| Age | Commit message (Collapse) | Author |
|---|
|
After commit 396eaf21ee17 ("blk-mq: improve DM's blk-mq IO merging via
blk_insert_cloned_request feedback"), map_request() will requeue the tio
when issued clone request return BLK_STS_RESOURCE or BLK_STS_DEV_RESOURCE.
Thus, if device driver status is error, a tio may be requeued multiple
times until the return value is not DM_MAPIO_REQUEUE. That means
type->start_io may be called multiple times, while type->end_io is only
called when IO complete.
In fact, even without commit 396eaf21ee17, setup_clone() failure can
also cause tio requeue and associated missed call to type->end_io.
The service-time path selector selects path based on in_flight_size,
which is increased by st_start_io() and decreased by st_end_io().
Missed calls to st_end_io() can lead to in_flight_size count error and
will cause the selector to make the wrong choice. In addition,
queue-length path selector will also be affected.
To fix the problem, call type->end_io in ->release_clone_rq before tio
requeue. map_info is passed to ->release_clone_rq() for map_request()
error path that result in requeue.
Fixes: 396eaf21ee17 ("blk-mq: improve DM's blk-mq IO merging via blk_insert_cloned_request feedback")
Cc: stable@vger.kernl.org
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
The flags field in 'struct shash_desc' never actually does anything.
The only ostensibly supported flag is CRYPTO_TFM_REQ_MAY_SLEEP.
However, no shash algorithm ever sleeps, making this flag a no-op.
With this being the case, inevitably some users who can't sleep wrongly
pass MAY_SLEEP. These would all need to be fixed if any shash algorithm
actually started sleeping. For example, the shash_ahash_*() functions,
which wrap a shash algorithm with the ahash API, pass through MAY_SLEEP
from the ahash API to the shash API. However, the shash functions are
called under kmap_atomic(), so actually they're assumed to never sleep.
Even if it turns out that some users do need preemption points while
hashing large buffers, we could easily provide a helper function
crypto_shash_update_large() which divides the data into smaller chunks
and calls crypto_shash_update() and cond_resched() for each chunk. It's
not necessary to have a flag in 'struct shash_desc', nor is it necessary
to make individual shash algorithms aware of this at all.
Therefore, remove shash_desc::flags, and document that the
crypto_shash_*() functions can be called from any context.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
|
|
CACHE_SYNC branch of run_cache_set
In the CACHE_SYNC branch of run_cache_set(), LIST_HEAD(journal) is used
to collect journal_replay(s) and filled by bch_journal_read().
If all goes well, bch_journal_replay() will release the list of
jounal_replay(s) at the end of the branch.
If something goes wrong, code flow will jump to the label "err:" and leave
the list unreleased.
This patch will release the list of journal_replay(s) in the case of
error detected.
v1 -> v2:
* Move the release code to the location after label 'err:' to
simply the change.
Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
of btree_gc_coalesce
Elements of keylist should be accessed before the list is freed.
Move bch_keylist_free() calling after the while loop to avoid wrong
content accessed.
Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
journal replay failed with messages:
Sep 10 19:10:43 ceph kernel: bcache: error on
bb379a64-e44e-4812-b91d-a5599871a3b1: bcache: journal entries
2057493-2057567 missing! (replaying 2057493-2076601), disabling
caching
The reason is in journal_reclaim(), when discard is enabled, we send
discard command and reclaim those journal buckets whose seq is old
than the last_seq_now, but before we write a journal with last_seq_now,
the machine is restarted, so the journal with the last_seq_now is not
written to the journal bucket, and the last_seq_wrote in the newest
journal is old than last_seq_now which we expect to be, so when we doing
replay, journals from last_seq_wrote to last_seq_now are missing.
It's hard to write a journal immediately after journal_reclaim(),
and it harmless if those missed journal are caused by discarding
since those journals are already wrote to btree node. So, if miss
seqs are started from the beginning journal, we treat it as normal,
and only print a message to show the miss journal, and point out
it maybe caused by discarding.
Patch v2 add a judgement condition to ignore the missed journal
only when discard enabled as Coly suggested.
(Coly Li: rebase the patch with other changes in bch_journal_replay())
Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com>
Tested-by: Dennis Schridde <devurandom@gmx.net>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
This patch tries to release mutex bch_register_lock early, to give
chance to stop cache set and bcache device early.
