mirrors/dolphin
0
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-09-28 20:28:46 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into jit_exit_addresses

Browse source
This commit is contained in:
Ryan Houdek 2013-11-21 21:17:58 -06:00
commit ea9ac07ec9
143 changed files with 1773 additions and 1093 deletions
Download patch file Download diff file Expand all files Collapse all files

4
Source/Core/AudioCommon/Src/AudioCommon.cpp
View file

@ -23,7 +23,7 @@ SoundStream *soundStream = nullptr;
namespace AudioCommon
{
SoundStream *InitSoundStream(CMixer *mixer)
SoundStream *InitSoundStream(CMixer *mixer, void *hWnd)
{
// TODO: possible memleak with mixer
@ -33,7 +33,7 @@ namespace AudioCommon
else if (backend == BACKEND_NULLSOUND && NullSound::isValid())
soundStream = new NullSound(mixer);
else if (backend == BACKEND_DIRECTSOUND && DSound::isValid())
soundStream = new DSound(mixer);
soundStream = new DSound(mixer, hWnd);
else if (backend == BACKEND_XAUDIO2)
{
if (XAudio2::isValid())

2
Source/Core/AudioCommon/Src/AudioCommon.h
View file

@ -40,7 +40,7 @@ union UDSPControl
namespace AudioCommon
{
SoundStream *InitSoundStream(CMixer *mixer);
SoundStream *InitSoundStream(CMixer *mixer, void *hWnd);
void ShutdownSoundStream();
std::vector<std::string> GetSoundBackends();
bool UseJIT();

4
Source/Core/AudioCommon/Src/DSoundStream.h
View file

@ -48,7 +48,7 @@ class DSound : public SoundStream
bool WriteDataToBuffer(DWORD dwOffset, char* soundData, DWORD dwSoundBytes);
public:
DSound(CMixer *mixer, void *_hWnd = NULL)
DSound(CMixer *mixer, void *_hWnd)
: SoundStream(mixer)
, bufferSize(0)
, currentPos(0)
@ -71,7 +71,7 @@ public:
#else
public:
DSound(CMixer *mixer)
DSound(CMixer *mixer, void *_hWnd)
: SoundStream(mixer)
{}
#endif

515
Source/Core/Common/Src/ArmEmitter.cpp
View file

@ -938,13 +938,13 @@ u32 EncodeVm(ARMReg Vm)
// Double/single, Neon
extern const VFPEnc VFPOps[16][2] = {
{{0xE0, 0xA0}, {0x20, 0xD1}}, // 0: VMLA
{{0xE0, 0xA0}, { -1, -1}}, // 0: VMLA
{{0xE1, 0xA4}, { -1, -1}}, // 1: VNMLA
{{0xE0, 0xA4}, {0x22, 0xD1}}, // 2: VMLS
{{0xE0, 0xA4}, { -1, -1}}, // 2: VMLS
{{0xE1, 0xA0}, { -1, -1}}, // 3: VNMLS
{{0xE3, 0xA0}, {0x20, 0xD0}}, // 4: VADD
{{0xE3, 0xA4}, {0x22, 0xD0}}, // 5: VSUB
{{0xE2, 0xA0}, {0x30, 0xD1}}, // 6: VMUL
{{0xE3, 0xA0}, { -1, -1}}, // 4: VADD
{{0xE3, 0xA4}, { -1, -1}}, // 5: VSUB
{{0xE2, 0xA0}, { -1, -1}}, // 6: VMUL
{{0xE2, 0xA4}, { -1, -1}}, // 7: VNMUL
{{0xEB, 0xAC}, { -1 /* 0x3B */, -1 /* 0x70 */}}, // 8: VABS(Vn(0x0) used for encoding)
{{0xE8, 0xA0}, { -1, -1}}, // 9: VDIV
@ -1237,7 +1237,7 @@ void ARMXEmitter::VCVT(ARMReg Dest, ARMReg Source, int flags)
}
}
void NEONXEmitter::VABA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VABA(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1248,7 +1248,7 @@ void NEONXEmitter::VABA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (encodedSize(Size) << 20) | EncodeVd(Vd) | (0x71 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VABAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VABAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= D0 && Vn < Q0, "Pass invalid register to " __FUNCTION__);
@ -1260,7 +1260,7 @@ void NEONXEmitter::VABAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (encodedSize(Size) << 20) | EncodeVd(Vd) | (0x50 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VABD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VABD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1273,7 +1273,7 @@ void NEONXEmitter::VABD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (encodedSize(Size) << 20) | EncodeVd(Vd) | (0x70 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VABDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VABDL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= D0 && Vn < Q0, "Pass invalid register to " __FUNCTION__);
@ -1285,7 +1285,7 @@ void NEONXEmitter::VABDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (encodedSize(Size) << 20) | EncodeVd(Vd) | (0x70 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VABS(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VABS(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1327,7 +1327,7 @@ void NEONXEmitter::VACLT(ARMReg Vd, ARMReg Vn, ARMReg Vm)
VACGT(Vd, Vn, Vm);
}
void NEONXEmitter::VADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1341,7 +1341,7 @@ void NEONXEmitter::VADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (0x8 << 8) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VADDHN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VADDHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd < Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= Q0, "Pass invalid register to " __FUNCTION__);
@ -1353,7 +1353,7 @@ void NEONXEmitter::VADDHN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| EncodeVd(Vd) | (0x80 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VADDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VADDL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= D0 && Vn < Q0, "Pass invalid register to " __FUNCTION__);
@ -1364,7 +1364,7 @@ void NEONXEmitter::VADDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 24) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) \
| EncodeVd(Vd) | EncodeVm(Vm));
}
void NEONXEmitter::VADDW(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VADDW(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= Q0, "Pass invalid register to " __FUNCTION__);
@ -1420,7 +1420,7 @@ void NEONXEmitter::VBSL(ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF3 << 24) | (1 << 20) | EncodeVn(Vn) | EncodeVd(Vd) | (0x11 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VCEQ(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1433,7 +1433,7 @@ void NEONXEmitter::VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
| (0x81 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCEQ(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1443,7 +1443,7 @@ void NEONXEmitter::VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF2 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 16) \
| EncodeVd(Vd) | ((Size & F_32 ? 1 : 0) << 10) | (0x10 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VCGE(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1455,7 +1455,7 @@ void NEONXEmitter::VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) \
| (0x31 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCGE(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1464,7 +1464,7 @@ void NEONXEmitter::VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 16) \
| EncodeVd(Vd) | ((Size & F_32 ? 1 : 0) << 10) | (0x8 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VCGT(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1476,7 +1476,7 @@ void NEONXEmitter::VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) \
| (0x30 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCGT(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1485,11 +1485,11 @@ void NEONXEmitter::VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) | (1 << 16) \
| EncodeVd(Vd) | ((Size & F_32 ? 1 : 0) << 10) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCLE(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VCLE(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
VCGE(Size, Vd, Vm, Vn);
}
void NEONXEmitter::VCLE(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCLE(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1498,7 +1498,7 @@ void NEONXEmitter::VCLE(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) | (1 << 16) \
| EncodeVd(Vd) | ((Size & F_32 ? 1 : 0) << 10) | (3 << 7) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCLS(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCLS(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1508,11 +1508,11 @@ void NEONXEmitter::VCLS(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) \
| EncodeVd(Vd) | (1 << 10) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCLT(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VCLT(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
VCGT(Size, Vd, Vm, Vn);
}
void NEONXEmitter::VCLT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCLT(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1521,7 +1521,7 @@ void NEONXEmitter::VCLT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) | (1 << 16) \
| EncodeVd(Vd) | ((Size & F_32 ? 1 : 0) << 10) | (0x20 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCLZ(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCLZ(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1530,7 +1530,7 @@ void NEONXEmitter::VCLZ(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) \
| EncodeVd(Vd) | (0x48 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VCNT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VCNT(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1540,7 +1540,7 @@ void NEONXEmitter::VCNT(NEONElementType Size, ARMReg Vd, ARMReg Vm)
Write32((0xF3 << 24) | (0xD << 20) | (encodedSize(Size) << 18) \
| EncodeVd(Vd) | (0x90 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VDUP(NEONElementType Size, ARMReg Vd, ARMReg Vm, u8 index)
void NEONXEmitter::VDUP(u32 Size, ARMReg Vd, ARMReg Vm, u8 index)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1562,7 +1562,7 @@ void NEONXEmitter::VDUP(NEONElementType Size, ARMReg Vd, ARMReg Vm, u8 index)
Write32((0xF3 << 24) | (0xD << 20) | (sizeEncoded << 16) | (indexEncoded << 16) \
| EncodeVd(Vd) | (0xC0 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VDUP(NEONElementType Size, ARMReg Vd, ARMReg Rt)
void NEONXEmitter::VDUP(u32 Size, ARMReg Vd, ARMReg Rt)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Rt < D0, "Pass invalid register to " __FUNCTION__);
@ -1616,7 +1616,7 @@ void NEONXEmitter::VFMS(ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | (1 << 21) | EncodeVn(Vn) | EncodeVd(Vd) | (0xC1 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VHADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VHADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1627,7 +1627,7 @@ void NEONXEmitter::VHADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 23) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VHSUB(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VHSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1638,7 +1638,7 @@ void NEONXEmitter::VHSUB(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 23) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (1 << 9) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMAX(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMAX(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1651,7 +1651,7 @@ void NEONXEmitter::VMAX(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 23) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (0x60 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMIN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMIN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1664,7 +1664,7 @@ void NEONXEmitter::VMIN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 23) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (0x61 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMLA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMLA(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1676,7 +1676,7 @@ void NEONXEmitter::VMLA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
else
Write32((0xF2 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | (0x90 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMLS(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMLS(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
@ -1688,7 +1688,7 @@ void NEONXEmitter::VMLS(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
else
Write32((0xF2 << 24) | (1 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | (0x90 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMLAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMLAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= Q0, "Pass invalid register to " __FUNCTION__);
@ -1699,7 +1699,7 @@ void NEONXEmitter::VMLAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 24) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (0x80 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VMLSL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMLSL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, Vn >= Q0, "Pass invalid register to " __FUNCTION__);
@ -1710,23 +1710,404 @@ void NEONXEmitter::VMLSL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
Write32((0xF2 << 24) | ((Size & I_UNSIGNED ? 1 : 0) << 24) | (encodedSize(Size) << 20) \
| EncodeVn(Vn) | EncodeVd(Vd) | (0xA0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VSUB(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
void NEONXEmitter::VMUL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to VSUB(integer)");
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use VSUB(integer) when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
// Gets encoded as a double register
bool register_quad = Vd >= Q0;
if (Size & F_32)
Write32((0xF3 << 24) | EncodeVn(Vn) | EncodeVd(Vd) | (0xD1 << 4) | (register_quad << 6) | EncodeVm(Vm));
else
Write32((0xF2 << 24) | ((Size & I_POLYNOMIAL) ? (1 << 24) : 0) | (encodedSize(Size) << 20) | \
EncodeVn(Vn) | EncodeVd(Vd) | (0x91 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VMULL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF2 << 24) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xC0 << 4) | ((Size & I_POLYNOMIAL) ? 1 << 9 : 0) | EncodeVm(Vm));
}
void NEONXEmitter::VNEG(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 16) | \
EncodeVd(Vd) | ((Size & F_32) ? 1 << 10 : 0) | (0xE << 6) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VORN(ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (3 << 20) | EncodeVn(Vn) | EncodeVd(Vd) | (0x11 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VORR(ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (2 << 20) | EncodeVn(Vn) | EncodeVd(Vd) | (0x11 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VPADAL(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | EncodeVd(Vd) | \
(0x60 << 4) | ((Size & I_UNSIGNED) ? 1 << 7 : 0) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VPADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
if (Size & F_32)
Write32((0xF3 << 24) | EncodeVn(Vn) | EncodeVd(Vd) | (0xD0 << 4) | EncodeVm(Vm));
else
Write32((0xF2 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xB1 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VPADDL(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | EncodeVd(Vd) | \
(0x20 << 4) | (Size & I_UNSIGNED ? 1 << 7 : 0) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VPMAX(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
if (Size & F_32)
Write32((0xF3 << 24) | EncodeVn(Vn) | EncodeVd(Vd) | (0xF0 << 4) | EncodeVm(Vm));
else
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xA0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VPMIN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
if (Size & F_32)
Write32((0xF3 << 24) | (1 << 21) | EncodeVn(Vn) | EncodeVd(Vd) | (0xF0 << 4) | EncodeVm(Vm));
else
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xA1 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQABS(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | EncodeVd(Vd) | \
(0x70 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VQADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x1 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VQDMLAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF2 << 24) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x90 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQDMLSL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF2 << 24) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xB0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQDMULH(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF2 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xB0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQDMULL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF2 << 24) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xD0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQNEG(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | EncodeVd(Vd) | \
(0x78 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VQRDMULH(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF3 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xB0 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VQRSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x51 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VQSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x41 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VQSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x21 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRADDHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF3 << 24) | (1 << 23) | ((encodedSize(Size) - 1) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x40 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VRECPE(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (0xB << 16) | EncodeVd(Vd) | \
(0x40 << 4) | (Size & F_32 ? 1 << 8 : 0) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRECPS(ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | EncodeVn(Vn) | EncodeVd(Vd) | (0xF1 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRHADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x10 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x50 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRSQRTE(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Vd = SubBase(Vd);
Vn = SubBase(Vn);
Vm = SubBase(Vm);
Write32((0xF3 << 24) | ((Vd & 0x10) << 18) | (encodedSize(Size) << 20) | ((Vn & 0xF) << 16) \
| ((Vd & 0xF) << 12) | (0x8 << 8) | ((Vn & 0x10) << 3) | (1 << 6) \
| ((Vm & 0x10) << 1) | (Vm & 0xF));
Write32((0xF3 << 24) | (0xB << 20) | ((Vd & 0x10) << 18) | (0xB << 16)
| ((Vd & 0xF) << 12) | (9 << 7) | (Size & F_32 ? (1 << 8) : 0) | (register_quad << 6)
| ((Vm & 0x10) << 1) | (Vm & 0xF));
}
void NEONXEmitter::VRSQRTS(ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
void NEONXEmitter::VLD1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (1 << 21) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xF1 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VRSUBHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
Write32((0xF3 << 24) | (1 << 23) | ((encodedSize(Size) - 1) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x60 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= D0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
_dbg_assert_msg_(DYNA_REC, !(Size & F_32), __FUNCTION__ " doesn't support float");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x40 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
if (Size & F_32)
Write32((0xF2 << 24) | (1 << 21) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0xD0 << 4) | (register_quad << 6) | EncodeVm(Vm));
else
Write32((0xF3 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x80 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VSUBHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
Write32((0xF2 << 24) | (1 << 23) | ((encodedSize(Size) - 1) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x60 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VSUBL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x20 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VSUBW(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
Write32((0xF2 << 24) | (Size & I_UNSIGNED ? 1 << 24 : 0) | (1 << 23) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x30 << 4) | EncodeVm(Vm));
}
void NEONXEmitter::VSWP(ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (1 << 17) | EncodeVd(Vd) | \
(register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VTRN(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 17) | EncodeVd(Vd) | \
(1 << 7) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VTST(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF2 << 24) | (encodedSize(Size) << 20) | EncodeVn(Vn) | EncodeVd(Vd) | \
(0x81 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VUZP(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 17) | EncodeVd(Vd) | \
(0x10 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VZIP(u32 Size, ARMReg Vd, ARMReg Vm)
{
_dbg_assert_msg_(DYNA_REC, Vd >= Q0, "Pass invalid register to " __FUNCTION__);
_dbg_assert_msg_(DYNA_REC, cpu_info.bNEON, "Can't use " __FUNCTION__ " when CPU doesn't support it");
bool register_quad = Vd >= Q0;
Write32((0xF3 << 24) | (0xB << 20) | (encodedSize(Size) << 18) | (1 << 17) | EncodeVd(Vd) | \
(0x18 << 4) | (register_quad << 6) | EncodeVm(Vm));
}
void NEONXEmitter::VLD1(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
{
u32 spacing = 0x7; // Only support loading to 1 reg
// Gets encoded as a double register
@ -1736,8 +2117,7 @@ void NEONXEmitter::VLD1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignmen
| ((Vd & 0xF) << 12) | (spacing << 8) | (encodedSize(Size) << 6)
| (align << 4) | Rm);
}
void NEONXEmitter::VLD2(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
void NEONXEmitter::VLD2(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
{
u32 spacing = 0x8; // Single spaced registers
// Gets encoded as a double register
@ -1747,8 +2127,7 @@ void NEONXEmitter::VLD2(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignmen
| ((Vd & 0xF) << 12) | (spacing << 8) | (encodedSize(Size) << 6)
| (align << 4) | Rm);
}
void NEONXEmitter::VST1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
void NEONXEmitter::VST1(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align, ARMReg Rm)
{
u32 spacing = 0x7; // Single spaced registers
// Gets encoded as a double register
@ -1759,8 +2138,7 @@ void NEONXEmitter::VST1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignmen
| (align << 4) | Rm);
}
void NEONXEmitter::VREVX(u32 size, NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VREVX(u32 size, u32 Size, ARMReg Vd, ARMReg Vm)
{
bool register_quad = Vd >= Q0;
Vd = SubBase(Vd);
@ -1771,44 +2149,19 @@ void NEONXEmitter::VREVX(u32 size, NEONElementType Size, ARMReg Vd, ARMReg Vm)
| (register_quad << 6) | ((Vm & 0x10) << 1) | (Vm & 0xF));
}
void NEONXEmitter::VREV64(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VREV64(u32 Size, ARMReg Vd, ARMReg Vm)
{
VREVX(0, Size, Vd, Vm);
}
void NEONXEmitter::VREV32(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VREV32(u32 Size, ARMReg Vd, ARMReg Vm)
{
VREVX(1, Size, Vd, Vm);
}
void NEONXEmitter::VREV16(NEONElementType Size, ARMReg Vd, ARMReg Vm)
void NEONXEmitter::VREV16(u32 Size, ARMReg Vd, ARMReg Vm)
{
VREVX(2, Size, Vd, Vm);
}
void NEONXEmitter::VRSQRTE(NEONElementType Size, ARMReg Vd, ARMReg Vm)
{
bool register_quad = Vd >= Q0;
Vd = SubBase(Vd);
Vm = SubBase(Vm);
Write32((0xF3 << 24) | (0xB << 20) | ((Vd & 0x10) << 18) | (0xB << 16)
| ((Vd & 0xF) << 12) | (9 << 7) | (Size & F_32 ? (1 << 8) : 0) | (register_quad << 6)
| ((Vm & 0x10) << 1) | (Vm & 0xF));
}
void NEONXEmitter::VORR(ARMReg Vd, ARMReg Vn, ARMReg Vm)
{
bool register_quad = Vd >= Q0;
Vd = SubBase(Vd);
Vn = SubBase(Vn);
Vm = SubBase(Vm);
Write32((0xF2 << 24) | (0x1 << 21) | ((Vd & 0x10) << 18) | ((Vn & 0xF) << 16)
| ((Vd & 0xF) << 12) | (1 << 8) | ((Vn & 0x10) << 3)
| (register_quad << 6) | ((Vm & 0x10) << 1) | (1 << 4) | (Vm & 0xF));
}
}

