It is possible to call fsync on a read-only handle (for example, fsck.ext2
does it when doing read-only check), and this call results in kernel
warning.

The patch b089cfd9 ("block: don't warn for flush on read-only device")
attempted to disable the warning, but it is buggy and it doesn't
(op_is_flush tests flags, but bio_op strips off the flags).
Signed-off-by: NMikulas Patocka <mpatocka@redhat.com>
Fixes: 721c7fc7 ("block: fail op_is_write() requests to read-only partitions")
Cc: stable@vger.kernel.org	# 4.18
Signed-off-by: NJens Axboe <axboe@kernel.dk>

There is a very small change a bio gets caught up in a really
unfortunate race between a task migration, cgroup exiting, and itself
trying to associate with a blkg. This is due to css offlining being
performed after the css->refcnt is killed which triggers removal of
blkgs that reach their blkg->refcnt of 0.

To avoid this, association with a blkg should use tryget and fallback to
using the root_blkg.

Fixes: 08e18eab ("block: add bi_blkg to the bio for cgroups")
Reviewed-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NDennis Zhou <dennisszhou@gmail.com>
Cc: Jiufei Xue <jiufei.xue@linux.alibaba.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Josef Bacik <josef@toxicpanda.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently, blkcg destruction relies on a sequence of events:
  1. Destruction starts. blkcg_css_offline() is called and blkgs
     release their reference to the blkcg. This immediately destroys
     the cgwbs (writeback).
  2. With blkgs giving up their reference, the blkcg ref count should
     become zero and eventually call blkcg_css_free() which finally
     frees the blkcg.

Jiufei Xue reported that there is a race between blkcg_bio_issue_check()
and cgroup_rmdir(). To remedy this, blkg destruction becomes contingent
on the completion of all writeback associated with the blkcg. A count of
the number of cgwbs is maintained and once that goes to zero, blkg
destruction can follow. This should prevent premature blkg destruction
related to writeback.

The new process for blkcg cleanup is as follows:
  1. Destruction starts. blkcg_css_offline() is called which offlines
     writeback. Blkg destruction is delayed on the cgwb_refcnt count to
     avoid punting potentially large amounts of outstanding writeback
     to root while maintaining any ongoing policies. Here, the base
     cgwb_refcnt is put back.
  2. When the cgwb_refcnt becomes zero, blkcg_destroy_blkgs() is called
     and handles destruction of blkgs. This is where the css reference
     held by each blkg is released.
  3. Once the blkcg ref count goes to zero, blkcg_css_free() is called.
     This finally frees the blkg.

It seems in the past blk-throttle didn't do the most understandable
things with taking data from a blkg while associating with current. So,
the simplification and unification of what blk-throttle is doing caused
this.

Fixes: 08e18eab ("block: add bi_blkg to the bio for cgroups")
Reviewed-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NDennis Zhou <dennisszhou@gmail.com>
Cc: Jiufei Xue <jiufei.xue@linux.alibaba.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Josef Bacik <josef@toxicpanda.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This reverts commit 4c699480.

Destroying blkgs is tricky because of the nature of the relationship. A
blkg should go away when either a blkcg or a request_queue goes away.
However, blkg's pin the blkcg to ensure they remain valid. To break this
cycle, when a blkcg is offlined, blkgs put back their css ref. This
eventually lets css_free() get called which frees the blkcg.

The above commit (4c699480) breaks this order of events by trying to
destroy blkgs in css_free(). As the blkgs still hold references to the
blkcg, css_free() is never called.

The race between blkcg_bio_issue_check() and cgroup_rmdir() will be
addressed in the following patch by delaying destruction of a blkg until
all writeback associated with the blkcg has been finished.