This patch also expends time out of stopping all bcache device from
2 seconds to 10 seconds, because stopping writeback rate update worker
may delay for 5 seconds, 2 seconds is not enough.
After this patch applied, stopping bcache devices during system reboot
or shutdown is very hard to be observed any more.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Add code comments to explain which call back function might be called
for the closure_queue(). This is an effort to make code to be more
understandable for readers.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Add comments to explain why in register_bcache() blkdev_put() won't
be called in two location. Add comments to explain why blkdev_put()
must be called in register_cache() when cache_alloc() failed.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
This patch adds return value to register_bdev(). Then if failure happens
inside register_bdev(), its caller register_bcache() may detect and
handle the failure more properly.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
When failure happens inside bch_journal_replay(), calling
cache_set_err_on() and handling the failure in async way is not a good
idea. Because after bch_journal_replay() returns, registering code will
continue to execute following steps, and unregistering code triggered
by cache_set_err_on() is running in same time. First it is unnecessary
to handle failure and unregister cache set in an async way, second there
might be potential race condition to run register and unregister code
for same cache set.
So in this patch, if failure happens in bch_journal_replay(), we don't
call cache_set_err_on(), and just print out the same error message to
kernel message buffer, then return -EIO immediately caller. Then caller
can detect such failure and handle it in synchrnozied way.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Bcache has several routines to release resources in implicit way, they
are called when the associated kobj released. This patch adds code
comments to notice when and which release callback will be called,
- When dc->disk.kobj released:
void bch_cached_dev_release(struct kobject *kobj)
- When d->kobj released:
void bch_flash_dev_release(struct kobject *kobj)
- When c->kobj released:
void bch_cache_set_release(struct kobject *kobj)
- When ca->kobj released
void bch_cache_release(struct kobject *kobj)
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Currently run_cache_set() has no return value, if there is failure in
bch_journal_replay(), the caller of run_cache_set() has no idea about
such failure and just continue to execute following code after
run_cache_set(). The internal failure is triggered inside
bch_journal_replay() and being handled in async way. This behavior is
inefficient, while failure handling inside bch_journal_replay(), cache
register code is still running to start the cache set. Registering and
unregistering code running as same time may introduce some rare race
condition, and make the code to be more hard to be understood.
This patch adds return value to run_cache_set(), and returns -EIO if
bch_journal_rreplay() fails. Then caller of run_cache_set() may detect
such failure and stop registering code flow immedidately inside
register_cache_set().
If journal replay fails, run_cache_set() can report error immediately
to register_cache_set(). This patch makes the failure handling for
bch_journal_replay() be in synchronized way, easier to understand and
debug, and avoid poetential race condition for register-and-unregister
in same time.
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
In journal_reclaim() ja->cur_idx of each cache will be update to
reclaim available journal buckets. Variable 'int n' is used to count how
many cache is successfully reclaimed, then n is set to c->journal.key
by SET_KEY_PTRS(). Later in journal_write_unlocked(), a for_each_cache()
loop will write the jset data onto each cache.
The problem is, if all jouranl buckets on each cache is full, the
following code in journal_reclaim(),
529 for_each_cache(ca, c, iter) {
530 struct journal_device *ja = &ca->journal;
531 unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets;
532
533 /* No space available on this device */
534 if (next == ja->discard_idx)
535 continue;
536
537 ja->cur_idx = next;
538 k->ptr[n++] = MAKE_PTR(0,
539 bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
540 ca->sb.nr_this_dev);
541 }
542
543 bkey_init(k);
544 SET_KEY_PTRS(k, n);
If there is no available bucket to reclaim, the if() condition at line
534 will always true, and n remains 0. Then at line 544, SET_KEY_PTRS()
will set KEY_PTRS field of c->journal.key to 0.
Setting KEY_PTRS field of c->journal.key to 0 is wrong. Because in
journal_write_unlocked() the journal data is written in following loop,
649 for (i = 0; i < KEY_PTRS(k); i++) {
650-671 submit journal data to cache device
672 }
If KEY_PTRS field is set to 0 in jouranl_reclaim(), the journal data
won't be written to cache device here. If system crahed or rebooted
before bkeys of the lost journal entries written into btree nodes, data
corruption will be reported during bcache reload after rebooting the
system.
Indeed there is only one cache in a cache set, there is no need to set
KEY_PTRS field in journal_reclaim() at all. But in order to keep the
for_each_cache() logic consistent for now, this patch fixes the above
problem by not setting 0 KEY_PTRS of journal key, if there is no bucket
available to reclaim.
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Cc: stable@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
'int ret' is defined as a local variable inside macro read_bucket().
Since this macro is called multiple times, and following patches will
use a 'int ret' variable in bch_journal_read(), this patch moves
definition of 'int ret' from macro read_bucket() to range of function
bch_journal_read().
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
There is a race between cache device register and cache set unregister.
For an already registered cache device, register_bcache will call
bch_is_open to iterate through all cachesets and check every cache
there. The race occurs if cache_set_free executes at the same time and
clears the caches right before ca is dereferenced in bch_is_open_cache.
To close the race, let's make sure the clean up work is protected by
the bch_register_lock as well.
This issue can be reproduced as follows,
while true; do echo /dev/XXX> /sys/fs/bcache/register ; done&
while true; do echo 1> /sys/block/XXX/bcache/set/unregister ; done &
and results in the following oops,
[ +0.000053] BUG: unable to handle kernel NULL pointer dereference at 0000000000000998
[ +0.000457] #PF error: [normal kernel read fault]
[ +0.000464] PGD 800000003ca9d067 P4D 800000003ca9d067 PUD 3ca9c067 PMD 0
[ +0.000388] Oops: 0000 [#1] SMP PTI
[ +0.000269] CPU: 1 PID: 3266 Comm: bash Not tainted 5.0.0+ #6
[ +0.000346] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-2.fc28 04/01/2014
[ +0.000472] RIP: 0010:register_bcache+0x1829/0x1990 [bcache]
[ +0.000344] Code: b0 48 83 e8 50 48 81 fa e0 e1 10 c0 0f 84 a9 00 00 00 48 89 c6 48 89 ca 0f b7 ba 54 04 00 00 4c 8b 82 60 0c 00 00 85 ff 74 2f <49> 3b a8 98 09 00 00 74 4e 44 8d 47 ff 31 ff 49 c1 e0 03 eb 0d
[ +0.000839] RSP: 0018:ffff92ee804cbd88 EFLAGS: 00010202
[ +0.000328] RAX: ffffffffc010e190 RBX: ffff918b5c6b5000 RCX: ffff918b7d8e0000
[ +0.000399] RDX: ffff918b7d8e0000 RSI: ffffffffc010e190 RDI: 0000000000000001
[ +0.000398] RBP: ffff918b7d318340 R08: 0000000000000000 R09: ffffffffb9bd2d7a
[ +0.000385] R10: ffff918b7eb253c0 R11: ffffb95980f51200 R12: ffffffffc010e1a0
[ +0.000411] R13: fffffffffffffff2 R14: 000000000000000b R15: ffff918b7e232620
[ +0.000384] FS: 00007f955bec2740(0000) GS:ffff918b7eb00000(0000) knlGS:0000000000000000
[ +0.000420] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ +0.000801] CR2: 0000000000000998 CR3: 000000003cad6000 CR4: 00000000001406e0
[ +0.000837] Call Trace:
[ +0.000682] ? _cond_resched+0x10/0x20
[ +0.000691] ? __kmalloc+0x131/0x1b0
[ +0.000710] kernfs_fop_write+0xfa/0x170
[ +0.000733] __vfs_write+0x2e/0x190
[ +0.000688] ? inode_security+0x10/0x30
[ +0.000698] ? selinux_file_permission+0xd2/0x120
[ +0.000752] ? security_file_permission+0x2b/0x100
[ +0.000753] vfs_write+0xa8/0x1a0
[ +0.000676] ksys_write+0x4d/0xb0
[ +0.000699] do_syscall_64+0x3a/0xf0
[ +0.000692] entry_SYSCALL_64_after_hwframe+0x44/0xa9
Signed-off-by: Liang Chen <liangchen.linux@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
There are a few nits in this function. They could in theory all
be separate patches, but that's probably taking small commits
too far.
1) I added a brief comment saying what it does.
2) I like to declare pointer parameters "const" where possible
for documentation reasons.
