avcodec/dss_sp: Fix multiple runtime error: signed integer overflow: -15699 * -164039...
authorMichael Niedermayer <michael@niedermayer.cc>
Mon, 8 May 2017 13:40:30 +0000 (15:40 +0200)
committerMichael Niedermayer <michael@niedermayer.cc>
Mon, 8 May 2017 13:41:15 +0000 (15:41 +0200)
Fixed: 1409/clusterfuzz-testcase-minimized-5237365020819456

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
libavcodec/dss_sp.c

index 93e54c5..2100936 100644 (file)
@@ -33,7 +33,7 @@
 
 #define DSS_SP_FRAME_SIZE        42
 #define DSS_SP_SAMPLE_COUNT     (66 * SUBFRAMES)
-#define DSS_SP_FORMULA(a, b, c) (((((a) * (1 << 15)) + (b) * (c)) + 0x4000) >> 15)
+#define DSS_SP_FORMULA(a, b, c) ((int)((((a) * (1 << 15)) + (b) * (unsigned)(c)) + 0x4000) >> 15)
 
 typedef struct DssSpSubframe {
     int16_t gain;
@@ -524,7 +524,7 @@ static void dss_sp_shift_sq_sub(const int32_t *filter_buf,
         tmp = dst[a] * filter_buf[0];
 
         for (i = 14; i > 0; i--)
-            tmp -= error_buf[i] * filter_buf[i];
+            tmp -= error_buf[i] * (unsigned)filter_buf[i];
 
         for (i = 14; i > 0; i--)
             error_buf[i] = error_buf[i - 1];