symbols, and move them from libavcodec/avcodec.h to libavutil/cpu.h. Originally committed as revision 25040 to svn://svn.ffmpeg.org/ffmpeg/trunktags/n0.8
@@ -75,7 +75,7 @@ static void simple_idct_arm_add(uint8_t *dest, int line_size, DCTELEM *block) | |||||
int mm_support(void) | int mm_support(void) | ||||
{ | { | ||||
return HAVE_IWMMXT * FF_MM_IWMMXT; | |||||
return HAVE_IWMMXT * AV_CPU_FLAG_IWMMXT; | |||||
} | } | ||||
void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx) | void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx) | ||||
@@ -19,6 +19,7 @@ | |||||
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||||
*/ | */ | ||||
#include "libavutil/cpu.h" | |||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
#define DEF(x, y) x ## _no_rnd_ ## y ##_iwmmxt | #define DEF(x, y) x ## _no_rnd_ ## y ##_iwmmxt | ||||
@@ -153,16 +154,16 @@ static void nop(uint8_t *block, const uint8_t *pixels, int line_size, int h) | |||||
void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx) | void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx) | ||||
{ | { | ||||
int mm_flags = FF_MM_IWMMXT; /* multimedia extension flags */ | |||||
int mm_flags = AV_CPU_FLAG_IWMMXT; /* multimedia extension flags */ | |||||
if (avctx->dsp_mask) { | if (avctx->dsp_mask) { | ||||
if (avctx->dsp_mask & FF_MM_FORCE) | |||||
if (avctx->dsp_mask & AV_CPU_FLAG_FORCE) | |||||
mm_flags |= (avctx->dsp_mask & 0xffff); | mm_flags |= (avctx->dsp_mask & 0xffff); | ||||
else | else | ||||
mm_flags &= ~(avctx->dsp_mask & 0xffff); | mm_flags &= ~(avctx->dsp_mask & 0xffff); | ||||
} | } | ||||
if (!(mm_flags & FF_MM_IWMMXT)) return; | |||||
if (!(mm_flags & AV_CPU_FLAG_IWMMXT)) return; | |||||
c->add_pixels_clamped = add_pixels_clamped_iwmmxt; | c->add_pixels_clamped = add_pixels_clamped_iwmmxt; | ||||
@@ -18,6 +18,7 @@ | |||||
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||||
*/ | */ | ||||
#include "libavutil/cpu.h" | |||||
#include "libavcodec/avcodec.h" | #include "libavcodec/avcodec.h" | ||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
#include "libavcodec/mpegvideo.h" | #include "libavcodec/mpegvideo.h" | ||||
@@ -111,7 +112,7 @@ static void dct_unquantize_h263_inter_iwmmxt(MpegEncContext *s, | |||||
void MPV_common_init_iwmmxt(MpegEncContext *s) | void MPV_common_init_iwmmxt(MpegEncContext *s) | ||||
{ | { | ||||
if (!(mm_flags & FF_MM_IWMMXT)) return; | |||||
if (!(mm_flags & AV_CPU_FLAG_IWMMXT)) return; | |||||
s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_iwmmxt; | s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_iwmmxt; | ||||
#if 0 | #if 0 | ||||
@@ -29,6 +29,7 @@ | |||||
*/ | */ | ||||
#include "libavutil/cpu.h" | |||||
#include "avcodec.h" | #include "avcodec.h" | ||||
@@ -93,7 +94,7 @@ typedef struct AVAudioConvert AVAudioConvert; | |||||
* @param in_fmt Input sample format | * @param in_fmt Input sample format | ||||
* @param in_channels Number of input channels | * @param in_channels Number of input channels | ||||
* @param[in] matrix Channel mixing matrix (of dimension in_channel*out_channels). Set to NULL to ignore. | * @param[in] matrix Channel mixing matrix (of dimension in_channel*out_channels). Set to NULL to ignore. | ||||
* @param flags See FF_MM_xx | |||||
* @param flags See AV_CPU_FLAG_xx | |||||
* @return NULL on error | * @return NULL on error | ||||
*/ | */ | ||||
AVAudioConvert *av_audio_convert_alloc(enum SampleFormat out_fmt, int out_channels, | AVAudioConvert *av_audio_convert_alloc(enum SampleFormat out_fmt, int out_channels, | ||||
@@ -28,10 +28,11 @@ | |||||
#include <errno.h> | #include <errno.h> | ||||
#include "libavutil/avutil.h" | #include "libavutil/avutil.h" | ||||
#include "libavutil/cpu.h" | |||||
#define LIBAVCODEC_VERSION_MAJOR 52 | #define LIBAVCODEC_VERSION_MAJOR 52 | ||||
#define LIBAVCODEC_VERSION_MINOR 87 | #define LIBAVCODEC_VERSION_MINOR 87 | ||||
#define LIBAVCODEC_VERSION_MICRO 0 | |||||
#define LIBAVCODEC_VERSION_MICRO 1 | |||||
#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \ | #define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \ | ||||
LIBAVCODEC_VERSION_MINOR, \ | LIBAVCODEC_VERSION_MINOR, \ | ||||
@@ -50,6 +51,7 @@ | |||||
#ifndef FF_API_PALETTE_CONTROL | #ifndef FF_API_PALETTE_CONTROL | ||||
#define FF_API_PALETTE_CONTROL (LIBAVCODEC_VERSION_MAJOR < 54) | #define FF_API_PALETTE_CONTROL (LIBAVCODEC_VERSION_MAJOR < 54) | ||||
#endif | #endif | ||||
#define FF_API_MM_FLAGS (LIBAVCODEC_VERSION_MAJOR < 53) | |||||
#define AV_NOPTS_VALUE INT64_C(0x8000000000000000) | #define AV_NOPTS_VALUE INT64_C(0x8000000000000000) | ||||
#define AV_TIME_BASE 1000000 | #define AV_TIME_BASE 1000000 | ||||
@@ -1657,27 +1659,25 @@ typedef struct AVCodecContext { | |||||
* result into program crash.) | * result into program crash.) | ||||
*/ | */ | ||||
unsigned dsp_mask; | unsigned dsp_mask; | ||||
#define FF_MM_FORCE 0x80000000 /* Force usage of selected flags (OR) */ | |||||
/* lower 16 bits - CPU features */ | |||||
#define FF_MM_MMX 0x0001 ///< standard MMX | |||||
#define FF_MM_3DNOW 0x0004 ///< AMD 3DNOW | |||||
#if LIBAVCODEC_VERSION_MAJOR < 53 | |||||
#define FF_MM_MMXEXT 0x0002 ///< SSE integer functions or AMD MMX ext | |||||
#if FF_API_MM_FLAGS | |||||
#define FF_MM_FORCE AV_CPU_FLAG_FORCE | |||||
#define FF_MM_MMX AV_CPU_FLAG_MMX | |||||
#define FF_MM_3DNOW AV_CPU_FLAG_3DNOW | |||||
#define FF_MM_MMXEXT AV_CPU_FLAG_MMX2 | |||||
#define FF_MM_MMX2 AV_CPU_FLAG_MMX2 | |||||
#define FF_MM_SSE AV_CPU_FLAG_SSE | |||||
#define FF_MM_SSE2 AV_CPU_FLAG_SSE2 | |||||
#define FF_MM_SSE2SLOW AV_CPU_FLAG_SSE2SLOW | |||||
#define FF_MM_3DNOWEXT AV_CPU_FLAG_3DNOWEXT | |||||
#define FF_MM_SSE3 AV_CPU_FLAG_SSE3 | |||||
#define FF_MM_SSE3SLOW AV_CPU_FLAG_SSE3SLOW | |||||
#define FF_MM_SSSE3 AV_CPU_FLAG_SSSE3 | |||||
#define FF_MM_SSE4 AV_CPU_FLAG_SSE4 | |||||
#define FF_MM_SSE42 AV_CPU_FLAG_SSE42 | |||||
#define FF_MM_IWMMXT AV_CPU_FLAG_IWMMXT | |||||
#define FF_MM_ALTIVEC AV_CPU_FLAG_ALTIVEC | |||||
#endif | #endif | ||||
#define FF_MM_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext | |||||
#define FF_MM_SSE 0x0008 ///< SSE functions | |||||
#define FF_MM_SSE2 0x0010 ///< PIV SSE2 functions | |||||
#define FF_MM_SSE2SLOW 0x40000000 ///< SSE2 supported, but usually not faster | |||||
///< than regular MMX/SSE (e.g. Core1) | |||||
#define FF_MM_3DNOWEXT 0x0020 ///< AMD 3DNowExt | |||||
#define FF_MM_SSE3 0x0040 ///< Prescott SSE3 functions | |||||
#define FF_MM_SSE3SLOW 0x20000000 ///< SSE3 supported, but usually not faster | |||||
///< than regular MMX/SSE (e.g. Core1) | |||||
#define FF_MM_SSSE3 0x0080 ///< Conroe SSSE3 functions | |||||
#define FF_MM_SSE4 0x0100 ///< Penryn SSE4.1 functions | |||||
#define FF_MM_SSE42 0x0200 ///< Nehalem SSE4.2 functions | |||||
#define FF_MM_IWMMXT 0x0100 ///< XScale IWMMXT | |||||
#define FF_MM_ALTIVEC 0x0001 ///< standard AltiVec | |||||
/** | /** | ||||
* bits per sample/pixel from the demuxer (needed for huffyuv). | * bits per sample/pixel from the demuxer (needed for huffyuv). | ||||
@@ -94,24 +94,24 @@ struct algo algos[] = { | |||||
{"SIMPLE-C", 1, ff_simple_idct, ff_ref_idct, NO_PERM}, | {"SIMPLE-C", 1, ff_simple_idct, ff_ref_idct, NO_PERM}, | ||||
#if HAVE_MMX | #if HAVE_MMX | ||||
{"MMX", 0, ff_fdct_mmx, ff_ref_fdct, NO_PERM, FF_MM_MMX}, | |||||
{"MMX", 0, ff_fdct_mmx, ff_ref_fdct, NO_PERM, AV_CPU_FLAG_MMX}, | |||||
#if HAVE_MMX2 | #if HAVE_MMX2 | ||||
{"MMX2", 0, ff_fdct_mmx2, ff_ref_fdct, NO_PERM, FF_MM_MMX2}, | |||||
{"SSE2", 0, ff_fdct_sse2, ff_ref_fdct, NO_PERM, FF_MM_SSE2}, | |||||
{"MMX2", 0, ff_fdct_mmx2, ff_ref_fdct, NO_PERM, AV_CPU_FLAG_MMX2}, | |||||
{"SSE2", 0, ff_fdct_sse2, ff_ref_fdct, NO_PERM, AV_CPU_FLAG_SSE2}, | |||||
#endif | #endif | ||||
#if CONFIG_GPL | #if CONFIG_GPL | ||||
{"LIBMPEG2-MMX", 1, ff_mmx_idct, ff_ref_idct, MMX_PERM, FF_MM_MMX}, | |||||
{"LIBMPEG2-MMX2", 1, ff_mmxext_idct, ff_ref_idct, MMX_PERM, FF_MM_MMX2}, | |||||
{"LIBMPEG2-MMX", 1, ff_mmx_idct, ff_ref_idct, MMX_PERM, AV_CPU_FLAG_MMX}, | |||||
{"LIBMPEG2-MMX2", 1, ff_mmxext_idct, ff_ref_idct, MMX_PERM, AV_CPU_FLAG_MMX2}, | |||||
#endif | #endif | ||||
{"SIMPLE-MMX", 1, ff_simple_idct_mmx, ff_ref_idct, MMX_SIMPLE_PERM, FF_MM_MMX}, | |||||
{"XVID-MMX", 1, ff_idct_xvid_mmx, ff_ref_idct, NO_PERM, FF_MM_MMX}, | |||||
{"XVID-MMX2", 1, ff_idct_xvid_mmx2, ff_ref_idct, NO_PERM, FF_MM_MMX2}, | |||||
{"XVID-SSE2", 1, ff_idct_xvid_sse2, ff_ref_idct, SSE2_PERM, FF_MM_SSE2}, | |||||
{"SIMPLE-MMX", 1, ff_simple_idct_mmx, ff_ref_idct, MMX_SIMPLE_PERM, AV_CPU_FLAG_MMX}, | |||||
{"XVID-MMX", 1, ff_idct_xvid_mmx, ff_ref_idct, NO_PERM, AV_CPU_FLAG_MMX}, | |||||
{"XVID-MMX2", 1, ff_idct_xvid_mmx2, ff_ref_idct, NO_PERM, AV_CPU_FLAG_MMX2}, | |||||
{"XVID-SSE2", 1, ff_idct_xvid_sse2, ff_ref_idct, SSE2_PERM, AV_CPU_FLAG_SSE2}, | |||||
#endif | #endif | ||||
#if HAVE_ALTIVEC | #if HAVE_ALTIVEC | ||||
{"altivecfdct", 0, fdct_altivec, ff_ref_fdct, NO_PERM, FF_MM_ALTIVEC}, | |||||
{"altivecfdct", 0, fdct_altivec, ff_ref_fdct, NO_PERM, AV_CPU_FLAG_ALTIVEC}, | |||||
#endif | #endif | ||||
#if ARCH_BFIN | #if ARCH_BFIN | ||||
@@ -187,7 +187,7 @@ DECLARE_ALIGNED(8, static DCTELEM, block_org)[64]; | |||||
static inline void mmx_emms(void) | static inline void mmx_emms(void) | ||||
{ | { | ||||
#if HAVE_MMX | #if HAVE_MMX | ||||
if (cpu_flags & FF_MM_MMX) | |||||
if (cpu_flags & AV_CPU_FLAG_MMX) | |||||
__asm__ volatile ("emms\n\t"); | __asm__ volatile ("emms\n\t"); | ||||
#endif | #endif | ||||
} | } | ||||
@@ -553,7 +553,7 @@ retry: | |||||
#endif | #endif | ||||
#if HAVE_MMX | #if HAVE_MMX | ||||
if(s->codec_id == CODEC_ID_MPEG4 && s->xvid_build>=0 && avctx->idct_algo == FF_IDCT_AUTO && (mm_support() & FF_MM_MMX)){ | |||||
if(s->codec_id == CODEC_ID_MPEG4 && s->xvid_build>=0 && avctx->idct_algo == FF_IDCT_AUTO && (mm_support() & AV_CPU_FLAG_MMX)){ | |||||
avctx->idct_algo= FF_IDCT_XVIDMMX; | avctx->idct_algo= FF_IDCT_XVIDMMX; | ||||
avctx->coded_width= 0; // force reinit | avctx->coded_width= 0; // force reinit | ||||
// dsputil_init(&s->dsp, avctx); | // dsputil_init(&s->dsp, avctx); | ||||
@@ -128,7 +128,7 @@ int main(int argc, char **argv) | |||||
AVCodecContext *ctx; | AVCodecContext *ctx; | ||||
int c; | int c; | ||||
DSPContext cctx, mmxctx; | DSPContext cctx, mmxctx; | ||||
int flags[2] = { FF_MM_MMX, FF_MM_MMX2 }; | |||||
int flags[2] = { AV_CPU_FLAG_MMX, AV_CPU_FLAG_MMX2 }; | |||||
int flags_size = HAVE_MMX2 ? 2 : 1; | int flags_size = HAVE_MMX2 ? 2 : 1; | ||||
for(;;) { | for(;;) { | ||||
@@ -145,11 +145,11 @@ int main(int argc, char **argv) | |||||
printf("ffmpeg motion test\n"); | printf("ffmpeg motion test\n"); | ||||
ctx = avcodec_alloc_context(); | ctx = avcodec_alloc_context(); | ||||
ctx->dsp_mask = FF_MM_FORCE; | |||||
ctx->dsp_mask = AV_CPU_FLAG_FORCE; | |||||
dsputil_init(&cctx, ctx); | dsputil_init(&cctx, ctx); | ||||
for (c = 0; c < flags_size; c++) { | for (c = 0; c < flags_size; c++) { | ||||
int x; | int x; | ||||
ctx->dsp_mask = FF_MM_FORCE | flags[c]; | |||||
ctx->dsp_mask = AV_CPU_FLAG_FORCE | flags[c]; | |||||
dsputil_init(&mmxctx, ctx); | dsputil_init(&mmxctx, ctx); | ||||
for (x = 0; x < 2; x++) { | for (x = 0; x < 2; x++) { | ||||
@@ -28,7 +28,7 @@ int mm_support(void) | |||||
int result = 0; | int result = 0; | ||||
#if HAVE_ALTIVEC | #if HAVE_ALTIVEC | ||||
if (has_altivec()) { | if (has_altivec()) { | ||||
result |= FF_MM_ALTIVEC; | |||||
result |= AV_CPU_FLAG_ALTIVEC; | |||||
} | } | ||||
#endif /* result */ | #endif /* result */ | ||||
return result; | return result; | ||||
@@ -474,6 +474,6 @@ void ff_cavsdsp_init_mmx(CAVSDSPContext *c, AVCodecContext *avctx) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX2) ff_cavsdsp_init_mmx2 (c, avctx); | |||||
if (mm_flags & FF_MM_3DNOW) ff_cavsdsp_init_3dnow(c, avctx); | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) ff_cavsdsp_init_mmx2 (c, avctx); | |||||
if (mm_flags & AV_CPU_FLAG_3DNOW) ff_cavsdsp_init_3dnow(c, avctx); | |||||
} | } |
@@ -79,21 +79,21 @@ int mm_support(void) | |||||
family = ((eax>>8)&0xf) + ((eax>>20)&0xff); | family = ((eax>>8)&0xf) + ((eax>>20)&0xff); | ||||
model = ((eax>>4)&0xf) + ((eax>>12)&0xf0); | model = ((eax>>4)&0xf) + ((eax>>12)&0xf0); | ||||
if (std_caps & (1<<23)) | if (std_caps & (1<<23)) | ||||
rval |= FF_MM_MMX; | |||||
rval |= AV_CPU_FLAG_MMX; | |||||
if (std_caps & (1<<25)) | if (std_caps & (1<<25)) | ||||
rval |= FF_MM_MMX2 | |||||
rval |= AV_CPU_FLAG_MMX2 | |||||
#if HAVE_SSE | #if HAVE_SSE | ||||
| FF_MM_SSE; | |||||
| AV_CPU_FLAG_SSE; | |||||
if (std_caps & (1<<26)) | if (std_caps & (1<<26)) | ||||
rval |= FF_MM_SSE2; | |||||
rval |= AV_CPU_FLAG_SSE2; | |||||
if (ecx & 1) | if (ecx & 1) | ||||
rval |= FF_MM_SSE3; | |||||
rval |= AV_CPU_FLAG_SSE3; | |||||
if (ecx & 0x00000200 ) | if (ecx & 0x00000200 ) | ||||
rval |= FF_MM_SSSE3; | |||||
rval |= AV_CPU_FLAG_SSSE3; | |||||
if (ecx & 0x00080000 ) | if (ecx & 0x00080000 ) | ||||
rval |= FF_MM_SSE4; | |||||
rval |= AV_CPU_FLAG_SSE4; | |||||
if (ecx & 0x00100000 ) | if (ecx & 0x00100000 ) | ||||
rval |= FF_MM_SSE42; | |||||
rval |= AV_CPU_FLAG_SSE42; | |||||
#endif | #endif | ||||
; | ; | ||||
} | } | ||||
@@ -103,13 +103,13 @@ int mm_support(void) | |||||
if(max_ext_level >= 0x80000001){ | if(max_ext_level >= 0x80000001){ | ||||
cpuid(0x80000001, eax, ebx, ecx, ext_caps); | cpuid(0x80000001, eax, ebx, ecx, ext_caps); | ||||
if (ext_caps & (1<<31)) | if (ext_caps & (1<<31)) | ||||
rval |= FF_MM_3DNOW; | |||||
rval |= AV_CPU_FLAG_3DNOW; | |||||
if (ext_caps & (1<<30)) | if (ext_caps & (1<<30)) | ||||
