linux-stable.git/block/blk-core.c, branch linux-4.4.y

blktrace: Fix potential deadlock between delete & sysfs ops

2020-05-20T06:11:37+00:00

commit 5acb3cc2c2e9d3020a4fee43763c6463767f1572 upstream.

The lockdep code had reported the following unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(s_active#228);
                               lock(&bdev->bd_mutex/1);
                               lock(s_active#228);
  lock(&bdev->bd_mutex);

 *** DEADLOCK ***

The deadlock may happen when one task (CPU1) is trying to delete a
partition in a block device and another task (CPU0) is accessing
tracing sysfs file (e.g. /sys/block/dm-1/trace/act_mask) in that
partition.

The s_active isn't an actual lock. It is a reference count (kn->count)
on the sysfs (kernfs) file. Removal of a sysfs file, however, require
a wait until all the references are gone. The reference count is
treated like a rwsem using lockdep instrumentation code.

The fact that a thread is in the sysfs callback method or in the
ioctl call means there is a reference to the opended sysfs or device
file. That should prevent the underlying block structure from being
removed.

Instead of using bd_mutex in the block_device structure, a new
blk_trace_mutex is now added to the request_queue structure to protect
access to the blk_trace structure.

Suggested-by: Christoph Hellwig 
Signed-off-by: Waiman Long 
Acked-by: Steven Rostedt (VMware) 

Fix typo in patch subject line, and prune a comment detailing how
the code used to work.

Signed-off-by: Jens Axboe 
Signed-off-by: Ben Hutchings 
Signed-off-by: Greg Kroah-Hartman

block: blk_init_allocated_queue() set q->fq as NULL in the fail case

2019-08-11T10:20:46+00:00

commit 54648cf1ec2d7f4b6a71767799c45676a138ca24 upstream.

We find the memory use-after-free issue in __blk_drain_queue()
on the kernel 4.14. After read the latest kernel 4.18-rc6 we
think it has the same problem.

Memory is allocated for q->fq in the blk_init_allocated_queue().
If the elevator init function called with error return, it will
run into the fail case to free the q->fq.

Then the __blk_drain_queue() uses the same memory after the free
of the q->fq, it will lead to the unpredictable event.

The patch is to set q->fq as NULL in the fail case of
blk_init_allocated_queue().

Fixes: commit 7c94e1c157a2 ("block: introduce blk_flush_queue to drive flush machinery")
Cc: 
Reviewed-by: Ming Lei 
Reviewed-by: Bart Van Assche 
Signed-off-by: xiao jin 
Signed-off-by: Jens Axboe 
[groeck: backport to v4.4.y/v4.9.y (context change)]
Signed-off-by: Guenter Roeck 
Signed-off-by: Alessio Balsini 
Signed-off-by: Greg Kroah-Hartman

block: do not use interruptible wait anywhere

2018-07-25T08:18:32+00:00

commit 1dc3039bc87ae7d19a990c3ee71cfd8a9068f428 upstream.

When blk_queue_enter() waits for a queue to unfreeze, or unset the
PREEMPT_ONLY flag, do not allow it to be interrupted by a signal.

The PREEMPT_ONLY flag was introduced later in commit 3a0a529971ec
("block, scsi: Make SCSI quiesce and resume work reliably").  Note the SCSI
device is resumed asynchronously, i.e. after un-freezing userspace tasks.

So that commit exposed the bug as a regression in v4.15.  A mysterious
SIGBUS (or -EIO) sometimes happened during the time the device was being
resumed.  Most frequently, there was no kernel log message, and we saw Xorg
or Xwayland killed by SIGBUS.[1]

[1] E.g. https://bugzilla.redhat.com/show_bug.cgi?id=1553979

Without this fix, I get an IO error in this test:

# dd if=/dev/sda of=/dev/null iflag=direct & \
  while killall -SIGUSR1 dd; do sleep 0.1; done & \
  echo mem > /sys/power/state ; \
  sleep 5; killall dd  # stop after 5 seconds

The interruptible wait was added to blk_queue_enter in
commit 3ef28e83ab15 ("block: generic request_queue reference counting").
Before then, the interruptible wait was only in blk-mq, but I don't think
it could ever have been correct.

Reviewed-by: Bart Van Assche 
Cc: stable@vger.kernel.org
Signed-off-by: Alan Jenkins 
Signed-off-by: Jens Axboe 
Signed-off-by: Sudip Mukherjee 
Signed-off-by: Greg Kroah-Hartman

block: wake up all tasks blocked in get_request()

2017-12-16T09:33:55+00:00

[ Upstream commit 34d9715ac1edd50285168dd8d80c972739a4f6a4 ]

Once blk_set_queue_dying() is done in blk_cleanup_queue(), we call
blk_freeze_queue() and wait for q->q_usage_counter becoming zero. But
if there are tasks blocked in get_request(), q->q_usage_counter can
never become zero. So we have to wake up all these tasks in
blk_set_queue_dying() first.

