[Xen-devel] Re : blkdev_issue_discard hung [solved ?]

[Olivier Bonvalet <xen.list@xxxxxxxxx>]

It seems to be a timeout problem. I can reproduce with a LVM device and
block throttling :

# ls -ld /dev/dm-3 
brw-rw---- 1 root disk 253, 3 sept. 26 12:17 /dev/dm-3

# echo '253:3 500' > /sys/fs/cgroup/blkio/blkio.throttle.read_iops_device
# echo '253:3 500' > /sys/fs/cgroup/blkio/blkio.throttle.write_iops_device
# echo '253:3 4194304' > /sys/fs/cgroup/blkio/blkio.throttle.write_bps_device
# echo '253:3 4194304' > /sys/fs/cgroup/blkio/blkio.throttle.read_bps_device

So it's probably just a timeout problem. 

I tried with a 4.13.3 kernel for Dom0, with last driver version for
xen-blkback, and it doesn't have the problem. Good !


Le lundi 25 septembre 2017 à 22:47 +0200, Olivier Bonvalet a écrit :
> Hi,
> 
> I have a problem with Xen 4.8.2 (or Linux 4.9.51 as dom0 kernel) :
> calls to DISCARD command from the DomU frequently hang.
> I also have the problem with Xen 4.8.1 and Linux 4.9 from Debian 9
> (stretch) repositories.
> 
> Here logs from dom0 :
> 
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540610] INFO: task 1.xvdo-
> 0:2740 blocked for more than 120 seconds.
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540667]       Not tainted
> 4.9-dae-dom0 #1
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540699] "echo 0 >
> /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540752] 1.xvdo-
> 0        D    0  2740      2 0x00000000
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540758]  ffff88027c6ac080
> 0000000000000000 ffff880285617e00 ffff88027ff4a280
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540763]  ffff88027c762fc0
> ffffc90011cdfbc0 ffffffff8153c720 ffff880279e40600
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540767]  ffff88027c762fc0
> 7fffffffffffffff ffff88027c762fc0 00000000007fe000
> Sep 25 22:32:44 hyp04-sbg kernel: [ 7468.540770] Call Trace:
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540783]  [<ffffffff8153c720>] ? __schedule+0x190/0x570
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540786]  [<ffffffff8153cb2d>] ? schedule+0x2d/0x80
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540793]  [<ffffffff8153fbce>] ? schedule_timeout+0x17e/0x2a0
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540799]  [<ffffffff8125a5c5>] ?
> generic_make_request_checks+0x145/0x430
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540804]  [<ffffffff810c2cf6>] ? ktime_get+0x36/0xa0
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540808]  [<ffffffff8153c528>] ? io_schedule_timeout+0x98/0x100
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540811]  [<ffffffff8153ddb4>] ?
> wait_for_completion_io+0xa4/0x120
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540815]  [<ffffffff81088040>] ? wake_up_q+0x70/0x70
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540823]  [<ffffffff81255ca7>] ? submit_bio_wait+0x47/0x60
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540826]  [<ffffffff81263fa5>] ?  0x65/0xa0
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540837]  [<ffffffffa02d9e3e>] ? __do_block_io_op+0x3be/0x610
> [xen_blkback]
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540841]  [<ffffffffa02da521>] ? xen_blkif_schedule+0x101/0x6f0
> [xen_blkback]
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540847]  [<ffffffff8109ca10>] ? wake_atomic_t_function+0x50/0x50
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540850]  [<ffffffffa02da420>] ? xen_blkif_be_int+0x30/0x30
> [xen_blkback]
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540854]  [<ffffffff81080a52>] ? kthread+0xc2/0xe0
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540856]  [<ffffffff81080990>] ? kthread_create_on_node+0x40/0x40
> Sep 25 22:32:44 hyp04-sbg kernel: [
> 7468.540859]  [<ffffffff81540e05>] ? ret_from_fork+0x25/0x30
> 
> Any ideas about what's happening, and how to fix that ?
> 
> Thanks,
> 
> Olivier
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> https://lists.xen.org/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel