[FFmpeg-devel] [PATCH 11/15] avformat/matroskadec: Check prebuffered_ns for overflow

[Michael Niedermayer]

Fixes: signed integer overflow: 9223372036630775808 + 1000000000 cannot be represented in type 'long'
Fixes: 51896/clusterfuzz-testcase-minimized-ffmpeg_dem_WEBM_DASH_MANIFEST_fuzzer-5406131992526848

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
---
 libavformat/matroskadec.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/libavformat/matroskadec.c b/libavformat/matroskadec.c
index 941c0bcdc9f..ae8823ae58e 100644
--- a/libavformat/matroskadec.c
+++ b/libavformat/matroskadec.c
@@ -4537,13 +4537,17 @@ static int64_t webm_dash_manifest_compute_bandwidth(AVFormatContext *s, int64_t
         int64_t prebuffer_ns = 1000000000;
         int64_t time_ns = sti->index_entries[i].timestamp * matroska->time_scale;
         double nano_seconds_per_second = 1000000000.0;
-        int64_t prebuffered_ns = time_ns + prebuffer_ns;
+        int64_t prebuffered_ns;
         double prebuffer_bytes = 0.0;
         int64_t temp_prebuffer_ns = prebuffer_ns;
         int64_t pre_bytes, pre_ns;
         double pre_sec, prebuffer, bits_per_second;
         CueDesc desc_beg = get_cue_desc(s, time_ns, cues_start);
 
+        if (time_ns > INT64_MAX - prebuffer_ns)
+            return -1;
+        prebuffered_ns = time_ns + prebuffer_ns;
+
         // Start with the first Cue.
         CueDesc desc_end = desc_beg;
 
-- 
2.17.1