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>
Move the "hctx stopped" test and the insert request calls into
blk_mq_direct_issue_request(). Rename that function into
blk_mq_try_issue_directly() to reflect its new semantics. Pass
the hctx pointer to that function instead of looking it up a
second time. These changes avoid that code has to be duplicated
in the next patch.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Jens Axboe
Thomas Gleixner
Gabriel Krisman Bertazi
Ming Lei
Shaohua Li
Bart Van Assche