Many throughput-sensitive workloads are made of several parallel I/O
flows, with all flows generated by the same application, or more
generically by the same task (e.g., system boot). The most
counterproductive action with these workloads is plugging I/O dispatch
when one of the bfq_queues associated with these flows remains
temporarily empty.
To avoid this plugging, BFQ has been using a burst-handling mechanism
for years now. This mechanism has proven effective for throughput, and
not detrimental for service guarantees. This commit pushes this
mechanism a little bit further, basing on the following two facts.
First, all the I/O flows of a the same application or task contribute
to the execution/completion of that common application or task. So the
performance figures that matter are total throughput of the flows and
task-wide I/O latency. In particular, these flows do not need to be
protected from each other, in terms of individual bandwidth or
latency.
Second, the above fact holds regardless of the number of flows.
Putting these two facts together, this commits merges stably the
bfq_queues associated with these I/O flows, i.e., with the processes
that generate these IO/ flows, regardless of how many the involved
processes are.
To decide whether a set of bfq_queues is actually associated with the
I/O flows of a common application or task, and to merge these queues
stably, this commit operates as follows: given a bfq_queue, say Q2,
currently being created, and the last bfq_queue, say Q1, created
before Q2, Q2 is merged stably with Q1 if
- very little time has elapsed since when Q1 was created
- Q2 has the same ioprio as Q1
- Q2 belongs to the same group as Q1
Merging bfq_queues also reduces scheduling overhead. A fio test with
ten random readers on /dev/nullb shows a throughput boost of 40%, with
a quadcore. Since BFQ's execution time amounts to ~50% of the total
per-request processing time, the above throughput boost implies that
BFQ's overhead is reduced by more than 50%.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-7-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Shared queues are likely to receive I/O at a high rate. This may
deceptively let them be considered as wakers of other queues. But a
false waker will unjustly steal bandwidth to its supposedly woken
queue. So considering also shared queues in the waking mechanism may
cause more control troubles than throughput benefits. This commit
keeps shared queues out of the waker-detection mechanism.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-6-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When the io_latency heuristic is off, bfq_queues must not start to be
weight-raised. Unfortunately, by mistake, this may happen when the
state of a previously weight-raised bfq_queue is resumed after a queue
split. This commit fixes this error.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-5-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Consider a bfq_queue bfqq that is about to be merged with another
bfq_queue new_bfqq. The processes associated with bfqq are cooperators
of the processes associated with new_bfqq. So, if bfqq has a waker,
then it is reasonable (and beneficial for throughput) to assume that
all these processes will be happy to let bfqq's waker freely inject
I/O when they have no I/O. So this commit makes new_bfqq inherit
bfqq's waker.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-4-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Consider a new I/O request that arrives for a bfq_queue bfqq. If, when
this happens, the only active bfq_queues are bfqq and either its waker
bfq_queue or one of its woken bfq_queues, then there is no point in
queueing this new I/O request in bfqq for service. In fact, the
in-service queue and bfqq agree on serving this new I/O request as
soon as possible. So this commit puts this new I/O request directly
into the dispatch list.
Tested-by: Jan Kara <jack@suse.cz>
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-3-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Suppose that I/O dispatch is plugged, to wait for new I/O for the
in-service bfq-queue, say bfqq. Suppose then that there is a further
bfq_queue woken by bfqq, and that this woken queue has pending I/O. A
woken queue does not steal bandwidth from bfqq, because it remains
soon without I/O if bfqq is not served. So there is virtually no risk
of loss of bandwidth for bfqq if this woken queue has I/O dispatched
while bfqq is waiting for new I/O. In contrast, this extra I/O
injection boosts throughput. This commit performs this extra
injection.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20210304174627.161-2-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Christoph reported that we'll likely trigger the WARN_ON_ONCE() checking
that we're not submitting a bvec with REQ_OP_ZONE_APPEND in
bio_iov_iter_get_pages() some time ago using zoned btrfs, but I couldn't
reproduce it back then.
Now Naohiro was able to trigger the bug as well with xfstests generic/095
on a zoned btrfs.
There is nothing that prevents bvec submissions via REQ_OP_ZONE_APPEND if
the hardware's zone append limit is met.
Reported-by: Naohiro Aota <naohiro.aota@wdc.com>
Reported-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/10bd414d9326c90cd69029077db63b363854eee5.1616600835.git.johannes.thumshirn@wdc.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When a stacked block device inserts a request into another block device
using blk_insert_cloned_request, the request's nr_phys_segments field gets
recalculated by a call to blk_recalc_rq_segments in
blk_cloned_rq_check_limits. But blk_recalc_rq_segments does not know how to
handle multi-segment discards. For disk types which can handle
multi-segment discards like nvme, this results in discard requests which
claim a single segment when it should report several, triggering a warning
in nvme and causing nvme to fail the discard from the invalid state.
WARNING: CPU: 5 PID: 191 at drivers/nvme/host/core.c:700 nvme_setup_discard+0x170/0x1e0 [nvme_core]
...
nvme_setup_cmd+0x217/0x270 [nvme_core]
nvme_loop_queue_rq+0x51/0x1b0 [nvme_loop]
__blk_mq_try_issue_directly+0xe7/0x1b0
blk_mq_request_issue_directly+0x41/0x70
? blk_account_io_start+0x40/0x50
dm_mq_queue_rq+0x200/0x3e0
blk_mq_dispatch_rq_list+0x10a/0x7d0
? __sbitmap_queue_get+0x25/0x90
? elv_rb_del+0x1f/0x30
? deadline_remove_request+0x55/0xb0
? dd_dispatch_request+0x181/0x210
__blk_mq_do_dispatch_sched+0x144/0x290
? bio_attempt_discard_merge+0x134/0x1f0
__blk_mq_sched_dispatch_requests+0x129/0x180
blk_mq_sched_dispatch_requests+0x30/0x60
__blk_mq_run_hw_queue+0x47/0xe0
__blk_mq_delay_run_hw_queue+0x15b/0x170
blk_mq_sched_insert_requests+0x68/0xe0
blk_mq_flush_plug_list+0xf0/0x170
blk_finish_plug+0x36/0x50
xlog_cil_committed+0x19f/0x290 [xfs]
xlog_cil_process_committed+0x57/0x80 [xfs]
xlog_state_do_callback+0x1e0/0x2a0 [xfs]
xlog_ioend_work+0x2f/0x80 [xfs]
process_one_work+0x1b6/0x350
worker_thread+0x53/0x3e0
? process_one_work+0x350/0x350
kthread+0x11b/0x140
? __kthread_bind_mask+0x60/0x60
ret_from_fork+0x22/0x30
This patch fixes blk_recalc_rq_segments to be aware of devices which can
have multi-segment discards. It calculates the correct discard segment
count by counting the number of bio as each discard bio is considered its
own segment.
