTotalThreads; + Lzma2EncProps_Normalize(&tp); + t1n = tp.numTotalThreads; + } + + t1 = p->lzma2Props.numTotalThreads; + t2 = p->numBlockThreads_Max; + t3 = p->numTotalThreads; + + if (t2 > MTCODER_THREADS_MAX) + t2 = MTCODER_THREADS_MAX; + + if (t3 <= 0) + { + if (t2 <= 0) + t2 = 1; + t3 = t1n * t2; + } + else if (t2 <= 0) + { + t2 = t3 / t1n; + if (t2 == 0) + { + t1 = 1; + t2 = t3; + } + if (t2 > MTCODER_THREADS_MAX) + t2 = MTCODER_THREADS_MAX; + } + else if (t1 <= 0) + { + t1 = t3 / t2; + if (t1 == 0) + t1 = 1; + } + else + t3 = t1n * t2; + + p->lzma2Props.numTotalThreads = t1; + + t2r = t2; + + fileSize = p->reduceSize; + + if ((p->blockSize < fileSize || fileSize == (UInt64)(Int64)-1)) + p->lzma2Props.lzmaProps.reduceSize = p->blockSize; + + Lzma2EncProps_Normalize(&p->lzma2Props); + + t1 = p->lzma2Props.numTotalThreads; + + { + if (t2 > 1 && fileSize != (UInt64)(Int64)-1) + { + UInt64 numBlocks = fileSize / p->blockSize; + if (numBlocks * p->blockSize != fileSize) + numBlocks++; + if (numBlocks < (unsigned)t2) + { + t2r = (int)numBlocks; + if (t2r == 0) + t2r = 1; + t3 = t1 * t2r; + } + } + } + + p->numBlockThreads_Max = t2; + p->numBlockThreads_Reduced = t2r; + p->numTotalThreads = t3; +} + + +static void XzProps_Normalize(CXzProps *p) +{ + /* we normalize xzProps properties, but we normalize only some of CXzProps::lzma2Props properties. + Lzma2Enc_SetProps() will normalize lzma2Props later. */ + + if (p->blockSize == XZ_PROPS_BLOCK_SIZE_SOLID) + { + p->lzma2Props.lzmaProps.reduceSize = p->reduceSize; + p->numBlockThreads_Reduced = 1; + p->numBlockThreads_Max = 1; + if (p->lzma2Props.numTotalThreads <= 0) + p->lzma2Props.numTotalThreads = p->numTotalThreads; + return; + } + else + { + CLzma2EncProps *lzma2 = &p->lzma2Props; + if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO) + { + // xz-auto + p->lzma2Props.lzmaProps.reduceSize = p->reduceSize; + + if (lzma2->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID) + { + // if (xz-auto && lzma2-solid) - we use solid for both + p->blockSize = XZ_PROPS_BLOCK_SIZE_SOLID; + p->numBlockThreads_Reduced = 1; + p->numBlockThreads_Max = 1; + if (p->lzma2Props.numTotalThreads <= 0) + p->lzma2Props.numTotalThreads = p->numTotalThreads; + } + else + { + // if (xz-auto && (lzma2-auto || lzma2-fixed_) + // we calculate block size for lzma2 and use that block size for xz, lzma2 uses single-chunk per block + CLzma2EncProps tp = p->lzma2Props; + if (tp.numTotalThreads <= 0) + tp.numTotalThreads = p->numTotalThreads; + + Lzma2EncProps_Normalize(&tp); + + p->blockSize = tp.blockSize; // fixed or solid + p->numBlockThreads_Reduced = tp.numBlockThreads_Reduced; + p->numBlockThreads_Max = tp.numBlockThreads_Max; + if (lzma2->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO) + lzma2->blockSize = tp.blockSize; // fixed or solid, LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID + if (lzma2->lzmaProps.reduceSize > tp.blockSize && tp.blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID) + lzma2->lzmaProps.reduceSize = tp.blockSize; + lzma2->numBlockThreads_Reduced = 1; + lzma2->numBlockThreads_Max = 1; + return; + } + } + else + { + // xz-fixed + // we can use xz::reduceSize or xz::blockSize as base for lzmaProps::reduceSize + + p->lzma2Props.lzmaProps.reduceSize = p->reduceSize; + { + UInt64 r = p->reduceSize; + if (r > p->blockSize || r == (UInt64)(Int64)-1) + r = p->blockSize; + lzma2->lzmaProps.reduceSize = r; + } + if (lzma2->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO) + lzma2->blockSize = LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID; + else if (lzma2->blockSize > p->blockSize && lzma2->blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID) + lzma2->blockSize = p->blockSize; + + XzEncProps_Normalize_Fixed(p); + } + } +} + + +/* ---------- CLzma2WithFilters ---------- */ + typedef struct { CLzma2EncHandle lzma2; + CSeqInFilter filter; + #ifdef USE_SUBBLOCK CSbEncInStream sb; #endif - CSeqInFilter filter; - ISzAlloc *alloc; - ISzAlloc *bigAlloc; } CLzma2WithFilters; -static void Lzma2WithFilters_Construct(CLzma2WithFilters *p, ISzAlloc *alloc, ISzAlloc *bigAlloc) +static void Lzma2WithFilters_Construct(CLzma2WithFilters *p) { - p->alloc = alloc; - p->bigAlloc = bigAlloc; p->lzma2 = NULL; + SeqInFilter_Construct(&p->filter); + #ifdef USE_SUBBLOCK SbEncInStream_Construct(&p->sb, alloc); #endif - SeqInFilter_Construct(&p->filter); } -static SRes Lzma2WithFilters_Create(CLzma2WithFilters *p) + +static SRes Lzma2WithFilters_Create(CLzma2WithFilters *p, ISzAllocPtr alloc, ISzAllocPtr bigAlloc) { - p->lzma2 = Lzma2Enc_Create(p->alloc, p->bigAlloc); if (!p->lzma2) - return SZ_ERROR_MEM; + { + p->lzma2 = Lzma2Enc_Create(alloc, bigAlloc); + if (!p->lzma2) + return SZ_ERROR_MEM; + } return SZ_OK; } -static void Lzma2WithFilters_Free(CLzma2WithFilters *p) + +static void Lzma2WithFilters_Free(CLzma2WithFilters *p, ISzAllocPtr alloc) { - SeqInFilter_Free(&p->filter); #ifdef USE_SUBBLOCK SbEncInStream_Free(&p->sb); #endif + + SeqInFilter_Free(&p->filter, alloc); if (p->lzma2) { Lzma2Enc_Destroy(p->lzma2); @@ -390,149 +725,638 @@ static void Lzma2WithFilters_Free(CLzma2WithFilters *p) } -void