Some functions in block/blk-core.c must only be used on blk-sq queues
while others are safe to use against any queue type. Document which
functions are intended for blk-sq queues and issue a warning if the
blk-sq API is misused. This does not only help block driver authors
but will also make it easier to remove the blk-sq code once that code
is declared obsolete.
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>
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>
blk_insert_flush should be using __blk_end_request to start with.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
User configures latency target, but the latency threshold for each
request size isn't fixed. For a SSD, the IO latency highly depends on
request size. To calculate latency threshold, we sample some data, eg,
average latency for request size 4k, 8k, 16k, 32k .. 1M. The latency
threshold of each request size will be the sample latency (I'll call it
base latency) plus latency target. For example, the base latency for
request size 4k is 80us and user configures latency target 60us. The 4k
latency threshold will be 80 + 60 = 140us.
To sample data, we calculate the order base 2 of rounded up IO sectors.
If the IO size is bigger than 1M, it will be accounted as 1M. Since the
calculation does round up, the base latency will be slightly smaller
than actual value. Also if there isn't any IO dispatched for a specific
IO size, we will use the base latency of smaller IO size for this IO
size.
But we shouldn't sample data at any time. The base latency is supposed
to be latency where disk isn't congested, because we use latency
threshold to schedule IOs between cgroups. If disk is congested, the
latency is higher, using it for scheduling is meaningless. Hence we only
do the sampling when block throttling is in the LOW limit, with
assumption disk isn't congested in such state. If the assumption isn't
true, eg, low limit is too high, calculated latency threshold will be
higher.
Hard disk is completely different. Latency depends on spindle seek
instead of request size. Currently this feature is SSD only, we probably
can use a fixed threshold like 4ms for hard disk though.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
A cgroup gets assigned a low limit, but the cgroup could never dispatch
enough IO to cross the low limit. In such case, the queue state machine
will remain in LIMIT_LOW state and all other cgroups will be throttled
according to low limit. This is unfair for other cgroups. We should
treat the cgroup idle and upgrade the state machine to lower state.
We also have a downgrade logic. If the state machine upgrades because of
cgroup idle (real idle), the state machine will downgrade soon as the
cgroup is below its low limit. This isn't what we want. A more
complicated case is cgroup isn't idle when queue is in LIMIT_LOW. But
when queue gets upgraded to lower state, other cgroups could dispatch
more IO and this cgroup can't dispatch enough IO, so the cgroup is below
its low limit and looks like idle (fake idle). In this case, the queue
should downgrade soon. The key to determine if we should do downgrade is
to detect if cgroup is truely idle.
Unfortunately it's very hard to determine if a cgroup is real idle. This
patch uses the 'think time check' idea from CFQ for the purpose. Please
note, the idea doesn't work for all workloads. For example, a workload
with io depth 8 has disk utilization 100%, hence think time is 0, eg,
not idle. But the workload can run higher bandwidth with io depth 16.
Compared to io depth 16, the io depth 8 workload is idle. We use the
idea to roughly determine if a cgroup is idle.
We treat a cgroup idle if its think time is above a threshold (by
default 1ms for SSD and 100ms for HD). The idea is think time above the
threshold will start to harm performance. HD is much slower so a longer
think time is ok.
The patch (and the latter patches) uses 'unsigned long' to track time.
We convert 'ns' to 'us' with 'ns >> 10'. This is fast but loses
precision, should not a big deal.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The throtl_slice is 100ms by default. This is a long time for SSD, a lot
of IO can run. To make cgroups have smoother throughput, we choose a
small value (20ms) for SSD.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
throtl_slice is important for blk-throttling. It's called slice
internally but it really is a time window blk-throttling samples data.
blk-throttling will make decision based on the samplings. An example is
bandwidth measurement. A cgroup's bandwidth is measured in the time
interval of throtl_slice.
A small throtl_slice meanse cgroups have smoother throughput but burn
more CPUs. It has 100ms default value, which is not appropriate for all
disks. A fast SSD can dispatch a lot of IOs in 100ms. This patch makes
it tunable.
Since throtl_slice isn't a time slice, the sysfs name
'throttle_sample_time' reflects its character better.
Signed-off-by: Shaohua Li <shli@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>
This makes it available outside of blk-merge.c, and inlining such a trivial
helper seems pretty useful to start with.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
When we attempt to merge request-to-request, we return a 0/1 if we
ended up merging or not. Change that to return the pointer to the
request that we freed. We will use this to move the freeing of
that request out of the merge logic, so that callers can drop
locks before freeing the request.
There should be no functional changes in this patch.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Omar Sandoval <osandov@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>
This can be used to check for fs vs non-fs requests and basically
removes all knowledge of BLOCK_PC specific from the block layer,
as well as preparing for removing the cmd_type field in struct request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
We want to use it outside of blk-core.c.
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>
Prep patch for adding MQ ops as well, since doing anon unions with
named initializers doesn't work on older compilers.
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: Omar Sandoval <osandov@fb.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>
A lot of the REQ_* flags are only used on struct requests, and only of
use to the block layer and a few drivers that dig into struct request
internals.
This patch adds a new req_flags_t rq_flags field to struct request for
them, and thus dramatically shrinks the number of common requests. It
also removes the unfortunate situation where we have to fit the fields
from the same enum into 32 bits for struct bio and 64 bits for
struct request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Shaun Tancheff <shaun.tancheff@seagate.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
All drivers use the default, so provide an inline version of it. If we
ever need other queue mapping we can add an optional method back,
although supporting will also require major changes to the queue setup
code.
This provides better code generation, and better debugability as well.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The target SCSI passthrough backend is much better served with the low-level
blk_rq_append_bio construct then the helpers built on top of it, so export it.
Also use the opportunity to remove the pointless request_queue argument and
make the code flow a little more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Timer context is not very useful for drivers to perform any meaningful abort
action from. So instead of calling the driver from this useless context
defer it to a workqueue as soon as possible.
Note that while a delayed_work item would seem the right thing here I didn't
dare to use it due to the magic in blk_add_timer that pokes deep into timer
internals. But maybe this encourages Tejun to add a sensible API for that to
the workqueue API and we'll all be fine in the end :)
Contains a major update from Keith Bush:
"This patch removes synchronizing the timeout work so that the timer can
start a freeze on its own queue. The timer enters the queue, so timer
context can only start a freeze, but not wait for frozen."
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull block integrity updates from Jens Axboe:
""This is the joint work of Dan and Martin, cleaning up and improving
the support for block data integrity"
* 'for-4.4/integrity' of git://git.kernel.dk/linux-block:
block, libnvdimm, nvme: provide a built-in blk_integrity nop profile
block: blk_flush_integrity() for bio-based drivers
block: move blk_integrity to request_queue
block: generic request_queue reference counting
nvme: suspend i/o during runtime blk_integrity_unregister
md: suspend i/o during runtime blk_integrity_unregister
md, dm, scsi, nvme, libnvdimm: drop blk_integrity_unregister() at shutdown
block: Inline blk_integrity in struct gendisk
block: Export integrity data interval size in sysfs
block: Reduce the size of struct blk_integrity
block: Consolidate static integrity profile properties
block: Move integrity kobject to struct gendisk
Request queues with merging disabled will not flush the plug list after
BLK_MAX_REQUEST_COUNT requests have been queued, since the code relies
on blk_attempt_plug_merge to compute the request_count. Fix this by
computing the number of queued requests even for nomerge queues.
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Since they lack requests to pin the request_queue active, synchronous
bio-based drivers may have in-flight integrity work from
bio_integrity_endio() that is not flushed by blk_freeze_queue(). Flush
that work to prevent races to free the queue and the final usage of the
blk_integrity profile.
This is temporary unless/until bio-based drivers start to generically
take a q_usage_counter reference while a bio is in-flight.
