Hi!
Attached patch fixes ticket #4672 for me.
My old valgrind reports many (!) invalid memory accesses with the patch,
but the line numbers make no sense so this may be a valgrind issue.
The memory footprint with the patch is still a magnitude lower than
for the sample from ticket #4648 (without patch).
Please comment, Carl Eugen
diff --git a/libavcodec/jpeg2000.h b/libavcodec/jpeg2000.h
index 24395a2..15fd6b5 100644
--- a/libavcodec/jpeg2000.h
+++ b/libavcodec/jpeg2000.h
@@ -67,8 +67,8 @@ enum Jpeg2000Quantsty { // quantization style
JPEG2000_QSTY_SE // scalar expounded
};
-#define JPEG2000_MAX_CBLKW 128
-#define JPEG2000_MAX_CBLKH 128
+#define JPEG2000_MAX_CBLKW 1024
+#define JPEG2000_MAX_CBLKH 1024
#define JPEG2000_MAX_DECLEVELS 32
diff --git a/libavcodec/jpeg2000dec.c b/libavcodec/jpeg2000dec.c
index 3d518e3..39e94ba 100644
--- a/libavcodec/jpeg2000dec.c
+++ b/libavcodec/jpeg2000dec.c
@@ -439,11 +439,6 @@ static int get_cox(Jpeg2000DecoderContext *s,
Jpeg2000CodingStyle *c)
return AVERROR_INVALIDDATA;
}
- if (c->log2_cblk_width > 7 || c->log2_cblk_height > 7) {
- avpriv_request_sample(s->avctx, "cblk size > 128");
- return AVERROR_PATCHWELCOME;
- }
-
c->cblk_style = bytestream2_get_byteu(&s->g);
if (c->cblk_style != 0) { // cblk style
av_log(s->avctx, AV_LOG_WARNING, "extra cblk styles %X\n",
c->cblk_style);
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
