| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
| |
* mb86235: One more use of util::sext
|
| |
|
|
|
|
| |
- Add mul_16x16 inline function to perform a signed 16x16-bit multiplication with 32-bit result. This was moved from cpu/e132xs to unite it with the analogous 32x32 operations.
- Add rotl_32, rotr_32, rotl_64 and rotr_64 inline functions to perform 32-bit and 64-bit circular shifts in either direction by the specified number of places, modulo 32 or 64. It is anticipated that these will eventually be replaced by standard functions in C++20's <bit> header, and so they have been given similar signatures and semantics (which are also validity-checked).
- Remove LSL, LSR, ROL and ROR macros from cpu/arm and cpu/arm7 to ameliorate unnecessary obfuscation.
|
| | |
|
| |
|
|
| |
tests in FuzzARM now pass. [Ryan Holtz]
|
| | |
|
| |\
| |
| |
| |
| |
| |
| | |
MAME 0.193
Conflicts:
src/devices/cpu/arm7/arm7thmb.cpp
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
(except for POP {Rlist} which is correct),
handle v5 behavior (lsb address bit is ignored in 16bit loads) in arm946es_cpu_device::arm7_cpu_read16 instead,
manually handle v4 and v5 difference for LDRSH signed halfword load.
note: it looks like all? v5 cores ignore least significant address bit during 16bit loads, so we should add something like
virtual uint32_t arm7_cpu_read16(uint32_t addr) override { return base_class::arm7_cpu_read16(addr & ~1)};
into theirs class declaration.
|
| | | |
|
| |/ |
|
| |
|
|
|
|
| |
ARMv5/ARM9 behavior."
This reverts commit 1bc2a2ab12757e21a90ef191bf9fe69a541132e8.
|
| |
|
|
|
|
| |
ARMv5/ARM9 behavior.
this needed to be changed or split when we will be emulate ARMv6/ARM11
|
| |
|
|
| |
This reverts commit 54155441e9ba9941e85d80c4834a66376a11e791.
|
| |
|
|
|
| |
This reverts commit f537428e5a40ba6dde8ca9bf0fe9ae6b1f189ac4, reversing
changes made to 0d70d798107d4e4e8fb9f230410aeb1e888d65c5.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
and Peter Wilhelmsen]
started adding some devices to the driver to support it [David Haywood]
derived IGS036 type in ARM7/9 core, because it seems like mmu could be different, and there are probably other internal devices [David Haywood]
fixed(?) BLX (LO) thumb opcode in ARM7/9 core, was storing a return address that skipped the following opcode [David Haywood]
I think there might still be ARM9 issues.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
Use standard uint64_t, uint32_t, uint16_t or uint8_t instead of UINT64, UINT32, UINT16 or UINT8
also use standard int64_t, int32_t, int16_t or int8_t instead of INT64, INT32, INT16 or INT8
|
| | |
|
| |
|