Previously we used separate boolean values to track quiescing and
copying actions. By switching to an atomic_t we can support blocks that
need a partial copy and partial zero.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
DM thinp already checks whether the discard_granularity of the data
device is a factor of the thin-pool block size. But when using the
dm-thin-pool's discard passdown support, DM thinp was not selecting the
max of the underlying data device's discard_granularity and the
thin-pool's block size.
Update set_discard_limits() to set discard_granularity to the max of
these values. This enables blkdev_issue_discard() to properly align the
discards that are sent to the DM thin device on a full block boundary.
As such each discard will now cover an entire DM thin-pool block and the
block will be reclaimed.
Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Update the DM thin provisioning target's allocation failure error to be
consistent with commit a9d6ceb8 ("[SCSI] return ENOSPC on thin
provisioning failure").
The DM thin target now returns -ENOSPC rather than -EIO when
block allocation fails due to the pool being out of data space (and
the 'error_if_no_space' thin-pool feature is enabled).
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-By: Joe Thornber <ejt@redhat.com>
Factor out a pool_work interface that noflush_work makes use of to wait
for and complete work items (in terms of a proper completion struct).
Allows discontinuing the use of a custom completion in terms of atomic_t
and wait_event.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Commit 85ad643b ("dm thin: add timeout to stop out-of-data-space mode
holding IO forever") introduced a fixed 60 second timeout. Users may
want to either disable or modify this timeout.
Allow the out-of-data-space timeout to be configured using the
'no_space_timeout' dm-thin-pool module param. Setting it to 0 will
disable the timeout, resulting in IO being queued until more data space
is added to the thin-pool.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org # 3.14+
If the pool runs out of data space, dm-thin can be configured to
either error IOs that would trigger provisioning, or hold those IOs
until the pool is resized. Unfortunately, holding IOs until the pool is
resized can result in a cascade of tasks hitting the hung_task_timeout,
which may render the system unavailable.
Add a fixed timeout so IOs can only be held for a maximum of 60 seconds.
If LVM is going to resize a thin-pool that is out of data space it needs
to be prompt about it.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org # 3.14+
Commit 3e1a0699 ("dm thin: fix out of data space handling") introduced
a regression in the metadata commit() method by returning an error if
the pool is in PM_OUT_OF_DATA_SPACE mode. This oversight caused a thin
device to return errors even if the default queue_if_no_space ENOSPC
handling mode is used.
Fix commit() to only fail if pool is in PM_READ_ONLY or PM_FAIL mode.
Reported-by: qindehua@163.com
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org # 3.14+
Use INIT_WORK_ONSTACK to silence "ODEBUG: object is on stack, but not
annotated".
Reported-by: Zdeněk Kabeláč <zkabelac@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Commit c140e1c4e2 ("dm thin: use per thin device deferred bio lists")
introduced the use of an rculist for all active thin devices. The use
of rcu_read_lock() in process_deferred_bios() can result in a BUG if a
dm_bio_prison_cell must be allocated as a side-effect of bio_detain():
BUG: sleeping function called from invalid context at mm/mempool.c:203
in_atomic(): 1, irqs_disabled(): 0, pid: 6, name: kworker/u8:0
3 locks held by kworker/u8:0/6:
#0: ("dm-" "thin"){.+.+..}, at: [<ffffffff8106be42>] process_one_work+0x192/0x550
#1: ((&pool->worker)){+.+...}, at: [<ffffffff8106be42>] process_one_work+0x192/0x550
#2: (rcu_read_lock){.+.+..}, at: [<ffffffff816360b5>] do_worker+0x5/0x4d0
We can't process deferred bios with the rcu lock held, since
dm_bio_prison_cell allocation may block if the bio-prison's cell mempool
is exhausted.
To fix:
- Introduce a refcount and completion field to each thin_c
- Add thin_get/put methods for adjusting the refcount. If the refcount
hits zero then the completion is triggered.
- Initialise refcount to 1 when creating thin_c
- When iterating the active_thins list we thin_get() whilst the rcu
lock is held.
- After the rcu lock is dropped we process the deferred bios for that
thin.
- When destroying a thin_c we thin_put() and then wait for the
completion -- to avoid a race between the worker thread iterating
from that thin_c and destroying the thin_c.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Commit c140e1c4e2 ("dm thin: use per thin device deferred bio lists")
incorrectly stopped disabling irqs when taking the pool's spinlock.
Irqs must be disabled when taking the pool's spinlock otherwise a thread
could spin_lock(), then get interrupted to service thin_endio() in
interrupt context, which would then deadlock in spin_lock_irqsave().
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
A thin-pool will allocate blocks using FIFO order for all thin devices
which share the thin-pool. Because of this simplistic allocation the
thin-pool's space can become fragmented quite easily; especially when
multiple threads are requesting blocks in parallel.
Sort each thin device's deferred_bio_list based on logical sector to
help reduce fragmentation of the thin-pool's ondisk layout.
The following tables illustrate the realized gains/potential offered by
sorting each thin device's deferred_bio_list. An "io size"-sized random
read of the device would result in "seeks/io" fragments being read, with
an average "distance/seek" between each fragment.
Data was written to a single thin device using multiple threads via
iozone (8 threads, 64K for both the block_size and io_size).
unsorted:
io size seeks/io distance/seek
--------------------------------------
4k 0.000 0b
16k 0.013 11m
64k 0.065 11m
256k 0.274 10m
1m 1.109 10m
4m 4.411 10m
16m 17.097 11m
64m 60.055 13m
256m 148.798 25m
1g 809.929 21m
sorted:
io size seeks/io distance/seek
--------------------------------------
4k 0.000 0b
16k 0.000 1g
64k 0.001 1g
256k 0.003 1g
1m 0.011 1g
4m 0.045 1g
16m 0.181 1g
64m 0.747 1011m
256m 3.299 1g
1g 14.373 1g
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
The thin-pool previously only had a single deferred_bios list that would
collect bios for all thin devices in the pool. Split this per-pool
deferred_bios list out to per-thin deferred_bios_list -- doing so
enables increased parallelism when processing deferred bios. And now
that each thin device has it's own deferred_bios_list we can sort all
bios in the list using logical sector. The requeue code in error
handling path is also cleaner as a side-effect.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
The pool is congested if the pool is in PM_OUT_OF_DATA_SPACE mode. This
is more explicit/clear/efficient than inferring whether or not the pool
is congested by checking if retry_on_resume_list is empty.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
If unable to ensure_next_mapping() we must add the current bio, which
was removed from the @bios list via bio_list_pop, back to the
deferred_bios list before all the remaining @bios.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
i) by the time DM core calls the postsuspend hook the dm_noflush flag
has been cleared. So the old thin_postsuspend did nothing. We need to
use the presuspend hook instead.
ii) There was a race between bios leaving DM core and arriving in the
deferred queue.
thin_presuspend now sets a 'requeue' flag causing all bios destined for
that thin to be requeued back to DM core. Then it requeues all held IO,
and all IO on the deferred queue (destined for that thin). Finally
postsuspend clears the 'requeue' flag.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
The spin lock in requeue_io() was held for too long, allowing deadlock.
