diff options
Diffstat (limited to 'src/devices/cpu/sh/sh2drc.cpp')
-rw-r--r-- | src/devices/cpu/sh/sh2drc.cpp | 2982 |
1 files changed, 2982 insertions, 0 deletions
diff --git a/src/devices/cpu/sh/sh2drc.cpp b/src/devices/cpu/sh/sh2drc.cpp new file mode 100644 index 00000000000..1003463b97a --- /dev/null +++ b/src/devices/cpu/sh/sh2drc.cpp @@ -0,0 +1,2982 @@ +// license:BSD-3-Clause +// copyright-holders:R. Belmont +/*************************************************************************** + + sh2drc.c + Universal machine language-based SH-2 emulator. + +***************************************************************************/ + +#include "emu.h" +#include "debugger.h" +#include "sh2.h" +#include "sh2comn.h" + +extern unsigned DasmSH2(std::ostream &stream, unsigned pc, uint16_t opcode); + +using namespace uml; + +/*************************************************************************** + DEBUGGING +***************************************************************************/ + +#define SET_EA (0) // makes slower but "shows work" in the EA fake register like the interpreter + +#define ADDSUBV_DIRECT (0) + +#if SET_EA +#define SETEA(x) UML_MOV(block, mem(&m_sh2_state->ea), ireg(x)) +#else +#define SETEA(x) +#endif + +/*************************************************************************** + CONSTANTS +***************************************************************************/ + +/* map variables */ +#define MAPVAR_PC M0 +#define MAPVAR_CYCLES M1 + +/* exit codes */ +#define EXECUTE_OUT_OF_CYCLES 0 +#define EXECUTE_MISSING_CODE 1 +#define EXECUTE_UNMAPPED_CODE 2 +#define EXECUTE_RESET_CACHE 3 + +#define PROBE_ADDRESS ~0 + + +/*************************************************************************** + MACROS +***************************************************************************/ + +#define R32(reg) m_regmap[reg] + +/*************************************************************************** + INLINE FUNCTIONS +***************************************************************************/ + +/*------------------------------------------------- + epc - compute the exception PC from a + descriptor +-------------------------------------------------*/ + +uint32_t sh2_device::epc(const opcode_desc *desc) +{ + return (desc->flags & OPFLAG_IN_DELAY_SLOT) ? (desc->pc - 1) : desc->pc; +} + +/*------------------------------------------------- + alloc_handle - allocate a handle if not + already allocated +-------------------------------------------------*/ + +void sh2_device::alloc_handle(drcuml_state *drcuml, code_handle **handleptr, const char *name) +{ + if (*handleptr == nullptr) + *handleptr = drcuml->handle_alloc(name); +} + +/*------------------------------------------------- + load_fast_iregs - load any fast integer + registers +-------------------------------------------------*/ + +void sh2_device::load_fast_iregs(drcuml_block *block) +{ + int regnum; + + for (regnum = 0; regnum < ARRAY_LENGTH(m_regmap); regnum++) + { + if (m_regmap[regnum].is_int_register()) + { + UML_MOV(block, uml::parameter::make_ireg(m_regmap[regnum].ireg()), mem(&m_sh2_state->r[regnum])); + } + } +} + + +/*------------------------------------------------- + save_fast_iregs - save any fast integer + registers +-------------------------------------------------*/ + +void sh2_device::save_fast_iregs(drcuml_block *block) +{ + int regnum; + + for (regnum = 0; regnum < ARRAY_LENGTH(m_regmap); regnum++) + { + if (m_regmap[regnum].is_int_register()) + { + UML_MOV(block, mem(&m_sh2_state->r[regnum]), uml::parameter::make_ireg(m_regmap[regnum].ireg())); + } + } +} + +/*------------------------------------------------- + cfunc_printf_probe - print the current CPU + state and return +-------------------------------------------------*/ + +static void cfunc_printf_probe(void *param) +{ + ((sh2_device *)param)->func_printf_probe(); +} + +void sh2_device::func_printf_probe() +{ + uint32_t pc = m_sh2_state->pc; + + printf(" PC=%08X r0=%08X r1=%08X r2=%08X\n", + pc, + (uint32_t)m_sh2_state->r[0], + (uint32_t)m_sh2_state->r[1], + (uint32_t)m_sh2_state->r[2]); + printf(" r3=%08X r4=%08X r5=%08X r6=%08X\n", + (uint32_t)m_sh2_state->r[3], + (uint32_t)m_sh2_state->r[4], + (uint32_t)m_sh2_state->r[5], + (uint32_t)m_sh2_state->r[6]); + printf(" r7=%08X r8=%08X r9=%08X r10=%08X\n", + (uint32_t)m_sh2_state->r[7], + (uint32_t)m_sh2_state->r[8], + (uint32_t)m_sh2_state->r[9], + (uint32_t)m_sh2_state->r[10]); + printf(" r11=%08X r12=%08X r13=%08X r14=%08X\n", + (uint32_t)m_sh2_state->r[11], + (uint32_t)m_sh2_state->r[12], + (uint32_t)m_sh2_state->r[13], + (uint32_t)m_sh2_state->r[14]); + printf(" r15=%08X macl=%08X mach=%08X gbr=%08X\n", + (uint32_t)m_sh2_state->r[15], + (uint32_t)m_sh2_state->macl, + (uint32_t)m_sh2_state->mach, + (uint32_t)m_sh2_state->gbr); + printf(" evec %x irqsr %x pc=%08x\n", + (uint32_t)m_sh2_state->evec, + (uint32_t)m_sh2_state->irqsr, (uint32_t)m_sh2_state->pc); +} + +/*------------------------------------------------- + cfunc_unimplemented - handler for + unimplemented opcdes +-------------------------------------------------*/ + +static void cfunc_unimplemented(void *param) +{ + ((sh2_device *)param)->func_unimplemented(); +} + +void sh2_device::func_unimplemented() +{ + // set up an invalid opcode exception + m_sh2_state->evec = RL( m_sh2_state->vbr + 4 * 4 ); + m_sh2_state->evec &= AM; + m_sh2_state->irqsr = m_sh2_state->sr; + // claim it's an NMI, because it pretty much is + m_sh2_state->pending_nmi = 1; +} + +/*------------------------------------------------- + cfunc_fastirq - checks for pending IRQs +-------------------------------------------------*/ +static void cfunc_fastirq(void *param) +{ + ((sh2_device *)param)->func_fastirq(); +} + +void sh2_device::func_fastirq() +{ + sh2_exception("fastirq",m_sh2_state->irqline); +} + +/*------------------------------------------------- + cfunc_MAC_W - implementation of MAC_W Rm,Rn +-------------------------------------------------*/ +static void cfunc_MAC_W(void *param) +{ + ((sh2_device *)param)->func_MAC_W(); +} + +void sh2_device::func_MAC_W() +{ + int32_t tempm, tempn, dest, src, ans; + uint32_t templ; + uint16_t opcode; + int n, m; + + // recover the opcode + opcode = m_sh2_state->arg0; + + // extract the operands + n = Rn; + m = Rm; + + tempn = (int32_t) RW( m_sh2_state->r[n] ); + m_sh2_state->r[n] += 2; + tempm = (int32_t) RW( m_sh2_state->r[m] ); + m_sh2_state->r[m] += 2; + templ = m_sh2_state->macl; + tempm = ((int32_t) (short) tempn * (int32_t) (short) tempm); + if ((int32_t) m_sh2_state->macl >= 0) + dest = 0; + else + dest = 1; + if ((int32_t) tempm >= 0) + { + src = 0; + tempn = 0; + } + else + { + src = 1; + tempn = 0xffffffff; + } + src += dest; + m_sh2_state->macl += tempm; + if ((int32_t) m_sh2_state->macl >= 0) + ans = 0; + else + ans = 1; + ans += dest; + if (m_sh2_state->sr & S) + { + if (ans == 1) + { + if ((m_cpu_type == CPU_TYPE_SH1) && ((src == 0) || (src == 2))) + { + m_sh2_state->mach |= 0x00000001; + } + + if (src == 0) + m_sh2_state->macl = 0x7fffffff; + if (src == 2) + m_sh2_state->macl = 0x80000000; + } + } + else + { + m_sh2_state->mach += tempn; + if (templ > m_sh2_state->macl) + m_sh2_state->mach += 1; + + // SH-1 has limited precision + if (m_cpu_type == CPU_TYPE_SH1) + { + if ((m_sh2_state->mach & 0x200) == 0) + { + m_sh2_state->mach &= 0x3ff; + } + else + { + m_sh2_state->mach |= 0xfffffc00; + } + } + + + } +} + +/*------------------------------------------------- + cfunc_MAC_L - implementation of MAC_L Rm,Rn +-------------------------------------------------*/ +static void cfunc_MAC_L(void *param) +{ + ((sh2_device *)param)->func_MAC_L(); +} + +void sh2_device::func_MAC_L() +{ + uint32_t RnL, RnH, RmL, RmH, Res0, Res1, Res2; + uint32_t temp0, temp1, temp2, temp3; + int32_t tempm, tempn, fnLmL; + uint16_t opcode; + int n, m; + + // recover the opcode + opcode = m_sh2_state->arg0; + + // extract the operands + n = Rn; + m = Rm; + + tempn = (int32_t) RL( m_sh2_state->r[n] ); + m_sh2_state->r[n] += 4; + tempm = (int32_t) RL( m_sh2_state->r[m] ); + m_sh2_state->r[m] += 4; + if ((int32_t) (tempn ^ tempm) < 0) + fnLmL = -1; + else + fnLmL = 0; + if (tempn < 0) + tempn = 0 - tempn; + if (tempm < 0) + tempm = 0 - tempm; + temp1 = (uint32_t) tempn; + temp2 = (uint32_t) tempm; + RnL = temp1 & 0x0000ffff; + RnH = (temp1 >> 16) & 0x0000ffff; + RmL = temp2 & 0x0000ffff; + RmH = (temp2 >> 16) & 0x0000ffff; + temp0 = RmL * RnL; + temp1 = RmH * RnL; + temp2 = RmL * RnH; + temp3 = RmH * RnH; + Res2 = 0; + Res1 = temp1 + temp2; + if (Res1 < temp1) + Res2 += 0x00010000; + temp1 = (Res1 << 16) & 0xffff0000; + Res0 = temp0 + temp1; + if (Res0 < temp0) + Res2++; + Res2 = Res2 + ((Res1 >> 16) & 0x0000ffff) + temp3; + if (fnLmL < 0) + { + Res2 = ~Res2; + if (Res0 == 0) + Res2++; + else + Res0 = (~Res0) + 1; + } + if (m_sh2_state->sr & S) + { + Res0 = m_sh2_state->macl + Res0; + if (m_sh2_state->macl > Res0) + Res2++; + Res2 += (m_sh2_state->mach & 0x0000ffff); + if (((int32_t) Res2 < 0) && (Res2 < 0xffff8000)) + { + Res2 = 0x00008000; + Res0 = 0x00000000; + } + else if (((int32_t) Res2 > 0) && (Res2 > 0x00007fff)) + { + Res2 = 0x00007fff; + Res0 = 0xffffffff; + } + m_sh2_state->mach = Res2; + m_sh2_state->macl = Res0; + } + else + { + Res0 = m_sh2_state->macl + Res0; + if (m_sh2_state->macl > Res0) + Res2++; + Res2 += m_sh2_state->mach; + m_sh2_state->mach = Res2; + m_sh2_state->macl = Res0; + } +} + +/*------------------------------------------------- + cfunc_DIV1 - implementation of DIV1 Rm,Rn +-------------------------------------------------*/ +static void cfunc_DIV1(void *param) +{ + ((sh2_device *)param)->func_DIV1(); +} + +void sh2_device::func_DIV1() +{ + uint32_t tmp0; + uint32_t old_q; + uint16_t opcode; + int n, m; + + // recover the opcode + opcode = m_sh2_state->arg0; + + // extract the operands + n = Rn; + m = Rm; + + old_q = m_sh2_state->sr & Q; + if (0x80000000 & m_sh2_state->r[n]) + m_sh2_state->sr |= Q; + else + m_sh2_state->sr &= ~Q; + + m_sh2_state->r[n] = (m_sh2_state->r[n] << 1) | (m_sh2_state->sr & T); + + if (!old_q) + { + if (!(m_sh2_state->sr & M)) + { + tmp0 = m_sh2_state->r[n]; + m_sh2_state->r[n] -= m_sh2_state->r[m]; + if(!(m_sh2_state->sr & Q)) + if(m_sh2_state->r[n] > tmp0) + m_sh2_state->sr |= Q; + else + m_sh2_state->sr &= ~Q; + else + if(m_sh2_state->r[n] > tmp0) + m_sh2_state->sr &= ~Q; + else + m_sh2_state->sr |= Q; + } + else + { + tmp0 = m_sh2_state->r[n]; + m_sh2_state->r[n] += m_sh2_state->r[m]; + if(!