Fixes: 4c699480 ("blk-throttle: fix race between blkcg_bio_issue_check() and cgroup_rmdir()")
Reviewed-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NDennis Zhou <dennisszhou@gmail.com>
Cc: Jiufei Xue <jiufei.xue@linux.alibaba.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently, variable ref_count within the bsg_device struct is of
type atomic_t.  For variables being used as reference counters,
the refcount API should be used instead of atomic.  The newer
refcount API works to prevent counter overflows and use-after-free
bugs.  So, move this varable from the atomic API to refcount,
potentially avoiding the issues mentioned.
Signed-off-by: NJohn Pittman <jpittman@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

kmem_cache_destroy() can handle NULL pointer correctly, so there is
no need to check e->icq_cache before calling kmem_cache_destroy().
Signed-off-by: NChengguang Xu <cgxu519@gmx.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We already note and mark discard and swap IO from bio_to_wbt_flags().
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We have two potential issues:

1) After commit 2887e41b, we only wake one process at the time when
   we finish an IO. We really want to wake up as many tasks as can
   queue IO. Before this commit, we woke up everyone, which could cause
   a thundering herd issue.

2) A task can potentially consume two wakeups, causing us to (in
   practice) miss a wakeup.

Fix both by providing our own wakeup function, which stops
__wake_up_common() from waking up more tasks if we fail to get a
queueing token. With the strict ordering we have on the wait list, this
wakes the right tasks and the right amount of tasks.

Based on a patch from Jianchao Wang <jianchao.w.wang@oracle.com>.
Tested-by: NAgarwal, Anchal <anchalag@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Prep patch for calling the handler from a different context,
no functional changes in this patch.
Tested-by: NAgarwal, Anchal <anchalag@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

A previous commit removed the ability to have per-rq flags. We used
those flags to maintain inflight counts. Since we don't have those
anymore, we have to always maintain inflight counts, even if wbt is
disabled. This is clearly suboptimal.

Add a queue quiesce around changing the wbt latency settings from sysfs
to work around this. With that, we can reliably put the enabled check in
our bio_to_wbt_flags(), since we know the WBT_TRACKED flag will be
consistent for the lifetime of the request.

Fixes: c1c80384 ("block: remove external dependency on wbt_flags")
Reviewed-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We need to do this inside the loop as well, or we can allow new
IO to supersede previous IO.
Tested-by: NAnchal Agarwal <anchalag@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We need the memory barrier before checking the list head,
use the appropriate helper for this. The matching queue
side memory barrier is provided by set_current_state().
Tested-by: NAnchal Agarwal <anchalag@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Check it in one place, instead of in multiple places.
Tested-by: NAnchal Agarwal <anchalag@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

For blk-mq, part_in_flight/rw will invoke blk_mq_in_flight/rw to
account the inflight requests. It will access the queue_hw_ctx and
nr_hw_queues w/o any protection. When updating nr_hw_queues and
blk_mq_in_flight/rw occur concurrently, panic comes up.

Before update nr_hw_queues, the q will be frozen. So we could use
q_usage_counter to avoid the race. percpu_ref_is_zero is used here
so that we will not miss any in-flight request. The access to
nr_hw_queues and queue_hw_ctx in blk_mq_queue_tag_busy_iter are
under rcu critical section, __blk_mq_update_nr_hw_queues could use
synchronize_rcu to ensure the zeroed q_usage_counter to be globally
visible.
Signed-off-by: NJianchao Wang <jianchao.w.wang@oracle.com>
Reviewed-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently, when update nr_hw_queues, IO scheduler's init_hctx will
be invoked before the mapping between ctx and hctx is adapted
correctly by blk_mq_map_swqueue. The IO scheduler init_hctx (kyber)
may depend on this mapping and get wrong result and panic finally.
A simply way to fix this is that switch the IO scheduler to 'none'
before update the nr_hw_queues, and then switch it back after
update nr_hw_queues. blk_mq_sched_init_/exit_hctx are removed due
to nobody use them any more.
Signed-off-by: NJianchao Wang <jianchao.w.wang@oracle.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This patch removes the duplicate initialization of q->queue_head
in the blk_alloc_queue_node(). This removes the 2nd initialization
so that we preserve the initialization order same as declaration
present in struct request_queue.
Reviewed-by: NOmar Sandoval <osandov@fb.com>
Signed-off-by: NChaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Because blk_do_io_stat() only does a judgement about the request
contributes to IO statistics, it better changes return type to bool.
Signed-off-by: NChengguang Xu <cgxu519@gmx.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The value that struct cftype .write() method returns is then directly
returned to userspace as the value returned by write() syscall, so it
should be the number of bytes actually written (or consumed) and not zero.