142
Source/Core/Common/Src/ArmEmitter.h
View file

@ -338,6 +338,15 @@ struct LiteralPool
};
typedef const u8* JumpTarget;
// XXX: Stop polluting the global namespace
const u32 I_8 = (1 << 0);
const u32 I_16 = (1 << 1);
const u32 I_32 = (1 << 2);
const u32 I_64 = (1 << 3);
const u32 I_SIGNED = (1 << 4);
const u32 I_UNSIGNED = (1 << 5);
const u32 F_32 = (1 << 6);
const u32 I_POLYNOMIAL = (1 << 7); // Only used in VMUL/VMULL
u32 EncodeVd(ARMReg Vd);
u32 EncodeVn(ARMReg Vn);
@ -572,17 +581,6 @@ public:
}; // class ARMXEmitter
enum NEONElementType
{
I_8 = (1 << 0),
I_16 = (1 << 1),
I_32 = (1 << 2),
I_64 = (1 << 3),
I_SIGNED = (1 << 4),
I_UNSIGNED = (1 << 5),
F_32 = (1 << 6)
};
enum NEONAlignment
{
ALIGN_NONE = 0,
@ -613,71 +611,105 @@ private:
return 0;
}
void VREVX(u32 size, NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VREVX(u32 size, u32 Size, ARMReg Vd, ARMReg Vm);
public:
public:
NEONXEmitter(ARMXEmitter *emit)
: _emit(emit)
{}
void VABA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABS(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VABA(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABDL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VABS(u32 Size, ARMReg Vd, ARMReg Vm);
void VACGE(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VACGT(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VACLE(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VACLT(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDHN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDW(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VADDW(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VAND(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VBIC(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VBIF(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VBIT(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VBSL(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCEQ(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCGE(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCGT(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCLE(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCLE(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCLS(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCLT(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCLT(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCLZ(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VCNT(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VDUP(NEONElementType Size, ARMReg Vd, ARMReg Vm, u8 index);
void VDUP(NEONElementType Size, ARMReg Vd, ARMReg Rt);
void VCEQ(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCEQ(u32 Size, ARMReg Vd, ARMReg Vm);
void VCGE(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCGE(u32 Size, ARMReg Vd, ARMReg Vm);
void VCGT(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCGT(u32 Size, ARMReg Vd, ARMReg Vm);
void VCLE(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCLE(u32 Size, ARMReg Vd, ARMReg Vm);
void VCLS(u32 Size, ARMReg Vd, ARMReg Vm);
void VCLT(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VCLT(u32 Size, ARMReg Vd, ARMReg Vm);
void VCLZ(u32 Size, ARMReg Vd, ARMReg Vm);
void VCNT(u32 Size, ARMReg Vd, ARMReg Vm);
void VDUP(u32 Size, ARMReg Vd, ARMReg Vm, u8 index);
void VDUP(u32 Size, ARMReg Vd, ARMReg Rt);
void VEOR(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VEXT(ARMReg Vd, ARMReg Vn, ARMReg Vm, u8 index);
void VFMA(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VFMS(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VHADD(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VHSUB(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMAX(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMIN(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLA(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLS(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLAL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLSL(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSUB(NEONElementType Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VREV64(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VREV32(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VREV16(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VRSQRTE(NEONElementType Size, ARMReg Vd, ARMReg Vm);
void VHADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VHSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMAX(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMIN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLA(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLS(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMLSL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMUL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VMULL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VNEG(u32 Size, ARMReg Vd, ARMReg Vm);
void VORN(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VORR(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VPADAL(u32 Size, ARMReg Vd, ARMReg Vm);
void VPADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VPADDL(u32 Size, ARMReg Vd, ARMReg Vm);
void VPMAX(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VPMIN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQABS(u32 Size, ARMReg Vd, ARMReg Vm);
void VQADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQDMLAL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQDMLSL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQDMULH(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQDMULL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQNEG(u32 Size, ARMReg Vd, ARMReg Vm);
void VQRDMULH(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQRSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VQSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRADDHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRECPE(u32 Size, ARMReg Vd, ARMReg Vm);
void VRECPS(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRHADD(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRSQRTE(u32 Size, ARMReg Vd, ARMReg Vm);
void VRSQRTS(ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VRSUBHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSHL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSUB(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSUBHN(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSUBL(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSUBW(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VSWP(ARMReg Vd, ARMReg Vm);
void VTRN(u32 Size, ARMReg Vd, ARMReg Vm);
void VTST(u32 Size, ARMReg Vd, ARMReg Vn, ARMReg Vm);
void VUZP(u32 Size, ARMReg Vd, ARMReg Vm);
void VZIP(u32 Size, ARMReg Vd, ARMReg Vm);
void VREV64(u32 Size, ARMReg Vd, ARMReg Vm);
void VREV32(u32 Size, ARMReg Vd, ARMReg Vm);
void VREV16(u32 Size, ARMReg Vd, ARMReg Vm);
void VLD1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
void VLD2(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
void VLD1(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
void VLD2(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
void VST1(NEONElementType Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
void VST1(u32 Size, ARMReg Vd, ARMReg Rn, NEONAlignment align = ALIGN_NONE, ARMReg Rm = _PC);
};
// Everything that needs to generate X86 code should inherit from this.

7
Source/Core/Common/Src/CPUDetect.h
View file

@ -41,7 +41,14 @@ struct CPUInfo
bool bLZCNT;
bool bSSE4A;
bool bAVX;
bool bFMA;
bool bAES;
// FXSAVE/FXRSTOR
bool bFXSR;
// This flag indicates that the hardware supports some mode
// in which denormal inputs _and_ outputs are automatically set to (signed) zero.
// TODO: ARM
bool bFlushToZero;
bool bLAHFSAHF64;
bool bLongMode;

4
Source/Core/Common/Src/ChunkFile.h
View file

@ -193,10 +193,12 @@ public:
void DoPointer(T*& x, T* const base)
{
// pointers can be more than 2^31 apart, but you're using this function wrong if you need that much range
s32 offset = x - base;
ptrdiff_t offset = x - base;
Do(offset);
if (mode == MODE_READ)
{
x = base + offset;
}
}
// Let's pretend std::list doesn't exist!

7
Source/Core/Common/Src/CommonFuncs.h
View file

@ -31,7 +31,12 @@ struct ArraySizeImpl : public std::extent<T>
#define b32(x) (b16(x) | (b16(x) >>16) )
#define ROUND_UP_POW2(x) (b32(x - 1) + 1)
#if defined __GNUC__ && !defined __SSSE3__ && !defined _M_GENERIC
#ifndef __GNUC_PREREQ
#define __GNUC_PREREQ(a, b) 0
#endif
#if (defined __GNUC__ && !__GNUC_PREREQ(4,9)) \
&& !defined __SSSE3__ && !defined _M_GENERIC
#include <emmintrin.h>
static __inline __m128i __attribute__((__always_inline__))
_mm_shuffle_epi8(__m128i a, __m128i mask)

2
Source/Core/Common/Src/FPURoundMode.h
View file

@ -36,7 +36,7 @@ namespace FPURoundMode
void SetPrecisionMode(u32 mode);
void SetSIMDMode(u32 mode);
void SetSIMDMode(u32 roundingMode, u32 nonIEEEMode);
/*
* There are two different flavors of float to int conversion:

2
Source/Core/Common/Src/GenericFPURoundMode.cpp
View file

@ -26,7 +26,7 @@ namespace FPURoundMode
void SetPrecisionMode(u32 mode)
{
}
void SetSIMDMode(u32 mode)
void SetSIMDMode(u32 mode, u32 nonIEEEMode)
{
}
void SaveSIMDState()

4
Source/Core/Common/Src/MathUtil.h
View file

@ -64,10 +64,10 @@ inline float FlushToZero(float f)
return x.f;
}
inline double FlushToZeroAsFloat(double d)
inline double FlushToZero(double d)
{
IntDouble x; x.d = d;
if ((x.i & DOUBLE_EXP) < 0x3800000000000000ULL)
if ((x.i & DOUBLE_EXP) == 0)
x.i &= DOUBLE_SIGN; // turn into signed zero
return x.d;
}

3
Source/Core/Common/Src/SDCardUtil.cpp
View file

@ -36,6 +36,7 @@
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <cinttypes>
#ifndef _WIN32
#include <unistd.h> // for unlink()
@ -196,7 +197,7 @@ bool SDCardCreate(u64 disk_size /*in MB*/, const char* filename)
disk_size *= 1024 * 1024;
if (disk_size < 0x800000 || disk_size > 0x800000000ULL) {
ERROR_LOG(COMMON, "Trying to create SD Card image of size %lliMB is out of range (8MB-32GB)", disk_size/(1024*1024));
ERROR_LOG(COMMON, "Trying to create SD Card image of size %" PRIu64 "MB is out of range (8MB-32GB)", disk_size/(1024*1024));
return false;
}

12
Source/Core/Common/Src/StringUtil.cpp
View file

@ -404,26 +404,30 @@ std::string UriEncode(const std::string & sSrc)
std::string UTF16ToUTF8(const std::wstring& input)
{
auto const size = WideCharToMultiByte(CP_UTF8, 0, input.data(), input.size(), nullptr, 0, nullptr, nullptr);
auto const size = WideCharToMultiByte(CP_UTF8, 0, input.data(), (int)input.size(), nullptr, 0, nullptr, nullptr);
std::string output;
output.resize(size);
if (size == 0 || size != WideCharToMultiByte(CP_UTF8, 0, input.data(), input.size(), &output[0], output.size(), nullptr, nullptr))
if (size == 0 || size != WideCharToMultiByte(CP_UTF8, 0, input.data(), (int)input.size(), &output[0], (int)output.size(), nullptr, nullptr))
{
output.clear();
}
return output;
}
std::wstring CPToUTF16(u32 code_page, const std::string& input)
{
auto const size = MultiByteToWideChar(code_page, 0, input.data(), input.size(), nullptr, 0);
auto const size = MultiByteToWideChar(code_page, 0, input.data(), (int)input.size(), nullptr, 0);
std::wstring output;
output.resize(size);
if (size == 0 || size != MultiByteToWideChar(code_page, 0, input.data(), input.size(), &output[0], output.size()))
if (size == 0 || size != MultiByteToWideChar(code_page, 0, input.data(), (int)input.size(), &output[0], (int)output.size()))
{
output.clear();
}
return output;
}

6
Source/Core/Common/Src/SysConf.cpp
View file

@ -5,6 +5,8 @@
#include "FileUtil.h"
#include "SysConf.h"
#include <cinttypes>
SysConf::SysConf()
: m_IsValid(false)
{
@ -42,7 +44,7 @@ bool SysConf::LoadFromFile(const char *filename)
u64 size = File::GetSize(filename);
if (size != SYSCONF_SIZE)
{
if (AskYesNoT("Your SYSCONF file is the wrong size.\nIt should be 0x%04x (but is 0x%04llx)\nDo you want to generate a new one?",
if (AskYesNoT("Your SYSCONF file is the wrong size.\nIt should be 0x%04x (but is 0x%04" PRIx64 ")\nDo you want to generate a new one?",
SYSCONF_SIZE, size))
{
GenerateSysConf();
@ -151,7 +153,7 @@ unsigned int create_item(SSysConfEntry &item, SysconfType type, const std::strin
{
item.offset = offset;
item.type = type;
item.nameLength = name.length();
item.nameLength = (u8)(name.length());
strncpy(item.name, name.c_str(), 32);
item.dataLength = data_length;
item.data = new u8[data_length];

42
Source/Core/Common/Src/Thread.h
View file

@ -38,7 +38,7 @@ class Event
public:
Event()
: is_set(false)
{};
{}
void Set()
{
@ -53,34 +53,20 @@ public:
void Wait()
{
std::unique_lock<std::mutex> lk(m_mutex);
m_condvar.wait(lk, IsSet(this));
m_condvar.wait(lk, [&]{ return is_set; });
is_set = false;
}
void Reset()
{
std::unique_lock<std::mutex> lk(m_mutex);
// no other action required, since wait loops on the predicate and any lingering signal will get cleared on the first iteration
// no other action required, since wait loops on
// the predicate and any lingering signal will get
// cleared on the first iteration
is_set = false;
}
private:
class IsSet
{
public:
IsSet(const Event* ev)
: m_event(ev)
{}
bool operator()()
{
return m_event->is_set;
}
private:
const Event* const m_event;
};
volatile bool is_set;
std::condition_variable m_condvar;
std::mutex m_mutex;
@ -110,28 +96,12 @@ public:
}
else
{
m_condvar.wait(lk, IsDoneWating(this));
m_condvar.wait(lk, [&]{ return (0 == m_waiting); });
return false;
}
}
private:
class IsDoneWating
{
public:
IsDoneWating(const Barrier* bar)
: m_bar(bar)
{}
bool operator()()
{
return (0 == m_bar->m_waiting);
}
private:
const Barrier* const m_bar;
};
std::condition_variable m_condvar;
std::mutex m_mutex;
const size_t m_count;

40
Source/Core/Common/Src/x64CPUDetect.cpp
View file

@ -162,6 +162,34 @@ void CPUInfo::Detect()
if ((cpu_id[2] >> 20) & 1) bSSE4_2 = true;
if ((cpu_id[2] >> 25) & 1) bAES = true;
// To check DAZ support, we first need to check FXSAVE support.
if ((cpu_id[3] >> 24) & 1)
{
// We can use FXSAVE.
bFXSR = true;
GC_ALIGNED16(u8 fx_state[512]);
memset(fx_state, 0, sizeof(fx_state));
#ifdef _WIN32
#ifdef _M_IX86
_fxsave(fx_state);
#elif defined (_M_X64)
_fxsave64(fx_state);
#endif
#else
__asm__("fxsave %0" : "=m" (fx_state));
#endif
// lowest byte of MXCSR_MASK
if ((fx_state[0x1C] >> 6) & 1)
{
// On x86, the FTZ field (supported since SSE1) only flushes denormal _outputs_ to zero,
// now that we checked DAZ support (flushing denormal _inputs_ to zero),
// we can set our generic flag.
bFlushToZero = true;
}
}
// AVX support requires 3 separate checks:
// - Is the AVX bit set in CPUID?
// - Is the XSAVE bit set in CPUID?
@ -169,7 +197,11 @@ void CPUInfo::Detect()
if (((cpu_id[2] >> 28) & 1) && ((cpu_id[2] >> 27) & 1))
{
if ((_xgetbv(_XCR_XFEATURE_ENABLED_MASK) & 0x6) == 0x6)
{
bAVX = true;
if ((cpu_id[2] >> 12) & 1)
bFMA = true;
}
}
}
if (max_ex_fn >= 0x80000004) {
@ -218,13 +250,19 @@ std::string CPUInfo::Summarize()
{
std::string sum(cpu_string);
if (bSSE) sum += ", SSE";
if (bSSE2) sum += ", SSE2";
if (bSSE2)
{
sum += ", SSE2";
if (!bFlushToZero)
sum += " (but not DAZ!)";
}
if (bSSE3) sum += ", SSE3";
if (bSSSE3) sum += ", SSSE3";
if (bSSE4_1) sum += ", SSE4.1";
if (bSSE4_2) sum += ", SSE4.2";
if (HTT) sum += ", HTT";
if (bAVX) sum += ", AVX";
if (bFMA) sum += ", FMA";
if (bAES) sum += ", AES";
if (bLongMode) sum += ", 64-bit support";
return sum;

57
Source/Core/Common/Src/x64Emitter.cpp
View file

@ -7,6 +7,8 @@
#include "x64ABI.h"
#include "CPUDetect.h"
#include <cinttypes>
namespace Gen
{
@ -154,6 +156,40 @@ void OpArg::WriteRex(XEmitter *emit, int opBits, int bits, int customOp) const
#endif
}
void OpArg::WriteVex(XEmitter* emit, int size, int packed, Gen::X64Reg regOp1, Gen::X64Reg regOp2) const
{
int R = !(regOp1 & 8);
int X = !(indexReg & 8);
int B = !(offsetOrBaseReg & 8);
// not so sure about this one...
int W = 0;
// aka map_select in AMD manuals
// only support VEX opcode map 1 for now (analog to secondary opcode map)
int mmmmm = 1;
int vvvv = (regOp2 == X64Reg::INVALID_REG) ? 0xf : (regOp2 ^ 0xf);
int L = size == 256;
int pp = (packed << 1) | (size == 64);
// do we need any VEX fields that only appear in the three-byte form?
if (X == 1 && B == 1 && W == 0 && mmmmm == 1)
{
u8 RvvvvLpp = (R << 7) | (vvvv << 3) | (L << 1) | pp;
emit->Write8(0xC5);
emit->Write8(RvvvvLpp);
}
else
{
u8 RXBmmmmm = (R << 7) | (X << 6) | (B << 5) | mmmmm;
u8 WvvvvLpp = (W << 7) | (vvvv << 3) | (L << 1) | pp;
emit->Write8(0xC4);
emit->Write8(RXBmmmmm);
emit->Write8(WvvvvLpp);
}
}
void OpArg::WriteRest(XEmitter *emit, int extraBytes, X64Reg _operandReg,
bool warn_64bit_offset) const
{
@ -176,7 +212,7 @@ void OpArg::WriteRest(XEmitter *emit, int extraBytes, X64Reg _operandReg,
_assert_msg_(DYNA_REC, (distance < 0x80000000LL
&& distance >= -0x80000000LL) ||
!warn_64bit_offset,
"WriteRest: op out of range (0x%llx uses 0x%llx)",
"WriteRest: op out of range (0x%" PRIx64 " uses 0x%" PRIx64 ")",
ripAddr, offset);
s32 offs = (s32)distance;
emit->Write32((u32)offs);
@ -1141,6 +1177,18 @@ void XEmitter::WriteSSEOp(int size, u8 sseOp, bool packed, X64Reg regOp, OpArg a
arg.WriteRest(this, extrabytes);
}
void XEmitter::WriteAVXOp(int size, u8 sseOp, bool packed, X64Reg regOp, OpArg arg, int extrabytes)
{
WriteAVXOp(size, sseOp, packed, regOp, X64Reg::INVALID_REG, arg, extrabytes);
}
void XEmitter::WriteAVXOp(int size, u8 sseOp, bool packed, X64Reg regOp1, X64Reg regOp2, OpArg arg, int extrabytes)
{
arg.WriteVex(this, size, packed, regOp1, regOp2);
Write8(sseOp);
arg.WriteRest(this, extrabytes, regOp1);
}
void XEmitter::MOVD_xmm(X64Reg dest, const OpArg &arg) {WriteSSEOp(64, 0x6E, true, dest, arg, 0);}
void XEmitter::MOVD_xmm(const OpArg &arg, X64Reg src) {WriteSSEOp(64, 0x7E, true, src, arg, 0);}
@ -1444,6 +1492,13 @@ void XEmitter::PMOVMSKB(X64Reg dest, OpArg arg) {WriteSSEOp(64, 0xD7, true, d
void XEmitter::PSHUFLW(X64Reg regOp, OpArg arg, u8 shuffle) {WriteSSEOp(64, 0x70, false, regOp, arg, 1); Write8(shuffle);}
// VEX
void XEmitter::VADDSD(X64Reg regOp1, X64Reg regOp2, OpArg arg) {WriteAVXOp(64, sseADD, false, regOp1, regOp2, arg);}
void XEmitter::VSUBSD(X64Reg regOp1, X64Reg regOp2, OpArg arg) {WriteAVXOp(64, sseSUB, false, regOp1, regOp2, arg);}
void XEmitter::VMULSD(X64Reg regOp1, X64Reg regOp2, OpArg arg) {WriteAVXOp(64, sseMUL, false, regOp1, regOp2, arg);}
void XEmitter::VDIVSD(X64Reg regOp1, X64Reg regOp2, OpArg arg) {WriteAVXOp(64, sseDIV, false, regOp1, regOp2, arg);}
void XEmitter::VSQRTSD(X64Reg regOp1, X64Reg regOp2, OpArg arg) {WriteAVXOp(64, sseSQRT, false, regOp1, regOp2, arg);}
// Prefixes
void XEmitter::LOCK() { Write8(0xF0); }

13
Source/Core/Common/Src/x64Emitter.h
View file

@ -33,6 +33,9 @@ enum X64Reg
XMM0=0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7,
XMM8, XMM9, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15,
YMM0=0, YMM1, YMM2, YMM3, YMM4, YMM5, YMM6, YMM7,
YMM8, YMM9, YMM10, YMM11, YMM12, YMM13, YMM14, YMM15,
INVALID_REG = 0xFFFFFFFF
};
@ -111,6 +114,7 @@ struct OpArg
offset = _offset;
}
void WriteRex(XEmitter *emit, int opBits, int bits, int customOp = -1) const;
void WriteVex(XEmitter* emit, int size, int packed, Gen::X64Reg regOp1, X64Reg regOp2) const;
void WriteRest(XEmitter *emit, int extraBytes=0, X64Reg operandReg=(X64Reg)0xFF, bool warn_64bit_offset = true) const;
void WriteSingleByteOp(XEmitter *emit, u8 op, X64Reg operandReg, int bits);
// This one is public - must be written to
@ -239,6 +243,8 @@ private:
void WriteBitTest(int bits, OpArg &dest, OpArg &index, int ext);
void WriteMXCSR(OpArg arg, int ext);
void WriteSSEOp(int size, u8 sseOp, bool packed, X64Reg regOp, OpArg arg, int extrabytes = 0);
void WriteAVXOp(int size, u8 sseOp, bool packed, X64Reg regOp, OpArg arg, int extrabytes = 0);
void WriteAVXOp(int size, u8 sseOp, bool packed, X64Reg regOp1, X64Reg regOp2, OpArg arg, int extrabytes = 0);
void WriteNormalOp(XEmitter *emit, int bits, NormalOp op, const OpArg &a1, const OpArg &a2);
protected:
@ -616,6 +622,13 @@ public:
void PSRAW(X64Reg reg, int shift);
void PSRAD(X64Reg reg, int shift);
// AVX
void VADDSD(X64Reg regOp1, X64Reg regOp2, OpArg arg);
void VSUBSD(X64Reg regOp1, X64Reg regOp2, OpArg arg);
void VMULSD(X64Reg regOp1, X64Reg regOp2, OpArg arg);
void VDIVSD(X64Reg regOp1, X64Reg regOp2, OpArg arg);
void VSQRTSD(X64Reg regOp1, X64Reg regOp2, OpArg arg);
void RTDSC();
// Utility functions