(m_sh2_state->sr & Q)) + { + if(m_sh2_state->r[n] < tmp0) + m_sh2_state->sr &= ~Q; + else + m_sh2_state->sr |= Q; + } + else + { + if(m_sh2_state->r[n] < tmp0) + m_sh2_state->sr |= Q; + else + m_sh2_state->sr &= ~Q; + } + } + } + else + { + if (!(m_sh2_state->sr & M)) + { + tmp0 = m_sh2_state->r[n]; + m_sh2_state->r[n] += m_sh2_state->r[m]; + if(!(m_sh2_state->sr & Q)) + if(m_sh2_state->r[n] < tmp0) + m_sh2_state->sr |= Q; + else + m_sh2_state->sr &= ~Q; + else + if(m_sh2_state->r[n] < tmp0) + m_sh2_state->sr &= ~Q; + else + m_sh2_state->sr |= Q; + } + else + { + tmp0 = m_sh2_state->r[n]; + m_sh2_state->r[n] -= m_sh2_state->r[m]; + if(!(m_sh2_state->sr & Q)) + if(m_sh2_state->r[n] > tmp0) + m_sh2_state->sr &= ~Q; + else + m_sh2_state->sr |= Q; + else + if(m_sh2_state->r[n] > tmp0) + m_sh2_state->sr |= Q; + else + m_sh2_state->sr &= ~Q; + } + } + + tmp0 = (m_sh2_state->sr & (Q | M)); + if((!tmp0) || (tmp0 == 0x300)) /* if Q == M set T else clear T */ + m_sh2_state->sr |= T; + else + m_sh2_state->sr &= ~T; +} + +#if (!ADDSUBV_DIRECT) +/*------------------------------------------------- + cfunc_ADDV - implementation of ADDV Rm,Rn +-------------------------------------------------*/ +static void cfunc_ADDV(void *param) +{ + ((sh2_device *)param)->func_ADDV(); +} + +void sh2_device::func_ADDV() +{ + int32_t dest, src, ans; + uint16_t opcode; + int n, m; + + // recover the opcode + opcode = m_sh2_state->arg0; + + // extract the operands + n = Rn; + m = Rm; + + if ((int32_t) m_sh2_state->r[n] >= 0) + dest = 0; + else + dest = 1; + if ((int32_t) m_sh2_state->r[m] >= 0) + src = 0; + else + src = 1; + src += dest; + m_sh2_state->r[n] += m_sh2_state->r[m]; + if ((int32_t) m_sh2_state->r[n] >= 0) + ans = 0; + else + ans = 1; + ans += dest; + if (src == 0 || src == 2) + { + if (ans == 1) + m_sh2_state->sr |= T; + else + m_sh2_state->sr &= ~T; + } + else + m_sh2_state->sr &= ~T; +} + +/*------------------------------------------------- + cfunc_SUBV - implementation of SUBV Rm,Rn +-------------------------------------------------*/ +static void cfunc_SUBV(void *param) +{ + ((sh2_device *)param)->func_SUBV(); +} + +void sh2_device::func_SUBV() +{ + int32_t dest, src, ans; + uint16_t opcode; + int n, m; + + // recover the opcode + opcode = m_sh2_state->arg0; + + // extract the operands + n = Rn; + m = Rm; + + if ((int32_t) m_sh2_state->r[n] >= 0) + dest = 0; + else + dest = 1; + if ((int32_t) m_sh2_state->r[m] >= 0) + src = 0; + else + src = 1; + src += dest; + m_sh2_state->r[n] -= m_sh2_state->r[m]; + if ((int32_t) m_sh2_state->r[n] >= 0) + ans = 0; + else + ans = 1; + ans += dest; + if (src == 1) + { + if (ans == 1) + m_sh2_state->sr |= T; + else + m_sh2_state->sr &= ~T; + } + else + m_sh2_state->sr &= ~T; +} +#else +void sh2_device::func_ADDV() {} +void sh2_device::func_SUBV() {} +#endif + +/*------------------------------------------------- + code_flush_cache - flush the cache and + regenerate static code +-------------------------------------------------*/ + +void sh2_device::code_flush_cache() +{ + drcuml_state *drcuml = m_drcuml.get(); + + /* empty the transient cache contents */ + drcuml->reset(); + + try + { + /* generate the entry point and out-of-cycles handlers */ + static_generate_nocode_handler(); + static_generate_out_of_cycles(); + static_generate_entry_point(); + + /* add subroutines for memory accesses */ + static_generate_memory_accessor(1, false, "read8", &m_read8); + static_generate_memory_accessor(1, true, "write8", &m_write8); + static_generate_memory_accessor(2, false, "read16", &m_read16); + static_generate_memory_accessor(2, true, "write16", &m_write16); + static_generate_memory_accessor(4, false, "read32", &m_read32); + static_generate_memory_accessor(4, true, "write32", &m_write32); + } + catch (drcuml_block::abort_compilation &) + { + fatalerror("Unable to generate SH2 static code\n"); + } + + m_cache_dirty = false; +} + +/* Execute cycles - returns number of cycles actually run */ +void sh2_device::execute_run_drc() +{ + drcuml_state *drcuml = m_drcuml.get(); + int execute_result; + + // run any active DMAs now +#ifndef USE_TIMER_FOR_DMA + for ( int i = 0; i < m_sh2_state->icount ; i++) + { + for( int dma=0;dma<1;dma++) + { + if (m_dma_timer_active[dma]) + sh2_do_dma(dma); + } + } +#endif + + /* reset the cache if dirty */ + if (m_cache_dirty) + code_flush_cache(); + + /* execute */ + do + { + /* run as much as we can */ + execute_result = drcuml->execute(*m_entry); + + /* if we need to recompile, do it */ + if (execute_result == EXECUTE_MISSING_CODE) + { + code_compile_block(0, m_sh2_state->pc); + } + else if (execute_result == EXECUTE_UNMAPPED_CODE) + { + fatalerror("Attempted to execute unmapped code at PC=%08X\n", m_sh2_state->pc); + } + else if (execute_result == EXECUTE_RESET_CACHE) + { + code_flush_cache(); + } + } while (execute_result != EXECUTE_OUT_OF_CYCLES); +} + +/*------------------------------------------------- + code_compile_block - compile a block of the + given mode at the specified pc +-------------------------------------------------*/ + +void sh2_device::code_compile_block(uint8_t mode, offs_t pc) +{ + drcuml_state *drcuml = m_drcuml.get(); + compiler_state compiler = { 0 }; + const opcode_desc *seqhead, *seqlast; + const opcode_desc *desclist; + bool override = false; + drcuml_block *block; + + g_profiler.start(PROFILER_DRC_COMPILE); + + /* get a description of this sequence */ + desclist = m_drcfe->describe_code(pc); + if (drcuml->logging() || drcuml->logging_native()) + log_opcode_desc(drcuml, desclist, 0); + + bool succeeded = false; + while (!succeeded) + { + try + { + /* start the block */ + block = drcuml->begin_block(4096); + + /* loop until we get through all instruction sequences */ + for (seqhead = desclist; seqhead != nullptr; seqhead = seqlast->next()) + { + const opcode_desc *curdesc; + uint32_t nextpc; + + /* add a code log entry */ + if (drcuml->logging()) + block->append_comment("-------------------------"); // comment + + /* determine the last instruction in this sequence */ + for (seqlast = seqhead; seqlast != nullptr; seqlast = seqlast->next()) + if (seqlast->flags & OPFLAG_END_SEQUENCE) + break; + assert(seqlast != nullptr); + + /* if we don't have a hash for this mode/pc, or if we are overriding all, add one */ + if (override || !drcuml->hash_exists(mode, seqhead->pc)) + UML_HASH(block, mode, seqhead->pc); // hash mode,pc + + /* if we already have a hash, and this is the first sequence, assume that we */ + /* are recompiling due to being out of sync and allow future overrides */ + else if (seqhead == desclist) + { + override = true; + UML_HASH(block, mode, seqhead->pc); // hash mode,pc + } + + /* otherwise, redispatch to that fixed PC and skip the rest of the processing */ + else + { + UML_LABEL(block, seqhead->pc | 0x80000000); // label seqhead->pc | 0x80000000 + UML_HASHJMP(block, 0, seqhead->pc, *m_nocode); + // hashjmp <mode>,seqhead->pc,nocode + continue; + } + + /* validate this code block if we're not pointing into ROM */ + if (m_program->get_write_ptr(seqhead->physpc) != nullptr) + generate_checksum_block(block, &compiler, seqhead, seqlast); + + /* label this instruction, if it may be jumped to locally */ + if (seqhead->flags & OPFLAG_IS_BRANCH_TARGET) + { + UML_LABEL(block, seqhead->pc | 0x80000000); // label seqhead->pc | 0x80000000 + } + + /* iterate over instructions in the sequence and compile them */ + for (curdesc = seqhead; curdesc != seqlast->next(); curdesc = curdesc->next()) + { + generate_sequence_instruction(block, &compiler, curdesc, 0xffffffff); + } + + /* if we need to return to the start, do it */ + if (seqlast->flags & OPFLAG_RETURN_TO_START) + { + nextpc = pc; + } + /* otherwise we just go to the next instruction */ + else + { + nextpc = seqlast->pc + (seqlast->skipslots + 1) * 2; + } + + /* count off cycles and go there */ + generate_update_cycles(block, &compiler, nextpc, true); // <subtract cycles> + + /* SH2 has no modes */ + if (seqlast->next() == nullptr || seqlast->next()->pc != nextpc) + { + UML_HASHJMP(block, 0, nextpc, *m_nocode); + } + // hashjmp <mode>,nextpc,nocode + } + + /* end the sequence */ + block->end(); + g_profiler.stop(); + succeeded = true; + } + catch (drcuml_block::abort_compilation &) + { + code_flush_cache(); + } + } +} + +/*------------------------------------------------- + static_generate_entry_point - generate a + static entry point +-------------------------------------------------*/ + +void sh2_device::static_generate_entry_point() +{ + drcuml_state *drcuml = m_drcuml.get(); + code_label skip = 1; + drcuml_block *block; + + /* begin generating */ + block = drcuml->begin_block(200); + + /* forward references */ + alloc_handle(drcuml, &m_nocode, "nocode"); + alloc_handle(drcuml, &m_write32, "write32"); // necessary? + alloc_handle(drcuml, &m_entry, "entry"); + UML_HANDLE(block, *m_entry); // handle entry + + /* load fast integer registers */ + load_fast_iregs(block); + + /* check for interrupts */ + UML_MOV(block, mem(&m_sh2_state->irqline), 0xffffffff); // mov irqline, #-1 + UML_CMP(block, mem(&m_sh2_state->pending_nmi), 0); // cmp pending_nmi, #0 + UML_JMPc(block, COND_Z, skip+2); // jz skip+2 + + UML_MOV(block, mem(&m_sh2_state->pending_nmi), 0); // zap pending_nmi + UML_JMP(block, skip+1); // and then go take it (evec is already set) + + UML_LABEL(block, skip+2); // skip+2: + UML_MOV(block, mem(&m_sh2_state->evec), 0xffffffff); // mov evec, -1 + UML_MOV(block, I0, 0xffffffff); // mov r0, -1 (r0 = irq) + UML_AND(block, I1, I0, 0xffff); // and r1, 0xffff + + UML_LZCNT(block, I1, mem(&m_sh2_state->pending_irq)); // lzcnt r1, r1 + UML_CMP(block, I1, 32); // cmp r1, #32 + UML_JMPc(block, COND_Z, skip+4); // jz skip+4 + + UML_SUB(block, mem(&m_sh2_state->irqline), 31, I1); // sub irqline, #31, r1 + + UML_LABEL(block, skip+4); // skip+4: + UML_CMP(block, mem(&m_sh2_state->internal_irq_level), 0xffffffff); // cmp internal_irq_level, #-1 + UML_JMPc(block, COND_Z, skip+3); // jz skip+3 + UML_CMP(block, mem(&m_sh2_state->internal_irq_level), mem(&m_sh2_state->irqline)); // cmp internal_irq_level, irqline + UML_JMPc(block, COND_LE, skip+3); // jle skip+3 + + UML_MOV(block, mem(&m_sh2_state->irqline), mem(&m_sh2_state->internal_irq_level)); // mov r0, internal_irq_level + + UML_LABEL(block, skip+3); // skip+3: + UML_CMP(block, mem(&m_sh2_state->irqline), 0xffffffff); // cmp irqline, #-1 + UML_JMPc(block, COND_Z, skip+1); // jz skip+1 + UML_CALLC(block, cfunc_fastirq, this); // callc fastirq + + UML_LABEL(block, skip+1); // skip+1: + + UML_CMP(block, mem(&m_sh2_state->evec), 0xffffffff); // cmp evec, 0xffffffff + UML_JMPc(block, COND_Z, skip); // jz skip + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, mem(&m_sh2_state->irqsr)); // mov r1, irqsr + UML_CALLH(block, *m_write32); // call write32 + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, mem(&m_sh2_state->pc)); // mov r1, pc + UML_CALLH(block, *m_write32); // call write32 + + UML_MOV(block, mem(&m_sh2_state->pc), mem(&m_sh2_state->evec)); // mov pc, evec + + UML_LABEL(block, skip); // skip: + + /* generate a hash jump via the current mode and PC */ + UML_HASHJMP(block, 0, mem(&m_sh2_state->pc), *m_nocode); // hashjmp <mode>,<pc>,nocode + + block->end(); +} + +/*------------------------------------------------- + static_generate_nocode_handler - generate an + exception handler for "out of code" +-------------------------------------------------*/ + +void sh2_device::static_generate_nocode_handler() +{ + drcuml_state *drcuml = m_drcuml.get(); + drcuml_block *block; + + /* begin generating */ + block = drcuml->begin_block(10); + + /* generate a hash jump via the current mode and PC */ + alloc_handle(drcuml, &m_nocode, "nocode"); + UML_HANDLE(block, *m_nocode); // handle nocode + UML_GETEXP(block, I0); // getexp i0 + UML_MOV(block, mem(&m_sh2_state->pc), I0); // mov [pc],i0 + save_fast_iregs(block); + UML_EXIT(block, EXECUTE_MISSING_CODE); // exit EXECUTE_MISSING_CODE + + block->end(); +} + + +/*------------------------------------------------- + static_generate_out_of_cycles - generate an + out of cycles exception handler +-------------------------------------------------*/ + +void sh2_device::static_generate_out_of_cycles() +{ + drcuml_state *drcuml = m_drcuml.get(); + drcuml_block *block; + + /* begin generating */ + block = drcuml->begin_block(10); + + /* generate a hash jump via the current mode and PC */ + alloc_handle(drcuml, &m_out_of_cycles, "out_of_cycles"); + UML_HANDLE(block, *m_out_of_cycles); // handle out_of_cycles + UML_GETEXP(block, I0); // getexp i0 + UML_MOV(block, mem(&m_sh2_state->pc), I0); // mov <pc>,i0 + save_fast_iregs(block); + UML_EXIT(block, EXECUTE_OUT_OF_CYCLES); // exit EXECUTE_OUT_OF_CYCLES + + block->end(); +} + +/*------------------------------------------------------------------ + static_generate_memory_accessor +------------------------------------------------------------------*/ + +void sh2_device::static_generate_memory_accessor(int size, int iswrite, const char *name, code_handle **handleptr) +{ + /* on entry, address is in I0; data for writes is in I1 */ + /* on exit, read result is in I0 */ + /* routine trashes I0 */ + drcuml_state *drcuml = m_drcuml.get(); + drcuml_block *block; + int label = 1; + + /* begin generating */ + block = drcuml->begin_block(1024); + + /* add a global entry for this */ + alloc_handle(drcuml, handleptr, name); + UML_HANDLE(block, **handleptr); // handle *handleptr + + // with internal handlers this becomes easier. + // if addr < 0x40000000 AND it with AM and do the read/write, else just do the read/write + UML_TEST(block, I0, 0x80000000); // test r0, #0x80000000 + UML_JMPc(block, COND_NZ, label); // if high bit is set, don't mask + + UML_CMP(block, I0, 0x40000000); // cmp #0x40000000, r0 + UML_JMPc(block, COND_AE, label); // bae label + + UML_AND(block, I0, I0, AM); // and r0, r0, #AM (0xc7ffffff) + + UML_LABEL(block, label++); // label: + + for (auto & elem : m_fastram) + { + if (elem.base != nullptr && (!iswrite || !elem.readonly)) + { + void *fastbase = (uint8_t *)elem.base - elem.start; + uint32_t skip = label++; + if (elem.end != 0xffffffff) + { + UML_CMP(block, I0, elem.end); // cmp i0,end + UML_JMPc(block, COND_A, skip); // ja skip + } + if (elem.start != 0x00000000) + { + UML_CMP(block, I0, elem.start);// cmp i0,fastram_start + UML_JMPc(block, COND_B, skip); // jb skip + } + + if (!iswrite) + { + if (size == 1) + { + UML_XOR(block, I0, I0, BYTE4_XOR_BE(0)); + UML_LOAD(block, I0, fastbase, I0, SIZE_BYTE, SCALE_x1); // load i0,fastbase,i0,byte + } + else if (size == 2) + { + UML_XOR(block, I0, I0, WORD_XOR_BE(0)); + UML_LOAD(block, I0, fastbase, I0, SIZE_WORD, SCALE_x1); // load i0,fastbase,i0,word_x1 + } + else if (size == 4) + { + UML_LOAD(block, I0, fastbase, I0, SIZE_DWORD, SCALE_x1); // load i0,fastbase,i0,dword_x1 + } + UML_RET(block); // ret + } + else + { + if (size == 1) + { + UML_XOR(block, I0, I0, BYTE4_XOR_BE(0)); + UML_STORE(block, fastbase, I0, I1, SIZE_BYTE, SCALE_x1);// store fastbase,i0,i1,byte + } + else if (size == 2) + { + UML_XOR(block, I0, I0, WORD_XOR_BE(0)); + UML_STORE(block, fastbase, I0, I1, SIZE_WORD, SCALE_x1);// store fastbase,i0,i1,word_x1 + } + else if (size == 4) + { + UML_STORE(block, fastbase, I0, I1, SIZE_DWORD, SCALE_x1); // store fastbase,i0,i1,dword_x1 + } + UML_RET(block); // ret + } + + UML_LABEL(block, skip); // skip: + } + } + + if (iswrite) + { + switch (size) + { + case 1: + UML_WRITE(block, I0, I1, SIZE_BYTE, SPACE_PROGRAM); // write r0, r1, program_byte + break; + + case 2: + UML_WRITE(block, I0, I1, SIZE_WORD, SPACE_PROGRAM); // write r0, r1, program_word + break; + + case 4: + UML_WRITE(block, I0, I1, SIZE_DWORD, SPACE_PROGRAM); // write r0, r1, program_dword + break; + } + } + else + { + switch (size) + { + case 1: + UML_READ(block, I0, I0, SIZE_BYTE, SPACE_PROGRAM); // read r0, program_byte + break; + + case 2: + UML_READ(block, I0, I0, SIZE_WORD, SPACE_PROGRAM); // read r0, program_word + break; + + case 4: + UML_READ(block, I0, I0, SIZE_DWORD, SPACE_PROGRAM); // read r0, program_dword + break; + } + } + + UML_RET(block); // ret + + block->end(); +} + +/*------------------------------------------------- + log_desc_flags_to_string - generate a string + representing the instruction description + flags +-------------------------------------------------*/ + +const char *sh2_device::log_desc_flags_to_string(uint32_t flags) +{ + static char tempbuf[30]; + char *dest = tempbuf; + + /* branches */ + if (flags & OPFLAG_IS_UNCONDITIONAL_BRANCH) + *dest++ = 'U'; + else if (flags & OPFLAG_IS_CONDITIONAL_BRANCH) + *dest++ = 'C'; + else + *dest++ = '.'; + + /* intrablock branches */ + *dest++ = (flags & OPFLAG_INTRABLOCK_BRANCH) ? 'i' : '.'; + + /* branch targets */ + *dest++ = (flags & OPFLAG_IS_BRANCH_TARGET) ? 'B' : '.'; + + /* delay slots */ + *dest++ = (flags & OPFLAG_IN_DELAY_SLOT) ? 'D' : '.'; + + /* exceptions */ + if (flags & OPFLAG_WILL_CAUSE_EXCEPTION) + *dest++ = 'E'; + else if (flags & OPFLAG_CAN_CAUSE_EXCEPTION) + *dest++ = 'e'; + else + *dest++ = '.'; + + /* read/write */ + if (flags & OPFLAG_READS_MEMORY) + *dest++ = 'R'; + else if (flags & OPFLAG_WRITES_MEMORY) + *dest++ = 'W'; + else + *dest++ = '.'; + + /* TLB validation */ + *dest++ = (flags & OPFLAG_VALIDATE_TLB) ? 'V' : '.'; + + /* TLB modification */ + *dest++ = (flags & OPFLAG_MODIFIES_TRANSLATION) ? 'T' : '.'; + + /* redispatch */ + *dest++ = (flags & OPFLAG_REDISPATCH) ? 'R' : '.'; + return tempbuf; +} + + +/*------------------------------------------------- + log_register_list - log a list of GPR registers +-------------------------------------------------*/ + +void sh2_device::log_register_list(drcuml_state *drcuml, const char *string, const uint32_t *reglist, const uint32_t *regnostarlist) +{ + int count = 0; + int regnum; + + /* skip if nothing */ + if (reglist[0] == 0 && reglist[1] == 0 && reglist[2] == 0) + return; + + drcuml->log_printf("[%s:", string); + + for (regnum = 0; regnum < 16; regnum++) + { + if (reglist[0] & REGFLAG_R(regnum)) + { + drcuml->log_printf("%sr%d", (count++ == 0) ? "" : ",", regnum); + if (regnostarlist != nullptr && !(regnostarlist[0] & REGFLAG_R(regnum))) + drcuml->log_printf("*"); + } + } + + if (reglist[1] & REGFLAG_PR) + { + drcuml->log_printf("%spr", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_PR)) + drcuml->log_printf("*"); + } + + if (reglist[1] & REGFLAG_SR) + { + drcuml->log_printf("%ssr", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_SR)) + drcuml->log_printf("*"); + } + + if (reglist[1] & REGFLAG_MACL) + { + drcuml->log_printf("%smacl", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_MACL)) + drcuml->log_printf("*"); + } + + if (reglist[1] & REGFLAG_MACH) + { + drcuml->log_printf("%smach", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_MACH)) + drcuml->log_printf("*"); + } + + if (reglist[1] & REGFLAG_GBR) + { + drcuml->log_printf("%sgbr", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_GBR)) + drcuml->log_printf("*"); + } + + if (reglist[1] & REGFLAG_VBR) + { + drcuml->log_printf("%svbr", (count++ == 0) ? "" : ","); + if (regnostarlist != nullptr && !(regnostarlist[1] & REGFLAG_VBR)) + drcuml->log_printf("*"); + } + + drcuml->log_printf("] "); +} + +/*------------------------------------------------- + log_opcode_desc - log a list of descriptions +-------------------------------------------------*/ + +void sh2_device::log_opcode_desc(drcuml_state *drcuml, const opcode_desc *desclist, int indent) +{ + /* open the file, creating it if necessary */ + if (indent == 0) + drcuml->log_printf("\nDescriptor list @ %08X\n", desclist->pc); + + /* output each descriptor */ + for ( ; desclist != nullptr; desclist = desclist->next()) + { + std::ostringstream stream; + + /* disassemle the current instruction and output it to the log */ + if (drcuml->logging() || drcuml->logging_native()) + { + if (desclist->flags & OPFLAG_VIRTUAL_NOOP) + stream << "<virtual nop>"; + else + DasmSH2(stream, desclist->pc, desclist->opptr.w[0]); + } + else + stream << "???"; + drcuml->log_printf("%08X [%08X] t:%08X f:%s: %-30s", desclist->pc, desclist->physpc, desclist->targetpc, log_desc_flags_to_string(desclist->flags), stream.str().c_str()); + + /* output register states */ + log_register_list(drcuml, "use", desclist->regin, nullptr); + log_register_list(drcuml, "mod", desclist->regout, desclist->regreq); + drcuml->log_printf("\n"); + + /* if we have a delay slot, output it recursively */ + if (desclist->delay.first() != nullptr) + log_opcode_desc(drcuml, desclist->delay.first(), indent + 1); + + /* at the end of a sequence add a dividing line */ + if (desclist->flags & OPFLAG_END_SEQUENCE) + drcuml->log_printf("-----\n"); + } +} + +/*------------------------------------------------- + log_add_disasm_comment - add a comment + including disassembly of an SH2 instruction +-------------------------------------------------*/ + +void sh2_device::log_add_disasm_comment(drcuml_block *block, uint32_t pc, uint32_t op) +{ + if (m_drcuml->logging()) + { + std::ostringstream