A previous commit wanted to make CFQ default to IOPS mode on
non-rotational storage, however it did so when the queue was
initialized and the non-rotational flag is only set later on
in the probe.
Add an elevator hook that gets called off the add_disk() path,
at that point we know that feature probing has finished, and
we can reliably check for the various flags that drivers can
set.
Fixes: 41c0126b ("block: Make CFQ default to IOPS mode on SSDs")
Tested-by: Romain Francoise <romain@orebokech.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Now blk_cleanup_queue() can be called before calling
del_gendisk()[1], inside which hctx->ctxs is touched
from blk_mq_unregister_hctx(), but the variable has
been freed by blk_cleanup_queue() at that time.
So this patch moves freeing of hctx->ctxs into queue's
release handler for fixing the oops reported by Stefan.
[1], 6cd18e711dd8075 (block: destroy bdi before blockdev is
unregistered)
Reported-by: Stefan Seyfried <stefan.seyfried@googlemail.com>
Cc: NeilBrown <neilb@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: stable@vger.kernel.org (v4.0)
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
CFQ idling causes reduced IOPS throughput on non-rotational disks.
Since disk head seeking is not applicable to SSDs, it doesn't really
help performance by anticipating future near-by IO requests.
By turning off idling (and switching to IOPS mode), we allow other
processes to dispatch IO requests down to the driver and so increase IO
throughput.
Following FIO benchmark results were taken on a cloud SSD offering with
idling on and off:
Idling iops avg-lat(ms) stddev bw
------------------------------------------------------
On 7054 90.107 38.697 28217KB/s
Off 29255 21.836 11.730 117022KB/s
fio --name=temp --size=100G --time_based --ioengine=libaio \
--randrepeat=0 --direct=1 --invalidate=1 --verify=0 \
--verify_fatal=0 --rw=randread --blocksize=4k --group_reporting=1 \
--filename=/dev/sdb --runtime=10 --iodepth=64 --numjobs=10
And the following is from a local SSD run:
Idling iops avg-lat(ms) stddev bw
------------------------------------------------------
On 19320 33.043 14.068 77281KB/s
Off 21626 29.465 12.662 86507KB/s
fio --name=temp --size=5G --time_based --ioengine=libaio \
--randrepeat=0 --direct=1 --invalidate=1 --verify=0 \
--verify_fatal=0 --rw=randread --blocksize=4k --group_reporting=1 \
--filename=/fio_data --runtime=10 --iodepth=64 --numjobs=10
Reviewed-by: Nauman Rafique <nauman@google.com>
Signed-off-by: Tahsin Erdogan <tahsin@google.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Now that bdi layer can handle per-blkcg bdi_writeback_congested state,
blk_{set|clear}_congested() can propagate non-root blkcg congestion
state to them.
This can be easily achieved by disabling the root_rl tests in
blk_{set|clear}_congested(). Note that we still need those tests when
!CONFIG_CGROUP_WRITEBACK as otherwise we'll end up flipping root blkcg
wb's congestion state for events happening on other blkcgs.
v2: Updated for bdi_writeback_congested.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_{set|clear}_queue_congested() take @q and set or clear,
respectively, the congestion state of its bdi's root wb. Because bdi
used to be able to handle congestion state only on the root wb, the
callers of those functions tested whether the congestion is on the
root blkcg and skipped if not.
This is cumbersome and makes implementation of per cgroup
bdi_writeback congestion state propagation difficult. This patch
renames blk_{set|clear}_queue_congested() to
blk_{set|clear}_congested(), and makes them take request_list instead
of request_queue and test whether the specified request_list is the
root one before updating bdi_writeback congestion state. This makes
the tests in the callers unnecessary and simplifies them.
As there are no external users of these functions, the definitions are
moved from include/linux/blkdev.h to block/blk-core.c.
This patch doesn't introduce any noticeable behavior difference.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
A blkg (blkcg_gq) can be congested and decongested independently from
other blkgs on the same request_queue. Accordingly, for cgroup
writeback support, the congestion status at bdi (backing_dev_info)
should be split and updated separately from matching blkg's.