Returning zero from write() syscall makes programs like /bin/echo or bash
spin.
Signed-off-by: NMaciej S. Szmigiero <mail@maciej.szmigiero.name>
Fixes: e21b7a0b ("block, bfq: add full hierarchical scheduling and cgroups support")
Cc: stable@vger.kernel.org
Signed-off-by: NJens Axboe <axboe@kernel.dk>

bfq_bfqq_charge_time contains some lengthy and redundant code. This
commit trims and condenses that code.
Signed-off-by: NPaolo Valente <paolo.valente@linaro.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

When a sync request is dispatched, the queue that contains that
request, and all the ancestor entities of that queue, are charged with
the number of sectors of the request. In constrast, if the request is
async, then the queue and its ancestor entities are charged with the
number of sectors of the request, multiplied by an overcharge
factor. This throttles the bandwidth for async I/O, w.r.t. to sync
I/O, and it is done to counter the tendency of async writes to steal
I/O throughput to reads.

On the opposite end, the lower this parameter, the stabler I/O
control, in the following respect.  The lower this parameter is, the
less the bandwidth enjoyed by a group decreases
- when the group does writes, w.r.t. to when it does reads;
- when other groups do reads, w.r.t. to when they do writes.

The fixes "block, bfq: always update the budget of an entity when
needed" and "block, bfq: readd missing reset of parent-entity service"
improved I/O control in bfq to such an extent that it has been
possible to revise this overcharge factor downwards.  This commit
introduces the resulting, new value.
Signed-off-by: NPaolo Valente <paolo.valente@linaro.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

When the next child entity to serve changes for a given parent entity,
the budget of that parent entity must be updated accordingly.
Unfortunately, this update is not performed, by mistake, for the
entities that happen to switch from having no child entity to serve,
to having one child entity to serve.
Signed-off-by: NPaolo Valente <paolo.valente@linaro.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The received-service counter needs to be equal to 0 when an entity is
set in service. Unfortunately, commit "block, bfq: fix service being
wrongly set to zero in case of preemption" mistakenly removed the
resetting of this counter for the parent entities of the bfq_queue
being set in service. This commit fixes this issue by resetting
service for parent entities, directly on the expiration of the
in-service bfq_queue.

Fixes: 9fae8dd5 ("block, bfq: fix service being wrongly set to zero in case of preemption")
Signed-off-by: NPaolo Valente <paolo.valente@linaro.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

On wbt invariant is that if one IO is tracked via WBT_TRACKED, rqw->inflight
should be updated for tracking this IO.

But commit c1c80384 ("block: remove external dependency on wbt_flags")
forgets to remove the early handling of !rwb_enabled(rwb) inside wbt_wait(),
then the inflight counter may not be increased in wbt_wait(), but decreased
in wbt_done() for this kind of IO, so this counter may become negative, then
wbt_wait() may wait forever.

This patch fixes the report in the following link:

	https://marc.info/?l=linux-block&m=153221542021033&w=2

Fixes: c1c80384 ("block: remove external dependency on wbt_flags")
Cc: Josef Bacik <jbacik@fb.com>
Reported-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Don't warn for a flush issued to a read-only device. It's not strictly
a writable command, as it doesn't change any on-media data by itself.
Reported-by: NStefan Agner <stefan@agner.ch>
Fixes: 721c7fc7 ("block: fail op_is_write() requests to read-only partitions")
Signed-off-by: NJens Axboe <axboe@kernel.dk>

For legacy queues the only call of blkg_root_lookup() happens after
bypass mode has been enabled. Since blkg_lookup() returns NULL for
queues in bypass mode, modify the blkg_root_lookup() such that it
no longer depends on bypass mode. Rename the function into
blk_queue_root_blkg() as suggested by Tejun.
Suggested-by: NTejun Heo <tj@kernel.org>
Fixes: 6bad9b21 ("blkcg: Introduce blkg_root_lookup()")
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

When an application's iops has exceeded its cgroup's iops limit, surely it
is throttled and kernel will set a timer for dispatching, thus IO latency
includes the delay.

