From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrea Righi Subject: Re: commit 66c1ca breaks fbdev mode switching Date: Tue, 7 Apr 2009 11:54:44 +0200 Message-ID: <20090407095443.GA9526@linux> References: <498817E2.3020008@gmail.com> <1233783983-28802-1-git-send-email-righi.andrea@gmail.com> <200903221748.42871.arvidjaar@gmail.com> <200903221857.20473.rjw@sisk.pl> <20090322145219.ad294615.akpm@linux-foundation.org> <20090406204718.ce4909b3.krzysztof.h1@poczta.fm> <20090406214528.GA17259@linux> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Geert Uytterhoeven Cc: Krzysztof Helt , Andrew Morton , "Rafael J. Wysocki" , Andrey Borzenkov , "Antonino A. Daplas" , Linux Frame Buffer Device Development , linux-pm@lists.linux-foundation.org, Linux Kernel Development , Dave Jones , Harvey Harrison , Johannes Weiner , Stefan Richter On Tue, Apr 07, 2009 at 11:17:00AM +0200, Geert Uytterhoeven wrote: > On Mon, 6 Apr 2009, Andrea Righi wrote: > > On Mon, Apr 06, 2009 at 08:47:18PM +0200, Krzysztof Helt wrote: > > > The commit 66c1ca breaks fbdev mode switching. The first mode > > > (selected with mode_option or during boot) is ok but every mode > > > switch with fbset locks up computer. If the commit is reverted > > > mode switching works correctly. > > > > > > I do not understand why. I would be grateful if someone else > > > can confirm the problem as well. > > > > > > The commit's details are: > > > > > > commit 66c1ca019078220dc1bf968f2bb18421100ef147 > > > Author: Andrea Righi > > > Date: Tue Mar 31 15:25:18 2009 -0700 > > > > > > fbmem: fix fb_info->lock and mm->mmap_sem circular locking dependency > > > > Any lockdep info or oops? it just locks up the system? > > # echo d > /proc/sysrq-trigger > > [ 187.654947] INFO: lockdep is turned off. > > Strange, I do have lockdep enabled: > > CONFIG_LOCKDEP_SUPPORT=y > CONFIG_LOCKDEP=y > CONFIG_DEBUG_LOCKDEP=y Probably because lockdep reported the first lock bug and disabled the checking for the next lock bugs, to avoid log flooding of follow-up bugs. Search the first lockdep occurrence in dmesg. -Andrea