Kashyap reports high CPU usage in blk_mq_queue_tag_busy_iter() and callees
using megaraid SAS RAID card since moving to shared tags [0].
Previously, when shared tags was shared sbitmap, this function was less
than optimum since we would iter through all tags for all hctx's,
yet only ever match upto tagset depth number of rqs.
Since the change to shared tags, things are even less efficient if we have
parallel callers of blk_mq_queue_tag_busy_iter(). This is because in
bt_iter() -> blk_mq_find_and_get_req() there would be more contention on
accessing each request ref and tags->lock since they are now shared among
all HW queues.
Optimise by having separate calls to bt_for_each() for when we're using
shared tags. In this case no longer pass a hctx, as it is no longer
relevant, and teach bt_iter() about this.
Ming suggested something along the lines of this change, apart from a
different implementation.
[0] https://lore.kernel.org/linux-block/e4e92abbe9d52bcba6b8cc6c91c442cc@mail.gmail.com/
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reported-and-tested-by: Kashyap Desai <kashyap.desai@broadcom.com>
Fixes: e155b0c238 ("blk-mq: Use shared tags for shared sbitmap support")
Link: https://lore.kernel.org/r/1638794990-137490-4-git-send-email-john.garry@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Typedefs busy_iter_fn and busy_tag_iter_fn are now identical, so delete
busy_iter_fn to reduce duplication.
It would be nicer to delete busy_tag_iter_fn, as the name busy_iter_fn is
less specific.
However busy_tag_iter_fn is used in many different parts of the tree,
unlike busy_iter_fn which is just use in block/, so just take the
straightforward path now, so that we could rename later treewide.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Tested-by: Kashyap Desai <kashyap.desai@broadcom.com>
Link: https://lore.kernel.org/r/1638794990-137490-3-git-send-email-john.garry@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The only user of blk_mq_hw_ctx blk_mq_hw_ctx argument is
blk_mq_rq_inflight().
Function blk_mq_rq_inflight() uses the hctx to find the associated request
queue to match against the request. However this same check is already
done in caller bt_iter(), so drop this check.
With that change there are no more users of busy_iter_fn blk_mq_hw_ctx
argument, so drop the argument.
Reviewed-by Hannes Reinecke <hare@suse.de>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Kashyap Desai <kashyap.desai@broadcom.com>
Link: https://lore.kernel.org/r/1638794990-137490-2-git-send-email-john.garry@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_run_dispatch_ops() is defined as one macro, and plug->mq_list
will be changed when running 'dispatch_ops', so add one local variable
for holding request queue.
Reported-and-tested-by: Yi Zhang <yi.zhang@redhat.com>
Fixes: 4cafe86c92 ("blk-mq: run dispatch lock once in case of issuing from list")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The operation protected via blk_mq_run_dispatch_ops() in blk_mq_run_hw_queue
won't sleep, so don't run might_sleep() for it.
Reported-and-tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It isn't necessary to call blk_mq_run_dispatch_ops() once for issuing
single request directly, and enough to do it one time when issuing from
whole list.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211203131534.3668411-5-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We have switched to allocate srcu into request queue, so it is fine
to pass request queue to blk_mq_run_dispatch_ops().
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211203131534.3668411-4-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In case of BLK_MQ_F_BLOCKING, per-hctx srcu is used to protect dispatch
critical area. However, this srcu instance stays at the end of hctx, and
it often takes standalone cacheline, often cold.
Inside srcu_read_lock() and srcu_read_unlock(), WRITE is always done on
the indirect percpu variable which is allocated from heap instead of
being embedded, srcu->srcu_idx is read only in srcu_read_lock(). It
doesn't matter if srcu structure stays in hctx or request queue.
So switch to per-request-queue srcu for protecting dispatch, and this
way simplifies quiesce a lot, not mention quiesce is always done on the
request queue wide.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211203131534.3668411-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove hctx_lock and hctx_unlock, and add one helper of
blk_mq_run_dispatch_ops() to run code block defined in dispatch_ops
with rcu/srcu read held.
Compared with hctx_lock()/hctx_unlock():
1) remove 2 branch to 1, so we just need to check
(hctx->flags & BLK_MQ_F_BLOCKING) once when running one dispatch_ops
2) srcu_idx needn't to be touched in case of non-blocking
3) might_sleep_if() can be moved to the blocking branch
Also put the added blk_mq_run_dispatch_ops() in private header, so that
the following patch can use it out of blk-mq.c.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211203131534.3668411-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
refcount_t is not as expensive as it used to be, but it's still more
expensive than the io_uring method of using atomic_t and just checking
for potential over/underflow.