32
Source/Core/Common/Src/x64FPURoundMode.cpp
View file

@ -4,6 +4,7 @@
#include "Common.h"
#include "FPURoundMode.h"
#include "CPUDetect.h"
#ifndef _WIN32
static const unsigned short FPU_ROUND_NEAR = 0 << 10;
@ -14,8 +15,11 @@ static const unsigned short FPU_ROUND_MASK = 3 << 10;
#include <xmmintrin.h>
#endif
const u32 MASKS = 0x1F80; // mask away the interrupts.
// OR-mask for disabling FPU exceptions (bits 7-12 in the MXCSR register)
const u32 EXCEPTION_MASK = 0x1F80;
// Denormals-Are-Zero (non-IEEE mode: denormal inputs are set to +/- 0)
const u32 DAZ = 0x40;
// Flush-To-Zero (non-IEEE mode: denormal outputs are set to +/- 0)
const u32 FTZ = 0x8000;
namespace FPURoundMode
@ -79,16 +83,28 @@ namespace FPURoundMode
//but still - set any useful sse options here
#endif
}
void SetSIMDMode(u32 mode)
void SetSIMDMode(u32 roundingMode, u32 nonIEEEMode)
{
static const u32 ssetable[4] =
// lookup table for FPSCR.RN-to-MXCSR.RC translation
static const u32 roundingModeLUT[4] =
{
(0 << 13) | MASKS,
(3 << 13) | MASKS,
(2 << 13) | MASKS,
(1 << 13) | MASKS,
(0 << 13) | EXCEPTION_MASK, // nearest
(3 << 13) | EXCEPTION_MASK, // -inf
(2 << 13) | EXCEPTION_MASK, // +inf
(1 << 13) | EXCEPTION_MASK, // zero
};
u32 csr = ssetable[mode];
u32 csr = roundingModeLUT[roundingMode];
static const u32 denormalLUT[2] =
{
FTZ, // flush-to-zero only
FTZ | DAZ, // flush-to-zero and denormals-are-zero (may not be supported)
};
if (nonIEEEMode)
{
csr |= denormalLUT[cpu_info.bFlushToZero];
}
_mm_setcsr(csr);
}

66
Source/Core/Core/Core.vcxproj.filters
View file

@ -133,6 +133,9 @@
<Filter Include="IPC HLE %28IOS/Starlet%29\USB/BT/Wiimote">
<UniqueIdentifier>{8352be4d-d37d-4f55-adec-b940a9712802}</UniqueIdentifier>
</Filter>
<Filter Include="PowerPC\JitILCommon">
<UniqueIdentifier>{827afa93-1a80-4835-93ae-b5516d95867f}</UniqueIdentifier>
</Filter>
</ItemGroup>
<ItemGroup>
<ClCompile Include="Src\BootManager.cpp" />
@ -636,39 +639,12 @@
<ClCompile Include="Src\PowerPC\JitCommon\JitCache.cpp">
<Filter>PowerPC\JitCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\IR.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\IR_X86.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_Branch.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_FloatingPoint.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_Integer.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_LoadStore.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_LoadStoreFloating.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_LoadStorePaired.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_Paired.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_SystemRegisters.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\Jit64IL\JitIL_Tables.cpp">
<Filter>PowerPC\JitIL</Filter>
</ClCompile>
@ -706,6 +682,33 @@
<Filter>PowerPC\Jit64</Filter>
</ClCompile>
<ClCompile Include="Src\stdafx.cpp" />
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_Branch.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_FloatingPoint.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_Integer.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_LoadStore.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_LoadStoreFloating.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_LoadStorePaired.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_Paired.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\JitILBase_SystemRegisters.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
<ClCompile Include="Src\PowerPC\JitILCommon\IR.cpp">
<Filter>PowerPC\JitILCommon</Filter>
</ClCompile>
</ItemGroup>
<ItemGroup>
<ClInclude Include="Src\BootManager.h" />
@ -1190,9 +1193,6 @@
<ClInclude Include="Src\PowerPC\JitCommon\JitCache.h">
<Filter>PowerPC\JitCommon</Filter>
</ClInclude>
<ClInclude Include="Src\PowerPC\Jit64IL\IR.h">
<Filter>PowerPC\JitIL</Filter>
</ClInclude>
<ClInclude Include="Src\PowerPC\Jit64IL\JitIL.h">
<Filter>PowerPC\JitIL</Filter>
</ClInclude>
@ -1209,6 +1209,12 @@
<Filter>PowerPC\Jit64</Filter>
</ClInclude>
<ClInclude Include="Src\stdafx.h" />
<ClInclude Include="Src\PowerPC\JitILCommon\JitILBase.h">
<Filter>PowerPC\JitILCommon</Filter>
</ClInclude>
<ClInclude Include="Src\PowerPC\JitILCommon\IR.h">
<Filter>PowerPC\JitILCommon</Filter>
</ClInclude>
</ItemGroup>
<ItemGroup>
<Text Include="CMakeLists.txt" />

9
Source/Core/Core/Src/Console.cpp
View file

@ -17,8 +17,8 @@
#include "PowerPCDisasm.h"
#include "Console.h"
#define CASE1(x) if (memcmp(cmd, x, 2*sizeof(TCHAR))==0)
#define CASE(x) else if (memcmp(cmd, x, 4*sizeof(TCHAR))==0)
#define CASE1(x) if (!strcmp(cmd, (x)))
#define CASE(x) else if (!strcmp(cmd, (x)))
void Console_Submit(const char *cmd)
{
@ -27,7 +27,7 @@ void Console_Submit(const char *cmd)
Core::StartTrace(false);
INFO_LOG(CONSOLE, "Read tracing started.");
}
CASE1("w")
CASE("w")
{
Core::StartTrace(true);
INFO_LOG(CONSOLE, "Write tracing started.");
@ -141,3 +141,6 @@ void Console_Submit(const char *cmd)
ERROR_LOG(CONSOLE, "Invalid command");
}
}
#undef CASE1
#undef CASE

4
Source/Core/Core/Src/Core.cpp
View file

@ -388,8 +388,8 @@ void EmuThread()
OSD::AddMessage("Dolphin " + g_video_backend->GetName() + " Video Backend.", 5000);
if (!DSP::GetDSPEmulator()->Initialize(_CoreParameter.bWii,
_CoreParameter.bDSPThread))
if (!DSP::GetDSPEmulator()->Initialize(g_pWindowHandle,
_CoreParameter.bWii, _CoreParameter.bDSPThread))
{
HW::Shutdown();
g_video_backend->Shutdown();

4
Source/Core/Core/Src/CoreParameter.cpp
View file

@ -18,6 +18,8 @@
#include "Core.h" // for bWii
#include "FifoPlayer/FifoDataFile.h"
#include <cinttypes>
SCoreStartupParameter::SCoreStartupParameter()
: hInstance(0),
bEnableDebugging(false), bAutomaticStart(false), bBootToPause(false),
@ -278,7 +280,7 @@ bool SCoreStartupParameter::AutoSetup(EBootBS2 _BootBS2)
// Use the TitleIDhex for name and/or unique ID if launching from nand folder
// or if it is not ascii characters (specifically sysmenu could potentially apply to other things)
char titleidstr[17];
snprintf(titleidstr, 17, "%016llx", ContentLoader.GetTitleID());
snprintf(titleidstr, 17, "%016" PRIx64, ContentLoader.GetTitleID());
if (!m_strName.length())
{

3
Source/Core/Core/Src/CoreTiming.cpp
View file

@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <vector>
#include <cinttypes>
#include "Thread.h"
#include "PowerPC/PowerPC.h"
@ -429,7 +430,7 @@ void LogPendingEvents()
Event *ptr = first;
while (ptr)
{
INFO_LOG(POWERPC, "PENDING: Now: %lld Pending: %lld Type: %d", globalTimer, ptr->time, ptr->type);
INFO_LOG(POWERPC, "PENDING: Now: %" PRId64 " Pending: %" PRId64 " Type: %d", globalTimer, ptr->time, ptr->type);
ptr = ptr->next;
}
}

3
Source/Core/Core/Src/DSPEmulator.h
View file

@ -15,7 +15,7 @@ public:
virtual bool IsLLE() = 0;
virtual bool Initialize(bool bWii, bool bDSPThread) = 0;
virtual bool Initialize(void *hWnd, bool bWii, bool bDSPThread) = 0;
virtual void Shutdown() = 0;
virtual void DoState(PointerWrap &p) = 0;
@ -35,6 +35,7 @@ public:
protected:
SoundStream *soundStream;
void *m_hWnd;
};
DSPEmulator *CreateDSPEmulator(bool HLE);

4
Source/Core/Core/Src/FifoPlayer/FifoDataFile.cpp
View file

@ -86,7 +86,7 @@ bool FifoDataFile::Save(const char *filename)
header.xfRegsSize = XF_REGS_SIZE;
header.frameListOffset = frameListOffset;
header.frameCount = m_Frames.size();
header.frameCount = (u32)m_Frames.size();
header.flags = m_Flags;
@ -111,7 +111,7 @@ bool FifoDataFile::Save(const char *filename)
dstFrame.fifoStart = srcFrame.fifoStart;
dstFrame.fifoEnd = srcFrame.fifoEnd;
dstFrame.memoryUpdatesOffset = memoryUpdatesOffset;
dstFrame.numMemoryUpdates = srcFrame.memoryUpdates.size();
dstFrame.numMemoryUpdates = (u32)srcFrame.memoryUpdates.size();
// Write frame info
u64 frameOffset = frameListOffset + (i * sizeof(FileFrameInfo));

2
Source/Core/Core/Src/FifoPlayer/FifoPlaybackAnalyzer.cpp
View file

@ -234,7 +234,7 @@ u32 FifoPlaybackAnalyzer::DecodeCommand(u8 *data)
break;
}
return data - dataStart;
return (u32)(data - dataStart);
}
void FifoPlaybackAnalyzer::StoreEfbCopyRegion()

10
Source/Core/Core/Src/FifoPlayer/FifoPlayer.cpp
View file

@ -105,7 +105,9 @@ bool FifoPlayer::Play()
u32 FifoPlayer::GetFrameObjectCount()
{
if (m_CurrentFrame < m_FrameInfo.size())
return m_FrameInfo[m_CurrentFrame].objectStarts.size();
{
return (u32)(m_FrameInfo[m_CurrentFrame].objectStarts.size());
}
return 0;
}
@ -172,7 +174,7 @@ void FifoPlayer::WriteFrame(const FifoFrameInfo &frame, const AnalyzedFrameInfo
m_FrameFifoSize = frame.fifoDataSize;
// Determine start and end objects
u32 numObjects = info.objectStarts.size();
u32 numObjects = (u32)(info.objectStarts.size());
u32 drawStart = std::min(numObjects, m_ObjectRangeStart);
u32 drawEnd = std::min(numObjects - 1, m_ObjectRangeEnd);
@ -181,7 +183,9 @@ void FifoPlayer::WriteFrame(const FifoFrameInfo &frame, const AnalyzedFrameInfo
// Skip memory updates during frame if true
if (m_EarlyMemoryUpdates)
memoryUpdate = frame.memoryUpdates.size();
{
memoryUpdate = (u32)(frame.memoryUpdates.size());
}
if (numObjects > 0)
{

6
Source/Core/Core/Src/FifoPlayer/FifoRecorder.cpp
View file

@ -83,9 +83,9 @@ void FifoRecorder::WriteGPCommand(u8 *data, u32 size)
if (m_FrameEnded && m_FifoData.size() > 0)
{
size_t dataSize = m_FifoData.size();
m_CurrentFrame.fifoDataSize = dataSize;
m_CurrentFrame.fifoDataSize = (u32)dataSize;
m_CurrentFrame.fifoData = new u8[dataSize];
memcpy(m_CurrentFrame.fifoData, &m_FifoData[0], dataSize);
memcpy(m_CurrentFrame.fifoData, m_FifoData.data(), dataSize);
sMutex.lock();
@ -129,7 +129,7 @@ void FifoRecorder::WriteMemory(u32 address, u32 size, MemoryUpdate::Type type)
// Record memory update
MemoryUpdate memUpdate;
memUpdate.address = address;
memUpdate.fifoPosition = m_FifoData.size();
memUpdate.fifoPosition = (u32)(m_FifoData.size());
memUpdate.size = size;
memUpdate.type = type;
memUpdate.data = new u8[size];

5
Source/Core/Core/Src/HW/DSPHLE/DSPHLE.cpp
View file

@ -42,8 +42,9 @@ struct DSPState
}
};
bool DSPHLE::Initialize(bool bWii, bool bDSPThread)
bool DSPHLE::Initialize(void *hWnd, bool bWii, bool bDSPThread)
{
m_hWnd = hWnd;
m_bWii = bWii;
m_pUCode = NULL;
m_lastUCode = NULL;
@ -265,7 +266,7 @@ void DSPHLE::InitMixer()
unsigned int AISampleRate, DACSampleRate;
AudioInterface::Callback_GetSampleRate(AISampleRate, DACSampleRate);
delete soundStream;
soundStream = AudioCommon::InitSoundStream(new HLEMixer(this, AISampleRate, DACSampleRate, 48000));
soundStream = AudioCommon::InitSoundStream(new HLEMixer(this, AISampleRate, DACSampleRate, 48000), m_hWnd);
if(!soundStream) PanicAlert("Error starting up sound stream");
// Mixer is initialized
m_InitMixer = true;

2
Source/Core/Core/Src/HW/DSPHLE/DSPHLE.h
View file

@ -16,7 +16,7 @@ class DSPHLE : public DSPEmulator {
public:
DSPHLE();
virtual bool Initialize(bool bWii, bool bDSPThread) override;
virtual bool Initialize(void *hWnd, bool bWii, bool bDSPThread) override;
virtual void Shutdown() override;
virtual bool IsLLE() override { return false ; }

60
Source/Core/Core/Src/HW/DSPHLE/UCodes/UCode_AX_Voice.h
View file

@ -469,36 +469,42 @@ void ProcessVoice(PB_TYPE& pb, const AXBuffers& buffers, u16 count, AXMixControl
// Mix LRS, AUXA and AUXB depending on mixer_control
// TODO: Handle DPL2 on AUXB.
if (mctrl & MIX_L)
MixAdd(buffers.left, samples, count, &pb.mixer.left, &pb.dpop.left, mctrl & MIX_L_RAMP);
if (mctrl & MIX_R)
MixAdd(buffers.right, samples, count, &pb.mixer.right, &pb.dpop.right, mctrl & MIX_R_RAMP);
if (mctrl & MIX_S)
MixAdd(buffers.surround, samples, count, &pb.mixer.surround, &pb.dpop.surround, mctrl & MIX_S_RAMP);
#define MIX_ON(C) (0 != (mctrl & MIX_##C))
#define RAMP_ON(C) (0 != (mctrl & MIX_##C##_RAMP))
if (mctrl & MIX_AUXA_L)
MixAdd(buffers.auxA_left, samples, count, &pb.mixer.auxA_left, &pb.dpop.auxA_left, mctrl & MIX_AUXA_L_RAMP);
if (mctrl & MIX_AUXA_R)
MixAdd(buffers.auxA_right, samples, count, &pb.mixer.auxA_right, &pb.dpop.auxA_right, mctrl & MIX_AUXA_R_RAMP);
if (mctrl & MIX_AUXA_S)
MixAdd(buffers.auxA_surround, samples, count, &pb.mixer.auxA_surround, &pb.dpop.auxA_surround, mctrl & MIX_AUXA_S_RAMP);
if (MIX_ON(L))
MixAdd(buffers.left, samples, count, &pb.mixer.left, &pb.dpop.left, RAMP_ON(L));
if (MIX_ON(R))
MixAdd(buffers.right, samples, count, &pb.mixer.right, &pb.dpop.right, RAMP_ON(R));
if (MIX_ON(S))
MixAdd(buffers.surround, samples, count, &pb.mixer.surround, &pb.dpop.surround, RAMP_ON(S));
if (mctrl & MIX_AUXB_L)
MixAdd(buffers.auxB_left, samples, count, &pb.mixer.auxB_left, &pb.dpop.auxB_left, mctrl & MIX_AUXB_L_RAMP);
if (mctrl & MIX_AUXB_R)
MixAdd(buffers.auxB_right, samples, count, &pb.mixer.auxB_right, &pb.dpop.auxB_right, mctrl & MIX_AUXB_R_RAMP);
if (mctrl & MIX_AUXB_S)
MixAdd(buffers.auxB_surround, samples, count, &pb.mixer.auxB_surround, &pb.dpop.auxB_surround, mctrl & MIX_AUXB_S_RAMP);
if (MIX_ON(AUXA_L))
MixAdd(buffers.auxA_left, samples, count, &pb.mixer.auxA_left, &pb.dpop.auxA_left, RAMP_ON(AUXA_L));
if (MIX_ON(AUXA_R))
MixAdd(buffers.auxA_right, samples, count, &pb.mixer.auxA_right, &pb.dpop.auxA_right, RAMP_ON(AUXA_R));
if (MIX_ON(AUXA_S))
MixAdd(buffers.auxA_surround, samples, count, &pb.mixer.auxA_surround, &pb.dpop.auxA_surround, RAMP_ON(AUXA_S));
if (MIX_ON(AUXB_L))
MixAdd(buffers.auxB_left, samples, count, &pb.mixer.auxB_left, &pb.dpop.auxB_left, RAMP_ON(AUXB_L));
if (MIX_ON(AUXB_R))
MixAdd(buffers.auxB_right, samples, count, &pb.mixer.auxB_right, &pb.dpop.auxB_right, RAMP_ON(AUXB_R));
if (MIX_ON(AUXB_S))
MixAdd(buffers.auxB_surround, samples, count, &pb.mixer.auxB_surround, &pb.dpop.auxB_surround, RAMP_ON(AUXB_S));
#ifdef AX_WII
if (mctrl & MIX_AUXC_L)
MixAdd(buffers.auxC_left, samples, count, &pb.mixer.auxC_left, &pb.dpop.auxC_left, mctrl & MIX_AUXC_L_RAMP);
if (mctrl & MIX_AUXC_R)
MixAdd(buffers.auxC_right, samples, count, &pb.mixer.auxC_right, &pb.dpop.auxC_right, mctrl & MIX_AUXC_R_RAMP);
if (mctrl & MIX_AUXC_S)
MixAdd(buffers.auxC_surround, samples, count, &pb.mixer.auxC_surround, &pb.dpop.auxC_surround, mctrl & MIX_AUXC_S_RAMP);
if (MIX_ON(AUXC_L))
MixAdd(buffers.auxC_left, samples, count, &pb.mixer.auxC_left, &pb.dpop.auxC_left, RAMP_ON(AUXC_L));
if (MIX_ON(AUXC_R))
MixAdd(buffers.auxC_right, samples, count, &pb.mixer.auxC_right, &pb.dpop.auxC_right, RAMP_ON(AUXC_R));
if (MIX_ON(AUXC_S))
MixAdd(buffers.auxC_surround, samples, count, &pb.mixer.auxC_surround, &pb.dpop.auxC_surround, RAMP_ON(AUXC_S));
#endif
#undef MIX_ON
#undef RAMP_ON
// Optionally, phase shift left or right channel to simulate 3D sound.
if (pb.initial_time_delay.on)
{
@ -524,8 +530,8 @@ void ProcessVoice(PB_TYPE& pb, const AXBuffers& buffers, u16 count, AXMixControl
pb.remote_src.cur_addr_frac = curr_pos & 0xFFFF;
// Mix to main[0-3] and aux[0-3]
#define WMCHAN_MIX_ON(n) ((pb.remote_mixer_control >> (2 * n)) & 3)
#define WMCHAN_MIX_RAMP(n) ((pb.remote_mixer_control >> (2 * n)) & 2)
#define WMCHAN_MIX_ON(n) (0 != ((pb.remote_mixer_control >> (2 * n)) & 3))
#define WMCHAN_MIX_RAMP(n) (0 != ((pb.remote_mixer_control >> (2 * n)) & 2))
if (WMCHAN_MIX_ON(0))
MixAdd(buffers.wm_main0, wm_samples, wm_count, &pb.remote_mixer.main0, &pb.remote_dpop.main0, WMCHAN_MIX_RAMP(0));
@ -544,6 +550,8 @@ void ProcessVoice(PB_TYPE& pb, const AXBuffers& buffers, u16 count, AXMixControl
if (WMCHAN_MIX_ON(7))
MixAdd(buffers.wm_aux3, wm_samples, wm_count, &pb.remote_mixer.aux3, &pb.remote_dpop.aux3, WMCHAN_MIX_RAMP(7));
}
#undef WMCHAN_MIX_RAMP
#undef WMCHAN_MIX_ON
#endif
}