stream; + DasmSH2(stream, pc, op); + block->append_comment("%08X: %s", pc, stream.str().c_str()); + } +} + +/*------------------------------------------------- + generate_update_cycles - generate code to + subtract cycles from the icount and generate + an exception if out +-------------------------------------------------*/ +void sh2_device::generate_update_cycles(drcuml_block *block, compiler_state *compiler, uml::parameter param, bool allow_exception) +{ + /* check full interrupts if pending */ + if (compiler->checkints) + { + code_label skip = compiler->labelnum++; + + compiler->checkints = false; + compiler->labelnum += 4; + + /* check for interrupts */ + UML_MOV(block, mem(&m_sh2_state->irqline), 0xffffffff); // mov irqline, #-1 + UML_CMP(block, mem(&m_sh2_state->pending_nmi), 0); // cmp pending_nmi, #0 + UML_JMPc(block, COND_Z, skip+2); // jz skip+2 + + UML_MOV(block, mem(&m_sh2_state->pending_nmi), 0); // zap pending_nmi + UML_JMP(block, skip+1); // and then go take it (evec is already set) + + UML_LABEL(block, skip+2); // skip+2: + UML_MOV(block, mem(&m_sh2_state->evec), 0xffffffff); // mov evec, -1 + UML_MOV(block, I0, 0xffffffff); // mov r0, -1 (r0 = irq) + UML_AND(block, I1, I0, 0xffff); // and r1, r0, 0xffff + + UML_LZCNT(block, I1, mem(&m_sh2_state->pending_irq)); // lzcnt r1, pending_irq + UML_CMP(block, I1, 32); // cmp r1, #32 + UML_JMPc(block, COND_Z, skip+4); // jz skip+4 + + UML_SUB(block, mem(&m_sh2_state->irqline), 31, I1); // sub irqline, #31, r1 + + UML_LABEL(block, skip+4); // skip+4: + UML_CMP(block, mem(&m_sh2_state->internal_irq_level), 0xffffffff); // cmp internal_irq_level, #-1 + UML_JMPc(block, COND_Z, skip+3); // jz skip+3 + UML_CMP(block, mem(&m_sh2_state->internal_irq_level), mem(&m_sh2_state->irqline)); // cmp internal_irq_level, irqline + UML_JMPc(block, COND_LE, skip+3); // jle skip+3 + + UML_MOV(block, mem(&m_sh2_state->irqline), mem(&m_sh2_state->internal_irq_level)); // mov r0, internal_irq_level + + UML_LABEL(block, skip+3); // skip+3: + UML_CMP(block, mem(&m_sh2_state->irqline), 0xffffffff); // cmp irqline, #-1 + UML_JMPc(block, COND_Z, skip+1); // jz skip+1 + UML_CALLC(block, cfunc_fastirq, this); // callc fastirq + + UML_LABEL(block, skip+1); // skip+1: + UML_CMP(block, mem(&m_sh2_state->evec), 0xffffffff); // cmp evec, 0xffffffff + UML_JMPc(block, COND_Z, skip); // jz skip + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, mem(&m_sh2_state->irqsr)); // mov r1, irqsr + UML_CALLH(block, *m_write32); // call write32 + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, param); // mov r1, nextpc + UML_CALLH(block, *m_write32); // call write32 + + UML_HASHJMP(block, 0, mem(&m_sh2_state->evec), *m_nocode); // hashjmp m_sh2_state->evec + + UML_LABEL(block, skip); // skip: + } + + /* account for cycles */ + if (compiler->cycles > 0) + { + UML_SUB(block, mem(&m_sh2_state->icount), mem(&m_sh2_state->icount), MAPVAR_CYCLES); // sub icount,icount,cycles + UML_MAPVAR(block, MAPVAR_CYCLES, 0); // mapvar cycles,0 + if (allow_exception) + UML_EXHc(block, COND_S, *m_out_of_cycles, param); + // exh out_of_cycles,nextpc + } + compiler->cycles = 0; +} + +/*------------------------------------------------- + generate_checksum_block - generate code to + validate a sequence of opcodes +-------------------------------------------------*/ + +void sh2_device::generate_checksum_block(drcuml_block *block, compiler_state *compiler, const opcode_desc *seqhead, const opcode_desc *seqlast) +{ + const opcode_desc *curdesc; + if (m_drcuml->logging()) + block->append_comment("[Validation for %08X]", seqhead->pc); // comment + + /* loose verify or single instruction: just compare and fail */ + if (!(m_drcoptions & SH2DRC_STRICT_VERIFY) || seqhead->next() == nullptr) + { + if (!(seqhead->flags & OPFLAG_VIRTUAL_NOOP)) + { + void *base = m_direct->read_ptr(seqhead->physpc, SH2_CODE_XOR(0)); + UML_LOAD(block, I0, base, 0, SIZE_WORD, SCALE_x2); // load i0,base,word + UML_CMP(block, I0, seqhead->opptr.w[0]); // cmp i0,*opptr + UML_EXHc(block, COND_NE, *m_nocode, epc(seqhead)); // exne nocode,seqhead->pc + } + } + + /* full verification; sum up everything */ + else + { +#if 0 + for (curdesc = seqhead->next(); curdesc != seqlast->next(); curdesc = curdesc->next()) + if (!(curdesc->flags & OPFLAG_VIRTUAL_NOOP)) + { + base = m_direct->read_ptr(curdesc->physpc, SH2_CODE_XOR(0)); + UML_LOAD(block, I0, curdesc->opptr.w, 0, SIZE_WORD, SCALE_x2); // load i0,*opptr,0,word + UML_CMP(block, I0, curdesc->opptr.w[0]); // cmp i0,*opptr + UML_EXHc(block, COND_NE, *m_nocode, epc(seqhead)); // exne nocode,seqhead->pc + } +#else + uint32_t sum = 0; + void *base = m_direct->read_ptr(seqhead->physpc, SH2_CODE_XOR(0)); + UML_LOAD(block, I0, base, 0, SIZE_WORD, SCALE_x4); // load i0,base,word + sum += seqhead->opptr.w[0]; + for (curdesc = seqhead->next(); curdesc != seqlast->next(); curdesc = curdesc->next()) + if (!(curdesc->flags & OPFLAG_VIRTUAL_NOOP)) + { + base = m_direct->read_ptr(curdesc->physpc, SH2_CODE_XOR(0)); + UML_LOAD(block, I1, base, 0, SIZE_WORD, SCALE_x2); // load i1,*opptr,word + UML_ADD(block, I0, I0, I1); // add i0,i0,i1 + sum += curdesc->opptr.w[0]; + } + UML_CMP(block, I0, sum); // cmp i0,sum + UML_EXHc(block, COND_NE, *m_nocode, epc(seqhead)); // exne nocode,seqhead->pc +#endif + } +} + + +/*------------------------------------------------- + generate_sequence_instruction - generate code + for a single instruction in a sequence +-------------------------------------------------*/ + +void sh2_device::generate_sequence_instruction(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint32_t ovrpc) +{ + offs_t expc; + + /* add an entry for the log */ + if (m_drcuml->logging() && !(desc->flags & OPFLAG_VIRTUAL_NOOP)) + log_add_disasm_comment(block, desc->pc, desc->opptr.w[0]); + + /* set the PC map variable */ + expc = (desc->flags & OPFLAG_IN_DELAY_SLOT) ? desc->pc - 1 : desc->pc; + UML_MAPVAR(block, MAPVAR_PC, expc); // mapvar PC,expc + + /* accumulate total cycles */ + compiler->cycles += desc->cycles; + + /* update the icount map variable */ + UML_MAPVAR(block, MAPVAR_CYCLES, compiler->cycles); // mapvar CYCLES,compiler->cycles + + /* if we want a probe, add it here */ + if (desc->pc == PROBE_ADDRESS) + { + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov [pc],desc->pc + UML_CALLC(block, cfunc_printf_probe, this); // callc cfunc_printf_probe,sh2 + } + + /* if we are debugging, call the debugger */ + if ((machine().debug_flags & DEBUG_FLAG_ENABLED) != 0) + { + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov [pc],desc->pc + save_fast_iregs(block); + UML_DEBUG(block, desc->pc); // debug desc->pc + } + else // not debug, see what other reasons there are for flushing the PC + { + if (m_drcoptions & SH2DRC_FLUSH_PC) // always flush? + { + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov m_sh2_state->pc, desc->pc + } + else // check for driver-selected flushes + { + int pcflush; + + for (pcflush = 0; pcflush < m_pcfsel; pcflush++) + { + if (desc->pc == m_pcflushes[pcflush]) + { + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov m_sh2_state->pc, desc->pc + } + } + } + } + + + /* if we hit an unmapped address, fatal error */ + if (desc->flags & OPFLAG_COMPILER_UNMAPPED) + { + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov [pc],desc->pc + save_fast_iregs(block); + UML_EXIT(block, EXECUTE_UNMAPPED_CODE); // exit EXECUTE_UNMAPPED_CODE + } + + /* if this is an invalid opcode, die */ + if (desc->flags & OPFLAG_INVALID_OPCODE) + { + fatalerror("SH2DRC: invalid opcode!\n"); + } + + /* otherwise, unless this is a virtual no-op, it's a regular instruction */ + else if (!(desc->flags & OPFLAG_VIRTUAL_NOOP)) + { + /* compile the instruction */ + if (!generate_opcode(block, compiler, desc, ovrpc)) + { + // handle an illegal op + UML_MOV(block, mem(&m_sh2_state->pc), desc->pc); // mov [pc],desc->pc + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); // mov [arg0],opcode + UML_CALLC(block, cfunc_unimplemented, this); // callc cfunc_unimplemented + } + } +} + +/*------------------------------------------------------------------ + generate_delay_slot +------------------------------------------------------------------*/ + +void sh2_device::generate_delay_slot(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint32_t ovrpc) +{ + compiler_state compiler_temp = *compiler; + + /* compile the delay slot using temporary compiler state */ + assert(desc->delay.first() != nullptr); + generate_sequence_instruction(block, &compiler_temp, desc->delay.first(), ovrpc); // <next instruction> + + /* update the label */ + compiler->labelnum = compiler_temp.labelnum; +} + +/*------------------------------------------------- + generate_opcode - generate code for a specific + opcode +-------------------------------------------------*/ + +bool sh2_device::generate_opcode(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint32_t ovrpc) +{ + uint32_t scratch, scratch2; + int32_t disp; + uint16_t opcode = desc->opptr.w[0]; + uint8_t opswitch = opcode >> 12; + int in_delay_slot = ((desc->flags & OPFLAG_IN_DELAY_SLOT) != 0); + + switch (opswitch) + { + case 0: + return generate_group_0(block, compiler, desc, opcode, in_delay_slot, ovrpc); + + case 1: // MOVLS4 + scratch = (opcode & 0x0f) * 4; + UML_ADD(block, I0, R32(Rn), scratch); // add r0, Rn, scratch + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + SETEA(0); // set ea for debug + UML_CALLH(block, *m_write32); + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 2: + return generate_group_2(block, compiler, desc, opcode, in_delay_slot, ovrpc); + case 3: + return generate_group_3(block, compiler, desc, opcode, ovrpc); + case 4: + return generate_group_4(block, compiler, desc, opcode, in_delay_slot, ovrpc); + + case 5: // MOVLL4 + scratch = (opcode & 0x0f) * 4; + UML_ADD(block, I0, R32(Rm), scratch); // add r0, Rm, scratch + SETEA(0); // set ea for debug + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, R32(Rn), I0); // mov Rn, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 6: + return generate_group_6(block, compiler, desc, opcode, in_delay_slot, ovrpc); + + case 7: // ADDI + scratch = opcode & 0xff; + scratch2 = (uint32_t)(int32_t)(int16_t)(int8_t)scratch; + UML_ADD(block, R32(Rn), R32(Rn), scratch2); // add Rn, Rn, scratch2 + return true; + + case 8: + return generate_group_8(block, compiler, desc, opcode, in_delay_slot, ovrpc); + + case 9: // MOVWI + if (ovrpc == 0xffffffff) + { + scratch = (desc->pc + 2) + ((opcode & 0xff) * 2) + 2; + } + else + { + scratch = (ovrpc + 2) + ((opcode & 0xff) * 2) + 2; + } + + if (m_drcoptions & SH2DRC_STRICT_PCREL) + { + UML_MOV(block, I0, scratch); // mov r0, scratch + SETEA(0); // set ea for debug + UML_CALLH(block, *m_read16); // read16(r0, r1) + UML_SEXT(block, R32(Rn), I0, SIZE_WORD); // sext Rn, r0, WORD + } + else + { + scratch2 = (uint32_t)(int32_t)(int16_t) RW(scratch); + UML_MOV(block, R32(Rn), scratch2); // mov Rn, scratch2 + } + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 10: // BRA + disp = ((int32_t)opcode << 20) >> 20; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = pc+4 + disp*2 + 2 + + generate_delay_slot(block, compiler, desc, m_sh2_state->ea-2); + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // hashjmp m_sh2_state->ea + return true; + + case 11: // BSR + // panicstr @ 403da22 relies on the delay slot clobbering the PR set by a BSR, so + // do this before running the delay slot + UML_ADD(block, mem(&m_sh2_state->pr), desc->pc, 4); // add m_pr, desc->pc, #4 (skip the current insn & delay slot) + + disp = ((int32_t)opcode << 20) >> 20; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = pc+4 + disp*2 + 2 + + generate_delay_slot(block, compiler, desc, m_sh2_state->ea-2); + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // hashjmp m_sh2_state->ea + return true; + + case 12: + return generate_group_12(block, compiler, desc, opcode, in_delay_slot, ovrpc); + + case 13: // MOVLI + if (ovrpc == 0xffffffff) + { + scratch = ((desc->pc + 4) & ~3) + ((opcode & 0xff) * 4); + } + else + { + scratch = ((ovrpc + 4) & ~3) + ((opcode & 0xff) * 4); + } + + if (m_drcoptions & SH2DRC_STRICT_PCREL) + { + UML_MOV(block, I0, scratch); // mov r0, scratch + UML_CALLH(block, *m_read32); // read32(r0, r1) + UML_MOV(block, R32(Rn), I0); // mov Rn, r0 + } + else + { + scratch2 = RL(scratch); + UML_MOV(block, R32(Rn), scratch2); // mov Rn, scratch2 + } + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 14: // MOVI + scratch = opcode & 0xff; + scratch2 = (uint32_t)(int32_t)(int16_t)(int8_t)scratch; + UML_MOV(block, R32(Rn), scratch2); + return true; + + case 15: + return false; + } + + return false; +} + +bool sh2_device::generate_group_0(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + switch (opcode & 0x3F) + { + case 0x00: // these are all illegal + case 0x01: + case 0x10: + case 0x11: + case 0x13: + case 0x20: + case 0x21: + case 0x30: + case 0x31: + case 0x32: + case 0x33: + case 0x38: + case 0x39: + case 0x3a: + case 0x3b: + return false; + + case 0x09: // NOP(); + return true; + + case 0x02: // STCSR(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->sr)); + return true; + + case 0x03: // BSRF(Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_ADD(block, mem(&m_sh2_state->target), R32(Rn), 4); // add target, Rm, #4 + UML_ADD(block, mem(&m_sh2_state->target), mem(&m_sh2_state->target), desc->pc); // add target, target, pc + + // 32x Cosmic Carnage @ 6002cb0 relies on the delay slot + // clobbering the calculated PR, so do it first + UML_ADD(block, mem(&m_sh2_state->pr), desc->pc, 4); // add m_pr, desc->pc, #4 (skip the current insn & delay slot) + + generate_delay_slot(block, compiler, desc, m_sh2_state->target); + + generate_update_cycles(block, compiler, mem(&m_sh2_state->target), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->target), *m_nocode); // jmp target + return true; + } + break; + + case 0x04: // MOVBS0(Rm, Rn); + case 0x14: // MOVBS0(Rm, Rn); + case 0x24: // MOVBS0(Rm, Rn); + case 0x34: // MOVBS0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rn)); // add r0, R0, Rn + UML_AND(block, I1, R32(Rm), 0x000000ff); // and r1, Rm, 0xff + UML_CALLH(block, *m_write8); // call write8 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x05: // MOVWS0(Rm, Rn); + case 0x15: // MOVWS0(Rm, Rn); + case 0x25: // MOVWS0(Rm, Rn); + case 0x35: // MOVWS0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rn)); // add r0, R0, Rn + UML_AND(block, I1, R32(Rm), 0x0000ffff); // and r1, Rm, 0xffff + UML_CALLH(block, *m_write16); // call write16 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x06: // MOVLS0(Rm, Rn); + case 0x16: // MOVLS0(Rm, Rn); + case 0x26: // MOVLS0(Rm, Rn); + case 0x36: // MOVLS0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rn)); // add r0, R0, Rn + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x07: // MULL(Rm, Rn); + case 0x17: // MULL(Rm, Rn); + case 0x27: // MULL(Rm, Rn); + case 0x37: // MULL(Rm, Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_MULU(block, mem(&m_sh2_state->macl), mem(&m_sh2_state->ea), R32(Rn), R32(Rm)); // mulu macl, ea, Rn, Rm + return true; + } + break; + + case 0x08: // CLRT(); + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + return true; + + case 0x0a: // STSMACH(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->mach)); // mov Rn, mach + return true; + + case 0x0b: // RTS(); + UML_MOV(block, mem(&m_sh2_state->target), mem(&m_sh2_state->pr)); // mov target, pr (in case of d-slot shenanigans) + + generate_delay_slot(block, compiler, desc, m_sh2_state->target); + + generate_update_cycles(block, compiler, mem(&m_sh2_state->target), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->target), *m_nocode); + return true; + + case 0x0c: // MOVBL0(Rm, Rn); + case 0x1c: // MOVBL0(Rm, Rn); + case 0x2c: // MOVBL0(Rm, Rn); + case 0x3c: // MOVBL0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rm)); // add r0, R0, Rm + UML_CALLH(block, *m_read8); // call read8 + UML_SEXT(block, R32(Rn), I0, SIZE_BYTE); // sext Rn, r0, BYTE + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x0d: // MOVWL0(Rm, Rn); + case 0x1d: // MOVWL0(Rm, Rn); + case 0x2d: // MOVWL0(Rm, Rn); + case 0x3d: // MOVWL0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rm)); // add r0, R0, Rm + UML_CALLH(block, *m_read16); // call read16 + UML_SEXT(block, R32(Rn), I0, SIZE_WORD); // sext Rn, r0, WORD + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x0e: // MOVLL0(Rm, Rn); + case 0x1e: // MOVLL0(Rm, Rn); + case 0x2e: // MOVLL0(Rm, Rn); + case 0x3e: // MOVLL0(Rm, Rn); + UML_ADD(block, I0, R32(0), R32(Rm)); // add r0, R0, Rm + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, R32(Rn), I0); // mov Rn, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x0f: // MAC_L(Rm, Rn); + case 0x1f: // MAC_L(Rm, Rn); + case 0x2f: // MAC_L(Rm, Rn); + case 0x3f: // MAC_L(Rm, Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + save_fast_iregs(block); + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); + UML_CALLC(block, cfunc_MAC_L, this); + load_fast_iregs(block); + return true; + } + break; + + case 0x12: // STCGBR(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->gbr)); // mov Rn, gbr + return true; + + case 0x18: // SETT(); + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + return true; + + case 0x19: // DIV0U(); + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~(M|Q|T)); // and sr, sr, ~(M|Q|T) + return true; + + case 0x1a: // STSMACL(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->macl)); // mov Rn, macl + return true; + + case 0x1b: // SLEEP(); + UML_MOV(block, I0, mem(&m_sh2_state->sleep_mode)); // mov i0, sleep_mode + UML_CMP(block, I0, 0x2); // cmp i0, #2 + UML_JMPc(block, COND_E, compiler->labelnum); // beq labelnum + // sleep mode != 2 + UML_MOV(block, mem(&m_sh2_state->sleep_mode), 0x1); // mov sleep_mode, #1 + generate_update_cycles(block, compiler, desc->pc, true); // repeat this insn + UML_JMP(block, compiler->labelnum+1); // jmp labelnum+1 + + UML_LABEL(block, compiler->labelnum++); // labelnum: + // sleep_mode == 2 + UML_MOV(block, mem(&m_sh2_state->sleep_mode), 0x0); // sleep_mode = 0 + generate_update_cycles(block, compiler, desc->pc+2, true); // go to next insn + + UML_LABEL(block, compiler->labelnum++); // labelnum+1: + return true; + + case 0x22: // STCVBR(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->vbr)); // mov Rn, vbr + return true; + + case 0x23: // BRAF(Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_ADD(block, mem(&m_sh2_state->target), R32(Rn), desc->pc+4); // add target, Rn, pc+4 + + generate_delay_slot(block, compiler, desc, m_sh2_state->target); + + generate_update_cycles(block, compiler, mem(&m_sh2_state->target), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->target), *m_nocode); // jmp target + return true; + } + break; + + case 0x28: // CLRMAC(); + UML_MOV(block, mem(&m_sh2_state->macl), 0); // mov macl, #0 + UML_MOV(block, mem(&m_sh2_state->mach), 0); // mov mach, #0 + return true; + + case 0x29: // MOVT(Rn); + UML_AND(block, R32(Rn), mem(&m_sh2_state->sr), T); // and Rn, sr, T + return true; + + case 0x2a: // STSPR(Rn); + UML_MOV(block, R32(Rn), mem(&m_sh2_state->pr)); // mov Rn, pr + return true; + + case 0x2b: // RTE(); + generate_delay_slot(block, compiler, desc, 0xffffffff); + + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, mem(&m_sh2_state->pc), I0); // mov pc, r0 + UML_ADD(block, R32(15), R32(15), 4); // add R15, R15, #4 + + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov sr, r0 + UML_ADD(block, R32(15), R32(15), 4); // add R15, R15, #4 + + compiler->checkints = true; + UML_MOV(block, mem(&m_sh2_state->ea), mem(&m_sh2_state->pc)); // mov ea, pc + generate_update_cycles(block, compiler, mem(&m_sh2_state->ea), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->pc), *m_nocode); // and jump to the "resume PC" + + return true; + } + + return false; +} + +bool sh2_device::generate_group_2(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + switch (opcode & 15) + { + case 0: // MOVBS(Rm, Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_AND(block, I1, R32(Rm), 0xff); // and r1, Rm, 0xff + UML_CALLH(block, *m_write8); + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 1: // MOVWS(Rm, Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_AND(block, I1, R32(Rm), 0xffff); // and r1, Rm, 0xffff + UML_CALLH(block, *m_write16); + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 2: // MOVLS(Rm, Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + UML_CALLH(block, *m_write32); + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 3: + return false; + + case 4: // MOVBM(Rm, Rn); + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + UML_SUB(block, R32(Rn), R32(Rn), 1); // sub Rn, Rn, 1 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_CALLH(block, *m_write8); // call write8 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 5: // MOVWM(Rm, Rn); + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + UML_SUB(block, R32(Rn), R32(Rn), 2); // sub Rn, Rn, 2 