[Motmot] Segmentation faults

[Andrew Straw]

John Schneider wrote:
> I get the same output from the backtrace for all three computers:
> ...
> (no debugging symbols found)
> [New Thread 0xaf6b8b90 (LWP 9591)]
> [New Thread 0xaeeb7b90 (LWP 9592)]
> (no debugging symbols found)
> [New Thread 0xae656b90 (LWP 9593)]
> (no debugging symbols found)
> [New Thread 0xade55b90 (LWP 9594)]
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 0xae656b90 (LWP 9593)]
> 0x00000000 in ?? ()
> (gdb) bt
> #0  0x00000000 in ?? ()
> #1  0xaf78860c in raw1394_loop_iterate () from /usr/lib/libraw1394.so.8
> #2  0xaf789348 in raw1394_write () from /usr/lib/libraw1394.so.8
> #3  0xaf7a7ff7 in platform_camera_write () from /usr/lib/libdc1394.so.22
> #4  0xaf79bdb9 in dc1394_set_control_registers () from /usr/lib/libdc1394.so.22
> #5  0xaf79727e in dc1394_video_set_transmission ()
>    from /usr/lib/libdc1394.so.22
> #6  0xaf930856 in CCdc1394_close () from /usr/lib/libcam_iface_dc1394.so
> #7  0xaf931522 in delete_CCdc1394 () from /usr/lib/libcam_iface_dc1394.so
> #8  0xb7bbe0b1 in delete_CamContext () from /usr/lib/libcam_iface_unity.so

Hmm, is this occurring with the same characteristics as the bug you 
wrote about earlier today? If this was happening right at camera 
initialization, I'd say to reduce the number of buffers and see if that 
helps. (Also re-plug the cameras and perhaps even reboot.) In fact, I'd 
still give that a try.

But if this isn't right at the beginning of acquisition, I'm at a loss. 
It looks like the software is in the middle of attempting to close the 
camera on purpose. Any idea why that would be?

-Andrew

-- 
Andrew D. Straw, Ph.D.
California Institute of Technology
http://www.its.caltech.edu/~astraw/