5
Source/Core/Core/Src/HW/DSPLLE/DSPLLE.cpp
View file

@ -130,8 +130,9 @@ void DSPLLE::dsp_thread(DSPLLE *dsp_lle)
}
}
bool DSPLLE::Initialize(bool bWii, bool bDSPThread)
bool DSPLLE::Initialize(void *hWnd, bool bWii, bool bDSPThread)
{
m_hWnd = hWnd;
m_bWii = bWii;
m_bDSPThread = bDSPThread;
m_InitMixer = false;
@ -184,7 +185,7 @@ void DSPLLE::InitMixer()
unsigned int AISampleRate, DACSampleRate;
AudioInterface::Callback_GetSampleRate(AISampleRate, DACSampleRate);
delete soundStream;
soundStream = AudioCommon::InitSoundStream(new CMixer(AISampleRate, DACSampleRate, 48000));
soundStream = AudioCommon::InitSoundStream(new CMixer(AISampleRate, DACSampleRate, 48000), m_hWnd);
if(!soundStream) PanicAlert("Error starting up sound stream");
// Mixer is initialized
m_InitMixer = true;

2
Source/Core/Core/Src/HW/DSPLLE/DSPLLE.h
View file

@ -14,7 +14,7 @@ class DSPLLE : public DSPEmulator {
public:
DSPLLE();
virtual bool Initialize(bool bWii, bool bDSPThread);
virtual bool Initialize(void *hWnd, bool bWii, bool bDSPThread);
virtual void Shutdown();
virtual bool IsLLE() { return true; }

5
Source/Core/Core/Src/HW/DSPLLE/DSPLLEGlobals.cpp
View file

@ -6,6 +6,7 @@
#include "FileUtil.h"
#include "DSP/DSPCore.h"
#include "DSPLLEGlobals.h"
#include <cinttypes>
#if PROFILE
@ -37,12 +38,12 @@ void ProfilerDump(u64 count)
File::IOFile pFile("DSP_Prof.txt", "wt");
if (pFile)
{
fprintf(pFile.GetHandle(), "Number of DSP steps: %llu\n\n", count);
fprintf(pFile.GetHandle(), "Number of DSP steps: %" PRIu64 "\n\n", count);
for (int i=0; i<PROFILE_MAP_SIZE;i++)
{
if (g_profileMap[i] > 0)
{
fprintf(pFile.GetHandle(), "0x%04X: %llu\n", i, g_profileMap[i]);
fprintf(pFile.GetHandle(), "0x%04X: %" PRIu64 "\n", i, g_profileMap[i]);
}
}
}

2
Source/Core/Core/Src/HW/DVDInterface.cpp
View file

@ -325,7 +325,7 @@ void ChangeDisc(const char* _newFileName)
{
Movie::g_bDiscChange = true;
std::string fileName = _newFileName;
int sizeofpath = fileName.find_last_of("/\\") + 1;
auto sizeofpath = fileName.find_last_of("/\\") + 1;
if (fileName.substr(sizeofpath).length() > 40)
{
PanicAlert("Saving iso filename to .dtm failed; max file name length is 40 characters.");

2
Source/Core/Core/Src/HW/EXI_DeviceMemoryCard.cpp
View file

@ -191,7 +191,7 @@ void CEXIMemoryCard::CmdDone()
void CEXIMemoryCard::CmdDoneLater(u64 cycles)
{
CoreTiming::RemoveEvent(et_cmd_done);
CoreTiming::ScheduleEvent(cycles, et_cmd_done, (u64)card_index);
CoreTiming::ScheduleEvent((int)cycles, et_cmd_done, (u64)card_index);
}
void CEXIMemoryCard::SetCS(int cs)

15
Source/Core/Core/Src/HW/GCMemcard.cpp
View file

@ -4,6 +4,9 @@
#include "GCMemcard.h"
#include "ColorUtil.h"
#include <cinttypes>
static void ByteSwap(u8 *valueA, u8 *valueB)
{
u8 tmp = *valueA;
@ -37,19 +40,19 @@ GCMemcard::GCMemcard(const char *filename, bool forceCreation, bool sjis)
PanicAlertT("File has the extension \"%s\"\nvalid extensions are (.raw/.gcp)", fileType.c_str());
return;
}
u32 size = mcdFile.GetSize();
auto size = mcdFile.GetSize();
if (size < MC_FST_BLOCKS*BLOCK_SIZE)
{
PanicAlertT("%s failed to load as a memorycard \nfile is not large enough to be a valid memory card file (0x%x bytes)", filename, size);
PanicAlertT("%s failed to load as a memorycard \nfile is not large enough to be a valid memory card file (0x%x bytes)", filename, (unsigned) size);
return;
}
if (size % BLOCK_SIZE)
{
PanicAlertT("%s failed to load as a memorycard \n Card file size is invalid (0x%x bytes)", filename, size);
PanicAlertT("%s failed to load as a memorycard \n Card file size is invalid (0x%x bytes)", filename, (unsigned) size);
return;
}
m_sizeMb = (size/BLOCK_SIZE) / MBIT_TO_BLOCKS;
m_sizeMb = (u16)((size/BLOCK_SIZE) / MBIT_TO_BLOCKS);
switch (m_sizeMb)
{
case MemCard59Mb:
@ -60,7 +63,7 @@ GCMemcard::GCMemcard(const char *filename, bool forceCreation, bool sjis)
case MemCard2043Mb:
break;
default:
PanicAlertT("%s failed to load as a memorycard \n Card size is invalid (0x%x bytes)", filename, size);
PanicAlertT("%s failed to load as a memorycard \n Card size is invalid (0x%x bytes)", filename, (unsigned) size);
return;
}
}
@ -173,7 +176,7 @@ GCMemcard::GCMemcard(const char *filename, bool forceCreation, bool sjis)
}
else
{
PanicAlertT("Failed to read block %d of the save data\nMemcard may be truncated\nFilePosition:%llx", i, mcdFile.Tell());
PanicAlertT("Failed to read block %d of the save data\nMemcard may be truncated\nFilePosition:%" PRIx64, i, mcdFile.Tell());
m_valid = false;
break;
}

2
Source/Core/Core/Src/HW/GCPadEmu.cpp
View file

@ -209,5 +209,5 @@ void GCPad::LoadDefaults(const ControllerInterface& ciface)
bool GCPad::GetMicButton() const
{
return m_buttons->controls.back()->control_ref->State();
return (0.0f != m_buttons->controls.back()->control_ref->State());
}

4
Source/Core/Core/Src/HW/WiimoteEmu/EmuSubroutines.cpp
View file

@ -32,7 +32,7 @@
namespace WiimoteEmu
{
void Spy(Wiimote* wm_, const void* data_, int size_)
void Spy(Wiimote* wm_, const void* data_, size_t size_)
{
#if 0
// enable log
@ -1275,7 +1275,7 @@ void Wiimote::DoState(PointerWrap& p)
else
{
std::queue<ReadRequest> tmp_queue(m_read_requests);
size = m_read_requests.size();
size = (u32)(m_read_requests.size());
p.Do(size);
while (!tmp_queue.empty())
{

2
Source/Core/Core/Src/HW/WiimoteEmu/WiimoteEmu.cpp
View file

@ -765,7 +765,7 @@ void Wiimote::Update()
if (-1 == rptf_size)
{
std::copy(rpt.begin(), rpt.end(), data);
rptf_size = rpt.size();
rptf_size = (s8)(rpt.size());
}
}
}

4
Source/Core/Core/Src/HW/WiimoteEmu/WiimoteEmu.h
View file

@ -96,7 +96,7 @@ inline double trim(double a)
class Wiimote : public ControllerEmu
{
friend class WiimoteReal::Wiimote;
friend void Spy(Wiimote* wm_, const void* data_, int size_);
friend void Spy(Wiimote* wm_, const void* data_, size_t size_);
public:
enum
@ -245,7 +245,7 @@ private:
} m_reg_speaker;
};
void Spy(Wiimote* wm_, const void* data_, int size_);
void Spy(Wiimote* wm_, const void* data_, size_t size_);
}

2
Source/Core/Core/Src/HW/WiimoteReal/IODummy.cpp
View file

@ -72,7 +72,7 @@ int Wiimote::IORead(u8* buf)
return 0;
}
int Wiimote::IOWrite(const u8* buf, int len)
int Wiimote::IOWrite(const u8* buf, size_t len)
{
return 0;
}

4
Source/Core/Core/Src/HW/WiimoteReal/IONix.cpp
View file

@ -263,9 +263,9 @@ int Wiimote::IORead(u8* buf)
return r;
}
int Wiimote::IOWrite(u8 const* buf, int len)
int Wiimote::IOWrite(u8 const* buf, size_t len)
{
return write(int_sock, buf, len);
return write(int_sock, buf, (int)len);
}
}; // WiimoteReal

10
Source/Core/Core/Src/HW/WiimoteReal/IOWin.cpp
View file

@ -140,7 +140,7 @@ namespace WiimoteReal
{
int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stack_t &stack, const u8* buf, int len);
int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stack_t &stack, const u8* buf, size_t len);
int _IORead(HANDLE &dev_handle, OVERLAPPED &hid_overlap_read, u8* buf, int index);
void _IOWakeup(HANDLE &dev_handle, OVERLAPPED &hid_overlap_read);
@ -247,7 +247,7 @@ void WiimoteScanner::FindWiimotes(std::vector<Wiimote*> & found_wiimotes, Wiimot
// SLEEP(2000);
}
int CheckDeviceType_Write(HANDLE &dev_handle, const u8* buf, int size, int attempts)
int CheckDeviceType_Write(HANDLE &dev_handle, const u8* buf, size_t size, int attempts)
{
OVERLAPPED hid_overlap_write = OVERLAPPED();
hid_overlap_write.hEvent = CreateEvent(NULL, true, false, NULL);
@ -641,7 +641,7 @@ int Wiimote::IORead(u8* buf)
}
int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stack_t &stack, const u8* buf, int len)
int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stack_t &stack, const u8* buf, size_t len)
{
WiimoteEmu::Spy(NULL, buf, len);
@ -663,7 +663,7 @@ int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stac
}
case MSBT_STACK_MS:
{
auto result = HidD_SetOutputReport(dev_handle, const_cast<u8*>(buf) + 1, len - 1);
auto result = HidD_SetOutputReport(dev_handle, const_cast<u8*>(buf) + 1, (ULONG)(len - 1));
//FlushFileBuffers(dev_handle);
if (!result)
@ -715,7 +715,7 @@ int _IOWrite(HANDLE &dev_handle, OVERLAPPED &hid_overlap_write, enum win_bt_stac
return 0;
}
int Wiimote::IOWrite(const u8* buf, int len)
int Wiimote::IOWrite(const u8* buf, size_t len)
{
return _IOWrite(dev_handle, hid_overlap_write, stack, buf, len);
}

4
Source/Core/Core/Src/HW/WiimoteReal/IOdarwin.mm
View file

@ -310,14 +310,14 @@ int Wiimote::IORead(unsigned char *buf)
return inputlen;
}
int Wiimote::IOWrite(const unsigned char *buf, int len)
int Wiimote::IOWrite(const unsigned char *buf, size_t len)
{
IOReturn ret;
if (!IsConnected())
return 0;
ret = [ichan writeAsync: const_cast<void*>((void *)buf) length: len refcon: nil];
ret = [ichan writeAsync: const_cast<void*>((void *)buf) length: (int)len refcon: nil];
if (ret == kIOReturnSuccess)
return len;

8
Source/Core/Core/Src/HW/WiimoteReal/WiimoteReal.cpp
View file

@ -239,7 +239,9 @@ bool Wiimote::Write()
IOWrite(rpt.data(), rpt.size());
if (is_speaker_data)
{
m_last_audio_report.Update();
}
m_write_reports.Pop();
return true;
@ -293,8 +295,10 @@ void Wiimote::Update()
// Send the report
if (!rpt.empty() && m_channel > 0)
Core::Callback_WiimoteInterruptChannel(index, m_channel,
rpt.data(), rpt.size());
{
Core::Callback_WiimoteInterruptChannel(index, m_channel,
rpt.data(), (u32)rpt.size());
}
}
void Wiimote::Prepare(int _index)

2
Source/Core/Core/Src/HW/WiimoteReal/WiimoteReal.h
View file

@ -106,7 +106,7 @@ private:
void WriteReport(Report rpt);
int IORead(u8* buf);
int IOWrite(u8 const* buf, int len);
int IOWrite(u8 const* buf, size_t len);
void IOWakeup();
void ThreadFunc();

6
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE.cpp
View file

@ -83,7 +83,7 @@ static u64 last_reply_time;
void EnqueReplyCallback(u64 userdata, int)
{
std::lock_guard<std::mutex> lk(s_reply_queue);
reply_queue.push_back(userdata);
reply_queue.push_back((u32)userdata);
}
void Init()
@ -546,7 +546,9 @@ void ExecuteCommand(u32 _Address)
const s64 ticks_til_last_reply = last_reply_time - CoreTiming::GetTicks();
if (ticks_til_last_reply > 0)
reply_delay = ticks_til_last_reply;
{
reply_delay = (int)ticks_til_last_reply;
}
last_reply_time = CoreTiming::GetTicks() + reply_delay;

14
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_DI.cpp
View file

@ -19,6 +19,8 @@
#include "../../DiscIO/Src/FileMonitor.h"
#include <cinttypes>
using namespace DVDInterface;
@ -108,7 +110,7 @@ bool CWII_IPC_HLE_Device_di::IOCtlV(u32 _CommandAddress)
// Get TMD offset for requested partition...
u64 const TMDOffset = ((u64)Memory::Read_U32(CommandBuffer.InBuffer[0].m_Address + 4) << 2 ) + 0x2c0;
INFO_LOG(WII_IPC_DVD, "DVDLowOpenPartition: TMDOffset 0x%016llx", TMDOffset);
INFO_LOG(WII_IPC_DVD, "DVDLowOpenPartition: TMDOffset 0x%016" PRIx64, TMDOffset);
static u32 const TMDsz = 0x208; //CommandBuffer.PayloadBuffer[0].m_Size;
u8 pTMD[TMDsz];
@ -204,13 +206,13 @@ u32 CWII_IPC_HLE_Device_di::ExecuteCommand(u32 _BufferIn, u32 _BufferInSize, u32
pFilename = m_pFileSystem->GetFileName(DVDAddress);
if (pFilename != NULL)
{
INFO_LOG(WII_IPC_DVD, "DVDLowRead: %s (0x%llx) - (DVDAddr: 0x%llx, Size: 0x%x)",
INFO_LOG(WII_IPC_DVD, "DVDLowRead: %s (0x%" PRIx64 ") - (DVDAddr: 0x%" PRIx64 ", Size: 0x%x)",
pFilename, m_pFileSystem->GetFileSize(pFilename), DVDAddress, Size);
FileMon::CheckFile(std::string(pFilename), (int)m_pFileSystem->GetFileSize(pFilename));
}
else
{
INFO_LOG(WII_IPC_DVD, "DVDLowRead: file unknown - (DVDAddr: 0x%llx, Size: 0x%x)",
INFO_LOG(WII_IPC_DVD, "DVDLowRead: file unknown - (DVDAddr: 0x%" PRIx64 ", Size: 0x%x)",
DVDAddress, Size);
}
}
@ -308,7 +310,7 @@ u32 CWII_IPC_HLE_Device_di::ExecuteCommand(u32 _BufferIn, u32 _BufferInSize, u32
u64 DVDAddress = (u64)DVDAddress32 << 2;
INFO_LOG(WII_IPC_DVD, "DVDLowUnencryptedRead: DVDAddr: 0x%08llx, Size: 0x%x", DVDAddress, Size);
INFO_LOG(WII_IPC_DVD, "DVDLowUnencryptedRead: DVDAddr: 0x%08" PRIx64 ", Size: 0x%x", DVDAddress, Size);
if (Size > _BufferOutSize)
{
@ -342,12 +344,12 @@ u32 CWII_IPC_HLE_Device_di::ExecuteCommand(u32 _BufferIn, u32 _BufferInSize, u32
pFilename = m_pFileSystem->GetFileName(DVDAddress);
if (pFilename != NULL)
{
INFO_LOG(WII_IPC_DVD, "DVDLowSeek: %s (0x%llx) - (DVDAddr: 0x%llx)",
INFO_LOG(WII_IPC_DVD, "DVDLowSeek: %s (0x%" PRIx64 ") - (DVDAddr: 0x%" PRIx64 ")",
pFilename, m_pFileSystem->GetFileSize(pFilename), DVDAddress);
}
else
{
INFO_LOG(WII_IPC_DVD, "DVDLowSeek: file unknown - (DVDAddr: 0x%llx)",
INFO_LOG(WII_IPC_DVD, "DVDLowSeek: file unknown - (DVDAddr: 0x%" PRIx64 ")",
DVDAddress);
}
}

12
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_es.cpp
View file

@ -35,6 +35,10 @@
#include "WII_IPC_HLE_Device_es.h"
// need to include this before polarssl/aes.h,
// otherwise we may not get __STDC_FORMAT_MACROS
#include <cinttypes>
#include "../PowerPC/PowerPC.h"
#include "../VolumeHandler.h"
#include "FileUtil.h"
@ -129,7 +133,7 @@ void CWII_IPC_HLE_Device_es::DoState(PointerWrap& p)
p.Do(m_AccessIdentID);
p.Do(m_TitleIDs);
u32 Count = m_ContentAccessMap.size();
u32 Count = (u32)(m_ContentAccessMap.size());
p.Do(Count);
u32 CFD, Position;
@ -205,7 +209,7 @@ u32 CWII_IPC_HLE_Device_es::OpenTitleContent(u32 CFD, u64 TitleID, u16 Index)
if (!Loader.IsValid())
{
WARN_LOG(WII_IPC_ES, "ES: loader not valid for %llx", TitleID);
WARN_LOG(WII_IPC_ES, "ES: loader not valid for %" PRIx64, TitleID);
return 0xffffffff;
}
@ -940,7 +944,7 @@ bool CWII_IPC_HLE_Device_es::IOCtlV(u32 _CommandAddress)
if (!bSuccess)
{
PanicAlertT("IOCTL_ES_LAUNCH: Game tried to reload a title that is not available in your NAND dump\n"
"TitleID %016llx.\n Dolphin will likely hang now.", TitleID);
"TitleID %016" PRIx64".\n Dolphin will likely hang now.", TitleID);
}
else
{
@ -983,7 +987,7 @@ bool CWII_IPC_HLE_Device_es::IOCtlV(u32 _CommandAddress)
//TODO: provide correct return code when bSuccess= false
Memory::Write_U32(0, _CommandAddress + 0x4);
ERROR_LOG(WII_IPC_ES, "IOCTL_ES_LAUNCH %016llx %08x %016llx %08x %016llx %04x", TitleID,view,ticketid,devicetype,titleid,access);
ERROR_LOG(WII_IPC_ES, "IOCTL_ES_LAUNCH %016" PRIx64 " %08x %016" PRIx64 " %08x %016" PRIx64 " %04x", TitleID,view,ticketid,devicetype,titleid,access);
// IOCTL_ES_LAUNCH 0001000248414341 00000001 0001c0fef3df2cfa 00000000 0001000248414341 ffff
// This is necessary because Reset(true) above deleted this object. Ew.

