From 7284ab789d5fe271b9d6a1666ab5ea6be8724cca Mon Sep 17 00:00:00 2001 From: Andreas Rheinhardt Date: Tue, 7 Jun 2022 02:06:41 +0200 Subject: [PATCH] avcodec/x86/mpegvideoenc: Remove obsolete MMX(EXT) functions x64 always has MMX, MMXEXT, SSE and SSE2 and this means that some functions for MMX, MMXEXT and 3dnow are always overridden by other functions (unless one e.g. explicitly disables SSE2) for x64. So given that the only systems that benefit from these functions are truely ancient 32bit x86s they are removed. Signed-off-by: Andreas Rheinhardt --- libavcodec/x86/mpegvideoenc.c | 95 +------------------------- libavcodec/x86/mpegvideoenc_template.c | 10 --- 2 files changed, 1 insertion(+), 104 deletions(-) diff --git a/libavcodec/x86/mpegvideoenc.c b/libavcodec/x86/mpegvideoenc.c index 3691cce26c..612e7ff758 100644 --- a/libavcodec/x86/mpegvideoenc.c +++ b/libavcodec/x86/mpegvideoenc.c @@ -42,34 +42,9 @@ DECLARE_ALIGNED(16, static const uint16_t, inv_zigzag_direct16)[64] = { #if HAVE_6REGS -#if HAVE_MMX_INLINE -#define COMPILE_TEMPLATE_MMXEXT 0 -#define COMPILE_TEMPLATE_SSE2 0 -#define COMPILE_TEMPLATE_SSSE3 0 -#define RENAME(a) a ## _mmx -#define RENAME_FDCT(a) a ## _mmx -#include "mpegvideoenc_template.c" -#endif /* HAVE_MMX_INLINE */ - -#if HAVE_MMXEXT_INLINE -#undef COMPILE_TEMPLATE_SSSE3 -#undef COMPILE_TEMPLATE_SSE2 -#undef COMPILE_TEMPLATE_MMXEXT -#define COMPILE_TEMPLATE_MMXEXT 1 -#define COMPILE_TEMPLATE_SSE2 0 -#define COMPILE_TEMPLATE_SSSE3 0 -#undef RENAME -#undef RENAME_FDCT -#define RENAME(a) a ## _mmxext -#define RENAME_FDCT(a) a ## _mmxext -#include "mpegvideoenc_template.c" -#endif /* HAVE_MMXEXT_INLINE */ - #if HAVE_SSE2_INLINE -#undef COMPILE_TEMPLATE_MMXEXT #undef COMPILE_TEMPLATE_SSE2 #undef COMPILE_TEMPLATE_SSSE3 -#define COMPILE_TEMPLATE_MMXEXT 0 #define COMPILE_TEMPLATE_SSE2 1 #define COMPILE_TEMPLATE_SSSE3 0 #undef RENAME @@ -80,10 +55,8 @@ DECLARE_ALIGNED(16, static const uint16_t, inv_zigzag_direct16)[64] = { #endif /* HAVE_SSE2_INLINE */ #if HAVE_SSSE3_INLINE -#undef COMPILE_TEMPLATE_MMXEXT #undef COMPILE_TEMPLATE_SSE2 #undef COMPILE_TEMPLATE_SSSE3 -#define COMPILE_TEMPLATE_MMXEXT 0 #define COMPILE_TEMPLATE_SSE2 1 #define COMPILE_TEMPLATE_SSSE3 1 #undef RENAME @@ -96,62 +69,6 @@ DECLARE_ALIGNED(16, static const uint16_t, inv_zigzag_direct16)[64] = { #endif /* HAVE_6REGS */ #if HAVE_INLINE_ASM -#if HAVE_MMX_INLINE -static void denoise_dct_mmx(MpegEncContext *s, int16_t *block){ - const int intra= s->mb_intra; - int *sum= s->dct_error_sum[intra]; - uint16_t *offset= s->dct_offset[intra]; - - s->dct_count[intra]++; - - __asm__ volatile( - "pxor %%mm7, %%mm7 \n\t" - "1: \n\t" - "pxor %%mm0, %%mm0 \n\t" - "pxor %%mm1, %%mm1 \n\t" - "movq (%0), %%mm2 \n\t" - "movq 8(%0), %%mm3 \n\t" - "pcmpgtw %%mm2, %%mm0 \n\t" - "pcmpgtw %%mm3, %%mm1 \n\t" - "pxor %%mm0, %%mm2 \n\t" - "pxor %%mm1, %%mm3 \n\t" - "psubw %%mm0, %%mm2 \n\t" - "psubw %%mm1, %%mm3 \n\t" - "movq %%mm2, %%mm4 \n\t" - "movq %%mm3, %%mm5 \n\t" - "psubusw (%2), %%mm2 \n\t" - "psubusw 8(%2), %%mm3 \n\t" - "pxor %%mm0, %%mm2 \n\t" - "pxor %%mm1, %%mm3 \n\t" - "psubw %%mm0, %%mm2 \n\t" - "psubw %%mm1, %%mm3 \n\t" - "movq %%mm2, (%0) \n\t" - "movq %%mm3, 8(%0) \n\t" - "movq %%mm4, %%mm2 \n\t" - "movq %%mm5, %%mm3 \n\t" - "punpcklwd %%mm7, %%mm4 \n\t" - "punpckhwd %%mm7, %%mm2 \n\t" - "punpcklwd %%mm7, %%mm5 \n\t" - "punpckhwd %%mm7, %%mm3 \n\t" - "paddd (%1), %%mm4 \n\t" - "paddd 8(%1), %%mm2 \n\t" - "paddd 16(%1), %%mm5 \n\t" - "paddd 24(%1), %%mm3 \n\t" - "movq %%mm4, (%1) \n\t" - "movq %%mm2, 8(%1) \n\t" - "movq %%mm5, 16(%1) \n\t" - "movq %%mm3, 24(%1) \n\t" - "add $16, %0 \n\t" - "add $32, %1 \n\t" - "add $16, %2 \n\t" - "cmp %3, %0 \n\t" - " jb 1b \n\t" - : "+r" (block), "+r" (sum), "+r" (offset) - : "r"(block+64) - ); -} -#endif /* HAVE_MMX_INLINE */ - #if HAVE_SSE2_INLINE static void denoise_dct_sse2(MpegEncContext *s, int16_t *block){ const int intra= s->mb_intra; @@ -218,17 +135,6 @@ av_cold void ff_dct_encode_init_x86(MpegEncContext *s) if (dct_algo == FF_DCT_AUTO || dct_algo == FF_DCT_MMX) { #if HAVE_MMX_INLINE int cpu_flags = av_get_cpu_flags(); - if (INLINE_MMX(cpu_flags)) { -#if HAVE_6REGS - s->dct_quantize = dct_quantize_mmx; -#endif - s->denoise_dct = denoise_dct_mmx; - } -#endif -#if HAVE_6REGS && HAVE_MMXEXT_INLINE - if (INLINE_MMXEXT(cpu_flags)) - s->dct_quantize = dct_quantize_mmxext; -#endif #if HAVE_SSE2_INLINE if (INLINE_SSE2(cpu_flags)) { #if HAVE_6REGS @@ -240,6 +146,7 @@ av_cold void ff_dct_encode_init_x86(MpegEncContext *s) #if HAVE_6REGS && HAVE_SSSE3_INLINE if (INLINE_SSSE3(cpu_flags)) s->dct_quantize = dct_quantize_ssse3; +#endif #endif } } diff --git a/libavcodec/x86/mpegvideoenc_template.c b/libavcodec/x86/mpegvideoenc_template.c index 30c06a6b2c..b5b0a5ffa2 100644 --- a/libavcodec/x86/mpegvideoenc_template.c +++ b/libavcodec/x86/mpegvideoenc_template.c @@ -57,15 +57,6 @@ #define MMREG_WIDTH "8" #define MM "%%mm" #define MOVQ "movq" -#if COMPILE_TEMPLATE_MMXEXT -#define SPREADW(a) "pshufw $0, "a", "a" \n\t" -#define PMAXW(a,b) "pmaxsw "a", "b" \n\t" -#define PMAX(a,b) \ - "pshufw $0x0E, "a", "b" \n\t"\ - PMAXW(b, a)\ - "pshufw $0x01, "a", "b" \n\t"\ - PMAXW(b, a) -#else #define SPREADW(a) \ "punpcklwd "a", "a" \n\t"\ "punpcklwd "a", "a" \n\t" @@ -81,7 +72,6 @@ PMAXW(b, a) #endif -#endif #if COMPILE_TEMPLATE_SSSE3 #define SAVE_SIGN(a,b) \