Currently we don't honour merge_bvec_fn in member devices so if there
is one, we force all requests to be single-page at most.
This is not ideal.
So enhance the raid10 merge_bvec_fn to check that function in children
as well.
This introduces a small problem. There is no locking around calls
the ->merge_bvec_fn and subsequent calls to ->make_request. So a
device added between these could end up getting a request which
violates its merge_bvec_fn.
Currently the best we can do is synchronize_sched(). This will work
providing no preemption happens. If there is preemption, we just
have to hope that new devices are largely consistent with old devices.
Signed-off-by: NeilBrown <neilb@suse.de>
md.h has an 'rdev_for_each()' macro for iterating the rdevs in an
mddev. However it uses the 'safe' version of list_for_each_entry,
and so requires the extra variable, but doesn't include 'safe' in the
name, which is useful documentation.
Consequently some places use this safe version without needing it, and
many use an explicity list_for_each entry.
So:
- rename rdev_for_each to rdev_for_each_safe
- create a new rdev_for_each which uses the plain
list_for_each_entry,
- use the 'safe' version only where needed, and convert all other
list_for_each_entry calls to use rdev_for_each.
Signed-off-by: NeilBrown <neilb@suse.de>
hot-replace is a feature being added to md which will allow a
device to be replaced without removing it from the array first.
With hot-replace a spare can be activated and recovery can start while
the original device is still in place, thus allowing a transition from
an unreliable device to a reliable device without leaving the array
degraded during the transition. It can also be use when the original
device is still reliable but it not wanted for some reason.
This will eventually be supported in RAID4/5/6 and RAID10.
This patch adds a super-block flag to distinguish the replacement
device. If an old kernel sees this flag it will reject the device.
It also adds two per-device flags which are viewable and settable via
sysfs.
"want_replacement" can be set to request that a device be replaced.
"replacement" is set to show that this device is replacing another
device.
The "rd%d" links in /sys/block/mdXx/md only apply to the original
device, not the replacement. We currently don't make links for the
replacement - there doesn't seem to be a need.
Signed-off-by: NeilBrown <neilb@suse.de>
Soon an array will be able to have multiple devices with the
same raid_disk number (an original and a replacement). So removing
a device based on the number won't work. So pass the actual device
handle instead.
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
* 'for-3.2/core' of git://git.kernel.dk/linux-block: (29 commits)
block: don't call blk_drain_queue() if elevator is not up
blk-throttle: use queue_is_locked() instead of lockdep_is_held()
blk-throttle: Take blkcg->lock while traversing blkcg->policy_list
blk-throttle: Free up policy node associated with deleted rule
block: warn if tag is greater than real_max_depth.
block: make gendisk hold a reference to its queue
blk-flush: move the queue kick into
blk-flush: fix invalid BUG_ON in blk_insert_flush
block: Remove the control of complete cpu from bio.
block: fix a typo in the blk-cgroup.h file
block: initialize the bounce pool if high memory may be added later
block: fix request_queue lifetime handling by making blk_queue_cleanup() properly shutdown
block: drop @tsk from attempt_plug_merge() and explain sync rules
block: make get_request[_wait]() fail if queue is dead
block: reorganize throtl_get_tg() and blk_throtl_bio()
block: reorganize queue draining
block: drop unnecessary blk_get/put_queue() in scsi_cmd_ioctl() and blk_get_tg()
block: pass around REQ_* flags instead of broken down booleans during request alloc/free
block: move blk_throtl prototypes to block/blk.h
block: fix genhd refcounting in blkio_policy_parse_and_set()
...
Fix up trivial conflicts due to "mddev_t" -> "struct mddev" conversion
and making the request functions be of type "void" instead of "int" in
- drivers/md/{faulty.c,linear.c,md.c,md.h,multipath.c,raid0.c,raid1.c,raid10.c,raid5.c}
- drivers/staging/zram/zram_drv.c
The typedefs are just annoying. 'mdk' probably refers to 'md_k.h'
which used to be an include file that defined this thing.
Signed-off-by: NeilBrown <neilb@suse.de>
Two related problems:
1/ some error paths call "md_unregister_thread(mddev->thread)"
without subsequently clearing ->thread. A subsequent call
to mddev_unlock will try to wake the thread, and crash.
2/ Most calls to md_wakeup_thread are protected against the thread
disappeared either by:
- holding the ->mutex
- having an active request, so something else must be keeping
the array active.
However mddev_unlock calls md_wakeup_thread after dropping the
mutex and without any certainty of an active request, so the
->thread could theoretically disappear.
So we need a spinlock to provide some protections.
So change md_unregister_thread to take a pointer to the thread
pointer, and ensure that it always does the required locking, and
clears the pointer properly.