This patch prepares by adding blkg->wb_congested and associating a
blkg with its matching per-blkcg bdi_writeback_congested on creation.
v2: Updated to associate bdi_writeback_congested instead of
bdi_writeback.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
For the planned cgroup writeback support, on each bdi
(backing_dev_info), each memcg will be served by a separate wb
(bdi_writeback). This patch updates bdi so that a bdi can host
multiple wbs (bdi_writebacks).
On the default hierarchy, blkcg implicitly enables memcg. This allows
using memcg's page ownership for attributing writeback IOs, and every
memcg - blkcg combination can be served by its own wb by assigning a
dedicated wb to each memcg. This means that there may be multiple
wb's of a bdi mapped to the same blkcg. As congested state is per
blkcg - bdi combination, those wb's should share the same congested
state. This is achieved by tracking congested state via
bdi_writeback_congested structs which are keyed by blkcg.
bdi->wb remains unchanged and will keep serving the root cgroup.
cgwb's (cgroup wb's) for non-root cgroups are created on-demand or
looked up while dirtying an inode according to the memcg of the page
being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree
by its memcg id. Once an inode is associated with its wb, it can be
retrieved using inode_to_wb().
Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all
pages will keep being associated with bdi->wb.
v3: inode_attach_wb() in account_page_dirtied() moved inside
mapping_cap_account_dirty() block where it's known to be !NULL.
Also, an unnecessary NULL check before kfree() removed. Both
detected by the kbuild bot.
v2: Updated so that wb association is per inode and wb is per memcg
rather than blkcg.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: kbuild test robot <fengguang.wu@intel.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@fb.com>
cgroup writeback requires support from both bdi and filesystem sides.
Add BDI_CAP_CGROUP_WRITEBACK and FS_CGROUP_WRITEBACK to indicate
support and enable BDI_CAP_CGROUP_WRITEBACK on block based bdi's by
default. Also, define CONFIG_CGROUP_WRITEBACK which is enabled if
both MEMCG and BLK_CGROUP are enabled.
inode_cgwb_enabled() which determines whether a given inode's both bdi
and fs support cgroup writeback is added.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@fb.com>
With the planned cgroup writeback support, backing-dev related
declarations will be more widely used across block and cgroup;
unfortunately, including backing-dev.h from include/linux/blkdev.h
makes cyclic include dependency quite likely.
This patch separates out backing-dev-defs.h which only has the
essential definitions and updates blkdev.h to include it. c files
which need access to more backing-dev details now include
backing-dev.h directly. This takes backing-dev.h off the common
include dependency chain making it a lot easier to use it across block
and cgroup.
v2: fs/fat build failure fixed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback)
and the role of the separation is unclear. For cgroup support for
writeback IOs, a bdi will be updated to host multiple wb's where each
wb serves writeback IOs of a different cgroup on the bdi. To achieve
that, a wb should carry all states necessary for servicing writeback
IOs for a cgroup independently.
This patch moves bdi->state into wb.
* enum bdi_state is renamed to wb_state and the prefix of all enums is
changed from BDI_ to WB_.
* Explicit zeroing of bdi->state is removed without adding zeoring of
wb->state as the whole data structure is zeroed on init anyway.
* As there's still only one bdi_writeback per backing_dev_info, all
uses of bdi->state are mechanically replaced with bdi->wb.state
introducing no behavior changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: drbd-dev@lists.linbit.com
Cc: Neil Brown <neilb@suse.de>
Cc: Alasdair Kergon <agk@redhat.com>
Cc: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently, a bio can only be associated with the io_context and blkcg
of %current using bio_associate_current(). This is too restrictive
for cgroup writeback support. Implement bio_associate_blkcg() which
associates a bio with the specified blkcg.
bio_associate_blkcg() leaves the io_context unassociated.
bio_associate_current() is updated so that it considers a bio as
already associated if it has a blkcg_css, instead of an io_context,
associated with it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
bio_associate_current() currently open codes task_css() and
css_tryget_online() to find and pin $current's blkcg css. Abstract it
into task_get_css() which is implemented from cgroup side. As a task
is always associated with an online css for every subsystem except
while the css_set update is propagating, task_get_css() retries till
css_tryget_online() succeeds.
This is a cleanup and shouldn't lead to noticeable behavior changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Add global constant blkcg_root_css which points to &blkcg_root.css.
This will be used by cgroup writeback support. If blkcg is disabled,
it's defined as ERR_PTR(-EINVAL).
v2: The declarations moved to include/linux/blk-cgroup.h as suggested
by Vivek.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently, blkcg does a minor optimization where the root blkcg is
created when the first blkcg policy is activated on a queue and
destroyed on the deactivation of the last. On systems where blkcg is
configured but not used, this saves one blkcg_gq struct per queue. On
systems where blkcg is actually used, there's no difference. The only
case where this can lead to any meaninful, albeit still minute, save
in memory consumption is when all blkcg policies are deactivated after
being widely used in the system, which is a hihgly unlikely scenario.
The conditional existence of root blkcg_gq has already created several
bugs in blkcg and became an issue once again for the new per-cgroup
wb_congested mechanism for cgroup writeback support leading to a NULL
dereference when no blkcg policy is active. This is really not worth
bothering with. This patch makes blkcg always allocate and link the
root blkcg_gq and release it only on queue destruction.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
cgroup aware writeback support will require exposing some of blkcg
details. In preprataion, move block/blk-cgroup.h to
include/linux/blk-cgroup.h. This patch is pure file move.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Storage controllers may expose multiple block devices that share hardware
resources managed by blk-mq. This patch enhances the shared tags so a
low-level driver can access the shared resources not tied to the unshared
h/w contexts. This way the LLD can dynamically add and delete disks and
request queues without having to track all the request_queue hctx's to
iterate outstanding tags.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
We can safely merge anything that wont generate an SG list entry,
so if the bio is data-less (discard), don't look at potential
SG gaps.
Signed-off-by: Jens Axboe <axboe@fb.com>
bdi_unregister() now contains very little functionality.
It contains a "WARN_ON" if bdi->dev is NULL. This warning is of no
real consequence as bdi->dev isn't needed by anything else in the function,
and it triggers if
blk_cleanup_queue() -> bdi_destroy()
is called before bdi_unregister, which happens since
Commit: 6cd18e711dd8 ("block: destroy bdi before blockdev is unregistered.")
So this isn't wanted.
It also calls bdi_set_min_ratio(). This needs to be called after
writes through the bdi have all been flushed, and before the bdi is destroyed.
Calling it early is better than calling it late as it frees up a global
resource.
Calling it immediately after bdi_wb_shutdown() in bdi_destroy()
perfectly fits these requirements.
So bdi_unregister() can be discarded with the important content moved to
bdi_destroy(), as can the
writeback_bdi_unregister
event which is already not used.
Reported-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org (v4.0)
Fixes: c4db59d31e39 ("fs: don't reassign dirty inodes to default_backing_dev_info")
Fixes: 6cd18e711dd8 ("block: destroy bdi before blockdev is unregistered.")
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Tested-by: Nicholas Moulin <nicholas.w.moulin@linux.intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently dm-multipath has to clone the bios for every request sent
to the lower devices, which wastes cpu cycles and ties down memory.
This patch instead adds a new REQ_CLONE flag that instructs req_bio_endio
to not complete bios attached to a request, which we set on clone
requests similar to bios in a flush sequence. With this change I/O
errors on a path failure only get propagated to dm-multipath, which
can then either resubmit the I/O or complete the bios on the original
request.