rval |= FF_MM_3DNOWEXT; | |||||
rval |= AV_CPU_FLAG_3DNOWEXT; | |||||
if (ext_caps & (1<<23)) | if (ext_caps & (1<<23)) | ||||
rval |= FF_MM_MMX; | |||||
rval |= AV_CPU_FLAG_MMX; | |||||
if (ext_caps & (1<<22)) | if (ext_caps & (1<<22)) | ||||
rval |= FF_MM_MMX2; | |||||
rval |= AV_CPU_FLAG_MMX2; | |||||
} | } | ||||
if (!strncmp(vendor.c, "GenuineIntel", 12) && | if (!strncmp(vendor.c, "GenuineIntel", 12) && | ||||
@@ -117,24 +117,24 @@ int mm_support(void) | |||||
/* 6/9 (pentium-m "banias"), 6/13 (pentium-m "dothan"), and 6/14 (core1 "yonah") | /* 6/9 (pentium-m "banias"), 6/13 (pentium-m "dothan"), and 6/14 (core1 "yonah") | ||||
* theoretically support sse2, but it's usually slower than mmx, | * theoretically support sse2, but it's usually slower than mmx, | ||||
* so let's just pretend they don't. */ | * so let's just pretend they don't. */ | ||||
if (rval & FF_MM_SSE2) rval ^= FF_MM_SSE2SLOW|FF_MM_SSE2; | |||||
if (rval & FF_MM_SSE3) rval ^= FF_MM_SSE3SLOW|FF_MM_SSE3; | |||||
if (rval & AV_CPU_FLAG_SSE2) rval ^= AV_CPU_FLAG_SSE2SLOW|AV_CPU_FLAG_SSE2; | |||||
if (rval & AV_CPU_FLAG_SSE3) rval ^= AV_CPU_FLAG_SSE3SLOW|AV_CPU_FLAG_SSE3; | |||||
} | } | ||||
#if 0 | #if 0 | ||||
av_log(NULL, AV_LOG_DEBUG, "%s%s%s%s%s%s%s%s%s%s%s%s\n", | av_log(NULL, AV_LOG_DEBUG, "%s%s%s%s%s%s%s%s%s%s%s%s\n", | ||||
(rval&FF_MM_MMX) ? "MMX ":"", | |||||
(rval&FF_MM_MMX2) ? "MMX2 ":"", | |||||
(rval&FF_MM_SSE) ? "SSE ":"", | |||||
(rval&FF_MM_SSE2) ? "SSE2 ":"", | |||||
(rval&FF_MM_SSE2SLOW) ? "SSE2(slow) ":"", | |||||
(rval&FF_MM_SSE3) ? "SSE3 ":"", | |||||
(rval&FF_MM_SSE3SLOW) ? "SSE3(slow) ":"", | |||||
(rval&FF_MM_SSSE3) ? "SSSE3 ":"", | |||||
(rval&FF_MM_SSE4) ? "SSE4.1 ":"", | |||||
(rval&FF_MM_SSE42) ? "SSE4.2 ":"", | |||||
(rval&FF_MM_3DNOW) ? "3DNow ":"", | |||||
(rval&FF_MM_3DNOWEXT) ? "3DNowExt ":""); | |||||
(rval&AV_CPU_FLAG_MMX) ? "MMX ":"", | |||||
(rval&AV_CPU_FLAG_MMX2) ? "MMX2 ":"", | |||||
(rval&AV_CPU_FLAG_SSE) ? "SSE ":"", | |||||
(rval&AV_CPU_FLAG_SSE2) ? "SSE2 ":"", | |||||
(rval&AV_CPU_FLAG_SSE2SLOW) ? "SSE2(slow) ":"", | |||||
(rval&AV_CPU_FLAG_SSE3) ? "SSE3 ":"", | |||||
(rval&AV_CPU_FLAG_SSE3SLOW) ? "SSE3(slow) ":"", | |||||
(rval&AV_CPU_FLAG_SSSE3) ? "SSSE3 ":"", | |||||
(rval&AV_CPU_FLAG_SSE4) ? "SSE4.1 ":"", | |||||
(rval&AV_CPU_FLAG_SSE42) ? "SSE4.2 ":"", | |||||
(rval&AV_CPU_FLAG_3DNOW) ? "3DNow ":"", | |||||
(rval&AV_CPU_FLAG_3DNOWEXT) ? "3DNowExt ":""); | |||||
#endif | #endif | ||||
return rval; | return rval; | ||||
} | } | ||||
@@ -52,7 +52,7 @@ static void get_pixels_8x4_sym_sse2(DCTELEM *block, const uint8_t *pixels, int l | |||||
void ff_dnxhd_init_mmx(DNXHDEncContext *ctx) | void ff_dnxhd_init_mmx(DNXHDEncContext *ctx) | ||||
{ | { | ||||
if (mm_support() & FF_MM_SSE2) { | |||||
if (mm_support() & AV_CPU_FLAG_SSE2) { | |||||
ctx->get_pixels_8x4_sym = get_pixels_8x4_sym_sse2; | ctx->get_pixels_8x4_sym = get_pixels_8x4_sym_sse2; | ||||
} | } | ||||
} | } |
@@ -22,6 +22,7 @@ | |||||
* MMX optimization by Nick Kurshev <nickols_k@mail.ru> | * MMX optimization by Nick Kurshev <nickols_k@mail.ru> | ||||
*/ | */ | ||||
#include "libavutil/cpu.h" | |||||
#include "libavutil/x86_cpu.h" | #include "libavutil/x86_cpu.h" | ||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
#include "libavcodec/h264dsp.h" | #include "libavcodec/h264dsp.h" | ||||
@@ -2525,7 +2526,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (avctx->dsp_mask) { | if (avctx->dsp_mask) { | ||||
if (avctx->dsp_mask & FF_MM_FORCE) | |||||
if (avctx->dsp_mask & AV_CPU_FLAG_FORCE) | |||||
mm_flags |= (avctx->dsp_mask & 0xffff); | mm_flags |= (avctx->dsp_mask & 0xffff); | ||||
else | else | ||||
mm_flags &= ~(avctx->dsp_mask & 0xffff); | mm_flags &= ~(avctx->dsp_mask & 0xffff); | ||||
@@ -2533,20 +2534,20 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
#if 0 | #if 0 | ||||
av_log(avctx, AV_LOG_INFO, "libavcodec: CPU flags:"); | av_log(avctx, AV_LOG_INFO, "libavcodec: CPU flags:"); | ||||
if (mm_flags & FF_MM_MMX) | |||||
if (mm_flags & AV_CPU_FLAG_MMX) | |||||
av_log(avctx, AV_LOG_INFO, " mmx"); | av_log(avctx, AV_LOG_INFO, " mmx"); | ||||
if (mm_flags & FF_MM_MMX2) | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) | |||||
av_log(avctx, AV_LOG_INFO, " mmx2"); | av_log(avctx, AV_LOG_INFO, " mmx2"); | ||||
if (mm_flags & FF_MM_3DNOW) | |||||
if (mm_flags & AV_CPU_FLAG_3DNOW) | |||||
av_log(avctx, AV_LOG_INFO, " 3dnow"); | av_log(avctx, AV_LOG_INFO, " 3dnow"); | ||||
if (mm_flags & FF_MM_SSE) | |||||
if (mm_flags & AV_CPU_FLAG_SSE) | |||||
av_log(avctx, AV_LOG_INFO, " sse"); | av_log(avctx, AV_LOG_INFO, " sse"); | ||||
if (mm_flags & FF_MM_SSE2) | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) | |||||
av_log(avctx, AV_LOG_INFO, " sse2"); | av_log(avctx, AV_LOG_INFO, " sse2"); | ||||
av_log(avctx, AV_LOG_INFO, "\n"); | av_log(avctx, AV_LOG_INFO, "\n"); | ||||
#endif | #endif | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
const int idct_algo= avctx->idct_algo; | const int idct_algo= avctx->idct_algo; | ||||
if(avctx->lowres==0){ | if(avctx->lowres==0){ | ||||
@@ -2557,7 +2558,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->idct_permutation_type= FF_SIMPLE_IDCT_PERM; | c->idct_permutation_type= FF_SIMPLE_IDCT_PERM; | ||||
#if CONFIG_GPL | #if CONFIG_GPL | ||||
}else if(idct_algo==FF_IDCT_LIBMPEG2MMX){ | }else if(idct_algo==FF_IDCT_LIBMPEG2MMX){ | ||||
if(mm_flags & FF_MM_MMX2){ | |||||
if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
c->idct_put= ff_libmpeg2mmx2_idct_put; | c->idct_put= ff_libmpeg2mmx2_idct_put; | ||||
c->idct_add= ff_libmpeg2mmx2_idct_add; | c->idct_add= ff_libmpeg2mmx2_idct_add; | ||||
c->idct = ff_mmxext_idct; | c->idct = ff_mmxext_idct; | ||||
@@ -2570,7 +2571,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
#endif | #endif | ||||
}else if((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_DECODER) && | }else if((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_DECODER) && | ||||
idct_algo==FF_IDCT_VP3 && HAVE_YASM){ | idct_algo==FF_IDCT_VP3 && HAVE_YASM){ | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->idct_put= ff_vp3_idct_put_sse2; | c->idct_put= ff_vp3_idct_put_sse2; | ||||
c->idct_add= ff_vp3_idct_add_sse2; | c->idct_add= ff_vp3_idct_add_sse2; | ||||
c->idct = ff_vp3_idct_sse2; | c->idct = ff_vp3_idct_sse2; | ||||
@@ -2584,12 +2585,12 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
}else if(idct_algo==FF_IDCT_CAVS){ | }else if(idct_algo==FF_IDCT_CAVS){ | ||||
c->idct_permutation_type= FF_TRANSPOSE_IDCT_PERM; | c->idct_permutation_type= FF_TRANSPOSE_IDCT_PERM; | ||||
}else if(idct_algo==FF_IDCT_XVIDMMX){ | }else if(idct_algo==FF_IDCT_XVIDMMX){ | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->idct_put= ff_idct_xvid_sse2_put; | c->idct_put= ff_idct_xvid_sse2_put; | ||||
