On Wed, 24 Aug 2016, Diego Biurrun wrote:

This avoids SIMD-optimized functions having to sign-extend their
line size argument manually to be able to do pointer arithmetic.
---

Assembly changed to match.

libavcodec/proresdsp.c          |  6 +++---
libavcodec/proresdsp.h          |  3 ++-
libavcodec/proresenc.c          | 19 ++++++++++---------
libavcodec/x86/proresdsp.asm    |  3 +--
libavcodec/x86/proresdsp_init.c |  6 +++---
5 files changed, 19 insertions(+), 18 deletions(-)

diff --git a/libavcodec/proresdsp.c b/libavcodec/proresdsp.c
index 3af2f0b..f782c90 100644
--- a/libavcodec/proresdsp.c
+++ b/libavcodec/proresdsp.c
@@ -36,11 +36,11 @@
/**
 * Add bias value, clamp and output pixels of a slice
 */
-static void put_pixels(uint16_t *dst, int stride, const int16_t *in)
+static void put_pixels(uint16_t *dst, ptrdiff_t linesize, const int16_t *in)
{
    int x, y, src_offset, dst_offset;

-    for (y = 0, dst_offset = 0; y < 8; y++, dst_offset += stride) {
+    for (y = 0, dst_offset = 0; y < 8; y++, dst_offset += linesize) {
        for (x = 0; x < 8; x++) {
            src_offset = (y << 3) + x;


This is a stray change unrelated to the patch at hand, caused by overzealous renaming.

The rest of the patch seems fine.

// Martin
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to