Don't worry, due to other issues addressed in the following "dm thin:
fix noflush suspend IO queueing" commit, this code was never called.
Fix this by taking the spin lock for a much shorter period of time.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Ideally a thin pool would never run out of data space; the low water
mark would trigger userland to extend the pool before we completely run
out of space. However, many small random IOs to unprovisioned space can
consume data space at an alarming rate. Adjust your low water mark if
you're frequently seeing "out-of-data-space" mode.
Before this fix, if data space ran out the pool would be put in
PM_READ_ONLY mode which also aborted the pool's current metadata
transaction (data loss for any changes in the transaction). This had a
side-effect of needlessly compromising data consistency. And retry of
queued unserviceable bios, once the data pool was resized, could
initiate changes to potentially inconsistent pool metadata.
Now when the pool's data space is exhausted transition to a new pool
mode (PM_OUT_OF_DATA_SPACE) that allows metadata to be changed but data
may not be allocated. This allows users to remove thin volumes or
discard data to recover data space.
The pool is no longer put in PM_READ_ONLY mode in response to the pool
running out of data space. And PM_READ_ONLY mode no longer aborts the
pool's current metadata transaction. Also, set_pool_mode() will now
notify userspace when the pool mode is changed.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
If a thin metadata operation fails the current transaction will abort,
whereby causing potential for IO layers up the stack (e.g. filesystems)
to have data loss. As such, set THIN_METADATA_NEEDS_CHECK_FLAG in the
thin metadata's superblock which:
1) requires the user verify the thin metadata is consistent (e.g. use
thin_check, etc)
2) suggests the user verify the thin data is consistent (e.g. use fsck)
The only way to clear the superblock's THIN_METADATA_NEEDS_CHECK_FLAG is
to run thin_repair.
On metadata operation failure: abort current metadata transaction, set
pool in read-only mode, and now set the needs_check flag.
As part of this change, constraints are introduced or relaxed:
* don't allow a pool to transition to write mode if needs_check is set
* don't allow data or metadata space to be resized if needs_check is set
* if a thin pool's metadata space is exhausted: the kernel will now
force the user to take the pool offline for repair before the kernel
will allow the metadata space to be extended.
Also, update Documentation to include information about when the thin
provisioning target commits metadata, how it handles metadata failures
and running out of space.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Commit b5330655 ("dm thin: handle metadata failures more consistently")
increased potential for the pool's mode to be changed in response to
metadata operation failures.
When the pool mode is changed it isn't synchronized with the mode in
pool_features stored in the target's context (ti->private) that is used
as the basis for (re)establishing the pool mode during resume via
bind_control_target.
It is important that we synchronize the pool mode when it is changed
otherwise the pool may experience and unexpected mode transition on the
next resume (especially if there was no new table load).
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
It was always intended that a user could provide a thin metadata device
that is larger than the max supported by the on-disk format. The extra
space would just go unused.
Unfortunately that never worked. If the user attempted to use a larger
metadata device on creation they would get an error like the following:
device-mapper: space map common: space map too large
device-mapper: transaction manager: couldn't create metadata space map
device-mapper: thin metadata: tm_create_with_sm failed
device-mapper: table: 252:17: thin-pool: Error creating metadata object
device-mapper: ioctl: error adding target to table
Fix this by allowing the initial metadata space map creation to cap its
size at the max number of blocks supported (DM_SM_METADATA_MAX_BLOCKS).
get_metadata_dev_size() must also impose DM_SM_METADATA_MAX_BLOCKS (via
THIN_METADATA_MAX_SECTORS), otherwise extending metadata would cap at
THIN_METADATA_MAX_SECTORS_WARNING (which is larger than supported).
Also, the calculation for THIN_METADATA_MAX_SECTORS didn't account for
the sizeof the disk_bitmap_header. So the supported maximum metadata
size is a bit smaller (reduced from 33423360 to 33292800 sectors).
Lastly, remove the "excess space will not be used" warning message from
get_metadata_dev_size(); it resulted in printing the warning multiple
times. Factor out warn_if_metadata_device_too_big(), call it from
pool_ctr() and maybe_resize_metadata_dev().
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
dm_pool_close_thin_device() must be called if dm_set_target_max_io_len()
fails in thin_ctr(). Otherwise __pool_destroy() will fail because the
pool will still have an open thin device:
device-mapper: thin metadata: attempt to close pmd when 1 device(s) are still open
device-mapper: thin: __pool_destroy: dm_pool_metadata_close() failed.
Also, must establish error code if failing thin_ctr() because the pool
is in fail_io mode.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
Commit 905e51b ("dm thin: commit outstanding data every second")
introduced a periodic commit. This commit occurs regardless of whether
any thin devices have made changes.
Fix the periodic commit to check if any of a pool's thin devices have
changed using dm_pool_changed_this_transaction().
Reported-by: Alexander Larsson <alexl@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
Pull core block IO changes from Jens Axboe:
"The major piece in here is the immutable bio_ve series from Kent, the
rest is fairly minor. It was supposed to go in last round, but
various issues pushed it to this release instead. The pull request
contains:
- Various smaller blk-mq fixes from different folks. Nothing major
here, just minor fixes and cleanups.
- Fix for a memory leak in the error path in the block ioctl code
from Christian Engelmayer.
- Header export fix from CaiZhiyong.
- Finally the immutable biovec changes from Kent Overstreet. This
enables some nice future work on making arbitrarily sized bios
possible, and splitting more efficient. Related fixes to immutable
bio_vecs:
- dm-cache immutable fixup from Mike Snitzer.
- btrfs immutable fixup from Muthu Kumar.
- bio-integrity fix from Nic Bellinger, which is also going to stable"
* 'for-3.14/core' of git://git.kernel.dk/linux-block: (44 commits)
xtensa: fixup simdisk driver to work with immutable bio_vecs
block/blk-mq-cpu.c: use hotcpu_notifier()
blk-mq: for_each_* macro correctness
block: Fix memory leak in rw_copy_check_uvector() handling
bio-integrity: Fix bio_integrity_verify segment start bug
block: remove unrelated header files and export symbol
blk-mq: uses page->list incorrectly
blk-mq: use __smp_call_function_single directly
btrfs: fix missing increment of bi_remaining
Revert "block: Warn and free bio if bi_end_io is not set"
block: Warn and free bio if bi_end_io is not set
blk-mq: fix initializing request's start time
block: blk-mq: don't export blk_mq_free_queue()
block: blk-mq: make blk_sync_queue support mq
block: blk-mq: support draining mq queue
dm cache: increment bi_remaining when bi_end_io is restored
block: fixup for generic bio chaining
block: Really silence spurious compiler warnings
block: Silence spurious compiler warnings
block: Kill bio_pair_split()
...
