Currently we use separate sbitmap pairs and active_queues atomic_t for
shared sbitmap support.
However a full sets of static requests are used per HW queue, which is
quite wasteful, considering that the total number of requests usable at
any given time across all HW queues is limited by the shared sbitmap depth.
As such, it is considerably more memory efficient in the case of shared
sbitmap to allocate a set of static rqs per tag set or request queue, and
not per HW queue.
So replace the sbitmap pairs and active_queues atomic_t with a shared
tags per tagset and request queue, which will hold a set of shared static
rqs.
Since there is now no valid HW queue index to be passed to the blk_mq_ops
.init and .exit_request callbacks, pass an invalid index token. This
changes the semantics of the APIs, such that the callback would need to
validate the HW queue index before using it. Currently no user of shared
sbitmap actually uses the HW queue index (as would be expected).
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/1633429419-228500-13-git-send-email-john.garry@huawei.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>
Pull SCSI updates from James Bottomley:
"This series consists of the usual driver updates (ufs, qla2xxx,
target, smartpqi, lpfc, mpt3sas).
The core change causing the most churn was replacing the command
request field request with a macro, allowing us to offset map to it
and remove the redundant field; the same was also done for the tag
field.
The most impactful change is the final removal of scsi_ioctl, which
has been deprecated for over a decade"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (293 commits)
scsi: ufs: Fix ufshcd_request_sense_async() for Samsung KLUFG8RHDA-B2D1
scsi: ufs: ufs-exynos: Fix static checker warning
scsi: mpt3sas: Use the proper SCSI midlayer interfaces for PI
scsi: lpfc: Use the proper SCSI midlayer interfaces for PI
scsi: lpfc: Copyright updates for 14.0.0.1 patches
scsi: lpfc: Update lpfc version to 14.0.0.1
scsi: lpfc: Add bsg support for retrieving adapter cmf data
scsi: lpfc: Add cmf_info sysfs entry
scsi: lpfc: Add debugfs support for cm framework buffers
scsi: lpfc: Add support for maintaining the cm statistics buffer
scsi: lpfc: Add rx monitoring statistics
scsi: lpfc: Add support for the CM framework
scsi: lpfc: Add cmfsync WQE support
scsi: lpfc: Add support for cm enablement buffer
scsi: lpfc: Add cm statistics buffer support
scsi: lpfc: Add EDC ELS support
scsi: lpfc: Expand FPIN and RDF receive logging
scsi: lpfc: Add MIB feature enablement support
scsi: lpfc: Add SET_HOST_DATA mbox cmd to pass date/time info to firmware
scsi: fc: Add EDC ELS definition
...
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>
Move the sg_timeout and sg_reserved_size fields into the bsg_device and
scsi_device structures as they have nothing to do with generic block I/O.
Note that these values are now separate for bsg vs. SCSI device node
access, but that just matches how /dev/sg vs the other nodes has always
behaved.
Link: https://lore.kernel.org/r/20210729064845.1044147-4-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull block fixes from Jens Axboe:
- NVMe pull request (Christoph):
- tracing fix (Keith Busch)
- fix multipath head refcounting (Hannes Reinecke)
- Write Zeroes vs PI fix (me)
- drop a bogus WARN_ON (Zhihao Cheng)
- Increase max blk-cgroup policy size, now that mq-deadline
uses it too (Oleksandr)
* tag 'block-5.14-2021-07-24' of git://git.kernel.dk/linux-block:
nvme: set the PRACT bit when using Write Zeroes with T10 PI
nvme: fix nvme_setup_command metadata trace event
nvme: fix refcounting imbalance when all paths are down
nvme-pci: don't WARN_ON in nvme_reset_work if ctrl.state is not RESETTING
block: increase BLKCG_MAX_POLS
Pull more block updates from Jens Axboe:
"A combination of changes that ended up depending on both the driver
and core branch (and/or the IDE removal), and a few late arriving
fixes. In detail:
- Fix io ticks wrap-around issue (Chunguang)
- nvme-tcp sock locking fix (Maurizio)
- s390-dasd fixes (Kees, Christoph)
- blk_execute_rq polling support (Keith)
- blk-cgroup RCU iteration fix (Yu)
- nbd backend ID addition (Prasanna)
- Partition deletion fix (Yufen)
- Use blk_mq_alloc_disk for mmc, mtip32xx, ubd (Christoph)
- Removal of now dead block request types due to IDE removal
(Christoph)
- Loop probing and control device cleanups (Christoph)
- Device uevent fix (Christoph)
- Misc cleanups/fixes (Tetsuo, Christoph)"
* tag 'block-5.14-2021-07-08' of git://git.kernel.dk/linux-block: (34 commits)
blk-cgroup: prevent rcu_sched detected stalls warnings while iterating blkgs
block: fix the problem of io_ticks becoming smaller
nvme-tcp: can't set sk_user_data without write_lock
loop: remove unused variable in loop_set_status()
block: remove the bdgrab in blk_drop_partitions
block: grab a device refcount in disk_uevent
s390/dasd: Avoid field over-reading memcpy()
dasd: unexport dasd_set_target_state
block: check disk exist before trying to add partition
ubd: remove dead code in ubd_setup_common
nvme: use return value from blk_execute_rq()
block: return errors from blk_execute_rq()
nvme: use blk_execute_rq() for passthrough commands
block: support polling through blk_execute_rq
block: remove REQ_OP_SCSI_{IN,OUT}
block: mark blk_mq_init_queue_data static
loop: rewrite loop_exit using idr_for_each_entry
loop: split loop_lookup
loop: don't allow deleting an unspecified loop device
loop: move loop_ctl_mutex locking into loop_add
...
