non meaningful output with pthread

Joël Krähemann weedlight@gmail.com
Tue Sep 23 08:03:00 GMT 2014


Am Dienstag, den 23.09.2014, 08:41 +0200 schrieb Joël Krähemann:
> Hi, I'm doing a sequencer and since I'm doing optical feedbacks I
> experience bad crashes. The engine becomes unreliable and gdb gives
> useless output.
> 
> I'm doing everything in view of thread safety from atomic operations to
> mutices, syncing ...
> 
> please take a look at http://ags.sf.net
> 
> 
> [New Thread 0x7fffb8db1700 (LWP 917)]
> [New Thread 0x7fffb85b0700 (LWP 918)]
> ** Message: short write, write 88 frames
> [New Thread 0x7fffb7daf700 (LWP 919)]
> ** Message: short write, write 240 frames
> [New Thread 0x7fffb75ae700 (LWP 920)]
> ** Message: short write, write 120 frames
> [New Thread 0x7fffb6dad700 (LWP 921)]
> ** Message: short write, write 521 frames
> [New Thread 0x7fffb65ac700 (LWP 922)]
> ** Message: short write, write 279 frames
> [New Thread 0x7fffb5dab700 (LWP 923)]
> [New Thread 0x7fffb55aa700 (LWP 924)]
> ** Message: short write, write 176 frames
> [New Thread 0x7fffb4da9700 (LWP 925)]
> ** Message: short write, write 160 frames
> [New Thread 0x7fffb45a8700 (LWP 926)]
> ** Message: short write, write 240 frames
> [Thread 0x7fffe3776700 (LWP 883) exited]
> ** Message: short write, write 400 frames
> [Thread 0x7fffb95b2700 (LWP 916) exited]
> [New Thread 0x7fffb3da7700 (LWP 927)]
> ** Message: short write, write 120 frames
> [Thread 0x7fffe0770700 (LWP 889) exited]
> ** Message: short write, write 80 frames
> ** Message: short write, write 160 frames
> [Thread 0x7fffb5dab700 (LWP 923) exited]
> [Thread 0x7fffb6dad700 (LWP 921) exited]
> ** Message: short write, write 80 frames
> ** Message: short write, write 80 frames
> 
> ...
> 
> Couldn't get registers: Kein passender Prozess gefunden.
> (gdb) bt
> #0  0x00007ffff3a42cef in start_thread (arg=<optimized out>) at
> pthread_create.c:349
> #1  0x00007ffff378ce6d in clone ()
> at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112
> #2  0x0000000000000000 in ?? ()
> 
> 

The bad crashes were probably caused by non atomic get where it had
should been atomic.






More information about the Gdb mailing list