Re: Threads related SIGSEGV in random.c

[Alexey Suslikov]

On Tuesday, September 18, 2012, Otto Moerbeek wrote:

> On Tue, Sep 18, 2012 at 03:32:04PM -0400, Ted Unangst wrote:
>
> > On Tue, Sep 18, 2012 at 19:40, Alexey Suslikov wrote:
> >
> > > This one being discovered by Roman Kravchuk using Kannel port (see
> > > https://github.com/jasperla/openbsd-wip/tree/master/net/kannel).
> > >
> > > While stress testing, Kannel components die with
> > >
> > > Program received signal SIGSEGV, Segmentation fault.
> > > [Switching to thread 1006387]
> > > 0x00000cb33345cf6e in random () at
> /usr/src/lib/libc/stdlib/random.c:387
> > > 387                     *fptr += *rptr;
> > >
> > > We suspect threading issue because of the following trace:
> >
> > > Is it normal to not have a mutex protected *random in OpenBSD?
> >
> > No, according to posix it should be thread safe.  I don't know why,
> > since rand() is one of the exempted functions, but random() is not.
> > Standards gods are capricious gods.
>
> I think you should stress *should* here. Looking at the
> implementation, I cannot believe it's actually thread-safe.
>
>
How about rand, srand, srand48 etc?