From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754472AbZJESLj (ORCPT ); Mon, 5 Oct 2009 14:11:39 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754449AbZJESLj (ORCPT ); Mon, 5 Oct 2009 14:11:39 -0400 Received: from smtp.zeugmasystems.com ([70.79.96.174]:31963 "EHLO zeugmasystems.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754443AbZJESLi convert rfc822-to-8bit (ORCPT ); Mon, 5 Oct 2009 14:11:38 -0400 X-MimeOLE: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 8BIT Subject: RE: futex question Date: Mon, 5 Oct 2009 11:11:00 -0700 Message-ID: In-Reply-To: <1254738974.26976.24.camel@twins> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: futex question Thread-Index: AcpFp1x/JKbp9PR2RPmNlfbRmdE46wAPs7AQ References: <20091001092218.GH15345@elte.hu> <4AC68F13.8050601@us.ibm.com> <4AC8CF32.8060108@anirban.org> <1254738974.26976.24.camel@twins> From: "Anirban Sinha" To: "Peter Zijlstra" , "Thomas Gleixner" Cc: "Anirban Sinha" , "Ingo Molnar" , , "Darren Hart" , "Kaz Kylheku" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > >The problem with the patch send my Ani is that it clears the robust >lists before the point of no return, so on a failing execve() we'd have >messed up the state. Ah! yes. I should have added the lines after load_binary() succeeds: fs/compat.c | 3 +++ fs/exec.c | 3 +++ 2 files changed, 6 insertions(+), 0 deletions(-) diff --git a/fs/compat.c b/fs/compat.c index 6d6f98f..7d1baf5 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -1539,6 +1539,9 @@ int compat_do_execve(char * filename, if (retval < 0) goto out; +#ifdef CONFIG_FUTEX + current->compat_robust_list = NULL; +#endif /* execve succeeded */ current->fs->in_exec = 0; current->in_execve = 0; diff --git a/fs/exec.c b/fs/exec.c index 172ceb6..d7b4ca3 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1354,6 +1354,9 @@ int do_execve(char * filename, if (retval < 0) goto out; +#ifdef CONFIG_FUTEX + current->robust_list = NULL; +#endif /* execve succeeded */ current->fs->in_exec = 0; current->in_execve = 0; btw, by the same token, shouldn't we call sched_exec() after the exec() actually succeeds and the process has a new mm (thus having a smallest effective memory and cache footprint)? I know that I could be missing something subtle. Ani