c->idct_add= ff_idct_xvid_sse2_add; | c->idct_add= ff_idct_xvid_sse2_add; | ||||
c->idct = ff_idct_xvid_sse2; | c->idct = ff_idct_xvid_sse2; | ||||
c->idct_permutation_type= FF_SSE2_IDCT_PERM; | c->idct_permutation_type= FF_SSE2_IDCT_PERM; | ||||
}else if(mm_flags & FF_MM_MMX2){ | |||||
}else if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
c->idct_put= ff_idct_xvid_mmx2_put; | c->idct_put= ff_idct_xvid_mmx2_put; | ||||
c->idct_add= ff_idct_xvid_mmx2_add; | c->idct_add= ff_idct_xvid_mmx2_add; | ||||
c->idct = ff_idct_xvid_mmx2; | c->idct = ff_idct_xvid_mmx2; | ||||
@@ -2606,7 +2607,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->add_pixels_clamped = ff_add_pixels_clamped_mmx; | c->add_pixels_clamped = ff_add_pixels_clamped_mmx; | ||||
c->clear_block = clear_block_mmx; | c->clear_block = clear_block_mmx; | ||||
c->clear_blocks = clear_blocks_mmx; | c->clear_blocks = clear_blocks_mmx; | ||||
if ((mm_flags & FF_MM_SSE) && | |||||
if ((mm_flags & AV_CPU_FLAG_SSE) && | |||||
!(CONFIG_MPEG_XVMC_DECODER && avctx->xvmc_acceleration > 1)){ | !(CONFIG_MPEG_XVMC_DECODER && avctx->xvmc_acceleration > 1)){ | ||||
/* XvMCCreateBlocks() may not allocate 16-byte aligned blocks */ | /* XvMCCreateBlocks() may not allocate 16-byte aligned blocks */ | ||||
c->clear_block = clear_block_sse; | c->clear_block = clear_block_sse; | ||||
@@ -2649,7 +2650,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->put_rv40_chroma_pixels_tab[1]= ff_put_rv40_chroma_mc4_mmx; | c->put_rv40_chroma_pixels_tab[1]= ff_put_rv40_chroma_mc4_mmx; | ||||
#endif | #endif | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
c->prefetch = prefetch_mmx2; | c->prefetch = prefetch_mmx2; | ||||
c->put_pixels_tab[0][1] = put_pixels16_x2_mmx2; | c->put_pixels_tab[0][1] = put_pixels16_x2_mmx2; | ||||
@@ -2740,7 +2741,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_mmx2; | c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_mmx2; | ||||
#endif | #endif | ||||
#if HAVE_7REGS && HAVE_TEN_OPERANDS | #if HAVE_7REGS && HAVE_TEN_OPERANDS | ||||
if( mm_flags&FF_MM_3DNOW ) | |||||
if( mm_flags&AV_CPU_FLAG_3DNOW ) | |||||
c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov; | c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov; | ||||
#endif | #endif | ||||
@@ -2748,7 +2749,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
ff_vc1dsp_init_mmx(c, avctx); | ff_vc1dsp_init_mmx(c, avctx); | ||||
c->add_png_paeth_prediction= add_png_paeth_prediction_mmx2; | c->add_png_paeth_prediction= add_png_paeth_prediction_mmx2; | ||||
} else if (mm_flags & FF_MM_3DNOW) { | |||||
} else if (mm_flags & AV_CPU_FLAG_3DNOW) { | |||||
c->prefetch = prefetch_3dnow; | c->prefetch = prefetch_3dnow; | ||||
c->put_pixels_tab[0][1] = put_pixels16_x2_3dnow; | c->put_pixels_tab[0][1] = put_pixels16_x2_3dnow; | ||||
@@ -2816,13 +2817,13 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->put_h264_qpel_pixels_tab[1][x+y*4] = put_h264_qpel8_mc##x##y##_##CPU;\ | c->put_h264_qpel_pixels_tab[1][x+y*4] = put_h264_qpel8_mc##x##y##_##CPU;\ | ||||
c->avg_h264_qpel_pixels_tab[0][x+y*4] = avg_h264_qpel16_mc##x##y##_##CPU;\ | c->avg_h264_qpel_pixels_tab[0][x+y*4] = avg_h264_qpel16_mc##x##y##_##CPU;\ | ||||
c->avg_h264_qpel_pixels_tab[1][x+y*4] = avg_h264_qpel8_mc##x##y##_##CPU; | c->avg_h264_qpel_pixels_tab[1][x+y*4] = avg_h264_qpel8_mc##x##y##_##CPU; | ||||
if((mm_flags & FF_MM_SSE2) && !(mm_flags & FF_MM_3DNOW)){ | |||||
if((mm_flags & AV_CPU_FLAG_SSE2) && !(mm_flags & AV_CPU_FLAG_3DNOW)){ | |||||
// these functions are slower than mmx on AMD, but faster on Intel | // these functions are slower than mmx on AMD, but faster on Intel | ||||
c->put_pixels_tab[0][0] = put_pixels16_sse2; | c->put_pixels_tab[0][0] = put_pixels16_sse2; | ||||
c->avg_pixels_tab[0][0] = avg_pixels16_sse2; | c->avg_pixels_tab[0][0] = avg_pixels16_sse2; | ||||
H264_QPEL_FUNCS(0, 0, sse2); | H264_QPEL_FUNCS(0, 0, sse2); | ||||
} | } | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
H264_QPEL_FUNCS(0, 1, sse2); | H264_QPEL_FUNCS(0, 1, sse2); | ||||
H264_QPEL_FUNCS(0, 2, sse2); | H264_QPEL_FUNCS(0, 2, sse2); | ||||
H264_QPEL_FUNCS(0, 3, sse2); | H264_QPEL_FUNCS(0, 3, sse2); | ||||
@@ -2837,7 +2838,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
H264_QPEL_FUNCS(3, 3, sse2); | H264_QPEL_FUNCS(3, 3, sse2); | ||||
} | } | ||||
#if HAVE_SSSE3 | #if HAVE_SSSE3 | ||||
if(mm_flags & FF_MM_SSSE3){ | |||||
if(mm_flags & AV_CPU_FLAG_SSSE3){ | |||||
H264_QPEL_FUNCS(1, 0, ssse3); | H264_QPEL_FUNCS(1, 0, ssse3); | ||||
H264_QPEL_FUNCS(1, 1, ssse3); | H264_QPEL_FUNCS(1, 1, ssse3); | ||||
H264_QPEL_FUNCS(1, 2, ssse3); | H264_QPEL_FUNCS(1, 2, ssse3); | ||||
@@ -2859,13 +2860,13 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_ssse3; | c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_ssse3; | ||||
c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_ssse3; | c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_ssse3; | ||||
c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_ssse3; | c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_ssse3; | ||||
if (mm_flags & FF_MM_SSE4) // not really sse4, just slow on Conroe | |||||
if (mm_flags & AV_CPU_FLAG_SSE4) // not really sse4, just slow on Conroe | |||||
c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_sse4; | c->add_hfyu_left_prediction = ff_add_hfyu_left_prediction_sse4; | ||||
#endif | #endif | ||||
} | } | ||||
#endif | #endif | ||||
if(mm_flags & FF_MM_3DNOW){ | |||||
if(mm_flags & AV_CPU_FLAG_3DNOW){ | |||||
c->vorbis_inverse_coupling = vorbis_inverse_coupling_3dnow; | c->vorbis_inverse_coupling = vorbis_inverse_coupling_3dnow; | ||||
c->vector_fmul = vector_fmul_3dnow; | c->vector_fmul = vector_fmul_3dnow; | ||||
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | ||||
@@ -2873,20 +2874,20 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->float_to_int16_interleave = float_to_int16_interleave_3dnow; | c->float_to_int16_interleave = float_to_int16_interleave_3dnow; | ||||
} | } | ||||
} | } | ||||
if(mm_flags & FF_MM_3DNOWEXT){ | |||||
if(mm_flags & AV_CPU_FLAG_3DNOWEXT){ | |||||
c->vector_fmul_reverse = vector_fmul_reverse_3dnow2; | c->vector_fmul_reverse = vector_fmul_reverse_3dnow2; | ||||
c->vector_fmul_window = vector_fmul_window_3dnow2; | c->vector_fmul_window = vector_fmul_window_3dnow2; | ||||
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | ||||
c->float_to_int16_interleave = float_to_int16_interleave_3dn2; | c->float_to_int16_interleave = float_to_int16_interleave_3dn2; | ||||
} | } | ||||
} | } | ||||
if(mm_flags & FF_MM_MMX2){ | |||||
if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
#if HAVE_YASM | #if HAVE_YASM | ||||
c->scalarproduct_int16 = ff_scalarproduct_int16_mmx2; | c->scalarproduct_int16 = ff_scalarproduct_int16_mmx2; | ||||
c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_mmx2; | c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_mmx2; | ||||
#endif | #endif | ||||
} | } | ||||
if(mm_flags & FF_MM_SSE){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE){ | |||||
c->vorbis_inverse_coupling = vorbis_inverse_coupling_sse; | c->vorbis_inverse_coupling = vorbis_inverse_coupling_sse; | ||||
c->ac3_downmix = ac3_downmix_sse; | c->ac3_downmix = ac3_downmix_sse; | ||||
c->vector_fmul = vector_fmul_sse; | c->vector_fmul = vector_fmul_sse; | ||||
@@ -2901,9 +2902,9 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->scalarproduct_float = ff_scalarproduct_float_sse; | c->scalarproduct_float = ff_scalarproduct_float_sse; | ||||
#endif | #endif | ||||
} | } | ||||
if(mm_flags & FF_MM_3DNOW) | |||||
if(mm_flags & AV_CPU_FLAG_3DNOW) | |||||
c->vector_fmul_add = vector_fmul_add_3dnow; // faster than sse | c->vector_fmul_add = vector_fmul_add_3dnow; // faster than sse | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse2; | c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse2; | ||||
c->float_to_int16 = float_to_int16_sse2; | c->float_to_int16 = float_to_int16_sse2; | ||||
c->float_to_int16_interleave = float_to_int16_interleave_sse2; | c->float_to_int16_interleave = float_to_int16_interleave_sse2; | ||||
@@ -2912,7 +2913,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_sse2; | c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_sse2; | ||||
#endif | #endif | ||||
} | } | ||||
if((mm_flags & FF_MM_SSSE3) && !(mm_flags & (FF_MM_SSE42|FF_MM_3DNOW)) && HAVE_YASM) // cachesplit | |||||
if((mm_flags & AV_CPU_FLAG_SSSE3) && !(mm_flags & (AV_CPU_FLAG_SSE42|AV_CPU_FLAG_3DNOW)) && HAVE_YASM) // cachesplit | |||||
c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_ssse3; | c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_ssse3; | ||||
} | } | ||||
@@ -1352,12 +1352,12 @@ void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
const int dct_algo = avctx->dct_algo; | const int dct_algo = avctx->dct_algo; | ||||
if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){ | if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){ | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->fdct = ff_fdct_sse2; | c->fdct = ff_fdct_sse2; | ||||
}else if(mm_flags & FF_MM_MMX2){ | |||||
}else if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
c->fdct = ff_fdct_mmx2; | c->fdct = ff_fdct_mmx2; | ||||
}else{ | }else{ | ||||
c->fdct = ff_fdct_mmx; | c->fdct = ff_fdct_mmx; | ||||
@@ -1375,7 +1375,7 @@ void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->hadamard8_diff[1]= hadamard8_diff_mmx; | c->hadamard8_diff[1]= hadamard8_diff_mmx; | ||||
c->pix_norm1 = pix_norm1_mmx; | c->pix_norm1 = pix_norm1_mmx; | ||||
c->sse[0] = (mm_flags & FF_MM_SSE2) ? sse16_sse2 : sse16_mmx; | |||||
c->sse[0] = (mm_flags & AV_CPU_FLAG_SSE2) ? sse16_sse2 : sse16_mmx; | |||||
c->sse[1] = sse8_mmx; | c->sse[1] = sse8_mmx; | ||||
c->vsad[4]= vsad_intra16_mmx; | c->vsad[4]= vsad_intra16_mmx; | ||||
@@ -1393,7 +1393,7 @@ void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->ssd_int8_vs_int16 = ssd_int8_vs_int16_mmx; | c->ssd_int8_vs_int16 = ssd_int8_vs_int16_mmx; | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
c->sum_abs_dctelem= sum_abs_dctelem_mmx2; | c->sum_abs_dctelem= sum_abs_dctelem_mmx2; | ||||
c->hadamard8_diff[0]= hadamard8_diff16_mmx2; | c->hadamard8_diff[0]= hadamard8_diff16_mmx2; | ||||
c->hadamard8_diff[1]= hadamard8_diff_mmx2; | c->hadamard8_diff[1]= hadamard8_diff_mmx2; | ||||
@@ -1406,19 +1406,19 @@ void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_mmx2; | c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_mmx2; | ||||
} | } | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->get_pixels = get_pixels_sse2; | c->get_pixels = get_pixels_sse2; | ||||
c->sum_abs_dctelem= sum_abs_dctelem_sse2; | c->sum_abs_dctelem= sum_abs_dctelem_sse2; | ||||
c->hadamard8_diff[0]= hadamard8_diff16_sse2; | c->hadamard8_diff[0]= hadamard8_diff16_sse2; | ||||
c->hadamard8_diff[1]= hadamard8_diff_sse2; | c->hadamard8_diff[1]= hadamard8_diff_sse2; | ||||
} | } | ||||
if (CONFIG_LPC && mm_flags & (FF_MM_SSE2|FF_MM_SSE2SLOW)) { | |||||
if (CONFIG_LPC && mm_flags & (AV_CPU_FLAG_SSE2|AV_CPU_FLAG_SSE2SLOW)) { | |||||
c->lpc_compute_autocorr = ff_lpc_compute_autocorr_sse2; | c->lpc_compute_autocorr = ff_lpc_compute_autocorr_sse2; | ||||
} | } | ||||
#if HAVE_SSSE3 | #if HAVE_SSSE3 | ||||
if(mm_flags & FF_MM_SSSE3){ | |||||
if(mm_flags & AV_CPU_FLAG_SSSE3){ | |||||
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | ||||
c->try_8x8basis= try_8x8basis_ssse3; | c->try_8x8basis= try_8x8basis_ssse3; | ||||
} | } | ||||
@@ -1429,7 +1429,7 @@ void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
} | } | ||||
#endif | #endif | ||||
if(mm_flags & FF_MM_3DNOW){ | |||||
if(mm_flags & AV_CPU_FLAG_3DNOW){ | |||||
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | if(!(avctx->flags & CODEC_FLAG_BITEXACT)){ | ||||
c->try_8x8basis= try_8x8basis_3dnow; | c->try_8x8basis= try_8x8basis_3dnow; | ||||
} | } | ||||
@@ -23,18 +23,18 @@ av_cold void ff_fft_init_mmx(FFTContext *s) | |||||
{ | { | ||||
#if HAVE_YASM | #if HAVE_YASM | ||||
int has_vectors = mm_support(); | int has_vectors = mm_support(); | ||||
if (has_vectors & FF_MM_SSE && HAVE_SSE) { | |||||
if (has_vectors & AV_CPU_FLAG_SSE && HAVE_SSE) { | |||||
/* SSE for P3/P4/K8 */ | /* SSE for P3/P4/K8 */ | ||||
s->imdct_calc = ff_imdct_calc_sse; | s->imdct_calc = ff_imdct_calc_sse; | ||||
s->imdct_half = ff_imdct_half_sse; | s->imdct_half = ff_imdct_half_sse; | ||||
s->fft_permute = ff_fft_permute_sse; | s->fft_permute = ff_fft_permute_sse; | ||||
s->fft_calc = ff_fft_calc_sse; | s->fft_calc = ff_fft_calc_sse; | ||||
} else if (has_vectors & FF_MM_3DNOWEXT && HAVE_AMD3DNOWEXT) { | |||||
} else if (has_vectors & AV_CPU_FLAG_3DNOWEXT && HAVE_AMD3DNOWEXT) { | |||||
/* 3DNowEx for K7 */ | /* 3DNowEx for K7 */ | ||||
s->imdct_calc = ff_imdct_calc_3dn2; | s->imdct_calc = ff_imdct_calc_3dn2; | ||||
s->imdct_half = ff_imdct_half_3dn2; | s->imdct_half = ff_imdct_half_3dn2; | ||||
s->fft_calc = ff_fft_calc_3dn2; | s->fft_calc = ff_fft_calc_3dn2; | ||||
} else if (has_vectors & FF_MM_3DNOW && HAVE_AMD3DNOW) { | |||||
} else if (has_vectors & AV_CPU_FLAG_3DNOW && HAVE_AMD3DNOW) { | |||||
/* 3DNow! for K6-2/3 */ | /* 3DNow! for K6-2/3 */ | ||||
s->imdct_calc = ff_imdct_calc_3dn; | s->imdct_calc = ff_imdct_calc_3dn; | ||||
s->imdct_half = ff_imdct_half_3dn; | s->imdct_half = ff_imdct_half_3dn; | ||||
@@ -47,7 +47,7 @@ av_cold void ff_fft_init_mmx(FFTContext *s) | |||||
av_cold void ff_dct_init_mmx(DCTContext *s) | av_cold void ff_dct_init_mmx(DCTContext *s) | ||||
{ | { | ||||