This borrows that same implementation, which in turn is based on the
mm implementation from Linus.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't call into generic_file_read_iter() if we know it's O_DIRECT, just
set it up ourselves and call our own handler. This avoids an indirect call
for O_DIRECT.
Fall back to filemap_read() if we fail.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
No functional changes in this patch, just in preparation for efficiently
calling this light function from the block O_DIRECT handling.
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When we attempt to merge off the cached request path, we return NULL
if successful. This makes the caller believe that it's should allocate
a new request, and hence we end up with the bio both merged and associated
with a new request. This, predictably, leads to all sorts of crashes.
Pass in a pointer to the bio pointer, and clear it for the merge case.
Then the caller knows that the bio is already queued, and no new requests
need to get allocated.
Fixes: 5b13bc8a3f ("blk-mq: cleanup request allocation")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Expected case is returning a request, just check for success and return
the request rather than having an error label.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Without checking q->poll_stat in queue_poll_stat_show(), kernel panic
may be caused if q->poll_stat isn't allocated.
Fixes: 48b5c1fbcd ("block: only allocate poll_stats if there's a user of them")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211202090716.3292244-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that blk_execute_rq does not take a gendisk argument there is no need
to pass it through the scsi_ioctl callchain either.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20211126121802.2090656-6-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove the gendisk aregument to blk_execute_rq and blk_execute_rq_nowait
given that it is unused now. Also convert the boolean at_head parameter
to actually use the bool type while touching the prototype.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20211126121802.2090656-5-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Just use the disk attached to the request_queue instead.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20211126121802.2090656-4-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There is a 1:1 relationship between request_queues and gendisks now, so
no need for these extra checks.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20211126121802.2090656-3-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The block layer already performs this check, no need to duplicate it in
the driver.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20211126121802.2090656-2-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The variable ret is being initialized with a value that is never
read, it is being updated later on. The assignment is redundant and
can be removed.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Link: https://lore.kernel.org/r/20211126230652.1175636-1-colin.i.king@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Grab a reference to the newly allocated or existing io_context in
create_task_io_context and return it. This simplifies the callers and
removes the need for double lookups.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-13-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In __copy_io we know that the newly allocate task_struct does not have
an I/O context yet and is not exiting. So just allocate the I/O context
struct and install it directly. There is no need to lock the task
either as it is just being created.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-12-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After the prepare side has been moved to the only I/O scheduler that
cares, do the same for the cleanup and the NULL initialization.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-9-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move blk_mq_sched_assign_ioc so that many interfaces from the file can
be marked static. Rename the function to ioc_find_get_icq as well and
return the icq to simplify the interface.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-8-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This reverts commit 4896c4e64ba5d5d5acdbcf68c5910dd4f6d8fa62.
The helper is not needed any more.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-6-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move the copying of the I/O context to the block layer as that is where
we can use the proper low-level interfaces.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211126115817.2087431-3-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio->bi_opf isn't finalized before checking the bio, so use it after
submit_bio_checks() returns.
Fixes: 5b13bc8a3f ("blk-mq: cleanup request allocation")
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit 7cc4ffc555 ("block, bfq: put reqs of waker and woken in
dispatch list") added a condition to bfq_insert_request() which added
waker's requests directly to dispatch list. The rationale was that
completing waker's IO is needed to get more IO for the current queue.
Although this rationale is valid, there is a hole in it. The waker does
not necessarily serve the IO only for the current queue and maybe it's
current IO is not needed for current queue to make progress. Furthermore
injecting IO like this completely bypasses any service accounting within
bfq and thus we do not properly track how much service is waker's queue
getting or that the waker is actually doing any IO. Depending on the
conditions this can result in the waker getting too much or too few
service.
Consider for example the following job file:
[global]
directory=/mnt/repro/
rw=write
size=8g
time_based
runtime=30
ramp_time=10
blocksize=1m
direct=0
ioengine=sync
[slowwriter]
numjobs=1
prioclass=2
prio=7
fsync=200
[fastwriter]
numjobs=1
prioclass=2
prio=0
fsync=200
Despite processes have very different IO priorities, they get the same
about of service. The reason is that bfq identifies these processes as
having waker-wakee relationship and once that happens, IO from
fastwriter gets injected during slowwriter's time slice. As a result bfq
is not aware that fastwriter has any IO to do and constantly schedules
only slowwriter's queue. Thus fastwriter is forced to compete with
slowwriter's IO all the time instead of getting its share of time based
on IO priority.