However, the dispatch delay which is calculated by the limit and the
elapsed jiffies is suboptimal.  As the dispatch delay is only calculated
once the application's iops is (iops limit + 1), it doesn't need to wait
any longer than the remaining time of the current slice.

The difference can be proved by the following fio job and cgroup iops
setting,
-----
$ echo 4 > /mnt/config/nullb/disk1/mbps    # limit nullb's bandwidth to 4MB/s for testing.
$ echo "253:1 riops=100 rbps=max" > /sys/fs/cgroup/unified/cg1/io.max
$ cat r2.job
[global]
name=fio-rand-read
filename=/dev/nullb1
rw=randread
bs=4k
direct=1
numjobs=1
time_based=1
runtime=60
group_reporting=1

[file1]
size=4G
ioengine=libaio
iodepth=1
rate_iops=50000
norandommap=1
thinktime=4ms
-----

wo patch:
file1: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=1
fio-3.7-66-gedfc
Starting 1 process

   read: IOPS=99, BW=400KiB/s (410kB/s)(23.4MiB/60001msec)
    slat (usec): min=10, max=336, avg=27.71, stdev=17.82
    clat (usec): min=2, max=28887, avg=5929.81, stdev=7374.29
     lat (usec): min=24, max=28901, avg=5958.73, stdev=7366.22
    clat percentiles (usec):
     |  1.00th=[    4],  5.00th=[    4], 10.00th=[    4], 20.00th=[    4],
     | 30.00th=[    4], 40.00th=[    4], 50.00th=[    6], 60.00th=[11731],
     | 70.00th=[11863], 80.00th=[11994], 90.00th=[12911], 95.00th=[22676],
     | 99.00th=[23725], 99.50th=[23987], 99.90th=[23987], 99.95th=[25035],
     | 99.99th=[28967]

w/ patch:
file1: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=1
fio-3.7-66-gedfc
Starting 1 process

   read: IOPS=100, BW=400KiB/s (410kB/s)(23.4MiB/60005msec)
    slat (usec): min=10, max=155, avg=23.24, stdev=16.79
    clat (usec): min=2, max=12393, avg=5961.58, stdev=5959.25
     lat (usec): min=23, max=12412, avg=5985.91, stdev=5951.92
    clat percentiles (usec):
     |  1.00th=[    3],  5.00th=[    3], 10.00th=[    4], 20.00th=[    4],
     | 30.00th=[    4], 40.00th=[    5], 50.00th=[   47], 60.00th=[11863],
     | 70.00th=[11994], 80.00th=[11994], 90.00th=[11994], 95.00th=[11994],
     | 99.00th=[11994], 99.50th=[11994], 99.90th=[12125], 99.95th=[12125],
     | 99.99th=[12387]
Signed-off-by: NLiu Bo <bo.liu@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Several block drivers call alloc_disk() followed by put_disk() if
something fails before device_add_disk() is called without calling
blk_cleanup_queue(). Make sure that also for this scenario a request
queue is dissociated from the cgroup controller. This patch avoids
that loading the parport_pc, paride and pf drivers triggers the
following kernel crash:

BUG: KASAN: null-ptr-deref in pi_init+0x42e/0x580 [paride]
Read of size 4 at addr 0000000000000008 by task modprobe/744
Call Trace:
dump_stack+0x9a/0xeb
kasan_report+0x139/0x350
pi_init+0x42e/0x580 [paride]
pf_init+0x2bb/0x1000 [pf]
do_one_initcall+0x8e/0x405
do_init_module+0xd9/0x2f2
load_module+0x3ab4/0x4700
SYSC_finit_module+0x176/0x1a0
do_syscall_64+0xee/0x2b0
entry_SYSCALL_64_after_hwframe+0x42/0xb7
Reported-by: NAlexandru Moise <00moses.alexander00@gmail.com>
Fixes: a063057d ("block: Fix a race between request queue removal and the block cgroup controller") # v4.17
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Tested-by: NAlexandru Moise <00moses.alexander00@gmail.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Alexandru Moise <00moses.alexander00@gmail.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This patch does not change any functionality.
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Alexandru Moise <00moses.alexander00@gmail.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently, we count the hctx as active after allocate driver tag
successfully. If a previously inactive hctx try to get tag first
time, it may fails and need to wait. However, due to the stale tag
->active_queues, the other shared-tags users are still able to
occupy all driver tags while there is someone waiting for tag.
Consequently, even if the previously inactive hctx is waked up, it
still may not be able to get a tag and could be starved.

To fix it, we count the hctx as active before try to allocate driver
tag, then when it is waiting the tag, the other shared-tag users
will reserve budget for it.
Reviewed-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJianchao Wang <jianchao.w.wang@oracle.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

In commit ed996a52 ("block: simplify and cleanup bvec pool
handling"), the value of the slab index is incremented by one in
bvec_alloc() after the allocation is done to indicate an index value of
0 does not need to be later freed.

bvec_nr_vecs() was not updated accordingly, and thus returns the wrong
value.  Decrement idx before performing the lookup.

Fixes: ed996a52 ("block: simplify and cleanup bvec pool handling")
Signed-off-by: NGreg Edwards <gedwards@ddn.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This patch does not change any functionality but avoids that gcc
reports the following warnings when building with W=1:

block/cfq-iosched.c: In function ?cfq_back_seek_max_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4756:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_back_seek_max_store, &cfqd->cfq_back_max, 0, UINT_MAX, 0);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_slice_idle_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4759:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_slice_idle_store, &cfqd->cfq_slice_idle, 0, UINT_MAX, 1);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_group_idle_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4760:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_group_idle_store, &cfqd->cfq_group_idle, 0, UINT_MAX, 1);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_low_latency_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4765:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_low_latency_store, &cfqd->cfq_latency, 0, 1, 0);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_slice_idle_us_store?:
block/cfq-iosched.c:4775:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4782:1: note: in expansion of macro ?USEC_STORE_FUNCTION?
 USEC_STORE_FUNCTION(cfq_slice_idle_us_store, &cfqd->cfq_slice_idle, 0, UINT_MAX);
 ^~~~~~~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_group_idle_us_store?:
block/cfq-iosched.c:4775:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4783:1: note: in expansion of macro ?USEC_STORE_FUNCTION?
 USEC_STORE_FUNCTION(cfq_group_idle_us_store, &cfqd->cfq_group_idle, 0, UINT_MAX);
 ^~~~~~~~~~~~~~~~~~~
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This patch avoids that gcc complains about fall-through when building
with W=1.
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

I am currently running a large bare metal instance (i3.metal)
on EC2 with 72 cores, 512GB of RAM and NVME drives, with a
4.18 kernel. I have a workload that simulates a database
workload and I am running into lockup issues when writeback
throttling is enabled,with the hung task detector also
kicking in.

Crash dumps show that most CPUs (up to 50 of them) are
all trying to get the wbt wait queue lock while trying to add
themselves to it in __wbt_wait (see stack traces below).

