There are several reasons for removing the check:
1) blk_mq_hw_queue_mapped() returns true always now since each hctx
may be mapped by one CPU at least
2) when there isn't any online CPU mapped to this hctx, there won't
be any IO queued to this CPU, blk_mq_run_hw_queue() only runs queue
if there is IO queued to this hctx
3) If __blk_mq_delay_run_hw_queue() is called by blk_mq_delay_run_hw_queue(),
which is run from blk_mq_dispatch_rq_list() or scsi_mq_get_budget(), and
the hctx to be handled has to be mapped.
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
No driver uses this interface any more, so remove it.
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch introduces helper of blk_mq_hw_queue_first_cpu() for
figuring out the hctx's first cpu, and code duplication can be
avoided.
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch figures out the final selected CPU, then writes
it to hctx->next_cpu once, then we can avoid to intermediate
next cpu observed from other dispatch paths.
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
From commit 20e4d81393 (blk-mq: simplify queue mapping & schedule
with each possisble CPU), one hctx can be mapped from all offline CPUs,
then hctx->next_cpu can be set as wrong.
This patch fixes this issue by making hctx->next_cpu pointing to the
first CPU in hctx->cpumask if all CPUs in hctx->cpumask are offline.
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Fixes: 20e4d81393 ("blk-mq: simplify queue mapping & schedule with each possisble CPU")
Cc: stable@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch orders getting budget and driver tag by making sure to acquire
driver tag after budget is got, this way can help to avoid the following
race:
1) before dispatch request from scheduler queue, get one budget first, then
dequeue a request, call it request A.
2) in another IO path for dispatching request B which is from hctx->dispatch,
driver tag is got, then try to get budget in blk_mq_dispatch_rq_list(),
unfortunately the budget is held by request A.
3) meantime blk_mq_dispatch_rq_list() is called for dispatching request
A, and try to get driver tag first, unfortunately no driver tag is
available because the driver tag is held by request B
4) both two IO pathes can't move on, and IO stall is caused.
This issue can be observed when running dbench on USB storage.
This patch fixes this issue by always getting budget before getting
driver tag.
Cc: stable@vger.kernel.org
Fixes: de14829740 ("blk-mq: introduce .get_budget and .put_budget in blk_mq_ops")
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Omar Sandoval <osandov@fb.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since the queue flags may be changed concurrently from multiple
contexts after a queue becomes visible in sysfs, make these changes
safe by protecting these with the queue lock.
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Introduce functions that modify the queue flags and that protect
these modifications with the request queue lock. Except for moving
one wake_up_all() call from inside to outside a critical section,
this patch does not change any functionality.
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Except for changing the atomic queue flag manipulations that are
protected by the queue lock into non-atomic manipulations, this
patch does not change any functionality.
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch does not change any functionality.
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When we insert a request, we set the software queue pending bit while
holding the software queue lock. However, we clear it outside of the
lock, so it's possible that a concurrent insert could reset the bit
after we clear it but before we empty the request list. Afterwards, the
bit would still be set but the software queue wouldn't have any requests
in it, leading us to do a spurious run in the future. This is mostly a
benign/theoretical issue, but it makes the following change easier to
justify.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
__blk_mq_requeue_request() covers two cases:
- one is that the requeued request is added to hctx->dispatch, such as
blk_mq_dispatch_rq_list()
- another case is that the request is requeued to io scheduler, such as
blk_mq_requeue_request().
We should call io sched's .requeue_request callback only for the 2nd
case.
Cc: Paolo Valente <paolo.valente@linaro.org>
Cc: Omar Sandoval <osandov@fb.com>
Fixes: bd166ef183 ("blk-mq-sched: add framework for MQ capable IO schedulers")
Cc: stable@vger.kernel.org
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Acked-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This removes the dependency on interrupts to wake up task. Set task
state as TASK_RUNNING, if need_resched() returns true,
while polling for IO completion.
Earlier, polling task used to sleep, relying on interrupt to wake it up.
This made some IO take very long when interrupt-coalescing is enabled in
NVMe.
Reference:
http://lists.infradead.org/pipermail/linux-nvme/2018-February/015435.html
Changes since v2->v3:
-using __set_current_state() instead of set_current_state()
Changes since v1->v2:
-setting task state once in blk_poll, instead of multiple
callers.
Signed-off-by: Nitesh Shetty <nj.shetty@samsung.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This status is returned from driver to block layer if device related
resource is unavailable, but driver can guarantee that IO dispatch
will be triggered in future when the resource is available.
Convert some drivers to return BLK_STS_DEV_RESOURCE. Also, if driver
returns BLK_STS_RESOURCE and SCHED_RESTART is set, rerun queue after
a delay (BLK_MQ_DELAY_QUEUE) to avoid IO stalls. BLK_MQ_DELAY_QUEUE is
3 ms because both scsi-mq and nvmefc are using that magic value.
If a driver can make sure there is in-flight IO, it is safe to return
BLK_STS_DEV_RESOURCE because:
1) If all in-flight IOs complete before examining SCHED_RESTART in
blk_mq_dispatch_rq_list(), SCHED_RESTART must be cleared, so queue
is run immediately in this case by blk_mq_dispatch_rq_list();
2) if there is any in-flight IO after/when examining SCHED_RESTART
in blk_mq_dispatch_rq_list():
- if SCHED_RESTART isn't set, queue is run immediately as handled in 1)
- otherwise, this request will be dispatched after any in-flight IO is
completed via blk_mq_sched_restart()
3) if SCHED_RESTART is set concurently in context because of
BLK_STS_RESOURCE, blk_mq_delay_run_hw_queue() will cover the above two
cases and make sure IO hang can be avoided.
One invariant is that queue will be rerun if SCHED_RESTART is set.
Suggested-by: Jens Axboe <axboe@kernel.dk>
Tested-by: Laurence Oberman <loberman@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull block updates from Jens Axboe:
"This is the main pull request for block IO related changes for the
4.16 kernel. Nothing major in this pull request, but a good amount of
improvements and fixes all over the map. This contains:
- BFQ improvements, fixes, and cleanups from Angelo, Chiara, and
Paolo.
- Support for SMR zones for deadline and mq-deadline from Damien and
Christoph.
- Set of fixes for bcache by way of Michael Lyle, including fixes
from himself, Kent, Rui, Tang, and Coly.
- Series from Matias for lightnvm with fixes from Hans Holmberg,
Javier, and Matias. Mostly centered around pblk, and the removing
rrpc 1.2 in preparation for supporting 2.0.
- A couple of NVMe pull requests from Christoph. Nothing major in
here, just fixes and cleanups, and support for command tracing from
Johannes.
- Support for blk-throttle for tracking reads and writes separately.
From Joseph Qi. A few cleanups/fixes also for blk-throttle from
Weiping.
- Series from Mike Snitzer that enables dm to register its queue more
logically, something that's alwways been problematic on dm since
it's a stacked device.
- Series from Ming cleaning up some of the bio accessor use, in
preparation for supporting multipage bvecs.
- Various fixes from Ming closing up holes around queue mapping and
quiescing.
- BSD partition fix from Richard Narron, fixing a problem where we
can't mount newer (10/11) FreeBSD partitions.
- Series from Tejun reworking blk-mq timeout handling. The previous
scheme relied on atomic bits, but it had races where we would think
a request had timed out if it to reused at the wrong time.
- null_blk now supports faking timeouts, to enable us to better
exercise and test that functionality separately. From me.
- Kill the separate atomic poll bit in the request struct. After
this, we don't use the atomic bits on blk-mq anymore at all. From
me.
- sgl_alloc/free helpers from Bart.
- Heavily contended tag case scalability improvement from me.
- Various little fixes and cleanups from Arnd, Bart, Corentin,
Douglas, Eryu, Goldwyn, and myself"
* 'for-4.16/block' of git://git.kernel.dk/linux-block: (186 commits)
block: remove smart1,2.h
nvme: add tracepoint for nvme_complete_rq
nvme: add tracepoint for nvme_setup_cmd
nvme-pci: introduce RECONNECTING state to mark initializing procedure
nvme-rdma: remove redundant boolean for inline_data
nvme: don't free uuid pointer before printing it
nvme-pci: Suspend queues after deleting them
bsg: use pr_debug instead of hand crafted macros
blk-mq-debugfs: don't allow write on attributes with seq_operations set
nvme-pci: Fix queue double allocations
block: Set BIO_TRACE_COMPLETION on new bio during split
blk-throttle: use queue_is_rq_based
block: Remove kblockd_schedule_delayed_work{,_on}()
blk-mq: Avoid that blk_mq_delay_run_hw_queue() introduces unintended delays
blk-mq: Rename blk_mq_request_direct_issue() into blk_mq_request_issue_directly()
lib/scatterlist: Fix chaining support in sgl_alloc_order()
blk-throttle: track read and write request individually
block: add bdev_read_only() checks to common helpers
block: fail op_is_write() requests to read-only partitions
blk-throttle: export io_serviced_recursive, io_service_bytes_recursive
...
Make sure that calling blk_mq_run_hw_queue() or
blk_mq_kick_requeue_list() triggers a queue run without delay even
if blk_mq_delay_run_hw_queue() has been called recently and if its
delay has not yet expired.
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Most blk-mq functions have a name that follows the pattern blk_mq_${action}.
However, the function name blk_mq_request_direct_issue is an exception.
Hence rename this function. This patch does not change any functionality.
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If we run into blk_mq_request_direct_issue(), when queue is busy, we
don't want to dispatch this request into hctx->dispatch_list, and
what we need to do is to return the queue busy info to caller, so
that caller can deal with it well.
Fixes: 396eaf21ee ("blk-mq: improve DM's blk-mq IO merging via blk_insert_cloned_request feedback")
Reported-by: Laurence Oberman <loberman@redhat.com>
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After commit:
923218f616 ("blk-mq: don't allocate driver tag upfront for flush rq")
we no longer use the 'can_block' argument in
blk_mq_sched_insert_request(). Kill it.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Added actual commit message as to why it's being removed.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_insert_cloned_request() is called in the fast path of a dm-rq driver
(e.g. blk-mq request-based DM mpath). blk_insert_cloned_request() uses
blk_mq_request_bypass_insert() to directly append the request to the
blk-mq hctx->dispatch_list of the underlying queue.
1) This way isn't efficient enough because the hctx spinlock is always
used.
2) With blk_insert_cloned_request(), we completely bypass underlying
queue's elevator and depend on the upper-level dm-rq driver's elevator
to schedule IO. But dm-rq currently can't get the underlying queue's
dispatch feedback at all. Without knowing whether a request was issued
or not (e.g. due to underlying queue being busy) the dm-rq elevator will
not be able to provide effective IO merging (as a side-effect of dm-rq
currently blindly destaging a request from its elevator only to requeue
it after a delay, which kills any opportunity for merging). This
obviously causes very bad sequential IO performance.
Fix this by updating blk_insert_cloned_request() to use
blk_mq_request_direct_issue(). blk_mq_request_direct_issue() allows a
request to be issued directly to the underlying queue and returns the
dispatch feedback (blk_status_t). If blk_mq_request_direct_issue()
returns BLK_SYS_RESOURCE the dm-rq driver will now use DM_MAPIO_REQUEUE
to _not_ destage the request. Whereby preserving the opportunity to
merge IO.
With this, request-based DM's blk-mq sequential IO performance is vastly
improved (as much as 3X in mpath/virtio-scsi testing).
Signed-off-by: Ming Lei <ming.lei@redhat.com>
[blk-mq.c changes heavily influenced by Ming Lei's initial solution, but
they were refactored to make them less fragile and easier to read/review]
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
No functional change. Just makes code flow more logically.
In following commit, __blk_mq_try_issue_directly() will be used to
return the dispatch result (blk_status_t) to DM. DM needs this
information to improve IO merging.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We know this WARN_ON is harmless and in reality it may be trigged,
so convert it to printk() and dump_stack() to avoid to confusing
people.
Also add comment about two releated races here.
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "jianchao.wang" <jianchao.w.wang@oracle.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When hctx->next_cpu is set from possible online CPUs, there is one
race in which hctx->next_cpu may be set as >= nr_cpu_ids, and finally
break workqueue.
The race can be triggered in the following two sitations:
1) when one CPU is becoming DEAD, blk_mq_hctx_notify_dead() is called
to dispatch requests from the DEAD cpu context, but at that
time, this DEAD CPU has been cleared from 'cpu_online_mask', so all
CPUs in hctx->cpumask may become offline, and cause hctx->next_cpu set
a bad value.
2) blk_mq_delay_run_hw_queue() is called from CPU B, and found the queue
should be run on the other CPU A, then CPU A may become offline at the
same time and all CPUs in hctx->cpumask become offline.
This patch deals with this issue by re-selecting next CPU, and making
sure it is set correctly.
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Reported-by: "jianchao.wang" <jianchao.w.wang@oracle.com>
Tested-by: "jianchao.wang" <jianchao.w.wang@oracle.com>
Fixes: 20e4d81393 ("blk-mq: simplify queue mapping & schedule with each possisble CPU")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A previous commit moved the clearing of rq->rq_flags later,
but we may have already set RQF_MQ_INFLIGHT when that happens.
Ensure that we correctly initialize rq->rq_flags to the
right value.
This is based on an original fix by Ming, just rewritten to not
require a conditional.
Fixes: 7c3fb70f03 ("block: rearrange a few request fields for better cache layout")
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The previous patch assigns interrupt vectors to all possible CPUs, so
now hctx can be mapped to possible CPUs, this patch applies this fact
to simplify queue mapping & schedule so that we don't need to handle
CPU hotplug for dealing with physical CPU plug & unplug. With this
simplication, we can work well on physical CPU plug & unplug, which
is a normal use case for VM at least.
Make sure we allocate blk_mq_ctx structures for all possible CPUs, and
set hctx->numa_node for possible CPUs which are mapped to this hctx. And
only choose the online CPUs for schedule.
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Stefan Haberland <sth@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Fixes: 4b855ad371 ("blk-mq: Create hctx for each present CPU")
(merged the three into one because any single one may not work, and fix
selecting online CPUs for scheduler)
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch does not change any functionality but makes the
blk_mq_mark_tag_wait() code slightly easier to read.
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move completion related items (like the call single data) near the
end of the struct, instead of mixing them in with the initial
queueing related fields.
Move queuelist below the bio structures. Then we have all
queueing related bits in the first cache line.
This yields a 1.5-2% increase in IOPS for a null_blk test, both for
sync and for high thread count access. Sync test goes form 975K to
992K, 32-thread case from 20.8M to 21.2M IOPS.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We only have one atomic flag left. Instead of using an entire
unsigned long for that, steal the bottom bit of the deadline
field that we already reserved.
