From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757952Ab2EANjY (ORCPT ); Tue, 1 May 2012 09:39:24 -0400 Received: from e7.ny.us.ibm.com ([32.97.182.137]:52403 "EHLO e7.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751869Ab2EANjX (ORCPT ); Tue, 1 May 2012 09:39:23 -0400 Date: Tue, 1 May 2012 06:39:00 -0700 From: "Paul E. McKenney" To: Benjamin Herrenschmidt Cc: Hugh Dickins , "Paul E. McKenney" , linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org Subject: Re: linux-next ppc64: RCU mods cause __might_sleep BUGs Message-ID: <20120501133900.GA4462@linux.vnet.ibm.com> Reply-To: paulmck@linux.vnet.ibm.com References: <1335832418.20866.95.camel@pasglop> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1335832418.20866.95.camel@pasglop> User-Agent: Mutt/1.5.21 (2010-09-15) X-Content-Scanned: Fidelis XPS MAILER x-cbid: 12050113-5806-0000-0000-000014C3E8BD Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, May 01, 2012 at 10:33:38AM +1000, Benjamin Herrenschmidt wrote: > On Mon, 2012-04-30 at 15:37 -0700, Hugh Dickins wrote: > > > > BUG: sleeping function called from invalid context at include/linux/pagemap.h:354 > > in_atomic(): 0, irqs_disabled(): 0, pid: 6886, name: cc1 > > Hrm ... in_atomic and irqs_disabled are both 0 ... so yeah it smells > like a preempt count problem... odd. All of the preempt-count patches are now in mainline. :-( The CONFIG_PROVE_RCU checks verify that either the task is new or it is the same task that was last context-switched to on this CPU. So the most likely suspect is a newly created task that starts running without schedule_tail() being invoked on the path from parent task to child task. If so, the fix would be to invoke rcu_switch_from() and rcu_switch_to() on that code path. So, does Power have a way of switching to a new task without involving schedule_tail()? I convinced myself that my old bugbear, usermode helpers, aren't causing this problem, but I could easily be missing something. > Did you get a specific bisect target yet ? On this one, Hugh is close enough. ;-) Thanx, Paul > Cheers, > Ben. > > > Call Trace: > > [c0000001a99f78e0] [c00000000000f34c] .show_stack+0x6c/0x16c (unreliable) > > [c0000001a99f7990] [c000000000077b40] .__might_sleep+0x11c/0x134 > > [c0000001a99f7a10] [c0000000000c6228] .filemap_fault+0x1fc/0x494 > > [c0000001a99f7af0] [c0000000000e7c9c] .__do_fault+0x120/0x684 > > [c0000001a99f7c00] [c000000000025790] .do_page_fault+0x458/0x664 > > [c0000001a99f7e30] [c000000000005868] handle_page_fault+0x10/0x30 > > > > I've plenty more examples, most of them from page faults or from kswapd; > > but I don't think there's any more useful information in them. > > > > Anything I can try later on? >