[gst-devel] Question: How do I obtain a better gdb back traceof segmentation faults?

[Benjamin Otte]

On Mon, 10 Jan 2005, David Parr wrote:

> RELATED QUESTION: what is the difference between ./.libs/gst-launch-0.8 and ./.libs/lt-gst-launch-0.8 ?
>
The first is the version that will get installed and is linked against the
libraries in /usr/lib, while the second is linked against the libraries in
the current package you're building.
So gst-launch-0.8 links to /usr/lib/libgstreamer.so while
lt-gst-launch-0.8 is linked to
/home/company/cvs/gstreamer/gst/.libs/libgstreamer.so.

> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 1078454208 (LWP 18665)]
> 0x4063ade5 in yuv420p_to_rgba32 (dst=0x80a1bf8, src=0x80a1bd8, width=660, height=4)
>     at imgconvert_template.h:43
> 43                  YUV_TO_RGB1_CCIR(cb_ptr[0], cr_ptr[0]);
> (gdb) i threads
> * 1 Thread 1078454208 (LWP 18665)  0x4063ade5 in yuv420p_to_rgba32 (dst=0x80a1bf8, src=0x80a1bd8, width=660,
>     height=4) at imgconvert_template.h:43
> (gdb) bt
> #0  0x4063ade5 in yuv420p_to_rgba32 (dst=0x80a1bf8, src=0x80a1bd8, width=660, height=4)
>     at imgconvert_template.h:43
> #1  0x4063ce15 in img_convert (dst=0x80a1bf8, dst_pix_fmt=6, src=0x80a1bd8, src_pix_fmt=0, src_width=660,
>     src_height=500) at imgconvert.c:1982
> #2  0x4054a6c9 in gst_ffmpegcsp_chain (pad=0x80a16f8, data=0x8064db0) at gstffmpegcolorspace.c:374
> #3  0x4005cb97 in gst_pad_call_chain_function (pad=0x80a16f8, data=0x8064db0) at gstpad.c:4430
> #4  0x40058e7d in gst_pad_push (pad=0x83b92b8, data=0x8064db0) at gstpad.c:3241
> #5  0x4052fef0 in gst_video_box_chain (pad=0x83b8ff8, _data=0x8064e98) at gstvideobox.c:699
> #6  0x4005cb97 in gst_pad_call_chain_function (pad=0x83b8ff8, data=0x8064e98) at gstpad.c:4430
> #7  0x40058e7d in gst_pad_push (pad=0x83b7a00, data=0x8064e98) at gstpad.c:3241
> #8  0x4050f69b in gst_mpeg2dec_chain (pad=0x83b7938, _data=0x8065150) at gstmpeg2dec.c:773
> #9  0x4005cb97 in gst_pad_call_chain_function (pad=0x83b7938, data=0x8065150) at gstpad.c:4430
> #10 0x40058e7d in gst_pad_push (pad=0x83ab240, data=0x8065150) at gstpad.c:3241
> #11 0x404f6f22 in gst_mpeg_demux_send_subbuffer (mpeg_demux=0x83b5c20, outstream=0x80a6bf0,
>     buffer=0x80650dc, timestamp=270000000, offset=12, size=19177) at gstmpegdemux.c:928
> #12 0x404f6158 in gst_mpeg_demux_parse_packet (mpeg_parse=0x83b5c20, buffer=0x80650dc) at gstmpegdemux.c:763
> #13 0x404f1b45 in gst_mpeg_parse_loop (element=0x83b5c20) at gstmpegparse.c:530
> #14 0x407eeeca in loop_group_schedule_function (argc=0, argv=0x80a2c98) at gstoptimalscheduler.c:1332
> #15 0x407ee703 in schedule_group (group=0x80a2c98) at gstoptimalscheduler.c:1154
> #16 0x407ee8a3 in gst_opt_scheduler_schedule_run_queue (osched=0x80a0520, only_group=0x0)
>     at gstoptimalscheduler.c:1205
> #17 0x407eeaea in schedule_chain (chain=0x80a2c78) at gstoptimalscheduler.c:1258
> #18 0x407f2605 in gst_opt_scheduler_iterate (sched=0x80a0520) at gstoptimalscheduler.c:2629
> #19 0x40066fd2 in gst_scheduler_iterate (sched=0x80a0520) at gstscheduler.c:738
> #20 0x400325cc in gst_bin_iterate_func (bin=0x809ffb0) at gstbin.c:1220
> #21 0x4008198d in gst_marshal_BOOLEAN__VOID (closure=0x8059e28, return_value=0xbfffedc0, n_param_values=1,
>     param_values=0xbfffef00, invocation_hint=0xbfffede8, marshal_data=0x40032566) at gstmarshal.c:509
> #22 0x40282d5d in g_cclosure_new_swap () from /usr/lib/libgobject-2.0.so.0
> #23 0x08059e28 in ?? ()
> #24 0xbfffedc0 in ?? ()
> #25 0x00000001 in ?? ()
> #26 0xbfffef00 in ?? ()
> #27 0xbfffede8 in ?? ()
> #28 0x40032566 in gst_bin_restore_thyself (object=0x8057438, self=0x8059d90) at gstbin.c:1204
> Previous frame inner to this frame (corrupt stack?)
> (gdb)
>
This looks like a stride issue in either the colorspace plugin or
videobox. Sounds useful to file that as a bug in bugzilla.

Benjamin