Remove ->atomic_flags, since it's now unused.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We reduce the resolution of request expiry, but since we're already
using jiffies for this where resolution depends on the kernel
configuration and since the timeout resolution is coarse anyway,
that should be fine.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't need this to be an atomic flag, it can be a regular
flag. We either end up on the same CPU for the polling, in which
case the state is sane, or we did the sleep which would imply
the needed barrier to ensure we see the right state.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It is nontrivial to derive from the blk-mq source code when
blk_mq_tags.active_queues is decremented. Hence add a comment that
explains this.
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Change "nedeing" into "needing" and "caes" into "cases".
Fixes: commit f906a6a0f4 ("blk-mq: improve tag waiting setup for non-shared tags")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In some stupider versions of gcc, it complains:
block/blk-mq.c: In function ‘blk_mq_complete_request’:
./include/linux/srcu.h:175:2: warning: ‘srcu_idx’ may be used uninitialized in this function [-Wmaybe-uninitialized]
__srcu_read_unlock(sp, idx);
^
block/blk-mq.c:620:6: note: ‘srcu_idx’ was declared here
int srcu_idx;
^
which is completely bogus, since we only use srcu_idx when
hctx->flags & BLK_MQ_F_BLOCKING is set, and that's the case where
hctx_lock() has initialized it.
Just set it to '0' in the normal path in hctx_lock() to silence
this annoying warning.
Fixes: 04ced159ce ("blk-mq: move hctx lock/unlock into a helper")
Fixes: 5197c05e16 ("blk-mq: protect completion path with RCU")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The RCU protection has been expanded to cover both queueing and
completion paths making ->queue_rq_srcu a misnomer. Rename it to
->srcu as suggested by Bart.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After the recent updates to use generation number and state based
synchronization, we can easily replace REQ_ATOM_STARTED usages by
adding an extra state to distinguish completed but not yet freed
state.
Add MQ_RQ_COMPLETE and replace REQ_ATOM_STARTED usages with
blk_mq_rq_state() tests. REQ_ATOM_STARTED no longer has any users
left and is removed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After the recent updates to use generation number and state based
synchronization, blk-mq no longer depends on REQ_ATOM_COMPLETE except
to avoid firing the same timeout multiple times.
Remove all REQ_ATOM_COMPLETE usages and use a new rq_flags flag
RQF_MQ_TIMEOUT_EXPIRED to avoid firing the same timeout multiple
times. This removes atomic bitops from hot paths too.
v2: Removed blk_clear_rq_complete() from blk_mq_rq_timed_out().
v3: Added RQF_MQ_TIMEOUT_EXPIRED flag.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "jianchao.wang" <jianchao.w.wang@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
With issue/complete and timeout paths now using the generation number
and state based synchronization, blk_abort_request() is the only one
which depends on REQ_ATOM_COMPLETE for arbitrating completion.
There's no reason for blk_abort_request() to be a completely separate
path. This patch makes blk_abort_request() piggyback on the timeout
path instead of trying to terminate the request directly.
This removes the last dependency on REQ_ATOM_COMPLETE in blk-mq.
Note that this makes blk_abort_request() asynchronous - it initiates
abortion but the actual termination will happen after a short while,
even when the caller owns the request. AFAICS, SCSI and ATA should be
fine with that and I think mtip32xx and dasd should be safe but not
completely sure. It'd be great if people who know the drivers take a
look.
v2: - Add comment explaining the lack of synchronization around
->deadline update as requested by Bart.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Asai Thambi SP <asamymuthupa@micron.com>
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Cc: Jan Hoeppner <hoeppner@linux.vnet.ibm.com>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_check_inflight() and blk_mq_poll_hybrid_sleep() test
REQ_ATOM_COMPLETE to determine the request state. Both uses are
speculative and we can test REQ_ATOM_STARTED and blk_mq_rq_state() for
equivalent results. Replace the tests. This will allow removing
REQ_ATOM_COMPLETE usages from blk-mq.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, blk-mq timeout path synchronizes against the usual
issue/completion path using a complex scheme involving atomic
bitflags, REQ_ATOM_*, memory barriers and subtle memory coherence
rules. Unfortunately, it contains quite a few holes.
There's a complex dancing around REQ_ATOM_STARTED and
REQ_ATOM_COMPLETE between issue/completion and timeout paths; however,
they don't have a synchronization point across request recycle
instances and it isn't clear what the barriers add.
blk_mq_check_expired() can easily read STARTED from N-2'th iteration,
deadline from N-1'th, blk_mark_rq_complete() against Nth instance.
In fact, it's pretty easy to make blk_mq_check_expired() terminate a
later instance of a request. If we induce 5 sec delay before
time_after_eq() test in blk_mq_check_expired(), shorten the timeout to
2s, and issue back-to-back large IOs, blk-mq starts timing out
requests spuriously pretty quickly. Nothing actually timed out. It
just made the call on a recycle instance of a request and then
terminated a later instance long after the original instance finished.
The scenario isn't theoretical either.
This patch replaces the broken synchronization mechanism with a RCU
and generation number based one.
1. Each request has a u64 generation + state value, which can be
updated only by the request owner. Whenever a request becomes
in-flight, the generation number gets bumped up too. This provides
the basis for the timeout path to distinguish different recycle
instances of the request.
Also, marking a request in-flight and setting its deadline are
protected with a seqcount so that the timeout path can fetch both
values coherently.
2. The timeout path fetches the generation, state and deadline. If
the verdict is timeout, it records the generation into a dedicated
request abortion field and does RCU wait.
3. The completion path is also protected by RCU (from the previous
patch) and checks whether the current generation number and state
match the abortion field. If so, it skips completion.
4. The timeout path, after RCU wait, scans requests again and
terminates the ones whose generation and state still match the ones
requested for abortion.
By now, the timeout path knows that either the generation number
and state changed if it lost the race or the completion will yield
to it and can safely timeout the request.
While it's more lines of code, it's conceptually simpler, doesn't
depend on direct use of subtle memory ordering or coherence, and
hopefully doesn't terminate the wrong instance.
While this change makes REQ_ATOM_COMPLETE synchronization unnecessary
between issue/complete and timeout paths, REQ_ATOM_COMPLETE isn't
removed yet as it's still used in other places. Future patches will
move all state tracking to the new mechanism and remove all bitops in
the hot paths.
Note that this patch adds a comment explaining a race condition in
BLK_EH_RESET_TIMER path. The race has always been there and this
patch doesn't change it. It's just documenting the existing race.
v2: - Fixed BLK_EH_RESET_TIMER handling as pointed out by Jianchao.
- s/request->gstate_seqc/request->gstate_seq/ as suggested by Peter.
- READ_ONCE() added in blk_mq_rq_update_state() as suggested by Peter.
v3: - Fixed possible extended seqcount / u64_stats_sync read looping
spotted by Peter.
- MQ_RQ_IDLE was incorrectly being set in complete_request instead
of free_request. Fixed.
v4: - Rebased on top of hctx_lock() refactoring patch.
- Added comment explaining the use of hctx_lock() in completion path.
v5: - Added comments requested by Bart.
- Note the addition of BLK_EH_RESET_TIMER race condition in the
commit message.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "jianchao.wang" <jianchao.w.wang@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, blk-mq protects only the issue path with RCU. This patch
puts the completion path under the same RCU protection. This will be
used to synchronize issue/completion against timeout by later patches,
which will also add the comments.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move the RCU vs SRCU logic into lock/unlock helpers, which makes
the actual functional bits within the locked region much easier
to read.
tj: Reordered in front of timeout revamp patches and added the missing
blk_mq_run_hw_queue() conversion.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In both elevator_switch_mq() and blk_mq_update_nr_hw_queues(), sched tags
can be allocated, and q->nr_hw_queue is used, and race is inevitable, for
example: blk_mq_init_sched() may trigger use-after-free on hctx, which is
freed in blk_mq_realloc_hw_ctxs() when nr_hw_queues is decreased.
This patch fixes the race be holding q->sysfs_lock.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reported-by: Yi Zhang <yi.zhang@redhat.com>
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Dispatch may still be in-progress after queue is frozen, so we have to
quiesce queue before switching IO scheduler and updating nr_requests.
Also when switching io schedulers, blk_mq_run_hw_queue() may still be
called somewhere(such as from nvme_reset_work()), and io scheduler's
per-hctx data may not be setup yet, so cause oops even inside
blk_mq_hctx_has_pending(), such as it can be run just between:
ret = e->ops.mq.init_sched(q, e);
AND
ret = e->ops.mq.init_hctx(hctx, i)
inside blk_mq_init_sched().
This reverts commit 7a148c2fcff8330(block: don't call blk_mq_quiesce_queue()
after queue is frozen) basically, and makes sure blk_mq_hctx_has_pending
won't be called if queue is quiesced.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Fixes: 7a148c2fcff83309(block: don't call blk_mq_quiesce_queue() after queue is frozen)
Reported-by: Yi Zhang <yi.zhang@redhat.com>
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now we track legacy requests with .q_usage_counter in commit 055f6e18e0
("block: Make q_usage_counter also track legacy requests"), but that
commit never runs and drains legacy queue before waiting for this counter
becoming zero, then IO hang is caused in the test of pulling disk during IO.
This patch fixes the issue by draining requests before waiting for
q_usage_counter becoming zero, both Mauricio and chenxiang reported this
issue, and observed that it can be fixed by this patch.
Link: https://marc.info/?l=linux-block&m=151192424731797&w=2
Fixes: 055f6e18e08f("block: Make q_usage_counter also track legacy requests")
Cc: Wen Xiong <wenxiong@us.ibm.com>
Tested-by: "chenxiang (M)" <chenxiang66@hisilicon.com>
Tested-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that we have a NUMA-aware version of kmalloc_array() we can use it
instead of kmalloc_node() without an overflow check in the size
calculation.
Link: http://lkml.kernel.org/r/20170927082038.3782-3-jthumshirn@suse.de
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Doug Ledford <dledford@redhat.com>
Cc: Hal Rosenstock <hal.rosenstock@gmail.com>
Cc: Mike Marciniszyn <infinipath@intel.com>
Cc: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Cc: Sean Hefty <sean.hefty@intel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If we run out of driver tags, we currently treat shared and non-shared
tags the same - both cases hook into the tag waitqueue. This is a bit
more costly than it needs to be on unshared tags, since we have to both
grab the hctx lock, and the waitqueue lock (and disable interrupts).
For the non-shared case, we can simply mark the queue as needing a
restart.
Split blk_mq_dispatch_wait_add() to account for both cases, and
rename it to blk_mq_mark_tag_wait() to better reflect what it
does now.
Without this patch, shared and non-shared performance is about the same
with 4 fio thread hammering on a single null_blk device (~410K, at 75%
sys). With the patch, the shared case is the same, but the non-shared
tags case runs at 431K at 71% sys.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently we are inconsistent in when we decide to run the queue. Using
blk_mq_run_hw_queues() we check if the hctx has pending IO before
running it, but we don't do that from the individual queue run function,
blk_mq_run_hw_queue(). This results in a lot of extra and pointless
queue runs, potentially, on flush requests and (much worse) on tag
starvation situations. This is observable just looking at top output,
with lots of kworkers active. For the !async runs, it just adds to the
CPU overhead of blk-mq.
Move the has-pending check into the run function instead of having
callers do it.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Several block layer and NVMe core functions accept a combination
of BLK_MQ_REQ_* flags through the 'flags' argument but there is
no verification at compile time whether the right type of block
layer flags is passed. Make it possible for sparse to verify this.
This patch does not change any functionality.
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Cc: linux-nvme@lists.infradead.org
Cc: Christoph Hellwig <hch@lst.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The contexts from which a SCSI device can be quiesced or resumed are:
* Writing into /sys/class/scsi_device/*/device/state.
* SCSI parallel (SPI) domain validation.
* The SCSI device power management methods. See also scsi_bus_pm_ops.
It is essential during suspend and resume that neither the filesystem
state nor the filesystem metadata in RAM changes. This is why while
the hibernation image is being written or restored that SCSI devices
are quiesced. The SCSI core quiesces devices through scsi_device_quiesce()
and scsi_device_resume(). In the SDEV_QUIESCE state execution of
non-preempt requests is deferred. This is realized by returning
BLKPREP_DEFER from inside scsi_prep_state_check() for quiesced SCSI
devices. Avoid that a full queue prevents power management requests
to be submitted by deferring allocation of non-preempt requests for
devices in the quiesced state. This patch has been tested by running
the following commands and by verifying that after each resume the
fio job was still running:
for ((i=0; i<10; i++)); do
(
cd /sys/block/md0/md &&
while true; do
[ "$(<sync_action)" = "idle" ] && echo check > sync_action
sleep 1
done
) &
pids=($!)
for d in /sys/class/block/sd*[a-z]; do
bdev=${d#/sys/class/block/}
hcil=$(readlink "$d/device")
hcil=${hcil#../../../}
echo 4 > "$d/queue/nr_requests"
echo 1 > "/sys/class/scsi_device/$hcil/device/queue_depth"
fio --name="$bdev" --filename="/dev/$bdev" --buffered=0 --bs=512 \
--rw=randread --ioengine=libaio --numjobs=4 --iodepth=16 \
--iodepth_batch=1 --thread --loops=$((2**31)) &
pids+=($!)
done
sleep 1
echo "$(date) Hibernating ..." >>hibernate-test-log.txt
systemctl hibernate
sleep 10
kill "${pids[@]}"
echo idle > /sys/block/md0/md/sync_action
wait
echo "$(date) Done." >>hibernate-test-log.txt
done
Reported-by: Oleksandr Natalenko <oleksandr@natalenko.name>
References: "I/O hangs after resuming from suspend-to-ram" (https://marc.info/?l=linux-block&m=150340235201348).
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Martin Steigerwald <martin@lichtvoll.de>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Set RQF_PREEMPT if BLK_MQ_REQ_PREEMPT is passed to
blk_get_request_flags().
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Martin Steigerwald <martin@lichtvoll.de>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch makes it possible to pause request allocation for
the legacy block layer by calling blk_mq_freeze_queue() and
blk_mq_unfreeze_queue().
Signed-off-by: Ming Lei <ming.lei@redhat.com>
[ bvanassche: Combined two patches into one, edited a comment and made sure
REQ_NOWAIT is handled properly in blk_old_get_request() ]
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Martin Steigerwald <martin@lichtvoll.de>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch attempts to make the case of hctx re-running on driver tag
failure more robust. Without this patch, it's pretty easy to trigger a
stall condition with shared tags. An example is using null_blk like
this:
modprobe null_blk queue_mode=2 nr_devices=4 shared_tags=1 submit_queues=1 hw_queue_depth=1
which sets up 4 devices, sharing the same tag set with a depth of 1.