Fixes: 3ef28e83ab157997 ("block: generic request_queue reference counting")
Signed-off-by: Ming Lei 
Signed-off-by: Jens Axboe 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

block: Relax a check in blk_start_queue()

2017-09-27T09:00:14+00:00

commit 4ddd56b003f251091a67c15ae3fe4a5c5c5e390a upstream.

Calling blk_start_queue() from interrupt context with the queue
lock held and without disabling IRQs, as the skd driver does, is
safe. This patch avoids that loading the skd driver triggers the
following warning:

WARNING: CPU: 11 PID: 1348 at block/blk-core.c:283 blk_start_queue+0x84/0xa0
RIP: 0010:blk_start_queue+0x84/0xa0
Call Trace:
 skd_unquiesce_dev+0x12a/0x1d0 [skd]
 skd_complete_internal+0x1e7/0x5a0 [skd]
 skd_complete_other+0xc2/0xd0 [skd]
 skd_isr_completion_posted.isra.30+0x2a5/0x470 [skd]
 skd_isr+0x14f/0x180 [skd]
 irq_forced_thread_fn+0x2a/0x70
 irq_thread+0x144/0x1a0
 kthread+0x125/0x140
 ret_from_fork+0x2a/0x40

Fixes: commit a038e2536472 ("[PATCH] blk_start_queue() must be called with irq disabled - add warning")
Signed-off-by: Bart Van Assche 
Cc: Paolo 'Blaisorblade' Giarrusso 
Cc: Andrew Morton 
Cc: Christoph Hellwig 
Cc: Hannes Reinecke 
Cc: Johannes Thumshirn 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

blk: Ensure users for current->bio_list can see the full list.

2017-04-08T07:53:32+00:00

commit f5fe1b51905df7cfe4fdfd85c5fb7bc5b71a094f upstream.

Commit 79bd99596b73 ("blk: improve order of bio handling in generic_make_request()")
changed current->bio_list so that it did not contain *all* of the
queued bios, but only those submitted by the currently running
make_request_fn.

There are two places which walk the list and requeue selected bios,
and others that check if the list is empty.  These are no longer
correct.

So redefine current->bio_list to point to an array of two lists, which
contain all queued bios, and adjust various code to test or walk both
lists.

Signed-off-by: NeilBrown 
Fixes: 79bd99596b73 ("blk: improve order of bio handling in generic_make_request()")
Signed-off-by: Jens Axboe 
[jwang: backport to 4.4]
Signed-off-by: Jack Wang 
Signed-off-by: Greg Kroah-Hartman 
[bwh: Restore changes in device-mapper from upstream version]
Signed-off-by: Ben Hutchings

blk: improve order of bio handling in generic_make_request()

2017-04-08T07:53:32+00:00

commit 79bd99596b7305ab08109a8bf44a6a4511dbf1cd upstream.

To avoid recursion on the kernel stack when stacked block devices
are in use, generic_make_request() will, when called recursively,
queue new requests for later handling.  They will be handled when the
make_request_fn for the current bio completes.

If any bios are submitted by a make_request_fn, these will ultimately
be handled seqeuntially.  If the handling of one of those generates
further requests, they will be added to the end of the queue.

This strict first-in-first-out behaviour can lead to deadlocks in
various ways, normally because a request might need to wait for a
previous request to the same device to complete.  This can happen when
they share a mempool, and can happen due to interdependencies
particular to the device.  Both md and dm have examples where this happens.

These deadlocks can be erradicated by more selective ordering of bios.
Specifically by handling them in depth-first order.  That is: when the
handling of one bio generates one or more further bios, they are
handled immediately after the parent, before any siblings of the
parent.  That way, when generic_make_request() calls make_request_fn
for some particular device, we can be certain that all previously
submited requests for that device have been completely handled and are
not waiting for anything in the queue of requests maintained in
generic_make_request().

An easy way to achieve this would be to use a last-in-first-out stack
instead of a queue.  However this will change the order of consecutive
bios submitted by a make_request_fn, which could have unexpected consequences.
Instead we take a slightly more complex approach.
A fresh queue is created for each call to a make_request_fn.  After it completes,
any bios for a different device are placed on the front of the main queue, followed
by any bios for the same device, followed by all bios that were already on
the queue before the make_request_fn was called.
This provides the depth-first approach without reordering bios on the same level.

This, by itself, it not enough to remove all deadlocks.  It just makes
it possible for drivers to take the extra step required themselves.

To avoid deadlocks, drivers must never risk waiting for a request
after submitting one to generic_make_request.  This includes never
allocing from a mempool twice in the one call to a make_request_fn.