XzProps_Init(CXzProps *p) +typedef struct { - p->lzma2Props = NULL; - p->filterProps = NULL; - p->checkId = XZ_CHECK_CRC32; + UInt64 unpackSize; + UInt64 totalSize; + size_t headerSize; +} CXzEncBlockInfo; + + +static SRes Xz_CompressBlock( + CLzma2WithFilters *lzmaf, + + ISeqOutStreamPtr outStream, + Byte *outBufHeader, + Byte *outBufData, size_t outBufDataLimit, + + ISeqInStreamPtr inStream, + // UInt64 expectedSize, + const Byte *inBuf, // used if (!inStream) + size_t inBufSize, // used if (!inStream), it's block size, props->blockSize is ignored + + const CXzProps *props, + ICompressProgressPtr progress, + int *inStreamFinished, /* only for inStream version */ + CXzEncBlockInfo *blockSizes, + ISzAllocPtr alloc, + ISzAllocPtr allocBig) +{ + CSeqCheckInStream checkInStream; + CSeqSizeOutStream seqSizeOutStream; + CXzBlock block; + unsigned filterIndex = 0; + CXzFilter *filter = NULL; + const CXzFilterProps *fp = &props->filterProps; + if (fp->id == 0) + fp = NULL; + + *inStreamFinished = False; + + RINOK(Lzma2WithFilters_Create(lzmaf, alloc, allocBig)) + + RINOK(Lzma2Enc_SetProps(lzmaf->lzma2, &props->lzma2Props)) + + // XzBlock_ClearFlags(&block) + XzBlock_ClearFlags_SetNumFilters(&block, 1 + (fp ? 1 : 0)) + + if (fp) + { + filter = &block.filters[filterIndex++]; + filter->id = fp->id; + filter->propsSize = 0; + + if (fp->id == XZ_ID_Delta) + { + filter->props[0] = (Byte)(fp->delta - 1); + filter->propsSize = 1; + } + else if (fp->ipDefined) + { + Byte *ptr = filter->props; + SetUi32(ptr, fp->ip) + filter->propsSize = 4; + } + } + + { + CXzFilter *f = &block.filters[filterIndex++]; + f->id = XZ_ID_LZMA2; + f->propsSize = 1; + f->props[0] = Lzma2Enc_WriteProperties(lzmaf->lzma2); + } + + seqSizeOutStream.vt.Write = SeqSizeOutStream_Write; + seqSizeOutStream.realStream = outStream; + seqSizeOutStream.outBuf = outBufData; + seqSizeOutStream.outBufLimit = outBufDataLimit; + seqSizeOutStream.processed = 0; + + /* + if (expectedSize != (UInt64)(Int64)-1) + { + block.unpackSize = expectedSize; + if (props->blockSize != (UInt64)(Int64)-1) + if (expectedSize > props->blockSize) + block.unpackSize = props->blockSize; + XzBlock_SetHasUnpackSize(&block) + } + */ + + if (outStream) + { + RINOK(XzBlock_WriteHeader(&block, &seqSizeOutStream.vt)) + } + + checkInStream.vt.Read = SeqCheckInStream_Read; + SeqCheckInStream_Init(&checkInStream, props->checkId); + + checkInStream.realStream = inStream; + checkInStream.data = inBuf; + checkInStream.limit = props->blockSize; + if (!inStream) + checkInStream.limit = inBufSize; + + if (fp) + { + #ifdef USE_SUBBLOCK + if (fp->id == XZ_ID_Subblock) + { + lzmaf->sb.inStream = &checkInStream.vt; + RINOK(SbEncInStream_Init(&lzmaf->sb)) + } + else + #endif + { + lzmaf->filter.realStream = &checkInStream.vt; + RINOK(SeqInFilter_Init(&lzmaf->filter, filter, alloc)) + } + } + + { + SRes res; + Byte *outBuf = NULL; + size_t outSize = 0; + BoolInt useStream = (fp || inStream); + // useStream = True; + + if (!useStream) + { + XzCheck_Update(&checkInStream.check, inBuf, inBufSize); + checkInStream.processed = inBufSize; + } + + if (!outStream) + { + outBuf = seqSizeOutStream.outBuf; // + (size_t)seqSizeOutStream.processed; + outSize = seqSizeOutStream.outBufLimit; // - (size_t)seqSizeOutStream.processed; + } + + res = Lzma2Enc_Encode2(lzmaf->lzma2, + outBuf ? NULL : &seqSizeOutStream.vt, + outBuf, + outBuf ? &outSize : NULL, + + useStream ? + (fp ? + ( + #ifdef USE_SUBBLOCK + (fp->id == XZ_ID_Subblock) ? &lzmaf->sb.vt: + #endif + &lzmaf->filter.vt) : + &checkInStream.vt) : NULL, + + useStream ? NULL : inBuf, + useStream ? 0 : inBufSize, + + progress); + + if (outBuf) + seqSizeOutStream.processed += outSize; + + RINOK(res) + blockSizes->unpackSize = checkInStream.processed; + } + { + Byte buf[4 + 64]; + unsigned padSize = XZ_GET_PAD_SIZE(seqSizeOutStream.processed); + UInt64 packSize = seqSizeOutStream.processed; + + buf[0] = 0; + buf[1] = 0; + buf[2] = 0; + buf[3] = 0; + + SeqCheckInStream_GetDigest(&checkInStream, buf + 4); + RINOK(WriteBytes(&seqSizeOutStream.vt, buf + (4 - padSize), padSize + XzFlags_GetCheckSize((CXzStreamFlags)props->checkId))) + + blockSizes->totalSize = seqSizeOutStream.processed - padSize; + + if (!outStream) + { + seqSizeOutStream.outBuf = outBufHeader; + seqSizeOutStream.outBufLimit = XZ_BLOCK_HEADER_SIZE_MAX; + seqSizeOutStream.processed = 0; + + block.unpackSize = blockSizes->unpackSize; + XzBlock_SetHasUnpackSize(&block) + + block.packSize = packSize; + XzBlock_SetHasPackSize(&block) + + RINOK(XzBlock_WriteHeader(&block, &seqSizeOutStream.vt)) + + blockSizes->headerSize = (size_t)seqSizeOutStream.processed; + blockSizes->totalSize += seqSizeOutStream.processed; + } + } + + if (inStream) + *inStreamFinished = checkInStream.realStreamFinished; + else + { + *inStreamFinished = False; + if (checkInStream.processed != inBufSize) + return SZ_ERROR_FAIL; + } + + return SZ_OK; } -void XzFilterProps_Init(CXzFilterProps *p) + + +typedef struct { - p->id = 0; - p->delta = 0; - p->ip = 0; - p->ipDefined = False; + ICompressProgress vt; + ICompressProgressPtr progress; + UInt64 inOffset; + UInt64 outOffset; +} CCompressProgress_XzEncOffset; + + +static