From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?utf-8?q?Pawe=C5=82_Sikora?= Subject: Re: [Bug #20472] [2.6.35.5 -> .7] INFO: task rpcbind:14163 blocked for more than 120 seconds. Date: Tue, 19 Oct 2010 18:03:59 +0200 Message-ID: <201010191804.00147.pluto@agmk.net> References: <201010181951.34620.pluto@agmk.net> <201010182336.51265.rjw@sisk.pl> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <201010182336.51265.rjw-KKrjLPT3xs0@public.gmane.org> Sender: kernel-testers-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: Text/Plain; charset="windows-1252" To: "Rafael J. Wysocki" Cc: Linux Kernel Mailing List , Kernel Testers List , Maciej Rutecki , Florian Mickler , Trond Myklebust , Arkadiusz Miskiewicz On Monday 18 October 2010 23:36:51 Rafael J. Wysocki wrote: > On Monday, October 18, 2010, Pawe=C5=82 Sikora wrote: > > On Sunday 17 October 2010 22:21:48 Rafael J. Wysocki wrote: > > > This message has been generated automatically as a part of a summ= ary report > > > of recent regressions. > > >=20 > > > The following bug entry is on the current list of known regressio= ns > > > from 2.6.35. Please verify if it still should be listed and let = the tracking team > > > know (either way). > >=20 > > hi, > >=20 > > hot news from the front, recent tests on the pld-linux.org vendor k= ernel > > show that the random task blocking in 2.6.35.7 during medium/heavy = load (~4..25) > > is related to the grsecurity patch. we'll forward problem report to= the > > grsec maintainer asap... >=20 > So, to make things clear, this is not a mainline kernel issue, is it? i thought that was a grsecurity issue but now i have a testcase that casues task blocking on my machine with the vanilla 2.6.35.7 kernel: steps to repoduce: on console 1: - run 'make -j32' in kernel tree and wait several second for make forks and medium system load (~15). on console 2: - run 'sync' from root account to bump load more. on console 3: - observe 'tail -f /var/log/kernel' for blocking issues. [ 360.517917] INFO: task sync:6712 blocked for more than 120 seconds. [ 360.517920] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disab= les this message. [ 360.517922] sync D 00000000ffff4c7c 0 6712 2770 0x00= 000000 [ 360.517926] ffff88021ac71d18 0000000000000086 ffff880200000000 0000= 000000004000 [ 360.517930] 0000000000013700 0000000000013700 ffff88021ac71fd8 ffff= 88021ac71fd8 [ 360.517933] ffff880205593020 0000000000013700 ffff88021ac71fd8 0000= 000000004000 [ 360.517936] Call Trace: [ 360.517944] [] schedule_timeout+0x20d/0x2f0 [ 360.517949] [] ? enqueue_entity+0xea/0x170 [ 360.517951] [] ? enqueue_task_fair+0x49/0x50 [ 360.517954] [] wait_for_common+0xc5/0x150 [ 360.517957] [] ? default_wake_function+0x0/0x10 [ 360.517959] [] wait_for_completion+0x18/0x20 [ 360.517963] [] sync_inodes_sb+0x83/0x170 [ 360.517967] [] ? sync_one_sb+0x0/0x20 [ 360.517969] [] __sync_filesystem+0x80/0x90 [ 360.517972] [] sync_one_sb+0x1b/0x20 [ 360.517975] [] iterate_supers+0x77/0xc0 [ 360.517978] [] sync_filesystems+0x1b/0x20 [ 360.517980] [] sys_sync+0x1c/0x40 [ 360.517984] [] system_call_fastpath+0x16/0x1b the 'make' is running on ext4 filesystem with underlying mdadm (raid10) created from 2 caviar raid edition disks. the cpu is intel quad core Q9300 with 8GB of ddr2 ram.