* Copyright (c) 2000, 2001 Fabrice Bellard
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
*
- * This file is part of FFmpeg.
+ * This file is part of Libav.
*
- * FFmpeg is free software; you can redistribute it and/or
+ * Libav 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,
+ * Libav 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
+ * License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*
* MMX optimization by Nick Kurshev <nickols_k@mail.ru>
*/
#include "libavutil/cpu.h"
-#include "libavutil/x86_cpu.h"
+#include "libavutil/x86/asm.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/mpegvideo.h"
#include "libavcodec/mathops.h"
#include "dsputil_mmx.h"
+#if HAVE_INLINE_ASM
+
static void get_pixels_mmx(DCTELEM *block, const uint8_t *pixels, int line_size)
{
__asm__ volatile(
"mov $-128, %%"REG_a" \n\t"
"pxor %%mm7, %%mm7 \n\t"
- ASMALIGN(4)
+ ".p2align 4 \n\t"
"1: \n\t"
"movq (%0), %%mm0 \n\t"
"movq (%0, %2), %%mm2 \n\t"
static void get_pixels_sse2(DCTELEM *block, const uint8_t *pixels, int line_size)
{
__asm__ volatile(
- "pxor %%xmm7, %%xmm7 \n\t"
+ "pxor %%xmm4, %%xmm4 \n\t"
"movq (%0), %%xmm0 \n\t"
"movq (%0, %2), %%xmm1 \n\t"
"movq (%0, %2,2), %%xmm2 \n\t"
"movq (%0, %3), %%xmm3 \n\t"
"lea (%0,%2,4), %0 \n\t"
- "punpcklbw %%xmm7, %%xmm0 \n\t"
- "punpcklbw %%xmm7, %%xmm1 \n\t"
- "punpcklbw %%xmm7, %%xmm2 \n\t"
- "punpcklbw %%xmm7, %%xmm3 \n\t"
+ "punpcklbw %%xmm4, %%xmm0 \n\t"
+ "punpcklbw %%xmm4, %%xmm1 \n\t"
+ "punpcklbw %%xmm4, %%xmm2 \n\t"
+ "punpcklbw %%xmm4, %%xmm3 \n\t"
"movdqa %%xmm0, (%1) \n\t"
"movdqa %%xmm1, 16(%1) \n\t"
"movdqa %%xmm2, 32(%1) \n\t"
"movq (%0, %2), %%xmm1 \n\t"
"movq (%0, %2,2), %%xmm2 \n\t"
"movq (%0, %3), %%xmm3 \n\t"
- "punpcklbw %%xmm7, %%xmm0 \n\t"
- "punpcklbw %%xmm7, %%xmm1 \n\t"
- "punpcklbw %%xmm7, %%xmm2 \n\t"
- "punpcklbw %%xmm7, %%xmm3 \n\t"
+ "punpcklbw %%xmm4, %%xmm0 \n\t"
+ "punpcklbw %%xmm4, %%xmm1 \n\t"
+ "punpcklbw %%xmm4, %%xmm2 \n\t"
+ "punpcklbw %%xmm4, %%xmm3 \n\t"
"movdqa %%xmm0, 64(%1) \n\t"
"movdqa %%xmm1, 80(%1) \n\t"
"movdqa %%xmm2, 96(%1) \n\t"
__asm__ volatile(
"pxor %%mm7, %%mm7 \n\t"
"mov $-128, %%"REG_a" \n\t"
- ASMALIGN(4)
+ ".p2align 4 \n\t"
"1: \n\t"
"movq (%0), %%mm0 \n\t"
"movq (%1), %%mm2 \n\t"
return tmp;
}
-int ff_sse16_sse2(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h);
-
static int hf_noise8_mmx(uint8_t * pix1, int line_size, int h) {
int tmp;
__asm__ volatile (
"paddusw "#t", "#a" \n\t"\
"movd "#a", "#dst" \n\t"\
-#define hadamard_func(cpu) \
-int ff_hadamard8_diff_##cpu (void *s, uint8_t *src1, uint8_t *src2, \
- int stride, int h); \
-int ff_hadamard8_diff16_##cpu(void *s, uint8_t *src1, uint8_t *src2, \
- int stride, int h);
-
-hadamard_func(mmx)
-hadamard_func(mmx2)
-hadamard_func(sse2)
-hadamard_func(ssse3)
-
#define DCT_SAD4(m,mm,o)\
"mov"#m" "#o"+ 0(%1), "#mm"2 \n\t"\
"mov"#m" "#o"+16(%1), "#mm"3 \n\t"\
#undef PHADDD
#endif //HAVE_SSSE3
+#endif /* HAVE_INLINE_ASM */
+
+int ff_sse16_sse2(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h);
+
+#define hadamard_func(cpu) \
+int ff_hadamard8_diff_##cpu (void *s, uint8_t *src1, uint8_t *src2, \
+ int stride, int h); \
+int ff_hadamard8_diff16_##cpu(void *s, uint8_t *src1, uint8_t *src2, \
