From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx2.suse.de ([195.135.220.15]:42109 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753061AbeAGLRW (ORCPT ); Sun, 7 Jan 2018 06:17:22 -0500 Subject: Re: hang in btrfs_async_reclaim_metadata_space To: Adam Borowski , linux-btrfs@vger.kernel.org References: <20180106051059.kpultyyfhgxjowf5@angband.pl> From: Nikolay Borisov Message-ID: <80c373ee-1bd7-d59b-a252-04309c8b0210@suse.com> Date: Sun, 7 Jan 2018 13:17:19 +0200 MIME-Version: 1.0 In-Reply-To: <20180106051059.kpultyyfhgxjowf5@angband.pl> Content-Type: text/plain; charset=utf-8 Sender: linux-btrfs-owner@vger.kernel.org List-ID: On 6.01.2018 07:10, Adam Borowski wrote: > Hi! > I got a reproducible infinite hang, reliably triggered by the testsuite of > "flatpak"; fails on at least 4.15-rc6, 4.9.75, and on another machine with > Debian's 4.14.2-1. > > [580632.355107] INFO: task kworker/u8:2:11105 blocked for more than 120 seconds. > [580632.355120] Not tainted 4.14.0-1-amd64 #1 Debian 4.14.2-1 > [580632.355124] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. > [580632.355129] kworker/u8:2 D 0 11105 2 0x80000000 > [580632.355176] Workqueue: events_unbound btrfs_async_reclaim_metadata_space [btrfs] > [580632.355179] Call Trace: > [580632.355192] __schedule+0x28e/0x880 > [580632.355196] schedule+0x2c/0x80 > [580632.355200] wb_wait_for_completion+0x64/0x90 > [580632.355205] ? finish_wait+0x80/0x80 > [580632.355207] __writeback_inodes_sb_nr+0xa1/0xd0 > [580632.355210] writeback_inodes_sb_nr+0x10/0x20 > [580632.355235] flush_space+0x3ed/0x520 [btrfs] > [580632.355238] ? pick_next_task_fair+0x158/0x590 > [580632.355242] ? __switch_to+0x1f3/0x460 > [580632.355267] btrfs_async_reclaim_metadata_space+0xf6/0x4a0 [btrfs] > [580632.355278] process_one_work+0x198/0x390 > [580632.355281] worker_thread+0x35/0x3c0 > [580632.355284] kthread+0x125/0x140 > [580632.355287] ? process_one_work+0x390/0x390 > [580632.355289] ? kthread_create_on_node+0x70/0x70 > [580632.355292] ? SyS_exit_group+0x14/0x20 > [580632.355295] ret_from_fork+0x25/0x30 > > The machines are distinct enough that this probably should happen > everywhere: > > AMD Phenom2, SSD, noatime,compress=lzo,space_cache=v2 > Intel Braswell, rust, noatime,autodefrag,space_cache=v2 Provide output of echo w > /proc/sysrq-trigger when the hang happens. > > > Meow! >