Builder ffmpeg64-solaris10-sparc Build #14540

Results:

Failed shell_2 shell_3 shell_4 shell_5

SourceStamp:

BuildSlave:

Reason:

The SingleBranchScheduler scheduler named 'schedule-ffmpeg64-solaris10-sparc' triggered this build

Steps and Logfiles:

1. git update ( 1 mins, 5 secs )
   1. stdio
2. shell 'gsed -i ...' ( 1 secs )
   1. stdio
3. shell_1 'gsed -i ...' ( 0 secs )
   1. stdio
4. shell_2 'gsed -i ...' failed ( 1 secs )
   1. stdio
5. shell_3 './configure --samples="../../../ffmpeg/fate-suite" ...' failed ( 1 mins, 28 secs )
   1. stdio
   2. config.log
6. shell_4 'gmake fate-rsync' failed ( 4 secs )
   1. stdio
7. shell_5 '../../../ffmpeg/fate.sh ../../../ffmpeg/fate_config_64.sh' failed ( 7 secs )
   1. stdio
   2. configure.log
   3. compile.log
   4. test.log

Build Properties:

Forced Build Properties:

Responsible Users:

1. Joshua Rogers

   joshua.rogersohnoyoudont@aisle.com
2. Michael Niedermayer

   michaelohnoyoudont@niedermayer.cc

Timing:

All Changes:

:

1. Change #286120

   Category	ffmpeg
   Changed by	Joshua Rogers <joshua.rogersohnoyoudont@aisle.com>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	0c058a08554af2ae962497e7453c9595bc12f998

   Comments

   avformat/rtpenc_mpv: Do not read a picture header past the end of the packet
   A packet ending in a picture start code made the temporal reference and
   picture type be read from after the end of the packet.
   
   This is not a security issue, the bytes read are in the padding that
   muxer input is required to have, see ce1f4c78d1.
   
   Fixes: read into padding
   Fixes: FbXPpSfFS7Zp
   Fixes: AISLE-2026-0111-00268
   Read past the packet end Replicated through API with ASAN, using an unpadded packet
   Wrong picture header Replicated through UnModified FFmpeg with the sample of trac ticket 6345
   Found-by: Joshua Rogers <joshua.rogers@aisle.com>

   Changed files

   • libavformat/rtpenc_mpv.c

2. Change #286121

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	e48278ec3b518326c2b9c5051cccfee06dd5d4ed

   Comments

   avformat/rtmppkt: Write the full basic header for continuation chunks
   Continuation chunks were marked with the single byte 0xC0 | channel_id,
   which is only valid for chunk stream ids below 64. Ids from 64 need the
   2 or 3 byte form of the basic header, as used for the first chunk, so
   the peer misparsed the rest of the stream. Repeat the basic header of
   the first chunk with chunk type 3.
   
   Corrupt chunk stream Replicated through UnModified FFmpeg
   Found during triage of the security report BvkwW4Yd19HN

   Changed files

   • libavformat/rtmppkt.c

3. Change #286122

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	9bd5ce2dcd7554829e7f93786324e4b763a32e8f

   Comments

   avformat/rtpdec_h261: Keep a single byte fragment with EBIT as end byte
   When a one byte payload completed the pending end byte and also had
   EBIT set, the merged byte was written as complete and then taken again
   from buf[-1] as the new pending end byte, adding a byte to the frame.
   Keep the merged byte pending with its 8 - EBIT valid bits instead.
   
   Wrong output Replicated through UnModified FFmpeg
   Found during triage of the security report TNp7g6QtpR0J

   Changed files

   • libavformat/rtpdec_h261.c

4. Change #286123

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	cdfad4aed8b8e3ddcaddee3f8f2a9398f9db9c34

   Comments

   avformat/rtpdec_rfc4175: Return no packet when the first field ends
   rfc4175_finalize_packet() returned success without a packet for the first
   field of interlaced video, so an empty packet was returned for every frame.
   
   This is not a security issue.
   
   Fixes: 1l09F1Fvpg1j
   Empty packets Replicated through API
   Found during triage of the security report V8a3e2cGQI7w, also raised by Forgejo Fairy while reviewing PR #24949

   Changed files

   • libavformat/rtpdec_rfc4175.c

5. Change #286124

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	ed314d886012bd52d3ea9705861b8c6743c200c4

   Comments

   avformat/rtpdec_rfc4175: Do not leak the frame finalized on a missed marker
   When a packet with a new timestamp put the previous frame into pkt, an
   invalid packet returned an error with the frame still in pkt, and a
   packet with the marker set put the new frame into the same pkt. In both
   cases the previous frame was lost and its buffer leaked, once per such
   packet, so a sender can exhaust the memory of the receiver.
   Return the previous frame in these cases, and the new one from the
   following call.
   
   Fixes: memleak
   Fixes: iF2DzMcFQavi
   Memory leak Replicated through UnModified FFmpeg with ASAN
   Found during triage of the security report V8a3e2cGQI7w

   Changed files

   • libavformat/rtpdec_rfc4175.c

6. Change #286125

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	b192fc29d662c2ada3fc9d9a944a20486ba4eff5

   Comments

   avformat/rtpdec_rfc4175: Return the recovered frame with its own timestamp
   When a packet of the next frame ended a frame whose marker was
   missed, that frame was returned with the timestamp of the next frame.
   
   This is not a security issue.
   
   Fixes: Gcg4mf6RNmEA
   Wrong timestamp Replicated through API
   Found-by: Forgejo Fairy while reviewing PR #24949

   Changed files

   • libavformat/rtpdec_rfc4175.c

7. Change #286126

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	d3576d034e111cc830c455b9244aa812981387b6

   Comments

   avformat/rdt: Allocate the RMStream of a stream only once
   A repeated ASMRuleBook line allocated a new RMStream for every stream it
   had already set up, leaking the old one.
   
   Memory leak Replicated through UnModified FFmpeg with ASAN
   Found during triage of the security report PXsIMPp0wPaW

   Changed files

   • libavformat/rdt.c

8. Change #286127

   Category	ffmpeg
   Changed by	Michael Niedermayer <michaelohnoyoudont@niedermayer.cc>
   Changed at	Thu 15 Oct 2026 02:58:38
   Repository	https://git.ffmpeg.org/ffmpeg.git
   Project	ffmpeg
   Branch	master
   Revision	50740a1144eb159d4e1408fc03049f8032f1fb11

   Comments

   avformat/rtpdec_rfc4175: Free the pending frame on close
   The handler had no close callback, so a frame that was not finalized
   when the stream was closed leaked.
   
   Memory leak Replicated through UnModified FFmpeg with ASAN
   Found during triage of the security report V8a3e2cGQI7w

   Changed files

   • libavformat/rtpdec_rfc4175.c