elevator_init_mq() is only called before adding disk, when there isn't
any FS I/O, only passthrough requests can be queued, so freezing queue
plus canceling dispatch work is enough to drain any dispatch activities,
then we can avoid synchronize_srcu() in blk_mq_quiesce_queue().
Long boot latency issue can be fixed in case of lots of disks added
during booting.
Fixes: 737eb78e82 ("block: Delay default elevator initialization")
Reported-by: yangerkun <yangerkun@huawei.com>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20211117115502.1600950-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Except for the features passed to blk_queue_required_elevator_features,
elevator.h is only needed internally to the block layer. Move the
ELEVATOR_F_* definitions to blkdev.h, and the move elevator.h to
block/, dropping all the spurious includes outside of that.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Link: https://lore.kernel.org/r/20210920123328.1399408-13-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When merging one bio to request, if they are discard IO and the queue
supports multi-range discard, we need to return ELEVATOR_DISCARD_MERGE
because both block core and related drivers(nvme, virtio-blk) doesn't
handle mixed discard io merge(traditional IO merge together with
discard merge) well.
Fix the issue by returning ELEVATOR_DISCARD_MERGE in this situation,
so both blk-mq and drivers just need to handle multi-range discard.
Reported-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Fixes: 2705dfb209 ("block: fix discard request merge")
Link: https://lore.kernel.org/r/20210729034226.1591070-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
elevator_get_default() uses the following algorithm to select an I/O
scheduler from inside add_disk():
- In case of a single hardware queue or if sharing hardware queues across
multiple request queues (BLK_MQ_F_TAG_HCTX_SHARED), use mq-deadline.
- Otherwise, use 'none'.
This is a good choice for most but not for all block drivers. Make it
possible to override the selection of mq-deadline with a new flag,
namely BLK_MQ_F_NO_SCHED_BY_DEFAULT.
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Martijn Coenen <maco@android.com>
Cc: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20210805174200.3250718-2-bvanassche@acm.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Lockdep complains about lock inversion between ioc->lock and bfqd->lock:
bfqd -> ioc:
put_io_context+0x33/0x90 -> ioc->lock grabbed
blk_mq_free_request+0x51/0x140
blk_put_request+0xe/0x10
blk_attempt_req_merge+0x1d/0x30
elv_attempt_insert_merge+0x56/0xa0
blk_mq_sched_try_insert_merge+0x4b/0x60
bfq_insert_requests+0x9e/0x18c0 -> bfqd->lock grabbed
blk_mq_sched_insert_requests+0xd6/0x2b0
blk_mq_flush_plug_list+0x154/0x280
blk_finish_plug+0x40/0x60
ext4_writepages+0x696/0x1320
do_writepages+0x1c/0x80
__filemap_fdatawrite_range+0xd7/0x120
sync_file_range+0xac/0xf0
ioc->bfqd:
bfq_exit_icq+0xa3/0xe0 -> bfqd->lock grabbed
put_io_context_active+0x78/0xb0 -> ioc->lock grabbed
exit_io_context+0x48/0x50
do_exit+0x7e9/0xdd0
do_group_exit+0x54/0xc0
To avoid this inversion we change blk_mq_sched_try_insert_merge() to not
free the merged request but rather leave that upto the caller similarly
to blk_mq_sched_try_merge(). And in bfq_insert_requests() we make sure
to free all the merged requests after dropping bfqd->lock.
Fixes: aee69d78de ("block, bfq: introduce the BFQ-v0 I/O scheduler as an extra scheduler")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20210623093634.27879-3-jack@suse.cz
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The insert_requests and dispatch_request elevator operations are
mandatory for the correct execution of an elevator, and all implemented
elevators (bfq, kyber and mq-deadline) implement them. As a result,
there is no need to check for these operations before calling them when
a queue has an elevator set. This simplifies the code in
__blk_mq_sched_dispatch_requests() and blk_mq_sched_insert_request().
To avoid out-of-tree elevators to crash the kernel in case of bad
implementation, add a check in elv_register() to verify that these
operations are implemented.
A small, probably not significant, IOPS improvement of 0.1% is observed
with this patch applied (4.117 MIOPS to 4.123 MIOPS, average of 20 fio
runs doing 4K random direct reads with psync and 32 jobs).
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Link: https://lore.kernel.org/r/20210618015922.713999-1-damien.lemoal@wdc.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Yanhui found that write performance is degraded a lot after applying
hctx shared tagset on one test machine with megaraid_sas. And turns out
it is caused by none scheduler which becomes default elevator caused by
hctx shared tagset patchset.
Given more scsi HBAs will apply hctx shared tagset, and the similar
performance exists for them too.
So keep previous behavior by still using default mq-deadline for queues
which apply hctx shared tagset, just like before.
Fixes: 32bc15afed ("blk-mq: Facilitate a shared sbitmap per tagset")
Reported-by: Yanhui Ma <yama@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: John Garry <john.garry@huawei.com>
Link: https://lore.kernel.org/r/20210406031933.767228-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After commit b89f625e28 ("block: don't release queue's sysfs
lock during switching elevator"), whole elevator register and
unregister function are covered by sysfs_lock. So, remove wrong
comment and add lockdep assert.
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We have defined common interface blk_queue_registered() to
test QUEUE_FLAG_REGISTERED. Just use it.
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
elv_support_iosched() will check queue_is_mq() for us. So, remove
the redundant check to clean code.
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We will register debugfs for scheduler no matter whether it have
defined callback funciton .exit_sched. So, blk_mq_exit_sched()
is always needed to unregister debugfs. Also, q->elevator should
be set as NULL after exiting scheduler.
For now, since all register scheduler have defined .exit_sched,
it will not cause any actual problem. But It will be more reasonable
to do this change.
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull on for-linus to resolve what otherwise would have been a conflict
with the cgroups rstat patchset from Tejun.
* for-linus: (942 commits)
blkcg: make blkcg_print_stat() print stats only for online blkgs
nvme: change nvme_passthru_cmd64 to explicitly mark rsvd
nvme-multipath: fix crash in nvme_mpath_clear_ctrl_paths
nvme-rdma: fix a segmentation fault during module unload
iocost: don't nest spin_lock_irq in ioc_weight_write()
io_uring: ensure we clear io_kiocb->result before each issue
um-ubd: Entrust re-queue to the upper layers
nvme-multipath: remove unused groups_only mode in ana log
nvme-multipath: fix possible io hang after ctrl reconnect
io_uring: don't touch ctx in setup after ring fd install
io_uring: Fix leaked shadow_req
Linux 5.4-rc5
riscv: cleanup do_trap_break
nbd: verify socket is supported during setup
ata: libahci_platform: Fix regulator_get_optional() misuse
nbd: handle racing with error'ed out commands
nbd: protect cmd->status with cmd->lock
io_uring: fix bad inflight accounting for SETUP_IOPOLL|SETUP_SQTHREAD
io_uring: used cached copies of sq->dropped and cq->overflow
ARM: dts: stm32: relax qspi pins slew-rate for stm32mp157
...
With transition to blk-mq, the elevator= kernel argument was removed as
it makes less and less sense with the current variety of devices. Since
this may surprise some users and there are advices on the Internet that
still suggest to use it, let's at least warn if the parameter is used.
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A BIO based request queue does not have a tag_set, which prevent testing
for the flag BLK_MQ_F_NO_SCHED indicating that the queue does not
require an elevator. This leads to an incorrect initialization of a
default elevator in some cases such as BIO based null_blk
(queue_mode == BIO) with zoned mode enabled as the default elevator in
this case is mq-deadline instead of "none".
Fix this by testing for a NULL queue mq_ops field which indicates that
the queue is BIO based and should not have an elevator.
Reported-by: Shinichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Reviewed-by: Bob Liu <bob.liu@oracle.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
cecf5d87ff ("block: split .sysfs_lock into two locks") starts to
release & acquire sysfs_lock before registering/un-registering elevator
queue during switching elevator for avoiding potential deadlock from
showing & storing 'queue/iosched' attributes and removing elevator's
kobject.
Turns out there isn't such deadlock because 'q->sysfs_lock' isn't
required in .show & .store of queue/iosched's attributes, and just
elevator's sysfs lock is acquired in elv_iosched_store() and
elv_iosched_show(). So it is safe to hold queue's sysfs lock when
registering/un-registering elevator queue.
The biggest issue is that commit cecf5d87ff assumes that concurrent
write on 'queue/scheduler' can't happen. However, this assumption isn't
true, because kernfs_fop_write() only guarantees that concurrent write
aren't called on the same open file, but the write could be from
different open on the file. So we can't release & re-acquire queue's
sysfs lock during switching elevator, otherwise use-after-free on
elevator could be triggered.