Commit 787a996cb2 ("dm thin: add error_if_no_space feature")
mistakenly forgot to increase the number of feature args supported.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
The pool mode must not be switched until after the corresponding pool
process_* methods have been established. Otherwise, because
set_pool_mode() isn't interlocked with the IO path for performance
reasons, the IO path can end up executing process_* operations that
don't match the mode. This patch eliminates problems like the following
(as seen on really fast PCIe SSD storage when transitioning the pool's
mode from PM_READ_ONLY to PM_WRITE):
kernel: device-mapper: thin: 253:2: reached low water mark for data device: sending event.
kernel: device-mapper: thin: 253:2: no free data space available.
kernel: device-mapper: thin: 253:2: switching pool to read-only mode
kernel: device-mapper: thin: 253:2: switching pool to write mode
kernel: ------------[ cut here ]------------
kernel: WARNING: CPU: 11 PID: 7564 at drivers/md/dm-thin.c:995 handle_unserviceable_bio+0x146/0x160 [dm_thin_pool]()
...
kernel: Workqueue: dm-thin do_worker [dm_thin_pool]
kernel: 00000000000003e3 ffff880308831cc8 ffffffff8152ebcb 00000000000003e3
kernel: 0000000000000000 ffff880308831d08 ffffffff8104c46c ffff88032502a800
kernel: ffff880036409000 ffff88030ec7ce00 0000000000000001 00000000ffffffc3
kernel: Call Trace:
kernel: [<ffffffff8152ebcb>] dump_stack+0x49/0x5e
kernel: [<ffffffff8104c46c>] warn_slowpath_common+0x8c/0xc0
kernel: [<ffffffff8104c4ba>] warn_slowpath_null+0x1a/0x20
kernel: [<ffffffffa001e2c6>] handle_unserviceable_bio+0x146/0x160 [dm_thin_pool]
kernel: [<ffffffffa001f276>] process_bio_read_only+0x136/0x180 [dm_thin_pool]
kernel: [<ffffffffa0020b75>] process_deferred_bios+0xc5/0x230 [dm_thin_pool]
kernel: [<ffffffffa0020d31>] do_worker+0x51/0x60 [dm_thin_pool]
kernel: [<ffffffff81067823>] process_one_work+0x183/0x490
kernel: [<ffffffff81068c70>] worker_thread+0x120/0x3a0
kernel: [<ffffffff81068b50>] ? manage_workers+0x160/0x160
kernel: [<ffffffff8106e86e>] kthread+0xce/0xf0
kernel: [<ffffffff8106e7a0>] ? kthread_freezable_should_stop+0x70/0x70
kernel: [<ffffffff8153b3ec>] ret_from_fork+0x7c/0xb0
kernel: [<ffffffff8106e7a0>] ? kthread_freezable_should_stop+0x70/0x70
kernel: ---[ end trace 3f00528e08ffa55c ]---
kernel: device-mapper: thin: pool mode is PM_WRITE not PM_READ_ONLY like expected!?
dm-thin.c:995 was the WARN_ON_ONCE(get_pool_mode(pool) != PM_READ_ONLY);
at the top of handle_unserviceable_bio(). And as the additional
debugging I had conveys: the pool mode was _not_ PM_READ_ONLY like
expected, it was already PM_WRITE, yet pool->process_bio was still set
to process_bio_read_only().
Also, while fixing this up, reduce logging of redundant pool mode
transitions by checking new_mode is different from old_mode.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
The pool's error_if_no_space flag can easily serve the same purpose that
no_free_space did, namely: control whether handle_unserviceable_bio()
will error a bio or requeue it.
This is cleaner since error_if_no_space is established when the pool's
features are processed during table load. So it avoids managing the
no_free_space flag by taking the pool's spinlock.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
If the pool runs out of data or metadata space, the pool can either
queue or error the IO destined to the data device. The default is to
queue the IO until more space is added.
An admin may now configure the pool to error IO when no space is
available by setting the 'error_if_no_space' feature when loading the
thin-pool table.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Now that we switch the pool to read-only mode when the data device runs
out of space it causes active writers to get IO errors once we resume
after resizing the data device.
If no_free_space is set, save bios to the 'retry_on_resume_list' and
requeue them on resume (once the data or metadata device may have been
resized).
With this patch the resize_io test passes again (on slower storage):
dmtest run --suite thin-provisioning -n /resize_io/
Later patches fix some subtle races associated with the pool mode
transitions done as part of the pool's -ENOSPC handling. These races
are exposed on fast storage (e.g. PCIe SSD).
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Factor out_of_data_space() out of alloc_data_block(). Eliminate the use
of 'no_free_space' as a latch in alloc_data_block() -- this is no longer
needed now that we switch to read-only mode when we run out of data or
metadata space. In a later patch, the 'no_free_space' flag will be
eliminated entirely (in favor of checking metadata rather than relying
on a transient flag).
Move no metdata space handling into metdata_operation_failed(). Set
no_free_space when metadata space is exhausted too. This is useful,
because it offers consistency, for the following patch that will requeue
data IOs if no_free_space.
Also, rename no_space() to retry_bios_on_resume().
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Introduce metadata_operation_failed() wrappers, around set_pool_mode(),
to assist with improving the consistency of how metadata failures are
handled. Logging is improved and metadata operation failures trigger
read-only mode immediately.
Also, eliminate redundant set_pool_mode() calls in the two
alloc_data_block() caller's error paths.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Factor check_low_water_mark() out of alloc_data_block().
Change a couple unsigned flags in the pool structure to bool.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Mappings could be processed in descending logical block order,
particularly if buffered IO is used. This could adversely affect the
latency of IO processing. Fix this by adding mappings to the end of the
'prepared_mappings' and 'prepared_discards' lists.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Also, move 'err' member in dm_thin_new_mapping structure to eliminate 4
byte hole (reduces size from 88 bytes to 80).
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
If a snapshot is created and later deleted the origin dm_thin_device's
snapshotted_time will have been updated to reflect the snapshot's
creation time. The 'shared' flag in the dm_thin_lookup_result struct
returned from dm_thin_find_block() is an approximation based on
snapshotted_time -- this is done to avoid 0(n), or worse, time
complexity. In this case, the shared flag would be true.
But because the 'shared' flag reflects an approximation a block can be
incorrectly assumed to be shared (e.g. false positive for 'shared'
because the snapshot no longer exists). This could result in discards
issued to a thin device not being passed down to the pool's underlying
data device.
To fix this we double check that a thin block is really still in-use
after a mapping is removed using dm_pool_block_is_used(). If the
reference count for a block is now zero the discard is allowed to be
passed down.
Also add a 'definitely_not_shared' member to the dm_thin_new_mapping
structure -- reflects that the 'shared' flag in the response from
dm_thin_find_block() can only be held as definitive if false is
returned.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1043527
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
As additional members are added to the dm_thin_new_mapping structure
care should be taken to make sure they get initialized before use.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJSwLfoAAoJEHm+PkMAQRiGi6QH/1U1B7lmHChDTw3jj1lfm9gA
189Si4QJlnxFWCKHvKEL+pcaVuACU+aMGI8+KyMYK4/JfuWVjjj5fr/SvyHH2/8m
LdSK8aHMhJ46uBS4WJ/l6v46qQa5e2vn8RKSBAyKm/h4vpt+hd6zJdoFrFai4th7
k/TAwOAEHI5uzexUChwLlUBRTvbq4U8QUvDu+DeifC8cT63CGaaJ4qVzjOZrx1an
eP6UXZrKDASZs7RU950i7xnFVDQu4PsjlZi25udsbeiKcZJgPqGgXz5ULf8ZH8RQ
YCi1JOnTJRGGjyIOyLj7pyB01h7XiSM2+eMQ0S7g54F2s7gCJ58c2UwQX45vRWU=
=/4/R
-----END PGP SIGNATURE-----
Merge tag 'v3.13-rc6' into for-3.14/core
Needed to bring blk-mq uptodate, since changes have been going in
since for-3.14/core was established.
Fixup merge issues related to the immutable biovec changes.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Conflicts:
block/blk-flush.c
fs/btrfs/check-integrity.c
fs/btrfs/extent_io.c
fs/btrfs/scrub.c
fs/logfs/dev_bdev.c
A thin-pool may be in read-only mode because the pool's data or metadata
space was exhausted. To allow for recovery, by adding more space to the
pool, we must allow a pool to transition from PM_READ_ONLY to PM_WRITE
mode. Otherwise, running out of space will render the pool permanently
read-only.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
If the thin-pool transitioned to fail mode and the thin-pool's table
were reloaded for some reason: the new table's default pool mode would
be read-write, though it will transition to fail mode during resume.
When the pool mode transitions directly from PM_WRITE to PM_FAIL we need
to re-establish the intermediate read-only state in both the metadata
and persistent-data block manager (as is usually done with the normal
pool mode transition sequence: PM_WRITE -> PM_READ_ONLY -> PM_FAIL).
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Rename commit_or_fallback() to commit(). Now all previous calls to
commit() will trigger the pool mode to fallback if the commit fails.
Also, check the error returned from commit() in alloc_data_block().
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Switch the thin pool to read-only mode in alloc_data_block() if
dm_pool_alloc_data_block() fails because the pool's metadata space is
exhausted.
Differentiate between data and metadata space in messages about no
free space available.
This issue was noticed with the device-mapper-test-suite using:
dmtest run --suite thin-provisioning -n /exhausting_metadata_space_causes_fail_mode/
The quantity of errors logged in this case must be reduced.
before patch:
device-mapper: thin: 253:4: reached low water mark for metadata device: sending event.
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
<snip ... these repeat for a _very_ long while ... >
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: 253:4: commit failed: error = -28
device-mapper: thin: 253:4: switching pool to read-only mode
after patch:
device-mapper: thin: 253:4: reached low water mark for metadata device: sending event.
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: 253:4: no free metadata space available.
device-mapper: thin: 253:4: switching pool to read-only mode
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
Switch the thin pool to read-only mode when dm_thin_insert_block() fails
since there is little reason to expect the cause of the failure to be
resolved without further action by user space.
This issue was noticed with the device-mapper-test-suite using:
dmtest run --suite thin-provisioning -n /exhausting_metadata_space_causes_fail_mode/
The quantity of errors logged in this case must be reduced.
before patch:
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: dm_thin_insert_block() failed
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map metadata: unable to allocate new metadata block
<snip ... these repeat for a long while ... >
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: space map common: dm_tm_shadow_block() failed
device-mapper: thin: 253:4: no free metadata space available.
device-mapper: thin: 253:4: switching pool to read-only mode
after patch:
device-mapper: space map metadata: unable to allocate new metadata block
device-mapper: thin: 253:4: dm_thin_insert_block() failed: error = -28
device-mapper: thin: 253:4: switching pool to read-only mode
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
This adds a generic mechanism for chaining bio completions. This is
going to be used for a bio_split() replacement, and it turns out to be
very useful in a fair amount of driver code - a fair number of drivers
were implementing this in their own roundabout ways, often painfully.
Note that this means it's no longer to call bio_endio() more than once
on the same bio! This can cause problems for drivers that save/restore
bi_end_io. Arguably they shouldn't be saving/restoring bi_end_io at all
- in all but the simplest cases they'd be better off just cloning the
bio, and immutable biovecs is making bio cloning cheaper. But for now,
we add a bio_endio_nodec() for these cases.
Signed-off-by: Kent Overstreet <kmo@daterainc.com>
Cc: Jens Axboe <axboe@kernel.dk>
Fix issue where the block layer would stack the discard limits of the
pool's data device even if the "ignore_discard" pool feature was
specified.
The pool and thin device(s) still had discards disabled because the
QUEUE_FLAG_DISCARD request_queue flag wasn't set. But to avoid user
confusion when "ignore_discard" is used: both the pool device and the
thin device(s) have zeroes for all discard limits.
Also, always set discard_zeroes_data_unsupported in targets because they
should never advertise the 'discard_zeroes_data' capability (even if the
pool's data device supports it).
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
If pool has 'no_free_space' set it means a previous allocation already
determined the pool has no free space (and failed that allocation with
-ENOSPC). By always returning -ENOSPC if 'no_free_space' is set, we do
not allow the pool to oscillate between allocating blocks and then not.
But a side-effect of this determinism is that if a user wants to be able
to allocate new blocks they'll need to reload the pool's table (to clear
the 'no_free_space' flag). This reload will happen automatically if the
pool's data volume is resized. But if the user takes action to free a
lot of space by deleting snapshot volumes, etc the pool will no longer
allow data allocations to continue without an intervening table reload.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
break_sharing() now handles an arbitrary alloc_data_block() error
the same way as provision_block(): marks pool read-only and errors the
cell.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Useful to know which pool is experiencing the error.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Do not blindly override the queue limits (specifically io_min and
io_opt). Allow traditional stacking of these limits if io_opt is a
factor of the thin-pool's data block size.
Without this patch mkfs.xfs does not recognize the thin device's
provided limits as a useful geometry (e.g. raid) so these hints are
ignored. This was due to setting io_min to a useless value.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Fix detection of the need to resize the dm thin metadata device.
The code incorrectly tried to extend the metadata device when it
didn't need to due to a merging error with patch 24347e9 ("dm thin:
detect metadata device resizing").
device-mapper: transaction manager: couldn't open metadata space map
device-mapper: thin metadata: tm_open_with_sm failed
device-mapper: thin: aborting transaction failed
device-mapper: thin: switching pool to failure mode
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Generate a dm event when the amount of remaining thin pool metadata
space falls below a certain level.
The threshold is taken to be a quarter of the size of the metadata
device with a minimum threshold of 4MB.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Allow the dm thin pool metadata device to be extended.
Whenever a pool is resumed, detect whether the size of the metadata
device has increased, and if so, extend the metadata to use the new
space.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
If a thin pool is created in read-only-metadata mode then only open the
metadata device read-only.
Previously it was always opened with FMODE_READ | FMODE_WRITE.
(Note that dm_get_device() still allows read-only dm devices to be used
read-write at the moment: If I create a read-only linear device for the
metadata, via dmsetup load --readonly, then I can still create a rw pool
out of it.)
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Refactor device size functions in preparation for similar metadata
device resizing functions.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix a discard granularity calculation to work for non power of 2 block sizes.
In order for thinp to passdown discard bios to the underlying data
device, the data device must have a discard granularity that is a
factor of the thinp block size. Originally this check was done by
using bitops since the block_size was known to be a power of two.
Introduced by commit f13945d757
("dm thin: support a non power of 2 discard_granularity").
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix a bug in dm_btree_remove that could leave leaf values with incorrect
reference counts. The effect of this was that removal of a shared block
could result in the space maps thinking the block was no longer used.
More concretely, if you have a thin device and a snapshot of it, sending
a discard to a shared region of the thin could corrupt the snapshot.
Thinp uses a 2-level nested btree to store it's mappings. This first
level is indexed by thin device, and the second level by logical
block.
Often when we're removing an entry in this mapping tree we need to
rebalance nodes, which can involve shadowing them, possibly creating a
copy if the block is shared. If we do create a copy then children of
that node need to have their reference counts incremented. In this
way reference counts percolate down the tree as shared trees diverge.
The rebalance functions were incrementing the children at the
appropriate time, but they were always assuming the children were
internal nodes. This meant the leaf values (in our case packed
block/flags entries) were not being incremented.
Cc: stable@vger.kernel.org
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch takes advantage of the new bio-prison interface where the
memory is now passed in rather than using a mempool in bio-prison.
This allows the map function to avoid performing potentially-blocking
allocations that could lead to deadlocks: We want to avoid the cell
allocation that is done in bio_detain.
(The potential for mempool deadlocks still remains in other functions
that use bio_detain.)
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Change the dm_bio_prison interface so that instead of allocating memory
internally, dm_bio_detain is supplied with a pre-allocated cell each
time it is called.
This enables a subsequent patch to move the allocation of the struct
dm_bio_prison_cell outside the thin target's mapping function so it can
no longer block there.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch allows the administrator to reduce the rate at which kcopyd
issues I/O.
Each module that uses kcopyd acquires a throttle parameter that can be
set in /sys/module/*/parameters.
We maintain a history of kcopyd usage by each module in the variables
io_period and total_period in struct dm_kcopyd_throttle. The actual
kcopyd activity is calculated as a percentage of time equal to
"(100 * io_period / total_period)". This is compared with the user-defined
throttle percentage threshold and if it is exceeded, we sleep.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Use 'bio' in the name of variables and functions that deal with
bios rather than 'request' to avoid confusion with the normal
block layer use of 'request'.
No functional changes.
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Use block_size_is_power_of_two() rather than checking
sectors_per_block_shift directly. Also introduce local pool variable in
get_bio_block() to eliminate redundant tc->pool dereferences.
No functional change.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Support a non-power-of-2 discard granularity in dm-thin, now that the block
layer supports this(via 8dd2cb7e88 "block:
discard granularity might not be power of 2" and
59771079c1 "blk: avoid divide-by-zero with zero
discard granularity").
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Avoid returning a truncated table or status string instead of setting
the DM_BUFFER_FULL_FLAG when the last target of a table fills the
buffer.
When processing a table or status request, the function retrieve_status
calls ti->type->status. If ti->type->status returns non-zero,
retrieve_status assumes that the buffer overflowed and sets
DM_BUFFER_FULL_FLAG.
However, targets don't return non-zero values from their status method
on overflow. Most targets returns always zero.
If a buffer overflow happens in a target that is not the last in the
table, it gets noticed during the next iteration of the loop in
retrieve_status; but if a buffer overflow happens in the last target, it
goes unnoticed and erroneously truncated data is returned.
In the current code, the targets behave in the following way:
* dm-crypt returns -ENOMEM if there is not enough space to store the
key, but it returns 0 on all other overflows.
* dm-thin returns errors from the status method if a disk error happened.
This is incorrect because retrieve_status doesn't check the error
code, it assumes that all non-zero values mean buffer overflow.
* all the other targets always return 0.
This patch changes the ti->type->status function to return void (because
most targets don't use the return code). Overflow is detected in
retrieve_status: if the status method fills up the remaining space
completely, it is assumed that buffer overflow happened.
Cc: stable@vger.kernel.org
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
thin_io_hints() is blindly copying the queue limits from the thin-pool
which can lead to incorrect limits being set. The fix here simply
deletes the thin_io_hints() hook which leaves the existing stacking
infrastructure to set the limits correctly.
When a thin-pool uses an MD device for the data device a thin device
from the thin-pool must respect MD's constraints about disallowing a bio
from spanning multiple chunks. Otherwise we can see problems. If the raid0
chunksize is 1152K and thin-pool chunksize is 256K I see the following
md/raid0 error (with extra debug tracing added to thin_endio) when
mkfs.xfs is executed against the thin device:
md/raid0:md99: make_request bug: can't convert block across chunks or bigger than 1152k 6688 127
device-mapper: thin: bio sector=2080 err=-5 bi_size=130560 bi_rw=17 bi_vcnt=32 bi_idx=0
This extra DM debugging shows that the failing bio is spanning across
the first and second logical 1152K chunk (sector 2080 + 255 takes the
bio beyond the first chunk's boundary of sector 2304). So the bio
splitting that DM is doing clearly isn't respecting the MD limits.
max_hw_sectors_kb is 127 for both the thin-pool and thin device
(queue_max_hw_sectors returns 255 so we'll excuse sysfs's lack of
precision). So this explains why bi_size is 130560.
But the thin device's max_hw_sectors_kb should be 4 (PAGE_SIZE) given
that it doesn't have a .merge function (for bio_add_page to consult
indirectly via dm_merge_bvec) yet the thin-pool does sit above an MD
device that has a compulsory merge_bvec_fn. This scenario is exactly
why DM must resort to sending single PAGE_SIZE bios to the underlying
layer. Some additional context for this is available in the header for
commit 8cbeb67a ("dm: avoid unsupported spanning of md stripe boundaries").
Long story short, the reason a thin device doesn't properly get
configured to have a max_hw_sectors_kb of 4 (PAGE_SIZE) is that
thin_io_hints() is blindly copying the queue limits from the thin-pool
device directly to the thin device's queue limits.
