Merge remote-tracking branch 'qatar/master'

* qatar/master:
  mov_chan: Only set the channel_layout if setting it to a nonzero value
  mov_chan: Reindent an incorrectly indented line
  mp2 muxer: mark as AVFMT_NOTIMESTAMPS.
  x86: float_dsp: fix ff_vector_fmac_scalar_avx() on Win64
  x86: more specific checks for availability of required assembly capabilities
  x86: avcodec: Drop silly "_mmx" suffix from dsputil template names
  fate: Drop redundant setting of FUZZ to 1
  cavsdsp: set idct permutation independently of dsputil
  x86: allow using add_hfyu_median_prediction_cmov on any cpu with cmov

Conflicts:
	libavcodec/x86/dsputil_mmx.c
	libavformat/mp3enc.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

12 files changed, 34 insertions, 28 deletions

diff --git a/libavcodec/cavsdec.c b/libavcodec/cavsdec.c
index 819f2d545f..0f4f3acfa3 100644
--- a/libavcodec/cavsdec.c
+++ b/libavcodec/cavsdec.c
@@ -921,9 +921,10 @@ static int decode_pic(AVSContext *h) {
     enum cavs_mb mb_type;
 
     if (!s->context_initialized) {
-        s->avctx->idct_algo = FF_IDCT_CAVS;
         if (ff_MPV_common_init(s) < 0)
             return -1;
+        ff_init_scantable_permutation(s->dsp.idct_permutation,
+                                      h->cdsp.idct_perm);
         ff_init_scantable(s->dsp.idct_permutation,&h->scantable,ff_zigzag_direct);
     }
     skip_bits(&s->gb,16);//bbv_dwlay
diff --git a/libavcodec/cavsdsp.c b/libavcodec/cavsdsp.c
index b1b0d67c05..7eb8eca4d0 100644
--- a/libavcodec/cavsdsp.c
+++ b/libavcodec/cavsdsp.c
@@ -544,6 +544,7 @@ av_cold void ff_cavsdsp_init(CAVSDSPContext* c, AVCodecContext *avctx) {
     c->cavs_filter_cv = cavs_filter_cv_c;
     c->cavs_filter_ch = cavs_filter_ch_c;
     c->cavs_idct8_add = cavs_idct8_add_c;
+    c->idct_perm = FF_NO_IDCT_PERM;
 
     if (HAVE_MMX) ff_cavsdsp_init_mmx(c, avctx);
 }
diff --git a/libavcodec/cavsdsp.h b/libavcodec/cavsdsp.h
index b41ad21bca..b281133ff1 100644
--- a/libavcodec/cavsdsp.h
+++ b/libavcodec/cavsdsp.h
@@ -33,6 +33,7 @@ typedef struct CAVSDSPContext {
     void (*cavs_filter_cv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
     void (*cavs_filter_ch)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
     void (*cavs_idct8_add)(uint8_t *dst, DCTELEM *block, int stride);
+    int idct_perm;
 } CAVSDSPContext;
 
 void ff_cavsdsp_init(CAVSDSPContext* c, AVCodecContext *avctx);
diff --git a/libavcodec/x86/cavsdsp.c b/libavcodec/x86/cavsdsp.c
index 40875449fc..03a8b87ec3 100644
--- a/libavcodec/x86/cavsdsp.c
+++ b/libavcodec/x86/cavsdsp.c
@@ -461,6 +461,7 @@ static void ff_cavsdsp_init_mmx2(CAVSDSPContext* c, AVCodecContext *avctx) {
     dspfunc(avg_cavs_qpel, 1, 8);
 #undef dspfunc
     c->cavs_idct8_add = cavs_idct8_add_mmx;
+    c->idct_perm = FF_TRANSPOSE_IDCT_PERM;
 }
 
 static void ff_cavsdsp_init_3dnow(CAVSDSPContext* c, AVCodecContext *avctx) {
@@ -477,6 +478,7 @@ static void ff_cavsdsp_init_3dnow(CAVSDSPContext* c, AVCodecContext *avctx) {
     dspfunc(avg_cavs_qpel, 1, 8);
 #undef dspfunc
     c->cavs_idct8_add = cavs_idct8_add_mmx;
+    c->idct_perm = FF_TRANSPOSE_IDCT_PERM;
 }
 
 #endif /* HAVE_INLINE_ASM */
diff --git a/libavcodec/x86/dnxhdenc.c b/libavcodec/x86/dnxhdenc.c
index 8b0c2ad225..b2ba894bf3 100644
--- a/libavcodec/x86/dnxhdenc.c
+++ b/libavcodec/x86/dnxhdenc.c
@@ -24,7 +24,7 @@
 #include "libavutil/x86/asm.h"
 #include "libavcodec/dnxhdenc.h"
 
-#if HAVE_INLINE_ASM
+#if HAVE_SSE2_INLINE
 
 static void get_pixels_8x4_sym_sse2(DCTELEM *block, const uint8_t *pixels, int line_size)
 {
@@ -52,14 +52,14 @@ static void get_pixels_8x4_sym_sse2(DCTELEM *block, const uint8_t *pixels, int l
     );
 }
 
