Re: RH6.0 (glibc 2.1.1-6) + kernel 2.2.10 SMP == SEGV in spinlock.c

Russ (leighton@mail.imake.com)
Thu, 22 Jul 1999 08:46:49 -0400

Yes, that _has_ happened!...recently I wrote
a program to tickle it harder...a a large number of
threads (e.g. 50) in tight loops hitting a mutexs.
Completelly craps out on SMP. This is still
a module in my program, probably the next step
is to write the 10 line proggy that exhibits this.

I don't believe you can code around it. It's a race
I bet and if you have a lot of mutex contention or
are just unluckly, wham.

Peter Desnoyers wrote:

> Russell Leighton (leighton@imake.com) wrote:
> >
> > Under stock RH6.0 + 2.2.10 SMP when I run
> > a program with threads very heavly competing for
> > a mutex I get segfault ... gdb sez in spinlock.c.
>
> Does gdb by chance say that thr=0 on line 82, in __pthread_unlock
> or something similar?
>
> We're hitting that bug, too. I'd be interested in working with you to
> hunt it down, although at the moment I'm just trying to change the
> behavior of my program to work around it.
>
> --
> ............................................................................
> Peter Desnoyers
> 162 Pleasant St. (617) 661-1979 pjd@fred.cambridge.ma.us
> Cambridge, Mass. 02139 (978) 461-0402 (work) pjd@giga-net.com

--


----------------------------------------------------------------------------------------
Russell Leighton    leighton@imake.com      http://www.imake.com
----------------------------------------------------------------------------------------






-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/