Pull device mapper updates from Mike Snitzer:
- Various DM persistent-data library improvements and fixes that
benefit both the DM thinp and cache targets.
- A few small DM kcopyd efficiency improvements.
- Significant zoned related block core, DM core and DM zoned target
changes that culminate with adding zoned append emulation (which is
required to properly fix DM crypt's zoned support).
- Various DM writecache target changes that improve efficiency. Adds an
optional "metadata_only" feature that only promotes bios flagged with
REQ_META. But the most significant improvement is writecache's
ability to pause writeback, for a confiurable time, if/when the
working set is larger than the cache (and the cache is full) -- this
ensures performance is no worse than the slower origin device.
* tag 'for-5.14/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (35 commits)
dm writecache: make writeback pause configurable
dm writecache: pause writeback if cache full and origin being written directly
dm io tracker: factor out IO tracker
dm btree remove: assign new_root only when removal succeeds
dm zone: fix dm_revalidate_zones() memory allocation
dm ps io affinity: remove redundant continue statement
dm writecache: add optional "metadata_only" parameter
dm writecache: add "cleaner" and "max_age" to Documentation
dm writecache: write at least 4k when committing
dm writecache: flush origin device when writing and cache is full
dm writecache: have ssd writeback wait if the kcopyd workqueue is busy
dm writecache: use list_move instead of list_del/list_add in writecache_writeback()
dm writecache: commit just one block, not a full page
dm writecache: remove unused gfp_t argument from wc_add_block()
dm crypt: Fix zoned block device support
dm: introduce zone append emulation
dm: rearrange core declarations for extended use from dm-zone.c
block: introduce BIO_ZONE_WRITE_LOCKED bio flag
block: introduce bio zone helpers
block: improve handling of all zones reset operation
...
With the legacy IDE driver gone drivers now use either REQ_OP_DRV_*
or REQ_OP_SCSI_*, so unify the two concepts of passthrough requests
into a single one.
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Introduce the helper functions bio_zone_no() and bio_zone_is_seq().
Both are the BIO counterparts of the request helpers blk_rq_zone_no()
and blk_rq_zone_is_seq(), respectively returning the number of the
target zone of a bio and true if the BIO target zone is sequential.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
The tags used for an IO scheduler are currently per hctx.
As such, when q->nr_hw_queues grows, so does the request queue total IO
scheduler tag depth.
This may cause problems for SCSI MQ HBAs whose total driver depth is
fixed.
Ming and Yanhui report higher CPU usage and lower throughput in scenarios
where the fixed total driver tag depth is appreciably lower than the total
scheduler tag depth:
https://lore.kernel.org/linux-block/440dfcfc-1a2c-bd98-1161-cec4d78c6dfc@huawei.com/T/#mc0d6d4f95275a2743d1c8c3e4dc9ff6c9aa3a76b
In that scenario, since the scheduler tag is got first, much contention
is introduced since a driver tag may not be available after we have got
the sched tag.
Improve this scenario by introducing request queue-wide tags for when
a tagset-wide sbitmap is used. The static sched requests are still
allocated per hctx, as requests are initialised per hctx, as in
blk_mq_init_request(..., hctx_idx, ...) ->
set->ops->init_request(.., hctx_idx, ...).