Cc: Martin K. Petersen <martin.petersen@oracle.com>
[martin: fix the CONFIG_BLK_DEV_INTEGRITY=n case]
Tested-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Allow pmem, and other synchronous/bio-based block drivers, to fallback
on a per-cpu reference count managed by the core for tracking queue
live/dead state.
The existing per-cpu reference count for the blk_mq case is promoted to
be used in all block i/o scenarios. This involves initializing it by
default, waiting for it to drop to zero at exit, and holding a live
reference over the invocation of q->make_request_fn() in
generic_make_request(). The blk_mq code continues to take its own
reference per blk_mq request and retains the ability to freeze the
queue, but the check that the queue is frozen is moved to
generic_make_request().
This fixes crash signatures like the following:
BUG: unable to handle kernel paging request at ffff880140000000
[..]
Call Trace:
[<ffffffff8145e8bf>] ? copy_user_handle_tail+0x5f/0x70
[<ffffffffa004e1e0>] pmem_do_bvec.isra.11+0x70/0xf0 [nd_pmem]
[<ffffffffa004e331>] pmem_make_request+0xd1/0x200 [nd_pmem]
[<ffffffff811c3162>] ? mempool_alloc+0x72/0x1a0
[<ffffffff8141f8b6>] generic_make_request+0xd6/0x110
[<ffffffff8141f966>] submit_bio+0x76/0x170
[<ffffffff81286dff>] submit_bh_wbc+0x12f/0x160
[<ffffffff81286e62>] submit_bh+0x12/0x20
[<ffffffff813395bd>] jbd2_write_superblock+0x8d/0x170
[<ffffffff8133974d>] jbd2_mark_journal_empty+0x5d/0x90
[<ffffffff813399cb>] jbd2_journal_destroy+0x24b/0x270
[<ffffffff810bc4ca>] ? put_pwq_unlocked+0x2a/0x30
[<ffffffff810bc6f5>] ? destroy_workqueue+0x225/0x250
[<ffffffff81303494>] ext4_put_super+0x64/0x360
[<ffffffff8124ab1a>] generic_shutdown_super+0x6a/0xf0
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Suggested-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull blk-cg updates from Jens Axboe:
"A bit later in the cycle, but this has been in the block tree for a a
while. This is basically four patchsets from Tejun, that improve our
buffered cgroup writeback. It was dependent on the other cgroup
changes, but they went in earlier in this cycle.
Series 1 is set of 5 patches that has cgroup writeback updates:
- bdi_writeback iteration fix which could lead to some wb's being
skipped or repeated during e.g. sync under memory pressure.
- Simplification of wb work wait mechanism.
- Writeback tracepoints updated to report cgroup.
Series 2 is is a set of updates for the CFQ cgroup writeback handling:
cfq has always charged all async IOs to the root cgroup. It didn't
have much choice as writeback didn't know about cgroups and there
was no way to tell who to blame for a given writeback IO.
writeback finally grew support for cgroups and now tags each
writeback IO with the appropriate cgroup to charge it against.
This patchset updates cfq so that it follows the blkcg each bio is
tagged with. Async cfq_queues are now shared across cfq_group,
which is per-cgroup, instead of per-request_queue cfq_data. This
makes all IOs follow the weight based IO resource distribution
implemented by cfq.
- Switched from GFP_ATOMIC to GFP_NOWAIT as suggested by Jeff.
- Other misc review points addressed, acks added and rebased.
Series 3 is the blkcg policy cleanup patches:
This patchset contains assorted cleanups for blkcg_policy methods
and blk[c]g_policy_data handling.
- alloc/free added for blkg_policy_data. exit dropped.
- alloc/free added for blkcg_policy_data.
- blk-throttle's async percpu allocation is replaced with direct
allocation.
- all methods now take blk[c]g_policy_data instead of blkcg_gq or
blkcg.
And finally, series 4 is a set of patches cleaning up the blkcg stats
handling:
blkcg's stats have always been somwhat of a mess. This patchset
tries to improve the situation a bit.
- The following patches added to consolidate blkcg entry point and
blkg creation. This is in itself is an improvement and helps
colllecting common stats on bio issue.
- per-blkg stats now accounted on bio issue rather than request
completion so that bio based and request based drivers can behave
the same way. The issue was spotted by Vivek.
- cfq-iosched implements custom recursive stats and blk-throttle
implements custom per-cpu stats. This patchset make blkcg core
support both by default.
- cfq-iosched and blk-throttle keep track of the same stats
multiple times. Unify them"
* 'for-4.3/blkcg' of git://git.kernel.dk/linux-block: (45 commits)
blkcg: use CGROUP_WEIGHT_* scale for io.weight on the unified hierarchy
blkcg: s/CFQ_WEIGHT_*/CFQ_WEIGHT_LEGACY_*/
blkcg: implement interface for the unified hierarchy
blkcg: misc preparations for unified hierarchy interface
blkcg: separate out tg_conf_updated() from tg_set_conf()
blkcg: move body parsing from blkg_conf_prep() to its callers
blkcg: mark existing cftypes as legacy
blkcg: rename subsystem name from blkio to io
blkcg: refine error codes returned during blkcg configuration
blkcg: remove unnecessary NULL checks from __cfqg_set_weight_device()
blkcg: reduce stack usage of blkg_rwstat_recursive_sum()
blkcg: remove cfqg_stats->sectors
blkcg: move io_service_bytes and io_serviced stats into blkcg_gq
blkcg: make blkg_[rw]stat_recursive_sum() to be able to index into blkcg_gq
blkcg: make blkcg_[rw]stat per-cpu
blkcg: add blkg_[rw]stat->aux_cnt and replace cfq_group->dead_stats with it
blkcg: consolidate blkg creation in blkcg_bio_issue_check()
blk-throttle: improve queue bypass handling
blkcg: move root blkg lookup optimization from throtl_lookup_tg() to __blkg_lookup()
blkcg: inline [__]blkg_lookup()
...
blkg (blkcg_gq) currently is created by blkcg policies invoking
blkg_lookup_create() which ends up repeating about the same code in
different policies. Theoretically, this can avoid the overhead of
looking and/or creating blkg's if blkcg is enabled but no policy is in
use; however, the cost of blkg lookup / creation is very low
especially if only the root blkcg is in use which is highly likely if
no blkcg policy is in active use - it boils down to a single very
predictable conditional and surrounding RCU protection.
This patch consolidates blkg creation to a new function
blkcg_bio_issue_check() which is called during bio issue from
generic_make_request_checks(). blkcg_bio_issue_check() is now the
only function which tries to create missing blkg's. The subsequent
policy and request_list operations just perform blkg_lookup() and if
missing falls back to the root.
* blk_get_rl() no longer tries to create blkg. It uses blkg_lookup()
instead of blkg_lookup_create().
* blk_throtl_bio() is now called from blkcg_bio_issue_check() with rcu
read locked and blkg already looked up. Both throtl_lookup_tg() and
throtl_lookup_create_tg() are dropped.
* cfq is similarly updated. cfq_lookup_create_cfqg() is replaced with
cfq_lookup_cfqg()which uses blkg_lookup().
This consolidates blkg handling and avoids unnecessary blkg creation
retries under memory pressure. In addition, this provides a common
bio entry point into blkcg where things like common accounting can be
performed.
v2: Build fixes for !CONFIG_CFQ_GROUP_IOSCHED and
!CONFIG_BLK_DEV_THROTTLING.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Last patch makes plug work for multiple queue case. However it only
works for single disk case, because it assumes only one request in the
plug list. If a task is accessing multiple disks, eg MD/DM, the
assumption is wrong. Let blk_attempt_plug_merge() record request from
the same queue.
V2: use NULL parameter in !mq case. Fix a bug. Add comments in
blk_attempt_plug_merge to make it less (hopefully) confusion.