int has_vectors = mm_support(); | int has_vectors = mm_support(); | ||||
if (has_vectors & FF_MM_SSE && HAVE_SSE) | |||||
if (has_vectors & AV_CPU_FLAG_SSE && HAVE_SSE) | |||||
s->dct32 = ff_dct32_float_sse; | s->dct32 = ff_dct32_float_sse; | ||||
} | } | ||||
#endif | #endif | ||||
@@ -51,7 +51,7 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id) | |||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
#if HAVE_YASM | #if HAVE_YASM | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
h->pred16x16[VERT_PRED8x8] = ff_pred16x16_vertical_mmx; | h->pred16x16[VERT_PRED8x8] = ff_pred16x16_vertical_mmx; | ||||
h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmx; | h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmx; | ||||
h->pred8x8 [VERT_PRED8x8] = ff_pred8x8_vertical_mmx; | h->pred8x8 [VERT_PRED8x8] = ff_pred8x8_vertical_mmx; | ||||
@@ -63,7 +63,7 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id) | |||||
} | } | ||||
} | } | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmxext; | h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmxext; | ||||
h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_mmxext; | h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_mmxext; | ||||
h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_mmxext; | h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_mmxext; | ||||
@@ -77,11 +77,11 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id) | |||||
} | } | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE) { | |||||
h->pred16x16[VERT_PRED8x8] = ff_pred16x16_vertical_sse; | h->pred16x16[VERT_PRED8x8] = ff_pred16x16_vertical_sse; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_sse2; | h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_sse2; | ||||
if (codec_id == CODEC_ID_VP8) { | if (codec_id == CODEC_ID_VP8) { | ||||
h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_tm_vp8_sse2; | h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_tm_vp8_sse2; | ||||
@@ -89,7 +89,7 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id) | |||||
} | } | ||||
} | } | ||||
if (mm_flags & FF_MM_SSSE3) { | |||||
if (mm_flags & AV_CPU_FLAG_SSSE3) { | |||||
h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_ssse3; | h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_ssse3; | ||||
h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_ssse3; | h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_ssse3; | ||||
h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_ssse3; | h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_ssse3; | ||||
@@ -744,7 +744,7 @@ void ff_h264dsp_init_x86(H264DSPContext *c) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
c->h264_idct_dc_add= | c->h264_idct_dc_add= | ||||
c->h264_idct_add= ff_h264_idct_add_mmx; | c->h264_idct_add= ff_h264_idct_add_mmx; | ||||
c->h264_idct8_dc_add= | c->h264_idct8_dc_add= | ||||
@@ -755,7 +755,7 @@ void ff_h264dsp_init_x86(H264DSPContext *c) | |||||
c->h264_idct_add8 = ff_h264_idct_add8_mmx; | c->h264_idct_add8 = ff_h264_idct_add8_mmx; | ||||
c->h264_idct_add16intra= ff_h264_idct_add16intra_mmx; | c->h264_idct_add16intra= ff_h264_idct_add16intra_mmx; | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
c->h264_idct_dc_add= ff_h264_idct_dc_add_mmx2; | c->h264_idct_dc_add= ff_h264_idct_dc_add_mmx2; | ||||
c->h264_idct8_dc_add= ff_h264_idct8_dc_add_mmx2; | c->h264_idct8_dc_add= ff_h264_idct8_dc_add_mmx2; | ||||
c->h264_idct_add16 = ff_h264_idct_add16_mmx2; | c->h264_idct_add16 = ff_h264_idct_add16_mmx2; | ||||
@@ -765,13 +765,13 @@ void ff_h264dsp_init_x86(H264DSPContext *c) | |||||
c->h264_loop_filter_strength= h264_loop_filter_strength_mmx2; | c->h264_loop_filter_strength= h264_loop_filter_strength_mmx2; | ||||
} | } | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
c->h264_idct8_add = ff_h264_idct8_add_sse2; | c->h264_idct8_add = ff_h264_idct8_add_sse2; | ||||
c->h264_idct8_add4= ff_h264_idct8_add4_sse2; | c->h264_idct8_add4= ff_h264_idct8_add4_sse2; | ||||
} | } | ||||
#if HAVE_YASM | #if HAVE_YASM | ||||
if (mm_flags & FF_MM_MMX2){ | |||||
if (mm_flags & AV_CPU_FLAG_MMX2){ | |||||
c->h264_v_loop_filter_chroma= ff_x264_deblock_v_chroma_mmxext; | c->h264_v_loop_filter_chroma= ff_x264_deblock_v_chroma_mmxext; | ||||
c->h264_h_loop_filter_chroma= ff_x264_deblock_h_chroma_mmxext; | c->h264_h_loop_filter_chroma= ff_x264_deblock_h_chroma_mmxext; | ||||
c->h264_v_loop_filter_chroma_intra= ff_x264_deblock_v_chroma_intra_mmxext; | c->h264_v_loop_filter_chroma_intra= ff_x264_deblock_v_chroma_intra_mmxext; | ||||
@@ -800,7 +800,7 @@ void ff_h264dsp_init_x86(H264DSPContext *c) | |||||
c->biweight_h264_pixels_tab[6]= ff_h264_biweight_4x4_mmx2; | c->biweight_h264_pixels_tab[6]= ff_h264_biweight_4x4_mmx2; | ||||
c->biweight_h264_pixels_tab[7]= ff_h264_biweight_4x2_mmx2; | c->biweight_h264_pixels_tab[7]= ff_h264_biweight_4x2_mmx2; | ||||
if( mm_flags&FF_MM_SSE2 ){ | |||||
if (mm_flags&AV_CPU_FLAG_SSE2) { | |||||
c->weight_h264_pixels_tab[0]= ff_h264_weight_16x16_sse2; | c->weight_h264_pixels_tab[0]= ff_h264_weight_16x16_sse2; | ||||
c->weight_h264_pixels_tab[1]= ff_h264_weight_16x8_sse2; | c->weight_h264_pixels_tab[1]= ff_h264_weight_16x8_sse2; | ||||
c->weight_h264_pixels_tab[2]= ff_h264_weight_8x16_sse2; | c->weight_h264_pixels_tab[2]= ff_h264_weight_8x16_sse2; | ||||
@@ -825,7 +825,7 @@ void ff_h264dsp_init_x86(H264DSPContext *c) | |||||
c->h264_idct_add16intra = ff_h264_idct_add16intra_sse2; | c->h264_idct_add16intra = ff_h264_idct_add16intra_sse2; | ||||
#endif | #endif | ||||
} | } | ||||
if ( mm_flags&FF_MM_SSSE3 ){ | |||||
if (mm_flags&AV_CPU_FLAG_SSSE3) { | |||||
c->biweight_h264_pixels_tab[0]= ff_h264_biweight_16x16_ssse3; | c->biweight_h264_pixels_tab[0]= ff_h264_biweight_16x16_ssse3; | ||||
c->biweight_h264_pixels_tab[1]= ff_h264_biweight_16x8_ssse3; | c->biweight_h264_pixels_tab[1]= ff_h264_biweight_16x8_ssse3; | ||||
c->biweight_h264_pixels_tab[2]= ff_h264_biweight_8x16_ssse3; | c->biweight_h264_pixels_tab[2]= ff_h264_biweight_8x16_ssse3; | ||||
@@ -429,7 +429,7 @@ void dsputil_init_pix_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
c->pix_abs[0][0] = sad16_mmx; | c->pix_abs[0][0] = sad16_mmx; | ||||
c->pix_abs[0][1] = sad16_x2_mmx; | c->pix_abs[0][1] = sad16_x2_mmx; | ||||
c->pix_abs[0][2] = sad16_y2_mmx; | c->pix_abs[0][2] = sad16_y2_mmx; | ||||
@@ -442,7 +442,7 @@ void dsputil_init_pix_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->sad[0]= sad16_mmx; | c->sad[0]= sad16_mmx; | ||||
c->sad[1]= sad8_mmx; | c->sad[1]= sad8_mmx; | ||||
} | } | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
c->pix_abs[0][0] = sad16_mmx2; | c->pix_abs[0][0] = sad16_mmx2; | ||||
c->pix_abs[1][0] = sad8_mmx2; | c->pix_abs[1][0] = sad8_mmx2; | ||||
@@ -458,7 +458,7 @@ void dsputil_init_pix_mmx(DSPContext* c, AVCodecContext *avctx) | |||||