3) It uses bitmap_weight(&rand, BITS_PER_LONG) to compute the Hamming
weight of a 32-bit random number (giving a random integer with
mean 16 and variance 8). Passing by reference in a 64-bit variable
is silly; just use hweight32().
4) Its helper function fract_exp_two is unnecessarily tangled.
Gcc can optimize the multiply by (1 << x) to a shift, but it can
be written in a much more straightforward way at the cost of one
more bit of internal precision. Some analysis reveals that this
bit is always available.
This shrinks the object code for fract_exp_two(x, 6) from 23 bytes:
0000000000000000 <foo1>:
0: 89 f9 mov %edi,%ecx
2: c1 e9 06 shr $0x6,%ecx
5: b8 01 00 00 00 mov $0x1,%eax
a: d3 e0 shl %cl,%eax
c: 83 e7 3f and $0x3f,%edi
f: d3 e7 shl %cl,%edi
11: c1 ef 06 shr $0x6,%edi
14: 01 f8 add %edi,%eax
16: c3 retq
To 19:
0000000000000017 <foo2>:
17: 89 f8 mov %edi,%eax
19: 83 e0 3f and $0x3f,%eax
1c: 83 c0 40 add $0x40,%eax
1f: 89 f9 mov %edi,%ecx
21: c1 e9 06 shr $0x6,%ecx
24: d3 e0 shl %cl,%eax
26: c1 e8 06 shr $0x6,%eax
29: c3 retq
(Verified with 0 <= frac_bits <= 8, 0 <= x < 16<<frac_bits;
both versions produce the same output.)
5) And finally, the call to bch_get_congested() in check_should_bypass()
is separated from the use of the value by multiple tests which
could moot the need to compute it. Move the computation down to
where it's needed. This also saves a local register to hold the
computed value.
Signed-off-by: George Spelvin <lkml@sdf.org>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
This patch uses kmemdup_nul to create a NUL-terminated string from
dc->sb.label. This is better than open coding it.
With this, we can move env[2] initialization into env[] array to make
code more elegant.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
clang has identified a code path in which it thinks a
variable may be unused:
drivers/md/bcache/alloc.c:333:4: error: variable 'bucket' is used uninitialized whenever 'if' condition is false
[-Werror,-Wsometimes-uninitialized]
fifo_pop(&ca->free_inc, bucket);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/md/bcache/util.h:219:27: note: expanded from macro 'fifo_pop'
#define fifo_pop(fifo, i) fifo_pop_front(fifo, (i))
^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/md/bcache/util.h:189:6: note: expanded from macro 'fifo_pop_front'
if (_r) { \
^~
drivers/md/bcache/alloc.c:343:46: note: uninitialized use occurs here
allocator_wait(ca, bch_allocator_push(ca, bucket));
^~~~~~
drivers/md/bcache/alloc.c:287:7: note: expanded from macro 'allocator_wait'
if (cond) \
^~~~
drivers/md/bcache/alloc.c:333:4: note: remove the 'if' if its condition is always true
fifo_pop(&ca->free_inc, bucket);
^
drivers/md/bcache/util.h:219:27: note: expanded from macro 'fifo_pop'
#define fifo_pop(fifo, i) fifo_pop_front(fifo, (i))
^
drivers/md/bcache/util.h:189:2: note: expanded from macro 'fifo_pop_front'
if (_r) { \
^
drivers/md/bcache/alloc.c:331:15: note: initialize the variable 'bucket' to silence this warning
long bucket;
^
This cannot happen in practice because we only enter the loop
if there is at least one element in the list.
Slightly rearranging the code makes this clearer to both the
reader and the compiler, which avoids the warning.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
To get the amount of unused buckets in sysfs_priority_stats, the code
count the buckets which GC_SECTORS_USED is zero. It's correct and should
not be overwritten by the count of buckets which prio is zero.
Signed-off-by: Guoju Fang <fangguoju@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
The bio from upper layer is considered completed when bio_complete()
returns. In most scenarios bio_complete() is called in search_free(),
but when read miss happens, the bio_compete() is called when backing
device reading completed, while the struct search is still in use until
cache inserting finished.
If someone stops the bcache device just then, the device may be closed
and released, but after cache inserting finished the struct search will
access a freed struct cached_dev.
This patch add the reference of bcache device before bio_complete() when
read miss happens, and put it after the search is not used.
