platform/kernel/linux-rpi.git
5 years agoblock: Introduce blk_exit_queue()
Bart Van Assche [Thu, 9 Aug 2018 14:53:37 +0000 (07:53 -0700)]
block: Introduce blk_exit_queue()

This patch does not change any functionality.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Alexandru Moise <00moses.alexander00@gmail.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblkcg: Introduce blkg_root_lookup()
Bart Van Assche [Thu, 9 Aug 2018 14:53:36 +0000 (07:53 -0700)]
blkcg: Introduce blkg_root_lookup()

This new function will be used in a later patch to verify whether a
queue has been dissociated from the cgroup controller before being
released.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Omar Sandoval <osandov@fb.com>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Alexandru Moise <00moses.alexander00@gmail.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: Remove two superfluous #include directives
Bart Van Assche [Thu, 9 Aug 2018 14:47:28 +0000 (07:47 -0700)]
block: Remove two superfluous #include directives

Commit 12f5b9314545 ("blk-mq: Remove generation seqeunce") removed the
only seqcount_t and u64_stats_sync instances from <linux/blkdev.h> but
did not remove the corresponding #include directives. Since these
include directives are no longer needed, remove them.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Jianchao Wang <jianchao.w.wang@oracle.com>
Cc: Hannes Reinecke <hare@suse.com>,
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-mq: count the hctx as active before allocating tag
Jianchao Wang [Thu, 9 Aug 2018 14:34:17 +0000 (08:34 -0600)]
blk-mq: count the hctx as active before allocating tag

Currently, we count the hctx as active after allocate driver tag
successfully. If a previously inactive hctx try to get tag first
time, it may fails and need to wait. However, due to the stale tag
->active_queues, the other shared-tags users are still able to
occupy all driver tags while there is someone waiting for tag.
Consequently, even if the previously inactive hctx is waked up, it
still may not be able to get a tag and could be starved.

To fix it, we count the hctx as active before try to allocate driver
tag, then when it is waiting the tag, the other shared-tag users
will reserve budget for it.

Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jianchao Wang <jianchao.w.wang@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: bvec_nr_vecs() returns value for wrong slab
Greg Edwards [Wed, 8 Aug 2018 19:27:53 +0000 (13:27 -0600)]
block: bvec_nr_vecs() returns value for wrong slab

In commit ed996a52c868 ("block: simplify and cleanup bvec pool
handling"), the value of the slab index is incremented by one in
bvec_alloc() after the allocation is done to indicate an index value of
0 does not need to be later freed.

bvec_nr_vecs() was not updated accordingly, and thus returns the wrong
value.  Decrement idx before performing the lookup.

Fixes: ed996a52c868 ("block: simplify and cleanup bvec pool handling")
Signed-off-by: Greg Edwards <gedwards@ddn.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoMerge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block
Jens Axboe [Thu, 9 Aug 2018 14:22:21 +0000 (08:22 -0600)]
Merge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block

Pull NVMe updates from Christoph:

"This should be the last round of NVMe updates before the 4.19 merge
 window opens.  It conatins support for write protected (aka read-only)
 namespaces from Chaitanya, two ANA fixes from Hannes and a fabrics
 fix from Tal Shorer."

* 'nvme-4.19' of git://git.infradead.org/nvme:
  nvme-fabrics: fix ctrl_loss_tmo < 0 to reconnect forever
  nvmet: add ns write protect support
  nvme: set gendisk read only based on nsattr
  nvme.h: add support for ns write protect definitions
  nvme.h: fixup ANA group descriptor format
  nvme: fixup crash on failed discovery

5 years agobcache: trivial - remove tailing backslash in macro BTREE_FLAG
Shenghui Wang [Thu, 9 Aug 2018 07:48:51 +0000 (15:48 +0800)]
bcache: trivial - remove tailing backslash in macro BTREE_FLAG

Remove the tailing backslash in macro BTREE_FLAG in btree.h

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: make the pr_err statement used for ENOENT only in sysfs_attatch section
Shenghui Wang [Thu, 9 Aug 2018 07:48:50 +0000 (15:48 +0800)]
bcache: make the pr_err statement used for ENOENT only in sysfs_attatch section

The pr_err statement in the code for sysfs_attatch section would run
for various error codes, which maybe confusing.

E.g,

Run the command twice:
   echo 796b5c05-b03c-4bc7-9cbd-a8df5e8be891 > \
/sys/block/bcache0/bcache/attach
   [the backing dev got attached on the first run]
   echo 796b5c05-b03c-4bc7-9cbd-a8df5e8be891 > \
/sys/block/bcache0/bcache/attach

In dmesg, after the command run twice, we can get:
bcache: bch_cached_dev_attach() Can't attach sda6: already attached
bcache: __cached_dev_store() Can't attach 796b5c05-b03c-4bc7-9cbd-\
a8df5e8be891
               : cache set not found
The first statement in the message was right, but the second was
confusing.

bch_cached_dev_attach has various pr_ statements for various error
codes, except ENOENT.

After the change, rerun above command twice:
echo 796b5c05-b03c-4bc7-9cbd-a8df5e8be891 > \
/sys/block/bcache0/bcache/attach
echo 796b5c05-b03c-4bc7-9cbd-a8df5e8be891 > \
/sys/block/bcache0/bcache/attach

In dmesg we only got:
bcache: bch_cached_dev_attach() Can't attach sda6: already attached
No confusing "cache set not found" message anymore.

And for some not exist SET-UUID:
echo 796b5c05-b03c-4bc7-9cbd-a8df5e8be898 > \
/sys/block/bcache0/bcache/attach
In dmesg we can get:
bcache: __cached_dev_store() Can't attach 796b5c05-b03c-4bc7-9cbd-\
a8df5e8be898
               : cache set not found

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: set max writeback rate when I/O request is idle
Coly Li [Thu, 9 Aug 2018 07:48:49 +0000 (15:48 +0800)]
bcache: set max writeback rate when I/O request is idle

Commit b1092c9af9ed ("bcache: allow quick writeback when backing idle")
allows the writeback rate to be faster if there is no I/O request on a
bcache device. It works well if there is only one bcache device attached
to the cache set. If there are many bcache devices attached to a cache
set, it may introduce performance regression because multiple faster
writeback threads of the idle bcache devices will compete the btree level
locks with the bcache device who have I/O requests coming.

This patch fixes the above issue by only permitting fast writebac when
all bcache devices attached on the cache set are idle. And if one of the
bcache devices has new I/O request coming, minimized all writeback
throughput immediately and let PI controller __update_writeback_rate()
to decide the upcoming writeback rate for each bcache device.

Also when all bcache devices are idle, limited wrieback rate to a small
number is wast of thoughput, especially when backing devices are slower
non-rotation devices (e.g. SATA SSD). This patch sets a max writeback
rate for each backing device if the whole cache set is idle. A faster
writeback rate in idle time means new I/Os may have more available space
for dirty data, and people may observe a better write performance then.

Please note bcache may change its cache mode in run time, and this patch
still works if the cache mode is switched from writeback mode and there
is still dirty data on cache.

Fixes: Commit b1092c9af9ed ("bcache: allow quick writeback when backing idle")
Cc: stable@vger.kernel.org #4.16+
Signed-off-by: Coly Li <colyli@suse.de>
Tested-by: Kai Krakow <kai@kaishome.de>
Tested-by: Stefan Priebe <s.priebe@profihost.ag>
Cc: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: add code comments for bset.c
Coly Li [Thu, 9 Aug 2018 07:48:48 +0000 (15:48 +0800)]
bcache: add code comments for bset.c

This patch tries to add code comments in bset.c, to make some
tricky code and designment to be more comprehensible. Most information
of this patch comes from the discussion between Kent and I, he
offers very informative details. If there is any mistake
of the idea behind the code, no doubt that's from me misrepresentation.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: fix mistaken comments in request.c
Coly Li [Thu, 9 Aug 2018 07:48:47 +0000 (15:48 +0800)]
bcache: fix mistaken comments in request.c

This patch updates code comment in bch_keylist_realloc() by fixing
incorrected function names, to make the code to be more comprehennsible.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: fix mistaken code comments in bcache.h
Coly Li [Thu, 9 Aug 2018 07:48:46 +0000 (15:48 +0800)]
bcache: fix mistaken code comments in bcache.h

This patch updates the code comment in struct cache with correct array
names, to make the code to be more comprehensible.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: add a comment in super.c
Coly Li [Thu, 9 Aug 2018 07:48:45 +0000 (15:48 +0800)]
bcache: add a comment in super.c

This patch adds a line of code comment in super.c:register_bdev(), to
make code to be more comprehensible.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: avoid unncessary cache prefetch bch_btree_node_get()
Coly Li [Thu, 9 Aug 2018 07:48:44 +0000 (15:48 +0800)]
bcache: avoid unncessary cache prefetch bch_btree_node_get()

In bch_btree_node_get() the read-in btree node will be partially
prefetched into L1 cache for following bset iteration (if there is).
But if the btree node read is failed, the perfetch operations will
waste L1 cache space. This patch checkes whether read operation and
only does cache prefetch when read I/O succeeded.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: display rate debug parameters to 0 when writeback is not running
Coly Li [Thu, 9 Aug 2018 07:48:43 +0000 (15:48 +0800)]
bcache: display rate debug parameters to 0 when writeback is not running

When writeback is not running, writeback rate should be 0, other value is
misleading. And the following dyanmic writeback rate debug parameters
should be 0 too,
rate, proportional, integral, change
otherwise they are misleading when writeback is not running.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: do not check return value of debugfs_create_dir()
Coly Li [Thu, 9 Aug 2018 07:48:42 +0000 (15:48 +0800)]
bcache: do not check return value of debugfs_create_dir()

Greg KH suggests that normal code should not care about debugfs. Therefore
no matter successful or failed of debugfs_create_dir() execution, it is
unncessary to check its return value.

There are two functions called debugfs_create_dir() and check the return
value, which are bch_debug_init() and closure_debug_init(). This patch
changes these two functions from int to void type, and ignore return values
of debugfs_create_dir().

This patch does not fix exact bug, just makes things work as they should.

Signed-off-by: Coly Li <colyli@suse.de>
Suggested-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: stable@vger.kernel.org
Cc: Kai Krakow <kai@kaishome.de>
Cc: Kent Overstreet <kent.overstreet@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agodrivers/block/drbd: remove the null check for kmem_cache_destroy
zhong jiang [Wed, 8 Aug 2018 15:22:47 +0000 (23:22 +0800)]
drivers/block/drbd: remove the null check for kmem_cache_destroy

kmem_cache_destroy has taken null pointer into account. So it is
safe to drop the null check before calling the function.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agodrivers/block/aoe/aoedev: NULL check is not needed for mempool_destroy
zhong jiang [Wed, 8 Aug 2018 15:00:35 +0000 (23:00 +0800)]
drivers/block/aoe/aoedev: NULL check is not needed for mempool_destroy

mempool_destroy has taken the null pointer into account. So it is safe
to remove the null check.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agodrivers/block/mtip32xx: remove the null check for debugfs_remove_recursive
zhong jiang [Wed, 8 Aug 2018 14:58:33 +0000 (22:58 +0800)]
drivers/block/mtip32xx: remove the null check for debugfs_remove_recursive

debugfs_remove_recursive has taken null pointer into account. So it is
safe to drop the null check before calling the function.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agonvme-fabrics: fix ctrl_loss_tmo < 0 to reconnect forever
Tal Shorer [Tue, 7 Aug 2018 20:42:39 +0000 (23:42 +0300)]
nvme-fabrics: fix ctrl_loss_tmo < 0 to reconnect forever

When the user supplies a ctrl_loss_tmo < 0, we warn them that this will
cause the fabrics layer to attempt reconnection forever.  However, in
reality the fabrics layer never attempts to reconnect because the
condition to test whether we should reconnect is backwards in this case.

Signed-off-by: Tal Shorer <tal.shorer@gmail.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvmet: add ns write protect support
Chaitanya Kulkarni [Wed, 8 Aug 2018 06:01:07 +0000 (23:01 -0700)]
nvmet: add ns write protect support

This patch implements the Namespace Write Protect feature described in
"NVMe TP 4005a Namespace Write Protect". In this version, we implement
No Write Protect and Write Protect states for target ns which can be
toggled by set-features commands from the host side.

For write-protect state transition, we need to flush the ns specified
as a part of command so we also add helpers for carrying out synchronous
flush operations.

Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
[hch: fixed an incorrect endianess conversion, minor cleanups]
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme: set gendisk read only based on nsattr
Chaitanya Kulkarni [Wed, 8 Aug 2018 06:01:06 +0000 (23:01 -0700)]
nvme: set gendisk read only based on nsattr

NVMe 1.3 TP 4005 introduces new filed (NSATTR). This field indicates
whether given namespace is write protected or not. This patch sets the
gendisk associated with the namespace to read only based on the identify
namespace nsattr field.

Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme.h: add support for ns write protect definitions
Chaitanya Kulkarni [Wed, 8 Aug 2018 06:01:05 +0000 (23:01 -0700)]
nvme.h: add support for ns write protect definitions

Add various definitions from NVMe 1.3 TP 4005.

Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme.h: fixup ANA group descriptor format
Hannes Reinecke [Wed, 8 Aug 2018 06:35:29 +0000 (08:35 +0200)]
nvme.h: fixup ANA group descriptor format

ANA Phase 3 draft had the 'reserved' field in the group descriptor
format set to '23:17' (so that the first namespace identifier started
at byte 24), but that got move with the approved TP to '31:17'
(so that the first namespace identifier started at byte 32).

Signed-off-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agocfq: Suppress compiler warnings about comparisons
Bart Van Assche [Tue, 7 Aug 2018 23:17:29 +0000 (16:17 -0700)]
cfq: Suppress compiler warnings about comparisons

This patch does not change any functionality but avoids that gcc
reports the following warnings when building with W=1:

block/cfq-iosched.c: In function ?cfq_back_seek_max_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4756:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_back_seek_max_store, &cfqd->cfq_back_max, 0, UINT_MAX, 0);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_slice_idle_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4759:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_slice_idle_store, &cfqd->cfq_slice_idle, 0, UINT_MAX, 1);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_group_idle_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4760:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_group_idle_store, &cfqd->cfq_group_idle, 0, UINT_MAX, 1);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_low_latency_store?:
block/cfq-iosched.c:4741:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4765:1: note: in expansion of macro ?STORE_FUNCTION?
 STORE_FUNCTION(cfq_low_latency_store, &cfqd->cfq_latency, 0, 1, 0);
 ^~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_slice_idle_us_store?:
block/cfq-iosched.c:4775:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4782:1: note: in expansion of macro ?USEC_STORE_FUNCTION?
 USEC_STORE_FUNCTION(cfq_slice_idle_us_store, &cfqd->cfq_slice_idle, 0, UINT_MAX);
 ^~~~~~~~~~~~~~~~~~~
block/cfq-iosched.c: In function ?cfq_group_idle_us_store?:
block/cfq-iosched.c:4775:13: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
  if (__data < (MIN))      \
             ^
block/cfq-iosched.c:4783:1: note: in expansion of macro ?USEC_STORE_FUNCTION?
 USEC_STORE_FUNCTION(cfq_group_idle_us_store, &cfqd->cfq_group_idle, 0, UINT_MAX);
 ^~~~~~~~~~~~~~~~~~~

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agocfq: Annotate fall-through in a switch statement
Bart Van Assche [Tue, 7 Aug 2018 23:17:28 +0000 (16:17 -0700)]
cfq: Annotate fall-through in a switch statement

This patch avoids that gcc complains about fall-through when building
with W=1.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-wbt: Avoid lock contention and thundering herd issue in wbt_wait
Anchal Agarwal [Tue, 7 Aug 2018 20:40:49 +0000 (14:40 -0600)]
blk-wbt: Avoid lock contention and thundering herd issue in wbt_wait

I am currently running a large bare metal instance (i3.metal)
on EC2 with 72 cores, 512GB of RAM and NVME drives, with a
4.18 kernel. I have a workload that simulates a database
workload and I am running into lockup issues when writeback
throttling is enabled,with the hung task detector also
kicking in.

Crash dumps show that most CPUs (up to 50 of them) are
all trying to get the wbt wait queue lock while trying to add
themselves to it in __wbt_wait (see stack traces below).

[    0.948118] CPU: 45 PID: 0 Comm: swapper/45 Not tainted 4.14.51-62.38.amzn1.x86_64 #1
[    0.948119] Hardware name: Amazon EC2 i3.metal/Not Specified, BIOS 1.0 10/16/2017
[    0.948120] task: ffff883f7878c000 task.stack: ffffc9000c69c000
[    0.948124] RIP: 0010:native_queued_spin_lock_slowpath+0xf8/0x1a0
[    0.948125] RSP: 0018:ffff883f7fcc3dc8 EFLAGS: 00000046
[    0.948126] RAX: 0000000000000000 RBX: ffff887f7709ca68 RCX: ffff883f7fce2a00
[    0.948128] RDX: 000000000000001c RSI: 0000000000740001 RDI: ffff887f7709ca68
[    0.948129] RBP: 0000000000000002 R08: 0000000000b80000 R09: 0000000000000000
[    0.948130] R10: ffff883f7fcc3d78 R11: 000000000de27121 R12: 0000000000000002
[    0.948131] R13: 0000000000000003 R14: 0000000000000000 R15: 0000000000000000
[    0.948132] FS:  0000000000000000(0000) GS:ffff883f7fcc0000(0000) knlGS:0000000000000000
[    0.948134] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.948135] CR2: 000000c424c77000 CR3: 0000000002010005 CR4: 00000000003606e0
[    0.948136] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    0.948137] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    0.948138] Call Trace:
[    0.948139]  <IRQ>
[    0.948142]  do_raw_spin_lock+0xad/0xc0
[    0.948145]  _raw_spin_lock_irqsave+0x44/0x4b
[    0.948149]  ? __wake_up_common_lock+0x53/0x90
[    0.948150]  __wake_up_common_lock+0x53/0x90
[    0.948155]  wbt_done+0x7b/0xa0
[    0.948158]  blk_mq_free_request+0xb7/0x110
[    0.948161]  __blk_mq_complete_request+0xcb/0x140
[    0.948166]  nvme_process_cq+0xce/0x1a0 [nvme]
[    0.948169]  nvme_irq+0x23/0x50 [nvme]
[    0.948173]  __handle_irq_event_percpu+0x46/0x300
[    0.948176]  handle_irq_event_percpu+0x20/0x50
[    0.948179]  handle_irq_event+0x34/0x60
[    0.948181]  handle_edge_irq+0x77/0x190
[    0.948185]  handle_irq+0xaf/0x120
[    0.948188]  do_IRQ+0x53/0x110
[    0.948191]  common_interrupt+0x87/0x87
[    0.948192]  </IRQ>
....
[    0.311136] CPU: 4 PID: 9737 Comm: run_linux_amd64 Not tainted 4.14.51-62.38.amzn1.x86_64 #1
[    0.311137] Hardware name: Amazon EC2 i3.metal/Not Specified, BIOS 1.0 10/16/2017
[    0.311138] task: ffff883f6e6a8000 task.stack: ffffc9000f1ec000
[    0.311141] RIP: 0010:native_queued_spin_lock_slowpath+0xf5/0x1a0
[    0.311142] RSP: 0018:ffffc9000f1efa28 EFLAGS: 00000046
[    0.311144] RAX: 0000000000000000 RBX: ffff887f7709ca68 RCX: ffff883f7f722a00
[    0.311145] RDX: 0000000000000035 RSI: 0000000000d80001 RDI: ffff887f7709ca68
[    0.311146] RBP: 0000000000000202 R08: 0000000000140000 R09: 0000000000000000
[    0.311147] R10: ffffc9000f1ef9d8 R11: 000000001a249fa0 R12: ffff887f7709ca68
[    0.311148] R13: ffffc9000f1efad0 R14: 0000000000000000 R15: ffff887f7709ca00
[    0.311149] FS:  000000c423f30090(0000) GS:ffff883f7f700000(0000) knlGS:0000000000000000
[    0.311150] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.311151] CR2: 00007feefcea4000 CR3: 0000007f7016e001 CR4: 00000000003606e0
[    0.311152] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    0.311153] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    0.311154] Call Trace:
[    0.311157]  do_raw_spin_lock+0xad/0xc0
[    0.311160]  _raw_spin_lock_irqsave+0x44/0x4b
[    0.311162]  ? prepare_to_wait_exclusive+0x28/0xb0
[    0.311164]  prepare_to_wait_exclusive+0x28/0xb0
[    0.311167]  wbt_wait+0x127/0x330
[    0.311169]  ? finish_wait+0x80/0x80
[    0.311172]  ? generic_make_request+0xda/0x3b0
[    0.311174]  blk_mq_make_request+0xd6/0x7b0
[    0.311176]  ? blk_queue_enter+0x24/0x260
[    0.311178]  ? generic_make_request+0xda/0x3b0
[    0.311181]  generic_make_request+0x10c/0x3b0
[    0.311183]  ? submit_bio+0x5c/0x110
[    0.311185]  submit_bio+0x5c/0x110
[    0.311197]  ? __ext4_journal_stop+0x36/0xa0 [ext4]
[    0.311210]  ext4_io_submit+0x48/0x60 [ext4]
[    0.311222]  ext4_writepages+0x810/0x11f0 [ext4]
[    0.311229]  ? do_writepages+0x3c/0xd0
[    0.311239]  ? ext4_mark_inode_dirty+0x260/0x260 [ext4]
[    0.311240]  do_writepages+0x3c/0xd0
[    0.311243]  ? _raw_spin_unlock+0x24/0x30
[    0.311245]  ? wbc_attach_and_unlock_inode+0x165/0x280
[    0.311248]  ? __filemap_fdatawrite_range+0xa3/0xe0
[    0.311250]  __filemap_fdatawrite_range+0xa3/0xe0
[    0.311253]  file_write_and_wait_range+0x34/0x90
[    0.311264]  ext4_sync_file+0x151/0x500 [ext4]
[    0.311267]  do_fsync+0x38/0x60
[    0.311270]  SyS_fsync+0xc/0x10
[    0.311272]  do_syscall_64+0x6f/0x170
[    0.311274]  entry_SYSCALL_64_after_hwframe+0x42/0xb7

In the original patch, wbt_done is waking up all the exclusive
processes in the wait queue, which can cause a thundering herd
if there is a large number of writer threads in the queue. The
original intention of the code seems to be to wake up one thread
only however, it uses wake_up_all() in __wbt_done(), and then
uses the following check in __wbt_wait to have only one thread
actually get out of the wait loop:

if (waitqueue_active(&rqw->wait) &&
            rqw->wait.head.next != &wait->entry)
                return false;

The problem with this is that the wait entry in wbt_wait is
define with DEFINE_WAIT, which uses the autoremove wakeup function.
That means that the above check is invalid - the wait entry will
have been removed from the queue already by the time we hit the
check in the loop.