c->pix_abs[1][3] = sad8_xy2_mmx2; | c->pix_abs[1][3] = sad8_xy2_mmx2; | ||||
} | } | ||||
} | } | ||||
if ((mm_flags & FF_MM_SSE2) && !(mm_flags & FF_MM_3DNOW) && avctx->codec_id != CODEC_ID_SNOW) { | |||||
if ((mm_flags & AV_CPU_FLAG_SSE2) && !(mm_flags & AV_CPU_FLAG_3DNOW) && avctx->codec_id != CODEC_ID_SNOW) { | |||||
c->sad[0]= sad16_sse2; | c->sad[0]= sad16_sse2; | ||||
} | } | ||||
} | } |
@@ -151,7 +151,7 @@ void ff_mpegaudiodec_init_mmx(MPADecodeContext *s) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
s->apply_window_mp3 = apply_window_mp3; | s->apply_window_mp3 = apply_window_mp3; | ||||
} | } | ||||
} | } |
@@ -627,7 +627,7 @@ void MPV_common_init_mmx(MpegEncContext *s) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
const int dct_algo = s->avctx->dct_algo; | const int dct_algo = s->avctx->dct_algo; | ||||
s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_mmx; | s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_mmx; | ||||
@@ -638,7 +638,7 @@ void MPV_common_init_mmx(MpegEncContext *s) | |||||
s->dct_unquantize_mpeg2_intra = dct_unquantize_mpeg2_intra_mmx; | s->dct_unquantize_mpeg2_intra = dct_unquantize_mpeg2_intra_mmx; | ||||
s->dct_unquantize_mpeg2_inter = dct_unquantize_mpeg2_inter_mmx; | s->dct_unquantize_mpeg2_inter = dct_unquantize_mpeg2_inter_mmx; | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
s->denoise_dct= denoise_dct_sse2; | s->denoise_dct= denoise_dct_sse2; | ||||
} else { | } else { | ||||
s->denoise_dct= denoise_dct_mmx; | s->denoise_dct= denoise_dct_mmx; | ||||
@@ -646,13 +646,13 @@ void MPV_common_init_mmx(MpegEncContext *s) | |||||
if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){ | if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){ | ||||
#if HAVE_SSSE3 | #if HAVE_SSSE3 | ||||
if(mm_flags & FF_MM_SSSE3){ | |||||
if(mm_flags & AV_CPU_FLAG_SSSE3){ | |||||
s->dct_quantize= dct_quantize_SSSE3; | s->dct_quantize= dct_quantize_SSSE3; | ||||
} else | } else | ||||
#endif | #endif | ||||
if(mm_flags & FF_MM_SSE2){ | |||||
if(mm_flags & AV_CPU_FLAG_SSE2){ | |||||
s->dct_quantize= dct_quantize_SSE2; | s->dct_quantize= dct_quantize_SSE2; | ||||
} else if(mm_flags & FF_MM_MMX2){ | |||||
} else if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
s->dct_quantize= dct_quantize_MMX2; | s->dct_quantize= dct_quantize_MMX2; | ||||
} else { | } else { | ||||
s->dct_quantize= dct_quantize_MMX; | s->dct_quantize= dct_quantize_MMX; | ||||
@@ -19,6 +19,7 @@ | |||||
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||||
*/ | */ | ||||
#include "libavutil/cpu.h" | |||||
#include "libavutil/x86_cpu.h" | #include "libavutil/x86_cpu.h" | ||||
#include "libavcodec/avcodec.h" | #include "libavcodec/avcodec.h" | ||||
#include "libavcodec/snow.h" | #include "libavcodec/snow.h" | ||||
@@ -876,8 +877,8 @@ void ff_dwt_init_x86(DWTContext *c) | |||||
{ | { | ||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if(mm_flags & FF_MM_SSE2 & 0){ | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
if(mm_flags & AV_CPU_FLAG_SSE2 & 0){ | |||||
c->horizontal_compose97i = ff_snow_horizontal_compose97i_sse2; | c->horizontal_compose97i = ff_snow_horizontal_compose97i_sse2; | ||||
#if HAVE_7REGS | #if HAVE_7REGS | ||||
c->vertical_compose97i = ff_snow_vertical_compose97i_sse2; | c->vertical_compose97i = ff_snow_vertical_compose97i_sse2; | ||||
@@ -885,7 +886,7 @@ void ff_dwt_init_x86(DWTContext *c) | |||||
c->inner_add_yblock = ff_snow_inner_add_yblock_sse2; | c->inner_add_yblock = ff_snow_inner_add_yblock_sse2; | ||||
} | } | ||||
else{ | else{ | ||||
if(mm_flags & FF_MM_MMX2){ | |||||
if(mm_flags & AV_CPU_FLAG_MMX2){ | |||||
c->horizontal_compose97i = ff_snow_horizontal_compose97i_mmx; | c->horizontal_compose97i = ff_snow_horizontal_compose97i_mmx; | ||||
#if HAVE_7REGS | #if HAVE_7REGS | ||||
c->vertical_compose97i = ff_snow_vertical_compose97i_mmx; | c->vertical_compose97i = ff_snow_vertical_compose97i_mmx; | ||||
@@ -736,7 +736,7 @@ void ff_vc1dsp_init_mmx(DSPContext* dsp, AVCodecContext *avctx) { | |||||
dsp->put_vc1_mspel_pixels_tab[11] = put_vc1_mspel_mc32_mmx; | dsp->put_vc1_mspel_pixels_tab[11] = put_vc1_mspel_mc32_mmx; | ||||
dsp->put_vc1_mspel_pixels_tab[15] = put_vc1_mspel_mc33_mmx; | dsp->put_vc1_mspel_pixels_tab[15] = put_vc1_mspel_mc33_mmx; | ||||
if (mm_flags & FF_MM_MMX2){ | |||||
if (mm_flags & AV_CPU_FLAG_MMX2){ | |||||
dsp->avg_vc1_mspel_pixels_tab[ 0] = ff_avg_vc1_mspel_mc00_mmx2; | dsp->avg_vc1_mspel_pixels_tab[ 0] = ff_avg_vc1_mspel_mc00_mmx2; | ||||
dsp->avg_vc1_mspel_pixels_tab[ 4] = avg_vc1_mspel_mc01_mmx2; | dsp->avg_vc1_mspel_pixels_tab[ 4] = avg_vc1_mspel_mc01_mmx2; | ||||
dsp->avg_vc1_mspel_pixels_tab[ 8] = avg_vc1_mspel_mc02_mmx2; | dsp->avg_vc1_mspel_pixels_tab[ 8] = avg_vc1_mspel_mc02_mmx2; | ||||
@@ -772,23 +772,23 @@ void ff_vc1dsp_init_mmx(DSPContext* dsp, AVCodecContext *avctx) { | |||||
dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_ ## EXT | dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_ ## EXT | ||||
#if HAVE_YASM | #if HAVE_YASM | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
ASSIGN_LF(mmx); | ASSIGN_LF(mmx); | ||||
} | } | ||||
return; | return; | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
ASSIGN_LF(mmx2); | ASSIGN_LF(mmx2); | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_sse2; | dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_sse2; | ||||
dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse2; | dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse2; | ||||
dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_sse2; | dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_sse2; | ||||
dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse2; | dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse2; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSSE3) { | |||||
if (mm_flags & AV_CPU_FLAG_SSSE3) { | |||||
ASSIGN_LF(ssse3); | ASSIGN_LF(ssse3); | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE4) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE4) { | |||||
dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse4; | dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse4; | ||||
dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse4; | dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse4; | ||||
} | } | ||||
@@ -35,11 +35,11 @@ av_cold void ff_vp56dsp_init_x86(VP56DSPContext* c, enum CodecID codec) | |||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
if (CONFIG_VP6_DECODER && codec == CODEC_ID_VP6) { | if (CONFIG_VP6_DECODER && codec == CODEC_ID_VP6) { | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
c->vp6_filter_diag4 = ff_vp6_filter_diag4_mmx; | c->vp6_filter_diag4 = ff_vp6_filter_diag4_mmx; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