Signed-off-by: Guoju Fang <fangguoju@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Otherwise, just activating a thin-pool and thin device and then
deactivating them will cause the thin-pool metadata to be changed
(e.g. superblock written) -- even without any metadata being changed.
Add 'in_service' flag to struct dm_pool_metadata and set it in
pmd_write_lock() because all on-disk metadata changes must take a write
lock of pmd->root_lock. Once 'in_service' is set it is never cleared.
__commit_transaction() will return 0 if 'in_service' is not set.
dm_pool_commit_metadata() is updated to use __pmd_write_lock() so that
it isn't the sole reason for putting a thin-pool in service.
Also fix dm_pool_commit_metadata() to open the next transaction if the
return from __commit_transaction() is 0. Not seeing why the early
return ever made since for a return of 0 given that dm-io's async_io(),
as used by bufio, always returns 0.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
No functional change, but this prepares to hook off of pmd_write_lock()
with additional functionality (as provided in next commit).
Suggested-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Fix __reserve_metadata_snap() to return early if __commit_transaction()
fails.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Otherwise, memory that is allocated (and potentially not previously
zeroed) will get written to disk as part of the space maps.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
In functions writecache_discard() and writecache_find_entry() there is a
high probablity that the pointer of structure rb_node won't equal NULL.
Add unlikely for the pointer node NULL.
Signed-off-by: Huaisheng Ye <yehs1@lenovo.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
bio is already available so there is no need to access it in terms of
the wb pointer.
Signed-off-by: Huaisheng Ye <yehs1@lenovo.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Substitute the global locking scheme with a fine grained one, employing
the read-write semaphore and the scalable exception tables with
per-bucket locks introduced by the previous two commits.
Summarizing, we now use a read-write semaphore to protect the mostly
read fields of the snapshot structure, e.g., valid, active, etc., and
per-bucket bit spinlocks to protect accesses to the complete and pending
exception tables.
Finally, we use an extra spinlock (pe_allocation_lock) to serialize the
allocation of new exceptions by the exception store. This allocation is
really fast, so the extra spinlock doesn't hurt the performance.
This scheme allows dm-snapshot to scale better, resulting in increased
IOPS and reduced latency.
Following are some benchmark results using the null_blk device:
modprobe null_blk gb=1024 bs=512 submit_queues=8 hw_queue_depth=4096 \
queue_mode=2 irqmode=1 completion_nsec=1 nr_devices=1
* Benchmark fio_origin_randwrite_throughput_N, from the device mapper
test suite [1] (direct IO, random 4K writes to origin device, IO
engine libaio):
+--------------+-------------+------------+
| # of workers | IOPS Before | IOPS After |
+--------------+-------------+------------+
| 1 | 57708 | 66421 |
| 2 | 63415 | 77589 |
| 4 | 67276 | 98839 |
| 8 | 60564 | 109258 |
+--------------+-------------+------------+
* Benchmark fio_origin_randwrite_latency_N, from the device mapper test
suite [1] (direct IO, random 4K writes to origin device, IO engine
psync):
+--------------+-----------------------+----------------------+
| # of workers | Latency (usec) Before | Latency (usec) After |
+--------------+-----------------------+----------------------+
| 1 | 16.25 | 13.27 |
| 2 | 31.65 | 25.08 |
| 4 | 55.28 | 41.08 |
| 8 | 121.47 | 74.44 |
+--------------+-----------------------+----------------------+
* Benchmark fio_snapshot_randwrite_throughput_N, from the device mapper
test suite [1] (direct IO, random 4K writes to snapshot device, IO
engine libaio):
+--------------+-------------+------------+
| # of workers | IOPS Before | IOPS After |
+--------------+-------------+------------+
| 1 | 72593 | 84938 |
| 2 | 97379 | 134973 |
| 4 | 90610 | 143077 |
| 8 | 90537 | 180085 |
+--------------+-------------+------------+
* Benchmark fio_snapshot_randwrite_latency_N, from the device mapper
test suite [1] (direct IO, random 4K writes to snapshot device, IO
engine psync):
+--------------+-----------------------+----------------------+
| # of workers | Latency (usec) Before | Latency (usec) After |
+--------------+-----------------------+----------------------+
| 1 | 12.53 | 10.6 |
| 2 | 19.78 | 14.89 |
| 4 | 40.37 | 23.47 |
| 8 | 89.32 | 48.48 |
+--------------+-----------------------+----------------------+
[1] https://github.com/jthornber/device-mapper-test-suite
Co-developed-by: Ilias Tsitsimpis <iliastsi@arrikto.com>
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Acked-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Use list_bl to implement the exception hash tables' buckets. This change
permits concurrent access, to distinct buckets, by multiple threads.
