[FFmpeg-devel] [PATCH] avcodec/takdec: Fix multiple runtime error: signed integer overflow: -512 * 4563386 cannot be represented in type 'int'

Michael Niedermayer michael at niedermayer.cc
Sat May 20 18:52:21 EEST 2017


Fixes: 1706/clusterfuzz-testcase-minimized-6112772670619648

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
---
 libavcodec/takdec.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/libavcodec/takdec.c b/libavcodec/takdec.c
index 31d703135a..65dbc060a8 100644
--- a/libavcodec/takdec.c
+++ b/libavcodec/takdec.c
@@ -447,12 +447,12 @@ static int decode_subframe(TAKDecContext *s, int32_t *decoded,
 
     tfilter[0] = s->predictors[0] * 64;
     for (i = 1; i < filter_order; i++) {
-        int32_t *p1 = &tfilter[0];
-        int32_t *p2 = &tfilter[i - 1];
+        uint32_t *p1 = &tfilter[0];
+        uint32_t *p2 = &tfilter[i - 1];
 
         for (j = 0; j < (i + 1) / 2; j++) {
-            x     = *p1 + (s->predictors[i] * *p2 + 256 >> 9);
-            *p2  += s->predictors[i] * *p1 + 256 >> 9;
+            x     = *p1 + ((int32_t)(s->predictors[i] * *p2 + 256) >> 9);
+            *p2  += (int32_t)(s->predictors[i] * *p1 + 256) >> 9;
             *p1++ = x;
             p2--;
         }
-- 
2.13.0



More information about the ffmpeg-devel mailing list