+ int stride, int h);
+
+hadamard_func(mmx)
+hadamard_func(mmx2)
+hadamard_func(sse2)
+hadamard_func(ssse3)
-void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx)
+void ff_dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx)
{
int mm_flags = av_get_cpu_flags();
+#if HAVE_INLINE_ASM
+ int bit_depth = avctx->bits_per_raw_sample;
+
if (mm_flags & AV_CPU_FLAG_MMX) {
const int dct_algo = avctx->dct_algo;
- if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){
+ if (avctx->bits_per_raw_sample <= 8 &&
+ (dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX)) {
if(mm_flags & AV_CPU_FLAG_SSE2){
c->fdct = ff_fdct_sse2;
- }else if(mm_flags & AV_CPU_FLAG_MMX2){
+ } else if (mm_flags & AV_CPU_FLAG_MMXEXT) {
c->fdct = ff_fdct_mmx2;
}else{
c->fdct = ff_fdct_mmx;
}
}
- c->get_pixels = get_pixels_mmx;
+ if (bit_depth <= 8)
+ c->get_pixels = get_pixels_mmx;
c->diff_pixels = diff_pixels_mmx;
c->pix_sum = pix_sum16_mmx;
c->diff_bytes= diff_bytes_mmx;
c->sum_abs_dctelem= sum_abs_dctelem_mmx;
- c->hadamard8_diff[0]= ff_hadamard8_diff16_mmx;
- c->hadamard8_diff[1]= ff_hadamard8_diff_mmx;
-
c->pix_norm1 = pix_norm1_mmx;
- c->sse[0] = (mm_flags & AV_CPU_FLAG_SSE2) ? ff_sse16_sse2 : sse16_mmx;
- c->sse[1] = sse8_mmx;
+ c->sse[0] = sse16_mmx;
+ c->sse[1] = sse8_mmx;
c->vsad[4]= vsad_intra16_mmx;
c->nsse[0] = nsse16_mmx;
c->ssd_int8_vs_int16 = ssd_int8_vs_int16_mmx;
-
- if (mm_flags & AV_CPU_FLAG_MMX2) {
+ if (mm_flags & AV_CPU_FLAG_MMXEXT) {
c->sum_abs_dctelem= sum_abs_dctelem_mmx2;
- c->hadamard8_diff[0]= ff_hadamard8_diff16_mmx2;
- c->hadamard8_diff[1]= ff_hadamard8_diff_mmx2;
c->vsad[4]= vsad_intra16_mmx2;
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
}
if(mm_flags & AV_CPU_FLAG_SSE2){
- c->get_pixels = get_pixels_sse2;
+ if (bit_depth <= 8)
+ c->get_pixels = get_pixels_sse2;
c->sum_abs_dctelem= sum_abs_dctelem_sse2;
- c->hadamard8_diff[0]= ff_hadamard8_diff16_sse2;
- c->hadamard8_diff[1]= ff_hadamard8_diff_sse2;
- }
-
- if (CONFIG_LPC && mm_flags & (AV_CPU_FLAG_SSE2|AV_CPU_FLAG_SSE2SLOW)) {
- c->lpc_compute_autocorr = ff_lpc_compute_autocorr_sse2;
}
#if HAVE_SSSE3
}
c->add_8x8basis= add_8x8basis_ssse3;
c->sum_abs_dctelem= sum_abs_dctelem_ssse3;
- c->hadamard8_diff[0]= ff_hadamard8_diff16_ssse3;
- c->hadamard8_diff[1]= ff_hadamard8_diff_ssse3;
}
#endif
c->add_8x8basis= add_8x8basis_3dnow;
}
}
+#endif /* HAVE_INLINE_ASM */
+
+#if HAVE_YASM
+ if (mm_flags & AV_CPU_FLAG_MMX) {
+ c->hadamard8_diff[0] = ff_hadamard8_diff16_mmx;
+ c->hadamard8_diff[1] = ff_hadamard8_diff_mmx;
+
+ if (mm_flags & AV_CPU_FLAG_MMXEXT) {
+ c->hadamard8_diff[0] = ff_hadamard8_diff16_mmx2;
+ c->hadamard8_diff[1] = ff_hadamard8_diff_mmx2;
+ }
+
+ if (mm_flags & AV_CPU_FLAG_SSE2){
+ c->sse[0] = ff_sse16_sse2;
+
+#if HAVE_ALIGNED_STACK
+ c->hadamard8_diff[0] = ff_hadamard8_diff16_sse2;
+ c->hadamard8_diff[1] = ff_hadamard8_diff_sse2;
+#endif
+ }
+
+#if HAVE_SSSE3 && HAVE_ALIGNED_STACK
+ if (mm_flags & AV_CPU_FLAG_SSSE3) {
+ c->hadamard8_diff[0] = ff_hadamard8_diff16_ssse3;
+ c->hadamard8_diff[1] = ff_hadamard8_diff_ssse3;
+ }
+#endif
+ }
+#endif /* HAVE_YASM */
- dsputil_init_pix_mmx(c, avctx);
+ ff_dsputil_init_pix_mmx(c, avctx);
}