Also, implement helper functions to lock and unlock the exception tables
based on the chunk number of the exception at hand.
We retain the global locking, by means of down_write(), which is
replaced by the next commit.
Still, we must acquire the per-bucket spinlocks when accessing the hash
tables, since list_bl does not allow modification on unlocked lists.
Co-developed-by: Ilias Tsitsimpis <iliastsi@arrikto.com>
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Acked-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
dm-snapshot uses a single mutex to serialize every access to the
snapshot state. This includes all accesses to the complete and pending
exception tables, which occur at every origin write, every snapshot
read/write and every exception completion.
The lock statistics indicate that this mutex is a bottleneck (average
wait time ~480 usecs for 8 processes doing random 4K writes to the
origin device) preventing dm-snapshot to scale as the number of threads
doing IO increases.
The major contention points are __origin_write()/snapshot_map() and
pending_complete(), i.e., the submission and completion of pending
exceptions.
Replace this mutex with a rw semaphore.
We essentially revert commit ae1093be5a0ef9 ("dm snapshot: use mutex
instead of rw_semaphore") and together with the next two patches we
substitute the single mutex with a fine-grained locking scheme, where we
use a read-write semaphore to protect the mostly read fields of the
snapshot structure, e.g., valid, active, etc., and per-bucket bit
spinlocks to protect accesses to the complete and pending exception
tables.
Co-developed-by: Ilias Tsitsimpis <iliastsi@arrikto.com>
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Acked-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
When completing a pending exception, pending_complete() waits for all
conflicting reads to drain, before inserting the final, completed
exception. Conflicting reads are snapshot reads redirected to the
origin, because the relevant chunk is not remapped to the COW device the
moment we receive the read.
The completed exception must be inserted into the exception table after
all conflicting reads drain to ensure snapshot reads don't return
corrupted data. This is required because inserting the completed
exception into the exception table signals that the relevant chunk is
remapped and both origin writes and snapshot merging will now overwrite
the chunk in origin.
This wait is done holding the snapshot lock to ensure that
pending_complete() doesn't starve if new snapshot reads keep coming for
this chunk.
In preparation for the next commit, where we use a spinlock instead of a
mutex to protect the exception tables, we remove the need for holding
the lock while waiting for conflicting reads to drain.
We achieve this in two steps:
1. pending_complete() inserts the completed exception before waiting for
conflicting reads to drain and removes the pending exception after
all conflicting reads drain.
This ensures that new snapshot reads will be redirected to the COW
device, instead of the origin, and thus pending_complete() will not
starve. Moreover, we use the existence of both a completed and
a pending exception to signify that the COW is done but there are
conflicting reads in flight.
2. In __origin_write() we check first if there is a pending exception
and then if there is a completed exception. If there is a pending
exception any submitted BIO is delayed on the pe->origin_bios list and
DM_MAPIO_SUBMITTED is returned. This ensures that neither writes to the
origin nor snapshot merging can overwrite the origin chunk, until all
conflicting reads drain, and thus snapshot reads will not return
corrupted data.
Summarizing, we now have the following possible combinations of pending
and completed exceptions for a chunk, along with their meaning:
A. No exceptions exist: The chunk has not been remapped yet.
B. Only a pending exception exists: The chunk is currently being copied
to the COW device.
C. Both a pending and a completed exception exist: COW for this chunk
has completed but there are snapshot reads in flight which had been
redirected to the origin before the chunk was remapped.
D. Only the completed exception exists: COW has been completed and there
are no conflicting reads in flight.
Co-developed-by: Ilias Tsitsimpis <iliastsi@arrikto.com>
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Acked-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Add missing dm_bitset_cursor_next() to properly advance the bitset
cursor.
Otherwise, the discarded state of all blocks is set according to the
discarded state of the first block.
