From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 11057 invoked by alias); 16 Aug 2007 05:38:09 -0000 Received: (qmail 10926 invoked by uid 22791); 16 Aug 2007 05:38:07 -0000 X-Spam-Check-By: sourceware.org Received: from smtp-out.google.com (HELO smtp-out.google.com) (216.239.45.13) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 16 Aug 2007 05:38:02 +0000 Received: from zps37.corp.google.com (zps37.corp.google.com [172.25.146.37]) by smtp-out.google.com with ESMTP id l7G5bvF4018941 for ; Wed, 15 Aug 2007 22:37:57 -0700 Received: from wa-out-1112.google.com (wahj4.prod.google.com [10.114.236.4]) by zps37.corp.google.com with ESMTP id l7G5bZGr016535 for ; Wed, 15 Aug 2007 22:37:36 -0700 Received: by wa-out-1112.google.com with SMTP id j4so169539wah for ; Wed, 15 Aug 2007 22:37:25 -0700 (PDT) Received: by 10.114.37.1 with SMTP id k1mr1393648wak.1187242644748; Wed, 15 Aug 2007 22:37:24 -0700 (PDT) Received: by 10.115.15.17 with HTTP; Wed, 15 Aug 2007 22:37:24 -0700 (PDT) Message-ID: <498552560708152237o76075b60idbbf3f3a8e4e5212@mail.gmail.com> Date: Thu, 16 Aug 2007 05:38:00 -0000 From: "=?BIG5?B?RG91ZyBLd2FuICjD9q62vHcp?=" To: "Jason Merrill" Subject: Re: ping [PATCH] Fix a deadlock bug in static variable initialization in libsupc++ Cc: gcc-patches@gcc.gnu.org, libstdc++@gcc.gnu.org In-Reply-To: <46C37623.1010308@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <498552560708021356k5206cebbmc070c183694700e3@mail.gmail.com> <46C37623.1010308@redhat.com> Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2007-08/txt/msg00980.txt.bz2 Hi Jason, That's reasonable. I can change that. Actually, we have don't need the skip list for system with 32-bit pthread ID. Is it desirable to support low-overhead recursive init detection in those systems? -Doug 2007/8/15, Jason Merrill : > Thanks for fixing this, and sorry about the delay. > > It seems like the skip list code is only used to support throwing the > recursive_init_exception. I think I'd prefer to just drop that and let > recursive initialization deadlock (which is allowed under undefined > behavior) rather than add that overhead to the common case. > > Jason >