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_CALLH(block, *m_write16); // call write16 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 6: // MOVLM(Rm, Rn); + UML_MOV(block, I1, R32(Rm)); // mov r1, Rm + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, 4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 13: // XTRCT(Rm, Rn); + UML_SHL(block, I0, R32(Rm), 16); // shl r0, Rm, #16 + UML_AND(block, I0, I0, 0xffff0000); // and r0, r0, #0xffff0000 + + UML_SHR(block, I1, R32(Rn), 16); // shr, r1, Rn, #16 + UML_AND(block, I1, I1, 0xffff); // and r1, r1, #0x0000ffff + + UML_OR(block, R32(Rn), I0, I1); // or Rn, r0, r1 + return true; + + case 7: // DIV0S(Rm, Rn); + UML_MOV(block, I0, mem(&m_sh2_state->sr)); // move r0, sr + UML_AND(block, I0, I0, ~(Q|M|T)); // and r0, r0, ~(Q|M|T) (clear the Q,M, and T bits) + + UML_TEST(block, R32(Rn), 0x80000000); // test Rn, #0x80000000 + UML_JMPc(block, COND_Z, compiler->labelnum); // jz labelnum + + UML_OR(block, I0, I0, Q); // or r0, r0, Q + UML_LABEL(block, compiler->labelnum++); // labelnum: + + UML_TEST(block, R32(Rm), 0x80000000); // test Rm, #0x80000000 + UML_JMPc(block, COND_Z, compiler->labelnum); // jz labelnum + + UML_OR(block, I0, I0, M); // or r0, r0, M + UML_LABEL(block, compiler->labelnum++); // labelnum: + + UML_XOR(block, I1, R32(Rn), R32(Rm)); // xor r1, Rn, Rm + UML_TEST(block, I1, 0x80000000); // test r1, #0x80000000 + UML_JMPc(block, COND_Z, compiler->labelnum); // jz labelnum + + UML_OR(block, I0, I0, T); // or r0, r0, T + UML_LABEL(block, compiler->labelnum++); // labelnum: + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov sr, r0 + return true; + + case 8: // TST(Rm, Rn); + UML_AND(block, I0, mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + UML_TEST(block, R32(Rm), R32(Rn)); // test Rm, Rn + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz compiler->labelnum + + UML_OR(block, I0, I0, T); // or r0, r0, T + UML_LABEL(block, compiler->labelnum++); // desc->pc: + + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov m_sh2_state->sr, r0 + return true; + + case 12: // CMPSTR(Rm, Rn); + UML_XOR(block, I0, R32(Rn), R32(Rm)); // xor r0, Rn, Rm (temp) + + UML_SHR(block, I1, I0, 24); // shr r1, r0, #24 (HH) + UML_AND(block, I1, I1, 0xff); // and r1, r1, #0xff + + UML_SHR(block, I2, I0, 16); // shr r2, r0, #16 (HL) + UML_AND(block, I2, I2, 0xff); // and r2, r2, #0xff + + UML_SHR(block, I3, I0, 8); // shr r3, r0, #8 (LH) + UML_AND(block, I3, I3, 0xff); // and r3, r3, #0xff + + UML_AND(block, I7, I0, 0xff); // and r7, r0, #0xff (LL) + + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T (clear the T bit) + + UML_CMP(block, I1, 0); // cmp r1, #0 + UML_JMPc(block, COND_Z, compiler->labelnum); // jnz labelnum + UML_CMP(block, I2, 0); // cmp r2, #0 + UML_JMPc(block, COND_Z, compiler->labelnum); // jnz labelnum + UML_CMP(block, I3, 0); // cmp r3, #0 + UML_JMPc(block, COND_Z, compiler->labelnum); // jnz labelnum + UML_CMP(block, I7, 0); // cmp r7, #0 + UML_JMPc(block, COND_NZ, compiler->labelnum+1); // jnz labelnum + + UML_LABEL(block, compiler->labelnum++); // labelnum: + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + + UML_LABEL(block, compiler->labelnum++); // labelnum+1: + return true; + + case 9: // AND(Rm, Rn); + UML_AND(block, R32(Rn), R32(Rn), R32(Rm)); // and Rn, Rn, Rm + return true; + + case 10: // XOR(Rm, Rn); + UML_XOR(block, R32(Rn), R32(Rn), R32(Rm)); // xor Rn, Rn, Rm + return true; + + case 11: // OR(Rm, Rn); + UML_OR(block, R32(Rn), R32(Rn), R32(Rm)); // or Rn, Rn, Rm + return true; + + case 14: // MULU(Rm, Rn); + UML_AND(block, I0, R32(Rm), 0xffff); // and r0, Rm, 0xffff + UML_AND(block, I1, R32(Rn), 0xffff); // and r1, Rn, 0xffff + UML_MULU(block, mem(&m_sh2_state->macl), mem(&m_sh2_state->ea), I0, I1); // mulu macl, ea, r0, r1 + return true; + + case 15: // MULS(Rm, Rn); + UML_SEXT(block, I0, R32(Rm), SIZE_WORD); // sext r0, Rm + UML_SEXT(block, I1, R32(Rn), SIZE_WORD); // sext r1, Rn + UML_MULS(block, mem(&m_sh2_state->macl), mem(&m_sh2_state->ea), I0, I1); // muls macl, ea, r0, r1 + return true; + } + + return false; +} + +bool sh2_device::generate_group_3(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, uint32_t ovrpc) +{ + switch (opcode & 15) + { + case 0: // CMPEQ(Rm, Rn); (equality) + UML_CMP(block, R32(Rn), R32(Rm)); // cmp Rn, Rm + UML_SETc(block, COND_E, I0); // set E, r0 + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, 1); // rolins sr, r0, 0, 1 + return true; + + case 2: // CMPHS(Rm, Rn); (unsigned greater than or equal) + UML_CMP(block, R32(Rn), R32(Rm)); // cmp Rn, Rm + UML_SETc(block, COND_AE, I0); // set AE, r0 + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, 1); // rolins sr, r0, 0, 1 + return true; + + case 3: // CMPGE(Rm, Rn); (signed greater than or equal) + UML_CMP(block, R32(Rn), R32(Rm)); // cmp Rn, Rm + UML_SETc(block, COND_GE, I0); // set GE, r0 + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, 1); // rolins sr, r0, 0, 1 + return true; + + case 6: // CMPHI(Rm, Rn); (unsigned greater than) + UML_CMP(block, R32(Rn), R32(Rm)); // cmp Rn, Rm + UML_SETc(block, COND_A, I0); // set A, r0 + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, 1); // rolins sr, r0, 0, 1 + return true; + + case 7: // CMPGT(Rm, Rn); (signed greater than) + UML_CMP(block, R32(Rn), R32(Rm)); // cmp Rn, Rm + UML_SETc(block, COND_G, I0); // set G, r0 + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, 1); // rolins sr, r0, 0, 1 + return true; + + case 1: + case 9: + return false; + + case 4: // DIV1(Rm, Rn); + save_fast_iregs(block); + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); + UML_CALLC(block, cfunc_DIV1, this); + load_fast_iregs(block); + return true; + + case 5: // DMULU(Rm, Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_MULU(block, mem(&m_sh2_state->macl), mem(&m_sh2_state->mach), R32(Rn), R32(Rm)); + return true; + } + break; + + case 13: // DMULS(Rm, Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_MULS(block, mem(&m_sh2_state->macl), mem(&m_sh2_state->mach), R32(Rn), R32(Rm)); + return true; + } + break; + + case 8: // SUB(Rm, Rn); + UML_SUB(block, R32(Rn), R32(Rn), R32(Rm)); // sub Rn, Rn, Rm + return true; + + case 12: // ADD(Rm, Rn); + UML_ADD(block, R32(Rn), R32(Rn), R32(Rm)); // add Rn, Rn, Rm + return true; + + case 10: // SUBC(Rm, Rn); + UML_CARRY(block, mem(&m_sh2_state->sr), 0); // carry = T (T is bit 0 of SR) + UML_SUBB(block, R32(Rn), R32(Rn), R32(Rm)); // addc Rn, Rn, Rm + UML_SETc(block, COND_C, I0); // setc i0, C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins sr,i0,0,T + return true; + + case 11: // SUBV(Rm, Rn); +#if ADDSUBV_DIRECT + UML_SUB(block, R32(Rn), R32(Rn), R32(Rm)); // sub Rn, Rn, Rm + UML_SETc(block, COND_V, I0); // setc i0, V + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T +#else + save_fast_iregs(block); + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); + UML_CALLC(block, cfunc_SUBV, this); + load_fast_iregs(block); +#endif + return true; + + case 14: // ADDC(Rm, Rn); + UML_CARRY(block, mem(&m_sh2_state->sr), 0); // carry = T (T is bit 0 of SR) + UML_ADDC(block, R32(Rn), R32(Rn), R32(Rm)); // addc Rn, Rn, Rm + UML_SETc(block, COND_C, I0); // setc i0, C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins sr,i0,0,T + return true; + + case 15: // ADDV(Rm, Rn); +#if ADDSUBV_DIRECT + UML_ADD(block, R32(Rn), R32(Rn), R32(Rm)); // add Rn, Rn, Rm + UML_SETc(block, COND_V, I0); // setc i0, V + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T +#else + save_fast_iregs(block); + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); + UML_CALLC(block, cfunc_ADDV, this); + load_fast_iregs(block); +#endif + return true; + } + return false; +} + +bool sh2_device::generate_group_4(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + switch (opcode & 0x3F) + { + case 0x00: // SHLL(Rn); + UML_SHL(block, R32(Rn), R32(Rn), 1); // shl Rn, Rn, 1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T + return true; + + case 0x01: // SHLR(Rn); + UML_SHR(block, R32(Rn), R32(Rn), 1); // shr Rn, Rn, 1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T + return true; + + case 0x04: // ROTL(Rn); + UML_ROL(block, R32(Rn), R32(Rn), 1); // rol Rn, Rn, 1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T + return true; + + case 0x05: // ROTR(Rn); + UML_ROR(block, R32(Rn), R32(Rn), 1); // ror Rn, Rn, 1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins [sr],i0,0,T + return true; + + case 0x02: // STSMMACH(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, #4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_MOV(block, I1, mem(&m_sh2_state->mach)); // mov r1, mach + SETEA(0); // set ea for debug + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x03: // STCMSR(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, #4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_MOV(block, I1, mem(&m_sh2_state->sr)); // mov r1, sr + SETEA(0); // set ea for debug + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x06: // LDSMMACH(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, #4 + UML_MOV(block, mem(&m_sh2_state->mach), I0); // mov mach, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x07: // LDCMSR(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, #4 + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov sr, r0 + + compiler->checkints = true; + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + + case 0x08: // SHLL2(Rn); + UML_SHL(block, R32(Rn), R32(Rn), 2); + return true; + + case 0x09: // SHLR2(Rn); + UML_SHR(block, R32(Rn), R32(Rn), 2); + return true; + + case 0x18: // SHLL8(Rn); + UML_SHL(block, R32(Rn), R32(Rn), 8); + return true; + + case 0x19: // SHLR8(Rn); + UML_SHR(block, R32(Rn), R32(Rn), 8); + return true; + + case 0x28: // SHLL16(Rn); + UML_SHL(block, R32(Rn), R32(Rn), 16); + return true; + + case 0x29: // SHLR16(Rn); + UML_SHR(block, R32(Rn), R32(Rn), 16); + return true; + + case 0x0a: // LDSMACH(Rn); + UML_MOV(block, mem(&m_sh2_state->mach), R32(Rn)); // mov mach, Rn + return true; + + case 0x0b: // JSR(Rn); + UML_MOV(block, mem(&m_sh2_state->target), R32(Rn)); // mov target, Rn + + UML_ADD(block, mem(&m_sh2_state->pr), desc->pc, 4); // add m_pr, desc->pc, #4 (skip the current insn & delay slot) + + generate_delay_slot(block, compiler, desc, m_sh2_state->target-4); + + generate_update_cycles(block, compiler, mem(&m_sh2_state->target), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->target), *m_nocode); // and do the jump + return true; + + case 0x0e: // LDCSR(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_AND(block, I0, I0, FLAGS); // and