For simplicity of resizing the request queue sbitmap when updating the
request queue depth, just init at the max possible size, so we don't need
to deal with the possibly with swapping out a new sbitmap for old if
we need to grow.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/1620907258-30910-3-git-send-email-john.garry@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull block fixes from Jens Axboe:
- Fix for shared tag set exit (Bart)
- Correct ioctl range for zoned ioctls (Damien)
- Removed dead/unused function (Lin)
- Fix perf regression for shared tags (Ming)
- Fix out-of-bounds issue with kyber and preemption (Omar)
- BFQ merge fix (Paolo)
- Two error handling fixes for nbd (Sun)
- Fix weight update in blk-iocost (Tejun)
- NVMe pull request (Christoph):
- correct the check for using the inline bio in nvmet (Chaitanya
Kulkarni)
- demote unsupported command warnings (Chaitanya Kulkarni)
- fix corruption due to double initializing ANA state (me, Hou Pu)
- reset ns->file when open fails (Daniel Wagner)
- fix a NULL deref when SEND is completed with error in nvmet-rdma
(Michal Kalderon)
- Fix kernel-doc warning (Bart)
* tag 'block-5.13-2021-05-14' of git://git.kernel.dk/linux-block:
block/partitions/efi.c: Fix the efi_partition() kernel-doc header
blk-mq: Swap two calls in blk_mq_exit_queue()
blk-mq: plug request for shared sbitmap
nvmet: use new ana_log_size instead the old one
nvmet: seset ns->file when open fails
nbd: share nbd_put and return by goto put_nbd
nbd: Fix NULL pointer in flush_workqueue
blkdev.h: remove unused codes blk_account_rq
block, bfq: avoid circular stable merges
blk-iocost: fix weight updates of inner active iocgs
nvmet: demote fabrics cmd parse err msg to debug
nvmet: use helper to remove the duplicate code
nvmet: demote discovery cmd parse err msg to debug
nvmet-rdma: Fix NULL deref when SEND is completed with error
nvmet: fix inline bio check for passthru
nvmet: fix inline bio check for bdev-ns
nvme-multipath: fix double initialization of ANA state
kyber: fix out of bounds access when preempted
block: uapi: fix comment about block device ioctl
Pull block fix from Jens Axboe:
"Turns out the bio max size change still has issues, so let's get it
reverted for 5.13-rc1. We'll shake out the issues there and defer it
to 5.14 instead"
* tag 'block-5.13-2021-05-09' of git://git.kernel.dk/linux-block:
Revert "bio: limit bio max size"
Pull block fixes from Jens Axboe:
- dasd spelling fixes (Bhaskar)
- Limit bio max size on multi-page bvecs to the hardware limit, to
avoid overly large bio's (and hence latencies). Originally queued for
the merge window, but needed a fix and was dropped from the initial
pull (Changheun)
- NVMe pull request (Christoph):
- reset the bdev to ns head when failover (Daniel Wagner)
- remove unsupported command noise (Keith Busch)
- misc passthrough improvements (Kanchan Joshi)
- fix controller ioctl through ns_head (Minwoo Im)
- fix controller timeouts during reset (Tao Chiu)
- rnbd fixes/cleanups (Gioh, Md, Dima)
- Fix iov_iter re-expansion (yangerkun)
* tag 'block-5.13-2021-05-07' of git://git.kernel.dk/linux-block:
block: reexpand iov_iter after read/write
nvmet: remove unsupported command noise
nvme-multipath: reset bdev to ns head when failover
nvme-pci: fix controller reset hang when racing with nvme_timeout
nvme: move the fabrics queue ready check routines to core
nvme: avoid memset for passthrough requests
nvme: add nvme_get_ns helper
nvme: fix controller ioctl through ns_head
bio: limit bio max size
RDMA/rtrs: fix uninitialized symbol 'cnt'
s390: dasd: Mundane spelling fixes
block/rnbd: Remove all likely and unlikely
block/rnbd-clt: Check the return value of the function rtrs_clt_query
block/rnbd: Fix style issues
block/rnbd-clt: Change queue_depth type in rnbd_clt_session to size_t
bio size can grow up to 4GB when muli-page bvec is enabled.
but sometimes it would lead to inefficient behaviors.
in case of large chunk direct I/O, - 32MB chunk read in user space -
all pages for 32MB would be merged to a bio structure if the pages
physical addresses are contiguous. it makes some delay to submit
until merge complete. bio max size should be limited to a proper size.
