linux-stable/block
Bob Liu 7996a8b551 blk-mq: fix hang caused by freeze/unfreeze sequence
The following is a description of a hang in blk_mq_freeze_queue_wait().
The hang happens on attempt to freeze a queue while another task does
queue unfreeze.

The root cause is an incorrect sequence of percpu_ref_resurrect() and
percpu_ref_kill() and as a result those two can be swapped:

 CPU#0                         CPU#1
 ----------------              -----------------
 q1 = blk_mq_init_queue(shared_tags)

                                q2 = blk_mq_init_queue(shared_tags):
                                  blk_mq_add_queue_tag_set(shared_tags):
                                    blk_mq_update_tag_set_depth(shared_tags):
				     list_for_each_entry()
                                      blk_mq_freeze_queue(q1)
                                       > percpu_ref_kill()
                                       > blk_mq_freeze_queue_wait()

 blk_cleanup_queue(q1)
  blk_mq_freeze_queue(q1)
   > percpu_ref_kill()
                 ^^^^^^ freeze_depth can't guarantee the order

                                      blk_mq_unfreeze_queue()
                                        > percpu_ref_resurrect()

   > blk_mq_freeze_queue_wait()
                 ^^^^^^ Hang here!!!!

This wrong sequence raises kernel warning:
percpu_ref_kill_and_confirm called more than once on blk_queue_usage_counter_release!
WARNING: CPU: 0 PID: 11854 at lib/percpu-refcount.c:336 percpu_ref_kill_and_confirm+0x99/0xb0

But the most unpleasant effect is a hang of a blk_mq_freeze_queue_wait(),
which waits for a zero of a q_usage_counter, which never happens
because percpu-ref was reinited (instead of being killed) and stays in
PERCPU state forever.

How to reproduce:
 - "insmod null_blk.ko shared_tags=1 nr_devices=0 queue_mode=2"
 - cpu0: python Script.py 0; taskset the corresponding process running on cpu0
 - cpu1: python Script.py 1; taskset the corresponding process running on cpu1

 Script.py:
 ------
 #!/usr/bin/python3

import os
import sys

while True:
    on = "echo 1 > /sys/kernel/config/nullb/%s/power" % sys.argv[1]
    off = "echo 0 > /sys/kernel/config/nullb/%s/power" % sys.argv[1]
    os.system(on)
    os.system(off)
------

This bug was first reported and fixed by Roman, previous discussion:
[1] Message id: 1443287365-4244-7-git-send-email-akinobu.mita@gmail.com
[2] Message id: 1443563240-29306-6-git-send-email-tj@kernel.org
[3] https://patchwork.kernel.org/patch/9268199/

Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Roman Pen <roman.penyaev@profitbricks.com>
Signed-off-by: Bob Liu <bob.liu@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-05-23 10:25:26 -06:00
..
partitions block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
badblocks.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
bfq-cgroup.c block: switch all files cleared marked as GPLv2 or later to SPDX tags 2019-04-30 16:11:59 -06:00
bfq-iosched.c block: switch all files cleared marked as GPLv2 or later to SPDX tags 2019-04-30 16:11:59 -06:00
bfq-iosched.h block: switch all files cleared marked as GPLv2 or later to SPDX tags 2019-04-30 16:11:59 -06:00
bfq-wf2q.c block: switch all files cleared marked as GPLv2 or later to SPDX tags 2019-04-30 16:11:59 -06:00
bio-integrity.c block/bio-integrity: use struct_size() in kmalloc() 2019-05-16 08:48:48 -06:00
bio.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-cgroup.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-core.c blk-mq: fix hang caused by freeze/unfreeze sequence 2019-05-23 10:25:26 -06:00
blk-exec.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-flush.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-integrity.c for-5.2/block-20190507 2019-05-07 18:14:36 -07:00
blk-ioc.c block: remove the queue_lock indirection 2018-11-15 12:17:28 -07:00
blk-iolatency.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-lib.c block: fix 32 bit overflow in __blkdev_issue_discard() 2018-11-14 08:17:18 -07:00
blk-map.c Merge branch 'for-4.16/block' of git://git.kernel.dk/linux-block 2018-01-29 11:51:49 -08:00
blk-merge.c block: remove the bi_seg_{front,back}_size fields in struct bio 2019-05-23 10:25:26 -06:00
blk-mq-cpumap.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-mq-debugfs-zoned.c block: Cleanup license notice 2019-01-17 21:21:40 -07:00
blk-mq-debugfs.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-mq-debugfs.h blk-mq-debugfs: support rq_qos 2018-12-16 19:53:47 -07:00
blk-mq-pci.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-mq-rdma.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-mq-sched.c blk-mq: grab .q_usage_counter when queuing request from plug code path 2019-05-04 07:24:02 -06:00
blk-mq-sched.h block: mq-deadline: Fix write completion handling 2018-12-17 11:19:39 -07:00
blk-mq-sysfs.c for-5.2/block-20190507 2019-05-07 18:14:36 -07:00
blk-mq-tag.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-mq-tag.h Merge branch 'for-4.15/block' of git://git.kernel.dk/linux-block 2017-11-14 15:32:19 -08:00
blk-mq-virtio.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
blk-mq.c blk-mq: fix hang caused by freeze/unfreeze sequence 2019-05-23 10:25:26 -06:00
blk-mq.h blk-mq: free hw queue's resource in hctx's release handler 2019-05-04 07:24:05 -06:00
blk-pm.c block: remove the queue_lock indirection 2018-11-15 12:17:28 -07:00
blk-pm.h block: remove the queue_lock indirection 2018-11-15 12:17:28 -07:00
blk-rq-qos.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-rq-qos.h block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-settings.c block: force an unlimited segment size on queues with a virt boundary 2019-05-23 10:25:26 -06:00
blk-softirq.c block: remove a few unused exports 2018-11-15 12:13:25 -07:00
blk-stat.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-stat.h block: deactivate blk_stat timer in wbt_disable_default() 2018-12-12 06:47:51 -07:00
blk-sysfs.c block: Replace all ktype default_attrs with groups 2019-04-25 22:06:11 +02:00
blk-throttle.c blkcg: consolidate bio_issue_init() to be a part of core 2018-12-07 22:26:37 -07:00
blk-timeout.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-wbt.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk-wbt.h block: remove external dependency on wbt_flags 2018-07-09 09:07:54 -06:00
blk-zoned.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
blk.h block: pass page to xen_biovec_phys_mergeable 2019-04-01 12:11:13 -06:00
bounce.c block: remove the i argument to bio_for_each_segment_all 2019-04-30 09:26:13 -06:00
bsg-lib.c block: switch all files cleared marked as GPLv2 or later to SPDX tags 2019-04-30 16:11:59 -06:00
bsg.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
cmdline-parser.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
compat_ioctl.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
elevator.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
genhd.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
ioctl.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
ioprio.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
Kconfig block: remove CONFIG_LBDAF 2019-04-06 10:48:35 -06:00
Kconfig.iosched block: remove legacy IO schedulers 2018-11-07 13:42:32 -07:00
kyber-iosched.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
Makefile block: remove legacy IO schedulers 2018-11-07 13:42:32 -07:00
mq-deadline.c block: add SPDX tags to block layer files missing licensing information 2019-04-30 16:12:03 -06:00
opal_proto.h block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
partition-generic.c block: fix use-after-free on gendisk 2019-04-22 09:48:12 -06:00
scsi_ioctl.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
sed-opal.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00
t10-pi.c block: switch all files cleared marked as GPLv2 to SPDX tags 2019-04-30 16:11:57 -06:00