SRes CompressProgress_XzEncOffset_Progress(ICompressProgressPtr pp, UInt64 inSize, UInt64 outSize) +{ + const CCompressProgress_XzEncOffset *p = Z7_CONTAINER_FROM_VTBL_CONST(pp, CCompressProgress_XzEncOffset, vt); + inSize += p->inOffset; + outSize += p->outOffset; + return ICompressProgress_Progress(p->progress, inSize, outSize); +} + + + + +struct CXzEnc +{ + ISzAllocPtr alloc; + ISzAllocPtr allocBig; + + CXzProps xzProps; + UInt64 expectedDataSize; + + CXzEncIndex xzIndex; + + CLzma2WithFilters lzmaf_Items[MTCODER_THREADS_MAX]; + + size_t outBufSize; /* size of allocated outBufs[i] */ + Byte *outBufs[MTCODER_BLOCKS_MAX]; + + #ifndef Z7_ST + unsigned checkType; + ISeqOutStreamPtr outStream; + BoolInt mtCoder_WasConstructed; + CMtCoder mtCoder; + CXzEncBlockInfo EncBlocks[MTCODER_BLOCKS_MAX]; + #endif +}; + + +static void XzEnc_Construct(CXzEnc *p) +{ + unsigned i; + + XzEncIndex_Construct(&p->xzIndex); + + for (i = 0; i < MTCODER_THREADS_MAX; i++) + Lzma2WithFilters_Construct(&p->lzmaf_Items[i]); + + #ifndef Z7_ST + p->mtCoder_WasConstructed = False; + { + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + p->outBufs[i] = NULL; + p->outBufSize = 0; + } + #endif +} + + +static void XzEnc_FreeOutBufs(CXzEnc *p) +{ + unsigned i; + for (i = 0; i < MTCODER_BLOCKS_MAX; i++) + if (p->outBufs[i]) + { + ISzAlloc_Free(p->alloc, p->outBufs[i]); + p->outBufs[i] = NULL; + } + p->outBufSize = 0; } -static SRes Xz_Compress(CXzStream *xz, CLzma2WithFilters *lzmaf, - ISeqOutStream *outStream, ISeqInStream *inStream, - const CXzProps *props, ICompressProgress *progress) +static void XzEnc_Free(CXzEnc *p, ISzAllocPtr alloc) { - xz->flags = (Byte)props->checkId; + unsigned i; - RINOK(Lzma2Enc_SetProps(lzmaf->lzma2, props->lzma2Props)); - RINOK(Xz_WriteHeader(xz->flags, outStream)); + XzEncIndex_Free(&p->xzIndex, alloc); + + for (i = 0; i < MTCODER_THREADS_MAX; i++) + Lzma2WithFilters_Free(&p->lzmaf_Items[i], alloc); + + #ifndef Z7_ST + if (p->mtCoder_WasConstructed) + { + MtCoder_Destruct(&p->mtCoder); + p->mtCoder_WasConstructed = False; + } + XzEnc_FreeOutBufs(p); + #endif +} + + +CXzEncHandle XzEnc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig) +{ + CXzEnc *p = (CXzEnc *)ISzAlloc_Alloc(alloc, sizeof(CXzEnc)); + if (!p) + return NULL; + XzEnc_Construct(p); + XzProps_Init(&p->xzProps); + XzProps_Normalize(&p->xzProps); + p->expectedDataSize = (UInt64)(Int64)-1; + p->alloc = alloc; + p->allocBig = allocBig; + return (CXzEncHandle)p; +} + +// #define GET_CXzEnc_p CXzEnc *p = (CXzEnc *)(void *)pp; + +void XzEnc_Destroy(CXzEncHandle p) +{ + // GET_CXzEnc_p + XzEnc_Free(p, p->alloc); + ISzAlloc_Free(p->alloc, p); +} + + +SRes XzEnc_SetProps(CXzEncHandle p, const CXzProps *props) +{ + // GET_CXzEnc_p + p->xzProps = *props; + XzProps_Normalize(&p->xzProps); + return SZ_OK; +} + + +void XzEnc_SetDataSize(CXzEncHandle p, UInt64 expectedDataSiize) +{ + // GET_CXzEnc_p + p->expectedDataSize = expectedDataSiize; +} + + + + +#ifndef Z7_ST + +static SRes XzEnc_MtCallback_Code(void *pp, unsigned coderIndex, unsigned outBufIndex, + const Byte *src, size_t srcSize, int finished) +{ + CXzEnc *me = (CXzEnc *)pp; + SRes res; + CMtProgressThunk progressThunk; + + Byte *dest = me->outBufs[outBufIndex]; + + UNUSED_VAR(finished) + + { + CXzEncBlockInfo *bInfo = &me->EncBlocks[outBufIndex]; + bInfo->totalSize = 0; + bInfo->unpackSize = 0; + bInfo->headerSize = 0; + } + + if (!dest) + { + dest = (Byte *)ISzAlloc_Alloc(me->alloc, me->outBufSize); + if (!dest) + return SZ_ERROR_MEM; + me->outBufs[outBufIndex] = dest; + } + + MtProgressThunk_CreateVTable(&progressThunk); + progressThunk.mtProgress = &me->mtCoder.mtProgress; + MtProgressThunk_INIT(&progressThunk) { - CSeqCheckInStream checkInStream; - CSeqSizeOutStream seqSizeOutStream; - CXzBlock block; - unsigned filterIndex = 0; - CXzFilter *filter = NULL; - const CXzFilterProps *fp = props->filterProps; + CXzEncBlockInfo blockSizes; + int inStreamFinished; + + res = Xz_CompressBlock( + &me->lzmaf_Items[coderIndex], + + NULL, + dest, + dest + XZ_BLOCK_HEADER_SIZE_MAX, me->outBufSize - XZ_BLOCK_HEADER_SIZE_MAX, + + NULL, + // srcSize, // expectedSize + src, srcSize, + + &me->xzProps, + &progressThunk.vt, + &inStreamFinished, + &blockSizes, + me->alloc, + me->allocBig); - XzBlock_ClearFlags(&block); - XzBlock_SetNumFilters(&block, 1 + (fp ? 1 : 0)); + if (res == SZ_OK) + me->EncBlocks[outBufIndex] = blockSizes; + + return res; + } +} + + +static SRes XzEnc_MtCallback_Write(void *pp, unsigned outBufIndex) +{ + CXzEnc *me = (CXzEnc *)pp; + + const CXzEncBlockInfo *bInfo = &me->EncBlocks[outBufIndex]; + const Byte *data = me->outBufs[outBufIndex]; + + RINOK(WriteBytes(me->outStream, data, bInfo->headerSize)) + + { + UInt64 totalPackFull = bInfo->totalSize + XZ_GET_PAD_SIZE(bInfo->totalSize); + RINOK(WriteBytes(me->outStream, data + XZ_BLOCK_HEADER_SIZE_MAX, (size_t)totalPackFull - bInfo->headerSize)) + } + + return XzEncIndex_AddIndexRecord(&me->xzIndex, bInfo->unpackSize, bInfo->totalSize, me->alloc); +} + +#endif + + + +SRes XzEnc_Encode(CXzEncHandle p, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, ICompressProgressPtr progress) +{ + // GET_CXzEnc_p + + const CXzProps *props = &p->xzProps; + + XzEncIndex_Init(&p->xzIndex); + { + UInt64 numBlocks = 1; + UInt64 blockSize = props->blockSize; - if (fp) + if (blockSize != XZ_PROPS_BLOCK_SIZE_SOLID + && props->reduceSize != (UInt64)(Int64)-1) { - filter = &block.filters[filterIndex++]; - filter->id = fp->id; - filter->propsSize = 0; - - if (fp->id == XZ_ID_Delta) - { - filter->props[0] = (Byte)(fp->delta - 1); - filter->propsSize = 1; - } - else if (fp->ipDefined) - { - SetUi32(filter->props, fp->ip); - filter->propsSize = 4; - } + numBlocks = props->reduceSize / blockSize; + if (numBlocks * blockSize != props->reduceSize) + numBlocks++; } + else + blockSize = (UInt64)1 << 62; + + RINOK(XzEncIndex_PreAlloc(&p->xzIndex, numBlocks, blockSize, XZ_GET_ESTIMATED_BLOCK_TOTAL_PACK_SIZE(blockSize), p->alloc)) + } + + RINOK(Xz_WriteHeader((CXzStreamFlags)props->checkId, outStream)) + + #ifndef Z7_ST + if (props->numBlockThreads_Reduced > 1) + { + IMtCoderCallback2 vt; + + if (!p->mtCoder_WasConstructed) { - CXzFilter *f = &block.filters[filterIndex++]; - f->id = XZ_ID_LZMA2; - f->propsSize = 1; - f->props[0] = Lzma2Enc_WriteProperties(lzmaf->lzma2); + p->mtCoder_WasConstructed = True; + MtCoder_Construct(&p->mtCoder); } - seqSizeOutStream.p.Write = MyWrite; - seqSizeOutStream.realStream = outStream; - seqSizeOutStream.processed = 0; + vt.Code = XzEnc_MtCallback_Code; + vt.Write = XzEnc_MtCallback_Write; + + p->checkType = props->checkId; + p->xzProps = *props; - RINOK(XzBlock_WriteHeader(&block, &seqSizeOutStream.p)); + p->outStream = outStream; + + p->mtCoder.allocBig = p->allocBig; + p->mtCoder.progress = progress; + p->mtCoder.inStream = inStream; + p->mtCoder.inData = NULL; + p->mtCoder.inDataSize = 0; + p->mtCoder.mtCallback = &vt; + p->mtCoder.mtCallbackObject = p; + + if ( props->blockSize == XZ_PROPS_BLOCK_SIZE_SOLID + || props->blockSize == XZ_PROPS_BLOCK_SIZE_AUTO) + return SZ_ERROR_FAIL; + + p->mtCoder.blockSize = (size_t)props->blockSize; + if (p->mtCoder.blockSize != props->blockSize) + return SZ_ERROR_PARAM; /* SZ_ERROR_MEM */ + + { + size_t destBlockSize = XZ_BLOCK_HEADER_SIZE_MAX + XZ_GET_MAX_BLOCK_PACK_SIZE(p->mtCoder.blockSize); + if (destBlockSize < p->mtCoder.blockSize) + return SZ_ERROR_PARAM; + if (p->outBufSize != destBlockSize) + XzEnc_FreeOutBufs(p); + p->outBufSize = destBlockSize; + } + + p->mtCoder.numThreadsMax = (unsigned)props->numBlockThreads_Max; + p->mtCoder.expectedDataSize = p->expectedDataSize; - checkInStream.p.Read = SeqCheckInStream_Read; - checkInStream.realStream = inStream; - SeqCheckInStream_Init(&checkInStream, XzFlags_GetCheckType(xz->flags)); + RINOK(MtCoder_Code(&p->mtCoder)) + } + else + #endif + { + int writeStartSizes; + CCompressProgress_XzEncOffset progress2; + Byte *bufData = NULL; + size_t bufSize = 0; + + progress2.vt.Progress = CompressProgress_XzEncOffset_Progress; + progress2.inOffset = 0; + progress2.outOffset = 0; + progress2.progress = progress; - if (fp) + writeStartSizes = 0; + + if (props->blockSize != XZ_PROPS_BLOCK_SIZE_SOLID) { - #ifdef USE_SUBBLOCK - if (fp->id == XZ_ID_Subblock) - { - lzmaf->sb.inStream = &checkInStream.p; - RINOK(SbEncInStream_Init(&lzmaf->sb)); - } - else - #endif + writeStartSizes = (props->forceWriteSizesInHeader > 0); + + if (writeStartSizes) { - lzmaf->filter.realStream = &checkInStream.p; - RINOK(SeqInFilter_Init(&lzmaf->filter, filter)); + size_t t2; + size_t t = (size_t)props->blockSize; + if (t != props->blockSize) + return SZ_ERROR_PARAM; + t = XZ_GET_MAX_BLOCK_PACK_SIZE(t); + if (t < props->blockSize) + return SZ_ERROR_PARAM; + t2 = XZ_BLOCK_HEADER_SIZE_MAX + t; + if (!p->outBufs[0] || t2 != p->outBufSize) + { + XzEnc_FreeOutBufs(p); + p->outBufs[0] = (Byte *)ISzAlloc_Alloc(p->alloc, t2); + if (!p->outBufs[0]) + return SZ_ERROR_MEM; + p->outBufSize = t2; + } + bufData = p->outBufs[0] + XZ_BLOCK_HEADER_SIZE_MAX; + bufSize = t; } } - + + for (;;) { - UInt64 packPos = seqSizeOutStream.processed; + CXzEncBlockInfo blockSizes; + int inStreamFinished; - SRes res = Lzma2Enc_Encode(lzmaf->lzma2, &seqSizeOutStream.p, - fp ? - #ifdef USE_SUBBLOCK - (fp->id == XZ_ID_Subblock) ? &lzmaf->sb.p: - #endif - &lzmaf->filter.p: - &checkInStream.p, - progress); + /* + UInt64 rem = (UInt64)(Int64)-1; + if (props->reduceSize != (UInt64)(Int64)-1 + && props->reduceSize >= progress2.inOffset) + rem = props->reduceSize - progress2.inOffset; + */ + + blockSizes.headerSize = 0; // for GCC - RINOK(res); - block.unpackSize = checkInStream.processed; - block.packSize = seqSizeOutStream.processed - packPos; - } + RINOK(Xz_CompressBlock( + &p->lzmaf_Items[0], + + writeStartSizes ? NULL : outStream, + writeStartSizes ? p->outBufs[0] : NULL, + bufData, bufSize, + + inStream, + // rem, + NULL, 0, + + props, + progress ? &progress2.vt : NULL, + &inStreamFinished, + &blockSizes, + p->alloc, + p->allocBig)) - { - unsigned padSize = 0; - Byte buf[128]; - while ((((unsigned)block.packSize + padSize) & 3) != 0) - buf[padSize++] = 0; - SeqCheckInStream_GetDigest(&checkInStream, buf + padSize); - RINOK(WriteBytes(&seqSizeOutStream.p, buf, padSize + XzFlags_GetCheckSize(xz->flags))); - RINOK(Xz_AddIndexRecord(xz, block.unpackSize, seqSizeOutStream.processed - padSize, &g_Alloc)); + { + UInt64 totalPackFull = blockSizes.totalSize + XZ_GET_PAD_SIZE(blockSizes.totalSize); + + if (writeStartSizes) + { + RINOK(WriteBytes(outStream, p->outBufs[0], blockSizes.headerSize)) + RINOK(WriteBytes(outStream, bufData, (size_t)totalPackFull - blockSizes.headerSize)) + } + + RINOK(XzEncIndex_AddIndexRecord(&p->xzIndex, blockSizes.unpackSize, blockSizes.totalSize, p->alloc)) + + progress2.inOffset += blockSizes.unpackSize; + progress2.outOffset += totalPackFull; + } + + if (inStreamFinished) + break; } } - return Xz_WriteFooter(xz, outStream); + + return XzEncIndex_WriteFooter(&p->xzIndex, (CXzStreamFlags)props->checkId, outStream); } -SRes Xz_Encode(ISeqOutStream *outStream, ISeqInStream *inStream, - const CXzProps *props, ICompressProgress *progress) +#include "Alloc.h" + +SRes Xz_Encode(ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, + const CXzProps *props, ICompressProgressPtr progress) { SRes res; - CXzStream xz; - CLzma2WithFilters lzmaf; - Xz_Construct(&xz); - Lzma2WithFilters_Construct(&lzmaf, &g_Alloc, &g_BigAlloc); - res = Lzma2WithFilters_Create(&lzmaf); + CXzEncHandle xz = XzEnc_Create(&g_Alloc, &g_BigAlloc); + if (!xz) + return SZ_ERROR_MEM; + res = XzEnc_SetProps(xz, props); if (res == SZ_OK) - res = Xz_Compress(&xz, &lzmaf, outStream, inStream, props, progress); - Lzma2WithFilters_Free(&lzmaf); - Xz_Free(&xz, &g_Alloc); + res = XzEnc_Encode(xz, outStream, inStream, progress); + XzEnc_Destroy(xz); return res; } -SRes Xz_EncodeEmpty(ISeqOutStream *outStream) +SRes Xz_EncodeEmpty(ISeqOutStreamPtr outStream) { SRes res; - CXzStream xz; - Xz_Construct(&xz); - res = Xz_WriteHeader(xz.flags, outStream); + CXzEncIndex xzIndex; + XzEncIndex_Construct(&xzIndex); + res = Xz_WriteHeader((CXzStreamFlags)0, outStream); if (res == SZ_OK) - res = Xz_WriteFooter(&xz, outStream); - Xz_Free(&xz, &g_Alloc); + res = XzEncIndex_WriteFooter(&xzIndex, (CXzStreamFlags)0, outStream); + XzEncIndex_Free(&xzIndex, NULL); // g_Alloc return res; } diff --git a/3rdparty/lzma/C/XzEnc.h b/3rdparty/lzma/C/XzEnc.h index c3c19eca008..77b78c014be 100644 --- a/3rdparty/lzma/C/XzEnc.h +++ b/3rdparty/lzma/C/XzEnc.h @@ -1,8 +1,8 @@ /* XzEnc.h -- Xz Encode -2011-02-07 : Igor Pavlov : Public domain */ +2023-04-13 : Igor Pavlov : Public domain */ -#ifndef __XZ_ENC_H -#define __XZ_ENC_H +#ifndef ZIP7_INC_XZ_ENC_H +#define ZIP7_INC_XZ_ENC_H #include "Lzma2Enc.h" @@ -10,6 +10,11 @@ EXTERN_C_BEGIN + +#define XZ_PROPS_BLOCK_SIZE_AUTO LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO +#define XZ_PROPS_BLOCK_SIZE_SOLID LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID + + typedef struct { UInt32 id; @@ -20,19 +25,36 @@ typedef struct void XzFilterProps_Init(CXzFilterProps *p); + typedef struct { - const CLzma2EncProps *lzma2Props; - const CXzFilterProps *filterProps; + CLzma2EncProps lzma2Props; + CXzFilterProps filterProps; unsigned checkId; + UInt64 blockSize; + int numBlockThreads_Reduced; + int numBlockThreads_Max; + int numTotalThreads; + int forceWriteSizesInHeader; + UInt64 reduceSize; } CXzProps; void XzProps_Init(CXzProps *p); -SRes Xz_Encode(ISeqOutStream *outStream, ISeqInStream *inStream, - const CXzProps *props, ICompressProgress *progress); +typedef struct CXzEnc CXzEnc; +typedef CXzEnc * CXzEncHandle; +// Z7_DECLARE_HANDLE(CXzEncHandle) + +CXzEncHandle XzEnc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig); +void XzEnc_Destroy(CXzEncHandle p); +SRes XzEnc_SetProps(CXzEncHandle p, const CXzProps *props); +void XzEnc_SetDataSize(CXzEncHandle p, UInt64 expectedDataSiize); +SRes XzEnc_Encode(CXzEncHandle p, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, ICompressProgressPtr progress); + +SRes Xz_Encode(ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream, + const CXzProps *props, ICompressProgressPtr progress); -SRes Xz_EncodeEmpty(ISeqOutStream *outStream); +SRes Xz_EncodeEmpty(ISeqOutStreamPtr outStream); EXTERN_C_END diff --git a/3rdparty/lzma/C/XzIn.c b/3rdparty/lzma/C/XzIn.c index 07b1cc39516..d0fc763679e 100644 --- a/3rdparty/lzma/C/XzIn.c +++ b/3rdparty/lzma/C/XzIn.c @@ -1,5 +1,5 @@ /* XzIn.c - Xz input -2015-11-08 : Igor Pavlov : Public domain */ +2023-04-02 : Igor Pavlov : Public domain */ #include "Precomp.h" @@ -9,26 +9,35 @@ #include "CpuArch.h" #include "Xz.h" -SRes Xz_ReadHeader(CXzStreamFlags *p, ISeqInStream *inStream) +/* +#define XZ_FOOTER_SIG_CHECK(p) (memcmp((p), XZ_FOOTER_SIG, XZ_FOOTER_SIG_SIZE) == 0) +*/ +#define XZ_FOOTER_SIG_CHECK(p) ((p)[0] == XZ_FOOTER_SIG_0 && (p)[1] == XZ_FOOTER_SIG_1) + + +SRes Xz_ReadHeader(CXzStreamFlags *p, ISeqInStreamPtr inStream) { Byte sig[XZ_STREAM_HEADER_SIZE]; - RINOK(SeqInStream_Read2(inStream, sig, XZ_STREAM_HEADER_SIZE, SZ_ERROR_NO_ARCHIVE)); - if (memcmp(sig, XZ_SIG, XZ_SIG_SIZE) != 0) + size_t processedSize = XZ_STREAM_HEADER_SIZE; + RINOK(SeqInStream_ReadMax(inStream, sig, &processedSize)) + if (processedSize != XZ_STREAM_HEADER_SIZE + || memcmp(sig, XZ_SIG, XZ_SIG_SIZE) != 0) return SZ_ERROR_NO_ARCHIVE; return Xz_ParseHeader(p, sig); } #define READ_VARINT_AND_CHECK(buf, pos, size, res) \ { unsigned s = Xz_ReadVarInt(buf + pos, size - pos, res); \ - if (s == 0) return SZ_ERROR_ARCHIVE; pos += s; } + if (s == 0) return SZ_ERROR_ARCHIVE; \ + pos += s; } -SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStream *inStream, Bool *isIndex, UInt32 *headerSizeRes) +SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStreamPtr inStream, BoolInt *isIndex, UInt32 *headerSizeRes) { Byte header[XZ_BLOCK_HEADER_SIZE_MAX]; unsigned headerSize; *headerSizeRes = 0; - RINOK(SeqInStream_ReadByte(inStream, &header[0])); - headerSize = ((unsigned)header[0] << 2) + 4; + RINOK(SeqInStream_ReadByte(inStream, &header[0])) + headerSize = (unsigned)header[0]; if (headerSize == 0) { *headerSizeRes = 1; @@ -37,12 +46,18 @@ SRes XzBlock_ReadHeader(CXzBlock *p, ISeqInStream *inStream, Bool *isIndex, UInt } *isIndex = False; + headerSize = (headerSize << 2) + 4; *headerSizeRes = headerSize; - RINOK(SeqInStream_Read(inStream, header + 1, headerSize - 1)); + { + size_t processedSize = headerSize - 1; + RINOK(SeqInStream_ReadMax(inStream, header + 1, &processedSize)) + if (processedSize != headerSize - 1) + return SZ_ERROR_INPUT_EOF; + } return XzBlock_Parse(p, header); } -#define ADD_SIZE_CHECH(size, val) \ +#define ADD_SIZE_CHECK(size, val) \ { UInt64 newSize = size + (val); if (newSize < size) return XZ_SIZE_OVERFLOW; size = newSize; } UInt64 Xz_GetUnpackSize(const CXzStream *p) @@ -50,7 +65,9 @@ UInt64 Xz_GetUnpackSize(const CXzStream *p) UInt64 size = 0; size_t i; for (i = 0; i < p->numBlocks; i++) - ADD_SIZE_CHECH(size, p->blocks[i].unpackSize); + { + ADD_SIZE_CHECK(size, p->blocks[i].unpackSize) + } return size; } @@ -59,18 +76,20 @@ UInt64 Xz_GetPackSize(const CXzStream *p) UInt64 size = 0; size_t i; for (i = 0; i < p->numBlocks; i++) - ADD_SIZE_CHECH(size, (p->blocks[i].totalSize + 3) & ~(UInt64)3); + { + ADD_SIZE_CHECK(size, (p->blocks[i].totalSize + 3) & ~(UInt64)3) + } return size; } /* -SRes XzBlock_ReadFooter(CXzBlock *p, CXzStreamFlags f, ISeqInStream *inStream) +SRes XzBlock_ReadFooter(CXzBlock *p, CXzStreamFlags f, ISeqInStreamPtr inStream) { return SeqInStream_Read(inStream, p->check, XzFlags_GetCheckSize(f)); } */ -static SRes Xz_ReadIndex2(CXzStream *p, const Byte *buf, size_t size, ISzAlloc *alloc) +static SRes Xz_ReadIndex2(CXzStream *p, const Byte *buf, size_t size, ISzAllocPtr alloc) { size_t numBlocks, pos = 1; UInt32 crc; @@ -85,7 +104,7 @@ static SRes Xz_ReadIndex2(CXzStream *p, const Byte *buf, size_t size, ISzAlloc * { UInt64 numBlocks64; - READ_VARINT_AND_CHECK(buf, pos, size, &numBlocks64); + READ_VARINT_AND_CHECK(buf, pos, size, &numBlocks64) numBlocks = (size_t)numBlocks64; if (numBlocks != numBlocks64 || numBlocks * 2 > size) return SZ_ERROR_ARCHIVE; @@ -96,15 +115,14 @@ static SRes Xz_ReadIndex2(CXzStream *p, const Byte *buf, size_t size, ISzAlloc * { size_t i; p->numBlocks = numBlocks; - p->numBlocksAllocated = numBlocks; - p->blocks = alloc->Alloc(alloc, sizeof(CXzBlockSizes) * numBlocks); - if (p->blocks == 0) + p->blocks = (CXzBlockSizes *)ISzAlloc_Alloc(alloc, sizeof(CXzBlockSizes) * numBlocks); + if (!p->blocks) return SZ_ERROR_MEM; for (i = 0; i < numBlocks; i++) { CXzBlockSizes *block = &p->blocks[i]; - READ_VARINT_AND_CHECK(buf, pos, size, &block->totalSize); - READ_VARINT_AND_CHECK(buf, pos, size, &block->unpackSize); + READ_VARINT_AND_CHECK(buf, pos, size, &block->totalSize) + READ_VARINT_AND_CHECK(buf, pos, size, &block->unpackSize) if (block->totalSize == 0) return SZ_ERROR_ARCHIVE; } @@ -115,7 +133,7 @@ static SRes Xz_ReadIndex2(CXzStream *p, const Byte *buf, size_t size, ISzAlloc * return (pos == size) ? SZ_OK : SZ_ERROR_ARCHIVE; } -static SRes Xz_ReadIndex(CXzStream *p, ILookInStream *stream, UInt64 indexSize, ISzAlloc *alloc) +static SRes Xz_ReadIndex(CXzStream *p, ILookInStreamPtr stream, UInt64 indexSize, ISzAllocPtr alloc) { SRes res; size_t size; @@ -125,36 +143,36 @@ static SRes Xz_ReadIndex(CXzStream *p, ILookInStream *stream, UInt64 indexSize, size = (size_t)indexSize; if (size != indexSize) return SZ_ERROR_UNSUPPORTED; - buf = alloc->Alloc(alloc, size); - if (buf == 0) + buf = (Byte *)ISzAlloc_Alloc(alloc, size); + if (!buf) return SZ_ERROR_MEM; res = LookInStream_Read2(stream, buf, size, SZ_ERROR_UNSUPPORTED); if (res == SZ_OK) res = Xz_ReadIndex2(p, buf, size, alloc); - alloc->Free(alloc, buf); + ISzAlloc_Free(alloc, buf); return res; } -static SRes LookInStream_SeekRead_ForArc(ILookInStream *stream, UInt64 offset, void *buf, size_t size) +static SRes LookInStream_SeekRead_ForArc(ILookInStreamPtr stream, UInt64 offset, void *buf, size_t size) { - RINOK(LookInStream_SeekTo(stream, offset)); + RINOK(LookInStream_SeekTo(stream, offset)) return LookInStream_Read(stream, buf, size); /* return LookInStream_Read2(stream, buf, size, SZ_ERROR_NO_ARCHIVE); */ } -static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOffset, ISzAlloc *alloc) +static SRes Xz_ReadBackward(CXzStream *p, ILookInStreamPtr stream, Int64 *startOffset, ISzAllocPtr alloc) { UInt64 indexSize; Byte buf[XZ_STREAM_FOOTER_SIZE]; - UInt64 pos = *startOffset; + UInt64 pos = (UInt64)*startOffset; if ((pos & 3) != 0 || pos < XZ_STREAM_FOOTER_SIZE) return SZ_ERROR_NO_ARCHIVE; pos -= XZ_STREAM_FOOTER_SIZE; - RINOK(LookInStream_SeekRead_ForArc(stream, pos, buf, XZ_STREAM_FOOTER_SIZE)); + RINOK(LookInStream_SeekRead_ForArc(stream, pos, buf, XZ_STREAM_FOOTER_SIZE)) - if (memcmp(buf + 10, XZ_FOOTER_SIG, XZ_FOOTER_SIG_SIZE) != 0) + if (!XZ_FOOTER_SIG_CHECK(buf + 10)) { UInt32 total = 0; pos += XZ_STREAM_FOOTER_SIZE; @@ -167,7 +185,7 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff i = (pos > TEMP_BUF_SIZE) ? TEMP_BUF_SIZE : (size_t)pos; pos -= i; - RINOK(LookInStream_SeekRead_ForArc(stream, pos, temp, i)); + RINOK(LookInStream_SeekRead_ForArc(stream, pos, temp, i)) total += (UInt32)i; for (; i != 0; i--) if (temp[i - 1] != 0) @@ -186,8 +204,8 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff if (pos < XZ_STREAM_FOOTER_SIZE) return SZ_ERROR_NO_ARCHIVE; pos -= XZ_STREAM_FOOTER_SIZE; - RINOK(LookInStream_SeekRead_ForArc(stream, pos, buf, XZ_STREAM_FOOTER_SIZE)); - if (memcmp(buf + 10, XZ_FOOTER_SIG, XZ_FOOTER_SIG_SIZE) != 0) + RINOK(LookInStream_SeekRead_ForArc(stream, pos, buf, XZ_STREAM_FOOTER_SIZE)) + if (!XZ_FOOTER_SIG_CHECK(buf + 10)) return SZ_ERROR_NO_ARCHIVE; } @@ -196,8 +214,13 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff if (!XzFlags_IsSupported(p->flags)) return SZ_ERROR_UNSUPPORTED; - if (GetUi32(buf) != CrcCalc(buf + 4, 6)) - return SZ_ERROR_ARCHIVE; + { + /* to eliminate GCC 6.3 warning: + dereferencing type-punned pointer will break strict-aliasing rules */ + const Byte *buf_ptr = buf; + if (GetUi32(buf_ptr) != CrcCalc(buf + 4, 6)) + return SZ_ERROR_ARCHIVE; + } indexSize = ((UInt64)GetUi32(buf + 4) + 1) << 2; @@ -205,8 +228,8 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff return SZ_ERROR_ARCHIVE; pos -= indexSize; - RINOK(LookInStream_SeekTo(stream, pos)); - RINOK(Xz_ReadIndex(p, stream, indexSize, alloc)); + RINOK(LookInStream_SeekTo(stream, pos)) + RINOK(Xz_ReadIndex(p, stream, indexSize, alloc)) { UInt64 totalSize = Xz_GetPackSize(p); @@ -215,8 +238,8 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff || pos < totalSize + XZ_STREAM_HEADER_SIZE) return SZ_ERROR_ARCHIVE; pos -= (totalSize + XZ_STREAM_HEADER_SIZE); - RINOK(LookInStream_SeekTo(stream, pos)); - *startOffset = pos; + RINOK(LookInStream_SeekTo(stream, pos)) + *startOffset = (Int64)pos; } { CXzStreamFlags headerFlags; @@ -224,7 +247,7 @@ static SRes Xz_ReadBackward(CXzStream *p, ILookInStream *stream, Int64 *startOff SecToRead_CreateVTable(&secToRead); secToRead.realStream = stream; - RINOK(Xz_ReadHeader(&headerFlags, &secToRead.s)); + RINOK(Xz_ReadHeader(&headerFlags, &secToRead.vt)) return (p->flags == headerFlags) ? SZ_OK : SZ_ERROR_ARCHIVE; } } @@ -238,12 +261,12 @@ void Xzs_Construct(CXzs *p) p->streams = 0; } -void Xzs_Free(CXzs *p, ISzAlloc *alloc) +void Xzs_Free(CXzs *p, ISzAllocPtr alloc) { size_t i; for (i = 0; i < p->num; i++) Xz_Free(&p->streams[i], alloc); - alloc->Free(alloc, p->streams); + ISzAlloc_Free(alloc, p->streams); p->num = p->numAllocated = 0; p->streams = 0; } @@ -262,7 +285,9 @@ UInt64 Xzs_GetUnpackSize(const CXzs *p) UInt64 size = 0; size_t i; for (i = 0; i < p->num; i++) - ADD_SIZE_CHECH(size, Xz_GetUnpackSize(&p->streams[i])); + { + ADD_SIZE_CHECK(size, Xz_GetUnpackSize(&p->streams[i])) + } return size; } @@ -272,15 +297,17 @@ UInt64 Xzs_GetPackSize(const CXzs *p) UInt64 size = 0; size_t i; for (i = 0; i < p->num; i++) - ADD_SIZE_CHECH(size, Xz_GetTotalSize(&p->streams[i])); + { + ADD_SIZE_CHECK(size, Xz_GetTotalSize(&p->streams[i])) + } return size; } */ -SRes Xzs_ReadBackward(CXzs *p, ILookInStream *stream, Int64 *startOffset, ICompressProgress *progress, ISzAlloc *alloc) +SRes Xzs_ReadBackward(CXzs *p, ILookInStreamPtr stream, Int64 *startOffset, ICompressProgressPtr progress, ISzAllocPtr alloc) { Int64 endOffset = 0; - RINOK(stream->Seek(stream, &endOffset, SZ_SEEK_END)); + RINOK(ILookInStream_Seek(stream, &endOffset, SZ_SEEK_END)) *startOffset = endOffset; for (;;) { @@ -288,25 +315,25 @@ SRes Xzs_ReadBackward(CXzs *p, ILookInStream *stream, Int64 *startOffset, ICompr SRes res; Xz_Construct(&st); res = Xz_ReadBackward(&st, stream, startOffset, alloc); - st.startOffset = *startOffset; - RINOK(res); + st.startOffset = (UInt64)*startOffset; + RINOK(res) if (p->num == p->numAllocated) { - size_t newNum = p->num + p->num / 4 + 1; - Byte *data = (Byte *)alloc->Alloc(alloc, newNum * sizeof(CXzStream)); - if (data == 0) + const size_t newNum = p->num + p->num / 4 + 1; + void *data = ISzAlloc_Alloc(alloc, newNum * sizeof(CXzStream)); + if (!data) return SZ_ERROR_MEM; p->numAllocated = newNum; if (p->num != 0) memcpy(data, p->streams, p->num * sizeof(CXzStream)); - alloc->Free(alloc, p->streams); + ISzAlloc_Free(alloc, p->streams); p->streams = (CXzStream *)data; } p->streams[p->num++] = st; if (*startOffset == 0) break; - RINOK(LookInStream_SeekTo(stream, *startOffset)); - if (progress && progress->Progress(progress, endOffset - *startOffset, (UInt64)(Int64)-1) != SZ_OK) + RINOK(LookInStream_SeekTo(stream, (UInt64)*startOffset)) + if (progress && ICompressProgress_Progress(progress, (UInt64)(endOffset - *startOffset), (UInt64)(Int64)-1) != SZ_OK) return SZ_ERROR_PROGRESS; } return SZ_OK; diff --git a/3rdparty/lzma/C/var_clang.mak b/3rdparty/lzma/C/var_clang.mak new file mode 100644 index 00000000000..a6df26e77c8 --- /dev/null +++ b/3rdparty/lzma/C/var_clang.mak @@ -0,0 +1,11 @@ +PLATFORM= +O=b/c +IS_X64= +IS_X86= +IS_ARM64= +CROSS_COMPILE= +MY_ARCH= +USE_ASM= +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/var_clang_arm64.mak b/3rdparty/lzma/C/var_clang_arm64.mak new file mode 100644 index 00000000000..4b354096956 --- /dev/null +++ b/3rdparty/lzma/C/var_clang_arm64.mak @@ -0,0 +1,11 @@ +PLATFORM=arm64 +O=b/c_$(PLATFORM) +IS_X64= +IS_X86= +IS_ARM64=1 +CROSS_COMPILE= +MY_ARCH= +USE_ASM=1 +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/var_clang_x64.mak b/3rdparty/lzma/C/var_clang_x64.mak new file mode 100644 index 00000000000..34e1b49c5e2 --- /dev/null +++ b/3rdparty/lzma/C/var_clang_x64.mak @@ -0,0 +1,11 @@ +PLATFORM=x64 +O=b/c_$(PLATFORM) +IS_X64=1 +IS_X86= +IS_ARM64= +CROSS_COMPILE= +MY_ARCH= +USE_ASM=1 +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/var_clang_x86.mak b/3rdparty/lzma/C/var_clang_x86.mak new file mode 100644 index 00000000000..bd2317c299f --- /dev/null +++ b/3rdparty/lzma/C/var_clang_x86.mak @@ -0,0 +1,11 @@ +PLATFORM=x86 +O=b/c_$(PLATFORM) +IS_X64= +IS_X86=1 +IS_ARM64= +CROSS_COMPILE= +MY_ARCH=-m32 +USE_ASM=1 +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/var_gcc.mak b/3rdparty/lzma/C/var_gcc.mak new file mode 100644 index 00000000000..664491cf2fc --- /dev/null +++ b/3rdparty/lzma/C/var_gcc.mak @@ -0,0 +1,12 @@ +PLATFORM= +O=b/g +IS_X64= +IS_X86= +IS_ARM64= +CROSS_COMPILE= +MY_ARCH= +USE_ASM= +CC=$(CROSS_COMPILE)gcc +CXX=$(CROSS_COMPILE)g++ + +# -march=armv8-a+crc+crypto diff --git a/3rdparty/lzma/C/var_gcc_arm64.mak b/3rdparty/lzma/C/var_gcc_arm64.mak new file mode 100644 index 00000000000..4bbb687dbb3 --- /dev/null +++ b/3rdparty/lzma/C/var_gcc_arm64.mak @@ -0,0 +1,12 @@ +PLATFORM=arm64 +O=b/g_$(PLATFORM) +IS_X64= +IS_X86= +IS_ARM64=1 +CROSS_COMPILE= +MY_ARCH=-mtune=cortex-a53 +USE_ASM=1 +CC=$(CROSS_COMPILE)gcc +CXX=$(CROSS_COMPILE)g++ + +# -march=armv8-a+crc+crypto diff --git a/3rdparty/lzma/C/var_gcc_x64.mak b/3rdparty/lzma/C/var_gcc_x64.mak new file mode 100644 index 00000000000..1acf604f5f2 --- /dev/null +++ b/3rdparty/lzma/C/var_gcc_x64.mak @@ -0,0 +1,10 @@ +PLATFORM=x64 +O=b/g_$(PLATFORM) +IS_X64=1 +IS_X86= +IS_ARM64= +CROSS_COMPILE= +MY_ARCH= +USE_ASM=1 +CC=$(CROSS_COMPILE)gcc +CXX=$(CROSS_COMPILE)g++ diff --git a/3rdparty/lzma/C/var_gcc_x86.mak b/3rdparty/lzma/C/var_gcc_x86.mak new file mode 100644 index 00000000000..f0718ec710e --- /dev/null +++ b/3rdparty/lzma/C/var_gcc_x86.mak @@ -0,0 +1,10 @@ +PLATFORM=x86 +O=b/g_$(PLATFORM) +IS_X64= +IS_X86=1 +IS_ARM64= +CROSS_COMPILE= +MY_ARCH=-m32 +USE_ASM=1 +CC=$(CROSS_COMPILE)gcc +CXX=$(CROSS_COMPILE)g++ diff --git a/3rdparty/lzma/C/var_mac_arm64.mak b/3rdparty/lzma/C/var_mac_arm64.mak new file mode 100644 index 00000000000..adf5fa1d703 --- /dev/null +++ b/3rdparty/lzma/C/var_mac_arm64.mak @@ -0,0 +1,11 @@ +PLATFORM=arm64 +O=b/m_$(PLATFORM) +IS_X64= +IS_X86= +IS_ARM64=1 +CROSS_COMPILE= +MY_ARCH=-arch arm64 +USE_ASM=1 +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/var_mac_x64.mak b/3rdparty/lzma/C/var_mac_x64.mak new file mode 100644 index 00000000000..13d7aa7fedf --- /dev/null +++ b/3rdparty/lzma/C/var_mac_x64.mak @@ -0,0 +1,11 @@ +PLATFORM=x64 +O=b/m_$(PLATFORM) +IS_X64=1 +IS_X86= +IS_ARM64= +CROSS_COMPILE= +MY_ARCH=-arch x86_64 +USE_ASM= +CC=$(CROSS_COMPILE)clang +CXX=$(CROSS_COMPILE)clang++ +USE_CLANG=1 diff --git a/3rdparty/lzma/C/warn_clang.mak b/3rdparty/lzma/C/warn_clang.mak new file mode 100644 index 00000000000..0d6446d2134 --- /dev/null +++ b/3rdparty/lzma/C/warn_clang.mak @@ -0,0 +1 @@ +CFLAGS_WARN = -Weverything -Wfatal-errors diff --git a/3rdparty/lzma/C/warn_clang_mac.mak b/3rdparty/lzma/C/warn_clang_mac.mak new file mode 100644 index 00000000000..44afc5374f7 --- /dev/null +++ b/3rdparty/lzma/C/warn_clang_mac.mak @@ -0,0 +1 @@ +CFLAGS_WARN = -Weverything -Wfatal-errors -Wno-poison-system-directories diff --git a/3rdparty/lzma/C/warn_gcc.mak b/3rdparty/lzma/C/warn_gcc.mak new file mode 100644 index 00000000000..7aab7a44937 --- /dev/null +++ b/3rdparty/lzma/C/warn_gcc.mak @@ -0,0 +1,51 @@ +CFLAGS_WARN_GCC_4_5 = \ + +CFLAGS_WARN_GCC_6 = \ + -Waddress \ + -Waggressive-loop-optimizations \ + -Wattributes \ + -Wbool-compare \ + -Wcast-align \ + -Wcomment \ + -Wdiv-by-zero \ + -Wduplicated-cond \ + -Wformat-contains-nul \ + -Winit-self \ + -Wint-to-pointer-cast \ + -Wunused \ + -Wunused-macros \ + +# -Wno-strict-aliasing + +CFLAGS_WARN_GCC_9 = \ + -Waddress \ + -Waddress-of-packed-member \ + -Waggressive-loop-optimizations \ + -Wattributes \ + -Wbool-compare \ + -Wbool-operation \ + -Wcast-align \ + -Wcast-align=strict \ + -Wcomment \ + -Wdangling-else \ + -Wdiv-by-zero \ + -Wduplicated-branches \ + -Wduplicated-cond \ + -Wformat-contains-nul \ + -Wimplicit-fallthrough=5 \ + -Winit-self \ + -Wint-in-bool-context \ + -Wint-to-pointer-cast \ + -Wunused \ + -Wunused-macros \ + -Wconversion \ + +# -Wno-sign-conversion \ + +CFLAGS_WARN_GCC_PPMD_UNALIGNED = \ + -Wno-strict-aliasing \ + + +CFLAGS_WARN = $(CFLAGS_WARN_GCC_9) \ + +# $(CFLAGS_WARN_GCC_PPMD_UNALIGNED) |