This is the mail archive of the
libc-alpha@sourceware.cygnus.com
mailing list for the glibc project.
Re: [artdodge@cs.bu.edu] libc/1220: The code in resolv/res_send.cis not thread-safe
- To: Ulrich Drepper <drepper@cygnus.com>
- Subject: Re: [artdodge@cs.bu.edu] libc/1220: The code in resolv/res_send.cis not thread-safe
- From: "Adam D. Bradley" <artdodge@cs.bu.edu>
- Date: Sun, 25 Jul 1999 00:22:04 -0400 (EDT)
- cc: Mark Kettenis <kettenis@wins.uva.nl>, aj@arthur.rhein-neckar.de, libc-alpha@sourceware.cygnus.com, Zack Weinberg <zack@bitmover.com>
On 24 Jul 1999, Ulrich Drepper wrote:
> Mark Kettenis <kettenis@wins.uva.nl> writes:
>
> > The resolver is indeed not thread safe. This is a known issue, and
> > not easy to solve.
>
> I think the existing res_send code can be made thread-safe.
I can see a straightforward hack that would do it. Basically, keep all
socket state (s, connected, vc) within the scope of res_send(), and be
sure to close it/clean it up at all exit points.
Alternately, I'm wondering if anyone can trade notes with me about what
issues would be involved plugging in the BIND 8.2.1 code? I built its
libbind_r.a and used that instead of the glibc2.0.6 libresolv.so, and it
seems to be more-or-less working out-of-the-box...
Thanks,
Adam
--
Your lives aren't small, but \ Adam Davenport Bradley, Grad Student
you're living them in a small \ Boston University Computer Science
way. Live openly and expansively! \ artdodge@cs.bu.edu 353-8921/MCS211
II Cor 6:12-13 (The Message) <>< \ http://www.netwinder.org/~artdodge
Hi! I'm a signature virus! Copy me into your signature so I can spread!!