2
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_fs.cpp
View file

@ -571,7 +571,7 @@ void CWII_IPC_HLE_Device_fs::DoState(PointerWrap& p)
}
else
{
u32 size = entry.size;
u32 size = (u32)entry.size;
p.Do(size);
File::IOFile handle(entry.physicalName, "rb");

12
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_hid.cpp
View file

@ -385,12 +385,12 @@ void CWII_IPC_HLE_Device_hid::FillOutDevices(u32 BufferOut, u32 BufferOutSize)
Memory::WriteBigEData((const u8*)&wii_device, OffsetBuffer, Align(wii_device.bLength, 4));
OffsetBuffer += Align(wii_device.bLength, 4);
bool deviceValid = true;
bool isHID = false;
for (c = 0; deviceValid && c < desc.bNumConfigurations; c++)
{
struct libusb_config_descriptor *config = NULL;
int cRet = libusb_get_config_descriptor(device, c, &config);
// do not try to use usb devices with more than one interface, games can crash
if(cRet == 0 && config->bNumInterfaces <= MAX_HID_INTERFACES)
{
@ -402,10 +402,14 @@ void CWII_IPC_HLE_Device_hid::FillOutDevices(u32 BufferOut, u32 BufferOutSize)
for (ic = 0; ic < config->bNumInterfaces; ic++)
{
const struct libusb_interface *interfaceContainer = &config->interface[ic];
for (i = 0; i < interfaceContainer->num_altsetting; i++)
{
const struct libusb_interface_descriptor *interface = &interfaceContainer->altsetting[i];
if (interface->bInterfaceClass == LIBUSB_CLASS_HID)
isHID = true;
WiiHIDInterfaceDescriptor wii_interface;
ConvertInterfaceToWii(&wii_interface, interface);
Memory::WriteBigEData((const u8*)&wii_interface, OffsetBuffer, Align(wii_interface.bLength, 4));
@ -435,6 +439,12 @@ void CWII_IPC_HLE_Device_hid::FillOutDevices(u32 BufferOut, u32 BufferOutSize)
}
} // configs
if (!isHID)
{
deviceValid = false;
OffsetBuffer = OffsetStart;
}
if (deviceValid)
{
Memory::Write_U32(OffsetBuffer-OffsetStart, OffsetStart); // fill in length

2
Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_usb.cpp
View file

@ -866,7 +866,7 @@ bool CWII_IPC_HLE_Device_usb_oh1_57e_305::SendEventRoleChange(bdaddr_t _bd, bool
bool CWII_IPC_HLE_Device_usb_oh1_57e_305::SendEventNumberOfCompletedPackets()
{
SQueuedEvent Event(sizeof(hci_event_hdr_t) + sizeof(hci_num_compl_pkts_ep) + sizeof(hci_num_compl_pkts_info) * m_WiiMotes.size(), 0);
SQueuedEvent Event((u32)(sizeof(hci_event_hdr_t) + sizeof(hci_num_compl_pkts_ep) + (sizeof(hci_num_compl_pkts_info) * m_WiiMotes.size())), 0);
INFO_LOG(WII_IPC_WIIMOTE, "Event: SendEventNumberOfCompletedPackets");

11
Source/Core/Core/Src/IPC_HLE/hci.h
View file

@ -84,14 +84,6 @@
// All structs in this file are packed
#pragma pack(push, 1)
/*
* Bluetooth Address Family Protocol Numbers
*/
#define BTPROTO_HCI 1
#define BTPROTO_L2CAP 2
#define BTPROTO_RFCOMM 3
#define BTPROTO_SCO 4
/* All sizes are in bytes */
#define BLUETOOTH_BDADDR_SIZE 6
@ -102,9 +94,8 @@
typedef struct {
uint8_t b[BLUETOOTH_BDADDR_SIZE];
} bdaddr_t;
#endif
#define BDADDR_ANY { { 0, 0, 0, 0, 0, 0 } }
#endif
/**************************************************************************
**************************************************************************

9
Source/Core/Core/Src/NetPlayClient.cpp
View file

@ -445,10 +445,11 @@ void NetPlayClient::SendWiimoteState(const PadMapping in_game_pad, const NetWiim
sf::Packet spac;
spac << (MessageId)NP_MSG_WIIMOTE_DATA;
spac << in_game_pad;
u8 size = nw.size();
spac << size;
for (unsigned int i = 0; i < size; ++i)
spac << nw.data()[i];
spac << (u8)nw.size();
for (auto it : nw)
{
spac << it;
}
std::lock_guard<std::recursive_mutex> lks(m_crit.send);
m_socket.Send(spac);

6
Source/Core/Core/Src/NetPlayServer.cpp
View file

@ -153,7 +153,7 @@ unsigned int NetPlayServer::OnConnect(sf::SocketTCP& socket)
rpac >> player.name;
// give new client first available id
player.pid = m_players.size() + 1;
player.pid = (PlayerId)(m_players.size() + 1);
// try to automatically assign new user a pad
for (unsigned int m = 0; m < 4; ++m)
@ -435,12 +435,14 @@ unsigned int NetPlayServer::OnData(sf::Packet& packet, sf::SocketTCP& socket)
case NP_MSG_PONG :
{
const u32 ping = m_ping_timer.GetTimeElapsed();
const u32 ping = (u32)m_ping_timer.GetTimeElapsed();
u32 ping_key = 0;
packet >> ping_key;
if (m_ping_key == ping_key)
{
player.ping = ping;
}
sf::Packet spac;
spac << (MessageId)NP_MSG_PLAYER_PING_DATA;

4
Source/Core/Core/Src/PowerPC/Interpreter/Interpreter.cpp
View file

@ -10,11 +10,11 @@
#include "../../Host.h"
#include "../../IPC_HLE/WII_IPC_HLE.h"
#ifdef USE_GDBSTUB
#include "../GDBStub.h"
#endif
#include <cinttypes>
namespace {
u32 last_pc;
@ -79,7 +79,7 @@ void Trace( UGeckoInstruction &instCode )
std::string fregs = "";
for (int i=0; i<32; i++)
{
sprintf(freg, "f%02d: %08llx %08llx ", i, PowerPC::ppcState.ps[i][0], PowerPC::ppcState.ps[i][1]);
sprintf(freg, "f%02d: %08" PRIx64 " %08" PRIx64 " ", i, PowerPC::ppcState.ps[i][0], PowerPC::ppcState.ps[i][1]);
fregs.append(freg);
}

29
Source/Core/Core/Src/PowerPC/Interpreter/Interpreter_FPUtils.h
View file

@ -5,6 +5,7 @@
#ifndef _INTERPRETER_FPUTILS_H
#define _INTERPRETER_FPUTILS_H
#include "CPUDetect.h"
#include "Interpreter.h"
#include "MathUtil.h"
@ -69,28 +70,22 @@ inline void UpdateFPSCR()
inline double ForceSingle(double _x)
{
//if (FPSCR.RN != 0)
// PanicAlert("RN = %d at %x", (int)FPSCR.RN, PC);
if (FPSCR.NI)
_x = FlushToZeroAsFloat(_x);
double x = static_cast<float>(_x);
// convert to float...
float x = _x;
if (!cpu_info.bFlushToZero && FPSCR.NI)
{
x = FlushToZero(x);
}
// ...and back to double:
return x;
}
inline double ForceDouble(double d)
{
//if (FPSCR.RN != 0)
// PanicAlert("RN = %d at %x", (int)FPSCR.RN, PC);
//if (FPSCR.NI)
//{
// IntDouble x; x.d = d;
//if ((x.i & DOUBLE_EXP) == 0)
// x.i &= DOUBLE_SIGN; // turn into signed zero
// return x.d;
//}
if (!cpu_info.bFlushToZero && FPSCR.NI)
{
d = FlushToZero(d);
}
return d;
}

11
Source/Core/Core/Src/PowerPC/Interpreter/Interpreter_SystemRegisters.cpp
View file

@ -48,15 +48,8 @@ static void FPSCRtoFPUSettings(UReg_FPSCR fp)
// Pokemon Colosseum does this. Gah.
}
// Also corresponding SSE rounding mode setting
if (FPSCR.NI)
{
// Either one of these two breaks Beyond Good & Evil.
// if (cpu_info.bSSSE3)
// csr |= DAZ;
// csr |= FTZ;
}
FPURoundMode::SetSIMDMode(FPSCR.RN);
// Set SSE rounding mode and denormal handling
FPURoundMode::SetSIMDMode(FPSCR.RN, FPSCR.NI);
}
void Interpreter::mtfsb0x(UGeckoInstruction _inst)

4
Source/Core/Core/Src/PowerPC/Jit64/Jit.h
View file

@ -119,7 +119,7 @@ public:
void tri_op(int d, int a, int b, bool reversible, void (XEmitter::*op)(Gen::X64Reg, Gen::OpArg));
typedef u32 (*Operation)(u32 a, u32 b);
void regimmop(int d, int a, bool binary, u32 value, Operation doop, void (XEmitter::*op)(int, const Gen::OpArg&, const Gen::OpArg&), bool Rc = false, bool carry = false);
void fp_tri_op(int d, int a, int b, bool reversible, bool dupe, void (XEmitter::*op)(Gen::X64Reg, Gen::OpArg));
void fp_tri_op(int d, int a, int b, bool reversible, bool dupe, void (XEmitter::*op_2)(Gen::X64Reg, Gen::OpArg), void (XEmitter::*op_3)(Gen::X64Reg, Gen::X64Reg, Gen::OpArg));
// OPCODES
void unknown_instruction(UGeckoInstruction _inst);
@ -182,7 +182,7 @@ public:
void ps_sum(UGeckoInstruction inst);
void ps_muls(UGeckoInstruction inst);
void fp_arith_s(UGeckoInstruction inst);
void fp_arith(UGeckoInstruction inst);
void frsqrtex(UGeckoInstruction inst);
void fcmpx(UGeckoInstruction inst);

16
Source/Core/Core/Src/PowerPC/Jit64/Jit64_Tables.cpp
View file

@ -320,12 +320,12 @@ static GekkoOPTemplate table31_2[] =
static GekkoOPTemplate table59[] =
{
{18, &Jit64::Default}, //{"fdivsx", OPTYPE_FPU, FL_RC_BIT_F, 16}},
{20, &Jit64::fp_arith_s}, //"fsubsx", OPTYPE_FPU, FL_RC_BIT_F}},
{21, &Jit64::fp_arith_s}, //"faddsx", OPTYPE_FPU, FL_RC_BIT_F}},
{18, &Jit64::fp_arith}, //{"fdivsx", OPTYPE_FPU, FL_RC_BIT_F, 16}},
{20, &Jit64::fp_arith}, //"fsubsx", OPTYPE_FPU, FL_RC_BIT_F}},
{21, &Jit64::fp_arith}, //"faddsx", OPTYPE_FPU, FL_RC_BIT_F}},
// {22, &Jit64::Default}, //"fsqrtsx", OPTYPE_FPU, FL_RC_BIT_F}}, // Not implemented on gekko
{24, &Jit64::Default}, //"fresx", OPTYPE_FPU, FL_RC_BIT_F}},
{25, &Jit64::fp_arith_s}, //"fmulsx", OPTYPE_FPU, FL_RC_BIT_F}},
{25, &Jit64::fp_arith}, //"fmulsx", OPTYPE_FPU, FL_RC_BIT_F}},
{28, &Jit64::fmaddXX}, //"fmsubsx", OPTYPE_FPU, FL_RC_BIT_F}},
{29, &Jit64::fmaddXX}, //"fmaddsx", OPTYPE_FPU, FL_RC_BIT_F}},
{30, &Jit64::fmaddXX}, //"fnmsubsx", OPTYPE_FPU, FL_RC_BIT_F}},
@ -354,12 +354,12 @@ static GekkoOPTemplate table63[] =
static GekkoOPTemplate table63_2[] =
{
{18, &Jit64::Default}, //"fdivx", OPTYPE_FPU, FL_RC_BIT_F, 30}},
{20, &Jit64::Default}, //"fsubx", OPTYPE_FPU, FL_RC_BIT_F}},
{21, &Jit64::Default}, //"faddx", OPTYPE_FPU, FL_RC_BIT_F}},
{18, &Jit64::fp_arith}, //"fdivx", OPTYPE_FPU, FL_RC_BIT_F, 30}},
{20, &Jit64::fp_arith}, //"fsubx", OPTYPE_FPU, FL_RC_BIT_F}},
{21, &Jit64::fp_arith}, //"faddx", OPTYPE_FPU, FL_RC_BIT_F}},
{22, &Jit64::Default}, //"fsqrtx", OPTYPE_FPU, FL_RC_BIT_F}},
{23, &Jit64::Default}, //"fselx", OPTYPE_FPU, FL_RC_BIT_F}},
{25, &Jit64::fp_arith_s}, //"fmulx", OPTYPE_FPU, FL_RC_BIT_F}},
{25, &Jit64::fp_arith}, //"fmulx", OPTYPE_FPU, FL_RC_BIT_F}},
{26, &Jit64::frsqrtex}, //"frsqrtex", OPTYPE_FPU, FL_RC_BIT_F}},
{28, &Jit64::fmaddXX}, //"fmsubx", OPTYPE_FPU, FL_RC_BIT_F}},
{29, &Jit64::fmaddXX}, //"fmaddx", OPTYPE_FPU, FL_RC_BIT_F}},

7
Source/Core/Core/Src/PowerPC/Jit64/JitRegCache.cpp
View file

@ -166,7 +166,7 @@ int RegCache::SanityCheck() const
void RegCache::DiscardRegContentsIfCached(int preg)
{
if (regs[preg].away && regs[preg].location.IsSimpleReg())
if (IsBound(preg))
{
X64Reg xr = regs[preg].location.GetSimpleReg();
xregs[xr].free = true;
@ -351,11 +351,12 @@ void FPURegCache::StoreFromRegister(int i)
{
X64Reg xr = regs[i].location.GetSimpleReg();
_assert_msg_(DYNA_REC, xr < NUMXREGS, "WTF - store - invalid reg");
OpArg newLoc = GetDefaultLocation(i);
if (xregs[xr].dirty)
emit->MOVAPD(newLoc, xr);
xregs[xr].free = true;
xregs[xr].dirty = false;
xregs[xr].ppcReg = -1;
OpArg newLoc = GetDefaultLocation(i);
emit->MOVAPD(newLoc, xr);
regs[i].location = newLoc;
regs[i].away = false;
}

7
Source/Core/Core/Src/PowerPC/Jit64/JitRegCache.h
View file

@ -93,7 +93,7 @@ public:
const OpArg &R(int preg) const {return regs[preg].location;}
X64Reg RX(int preg) const
{
if (regs[preg].away && regs[preg].location.IsSimpleReg())
if (IsBound(preg))
return regs[preg].location.GetSimpleReg();
PanicAlert("Not so simple - %i", preg);
return (X64Reg)-1;
@ -111,6 +111,11 @@ public:
return xregs[xreg].free && !xlocks[xreg];
}
bool IsBound(int preg) const
{
return regs[preg].away && regs[preg].location.IsSimpleReg();
}
X64Reg GetFreeXReg();

120
Source/Core/Core/Src/PowerPC/Jit64/Jit_FloatingPoint.cpp
View file

@ -13,37 +13,62 @@ static const u64 GC_ALIGNED16(psAbsMask2[2]) = {0x7FFFFFFFFFFFFFFFULL, 0x7FFFFF
static const double GC_ALIGNED16(psOneOne2[2]) = {1.0, 1.0};
static const double one_const = 1.0f;
void Jit64::fp_tri_op(int d, int a, int b, bool reversible, bool dupe, void (XEmitter::*op)(Gen::X64Reg, Gen::OpArg))
void Jit64::fp_tri_op(int d, int a, int b, bool reversible, bool single,
void (XEmitter::*op_2)(Gen::X64Reg, Gen::OpArg),
void (XEmitter::*op_3)(Gen::X64Reg, Gen::X64Reg, Gen::OpArg))
{
if (!cpu_info.bAVX)
{
op_3 = nullptr;
}
fpr.Lock(d, a, b);
if (d == a)
{
fpr.BindToRegister(d, true);
(this->*op)(fpr.RX(d), fpr.R(b));
fpr.BindToRegister(d);
(this->*op_2)(fpr.RX(d), fpr.R(b));
}
else if (d == b)
{
if (reversible)
{
fpr.BindToRegister(d, true);
(this->*op)(fpr.RX(d), fpr.R(a));
fpr.BindToRegister(d);
(this->*op_2)(fpr.RX(d), fpr.R(a));
}
else
{
MOVSD(XMM0, fpr.R(b));
fpr.BindToRegister(d, !dupe);
MOVSD(fpr.RX(d), fpr.R(a));
(this->*op)(fpr.RX(d), Gen::R(XMM0));
if (op_3)
{
fpr.BindToRegister(d);
fpr.BindToRegister(a, true, false);
(this->*op_3)(fpr.RX(d), fpr.RX(a), fpr.R(b));
}
else
{
MOVSD(XMM0, fpr.R(b));
fpr.BindToRegister(d, false);
MOVSD(fpr.RX(d), fpr.R(a));
(this->*op_2)(fpr.RX(d), Gen::R(XMM0));
}
}
}
else
{
// Sources different from d, can use rather quick solution
fpr.BindToRegister(d, !dupe);
MOVSD(fpr.RX(d), fpr.R(a));
(this->*op)(fpr.RX(d), fpr.R(b));
if (op_3)
{
fpr.BindToRegister(d, false);
fpr.BindToRegister(a);
(this->*op_3)(fpr.RX(d), fpr.RX(a), fpr.R(b));
}
else
{
fpr.BindToRegister(d, false);
MOVSD(fpr.RX(d), fpr.R(a));
(this->*op_2)(fpr.RX(d), fpr.R(b));
}
}
if (dupe)
if (single)
{
ForceSinglePrecisionS(fpr.RX(d));
if (cpu_info.bSSE3)
@ -60,7 +85,7 @@ void Jit64::fp_tri_op(int d, int a, int b, bool reversible, bool dupe, void (XEm
fpr.UnlockAll();
}
void Jit64::fp_arith_s(UGeckoInstruction inst)
void Jit64::fp_arith(UGeckoInstruction inst)
{
INSTRUCTION_START
JITDISABLE(bJITFloatingPointOff)
@ -73,31 +98,38 @@ void Jit64::fp_arith_s(UGeckoInstruction inst)
Default(inst); return;
}
bool dupe = inst.OPCD == 59;
bool single = inst.OPCD == 59;
switch (inst.SUBOP5)
{
case 18: fp_tri_op(inst.FD, inst.FA, inst.FB, false, dupe, &XEmitter::DIVSD); break; //div
case 20: fp_tri_op(inst.FD, inst.FA, inst.FB, false, dupe, &XEmitter::SUBSD); break; //sub
case 21: fp_tri_op(inst.FD, inst.FA, inst.FB, true, dupe, &XEmitter::ADDSD); break; //add
case 25: fp_tri_op(inst.FD, inst.FA, inst.FC, true, dupe, &XEmitter::MULSD); break; //mul
case 18: fp_tri_op(inst.FD, inst.FA, inst.FB, false, single, &XEmitter::DIVSD, &XEmitter::VDIVSD); break; //div
case 20: fp_tri_op(inst.FD, inst.FA, inst.FB, false, single, &XEmitter::SUBSD, &XEmitter::VSUBSD); break; //sub
case 21: fp_tri_op(inst.FD, inst.FA, inst.FB, true, single, &XEmitter::ADDSD, &XEmitter::VADDSD); break; //add
case 25: fp_tri_op(inst.FD, inst.FA, inst.FC, true, single, &XEmitter::MULSD, &XEmitter::VMULSD); break; //mul
default:
_assert_msg_(DYNA_REC, 0, "fp_arith_s WTF!!!");
_assert_msg_(DYNA_REC, 0, "fp_arith WTF!!!");
}
}
void Jit64::frsqrtex(UGeckoInstruction inst)
{
INSTRUCTION_START
JITDISABLE(bJITFloatingPointOff)
int d = inst.FD;
int b = inst.FB;
fpr.Lock(b, d);
fpr.BindToRegister(d, true, true);
MOVSD(XMM0, M((void *)&one_const));
SQRTSD(XMM1, fpr.R(b));
DIVSD(XMM0, R(XMM1));
MOVSD(fpr.R(d), XMM0);
fpr.UnlockAll();
INSTRUCTION_START
JITDISABLE(bJITFloatingPointOff)
int d = inst.FD;
int b = inst.FB;
fpr.Lock(b, d);
fpr.BindToRegister(d, d == b, true);
MOVSD(XMM0, M((void *)&one_const));
SQRTSD(XMM1, fpr.R(b));
if (cpu_info.bAVX)
{
VDIVSD(fpr.RX(d), XMM0, R(XMM1));
}
else
{
DIVSD(XMM0, R(XMM1));
MOVSD(fpr.R(d), XMM0);
}
fpr.UnlockAll();
}
void Jit64::fmaddXX(UGeckoInstruction inst)
@ -192,16 +224,28 @@ void Jit64::fmrx(UGeckoInstruction inst)
{
INSTRUCTION_START
JITDISABLE(bJITFloatingPointOff)
if (inst.Rc) {
if (inst.Rc)
{
Default(inst); return;
}
int d = inst.FD;
int b = inst.FB;
fpr.Lock(b, d);
fpr.BindToRegister(d, true, true);
MOVSD(XMM0, fpr.R(b));
MOVSD(fpr.R(d), XMM0);
fpr.UnlockAll();
if (d != b)
{
fpr.Lock(b, d);
// we don't need to load d, but if it already is, it must be marked as dirty
if (fpr.IsBound(d))
{
fpr.BindToRegister(d);
}
fpr.BindToRegister(b, true, false);
// caveat: the order of ModRM:r/m, ModRM:reg is deliberate!
// "MOVSD reg, mem" zeros out the upper half of the destination register
MOVSD(fpr.R(d), fpr.RX(b));
fpr.UnlockAll();
}
}
void Jit64::fcmpx(UGeckoInstruction inst)

8
Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStorePaired.cpp
View file

@ -122,12 +122,6 @@ void Jit64::psq_l(UGeckoInstruction inst)
const UGQR gqr(rSPR(SPR_GQR0 + inst.I));
if (inst.W) {
// PanicAlert("Single ps load: %i %i", gqr.ST_TYPE, gqr.ST_SCALE);
Default(inst);
return;
}
bool update = inst.OPCD == 57;
int offset = inst.SIMM_12;
@ -143,6 +137,8 @@ void Jit64::psq_l(UGeckoInstruction inst)
MOV(32, gpr.R(inst.RA), R(ECX));
MOVZX(32, 16, EAX, M(((char *)&GQR(inst.I)) + 2));
MOVZX(32, 8, EDX, R(AL));
if (inst.W)
OR(32, R(EDX), Imm8(8));
#ifdef _M_IX86
int addr_scale = SCALE_4;
#else

