SEGV in current svn trunk

[Jules Colding]

Hi,

I'm getting segmentation faults repeatedly on Ubuntu Gutsy with my  
small ORBit2 (svn trunk) based daemon. I caught the backtrace below  
with gdb. Does anyone spot the bug? I think I've starred on my monitor  
to long...

Thanks,
   jules


Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1211815024 (LWP 2328)]
link_source_set_condition (source=0x31204e4f, condition=0) at linc-
source.c:192
192 watch->pollfd.events = condition;
(gdb) bt f
#0  link_source_set_condition (source=0x31204e4f, condition=0) at linc-
source.c:192
No locals.
#1  0xb7e668bb in link_watch_unlisten (w=0x8063240) at linc-source.c:322
No locals.
#2  0xb7e66935 in link_io_remove_watch (w=0x8063240) at linc-source.c:
357
No locals.
#3  0xb7e63429 in link_source_remove (cnx=<value optimized out>) at  
linc-connection.c:176
No locals.
#4  0xb7e63451 in link_connection_dispose (obj=0x8062058) at linc-
connection.c:1247
        cnx = (LinkConnection *) 0x31204e4f
#5  0xb7e420fb in giop_connection_dispose (obj=0x8062058) at giop-
connection.c:72
        __PRETTY_FUNCTION__ = "giop_connection_dispose"
#6  0xb7c8da4b in g_object_unref () from /usr/lib/libgobject-2.0.so.0
No symbol table info available.
#7  0xb7e6415c in link_connection_exec_cnx_unref (cmd=0xb7c52234,  
immediate=1) at linc-connection.c:103
No locals.
#8  0xb7e626a2 in link_dispatch_command (data=0x31204e4f, immediate=0)  
at linc.c:505
No locals.
#9  0xb7e6298f in link_exec_command (cmd=0xb7c52234) at linc.c:135
        res = <value optimized out>
#10 0xb7e63c51 in link_connection_unref_unlock (cnx=0x8062058) at linc-
connection.c:93
        cmd = {{cmd = {cmd = {type = LINK_COMMAND_CNX_UNREF}, complete = 0},  
cnx = 0x8062058}}
#11 0xb7e44fe4 in giop_timeout (data=0x8062058) at giop-recv-buffer.c:
1388
        retv = <value optimized out>
        __PRETTY_FUNCTION__ = "giop_timeout"
#12 0xb7ea98d6 in ?? () from /usr/lib/libglib-2.0.so.0
No symbol table info available.
#13 0x08062058 in ?? ()
No symbol table info available.
#14 0xb7e178ac in __pthread_mutex_unlock_usercnt () from /lib/tls/i686/
cmov/libpthread.so.0
No symbol table info available.
#15 0xb7ea911c in g_main_context_dispatch () from /usr/lib/
libglib-2.0.so.0
No symbol table info available.
#16 0xb7eac55f in ?? () from /usr/lib/libglib-2.0.so.0
No symbol table info available.
#17 0x0805b470 in ?? ()
No symbol table info available.
#18 0x00000064 in ?? ()
No symbol table info available.
#19 0x080644a8 in ?? ()
No symbol table info available.
#20 0x00000008 in ?? ()
No symbol table info available.
#21 0x00000008 in ?? ()
No symbol table info available.
#22 0xb7e112d6 in ?? () from /lib/tls/i686/cmov/libpthread.so.0
No symbol table info available.
#23 0xb7e16541 in pthread_mutex_lock () from /lib/tls/i686/cmov/
libpthread.so.0
No symbol table info available.
#24 0xb7eac909 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
No symbol table info available.
#25 0xb7e62a70 in link_io_thread_fn (data=0x0) at linc.c:396
No locals.
#26 0xb7ecc5af in ?? () from /usr/lib/libglib-2.0.so.0
No symbol table info available.
#27 0x00000000 in ?? ()
No symbol table info available.

_______________________________________________
orbit-list mailing list
orbit-list@...
http://mail.gnome.org/mailman/listinfo/orbit-list

[Tor Lillqvist]

>  I'm getting segmentation faults repeatedly on Ubuntu Gutsy with my
>  small ORBit2 (svn trunk) based daemon.

When did you start getting the segfaults? Was it after my commit to
linc2/src/linc.c on 2008-04-23? If so that change will have to be
reverted. Or correspondingly, if it was after some other recent
change, you need to check if reverting that change helps. There hasn't
really been much done to ORBit2 lately, so if your code used to work
fine earlier, finding what broke it shouldn't be hard.

--tml
_______________________________________________
orbit-list mailing list
orbit-list@...
http://mail.gnome.org/mailman/listinfo/orbit-list

[Jules Colding]

On 25/04/2008, at 09.21, Tor Lillqvist wrote:

>> I'm getting segmentation faults repeatedly on Ubuntu Gutsy with my
>> small ORBit2 (svn trunk) based daemon.
>
> When did you start getting the segfaults? Was it after my commit to
> linc2/src/linc.c on 2008-04-23? If so that change will have to be
> reverted. Or correspondingly, if it was after some other recent
> change, you need to check if reverting that change helps. There hasn't
> really been much done to ORBit2 lately, so if your code used to work
> fine earlier, finding what broke it shouldn't be hard.


I saw the segfault for the first time yesterday. Fortunately I hope it  
is due to a mix-up of libraries from 2.14.9 and 2.14.<svn trunk>. This  
morning I discovered that I had old libraries lying around which I've  
now cleaned up. I strongly suspect this to be the real reason as your  
fixes looks really innocent. I'll make some further testing to confirm  
it and will let you know later today.


Best regards,
   jules



_______________________________________________
orbit-list mailing list
orbit-list@...
http://mail.gnome.org/mailman/listinfo/orbit-list

[Jules Colding]

On 25/04/2008, at 10.02, Jules Colding wrote:

>
> On 25/04/2008, at 09.21, Tor Lillqvist wrote:
>
>>> I'm getting segmentation faults repeatedly on Ubuntu Gutsy with my
>>> small ORBit2 (svn trunk) based daemon.
>>
>> When did you start getting the segfaults? Was it after my commit to
>> linc2/src/linc.c on 2008-04-23? If so that change will have to be
>> reverted. Or correspondingly, if it was after some other recent
>> change, you need to check if reverting that change helps. There  
>> hasn't
>> really been much done to ORBit2 lately, so if your code used to work
>> fine earlier, finding what broke it shouldn't be hard.
>
>
> I saw the segfault for the first time yesterday. Fortunately I hope it
> is due to a mix-up of libraries from 2.14.9 and 2.14.<svn trunk>. This
> morning I discovered that I had old libraries lying around which I've
> now cleaned up. I strongly suspect this to be the real reason as your
> fixes looks really innocent. I'll make some further testing to confirm
> it and will let you know later today.

No segfaults yet. I think we can safely conclude that the library mess  
was the culprit.

Best regards,
   jules

_______________________________________________
orbit-list mailing list
orbit-list@...
http://mail.gnome.org/mailman/listinfo/orbit-list