[    0.948118] CPU: 45 PID: 0 Comm: swapper/45 Not tainted 4.14.51-62.38.amzn1.x86_64 #1
[    0.948119] Hardware name: Amazon EC2 i3.metal/Not Specified, BIOS 1.0 10/16/2017
[    0.948120] task: ffff883f7878c000 task.stack: ffffc9000c69c000
[    0.948124] RIP: 0010:native_queued_spin_lock_slowpath+0xf8/0x1a0
[    0.948125] RSP: 0018:ffff883f7fcc3dc8 EFLAGS: 00000046
[    0.948126] RAX: 0000000000000000 RBX: ffff887f7709ca68 RCX: ffff883f7fce2a00
[    0.948128] RDX: 000000000000001c RSI: 0000000000740001 RDI: ffff887f7709ca68
[    0.948129] RBP: 0000000000000002 R08: 0000000000b80000 R09: 0000000000000000
[    0.948130] R10: ffff883f7fcc3d78 R11: 000000000de27121 R12: 0000000000000002
[    0.948131] R13: 0000000000000003 R14: 0000000000000000 R15: 0000000000000000
[    0.948132] FS:  0000000000000000(0000) GS:ffff883f7fcc0000(0000) knlGS:0000000000000000
[    0.948134] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.948135] CR2: 000000c424c77000 CR3: 0000000002010005 CR4: 00000000003606e0
[    0.948136] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    0.948137] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    0.948138] Call Trace:
[    0.948139]  <IRQ>
[    0.948142]  do_raw_spin_lock+0xad/0xc0
[    0.948145]  _raw_spin_lock_irqsave+0x44/0x4b
[    0.948149]  ? __wake_up_common_lock+0x53/0x90
[    0.948150]  __wake_up_common_lock+0x53/0x90
[    0.948155]  wbt_done+0x7b/0xa0
[    0.948158]  blk_mq_free_request+0xb7/0x110
[    0.948161]  __blk_mq_complete_request+0xcb/0x140
[    0.948166]  nvme_process_cq+0xce/0x1a0 [nvme]
[    0.948169]  nvme_irq+0x23/0x50 [nvme]
[    0.948173]  __handle_irq_event_percpu+0x46/0x300
[    0.948176]  handle_irq_event_percpu+0x20/0x50
[    0.948179]  handle_irq_event+0x34/0x60
[    0.948181]  handle_edge_irq+0x77/0x190
[    0.948185]  handle_irq+0xaf/0x120
[    0.948188]  do_IRQ+0x53/0x110
[    0.948191]  common_interrupt+0x87/0x87
[    0.948192]  </IRQ>
....
[    0.311136] CPU: 4 PID: 9737 Comm: run_linux_amd64 Not tainted 4.14.51-62.38.amzn1.x86_64 #1
[    0.311137] Hardware name: Amazon EC2 i3.metal/Not Specified, BIOS 1.0 10/16/2017
[    0.311138] task: ffff883f6e6a8000 task.stack: ffffc9000f1ec000
[    0.311141] RIP: 0010:native_queued_spin_lock_slowpath+0xf5/0x1a0
[    0.311142] RSP: 0018:ffffc9000f1efa28 EFLAGS: 00000046
[    0.311144] RAX: 0000000000000000 RBX: ffff887f7709ca68 RCX: ffff883f7f722a00
[    0.311145] RDX: 0000000000000035 RSI: 0000000000d80001 RDI: ffff887f7709ca68
[    0.311146] RBP: 0000000000000202 R08: 0000000000140000 R09: 0000000000000000
[    0.311147] R10: ffffc9000f1ef9d8 R11: 000000001a249fa0 R12: ffff887f7709ca68
[    0.311148] R13: ffffc9000f1efad0 R14: 0000000000000000 R15: ffff887f7709ca00
[    0.311149] FS:  000000c423f30090(0000) GS:ffff883f7f700000(0000) knlGS:0000000000000000
[    0.311150] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.311151] CR2: 00007feefcea4000 CR3: 0000007f7016e001 CR4: 00000000003606e0
[    0.311152] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    0.311153] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    0.311154] Call Trace:
[    0.311157]  do_raw_spin_lock+0xad/0xc0
[    0.311160]  _raw_spin_lock_irqsave+0x44/0x4b
[    0.311162]  ? prepare_to_wait_exclusive+0x28/0xb0
[    0.311164]  prepare_to_wait_exclusive+0x28/0xb0
[    0.311167]  wbt_wait+0x127/0x330
[    0.311169]  ? finish_wait+0x80/0x80
[    0.311172]  ? generic_make_request+0xda/0x3b0
[    0.311174]  blk_mq_make_request+0xd6/0x7b0
[    0.311176]  ? blk_queue_enter+0x24/0x260
[    0.311178]  ? generic_make_request+0xda/0x3b0
[    0.311181]  generic_make_request+0x10c/0x3b0
[    0.311183]  ? submit_bio+0x5c/0x110
[    0.311185]  submit_bio+0x5c/0x110
[    0.311197]  ? __ext4_journal_stop+0x36/0xa0 [ext4]
[    0.311210]  ext4_io_submit+0x48/0x60 [ext4]
[    0.311222]  ext4_writepages+0x810/0x11f0 [ext4]
[    0.311229]  ? do_writepages+0x3c/0xd0
[    0.311239]  ? ext4_mark_inode_dirty+0x260/0x260 [ext4]
[    0.311240]  do_writepages+0x3c/0xd0
[    0.311243]  ? _raw_spin_unlock+0x24/0x30
[    0.311245]  ? wbc_attach_and_unlock_inode+0x165/0x280
[    0.311248]  ? __filemap_fdatawrite_range+0xa3/0xe0
[    0.311250]  __filemap_fdatawrite_range+0xa3/0xe0
[    0.311253]  file_write_and_wait_range+0x34/0x90
[    0.311264]  ext4_sync_file+0x151/0x500 [ext4]
[    0.311267]  do_fsync+0x38/0x60
[    0.311270]  SyS_fsync+0xc/0x10
[    0.311272]  do_syscall_64+0x6f/0x170
[    0.311274]  entry_SYSCALL_64_after_hwframe+0x42/0xb7