Fixes: 1e739730c5 ("block: optionally merge discontiguous discard bios into a single request")
Signed-off-by: David Jeffery <djeffery@redhat.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Laurence Oberman <loberman@redhat.com>
Link: https://lore.kernel.org/r/20210211143807.GA115624@redhat
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When zone reset ioctl and data read race for a same zone on zoned block
devices, the data read leaves stale page cache even though the zone
reset ioctl zero clears all the zone data on the device. To avoid
non-zero data read from the stale page cache after zone reset, discard
page cache of reset target zones in blkdev_zone_mgmt_ioctl(). Introduce
the helper function blkdev_truncate_zone_range() to discard the page
cache. Ensure the page cache discarded by calling the helper function
before and after zone reset in same manner as fallocate does.
This patch can be applied back to the stable kernel version v5.10.y.
Rework is needed for older stable kernels.
Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Fixes: 3ed05a987e ("blk-zoned: implement ioctls")
Cc: <stable@vger.kernel.org> # 5.10+
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Link: https://lore.kernel.org/r/20210311072546.678999-1-shinichiro.kawasaki@wdc.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
register_disk() suppress uevents for devices with the GENHD_FL_HIDDEN
but enables uevents at the end again in order to announce disk after
possible partitions are created.
When the device is removed the uevents are still on and user land sees
'remove' messages for devices which were never 'add'ed to the system.
KERNEL[95481.571887] remove /devices/virtual/nvme-fabrics/ctl/nvme5/nvme0c5n1 (block)
Let's suppress the uevents for GENHD_FL_HIDDEN by not enabling the
uevents at all.
Signed-off-by: Daniel Wagner <dwagner@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin Wilck <mwilck@suse.com>
Link: https://lore.kernel.org/r/20210311151917.136091-1-dwagner@suse.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ever since the addition of multipage bio_vecs BIO_MAX_PAGES has been
horribly confusingly misnamed. Rename it to BIO_MAX_VECS to stop
confusing users of the bio API.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20210311110137.1132391-2-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Similarly to a single zone reset operation (REQ_OP_ZONE_RESET), execute
REQ_OP_ZONE_RESET_ALL operations with REQ_SYNC set.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The current blkio.throttle.io_service_bytes_recursive doesn't
work correctly.
As an example, for the following blkcg hierarchy:
(Made 1GB READ in test1, 512MB READ in test2)
test
/ \
test1 test2
$ head -n 1 test/test1/blkio.throttle.io_service_bytes_recursive
8:0 Read 1073684480
$ head -n 1 test/test2/blkio.throttle.io_service_bytes_recursive
8:0 Read 537448448
$ head -n 1 test/blkio.throttle.io_service_bytes_recursive
8:0 Read 537448448
Clearly, above data of "test" reflects "test2" not "test1"+"test2".
Do the correct summary in blkg_rwstat_recursive_sum().
Signed-off-by: Xunlei Pang <xlpang@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
SCSI uses a global atomic variable to track queue depth for each
LUN/request queue.
This doesn't scale well when there are lots of CPU cores and the disk is
very fast. It has been observed that IOPS is affected a lot by tracking
queue depth via sdev->device_busy in the I/O path.
Return budget token from .get_budget callback. The budget token can be
passed to driver so that we can replace the atomic variable with
sbitmap_queue and alleviate the scaling problems that way.
Link: https://lore.kernel.org/r/20210122023317.687987-9-ming.lei@redhat.com
Cc: Omar Sandoval <osandov@fb.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Cc: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Tested-by: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Allocation hint should have belonged to sbitmap. Also, when sbitmap's depth
is high and there is no need to use mulitple wakeup queues, user can
benefit from percpu allocation hint too.
Move allocation hint into sbitmap, then SCSI device queue can benefit from
allocation hint when converting to plain sbitmap.
Convert vhost/scsi.c to use sbitmap allocation with percpu alloc hint. This
is more efficient than the previous approach.
Link: https://lore.kernel.org/r/20210122023317.687987-5-ming.lei@redhat.com
Cc: Omar Sandoval <osandov@fb.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Cc: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Cc: Mike Christie <michael.christie@oracle.com>
Cc: virtualization@lists.linux-foundation.org
Tested-by: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently the allocation round_robin info is maintained by sbitmap_queue.
However, bit allocation really belongs to sbitmap. Move it there.
Link: https://lore.kernel.org/r/20210122023317.687987-3-ming.lei@redhat.com
Cc: Omar Sandoval <osandov@fb.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Cc: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: virtualization@lists.linux-foundation.org
Tested-by: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Correct the comments since bfq_fifo_expire[0] is for async request,
while bfq_fifo_expire[1] is for sync request.
Also update docs, according the source code, the default
fifo_expire_async is 250ms, and fifo_expire_sync is 125ms.
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit a1ce35fa49 ("block: remove dead
elevator code") removed all users of RQF_SORTED. However it is still
defined, and there is one reference left to it (which in effect is
dead code). Clear it all up.
Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
With the removal of the skd driver, using IRQ safe locking of a bdev
bd_size_lock spinlock to protect the bdev inode size is not necessary
anymore as there is no other known driver using this lock under an IRQ
disabled context (e.g. calling set_capacity() with IRQ disabled).
Revert commit 0fe37724f8 ("block: fix bd_size_lock use") which
introduced the IRQ safe change.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=ccMs
-----END PGP SIGNATURE-----
Merge tag 'block-5.12-2021-02-27' of git://git.kernel.dk/linux-block
Pull more block updates from Jens Axboe:
"A few stragglers (and one due to me missing it originally), and fixes
for changes in this merge window mostly. In particular:
- blktrace cleanups (Chaitanya, Greg)
- Kill dead blk_pm_* functions (Bart)
- Fixes for the bio alloc changes (Christoph)
- Fix for the partition changes (Christoph, Ming)
- Fix for turning off iopoll with polled IO inflight (Jeffle)
- nbd disconnect fix (Josef)
- loop fsync error fix (Mauricio)
- kyber update depth fix (Yang)
- max_sectors alignment fix (Mikulas)
- Add bio_max_segs helper (Matthew)"
* tag 'block-5.12-2021-02-27' of git://git.kernel.dk/linux-block: (21 commits)
block: Add bio_max_segs
blktrace: fix documentation for blk_fill_rw()
block: memory allocations in bounce_clone_bio must not fail
block: remove the gfp_mask argument to bounce_clone_bio
block: fix bounce_clone_bio for passthrough bios
block-crypto-fallback: use a bio_set for splitting bios
block: fix logging on capacity change
blk-settings: align max_sectors on "logical_block_size" boundary
block: reopen the device in blkdev_reread_part
block: don't skip empty device in in disk_uevent
blktrace: remove debugfs file dentries from struct blk_trace
nbd: handle device refs for DESTROY_ON_DISCONNECT properly
kyber: introduce kyber_depth_updated()
loop: fix I/O error on fsync() in detached loop devices
block: fix potential IO hang when turning off io_poll
block: get rid of the trace rq insert wrapper
blktrace: fix blk_rq_merge documentation
blktrace: fix blk_rq_issue documentation
blktrace: add blk_fill_rwbs documentation comment
block: remove superfluous param in blk_fill_rwbs()
...