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
 
 void ff_dnxhdenc_init_x86(DNXHDEncContext *ctx)
 {
-#if HAVE_INLINE_ASM
+#if HAVE_SSE2_INLINE
     if (av_get_cpu_flags() & AV_CPU_FLAG_SSE2) {
         if (ctx->cid_table->bit_depth == 8)
             ctx->get_pixels_8x4_sym = get_pixels_8x4_sym_sse2;
     }
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
 }
diff --git a/libavcodec/x86/dsputil_mmx_avg_template.c b/libavcodec/x86/dsputil_avg_template.c
index 6f768595c0..6f768595c0 100644
--- a/libavcodec/x86/dsputil_mmx_avg_template.c
+++ b/libavcodec/x86/dsputil_avg_template.c
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index 93f09fb8b6..3109c70fb5 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -171,7 +171,7 @@ DECLARE_ALIGNED(16, const double, ff_pd_2)[2] = { 2.0, 2.0 };
 #define PAVGB(a, b, c, e)               PAVGB_MMX_NO_RND(a, b, c, e)
 #define OP_AVG(a, b, c, e)              PAVGB_MMX(a, b, c, e)
 
-#include "dsputil_mmx_rnd_template.c"
+#include "dsputil_rnd_template.c"
 
 #undef DEF
 #undef SET_RND
@@ -185,7 +185,7 @@ DECLARE_ALIGNED(16, const double, ff_pd_2)[2] = { 2.0, 2.0 };
 #define PAVGBP(a, b, c, d, e, f)        PAVGBP_MMX(a, b, c, d, e, f)
 #define PAVGB(a, b, c, e)               PAVGB_MMX(a, b, c, e)
 
-#include "dsputil_mmx_rnd_template.c"
+#include "dsputil_rnd_template.c"
 
 #undef DEF
 #undef SET_RND
@@ -200,7 +200,7 @@ DECLARE_ALIGNED(16, const double, ff_pd_2)[2] = { 2.0, 2.0 };
 #define PAVGB "pavgusb"
 #define OP_AVG PAVGB
 
-#include "dsputil_mmx_avg_template.c"
+#include "dsputil_avg_template.c"
 
 #undef DEF
 #undef PAVGB
@@ -215,7 +215,7 @@ DECLARE_ALIGNED(16, const double, ff_pd_2)[2] = { 2.0, 2.0 };
 #define PAVGB "pavgb"
 #define OP_AVG PAVGB
 
-#include "dsputil_mmx_avg_template.c"
+#include "dsputil_avg_template.c"
 
 #undef DEF
 #undef PAVGB
@@ -2847,7 +2847,9 @@ static void dsputil_init_mmx2(DSPContext *c, AVCodecContext *avctx,
         c->avg_h264_chroma_pixels_tab[1] = ff_avg_h264_chroma_mc4_10_mmx2;
     }
 
-    c->add_hfyu_median_prediction   = ff_add_hfyu_median_prediction_mmx2;
+    /* slower than cmov version on AMD */
+    if (!(mm_flags & AV_CPU_FLAG_3DNOW))
+        c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_mmx2;
 
     c->scalarproduct_int16          = ff_scalarproduct_int16_mmx2;
     c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_mmx2;
@@ -2924,11 +2926,6 @@ static void dsputil_init_3dnow(DSPContext *c, AVCodecContext *avctx,
     }
 
     c->vorbis_inverse_coupling = vorbis_inverse_coupling_3dnow;
-
-#if HAVE_7REGS
-    if (mm_flags & AV_CPU_FLAG_CMOV)
-        c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov;
-#endif
 #endif /* HAVE_INLINE_ASM */
 
 #if HAVE_YASM
@@ -3139,6 +3136,11 @@ void ff_dsputil_init_mmx(DSPContext *c, AVCodecContext *avctx)
 {
     int mm_flags = av_get_cpu_flags();
 
