Getting stacktrace for userspace applications on crash(SIGSEGV)

Johannes Thoma johannes at
Sun Sep 11 13:39:12 EDT 2016


Am 10.09.16 um 05:18 schrieb Daniel.:
> Yeap it is. The protocol takes 4 threads the other are user's threads.
> Still, when getting some core dump and fatal error log I was expecting
> to get backtrace from the thread responsible for the segmentation
> fault? Am I wrong?
AFAIK a signal (like SIGSEGV) is always sent to the whole process not to
a specific thread. Therefore gdb cannot know which thread caused the
segfault (except if it would analyse the assembly which it doesn't). So,
unfortunately you'll have to find the misbehaving thread by yourself ..


- Johannes
> Thanks for the help and best regards!!!

More information about the Kernelnewbies mailing list