[FFmpeg-cvslog] apng: use correct size for output buffer

Andreas Cadhalpun git at videolan.org
Sat Nov 7 13:18:36 CET 2015


ffmpeg | branch: master | Andreas Cadhalpun <Andreas.Cadhalpun at googlemail.com> | Fri Nov  6 23:44:01 2015 +0100| [3e8e1a660ea182111057d56ec1cfad2c62250f4c] | committer: Andreas Cadhalpun

apng: use correct size for output buffer

The buffer needs s->bpp bytes, at maximum currently 10.
Assert that s->bpp is not larger.

This fixes a stack buffer overflow.

Reviewed-by: wm4 <nfxjfg at googlemail.com>
Reviewed-by: Paul B Mahol <onemda at gmail.com>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun at googlemail.com>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3e8e1a660ea182111057d56ec1cfad2c62250f4c
---

 libavcodec/pngdec.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/libavcodec/pngdec.c b/libavcodec/pngdec.c
index 689aa2b..feb1763 100644
--- a/libavcodec/pngdec.c
+++ b/libavcodec/pngdec.c
@@ -1016,7 +1016,7 @@ static int handle_p_frame_apng(AVCodecContext *avctx, PNGDecContext *s,
             for (x = s->x_offset; x < s->x_offset + s->cur_w; ++x, foreground += s->bpp, background += s->bpp) {
                 size_t b;
                 uint8_t foreground_alpha, background_alpha, output_alpha;
-                uint8_t output[4];
+                uint8_t output[10];
 
                 // Since we might be blending alpha onto alpha, we use the following equations:
                 // output_alpha = foreground_alpha + (1 - foreground_alpha) * background_alpha
@@ -1056,6 +1056,8 @@ static int handle_p_frame_apng(AVCodecContext *avctx, PNGDecContext *s,
 
                 output_alpha = foreground_alpha + FAST_DIV255((255 - foreground_alpha) * background_alpha);
 
+                av_assert0(s->bpp <= 10);
+
                 for (b = 0; b < s->bpp - 1; ++b) {
                     if (output_alpha == 0) {
                         output[b] = 0;



More information about the ffmpeg-cvslog mailing list