Secondly, auto-removing the wait entries also means that the wait
queue essentially gets reordered "randomly" (e.g. threads re-add
themselves in the order they got to run after being woken up).
Additionally, new requests entering wbt_wait might overtake requests
that were queued earlier, because the wait queue will be
(temporarily) empty after the wake_up_all, so the waitqueue_active
check will not stop them. This can cause certain threads to starve
under high load.

The fix is to leave the woken up requests in the queue and remove
them in finish_wait() once the current thread breaks out of the
wait loop in __wbt_wait. This will ensure new requests always
end up at the back of the queue, and they won't overtake requests
that are already in the wait queue. With that change, the loop
in wbt_wait is also in line with many other wait loops in the kernel.
Waking up just one thread drastically reduces lock contention, as
does moving the wait queue add/remove out of the loop.

A significant drop in lockdep's lock contention numbers is seen when
running the test application on the patched kernel.

Signed-off-by: Anchal Agarwal <anchalag@amazon.com>
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agotarget/loop: depend on SCSI
Christoph Hellwig [Tue, 7 Aug 2018 07:42:22 +0000 (09:42 +0200)]
target/loop: depend on SCSI

The target loopback driver is a low-level driver for the SCSI subsystem,
and as such needs to depend on it.

Fixes: 8a39a047 ("target: don't depend on SCSI")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agonvme: fixup crash on failed discovery
Hannes Reinecke [Tue, 7 Aug 2018 10:43:42 +0000 (12:43 +0200)]
nvme: fixup crash on failed discovery

When the initial discovery fails the subsystem hasn't been setup yet
in nvme_mpath_stop, and we can't dereference ctrl->subsys.

Fixes: 0d0b660f ("nvme: add ANA support")
Signed-off-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agoxen-blkfront: use true and false for boolean values
Gustavo A. R. Silva [Mon, 6 Aug 2018 14:14:55 +0000 (08:14 -0600)]
xen-blkfront: use true and false for boolean values

Return statements in functions returning bool should use true or false
instead of an integer value.

This code was detected with the help of Coccinelle.

Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agolightnvm: remove minor version check for 2.0
Matias Bjørling [Sun, 5 Aug 2018 14:25:19 +0000 (08:25 -0600)]
lightnvm: remove minor version check for 2.0

A minor version number increase should not break backwards
compatibility.

Fixes: 3cb98f84d368b ("lightnvm: add minor version to generic geometry")
Reviewed-by: Javier GonzĂ¡lez <javier@cnexlabs.com>
Signed-off-by: Matias Bjørling <mb@lightnvm.io>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoMerge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block2
Jens Axboe [Mon, 6 Aug 2018 01:34:09 +0000 (19:34 -0600)]
Merge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block2

Pull NVMe changes from Christoph:

"This contains the support for TP4004, Asymmetric Namespace Access,
 which makes NVMe multipathing usable in practice."

* 'nvme-4.19' of git://git.infradead.org/nvme:
  nvmet: use Retain Async Event bit to clear AEN
  nvmet: support configuring ANA groups
  nvmet: add minimal ANA support
  nvmet: track and limit the number of namespaces per subsystem
  nvmet: keep a port pointer in nvmet_ctrl
  nvme: add ANA support
  nvme: remove nvme_req_needs_failover
  nvme: simplify the API for getting log pages
  nvme.h: add ANA definitions
  nvme.h: add support for the log specific field

Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoMerge tag 'v4.18-rc6' into for-4.19/block2
Jens Axboe [Mon, 6 Aug 2018 01:32:09 +0000 (19:32 -0600)]
Merge tag 'v4.18-rc6' into for-4.19/block2

Pull in 4.18-rc6 to get the NVMe core AEN change to avoid a
merge conflict down the line.

Signed-of-by: Jens Axboe <axboe@kernel.dk>
5 years agoscsi: Check sense buffer size at build time
Kees Cook [Tue, 31 Jul 2018 19:51:54 +0000 (12:51 -0700)]
scsi: Check sense buffer size at build time

To avoid introducing problems like those fixed in commit f7068114d45e
("sr: pass down correctly sized SCSI sense buffer"), this creates a macro
wrapper for scsi_execute() that verifies the size of the sense buffer
similar to what was done for command string sizes in commit 3756f6401c30
("exec: avoid gcc-8 warning for get_task_comm").

Another solution could be to add a length argument to scsi_execute(),
but this function already takes a lot of arguments and Jens was not fond
of that approach.

Additionally, this moves the SCSI_SENSE_BUFFERSIZE definition into
scsi_device.h, and removes a redundant include for scsi_device.h from
scsi_cmnd.h.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agolibata-scsi: Move sense buffers onto stack
Kees Cook [Tue, 31 Jul 2018 19:51:53 +0000 (12:51 -0700)]
libata-scsi: Move sense buffers onto stack

To support future compile-time sizeof() checks that will be able to
validate the length of sense buffers, this removes the only dynamically
allocated sense buffers in the tree by putting the 96 byte sense buffers
on the stack.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agocdrom: Use struct scsi_sense_hdr internally
Kees Cook [Tue, 31 Jul 2018 19:51:52 +0000 (12:51 -0700)]
cdrom: Use struct scsi_sense_hdr internally

This removes more casts of struct request_sense and uses the standard
struct scsi_sense_hdr instead. This also fixes any possible stale values
since the prior code did not check the sense length.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoide-cd: Remove redundant sense buffer
Kees Cook [Tue, 31 Jul 2018 19:51:51 +0000 (12:51 -0700)]
ide-cd: Remove redundant sense buffer

This is already able to process the sense buffer, so remove the redundant
parsing during the failure path. This also fixes any possible stale values
since the prior code did not check the sense length.

Acked-by: David S. Miller <davem@davemloft.net>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: Switch struct packet_command to use struct scsi_sense_hdr
Kees Cook [Thu, 2 Aug 2018 21:22:13 +0000 (15:22 -0600)]
block: Switch struct packet_command to use struct scsi_sense_hdr

There is a lot of needless struct request_sense usage in the CDROM
code. These can all be struct scsi_sense_hdr instead, to avoid any
confusion over their respective structure sizes. This patch is a lot
of noise changing "sense" to "sshdr", but the final code is more
readable to distinguish between "sense" meaning "struct request_sense"
and "sshdr" meaning "struct scsi_sense_hdr".

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agotarget: don't depend on SCSI
Christoph Hellwig [Tue, 31 Jul 2018 19:51:49 +0000 (12:51 -0700)]
target: don't depend on SCSI

The core target code only needs code from scsi_common.c, which is now
separately selectable.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoscsi: build scsi_common.o for all scsi passthrough request users
Christoph Hellwig [Tue, 31 Jul 2018 19:51:48 +0000 (12:51 -0700)]
scsi: build scsi_common.o for all scsi passthrough request users

Split scsi_common.o out of SCSI so that non-SCSI users can pull it in
easily for future sense buffer helper usage.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoscsi: cxlflash: Drop unused sense buffers
Kees Cook [Tue, 31 Jul 2018 19:51:47 +0000 (12:51 -0700)]
scsi: cxlflash: Drop unused sense buffers

This removes the unused sense buffer in read_cap16() and write_same16().

Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoide-cd: Drop unused sense buffers
Kees Cook [Tue, 31 Jul 2018 19:51:46 +0000 (12:51 -0700)]
ide-cd: Drop unused sense buffers

This drops unused sense buffers from:

cdrom_eject()
cdrom_read_capacity()
cdrom_read_tocentry()
ide_cd_lockdoor()
ide_cd_read_toc()

Acked-by: David S. Miller <davem@davemloft.net>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-mq: fix updating tags depth
Ming Lei [Thu, 2 Aug 2018 10:23:26 +0000 (18:23 +0800)]
blk-mq: fix updating tags depth

The passed 'nr' from userspace represents the total depth, meantime
inside 'struct blk_mq_tags', 'nr_tags' stores the total tag depth,
and 'nr_reserved_tags' stores the reserved part.

