From mboxrd@z Thu Jan 1 00:00:00 1970 From: Josh Boyer Subject: Re: [PATCH] epoll: Fix spurious lockdep warnings. Date: Tue, 9 Aug 2011 11:11:55 -0400 Message-ID: <20110809151154.GA2411@zod.bos.redhat.com> References: <1312065022-17890-1-git-send-email-nelhage@nelhage.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Paul Bolle , Alexander Viro , linux-fsdevel@vger.kernel.org, Davide Libenzi , Nelson Elhage To: Nelson Elhage Return-path: Received: from mx1.redhat.com ([209.132.183.28]:11533 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753028Ab1HIPMC (ORCPT ); Tue, 9 Aug 2011 11:12:02 -0400 Content-Disposition: inline In-Reply-To: <1312065022-17890-1-git-send-email-nelhage@nelhage.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Sat, Jul 30, 2011 at 06:30:22PM -0400, Nelson Elhage wrote: > epoll can acquire multiple ep->mutex on multiple "struct eventpoll"s > at once in the case where one epoll fd is monitoring another epoll > fd. This is perfectly OK, since we're careful about the lock ordering, > but causes spurious lockdep warnings. Annotate the recursion using > mutex_lock_nested, and add a comment explaining the nesting rules for > good measure. > > Reported-by: Paul Bolle > Signed-off-by: Nelson Elhage We were seeing lockdep warnings for this in rawhide until I applied the patch. Would it be possible to get it into 3.1 still? josh