A common pattern in drivers is to call bio_split() in a loop, handling
the first part and then looping around to possibly split the next part.
Instead, a driver that finds it needs to split a bio should queue
(with generic_make_request) the second part, handle the first part,
and then return.  The new code in generic_make_request will ensure the
requests to underlying bios are processed first, then the second bio
that was split off.  If it splits again, the same process happens.  In
each case one bio will be completely handled before the next one is attempted.

With this is place, it should be possible to disable the
punt_bios_to_recover() recovery thread for many block devices, and
eventually it may be possible to remove it completely.

Ref: http://www.spinics.net/lists/raid/msg54680.html
Tested-by: Jinpu Wang 
Inspired-by: Lars Ellenberg 
Signed-off-by: NeilBrown 
Signed-off-by: Jens Axboe 
[jwang: backport to 4.4]
Signed-off-by: Jack Wang 
Signed-off-by: Greg Kroah-Hartman

block: Fix race triggered by blk_set_queue_dying()

2016-09-15T06:27:51+00:00

commit 1b856086813be9371929b6cc62045f9fd470f5a0 upstream.

blk_set_queue_dying() can be called while another thread is
submitting I/O or changing queue flags, e.g. through dm_stop_queue().
Hence protect the QUEUE_FLAG_DYING flag change with locking.

Signed-off-by: Bart Van Assche 
Cc: Christoph Hellwig 
Cc: Mike Snitzer 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

dm: fix excessive dm-mq context switching

2016-04-12T16:08:40+00:00

commit 6acfe68bac7e6f16dc312157b1fa6e2368985013 upstream.

Request-based DM's blk-mq support (dm-mq) was reported to be 50% slower
than if an underlying null_blk device were used directly.  One of the
reasons for this drop in performance is that blk_insert_clone_request()
was calling blk_mq_insert_request() with @async=true.  This forced the
use of kblockd_schedule_delayed_work_on() to run the blk-mq hw queues
which ushered in ping-ponging between process context (fio in this case)
and kblockd's kworker to submit the cloned request.  The ftrace
function_graph tracer showed:

  kworker-2013  =>   fio-12190
  fio-12190    =>  kworker-2013
  ...
  kworker-2013  =>   fio-12190
  fio-12190    =>  kworker-2013
  ...

Fixing blk_insert_clone_request()'s blk_mq_insert_request() call to
_not_ use kblockd to submit the cloned requests isn't enough to
eliminate the observed context switches.

In addition to this dm-mq specific blk-core fix, there are 2 DM core
fixes to dm-mq that (when paired with the blk-core fix) completely
eliminate the observed context switching:

1)  don't blk_mq_run_hw_queues in blk-mq request completion

    Motivated by desire to reduce overhead of dm-mq, punting to kblockd
    just increases context switches.

    In my testing against a really fast null_blk device there was no benefit
    to running blk_mq_run_hw_queues() on completion (and no other blk-mq
    driver does this).  So hopefully this change doesn't induce the need for
    yet another revert like commit 621739b00e16ca2d !

2)  use blk_mq_complete_request() in dm_complete_request()

    blk_complete_request() doesn't offer the traditional q->mq_ops vs
    .request_fn branching pattern that other historic block interfaces
    do (e.g. blk_get_request).  Using blk_mq_complete_request() for
    blk-mq requests is important for performance.  It should be noted
    that, like blk_complete_request(), blk_mq_complete_request() doesn't
    natively handle partial completions -- but the request-based
    DM-multipath target does provide the required partial completion
    support by dm.c:end_clone_bio() triggering requeueing of the request
    via dm-mpath.c:multipath_end_io()'s return of DM_ENDIO_REQUEUE.

dm-mq fix #2 is _much_ more important than #1 for eliminating the
context switches.
Before: cpu          : usr=15.10%, sys=59.39%, ctx=7905181, majf=0, minf=475
After:  cpu          : usr=20.60%, sys=79.35%, ctx=2008, majf=0, minf=472

With these changes multithreaded async read IOPs improved from ~950K
to ~1350K for this dm-mq stacked on null_blk test-case.  The raw read
IOPs of the underlying null_blk device for the same workload is ~1950K.

Fixes: 7fb4898e0 ("block: add blk-mq support to blk_insert_cloned_request()")
Fixes: bfebd1cdb ("dm: add full blk-mq support to request-based DM")
Reported-by: Sagi Grimberg 
Signed-off-by: Mike Snitzer 
Acked-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: add blk_start_queue_async()

2015-12-28T20:07:07+00:00

We currently only have an inline/sync helper to restart a stopped
queue. If drivers need an async version, they have to roll their
own. Add a generic helper instead.

Signed-off-by: Jens Axboe