There are two issues in blk_mq_tag_update_depth() now:

1) for growing tags, we should have used the passed 'nr', and keep the
number of reserved tags not changed.

2) the passed 'nr' should have been used for checking against
'tags->nr_tags', instead of number of the normal part.

This patch fixes the above two cases, and avoids kernel crash caused
by wrong resizing sbitmap queue.

Cc: "Ewan D. Milne" <emilne@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Omar Sandoval <osandov@fb.com>
Tested by: Marco Patalano <mpatalan@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: really disable runtime-pm for blk-mq
Ming Lei [Mon, 30 Jul 2018 12:02:19 +0000 (20:02 +0800)]
block: really disable runtime-pm for blk-mq

Runtime PM isn't ready for blk-mq yet, and commit 765e40b675a9 ("block:
disable runtime-pm for blk-mq") tried to disable it. Unfortunately,
it can't take effect in that way since user space still can switch
it on via 'echo auto > /sys/block/sdN/device/power/control'.

This patch disables runtime-pm for blk-mq really by pm_runtime_disable()
and fixes all kinds of PM related kernel crash.

Cc: Tomas Janousek <tomi@nomi.cz>
Cc: Przemek Socha <soprwa@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: <stable@vger.kernel.org>
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoaoe: mark expected switch fall-through
Gustavo A. R. Silva [Wed, 1 Aug 2018 22:30:20 +0000 (17:30 -0500)]
aoe: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 114722 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: make iolatency avg_lat exponentially decay
Dennis Zhou (Facebook) [Thu, 2 Aug 2018 06:15:41 +0000 (23:15 -0700)]
block: make iolatency avg_lat exponentially decay

Currently, avg_lat is calculated by accumulating the mean of every
window in a long running cumulative average. As time goes on, the metric
becomes less and less useful due to the accumulated history.

This patch reuses the same calculation done in load averages to make the
avg_lat metric more lively. Unlike load averages, the avg only advances
when a window elapses (due to an io). Idle periods extend the most
recent window. Bucketing is used to limit the history of avg_lat by
binding it to the window size. So, the window range for 1/exp (decay
rate) is [1 min, 2.5 min) when windows elapse immediately.

The current sample window size is exposed in the debug info to enable
calculation of the window range.

Signed-off-by: Dennis Zhou <dennisszhou@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-cgroup: clear the throttle queue on fork
Josef Bacik [Tue, 31 Jul 2018 16:39:04 +0000 (12:39 -0400)]
blk-cgroup: clear the throttle queue on fork

We were hitting a panic in production where we put too many times on the
request queue.  This is because we'd get the throttle_queue of the
parent if we fork()'ed while we needed to be throttled, but we didn't
have a reference on it.  Instead just clear these flags on fork so the
child doesn't pay for the sins of its father.

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-cgroup: hold the queue ref during throttling
Josef Bacik [Tue, 31 Jul 2018 16:39:03 +0000 (12:39 -0400)]
blk-cgroup: hold the queue ref during throttling

The blkg lifetime is protected by the queue lifetime, so we need to put
the queue _after_ we're done using the blkg.

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblk-iolatency: fix blkg leak in timer_fn
Josef Bacik [Tue, 31 Jul 2018 16:39:02 +0000 (12:39 -0400)]
blk-iolatency: fix blkg leak in timer_fn

At this point we have a ref on the blkg, we need to drop it if we don't
have a iolat.

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock/bsg-lib: use PTR_ERR_OR_ZERO to simplify the flow path
zhong jiang [Tue, 31 Jul 2018 16:13:14 +0000 (00:13 +0800)]
block/bsg-lib: use PTR_ERR_OR_ZERO to simplify the flow path

Simplify the code by using the PTR_ERR_OR_ZERO, instead of the
open code. It is better.

Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agot10-pi: provide empty t10_pi_complete() for !CONFIG_BLK_DEV_INTEGRITY
Jens Axboe [Tue, 31 Jul 2018 15:10:26 +0000 (09:10 -0600)]
t10-pi: provide empty t10_pi_complete() for !CONFIG_BLK_DEV_INTEGRITY

Fixes a link failure whtn BLK_DEV_INTEGRITY isn't defined.

Fixes: 10c41ddd6132 ("block: move dif_prepare/dif_complete functions to block layer")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: blk_init_allocated_queue() set q->fq as NULL in the fail case
xiao jin [Mon, 30 Jul 2018 06:11:12 +0000 (14:11 +0800)]
block: blk_init_allocated_queue() set q->fq as NULL in the fail case

We find the memory use-after-free issue in __blk_drain_queue()
on the kernel 4.14. After read the latest kernel 4.18-rc6 we
think it has the same problem.

Memory is allocated for q->fq in the blk_init_allocated_queue().
If the elevator init function called with error return, it will
run into the fail case to free the q->fq.

Then the __blk_drain_queue() uses the same memory after the free
of the q->fq, it will lead to the unpredictable event.

The patch is to set q->fq as NULL in the fail case of
blk_init_allocated_queue().

Fixes: commit 7c94e1c157a2 ("block: introduce blk_flush_queue to drive flush machinery")
Cc: <stable@vger.kernel.org>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: xiao jin <jin.xiao@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agonvme: use blk API to remap ref tags for IOs with metadata
Max Gurtovoy [Sun, 29 Jul 2018 21:15:33 +0000 (00:15 +0300)]
nvme: use blk API to remap ref tags for IOs with metadata

Also moved the logic of the remapping to the nvme core driver instead
of implementing it in the nvme pci driver. This way all the other nvme
transport drivers will benefit from it (in case they'll implement metadata
support).

Suggested-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: move dif_prepare/dif_complete functions to block layer
Max Gurtovoy [Sun, 29 Jul 2018 21:15:32 +0000 (00:15 +0300)]
block: move dif_prepare/dif_complete functions to block layer

Currently these functions are implemented in the scsi layer, but their
actual place should be the block layer since T10-PI is a general data
integrity feature that is used in the nvme protocol as well. Also, use
the tuple size from the integrity profile since it may vary between
integrity types.

Suggested-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: move ref_tag calculation func to the block layer
Max Gurtovoy [Sun, 29 Jul 2018 21:15:31 +0000 (00:15 +0300)]
block: move ref_tag calculation func to the block layer

Currently this function is implemented in the scsi layer, but it's
actual place should be the block layer since T10-PI is a general
data integrity feature that is used in the nvme protocol as well.

Suggested-by: Christoph Hellwig <hch@lst.de>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: don't account for split bio's size in cgroup stats
Josef Bacik [Mon, 30 Jul 2018 14:10:01 +0000 (10:10 -0400)]
block: don't account for split bio's size in cgroup stats

We need to check in blkcg_bio_issue_check if the bio is flagged as
QUEUE_ENTERED, because if it is then we've already accounted for the
size of the IO in the cgroup stats.  We can still however account for
the extra IO since it'll be another request.

Reported-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agopktcdvd: Fix possible Spectre-v1 for pkt_devs
Jinbum Park [Sat, 28 Jul 2018 04:20:44 +0000 (13:20 +0900)]
pktcdvd: Fix possible Spectre-v1 for pkt_devs

User controls @dev_minor which to be used as index of pkt_devs.
So, It can be exploited via Spectre-like attack. (speculative execution)

This kind of attack leaks address of pkt_devs, [1]
It leads an attacker to bypass security mechanism such as KASLR.

So sanitize @dev_minor before using it to prevent attack.