12
Source/Core/Core/Src/PowerPC/Jit64/Jit_Paired.cpp
View file

@ -14,10 +14,9 @@
// cmppd, andpd, andnpd, or
// lfsx, ps_merge01 etc
const u64 GC_ALIGNED16(psSignBits[2]) = {0x8000000000000000ULL, 0x8000000000000000ULL};
const u64 GC_ALIGNED16(psAbsMask[2]) = {0x7FFFFFFFFFFFFFFFULL, 0x7FFFFFFFFFFFFFFFULL};
const double GC_ALIGNED16(psOneOne[2]) = {1.0, 1.0};
const double GC_ALIGNED16(psZeroZero[2]) = {0.0, 0.0};
static const u64 GC_ALIGNED16(psSignBits[2]) = {0x8000000000000000ULL, 0x8000000000000000ULL};
static const u64 GC_ALIGNED16(psAbsMask[2]) = {0x7FFFFFFFFFFFFFFFULL, 0x7FFFFFFFFFFFFFFFULL};
static const double GC_ALIGNED16(psOneOne[2]) = {1.0, 1.0};
void Jit64::ps_mr(UGeckoInstruction inst)
{
@ -52,14 +51,15 @@ void Jit64::ps_sel(UGeckoInstruction inst)
fpr.Lock(a, b, c, d);
MOVAPD(XMM0, fpr.R(a));
XORPD(XMM1, R(XMM1));
// XMM0 = XMM0 < 0 ? all 1s : all 0s
CMPPD(XMM0, M((void*)psZeroZero), LT);
CMPPD(XMM0, R(XMM1), LT);
MOVAPD(XMM1, R(XMM0));
ANDPD(XMM0, fpr.R(b));
ANDNPD(XMM1, fpr.R(c));
ORPD(XMM0, R(XMM1));
fpr.BindToRegister(d, false);
MOVAPD(fpr.RX(d), R(XMM0));
ORPD(fpr.RX(d), R(XMM1));
fpr.UnlockAll();
}

10
Source/Core/Core/Src/PowerPC/Jit64IL/IR_X86.cpp
View file

@ -470,7 +470,7 @@ static void regEmitMemLoad(RegInfo& RI, InstLoc I, unsigned Size) {
X64Reg reg;
auto info = regBuildMemAddress(RI, I, getOp1(I), 1, Size, &reg);
RI.Jit->SafeLoadToReg(reg, info.first, Size, info.second, regsInUse(RI), false);
RI.Jit->SafeLoadToReg(reg, info.first, Size, info.second, regsInUse(RI), false, EmuCodeBlock::SAFE_LOADSTORE_NO_FASTMEM);
if (regReadUse(RI, I))
RI.regs[reg] = I;
}
@ -498,7 +498,7 @@ static void regEmitMemStore(RegInfo& RI, InstLoc I, unsigned Size) {
} else {
RI.Jit->MOV(32, R(EAX), regLocForInst(RI, getOp1(I)));
}
RI.Jit->SafeWriteRegToReg(EAX, ECX, Size, 0, regsInUse(RI));
RI.Jit->SafeWriteRegToReg(EAX, ECX, Size, 0, regsInUse(RI), EmuCodeBlock::SAFE_LOADSTORE_NO_FASTMEM);
if (RI.IInfo[I - RI.FirstI] & 4)
regClearInst(RI, getOp1(I));
}
@ -1188,7 +1188,7 @@ static void DoWriteCode(IRBuilder* ibuild, JitIL* Jit, u32 exitAddress) {
Jit->MOV(32, R(EAX), loc1);
}
Jit->MOV(32, R(ECX), regLocForInst(RI, getOp2(I)));
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 0, regsInUse(RI));
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 0, regsInUse(RI), EmuCodeBlock::SAFE_LOADSTORE_NO_FASTMEM);
if (RI.IInfo[I - RI.FirstI] & 4)
fregClearInst(RI, getOp1(I));
if (RI.IInfo[I - RI.FirstI] & 8)
@ -1251,12 +1251,12 @@ static void DoWriteCode(IRBuilder* ibuild, JitIL* Jit, u32 exitAddress) {
Jit->PSRLQ(XMM0, 32);
Jit->MOVD_xmm(R(EAX), XMM0);
Jit->MOV(32, R(ECX), address);
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 0, regsInUse(RI));
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 0, regsInUse(RI), EmuCodeBlock::SAFE_LOADSTORE_NO_FASTMEM);
Jit->MOVAPD(XMM0, value);
Jit->MOVD_xmm(R(EAX), XMM0);
Jit->MOV(32, R(ECX), address);
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 4, regsInUse(RI));
RI.Jit->SafeWriteRegToReg(EAX, ECX, 32, 4, regsInUse(RI), EmuCodeBlock::SAFE_LOADSTORE_NO_FASTMEM);
Jit->SetJumpTarget(exit);
if (RI.IInfo[I - RI.FirstI] & 4)

8
Source/Core/Core/Src/PowerPC/Jit64IL/JitIL.cpp
View file

@ -3,6 +3,8 @@
// Refer to the license.txt file included.
#include <map>
#include <memory>
#include <cinttypes>
#include "Common.h"
#include "../../HLE/HLE.h"
@ -217,14 +219,14 @@ namespace JitILProfiler
const u64 totalElapsed = block.totalElapsed;
const u64 numberOfCalls = block.numberOfCalls;
const double elapsedPerCall = totalElapsed / (double)numberOfCalls;
fprintf(file.GetHandle(), "%016llx,%lld,%lld,%f\n", codeHash, totalElapsed, numberOfCalls, elapsedPerCall);
fprintf(file.GetHandle(), "%016" PRIx64 ",%" PRId64 ",%" PRId64 ",%f\n", codeHash, totalElapsed, numberOfCalls, elapsedPerCall);
}
}
};
std::auto_ptr<JitILProfilerFinalizer> finalizer;
std::unique_ptr<JitILProfilerFinalizer> finalizer;
static void Init()
{
finalizer = std::auto_ptr<JitILProfilerFinalizer>(new JitILProfilerFinalizer);
finalizer = std::unique_ptr<JitILProfilerFinalizer>(new JitILProfilerFinalizer);
}
static void Shutdown()
{

8
Source/Core/Core/Src/PowerPC/Jit64IL/JitIL.h
View file

@ -58,14 +58,10 @@ private:
JitBlockCache blocks;
TrampolineCache trampolines;
// The default code buffer. We keep it around to not have to alloc/dealloc a
// large chunk of memory for each recompiled block.
PPCAnalyst::CodeBuffer code_buffer;
public:
JitILAsmRoutineManager asm_routines;
JitIL() : code_buffer(32000) {}
JitIL() {}
~JitIL() {}
// Initialization, etc
@ -140,6 +136,4 @@ public:
void DynaRunTable63(UGeckoInstruction _inst) override;
};
void Jit(u32 em_address);
#endif // _JITIL_H

22
Source/Core/Core/Src/PowerPC/JitCommon/JitAsmCommon.cpp
View file

@ -196,7 +196,7 @@ void CommonAsmRoutines::GenQuantizedStores()
PACKSSDW(XMM0, R(XMM0));
PACKUSWB(XMM0, R(XMM0));
MOVD_xmm(R(EAX), XMM0);
SafeWriteRegToReg(AX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(AX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
@ -215,7 +215,7 @@ void CommonAsmRoutines::GenQuantizedStores()
PACKSSWB(XMM0, R(XMM0));
MOVD_xmm(R(EAX), XMM0);
SafeWriteRegToReg(AX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(AX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
@ -241,7 +241,7 @@ void CommonAsmRoutines::GenQuantizedStores()
MOV(16, R(AX), M((char*)psTemp + 4));
BSWAP(32, EAX);
SafeWriteRegToReg(EAX, ECX, 32, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 32, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
@ -261,7 +261,7 @@ void CommonAsmRoutines::GenQuantizedStores()
MOVD_xmm(R(EAX), XMM0);
BSWAP(32, EAX);
ROL(32, R(EAX), Imm8(16));
SafeWriteRegToReg(EAX, ECX, 32, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 32, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
@ -286,7 +286,7 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
// Easy!
const u8* storeSingleFloat = AlignCode4();
SafeWriteFloatToReg(XMM0, ECX, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteFloatToReg(XMM0, ECX, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
/*
if (cpu_info.bSSSE3) {
@ -294,11 +294,11 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
// TODO: SafeWriteFloat
MOVSS(M(&psTemp[0]), XMM0);
MOV(32, R(EAX), M(&psTemp[0]));
SafeWriteRegToReg(EAX, ECX, 32, 0, SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 32, 0, SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
} else {
MOVSS(M(&psTemp[0]), XMM0);
MOV(32, R(EAX), M(&psTemp[0]));
SafeWriteRegToReg(EAX, ECX, 32, 0, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 32, 0, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
}*/
const u8* storeSingleU8 = AlignCode4(); // Used by MKWii
@ -309,7 +309,7 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
MAXSS(XMM0, R(XMM1));
MINSS(XMM0, M((void *)&m_255));
CVTTSS2SI(EAX, R(XMM0));
SafeWriteRegToReg(AL, ECX, 8, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(AL, ECX, 8, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
const u8* storeSingleS8 = AlignCode4();
@ -319,7 +319,7 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
MAXSS(XMM0, M((void *)&m_m128));
MINSS(XMM0, M((void *)&m_127));
CVTTSS2SI(EAX, R(XMM0));
SafeWriteRegToReg(AL, ECX, 8, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(AL, ECX, 8, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
const u8* storeSingleU16 = AlignCode4(); // Used by MKWii
@ -330,7 +330,7 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
MAXSS(XMM0, R(XMM1));
MINSS(XMM0, M((void *)&m_65535));
CVTTSS2SI(EAX, R(XMM0));
SafeWriteRegToReg(EAX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
const u8* storeSingleS16 = AlignCode4();
@ -340,7 +340,7 @@ void CommonAsmRoutines::GenQuantizedSingleStores()
MAXSS(XMM0, M((void *)&m_m32768));
MINSS(XMM0, M((void *)&m_32767));
CVTTSS2SI(EAX, R(XMM0));
SafeWriteRegToReg(EAX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_WRITE_NO_PROLOG | SAFE_WRITE_NO_FASTMEM);
SafeWriteRegToReg(EAX, ECX, 16, 0, QUANTIZED_REGS_TO_SAVE, SAFE_LOADSTORE_NO_PROLOG | SAFE_LOADSTORE_NO_FASTMEM);
RET();
singleStoreQuantized = reinterpret_cast<const u8**>(const_cast<u8*>(AlignCode16()));

5
Source/Core/Core/Src/PowerPC/JitCommon/JitBackpatch.cpp
View file

@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <string>
#include <cinttypes>
#include "Common.h"
#include "disasm.h"
@ -32,7 +33,7 @@ static void BackPatchError(const std::string &text, u8 *codePtr, u32 emAddress)
#endif
PanicAlert("%s\n\n"
"Error encountered accessing emulated address %08x.\n"
"Culprit instruction: \n%s\nat %#llx",
"Culprit instruction: \n%s\nat %#" PRIx64,
text.c_str(), emAddress, disbuf, code_addr);
return;
}
@ -233,7 +234,7 @@ const u8 *Jitx86Base::BackPatch(u8 *codePtr, u32 emAddress, void *ctx_void)
XEmitter emitter(start);
const u8 *trampoline = trampolines.GetWriteTrampoline(info, registersInUse);
emitter.CALL((void *)trampoline);
emitter.NOP(codePtr + info.instructionSize - emitter.GetCodePtr());
emitter.NOP((int)(codePtr + info.instructionSize - emitter.GetCodePtr()));
return start;
}
#else

18
Source/Core/Core/Src/PowerPC/JitCommon/Jit_Util.cpp
View file

@ -117,18 +117,20 @@ u8 *EmuCodeBlock::UnsafeLoadToReg(X64Reg reg_value, Gen::OpArg opAddress, int ac
return result;
}
void EmuCodeBlock::SafeLoadToReg(X64Reg reg_value, const Gen::OpArg & opAddress, int accessSize, s32 offset, u32 registersInUse, bool signExtend)
void EmuCodeBlock::SafeLoadToReg(X64Reg reg_value, const Gen::OpArg & opAddress, int accessSize, s32 offset, u32 registersInUse, bool signExtend, int flags)
{
if (!jit->js.memcheck)
{
registersInUse &= ~(1 << RAX | 1 << reg_value);
}
#if defined(_M_X64)
if (!Core::g_CoreStartupParameter.bMMU &&
Core::g_CoreStartupParameter.bFastmem &&
!(flags & (SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_FASTMEM))
#ifdef ENABLE_MEM_CHECK
if (!Core::g_CoreStartupParameter.bMMU && !Core::g_CoreStartupParameter.bEnableDebugging && Core::g_CoreStartupParameter.bFastmem)
#else
if (!Core::g_CoreStartupParameter.bMMU && Core::g_CoreStartupParameter.bFastmem)
&& !Core::g_CoreStartupParameter.bEnableDebugging
#endif
)
{
u8 *mov = UnsafeLoadToReg(reg_value, opAddress, accessSize, offset, signExtend);
@ -282,14 +284,14 @@ void EmuCodeBlock::SafeWriteRegToReg(X64Reg reg_value, X64Reg reg_addr, int acce
#if defined(_M_X64)
if (!Core::g_CoreStartupParameter.bMMU &&
Core::g_CoreStartupParameter.bFastmem &&
!(flags & (SAFE_WRITE_NO_SWAP | SAFE_WRITE_NO_FASTMEM))
!(flags & (SAFE_LOADSTORE_NO_SWAP | SAFE_LOADSTORE_NO_FASTMEM))
#ifdef ENABLE_MEM_CHECK
&& !Core::g_CoreStartupParameter.bEnableDebugging
#endif
)
{
MOV(32, M(&PC), Imm32(jit->js.compilerPC)); // Helps external systems know which instruction triggered the write
u8 *mov = UnsafeWriteRegToReg(reg_value, reg_addr, accessSize, offset, !(flags & SAFE_WRITE_NO_SWAP));
u8 *mov = UnsafeWriteRegToReg(reg_value, reg_addr, accessSize, offset, !(flags & SAFE_LOADSTORE_NO_SWAP));
if (accessSize == 8)
{
NOP(1);
@ -321,8 +323,8 @@ void EmuCodeBlock::SafeWriteRegToReg(X64Reg reg_value, X64Reg reg_addr, int acce
MOV(32, M(&PC), Imm32(jit->js.compilerPC)); // Helps external systems know which instruction triggered the write
TEST(32, R(reg_addr), Imm32(mem_mask));
FixupBranch fast = J_CC(CC_Z, true);
bool noProlog = flags & SAFE_WRITE_NO_PROLOG;
bool swap = !(flags & SAFE_WRITE_NO_SWAP);
bool noProlog = (0 != (flags & SAFE_LOADSTORE_NO_PROLOG));
bool swap = !(flags & SAFE_LOADSTORE_NO_SWAP);
ABI_PushRegistersAndAdjustStack(registersInUse, noProlog);
switch (accessSize)
{

10
Source/Core/Core/Src/PowerPC/JitCommon/Jit_Util.h
View file

@ -28,13 +28,13 @@ public:
// these return the address of the MOV, for backpatching
u8 *UnsafeWriteRegToReg(Gen::X64Reg reg_value, Gen::X64Reg reg_addr, int accessSize, s32 offset = 0, bool swap = true);
u8 *UnsafeLoadToReg(Gen::X64Reg reg_value, Gen::OpArg opAddress, int accessSize, s32 offset, bool signExtend);
void SafeLoadToReg(Gen::X64Reg reg_value, const Gen::OpArg & opAddress, int accessSize, s32 offset, u32 registersInUse, bool signExtend);
enum SafeWriteFlags
enum SafeLoadStoreFlags
{
SAFE_WRITE_NO_SWAP = 1,
SAFE_WRITE_NO_PROLOG = 2,
SAFE_WRITE_NO_FASTMEM = 4
SAFE_LOADSTORE_NO_SWAP = 1,
SAFE_LOADSTORE_NO_PROLOG = 2,
SAFE_LOADSTORE_NO_FASTMEM = 4
};
void SafeLoadToReg(Gen::X64Reg reg_value, const Gen::OpArg & opAddress, int accessSize, s32 offset, u32 registersInUse, bool signExtend, int flags = 0);
void SafeWriteRegToReg(Gen::X64Reg reg_value, Gen::X64Reg reg_addr, int accessSize, s32 offset, u32 registersInUse, int flags = 0);
// Trashes both inputs and EAX.

7
Source/Core/Core/Src/PowerPC/JitILCommon/IR.cpp
View file

@ -118,6 +118,7 @@ Fix profiled loads/stores to work safely. On 32-bit, one solution is to
#include <algorithm>
#include <memory>
#include <cinttypes>
#include <ctime>
#include <set>
#include "IR.h"
@ -1223,7 +1224,7 @@ struct Writer
virtual ~Writer() {}
};
static std::auto_ptr<Writer> writer;
static std::unique_ptr<Writer> writer;
static const std::string opcodeNames[] = {
"Nop", "LoadGReg", "LoadLink", "LoadCR", "LoadCarry", "LoadCTR",
@ -1275,11 +1276,11 @@ void IRBuilder::WriteToFile(u64 codeHash) {
_assert_(sizeof(opcodeNames) / sizeof(opcodeNames[0]) == Int3 + 1);
if (!writer.get()) {
writer = std::auto_ptr<Writer>(new Writer);
writer = std::unique_ptr<Writer>(new Writer);
}
FILE* const file = writer->file.GetHandle();
fprintf(file, "\ncode hash:%016llx\n", codeHash);
fprintf(file, "\ncode hash:%016" PRIx64 "\n", codeHash);
const InstLoc lastCurReadPtr = curReadPtr;
StartForwardPass();

5
Source/Core/Core/Src/PowerPC/JitInterface.cpp
View file

@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <algorithm>
#include <cinttypes>
#ifdef _WIN32
#include <windows.h>
@ -171,12 +172,12 @@ namespace JitInterface
double percent = 100.0 * (double)stat.cost / (double)cost_sum;
#ifdef _WIN32
double timePercent = 100.0 * (double)block->ticCounter / (double)timecost_sum;
fprintf(f.GetHandle(), "%08x\t%s\t%llu\t%llu\t%.2lf\t%llf\t%lf\t%i\n",
fprintf(f.GetHandle(), "%08x\t%s\t%" PRIu64 "\t%" PRIu64 "\t%.2lf\t%llf\t%lf\t%i\n",
block->originalAddress, name.c_str(), stat.cost,
block->ticCounter, percent, timePercent,
(double)block->ticCounter*1000.0/(double)countsPerSec, block->codeSize);
#else
fprintf(f.GetHandle(), "%08x\t%s\t%llu\t???\t%.2lf\t???\t???\t%i\n",
fprintf(f.GetHandle(), "%08x\t%s\t%" PRIu64 "\t???\t%.2lf\t???\t???\t%i\n",
block->originalAddress, name.c_str(), stat.cost, percent, block->codeSize);
#endif
}

70
Source/Core/Core/Src/PowerPC/PPCTables.cpp
View file

@ -4,6 +4,7 @@
#include <algorithm>
#include <vector>
#include <cinttypes>
#include "Common.h"
#include "PPCTables.h"
@ -13,24 +14,15 @@
#include "Interpreter/Interpreter_Tables.h"
#include "JitInterface.h"
struct op_inf
{
const char *name;
int count;
bool operator < (const op_inf &o) const
{
return count > o.count;
}
};
GekkoOPInfo *m_infoTable[64];
GekkoOPInfo *m_infoTable4[1024];
GekkoOPInfo *m_infoTable19[1024];
GekkoOPInfo *m_infoTable31[1024];
GekkoOPInfo *m_infoTable59[32];
GekkoOPInfo *m_infoTable63[1024];
GekkoOPInfo *m_infoTable[64];
GekkoOPInfo *m_infoTable4[1024];
GekkoOPInfo *m_infoTable19[1024];
GekkoOPInfo *m_infoTable31[1024];
GekkoOPInfo *m_infoTable59[32];
GekkoOPInfo *m_infoTable63[1024];
GekkoOPInfo *m_allInstructions[512];
int m_numInstructions;
GekkoOPInfo *m_allInstructions[512];
int m_numInstructions;
GekkoOPInfo *GetOpInfo(UGeckoInstruction _inst)
{
@ -181,26 +173,34 @@ void CountInstruction(UGeckoInstruction _inst)
{
GekkoOPInfo *info = GetOpInfo(_inst);
if (info)
{
info->runCount++;
}
}
void PrintInstructionRunCounts()
{
std::vector<op_inf> temp;
for (int i = 0; i < m_numInstructions; i++)
typedef std::pair<const char*, u64> OpInfo;
std::vector<OpInfo> temp;
temp.reserve(m_numInstructions);
for (int i = 0; i < m_numInstructions; ++i)
{
op_inf x;
x.name = m_allInstructions[i]->opname;
x.count = m_allInstructions[i]->runCount;
temp.push_back(x);
GekkoOPInfo *pInst = m_allInstructions[i];
temp.emplace_back(pInst->opname, pInst->runCount);
}
std::sort(temp.begin(), temp.end());
for (int i = 0; i < m_numInstructions; i++)
std::sort(temp.begin(), temp.end(),
[](const OpInfo &a, const OpInfo &b)
{
return a.second > b.second;
});
for (auto &inst : temp)
{
if (temp[i].count == 0)
if (inst.second == 0)
break;
DEBUG_LOG(POWERPC, "%s : %i", temp[i].name,temp[i].count);
//PanicAlert("%s : %i", temp[i].name,temp[i].count);
DEBUG_LOG(POWERPC, "%s : %llu", inst.first, inst.second);
//PanicAlert("%s : %llu", inst.first, inst.second);
}
}
@ -211,20 +211,22 @@ void LogCompiledInstructions()
File::IOFile f(StringFromFormat("%sinst_log%i.txt", File::GetUserPath(D_LOGS_IDX).c_str(), time), "w");
for (int i = 0; i < m_numInstructions; i++)
{
if (m_allInstructions[i]->compileCount > 0)
GekkoOPInfo *pInst = m_allInstructions[i];
if (pInst->compileCount > 0)
{
fprintf(f.GetHandle(), "%s\t%i\t%lld\t%08x\n", m_allInstructions[i]->opname,
m_allInstructions[i]->compileCount, m_allInstructions[i]->runCount, m_allInstructions[i]->lastUse);
fprintf(f.GetHandle(), "%s\t%i\t%" PRId64 "\t%08x\n", pInst->opname,
pInst->compileCount, pInst->runCount, pInst->lastUse);
}
}
f.Open(StringFromFormat("%sinst_not%i.txt", File::GetUserPath(D_LOGS_IDX).c_str(), time), "w");
for (int i = 0; i < m_numInstructions; i++)
{
if (m_allInstructions[i]->compileCount == 0)
GekkoOPInfo *pInst = m_allInstructions[i];
if (pInst->compileCount == 0)
{
fprintf(f.GetHandle(), "%s\t%i\t%lld\n", m_allInstructions[i]->opname,
m_allInstructions[i]->compileCount, m_allInstructions[i]->runCount);
fprintf(f.GetHandle(), "%s\t%i\t%" PRId64 "\n", pInst->opname,
pInst->compileCount, pInst->runCount);
}
}