Running a fio job ala:
[global]
bs=4k
rw=randread
norandommap
direct=1
ioengine=libaio
iodepth=4
[nullb0]
filename=/dev/nullb0
[nullb1]
filename=/dev/nullb1
[nullb2]
filename=/dev/nullb2
[nullb3]
filename=/dev/nullb3
will inevitably end with one or more threads being stuck waiting for a
scheduler tag. That IO is then stuck forever, until someone else
triggers a run of the queue.
Ensure that we always re-run the hardware queue, if the driver tag we
were waiting for got freed before we added our leftover request entries
back on the dispatch list.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Tested-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We have to put the driver tag if dispatch budget can't be got, otherwise
it might cause IO deadlock, especially in case that size of tags is very
small.
Fixes: de1482974080(blk-mq: introduce .get_budget and .put_budget in blk_mq_ops)
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The idea behind it is simple:
1) for none scheduler, driver tag has to be borrowed for flush rq,
otherwise we may run out of tag, and that causes an IO hang. And
get/put driver tag is actually noop for none, so reordering tags
isn't necessary at all.
2) for a real I/O scheduler, we need not allocate a driver tag upfront
for flush rq. It works just fine to follow the same approach as
normal requests: allocate driver tag for each rq just before calling
->queue_rq().
One driver visible change is that the driver tag isn't shared in the
flush request sequence. That won't be a problem, since we always do that
in legacy path.
Then flush rq need not be treated specially wrt. get/put driver tag.
This cleans up the code - for instance, reorder_tags_to_front() can be
removed, and we needn't worry about request ordering in dispatch list
for avoiding I/O deadlock.
Also we have to put the driver tag before requeueing.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We need this helper to put the driver tag for flush rq, since we will
not share tag in the flush request sequence in the following patch
in case that I/O scheduler is applied.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Block flush need this function without running the queue, so add a
parameter controlling whether we run it or not.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When freeing the driver tag of the next rq with an I/O scheduler
configured, we get the first entry of the list. However, this can
race with requeue of a request, and we end up getting the wrong request
from the head of the list. Free the driver tag of next rq before the
failed one is requeued in the failure branch of queue_rq callback.
Signed-off-by: Jianchao Wang <jianchao.w.wang@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It is enough to just check if we can get the budget via .get_budget().
And we don't need to deal with device state change in .get_budget().
For SCSI, one issue to be fixed is that we have to call
scsi_mq_uninit_cmd() to free allocated ressources if SCSI device fails
to handle the request. And it isn't enough to simply call
blk_mq_end_request() to do that if this request is marked as
RQF_DONTPREP.
Fixes: 0df21c86bdbf(scsi: implement .get_budget and .put_budget for blk-mq)
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_get_tag() can modify data->ctx. This means that in the
error path of blk_mq_get_request() data->ctx should be passed to
blk_mq_put_ctx() instead of local_ctx. Note: since blk_mq_put_ctx()
ignores its argument, this patch does not change any functionality.
References: commit 1ad43c0078 ("blk-mq: don't leak preempt counter/q_usage_counter when allocating rq failed")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
if blk-mq use "none" io scheduler, nr_request get a wrong value when
input a number > tag_set->queue_depth. blk_mq_tag_update_depth will get
the smaller one min(nr, set->queue_depth), and then q->nr_request get a
wrong value.
Reproduce:
echo none > /sys/block/nvme0n1/queue/scheduler
echo 1000000 > /sys/block/nvme0n1/queue/nr_requests
cat /sys/block/nvme0n1/queue/nr_requests
1000000
Signed-off-by: weiping zhang <zhangweiping@didichuxing.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
That we we can also poll non blk-mq queues. Mostly needed for
the NVMe multipath code, but could also be useful elsewhere.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
SCSI restarts its queue in scsi_end_request() automatically, so we don't
need to handle this case in blk-mq.
Especailly any request won't be dequeued in this case, we needn't to
worry about IO hang caused by restart vs. dispatch.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
SCSI devices use host-wide tagset, and the shared driver tag space is
often quite big. However, there is also a queue depth for each lun(
.cmd_per_lun), which is often small, for example, on both lpfc and
qla2xxx, .cmd_per_lun is just 3.
So lots of requests may stay in sw queue, and we always flush all
belonging to same hw queue and dispatch them all to driver.
Unfortunately it is easy to cause queue busy because of the small
.cmd_per_lun. Once these requests are flushed out, they have to stay in
hctx->dispatch, and no bio merge can happen on these requests, and
sequential IO performance is harmed.
This patch introduces blk_mq_dequeue_from_ctx for dequeuing a request
from a sw queue, so that we can dispatch them in scheduler's way. We can
then avoid dequeueing too many requests from sw queue, since we don't
flush ->dispatch completely.
This patch improves dispatching from sw queue by using the .get_budget
and .put_budget callbacks.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For SCSI devices, there is often a per-request-queue depth, which needs
to be respected before queuing one request.
Currently blk-mq always dequeues the request first, then calls
.queue_rq() to dispatch the request to lld. One obvious issue with this
approach is that I/O merging may not be successful, because when the
per-request-queue depth can't be respected, .queue_rq() has to return
BLK_STS_RESOURCE, and then this request has to stay in hctx->dispatch
list. This means it never gets a chance to be merged with other IO.
This patch introduces .get_budget and .put_budget callback in blk_mq_ops,
then we can try to get reserved budget first before dequeuing request.
If the budget for queueing I/O can't be satisfied, we don't need to
dequeue request at all. Hence the request can be left in the IO
scheduler queue, for more merging opportunities.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Legacy queue sets request's request_list, mq doesn't. This makes mq does
the same thing, so we can find cgroup of a request. Note, we really
only use blkg field of request_list, it's pointless to allocate mempool
for request_list in mq case.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For memory ordering guarantees on stores, we need to ensure that
these two bits share the same byte of storage in the unsigned
long. Add a comment as to why, and a BUILD_BUG_ON() to ensure that
we don't violate this requirement.
Suggested-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Attempt to untangle the ordering in blk-mq. The patch introducing the
single smp_mb__before_atomic() is obviously broken in that it doesn't
clearly specify a pairing barrier and an obtained guarantee.
The comment is further misleading in that it hints that the
deadline store and the COMPLETE store also need to be ordered, but
AFAICT there is no such dependency. However what does appear to be
important is the clear happening _after_ the store, and that worked by
pure accident.
This clarifies blk_mq_start_request() -- we should not get there with
STARTING set -- this simplifies the code and makes the barrier usage
sane (the old code could be read to allow not having _any_ atomic after
the barrier, in which case the barrier hasn't got anything to order). We
then also introduce the missing pairing barrier for it.
Also down-grade the barrier to smp_wmb(), this is cheaper for
PowerPC/ARM and doesn't cost anything extra on x86.
And it documents the STARTING vs COMPLETE ordering. Although I've not
been entirely successful in reverse engineering the blk-mq state
machine so there might still be more funnies around timeout vs
requeue.
If I got anything wrong, feel free to educate me by adding comments to
clarify things ;-)
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Ming Lei <tom.leiming@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Andrea Parri <parri.andrea@gmail.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Bart Van Assche <bart.vanassche@wdc.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Fixes: 538b753418 ("blk-mq: request deadline must be visible before marking rq as started")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For some reason, the laptop mode IO completion notified was never wired
up for blk-mq. Ensure that we trigger the callback appropriately, to arm
the laptop mode flush timer.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
since 9bddeb2a5b "blk-mq: make per-sw-queue bio merge as default .bio_merge"
there is no caller for this function.
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: weiping zhang <zhangweiping@didichuxing.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A NULL pointer crash was reported for the case of having the BFQ IO
scheduler attached to the underlying blk-mq paths of a DM multipath
device. The crash occured in blk_mq_sched_insert_request()'s call to
e->type->ops.mq.insert_requests().
Paolo Valente correctly summarized why the crash occured with:
"the call chain (dm_mq_queue_rq -> map_request -> setup_clone ->
blk_rq_prep_clone) creates a cloned request without invoking
e->type->ops.mq.prepare_request for the target elevator e. The cloned
request is therefore not initialized for the scheduler, but it is
however inserted into the scheduler by blk_mq_sched_insert_request."
All said, a request-based DM multipath device's IO scheduler should be
the only one used -- when the original requests are issued to the
underlying paths as cloned requests they are inserted directly in the
underlying dispatch queue(s) rather than through an additional elevator.
But commit bd166ef18 ("blk-mq-sched: add framework for MQ capable IO
schedulers") switched blk_insert_cloned_request() from using
blk_mq_insert_request() to blk_mq_sched_insert_request(). Which
incorrectly added elevator machinery into a call chain that isn't
supposed to have any.
To fix this introduce a blk-mq private blk_mq_request_bypass_insert()
that blk_insert_cloned_request() calls to insert the request without
involving any elevator that may be attached to the cloned request's
request_queue.
Fixes: bd166ef183 ("blk-mq-sched: add framework for MQ capable IO schedulers")
Cc: stable@vger.kernel.org
Reported-by: Bart Van Assche <Bart.VanAssche@wdc.com>
Tested-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull block layer updates from Jens Axboe:
"This is the first pull request for 4.14, containing most of the code
changes. It's a quiet series this round, which I think we needed after
the churn of the last few series. This contains:
- Fix for a registration race in loop, from Anton Volkov.
- Overflow complaint fix from Arnd for DAC960.
- Series of drbd changes from the usual suspects.
- Conversion of the stec/skd driver to blk-mq. From Bart.
- A few BFQ improvements/fixes from Paolo.
- CFQ improvement from Ritesh, allowing idling for group idle.
- A few fixes found by Dan's smatch, courtesy of Dan.
- A warning fixup for a race between changing the IO scheduler and
device remova. From David Jeffery.
- A few nbd fixes from Josef.
- Support for cgroup info in blktrace, from Shaohua.
- Also from Shaohua, new features in the null_blk driver to allow it
to actually hold data, among other things.
- Various corner cases and error handling fixes from Weiping Zhang.
- Improvements to the IO stats tracking for blk-mq from me. Can
drastically improve performance for fast devices and/or big
machines.
- Series from Christoph removing bi_bdev as being needed for IO
submission, in preparation for nvme multipathing code.
- Series from Bart, including various cleanups and fixes for switch
fall through case complaints"
* 'for-4.14/block' of git://git.kernel.dk/linux-block: (162 commits)
kernfs: checking for IS_ERR() instead of NULL
drbd: remove BIOSET_NEED_RESCUER flag from drbd_{md_,}io_bio_set
drbd: Fix allyesconfig build, fix recent commit
drbd: switch from kmalloc() to kmalloc_array()
drbd: abort drbd_start_resync if there is no connection
drbd: move global variables to drbd namespace and make some static
drbd: rename "usermode_helper" to "drbd_usermode_helper"
drbd: fix race between handshake and admin disconnect/down
drbd: fix potential deadlock when trying to detach during handshake
drbd: A single dot should be put into a sequence.
drbd: fix rmmod cleanup, remove _all_ debugfs entries
drbd: Use setup_timer() instead of init_timer() to simplify the code.
drbd: fix potential get_ldev/put_ldev refcount imbalance during attach
drbd: new disk-option disable-write-same
drbd: Fix resource role for newly created resources in events2
drbd: mark symbols static where possible
drbd: Send P_NEG_ACK upon write error in protocol != C
drbd: add explicit plugging when submitting batches
drbd: change list_for_each_safe to while(list_first_entry_or_null)
drbd: introduce drbd_recv_header_maybe_unplug
...
Since patch "blk-mq: switch .queue_rq return value to blk_status_t"
.queue_rq() returns a BLK_STS_* value instead of a BLK_MQ_RQ_*
value. Hence refer to the former in comments about .queue_rq()
return values.
Fixes: commit 39a70c76b8 ("blk-mq: clarify dispatch may not be drained/blocked by stopping queue")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_get_request() does not release the callers queue usage counter
when allocation fails. The caller still needs to account for its own
queue usage when it is unable to allocate a request.
Fixes: 1ad43c0078 ("blk-mq: don't leak preempt counter/q_usage_counter when allocating rq failed")
Reported-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Tested-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The blk_mq_delay_kick_requeue_list() function is used by the device
mapper and only by the device mapper to rerun the queue and requeue
list after a delay. This function is called once per request that
gets requeued. Modify this function such that the queue is run once
per path change event instead of once per request that is requeued.
Fixes: commit 2849450ad3 ("blk-mq: introduce blk_mq_delay_kick_requeue_list()")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Mike Snitzer <snitzer@redhat.com>
Cc: Laurence Oberman <loberman@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Modify blk_mq_in_flight() to count both a partition and root at
the same time. Then we only have to call it once, instead of
potentially looping the tags twice.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't have to inc/dec some counter, since we can just
iterate the tags. That makes inc/dec a noop, but means we
have to iterate busy tags to get an in-flight count.
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When blk_mq_get_request() failed, preempt counter isn't
released, and blk_mq_make_request() doesn't release the counter
too.
This patch fixes the issue, and makes sure that preempt counter
is only held if rq is allocated successfully. The same policy is
applied on .q_usage_counter too.
Signed-off-by: Ming Lei <minlei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We recently had a bug in the IPR SCSI driver, where it would end up
making the SCSI mid layer run the mq hardware queue with interrupts
disabled. This isn't legal, since the software queue locking relies
on never being grabbed from interrupt context. Additionally, drivers
that set BLK_MQ_F_BLOCKING may schedule from this context.
Add a WARN_ON_ONCE() to catch bad users up front.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull more block updates from Jens Axboe:
"This is a followup for block changes, that didn't make the initial
pull request. It's a bit of a mixed bag, this contains:
- A followup pull request from Sagi for NVMe. Outside of fixups for
NVMe, it also includes a series for ensuring that we properly
quiesce hardware queues when browsing live tags.
- Set of integrity fixes from Dmitry (mostly), fixing various issues
for folks using DIF/DIX.
- Fix for a bug introduced in cciss, with the req init changes. From
Christoph.
- Fix for a bug in BFQ, from Paolo.
- Two followup fixes for lightnvm/pblk from Javier.
- Depth fix from Ming for blk-mq-sched.
- Also from Ming, performance fix for mtip32xx that was introduced
with the dynamic initialization of commands"
* 'for-linus' of git://git.kernel.dk/linux-block: (44 commits)
block: call bio_uninit in bio_endio
nvmet: avoid unneeded assignment of submit_bio return value
nvme-pci: add module parameter for io queue depth
nvme-pci: compile warnings in nvme_alloc_host_mem()
nvmet_fc: Accept variable pad lengths on Create Association LS
nvme_fc/nvmet_fc: revise Create Association descriptor length
lightnvm: pblk: remove unnecessary checks
lightnvm: pblk: control I/O flow also on tear down
cciss: initialize struct scsi_req
null_blk: fix error flow for shared tags during module_init
block: Fix __blkdev_issue_zeroout loop
nvme-rdma: unconditionally recycle the request mr
nvme: split nvme_uninit_ctrl into stop and uninit
virtio_blk: quiesce/unquiesce live IO when entering PM states
mtip32xx: quiesce request queues to make sure no submissions are inflight
nbd: quiesce request queues to make sure no submissions are inflight
nvme: kick requeue list when requeueing a request instead of when starting the queues
nvme-pci: quiesce/unquiesce admin_q instead of start/stop its hw queues
nvme-loop: quiesce/unquiesce admin_q instead of start/stop its hw queues
nvme-fc: quiesce/unquiesce admin_q instead of start/stop its hw queues
...
