virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-10-06 05:47:18 +02:00
Code Issues Releases Activity

avcodec/x86/h264_qpel: Remove MMX(EXT) funcs overridden by SSSE3

Browse Source 
SSSE3 is already quite old (introduced 2006 for Intel, 2011 for AMD),
so that the overwhelming majority of our users (particularly those
that actually update their FFmpeg) will be using the SSSE3 versions.
This commit therefore removes the MMX(EXT) functions overridden
by them (which don't abide by the ABI) to get closer to a removal
of emms_c.

Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit is contained in:
Andreas Rheinhardt
2025-09-21 15:12:49 +02:00
parent 4fc05c28f4
commit c7161befb4
2 changed files with 3 additions and 91 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
libavcodec/x86/h264_qpel.c
View File

@@ -46,12 +46,10 @@ void ff_avg_pixels16_l2_mmxext(uint8_t *dst, const uint8_t *src1, const uint8_t
#define ff_avg_pixels8_l2_sse2 ff_avg_pixels8_l2_mmxext
#define ff_put_pixels16_l2_sse2 ff_put_pixels16_l2_mmxext
#define ff_avg_pixels16_l2_sse2 ff_avg_pixels16_l2_mmxext
#define ff_put_pixels8_mmxext(...)
#define ff_put_pixels4_mmxext(...)
#define DEF_QPEL(OPNAME)\
void ff_ ## OPNAME ## _h264_qpel4_h_lowpass_mmxext(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride);\
void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_mmxext(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride);\
void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_ssse3(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride);\
void ff_ ## OPNAME ## _h264_qpel4_h_lowpass_l2_mmxext(uint8_t *dst, const uint8_t *src, const uint8_t *src2, int dstStride, int src2Stride);\
void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_l2_mmxext(uint8_t *dst, const uint8_t *src, const uint8_t *src2, int dstStride, int src2Stride);\
@@ -91,15 +89,6 @@ static av_always_inline void ff_ ## OPNAME ## h264_qpel8or16_hv2_lowpass_ ## MMX
}while(w--);\
}\
\
static av_always_inline void ff_ ## OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride){\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
src += 8*srcStride;\
dst += 8*dstStride;\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
}\
\
static av_always_inline void ff_ ## OPNAME ## h264_qpel16_h_lowpass_l2_ ## MMX(uint8_t *dst, const uint8_t *src, const uint8_t *src2, int dstStride, int src2Stride){\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst , src , src2 , dstStride, src2Stride);\
ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst+8, src+8, src2+8, dstStride, src2Stride);\
@@ -196,10 +185,6 @@ static av_always_inline void ff_ ## OPNAME ## h264_qpel16_hv_lowpass_ ## MMX(uin
#define ff_put_h264_qpel8or16_hv2_lowpass_sse2 ff_put_h264_qpel8or16_hv2_lowpass_mmxext
#define ff_avg_h264_qpel8or16_hv2_lowpass_sse2 ff_avg_h264_qpel8or16_hv2_lowpass_mmxext
#define H264_MC_C_H(OPNAME, SIZE, MMX, ALIGN) \
H264_MC_C(OPNAME, SIZE, MMX, ALIGN)\
H264_MC_H(OPNAME, SIZE, MMX, ALIGN)\
#define H264_MC_C_V_H_HV(OPNAME, SIZE, MMX, ALIGN) \
H264_MC_C(OPNAME, SIZE, MMX, ALIGN)\
H264_MC_V(OPNAME, SIZE, MMX, ALIGN)\
@@ -356,8 +341,7 @@ QPEL_H264_HV_XMM(put_, PUT_OP, ssse3)
QPEL_H264_HV_XMM(avg_,AVG_MMXEXT_OP, ssse3)
H264_MC(H264_MC_C_V_H_HV, 4, mmxext, 8)
H264_MC(H264_MC_C_H, 8, mmxext, 8)
H264_MC(H264_MC_H, 16, mmxext, 8)
H264_MC_C(avg_, 8, mmxext, 8)
H264_MC_816(H264_MC_V, sse2)
H264_MC_816(H264_MC_HV, sse2)
H264_MC_816(H264_MC_H, ssse3)
@@ -421,20 +405,11 @@ LUMA_MC_816(10, mc33, sse2)
#endif /* HAVE_X86ASM */
#define SET_QPEL_FUNCS123(PFX, IDX, SIZE, CPU, PREFIX) \
#define SET_QPEL_FUNCS_1PP(PFX, IDX, SIZE, CPU, PREFIX) \
do { \
c->PFX ## _pixels_tab[IDX][ 1] = PREFIX ## PFX ## SIZE ## _mc10_ ## CPU; \
c->PFX ## _pixels_tab[IDX][ 2] = PREFIX ## PFX ## SIZE ## _mc20_ ## CPU; \
c->PFX ## _pixels_tab[IDX][ 3] = PREFIX ## PFX ## SIZE ## _mc30_ ## CPU; \
} while (0)
#define SET_QPEL_FUNCS0123(PFX, IDX, SIZE, CPU, PREFIX) \
do { \
c->PFX ## _pixels_tab[IDX][ 0] = PREFIX ## PFX ## SIZE ## _mc00_ ## CPU; \
SET_QPEL_FUNCS123(PFX, IDX, SIZE, CPU, PREFIX); \
} while (0)
#define SET_QPEL_FUNCS_1PP(PFX, IDX, SIZE, CPU, PREFIX) \
do { \
SET_QPEL_FUNCS123(PFX, IDX, SIZE, CPU, PREFIX); \
c->PFX ## _pixels_tab[IDX][ 4] = PREFIX ## PFX ## SIZE ## _mc01_ ## CPU; \
c->PFX ## _pixels_tab[IDX][ 5] = PREFIX ## PFX ## SIZE ## _mc11_ ## CPU; \
c->PFX ## _pixels_tab[IDX][ 6] = PREFIX ## PFX ## SIZE ## _mc21_ ## CPU; \
@@ -478,11 +453,8 @@ av_cold void ff_h264qpel_init_x86(H264QpelContext *c, int bit_depth)
if (EXTERNAL_MMXEXT(cpu_flags)) {
if (!high_bit_depth) {
SET_QPEL_FUNCS123 (put_h264_qpel, 0, 16, mmxext, );
SET_QPEL_FUNCS123 (put_h264_qpel, 1, 8, mmxext, );
SET_QPEL_FUNCS_1PP(put_h264_qpel, 2, 4, mmxext, );
SET_QPEL_FUNCS123 (avg_h264_qpel, 0, 16, mmxext, );
SET_QPEL_FUNCS0123(avg_h264_qpel, 1, 8, mmxext, );
c->avg_h264_qpel_pixels_tab[1][0] = avg_h264_qpel8_mc00_mmxext;
SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, mmxext, );
} else if (bit_depth == 10) {
SET_QPEL_FUNCS(put_h264_qpel, 2, 4, 10_mmxext, ff_);