Fix this by eliminating thin_io_hints. Doing so is safe because the
block layer's queue limits stacking already enables the upper level thin
device to inherit the thin-pool device's discard and minimum_io_size and
optimal_io_size limits that get set in pool_io_hints. But avoiding the
queue limits copy allows the thin and thin-pool limits to be different
where it is important, namely max_hw_sectors_kb.
Reported-by: Daniel Browning <db@kavod.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch removes map_info from bio-based device mapper targets.
map_info is still used for request-based targets.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch removes endio_hook_pool from dm-thin and uses per-bio data instead.
This patch removes any use of map_info in preparation for the next patch
that removes map_info from bio-based device mapper.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Add WRITE SAME support to dm-io and make it accessible to
dm_kcopyd_zero(). dm_kcopyd_zero() provides an asynchronous interface
whereas the blkdev_issue_write_same() interface is synchronous.
WRITE SAME is a SCSI command that can be leveraged for more efficient
zeroing of a specified logical extent of a device which supports it.
Only a single zeroed logical block is transfered to the target for each
WRITE SAME and the target then writes that same block across the
specified extent.
The dm thin target uses this.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Throttle all errors logged from the IO path by dm thin.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Remove unused @data_block parameter from cell_defer.
Change thin_bio_map to use many returns rather than setting a variable.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Rename cell_defer_except() to cell_defer_no_holder() which describes
its function more clearly.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
If "ignore_discard" is specified when creating the thin pool device then
discard support is disabled for that device. The pool device's status
should reflect this fact rather than stating "no_discard_passdown"
(which implies discards are enabled but passdown is disabled).
Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
When discards are prepared it is best to directly wake the worker that
will process them. The worker will be woken anyway, via periodic
commit, but there is no reason to not wake_worker here.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
There is a race when discard bios and non-discard bios are issued
simultaneously to the same block.
Discard support is expensive for all thin devices precisely because you
have to be careful to quiesce the area you're discarding. DM thin must
handle this conflicting IO pattern (simultaneous non-discard vs discard)
even though a sane application shouldn't be issuing such IO.
The race manifests as follows:
1. A non-discard bio is mapped in thin_bio_map.
This doesn't lock out parallel activity to the same block.
2. A discard bio is issued to the same block as the non-discard bio.
3. The discard bio is locked in a dm_bio_prison_cell in process_discard
to lock out parallel activity against the same block.
4. The non-discard bio's mapping continues and its all_io_entry is
incremented so the bio is accounted for in the thin pool's all_io_ds
which is a dm_deferred_set used to track time locality of non-discard IO.
5. The non-discard bio is finally locked in a dm_bio_prison_cell in
process_bio.
The race can result in deadlock, leaving the block layer hanging waiting
for completion of a discard bio that never completes, e.g.:
INFO: task ruby:15354 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
ruby D ffffffff8160f0e0 0 15354 15314 0x00000000
ffff8802fb08bc58 0000000000000082 ffff8802fb08bfd8 0000000000012900
ffff8802fb08a010 0000000000012900 0000000000012900 0000000000012900
ffff8802fb08bfd8 0000000000012900 ffff8803324b9480 ffff88032c6f14c0
Call Trace:
[<ffffffff814e5a19>] schedule+0x29/0x70
[<ffffffff814e3d85>] schedule_timeout+0x195/0x220
[<ffffffffa06b9bc1>] ? _dm_request+0x111/0x160 [dm_mod]
[<ffffffff814e589e>] wait_for_common+0x11e/0x190
[<ffffffff8107a170>] ? try_to_wake_up+0x2b0/0x2b0
[<ffffffff814e59ed>] wait_for_completion+0x1d/0x20
[<ffffffff81233289>] blkdev_issue_discard+0x219/0x260
[<ffffffff81233e79>] blkdev_ioctl+0x6e9/0x7b0
[<ffffffff8119a65c>] block_ioctl+0x3c/0x40
[<ffffffff8117539c>] do_vfs_ioctl+0x8c/0x340
[<ffffffff8119a547>] ? block_llseek+0x67/0xb0
[<ffffffff811756f1>] sys_ioctl+0xa1/0xb0
[<ffffffff810561f6>] ? sys_rt_sigprocmask+0x86/0xd0
[<ffffffff814ef099>] system_call_fastpath+0x16/0x1b
The thinp-test-suite's test_discard_random_sectors reliably hits this
deadlock on fast SSD storage.
The fix for this race is that the all_io_entry for a bio must be
incremented whilst the dm_bio_prison_cell is held for the bio's
associated virtual and physical blocks. That cell locking wasn't
occurring early enough in thin_bio_map. This patch fixes this.
Care is taken to always call the new function inc_all_io_entry() with
the relevant cells locked, but they are generally unlocked before
calling issue() to try to avoid holding the cells locked across
generic_submit_request.
Also, now that thin_bio_map may lock bios in a cell, process_bio() is no
longer the only thread that will do so. Because of this we must be sure
to use cell_defer_except() to release all non-holder entries, that
were added by the other thread, because they must be deferred.
This patch depends on "dm thin: replace dm_cell_release_singleton with
cell_defer_except".
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Cc: stable@vger.kernel.org
Change existing users of the function dm_cell_release_singleton to share
cell_defer_except instead, and then remove the now-unused function.
Everywhere that calls dm_cell_release_singleton, the bio in question
is the holder of the cell.
If there are no non-holder entries in the cell then cell_defer_except
behaves exactly like dm_cell_release_singleton. Conversely, if there
*are* non-holder entries then dm_cell_release_singleton must not be used
because those entries would need to be deferred.
Consequently, it is safe to replace use of dm_cell_release_singleton
with cell_defer_except.
This patch is a pre-requisite for "dm thin: fix race between
simultaneous io and discards to same block".
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The bio prison code will be useful to other future DM targets so
move it to a separate module.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The bio prison code will be useful to share with future DM targets.
Prepare to move this code into a separate module, adding a dm prefix
to structures and functions that will be exported.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Support discards when the pool's block size is not a power of 2.
The block layer assumes discard_granularity is a power of 2 (in
blkdev_issue_discard), so we set this to the largest power of 2 that is
a divides into the number of sectors in each block, but never less than
DATA_DEV_BLOCK_SIZE_MIN_SECTORS.
This patch eliminates the "Discard support must be disabled when the
block size is not a power of 2" constraint that was imposed in commit
55f2b8b ("dm thin: support for non power of 2 pool blocksize"). That
commit was incomplete: using a block size that is not a power of 2
shouldn't mean disabling discard support on the device completely.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The discard limits that get established for a thin-pool or thin device
may be incompatible with the pool's data device. Avoid this by checking
the discard limits of the pool's data device. If an incompatibility is
found then the pool's 'discard passdown' feature is disabled.
Change thin_io_hints to ensure that a thin device always uses the same
queue limits as its pool device.
