From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752016AbbI0Rlq (ORCPT ); Sun, 27 Sep 2015 13:41:46 -0400 Received: from e31.co.us.ibm.com ([32.97.110.149]:58011 "EHLO e31.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751241AbbI0Rlp (ORCPT ); Sun, 27 Sep 2015 13:41:45 -0400 X-IBM-Helo: d03dlp02.boulder.ibm.com X-IBM-MailFrom: paulmck@linux.vnet.ibm.com X-IBM-RcptTo: linux-kernel@vger.kernel.org Date: Sun, 27 Sep 2015 10:41:39 -0700 From: "Paul E. McKenney" To: Sedat Dilek Cc: Ingo Molnar , Linus Torvalds , LKML Subject: Re: [GIT PULL rcu/urgent] Fix two more 4.3 regressions Message-ID: <20150927174139.GJ30373@linux.vnet.ibm.com> Reply-To: paulmck@linux.vnet.ibm.com References: <20150927073219.GE30373@linux.vnet.ibm.com> <20150927154929.GF30373@linux.vnet.ibm.com> <20150927164210.GH30373@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 15092717-8236-0000-0000-0000124187F0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sun, Sep 27, 2015 at 07:24:22PM +0200, Sedat Dilek wrote: > On Sun, Sep 27, 2015 at 6:42 PM, Paul E. McKenney > wrote: > > On Sun, Sep 27, 2015 at 05:55:43PM +0200, Sedat Dilek wrote: > >> On Sun, Sep 27, 2015 at 5:49 PM, Paul E. McKenney > >> wrote: > >> > On Sun, Sep 27, 2015 at 09:37:05AM +0200, Sedat Dilek wrote: > >> >> On Sun, Sep 27, 2015 at 9:32 AM, Paul E. McKenney > >> >> wrote: > >> >> > On Sun, Sep 27, 2015 at 08:28:39AM +0200, Sedat Dilek wrote: > > > > [ . . . ] > > > >> Hi Paul :-), > >> > >> Here is the stack trace. > >> > >> [ 23.045871] BUG: sleeping function called from invalid context at > >> kernel/workqueue.c:2678 > >> [ 23.045982] in_atomic(): 0, irqs_disabled(): 1, pid: 1399, name: acpid > >> [ 23.046064] 3 locks held by acpid/1399: > >> [ 23.046066] #0: (&evdev->mutex){+.+...}, at: [] > >> evdev_release+0xbc/0xf0 > >> [ 23.046081] #1: (&dev->mutex#2){+.+...}, at: [] > >> input_close_device+0x27/0x70 > >> [ 23.046093] #2: (hid_open_mut){+.+...}, at: [] > >> usbhid_close+0x28/0xb0 [usbhid] > >> [ 23.046106] irq event stamp: 3306 > >> [ 23.046109] hardirqs last enabled at (3305): [] > >> _raw_spin_unlock_irq+0x32/0x60 > >> [ 23.046115] hardirqs last disabled at (3306): [] > >> del_timer_sync+0x37/0x110 > >> [ 23.046122] softirqs last enabled at (2704): [] > >> local_bh_enable+0x9/0x20 > >> [ 23.046128] softirqs last disabled at (2702): [] > >> local_bh_disable+0x9/0x20 > >> [ 23.046136] CPU: 2 PID: 1399 Comm: acpid Not tainted > >> 4.3.0-rc3-3-llvmlinux-amd64 #1 > >> [ 23.046139] Hardware name: SAMSUNG ELECTRONICS CO., LTD. > >> 530U3BI/530U4BI/530U4BH/530U3BI/530U4BI/530U4BH, BIOS 13XK 03/28/2013 > >> [ 23.046143] ffff8800d36ee948 0000000000000092 0000000000000000 > >> ffff8800bbacfae8 > >> [ 23.046151] ffffffff8149adad ffff8800bbacfb18 ffffffff810cd1ea > >> ffffffff81c56f0a > >> [ 23.046158] ffff8800c22dc400 0000000000000000 