This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [RFC] mutex destruction (#13690): problem description and workarounds
- From: Rich Felker <dalias at libc dot org>
- To: Torvald Riegel <triegel at redhat dot com>
- Cc: GLIBC Devel <libc-alpha at sourceware dot org>
- Date: Tue, 2 Dec 2014 18:06:00 -0500
- Subject: Re: [RFC] mutex destruction (#13690): problem description and workarounds
- Authentication-results: sourceware.org; auth=none
- References: <1396621230 dot 10643 dot 7191 dot camel at triegel dot csb> <1417101135 dot 1771 dot 340 dot camel at triegel dot csb> <1417553454 dot 3930 dot 18 dot camel at triegel dot csb>
On Tue, Dec 02, 2014 at 09:50:54PM +0100, Torvald Riegel wrote:
> On Thu, 2014-11-27 at 16:12 +0100, Torvald Riegel wrote:
> > I have brought up this issue on LKML, asking for the kernel community's
> > feedback on whether this is best solved by, in some way, allowing
> > spurious wake-ups:
> > https://lkml.org/lkml/2014/11/27/472
> >
>
> See https://lkml.org/lkml/2014/12/1/723
Is Linus's position that the existing interfaces don't need to be
changed or have new additions, and that users of FUTEX_WAIT should
live with spurious wakes with a return value of 0? If so I'm fine with
that.
> If the kernel wants that, I think that's what we should do.
>
> In particular, I propose we:
> * Make sure we point out the issue in the docs: futex man pages for
> sure; maybe in high-level documentation, an article, or something like
> that that can be referenced, too.
> * Allow spurious wake-up on return of 0 in any wrapper glibc might
> offer.
Sounds good.
Rich