It's often inconvenient to use BIO_MAX_PAGES due to min() requiring the
sign to be the same. Introduce bio_max_segs() and change BIO_MAX_PAGES to
be unsigned to make it easier for the users.
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The caller can't cope with a failure from bounce_clone_bio, so
use __GFP_NOFAIL for the passthrough case. bio_alloc_bioset already
won't fail due to the use of mempools.
And yes, we need to get rid of this bock layer bouncing code entirely
sooner or later..
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that bio_alloc_bioset does not fall back to kmalloc for a NULL
bio_set, handle that case explicitly and simplify the calling
conventions.
Based on an earlier patch from Chaitanya Kulkarni.
Fixes: 3175199ab0 ("block: split bio_kmalloc from bio_alloc_bioset")
Reported-by: Chaitanya Kulkarni <Chaitanya.Kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio_split with a NULL bs argumen used to fall back to kmalloc the
bio, which does not guarantee forward progress and could to deadlocks.
Now that the overloading of the NULL bs argument to bio_alloc_bioset
has been removed it crashes instead. Fix all that by using a special
crafted bioset.
Fixes: 3175199ab0 ("block: split bio_kmalloc from bio_alloc_bioset")
Reported-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Local variable of 'capacity' stores the previous disk capacity, and
'size' variable records the latest disk capacity, so swap them for
fixing logging on capacity change.
Cc: Christoph Hellwig <hch@lst.de>
Fixes: a782483cc1 ("block: remove the nr_sects field in struct hd_struct")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We get I/O errors when we run md-raid1 on the top of dm-integrity on the
top of ramdisk.
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8048, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8147, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8246, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8345, 0xbb
The ramdisk device has logical_block_size 512 and max_sectors 255. The
dm-integrity device uses logical_block_size 4096 and it doesn't affect the
"max_sectors" value - thus, it inherits 255 from the ramdisk. So, we have
a device with max_sectors not aligned on logical_block_size.
The md-raid device sees that the underlying leg has max_sectors 255 and it
will split the bios on 255-sector boundary, making the bios unaligned on
logical_block_size.
In order to fix the bug, we round down max_sectors to logical_block_size.
Cc: stable@vger.kernel.org
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Historically the BLKRRPART ioctls called into the now defunct ->revalidate
method, which caused the sd driver to check if any media is present.
When the ->revalidate method was removed this revalidation was lost,
leading to lots of I/O errors when using the eject command. Fix this by
reopening the device to rescan the partitions, and thus calling the
revalidation logic in the sd driver.
Fixes: 471bd0af54 ("sd: use bdev_check_media_change")
Reported--by: Tom Seewald <tseewald@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Tom Seewald <tseewald@gmail.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Restore the previous behavior by using the correct flag for the whole device
("part0").
Fixes: 99dfc43ecb ("block: use ->bi_bdev for bio based I/O accounting")
Reported-by: John Stultz <john.stultz@linaro.org>
Tested-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Hang occurs when user changes the scheduler queue depth, by writing to
the 'nr_requests' sysfs file of that device.
The details of the environment that we found the problem are as follows:
an eMMC block device
total driver tags: 16
default queue_depth: 32
kqd->async_depth initialized in kyber_init_sched() with queue_depth=32
Then we change queue_depth to 256, by writing to the 'nr_requests' sysfs
file. But kqd->async_depth don't be updated after queue_depth changes.
Now the value of async depth is too small for queue_depth=256, this may
cause hang.
This patch introduces kyber_depth_updated(), so that kyber can update
async depth when queue depth changes.
Signed-off-by: Yang Yang <yang.yang@vivo.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
internal_hash and journal_mac capabilities.
- Various DM writecache fixes to address performance, fix table output
to match what was provided at table creation, fix writing beyond end
of device when shrinking underlying data device, and a couple other
small cleanups.
- Add DM crypt support for using trusted keys.
- Fix deadlock when swapping to DM crypt device by throttling number
of in-flight REQ_SWAP bios. Implemented in DM core so that other
bio-based targets can opt-in by setting ti->limit_swap_bios.
- Fix various inverted logic bugs in the .iterate_devices callout
functions that are used to assess if specific feature or capability
is supported across all devices being combined/stacked by DM.
- Fix DM era target bugs that exposed users to lost writes or memory
leaks.
- Add DM core support for passing through inline crypto support of
underlying devices. Includes block/keyslot-manager changes that
enable extending this support to DM.
- Various small fixes and cleanups (spelling fixes, front padding
calculation cleanup, cleanup conditional zoned support in targets,
etc).
-----BEGIN PGP SIGNATURE-----
iQFHBAABCAAxFiEEJfWUX4UqZ4x1O2wixSPxCi2dA1oFAmAqxggTHHNuaXR6ZXJA
cmVkaGF0LmNvbQAKCRDFI/EKLZ0DWjVOCACkZKleQhsCEYHNtjZ40Du+4PPBvESA
O+ScdUCeik4YUXvQtlFRPcYxxOH0zL0CUivLnNlsKzGTTgulw5azgFNuUTzIhH5y
a86Q+DReigPegzVCCOenInU18pYa03rLtYOAb6SK49IqVeMWMFSJVBv73HWS7OFV
slMlsQCN46YgbviYsGUXk5+uKMET4ijJZVW+8zSYg0GsWLHdgQtBkEoojO1n9H2B
jio2Nvhto0bJ4dV482lmd3G+LABmaBbLs0Xx/a7iHVigkIYZz4BHwDYNz/EQnNEi
dYlOrSL9a6ur+DFR6vxShzG40LbK7KVr8jHiXyKv2WZA7FMK0l4fyEFV
=E+n3
-----END PGP SIGNATURE-----
Merge tag 'for-5.12/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm
Pull device mapper updates from Mike Snitzer:
- Fix DM integrity's HMAC support to provide enhanced security of
internal_hash and journal_mac capabilities.
- Various DM writecache fixes to address performance, fix table output
to match what was provided at table creation, fix writing beyond end
of device when shrinking underlying data device, and a couple other
small cleanups.
- Add DM crypt support for using trusted keys.
- Fix deadlock when swapping to DM crypt device by throttling number of
in-flight REQ_SWAP bios. Implemented in DM core so that other
bio-based targets can opt-in by setting ti->limit_swap_bios.
- Fix various inverted logic bugs in the .iterate_devices callout
functions that are used to assess if specific feature or capability
is supported across all devices being combined/stacked by DM.
- Fix DM era target bugs that exposed users to lost writes or memory
leaks.
- Add DM core support for passing through inline crypto support of
underlying devices. Includes block/keyslot-manager changes that
enable extending this support to DM.
- Various small fixes and cleanups (spelling fixes, front padding
calculation cleanup, cleanup conditional zoned support in targets,
etc).