r0, r0, FLAGS + UML_MOV(block, mem(&m_sh2_state->sr), I0); + + compiler->checkints = true; + return true; + + case 0x0f: // MAC_W(Rm, Rn); + case 0x1f: // MAC_W(Rm, Rn); + case 0x2f: // MAC_W(Rm, Rn); + case 0x3f: // MAC_W(Rm, Rn); + save_fast_iregs(block); + UML_MOV(block, mem(&m_sh2_state->arg0), desc->opptr.w[0]); + UML_CALLC(block, cfunc_MAC_W, this); + load_fast_iregs(block); + return true; + + case 0x10: // DT(Rn); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_AND(block, I0, mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + UML_SUB(block, R32(Rn), R32(Rn), 1); // sub Rn, Rn, 1 + UML_JMPc(block, COND_NZ, compiler->labelnum); // jz compiler->labelnum + + UML_OR(block, I0, I0, T); // or r0, r0, T + UML_LABEL(block, compiler->labelnum++); // desc->pc: + + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov m_sh2_state->sr, r0 + return true; + } + break; + + case 0x11: // CMPPZ(Rn); + UML_AND(block, I0, mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + + UML_CMP(block, R32(Rn), 0); // cmp Rn, 0 + UML_JMPc(block, COND_S, compiler->labelnum); // js compiler->labelnum (if negative) + + UML_OR(block, I0, I0, T); // or r0, r0, T + UML_LABEL(block, compiler->labelnum++); // desc->pc: + + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov m_sh2_state->sr, r0 + return true; + + case 0x15: // CMPPL(Rn); + UML_AND(block, I0, mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + + UML_CMP(block, R32(Rn), 0); // cmp Rn, 0 + + UML_JMPc(block, COND_S, compiler->labelnum); // js compiler->labelnum (if negative) + UML_JMPc(block, COND_Z, compiler->labelnum); // jz compiler->labelnum (if zero) + + UML_OR(block, I0, I0, T); // or r0, r0, T + + UML_LABEL(block, compiler->labelnum++); // desc->pc: + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov m_sh2_state->sr, r0 + return true; + + case 0x12: // STSMMACL(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, #4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_MOV(block, I1, mem(&m_sh2_state->macl)); // mov r1, macl + SETEA(0); // set ea for debug + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x13: // STCMGBR(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, #4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_MOV(block, I1, mem(&m_sh2_state->gbr)); // mov r1, gbr + SETEA(0); // set ea for debug + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x16: // LDSMMACL(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, #4 + UML_MOV(block, mem(&m_sh2_state->macl), I0); // mov macl, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x17: // LDCMGBR(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, #4 + UML_MOV(block, mem(&m_sh2_state->gbr), I0); // mov gbr, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x1a: // LDSMACL(Rn); + UML_MOV(block, mem(&m_sh2_state->macl), R32(Rn)); // mov macl, Rn + return true; + + case 0x1b: // TAS(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read8); // call read8 + + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T + + UML_CMP(block, I0, 0); // cmp r0, #0 + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz labelnum + + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + + UML_LABEL(block, compiler->labelnum++); // labelnum: + + UML_OR(block, I1, I0, 0x80); // or r1, r0, #0x80 + + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + UML_CALLH(block, *m_write8); // write the value back + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x1e: // LDCGBR(Rn); + UML_MOV(block, mem(&m_sh2_state->gbr), R32(Rn)); // mov gbr, Rn + return true; + + case 0x20: // SHAL(Rn); + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T + UML_SHR(block, I0, R32(Rn), 31); // shr r0, Rn, 31 + UML_AND(block, I0, I0, T); // and r0, r0, T + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), I0); // or sr, sr, r0 + UML_SHL(block, R32(Rn), R32(Rn), 1); // shl Rn, Rn, 1 + return true; + + case 0x21: // SHAR(Rn); + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T + UML_AND(block, I0, R32(Rn), T); // and r0, Rn, T + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), I0); // or sr, sr, r0 + UML_SAR(block, R32(Rn), R32(Rn), 1); // sar Rn, Rn, 1 + return true; + + case 0x22: // STSMPR(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, 4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_MOV(block, I1, mem(&m_sh2_state->pr)); // mov r1, pr + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x23: // STCMVBR(Rn); + UML_SUB(block, R32(Rn), R32(Rn), 4); // sub Rn, Rn, 4 + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_MOV(block, I1, mem(&m_sh2_state->vbr)); // mov r1, vbr + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x24: // ROTCL(Rn); + UML_CARRY(block, mem(&m_sh2_state->sr), 0); // carry sr,0 + UML_ROLC(block, R32(Rn), R32(Rn), 1); // rolc Rn,Rn,1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins sr,i0,0,T + return true; + + case 0x25: // ROTCR(Rn); + UML_CARRY(block, mem(&m_sh2_state->sr), 0); // carry sr,0 + UML_RORC(block, R32(Rn), R32(Rn), 1); // rorc Rn,Rn,1 + UML_SETc(block, COND_C, I0); // set i0,C + UML_ROLINS(block, mem(&m_sh2_state->sr), I0, 0, T); // rolins sr,i0,0,T + return true; + + case 0x26: // LDSMPR(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, mem(&m_sh2_state->pr), I0); // mov m_pr, r0 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, Rn, #4 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x27: // LDCMVBR(Rn); + UML_MOV(block, I0, R32(Rn)); // mov r0, Rn + SETEA(0); + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, mem(&m_sh2_state->vbr), I0); // mov m_sh2_state->vbr, r0 + UML_ADD(block, R32(Rn), R32(Rn), 4); // add Rn, Rn, #4 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 0x2a: // LDSPR(Rn); + UML_MOV(block, mem(&m_sh2_state->pr), R32(Rn)); // mov m_pr, Rn + return true; + + case 0x2b: // JMP(Rn); + UML_MOV(block, mem(&m_sh2_state->target), R32(Rn)); // mov target, Rn + + generate_delay_slot(block, compiler, desc, m_sh2_state->target); + + generate_update_cycles(block, compiler, mem(&m_sh2_state->target), true); // <subtract cycles> + UML_HASHJMP(block, 0, mem(&m_sh2_state->target), *m_nocode); // jmp (target) + return true; + + case 0x2e: // LDCVBR(Rn); + UML_MOV(block, mem(&m_sh2_state->vbr), R32(Rn)); // mov vbr, Rn + return true; + + case 0x0c: + case 0x0d: + case 0x14: + case 0x1c: + case 0x1d: + case 0x2c: + case 0x2d: + case 0x30: + case 0x31: + case 0x32: + case 0x33: + case 0x34: + case 0x35: + case 0x36: + case 0x37: + case 0x38: + case 0x39: + case 0x3a: + case 0x3b: + case 0x3c: + case 0x3d: + case 0x3e: + return false; + } + + return false; +} + +bool sh2_device::generate_group_6(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + switch (opcode & 15) + { + case 0: // MOVBL(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + SETEA(0); // debug: ea = r0 + UML_CALLH(block, *m_read8); // call read8 + UML_SEXT(block, R32(Rn), I0, SIZE_BYTE); // sext Rn, r0, BYTE + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 1: // MOVWL(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + SETEA(0); // debug: ea = r0 + UML_CALLH(block, *m_read16); // call read16 + UML_SEXT(block, R32(Rn), I0, SIZE_WORD); // sext Rn, r0, WORD + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 2: // MOVLL(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + SETEA(0); // debug: ea = r0 + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, R32(Rn), I0); // mov Rn, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 3: // MOV(Rm, Rn); + UML_MOV(block, R32(Rn), R32(Rm)); // mov Rn, Rm + return true; + + case 7: // NOT(Rm, Rn); + UML_XOR(block, R32(Rn), R32(Rm), 0xffffffff); // xor Rn, Rm, 0xffffffff + return true; + + case 9: // SWAPW(Rm, Rn); + UML_ROL(block, R32(Rn), R32(Rm), 16); // rol Rn, Rm, 16 + return true; + + case 11: // NEG(Rm, Rn); + UML_SUB(block, R32(Rn), 0, R32(Rm)); // sub Rn, 0, Rm + return true; + + case 12: // EXTUB(Rm, Rn); + UML_AND(block, R32(Rn), R32(Rm), 0x000000ff); // and Rn, Rm, 0xff + return true; + + case 13: // EXTUW(Rm, Rn); + UML_AND(block, R32(Rn), R32(Rm), 0x0000ffff); // and Rn, Rm, 0xffff + return true; + + case 14: // EXTSB(Rm, Rn); + UML_SEXT(block, R32(Rn), R32(Rm), SIZE_BYTE); // sext Rn, Rm, BYTE + return true; + + case 15: // EXTSW(Rm, Rn); + UML_SEXT(block, R32(Rn), R32(Rm), SIZE_WORD); // sext Rn, Rm, WORD + return true; + + case 4: // MOVBP(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + UML_CALLH(block, *m_read8); // call read8 + UML_SEXT(block, R32(Rn), I0, SIZE_BYTE); // sext Rn, r0, BYTE + + if (Rm != Rn) + UML_ADD(block, R32(Rm), R32(Rm), 1); // add Rm, Rm, #1 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 5: // MOVWP(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + UML_CALLH(block, *m_read16); // call read16 + UML_SEXT(block, R32(Rn), I0, SIZE_WORD); // sext Rn, r0, WORD + + if (Rm != Rn) + UML_ADD(block, R32(Rm), R32(Rm), 2); // add Rm, Rm, #2 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 6: // MOVLP(Rm, Rn); + UML_MOV(block, I0, R32(Rm)); // mov r0, Rm + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, R32(Rn), I0); // mov Rn, r0 + + if (Rm != Rn) + UML_ADD(block, R32(Rm), R32(Rm), 4); // add Rm, Rm, #4 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 8: // SWAPB(Rm, Rn); + UML_AND(block, I0, R32(Rm), 0xffff0000); // and r0, Rm, #0xffff0000 + UML_AND(block, I1, R32(Rm), 0x000000ff); // and r0, Rm, #0x000000ff + UML_AND(block, I2, R32(Rm), 0x0000ff00); // and r0, Rm, #0x0000ff00 + UML_SHL(block, I1, I1, 8); // shl r1, r1, #8 + UML_SHR(block, I2, I2, 8); // shr r2, r2, #8 + UML_OR(block, I0, I0, I1); // or r0, r0, r1 + UML_OR(block, R32(Rn), I0, I2); // or Rn, r0, r2 + return true; + + case 10: // NEGC(Rm, Rn); + UML_MOV(block, I0, mem(&m_sh2_state->sr)); // mov r0, sr (save SR) + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T (clear the T bit) + UML_CARRY(block, I0, 0); // carry = T (T is bit 0 of SR) + UML_SUBB(block, R32(Rn), 0, R32(Rm)); // subb Rn, #0, Rm + + UML_JMPc(block, COND_NC, compiler->labelnum); // jnc labelnum + + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + + UML_LABEL(block, compiler->labelnum++); // labelnum: + + return true; + } + + return false; +} + +bool sh2_device::generate_group_8(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + int32_t disp; + uint32_t udisp; + code_label templabel; + + switch ( opcode & (15<<8) ) + { + case 0 << 8: // MOVBS4(opcode & 0x0f, Rm); + udisp = (opcode & 0x0f); + UML_ADD(block, I0, R32(Rm), udisp); // add r0, Rm, udisp + UML_MOV(block, I1, R32(0)); // mov