Reported-by: "Moshe Melnikov" <moshe@zadarastorage.com>
Signed-off-by: NeilBrown <neilb@suse.de>
cc: stable@kernel.org
There is very little benefit in allowing to let a ->make_request
instance update the bios device and sector and loop around it in
__generic_make_request when we can archive the same through calling
generic_make_request from the driver and letting the loop in
generic_make_request handle it.
Note that various drivers got the return value from ->make_request and
returned non-zero values for errors.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: NeilBrown <neilb@suse.de>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
It is only safe to choose not to write to a bad block if that bad
block is safely recorded in metadata - i.e. if it has been
'acknowledged'.
If it hasn't we need to wait for the acknowledgement.
We support that using rdev->blocked wait and
md_wait_for_blocked_rdev by introducing a new device flag
'BlockedBadBlock'.
This flag is only advisory.
It is cleared whenever we acknowledge a bad block, so that a waiter
can re-check the particular bad blocks that it is interested it.
It should be set by a caller when they find they need to wait.
This (set after test) is inherently racy, but as
md_wait_for_blocked_rdev already has a timeout, losing the race will
have minimal impact.
When we clear "Blocked" was also clear "BlockedBadBlocks" incase it
was set incorrectly (see above race).
We also modify the way we manage 'Blocked' to fit better with the new
handling of 'BlockedBadBlocks' and to make it consistent between
externally managed and internally managed metadata. This requires
that each raidXd loop checks if the metadata needs to be written and
triggers a write (md_check_recovery) if needed. Otherwise a queued
write request might cause raidXd to wait for the metadata to write,
and only that thread can write it.
Before writing metadata, we set FaultRecorded for all devices that
are Faulty, then after writing the metadata we clear Blocked for any
device for which the Fault was certainly Recorded.
The 'faulty' device flag now appears in sysfs if the device is faulty
*or* it has unacknowledged bad blocks. So user-space which does not
understand bad blocks can continue to function correctly.
User space which does, should not assume a device is faulty until it
sees the 'faulty' flag, and then sees the list of unacknowledged bad
blocks is empty.
Signed-off-by: NeilBrown <neilb@suse.de>
If a device has ever seen a write error, we will want to handle
known-bad-blocks differently.
So create an appropriate state flag and export it via sysfs.
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
Now that we have a bad block list, we should not read from those
blocks.
There are several main parts to this:
1/ read_balance needs to check for bad blocks, and return not only
the chosen device, but also how many good blocks are available
there.
2/ fix_read_error needs to avoid trying to read from bad blocks.
3/ read submission must be ready to issue multiple reads to
different devices as different bad blocks on different devices
could mean that a single large read cannot be served by any one
device, but can still be served by the array.
This requires keeping count of the number of outstanding requests
per bio. This count is stored in 'bi_phys_segments'
4/ retrying a read needs to also be ready to submit a smaller read
and queue another request for the rest.
This does not yet handle bad blocks when reading to perform resync,
recovery, or check.
'md_trim_bio' will also be used for RAID10, so put it in md.c and
export it.
Signed-off-by: NeilBrown <neilb@suse.de>
Space must have been allocated when array was created.
A feature flag is set when the badblock list is non-empty, to
ensure old kernels don't load and trust the whole device.
We only update the on-disk badblocklist when it has changed.
If the badblocklist (or other metadata) is stored on a bad block, we
don't cope very well.
If metadata has no room for bad block, flag bad-blocks as disabled,
and do the same for 0.90 metadata.
Signed-off-by: NeilBrown <neilb@suse.de>
This the first step in allowing md to track bad-blocks per-device so
that we can fail individual blocks rather than the whole device.
This patch just adds a data structure for recording bad blocks, with
routines to add, remove, search the list.
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
Revert most of commit e384e58549
md/bitmap: prepare for storing write-intent-bitmap via dm-dirty-log.
MD should not need to use DM's dirty log - we decided to use md's
bitmaps instead.
Keeping the DIV_ROUND_UP clean-ups that were part of commit
e384e58549, however.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
If we hit a read error while recovering a mirror, we want to abort the
recovery without necessarily failing the disk - as having a disk this
a read error is better than not having an array at all.
Currently this is managed with a per-array flag "recovery_disabled"
and is only implemented for RAID1. For RAID10 we will need finer
grained control as we might want to disable recovery for individual
devices separately.
So push more of the decision making into the personality.
'recovery_disabled' is now a 'cookie' which is copied when the
personality want to disable recovery and is changed when a device is
added to the array as this is used as a trigger to 'try recovery
again'.
This will allow RAID10 to get the control that it needs.
Signed-off-by: NeilBrown <neilb@suse.de>
There are places where sysfs links to rdev are handled
in a same way. Add the helper functions to consolidate
them.
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Add bitmap support to the device-mapper specific metadata area.