Pull irq updates from Thomas Gleixner:
"The irq department delivers:
- Expand the generic infrastructure handling the irq migration on CPU
hotplug and convert X86 over to it. (Thomas Gleixner)
Aside of consolidating code this is a preparatory change for:
- Finalizing the affinity management for multi-queue devices. The
main change here is to shut down interrupts which are affine to a
outgoing CPU and reenabling them when the CPU comes online again.
That avoids moving interrupts pointlessly around and breaking and
reestablishing affinities for no value. (Christoph Hellwig)
Note: This contains also the BLOCK-MQ and NVME changes which depend
on the rework of the irq core infrastructure. Jens acked them and
agreed that they should go with the irq changes.
- Consolidation of irq domain code (Marc Zyngier)
- State tracking consolidation in the core code (Jeffy Chen)
- Add debug infrastructure for hierarchical irq domains (Thomas
Gleixner)
- Infrastructure enhancement for managing generic interrupt chips via
devmem (Bartosz Golaszewski)
- Constification work all over the place (Tobias Klauser)
- Two new interrupt controller drivers for MVEBU (Thomas Petazzoni)
- The usual set of fixes, updates and enhancements all over the
place"
* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (112 commits)
irqchip/or1k-pic: Fix interrupt acknowledgement
irqchip/irq-mvebu-gicp: Allocate enough memory for spi_bitmap
irqchip/gic-v3: Fix out-of-bound access in gic_set_affinity
nvme: Allocate queues for all possible CPUs
blk-mq: Create hctx for each present CPU
blk-mq: Include all present CPUs in the default queue mapping
genirq: Avoid unnecessary low level irq function calls
genirq: Set irq masked state when initializing irq_desc
genirq/timings: Add infrastructure for estimating the next interrupt arrival time
genirq/timings: Add infrastructure to track the interrupt timings
genirq/debugfs: Remove pointless NULL pointer check
irqchip/gic-v3-its: Don't assume GICv3 hardware supports 16bit INTID
irqchip/gic-v3-its: Add ACPI NUMA node mapping
irqchip/gic-v3-its-platform-msi: Make of_device_ids const
irqchip/gic-v3-its: Make of_device_ids const
irqchip/irq-mvebu-icu: Add new driver for Marvell ICU
irqchip/irq-mvebu-gicp: Add new driver for Marvell GICP
dt-bindings/interrupt-controller: Add DT binding for the Marvell ICU
genirq/irqdomain: Remove auto-recursive hierarchy support
irqchip/MSI: Use irq_domain_update_bus_token instead of an open coded access
...
Currently all integrity prep hooks are open-coded, and if prepare fails
we ignore it's code and fail bio with EIO. Let's return real error to
upper layer, so later caller may react accordingly.
In fact no one want to use bio_integrity_prep() w/o bio_integrity_enabled,
so it is reasonable to fold it in to one function.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
[hch: merged with the latest block tree,
return bool from bio_integrity_prep]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull scheduler updates from Ingo Molnar:
"The main changes in this cycle were:
- Add the SYSTEM_SCHEDULING bootup state to move various scheduler
debug checks earlier into the bootup. This turns silent and
sporadically deadly bugs into nice, deterministic splats. Fix some
of the splats that triggered. (Thomas Gleixner)
- A round of restructuring and refactoring of the load-balancing and
topology code (Peter Zijlstra)
- Another round of consolidating ~20 of incremental scheduler code
history: this time in terms of wait-queue nomenclature. (I didn't
get much feedback on these renaming patches, and we can still
easily change any names I might have misplaced, so if anyone hates
a new name, please holler and I'll fix it.) (Ingo Molnar)
- sched/numa improvements, fixes and updates (Rik van Riel)
- Another round of x86/tsc scheduler clock code improvements, in hope
of making it more robust (Peter Zijlstra)
- Improve NOHZ behavior (Frederic Weisbecker)
- Deadline scheduler improvements and fixes (Luca Abeni, Daniel
Bristot de Oliveira)
- Simplify and optimize the topology setup code (Lauro Ramos
Venancio)
- Debloat and decouple scheduler code some more (Nicolas Pitre)
- Simplify code by making better use of llist primitives (Byungchul
Park)
- ... plus other fixes and improvements"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (103 commits)
sched/cputime: Refactor the cputime_adjust() code
sched/debug: Expose the number of RT/DL tasks that can migrate
sched/numa: Hide numa_wake_affine() from UP build
sched/fair: Remove effective_load()
sched/numa: Implement NUMA node level wake_affine()
sched/fair: Simplify wake_affine() for the single socket case
sched/numa: Override part of migrate_degrades_locality() when idle balancing
sched/rt: Move RT related code from sched/core.c to sched/rt.c
sched/deadline: Move DL related code from sched/core.c to sched/deadline.c
sched/cpuset: Only offer CONFIG_CPUSETS if SMP is enabled
sched/fair: Spare idle load balancing on nohz_full CPUs
nohz: Move idle balancer registration to the idle path
sched/loadavg: Generalize "_idle" naming to "_nohz"
sched/core: Drop the unused try_get_task_struct() helper function
sched/fair: WARN() and refuse to set buddy when !se->on_rq
sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming
sched/wait: Move bit_wait_table[] and related functionality from sched/core.c to sched/wait_bit.c
sched/wait: Split out the wait_bit*() APIs from <linux/wait.h> into <linux/wait_bit.h>
sched/wait: Re-adjust macro line continuation backslashes in <linux/wait.h>
...
Currently we only create hctx for online CPUs, which can lead to a lot
of churn due to frequent soft offline / online operations. Instead
allocate one for each present CPU to avoid this and dramatically simplify
the code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Cc: Keith Busch <keith.busch@intel.com>
Cc: linux-block@vger.kernel.org
Cc: linux-nvme@lists.infradead.org
Link: http://lkml.kernel.org/r/20170626102058.10200-3-hch@lst.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
For historical reasons we default to bouncing highmem pages for all block
queues. But the blk-mq drivers are easy to audit to ensure that we don't
need this - scsi and mtip32xx set explicit limits and everyone else doesn't
have any particular ones.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We only call blk_queue_bounce for request-based drivers, so stop messing
with it for make_request based drivers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull in the fix for shared tags, as it conflicts with the pending
changes in for-4.13/block. We already pulled in v4.12-rc5 to solve
other conflicts or get fixes that went into 4.12, so not a lot
of changes in this merge.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
hwctx's queue_num has been set prior call blk_mq_init_hctx, so no need
set it again.
Signed-off-by: weiping <zhangweiping@didichuxing.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since blk_mq_quiesce_queue_nowait() can be called from interrupt
context, make this safe. Since this function is not in the hot
path, uninline it.
Fixes: commit f4560ffe8c ("blk-mq: use QUEUE_FLAG_QUIESCED to quiesce queue")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If we have shared tags enabled, then every IO completion will trigger
a full loop of every queue belonging to a tag set, and every hardware
queue for each of those queues, even if nothing needs to be done.
This causes a massive performance regression if you have a lot of
shared devices.
Instead of doing this huge full scan on every IO, add an atomic
counter to the main queue that tracks how many hardware queues have
been marked as needing a restart. With that, we can avoid looking for
restartable queues, if we don't have to.
Max reports that this restores performance. Before this patch, 4K
IOPS was limited to 22-23K IOPS. With the patch, we are running at
950-970K IOPS.
Fixes: 6d8c6c0f97 ("blk-mq: Restart a single queue if tag sets are shared")
Reported-by: Max Gurtovoy <maxg@mellanox.com>
Tested-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Tested-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A queue must be frozen while the mapped state of a hardware queue
is changed. Additionally, any change of the mapped state is
followed by a call to blk_mq_map_swqueue() (see also
blk_mq_init_allocated_queue() and blk_mq_update_nr_hw_queues()).
Since blk_mq_map_swqueue() does not map any unmapped hardware
queue onto any software queue, no attempt will be made to run
an unmapped hardware queue. Hence issue a warning upon attempts
to run an unmapped hardware queue.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Document the locking assumptions in functions that modify
blk_mq_ctx.rq_list to make it easier for humans to verify
this code.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Initialization of blk-mq requests is a bit weird: blk_mq_rq_ctx_init()
is called after a value has been assigned to .rq_flags and .rq_flags
is initialized in __blk_mq_finish_request(). Initialize .rq_flags in
blk_mq_rq_ctx_init() instead of relying on __blk_mq_finish_request().
Moving the initialization of .rq_flags is fine because all changes
and tests of .rq_flags occur between blk_get_request() and finishing
a request.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of declaring the second argument of blk_*_get_request()
as int and passing it to functions that expect an unsigned int,
declare that second argument as unsigned int. Also because of
consistency, rename that second argument from 'rw' into 'op'.
This patch does not change any functionality.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since the srcu structure is rather large (184 bytes on an x86-64
system with kernel debugging disabled), only allocate it if needed.
Reported-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A new bio operation flag REQ_NOWAIT is introduced to identify bio's
orignating from iocb with IOCB_NOWAIT. This flag indicates
to return immediately if a request cannot be made instead
of retrying.
Stacked devices such as md (the ones with make_request_fn hooks)
currently are not supported because it may block for housekeeping.
For example, an md can have a part of the device suspended.
For this reason, only request based devices are supported.
In the future, this feature will be expanded to stacked devices
by teaching them how to handle the REQ_NOWAIT flags.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
So I've noticed a number of instances where it was not obvious from the
code whether ->task_list was for a wait-queue head or a wait-queue entry.
Furthermore, there's a number of wait-queue users where the lists are
not for 'tasks' but other entities (poll tables, etc.), in which case
the 'task_list' name is actively confusing.
To clear this all up, name the wait-queue head and entry list structure
fields unambiguously:
struct wait_queue_head::task_list => ::head
struct wait_queue_entry::task_list => ::entry
For example, this code:
rqw->wait.task_list.next != &wait->task_list
... is was pretty unclear (to me) what it's doing, while now it's written this way:
rqw->wait.head.next != &wait->entry
... which makes it pretty clear that we are iterating a list until we see the head.
Other examples are:
list_for_each_entry_safe(pos, next, &x->task_list, task_list) {
list_for_each_entry(wq, &fence->wait.task_list, task_list) {
... where it's unclear (to me) what we are iterating, and during review it's
hard to tell whether it's trying to walk a wait-queue entry (which would be
a bug), while now it's written as:
list_for_each_entry_safe(pos, next, &x->head, entry) {
list_for_each_entry(wq, &fence->wait.head, entry) {
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Rename:
wait_queue_t => wait_queue_entry_t
'wait_queue_t' was always a slight misnomer: its name implies that it's a "queue",
but in reality it's a queue *entry*. The 'real' queue is the wait queue head,
which had to carry the name.
Start sorting this out by renaming it to 'wait_queue_entry_t'.
This also allows the real structure name 'struct __wait_queue' to
lose its double underscore and become 'struct wait_queue_entry',
which is the more canonical nomenclature for such data types.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch reverts commit 2719aa217e0d02(blk-mq: don't use
sync workqueue flushing from drivers) because only
blk_mq_quiesce_queue() need the sync flush, and now
we don't need to stop queue any more, so revert it.
Also changes to cancel_delayed_work() in blk_mq_stop_hw_queue().
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
BLK_MQ_S_STOPPED may not be observed in other concurrent I/O paths,
we can't guarantee that dispatching won't happen after returning
from the APIs of stopping queue.
So clarify the fact and avoid potential misuse.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Queue can be started by other blk-mq APIs and can be used in
different cases, this limits uses of blk_mq_quiesce_queue()
if it is based on stopping queue, and make its usage very
difficult, especially users have to use the stop queue APIs
carefully for avoiding to break blk_mq_quiesce_queue().
We have applied the QUIESCED flag for draining and blocking
dispatch, so it isn't necessary to stop queue any more.
After stopping queue is removed, blk_mq_quiesce_queue() can
be used safely and easily, then users won't worry about queue
restarting during quiescing at all.
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Actually what we want to get from blk_mq_quiesce_queue()
isn't only to wait for completion of all ongoing .queue_rq().
In the typical context of canceling requests, we need to
make sure that the following is done in the dispatch path
before starting to cancel requests:
- failed dispatched request is finished
- busy dispatched request is requeued, and the STARTED
flag is cleared
So update comment to keep code, doc and our expection consistent.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It is required that no dispatch can happen any more once
blk_mq_quiesce_queue() returns, and we don't have such requirement
on APIs of stopping queue.
But blk_mq_quiesce_queue() still may not block/drain dispatch in the
the case of BLK_MQ_S_START_ON_RUN, so use the new introduced flag of
QUEUE_FLAG_QUIESCED and evaluate it inside RCU read-side critical
sections for fixing this issue.
Also blk_mq_quiesce_queue() is implemented via stopping queue, which
limits its uses, and easy to cause race, because any queue restart in
other paths may break blk_mq_quiesce_queue(). With the introduced
flag of QUEUE_FLAG_QUIESCED, we don't need to depend on stopping queue
for quiescing any more.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_start_stopped_hw_queues() is used implictly
as counterpart of blk_mq_quiesce_queue() for unquiescing queue,
so we introduce blk_mq_unquiesce_queue() and make it
as counterpart of blk_mq_quiesce_queue() explicitly.
This function is for improving the current quiescing mechanism
in the following patches.
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_queue_split() is always called with the last arg being q->bio_split,
where 'q' is the first arg.
Also blk_queue_split() sometimes uses the passed-in 'bs' and sometimes uses
q->bio_split.
This is inconsistent and unnecessary. Remove the last arg and always use
q->bio_split inside blk_queue_split()
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Credit-to: Javier González <jg@lightnvm.io> (Noticed that lightnvm was missed)
Reviewed-by: Javier González <javier@cnexlabs.com>
Tested-by: Javier González <javier@cnexlabs.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move most code into blk_mq_rq_ctx_init, and the rest into
blk_mq_get_request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This patch makes sure we always allocate requests in the core blk-mq
code and use a common prepare_request method to initialize them for
both mq I/O schedulers. For Kyber and additional limit_depth method
is added that is called before allocating the request.
Also because none of the intializations can really fail the new method
does not return an error - instead the bfq finish method is hardened
to deal with the no-IOC case.