* tag 'for-5.12/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (31 commits)
dm: fix deadlock when swapping to encrypted device
dm: simplify target code conditional on CONFIG_BLK_DEV_ZONED
dm: set DM_TARGET_PASSES_CRYPTO feature for some targets
dm: support key eviction from keyslot managers of underlying devices
dm: add support for passing through inline crypto support
block/keyslot-manager: Introduce functions for device mapper support
block/keyslot-manager: Introduce passthrough keyslot manager
dm era: only resize metadata in preresume
dm era: Use correct value size in equality function of writeset tree
dm era: Fix bitset memory leaks
dm era: Verify the data block size hasn't changed
dm era: Reinitialize bitset cache before digesting a new writeset
dm era: Update in-core bitset after committing the metadata
dm era: Recover committed writeset after crash
dm writecache: use bdev_nr_sectors() instead of open-coded equivalent
dm writecache: fix writing beyond end of underlying device when shrinking
dm table: remove needless request_queue NULL pointer checks
dm table: fix zoned iterate_devices based device capability checks
dm table: fix DAX iterate_devices based device capability checks
dm table: fix iterate_devices based device capability checks
...
- Add support for eMMC inline encryption
- Add a helper function to parse DT properties for clock phases
- Some improvements and cleanups for the mmc_test module
MMC host:
- android-goldfish: Remove driver
- cqhci: Add support for eMMC inline encryption
- dw_mmc-zx: Remove driver
- meson-gx: Extend support for scatter-gather to allow SD_IO_RW_EXTENDED
- mmci: Add support for probing bus voltage level translator
- mtk-sd: Address race condition for request timeouts
- sdhci_am654: Add Support for the variant on TI's AM64 SoC
- sdhci-esdhc-imx: Prevent kernel panic at ->remove()
- sdhci-iproc: Add ACPI bindings for the RPi to enable SD and WiFi on RPi4
- sdhci-msm: Add Inline Crypto Engine support
- sdhci-msm: Use actual_clock to improve timeout calculations
- sdhci-of-aspeed: Add Andrew Jeffery as maintainer
- sdhci-of-aspeed: Extend clock support for the AST2600 variant
- sdhci-pci-gli: Increase idle period for low power state for GL9763E
- sdhci-pci-o2micro: Make tuning for SDR104 HW more robust
- sdhci-sirf: Remove driver
- sdhci-xenon: Add support for the AP807 variant
- sunxi-mmc: Add support for the A100 variant
- sunxi-mmc: Ensure host is suspended during system sleep
- tmio: Add detection of data timeout errors
- tmio/renesas_sdhi: Extend support for retuning
- renesas_sdhi_internal_dmac: Add support for the ->pre|post_req() ops
-----BEGIN PGP SIGNATURE-----
iQJLBAABCgA1FiEEugLDXPmKSktSkQsV/iaEJXNYjCkFAmAqZ+YXHHVsZi5oYW5z
c29uQGxpbmFyby5vcmcACgkQ/iaEJXNYjCmB8RAAwIwsLZXLQhNfzFGZvs+gbihd
hgHnPXCHxLge2VDph1KKCCkxnjKTKHfB3McZOrlNZI0wiCBMSI+ZuoxIT0UlWqsy
IyZ1s3u1YT30tPpyZ8UqYDzft/9/vX1TiZlLqsYnN05ykIe7siaeBb/w/8py0ip3
rVptRn49V2TRHiu/J8FvVF1diPMKNn1S063Xyrxu4yWnQX040xHC+tyDCl9xclWR
oUDl1eqJmZRomijPu+AKte3uDppcr4ejDxfvzPrx4bpJNS7ZgJ5TG5MtDm+j+0Vu
aJT0bPcfn/jvHahCcpcKkYcuesKIw2CFbglv9aIxbvOfEUkTSL4zO+VCvKD9r+wk
WSXrPZ27ukTJmZIA6JqdUfs3/4oi5/80uA2kQkhfyYmlA7sLJcdRmBzSgltpJWp5
bmno/grpEXUgN59F5xe3/gINQNgAt319vmOPQg2LFF/uiOWfRytunNgXCCYMJQX8
1U9q6RHQWfdcasCOhAA9U9NxM1zecuIYb/2ecDhmovSmpElxdUFuN+TW1Om/xKHh
o0xxu+/654dcehyHdW8/3kq9Oz9fhBoorC3F/OUm0k0DBL50G+476hl1rbhTMVl+
SlPIUvDxCu2GRwAuprQ9vu+jUGUSvC8mfxswQkrcLak/iPeOyYNrLDFN9nA80Kve
G40UQwsDC3u7X1Z7rkY=
=rwB7
-----END PGP SIGNATURE-----
Merge tag 'mmc-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Pull MMC updates from Ulf Hansson:
"MMC core:
- Add support for eMMC inline encryption
- Add a helper function to parse DT properties for clock phases
- Some improvements and cleanups for the mmc_test module
MMC host:
- android-goldfish: Remove driver
- cqhci: Add support for eMMC inline encryption
- dw_mmc-zx: Remove driver
- meson-gx: Extend support for scatter-gather to allow SD_IO_RW_EXTENDED
- mmci: Add support for probing bus voltage level translator
- mtk-sd: Address race condition for request timeouts
- sdhci_am654: Add Support for the variant on TI's AM64 SoC
- sdhci-esdhc-imx: Prevent kernel panic at ->remove()
- sdhci-iproc: Add ACPI bindings for the RPi to enable SD and WiFi on RPi4
- sdhci-msm: Add Inline Crypto Engine support
- sdhci-msm: Use actual_clock to improve timeout calculations
- sdhci-of-aspeed: Add Andrew Jeffery as maintainer
- sdhci-of-aspeed: Extend clock support for the AST2600 variant
- sdhci-pci-gli: Increase idle period for low power state for GL9763E
- sdhci-pci-o2micro: Make tuning for SDR104 HW more robust
- sdhci-sirf: Remove driver
- sdhci-xenon: Add support for the AP807 variant
- sunxi-mmc: Add support for the A100 variant
- sunxi-mmc: Ensure host is suspended during system sleep
- tmio: Add detection of data timeout errors
- tmio/renesas_sdhi: Extend support for retuning
- renesas_sdhi_internal_dmac: Add support for the ->pre|post_req() ops"
* tag 'mmc-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc: (86 commits)
mmc: sdhci-esdhc-imx: fix kernel panic when remove module
mmc: host: Retire MMC_GOLDFISH
mmc: cb710: Use new tasklet API
mmc: sdhci-pci-o2micro: Bug fix for SDR104 HW tuning failure
mmc: mmc_test: use erase_arg for mmc_erase command
mmc: wbsd: Use new tasklet API
mmc: via-sdmmc: Use new tasklet API
mmc: uniphier-sd: Use new tasklet API
mmc: tifm_sd: Use new tasklet API
mmc: s3cmci: Use new tasklet API
mmc: omap: Use new tasklet API
mmc: dw_mmc: Use new tasklet API
mmc: au1xmmc: Use new tasklet API
mmc: atmel-mci: Use new tasklet API
mmc: cavium: Replace spin_lock_irqsave with spin_lock in hard IRQ
mmc: queue: Remove unused define
mmc: core: Drop redundant bouncesz from struct mmc_card
mmc: core: Drop redundant member in struct mmc host
mmc: core: Use host instead of card argument to mmc_spi_send_csd()
mmc: core: Exclude unnecessary header file
...
QUEUE_FLAG_POLL flag will be cleared when turning off 'io_poll', while
at that moment there may be IOs stuck in hw queue uncompleted. The
following polling routine won't help reap these IOs, since blk_poll()
will return immediately because of cleared QUEUE_FLAG_POLL flag. Thus
these IOs will hang until they finnaly time out. The hang out can be
observed by 'fio --engine=io_uring iodepth=1', while turning off
'io_poll' at the same time.
To fix this, freeze and flush the request queue first when turning off
'io_poll'.
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Get rid of the wrapper for trace_block_rq_insert() and call the function
directly.
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit a1ce35fa49 ("block: remove dead elevator code") removed the last
callers of blk_pm_requeue_request(), blk_pm_add_request() and
blk_pm_put_request(). Hence remove the definitions of these functions.
Removing these functions removes all users of the struct request nr_pending
member. Hence also remove 'nr_pending'. Note: 'nr_pending' is no longer
used since commit 7cedffec8e ("block: Make blk_get_request() block for
non-PM requests while suspended").
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=os1E
-----END PGP SIGNATURE-----
Merge tag 'for-5.12/block-2021-02-17' of git://git.kernel.dk/linux-block
Pull core block updates from Jens Axboe:
"Another nice round of removing more code than what is added, mostly
due to Christoph's relentless pursuit of tech debt removal/cleanups.
This pull request contains:
- Two series of BFQ improvements (Paolo, Jan, Jia)
- Block iov_iter improvements (Pavel)
- bsg error path fix (Pan)
- blk-mq scheduler improvements (Jan)
- -EBUSY discard fix (Jan)
- bvec allocation improvements (Ming, Christoph)
- bio allocation and init improvements (Christoph)
- Store bdev pointer in bio instead of gendisk + partno (Christoph)
- Block trace point cleanups (Christoph)
- hard read-only vs read-only split (Christoph)
- Block based swap cleanups (Christoph)
- Zoned write granularity support (Damien)
- Various fixes/tweaks (Chunguang, Guoqing, Lei, Lukas, Huhai)"
* tag 'for-5.12/block-2021-02-17' of git://git.kernel.dk/linux-block: (104 commits)
mm: simplify swapdev_block
sd_zbc: clear zone resources for non-zoned case
block: introduce blk_queue_clear_zone_settings()
zonefs: use zone write granularity as block size
block: introduce zone_write_granularity limit
block: use blk_queue_set_zoned in add_partition()
nullb: use blk_queue_set_zoned() to setup zoned devices
nvme: cleanup zone information initialization
block: document zone_append_max_bytes attribute
block: use bi_max_vecs to find the bvec pool
md/raid10: remove dead code in reshape_request
block: mark the bio as cloned in bio_iov_bvec_set
block: set BIO_NO_PAGE_REF in bio_iov_bvec_set
block: remove a layer of indentation in bio_iov_iter_get_pages
block: turn the nr_iovecs argument to bio_alloc* into an unsigned short
block: remove the 1 and 4 vec bvec_slabs entries
block: streamline bvec_alloc
block: factor out a bvec_alloc_gfp helper
block: move struct biovec_slab to bio.c
block: reuse BIO_INLINE_VECS for integrity bvecs
...
With llist_head it is possible to avoid the locking (the irq-off region)
when items are added. This makes it possible to add items on a remote
CPU without additional locking.
llist_add() returns true if the list was previously empty. This can be
used to invoke the SMP function call / raise sofirq only if the first
item was added (otherwise it is already pending).
This simplifies the code a little and reduces the IRQ-off regions.
blk_mq_raise_softirq() needs a preempt-disable section to ensure the
request is enqueued on the same CPU as the softirq is raised.
Some callers (USB-storage) invoke this path in preemptible context.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Controllers with multiple queues have their IRQ-handelers pinned to a
CPU. The core shouldn't need to complete the request on a remote CPU.
Remove this case and always raise the softirq to complete the request.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that device mapper supports inline encryption, add the ability to
evict keys from all underlying devices. When an upper layer requests
a key eviction, we simply iterate through all underlying devices
and evict that key from each device.
Co-developed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Introduce blk_ksm_update_capabilities() to update the capabilities of
a keyslot manager (ksm) in-place. The pointer to a ksm in a device's
request queue may not be easily replaced, because upper layers like
the filesystem might access it (e.g. for programming keys/checking
capabilities) at the same time the device wants to replace that
request queue's ksm (and free the old ksm's memory). This function
allows the device to update the capabilities of the ksm in its request
queue directly. Devices can safely update the ksm this way without any
synchronization with upper layers *only* if the updated (new) ksm
continues to support all the crypto capabilities that the old ksm did
(see description below for blk_ksm_is_superset() for why this is so).
Also introduce blk_ksm_is_superset() which checks whether one ksm's
capabilities are a (not necessarily strict) superset of another ksm's.
The blk-crypto framework requires that crypto capabilities that were
advertised when a bio was created continue to be supported by the
device until that bio is ended - in practice this probably means that
a device's advertised crypto capabilities can *never* "shrink" (since
there's no synchronization between bio creation and when a device may
want to change its advertised capabilities) - so a previously
advertised crypto capability must always continue to be supported.
This function can be used to check that a new ksm is a valid
replacement for an old ksm.
Signed-off-by: Satya Tangirala <satyat@google.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
The device mapper may map over devices that have inline encryption
capabilities, and to make use of those capabilities, the DM device must
itself advertise those inline encryption capabilities. One way to do this
would be to have the DM device set up a keyslot manager with a
"sufficiently large" number of keyslots, but that would use a lot of
memory. Also, the DM device itself has no "keyslots", and it doesn't make
much sense to talk about "programming a key into a DM device's keyslot
manager", so all that extra memory used to represent those keyslots is just
wasted. All a DM device really needs to be able to do is advertise the
crypto capabilities of the underlying devices in a coherent manner and
expose a way to evict keys from the underlying devices.
There are also devices with inline encryption hardware that do not
have a limited number of keyslots. One can send a raw encryption key along
with a bio to these devices (as opposed to typical inline encryption
hardware that require users to first program a raw encryption key into a
keyslot, and send the index of that keyslot along with the bio). These
devices also only need the same things from the keyslot manager that DM
devices need - a way to advertise crypto capabilities and potentially a way
to expose a function to evict keys from hardware.
So we introduce a "passthrough" keyslot manager that provides a way to
represent a keyslot manager that doesn't have just a limited number of
keyslots, and for which do not require keys to be programmed into keyslots.
DM devices can set up a passthrough keyslot manager in their request
queues, and advertise appropriate crypto capabilities based on those of the
underlying devices. Blk-crypto does not attempt to program keys into any
keyslots in the passthrough keyslot manager. Instead, if/when the bio is
resubmitted to the underlying device, blk-crypto will try to program the
key into the underlying device's keyslot manager.
Signed-off-by: Satya Tangirala <satyat@google.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Introduce the internal function blk_queue_clear_zone_settings() to
cleanup all limits and resources related to zoned block devices. This
new function is called from blk_queue_set_zoned() when a disk zoned
model is set to BLK_ZONED_NONE. This particular case can happens when a
partition is created on a host-aware scsi disk.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@edc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Per ZBC and ZAC specifications, host-managed SMR hard-disks mandate that
all writes into sequential write required zones be aligned to the device
physical block size. However, NVMe ZNS does not have this constraint and
allows write operations into sequential zones to be aligned to the
device logical block size. This inconsistency does not help with
software portability across device types.
To solve this, introduce the zone_write_granularity queue limit to
indicate the alignment constraint, in bytes, of write operations into
zones of a zoned block device. This new limit is exported as a
read-only sysfs queue attribute and the helper
blk_queue_zone_write_granularity() introduced for drivers to set this
limit.
The function blk_queue_set_zoned() is modified to set this new limit to
the device logical block size by default. NVMe ZNS devices as well as
zoned nullb devices use this default value as is. The scsi disk driver
is modified to execute the blk_queue_zone_write_granularity() helper to
set the zone write granularity of host-managed SMR disks to the disk
physical block size.
The accessor functions queue_zone_write_granularity() and
bdev_zone_write_granularity() are also introduced.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@edc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When changing the zoned model of host-aware zoned block devices, use
blk_queue_set_zoned() instead of directly assigning the gendisk queue
zoned limit.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@edc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add bio_add_zone_append_page(), a wrapper around bio_add_hw_page() which
is intended to be used by file systems that directly add pages to a bio
instead of using bio_iov_iter_get_pages().
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Instead of encoding of the bvec pool using magic bio flags, just use
a helper to find the pool based on the max_vecs value.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio_iov_bvec_set clones the bio_vecs from the iter, and thus should be
treated like a cloned bio in every respect. That also includes not
touching bi_max_vecs as that is a property of the bio allocation and not
its current payload.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio_iov_bvec_set assigns the foreign bvec, so setting the NO_PAGE_REF
directly there seems like the best fit.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove a pointless layer of indentation after a return statement.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The bi_max_vecs and bi_vcnt fields are defined as unsigned short, so
don't allow passing larger values in.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
All bios with up to 4 bvecs use the inline bvecs in the bio itself, so
don't bother to define bvec_slabs entries for them. Also decruftify
the bvec_slabs definition and initialization while we're at it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Avoid the pointless goto by trying the slab allocation first and falling
through to the mempool.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Clean up bvec_alloc a little by factoring out a helper for the gfp_t
manipulations.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
struct biovec_slab is only used inside of bio.c, so move it there.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bvec_alloc always uses biovec_slabs, and thus always needs to use the
same number of inline vecs. Share a single definition for the data
and integrity bvecs.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=fKCL
-----END PGP SIGNATURE-----
Merge tag 'block-5.11-2021-02-05' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
"A few small regression fixes:
- NVMe pull request from Christoph:
- more quirks for buggy devices (Thorsten Leemhuis, Claus Stovgaard)
- update the email address for Keith (Keith Busch)
- fix an out of bounds access in nvmet-tcp (Sagi Grimberg)
- Regression fix for BFQ shallow depth calculations introduced in
this merge window (Lin)"
* tag 'block-5.11-2021-02-05' of git://git.kernel.dk/linux-block:
nvmet-tcp: fix out-of-bounds access when receiving multiple h2cdata PDUs
bfq-iosched: Revert "bfq: Fix computation of shallow depth"
update the email address for Keith Bush
nvme-pci: ignore the subsysem NQN on Phison E16
nvme-pci: avoid the deepest sleep state on Kingston A2000 SSDs
This reverts commit 6d4d273588.
bfq.limit_depth passes word_depths[] as shallow_depth down to sbitmap core
sbitmap_get_shallow, which uses just the number to limit the scan depth of
each bitmap word, formula:
scan_percentage_for_each_word = shallow_depth / (1 << sbimap->shift) * 100%
That means the comments's percentiles 50%, 75%, 18%, 37% of bfq are correct.
But after commit patch 'bfq: Fix computation of shallow depth', we use
sbitmap.depth instead, as a example in following case:
sbitmap.depth = 256, map_nr = 4, shift = 6; sbitmap_word.depth = 64.
The resulsts of computed bfqd->word_depths[] are {128, 192, 48, 96}, and
three of the numbers exceed core dirver's 'sbitmap_word.depth=64' limit
nothing.
Signed-off-by: Lin Feng <linf@wangsu.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio_init() clears bio instance, so the bvec index has to be set after
bio_init(), otherwise bio->bi_io_vec may be leaked.
Fixes: 3175199ab0 ("block: split bio_kmalloc from bio_alloc_bioset")
Cc: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Cc: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add a resource-managed variant of blk_ksm_init() so that drivers don't
have to worry about calling blk_ksm_destroy().
Note that the implementation uses a custom devres action to call
blk_ksm_destroy() rather than switching the two allocations to be
directly devres-managed, e.g. with devm_kmalloc(). This is because we
need to keep zeroing the memory containing the keyslots when it is
freed, and also because we want to continue using kvmalloc() (and there
is no devm_kvmalloc()).
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Satya Tangirala <satyat@google.com>
Acked-by: Jens Axboe <axboe@kernel.dk>
Link: https://lore.kernel.org/r/20210121082155.111333-2-ebiggers@kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
-----BEGIN PGP SIGNATURE-----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=hb7W
-----END PGP SIGNATURE-----
Merge tag 'block-5.11-2021-01-29' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
"All over the place fixes for this release:
- blk-cgroup iteration teardown resched fix (Baolin)
- NVMe pull request from Christoph:
- add another Write Zeroes quirk (Chaitanya Kulkarni)
- handle a no path available corner case (Daniel Wagner)
- use the proper RCU aware list_add helper (Chao Leng)
- bcache regression fix (Coly)
- bdev->bd_size_lock IRQ fix. This will be fixed in drivers for 5.12,
but for now, we'll make it IRQ safe (Damien)
- null_blk zoned init fix (Damien)
- add_partition() error handling fix (Dinghao)
- s390 dasd kobject fix (Jan)
- nbd fix for freezing queue while adding connections (Josef)
- tag queueing regression fix (Ming)
- revert of a patch that inadvertently meant that we regressed write
performance on raid (Maxim)"
* tag 'block-5.11-2021-01-29' of git://git.kernel.dk/linux-block:
null_blk: cleanup zoned mode initialization
nvme-core: use list_add_tail_rcu instead of list_add_tail for nvme_init_ns_head
nvme-multipath: Early exit if no path is available
nvme-pci: add the DISABLE_WRITE_ZEROES quirk for a SPCC device
bcache: only check feature sets when sb->version >= BCACHE_SB_VERSION_CDEV_WITH_FEATURES
block: fix bd_size_lock use
blk-cgroup: Use cond_resched() when destroy blkgs
Revert "block: simplify set_init_blocksize" to regain lost performance
nbd: freeze the queue while we're adding connections
s390/dasd: Fix inconsistent kobject removal
block: Fix an error handling in add_partition
blk-mq: test QUEUE_FLAG_HCTX_ACTIVE for sbitmap_shared in hctx_may_queue
Commit 684da7628d ("block: remove unnecessary argument from
blk_execute_rq") changes the signature of blk_execute_rq(), but misses
to adjust its kernel-doc.
Hence, make htmldocs warns on ./block/blk-exec.c:78:
warning: Excess function parameter 'q' description in 'blk_execute_rq'
Drop removed argument from kernel-doc of blk_execute_rq() as well.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Acked-by: Guoqing Jiang <Guoqing.jiang@cloud.ionos.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit 52f019d43c ("block: add a hard-readonly flag to struct gendisk")
provides some kernel-doc for set_disk_ro(), but introduces a small typo.
Hence, make htmldocs warns on ./block/genhd.c:1441:
warning: Function parameter or member 'read_only' not described in 'set_disk_ro'
warning: Excess function parameter 'ready_only' description in 'set_disk_ro'
Remove that typo in the kernel-doc for set_disk_ro().
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Some block device drivers, e.g. the skd driver, call set_capacity() with
IRQ disabled. This results in lockdep ito complain about inconsistent
lock states ("inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage")
because set_capacity takes a block device bd_size_lock using the
functions spin_lock() and spin_unlock(). Ensure a consistent locking
state by replacing these calls with spin_lock_irqsave() and
spin_lock_irqrestore(). The same applies to bdev_set_nr_sectors().
With this fix, all lockdep complaints are resolved.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
On !PREEMPT kernel, we can get below softlockup when doing stress
testing with creating and destroying block cgroup repeatly. The
reason is it may take a long time to acquire the queue's lock in
the loop of blkcg_destroy_blkgs(), or the system can accumulate a
huge number of blkgs in pathological cases. We can add a need_resched()
check on each loop and release locks and do cond_resched() if true
to avoid this issue, since the blkcg_destroy_blkgs() is not called
from atomic contexts.
[ 4757.010308] watchdog: BUG: soft lockup - CPU#11 stuck for 94s!
[ 4757.010698] Call trace:
[ 4757.010700] blkcg_destroy_blkgs+0x68/0x150
[ 4757.010701] cgwb_release_workfn+0x104/0x158
[ 4757.010702] process_one_work+0x1bc/0x3f0
[ 4757.010704] worker_thread+0x164/0x468
[ 4757.010705] kthread+0x108/0x138
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There is no point in allocating memory for a synchronous flush.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Acked-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bio_kmalloc shares almost no logic with the bio_set based fast path
in bio_alloc_bioset. Split it into an entirely separate implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Acked-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Use bio_kmalloc instead of open coding it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Acked-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently whenever bfq queue has a request queued we add now -
last_completion_time to the think time statistics. This is however
misleading in case the process is able to submit several requests in
parallel because e.g. if the queue has request completed at time T0 and
then queues new requests at times T1, T2, then we will add T1-T0 and
T2-T0 to think time statistics which just doesn't make any sence (the
queue's think time is penalized by the queue being able to submit more
IO). So add to think time statistics only time intervals when the queue
had no IO pending.
Signed-off-by: Jan Kara <jack@suse.cz>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
[axboe: fix whitespace on empty line]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Use local variable 'ttime' instead of dereferencing bfqq.
Signed-off-by: Jan Kara <jack@suse.cz>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bfq_setup_cooperator() uses bfqd->in_serv_last_pos so detect whether it
makes sense to merge current bfq queue with the in-service queue.
However if the in-service queue is freshly scheduled and didn't dispatch
any requests yet, bfqd->in_serv_last_pos is stale and contains value
from the previously scheduled bfq queue which can thus result in a bogus
decision that the two queues should be merged. This bug can be observed
for example with the following fio jobfile:
[global]
direct=0
ioengine=sync
invalidate=1
size=1g
rw=read
[reader]
numjobs=4
directory=/mnt
where the 4 processes will end up in the one shared bfq queue although
they do IO to physically very distant files (for some reason I was able to
observe this only with slice_idle=1ms setting).
Fix the problem by invalidating bfqd->in_serv_last_pos when switching
in-service queue.
Fixes: 058fdecc6d ("block, bfq: fix in-service-queue check for queue merging")
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When calling blkcg_schedule_throttle(), for the same queue,
redundant get/put operations can be removed.
Signed-off-by: Chunguang Xu <brookxu@tencent.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The first parameter rwb is not used for this function.
So just remove it.
Signed-off-by: Lei Chen <lennychen@tencent.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Cloned bios are can be used to on the same device, in which case we need
to inherit the BIO_REMAPPED flag to avoid a double partition remap. When
the cloned bios are used on another device, bio_set_dev will clear the flag.
Fixes: 309dca309f ("block: store a block_device pointer in struct bio")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It's only used in the same file, mark is appropriately static.
Fixes: 71217df39d ("block, bfq: make waker-queue detection more robust")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In the presence of many parallel I/O flows, the detection of waker
bfq_queues suffers from false positives. This commits addresses this
issue by making the filtering of actual wakers more selective. In more
detail, a candidate waker must be found to meet waker requirements
three times before being promoted to actual waker.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
To prevent injection information from being lost on bfq_queue merging,
also the amount of service that a bfq_queue receives must be saved and
restored when the bfq_queue is merged and split, respectively.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
To prevent weight-raising information from being lost on bfq_queue merging,
also the amount of service that a bfq_queue receives must be saved and
restored when the bfq_queue is merged and split, respectively.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A bfq_queue may happen to be deemed as soft real-time while it is
still enjoying interactive weight-raising. If this happens because of
a false positive, then the bfq_queue is likely to loose its soft
real-time status soon. Upon losing such a status, the bfq_queue must
get back its interactive weight-raising, if its interactive period is
not over yet. But this case is not handled. This commit corrects this
error.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Upon an I/O-dispatch attempt, BFQ may detect that it was better to
plug I/O dispatch, and to wait for a new request to arrive for the
currently in-service queue. But the arrival of a new request for an
empty bfq_queue, and thus the switch from idle to busy of the
bfq_queue, may cause the scenario to change, and make plugging no
longer needed for service guarantees, or more convenient for
throughput. In this case, keeping I/O-dispatch plugged would certainly
lower throughput.