8
Source/Core/Core/Src/State.cpp
View file

@ -247,7 +247,7 @@ void CompressAndDumpState(CompressAndDumpState_args save_args)
// Setting up the header
StateHeader header;
memcpy(header.gameID, SConfig::GetInstance().m_LocalCoreStartupParameter.GetUniqueID().c_str(), 6);
header.size = g_use_compression ? buffer_size : 0;
header.size = g_use_compression ? (u32)buffer_size : 0;
header.time = Common::Timer::GetDoubleTime();
f.WriteArray(&header, 1);
@ -261,9 +261,13 @@ void CompressAndDumpState(CompressAndDumpState_args save_args)
lzo_uint out_len = 0;
if ((i + IN_LEN) >= buffer_size)
cur_len = buffer_size - i;
{
cur_len = (lzo_uint32)(buffer_size - i);
}
else
{
cur_len = IN_LEN;
}
if (lzo1x_1_compress(buffer_data + i, cur_len, out, &out_len, wrkmem) != LZO_E_OK)
PanicAlertT("Internal LZO Error - compression failed");

6
Source/Core/DiscIO/Src/CompressedBlob.cpp
View file

@ -9,6 +9,8 @@
#include <unistd.h>
#endif
#include <cinttypes>
#include "CompressedBlob.h"
#include "DiscScrubber.h"
#include "FileUtil.h"
@ -99,7 +101,7 @@ void CompressedBlobReader::GetBlock(u64 block_num, u8 *out_ptr)
// First, check hash.
u32 block_hash = HashAdler32(source, comp_block_size);
if (block_hash != hashes[block_num])
PanicAlert("Hash of block %lli is %08x instead of %08x.\n"
PanicAlert("Hash of block %" PRIu64 " is %08x instead of %08x.\n"
"Your ISO, %s, is corrupt.",
block_num, block_hash, hashes[block_num],
file_name.c_str());
@ -127,7 +129,7 @@ void CompressedBlobReader::GetBlock(u64 block_num, u8 *out_ptr)
{
// this seem to fire wrongly from time to time
// to be sure, don't use compressed isos :P
PanicAlert("Failure reading block %lli - out of data and not at end.", block_num);
PanicAlert("Failure reading block %" PRIu64 " - out of data and not at end.", block_num);
}
inflateEnd(&z);
if (uncomp_size != header.block_size)

8
Source/Core/DiscIO/Src/DiscScrubber.cpp
View file

@ -7,6 +7,8 @@
#include "FileUtil.h"
#include "DiscScrubber.h"
#include <cinttypes>
namespace DiscIO
{
@ -121,13 +123,13 @@ void GetNextBlock(File::IOFile& in, u8* buffer)
if (m_isScrubbing && m_FreeTable[i])
{
DEBUG_LOG(DISCIO, "Freeing 0x%016llx", CurrentOffset);
DEBUG_LOG(DISCIO, "Freeing 0x%016" PRIx64, CurrentOffset);
std::fill(buffer, buffer + m_BlockSize, 0xFF);
in.Seek(m_BlockSize, SEEK_CUR);
}
else
{
DEBUG_LOG(DISCIO, "Used 0x%016llx", CurrentOffset);
DEBUG_LOG(DISCIO, "Used 0x%016" PRIx64, CurrentOffset);
in.ReadBytes(buffer, m_BlockSize);
}
@ -150,7 +152,7 @@ void MarkAsUsed(u64 _Offset, u64 _Size)
u64 CurrentOffset = _Offset;
u64 EndOffset = CurrentOffset + _Size;
DEBUG_LOG(DISCIO, "Marking 0x%016llx - 0x%016llx as used", _Offset, EndOffset);
DEBUG_LOG(DISCIO, "Marking 0x%016" PRIx64 " - 0x%016" PRIx64 " as used", _Offset, EndOffset);
while ((CurrentOffset < EndOffset) && (CurrentOffset < m_FileSize))
{

3
Source/Core/DiscIO/Src/FileSystemGCWii.cpp
View file

@ -8,6 +8,7 @@
#include <string>
#include <vector>
#include <algorithm>
#include <cinttypes>
#include "FileSystemGCWii.h"
#include "StringUtil.h"
@ -70,7 +71,7 @@ u64 CFileSystemGCWii::ReadFile(const char* _rFullPath, u8* _pBuffer, size_t _Max
if (pFileInfo->m_FileSize > _MaxBufferSize)
return 0;
DEBUG_LOG(DISCIO, "Filename: %s. Offset: %llx. Size: %llx",_rFullPath,
DEBUG_LOG(DISCIO, "Filename: %s. Offset: %" PRIx64 ". Size: %" PRIx64, _rFullPath,
pFileInfo->m_Offset, pFileInfo->m_FileSize);
m_rVolume->Read(pFileInfo->m_Offset, pFileInfo->m_FileSize, _pBuffer);

12
Source/Core/DiscIO/Src/VolumeCreator.cpp
View file

@ -183,8 +183,18 @@ static IVolume* CreateVolumeFromCryptedWiiImage(IBlobReader& _rReader, u32 _Part
memset(IV, 0, 16);
_rReader.Read(rPartition.Offset + 0x44c, 8, IV);
bool usingKoreanKey = false;
// Issue: 6813
// Magic value is at 0x501f1 (1byte)
// If encrypted with the Korean key, the magic value would be 1
// Otherwise it is zero
if (Korean && Reader.Read32(0x501ee) != 0)
{
usingKoreanKey = true;
}
aes_context AES_ctx;
aes_setkey_dec(&AES_ctx, (Korean ? g_MasterKeyK : g_MasterKey), 128);
aes_setkey_dec(&AES_ctx, (usingKoreanKey ? g_MasterKeyK : g_MasterKey), 128);
u8 VolumeKey[16];
aes_crypt_cbc(&AES_ctx, AES_DECRYPT, 16, IV, SubKey, VolumeKey);

3
Source/Core/DolphinWX/Src/GameListCtrl.cpp
View file

@ -9,6 +9,7 @@
#include <wx/filename.h>
#include <algorithm>
#include <cinttypes>
#include <memory>
#include "FileSearch.h"
@ -383,7 +384,7 @@ wxString NiceSizeFormat(u64 _size)
auto const value = (_size + unit_size / 2) / unit_size;
auto const frac = (_size % unit_size * 10 + unit_size / 2) / unit_size % 10;
return StrToWxStr(StringFromFormat("%llu.%llu %s", value, frac, unit_symbols[unit]));
return StrToWxStr(StringFromFormat("%" PRIu64 ".%" PRIu64 " %s", value, frac, unit_symbols[unit]));
}
void CGameListCtrl::InsertItemInReportView(long _Index)

2
Source/Core/DolphinWX/Src/ISOFile.cpp
View file

@ -188,7 +188,7 @@ std::string GameListItem::CreateCacheFilename()
// Filename.extension_HashOfFolderPath_Size.cache
// Append hash to prevent ISO name-clashing in different folders.
Filename.append(StringFromFormat("%s_%x_%llx.cache",
Filename.append(StringFromFormat("%s_%x_%zx.cache",
extension.c_str(), HashFletcher((const u8 *)LegalPathname.c_str(), LegalPathname.size()),
File::GetSize(m_FileName)));

3
Source/Core/DolphinWX/Src/ISOProperties.cpp
View file

@ -7,6 +7,7 @@
#endif
#include <type_traits>
#include <cinttypes>
#include "Common.h"
#include "CommonPaths.h"
@ -118,7 +119,7 @@ CISOProperties::CISOProperties(const std::string fileName, wxWindow* parent, wxW
u8 _tTitleID[8];
if(OpenISO->GetTitleID(_tTitleID))
{
snprintf(tmp, 17, "%016llx", Common::swap64(_tTitleID));
snprintf(tmp, 17, "%016" PRIx64, Common::swap64(_tTitleID));
_iniFilename = tmp;
}
}

12
Source/Core/DolphinWX/Src/MainAndroid.cpp
View file

@ -278,10 +278,22 @@ JNIEXPORT jstring JNICALL Java_org_dolphinemu_dolphinemu_NativeLibrary_GetTitle(
env->ReleaseStringUTFChars(jFile, File);
return env->NewStringUTF(Name.c_str());
}
JNIEXPORT jstring JNICALL Java_org_dolphinemu_dolphinemu_NativeLibrary_GetVersionString(JNIEnv *env, jobject obj)
{
return env->NewStringUTF(scm_rev_str);
}
JNIEXPORT jboolean JNICALL Java_org_dolphinemu_dolphinemu_NativeLibrary_SupportsNEON(JNIEnv *env, jobject obj)
{
return cpu_info.bNEON;
}
JNIEXPORT void JNICALL Java_org_dolphinemu_dolphinemu_NativeLibrary_SaveScreenShot(JNIEnv *env, jobject obj)
{
Core::SaveScreenShot();
}
JNIEXPORT jstring JNICALL Java_org_dolphinemu_dolphinemu_NativeLibrary_GetConfig(JNIEnv *env, jobject obj, jstring jFile, jstring jKey, jstring jValue, jstring jDefault)
{
IniFile ini;

8
Source/Core/DolphinWX/Src/MemoryCards/WiiSaveCrypted.cpp
View file

@ -7,12 +7,14 @@
// Licensed under the terms of the GNU GPL, version 2
// http://www.gnu.org/licenses/old-licenses/gpl-2.0.txt
#include <algorithm>
#include <cinttypes>
#include "WiiSaveCrypted.h"
#include "FileUtil.h"
#include "MathUtil.h"
#include "NandPaths.h"
#include "FileUtil.h"
#include <algorithm>
static Common::replace_v replacements;
@ -152,7 +154,7 @@ void CWiiSaveCrypted::ReadHDR()
md5((u8*)&_header, HEADER_SZ, md5_calc);
if (memcmp(md5_file, md5_calc, 0x10))
{
PanicAlertT("MD5 mismatch\n %016llx%016llx != %016llx%016llx", Common::swap64(md5_file),Common::swap64(md5_file+8), Common::swap64(md5_calc), Common::swap64(md5_calc+8));
PanicAlertT("MD5 mismatch\n %016" PRIx64 "%016" PRIx64 " != %016" PRIx64 "%016" PRIx64, Common::swap64(md5_file),Common::swap64(md5_file+8), Common::swap64(md5_calc), Common::swap64(md5_calc+8));
b_valid= false;
}
@ -244,7 +246,7 @@ void CWiiSaveCrypted::ReadBKHDR()
if (_sizeOfFiles + FULL_CERT_SZ != _totalSize)
WARN_LOG(CONSOLE, "Size(%x) + cert(%x) does not equal totalsize(%x)", _sizeOfFiles, FULL_CERT_SZ, _totalSize);
if (m_TitleID != Common::swap64(bkhdr.SaveGameTitle))
WARN_LOG(CONSOLE, "Encrypted title (%llx) does not match unencrypted title (%llx)", m_TitleID, Common::swap64(bkhdr.SaveGameTitle));
WARN_LOG(CONSOLE, "Encrypted title (%" PRIx64 ") does not match unencrypted title (%" PRIx64 ")", m_TitleID, Common::swap64(bkhdr.SaveGameTitle));
}
void CWiiSaveCrypted::WriteBKHDR()

11
Source/Core/DolphinWX/Src/WiimoteConfigDiag.cpp
View file

@ -3,6 +3,7 @@
#include "HW/Wiimote.h"
#include "HW/WiimoteReal/WiimoteReal.h"
#include "Frame.h"
#include "NetPlayProto.h"
WiimoteConfigDiag::WiimoteConfigDiag(wxWindow* const parent, InputPlugin& plugin)
: wxDialog(parent, -1, _("Dolphin Wiimote Configuration"), wxDefaultPosition, wxDefaultSize)
@ -134,6 +135,16 @@ WiimoteConfigDiag::WiimoteConfigDiag(wxWindow* const parent, InputPlugin& plugin
WiimoteSpkVolumeText->Disable();
WiimoteSpkVolumeMinText->Disable();
WiimoteSpkVolumeMaxText->Disable();
if (NetPlay::IsNetPlayRunning())
{
bb_source->Disable();
for (int i = 0; i < 4; ++i)
{
wiimote_label[i]->Disable();
wiimote_source_ch[i]->Disable();
}
}
}

52
Source/Core/InputCommon/Src/ControllerInterface/SDL/SDL.cpp
View file

@ -154,12 +154,13 @@ Joystick::~Joystick()
{
// stop/destroy all effects
SDL_HapticStopAll(m_haptic);
std::list<EffectIDState>::iterator
i = m_state_out.begin(),
e = m_state_out.end();
for ( ; i != e; ++i)
if (i->id != -1)
SDL_HapticDestroyEffect(m_haptic, i->id);
for (auto &i : m_state_out)
{
if (i.id != -1)
{
SDL_HapticDestroyEffect(m_haptic, i.id);
}
}
// close haptic first
SDL_HapticClose(m_haptic);
}
@ -210,7 +211,7 @@ void Joystick::ConstantEffect::SetState(ControlState state)
}
const Sint16 old = m_effect.effect.constant.level;
m_effect.effect.constant.level = state * 0x7FFF;
m_effect.effect.constant.level = (Sint16)(state * 0x7FFF);
if (old != m_effect.effect.constant.level)
m_effect.changed = true;
}
@ -228,7 +229,7 @@ void Joystick::RampEffect::SetState(ControlState state)
}
const Sint16 old = m_effect.effect.ramp.start;
m_effect.effect.ramp.start = state * 0x7FFF;
m_effect.effect.ramp.start = (Sint16)(state * 0x7FFF);
if (old != m_effect.effect.ramp.start)
m_effect.changed = true;
}
@ -247,7 +248,7 @@ void Joystick::SineEffect::SetState(ControlState state)
const Sint16 old = m_effect.effect.periodic.magnitude;
m_effect.effect.periodic.period = 5;
m_effect.effect.periodic.magnitude = state * 0x5000;
m_effect.effect.periodic.magnitude = (Sint16)(state * 0x5000);
m_effect.effect.periodic.attack_length = 0;
m_effect.effect.periodic.fade_length = 500;
@ -293,7 +294,7 @@ void Joystick::TriangleEffect::SetState(ControlState state)
const Sint16 old = m_effect.effect.periodic.magnitude;
m_effect.effect.periodic.period = 5;
m_effect.effect.periodic.magnitude = state * 0x5000;
m_effect.effect.periodic.magnitude = (Sint16)(state * 0x5000);
m_effect.effect.periodic.attack_length = 0;
m_effect.effect.periodic.fade_length = 100;
@ -313,34 +314,35 @@ bool Joystick::UpdateInput()
bool Joystick::UpdateOutput()
{
#ifdef USE_SDL_HAPTIC
std::list<EffectIDState>::iterator
i = m_state_out.begin(),
e = m_state_out.end();
for ( ; i != e; ++i)
for (auto &i : m_state_out)
{
if (i->changed) // if SetState was called on this output
if (i.changed) // if SetState was called on this output
{
if (-1 == i->id) // effect isn't currently uploaded
if (-1 == i.id) // effect isn't currently uploaded
{
if (i->effect.type) // if outputstate is >0 this would be true
if ((i->id = SDL_HapticNewEffect( m_haptic, &i->effect )) > -1) // upload the effect
SDL_HapticRunEffect(m_haptic, i->id, 1); // run the effect
if (i.effect.type) // if outputstate is >0 this would be true
{
if ((i.id = SDL_HapticNewEffect(m_haptic, &i.effect)) > -1) // upload the effect
{
SDL_HapticRunEffect(m_haptic, i.id, 1); // run the effect
}
}
}
else // effect is already uploaded
{
if (i->effect.type) // if ouputstate >0
if (i.effect.type) // if ouputstate >0
{
SDL_HapticUpdateEffect(m_haptic, i->id, &i->effect); // update the effect
SDL_HapticUpdateEffect(m_haptic, i.id, &i.effect); // update the effect
}
else
{
SDL_HapticStopEffect(m_haptic, i->id); // else, stop and remove the effect
SDL_HapticDestroyEffect(m_haptic, i->id);
i->id = -1; // mark it as not uploaded
SDL_HapticStopEffect(m_haptic, i.id); // else, stop and remove the effect
SDL_HapticDestroyEffect(m_haptic, i.id);
i.id = -1; // mark it as not uploaded
}
}
i->changed = false;
i.changed = false;
}
}
#endif

4
Source/Core/VideoBackends/D3D/Src/PerfQuery.cpp
View file

@ -116,7 +116,7 @@ void PerfQuery::FlushOne()
}
// NOTE: Reported pixel metrics should be referenced to native resolution
m_results[entry.query_type] += (u64)result * EFB_WIDTH / g_renderer->GetTargetWidth() * EFB_HEIGHT / g_renderer->GetTargetHeight();
m_results[entry.query_type] += (u32)(result * EFB_WIDTH / g_renderer->GetTargetWidth() * EFB_HEIGHT / g_renderer->GetTargetHeight());
m_query_read_pos = (m_query_read_pos + 1) % ArraySize(m_query_buffer);
--m_query_count;
@ -147,7 +147,7 @@ void PerfQuery::WeakFlush()
if (hr == S_OK)
{
// NOTE: Reported pixel metrics should be referenced to native resolution
m_results[entry.query_type] += (u64)result * EFB_WIDTH / g_renderer->GetTargetWidth() * EFB_HEIGHT / g_renderer->GetTargetHeight();
m_results[entry.query_type] += (u32)(result * EFB_WIDTH / g_renderer->GetTargetWidth() * EFB_HEIGHT / g_renderer->GetTargetHeight());
m_query_read_pos = (m_query_read_pos + 1) % ArraySize(m_query_buffer);
--m_query_count;

