Debugging tipps anyone?

[Michael Schwendt]

This is the result of last night's attempt at testing whether an updated
version of "wesnoth" 0.9.2 (Fedora Extras) would seem to work with FC4.
Clicking some buttons in its network related menus or connecting to the
official server lead to this fully reproducible crash:

  Program received signal SIGSEGV, Segmentation fault.
  [Switching to Thread 54066096 (zombie)]
  0x00a353a8 in ?? ()
  (gdb) bt
  #0  0x00a353a8 in ?? ()
  #1  0x0026ab7a in __nptl_deallocate_tsd () from /lib/libpthread.so.0
  #2  0x0026bb8e in start_thread () from /lib/libpthread.so.0
  #3  0x00c79dee in clone () from /lib/libc.so.6
  (gdb) t
  [Current thread is 4 (Thread 54066096 (zombie))]

It's not reproducible with FC3.

Backtraces for the threads (attached) don't look suspicious IMO. Adding
more debuginfo packages didn't give more detailed output. Two other things
I tried last night was to build without -O2 and build SDL with default optflags
instead of -O3. No change.

Suggestions, ideas, or hints much appreciated.

-- 
Fedora Core release 4 (Stentz) - Linux 2.6.11-1.1369_FC4
loadavg: 1.45 1.33 1.50

[Enrico Scholz]

fedora(a)wir-sind-cool.org (Michael Schwendt) writes:

>  This is the result of last night's attempt at testing whether an
> updated
>  version of "wesnoth" 0.9.2 (Fedora Extras) would seem to work with FC4.
>  ...
>    Program received signal SIGSEGV, Segmentation fault.
>    [Switching to Thread 54066096 (zombie)]
>    0x00a353a8 in ?? ()
>  ...
>  Suggestions, ideas, or hints much appreciated. 
'valgrind --db-attach=yes --tool=memcheck wesnoth' would be my first
idea...


Enrico

[Michael Schwendt]

On Tue, 14 Jun 2005 13:43:39 +0200, Enrico Scholz wrote:

>  > This is the result of last night's attempt at testing
> whether an updated
>  > version of "wesnoth" 0.9.2 (Fedora Extras) would seem to work with FC4.
>  > ...
>  >   Program received signal SIGSEGV, Segmentation fault.
>  >   [Switching to Thread 54066096 (zombie)]
>  >   0x00a353a8 in ?? ()
>  > ...
>  > Suggestions, ideas, or hints much appreciated.
>  
>  'valgrind --db-attach=yes --tool=memcheck wesnoth' would be my first
>  idea... 
Indicates a similar direction (SDL->pthreads usage)...

==9255== 
==9255== Thread 4:
==9255== Jump to the invalid address stated on the next line
==9255==    at 0x1BAC23A8: ???
==9255==    by 0x26BB8D: start_thread (in /lib/libpthread-2.3.5.so)
==9255==    by 0xC79DED: clone (in /lib/libc-2.3.5.so)
==9255==  Address 0x1BAC23A8 is not stack'd, malloc'd or (recently) free'd
==9255== 
==9255== ---- Attach to debugger ? --- [Return/N/n/Y/y/C/c] ---- 



-- 
Fedora Core release 4 (Stentz) - Linux 2.6.11-1.1369_FC4
loadavg: 1.26 1.49 2.20

[Michael Schwendt]

On Tue, 14 Jun 2005 12:16:38 +0200, Michael Schwendt wrote:

>  This is the result of last night's attempt at testing whether an
> updated
>  version of "wesnoth" 0.9.2 (Fedora Extras) would seem to work with FC4.
>  Clicking some buttons in its network related menus or connecting to the
>  official server lead to this fully reproducible crash:
>  
>    Program received signal SIGSEGV, Segmentation fault.
>    [Switching to Thread 54066096 (zombie)]
>    0x00a353a8 in ?? ()
>    (gdb) bt
>    #0  0x00a353a8 in ?? ()
>    #1  0x0026ab7a in __nptl_deallocate_tsd () from /lib/libpthread.so.0
>    #2  0x0026bb8e in start_thread () from /lib/libpthread.so.0
>    #3  0x00c79dee in clone () from /lib/libc.so.6
>    (gdb) t
>    [Current thread is 4 (Thread 54066096 (zombie))]
>  
>  It's not reproducible with FC3.
>  
>  Backtraces for the threads (attached) don't look suspicious IMO. Adding
>  more debuginfo packages didn't give more detailed output. Two other things
>  I tried last night was to build without -O2 and build SDL with default optflags
>  instead of -O3. No change.
>  
>  Suggestions, ideas, or hints much appreciated. 
It appears that the first version of wesnoth (0.8.6), which introduced SDL
based threads for networking functions, and all newer versions cause
segfaults on FC4. When a thread function is asked to return (e.g. prior to
SDL_ThreadWait), a crash as above is the result as soon as it terminates.
A simple SDL thread test does not show such symptoms. Something's special
in the context of wesnoth.

-- 
Fedora Core release 4 (Stentz) - Linux 2.6.11-1.1369_FC4
loadavg: 2.72 2.46 2.33