Last but not least this removes the abuse of RQF_QUEUE by the blk-mq
scheduling code as RQF_ELFPRIV is all that is needed now.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_sched_assign_ioc now only handles the assigned of the ioc if
the schedule needs it (bfq only at the moment). The caller to the
per-request initializer is moved out so that it can be merged with
a similar call for the kyber I/O scheduler.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Merge three functions only tail-called by blk_mq_free_request into
blk_mq_free_request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Having these as separate helpers in a header really does not help
readability, or my chances to refactor this code sanely.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Having them out of line in blk-mq-sched.c just makes the code flow
unnecessarily complicated.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----
iQEcBAABAgAGBQJZPdbLAAoJEHm+PkMAQRiGx4wH/1nCjfnl6fE8oJ24/1gEAOUh
biFdqJkYZmlLYHVtYfLm4Ueg4adJdg0wx6qM/4RaAzmQVvLfDV34bc1qBf1+P95G
kVF+osWyXrZo5cTwkwapHW/KNu4VJwAx2D1wrlxKDVG5AOrULH1pYOYGOpApEkZU
4N+q5+M0ce0GJpqtUZX+UnI33ygjdDbBxXoFKsr24B7eA0ouGbAJ7dC88WcaETL+
2/7tT01SvDMo0jBSV0WIqlgXwZ5gp3yPGnklC3F4159Yze6VFrzHMKS/UpPF8o8E
W9EbuzwxsKyXUifX2GY348L1f+47glen/1sedbuKnFhP6E9aqUQQJXvEO7ueQl4=
=m2Gx
-----END PGP SIGNATURE-----
Merge tag 'v4.12-rc5' into for-4.13/block
We've already got a few conflicts and upcoming work depends on some of the
changes that have gone into mainline as regression fixes for this series.
Pull in 4.12-rc5 to resolve these conflicts and make it easier on down stream
trees to continue working on 4.13 changes.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Use the same values for use for request completion errors as the return
value from ->queue_rq. BLK_STS_RESOURCE is special cased to cause
a requeue, and all the others are completed as-is.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently we use nornal Linux errno values in the block layer, and while
we accept any error a few have overloaded magic meanings. This patch
instead introduces a new blk_status_t value that holds block layer specific
status codes and explicitly explains their meaning. Helpers to convert from
and to the previous special meanings are provided for now, but I suspect
we want to get rid of them in the long run - those drivers that have a
errno input (e.g. networking) usually get errnos that don't know about
the special block layer overloads, and similarly returning them to userspace
will usually return somethings that strictly speaking isn't correct
for file system operations, but that's left as an exercise for later.
For now the set of errors is a very limited set that closely corresponds
to the previous overloaded errno values, but there is some low hanging
fruite to improve it.
blk_status_t (ab)uses the sparse __bitwise annotations to allow for sparse
typechecking, so that we can easily catch places passing the wrong values.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
When direct issue is done on request picked up from plug list,
the hctx need to be updated with the actual hw queue, otherwise
wrong hctx is used and may hurt performance, especially when
wrong SRCU readlock is acquired/released
Reported-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The tagset lock needs to be held when iterating the tag_list, so a
lockdep assert was added when updating number of hardware queues. The
drivers calling this API, however, were unaware of the new requirement,
so are failing the assertion.
This patch takes the lock within the blk-mq function so the drivers do
not have to be modified in order to be safe.
Fixes: 705cda97e ("blk-mq: Make it safe to use RCU to iterate over blk_mq_tag_set.tag_list")
Reported-by: Gabriel Krisman Bertazi <krisman@collabora.co.uk>
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Because what the per-sw-queue bio merge does is basically same with
scheduler's .bio_merge(), this patch makes per-sw-queue bio merge
as the default .bio_merge if no scheduler is used or io scheduler
doesn't provide .bio_merge().
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Before blk-mq is introduced, I/O is merged to elevator
before being putted into plug queue, but blk-mq changed the
order and makes merging to sw queue basically impossible.
Then it is observed that throughput of sequential I/O is degraded
about 10%~20% on virtio-blk in the test[1] if mq-deadline isn't used.
This patch moves the bio merging per sw queue before plugging,
like what blk_queue_bio() does, and the performance regression is
fixed under this situation.
[1]. test script:
sudo fio --direct=1 --size=128G --bsrange=4k-4k --runtime=40 --numjobs=16 --ioengine=libaio --iodepth=64 --group_reporting=1 --filename=/dev/vdb --name=virtio_blk-test-$RW --rw=$RW --output-format=json
RW=read or write
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
No one uses it any more, so remove it.
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
When formatting NVMe to 512B/4K + T10 DIf/DIX, dd with split op returns
"Input/output error". Looks block layer split the bio after calling
bio_integrity_prep(bio). This patch fixes the issue.
Below is how we debug this issue:
(1)format nvme to 4K block # size with type 2 DIF
(2)dd with block size bigger than 1024k.
oflag=direct
dd: error writing '/dev/nvme0n1': Input/output error
We added some debug code in nvme device driver. It showed us the first
op and the second op have the same bi and pi address. This is not
correct.
1st op: nvme0n1 Op:Wr slba 0x505 length 0x100, PI ctrl=0x1400,
dsmgmt=0x0, AT=0x0 & RT=0x505
Guard 0x00b1, AT 0x0000, RT physical 0x00000505 RT virtual 0x00002828
2nd op: nvme0n1 Op:Wr slba 0x605 length 0x1, PI ctrl=0x1400, dsmgmt=0x0,
AT=0x0 & RT=0x605 ==> This op fails and subsequent 5 retires..
Guard 0x00b1, AT 0x0000, RT physical 0x00000605 RT virtual 0x00002828
With the fix, It showed us both of the first op and the second op have
correct bi and pi address.
1st op: nvme2n1 Op:Wr slba 0x505 length 0x100, PI ctrl=0x1400,
dsmgmt=0x0, AT=0x0 & RT=0x505
Guard 0x5ccb, AT 0x0000, RT physical 0x00000505 RT virtual
0x00002828
2nd op: nvme2n1 Op:Wr slba 0x605 length 0x1, PI ctrl=0x1400, dsmgmt=0x0,
AT=0x0 & RT=0x605
Guard 0xab4c, AT 0x0000, RT physical 0x00000605 RT virtual
0x00003028
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Making __blk_mq_stop_hw_queues static fixes sparse warning:
block/blk-mq.c:6: warning: symbol '__blk_mq_stop_hw_queues' was not
declared. Should it be static?
Fixes: 2719aa217e ("blk-mq: don't use sync workqueue flushing from drivers")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This can be triggered by hot-unplug one cpu.
======================================================
[ INFO: possible circular locking dependency detected ]
4.11.0+ #17 Not tainted
-------------------------------------------------------
step_after_susp/2640 is trying to acquire lock:
(all_q_mutex){+.+...}, at: [<ffffffffb33f95b8>] blk_mq_queue_reinit_work+0x18/0x110
but task is already holding lock:
(cpu_hotplug.lock){+.+.+.}, at: [<ffffffffb306d04f>] cpu_hotplug_begin+0x7f/0xe0
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (cpu_hotplug.lock){+.+.+.}:
lock_acquire+0x11c/0x230
__mutex_lock+0x92/0x990
mutex_lock_nested+0x1b/0x20
get_online_cpus+0x64/0x80
blk_mq_init_allocated_queue+0x3a0/0x4e0
blk_mq_init_queue+0x3a/0x60
loop_add+0xe5/0x280
loop_init+0x124/0x177
do_one_initcall+0x53/0x1c0
kernel_init_freeable+0x1e3/0x27f
kernel_init+0xe/0x100
ret_from_fork+0x31/0x40
-> #0 (all_q_mutex){+.+...}:
__lock_acquire+0x189a/0x18a0
lock_acquire+0x11c/0x230
__mutex_lock+0x92/0x990
mutex_lock_nested+0x1b/0x20
blk_mq_queue_reinit_work+0x18/0x110
blk_mq_queue_reinit_dead+0x1c/0x20
cpuhp_invoke_callback+0x1f2/0x810
cpuhp_down_callbacks+0x42/0x80
_cpu_down+0xb2/0xe0
freeze_secondary_cpus+0xb6/0x390
suspend_devices_and_enter+0x3b3/0xa40
pm_suspend+0x129/0x490
state_store+0x82/0xf0
kobj_attr_store+0xf/0x20
sysfs_kf_write+0x45/0x60
kernfs_fop_write+0x135/0x1c0
__vfs_write+0x37/0x160
vfs_write+0xcd/0x1d0
SyS_write+0x58/0xc0
do_syscall_64+0x8f/0x710
return_from_SYSCALL_64+0x0/0x7a
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(cpu_hotplug.lock);
lock(all_q_mutex);
lock(cpu_hotplug.lock);
lock(all_q_mutex);
*** DEADLOCK ***
8 locks held by step_after_susp/2640:
#0: (sb_writers#6){.+.+.+}, at: [<ffffffffb3244aed>] vfs_write+0x1ad/0x1d0
#1: (&of->mutex){+.+.+.}, at: [<ffffffffb32d3a51>] kernfs_fop_write+0x101/0x1c0
#2: (s_active#166){.+.+.+}, at: [<ffffffffb32d3a59>] kernfs_fop_write+0x109/0x1c0
#3: (pm_mutex){+.+...}, at: [<ffffffffb30d2ecd>] pm_suspend+0x21d/0x490
#4: (acpi_scan_lock){+.+.+.}, at: [<ffffffffb34dc3d7>] acpi_scan_lock_acquire+0x17/0x20
#5: (cpu_add_remove_lock){+.+.+.}, at: [<ffffffffb306d6d7>] freeze_secondary_cpus+0x27/0x390
#6: (cpu_hotplug.dep_map){++++++}, at: [<ffffffffb306cfd5>] cpu_hotplug_begin+0x5/0xe0
#7: (cpu_hotplug.lock){+.+.+.}, at: [<ffffffffb306d04f>] cpu_hotplug_begin+0x7f/0xe0
stack backtrace:
CPU: 3 PID: 2640 Comm: step_after_susp Not tainted 4.11.0+ #17
Hardware name: Dell Inc. OptiPlex 7040/0JCTF8, BIOS 1.4.9 09/12/2016
Call Trace:
dump_stack+0x99/0xce
print_circular_bug+0x1fa/0x270
__lock_acquire+0x189a/0x18a0
lock_acquire+0x11c/0x230
? lock_acquire+0x11c/0x230
? blk_mq_queue_reinit_work+0x18/0x110
? blk_mq_queue_reinit_work+0x18/0x110
__mutex_lock+0x92/0x990
? blk_mq_queue_reinit_work+0x18/0x110
? kmem_cache_free+0x2cb/0x330
? anon_transport_class_unregister+0x20/0x20
? blk_mq_queue_reinit_work+0x110/0x110
mutex_lock_nested+0x1b/0x20
? mutex_lock_nested+0x1b/0x20
blk_mq_queue_reinit_work+0x18/0x110
blk_mq_queue_reinit_dead+0x1c/0x20
cpuhp_invoke_callback+0x1f2/0x810
? __flow_cache_shrink+0x160/0x160
cpuhp_down_callbacks+0x42/0x80
_cpu_down+0xb2/0xe0
freeze_secondary_cpus+0xb6/0x390
suspend_devices_and_enter+0x3b3/0xa40
? rcu_read_lock_sched_held+0x79/0x80
pm_suspend+0x129/0x490
state_store+0x82/0xf0
kobj_attr_store+0xf/0x20
sysfs_kf_write+0x45/0x60
kernfs_fop_write+0x135/0x1c0
__vfs_write+0x37/0x160
? rcu_read_lock_sched_held+0x79/0x80
? rcu_sync_lockdep_assert+0x2f/0x60
? __sb_start_write+0xd9/0x1c0
? vfs_write+0x1ad/0x1d0
vfs_write+0xcd/0x1d0
SyS_write+0x58/0xc0
? rcu_read_lock_sched_held+0x79/0x80
do_syscall_64+0x8f/0x710
? trace_hardirqs_on_thunk+0x1a/0x1c
entry_SYSCALL64_slow_path+0x25/0x25
The cpu hotplug path will hold cpu_hotplug.lock and then reinit all exiting
queues for blk mq w/ all_q_mutex, however, blk_mq_init_allocated_queue() will
contend these two locks in the inversion order. This is due to commit eabe06595d
(blk/mq: Cure cpu hotplug lock inversion), it fixes a cpu hotplug lock inversion
issue because of hotplug rework, however the hotplug rework is still work-in-progress
and lives in a -tip branch and mainline cannot yet trigger that splat. The commit
breaks the linus's tree in the merge window, so this patch reverts the lock order
and avoids to splat linus's tree.
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Originally, I tied debugfs registration/unregistration together with
sysfs. There's no reason to do this, and it's getting in the way of
letting schedulers define their own debugfs attributes. Instead, tie the
debugfs registration to the lifetime of the structures themselves.
The saner lifetimes mean we can also get rid of the extra mq directory
and move everything one level up. I.e., nvme0n1/mq/hctx0/tags is now
just nvme0n1/hctx0/tags.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
A previous commit introduced the sync flush, which we need from
internal callers like blk_mq_quiesce_queue(). However, we also
call the stop helpers from drivers, particularly from ->queue_rq()
when we have to stop processing for a bit. We can't block from
those locations, and we don't have to guarantee that we're
fully flushed.
Fixes: 9f99373790 ("blk-mq: unify hctx delayed_run_work and run_work")
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
After queue is frozen, no request in this queue can be in use at all, so
there can't be any .queue_rq() running on this queue. It isn't
necessary to call blk_mq_quiesce_queue() any more, so remove it in both
elevator_switch_mq() and blk_mq_update_nr_requests().
Cc: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Fixed up the description a bit.
Signed-off-by: Jens Axboe <axboe@fb.com>
Remove the request_idx parameter, which can't be used safely now that we
support I/O schedulers with blk-mq. Except for a superflous check in
mtip32xx it was unused anyway.
Also pass the tag_set instead of just the driver data - this allows drivers
to avoid some code duplication in a follow on cleanup.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull block layer updates from Jens Axboe:
- Add BFQ IO scheduler under the new blk-mq scheduling framework. BFQ
was initially a fork of CFQ, but subsequently changed to implement
fairness based on B-WF2Q+, a modified variant of WF2Q. BFQ is meant
to be used on desktop type single drives, providing good fairness.
From Paolo.
- Add Kyber IO scheduler. This is a full multiqueue aware scheduler,
using a scalable token based algorithm that throttles IO based on
live completion IO stats, similary to blk-wbt. From Omar.
- A series from Jan, moving users to separately allocated backing
devices. This continues the work of separating backing device life
times, solving various problems with hot removal.
- A series of updates for lightnvm, mostly from Javier. Includes a
'pblk' target that exposes an open channel SSD as a physical block
device.
- A series of fixes and improvements for nbd from Josef.