Introduce requested_pf to track whether or not the table line originally
contained the no_discard_passdown flag and use this directly for table
output. We prepare the correct setting for discard_passdown directly in
bind_control_target (called from pool_io_hints) and store it in
adjusted_pf rather than waiting until we have access to pool->pf in
pool_preresume.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
A little thin discard code refactoring to make the next patch (dm thin:
fix discard support for data devices) more readable.
Pull out a couple of functions (and uses bools instead of unsigned for
features).
No functional changes.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The dm thin pool target claims to support the zeroing of discarded
data areas. This turns out to be incorrect when processing discards
that do not exactly cover a complete number of blocks, so the target
must always set discard_zeroes_data_unsupported.
The thin pool target will zero blocks when they are allocated if the
skip_block_zeroing feature is not specified. The block layer
may send a discard that only partly covers a block. If a thin pool
block is partially discarded then there is no guarantee that the
discarded data will get zeroed before it is accessed again.
Due to this, thin devices cannot claim discards will always zero data.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org # 3.4+
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Commit outstanding metadata before returning the status for a dm thin
pool so that the numbers reported are as up-to-date as possible.
The commit is not performed if the device is suspended or if
the DM_NOFLUSH_FLAG is supplied by userspace and passed to the target
through a new 'status_flags' parameter in the target's dm_status_fn.
The userspace dmsetup tool will support the --noflush flag with the
'dmsetup status' and 'dmsetup wait' commands from version 1.02.76
onwards.
Tested-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Add read-only and fail-io modes to thin provisioning.
If a transaction commit fails the pool's metadata device will transition
to "read-only" mode. If a commit fails once already in read-only mode
the transition to "fail-io" mode occurs.
Once in fail-io mode the pool and all associated thin devices will
report a status of "Fail".
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Reduce the number of metadata commits by using
dm_thin_changed_this_transaction to check if metadata was changed on a
per thin device granularity.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Add a parameter to dm_pool_metadata_open to indicate whether or not an
unformatted metadata area should be formatted.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The thin provisioning target commits internal metadata on flush. So it
should receive flushes regardless of whether the underlying devices
support them.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
There's no need to break sharing, triggering a copy, for a write that has no
data (i.e. a flush).
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix memory leak in process_prepared_mapping by always freeing
the dm_thin_new_mapping structs from the mapping_pool mempool on
the error paths.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
dm-thin will be most likely used with a block size that is a power of
two. So it should be optimized for this case.
This patch changes division and modulo operations to shifts and bit
masks if block size is a power of two.
A test that bi_sector is divisible by a block size is removed from
io_overlaps_block. Device mapper never sends bios that span a block
boundary. Consequently, if we tested that bi_size is equivalent to block
size, bi_sector must already be on a block boundary.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch sets the variable "ti->split_discard_requests" for the dm thin
target so that device mapper core splits discard requests on a block
boundary.
Consequently, a discard request that spans multiple blocks is never sent
to dm-thin. The patch also removes some code in process_discard that
deals with discards that span multiple blocks.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Non power of 2 blocksize support is needed to properly align thinp IO
on storage that has non power of 2 optimal IO sizes (e.g. RAID6 10+2).
Use sector_div to support non power of 2 blocksize for the pool's
data device. This provides comparable performance to the power of 2
math that was performed until now (as tested on modern x86_64 hardware).
The kernel currently assumes that limits->discard_granularity is a power
of two so the thin target only enables discard support if the block
size is a power of two.
Eliminate pool structure's 'block_shift', 'offset_mask' and
remaining 4 byte holes.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Remove the restriction that limits a target's specified maximum incoming
I/O size to be a power of 2.
Rename this setting from 'split_io' to the less-ambiguous 'max_io_len'.
Change it from sector_t to uint32_t, which is plenty big enough, and
introduce a wrapper function dm_set_target_max_io_len() to set it.
Use sector_div() to process it now that it is not necessarily a power of 2.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Provide specific error message strings for two pool_ctr() failure cases
that currently give just "Unknown error".
Reference: test_two_pools_pointing_to_the_same_metadata_fails and
test_different_pool_cant_replace_pool in thinp-test-suite.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Clean up "warning: dubious: !x & y". Also make it clear that
__snapshotted_since() returns a bool and that dm_thin_lookup_result's
'shared' member is a flag.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Reduce the slab size used for the dm_thin_endio_hook mempool.
Allocation has been seen to fail on machines with smaller amounts
of memory due to fragmentation.
lvm: page allocation failure. order:5, mode:0xd0
device-mapper: table: 253:38: thin-pool: Error creating pool's endio_hook mempool
Cc: stable@vger.kernel.org
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
When process_discard receives a partial discard that doesn't cover a
full block, it sends this discard down to that block. Unfortunately, the
block can be shared and the discard would corrupt the other snapshots
sharing this block.
This patch detects block sharing and ends the discard with success when
sending it to the shared block.
The above change means that if the device supports discard it can't be
guaranteed that a discard request zeroes data. Therefore, we set
ti->discard_zeroes_data_unsupported.
Thin target discard support with this bug arrived in commit
104655fd4d (dm thin: support discards).
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Userland sometimes sees a corrupt metadata block if metadata is changing
rapidly when a metadata snapshot is reserved for userland, To make the
problem go away, commit before we take the metadata snapshot (which is a
sensible thing to do anyway).
The checksums mean userland spots this corruption immediately so there's
no risk of acting on incorrect data. No corruption exists from the
kernel's point of view, and thin_check passes after pool shutdown.
I believe this is to do with shared blocks at the first level of the
{device, mapping} btree. Prior to the metadata-snap support no sharing
at this level was possible, so this patch is only required after commit
cc8394d86f ("dm thin: provide userspace
access to pool metadata").
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
This patch implements two new messages that can be sent to the thin
pool target allowing it to take a snapshot of the _metadata_. This,
read-only snapshot can be accessed by userland, concurrently with the
live target.
Only one metadata snapshot can be held at a time. The pool's status
line will give the block location for the current msnap.
Since version 0.1.5 of the userland thin provisioning tools, the
thin_dump program displays the msnap as follows:
thin_dump -m <msnap root> <metadata dev>
Available here: https://github.com/jthornber/thin-provisioning-tools
Now that userland can access the metadata we can do various things
that have traditionally been kernel side tasks:
i) Incremental backups.
By using metadata snapshots we can work out what blocks have
changed over time. Combined with data snapshots we can ensure
the data doesn't change while we back it up.
A short proof of concept script can be found here:
https://github.com/jthornber/thinp-test-suite/blob/master/incremental_backup_example.rb
ii) Migration of thin devices from one pool to another.
iii) Merging snapshots back into an external origin.
iv) Asyncronous replication.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Use dedicated caches prefixed with a "dm_" name rather than relying on
kmalloc mempools backed by generic slab caches so the memory usage of
thin provisioning (and any leaks) can be accounted for independently.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
When the thin pool target clears the discard_passdown parameter
internally, it incorrectly changes the table line reported to userspace.
