[FFmpeg-cvslog] pnmdec: make sure v is capped by maxval

Andreas Cadhalpun git at videolan.org
Fri Nov 18 01:08:42 EET 2016


ffmpeg | branch: release/3.2 | Andreas Cadhalpun <Andreas.Cadhalpun at googlemail.com> | Wed Nov  9 01:09:35 2016 +0100| [039a3e6db89289ea8fc74db52e45597c8b2f382b] | committer: Andreas Cadhalpun

pnmdec: make sure v is capped by maxval

Otherwise put_bits can be called with a value that doesn't fit in the
sample_len, causing an assertion failure.

Reviewed-by: Michael Niedermayer <michael at niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun at googlemail.com>
(cherry picked from commit cdb5479c9ddc886f0b8661db585405ebab343e80)
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun at googlemail.com>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=039a3e6db89289ea8fc74db52e45597c8b2f382b
---

 libavcodec/pnmdec.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/libavcodec/pnmdec.c b/libavcodec/pnmdec.c
index ca97cc3..958c5e4 100644
--- a/libavcodec/pnmdec.c
+++ b/libavcodec/pnmdec.c
@@ -43,7 +43,7 @@ static int pnm_decode_frame(AVCodecContext *avctx, void *data,
     int buf_size         = avpkt->size;
     PNMContext * const s = avctx->priv_data;
     AVFrame * const p    = data;
-    int i, j, n, linesize, h, upgrade = 0, is_mono = 0;
+    int i, j, k, n, linesize, h, upgrade = 0, is_mono = 0;
     unsigned char *ptr;
     int components, sample_len, ret;
 
@@ -143,10 +143,14 @@ static int pnm_decode_frame(AVCodecContext *avctx, void *data,
                         v = (*s->bytestream++)&1;
                     } else {
                         /* read a sequence of digits */
-                        do {
+                        for (k = 0; k < 5 && c <= 9; k += 1) {
                             v = 10*v + c;
                             c = (*s->bytestream++) - '0';
-                        } while (c <= 9);
+                        }
+                        if (v > s->maxval) {
+                            av_log(avctx, AV_LOG_ERROR, "value %d larger than maxval %d\n", v, s->maxval);
+                            return AVERROR_INVALIDDATA;
+                        }
                     }
                     if (sample_len == 16) {
                         ((uint16_t*)ptr)[j] = (((1<<sample_len)-1)*v + (s->maxval>>1))/s->maxval;



More information about the ffmpeg-cvslog mailing list