Fixes the issue by not releasing queue's sysfs lock during switching
elevator.
Fixes: cecf5d87ff ("block: split .sysfs_lock into two locks")
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Greg KH <gregkh@linuxfoundation.org>
Cc: Mike Snitzer <snitzer@redhat.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The lookup logic is broken - 'e' will never be NULL, even if the
list is empty. Maintain lookup hit in a separate variable instead.
Fixes: a0958ba7fc ("block: Improve default elevator selection")
Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When elevator_init_mq() is called from blk_mq_init_allocated_queue(),
the only information known about the device is the number of hardware
queues as the block device scan by the device driver is not completed
yet for most drivers. The device type and elevator required features
are not set yet, preventing to correctly select the default elevator
most suitable for the device.
This currently affects all multi-queue zoned block devices which default
to the "none" elevator instead of the required "mq-deadline" elevator.
These drives currently include host-managed SMR disks connected to a
smartpqi HBA and null_blk block devices with zoned mode enabled.
Upcoming NVMe Zoned Namespace devices will also be affected.
Fix this by adding the boolean elevator_init argument to
blk_mq_init_allocated_queue() to control the execution of
elevator_init_mq(). Two cases exist:
1) elevator_init = false is used for calls to
blk_mq_init_allocated_queue() within blk_mq_init_queue(). In this
case, a call to elevator_init_mq() is added to __device_add_disk(),
resulting in the delayed initialization of the queue elevator
after the device driver finished probing the device information. This
effectively allows elevator_init_mq() access to more information
about the device.
2) elevator_init = true preserves the current behavior of initializing
the elevator directly from blk_mq_init_allocated_queue(). This case
is used for the special request based DM devices where the device
gendisk is created before the queue initialization and device
information (e.g. queue limits) is already known when the queue
initialization is executed.
Additionally, to make sure that the elevator initialization is never
done while requests are in-flight (there should be none when the device
driver calls device_add_disk()), freeze and quiesce the device request
queue before calling blk_mq_init_sched() in elevator_init_mq().
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For block devices that do not specify required features, preserve the
current default elevator selection (mq-deadline for single queue
devices, none for multi-queue devices). However, for devices specifying
required features (e.g. zoned block devices ELEVATOR_F_ZBD_SEQ_WRITE
feature), select the first available elevator providing the required
features.
In all cases, default to "none" if no elevator is available or if the
initialization of the default elevator fails.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Introduce the definition of elevator features through the
elevator_features flags in the elevator_type structure. Each flag can
represent a feature supported by an elevator. The first feature defined
by this patch is support for zoned block device sequential write
constraint with the flag ELEVATOR_F_ZBD_SEQ_WRITE, which is implemented
by the mq-deadline elevator using zone write locking.
Other possible features are IO priorities, write hints, latency targets
or single-LUN dual-actuator disks (for which the elevator could maintain
one LBA ordered list per actuator).
The required_elevator_features field is also added to the request_queue
structure to allow a device driver to specify elevator feature flags
that an elevator must support for the correct operation of the device
(e.g. device drivers for zoned block devices can have the
ELEVATOR_F_ZBD_SEQ_WRITE flag as a required feature).
The helper function blk_queue_required_elevator_features() is
defined for setting this new field.
With these two new fields in place, the elevator functions
elevator_match() and elevator_find() are modified to allow a user to set
only an elevator with a set of features that satisfies the device
required features. Elevators not matching the device requirements are
not shown in the device sysfs queue/scheduler file to prevent their use.
The "none" elevator can always be selected as before.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If the default elevator chosen is mq-deadline, elevator_init_mq() may
return an error if mq-deadline initialization fails, leading to
blk_mq_init_allocated_queue() returning an error, which in turn will
cause the block device initialization to fail and the device not being
exposed.
Instead of taking such extreme measure, handle mq-deadline
initialization failures in the same manner as when mq-deadline is not
available (no module to load), that is, default to the "none" scheduler.
With this change, elevator_init_mq() return type can be changed to void.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of checking a queue tag_set BLK_MQ_F_NO_SCHED flag before
calling elevator_init_mq() to make sure that the queue supports IO
scheduling, use the elevator.c function elv_support_iosched() in
elevator_init_mq(). This does not introduce any functional change but
ensure that elevator_init_mq() does the right thing based on the queue
settings.
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
