[FFmpeg-cvslog] avcodec/x86/lossless_audiodsp: support len %16 == 8 in scalarproduct_and_madd_int16()

[Michael Niedermayer]

ffmpeg | branch: master | Michael Niedermayer <michaelni at gmx.at> | Sat Nov 22 20:38:30 2014 +0100| [4327088da355509b9f97e4bdb658456ee56de72b] | committer: Michael Niedermayer

avcodec/x86/lossless_audiodsp: support len %16 == 8 in scalarproduct_and_madd_int16()

Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4327088da355509b9f97e4bdb658456ee56de72b
---

 libavcodec/lossless_audiodsp.h       |    2 +-
 libavcodec/x86/lossless_audiodsp.asm |    6 ++++++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/libavcodec/lossless_audiodsp.h b/libavcodec/lossless_audiodsp.h
index 4c27502..c3ee2be 100644
--- a/libavcodec/lossless_audiodsp.h
+++ b/libavcodec/lossless_audiodsp.h
@@ -29,7 +29,7 @@ typedef struct LLAudDSPContext {
     /**
      * Calculate scalar product of v1 and v2,
      * and v1[i] += v3[i] * mul
-     * @param len length of vectors, should be multiple of 16
+     * @param len length of vectors, should be multiple of 8
      */
     int32_t (*scalarproduct_and_madd_int16)(int16_t *v1 /* align 16 */,
                                             const int16_t *v2,
diff --git a/libavcodec/x86/lossless_audiodsp.asm b/libavcodec/x86/lossless_audiodsp.asm
index 64b769f..cc5e5e5 100644
--- a/libavcodec/x86/lossless_audiodsp.asm
+++ b/libavcodec/x86/lossless_audiodsp.asm
@@ -26,6 +26,10 @@ SECTION_TEXT
 ; int ff_scalarproduct_and_madd_int16(int16_t *v1, int16_t *v2, int16_t *v3,
 ;                                     int order, int mul)
 cglobal scalarproduct_and_madd_int16, 4,4,8, v1, v2, v3, order, mul
+%if mmsize == 16
+    test orderq, 8
+        jnz scalarproduct_and_madd_int16_mmxext;
+%endif
     shl orderq, 1
     movd    m7, mulm
 %if mmsize == 16
@@ -117,6 +121,8 @@ align 16
 ;                                     int order, int mul)
 INIT_XMM ssse3
 cglobal scalarproduct_and_madd_int16, 4,5,10, v1, v2, v3, order, mul
+    test orderq, 8
+        jnz scalarproduct_and_madd_int16_mmxext;
     shl orderq, 1
     movd    m7, mulm
     pshuflw m7, m7, 0