+#if HAVE_7REGS && HAVE_INLINE_ASM
+    if (mm_flags & AV_CPU_FLAG_CMOV)
+        c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov;
+#endif
+
     if (mm_flags & AV_CPU_FLAG_MMX) {
 #if HAVE_INLINE_ASM
         const int idct_algo = avctx->idct_algo;
@@ -3162,8 +3164,6 @@ void ff_dsputil_init_mmx(DSPContext *c, AVCodecContext *avctx)
                 }
                 c->idct_permutation_type = FF_LIBMPEG2_IDCT_PERM;
 #endif
-            } else if (idct_algo == FF_IDCT_CAVS) {
-                    c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
             } else if (idct_algo == FF_IDCT_XVIDMMX) {
                 if (mm_flags & AV_CPU_FLAG_SSE2) {
                     c->idct_put              = ff_idct_xvid_sse2_put;
diff --git a/libavcodec/x86/dsputil_mmx_qns_template.c b/libavcodec/x86/dsputil_qns_template.c
index 77a41b9dcb..77a41b9dcb 100644
--- a/libavcodec/x86/dsputil_mmx_qns_template.c
+++ b/libavcodec/x86/dsputil_qns_template.c
diff --git a/libavcodec/x86/dsputil_mmx_rnd_template.c b/libavcodec/x86/dsputil_rnd_template.c
index e4c91381fa..e4c91381fa 100644
--- a/libavcodec/x86/dsputil_mmx_rnd_template.c
+++ b/libavcodec/x86/dsputil_rnd_template.c
diff --git a/libavcodec/x86/dsputilenc_mmx.c b/libavcodec/x86/dsputilenc_mmx.c
index 9f9680b596..325d55062b 100644
--- a/libavcodec/x86/dsputilenc_mmx.c
+++ b/libavcodec/x86/dsputilenc_mmx.c
@@ -1042,7 +1042,7 @@ static int ssd_int8_vs_int16_mmx(const int8_t *pix1, const int16_t *pix2, int si
 #define SET_RND MOVQ_WONE
 #define SCALE_OFFSET 1
 
-#include "dsputil_mmx_qns_template.c"
+#include "dsputil_qns_template.c"
 
 #undef DEF
 #undef SET_RND
@@ -1056,7 +1056,7 @@ static int ssd_int8_vs_int16_mmx(const int8_t *pix1, const int16_t *pix2, int si
     "pmulhrw " #s ", "#x "           \n\t"\
     "pmulhrw " #s ", "#y "           \n\t"
 
-#include "dsputil_mmx_qns_template.c"
+#include "dsputil_qns_template.c"
 
 #undef DEF
 #undef SET_RND
@@ -1075,7 +1075,7 @@ static int ssd_int8_vs_int16_mmx(const int8_t *pix1, const int16_t *pix2, int si
     "pmulhrsw " #s ", "#x "          \n\t"\
     "pmulhrsw " #s ", "#y "          \n\t"
 
-#include "dsputil_mmx_qns_template.c"
+#include "dsputil_qns_template.c"
 
 #undef DEF
 #undef SET_RND
diff --git a/libavcodec/x86/lpc.c b/libavcodec/x86/lpc.c
index e0e6f8ba8a..a66f1e0da0 100644
--- a/libavcodec/x86/lpc.c
+++ b/libavcodec/x86/lpc.c
@@ -23,7 +23,7 @@
 #include "libavutil/cpu.h"
 #include "libavcodec/lpc.h"
 
-#if HAVE_INLINE_ASM
+#if HAVE_SSE2_INLINE
 
 static void lpc_apply_welch_window_sse2(const int32_t *data, int len,
                                         double *w_data)
@@ -138,16 +138,16 @@ static void lpc_compute_autocorr_sse2(const double *data, int len, int lag,
     }
 }
 
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
 
 av_cold void ff_lpc_init_x86(LPCContext *c)
 {
+#if HAVE_SSE2_INLINE
     int mm_flags = av_get_cpu_flags();
 
-#if HAVE_INLINE_ASM
     if (mm_flags & (AV_CPU_FLAG_SSE2|AV_CPU_FLAG_SSE2SLOW)) {
         c->lpc_apply_welch_window = lpc_apply_welch_window_sse2;
         c->lpc_compute_autocorr   = lpc_compute_autocorr_sse2;
     }
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
 }
diff --git a/libavcodec/x86/mpegaudiodec.c b/libavcodec/x86/mpegaudiodec.c
index c716af2ea2..8a6d59ad6a 100644
--- a/libavcodec/x86/mpegaudiodec.c
+++ b/libavcodec/x86/mpegaudiodec.c
@@ -36,7 +36,7 @@ void ff_four_imdct36_float_avx(float *out, float *buf, float *in, float *win,
 
 DECLARE_ALIGNED(16, static float, mdct_win_sse)[2][4][4*40];
 
-#if HAVE_INLINE_ASM
+#if HAVE_SSE2_INLINE
 
 #define MACS(rt, ra, rb) rt+=(ra)*(rb)
 #define MLSS(rt, ra, rb) rt-=(ra)*(rb)
@@ -180,7 +180,7 @@ static void apply_window_mp3(float *in, float *win, int *unused, float *out,
     *out = sum;
 }
 
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
 
 #if HAVE_YASM
 #define DECL_IMDCT_BLOCKS(CPU1, CPU2)                                       \
@@ -244,11 +244,12 @@ void ff_mpadsp_init_mmx(MPADSPContext *s)
         }
     }
 
-#if HAVE_INLINE_ASM
+#if HAVE_SSE2_INLINE
     if (mm_flags & AV_CPU_FLAG_SSE2) {
         s->apply_window_float = apply_window_mp3;
     }
-#endif /* HAVE_INLINE_ASM */
+#endif /* HAVE_SSE2_INLINE */
+
 #if HAVE_YASM
     if (0) {
 #if HAVE_AVX_EXTERNAL