From: Dave Chinner <david@fromorbit.com>
To: Jens Axboe <axboe@kernel.dk>
Cc: Ming Lei <tom.leiming@gmail.com>, linux-kernel@vger.kernel.org
Subject: [HANG 3.13-rc3] blk-mq/virtio: mkfs.ext4 hangs in blk_mq_wait_for_tags
Date: Tue, 10 Dec 2013 09:39:05 +1100 [thread overview]
Message-ID: <20131209223905.GU31386@dastard> (raw)
Hi Jens,
Another day, another blkmq/virtio problem. Running mkfs.ext4 on a
sparse 100TB VM file image, it hangs hard while writing superblock
information:
$ tests/fsmark-50-test-ext4.sh
mke2fs 1.43-WIP (20-Jun-2013)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=0 blocks, Stripe width=0 blocks
1677721600 inodes, 26843545600 blocks
1342177280 blocks (5.00%) reserved for the super user
First data block=0
819200 block groups
32768 blocks per group, 32768 fragments per group
2048 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
102400000, 214990848, 512000000, 550731776, 644972544, 1934917632,
2560000000, 3855122432, 5804752896, 12800000000, 17414258688
Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information:
It writes a few superblocks, then hangs. Immediately after it stops
updating that last line, I see this:
root@test4:~# echo w > /proc/sysrq-trigger
[ 79.408153] SysRq : Show Blocked State
[ 79.408832] task PC stack pid father
[ 79.409860] mke2fs D ffff88011bc13100 3904 4242 4241 0x00000002
[ 79.411074] ffff88021a737978 0000000000000086 ffff8800dbb9de40 0000000000013100
[ 79.412009] ffff88021a737fd8 0000000000013100 ffff88011ac7af20 ffff8800dbb9de40
[ 79.412009] ffff88021a737988 ffffe8fcfbc038d0 ffff88011b39c058 ffff88011b39c040
[ 79.412009] Call Trace:
[ 79.412009] [<ffffffff81ae36d9>] schedule+0x29/0x70
[ 79.412009] [<ffffffff8178863e>] percpu_ida_alloc+0x16e/0x330
[ 79.412009] [<ffffffff810cf393>] ? finish_wait+0x63/0x80
[ 79.412009] [<ffffffff810cf3f0>] ? __init_waitqueue_head+0x40/0x40
[ 79.412009] [<ffffffff8175f30f>] blk_mq_wait_for_tags+0x1f/0x40
[ 79.412009] [<ffffffff8175e28e>] blk_mq_alloc_request_pinned+0x4e/0x110
[ 79.412009] [<ffffffff8175eacb>] blk_mq_make_request+0x41b/0x500
[ 79.412009] [<ffffffff81753552>] generic_make_request+0xc2/0x110
[ 79.412009] [<ffffffff81754a1c>] submit_bio+0x6c/0x120
[ 79.412009] [<ffffffff811d1dd3>] _submit_bh+0x133/0x200
[ 79.412009] [<ffffffff811d1eb0>] submit_bh+0x10/0x20
[ 79.412009] [<ffffffff811d5298>] __block_write_full_page+0x1b8/0x370
[ 79.412009] [<ffffffff811d3e30>] ? block_read_full_page+0x320/0x320
[ 79.412009] [<ffffffff811d8450>] ? I_BDEV+0x10/0x10
[ 79.412009] [<ffffffff811d8450>] ? I_BDEV+0x10/0x10
[ 79.412009] [<ffffffff811d5541>] block_write_full_page_endio+0xf1/0x100
[ 79.412009] [<ffffffff811d5565>] block_write_full_page+0x15/0x20
[ 79.412009] [<ffffffff811d8908>] blkdev_writepage+0x18/0x20
[ 79.412009] [<ffffffff8115668a>] __writepage+0x1a/0x50
[ 79.412009] [<ffffffff81157055>] write_cache_pages+0x225/0x470
[ 79.412009] [<ffffffff81156670>] ? mapping_tagged+0x20/0x20
[ 79.412009] [<ffffffff811572ed>] generic_writepages+0x4d/0x70
[ 79.412009] [<ffffffff810c4d0f>] ? __dequeue_entity+0x2f/0x50
[ 79.412009] [<ffffffff81158bd1>] do_writepages+0x21/0x50
[ 79.412009] [<ffffffff8114e199>] __filemap_fdatawrite_range+0x59/0x60
[ 79.412009] [<ffffffff81ae7e8e>] ? _raw_spin_unlock_irq+0xe/0x20
[ 79.412009] [<ffffffff8114e1da>] filemap_write_and_wait_range+0x3a/0x80
[ 79.412009] [<ffffffff811d8b14>] blkdev_fsync+0x24/0x50
[ 79.412009] [<ffffffff811cf898>] do_fsync+0x58/0x80
[ 79.412009] [<ffffffff81aeb8e5>] ? do_async_page_fault+0x35/0xc0
[ 79.412009] [<ffffffff811cfb30>] SyS_fsync+0x10/0x20
[ 79.412009] [<ffffffff81af08e9>] system_call_fastpath+0x16/0x1b
And a coupel of seconds later the VM hangs hard - console,
networking, everything just stops dead and it doesn't even respond
ot an nmi from the qemu command console.
The test is exactly the same as described in the previous problem I
had:
http://marc.info/?l=linux-kernel&m=138621901319333&w=2
The only difference is that I'm trying to run the concurrent create
workload on ext4 now, not XFS, and it's failing in mkfs.ext4 during
the setup code....
At this point, I have to ask: is anyone doing high IOPS testing on
virtio/blk_mq? This is the third regression I've hit since it was
merged, and I'm really not stressing this code nearly as much as
some of the hardware out there is capable of doing....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
next reply other threads:[~2013-12-09 22:39 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-12-09 22:39 Dave Chinner [this message]
2013-12-09 22:40 ` [HANG 3.13-rc3] blk-mq/virtio: mkfs.ext4 hangs in blk_mq_wait_for_tags Jens Axboe
2013-12-12 23:56 ` Dave Chinner
2013-12-13 1:57 ` Ming Lei
2013-12-13 10:58 ` Dave Chinner
2013-12-22 3:56 ` Jens Axboe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20131209223905.GU31386@dastard \
--to=david@fromorbit.com \
--cc=axboe@kernel.dk \
--cc=linux-kernel@vger.kernel.org \
--cc=tom.leiming@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox