Introduce the BLK_MQ_REQ_PM flag. This flag makes the request allocation
functions set RQF_PM. This is the first step towards removing
BLK_MQ_REQ_PREEMPT.

Link: https://lore.kernel.org/r/20201209052951.16136-3-bvanassche@acm.org
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Stanley Chu <stanley.chu@mediatek.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Can Guo <cang@codeaurora.org>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NJens Axboe <axboe@kernel.dk>
Reviewed-by: NCan Guo <cang@codeaurora.org>
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

With the current implementation the following race can happen:

 * blk_pre_runtime_suspend() calls blk_freeze_queue_start() and
   blk_mq_unfreeze_queue().

 * blk_queue_enter() calls blk_queue_pm_only() and that function returns
   true.

 * blk_queue_enter() calls blk_pm_request_resume() and that function does
   not call pm_request_resume() because the queue runtime status is
   RPM_ACTIVE.

 * blk_pre_runtime_suspend() changes the queue status into RPM_SUSPENDING.

Fix this race by changing the queue runtime status into RPM_SUSPENDING
before switching q_usage_counter to atomic mode.

Link: https://lore.kernel.org/r/20201209052951.16136-2-bvanassche@acm.org
Fixes: 986d413b ("blk-mq: Enable support for runtime power management")
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: stable <stable@vger.kernel.org>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NJens Axboe <axboe@kernel.dk>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Acked-by: NStanley Chu <stanley.chu@mediatek.com>
Co-developed-by: NCan Guo <cang@codeaurora.org>
Signed-off-by: NCan Guo <cang@codeaurora.org>
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

blk_mq_end_request() will use the block status returned from queue_rq() as
argument, except in one instance in blk_mq_dispatch_rq_list(), where the
generic BLK_STS_IOERR is used.

Link: https://lore.kernel.org/r/20200930080256.90964-2-hare@suse.deReviewed-by: NEwan D. Milne <emilne@redhat.com>
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Fix a typo:
	blk_mq_run_hw_queue -> blk_mq_run_hw_queues
Signed-off-by: NMauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We don't need to check whether the node is memoryless numa node before
calling allocator interface. SLUB(and SLAB,SLOB) relies on the page
allocator to pick a node. Page allocator should deal with memoryless
nodes just fine. It has zonelists constructed for each possible nodes.
And it will automatically fall back into a node which is closest to the
requested node. As long as __GFP_THISNODE is not enforced of course.

The code comments of kmem_cache_alloc_node() of SLAB also showed this:
 * Fallback to other node is possible if __GFP_THISNODE is not set.

blk-mq code doesn't set __GFP_THISNODE, so we can remove the calling
of local_memory_node().
Signed-off-by: NXianting Tian <tian.xianting@h3c.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Fix this warning:

	./block/bio.c:1098: WARNING: Inline emphasis start-string without end-string.

The thing is that *iter is not a valid markup.

That seems to be a typo:
	*iter -> @iter
Signed-off-by: NMauro Carvalho Chehab <mchehab+huawei@kernel.org>

Using "@bio's parent" causes the following waring:
	./block/bio.c:10: WARNING: Inline emphasis start-string without end-string.

The main problem here is that this would be converted into:

	**bio**'s parent

By kernel-doc, which is not a valid notation. It would be
possible to use, instead, this kernel-doc markup:

	``bio's`` parent

Yet, here, is probably simpler to just use an altenative language:

	the parent of @bioSigned-off-by: NMauro Carvalho Chehab <mchehab+huawei@kernel.org>

A zoned device with limited resources to open or activate zones may
return an error when the host exceeds those limits. The same command may
be successful if retried later, but the host needs to wait for specific
zone states before it should expect a retry to succeed. Have the block
layer provide an appropriate status for these conditions so applications
can distinuguish this error for special handling.

Cc: linux-api@vger.kernel.org
Cc: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDamien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: NJohannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: NKeith Busch <kbusch@kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

blk_exit_queue will free elevator_data, while blk_mq_run_work_fn
will access it. Move cancel of hctx->run_work to the front of
blk_exit_queue to avoid use-after-free.

Fixes: 1b97871b ("blk-mq: move cancel of hctx->run_work into blk_mq_hw_sysfs_release")
Signed-off-by: NYang Yang <yang.yang@vivo.com>
Reviewed-by: NMing Lei <ming.lei@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

After commit 923218f6 ("blk-mq: don't allocate driver tag upfront
for flush rq"), blk_mq_submit_bio() will call blk_insert_flush()
directly to handle flush request rather than blk_mq_sched_insert_request()
in the case of elevator.

Then, all flush request either have set RQF_FLUSH_SEQ flag when call
blk_mq_sched_insert_request(), or have inserted into hctx->dispatch.
So, remove the dead code path.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Since whole elevator register is protectd by sysfs_lock, we
don't need extras 'has_elevator'. Just use q->elevator directly.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

After commit b89f625e ("block: don't release queue's sysfs
lock during switching elevator"), whole elevator register and
unregister function are covered by sysfs_lock. So, remove wrong
comment and add lockdep assert.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We have introduced helper function blk_mq_hctx_stopped() to test
BLK_MQ_S_STOPPED.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We have defined common interface blk_queue_registered() to
test QUEUE_FLAG_REGISTERED. Just use it.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

elv_support_iosched() will check queue_is_mq() for us. So, remove
the redundant check to clean code.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We will register debugfs for scheduler no matter whether it have
defined callback funciton .exit_sched. So, blk_mq_exit_sched()
is always needed to unregister debugfs. Also, q->elevator should
be set as NULL after exiting scheduler.

For now, since all register scheduler have defined .exit_sched,
it will not cause any actual problem. But It will be more reasonable
to do this change.
Signed-off-by: NYufen Yu <yuyufen@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

syzbot is reporting unkillable task [1], for the caller is failing to
handle a corrupted filesystem image which attempts to access beyond
the end of the device. While we need to fix the caller, flooding the
console with handle_bad_sector() message is unlikely useful.

[1] https://syzkaller.appspot.com/bug?id=f1f49fb971d7a3e01bd8ab8cff2ff4572ccf3092Signed-off-by: NTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Re-use throtl_set_slice_end() to remove duplicate code.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The __throtl_de/enqueue_tg() functions are only be called by
throtl_de/enqueue_tg(), thus we can just open code them to
make code more readable.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The throtl_schedule_next_dispatch() will validate if the service queue
is empty before calling update_min_dispatch_time(), and the
update_min_dispatch_time() will call throtl_rb_first(), which will
validate service queue again.

Thus we can move the service queue validation out of the
throtl_rb_first() to remove the redundant validation in the fast path.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We should move the list operation after validation.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

It can not scale up in throtl_adjusted_limit() if we set bps or iops is
1, which will cause IO hang when enable low limit. Thus we should treat
1 as a illegal value to avoid this issue.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The IO latency tracking is only for LOW limit, so we should add a
validation to avoid redundant latency tracking if the LOW limit
is not valid.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We only update the tg->last_finish_time when the low limitaion is
enabled, so we can move the tg->last_finish_time validation a little
forward to avoid getting the unnecessary current time stamp if the
the low limitation is not enabled.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The throtl_downgrade_state() is always used to change to LIMIT_LOW
limitation, thus remove the latter meaningless parameter which
indicates the limitation index.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Remove redundant 'return' statement for 'void' functions.
Signed-off-by: NBaolin Wang <baolin.wang@linux.alibaba.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

It is unnecessary to force request-based DM to call into bio-based
dm_submit_bio (via indirect disk->fops->submit_bio) only to have it then
call blk_mq_submit_bio().

Fix this by establishing a request-based DM block_device_operations
(dm_rq_blk_dops, which doesn't have .submit_bio) and update
dm_setup_md_queue() to set md->disk->fops to it for
DM_TYPE_REQUEST_BASED.

Remove DM_TYPE_REQUEST_BASED conditional in dm_submit_bio and unexport
blk_mq_submit_bio.

Fixes: c62b37d9 ("block: move ->make_request_fn to struct block_device_operations")
Signed-off-by: NMike Snitzer <snitzer@redhat.com>

Don't error out if the dasd_biodasdinfo symbol is not available.

Cc: stable@vger.kernel.org
Fixes: 26d7e28e ("s390/dasd: remove ioctl_by_bdev calls")
Reported-by: NChristian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Tested-by: NChristian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: NStefan Haberland <sth@linux.ibm.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

According to Documentation/block/stat.rst, inflight should not include
I/O requests that are in the queue but not yet dispatched to the device,
but blk-mq identifies as inflight any request that has a tag allocated,
which, for queues without elevator, happens at request allocation time
and before it is queued in the ctx (default case in blk_mq_submit_bio).

In addition, current behavior is different for queues with elevator from
queues without it, since for the former the driver tag is allocated at
dispatch time.  A more precise approach would be to only consider
requests with state MQ_RQ_IN_FLIGHT.

This effectively reverts commit 6131837b ("blk-mq: count allocated
but not started requests in iostats inflight") to consolidate blk-mq
behavior with itself (elevator case) and with original documentation,
but it differs from the behavior used by the legacy path.

This version differs from v1 by using blk_mq_rq_state to access the
state attribute.  Avoid using blk_mq_request_started, which was
suggested, since we don't want to include MQ_RQ_COMPLETE.
Signed-off-by: NGabriel Krisman Bertazi <krisman@collabora.com>
Cc: Omar Sandoval <osandov@fb.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Move blk_mq_sched_try_merge to blk-merge.c, which allows to mark
a lot of the merge infrastructure static there.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Also move the definition from the public blkdev.h to the private
block/blk.h header.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Also move the definition from the public blkdev.h to the private
block/blk.h header.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

bio_crypt_set_ctx() assumes its gfp_mask argument always includes
__GFP_DIRECT_RECLAIM, so that the mempool_alloc() will always succeed.