Drop the special injection condition from bfq_insert_request(). As a
result, requests will be tracked and queued in a normal way and on next
dispatch bfq_select_queue() can decide whether the waker's inserted
requests should be injected during the current queue's timeslice or not.
Fixes: 7cc4ffc555 ("block, bfq: put reqs of waker and woken in dispatch list")
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-8-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Waker - wakee relationships are important in deciding whether one queue
can preempt the other one. Print information about detected waker-wakee
relationships so that scheduling decisions can be better understood from
block traces.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-7-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of having helper formating bfqq pid, provide a helper to
generate full bfqq name as used in the traces. It saves some code
duplication and will save more in the coming tracepoints.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-6-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, when process A starts issuing requests shortly after process
B has completed some IO three times in a row, we decide that B is a
"waker" of A meaning that completing IO of B is needed for A to make
progress and generally stop separating A's and B's IO much. This logic
is useful to avoid unnecessary idling and thus throughput loss for cases
where workload needs to switch e.g. between the process and the
journaling thread doing IO. However the detection heuristic tends to
frequently give false positives when A and B are fighting IO bandwidth
and other processes aren't doing much IO as we are basically deemed to
eventually accumulate three occurences of a situation where one process
starts issuing requests after the other has completed some IO. To reduce
these false positives, cancel the waker detection also if we didn't
accumulate three detected wakeups within given timeout. The rationale is
that if wakeups are really rare, the pointless idling doesn't hurt
throughput that much anyway.
This significantly reduces false waker detection for workload like:
[global]
directory=/mnt/repro/
rw=write
size=8g
time_based
runtime=30
ramp_time=10
blocksize=1m
direct=0
ioengine=sync
[slowwriter]
numjobs=1
fsync=200
[fastwriter]
numjobs=1
fsync=200
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-5-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When cgroup IO scheduling is used with BFQ it does not really provide
service differentiation if the cgroup drives a big IO depth. That for
example happens with writeback which asynchronously submits lots of IO
but it can happen with AIO as well. The problem is that if we have two
cgroups that submit IO with different weights, the cgroup with higher
weight properly gets more IO time and is able to dispatch more IO.
However this causes lower weight cgroup to accumulate more requests
inside BFQ and eventually lower weight cgroup consumes most of IO
scheduler tags. At that point higher weight cgroup stops getting better
service as it is mostly blocked waiting for a scheduler tag while its
queues inside BFQ are empty and thus lower weight cgroup gets served.
Check how many requests submitting cgroup has allocated in
bfq_limit_depth() and if it consumes more requests than what would
correspond to its weight limit available depth to 1 so that the cgroup
cannot consume many more requests. With this limitation the higher
weight cgroup gets proper service even with writeback.
Reviewed-by: Michal Koutný <mkoutny@suse.com>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-4-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Store bitmap depth shift inside bfq_data so that we can use it in
bfq_limit_depth() for proportioning when limiting number of available
request tags for a cgroup.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-3-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When we want to limit number of requests used by each bfqq and also
cgroup, we need to track also number of requests used by each cgroup.
So track number of allocated requests for each bfq_entity.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-2-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently we lookup ICQ only after the request is allocated. However BFQ
will want to decide how many scheduler tags it allows a given bfq queue
(effectively a process) to consume based on cgroup weight. So provide a
function blk_mq_sched_get_icq() so that BFQ can lookup ICQ earlier.
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20211125133645.27483-1-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The completion callback for the sdhci-pci device is invoked from a
kworker.
I couldn't identify in which context is mmc_blk_mq_req_done() invoke but
the remaining caller are from invoked from preemptible context. Here it
would make sense to complete the request directly instead scheduling
ksoftirqd for its completion.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20211025070658.1565848-3-bigeasy@linutronix.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add blk_mq_complete_request_direct() which completes the block request
directly instead deferring it to softirq for single queue devices.
This is useful for devices which complete the requests in preemptible
context and raising softirq from means scheduling ksoftirqd.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211025070658.1565848-2-bigeasy@linutronix.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This function is trivial and is only used in one place. Having this
function is misleading because it implies that blk_crypto_register()
needs to be paired with blk_crypto_unregister(), which is not the case.
Just set disk->queue->crypto_profile to NULL directly.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211124013733.347612-1-ebiggers@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Refactor the request alloction so that blk_mq_get_cached_request tries
to find a cached request first, and the entirely separate and now
self contained blk_mq_get_new_requests allocates one or more requests
if that is not possible.
There is a small change in behavior as submit_bio_checks is called
twice now if a cached request is present but can't be used, but that
is a small price to pay for unwinding this code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211124062856.1444266-1-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