To address this issue, this commit makes such a check, and stops
plugging I/O if it is better to stop plugging I/O.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Some BFQ mechanisms make their decisions on a bfq_queue basing also on
whether the bfq_queue is I/O bound. In this respect, the current logic
for evaluating whether a bfq_queue is I/O bound is rather rough. This
commits replaces this logic with a more effective one.
The new logic measures the percentage of time during which a bfq_queue
is active, and marks the bfq_queue as I/O bound if the latter if this
percentage is above a fixed threshold.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When an already remapped bio is resubmitted (e.g. by blk_queue_split),
bio_check_eod will compare the remapped bi_sector against the size
of the partition, leading to spurious I/O failures.
Skip the EOD check in this case.
Fixes: 309dca309f ("block: store a block_device pointer in struct bio")
Reported-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The block layer spends quite a while in blkdev_direct_IO() to copy and
initialise bio's bvec. However, if we've already got a bvec in the input
iterator it might be reused in some cases, i.e. when new
ITER_BVEC_FLAG_FIXED flag is set. Simple tests show considerable
performance boost, and it also reduces memory footprint.
Suggested-by: Matthew Wilcox <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Direct IO does not operate on the current working set of pages managed
by the kernel, so it should not be accounted as memory stall to PSI
infrastructure.
The block layer and iomap direct IO use bio_iov_iter_get_pages()
to build bios, and they are the only users of it, so to avoid PSI
tracking for them clear out BIO_WORKINGSET flag. Do same for
dio_bio_submit() because fs/direct_io constructs bios by hand directly
calling bio_add_page().
Reported-by: Christoph Hellwig <hch@infradead.org>
Suggested-by: Christoph Hellwig <hch@infradead.org>
Suggested-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Once we have called device_initialize(), we should use put_device() to
give up the reference on error, just like what we have done on failure
of device_add().
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In case of blk_mq_is_sbitmap_shared(), we should test QUEUE_FLAG_HCTX_ACTIVE against
q->queue_flags instead of BLK_MQ_S_TAG_ACTIVE.
So fix it.
Cc: John Garry <john.garry@huawei.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Fixes: f1b49fdc1c ("blk-mq: Record active_queues_shared_sbitmap per tag_set for when using shared sbitmap")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: John Garry <john.garry@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bvec_alloc(), bvec_free() and bvec_nr_vecs() are only used inside block
layer core functions, no need to declare them in public header.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
bvec_alloc() may allocate more bio vectors than requested, so set
.bi_max_vecs as actual allocated vector number, instead of the requested
number. This way can help fs build bigger bio because new bio often won't
be allocated until the current one becomes full.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The inline bvecs won't be used if user needn't bvecs by not passing
BIOSET_NEED_BVECS, so don't allocate bvecs in this situation.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Tested-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
q->bio_split is only used by bio_split() for fast cloning bio, and no
need to allocate bvecs, so remove this flag.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Tested-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Managing bio slab cache via xarray by using slab cache size as xarray
index, and storing 'struct bio_slab' instance into xarray.
So code is simplified a lot, meantime it becomes more readable than before.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Tested-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
As we can see, returns parent_sched_may_change whether
sd->next_in_service changes or not, so remove this judgment.
Signed-off-by: huhai <huhai@tj.kylinos.cn>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently when non-mq aware IO scheduler (BFQ, mq-deadline) is used for
a queue with multiple HW queues, the performance it rather bad. The
problem is that these IO schedulers use queue-wide locking and their
dispatch function does not respect the hctx it is passed in and returns
any request it finds appropriate. Thus locality of request access is
broken and dispatch from multiple CPUs just contends on IO scheduler
locks. For these IO schedulers there's little point in dispatching from
multiple CPUs. Instead dispatch always only from a single CPU to limit
contention.
Below is a comparison of dbench runs on XFS filesystem where the storage
is a raid card with 64 HW queues and to it attached a single rotating
disk. BFQ is used as IO scheduler:
clients MQ SQ MQ-Patched
Amean 1 39.12 (0.00%) 43.29 * -10.67%* 36.09 * 7.74%*
Amean 2 128.58 (0.00%) 101.30 * 21.22%* 96.14 * 25.23%*
Amean 4 577.42 (0.00%) 494.47 * 14.37%* 508.49 * 11.94%*
Amean 8 610.95 (0.00%) 363.86 * 40.44%* 362.12 * 40.73%*
Amean 16 391.78 (0.00%) 261.49 * 33.25%* 282.94 * 27.78%*
Amean 32 324.64 (0.00%) 267.71 * 17.54%* 233.00 * 28.23%*
Amean 64 295.04 (0.00%) 253.02 * 14.24%* 242.37 * 17.85%*
Amean 512 10281.61 (0.00%) 10211.16 * 0.69%* 10447.53 * -1.61%*
Numbers are times so lower is better. MQ is stock 5.10-rc6 kernel. SQ is
the same kernel with megaraid_sas.host_tagset_enable=0 so that the card
advertises just a single HW queue. MQ-Patched is a kernel with this
patch applied.
You can see multiple hardware queues heavily hurt performance in
combination with BFQ. The patch restores the performance.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This reverts commit b445547ec1.
Since both mq-deadline and BFQ completely ignore hctx they are passed to
their dispatch function and dispatch whatever request they deem fit
checking whether any request for a particular hctx is queued is just
pointless since we'll very likely get a request from a different hctx
anyway. In the following commit we'll deal with lock contention in these
IO schedulers in presence of multiple HW queues in a different way.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commits preserves I/O-dispatch plugging for a special symmetric
case that may suddenly turn into asymmetric: the case where only one
bfq_queue, say bfqq, is busy. In this case, not expiring bfqq does not
cause any harm to any other queues in terms of service guarantees. In
contrast, it avoids the following unlucky sequence of events: (1) bfqq
is expired, (2) a new queue with a lower weight than bfqq becomes busy
(or more queues), (3) the new queue is served until a new request
arrives for bfqq, (4) when bfqq is finally served, there are so many
requests of the new queue in the drive that the pending requests for
bfqq take a lot of time to be served. In particular, event (2) may
case even already dispatched requests of bfqq to be delayed, inside
the drive. So, to avoid this series of events, the scenario is
preventively declared as asymmetric also if bfqq is the only busy
queues. By doing so, I/O-dispatch plugging is performed for bfqq.
Tested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Paolo Valente
Bhaskar Chowdhury
Johannes Thumshirn
David Jeffery
Shin'ichiro Kawasaki
Daniel Wagner
Christoph Hellwig
Damien Le Moal
Xunlei Pang
Ming Lei
Joseph Qi
Jean Delvare
Linus Torvalds
Matthew Wilcox (Oracle)
Mikulas Patocka
Yang Yang
Jeffle Xu
Chaitanya Kulkarni
Bart Van Assche
Sebastian Andrzej Siewior
Satya Tangirala
Lin Feng
Eric Biggers
Lukas Bulwahn
Baolin Wang
Jan Kara
Chunguang Xu
Lei Chen
Jens Axboe
Pavel Begunkov
Guoqing Jiang
Pan Bian
Dinghao Liu
huhai