- A series from Omar, removing queue sharing between devices on mostly
legacy drivers. This helps us clean up other bits, if we know that a
queue only has a single device backing. This has been overdue for
more than a decade.
- Fixes for the blk-stats, and improvements to unify the stats and user
windows. This both improves blk-wbt, and enables other users to
register a need to receive IO stats for a device. From Omar.
- blk-throttle improvements from Shaohua. This provides a scalable
framework for implementing scalable priotization - particularly for
blk-mq, but applicable to any type of block device. The interface is
marked experimental for now.
- Bucketized IO stats for IO polling from Stephen Bates. This improves
efficiency of polled workloads in the presence of mixed block size
IO.
- A few fixes for opal, from Scott.
- A few pulls for NVMe, including a lot of fixes for NVMe-over-fabrics.
From a variety of folks, mostly Sagi and James Smart.
- A series from Bart, improving our exposed info and capabilities from
the blk-mq debugfs support.
- A series from Christoph, cleaning up how handle WRITE_ZEROES.
- A series from Christoph, cleaning up the block layer handling of how
we track errors in a request. On top of being a nice cleanup, it also
shrinks the size of struct request a bit.
- Removal of mg_disk and hd (sorry Linus) by Christoph. The former was
never used by platforms, and the latter has outlived it's usefulness.
- Various little bug fixes and cleanups from a wide variety of folks.
* 'for-4.12/block' of git://git.kernel.dk/linux-block: (329 commits)
block: hide badblocks attribute by default
blk-mq: unify hctx delay_work and run_work
block: add kblock_mod_delayed_work_on()
blk-mq: unify hctx delayed_run_work and run_work
nbd: fix use after free on module unload
MAINTAINERS: bfq: Add Paolo as maintainer for the BFQ I/O scheduler
blk-mq-sched: alloate reserved tags out of normal pool
mtip32xx: use runtime tag to initialize command header
scsi: Implement blk_mq_ops.show_rq()
blk-mq: Add blk_mq_ops.show_rq()
blk-mq: Show operation, cmd_flags and rq_flags names
blk-mq: Make blk_flags_show() callers append a newline character
blk-mq: Move the "state" debugfs attribute one level down
blk-mq: Unregister debugfs attributes earlier
blk-mq: Only unregister hctxs for which registration succeeded
blk-mq-debugfs: Rename functions for registering and unregistering the mq directory
blk-mq: Let blk_mq_debugfs_register() look up the queue name
blk-mq: Register <dev>/queue/mq after having registered <dev>/queue
ide-pm: always pass 0 error to ide_complete_rq in ide_do_devset
ide-pm: always pass 0 error to __blk_end_request_all
..
The only difference between ->run_work and ->delay_work, is that
the latter is used to defer running a queue. This is done by
marking the queue stopped, and scheduling ->delay_work to run
sometime in the future. While the queue is stopped, direct runs
or runs through ->run_work will not run the queue.
If we combine the handlers, then we need to handle two things:
1) If a delayed/stopped run is scheduled, then we should not run
the queue before that has been completed.
2) If a queue is delayed/stopped, the handler needs to restart
the queue. Normally a run of a queue with the stopped bit set
would be a no-op.
Case 1 is handled by modifying a currently pending queue run
to the deadline set by the caller of blk_mq_delay_queue().
Subsequent attempts to queue a queue run will find the work
item already pending, and direct runs will see a stopped queue
as before.
Case 2 is handled by adding a new bit, BLK_MQ_S_START_ON_RUN,
that tells the work handler that it should clear a stopped
queue and run the handler.
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
They serve the exact same purpose. Get rid of the non-delayed
work variant, and just run it without delay for the normal case.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Avoid that the following kernel bug gets triggered:
BUG: sleeping function called from invalid context at ./include/linux/buffer_head.h:349
in_atomic(): 1, irqs_disabled(): 0, pid: 8019, name: find
CPU: 10 PID: 8019 Comm: find Tainted: G W I 4.11.0-rc4-dbg+ #2
Call Trace:
dump_stack+0x68/0x93
___might_sleep+0x16e/0x230
__might_sleep+0x4a/0x80
__ext4_get_inode_loc+0x1e0/0x4e0
ext4_iget+0x70/0xbc0
ext4_iget_normal+0x2f/0x40
ext4_lookup+0xb6/0x1f0
lookup_slow+0x104/0x1e0
walk_component+0x19a/0x330
path_lookupat+0x4b/0x100
filename_lookup+0x9a/0x110
user_path_at_empty+0x36/0x40
vfs_statx+0x67/0xc0
SYSC_newfstatat+0x20/0x40
SyS_newfstatat+0xe/0x10
entry_SYSCALL_64_fastpath+0x18/0xad
This happens since the big if/else in blk_mq_make_request() doesn't
have final else section that also drops the ctx. Add that.
Fixes: b00c53e8f4 ("blk-mq: fix schedule-while-atomic with scheduler attached")
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Omar Sandoval <osandov@fb.com>
Added a bit more to the commit log.
Signed-off-by: Jens Axboe <axboe@fb.com>
No point in providing and exporting this helper. There's just
one (real) user of it, just use rq_data_dir().
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
If the caller passes in wait=true, it has to be able to block
for a driver tag. We just had a bug where flush insertion
would block on tag allocation, while we had preempt disabled.
Ensure that we catch cases like that earlier next time.
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Fixes an issue where the size of the poll_stat array in request_queue
does not match the size expected by the new size based bucketing for
IO completion polling.
Fixes: 720b8ccc45 ("blk-mq: Add a polling specific stats function")
Signed-off-by: Stephen Bates <sbates@raithlin.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Rather than bucketing IO statisics based on direction only we also
bucket based on the IO size. This leads to improved polling
performance. Update the bucket callback function and use it in the
polling latency estimation.
Signed-off-by: Stephen Bates <sbates@raithlin.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If we have a scheduler attached, blk_mq_tag_to_rq() on the
scheduled tags will return NULL if a request is no longer
in flight. This is different than using the normal tags,
where it will always return the fixed request. Check for
this condition for polling, in case we happen to enter
polling for a completed request.
The request address remains valid, so this check and return
should be perfectly safe.
Fixes: bd166ef183 ("blk-mq-sched: add framework for MQ capable IO schedulers")
Tested-by: Stephen Bates <sbates@raithlin.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Merge blk_mq_ipi_complete_request and blk_mq_stat_add into their only
caller.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Now that all drivers that call blk_mq_complete_requests have a
->complete callback we can remove the direct call to blk_mq_end_request,
as well as the error argument to blk_mq_complete_request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Export this function such that it becomes available to block
drivers.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Matias Bjørling <m@bjorling.me>
Cc: Adam Manzanares <adam.manzanares@wdc.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently, this callback is called right after put_request() and has no
distinguishable purpose. Instead, let's call it before put_request() as
soon as I/O has completed on the request, before we account it in
blk-stat. With this, Kyber can enable stats when it sees a latency
outlier and make sure the outlier gets accounted.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_mq_finish_request() is required for schedulers that define their own
put_request(). blk_mq_run_hw_queue() is required for schedulers that
hold back requests to be run later.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The blk_mq_dispatch_rq_list() implementation got modified several
times but the comments in that function were not updated every
time. Since it is nontrivial what is going on, update the comments
in blk_mq_dispatch_rq_list().
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Since the next patch in this series will use RCU to iterate over
tag_list, make this safe. Add lockdep_assert_held() statements
in functions that iterate over tag_list to make clear that using
list_for_each_entry() instead of list_for_each_entry_rcu() is
fine in these functions.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Minor cleanup that makes it easier to figure out what's going on in the
driver tag allocation failure path of blk_mq_dispatch_rq_list().
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
We've added a considerable amount of fixes for stalls and issues
with the blk-mq scheduling in the 4.11 series since forking
off the for-4.12/block branch. We need to do improvements on
top of that for 4.12, so pull in the previous fixes to make
our lives easier going forward.
Signed-off-by: Jens Axboe <axboe@fb.com>
To improve scalability, if hardware queues are shared, restart
a single hardware queue in round-robin fashion. Rename
blk_mq_sched_restart_queues() to reflect the new semantics.
Remove blk_mq_sched_mark_restart_queue() because this function
has no callers. Remove flag QUEUE_FLAG_RESTART because this
patch removes the code that uses this flag.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Introduce a function that runs a hardware queue unconditionally
after a delay. Note: there is already a function that stops and
restarts a hardware queue after a delay, namely blk_mq_delay_queue().
This function will be used in the next patch in this series.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Long Li <longli@microsoft.com>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_mq_update_nr_hw_queues() used to remap hardware queues, which is the
behavior that drivers expect. However, commit 4e68a01142 changed
blk_mq_queue_reinit() to not remap queues for the case of CPU
hotplugging, inadvertently making blk_mq_update_nr_hw_queues() not remap
queues as well. This breaks, for example, NBD's multi-connection mode,
leaving the added hardware queues unused. Fix it by making
blk_mq_update_nr_hw_queues() explicitly remap the queues.
Fixes: 4e68a01142 ("blk-mq: don't redistribute hardware queues on a CPU hotplug event")
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
In elevator_switch(), if blk_mq_init_sched() fails, we attempt to fall
back to the original scheduler. However, at this point, we've already
torn down the original scheduler's tags, so this causes a crash. Doing
the fallback like the legacy elevator path is much harder for mq, so fix
it by just falling back to none, instead.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If a new hardware queue is added at runtime, we don't allocate scheduler
tags for it, leading to a crash. This hooks up the scheduler framework
to blk_mq_{init,exit}_hctx() to make sure everything gets properly
initialized/freed.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
While dispatching requests, if we fail to get a driver tag, we mark the
hardware queue as waiting for a tag and put the requests on a
hctx->dispatch list to be run later when a driver tag is freed. However,
blk_mq_dispatch_rq_list() may dispatch requests from multiple hardware
queues if using a single-queue scheduler with a multiqueue device. If
blk_mq_get_driver_tag() fails, it doesn't update the hardware queue we
are processing. This means we end up using the hardware queue of the
previous request, which may or may not be the same as that of the
current request. If it isn't, the wrong hardware queue will end up
waiting for a tag, and the requests will be on the wrong dispatch list,
leading to a hang.
The fix is twofold:
1. Make sure we save which hardware queue we were trying to get a
request for in blk_mq_get_driver_tag() regardless of whether it
succeeds or not.
2. Make blk_mq_dispatch_rq_list() take a request_queue instead of a
blk_mq_hw_queue to make it clear that it must handle multiple
hardware queues, since I've already messed this up on a couple of
occasions.
This didn't appear in testing with nvme and mq-deadline because nvme has
more driver tags than the default number of scheduler tags. However,
with the blk_mq_update_nr_hw_queues() fix, it showed up with nbd.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The block layer core sets blk_mq_queue_data.list but no block
drivers read that member. Hence remove it and also the code that
is used to set this member.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
Commit a4d907b6a3 unified the single and multi queue request handlers,
but in the process, it also screwed up the locking balance and calls
blk_mq_try_issue_directly() with the ctx preempt lock held. This is a
problem for drivers that have set BLK_MQ_F_BLOCKING, since now they
can't reliably sleep.
While in there, protect against similar issues in the future, by adding
a might_sleep() trigger in the BLOCKING path for direct issue or queue
run.
Reported-by: Josef Bacik <josef@toxicpanda.com>
Tested-by: Josef Bacik <josef@toxicpanda.com>
Fixes: a4d907b6a3 ("blk-mq: streamline blk_mq_make_request")
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
In blk_mq_alloc_request_hctx, blk_mq_sched_get_request doesn't
get sw context so we don't need to put the context with
blk_mq_put_ctx. Unless, we will see preempt counter underflow.
Cc: Omar Sandoval <osandov@fb.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
In blk_mq_alloc_request_hctx, blk_mq_sched_get_request doesn't
get sw context so we don't need to put the context with
blk_mq_put_ctx. Unless, we will see preempt counter underflow.
Cc: Omar Sandoval <osandov@fb.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently we return true in blk_mq_dispatch_rq_list() if we queued IO
successfully, but we really want to return whether or not the we made
progress. Progress includes if we got an error return. If we don't,
this can lead to a hang in blk_mq_sched_dispatch_requests() when a
driver is draining IO by returning BLK_MQ_QUEUE_ERROR instead of
manually ending the IO in error and return BLK_MQ_QUEUE_OK.
Tested-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
When try to issue a request directly and we fail we will requeue the
request, but call blk_mq_end_request() as well. This leads to the
completed request being on a queuelist and getting ended twice, which
causes list corruption in schedulers and other shenanigans.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Ming Lei <tom.leiming@gmail.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_alloc_queue_node() already allocates q->stats, so
blk_mq_init_allocated_queue() is overwriting it with a new allocation.
Fixes: a83b576c9c ("block: fix stacked driver stats init and free")
Reviewed-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
As the .q_usage_counter is used by both legacy and
mq path, we need to block new I/O if queue becomes
dead in blk_queue_enter().
So rename it and we can use this function in both
paths.
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch adds comment on two races related with
timeout handler:
- requeue from queue busy vs. timeout
- rq free & reallocation vs. timeout
Both the races themselves and current solution aren't
explicit enough, so add comments on them.
Cc: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently there is no way to know the request size when the request is
finished. Next patch will need this info. We could add extra field to
record the size, but blk_issue_stat has enough space to record it, so
this patch just overloads blk_issue_stat. With this, we will have 49bits
to track time, which still is very long time.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently we return true in blk_mq_dispatch_rq_list() if we queued IO
successfully, but we really want to return whether or not the we made
progress. Progress includes if we got an error return. If we don't,
this can lead to a hang in blk_mq_sched_dispatch_requests() when a
driver is draining IO by returning BLK_MQ_QUEUE_ERROR instead of
manually ending the IO in error and return BLK_MQ_QUEUE_OK.
Tested-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Turn the different ways of merging or issuing I/O into a series of if/else
statements instead of the current maze of gotos. Note that this means we
pin the CPU a little longer for some cases as the CTX put is moved to
common code at the end of the function.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Now that we have a nice direct issue heper this helps simplifying
the code a bit, and also gets rid of the old_rq variable.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Rename blk_mq_try_issue_directly to __blk_mq_try_issue_directly and add a
new wrapper that takes care of RCU / SRCU locking to avoid having
boileplate code in the caller which would get duplicated with new callers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
They are mostly the same code anyway - this just one small conditional
for the plug case that is different for both variants.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This flag was never used since it was introduced.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently, statistics are gathered in ~0.13s windows, and users grab the
statistics whenever they need them. This is not ideal for both in-tree
users:
1. Writeback throttling wants its own dynamically sized window of
statistics. Since the blk-stats statistics are reset after every
window and the wbt windows don't line up with the blk-stats windows,
wbt doesn't see every I/O.
