This is the mail archive of the
mailing list for the GDB project.
Re: [patch] Fix for PR gdb/10757
On Monday 26 October 2009 21:25:39, Paul Pluzhnikov wrote:
> I wonder if it *could* take some lock that would make all other threads
> wait in pthread_create -- after all, thread_db does have access to inferior
> via ps_pdwrite. That would certainly make the interface cleaner, at a cost
> of an extra global lock on pthread_create path.
the problem with that is that gdb can stop all threads while one
of them is holding the lock --- you'd have to consider
the possibility of gdb deadlocking.