On Tue, 14 May 2024, Marton Balint wrote:

Otherwise the subsequent ffio_ensure_seekback calls destroy the buffer of the
earlier. The worst case ~66kB seekback is so small it is easier to request it
entirely.

Fixes ticket #10837, a regression since
0d17f5228f4d3854066ec1001f69c7d1714b0df9.

Will apply the series in 1-2 days.

Regards,
Marton


Signed-off-by: Marton Balint <c...@passwd.hu>
---
libavformat/mp3dec.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/libavformat/mp3dec.c b/libavformat/mp3dec.c
index ec6cf567bc..78d6c8c71c 100644
--- a/libavformat/mp3dec.c
+++ b/libavformat/mp3dec.c
@@ -32,6 +32,7 @@
#include "replaygain.h"

#include "libavcodec/codec_id.h"
+#include "libavcodec/mpegaudio.h"
#include "libavcodec/mpegaudiodecheader.h"

#define XING_FLAG_FRAMES 0x01
@@ -400,15 +401,16 @@ static int mp3_read_header(AVFormatContext *s)
    if (ret < 0)
        return ret;

+    ret = ffio_ensure_seekback(s->pb, 64 * 1024 + MPA_MAX_CODED_FRAME_SIZE + 
4);
+    if (ret < 0)
+        return ret;
+
    off = avio_tell(s->pb);
    for (i = 0; i < 64 * 1024; i++) {
        uint32_t header, header2;
        int frame_size;
-        if (!(i&1023))
-            ffio_ensure_seekback(s->pb, i + 1024 + 4);
        frame_size = check(s->pb, off + i, &header);
        if (frame_size > 0) {
-            ffio_ensure_seekback(s->pb, i + 1024 + frame_size + 4);
            ret = check(s->pb, off + i + frame_size, &header2);
            if (ret >= 0 &&
                (header & MP3_MASK) == (header2 & MP3_MASK))
--
2.35.3

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

Reply via email to