41
Source/Core/VideoBackends/D3D/Src/Render.cpp
View file

@ -2,7 +2,8 @@
// Licensed under GPLv2
// Refer to the license.txt file included.
#include <math.h>
#include <cinttypes>
#include <cmath>
#include "Timer.h"
@ -33,6 +34,7 @@
#include "FPSCounter.h"
#include "ConfigManager.h"
#include <strsafe.h>
#include "ImageWrite.h"
namespace DX11
{
@ -680,7 +682,7 @@ void Renderer::SetBlendMode(bool forceUpdate)
}
}
void Renderer::TakeScreenshot(const TargetRectangle &rc, std::string filename)
bool Renderer::SaveScreenshot(const std::string &filename, const TargetRectangle& rc)
{
if (!s_screenshot_texture)
CreateScreenshotTexture(rc);
@ -689,26 +691,25 @@ void Renderer::TakeScreenshot(const TargetRectangle &rc, std::string filename)
D3D11_BOX box = CD3D11_BOX(rc.left, rc.top, 0, rc.right, rc.bottom, 1);
D3D::context->CopySubresourceRegion(s_screenshot_texture, 0, 0, 0, 0, (ID3D11Resource*)D3D::GetBackBuffer()->GetTex(), 0, &box);
u8* __restrict dest = (u8*) malloc(rc.GetWidth() * rc.GetHeight() * 3);
D3D11_MAPPED_SUBRESOURCE map;
D3D::context->Map(s_screenshot_texture, 0, D3D11_MAP_READ_WRITE, 0, &map);
u8* src = (u8*) map.pData;
for (int y = 0; y < rc.GetHeight(); ++y)
{
u8* __restrict row = src;
for (int x = 0; x < rc.GetWidth(); ++x)
{
*dest++ = *row++;
*dest++ = *row++;
*dest++ = *row++;
row++;
}
src += map.RowPitch;
}
bool saved_png = TextureToPng((u8*)map.pData, map.RowPitch, filename, rc.GetWidth(), rc.GetHeight(), false);
D3D::context->Unmap(s_screenshot_texture, 0);
SaveScreenshot(dest, rc.GetWidth(), rc.GetHeight(), filename);
if (saved_png)
{
OSD::AddMessage(StringFromFormat("Saved %i x %i %s", rc.GetWidth(),
rc.GetHeight(), filename.c_str()));
}
else
{
OSD::AddMessage(StringFromFormat("Error saving %s", filename.c_str()));
}
return saved_png;
}
void formatBufferDump(const u8* in, u8* out, int w, int h, int p)
@ -846,7 +847,7 @@ void Renderer::Swap(u32 xfbAddr, u32 fbWidth, u32 fbHeight,const EFBRectangle& r
// done with drawing the game stuff, good moment to save a screenshot
if (s_bScreenshot)
{
TakeScreenshot(GetTargetRectangle(), s_sScreenshotName);
SaveScreenshot(s_sScreenshotName, GetTargetRectangle());
s_bScreenshot = false;
}
@ -921,7 +922,7 @@ void Renderer::Swap(u32 xfbAddr, u32 fbWidth, u32 fbHeight,const EFBRectangle& r
if (SConfig::GetInstance().m_ShowLag)
{
char lag[10];
StringCchPrintfA(lag, 10, "Lag: %llu\n", Movie::g_currentLagCount);
StringCchPrintfA(lag, 10, "Lag: %" PRIu64 "\n", Movie::g_currentLagCount);
D3D::font.DrawTextScaled(0, 18, 20, 0.0f, 0xFF00FFFF, lag);
}

3
Source/Core/VideoBackends/D3D/Src/Render.h
View file

@ -48,10 +48,9 @@ public:
void UpdateViewport();
static void TakeScreenshot(const TargetRectangle &rc, std::string filename);
bool SaveScreenshot(const std::string &filename, const TargetRectangle &rc);
static bool CheckForResize();
};
}

34
Source/Core/VideoBackends/D3D/Src/TextureCache.cpp
View file

@ -14,6 +14,7 @@
#include "PSTextureEncoder.h"
#include "HW/Memmap.h"
#include "VideoConfig.h"
#include "ImageWrite.h"
namespace DX11
{
@ -32,7 +33,7 @@ void TextureCache::TCacheEntry::Bind(unsigned int stage)
D3D::context->PSSetShaderResources(stage, 1, &texture->GetSRV());
}
bool TextureCache::TCacheEntry::Save(const char filename[], unsigned int level)
bool TextureCache::TCacheEntry::Save(const std::string filename, unsigned int level)
{
// TODO: Somehow implement this (D3DX11 doesn't support dumping individual LODs)
static bool warn_once = true;
@ -42,8 +43,35 @@ bool TextureCache::TCacheEntry::Save(const char filename[], unsigned int level)
warn_once = false;
return false;
}
//return SUCCEEDED(PD3DX11SaveTextureToFileA(D3D::context, texture->GetTex(), D3DX11_IFF_PNG, filename));
return true;
ID3D11Texture2D* pNewTexture = NULL;
ID3D11Texture2D* pSurface = texture->GetTex();
D3D11_TEXTURE2D_DESC desc;
pSurface->GetDesc(&desc);
desc.BindFlags = 0;
desc.CPUAccessFlags = D3D11_CPU_ACCESS_READ | D3D11_CPU_ACCESS_WRITE;
desc.Usage = D3D11_USAGE_STAGING;
HRESULT hr = D3D::device->CreateTexture2D(&desc, NULL, &pNewTexture);
bool saved_png = false;
if (SUCCEEDED(hr) && pNewTexture)
{
D3D::context->CopyResource(pNewTexture, pSurface);
D3D11_MAPPED_SUBRESOURCE map;
HRESULT hr = D3D::context->Map(pNewTexture, 0, D3D11_MAP_READ_WRITE, 0, &map);
if (SUCCEEDED(hr))
{
saved_png = TextureToPng((u8*)map.pData, map.RowPitch, filename, desc.Width, desc.Height);
D3D::context->Unmap(pNewTexture, 0);
}
SAFE_RELEASE(pNewTexture);
}
return saved_png;
}
void TextureCache::TCacheEntry::Load(unsigned int width, unsigned int height,

2
Source/Core/VideoBackends/D3D/Src/TextureCache.h
View file

@ -36,7 +36,7 @@ private:
const float *colmat);
void Bind(unsigned int stage);
bool Save(const char filename[], unsigned int level);
bool Save(const std::string filename, unsigned int level);
};
TCacheEntryBase* CreateTexture(unsigned int width, unsigned int height,

4
Source/Core/VideoBackends/D3D/Src/VertexManager.cpp
View file

@ -222,9 +222,9 @@ void VertexManager::vFlush()
tex.texImage0[i&3].width + 1, tex.texImage0[i&3].height + 1,
tex.texImage0[i&3].format, tex.texTlut[i&3].tmem_offset<<9,
tex.texTlut[i&3].tlut_format,
(tex.texMode0[i&3].min_filter & 3),
((tex.texMode0[i&3].min_filter & 3) != 0),
(tex.texMode1[i&3].max_lod + 0xf) / 0x10,
tex.texImage1[i&3].image_type);
(tex.texImage1[i&3].image_type != 0));
if (tentry)
{

2
Source/Core/VideoBackends/OGL/Src/RasterFont.cpp
View file

@ -144,7 +144,7 @@ RasterFont::RasterFont()
for(u32 y=0; y<char_height; y++) {
for(u32 c=0; c<char_count; c++) {
for(u32 x=0; x<char_width; x++) {
bool pixel = rasters[c][y] & (1<<(char_width-x-1));
bool pixel = (0 != (rasters[c][y] & (1<<(char_width-x-1))));
texture_data[char_width*char_count*y+char_width*c+x] = pixel ? -1 : 0;
}
}

91
Source/Core/VideoBackends/OGL/Src/Render.cpp
View file

@ -9,8 +9,12 @@
#include <vector>
#include <cmath>
#include <cstdio>
#include <cinttypes>
#include "GLUtil.h"
#if defined(HAVE_WX) && HAVE_WX
#include "WxUtils.h"
#endif
#include "FileUtil.h"
@ -22,6 +26,7 @@
#include "DriverDetails.h"
#include "VideoConfig.h"
#include "Statistics.h"
#include "ImageWrite.h"
#include "PixelEngine.h"
#include "Render.h"
#include "BPStructs.h"
@ -69,7 +74,6 @@ void VideoConfig::UpdateProjectionHack()
::UpdateProjectionHack(g_Config.iPhackvalue, g_Config.sPhackvalue);
}
int OSDInternalW, OSDInternalH;
namespace OGL
@ -108,6 +112,10 @@ static u32 s_blendMode;
static bool s_vsync;
#if defined(HAVE_WX) && HAVE_WX
static std::thread scrshotThread;
#endif
// EFB cache related
static const u32 EFB_CACHE_RECT_SIZE = 64; // Cache 64x64 blocks.
static const u32 EFB_CACHE_WIDTH = (EFB_WIDTH + EFB_CACHE_RECT_SIZE - 1) / EFB_CACHE_RECT_SIZE; // round up
@ -461,19 +469,23 @@ Renderer::Renderer()
}
g_Config.backend_info.bSupportsDualSourceBlend = GLEW_ARB_blend_func_extended;
g_Config.backend_info.bSupportsGLSLUBO = GLEW_ARB_uniform_buffer_object;
g_Config.backend_info.bSupportsPrimitiveRestart = GLEW_VERSION_3_1 || GLEW_NV_primitive_restart;
g_Config.backend_info.bSupportsEarlyZ = GLEW_ARB_shader_image_load_store;
#define TO_BOOL(c) (0 != (c))
g_ogl_config.bSupportsGLSLCache = GLEW_ARB_get_program_binary;
g_ogl_config.bSupportsGLPinnedMemory = GLEW_AMD_pinned_memory;
g_ogl_config.bSupportsGLSync = GLEW_ARB_sync;
g_ogl_config.bSupportsGLBaseVertex = GLEW_ARB_draw_elements_base_vertex;
g_ogl_config.bSupportCoverageMSAA = GLEW_NV_framebuffer_multisample_coverage;
g_ogl_config.bSupportSampleShading = GLEW_ARB_sample_shading;
g_ogl_config.bSupportOGL31 = GLEW_VERSION_3_1;
g_ogl_config.bSupportViewportFloat = GLEW_ARB_viewport_array;
g_Config.backend_info.bSupportsDualSourceBlend = TO_BOOL(GLEW_ARB_blend_func_extended);
g_Config.backend_info.bSupportsGLSLUBO = TO_BOOL(GLEW_ARB_uniform_buffer_object);
g_Config.backend_info.bSupportsPrimitiveRestart = TO_BOOL(GLEW_VERSION_3_1) || TO_BOOL(GLEW_NV_primitive_restart);
g_Config.backend_info.bSupportsEarlyZ = TO_BOOL(GLEW_ARB_shader_image_load_store);
g_ogl_config.bSupportsGLSLCache = TO_BOOL(GLEW_ARB_get_program_binary);
g_ogl_config.bSupportsGLPinnedMemory = TO_BOOL(GLEW_AMD_pinned_memory);
g_ogl_config.bSupportsGLSync = TO_BOOL(GLEW_ARB_sync);
g_ogl_config.bSupportsGLBaseVertex = TO_BOOL(GLEW_ARB_draw_elements_base_vertex);
g_ogl_config.bSupportCoverageMSAA = TO_BOOL(GLEW_NV_framebuffer_multisample_coverage);
g_ogl_config.bSupportSampleShading = TO_BOOL(GLEW_ARB_sample_shading);
g_ogl_config.bSupportOGL31 = TO_BOOL(GLEW_VERSION_3_1);
g_ogl_config.bSupportViewportFloat = TO_BOOL(GLEW_ARB_viewport_array);
#undef TO_BOOL
if(strstr(g_ogl_config.glsl_version, "1.00") || strstr(g_ogl_config.glsl_version, "1.10") || strstr(g_ogl_config.glsl_version, "1.20"))
{
@ -616,6 +628,11 @@ Renderer::Renderer()
Renderer::~Renderer()
{
#if defined(HAVE_WX) && HAVE_WX
if (scrshotThread.joinable())
scrshotThread.join();
#endif
}
void Renderer::Shutdown()
@ -681,7 +698,7 @@ void Renderer::DrawDebugInfo()
p+=sprintf(p, "FPS: %d\n", s_fps);
if (SConfig::GetInstance().m_ShowLag)
p+=sprintf(p, "Lag: %llu\n", Movie::g_currentLagCount);
p+=sprintf(p, "Lag: %" PRIu64 "\n", Movie::g_currentLagCount);
if (g_ActiveConfig.bShowInputDisplay)
p+=sprintf(p, "%s", Movie::GetInputDisplay().c_str());
@ -1391,9 +1408,11 @@ void Renderer::Swap(u32 xfbAddr, u32 fbWidth, u32 fbHeight,const EFBRectangle& r
// Save screenshot
if (s_bScreenshot)
{
TakeScreenshot(flipped_trc, s_sScreenshotName);
s_bScreenshot = false;
std::lock_guard<std::mutex> lk(s_criticalScreenshot);
SaveScreenshot(s_sScreenshotName, flipped_trc);
// Reset settings
s_sScreenshotName.clear();
s_bScreenshot = false;
}
// Frame dumps are handled a little differently in Windows
@ -1608,7 +1627,7 @@ void Renderer::Swap(u32 xfbAddr, u32 fbWidth, u32 fbHeight,const EFBRectangle& r
// For testing zbuffer targets.
// Renderer::SetZBufferRender();
// SaveTexture("tex.tga", GL_TEXTURE_2D, s_FakeZTarget,
// SaveTexture("tex.png", GL_TEXTURE_2D, s_FakeZTarget,
// GetTargetWidth(), GetTargetHeight());
Core::Callback_VideoCopiedToXFB(XFBWrited || (g_ActiveConfig.bUseXFB && g_ActiveConfig.bUseRealXFB));
XFBWrited = false;
@ -1769,44 +1788,48 @@ void Renderer::SetInterlacingMode()
// TODO
}
void Renderer::FlipImageData(u8 *data, int w, int h)
void Renderer::FlipImageData(u8 *data, int w, int h, int pixel_width)
{
// XXX make this faster
u8* __restrict top = data;
u8* bot = data + w * h * 3;
for (int y = 0; y < h / 2; y++)
// Flip image upside down. Damn OpenGL.
for (int y = 0; y < h / 2; ++y)
{
size_t stride = w * 3;
bot -= stride;
u8* __restrict brow = bot;
for(size_t x = 0; x < stride; x++)
for(int x = 0; x < w; ++x)
{
std::swap(*top++, *brow++);
for (auto delta = 0; delta < pixel_width; ++delta)
std::swap(data[(y * w + x) * pixel_width + delta], data[((h - 1 - y) * w + x) * pixel_width + delta]);
}
}
}
void Renderer::TakeScreenshot(const TargetRectangle &back_rc, std::string filename)
}
namespace OGL
{
bool Renderer::SaveScreenshot(const std::string &filename, const TargetRectangle &back_rc)
{
u32 W = back_rc.GetWidth();
u32 H = back_rc.GetHeight();
u8 *data = (u8 *)malloc((sizeof(u8) * 3 * W * H));
u8 *data = new u8[W * 4 * H];
glPixelStorei(GL_PACK_ALIGNMENT, 1);
glReadPixels(back_rc.left, back_rc.bottom, W, H, GL_RGB, GL_UNSIGNED_BYTE, data);
glReadPixels(back_rc.left, back_rc.bottom, W, H, GL_RGBA, GL_UNSIGNED_BYTE, data);
// Show failure message
if (GL_REPORT_ERROR() != GL_NO_ERROR)
{
free(data);
delete[] data;
OSD::AddMessage("Error capturing or saving screenshot.", 2000);
return;
return false;
}
// Turn image upside down
FlipImageData(data, W, H);
FlipImageData(data, W, H, 4);
bool success = TextureToPng(data, W*4, filename, W, H, false);
delete[] data;
return success;
SaveScreenshot(data, W, H, filename);
}
}

4
Source/Core/VideoBackends/OGL/Src/Render.h
View file

@ -63,7 +63,7 @@ public:
void RenderText(const char* pstr, int left, int top, u32 color) override;
void DrawDebugInfo();
static void FlipImageData(u8 *data, int w, int h);
void FlipImageData(u8 *data, int w, int h, int pixel_width = 3);
u32 AccessEFB(EFBAccessType type, u32 x, u32 y, u32 poke_data) override;
@ -80,7 +80,7 @@ public:
void UpdateViewport() override;
static void TakeScreenshot(const TargetRectangle &rc, std::string filename);
bool SaveScreenshot(const std::string &filename, const TargetRectangle &rc);
private:
void UpdateEFBCache(EFBAccessType type, u32 cacheRectIdx, const EFBRectangle& efbPixelRc, const TargetRectangle& targetPixelRc, const u32* data);

18
Source/Core/VideoBackends/OGL/Src/StreamBuffer.cpp
View file

@ -59,7 +59,7 @@ StreamBuffer::~StreamBuffer()
glDeleteBuffers(1, &m_buffer);
}
#define SLOT(x) (x)*SYNC_POINTS/m_size
#define SLOT(x) ((x)*SYNC_POINTS/m_size)
void StreamBuffer::Alloc ( size_t size, u32 stride )
{
@ -81,14 +81,14 @@ void StreamBuffer::Alloc ( size_t size, u32 stride )
case PINNED_MEMORY:
// insert waiting slots for used memory
for(u32 i=SLOT(m_used_iterator); i<SLOT(m_iterator); i++)
for(size_t i=SLOT(m_used_iterator); i<SLOT(m_iterator); i++)
{
fences[i] = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
}
m_used_iterator = m_iterator;
// wait for new slots to end of buffer
for(u32 i=SLOT(m_free_iterator)+1; i<=SLOT(iter_end) && i < SYNC_POINTS; i++)
for (size_t i = SLOT(m_free_iterator) + 1; i <= SLOT(iter_end) && i < SYNC_POINTS; i++)
{
glClientWaitSync(fences[i], GL_SYNC_FLUSH_COMMANDS_BIT, GL_TIMEOUT_IGNORED);
glDeleteSync(fences[i]);
@ -99,8 +99,10 @@ void StreamBuffer::Alloc ( size_t size, u32 stride )
if(iter_end >= m_size) {
// insert waiting slots in unused space at the end of the buffer
for(u32 i=SLOT(m_used_iterator); i < SYNC_POINTS; i++)
for (size_t i = SLOT(m_used_iterator); i < SYNC_POINTS; i++)
{
fences[i] = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);
}
// move to the start
m_used_iterator = m_iterator_aligned = m_iterator = 0; // offset 0 is always aligned
@ -244,10 +246,14 @@ void StreamBuffer::Shutdown()
void StreamBuffer::DeleteFences()
{
for(u32 i=SLOT(m_free_iterator)+1; i < SYNC_POINTS; i++)
for (size_t i = SLOT(m_free_iterator) + 1; i < SYNC_POINTS; i++)
{
glDeleteSync(fences[i]);
for(u32 i=0; i<SLOT(m_iterator); i++)
}
for (size_t i = 0; i < SLOT(m_iterator); i++)
{
glDeleteSync(fences[i]);
}
delete [] fences;
}

24
Source/Core/VideoBackends/OGL/Src/TextureCache.cpp
View file

@ -59,15 +59,15 @@ struct VBOCache {
};
static std::map<u64,VBOCache> s_VBO;
bool SaveTexture(const char* filename, u32 textarget, u32 tex, int virtual_width, int virtual_height, unsigned int level)
bool SaveTexture(const std::string filename, u32 textarget, u32 tex, int virtual_width, int virtual_height, unsigned int level)
{
#ifndef USE_GLES3
int width = std::max(virtual_width >> level, 1);
int height = std::max(virtual_height >> level, 1);
std::vector<u32> data(width * height);
u8* data = new u8[width * height * 4];
glActiveTexture(GL_TEXTURE0+9);
glBindTexture(textarget, tex);
glGetTexImage(textarget, level, GL_BGRA, GL_UNSIGNED_BYTE, &data[0]);
glGetTexImage(textarget, level, GL_RGBA, GL_UNSIGNED_BYTE, data);
glBindTexture(textarget, 0);
TextureCache::SetStage();
@ -75,10 +75,12 @@ bool SaveTexture(const char* filename, u32 textarget, u32 tex, int virtual_width
if (GL_NO_ERROR != err)
{
PanicAlert("Can't save texture, GL Error: %s", gluErrorString(err));
delete[] data;
return false;
}
return SaveTGA(filename, width, height, &data[0]);
bool success = TextureToPng(data, width * 4, filename, width, height, true);
delete[] data;
return success;
#else
return false;
#endif
@ -125,13 +127,9 @@ void TextureCache::TCacheEntry::Bind(unsigned int stage)
}
}
bool TextureCache::TCacheEntry::Save(const char filename[], unsigned int level)
bool TextureCache::TCacheEntry::Save(const std::string filename, unsigned int level)
{
// TODO: make ogl dump PNGs
std::string tga_filename(filename);
tga_filename.replace(tga_filename.size() - 3, 3, "tga");
return SaveTexture(tga_filename.c_str(), GL_TEXTURE_2D, texture, virtual_width, virtual_height, level);
return SaveTexture(filename, GL_TEXTURE_2D, texture, virtual_width, virtual_height, level);
}
TextureCache::TCacheEntryBase* TextureCache::CreateTexture(unsigned int width,
@ -395,8 +393,8 @@ void TextureCache::TCacheEntry::FromRenderTarget(u32 dstAddr, unsigned int dstFo
if (g_ActiveConfig.bDumpEFBTarget)
{
static int count = 0;
SaveTexture(StringFromFormat("%sefb_frame_%i.tga", File::GetUserPath(D_DUMPTEXTURES_IDX).c_str(),
count++).c_str(), GL_TEXTURE_2D, texture, virtual_width, virtual_height, 0);
SaveTexture(StringFromFormat("%sefb_frame_%i.png", File::GetUserPath(D_DUMPTEXTURES_IDX).c_str(),
count++), GL_TEXTURE_2D, texture, virtual_width, virtual_height, 0);
}
g_renderer->RestoreAPIState();

Some files were not shown because too many files have changed in this diff Show more