I've done some basic testing of this on a Linux target with ALUA support,
and it survives path failures during I/O nicely.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Commit c4cf5261 ("bio: skip atomic inc/dec of ->bi_remaining for
non-chains") regressed all existing callers that followed this pattern:
1) saving a bio's original bi_end_io
2) wiring up an intermediate bi_end_io
3) restoring the original bi_end_io from intermediate bi_end_io
4) calling bio_endio() to execute the restored original bi_end_io
The regression was due to BIO_CHAIN only ever getting set if
bio_inc_remaining() is called. For the above pattern it isn't set until
step 3 above (step 2 would've needed to establish BIO_CHAIN). As such
the first bio_endio(), in step 2 above, never decremented __bi_remaining
before calling the intermediate bi_end_io -- leaving __bi_remaining with
the value 1 instead of 0. When bio_inc_remaining() occurred during step
3 it brought it to a value of 2. When the second bio_endio() was
called, in step 4 above, it should've called the original bi_end_io but
it didn't because there was an extra reference that wasn't dropped (due
to atomic operations being optimized away since BIO_CHAIN wasn't set
upfront).
Fix this issue by removing the __bi_remaining management complexity for
all callers that use the above pattern -- bio_chain() is the only
interface that _needs_ to be concerned with __bi_remaining. For the
above pattern callers just expect the bi_end_io they set to get called!
Remove bio_endio_nodec() and also remove all bio_inc_remaining() calls
that aren't associated with the bio_chain() interface.
Also, the bio_inc_remaining() interface has been moved local to bio.c.
Fixes: c4cf5261 ("bio: skip atomic inc/dec of ->bi_remaining for non-chains")
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The only possible problem of using mutex_lock() instead of trylock
is about deadlock.
If there aren't any locks held before calling blkdev_reread_part(),
deadlock can't be caused by this conversion.
If there are locks held before calling blkdev_reread_part(),
and if these locks arn't required in open, close handler and I/O
path, deadlock shouldn't be caused too.
Both user space's ioctl(BLKRRPART) and md_setup_drive() from
init/do_mounts_md.c belongs to the 1st case, so the conversion is safe
for the two cases.
For loop, the previous patches in this pathset has fixed the ABBA lock
dependency, so the conversion is OK.
For nbd, tx_lock is held when calling the function:
- both open and release won't hold the lock
- when blkdev_reread_part() is run, I/O thread has been stopped
already, so tx_lock won't be acquired in I/O path at that time.
- so the conversion won't cause deadlock for nbd
For dasd, both dasd_open(), dasd_release() and request function don't
acquire any mutex/semphone, so the conversion should be safe.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch exports blkdev_reread_part() for block drivers, also
introduce __blkdev_reread_part().
For some drivers, such as loop, reread of partitions can be run
from the release path, and bd_mutex may already be held prior to
calling ioctl_by_bdev(bdev, BLKRRPART, 0), so introduce
__blkdev_reread_part for use in such cases.
CC: Christoph Hellwig <hch@lst.de>
CC: Jens Axboe <axboe@kernel.dk>
CC: Tejun Heo <tj@kernel.org>
CC: Alexander Viro <viro@zeniv.linux.org.uk>
CC: Markus Pargmann <mpa@pengutronix.de>
CC: Stefan Weinhuber <wein@de.ibm.com>
CC: Stefan Haberland <stefan.haberland@de.ibm.com>
CC: Sebastian Ott <sebott@linux.vnet.ibm.com>
CC: Fabian Frederick <fabf@skynet.be>
CC: Ming Lei <ming.lei@canonical.com>
CC: David Herrmann <dh.herrmann@gmail.com>
CC: Andrew Morton <akpm@linux-foundation.org>
CC: Peter Zijlstra <peterz@infradead.org>
CC: nbd-general@lists.sourceforge.net
CC: linux-s390@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Since the big barrier rewrite/removal in 2007 we never fail FLUSH or
FUA requests, which means we can remove the magic BIO_EOPNOTSUPP flag
to help propagating those to the buffer_head layer.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
lockdep gets unhappy about the not disabling irqs when using the queue_lock
around it. Instead of trying to fix that up just switch to an atomic_t
and get rid of the lock.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
With commit ff36ab345 ("dm: remove request-based logic from
make_request_fn wrapper") DM no longer calls blk_queue_bio() directly,
so remove its export. Doing so required a forward declaration in
blk-core.c.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Last patch makes plug work for multiple queue case. However it only
works for single disk case, because it assumes only one request in the
plug list. If a task is accessing multiple disks, eg MD/DM, the
assumption is wrong. Let blk_attempt_plug_merge() record request from
the same queue.
