From: Jan Kara <jack@suse.cz>
Subject: Re: ftest01:16876 blocked for more than 120 seconds
Date: Wed, 4 Jun 2008 18:56:24 +0200
Message-ID: <20080604165624.GB5228@atrey.karlin.mff.cuni.cz>
References: <20080603200006.GA9823@martell.zuzino.mipt.ru>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: linux-ext4@vger.kernel.org
To: Alexey Dobriyan <adobriyan@gmail.com>
Return-path: <linux-ext4-owner@vger.kernel.org>
Received: from atrey.karlin.mff.cuni.cz ([195.113.26.193]:36068 "EHLO
	atrey.karlin.mff.cuni.cz" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752318AbYFDQ4Z (ORCPT
	<rfc822;linux-ext4@vger.kernel.org>); Wed, 4 Jun 2008 12:56:25 -0400
Content-Disposition: inline
In-Reply-To: <20080603200006.GA9823@martell.zuzino.mipt.ru>
Sender: linux-ext4-owner@vger.kernel.org
List-ID: <linux-ext4.vger.kernel.org>

  Hi,

> This happened during LTP run together with mass-compiling kernels.
> Kernel is 2.6.26-rc4-1beee8dc8cf58e3f605bd7b34d7a39939be7d8d2
> fs is ext3 mounted with noatime.
  Thanks for report.

> INFO: task ftest01:16876 blocked for more than 120 seconds.
> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> ftest01       D 0000000000000000     0 16876  16863
>  ffff810066fd3d68 0000000000000046 ffff810066fd3d08 ffffffff80254aae
>  ffff81005e7ab3c0 ffff810075530000 ffff81005e7ab600 0000000000000002
>  ffff810066fd3d28 ffffffff80254c55 ffff81000c043440 0000000000000282
> Call Trace:
>  [<ffffffff80254aae>] ? mark_held_locks+0x3e/0x80
>  [<ffffffff80254c55>] ? trace_hardirqs_on+0xd5/0x160
>  [<ffffffff80460722>] ? _spin_unlock_irqrestore+0x42/0x80
>  [<ffffffff802a54a9>] inode_wait+0x9/0x10
>  [<ffffffff8045dfd5>] __wait_on_bit+0x55/0x80
>  [<ffffffff802a54a0>] ? inode_wait+0x0/0x10
>  [<ffffffff802b164a>] __writeback_single_inode+0x21a/0x320
>  [<ffffffff80247860>] ? wake_bit_function+0x0/0x30
>  [<ffffffff802b1bff>] sync_sb_inodes+0x1df/0x2d0
>  [<ffffffff802b1d8f>] sync_inodes_sb+0x9f/0xc0
>  [<ffffffff802b1e3a>] __sync_inodes+0x8a/0xe0
>  [<ffffffff802b1ebe>] sync_inodes+0x2e/0x40
>  [<ffffffff802b4f54>] do_sync+0x34/0x70
>  [<ffffffff802b4f9e>] sys_sync+0xe/0x20
>  [<ffffffff8020b68b>] system_call_after_swapgs+0x7b/0x80
> 
> 1 lock held by ftest01/16876:
>  #0:  (&type->s_umount_key#3){----}, at: [<ffffffff802b1e28>] __sync_inodes+0x78/0xe0
  We are waiting for __I_SYNC - i.e., some other process is syncing
the inode and it took it longer than 120 sec to write this inode which
is well possible on heavily loaded system. I'm not sure we can do much here...

									Honza
-- 
Jan Kara <jack@suse.cz>
SuSE CR Labs