Emulator-Archive/ppsspp
1
0
Fork 0
mirror of https://github.com/hrydgard/ppsspp.git synced 2024-11-23 13:30:02 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix recent texture scaling bug. Should help #9250

Browse Source 
Not all cases tested yet.
This commit is contained in:
Henrik Rydgard 2017-01-29 14:42:21 +01:00
parent 5b1b4329fd
commit 59f5c53040
2 changed files with 12 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
GPU/Common/VertexDecoderArm.cpp
View File

@ -656,6 +656,8 @@ void VertexDecoderJitCache::Jit_TcU8ToFloat() {
VMOVL(I_8 | I_UNSIGNED, neonScratchRegQ, neonScratchReg); // Widen to 16-bit
VMOVL(I_16 | I_UNSIGNED, neonScratchRegQ, neonScratchReg); // Widen to 32-bit
VCVT(F_32 | I_UNSIGNED, neonScratchRegQ, neonScratchRegQ);
VMOV_neon(F_32, neonScratchReg2, by128);
VMUL(F_32, neonScratchReg, neonScratchReg, neonScratchReg2);
ADD(scratchReg2, dstReg, dec_->decFmt.uvoff);
VST1(F_32, neonScratchReg, scratchReg2, 1, ALIGN_NONE);
} else {
@ -665,6 +667,9 @@ void VertexDecoderJitCache::Jit_TcU8ToFloat() {
VMOV(fpScratchReg2, tempReg2);
VCVT(fpScratchReg, fpScratchReg, TO_FLOAT);
VCVT(fpScratchReg2, fpScratchReg2, TO_FLOAT);
MOVI2F(S15, by128, scratchReg);
VMUL(fpScratchReg, fpScratchReg, S15);
VMUL(fpScratchReg2, fpScratchReg2, S15);
VSTR(fpScratchReg, dstReg, dec_->decFmt.uvoff);
VSTR(fpScratchReg2, dstReg, dec_->decFmt.uvoff + 4);
}
@ -705,6 +710,8 @@ void VertexDecoderJitCache::Jit_TcU16ToFloat() {
VMOVL(I_16 | I_UNSIGNED, neonScratchRegQ, neonScratchReg); // Widen to 32-bit
VCVT(F_32 | I_UNSIGNED, neonScratchRegQ, neonScratchRegQ);
ADD(scratchReg2, dstReg, dec_->decFmt.uvoff);
VMOV_neon(F_32, neonScratchReg2, by32768);
VMUL(F_32, neonScratchReg, neonScratchReg, neonScratchReg2);
VST1(F_32, neonScratchReg, scratchReg2, 1, ALIGN_NONE);
} else {
LDRH(tempReg1, srcReg, dec_->tcoff);
@ -713,6 +720,9 @@ void VertexDecoderJitCache::Jit_TcU16ToFloat() {
VMOV(fpScratchReg2, tempReg2);
VCVT(fpScratchReg, fpScratchReg, TO_FLOAT);
VCVT(fpScratchReg2, fpScratchReg2, TO_FLOAT);
MOVI2F(S15, by32768, scratchReg);
VMUL(fpScratchReg, fpScratchReg, S15);
VMUL(fpScratchReg2, fpScratchReg2, S15);
VSTR(fpScratchReg, dstReg, dec_->decFmt.uvoff);
VSTR(fpScratchReg2, dstReg, dec_->decFmt.uvoff + 4);
}

4
GPU/Common/VertexDecoderArm64.cpp
View File

@ -637,14 +637,14 @@ void VertexDecoderJitCache::Jit_TcU8ToFloat() {
fp.LDUR(16, neonScratchRegD, srcReg, dec_->tcoff);
fp.UXTL(8, neonScratchRegQ, neonScratchRegD); // Widen to 16-bit
fp.UXTL(16, neonScratchRegQ, neonScratchRegD); // Widen to 32-bit
fp.UCVTF(32, neonScratchRegD, neonScratchRegD);
fp.UCVTF(32, neonScratchRegD, neonScratchRegD, 7);
fp.STUR(64, neonScratchRegD, dstReg, dec_->decFmt.uvoff);
}
void VertexDecoderJitCache::Jit_TcU16Prescale() {
fp.LDUR(32, neonScratchRegD, srcReg, dec_->tcoff);
fp.UXTL(16, neonScratchRegQ, neonScratchRegD); // Widen to 32-bit
fp.UCVTF(32, neonScratchRegD, neonScratchRegD);
fp.UCVTF(32, neonScratchRegD, neonScratchRegD, 15);
fp.FMUL(32, neonScratchRegD, neonScratchRegD, neonUVScaleReg); // TODO: FMLA
fp.FADD(32, neonScratchRegD, neonScratchRegD, neonUVOffsetReg);
fp.STUR(64, neonScratchRegD, dstReg, dec_->decFmt.uvoff);