Cc: Jens Axboe <axboe@fb.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This removes the request types and hacks from the block code and into the
old IDE driver. There is a small amunt of code duplication due to this,
but it's not too bad.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch supports to run one single flush machinery for
each blk-mq dispatch queue, so that:
- current init_request and exit_request callbacks can
cover flush request too, then the buggy copying way of
initializing flush request's pdu can be fixed
- flushing performance gets improved in case of multi hw-queue
In fio sync write test over virtio-blk(4 hw queues, ioengine=sync,
iodepth=64, numjobs=4, bs=4K), it is observed that througput gets
increased a lot over my test environment:
- throughput: +70% in case of virtio-blk over null_blk
- throughput: +30% in case of virtio-blk over SSD image
The multi virtqueue feature isn't merged to QEMU yet, and patches for
the feature can be found in below tree:
git://kernel.ubuntu.com/ming/qemu.git v2.1.0-mq.4
And simply passing 'num_queues=4 vectors=5' should be enough to
enable multi queue(quad queue) feature for QEMU virtio-blk.
Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch adds 'blk_mq_ctx' parameter to blk_get_flush_queue(),
so that this function can find the corresponding blk_flush_queue
bound with current mq context since the flush queue will become
per hw-queue.
For legacy queue, the parameter can be simply 'NULL'.
For multiqueue case, the parameter should be set as the context
from which the related request is originated. With this context
info, the hw queue and related flush queue can be found easily.
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Now mission of the two helpers is over, and just call
blk_alloc_flush_queue() and blk_free_flush_queue() directly.
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch introduces 'struct blk_flush_queue' and puts all
flush machinery related fields into this structure, so that
- flush implementation details aren't exposed to driver
- it is easy to convert to per dispatch-queue flush machinery
This patch is basically a mechanical replacement.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
These two temporary functions are introduced for holding flush
initialization and de-initialization, so that we can
introduce 'flush queue' easier in the following patch. And
once 'flush queue' and its allocation/free functions are ready,
they will be removed for sake of code readability.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Duplicate the (small) timeout handler in blk-mq so that we can pass
arguments more easily to the driver timeout handler. This enables
the next patch.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
elv_abort_queue has no callers, and blk_abort_flushes is only called by
elv_abort_queue.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
For request_fn based devices, the block layer exports a 'nr_requests'
file through sysfs to allow adjusting of queue depth on the fly.
Currently this returns -EINVAL for blk-mq, since it's not wired up.
Wire this up for blk-mq, so that it now also always dynamic
adjustments of the allowed queue depth for any given block device
managed by blk-mq.
Signed-off-by: Jens Axboe <axboe@fb.com>
This adds support for active queue tracking, meaning that the
blk-mq tagging maintains a count of active users of a tag set.
This allows us to maintain a notion of fairness between users,
so that we can distribute the tag depth evenly without starving
some users while allowing others to try unfair deep queues.
If sharing of a tag set is detected, each hardware queue will
track the depth of its own queue. And if this exceeds the total
depth divided by the number of active queues, the user is actively
throttled down.
The active queue count is done lazily to avoid bouncing that data
between submitter and completer. Each hardware queue gets marked
active when it allocates its first tag, and gets marked inactive
when 1) the last tag is cleared, and 2) the queue timeout grace
period has passed.
Signed-off-by: Jens Axboe <axboe@fb.com>
If a requeue event races with a timeout, we can get into the
situation where we attempt to complete a request from the
timeout handler when it's not start anymore. This causes a crash.
So have the timeout handler check that REQ_ATOM_STARTED is still
set on the request - if not, we ignore the event. If this happens,
the request has now been marked as complete. As a consequence, we
need to ensure to clear REQ_ATOM_COMPLETE in blk_mq_start_request(),
as to maintain proper request state.
Signed-off-by: Jens Axboe <axboe@fb.com>
Martin reported that his test system would not boot with
current git, it oopsed with this:
BUG: unable to handle kernel paging request at ffff88046c6c9e80
IP: [<ffffffff812971e0>] blk_queue_start_tag+0x90/0x150
PGD 1ddf067 PUD 1de2067 PMD 47fc7d067 PTE 800000046c6c9060
Oops: 0002 [#1] SMP DEBUG_PAGEALLOC
Modules linked in: sd_mod lpfc(+) scsi_transport_fc scsi_tgt oracleasm
rpcsec_gss_krb5 ipv6 igb dca i2c_algo_bit i2c_core hwmon
CPU: 3 PID: 87 Comm: kworker/u17:1 Not tainted 3.14.0+ #246
Hardware name: Supermicro X9DRX+-F/X9DRX+-F, BIOS 3.00 07/09/2013
Workqueue: events_unbound async_run_entry_fn
task: ffff8802743c2150 ti: ffff880273d02000 task.ti: ffff880273d02000
RIP: 0010:[<ffffffff812971e0>] [<ffffffff812971e0>]
blk_queue_start_tag+0x90/0x150
RSP: 0018:ffff880273d03a58 EFLAGS: 00010092
RAX: ffff88046c6c9e78 RBX: ffff880077208e78 RCX: 00000000fffc8da6
RDX: 00000000fffc186d RSI: 0000000000000009 RDI: 00000000fffc8d9d
RBP: ffff880273d03a88 R08: 0000000000000001 R09: ffff8800021c2410
R10: 0000000000000005 R11: 0000000000015b30 R12: ffff88046c5bb8a0
R13: ffff88046c5c0890 R14: 000000000000001e R15: 000000000000001e
FS: 0000000000000000(0000) GS:ffff880277b00000(0000)
knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffff88046c6c9e80 CR3: 00000000018f6000 CR4: 00000000000407e0
Stack:
ffff880273d03a98 ffff880474b18800 0000000000000000 ffff880474157000
ffff88046c5c0890 ffff880077208e78 ffff880273d03ae8 ffffffff813b9e62
ffff880200000010 ffff880474b18968 ffff880474b18848 ffff88046c5c0cd8
Call Trace:
[<ffffffff813b9e62>] scsi_request_fn+0xf2/0x510
[<ffffffff81293167>] __blk_run_queue+0x37/0x50
[<ffffffff8129ac43>] blk_execute_rq_nowait+0xb3/0x130
[<ffffffff8129ad24>] blk_execute_rq+0x64/0xf0
[<ffffffff8108d2b0>] ? bit_waitqueue+0xd0/0xd0
[<ffffffff813bba35>] scsi_execute+0xe5/0x180
[<ffffffff813bbe4a>] scsi_execute_req_flags+0x9a/0x110
[<ffffffffa01b1304>] sd_spinup_disk+0x94/0x460 [sd_mod]
[<ffffffff81160000>] ? __unmap_hugepage_range+0x200/0x2f0
[<ffffffffa01b2b9a>] sd_revalidate_disk+0xaa/0x3f0 [sd_mod]
[<ffffffffa01b2fb8>] sd_probe_async+0xd8/0x200 [sd_mod]
[<ffffffff8107703f>] async_run_entry_fn+0x3f/0x140
[<ffffffff8106a1c5>] process_one_work+0x175/0x410
[<ffffffff8106b373>] worker_thread+0x123/0x400
[<ffffffff8106b250>] ? manage_workers+0x160/0x160
[<ffffffff8107104e>] kthread+0xce/0xf0
[<ffffffff81070f80>] ? kthread_freezable_should_stop+0x70/0x70
[<ffffffff815f0bac>] ret_from_fork+0x7c/0xb0
[<ffffffff81070f80>] ? kthread_freezable_should_stop+0x70/0x70
Code: 48 0f ab 11 72 db 48 81 4b 40 00 00 10 00 89 83 08 01 00 00 48 89
df 49 8b 04 24 48 89 1c d0 e8 f7 a8 ff ff 49 8b 85 28 05 00 00 <48> 89
58 08 48 89 03 49 8d 85 28 05 00 00 48 89 43 08 49 89 9d
RIP [<ffffffff812971e0>] blk_queue_start_tag+0x90/0x150
RSP <ffff880273d03a58>
CR2: ffff88046c6c9e80
Martin bisected and found this to be the problem patch;
commit 6d113398dc
Author: Jan Kara <jack@suse.cz>
Date: Mon Feb 24 16:39:54 2014 +0100
block: Stop abusing rq->csd.list in blk-softirq
and the problem was immediately apparent. The patch states that
it is safe to reuse queuelist at completion time, since it is
no longer used. However, that is not true if a device is using
block enabled tagging. If that is the case, then the queuelist
is reused to keep track of busy tags. If a device also ended
up using softirq completions, we'd reuse ->queuelist for the
IPI handling while block tagging was still using it. Boom.
