After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 104828 - mpeg movies segfault with CVS 0.6.0
mpeg movies segfault with CVS 0.6.0
Status: RESOLVED DUPLICATE of bug 104892
Product: GStreamer
Classification: Platform
Component: gst-plugins
0.6.0
Other Linux
: Normal normal
: 0.6.x
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2003-01-30 17:19 UTC by Christian Fredrik Kalager Schaller
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: 2.1/2.2

Attachments
gst-mask=-1 log (775.35 KB, application/octet-stream)
2003-01-30 17:29 UTC, Christian Fredrik Kalager Schaller 		  Details
Patch to disable makecontext() on i386 for now. Please test. (1.06 KB, patch)
2003-01-31 01:26 UTC, Erik Walthinsen 		none Details | Review

Description Christian Fredrik Kalager Schaller 2003-01-30 17:19:34 UTC 
Using todays CVS of the 0.6 branch and gst-launch-ext I get:

/pipeline0/thread0/mpeg2dec0: streaminfo = (GstCaps *) {
application/x-gst-streaminfo; bitrate=(int) 1180000; framerate=(float)
29.970030 }
Caught SIGSEGV accessing address 0x409a1000

+ Trace 33321

  • #0 ??
  • #1 fault_handler
  • #2 <signal handler called>
  • #3 find_start_code
  • #4 gst_mpeg_packetize_read
  • #5 gst_mpeg_parse_loop
    at gstmpegparse.c line 419
  • #6 gst_basic_scheduler_loopfunc_wrapper
  • #7 cothread_stub
    at cothreads.c line 462
  • #8 __makecontext
  • #9 gst_basic_scheduler_iterate
  • #10 gst_scheduler_iterate
    at gstscheduler.c line 732
  • #11 gst_bin_iterate_func
    at gstbin.c line 892
  • #12 gst_bin_iterate
    at gstbin.c line 935
  • #13 idle_func
  • #14 g_idle_dispatch
  • #15 g_main_dispatch
    at gmain.c line 1653
  • #16 g_main_context_dispatch
    at gmain.c line 2197
  • #17 g_main_context_iterate
  • #18 g_main_loop_run
    at gmain.c line 2498
  • #19 gst_main
    at gst.c line 694
  • #20 main
  • #21 __libc_start_main
    from /lib/tls/libc.so.6 






Comment 1


Christian Fredrik Kalager Schaller





          2003-01-30 17:29:22 UTC
        



Created attachment 13958 [details]
gst-mask=-1 log






Comment 2


Christian Fredrik Kalager Schaller





          2003-01-31 00:03:07 UTC
        



Ok, adding #undef HAVE_MAKECONTEXT
 to line 704 as Wim suggested fixes the issue.






Comment 3


Christian Fredrik Kalager Schaller





          2003-01-31 00:05:06 UTC
        



line 705 of cothreads.c that is






Comment 4


Erik Walthinsen





          2003-01-31 01:26:33 UTC
        



Created attachment 13978 [details] [review]
Patch to disable makecontext() on i386 for now.  Please test.






Comment 5


Christian Fredrik Kalager Schaller





          2003-01-31 12:23:26 UTC
        



Do not commit this patch yet. It seems the NTPL implementation in
phoebe2 is borked, upgrading to Rawhide seems to have fixed most
issues. Will close/report back when I know more.






Comment 6


Christian Fredrik Kalager Schaller





          2003-01-31 15:12:34 UTC
        



Correction, it was only the temporary lack of sound support that made
it look like things work with newer glibc packages. This patch is
indeed essential to get mpeg working with RH8.1 (vob and avi is still
broken however unless I use opt)






Comment 7


Christian Fredrik Kalager Schaller





          2003-01-31 15:29:39 UTC
        



I  now get this segfault with gst-launch-ext even using the patch.

RUNNING pipeline
Caught SIGSEGV accessing address 0x409a1000




+
Trace
    33359






    

  • 
#0
??

    • 

  • 
#1
fault_handler

    • 

  • 
#2
<signal handler called>

    • 

  • 
#3
find_start_code

    • 

  • 
#4
gst_mpeg_packetize_read

    • 

  • 
#5
gst_mpeg_parse_loop

    
                at gstmpegparse.c
                  line
                  419

    

    • 

  • 
#6
gst_basic_scheduler_loopfunc_wrapper

    • 

  • 
#7
cothread_stub

    
                at cothreads.c
                  line
                  462

    

    • 

  • 
#8
cothread_switch

    
                at cothreads.c
                  line
                  645

    

    • 

  • 
#9
gst_basic_scheduler_iterate

    • 

  • 
#10
gst_scheduler_iterate

    
                at gstscheduler.c
                  line
                  732

    

    • 

  • 
#11
gst_bin_iterate_func

    
                at gstbin.c
                  line
                  892

    

    • 

  • 
#12
gst_bin_iterate

    
                at gstbin.c
                  line
                  935

    

    • 

  • 
#13
idle_func

    • 

  • 
#14
g_idle_dispatch

    • 

  • 
#15
g_main_dispatch

    
                at gmain.c
                  line
                  1653

    

    • 

  • 
#16
g_main_context_dispatch

    
                at gmain.c
                  line
                  2197

    

    • 

  • 
#17
g_main_context_iterate

    • 

  • 
#18
g_main_loop_run

    
                at gmain.c
                  line
                  2498

    

    • 

  • 
#19
gst_main

    
                at gst.c
                  line
                  694

    

    • 

  • 
#20
main

    • 

  • 
#21
__libc_start_main

    • 













Comment 8


Thomas Vander Stichele





          2003-02-01 21:09:56 UTC
        



I added some extra checks in 0.6 branch to check for NULLity of
pointers, but this doesn't solve the root problem.

The patch isn't useful since all it does is remove makecontext, which
doesn't seem to be the cause of the problem.  So there's no point in
applying it, we're not even sure if it fixes anything, let alone this
bug :)






Comment 9


Erik Walthinsen





          2003-02-01 22:53:55 UTC
        



Deemed unfixable in 0.6.0, will have to be fixed in 0.6.x or even later.






Comment 10


Erik Walthinsen





          2003-02-01 22:58:04 UTC
        



Since all these bugs are listed as needing to be fixed in 0.6.x, they are by
definition bugs in 0.6.0






Comment 11


Christian Fredrik Kalager Schaller





          2003-02-03 11:49:57 UTC
        



Tried compiling 0.6.0 on my laptop today (running Pentium CPU).
Managed to play an mpeg. The AVI I tried still segfaulted with the
same segfault seen here and in my AVI and VOB bug reports.






Comment 12


Christian Fredrik Kalager Schaller





          2003-02-24 12:34:32 UTC
        




*** This bug has been marked as a duplicate of 104892 ***