0000000000000a76 > >> ffff8800bbacfb58 > >> [ 23.046165] Call Trace: > >> [ 23.046172] [] dump_stack+0x7d/0xa0 > >> [ 23.046177] [] ___might_sleep+0x28a/0x2a0 > >> [ 23.046182] [] __might_sleep+0x4f/0xc0 > >> [ 23.046187] [] start_flush_work+0x2f/0x290 > >> [ 23.046192] [] flush_work+0x5c/0x80 > >> [ 23.046195] [] ? flush_work+0x1a/0x80 > >> [ 23.046202] [] ? trace_hardirqs_off+0xd/0x10 > >> [ 23.046206] [] ? try_to_grab_pending+0x48/0x360 > >> [ 23.046211] [] ? _raw_spin_lock_irqsave+0x73/0x80 > >> [ 23.046216] [] __cancel_work_timer+0x179/0x260 > >> [ 23.046221] [] ? _raw_spin_unlock_irqrestore+0x52/0x80 > >> [ 23.046226] [] ? try_to_del_timer_sync+0xad/0xc0 > >> [ 23.046230] [] cancel_work_sync+0x18/0x20 > >> [ 23.046237] [] usbhid_close+0x75/0xb0 [usbhid] > >> [ 23.046245] [] hidinput_close+0x31/0x40 [hid] > >> [ 23.046251] [] ? hidinput_open+0x40/0x40 [hid] > >> [ 23.046256] [] input_close_device+0x48/0x70 > >> [ 23.046261] [] evdev_release+0xd6/0xf0 > >> [ 23.046267] [] __fput+0x107/0x240 > >> [ 23.046271] [] ____fput+0x16/0x20 > >> [ 23.046276] [] task_work_run+0x6c/0xe0 > >> [ 23.046282] [] prepare_exit_to_usermode+0x117/0x120 > >> [ 23.046287] [] syscall_return_slowpath+0x231/0x2a0 > >> [ 23.046292] [] ? filp_close+0x65/0x90 > >> [ 23.046298] [] ? trace_hardirqs_on_caller+0x19/0x290 > >> [ 23.046303] [] ? trace_hardirqs_on_thunk+0x17/0x19 > >> [ 23.046308] [] int_ret_from_sys_call+0x25/0x9f > >> > >> Can you give help on how to "debug" this? > >> > >> I switched from full-dynticks to simple cpu-accounting which did not help. > >> But this was only a suspicion as Jiri pointed to the possibility > >> del_timer_sync() could have get some mis-optimization. > >> > >> So, more empty head here. > > > > I am not familiar with the "hardirqs last enabled" debug output, but > > I am guessing that hardirqs are disabled because the "last disabled" > > number is greater than the "last enabled" number. So if I understand > > correctly, something somewhere up the call stack has irqs disabled, > > but is calling a function that needs them enabled. > > > > A quick look at usbhid_close() shows no problem, but then again, I am > > looking at the v4.2 version rather than whatever you are testing. > > But v4.3-rc2 looks the same. > > > > Looks like you are already in contact with Jiri, which is good. > > > > Another thing to try would be to insert might_sleep() calls, after > > usbhid_close(), since it enables irqs explicitly, but before the call > > to flush_work(). Jiri's comment can be interpreted as a suggestion > > to insert a might_sleep() in try_to_del_timer_sync(), though > > __cancel_work_timer() wouldn't be a bad place to try as well, > > just before its call to flush_work(). > > With... > > --- a/kernel/time/timer.c > +++ b/kernel/time/timer.c > @@ -1026,6 +1026,8 @@ int