2. Polling currently grabs the statistics on every I/O. Again, depending
on how the window lines up, we may miss some I/Os. It's also
unnecessary overhead to get the statistics on every I/O; the hybrid
polling heuristic would be just as happy with the statistics from the
previous full window.
This reworks the blk-stats infrastructure to be callback-based: users
register a callback that they want called at a given time with all of
the statistics from the window during which the callback was active.
Users can dynamically bucketize the statistics. wbt and polling both
currently use read vs. write, but polling can be extended to further
subdivide based on request size.
The callbacks are kept on an RCU list, and each callback has percpu
stats buffers. There will only be a few users, so the overhead on the
I/O completion side is low. The stats flushing is also simplified
considerably: since the timer function is responsible for clearing the
statistics, we don't have to worry about stale statistics.
wbt is a trivial conversion. After the conversion, the windowing problem
mentioned above is fixed.
For polling, we register an extra callback that caches the previous
window's statistics in the struct request_queue for the hybrid polling
heuristic to use.
Since we no longer have a single stats buffer for the request queue,
this also removes the sysfs and debugfs stats entries. To replace those,
we add a debugfs entry for the poll statistics.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The stats buckets will become generic soon, so make the existing users
use the common READ and WRITE definitions instead of one internal to
blk-stat.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
We always call wbt_exit() from blk_release_queue(), so these are
unnecessary.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If we have scheduling enabled, we jump directly to insert-and-run.
That's fine, but we run the queue async and we don't pass in information
on whether we can block from this context or not. Fixup both these
cases.
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
It is obviously that hctx->cpumask is per hctx, and both
share same lifetime, so this patch moves freeing of hctx->cpumask
into release handler of hctx's kobject.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch removes kobject_put() over hctx in __blk_mq_unregister_dev(),
and trys to keep lifetime consistent between hctx and hctx's kobject.
Now blk_mq_sysfs_register() and blk_mq_sysfs_unregister() become
totally symmetrical, and kobject's refcounter drops to zero just
when the hctx is freed.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently from kobject view, both q->mq_kobj and ctx->kobj can
be released during one cycle of blk_mq_register_dev() and
blk_mq_unregister_dev(). Actually, sw queue's lifetime is
same with its request queue's, which is covered by request_queue->kobj.
So we don't need to call kobject_put() for the two kinds of
kobject in __blk_mq_unregister_dev(), instead we do that
in release handler of request queue.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull block layer fixes from Jens Axboe:
"A collection of fixes for this merge window, either fixes for existing
issues, or parts that were waiting for acks to come in. This pull
request contains:
- Allocation of nvme queues on the right node from Shaohua.
This was ready long before the merge window, but waiting on an ack
from Bjorn on the PCI bit. Now that we have that, the three patches
can go in.
- Two fixes for blk-mq-sched with nvmeof, which uses hctx specific
request allocations. This caused an oops. One part from Sagi, one
part from Omar.
- A loop partition scan deadlock fix from Omar, fixing a regression
in this merge window.
- A three-patch series from Keith, closing up a hole on clearing out
requests on shutdown/resume.
- A stable fix for nbd from Josef, fixing a leak of sockets.
- Two fixes for a regression in this window from Jan, fixing a
problem with one of his earlier patches dealing with queue vs bdi
life times.
- A fix for a regression with virtio-blk, causing an IO stall if
scheduling is used. From me.
- A fix for an io context lock ordering problem. From me"
* 'for-linus' of git://git.kernel.dk/linux-block:
block: Move bdi_unregister() to del_gendisk()
blk-mq: ensure that bd->last is always set correctly
block: don't call ioc_exit_icq() with the queue lock held for blk-mq
block: Initialize bd_bdi on inode initialization
loop: fix LO_FLAGS_PARTSCAN hang
nvme: Complete all stuck requests
blk-mq: Provide freeze queue timeout
blk-mq: Export blk_mq_freeze_queue_wait
nbd: stop leaking sockets
blk-mq: move update of tags->rqs to __blk_mq_alloc_request()
blk-mq: kill blk_mq_set_alloc_data()
blk-mq: make blk_mq_alloc_request_hctx() allocate a scheduler request
blk-mq-sched: Allocate sched reserved tags as specified in the original queue tagset
nvme: allocate nvme_queue in correct node
PCI: add an API to get node from vector
blk-mq: allocate blk_mq_tags and requests in correct node
When drivers are called with a request in blk-mq, blk-mq flags the
state such that the driver knows if this is the last request in
this call chain or not. The driver can then use that information
to defer kicking off IO until bd->last is true. However, with blk-mq
and scheduling, we need to allocate a driver tag for a request before
it can be issued. If we fail to allocate such a tag, we could end up
in the situation where the last request issued did not have
bd->last == true set. This can then cause a driver hang.
This fixes a hang with virtio-blk, which uses bd->last as a hint
on whether to kick the queue or not.
Reported-by: Chris Mason <clm@fb.com>
Tested-by: Chris Mason <clm@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
A driver may wish to take corrective action if queued requests do not
complete within a set time.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Drivers can start a freeze, so this provides a way to wait for frozen.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
No functional difference, it just makes a little more sense to update
the tag map where we actually allocate the tag.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Tested-by: Sagi Grimberg <sagi@grimberg.me>
blk_mq_alloc_request_hctx() allocates a driver request directly, unlike
its blk_mq_alloc_request() counterpart. It also crashes because it
doesn't update the tags->rqs map.
Fix it by making it allocate a scheduler request.
Reported-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Tested-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Modified by me to also check at driver tag allocation time if the
original request was reserved, so we can be sure to allocate a
properly reserved tag at that point in time, too.
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_mq_tags/requests of specific hardware queue are mostly used in
specific cpus, which might not be in the same numa node as disk. For
example, a nvme card is in node 0. half hardware queue will be used by
node 0, the other node 1.
Signed-off-by: Shaohua Li <shli@fb.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Fix up affected files that include this signal functionality via sched.h.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We are going to split <linux/sched/topology.h> out of <linux/sched.h>, which
will have to be picked up from other headers and a couple of .c files.
Create a trivial placeholder <linux/sched/topology.h> file that just
maps to <linux/sched.h> to make this patch obviously correct and
bisectable.
Include the new header in the files that are going to need it.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit 50e1dab86a ("blk-mq-sched: fix starvation for multiple hardware
queues and shared tags") fixed one starvation issue for shared tags.
However, we can still get into a situation where we fail to allocate a
tag because all tags are allocated but we don't have any pending
requests on any hardware queue.
One solution for this would be to restart all queues that share a tag
map, but that really sucks. Ideally, we could just block and wait for a
tag, but that isn't always possible from blk_mq_dispatch_rq_list().
However, we can still use the struct sbitmap_queue wait queues with a
custom callback instead of blocking. This has a few benefits:
1. It avoids iterating over all hardware queues when completing an I/O,
which the current restart code has to do.
2. It benefits from the existing rolling wakeup code.
3. It avoids punting to another thread just to have it block.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The current request insertion machinery works just fine for
directly inserting flushes, so no need to special case
this anymore.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Currently we're almost there, but if we dispatch nothing, then we
still return success.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Calling blk_queue_make_request resets a bunch of settings on the
request_queue, but all we really want is to update the make_request_fn,
so do this directly so we don't lose things like the logical and
physical block sizes.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Add a new merge strategy that merges discard bios into a request until the
maximum number of discard ranges (or the maximum discard size) is reached
from the plug merging code. I/O scheduler merging is not wired up yet
but might also be useful, although not for fast devices like NVMe which
are the only user for now.
Note that for now we don't support limiting the size of each discard range,
but if needed that can be added later.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Switch these constants to an enum, and make let the compiler ensure that
all callers of blk_try_merge and elv_merge handle all potential values.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
When I added the blk-mq debugging information to debugfs, I didn't
notice that blktrace also creates a "block" directory in debugfs. Make
them use the same dentry, now created in the core block code. Based on a
patch from Jens.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
We rely on blk_mq_get_driver_tag() not failing if 'wait' is true,
but it currently fails in that case if the queue happens to be
stopped at the time of the call.
We don't need to check for stopped here, it's just assigning
the tag. If the queue is stopped, we'll handle it when
attempting to run the queue.
This fixes a stall/crash on flush intensive workloads, where
we proceed to process a flush that doesn't have a valid tag
assigned.
Signed-off-by: Jens Axboe <axboe@fb.com>
And require all drivers that want to support BLOCK_PC to allocate it
as the first thing of their private data. To support this the legacy
IDE and BSG code is switched to set cmd_size on their queues to let
the block layer allocate the additional space.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Instead of letting the caller check this and handle the details
of inserting a flush request, put the logic in the scheduler
insertion function. This fixes direct flush insertion outside
of the usual make_request_fn calls, like from dm via
blk_insert_cloned_request().
Signed-off-by: Jens Axboe <axboe@fb.com>
This centralizes the checks for bios that needs to be go into the flush
state machine.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
When we invoke dispatch_requests(), the scheduler empties everything
into the passed in list. This isn't always a good thing, since it
means that we remove items that we could have potentially merged
with.
Change the function to dispatch single requests at the time. If
we do that, we can backoff exactly at the point where the device
can't consume more IO, and leave the rest with the scheduler for
better merging and future dispatch decision making.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Tested-by: Hannes Reinecke <hare@suse.com>
If we have both multiple hardware queues and shared tag map between
devices, we need to ensure that we propagate the hardware queue
restart bit higher up. This is because we can get into a situation
where we don't have any IO pending on a hardware queue, yet we fail
getting a tag to start new IO. If that happens, it's not enough to
mark the hardware queue as needing a restart, we need to bubble
that up to the higher level queue as well.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Tested-by: Hannes Reinecke <hare@suse.com>
We don't want to hold on to this resource when we have a scheduler
attached.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Tested-by: Hannes Reinecke <hare@suse.com>
Once we mark the queue as needing a restart, re-check if we can
get a driver tag. This fixes a theoretical issue where the needed
IO completes _after_ blk_mq_get_driver_tag() fails, but before we
manage to set the restart bit.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Tested-by: Hannes Reinecke <hare@suse.com>
We'll use the same criteria for whether we need to run the queue sync
or async when we have a scheduler, as we do without one.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Tested-by: Hannes Reinecke <hare@suse.com>
In preparation for putting blk-mq debugging information in debugfs,
create a directory tree mirroring the one in sysfs:
# tree -d /sys/kernel/debug/block
/sys/kernel/debug/block
|-- nvme0n1
| `-- mq
| |-- 0
| | `-- cpu0
| |-- 1
| | `-- cpu1
| |-- 2
| | `-- cpu2
| `-- 3
| `-- cpu3
`-- vda
`-- mq
`-- 0
|-- cpu0
|-- cpu1
|-- cpu2
`-- cpu3
Also add the scaffolding for the actual files that will go in here,
either under the hardware queue or software queue directories.
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If we come in from blk_mq_alloc_requst() with NOWAIT set in flags,
we must ensure that we don't later overwrite that in
blk_mq_sched_get_request(). Initialize alloc_data->flags before
passing it in.
Reported-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
If we have a scheduler attached, we have two sets of tags. We don't
want to apply our active queue throttling for the scheduler side
of tags, that only applies to driver tags since that's the resource
we need to dispatch an IO.
Signed-off-by: Jens Axboe <axboe@fb.com>
Add support for growing the tags associated with a hardware queue, for
the scheduler tags. Currently we only support resizing within the
limits of the original depth, change that so we can grow it as well by
allocating and replacing the existing scheduler tag set.
This is similar to how we could increase the software queue depth with
the legacy IO stack and schedulers.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
The run handler we register for the delayed work requires that the
queue be stopped, yet we leave that up to the caller. Let's move
it into blk_mq_delay_queue() itself, so that the API is sane.
This fixes a stall with SCSI, where it calls blk_mq_delay_queue()
without having stopped the queue. Hence the queue is never run.
Reported-by: Hannes Reinecke <hare@suse.com>
Fixes: 70f4db639c ("blk-mq: add blk_mq_delay_queue")
Signed-off-by: Jens Axboe <axboe@fb.com>
Add Kconfig entries to manage what devices get assigned an MQ
scheduler, and add a blk-mq flag for drivers to opt out of scheduling.
The latter is useful for admin type queues that still allocate a blk-mq
queue and tag set, but aren't use for normal IO.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
This adds a set of hooks that intercepts the blk-mq path of
allocating/inserting/issuing/completing requests, allowing
us to develop a scheduler within that framework.
We reuse the existing elevator scheduler API on the registration
side, but augment that with the scheduler flagging support for
the blk-mq interfce, and with a separate set of ops hooks for MQ
devices.
We split driver and scheduler tags, so we can run the scheduling
independently of device queue depth.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
This is in preparation for having two sets of tags available. For
that we need a static index, and a dynamically assignable one.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
No functional change in this patch, just in preparation for having
two types of tags available to the block layer for a single request.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Prep patch for adding an extra tag map for scheduler requests.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
This is in preparation for having another tag set available. Cleanup
the parameters, and allow passing in of tags for blk_mq_put_tag().
Signed-off-by: Jens Axboe <axboe@fb.com>
[hch: even more cleanups]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
It's only used in blk-mq, kill it from the main exported header
and kill the symbol export as well.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
ktime_set(S,N) was required for the timespec storage type and is still
useful for situations where a Seconds and Nanoseconds part of a time value
needs to be converted. For anything where the Seconds argument is 0, this
is pointless and can be replaced with a simple assignment.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
In blk_mq_map_swqueue, there is a memory optimization that frees the
tags of a queue that has gone unmapped. Later, if that hctx is remapped
after another topology change, the tags need to be reallocated.
If this allocation fails, a simple WARN_ON triggers, but the block layer
ends up with an active hctx without any corresponding set of tags.
Then, any income IO to that hctx can trigger an Oops.
I can reproduce it consistently by running IO, flipping CPUs on and off
and eventually injecting a memory allocation failure in that path.
In the fix below, if the system experiences a failed allocation of any
hctx's tags, we remap all the ctxs of that queue to the hctx_0, which
should always keep it's tags. There is a minor performance hit, since
our mapping just got worse after the error path, but this is
the simplest solution to handle this error path. The performance hit
will disappear after another successful remap.
I considered dropping the memory optimization all together, but it
seemed a bad trade-off to handle this very specific error case.
This should apply cleanly on top of Jens' for-next branch.