V2: use NULL parameter in !mq case. Fix a bug. Add comments in
blk_attempt_plug_merge to make it less (hopefully) confusion.
Cc: Jens Axboe <axboe@fb.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
plug is still helpful for workload with IO merge, but it can be harmful
otherwise especially with multiple hardware queues, as there is
(supposed) no lock contention in this case and plug can introduce
latency. For multiple queues, we do limited plug, eg plug only if there
is request merge. If a request doesn't have merge with following
request, the requet will be dispatched immediately.
V2: check blk_queue_nomerges() as suggested by Jeff.
Cc: Jens Axboe <axboe@fb.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If we directly issue a request and it fails, we use
blk_mq_merge_queue_io(). But we already assigned bio to a request in
blk_mq_bio_to_request. blk_mq_merge_queue_io shouldn't run
blk_mq_bio_to_request again.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The following appears in blk_sq_make_request:
/*
* If we have multiple hardware queues, just go directly to
* one of those for sync IO.
*/
We clearly don't have multiple hardware queues, here! This comment was
introduced with this commit 07068d5b8e (blk-mq: split make request
handler for multi and single queue):
We want slightly different behavior from them:
- On single queue devices, we currently use the per-process plug
for deferred IO and for merging.
- On multi queue devices, we don't use the per-process plug, but
we want to go straight to hardware for SYNC IO.
The old code had this:
use_plug = !is_flush_fua && ((q->nr_hw_queues == 1) || !is_sync);
and that was converted to:
use_plug = !is_flush_fua && !is_sync;
which is not equivalent. For the single queue case, that second half of
the && expression is always true. So, what I think was actually inteded
follows (and this more closely matches what is done in blk_queue_bio).
V2: delete the 'likely', which should not be a big deal
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Current code looks like inner plug gets flushed with a
blk_finish_plug(). Actually it's a nop. All requests/callbacks are added
to current->plug, while only outmost plug is assigned to current->plug.
So inner plug always has empty request/callback list, which makes
blk_flush_plug_list() a nop. This tries to make the code more clear.
Signed-off-by: Shaohua Li <shli@fb.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This removes the request types and hacks from the block code and into the
old IDE driver. There is a small amunt of code duplication due to this,
but it's not too bad.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Struct bio has a reference count that controls when it can be freed.
Most uses cases is allocating the bio, which then returns with a
single reference to it, doing IO, and then dropping that single
reference. We can remove this atomic_dec_and_test() in the completion
path, if nobody else is holding a reference to the bio.
If someone does call bio_get() on the bio, then we flag the bio as
now having valid count and that we must properly honor the reference
count when it's being put.
Tested-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Struct bio has an atomic ref count for chained bio's, and we use this
to know when to end IO on the bio. However, most bio's are not chained,
so we don't need to always introduce this atomic operation as part of
ending IO.
Add a helper to elevate the bi_remaining count, and flag the bio as
now actually needing the decrement at end_io time. Rename the field
to __bi_remaining to catch any current users of this doing the
incrementing manually.
For high IOPS workloads, this reduces the overhead of bio_endio()
substantially.
Tested-by: Robert Elliott <elliott@hp.com>
Acked-by: Kent Overstreet <kent.overstreet@gmail.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@fb.com>
Normally if driver is busy to dispatch a request the logic is like below:
block layer: driver:
__blk_mq_run_hw_queue
a. blk_mq_stop_hw_queue
b. rq add to ctx->dispatch
later:
1. blk_mq_start_hw_queue
2. __blk_mq_run_hw_queue
But it's possible step 1-2 runs between a and b. And since rq isn't in
ctx->dispatch yet, step 2 will not run rq. The rq might get lost if
there are no subsequent requests kick in.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Because of the peculiar way that md devices are created (automatically
when the device node is opened), a new device can be created and
registered immediately after the
blk_unregister_region(disk_devt(disk), disk->minors);
call in del_gendisk().
