From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [REGRESSION] commit 66c1ca0: {fbmem: fix fb_info->lock and mm->mmap_sem ...} causes Xfbdev not working Date: Fri, 10 Apr 2009 14:16:55 -0700 Message-ID: <20090410141655.832e287d.akpm@linux-foundation.org> References: <20090409125850.GA2727@linux> <20090410222122.87be60f6.krzysztof.h1@poczta.fm> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20090410222122.87be60f6.krzysztof.h1@poczta.fm> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-pm-bounces@lists.linux-foundation.org Errors-To: linux-pm-bounces@lists.linux-foundation.org To: Krzysztof Helt Cc: adaplas@gmail.com, arvidjaar@gmail.com, hannes@cmpxchg.org, linux-kernel@vger.kernel.org, stefanr@s5r6.in-berlin.de, geert@linux-m68k.org, linux-fbdev-devel@lists.sourceforge.net, linux-pm@lists.linux-foundation.org, harvey.harrison@gmail.com, righi.andrea@gmail.com On Fri, 10 Apr 2009 22:21:22 +0200 Krzysztof Helt wrote: > On Thu, 9 Apr 2009 14:58:51 +0200 > Andrea Righi wrote: > > > On Thu, Apr 09, 2009 at 07:36:24PM +0800, Eric Miao wrote: > > > This happens on my Marvell PXA310-based Littleton platform with > > > Angstrom Distribution. The offending paths are many: > > > > > > FBIOPUT_VSCREENINFO: > > > lock_fb_info() > > > --> fb_set_var() > > > --> fb_notifier_call_chain() [FBINFO_MISC_USEREVENT] > > > --> fbcon_event_notifier() [FB_EVENT_MODE_CHANGE] > > > --> lock_fb_info() > > > > > > OK, now hang. I'd suggest a clean fix to the original assumption of > > > circular locking > > > issue and revert this commit first. > > > > > > -- > > > Cheers > > > - eric > > > > I can agree to revert 66c1ca019078220dc1bf968f2bb18421100ef147, since I > > don't have a clean fix for this. Pushing down fb_info->lock in > > fb_set_var() excluding to call fb_notifier_call_chain with fb_info->lock > > held doesn't seem to be so trivial... > > > > However, reverting this will re-introduce the circular locking > > dependency fb_info->lock => mm->mmap_sem => fb_info->lock. > > > > If anybody is interested I have looked into the code of the fb_mmap() > and have found that the problem Andrea tried to fix is caused > only by drivers which implements their own fb_mmap(). > > The correct solution is to push the info->lock mutex into > the fb_mmap() implemented inside the drivers. Some drivers > may not need it (the ones which only calculate offsets > and do not "real" iommaping). > > I'll try to prepare a patch after the Easters. Thanks. > This will require a revert > of the 66c1ca commit. yes, I'll be queueing a revert of that when I get that far through my backlog. Today or tomorrow.