BUG: Segfault in Motion 3.2.4 snap 5

Getting a segfault after a time using Motion 3.2.4 snap 5 accessing a new Axis 206M Megapixel webcam, pic size is 1280x1024.

Standard RH AS kernel version 2.4.21-27.0.2.ELsmp on RedHat AS 3 (Taroon Update 6)

ffmpeg-0.4.8-3.1.fc2.fr

motion.conf changes:

width 1280
height 1024
framerate 5
netcam_url http://webcam-ip/jpg/image.jpg
netcam_userpass password

Test case

I ran it as gdb --args ./motion -n -d 100

Last lines were as follows

[1] Content-length present
[1] processing jpeg image - content length 336211
[1] Calculated frame time 422756.687500
[1] ***new pic delay successful***
[1] Received first header
[1] Received first header
[1] Received first header
[1] Received first header
[1] Received first header
[1] Non-streaming camera
[1] Received first header
[1] Content-length present
[1] processing jpeg image - content length 339703
[1] Calculated frame time 425997.125000
[1] ***new pic delay successful***
[1] Received first header
[1] Received first header
[1] Received first header
[1] Received first header
[1] Received first header
[1] Non-streaming camera
[1] Received first header
[1] Content-length present
[1] Calculated frame time 402952.625000
[1] processing jpeg image - content length 330787
[1] Received first header
[1] Received first header
[1] Received first header

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 27130800 (LWP 6962)]
0x001832a7 in _int_free () from /lib/tls/libc.so.6
(gdb) bt
#0  0x001832a7 in _int_free () from /lib/tls/libc.so.6
#1  0x00182278 in free () from /lib/tls/libc.so.6
#2  0x00b944a6 in __av_freep () from /usr/lib/libavcodec.so
#3  0x00b95f36 in MPV_common_end () from /usr/lib/libavcodec.so
#4  0x00b9cdf5 in MPV_encode_end () from /usr/lib/libavcodec.so
#5  0x00b94a86 in avcodec_close () from /usr/lib/libavcodec.so
#6  0x0805f618 in ffmpeg_close (ffmpeg=0x86a4928) at ffmpeg.c:443
#7  0x08058a4d in event_ffmpeg_closefile (cnt=0x869d018, type=8, dummy1=0x0, dummy2=0x0, dummy3=0x0, tm=0x86a3d28) at event.c:501
#8  0x08058be5 in event (cnt=0x869d018, type=8, image=0x0, filename=0x0, eventdata=0x0, tm=0x86a3d28) at event.c:645
#9  0x0804c56b in motion_loop (arg=0x869d018) at motion.c:1139
#10 0x008eadd8 in start_thread () from /lib/tls/libpthread.so.0
#11 0x001edd1a in clone () from /lib/tls/libc.so.6
(gdb)

Environment

Motion version:	3.2.3snap5
ffmpeg version:	0.4.8
Shared libraries:	ffmpeg
Server OS:	RedHat AS 3, kernel 2.4.21-27.0.2.ELsmp

-- PeterS - 08 Dec 2005

Follow up

Looks like it is ffmpeg that fails. You should try with a newer ffmpeg. 0.4.8 is totally out of date now.

-- KennethLavrsen - 08 Dec 2005

---

According to the Motion Guide, Motion supports FFMpeg 0.4.8 and 0.4.9. Are you saying it does not?

-- PeterS - 08 Dec 2005

---

It should. But ffmpeg then again depends on other libraries that may be newer than when 0.4.8 was made, so in a new relatively new and updated distribution ffmpeg 0.4.8 could start showing problems. And it is not practically possible to test all combinations. And it may have nothing to do with how Motion uses ffmpeg.

-- KennethLavrsen - 08 Dec 2005

---

Sorry to hassle you so much, I'm just trying to chase down as many problems as possible. I will alternately try it with the newer ffmpeg on a Fedora Core 4 system.

-- PeterS - 08 Dec 2005

---

It ran fine through the weekend on a system utilising ffmpeg 0.4.9 . I'm not sure if this is a bug or if perhaps the documentation needs to be altered to reflect the fact motion no longer works with ffmpeg 0.4.8 .

-- PeterS - 12 Dec 2005

Fix record