try_to_del_timer_sync(struct timer_list *timer) > unsigned long flags; > int ret = -1; > > + might_sleep(); > + > debug_assert_init(timer); > > base = lock_timer_base(timer, &flags); > > ...I get now... > > [ 53.104411] BUG: sleeping function called from invalid context at > kernel/time/timer.c:1029 > [ 53.104417] in_atomic(): 0, irqs_disabled(): 1, pid: 2100, name: fusermount > [ 53.104419] 1 lock held by fusermount/2100: > [ 53.104420] #0: (&type->s_umount_key#33){+.+...}, at: > [] deactivate_super+0x5b/0x70 > [ 53.104432] irq event stamp: 3734 > [ 53.104433] hardirqs last enabled at (3733): [] > mod_delayed_work_on+0x78/0xa0 > [ 53.104438] hardirqs last disabled at (3734): [] > flush_delayed_work+0x1f/0x70 > [ 53.104441] softirqs last enabled at (3730): [] > wb_shutdown+0x47/0xb0 > [ 53.104444] softirqs last disabled at (3728): [] > wb_shutdown+0x26/0xb0 > [ 53.104449] CPU: 3 PID: 2100 Comm: fusermount Not tainted > 4.3.0-rc3-4-llvmlinux-amd64 #2 > [ 53.104451] Hardware name: SAMSUNG ELECTRONICS CO., LTD. > 530U3BI/530U4BI/530U4BH/530U3BI/530U4BI/530U4BH, BIOS 13XK 03/28/2013 > [ 53.104453] 000000000000000a 0000000000000096 0000000000000000 > ffff8800af39bb28 > [ 53.104457] ffffffff8149cadd ffff8800af39bb58 ffffffff810cd5aa > ffffffff81c5f2e8 > [ 53.104461] ffff8800c5785940 0000000000000000 0000000000000405 > ffff8800af39bb98 > [ 53.104465] Call Trace: > [ 53.104470] [] dump_stack+0x7d/0xa0 > [ 53.104473] [] ___might_sleep+0x28a/0x2a0 > [ 53.104475] [] __might_sleep+0x4f/0xc0 > [ 53.104480] [] try_to_del_timer_sync+0x25/0xe0 > [ 53.104483] [] del_timer_sync+0xf4/0x110 > [ 53.104486] [] ? del_timer_sync+0x3b/0x110 > [ 53.104497] [] flush_delayed_work+0x28/0x70 > [ 53.104499] [] wb_shutdown+0x7b/0xb0 > [ 53.104501] [] bdi_destroy+0x7d/0x2d0 > [ 53.104505] [] ? trace_hardirqs_on+0xd/0x10 > [ 53.104509] [] fuse_put_super+0xf8/0x140 > [ 53.104511] [] ? fuse_evict_inode+0x80/0x80 > [ 53.104513] [] generic_shutdown_super+0x68/0x120 > [ 53.104515] [] kill_anon_super+0x20/0x70 > [ 53.104518] [] fuse_kill_sb_anon+0x4d/0x60 > [ 53.104520] [] deactivate_locked_super+0x4d/0xc0 > [ 53.104522] [] deactivate_super+0x63/0x70 > [ 53.104525] [] cleanup_mnt+0xb2/0x140 > [ 53.104528] [] __cleanup_mnt+0x1a/0x30 > [ 53.104531] [] task_work_run+0x6c/0xe0 > [ 53.104534] [] prepare_exit_to_usermode+0x13a/0x140 > [ 53.104537] [] syscall_return_slowpath+0x281/0x2f0 > [ 53.104540] [] ? SyS_umount+0x341/0x620 > [ 53.104542] [] ? trace_hardirqs_on_caller+0x19/0x290 > [ 53.104545] [] ? trace_hardirqs_on_thunk+0x17/0x19 > [ 53.104550] [] int_ret_from_sys_call+0x25/0x9f > > A problem in fuse? Or there is a need to qualify might_sleep() by its caller. So perhaps set a per-CPU variable in the caller and only call might_sleep() if that per-CPU variable is set. Of course, have the caller clear it on return. Thanx, Paul