Fix this by adding a new ipi_list list head, and share the
memory used with the request hash table. The hash table is
never used after the request is moved to the dispatch list,
which happens long before any potential completion of the
request. Add a new request bit for this, so we don't have
cases that check rq->hash while it could potentially have
been reused for the IPI completion.
Reported-by: Martin K. Petersen <martin.petersen@oracle.com>
Tested-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
request_queue bypassing is used to suppress higher-level function of a
request_queue so that they can be switched, reconfigured and shut
down. A request_queue does the followings while bypassing.
* bypasses elevator and io_cq association and queues requests directly
to the FIFO dispatch queue.
* bypasses block cgroup request_list lookup and always uses the root
request_list.
Once confirmed to be bypassing, specific elevator and block cgroup
policy implementations can assume that nothing is in flight for them
and perform various operations which would be dangerous otherwise.
Such confirmation is acheived by short-circuiting all new requests
directly to the dispatch queue and waiting for all the requests which
were issued before to finish. Unfortunately, while the request
allocating and draining sides were properly handled, we forgot to
actually plug the request dispatch path. Even after bypassing mode is
confirmed, if the attached driver tries to fetch a request and the
dispatch queue is empty, __elv_next_request() would invoke the current
elevator's elevator_dispatch_fn() callback. As all in-flight requests
were drained, the elevator wouldn't contain any request but once
bypass is confirmed we don't even know whether the elevator is even
there. It might be in the process of being switched and half torn
down.
Frank Mayhar reports that this actually happened while switching
elevators, leading to an oops.
Let's fix it by making __elv_next_request() avoid invoking the
elevator_dispatch_fn() callback if the queue is bypassing. It already
avoids invoking the callback if the queue is dying. As a dying queue
is guaranteed to be bypassing, we can simply replace blk_queue_dying()
check with blk_queue_bypass().
Reported-by: Frank Mayhar <fmayhar@google.com>
References: http://lkml.kernel.org/g/1390319905.20232.38.camel@bobble.lax.corp.google.com
Cc: stable@vger.kernel.org
Tested-by: Frank Mayhar <fmayhar@google.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Linux currently has two models for block devices:
- The classic request_fn based approach, where drivers use struct
request units for IO. The block layer provides various helper
functionalities to let drivers share code, things like tag
management, timeout handling, queueing, etc.
- The "stacked" approach, where a driver squeezes in between the
block layer and IO submitter. Since this bypasses the IO stack,
driver generally have to manage everything themselves.
With drivers being written for new high IOPS devices, the classic
request_fn based driver doesn't work well enough. The design dates
back to when both SMP and high IOPS was rare. It has problems with
scaling to bigger machines, and runs into scaling issues even on
smaller machines when you have IOPS in the hundreds of thousands
per device.
The stacked approach is then most often selected as the model
for the driver. But this means that everybody has to re-invent
everything, and along with that we get all the problems again
that the shared approach solved.
This commit introduces blk-mq, block multi queue support. The
design is centered around per-cpu queues for queueing IO, which
then funnel down into x number of hardware submission queues.
