[FFmpeg-devel] [PATCH 5/6] x86/simple_idct: add explicit sse2 simple_idct_put/add versions.

[Ronald S. Bultje]

These use the mmx IDCT, but sse2 put/add_pixels_clamped implementations.
This way we don't need to use the ff_put/add_pixels_clamped function
pointers.
---
 libavcodec/x86/idctdsp_init.c | 10 ++++++++++
 libavcodec/x86/simple_idct.c  | 15 +++++++++++++--
 libavcodec/x86/simple_idct.h  |  3 +++
 3 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/libavcodec/x86/idctdsp_init.c b/libavcodec/x86/idctdsp_init.c
index bcf7e5b..579c5e7 100644
--- a/libavcodec/x86/idctdsp_init.c
+++ b/libavcodec/x86/idctdsp_init.c
@@ -87,6 +87,16 @@ av_cold void ff_idctdsp_init_x86(IDCTDSPContext *c, AVCodecContext *avctx,
     }
 
     if (ARCH_X86_64 && avctx->lowres == 0) {
+        if (!high_bit_depth &&
+            avctx->lowres == 0 &&
+            (avctx->idct_algo == FF_IDCT_AUTO ||
+             avctx->idct_algo == FF_IDCT_SIMPLEAUTO ||
+             avctx->idct_algo == FF_IDCT_SIMPLEMMX)) {
+                c->idct_put  = ff_simple_idct_put_sse2;
+                c->idct_add  = ff_simple_idct_add_sse2;
+                c->perm_type = FF_IDCT_PERM_SIMPLE;
+        }
+
         if (avctx->bits_per_raw_sample == 10 &&
         (avctx->idct_algo == FF_IDCT_AUTO ||
          avctx->idct_algo == FF_IDCT_SIMPLEAUTO ||
diff --git a/libavcodec/x86/simple_idct.c b/libavcodec/x86/simple_idct.c
index d3a19fa..1155920 100644
--- a/libavcodec/x86/simple_idct.c
+++ b/libavcodec/x86/simple_idct.c
@@ -24,6 +24,7 @@
 #include "libavutil/x86/asm.h"
 
 #include "libavcodec/idctdsp.h"
+#include "libavcodec/x86/idctdsp.h"
 
 #include "idctdsp.h"
 #include "simple_idct.h"
@@ -907,12 +908,22 @@ void ff_simple_idct_mmx(int16_t *block)
 void ff_simple_idct_put_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
 {
     idct(block);
-    ff_put_pixels_clamped(block, dest, line_size);
+    ff_put_pixels_clamped_mmx(block, dest, line_size);
 }
 void ff_simple_idct_add_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
 {
     idct(block);
-    ff_add_pixels_clamped(block, dest, line_size);
+    ff_add_pixels_clamped_mmx(block, dest, line_size);
+}
+void ff_simple_idct_put_sse2(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+    idct(block);
+    ff_put_pixels_clamped_sse2(block, dest, line_size);
+}
+void ff_simple_idct_add_sse2(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+    idct(block);
+    ff_add_pixels_clamped_sse2(block, dest, line_size);
 }
 
 #endif /* HAVE_INLINE_ASM */
diff --git a/libavcodec/x86/simple_idct.h b/libavcodec/x86/simple_idct.h
index ad76baf..d17ef6a 100644
--- a/libavcodec/x86/simple_idct.h
+++ b/libavcodec/x86/simple_idct.h
@@ -26,6 +26,9 @@ void ff_simple_idct_mmx(int16_t *block);
 void ff_simple_idct_add_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block);
 void ff_simple_idct_put_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block);
 
+void ff_simple_idct_add_sse2(uint8_t *dest, ptrdiff_t line_size, int16_t *block);
+void ff_simple_idct_put_sse2(uint8_t *dest, ptrdiff_t line_size, int16_t *block);
+
 void ff_simple_idct10_sse2(int16_t *block);
 void ff_simple_idct10_avx(int16_t *block);
 
-- 
2.8.1