Fixes: ae4a46a1f6 ("dm cache metadata: use bitset cursor api to load discard bitset")
Cc: stable@vger.kernel.org
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
The function blkdev_report_zones() returns success even if no zone
information is reported (empty report). Empty zone reports can only
happen if the report start sector passed exceeds the device capacity.
The conditions for this to happen are either a bug in the caller code,
or, a change in the device that forced the low level driver to change
the device capacity to a value that is lower than the report start
sector. This situation includes a failed disk revalidation resulting in
the disk capacity being changed to 0.
If this change happens while dm-zoned is in its initialization phase
executing dmz_init_zones(), this function may enter an infinite loop
and hang the system. To avoid this, add a check to disallow empty zone
reports and bail out early. Also fix the function dmz_update_zone() to
make sure that the report for the requested zone was correctly obtained.
Fixes: 3b1a94c88b79 ("dm zoned: drive-managed zoned block device target")
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Shaun Tancheff <shaun@tancheff.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
My static checker complains about this line from dmz_get_zoned_device()
aligned_capacity = dev->capacity & ~(blk_queue_zone_sectors(q) - 1);
The problem is that "aligned_capacity" and "dev->capacity" are sector_t
type (which is a u64 under most configs) but blk_queue_zone_sectors(q)
returns a u32 so the higher 32 bits in aligned_capacity are cleared to
zero. This patch adds a cast to address the issue.
Fixes: 114e025968b5 ("dm zoned: ignore last smaller runt zone")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
The sector used here is a little endian value, so use the right
type for it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
The problem is that any 'uptodate' vs 'disks' check is not precise
in this path. Put a "WARN_ON(!test_bit(R5_UPTODATE, &dev->flags)" on the
device that might try to kick off writes and then skip the action.
Better to prevent the raid driver from taking unexpected action *and* keep
the system alive vs killing the machine with BUG_ON.
Note: fixed warning reported by kbuild test robot <lkp@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Nigel Croxon <ncroxon@redhat.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
This reverts commit 4f4fd7c5798bbdd5a03a60f6269cf1177fbd11ef.
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Nigel Croxon <ncroxon@redhat.com>
Cc: Xiao Ni <xni@redhat.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
Mdadm expects that setting drive as faulty will fail with -EBUSY only if
this operation will cause RAID to be failed. If this happens, it will
try to stop the array. Currently -EBUSY might also be returned if rdev
is in the middle of the removal process - for example there is a race
with mdmon that already requested the drive to be failed/removed.
If rdev does not contain mddev, return -ENODEV instead, so the caller
can distinguish between those two cases and behave accordingly.
Reviewed-by: NeilBrown <neilb@suse.com>
Signed-off-by: Pawel Baldysiak <pawel.baldysiak@intel.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
Pull in v5.1-rc5 to resolve two conflicts. One is in BFQ, in just
a comment, and is trivial. The other one is a conflict due to a
later fix in the bio multi-page work, and needs a bit more care.
* tag 'v5.1-rc5': (476 commits)
Linux 5.1-rc5
fs: prevent page refcount overflow in pipe_buf_get
mm: prevent get_user_pages() from overflowing page refcount
mm: add 'try_get_page()' helper function
mm: make page ref count overflow check tighter and more explicit
clk: imx: Fix PLL_1416X not rounding rates
clk: mediatek: fix clk-gate flag setting
arm64: futex: Fix FUTEX_WAKE_OP atomic ops with non-zero result value
iommu/amd: Set exclusion range correctly
clang-format: Update with the latest for_each macro list
perf/core: Fix perf_event_disable_inatomic() race
block: fix the return errno for direct IO
Revert "SUNRPC: Micro-optimise when the task is known not to be sleeping"
NFSv4.1 fix incorrect return value in copy_file_range
xprtrdma: Fix helper that drains the transport
NFS: Fix handling of reply page vector
NFS: Forbid setting AF_INET6 to "struct sockaddr_in"->sin_family.
dma-debug: only skip one stackframe entry
platform/x86: pmc_atom: Drop __initconst on dmi table
nvmet: fix discover log page when offsets are used
...
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
This tells sparse that we release and reacquire the device_lock and
avoids a warning.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
This tells sparse that we acquire/release the two stripe locks and
avoids a warning.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
Sparse complains that it has no external declaration, and it turns out
that it is never even used outside of md.c. So just mark it static
and drop the export.