We might have a 1:1 mapping between the two, or it might be
an N:M mapping. That all depends on what the hardware supports.
blk-mq provides various helper functions, which include:
- Scalable support for request tagging. Most devices need to
be able to uniquely identify a request both in the driver and
to the hardware. The tagging uses per-cpu caches for freed
tags, to enable cache hot reuse.
- Timeout handling without tracking request on a per-device
basis. Basically the driver should be able to get a notification,
if a request happens to fail.
- Optional support for non 1:1 mappings between issue and
submission queues. blk-mq can redirect IO completions to the
desired location.
- Support for per-request payloads. Drivers almost always need
to associate a request structure with some driver private
command structure. Drivers can tell blk-mq this at init time,
and then any request handed to the driver will have the
required size of memory associated with it.
- Support for merging of IO, and plugging. The stacked model
gets neither of these. Even for high IOPS devices, merging
sequential IO reduces per-command overhead and thus
increases bandwidth.
For now, this is provided as a potential 3rd queueing model, with
the hope being that, as it matures, it can replace both the classic
and stacked model. That would get us back to having just 1 real
model for block devices, leaving the stacked approach to dm/md
devices (as it was originally intended).
Contributions in this patch from the following people:
Shaohua Li <shli@fusionio.com>
Alexander Gordeev <agordeev@redhat.com>
Christoph Hellwig <hch@infradead.org>
Mike Christie <michaelc@cs.wisc.edu>
Matias Bjorling <m@bjorling.me>
Jeff Moyer <jmoyer@redhat.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Switch elevator to use the new hashtable implementation. This reduces the
amount of generic unrelated code in the elevator.
This also removes the dymanic allocation of the hash table. The size of the table is
constant so there's no point in paying the price of an extra dereference when accessing
it.
This patch depends on d9b482c ("hashtable: introduce a small and naive
hashtable") which was merged in v3.6.
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A block driver may start cleaning up resources needed by its
request_fn as soon as blk_cleanup_queue() finished, so request_fn
must not be invoked after draining finished. This is important
when blk_run_queue() is invoked without any requests in progress.
As an example, if blk_drain_queue() and scsi_run_queue() run in
parallel, blk_drain_queue() may have finished all requests after
scsi_run_queue() has taken a SCSI device off the starved list but
before that last function has had a chance to run the queue.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Cc: James Bottomley <JBottomley@Parallels.com>
Cc: Mike Christie <michaelc@cs.wisc.edu>
Cc: Chanho Min <chanho.min@lge.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
QUEUE_FLAG_DEAD is used to indicate that queuing new requests must
stop. After this flag has been set queue draining starts. However,
during the queue draining phase it is still safe to invoke the
queue's request_fn, so QUEUE_FLAG_DYING is a better name for this
flag.
This patch has been generated by running the following command
over the kernel source tree:
git grep -lEw 'blk_queue_dead|QUEUE_FLAG_DEAD' |
xargs sed -i.tmp -e 's/blk_queue_dead/blk_queue_dying/g' \
-e 's/QUEUE_FLAG_DEAD/QUEUE_FLAG_DYING/g'; \
sed -i.tmp -e "s/QUEUE_FLAG_DYING$(printf \\t)*5/QUEUE_FLAG_DYING$(printf \\t)5/g" \
include/linux/blkdev.h; \
sed -i.tmp -e 's/ DEAD/ DYING/g' -e 's/dead queue/a dying queue/' \
-e 's/Dead queue/A dying queue/' block/blk-core.c
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: James Bottomley <JBottomley@Parallels.com>
Cc: Mike Christie <michaelc@cs.wisc.edu>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Chanho Min <chanho.min@lge.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove special-casing of non-rw fs style requests (discard). The nomerge
flags are consolidated in blk_types.h, and rq_mergeable() and
bio_mergeable() have been modified to use them.
bio_is_rw() is used in place of bio_has_data() a few places. This is
done to to distinguish true reads and writes from other fs type requests
that carry a payload (e.g. write same).
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
__generic_unplug_device() function is removed with commit
7eaceaccab, which forgot to
remove the declaration at meantime. Here remove it.
Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