[1] https://github.com/jinb-park/linux-exploit/
tree/master/exploit-remaining-spectre-gadget/leak_pkt_devs.c

Signed-off-by: Jinbum Park <jinb.park7@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agonvmet: use Retain Async Event bit to clear AEN
Chaitanya Kulkarni [Thu, 26 Jul 2018 21:00:41 +0000 (14:00 -0700)]
nvmet: use Retain Async Event bit to clear AEN

In the current implementation, we clear the AEN bit when we get the
"get log page" command if given log page is associated with AEN.
This patch allows optionally retaining the AEN for the ctrl
under consideration when Retain Asynchronous Event (RAE) bit is set
as a part of "get log page" command.

This allows the host to read the Log page and optionally retaining the
AEN associated with this log page when using userspace tools like
nvme-cli.

Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
[hch: also use the new helper in the just merged ANA code]
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvmet: support configuring ANA groups
Christoph Hellwig [Fri, 1 Jun 2018 06:59:25 +0000 (08:59 +0200)]
nvmet: support configuring ANA groups

Allow creating non-default ANA groups (group ID > 1).  Groups are created
either by assigning the group ID to a namespace, or by creating a configfs
group object under a specific port.  All namespaces assigned to a group
that doesn't have a configfs object for a given port are marked as
inaccessible.

Allow changing the ANA state on a per-port basis by creating an
ana_groups directory under each port, and another directory with an
ana_state file in it.  The default ANA group 1 directory is created
automatically for each port.

For all changes in ANA configuration the ANA change AEN is sent.  We only
keep a global changecount instead of additional per-group changecounts to
keep the implementation as simple as possible.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvmet: add minimal ANA support
Christoph Hellwig [Thu, 19 Jul 2018 14:35:20 +0000 (07:35 -0700)]
nvmet: add minimal ANA support

Add support for Asynchronous Namespace Access as specified in NVMe 1.3
TP 4004.

Just add a default ANA group 1 that is optimized on all ports.  This is
(and will remain) the default assignment for any namespace not epxlicitly
assigned to another ANA group.  The ANA state can be manually changed
through the configfs interface, including the change state.

Includes fixes and improvements from Hannes Reinecke.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvmet: track and limit the number of namespaces per subsystem
Christoph Hellwig [Sun, 13 May 2018 17:00:13 +0000 (19:00 +0200)]
nvmet: track and limit the number of namespaces per subsystem

TP 4004 introduces a new 'Maximum Number of Allocated Namespaces' field
in the Identify controller data to help the host size resources.  Put
an upper limit on the supported namespaces to be able to support this
value as supporting 32-bits worth of namespaces would lead to very
large buffers.  The limit is completely arbitrary at this point.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvmet: keep a port pointer in nvmet_ctrl
Christoph Hellwig [Thu, 7 Jun 2018 13:09:50 +0000 (15:09 +0200)]
nvmet: keep a port pointer in nvmet_ctrl

This will be needed for the ANA AEN code.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvme: add ANA support
Christoph Hellwig [Mon, 14 May 2018 06:48:54 +0000 (08:48 +0200)]
nvme: add ANA support

Add support for Asynchronous Namespace Access as specified in NVMe 1.3
TP 4004.  With ANA each namespace attached to a controller belongs to an
ANA group that describes the characteristics of accessing the namespaces
through this controller.  In the optimized and non-optimized states
namespaces can be accessed regularly, although in a multi-pathing
environment we should always prefer to access a namespace through a
controller where an optimized relationship exists.  Namespaces in
Inaccessible, Permanent-Loss or Change state for a given controller
should not be accessed.

The states are updated through reading the ANA log page, which is read
once during controller initialization, whenever the ANA change notice
AEN is received, or when one of the ANA specific status codes that
signal a state change is received on a command.

The ANA state is kept in the nvme_ns structure, which makes the checks in
the fast path very simple.  Updating the ANA state when reading the log
page is also very simple, the only downside is that finding the initial
ANA state when scanning for namespaces is a bit cumbersome.

The gendisk for a ns_head is only registered once a live path for it
exists.  Without that the kernel would hang during partition scanning.

Includes fixes and improvements from Hannes Reinecke.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvme: remove nvme_req_needs_failover
Christoph Hellwig [Mon, 4 Jun 2018 06:43:00 +0000 (08:43 +0200)]
nvme: remove nvme_req_needs_failover

Now that we just call out to blk_path_error there isn't really any good
reason to not merge it into the only caller.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvme: simplify the API for getting log pages
Christoph Hellwig [Wed, 6 Jun 2018 12:39:00 +0000 (14:39 +0200)]
nvme: simplify the API for getting log pages

Merge nvme_get_log and nvme_get_log_ext into a single helper, which takes
a plain nsid instead of the nvme_ns pointer.  Also add support for the
log specific field while we're at it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvme.h: add ANA definitions
Christoph Hellwig [Sun, 13 May 2018 16:53:57 +0000 (18:53 +0200)]
nvme.h: add ANA definitions

Add various defintions from NVMe 1.3 TP 4004.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agonvme.h: add support for the log specific field
Christoph Hellwig [Sat, 12 May 2018 16:18:12 +0000 (18:18 +0200)]
nvme.h: add support for the log specific field

NVMe 1.3 added a new log specific field to the get log page CQ
defintion, add it to our get_log_page SQ structure.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
5 years agopartitions/aix: append null character to print data from disk
Mauricio Faria de Oliveira [Thu, 26 Jul 2018 01:46:29 +0000 (22:46 -0300)]
partitions/aix: append null character to print data from disk

Even if properly initialized, the lvname array (i.e., strings)
is read from disk, and might contain corrupt data (e.g., lack
the null terminating character for strings).

So, make sure the partition name string used in pr_warn() has
the null terminating character.

Fixes: 6ceea22bbbc8 ("partitions: add aix lvm partition support files")
Suggested-by: Daniel J. Axtens <daniel.axtens@canonical.com>
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agopartitions/aix: fix usage of uninitialized lv_info and lvname structures
Mauricio Faria de Oliveira [Thu, 26 Jul 2018 01:46:28 +0000 (22:46 -0300)]
partitions/aix: fix usage of uninitialized lv_info and lvname structures

The if-block that sets a successful return value in aix_partition()
uses 'lvip[].pps_per_lv' and 'n[].name' potentially uninitialized.

For example, if 'numlvs' is zero or alloc_lvn() fails, neither is
initialized, but are used anyway if alloc_pvd() succeeds after it.

So, make the alloc_pvd() call conditional on their initialization.

This has been hit when attaching an apparently corrupted/stressed
AIX LUN, misleading the kernel to pr_warn() invalid data and hang.

    [...] partition (null) (11 pp's found) is not contiguous
    [...] partition (null) (2 pp's found) is not contiguous
    [...] partition (null) (3 pp's found) is not contiguous
    [...] partition (null) (64 pp's found) is not contiguous

Fixes: 6ceea22bbbc8 ("partitions: add aix lvm partition support files")
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: stop using the deprecated get_seconds()
Arnd Bergmann [Thu, 26 Jul 2018 04:17:41 +0000 (12:17 +0800)]
bcache: stop using the deprecated get_seconds()

The get_seconds function is deprecated now since it returns a 32-bit
value that will eventually overflow, and we are replacing it throughout
the kernel with ktime_get_seconds() or ktime_get_real_seconds() that
return a time64_t.

bcache uses get_seconds() to read the current system time and store it in
the superblock as well as in uuid_entry structures that are user visible.

