linux-stable.git/block/blk-sysfs.c, branch linux-5.2.y

blk-mq: move cancel of requeue_work to the front of blk_exit_queue

2019-08-25T14:10:24+00:00

commit e26cc08265dda37d2acc8394604f220ef412299d upstream.

blk_exit_queue will free elevator_data, while blk_mq_requeue_work
will access it. Move cancel of requeue_work to the front of
blk_exit_queue to avoid use-after-free.

blk_exit_queue                blk_mq_requeue_work
  __elevator_exit               blk_mq_run_hw_queues
    blk_mq_exit_sched             blk_mq_run_hw_queue
      dd_exit_queue                 blk_mq_hctx_has_pending
        kfree(elevator_data)          blk_mq_sched_has_work
                                        dd_has_work

Fixes: fbc2a15e3433 ("blk-mq: move cancel of requeue_work into blk_mq_release")
Cc: stable@vger.kernel.org
Reviewed-by: Ming Lei 
Signed-off-by: zhengbin 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: free sched's request pool in blk_cleanup_queue

2019-06-07T04:39:39+00:00

In theory, IO scheduler belongs to request queue, and the request pool
of sched tags belongs to the request queue too.

However, the current tags allocation interfaces are re-used for both
driver tags and sched tags, and driver tags is definitely host wide,
and doesn't belong to any request queue, same with its request pool.
So we need tagset instance for freeing request of sched tags.

Meantime, blk_mq_free_tag_set() often follows blk_cleanup_queue() in case
of non-BLK_MQ_F_TAG_SHARED, this way requires that request pool of sched
tags to be freed before calling blk_mq_free_tag_set().

Commit 47cdee29ef9d94e ("block: move blk_exit_queue into __blk_release_queue")
moves blk_exit_queue into __blk_release_queue for simplying the fast
path in generic_make_request(), then causes oops during freeing requests
of sched tags in __blk_release_queue().

Fix the above issue by move freeing request pool of sched tags into
blk_cleanup_queue(), this way is safe becasue queue has been frozen and no any
in-queue requests at that time. Freeing sched tags has to be kept in queue's
release handler becasue there might be un-completed dispatch activity
which might refer to sched tags.

Cc: Bart Van Assche 
Cc: Christoph Hellwig 
Fixes: 47cdee29ef9d94e485eb08f962c74943023a5271 ("block: move blk_exit_queue into __blk_release_queue")
Tested-by: Yi Zhang 
Reported-by: kernel test robot 
Signed-off-by: Ming Lei 
Signed-off-by: Jens Axboe

block: move blk_exit_queue into __blk_release_queue

2019-05-29T12:09:09+00:00

Commit 498f6650aec8 ("block: Fix a race between the cgroup code and
request queue initialization") moves what blk_exit_queue does into
blk_cleanup_queue() for fixing issue caused by changing back
queue lock.

However, after legacy request IO path is killed, driver queue lock
won't be used at all, and there isn't story for changing back
queue lock. Then the issue addressed by Commit 498f6650aec8 doesn't
exist any more.

So move move blk_exit_queue into __blk_release_queue.

This patch basically reverts the following two commits:

	498f6650aec8 block: Fix a race between the cgroup code and request queue initialization
	24ecc3585348 block: Ensure that a request queue is dissociated from the cgroup controller

Cc: Bart Van Assche 
Reviewed-by: Christoph Hellwig 
Signed-off-by: Ming Lei 
Signed-off-by: Jens Axboe

block: don't show io_timeout if driver has no timeout handler

2019-04-22T15:48:12+00:00

If the low level driver has no timeout handler, the
/sys/block//queue/io_timeout will not be displayed.

Reviewed-by: Bart Van Assche 
Signed-off-by: Weiping Zhang 
Signed-off-by: Jens Axboe

block: add BLK_MQ_POLL_CLASSIC for hybrid poll and return EINVAL for unexpected value

2019-03-20T20:02:07+00:00

For q->poll_nsec == -1, means doing classic poll, not hybrid poll.
We introduce a new flag BLK_MQ_POLL_CLASSIC to replace -1, which
may make code much easier to read.

Additionally, since val is an int obtained with kstrtoint(), val can be
a negative value other than -1, so return -EINVAL for that case.

Thanks to Damien Le Moal for some good suggestion.

Reviewed-by: Damien Le Moal 
Signed-off-by: Yufen Yu 
Signed-off-by: Jens Axboe

block: avoid setting wbt_lat_usec to current value

2019-02-11T15:20:14+00:00

There's no reason to set wbt min lat and freeze request queue
if current value is the same.

Reviewed-by: Johannes Thumshirn 
Signed-off-by: Aleksei Zakharov 
Signed-off-by: Jens Axboe

blk-sysfs: Rework documention of __blk_release_queue

2019-02-10T17:23:29+00:00

The Notes section of the comment was removed, because now
blk_release_queue can only be executed from blk_cleanup_queue (being
called when the q->kobj reaches zero), and also blk_init_queue was removed
in a1ce35fa4985.

Signed-off-by: Marcos Paulo de Souza 
Signed-off-by: Jens Axboe

Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

2018-12-28T22:48:06+00:00

Pull SCSI updates from James Bottomley:
 "This is mostly update of the usual drivers: smarpqi, lpfc, qedi,
  megaraid_sas, libsas, zfcp, mpt3sas, hisi_sas.

  Additionally, we have a pile of annotation, unused variable and minor
  updates.

  The big API change is the updates for Christoph's DMA rework which
  include removing the DISABLE_CLUSTERING flag.

  And finally there are a couple of target tree updates"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (259 commits)
  scsi: isci: request: mark expected switch fall-through
  scsi: isci: remote_node_context: mark expected switch fall-throughs
  scsi: isci: remote_device: Mark expected switch fall-throughs
  scsi: isci: phy: Mark expected switch fall-through
  scsi: iscsi: Capture iscsi debug messages using tracepoints
  scsi: myrb: Mark expected switch fall-throughs
  scsi: megaraid: fix out-of-bound array accesses
  scsi: mpt3sas: mpt3sas_scsih: Mark expected switch fall-through
  scsi: fcoe: remove set but not used variable 'port'
  scsi: smartpqi: call pqi_free_interrupts() in pqi_shutdown()
  scsi: smartpqi: fix build warnings
  scsi: smartpqi: update driver version
  scsi: smartpqi: add ofa support
  scsi: smartpqi: increase fw status register read timeout
  scsi: smartpqi: bump driver version
  scsi: smartpqi: add smp_utils support
  scsi: smartpqi: correct lun reset issues
  scsi: smartpqi: correct volume status
  scsi: smartpqi: do not offline disks for transient did no connect conditions
  scsi: smartpqi: allow for larger raid maps
  ...

scsi: block: remove the cluster flag

2018-12-19T04:39:26+00:00

Now that the the SCSI layer replaced the use of the cluster flag with
segment size limits and the DMA boundary we can remove the cluster flag
from the block layer.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Jens Axboe 
Signed-off-by: Martin K. Petersen

blk-mq: enable IO poll if .nr_queues of type poll > 0

2018-12-18T04:35:07+00:00

The queue mapping of type poll only exists when set->map[HCTX_TYPE_POLL].nr_queues
is bigger than zero, so enhance the constraint by checking .nr_queues of type poll
before enabling IO poll.

Otherwise IO race & timeout can be observed when running block/007.

Cc: Jeff Moyer 
Cc: Christoph Hellwig 
Signed-off-by: Ming Lei 
Signed-off-by: Jens Axboe