--- Begin Message ---
- From: Steven Stewart-Gallus <sstewartgallus00 at mylangara dot bc dot ca>
- To: Joseph Myers <joseph at codesourcery dot com>
- Date: Sun, 25 Jan 2015 22:42:11 +0000 (GMT)
- Subject: Re: Can someone here consider wrapping the futex system calls?
- References: <fa9edc5f1057b dot 54c41cba at langara dot bc dot ca> <alpine dot DEB dot 2 dot 10 dot 1501242242340 dot 8829 at digraph dot polyomino dot org dot uk>
This is a different case. I need FUTEX_WAIT to be cancellable so I can kill
threads that use it. As such GLibc needs to wrap the system call. Other than
wrapping cancellable system calls I have no opinion on the matter.
----- Original Message -----
From: Joseph Myers <joseph@codesourcery.com>
Date: Saturday, January 24, 2015 2:47 pm
Subject: Re: Can someone here consider wrapping the futex system calls?
To: Steven Stewart-Gallus <sstewartgallus00@mylangara.bc.ca>
Cc: libc-alpha@sourceware.org
> On Sat, 24 Jan 2015, Steven Stewart-Gallus wrote:
>
> > Hello,
> >
> > I don't really have a problem with wrapping futex system calls
> myself but I need
> > cancellable system calls so GLibc has to do it (unfortunately,
> FUTEX_FD which
> > could allow workarounds never worked and was removed). Can
> someone here consider
> > wrapping the futex system calls so I can cancel FUTEX_WAIT?
>
> Please join in the discussion of principles for when there should
> be
> wrappers for a system call, and work on driving that discussion to
> consensus (e.g. seeking further views, analysing the common ground
> between
> views expressed, looking at issues with different approaches for
> system
> call wrappers, looking at how different suggestions from the
> discussion
> would handle various existing system calls that currently lack
> wrappers
> (or for that matter, those that currently have wrappers)).
>
> See various references from
> <https://sourceware.org/ml/libc-alpha/2015-01/msg00315.html>.
>
> --
> Joseph S. Myers
> joseph@codesourcery.com
>
--- End Message ---