c->vp6_filter_diag4 = ff_vp6_filter_diag4_sse2; | c->vp6_filter_diag4 = ff_vp6_filter_diag4_sse2; | ||||
} | } | ||||
} | } | ||||
@@ -285,7 +285,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
int mm_flags = mm_support(); | int mm_flags = mm_support(); | ||||
#if HAVE_YASM | #if HAVE_YASM | ||||
if (mm_flags & FF_MM_MMX) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX) { | |||||
c->vp8_idct_dc_add = ff_vp8_idct_dc_add_mmx; | c->vp8_idct_dc_add = ff_vp8_idct_dc_add_mmx; | ||||
c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_mmx; | c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_mmx; | ||||
c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx; | c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx; | ||||
@@ -312,7 +312,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
/* note that 4-tap width=16 functions are missing because w=16 | /* note that 4-tap width=16 functions are missing because w=16 | ||||
* is only used for luma, and luma is always a copy or sixtap. */ | * is only used for luma, and luma is always a copy or sixtap. */ | ||||
if (mm_flags & FF_MM_MMX2) { | |||||
if (mm_flags & AV_CPU_FLAG_MMX2) { | |||||
VP8_LUMA_MC_FUNC(0, 16, mmxext); | VP8_LUMA_MC_FUNC(0, 16, mmxext); | ||||
VP8_MC_FUNC(1, 8, mmxext); | VP8_MC_FUNC(1, 8, mmxext); | ||||
VP8_MC_FUNC(2, 4, mmxext); | VP8_MC_FUNC(2, 4, mmxext); | ||||
@@ -334,14 +334,14 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmxext; | c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmxext; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE) { | |||||
c->vp8_idct_add = ff_vp8_idct_add_sse; | c->vp8_idct_add = ff_vp8_idct_add_sse; | ||||
c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_sse; | c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_sse; | ||||
c->put_vp8_epel_pixels_tab[0][0][0] = | c->put_vp8_epel_pixels_tab[0][0][0] = | ||||
c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse; | c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse; | ||||
} | } | ||||
if (mm_flags & (FF_MM_SSE2|FF_MM_SSE2SLOW)) { | |||||
if (mm_flags & (AV_CPU_FLAG_SSE2|AV_CPU_FLAG_SSE2SLOW)) { | |||||
VP8_LUMA_MC_FUNC(0, 16, sse2); | VP8_LUMA_MC_FUNC(0, 16, sse2); | ||||
VP8_MC_FUNC(1, 8, sse2); | VP8_MC_FUNC(1, 8, sse2); | ||||
VP8_BILINEAR_MC_FUNC(0, 16, sse2); | VP8_BILINEAR_MC_FUNC(0, 16, sse2); | ||||
@@ -356,7 +356,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_sse2; | c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_sse2; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE2) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE2) { | |||||
c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_sse2; | c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_sse2; | ||||
c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2; | c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2; | ||||
@@ -368,7 +368,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_sse2; | c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_sse2; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSSE3) { | |||||
if (mm_flags & AV_CPU_FLAG_SSSE3) { | |||||
VP8_LUMA_MC_FUNC(0, 16, ssse3); | VP8_LUMA_MC_FUNC(0, 16, ssse3); | ||||
VP8_MC_FUNC(1, 8, ssse3); | VP8_MC_FUNC(1, 8, ssse3); | ||||
VP8_MC_FUNC(2, 4, ssse3); | VP8_MC_FUNC(2, 4, ssse3); | ||||
@@ -390,7 +390,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) | |||||
c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_ssse3; | c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_ssse3; | ||||
} | } | ||||
if (mm_flags & FF_MM_SSE4) { | |||||
if (mm_flags & AV_CPU_FLAG_SSE4) { | |||||
c->vp8_idct_dc_add = ff_vp8_idct_dc_add_sse4; | c->vp8_idct_dc_add = ff_vp8_idct_dc_add_sse4; | ||||
c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse4; | c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse4; | ||||
@@ -9,6 +9,7 @@ HEADERS = adler32.h \ | |||||
base64.h \ | base64.h \ | ||||
bswap.h \ | bswap.h \ | ||||
common.h \ | common.h \ | ||||
cpu.h \ | |||||
crc.h \ | crc.h \ | ||||
error.h \ | error.h \ | ||||
eval.h \ | eval.h \ | ||||
@@ -40,7 +40,7 @@ | |||||
#define AV_VERSION(a, b, c) AV_VERSION_DOT(a, b, c) | #define AV_VERSION(a, b, c) AV_VERSION_DOT(a, b, c) | ||||
#define LIBAVUTIL_VERSION_MAJOR 50 | #define LIBAVUTIL_VERSION_MAJOR 50 | ||||
#define LIBAVUTIL_VERSION_MINOR 24 | |||||
#define LIBAVUTIL_VERSION_MINOR 25 | |||||
#define LIBAVUTIL_VERSION_MICRO 0 | #define LIBAVUTIL_VERSION_MICRO 0 | ||||
#define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \ | #define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \ | ||||
@@ -0,0 +1,44 @@ | |||||
/* | |||||
* Copyright (c) 2000, 2001, 2002 Fabrice Bellard | |||||
* | |||||
* This file is part of FFmpeg. | |||||
* | |||||
* FFmpeg is free software; you can redistribute it and/or | |||||
* modify it under the terms of the GNU Lesser General Public | |||||
* License as published by the Free Software Foundation; either | |||||
* version 2.1 of the License, or (at your option) any later version. | |||||
* | |||||
* FFmpeg is distributed in the hope that it will be useful, | |||||
* but WITHOUT ANY WARRANTY; without even the implied warranty of | |||||
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |||||
* Lesser General Public License for more details. | |||||
* | |||||
* You should have received a copy of the GNU Lesser General Public | |||||
* License along with FFmpeg; if not, write to the Free Software | |||||
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |||||
*/ | |||||
#ifndef AVUTIL_CPU_H | |||||
#define AVUTIL_CPU_H | |||||
#include "avutil.h" | |||||
#define AV_CPU_FLAG_FORCE 0x80000000 /* force usage of selected flags (OR) */ | |||||
/* lower 16 bits - CPU features */ | |||||
#define AV_CPU_FLAG_MMX 0x0001 ///< standard MMX | |||||
#define AV_CPU_FLAG_3DNOW 0x0004 ///< AMD 3DNOW | |||||
#define AV_CPU_FLAG_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext | |||||
#define AV_CPU_FLAG_SSE 0x0008 ///< SSE functions | |||||
#define AV_CPU_FLAG_SSE2 0x0010 ///< PIV SSE2 functions | |||||
#define AV_CPU_FLAG_SSE2SLOW 0x40000000 ///< SSE2 supported, but usually not faster | |||||
#define AV_CPU_FLAG_3DNOWEXT 0x0020 ///< AMD 3DNowExt | |||||
#define AV_CPU_FLAG_SSE3 0x0040 ///< Prescott SSE3 functions | |||||
#define AV_CPU_FLAG_SSE3SLOW 0x20000000 ///< SSE3 supported, but usually not faster t | |||||
#define AV_CPU_FLAG_SSSE3 0x0080 ///< Conroe SSSE3 functions | |||||
#define AV_CPU_FLAG_SSE4 0x0100 ///< Penryn SSE4.1 functions | |||||
#define AV_CPU_FLAG_SSE42 0x0200 ///< Nehalem SSE4.2 functions | |||||
#define AV_CPU_FLAG_IWMMXT 0x0100 ///< XScale IWMMXT | |||||
#define AV_CPU_FLAG_ALTIVEC 0x0001 ///< standard | |||||
#endif /* AVUTIL_CPU_H */ |