60
libavcodec/x86/h264_qpel_8bit.asm
View File

@@ -96,66 +96,6 @@ INIT_MMX mmxext
QPEL4_H_LOWPASS_OP put
QPEL4_H_LOWPASS_OP avg
%macro QPEL8_H_LOWPASS_OP 1
cglobal %1_h264_qpel8_h_lowpass, 4,5 ; dst, src, dstStride, srcStride
movsxdifnidn r2, r2d
movsxdifnidn r3, r3d
mov r4d, 8
pxor m7, m7
mova m6, [pw_5]
.loop:
mova m0, [r1]
mova m2, [r1+1]
mova m1, m0
mova m3, m2
punpcklbw m0, m7
punpckhbw m1, m7
punpcklbw m2, m7
punpckhbw m3, m7
paddw m0, m2
paddw m1, m3
psllw m0, 2
psllw m1, 2
mova m2, [r1-1]
mova m4, [r1+2]
mova m3, m2
mova m5, m4
punpcklbw m2, m7
punpckhbw m3, m7
punpcklbw m4, m7
punpckhbw m5, m7
paddw m2, m4
paddw m5, m3
psubw m0, m2
psubw m1, m5
pmullw m0, m6
pmullw m1, m6
movd m2, [r1-2]
movd m5, [r1+7]
punpcklbw m2, m7
punpcklbw m5, m7
paddw m2, m3
paddw m4, m5
mova m5, [pw_16]
paddw m2, m5
paddw m4, m5
paddw m0, m2
paddw m1, m4
psraw m0, 5
psraw m1, 5
packuswb m0, m1
op_%1 m0, [r0], m4
add r0, r2
add r1, r3
dec r4d
jg .loop
RET
%endmacro
INIT_MMX mmxext
QPEL8_H_LOWPASS_OP put
QPEL8_H_LOWPASS_OP avg
%macro QPEL8_H_LOWPASS_OP_XMM 1
cglobal %1_h264_qpel8_h_lowpass, 4,5,8 ; dst, src, dstStride, srcStride
movsxdifnidn r2, r2d