From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S932275AbcLLWDx (ORCPT <rfc822;w@1wt.eu>);
        Mon, 12 Dec 2016 17:03:53 -0500
Received: from mail-yw0-f194.google.com ([209.85.161.194]:34152 "EHLO
        mail-yw0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751702AbcLLWDw (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Mon, 12 Dec 2016 17:03:52 -0500
Date: Mon, 12 Dec 2016 17:03:49 -0500
From: Tejun Heo <tj@kernel.org>
To: Michal Hocko <mhocko@kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>, Shaohua Li <shli@fb.com>
Subject: Re: WARNING: CPU: 3 PID: 1568 at kernel/sched/core.c:7738
 __might_sleep+0x69/0x7e
Message-ID: <20161212220349.GK13864@htj.duckdns.org>
References: <20161209151229.GD4334@dhcp22.suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161209151229.GD4334@dhcp22.suse.cz>
User-Agent: Mutt/1.7.1 (2016-10-04)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Hello, Michal.

On Fri, Dec 09, 2016 at 04:12:29PM +0100, Michal Hocko wrote:
> Hi Tejun,
> I am seeing the following warning during boot in my kvm testing
> environment:
> [    3.549649] WARNING: CPU: 3 PID: 1568 at kernel/sched/core.c:7738 __might_sleep+0x69/0x7e
> [    3.551279] do not call blocking ops when !TASK_RUNNING; state=2 set at [<ffffffff81098621>] prepare_to_wait+0x65/0xac
> [    3.553354] Modules linked in:
> [    3.554032] CPU: 3 PID: 1568 Comm: modprobe Not tainted 4.9.0-rc8-nofstest5-next-20161209-00013-g3134766e941a #1018
> [    3.556099] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.3-20161025_171302-gandalf 04/01/2014
> [    3.557330] Call Trace:
> [    3.557330]  dump_stack+0x68/0x92
> [    3.557330]  __warn+0xc2/0xdd
> [    3.557330]  warn_slowpath_fmt+0x4b/0x53
> [    3.557330]  ? prepare_to_wait+0x65/0xac
> [    3.557330]  ? prepare_to_wait+0x65/0xac
> [    3.557330]  __might_sleep+0x69/0x7e
> [    3.557330]  do_ide_request+0x47/0x540
> [    3.557330]  ? __lock_is_held+0x47/0x7a
> [    3.557330]  ? hrtimer_try_to_cancel+0x15f/0x1c6
> [    3.557330]  ? _raw_spin_unlock_irqrestore+0x4c/0x61
> [    3.557330]  __blk_run_queue+0x31/0x3a
> [    3.557330]  cfq_insert_request+0x331/0x429
> [    3.557330]  __elv_add_request+0x2bb/0x346
> [    3.557330]  blk_flush_plug_list+0x1da/0x21f
> [    3.557330]  ? bit_wait+0x50/0x50
> [    3.557330]  io_schedule_timeout+0x47/0x102
> [    3.557330]  bit_wait_io+0x1b/0x5a
> [    3.557330]  __wait_on_bit+0x4d/0x7b
> [    3.557330]  out_of_line_wait_on_bit+0x72/0x7d
> [    3.557330]  ? bit_wait+0x50/0x50
> [    3.557330]  ? autoremove_wake_function+0x4e/0x4e
> [    3.557330]  __wait_on_buffer+0x40/0x43
> [    3.557330]  bh_submit_read+0x60/0x72
> [    3.557330]  ext4_get_branch+0x91/0x115
> [    3.557330]  ext4_ind_map_blocks+0x181/0xa16
> [    3.557330]  ? __lock_is_held+0x47/0x7a
> [    3.557330]  ? lock_acquire+0x133/0x1c7
> [    3.557330]  ext4_map_blocks+0x120/0x4ca
> [    3.557330]  ext4_mpage_readpages+0x2a7/0x70a
> [    3.557330]  ext4_readpages+0x33/0x35
> [    3.557330]  __do_page_cache_readahead+0x239/0x32a
> [    3.557330]  filemap_fault+0x180/0x602
> [    3.557330]  ? filemap_fault+0x180/0x602
> [    3.557330]  ext4_filemap_fault+0x37/0x4b
> [    3.557330]  __do_fault+0x1f/0x94
> [    3.557330]  handle_mm_fault+0x946/0xfca
> [    3.557330]  __do_page_fault+0x361/0x518
> [    3.557330]  do_page_fault+0x42/0x5f
> [    3.557330]  page_fault+0x22/0x30
> 
> At first I thought this would be related to you recent work
> http://lkml.kernel.org/r/1477673892-28940-1-git-send-email-tj@kernel.org
> but that doesn't seem to be in the linux-next I am using as a base.
> io_schedule_timeout is on the stack though so you might have an idea
> what is going on here. That might_sleep check has been added a long ago
> 02e7cf8f8488 ("ide-cd,atapi: use bio for internal commands") in 2.6.31
> so I am wondering what has changed since than.

10d784eae2b4 ("sched: always use blk_schedule_flush_plug in
io_schedule_out") seems related.  might_sleep is complaining that it's
being called while !TASK_RUNNING.  Shaohua, any ideas?

Thanks.

-- 
tejun