In the original patch, wbt_done is waking up all the exclusive
processes in the wait queue, which can cause a thundering herd
if there is a large number of writer threads in the queue. The
original intention of the code seems to be to wake up one thread
only however, it uses wake_up_all() in __wbt_done(), and then
uses the following check in __wbt_wait to have only one thread
actually get out of the wait loop:

if (waitqueue_active(&rqw->wait) &&
            rqw->wait.head.next != &wait->entry)
                return false;

The problem with this is that the wait entry in wbt_wait is
define with DEFINE_WAIT, which uses the autoremove wakeup function.
That means that the above check is invalid - the wait entry will
have been removed from the queue already by the time we hit the
check in the loop.

Secondly, auto-removing the wait entries also means that the wait
queue essentially gets reordered "randomly" (e.g. threads re-add
themselves in the order they got to run after being woken up).
Additionally, new requests entering wbt_wait might overtake requests
that were queued earlier, because the wait queue will be
(temporarily) empty after the wake_up_all, so the waitqueue_active
check will not stop them. This can cause certain threads to starve
under high load.

The fix is to leave the woken up requests in the queue and remove
them in finish_wait() once the current thread breaks out of the
wait loop in __wbt_wait. This will ensure new requests always
end up at the back of the queue, and they won't overtake requests
that are already in the wait queue. With that change, the loop
in wbt_wait is also in line with many other wait loops in the kernel.
Waking up just one thread drastically reduces lock contention, as
does moving the wait queue add/remove out of the loop.

A significant drop in lockdep's lock contention numbers is seen when
running the test application on the patched kernel.
Signed-off-by: NAnchal Agarwal <anchalag@amazon.com>
Signed-off-by: NFrank van der Linden <fllinden@amazon.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

It turns out that commit 721c7fc7 ("block: fail op_is_write()
requests to read-only partitions"), while obviously correct, causes
problems for some older lvm2 installations.

The reason is that the lvm snapshotting will continue to write to the
snapshow COW volume, even after the volume has been marked read-only.
End result: snapshot failure.

This has actually been fixed in newer version of the lvm2 tool, but the
old tools still exist, and the breakage was reported both in the kernel
bugzilla and in the Debian bugzilla:

  https://bugzilla.kernel.org/show_bug.cgi?id=200439
  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=900442

The lvm2 fix is here

  https://sourceware.org/git/?p=lvm2.git;a=commit;h=a6fdb9d9d70f51c49ad11a87ab4243344e6701a3

but until everybody has updated to recent versions, we'll have to weaken
the "never write to read-only partitions" check.  It now allows the
write to happen, but causes a warning, something like this:

  generic_make_request: Trying to write to read-only block-device dm-3 (partno X)
  Modules linked in: nf_tables xt_cgroup xt_owner kvm_intel iwlmvm kvm irqbypass iwlwifi
  CPU: 1 PID: 77 Comm: kworker/1:1 Not tainted 4.17.9-gentoo #3
  Hardware name: LENOVO 20B6A019RT/20B6A019RT, BIOS GJET91WW (2.41 ) 09/21/2016
  Workqueue: ksnaphd do_metadata
  RIP: 0010:generic_make_request_checks+0x4ac/0x600
  ...
  Call Trace:
   generic_make_request+0x64/0x400
   submit_bio+0x6c/0x140
   dispatch_io+0x287/0x430
   sync_io+0xc3/0x120
   dm_io+0x1f8/0x220
   do_metadata+0x1d/0x30
   process_one_work+0x1b9/0x3e0
   worker_thread+0x2b/0x3c0
   kthread+0x113/0x130
   ret_from_fork+0x35/0x40

Note that this is a "revert" in behavior only.  I'm leaving alone the
actual code cleanups in commit 721c7fc7, but letting the previously
uncaught request go through with a warning instead of stopping it.