r1, R0 + UML_CALLH(block, *m_write8); // call write8 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 1 << 8: // MOVWS4(opcode & 0x0f, Rm); + udisp = (opcode & 0x0f) * 2; + UML_ADD(block, I0, R32(Rm), udisp); // add r0, Rm, udisp + UML_MOV(block, I1, R32(0)); // mov r1, R0 + UML_CALLH(block, *m_write16); // call write16 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 2<< 8: + case 3<< 8: + case 6<< 8: + case 7<< 8: + case 10<< 8: + case 12<< 8: + case 14<< 8: + return false; + + case 4<< 8: // MOVBL4(Rm, opcode & 0x0f); + udisp = opcode & 0x0f; + UML_ADD(block, I0, R32(Rm), udisp); // add r0, Rm, udisp + SETEA(0); + UML_CALLH(block, *m_read8); // call read8 + UML_SEXT(block, R32(0), I0, SIZE_BYTE); // sext R0, r0, BYTE + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 5<< 8: // MOVWL4(Rm, opcode & 0x0f); + udisp = (opcode & 0x0f)*2; + UML_ADD(block, I0, R32(Rm), udisp); // add r0, Rm, udisp + SETEA(0); + UML_CALLH(block, *m_read16); // call read16 + UML_SEXT(block, R32(0), I0, SIZE_WORD); // sext R0, r0, WORD + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 8<< 8: // CMPIM(opcode & 0xff); + UML_AND(block, I0, mem(&m_sh2_state->sr), ~T); // and r0, sr, ~T (clear the T bit) + + UML_SEXT(block, I1, opcode&0xff, SIZE_BYTE); // sext r1, opcode&0xff, BYTE + UML_CMP(block, I1, R32(0)); // cmp r1, R0 + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz compiler->labelnum (if negative) + + UML_OR(block, I0, I0, T); // or r0, r0, T + + UML_LABEL(block, compiler->labelnum++); // labelnum: + UML_MOV(block, mem(&m_sh2_state->sr), I0); // mov m_sh2_state->sr, r0 + return true; + + case 9<< 8: // BT(opcode & 0xff); + UML_TEST(block, mem(&m_sh2_state->sr), T); // test m_sh2_state->sr, T + UML_JMPc(block, COND_Z, compiler->labelnum); // jz compiler->labelnum + + disp = ((int32_t)opcode << 24) >> 24; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = destination + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // jmp m_sh2_state->ea + + UML_LABEL(block, compiler->labelnum++); // labelnum: + return true; + + case 11<< 8: // BF(opcode & 0xff); + UML_TEST(block, mem(&m_sh2_state->sr), T); // test m_sh2_state->sr, T + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz compiler->labelnum + + disp = ((int32_t)opcode << 24) >> 24; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = destination + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // jmp m_sh2_state->ea + + UML_LABEL(block, compiler->labelnum++); // labelnum: + return true; + + case 13<< 8: // BTS(opcode & 0xff); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_TEST(block, mem(&m_sh2_state->sr), T); // test m_sh2_state->sr, T + UML_JMPc(block, COND_Z, compiler->labelnum); // jz compiler->labelnum + + disp = ((int32_t)opcode << 24) >> 24; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = destination + + templabel = compiler->labelnum; // save our label + compiler->labelnum++; // make sure the delay slot doesn't use it + generate_delay_slot(block, compiler, desc, m_sh2_state->ea-2); + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // jmp m_sh2_state->ea + + UML_LABEL(block, templabel); // labelnum: + return true; + } + break; + + case 15<< 8: // BFS(opcode & 0xff); + if (m_cpu_type > CPU_TYPE_SH1) + { + UML_TEST(block, mem(&m_sh2_state->sr), T); // test m_sh2_state->sr, T + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz compiler->labelnum + + disp = ((int32_t)opcode << 24) >> 24; + m_sh2_state->ea = (desc->pc + 2) + disp * 2 + 2; // m_sh2_state->ea = destination + + templabel = compiler->labelnum; // save our label + compiler->labelnum++; // make sure the delay slot doesn't use it + generate_delay_slot(block, compiler, desc, m_sh2_state->ea-2); // delay slot only if the branch is taken + + generate_update_cycles(block, compiler, m_sh2_state->ea, true); // <subtract cycles> + UML_HASHJMP(block, 0, m_sh2_state->ea, *m_nocode); // jmp m_sh2_state->ea + + UML_LABEL(block, templabel); // labelnum: + return true; + } + break; + } + + return false; +} + +bool sh2_device::generate_group_12(drcuml_block *block, compiler_state *compiler, const opcode_desc *desc, uint16_t opcode, int in_delay_slot, uint32_t ovrpc) +{ + uint32_t scratch; + + switch (opcode & (15<<8)) + { + case 0<<8: // MOVBSG(opcode & 0xff); + scratch = (opcode & 0xff); + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_AND(block, I1, R32(0), 0xff); // and r1, R0, 0xff + UML_CALLH(block, *m_write8); // call write8 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 1<<8: // MOVWSG(opcode & 0xff); + scratch = (opcode & 0xff) * 2; + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_AND(block, I1, R32(0), 0xffff); // and r1, R0, 0xffff + UML_CALLH(block, *m_write16); // call write16 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 2<<8: // MOVLSG(opcode & 0xff); + scratch = (opcode & 0xff) * 4; + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_MOV(block, I1, R32(0)); // mov r1, R0 + UML_CALLH(block, *m_write32); // call write32 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 3<<8: // TRAPA(opcode & 0xff); + scratch = (opcode & 0xff) * 4; + UML_ADD(block, mem(&m_sh2_state->ea), mem(&m_sh2_state->vbr), scratch); // add ea, vbr, scratch + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, mem(&m_sh2_state->sr)); // mov r1, sr + UML_CALLH(block, *m_write32); // write32 + + UML_SUB(block, R32(15), R32(15), 4); // sub R15, R15, #4 + UML_MOV(block, I0, R32(15)); // mov r0, R15 + UML_MOV(block, I1, desc->pc+2); // mov r1, pc+2 + UML_CALLH(block, *m_write32); // write32 + + UML_MOV(block, I0, mem(&m_sh2_state->ea)); // mov r0, ea + UML_CALLH(block, *m_read32); // read32 + UML_HASHJMP(block, 0, I0, *m_nocode); // jmp (r0) + + return true; + + case 4<<8: // MOVBLG(opcode & 0xff); + scratch = (opcode & 0xff); + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_CALLH(block, *m_read8); // call read16 + UML_SEXT(block, R32(0), I0, SIZE_BYTE); // sext R0, r0, BYTE + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 5<<8: // MOVWLG(opcode & 0xff); + scratch = (opcode & 0xff) * 2; + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_CALLH(block, *m_read16); // call read16 + UML_SEXT(block, R32(0), I0, SIZE_WORD); // sext R0, r0, WORD + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 6<<8: // MOVLLG(opcode & 0xff); + scratch = (opcode & 0xff) * 4; + UML_ADD(block, I0, mem(&m_sh2_state->gbr), scratch); // add r0, gbr, scratch + UML_CALLH(block, *m_read32); // call read32 + UML_MOV(block, R32(0), I0); // mov R0, r0 + + if (!in_delay_slot) + generate_update_cycles(block, compiler, desc->pc + 2, true); + return true; + + case 7<<8: // MOVA(opcode & 0xff); + scratch = (opcode & 0xff) * 4; + scratch += ((desc->pc + 4) & ~3); + + UML_MOV(block, R32(0), scratch); // mov R0, scratch + return true; + + case 8<<8: // TSTI(opcode & 0xff); + scratch = opcode & 0xff; + + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T (clear the T bit) + UML_AND(block, I0, R32(0), scratch); // and r0, R0, scratch + UML_CMP(block, I0, 0); // cmp r0, #0 + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz labelnum + + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + + UML_LABEL(block, compiler->labelnum++); // labelnum: + return true; + + case 9<<8: // ANDI(opcode & 0xff); + UML_AND(block, R32(0), R32(0), opcode & 0xff); // and r0, r0, opcode & 0xff + return true; + + case 10<<8: // XORI(opcode & 0xff); + UML_XOR(block, R32(0), R32(0), opcode & 0xff); // xor r0, r0, opcode & 0xff + return true; + + case 11<<8: // ORI(opcode & 0xff); + UML_OR(block, R32(0), R32(0), opcode & 0xff); // or r0, r0, opcode & 0xff + return true; + + case 12<<8: // TSTM(opcode & 0xff); + UML_AND(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), ~T); // and sr, sr, ~T (clear the T bit) + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + UML_CALLH(block, *m_read8); // read8 + + UML_AND(block, I0, I0, opcode & 0xff); + UML_CMP(block, I0, 0); // cmp r0, #0 + UML_JMPc(block, COND_NZ, compiler->labelnum); // jnz labelnum + + UML_OR(block, mem(&m_sh2_state->sr), mem(&m_sh2_state->sr), T); // or sr, sr, T + + UML_LABEL(block, compiler->labelnum++); // labelnum: + return true; + + case 13<<8: // ANDM(opcode & 0xff); + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + UML_CALLH(block, *m_read8); // read8 + + UML_AND(block, I1, I0, opcode&0xff); // and r1, r0, #opcode&0xff + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + SETEA(0); + UML_CALLH(block, *m_write8); // write8 + return true; + + case 14<<8: // XORM(opcode & 0xff); + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + UML_CALLH(block, *m_read8); // read8 + + UML_XOR(block, I1, I0, opcode&0xff); // xor r1, r0, #opcode&0xff + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + SETEA(0); + UML_CALLH(block, *m_write8); // write8 + return true; + + case 15<<8: // ORM(opcode & 0xff); + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + UML_CALLH(block, *m_read8); // read8 + + UML_OR(block, I1, I0, opcode&0xff); // or r1, r0, #opcode&0xff + UML_ADD(block, I0, R32(0), mem(&m_sh2_state->gbr)); // add r0, R0, gbr + SETEA(0); + UML_CALLH(block, *m_write8); // write8 + return true; + } + + return false; +} + +/*************************************************************************** + CORE CALLBACKS +***************************************************************************/ + +/*------------------------------------------------- + sh2drc_set_options - configure DRC options +-------------------------------------------------*/ + +void sh2_device::sh2drc_set_options(uint32_t options) +{ + if (!allow_drc()) return; + m_drcoptions = options; +} + + +/*------------------------------------------------- + sh2drc_add_pcflush - add a new address where + the PC must be flushed for speedups to work +-------------------------------------------------*/ + +void sh2_device::sh2drc_add_pcflush(offs_t address) +{ + if (!allow_drc()) return; + + if (m_pcfsel < ARRAY_LENGTH(m_pcflushes)) + m_pcflushes[m_pcfsel++] = address; +} + + +/*------------------------------------------------- + sh2drc_add_fastram - add a new fastram + region +-------------------------------------------------*/ + +void sh2_device::sh2drc_add_fastram(offs_t start, offs_t end, uint8_t readonly, void *base) +{ + if (m_fastram_select < ARRAY_LENGTH(m_fastram)) + { + m_fastram[m_fastram_select].start = start; + m_fastram[m_fastram_select].end = end; + m_fastram[m_fastram_select].readonly = readonly; + m_fastram[m_fastram_select].base = base; + m_fastram_select++; + } +} |