When 32MB chunk read with direct I/O option is coming from userspace,
kernel behavior is below now in do_direct_IO() loop. it's timeline.
| bio merge for 32MB. total 8,192 pages are merged.
| total elapsed time is over 2ms.
|------------------ ... ----------------------->|
| 8,192 pages merged a bio.
| at this time, first bio submit is done.
| 1 bio is split to 32 read request and issue.
|--------------->
|--------------->
|--------------->
......
|--------------->
|--------------->|
total 19ms elapsed to complete 32MB read done from device. |
If bio max size is limited with 1MB, behavior is changed below.
| bio merge for 1MB. 256 pages are merged for each bio.
| total 32 bio will be made.
| total elapsed time is over 2ms. it's same.
| but, first bio submit timing is fast. about 100us.
|--->|--->|--->|---> ... -->|--->|--->|--->|--->|
| 256 pages merged a bio.
| at this time, first bio submit is done.
| and 1 read request is issued for 1 bio.
|--------------->
|--------------->
|--------------->
......
|--------------->
|--------------->|
total 17ms elapsed to complete 32MB read done from device. |
As a result, read request issue timing is faster if bio max size is limited.
Current kernel behavior with multipage bvec, super large bio can be created.
And it lead to delay first I/O request issue.
Signed-off-by: Changheun Lee <nanich.lee@samsung.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20210503095203.29076-1-nanich.lee@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull block updates from Jens Axboe:
"Pretty quiet round this time, which is nice. In detail:
- Series revamping bounce buffer support (Christoph)
- Dead code removal (Christoph, Bart)
- Partition iteration revamp, now using xarray (Christoph)
- Passthrough request scheduler improvements (Lin)
- Series of BFQ improvements (Paolo)
- Fix ioprio task iteration (Peter)
- Various little tweaks and fixes (Tejun, Saravanan, Bhaskar, Max,
Nikolay)"
* tag 'for-5.13/block-2021-04-27' of git://git.kernel.dk/linux-block: (41 commits)
blk-iocost: don't ignore vrate_min on QD contention
blk-mq: Fix spurious debugfs directory creation during initialization
bfq/mq-deadline: remove redundant check for passthrough request
blk-mq: bypass IO scheduler's limit_depth for passthrough request
block: Remove an obsolete comment from sg_io()
block: move bio_list_copy_data to pktcdvd
block: remove zero_fill_bio_iter
block: add queue_to_disk() to get gendisk from request_queue
block: remove an incorrect check from blk_rq_append_bio
block: initialize ret in bdev_disk_changed
block: Fix sys_ioprio_set(.which=IOPRIO_WHO_PGRP) task iteration
block: remove disk_part_iter
block: simplify diskstats_show
block: simplify show_partition
block: simplify printk_all_partitions
block: simplify partition_overlaps
block: simplify partition removal
block: take bd_mutex around delete_partitions in del_gendisk
block: refactor blk_drop_partitions
block: move more syncing and invalidation to delete_partition
...
Commit 01e99aeca3 ("blk-mq: insert passthrough request into
hctx->dispatch directly") gives high priority to passthrough requests and
bypass underlying IO scheduler. But as we allocate tag for such request it
still runs io-scheduler's callback limit_depth, while we really want is to
give full sbitmap-depth capabity to such request for acquiring available
tag.
blktrace shows PC requests(dmraid -s -c -i) hit bfq's limit_depth:
8,0 2 0 0.000000000 39952 1,0 m N bfq [bfq_limit_depth] wr_busy 0 sync 0 depth 8
8,0 2 1 0.000008134 39952 D R 4 [dmraid]
8,0 2 2 0.000021538 24 C R [0]
8,0 2 0 0.000035442 39952 1,0 m N bfq [bfq_limit_depth] wr_busy 0 sync 0 depth 8
8,0 2 3 0.000038813 39952 D R 24 [dmraid]
8,0 2 4 0.000044356 24 C R [0]
This patch introduce a new wrapper to make code not that ugly.
Signed-off-by: Lin Feng <linf@wangsu.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20210415033920.213963-1-linf@wangsu.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Sometimes we need to get the corresponding gendisk from request_queue.
It is preferred that block drivers store private data in
gendisk->private_data rather than request_queue->queuedata, e.g. see:
commit c4a59c4e5d ("dm: stop using ->queuedata").
So if only request_queue is given, we need to get its corresponding
gendisk to get the private data stored in that gendisk.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of overloading the passthrough fast path with the deprecated
block layer bounce buffering let the users that combine an old
undermaintained driver with a highmem system pay the price by always
falling back to copies in that case.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Link: https://lore.kernel.org/r/20210331073001.46776-9-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