Unfortunately, the two structures in are still limited to 32 bits, so this
won't fix any real problems but will still overflow in year 2106. Let's
at least document that properly, in case we get an updated format in the
future it can be fixed. We still have a long time before the overflow
and checking the tools at https://github.com/koverstreet/bcache-tools
reveals no access to any of them.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: do not assign in if condition in bcache_device_init()
Florian Schmaus [Thu, 26 Jul 2018 04:17:40 +0000 (12:17 +0800)]
bcache: do not assign in if condition in bcache_device_init()

Fixes an error condition reported by checkpatch.pl which is caused by
assigning a variable in an if condition.

Signed-off-by: Florian Schmaus <flo@geekplace.eu>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: do not assign in if condition in bcache_init()
Florian Schmaus [Thu, 26 Jul 2018 04:17:39 +0000 (12:17 +0800)]
bcache: do not assign in if condition in bcache_init()

Fixes an error condition reported by checkpatch.pl which is caused by
assigning a variable in an if condition.

Signed-off-by: Florian Schmaus <flo@geekplace.eu>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: free heap cache_set->flush_btree in bch_journal_free
Shenghui Wang [Thu, 26 Jul 2018 04:17:38 +0000 (12:17 +0800)]
bcache: free heap cache_set->flush_btree in bch_journal_free

Free the cache_set->flush_bree heap memory on journal free.

Signed-off-by: Wang Sheng-Hui <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: do not assign in if condition register_bcache()
Florian Schmaus [Thu, 26 Jul 2018 04:17:37 +0000 (12:17 +0800)]
bcache: do not assign in if condition register_bcache()

Fixes an error condition reported by checkpatch.pl which is caused by
assigning a variable in an if condition.

Signed-off-by: Florian Schmaus <flo@geekplace.eu>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: fix I/O significant decline while backend devices registering
Tang Junhui [Thu, 26 Jul 2018 04:17:36 +0000 (12:17 +0800)]
bcache: fix I/O significant decline while backend devices registering

I attached several backend devices in the same cache set, and produced lots
of dirty data by running small rand I/O writes in a long time, then I
continue run I/O in the others cached devices, and stopped a cached device,
after a mean while, I register the stopped device again, I see the running
I/O in the others cached devices dropped significantly, sometimes even
jumps to zero.

In currently code, bcache would traverse each keys and btree node to count
the dirty data under read locker, and the writes threads can not get the
btree write locker, and when there is a lot of keys and btree node in the
registering device, it would last several seconds, so the write I/Os in
others cached device are blocked and declined significantly.

In this patch, when a device registering to a ache set, which exist others
cached devices with running I/Os, we get the amount of dirty data of the
device in an incremental way, and do not block other cached devices all the
time.

Patch v2: Rename some variables and macros name as Coly suggested.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: calculate the number of incremental GC nodes according to the total of btree...
Tang Junhui [Thu, 26 Jul 2018 04:17:35 +0000 (12:17 +0800)]
bcache: calculate the number of incremental GC nodes according to the total of btree nodes

This patch base on "[PATCH] bcache: finish incremental GC".

Since incremental GC would stop 100ms when front side I/O comes, so when
there are many btree nodes, if GC only processes constant (100) nodes each
time, GC would last a long time, and the front I/Os would run out of the
buckets (since no new bucket can be allocated during GC), and I/Os be
blocked again.

So GC should not process constant nodes, but varied nodes according to the
number of btree nodes. In this patch, GC is divided into constant (100)
times, so when there are many btree nodes, GC can process more nodes each
time, otherwise GC will process less nodes each time (but no less than
MIN_GC_NODES).

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: finish incremental GC
Tang Junhui [Thu, 26 Jul 2018 04:17:34 +0000 (12:17 +0800)]
bcache: finish incremental GC

In GC thread, we record the latest GC key in gc_done, which is expected
to be used for incremental GC, but in currently code, we didn't realize
it. When GC runs, front side IO would be blocked until the GC over, it
would be a long time if there is a lot of btree nodes.

This patch realizes incremental GC, the main ideal is that, when there
are front side I/Os, after GC some nodes (100), we stop GC, release locker
of the btree node, and go to process the front side I/Os for some times
(100 ms), then go back to GC again.

By this patch, when we doing GC, I/Os are not blocked all the time, and
there is no obvious I/Os zero jump problem any more.

Patch v2: Rename some variables and macros name as Coly suggested.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: simplify the calculation of the total amount of flash dirty data
Tang Junhui [Thu, 26 Jul 2018 04:17:33 +0000 (12:17 +0800)]
bcache: simplify the calculation of the total amount of flash dirty data

Currently we calculate the total amount of flash only devices dirty data
by adding the dirty data of each flash only device under registering
locker. It is very inefficient.

In this patch, we add a member flash_dev_dirty_sectors in struct cache_set
to record the total amount of flash only devices dirty data in real time,
so we didn't need to calculate the total amount of dirty data any more.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoreadahead: stricter check for bdi io_pages
Markus Stockhausen [Fri, 27 Jul 2018 15:09:53 +0000 (09:09 -0600)]
readahead: stricter check for bdi io_pages

ondemand_readahead() checks bdi->io_pages to cap the maximum pages
that need to be processed. This works until the readit section. If
we would do an async only readahead (async size = sync size) and
target is at beginning of window we expand the pages by another
get_next_ra_size() pages. Btrace for large reads shows that kernel
always issues a doubled size read at the beginning of processing.
Add an additional check for io_pages in the lower part of the func.
The fix helps devices that hard limit bio pages and rely on proper
handling of max_hw_read_sectors (e.g. older FusionIO cards). For
that reason it could qualify for stable.

Fixes: 9491ae4a ("mm: don't cap request size based on read-ahead setting")
Cc: stable@vger.kernel.org
Signed-off-by: Markus Stockhausen stockhausen@collogia.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoscsi: virtio_scsi: fix pi_bytes{out,in} on 4 KiB block size devices
Greg Edwards [Thu, 26 Jul 2018 19:52:54 +0000 (15:52 -0400)]
scsi: virtio_scsi: fix pi_bytes{out,in} on 4 KiB block size devices

When the underlying device is a 4 KiB logical block size device with a
protection interval exponent of 0, i.e. 4096 bytes data + 8 bytes PI, the
driver miscalculates the pi_bytes{out,in} by a factor of 8x (64 bytes).

This leads to errors on all reads and writes on 4 KiB logical block size
devices when CONFIG_BLK_DEV_INTEGRITY is enabled and the
VIRTIO_SCSI_F_T10_PI feature bit has been negotiated.

Fixes: e6dc783a38ec0 ("virtio-scsi: Enable DIF/DIX modes in SCSI host LLD")
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Edwards <gedwards@ddn.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: move bio_integrity_{intervals,bytes} into blkdev.h
Greg Edwards [Wed, 25 Jul 2018 14:22:58 +0000 (10:22 -0400)]
block: move bio_integrity_{intervals,bytes} into blkdev.h

This allows bio_integrity_bytes() to be called from drivers instead of
open coding it.

Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Edwards <gedwards@ddn.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoxen/blkfront: remove unused macros
Juergen Gross [Wed, 25 Jul 2018 07:42:07 +0000 (09:42 +0200)]
xen/blkfront: remove unused macros

Remove some macros not used anywhere.

Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoMerge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block
Jens Axboe [Wed, 25 Jul 2018 14:43:30 +0000 (08:43 -0600)]
Merge branch 'nvme-4.19' of git://git.infradead.org/nvme into for-4.19/block

Pull NVMe updates from Christoph:

"Highlights:

 - massively improved tracepoints (Keith Busch)
 - support for larger inline data in the RDMA host and target
   (Steve Wise)
 - RDMA setup/teardown path fixes and refactor (Sagi Grimberg)
 - Command Supported and Effects log support for the NVMe target
   (Chaitanya Kulkarni)
 - buffered I/O support for the NVMe target (Chaitanya Kulkarni)

 plus the usual set of cleanups and small enhancements."