Fixes: 721c7fc7 ("block: fail op_is_write() requests to read-only partitions")
Reported-and-tested-by: NWGH <wgh@torlan.ru>
Acked-by: NMike Snitzer <snitzer@redhat.com>
Cc: Sagi Grimberg <sagi@grimberg.me>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Zdenek Kabelac <zkabelac@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Commit d250bf4e("blk-mq: only iterate over inflight requests
in blk_mq_tagset_busy_iter") uses 'blk_mq_rq_state(rq) == MQ_RQ_IN_FLIGHT'
to replace 'blk_mq_request_started(req)', this way is wrong, and causes
lots of test system hang during booting.

Fix the issue by using blk_mq_request_started(req) inside bt_tags_iter().

Fixes: d250bf4e ("blk-mq: only iterate over inflight requests in blk_mq_tagset_busy_iter")
Cc: Josef Bacik <josef@toxicpanda.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Mark Brown <broonie@kernel.org>
Cc: Matt Hart <matthew.hart@linaro.org>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: John Garry <john.garry@huawei.com>
Cc: Hannes Reinecke <hare@suse.com>,
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>,
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Cc: linux-scsi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Tested-by: NGuenter Roeck <linux@roeck-us.net>
Reported-by: NMark Brown <broonie@kernel.org>
Reported-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The passed 'nr' from userspace represents the total depth, meantime
inside 'struct blk_mq_tags', 'nr_tags' stores the total tag depth,
and 'nr_reserved_tags' stores the reserved part.

There are two issues in blk_mq_tag_update_depth() now:

1) for growing tags, we should have used the passed 'nr', and keep the
number of reserved tags not changed.

2) the passed 'nr' should have been used for checking against
'tags->nr_tags', instead of number of the normal part.

This patch fixes the above two cases, and avoids kernel crash caused
by wrong resizing sbitmap queue.

Cc: "Ewan D. Milne" <emilne@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Omar Sandoval <osandov@fb.com>
Tested by: Marco Patalano <mpatalan@redhat.com>
Signed-off-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Runtime PM isn't ready for blk-mq yet, and commit 765e40b6 ("block:
disable runtime-pm for blk-mq") tried to disable it. Unfortunately,
it can't take effect in that way since user space still can switch
it on via 'echo auto > /sys/block/sdN/device/power/control'.

This patch disables runtime-pm for blk-mq really by pm_runtime_disable()
and fixes all kinds of PM related kernel crash.

Cc: Tomas Janousek <tomi@nomi.cz>
Cc: Przemek Socha <soprwa@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: <stable@vger.kernel.org>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Tested-by: NPatrick Steinhardt <ps@pks.im>
Signed-off-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently, avg_lat is calculated by accumulating the mean of every
window in a long running cumulative average. As time goes on, the metric
becomes less and less useful due to the accumulated history.

This patch reuses the same calculation done in load averages to make the
avg_lat metric more lively. Unlike load averages, the avg only advances
when a window elapses (due to an io). Idle periods extend the most
recent window. Bucketing is used to limit the history of avg_lat by
binding it to the window size. So, the window range for 1/exp (decay
rate) is [1 min, 2.5 min) when windows elapse immediately.

The current sample window size is exposed in the debug info to enable
calculation of the window range.
Signed-off-by: NDennis Zhou <dennisszhou@gmail.com>
Acked-by: NTejun Heo <tj@kernel.org>
Acked-by: NJohannes Weiner <hannes@cmpxchg.org>
Acked-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The blkg lifetime is protected by the queue lifetime, so we need to put
the queue _after_ we're done using the blkg.
Signed-off-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

At this point we have a ref on the blkg, we need to drop it if we don't
have a iolat.
Signed-off-by: NJosef Bacik <josef@toxicpanda.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>