For now this assumption is still fine, since no callers violate it.
Making bio_crypt_set_ctx() able to fail would add unneeded complexity.

However, if a caller didn't use __GFP_DIRECT_RECLAIM, it would be very
hard to notice the bug.  Make it easier by adding a WARN_ON_ONCE().
Signed-off-by: NEric Biggers <ebiggers@google.com>
Reviewed-by: NSatya Tangirala <satyat@google.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Satya Tangirala <satyat@google.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

blk_crypto_rq_bio_prep() assumes its gfp_mask argument always includes
__GFP_DIRECT_RECLAIM, so that the mempool_alloc() will always succeed.

However, blk_crypto_rq_bio_prep() might be called with GFP_ATOMIC via
setup_clone() in drivers/md/dm-rq.c.

This case isn't currently reachable with a bio that actually has an
encryption context.  However, it's fragile to rely on this.  Just make
blk_crypto_rq_bio_prep() able to fail.
Suggested-by: NSatya Tangirala <satyat@google.com>
Signed-off-by: NEric Biggers <ebiggers@google.com>
Reviewed-by: NMike Snitzer <snitzer@redhat.com>
Reviewed-by: NSatya Tangirala <satyat@google.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

bio_crypt_clone() assumes its gfp_mask argument always includes
__GFP_DIRECT_RECLAIM, so that the mempool_alloc() will always succeed.

However, bio_crypt_clone() might be called with GFP_ATOMIC via
setup_clone() in drivers/md/dm-rq.c, or with GFP_NOWAIT via
kcryptd_io_read() in drivers/md/dm-crypt.c.

Neither case is currently reachable with a bio that actually has an
encryption context.  However, it's fragile to rely on this.  Just make
bio_crypt_clone() able to fail, analogous to bio_integrity_clone().
Reported-by: NMiaohe Lin <linmiaohe@huawei.com>
Signed-off-by: NEric Biggers <ebiggers@google.com>
Reviewed-by: NMike Snitzer <snitzer@redhat.com>
Reviewed-by: NSatya Tangirala <satyat@google.com>
Cc: Satya Tangirala <satyat@google.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

All remaining callers of bdget() outside of fs/block_dev.c want to get a
reference to the struct block_device for a given struct hd_struct.  Add
a helper just for that and then mark bdget static.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Use in compat_syscall to import either native or the compat iovecs, and
remove the now superflous compat_import_iovec.

This removes the need for special compat logic in most callers, and
the remaining ones can still be simplified by using __import_iovec
with a bool compat parameter.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

One-element arrays are being deprecated[1]. Replace the one-element array
with a simple object of type compat_caddr_t: 'compat_caddr_t unused'[2],
once it seems this field is actually never used.

Also, update struct cdrom_generic_command in UAPI by adding an
anonimous union to avoid using the one-element array _reserved_.

[1] https://www.kernel.org/doc/html/v5.9-rc1/process/deprecated.html#zero-length-and-one-element-arrays
[2] https://github.com/KSPP/linux/issues/86Signed-off-by: NGustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/lkml/5f76f5d0.qJ4t%2FHWuRzSW7bTa%25lkp@intel.com/Build-tested-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

scsi_put_cdrom_generic_arg() is copying uninitialized stack memory to
userspace, since the compiler may leave a 3-byte hole in the middle of
`cgc32`. Fix it by adding a padding field to `struct
compat_cdrom_generic_command`.

Cc: stable@vger.kernel.org
Fixes: f3ee6e63 ("compat_ioctl: move CDROM_SEND_PACKET handling into scsi")
Suggested-by: NDan Carpenter <dan.carpenter@oracle.com>
Suggested-by: NArnd Bergmann <arnd@arndb.de>
Reported-by: syzbot+85433a479a646a064ab3@syzkaller.appspotmail.com
Signed-off-by: NPeilin Ye <yepeilin.cs@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

'f5bbbbe4 ("blk-mq: sync the update nr_hw_queues with
blk_mq_queue_tag_busy_iter")' introduce a bug what we may sleep between
rcu lock. Then '530ca2c9 ("blk-mq: Allow blocking queue tag iter
callbacks")' fix it by get request_queue's ref. And 'a9a80808 ("block:
Remove the synchronize_rcu() call from __blk_mq_update_nr_hw_queues()")'
remove the synchronize_rcu in __blk_mq_update_nr_hw_queues. We need
update the confused comments in blk_mq_queue_tag_busy_iter.
Signed-off-by: Nyangerkun <yangerkun@huawei.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>