* 'nvme-4.19' of git://git.infradead.org/nvme:
  nvmet: don't use uuid_le type
  nvmet: check fileio lba range access boundaries
  nvmet: fix file discard return status
  nvme-rdma: centralize admin/io queue teardown sequence
  nvme-rdma: centralize controller setup sequence
  nvme-rdma: unquiesce queues when deleting the controller
  nvme-rdma: mark expected switch fall-through
  nvme: add disk name to trace events
  nvme: add controller name to trace events
  nvme: use hw qid in trace events
  nvme: cache struct nvme_ctrl reference to struct nvme_request
  nvmet-rdma: add an error flow for post_recv failures
  nvmet-rdma: add unlikely check in the fast path
  nvmet-rdma: support max(16KB, PAGE_SIZE) inline data
  nvme-rdma: support up to 4 segments of inline data
  nvmet: add buffered I/O support for file backed ns
  nvmet: add commands supported and effects log page
  nvme: move init of keep_alive work item to controller initialization
  nvme.h: resync with nvme-cli

5 years agoblock: allow max_discard_segments to be stacked
Mike Snitzer [Fri, 20 Jul 2018 18:57:38 +0000 (14:57 -0400)]
block: allow max_discard_segments to be stacked

Set max_discard_segments to USHRT_MAX in blk_set_stacking_limits() so
that blk_stack_limits() can stack up this limit for stacked devices.

before:

$ cat /sys/block/nvme0n1/queue/max_discard_segments
256
$ cat /sys/block/dm-0/queue/max_discard_segments
1

after:

$ cat /sys/block/nvme0n1/queue/max_discard_segments
256
$ cat /sys/block/dm-0/queue/max_discard_segments
256

Fixes: 1e739730c5b9e ("block: optionally merge discontiguous discard bios into a single request")
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: unexport bio_clone_bioset
Christoph Hellwig [Tue, 24 Jul 2018 07:52:34 +0000 (09:52 +0200)]
block: unexport bio_clone_bioset

Now only used by the bounce code, so move it there and mark the function
static.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agomd: remove a bogus comment
Christoph Hellwig [Tue, 24 Jul 2018 07:52:33 +0000 (09:52 +0200)]
md: remove a bogus comment

The function name mentioned doesn't exist, and the code next to it
doesn't match the description either.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: remove bio_clone_kmalloc
Christoph Hellwig [Tue, 24 Jul 2018 07:52:32 +0000 (09:52 +0200)]
block: remove bio_clone_kmalloc

Unused now.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoexofs: use bio_clone_fast in _write_mirror
Christoph Hellwig [Tue, 24 Jul 2018 07:52:31 +0000 (09:52 +0200)]
exofs: use bio_clone_fast in _write_mirror

The mirroring code never changes the bio data or biovecs.  This means
we can reuse the biovec allocation easily instead of duplicating it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by Boaz Harrosh <ooo@electrozaur.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agobcache: don't clone bio in bch_data_verify
Christoph Hellwig [Tue, 24 Jul 2018 07:52:30 +0000 (09:52 +0200)]
bcache: don't clone bio in bch_data_verify

We immediately overwrite the biovec array, so instead just allocate
a new bio and copy over the disk, setor and size.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: bio_set_pages_dirty can't see NULL bv_page in a valid bio_vec
Christoph Hellwig [Tue, 24 Jul 2018 12:04:13 +0000 (14:04 +0200)]
block: bio_set_pages_dirty can't see NULL bv_page in a valid bio_vec

So don't bother handling it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: simplify bio_check_pages_dirty
Christoph Hellwig [Tue, 24 Jul 2018 12:04:12 +0000 (14:04 +0200)]
block: simplify bio_check_pages_dirty

bio_check_pages_dirty currently inviolates the invariant that bv_page of
a bio_vec inside bi_vcnt shouldn't be zero, and that is going to become
really annoying with multpath biovecs.  Fortunately there isn't any
all that good reason for it - once we decide to defer freeing the bio
to a workqueue holding onto a few additional pages isn't really an
issue anymore.  So just check if there is a clean page that needs
dirtying in the first path, and do a second pass to free them if there
was none, while the cache is still hot.

Also use the chance to micro-optimize bio_dirty_fn a bit by not saving
irq state - we know we are called from a workqueue.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agoblock: Rename the null_blk_mod kernel module back into null_blk
Bart Van Assche [Mon, 23 Jul 2018 21:18:33 +0000 (14:18 -0700)]
block: Rename the null_blk_mod kernel module back into null_blk

Commit ca4b2a011948 ("null_blk: add zone support") breaks several
blktests scripts because it renamed the null_blk kernel module into
null_blk_mod. Hence rename null_blk_mod back into null_blk.

Fixes: ca4b2a011948 ("null_blk: add zone support")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Matias Bjorling <matias.bjorling@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
5 years agonvmet: don't use uuid_le type
Andy Shevchenko [Tue, 17 Jul 2018 14:17:36 +0000 (17:17 +0300)]
nvmet: don't use uuid_le type

Don't use sizeof(uuid_le) where none of the parameters is type of uuid_le.
Since both arguments are u8 [16], use size of destination there.

Moreover, uuid_le is a deprecated type, and nvmet is using uuid_t
already.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvmet: check fileio lba range access boundaries
Sagi Grimberg [Wed, 11 Jul 2018 13:13:04 +0000 (16:13 +0300)]
nvmet: check fileio lba range access boundaries

Fail out-of-bounds with a proper status code.

Fixes: d5eff33ee6f8 ("nvmet: add simple file backed ns support")
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvmet: fix file discard return status
Sagi Grimberg [Wed, 11 Jul 2018 09:43:16 +0000 (12:43 +0300)]
nvmet: fix file discard return status

If nvmet_copy_from_sgl failed, we falsly return successful
completion status.

Fixes: d5eff33ee6f8 ("nvmet: add simple file backed ns support")
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme-rdma: centralize admin/io queue teardown sequence
Sagi Grimberg [Mon, 9 Jul 2018 09:49:07 +0000 (12:49 +0300)]
nvme-rdma: centralize admin/io queue teardown sequence

We follow the same queue teardown sequence in delete, reset and error
recovery. Centralize the logic.  This patch does not change any
functionality.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme-rdma: centralize controller setup sequence
Sagi Grimberg [Mon, 9 Jul 2018 09:49:06 +0000 (12:49 +0300)]
nvme-rdma: centralize controller setup sequence

Centralize controller sequence to a single routine that correctly cleans
up after failures instead of having multiple apperances in several flows
(create, reset, reconnect).

One thing that we also gain here are the sanity/boundary checks also
when connecting back to a dynamic controller.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme-rdma: unquiesce queues when deleting the controller
Sagi Grimberg [Mon, 9 Jul 2018 09:49:05 +0000 (12:49 +0300)]
nvme-rdma: unquiesce queues when deleting the controller

If the controller is going away, we need to unquiesce the IO queues so
that all pending request can fail gracefully before moving forward with
controller deletion. Do that before we destroy the IO queues so
blk_cleanup_queue won't block in freeze.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme-rdma: mark expected switch fall-through
Gustavo A. R. Silva [Thu, 5 Jul 2018 13:14:00 +0000 (08:14 -0500)]
nvme-rdma: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agonvme: add disk name to trace events
Keith Busch [Fri, 29 Jun 2018 22:50:03 +0000 (16:50 -0600)]
nvme: add disk name to trace events

This will print the disk name to the nvme event trace for io requests so
a user can better distinguish traffic to different disks. This can be used
to  create disk based filters. For example, to see only nvme0n2 traffic:

  echo "disk == \"nvme0n2\"" > /sys/kernel/debug/tracing/events/nvme/filter

Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
[hch: turned __assign_disk_name into an inline function]
Signed-off-by: Christoph Hellwig <hch@lst.de>