Acked-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
If we want to convert from a little endian format we need to cast
to a little endian type, otherwise sparse will be unhappy.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
If we want to convert from a little endian format we need to cast
to a little endian type, otherwise sparse will be unhappy.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
If we want to convert from a little endian format we need to cast
to a little endian type, otherwise sparse will be unhappy.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
The on-disk value is little endian and we need to convert it to
native endian before storing the value in the in-core structure.
Fixes: 7564beda19b36 ("md-cluster/raid10: support add disk under grow mode")
Cc: <stable@vger.kernel.org> # 4.20+
Acked-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
When doing re-add, we need to ensure rdev->mddev->pers is not NULL,
which can avoid potential NULL pointer derefence in fallowing
add_bound_rdev().
Fixes: a6da4ef85cef ("md: re-add a failed disk")
Cc: Xiao Ni <xni@redhat.com>
Cc: NeilBrown <neilb@suse.com>
Cc: <stable@vger.kernel.org> # 4.4+
Reviewed-by: NeilBrown <neilb@suse.com>
Signed-off-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
|
|
Currently support for 64-bit sector_t and blkcnt_t is optional on 32-bit
architectures. These types are required to support block device and/or
file sizes larger than 2 TiB, and have generally defaulted to on for
a long time. Enabling the option only increases the i386 tinyconfig
size by 145 bytes, and many data structures already always use
64-bit values for their in-core and on-disk data structures anyway,
so there should not be a large change in dynamic memory usage either.
Dropping this option removes a somewhat weird non-default config that
has cause various bugs or compiler warnings when actually used.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
dm-integrity will deadlock if overlapping I/O is issued to it, the bug
was introduced by commit 724376a04d1a ("dm integrity: implement fair
range locks"). Users rarely use overlapping I/O so this bug went
undetected until now.
Fix this bug by correcting, likely cut-n-paste, typos in
ranges_overlap() and also remove a flawed ranges_overlap() check in
remove_range_unlocked(). This condition could leave unprocessed bios
hanging on wait_list forever.
Cc: stable@vger.kernel.org # v4.19+
Fixes: 724376a04d1a ("dm integrity: implement fair range locks")
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Storage devices which report supporting discard commands like
WRITE_SAME_16 with unmap, but reject discard commands sent to the
storage device. This is a clear storage firmware bug but it doesn't
change the fact that should a program cause discards to be sent to a
multipath device layered on this buggy storage, all paths can end up
failed at the same time from the discards, causing possible I/O loss.
The first discard to a path will fail with Illegal Request, Invalid
field in cdb, e.g.:
kernel: sd 8:0:8:19: [sdfn] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
kernel: sd 8:0:8:19: [sdfn] tag#0 Sense Key : Illegal Request [current]
kernel: sd 8:0:8:19: [sdfn] tag#0 Add. Sense: Invalid field in cdb
kernel: sd 8:0:8:19: [sdfn] tag#0 CDB: Write same(16) 93 08 00 00 00 00 00 a0 08 00 00 00 80 00 00 00
kernel: blk_update_request: critical target error, dev sdfn, sector 10487808
The SCSI layer converts this to the BLK_STS_TARGET error number, the sd
device disables its support for discard on this path, and because of the
BLK_STS_TARGET error multipath fails the discard without failing any
path or retrying down a different path. But subsequent discards can
cause path failures. Any discards sent to the path which already failed
a discard ends up failing with EIO from blk_cloned_rq_check_limits with
an "over max size limit" error since the discard limit was set to 0 by
the sd driver for the path. As the error is EIO, this now fails the
path and multipath tries to send the discard down the next path. This
cycle continues as discards are sent until all paths fail.
Fix this by training DM core to disable DISCARD if the underlying
storage already did so.
Also, fix branching in dm_done() and clone_endio() to reflect the
mutually exclussive nature of the IO operations in question.
Cc: stable@vger.kernel.org
Reported-by: David Jeffery <djeffery@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Some devices don't use blk_integrity but still want stable pages
because they do their own checksumming. Examples include rbd and iSCSI
when data digests are negotiated. Stacking DM (and thus LVM) on top of
these devices results in sporadic checksum errors.
Set BDI_CAP_STABLE_WRITES if any underlying device has it set.
Cc: stable@vger.kernel.org
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
