From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <libc-hacker-return-6431-listarch-libc-hacker=sources.redhat.com@sources.redhat.com>
Received: (qmail 31846 invoked by alias); 1 Apr 2003 20:17:54 -0000
Mailing-List: contact libc-hacker-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:libc-hacker-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/libc-hacker/>
List-Post: <mailto:libc-hacker@sources.redhat.com>
List-Help: <mailto:libc-hacker-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: libc-hacker-owner@sources.redhat.com
Received: (qmail 31829 invoked from network); 1 Apr 2003 20:17:53 -0000
Received: from unknown (HELO localhost.localdomain) (195.113.19.66)
  by sources.redhat.com with SMTP; 1 Apr 2003 20:17:53 -0000
Received: from sunsite.mff.cuni.cz (localhost.localdomain [127.0.0.1])
	by localhost.localdomain (8.12.8/8.12.8) with ESMTP id h31KHjqO019826;
	Tue, 1 Apr 2003 22:17:45 +0200
Received: (from jakub@localhost)
	by sunsite.mff.cuni.cz (8.12.8/8.12.8/Submit) id h31KHjW7019824;
	Tue, 1 Apr 2003 22:17:45 +0200
Date: Tue, 01 Apr 2003 20:17:00 -0000
From: Jakub Jelinek <jakub@redhat.com>
To: Roland McGrath <roland@redhat.com>
Cc: Glibc hackers <libc-hacker@sources.redhat.com>
Subject: Re: [PATCH] Avoid cancellation point in pthread_mutex_lock etc. in linuxthreads
Message-ID: <20030401201745.GM16629@sunsite.ms.mff.cuni.cz>
Reply-To: Jakub Jelinek <jakub@redhat.com>
References: <20030401151926.GI16629@sunsite.ms.mff.cuni.cz> <200304011950.h31JoPZ23056@magilla.sf.frob.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <200304011950.h31JoPZ23056@magilla.sf.frob.com>
User-Agent: Mutt/1.4i
X-SW-Source: 2003-04/txt/msg00009.txt.bz2

On Tue, Apr 01, 2003 at 11:50:25AM -0800, Roland McGrath wrote:
> Why __pthread_sigsuspend instead of just using INTERNAL_SYSCALL directly?
> kernel-features.h can define something about rt_sigsuspend vs sigsuspend
> for each machine.

It differs between arches.
On most arches it is
INTERNAL_SYSCALL (rt_sigsuspend, err, 2, set, _NSIG / 8);
on some for compatibility
INTERNAL_SYSCALL (sigsuspend, err, 3, 0, 0, set->__val[0]);
(with possible runtime choice) but on Alpha it is
INTERNAL_SYSCALL (sigsuspend, err, 1, set->__val[0]);
(well, looks like the same on sparc32 for pre-rtsig kernels,
apparently nobody is running < 2.2.x kernels on sparc any more).
I think best would be a special header in
linuxthreads/sysdeps/unix/sysv/linux/*
and inline the call to avoid the call/return overhead.

	Jakub