The Oops is the one below:
SP (3fff935ce4d0) is in userspace
1:mon> e
cpu 0x1: Vector: 300 (Data Access) at [c000000fe99eb110]
pc: c0000000005e868c: __sbitmap_queue_get+0x2c/0x180
lr: c000000000575328: __bt_get+0x48/0xd0
sp: c000000fe99eb390
msr: 900000010280b033
dar: 28
dsisr: 40000000
current = 0xc000000fe9966800
paca = 0xc000000007e80300 softe: 0 irq_happened: 0x01
pid = 11035, comm = aio-stress
Linux version 4.8.0-rc6+ (root@bean) (gcc version 5.4.0 20160609
(Ubuntu/IBM 5.4.0-6ubuntu1~16.04.2) ) #3 SMP Mon Oct 10 20:16:53 CDT 2016
1:mon> s
[c000000fe99eb3d0] c000000000575328 __bt_get+0x48/0xd0
[c000000fe99eb400] c000000000575838 bt_get.isra.1+0x78/0x2d0
[c000000fe99eb480] c000000000575cb4 blk_mq_get_tag+0x44/0x100
[c000000fe99eb4b0] c00000000056f6f4 __blk_mq_alloc_request+0x44/0x220
[c000000fe99eb500] c000000000570050 blk_mq_map_request+0x100/0x1f0
[c000000fe99eb580] c000000000574650 blk_mq_make_request+0xf0/0x540
[c000000fe99eb640] c000000000561c44 generic_make_request+0x144/0x230
[c000000fe99eb690] c000000000561e00 submit_bio+0xd0/0x200
[c000000fe99eb740] c0000000003ef740 ext4_io_submit+0x90/0xb0
[c000000fe99eb770] c0000000003e95d8 ext4_writepages+0x588/0xdd0
[c000000fe99eb910] c00000000025a9f0 do_writepages+0x60/0xc0
[c000000fe99eb940] c000000000246c88 __filemap_fdatawrite_range+0xf8/0x180
[c000000fe99eb9e0] c000000000246f90 filemap_write_and_wait_range+0x70/0xf0
[c000000fe99eba20] c0000000003dd844 ext4_sync_file+0x214/0x540
[c000000fe99eba80] c000000000364718 vfs_fsync_range+0x78/0x130
[c000000fe99ebad0] c0000000003dd46c ext4_file_write_iter+0x35c/0x430
[c000000fe99ebb90] c00000000038c280 aio_run_iocb+0x3b0/0x450
[c000000fe99ebce0] c00000000038dc28 do_io_submit+0x368/0x730
[c000000fe99ebe30] c000000000009404 system_call+0x38/0xec
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Cc: Brian King <brking@linux.vnet.ibm.com>
Cc: Douglas Miller <dougmill@linux.vnet.ibm.com>
Cc: linux-block@vger.kernel.org
Cc: linux-scsi@vger.kernel.org
Reviewed-by: Douglas Miller <dougmill@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
While stressing memory and IO at the same time we changed SMT settings,
we were able to consistently trigger deadlocks in the mm system, which
froze the entire machine.
I think that under memory stress conditions, the large allocations
performed by blk_mq_init_rq_map may trigger a reclaim, which stalls
waiting on the block layer remmaping completion, thus deadlocking the
system. The trace below was collected after the machine stalled,
waiting for the hotplug event completion.
The simplest fix for this is to make allocations in this path
non-reclaimable, with GFP_NOIO. With this patch, We couldn't hit the
issue anymore.
This should apply on top of Jens's for-next branch cleanly.
Changes since v1:
- Use GFP_NOIO instead of GFP_NOWAIT.
Call Trace:
[c000000f0160aaf0] [c000000f0160ab50] 0xc000000f0160ab50 (unreliable)
[c000000f0160acc0] [c000000000016624] __switch_to+0x2e4/0x430
[c000000f0160ad20] [c000000000b1a880] __schedule+0x310/0x9b0
[c000000f0160ae00] [c000000000b1af68] schedule+0x48/0xc0
[c000000f0160ae30] [c000000000b1b4b0] schedule_preempt_disabled+0x20/0x30
[c000000f0160ae50] [c000000000b1d4fc] __mutex_lock_slowpath+0xec/0x1f0
[c000000f0160aed0] [c000000000b1d678] mutex_lock+0x78/0xa0
[c000000f0160af00] [d000000019413cac] xfs_reclaim_inodes_ag+0x33c/0x380 [xfs]
[c000000f0160b0b0] [d000000019415164] xfs_reclaim_inodes_nr+0x54/0x70 [xfs]
[c000000f0160b0f0] [d0000000194297f8] xfs_fs_free_cached_objects+0x38/0x60 [xfs]
[c000000f0160b120] [c0000000003172c8] super_cache_scan+0x1f8/0x210
[c000000f0160b190] [c00000000026301c] shrink_slab.part.13+0x21c/0x4c0
[c000000f0160b2d0] [c000000000268088] shrink_zone+0x2d8/0x3c0
[c000000f0160b380] [c00000000026834c] do_try_to_free_pages+0x1dc/0x520
[c000000f0160b450] [c00000000026876c] try_to_free_pages+0xdc/0x250
[c000000f0160b4e0] [c000000000251978] __alloc_pages_nodemask+0x868/0x10d0
[c000000f0160b6f0] [c000000000567030] blk_mq_init_rq_map+0x160/0x380
[c000000f0160b7a0] [c00000000056758c] blk_mq_map_swqueue+0x33c/0x360
[c000000f0160b820] [c000000000567904] blk_mq_queue_reinit+0x64/0xb0
[c000000f0160b850] [c00000000056a16c] blk_mq_queue_reinit_notify+0x19c/0x250
[c000000f0160b8a0] [c0000000000f5d38] notifier_call_chain+0x98/0x100
[c000000f0160b8f0] [c0000000000c5fb0] __cpu_notify+0x70/0xe0
[c000000f0160b930] [c0000000000c63c4] notify_prepare+0x44/0xb0
[c000000f0160b9b0] [c0000000000c52f4] cpuhp_invoke_callback+0x84/0x250
[c000000f0160ba10] [c0000000000c570c] cpuhp_up_callbacks+0x5c/0x120
[c000000f0160ba60] [c0000000000c7cb8] _cpu_up+0xf8/0x1d0
[c000000f0160bac0] [c0000000000c7eb0] do_cpu_up+0x120/0x150
[c000000f0160bb40] [c0000000006fe024] cpu_subsys_online+0x64/0xe0
[c000000f0160bb90] [c0000000006f5124] device_online+0xb4/0x120
[c000000f0160bbd0] [c0000000006f5244] online_store+0xb4/0xc0
[c000000f0160bc20] [c0000000006f0a68] dev_attr_store+0x68/0xa0
[c000000f0160bc60] [c0000000003ccc30] sysfs_kf_write+0x80/0xb0
[c000000f0160bca0] [c0000000003cbabc] kernfs_fop_write+0x17c/0x250
[c000000f0160bcf0] [c00000000030fe6c] __vfs_write+0x6c/0x1e0
[c000000f0160bd90] [c000000000311490] vfs_write+0xd0/0x270
[c000000f0160bde0] [c0000000003131fc] SyS_write+0x6c/0x110
[c000000f0160be30] [c000000000009204] system_call+0x38/0xec
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Cc: Brian King <brking@linux.vnet.ibm.com>
Cc: Douglas Miller <dougmill@linux.vnet.ibm.com>
Cc: linux-block@vger.kernel.org
Cc: linux-scsi@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
Takes a list of requests, and dispatches it. Moves any residual
requests to the dispatch list.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
We have a variant for all hardware queues, but not one for a single
hardware queue.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
After commit 287922eb0b ("block: defer timeouts to a workqueue"),
deleting the timeout work after freezing the queue shouldn't be
necessary, since the synchronization is already enforced by the
acquisition of a q_usage_counter reference in blk_mq_timeout_work.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Reviewed-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The previous commit introduced the hybrid sleep/poll mode. Take
that one step further, and use the completion latencies to
automatically sleep for half the mean completion time. This is
a good approximation.
This changes the 'io_poll_delay' sysfs file a bit to expose the
various options. Depending on the value, the polling code will
behave differently:
-1 Never enter hybrid sleep mode
0 Use half of the completion mean for the sleep delay
>0 Use this specific value as the sleep delay
Signed-off-by: Jens Axboe <axboe@fb.com>
Tested-By: Stephen Bates <sbates@raithlin.com>
Reviewed-By: Stephen Bates <sbates@raithlin.com>
This patch enables a hybrid polling mode. Instead of polling after IO
submission, we can induce an artificial delay, and then poll after that.
For example, if the IO is presumed to complete in 8 usecs from now, we
can sleep for 4 usecs, wake up, and then do our polling. This still puts
a sleep/wakeup cycle in the IO path, but instead of the wakeup happening
after the IO has completed, it'll happen before. With this hybrid
scheme, we can achieve big latency reductions while still using the same
(or less) amount of CPU.
Signed-off-by: Jens Axboe <axboe@fb.com>
Tested-By: Stephen Bates <sbates@raithlin.com>
Reviewed-By: Stephen Bates <sbates@raithlin.com>
In both legacy and mq path, req count of plug list is computed
before allocating request, so the number can be stale when falling
back to slept allocation, also the new introduced wbt can sleep
too.
This patch deals with the case by checking if plug list becomes
empty, and fixes the KASAN report of 'BUG: KASAN: stack-out-of-bounds'
which is introduced by Shaohua's patches of dispatching big request.
Fixes: 600271d900002(blk-mq: immediately dispatch big size request)
Fixes: 50d24c34403c6(block: immediately dispatch big size request)
Cc: Shaohua Li <shli@fb.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The poll code is blk-mq specific, let's move it to blk-mq.c. This
is a prep patch for improving the polling code.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
A previous commit changed this to pass in the hardware queue, but
it was using the wrong hardware queue. Hence a request that was
allocated on one hardware queue ended up being issued on another
one, and that caused IO timeouts and oopses on some drivers. Since
the request holds hardware queue private resources, like a tag,
we can't just issue it on a different hardware queue.
Fixes: 2253efc850 ("blk-mq: Move more code into blk_mq_direct_issue_request()")
Signed-off-by: Jens Axboe <axboe@fb.com>
Enable throttling of buffered writeback to make it a lot
more smooth, and has way less impact on other system activity.
Background writeback should be, by definition, background
activity. The fact that we flush huge bundles of it at the time
means that it potentially has heavy impacts on foreground workloads,
which isn't ideal. We can't easily limit the sizes of writes that
we do, since that would impact file system layout in the presence
of delayed allocation. So just throttle back buffered writeback,
unless someone is waiting for it.
The algorithm for when to throttle takes its inspiration in the
CoDel networking scheduling algorithm. Like CoDel, blk-wb monitors
the minimum latencies of requests over a window of time. In that
window of time, if the minimum latency of any request exceeds a
given target, then a scale count is incremented and the queue depth
is shrunk. The next monitoring window is shrunk accordingly. Unlike
CoDel, if we hit a window that exhibits good behavior, then we
simply increment the scale count and re-calculate the limits for that
scale value. This prevents us from oscillating between a
close-to-ideal value and max all the time, instead remaining in the
windows where we get good behavior.
Unlike CoDel, blk-wb allows the scale count to to negative. This
happens if we primarily have writes going on. Unlike positive
scale counts, this doesn't change the size of the monitoring window.
When the heavy writers finish, blk-bw quickly snaps back to it's
stable state of a zero scale count.
The patch registers a sysfs entry, 'wb_lat_usec'. This sets the latency
target to me met. It defaults to 2 msec for non-rotational storage, and
75 msec for rotational storage. Setting this value to '0' disables
blk-wb. Generally, a user would not have to touch this setting.
We don't enable WBT on devices that are managed with CFQ, and have
a non-root block cgroup attached. If we have a proportional share setup
on this particular disk, then the wbt throttling will interfere with
that. We don't have a strong need for wbt for that case, since we will
rely on CFQ doing that for us.
Signed-off-by: Jens Axboe <axboe@fb.com>
For legacy block, we simply track them in the request queue. For
blk-mq, we track them on a per-sw queue basis, which we can then
sum up through the hardware queues and finally to a per device
state.
The stats are tracked in, roughly, 0.1s interval windows.
Add sysfs files to display the stats.
The feature is off by default, to avoid any extra overhead. In-kernel
users of it can turn it on by setting QUEUE_FLAG_STATS in the queue
flags. We currently don't turn it on if someone just reads any of
the stats files, that is something we could add as well.
Signed-off-by: Jens Axboe <axboe@fb.com>
Commit 0e87e58bf6 ("blk-mq: improve warning for running a queue on the
wrong CPU") attempts to avoid triggering the WARN_ON in
__blk_mq_run_hw_queue when the expected CPU is dead. Problem is, in the
last batch execution before round robin, blk_mq_hctx_next_cpu can
schedule a dead CPU and also update next_cpu to the next alive CPU in
the mask, which will trigger the WARN_ON despite the previous
workaround.
The following patch fixes this scenario by always scheduling the value
in hctx->next_cpu. This changes the moment when we round-robin the CPU
running the hctx, but it really doesn't matter, since it still executes
BLK_MQ_CPU_WORK_BATCH times in a row before switching to another CPU.
Fixes: 0e87e58bf6 ("blk-mq: improve warning for running a queue on the wrong CPU")
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This is corresponding part for blk-mq. Disk with multiple hardware
queues doesn't need this as we only hold 1 request at most.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Most blk_mq_requeue_request() and blk_mq_add_to_requeue_list() calls
are followed by kicking the requeue list. Hence add an argument to
these two functions that allows to kick the requeue list. This was
proposed by Christoph Hellwig.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@fb.com>
blk_mq_quiesce_queue() waits until ongoing .queue_rq() invocations
have finished. This function does *not* wait until all outstanding
requests have finished (this means invocation of request.end_io()).
The algorithm used by blk_mq_quiesce_queue() is as follows:
* Hold either an RCU read lock or an SRCU read lock around
.queue_rq() calls. The former is used if .queue_rq() does not
block and the latter if .queue_rq() may block.
* blk_mq_quiesce_queue() first calls blk_mq_stop_hw_queues()
followed by synchronize_srcu() or synchronize_rcu(). The latter
call waits for .queue_rq() invocations that started before
blk_mq_quiesce_queue() was called.
* The blk_mq_hctx_stopped() calls that control whether or not
.queue_rq() will be called are called with the (S)RCU read lock
held. This is necessary to avoid race conditions against
blk_mq_quiesce_queue().
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Ming Lei <tom.leiming@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Ming Lei
Bart Van Assche
Omar Sandoval
Nitesh Shetty
Linus Torvalds
Mike Snitzer
Jens Axboe
Christoph Hellwig
Tejun Heo
Johannes Thumshirn
Jianchao Wang
weiping zhang
Shaohua Li
Peter Zijlstra
Keith Busch
Ming Lei
Dmitry Monakhov
Ingo Molnar
Bart Van Assche
Goldwyn Rodrigues
NeilBrown
Wen Xiong
Colin Ian King
Wanpeng Li
Jens Axboe
Stephen Bates
Minchan Kim
Josef Bacik
Ming Lei
Sagi Grimberg
Josef Bacik
Thomas Gleixner
Gabriel Krisman Bertazi
Ming Lei