Therefore it is important that all visible artifacts of the previous
device are removed before this call. In particular, the 'bdi'.
Since:
commit c4db59d31e39ea067c32163ac961e9c80198fd37
Author: Christoph Hellwig <hch@lst.de>
fs: don't reassign dirty inodes to default_backing_dev_info
moved the
device_unregister(bdi->dev);
call from bdi_unregister() to bdi_destroy() it has been quite easy to
lose a race and have a new (e.g.) "md127" be created after the
blk_unregister_region() call and before bdi_destroy() is ultimately
called by the final 'put_disk', which must come after del_gendisk().
The new device finds that the bdi name is already registered in sysfs
and complains
> [ 9627.630029] WARNING: CPU: 18 PID: 3330 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x5a/0x70()
> [ 9627.630032] sysfs: cannot create duplicate filename '/devices/virtual/bdi/9:127'
We can fix this by moving the bdi_destroy() call out of
blk_release_queue() (which can happen very late when a refcount
reaches zero) and into blk_cleanup_queue() - which happens exactly when the md
device driver calls it.
Then it is only necessary for md to call blk_cleanup_queue() before
del_gendisk(). As loop.c devices are also created on demand by
opening the device node, we make the same change there.
Fixes: c4db59d31e39ea067c32163ac961e9c80198fd37
Reported-by: Azat Khuzhin <a3at.mail@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: stable@vger.kernel.org (v4.0)
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Our issue is descripted in below call path:
->elevator_init
->elevator_init_fn
->{cfq,deadline,noop}_init_queue
->elevator_alloc
->kzalloc_node
fail to call kzalloc_node and then put module in elevator_alloc;
fail to call elevator_init_fn and then put module again in elevator_init.
Remove elevator_put invoking in error path of elevator_alloc to avoid
double release issue.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
hctx->tags has to be set as NULL in case that it is to be unmapped
no matter if set->tags[hctx->queue_num] is NULL or not in blk_mq_map_swqueue()
because shared tags can be freed already from another request queue.
The same situation has to be considered during handling CPU online too.
Unmapped hw queue can be remapped after CPU topo is changed, so we need
to allocate tags for the hw queue in blk_mq_map_swqueue(). Then tags
allocation for hw queue can be removed in hctx cpu online notifier, and it
is reasonable to do that after mapping is updated.
Cc: <stable@vger.kernel.org>
Reported-by: Dongsu Park <dongsu.park@profitbricks.com>
Tested-by: Dongsu Park <dongsu.park@profitbricks.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Firstly during CPU hotplug, even queue is freezed, timeout
handler still may come and access hctx->tags, which may cause
use after free, so this patch deactivates timeout handler
inside CPU hotplug notifier.
Secondly, tags can be shared by more than one queues, so we
have to check if the hctx has been unmapped, otherwise
still use-after-free on tags can be triggered.
Cc: <stable@vger.kernel.org>
Reported-by: Dongsu Park <dongsu.park@profitbricks.com>
Tested-by: Dongsu Park <dongsu.park@profitbricks.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Commit 889fa31f00b2 was a bit too eager in reducing the loop count,
so we ended up missing queues in some configurations. Ensure that
our division rounds up, so that's not the case.
Reported-by: Guenter Roeck <linux@roeck-us.net>
Fixes: 889fa31f00b2 ("blk-mq: reduce unnecessary software queue looping")
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull block layer core bits from Jens Axboe:
"This is the core pull request for 4.1. Not a lot of stuff in here for
this round, mostly little fixes or optimizations. This pull request
contains:
- An optimization that speeds up queue runs on blk-mq, especially for
the case where there's a large difference between nr_cpu_ids and
the actual mapped software queues on a hardware queue. From Chong
Yuan.
- Honor node local allocations for requests on legacy devices. From
David Rientjes.
- Cleanup of blk_mq_rq_to_pdu() from me.
- exit_aio() fixup from me, greatly speeding up exiting multiple IO
contexts off exit_group(). For my particular test case, fio exit
took ~6 seconds. A typical case of both exposing RCU grace periods
to user space, and serializing exit of them.
- Make blk_mq_queue_enter() honor the gfp mask passed in, so we only
wait if __GFP_WAIT is set. From Keith Busch.
- blk-mq exports and two added helpers from Mike Snitzer, which will
be used by the dm-mq code.
- Cleanups of blk-mq queue init from Wei Fang and Xiaoguang Wang"
* 'for-4.1/core' of git://git.kernel.dk/linux-block:
blk-mq: reduce unnecessary software queue looping
aio: fix serial draining in exit_aio()
blk-mq: cleanup blk_mq_rq_to_pdu()
blk-mq: put blk_queue_rq_timeout together in blk_mq_init_queue()
block: remove redundant check about 'set->nr_hw_queues' in blk_mq_alloc_tag_set()
block: allocate request memory local to request queue
blk-mq: don't wait in blk_mq_queue_enter() if __GFP_WAIT isn't set
blk-mq: export blk_mq_run_hw_queues
blk-mq: add blk_mq_init_allocated_queue and export blk_mq_register_disk
In flush_busy_ctxs() and blk_mq_hctx_has_pending(), regardless of how many
ctxs assigned to one hctx, they will all loop hctx->ctx_map.map_size
times. Here hctx->ctx_map.map_size is a const ALIGN(nr_cpu_ids, 8) / 8.
Especially, flush_busy_ctxs() is in hot code path. And it's unnecessary.
Change ->map_size to contain the actually mapped software queues, so we
only loop for as many iterations as we have to.
And remove cpumask setting and nr_ctx count in blk_mq_init_cpu_queues()
since they are all re-done in blk_mq_map_swqueue().
blk_mq_map_swqueue().
Signed-off-by: Chong Yuan <chong.yuan@memblaze.com>
Reviewed-by: Wenbo Wang <wenbo.wang@memblaze.com>
Updated by me for formatting and commenting.
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull vfs update from Al Viro:
"Part one:
- struct filename-related cleanups
- saner iov_iter_init() replacements (and switching the syscalls to
use of those)
- ntfs switch to ->write_iter() (Anton)
- aio cleanups and splitting iocb into common and async parts
(Christoph)
- assorted fixes (me, bfields, Andrew Elble)
There's a lot more, including the completion of switchover to
->{read,write}_iter(), d_inode/d_backing_inode annotations, f_flags
race fixes, etc, but that goes after #for-davem merge. David has
pulled it, and once it's in I'll send the next vfs pull request"
* 'for-linus-1' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (35 commits)
sg_start_req(): use import_iovec()
sg_start_req(): make sure that there's not too many elements in iovec
blk_rq_map_user(): use import_single_range()
sg_io(): use import_iovec()
process_vm_access: switch to {compat_,}import_iovec()
switch keyctl_instantiate_key_common() to iov_iter
switch {compat_,}do_readv_writev() to {compat_,}import_iovec()
aio_setup_vectored_rw(): switch to {compat_,}import_iovec()
vmsplice_to_user(): switch to import_iovec()
kill aio_setup_single_vector()
aio: simplify arguments of aio_setup_..._rw()
aio: lift iov_iter_init() into aio_setup_..._rw()
lift iov_iter into {compat_,}do_readv_writev()
NFS: fix BUG() crash in notify_change() with patch to chown_common()
dcache: return -ESTALE not -EBUSY on distributed fs race
NTFS: Version 2.1.32 - Update file write from aio_write to write_iter.
VFS: Add iov_iter_fault_in_multipages_readable()
drop bogus check in file_open_root()
switch security_inode_getattr() to struct path *
constify tomoyo_realpath_from_path()
...