This breaks dumb string comparisons on these table lines in generic
userspace device-mapper library code and leads to tables being reloaded
repeatedly when nothing is actually meant to be changing.
This patch corrects this by no longer changing the table line when
discard passdown was disabled.
We can still tell when discard passdown is overridden by looking for the
message "Discard unsupported by data device (sdX): Disabling discard passdown."
This automatic detection is also moved from the 'load' to the 'resume'
so that it is re-evaluated should the properties of underlying devices
change.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix two places in commit 104655fd4d ("dm thin: support discards") that
didn't use pool->lock to protect against concurrent changes to the
prepared_discards list.
Without this fix, thin_endio() can race with process_discard(), leading
to concurrent list_add()s that result in the processes locking up with
an error like the following:
WARNING: at lib/list_debug.c:32 __list_add+0x8f/0xa0()
...
list_add corruption. next->prev should be prev (ffff880323b96140), but was ffff8801d2c48440. (next=ffff8801d2c485c0).
...
Pid: 17205, comm: kworker/u:1 Tainted: G W O 3.4.0-rc3.snitm+ #1
Call Trace:
[<ffffffff8103ca1f>] warn_slowpath_common+0x7f/0xc0
[<ffffffff8103cb16>] warn_slowpath_fmt+0x46/0x50
[<ffffffffa04f6ce6>] ? bio_detain+0xc6/0x210 [dm_thin_pool]
[<ffffffff8124ff3f>] __list_add+0x8f/0xa0
[<ffffffffa04f70d2>] process_discard+0x2a2/0x2d0 [dm_thin_pool]
[<ffffffffa04f6a78>] ? remap_and_issue+0x38/0x50 [dm_thin_pool]
[<ffffffffa04f7c3b>] process_deferred_bios+0x7b/0x230 [dm_thin_pool]
[<ffffffffa04f7df0>] ? process_deferred_bios+0x230/0x230 [dm_thin_pool]
[<ffffffffa04f7e42>] do_worker+0x52/0x60 [dm_thin_pool]
[<ffffffff81056fa9>] process_one_work+0x129/0x450
[<ffffffff81059b9c>] worker_thread+0x17c/0x3c0
[<ffffffff81059a20>] ? manage_workers+0x120/0x120
[<ffffffff8105eabe>] kthread+0x9e/0xb0
[<ffffffff814ceda4>] kernel_thread_helper+0x4/0x10
[<ffffffff8105ea20>] ? kthread_freezable_should_stop+0x70/0x70
[<ffffffff814ceda0>] ? gs_change+0x13/0x13
---[ end trace 7e0a523bc5e52692 ]---
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix a significant memory leak inadvertently introduced during
simplification of cell_release_singleton() in commit
6f94a4c45a ("dm thin: fix stacked bi_next
usage").
A cell's hlist_del() must be accompanied by a mempool_free().
Use __cell_release() to do this, like before.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Add dm thin target arguments to control discard support.
ignore_discard: Disables discard support
no_discard_passdown: Don't pass discards down to the underlying data
device, but just remove the mapping within the thin provisioning target.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Support discards in the thin target.
On discard the corresponding mapping(s) are removed from the thin
device. If the associated block(s) are no longer shared the discard
is passed to the underlying device.
All bios other than discards now have an associated deferred_entry
that is saved to the 'all_io_entry' in endio_hook. When non-discard
IO completes and associated mappings are quiesced any discards that
were deferred, via ds_add_work() in process_discard(), will be queued
for processing by the worker thread.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
drivers/md/dm-thin.c | 173 ++++++++++++++++++++++++++++++++++++++++++++++----
drivers/md/dm-thin.c | 172 ++++++++++++++++++++++++++++++++++++++++++++++-----
1 file changed, 158 insertions(+), 14 deletions(-)
This patch contains the ground work needed for dm-thin to support discard.
- Adds endio function that replaces shared_read_endio.
- Introduce an explicit 'quiesced' flag into the new_mapping structure.
Before, this was implicitly indicated by m->list being empty.
- The map_info->ptr remains constant for the duration of a bio's trip
through the thin target. Make it easier to reason about it.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Support the use of an external _read only_ device as an origin for a thin
device.
Any read to an unprovisioned area of the thin device will be passed
through to the origin. Writes trigger allocation of new blocks as
usual.
One possible use case for this would be VM hosts that want to run
guests on thinly-provisioned volumes but have the base image on another
device (possibly shared between many VMs).
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
The thin metadata format can only make use of a device that is <=
THIN_METADATA_MAX_SECTORS (currently 15.9375 GB). Therefore, there is no
practical benefit to using a larger device.
However, it may be that other factors impose a certain granularity for
the space that is allocated to a device (E.g. lvm2 can impose a coarse
granularity through the use of large, >= 1 GB, physical extents).
Rather than reject a larger metadata device, during thin-pool device
construction, switch to allowing it but issue a warning if a device
larger than THIN_METADATA_MAX_SECTORS_WARNING (16 GB) is
provided. Any space over 15.9375 GB will not be used.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Commit unwritten data every second to prevent too much building up.
Released blocks don't become available until after the next commit
(for crash resilience). Prior to this patch commits were only
triggered by a message to the target or a REQ_{FLUSH,FUA} bio. This
allowed far too big a position to build up.
The interval is hard-coded to 1 second. This is a sensible setting.
I'm not making this user configurable, since there isn't much to be
gained by tweaking this - and a lot lost by setting it far too high.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Remove documentation for unimplemented 'trim' message.
I'd planned a 'trim' target message for shrinking thin devices, but
this is better handled via the discard ioctl.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Avoid using the bi_next field for the holder of a cell when deferring
bios because a stacked device below might change it. Store the
holder in a new field in struct cell instead.
When a cell is created, the bio that triggered creation (the holder) was
added to the same bio list as subsequent bios. In some cases we pass
this holder bio directly to devices underneath. If those devices use
the bi_next field there will be trouble...
This also simplifies some code that had to work out which bio was the
holder.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Initial EXPERIMENTAL implementation of device-mapper thin provisioning
with snapshot support. The 'thin' target is used to create instances of
the virtual devices that are hosted in the 'thin-pool' target. The
thin-pool target provides data sharing among devices. This sharing is
made possible using the persistent-data library in the previous patch.
The main highlight of this implementation, compared to the previous
implementation of snapshots, is that it allows many virtual devices to
be stored on the same data volume, simplifying administration and
allowing sharing of data between volumes (thus reducing disk usage).
Another big feature is support for arbitrary depth of recursive
snapshots (snapshots of snapshots of snapshots ...). The previous
implementation of snapshots did this by chaining together lookup tables,
and so performance was O(depth). This new implementation uses a single
data structure so we don't get this degradation with depth.
For further information and examples of how to use this, please read
Documentation/device-mapper/thin-provisioning.txt
Signed-off-by: Joe Thornber <thornber@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Joe Thornber