This patch allows the creation of the bitmap metadata area upon
initial array creation via device-mapper.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Add the 'sync_super' function pointer to MD array structure (struct mddev_s)
If device-mapper (dm-raid.c) is to define its own on-disk superblock and be
able to load it, there must still be a way for MD to initiate superblock
updates. The simplest way to make this happen is to provide a pointer in
the MD array structure that can be set by device-mapper (or other module)
with a function to do this. If the function has been set, it will be used;
otherwise, the method with be looked up via 'super_types' as usual.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
When an md device adds a request to a queue, it can call
mddev_check_plugged.
If this succeeds then we know that the md thread will be woken up
shortly, and ->plug_cnt will be non-zero until then, so some
processing can be delayed.
If it fails, then no unplug callback is expected and the make_request
function needs to do whatever is required to make the request happen.
Signed-off-by: NeilBrown <neilb@suse.de>
md has some plugging infrastructure for RAID5 to use because the
normal plugging infrastructure required a 'request_queue', and when
called from dm, RAID5 doesn't have one of those available.
This relied on the ->unplug_fn callback which doesn't exist any more.
So remove all of that code, both in md and raid5. Subsequent patches
with restore the plugging functionality.
Signed-off-by: NeilBrown <neilb@suse.de>
Revert
b821eaa572
and
f3b99be19d
When I wrote the first of these I had a wrong idea about the
lifetime of 'struct block_device'. It can disappear at any time that
the block device is not open if it falls out of the inode cache.
So relying on the 'size' recorded with it to detect when the
device size has changed and so we need to revalidate, is wrong.
Rather, we really do need the 'changed' attribute stored directly in
the mddev and set/tested as appropriate.
Without this patch, a sequence of:
mknod / open / close / unlink
(which can cause a block_device to be created and then destroyed)
will result in a rescan of the partition table and consequence removal
and addition of partitions.
Several of these in a row can get udev racing to create and unlink and
other code can get confused.
With the patch, the rescan is only performed when needed and so there
are no races.
This is suitable for any stable kernel from 2.6.35.
Reported-by: "Wojcik, Krzysztof" <krzysztof.wojcik@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Cc: stable@kernel.org
This flag is not needed and is used badly.
Devices that are included in a native-metadata array are reserved
exclusively for that array - and currently have AllReserved set.
They all are bd_claimed for the rdev and so cannot be shared.
Devices that are included in external-metadata arrays can be shared
among multiple arrays - providing there is no overlap.
These are bd_claimed for md in general - not for a particular rdev.
When changing the amount of a device that is used in an array we need
to check for overlap. This currently includes a check on AllReserved
So even without overlap, sharing with an AllReserved device is not
allowed.
However the bd_claim usage already precludes sharing with these
devices, so the test on AllReserved is not needed. And in fact it is
wrong.
As this is the only use of AllReserved, simply remove all usage and
definition of AllReserved.
Signed-off-by: NeilBrown <neilb@suse.de>
Allow the metadata to be on a separate device from the
data.
This doesn't mean the data and metadata will by on separate
physical devices - it simply gives device-mapper and userspace
tools more flexibility.
Signed-off-by: NeilBrown <neilb@suse.de>
Add new parameter to 'sync_page_io'.
The new parameter allows us to distinguish between metadata and data
operations. This becomes important later when we add the ability to
use separate devices for data and metadata.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
When an md device is in the process of coming on line it is possible
for an IO request (typically a partition table probe) to get through
before the array is fully initialised, which can cause unexpected
behaviour (e.g. a crash).
So explicitly record when the array is ready for IO and don't allow IO
through until then.
There is no possibility for a similar problem when the array is going
off-line as there must only be one 'open' at that time, and it is busy
off-lining the array and so cannot send IO requests. So no memory
barrier is needed in md_stop()
This has been a bug since commit 409c57f380 in 2.6.30 which
introduced md_make_request. Before then, each personality would
register its own make_request_fn when it was ready.
This is suitable for any stable kernel from 2.6.30.y onwards.
Cc: <stable@kernel.org>
Signed-off-by: NeilBrown <neilb@suse.de>
Reported-by: "Hawrylewicz Czarnowski, Przemyslaw" <przemyslaw.hawrylewicz.czarnowski@intel.com>
bio_clone and bio_alloc allocate from a common bio pool.
If an md device is stacked with other devices that use this pool, or under
something like swap which uses the pool, then the multiple calls on
the pool can cause deadlocks.
So allocate a local bio pool for each md array and use that rather
than the common pool.
This pool is used both for regular IO and metadata updates.
Signed-off-by: NeilBrown <neilb@suse.de>
Currently sync_page_io takes a 'bdev'.
Every caller passes 'rdev->bdev'.
We will soon want another field out of the rdev in sync_page_io,
So just pass the rdev instead of the bdev out of it.
Signed-off-by: NeilBrown <neilb@suse.de>
This patch converts md to support REQ_FLUSH/FUA instead of now
deprecated REQ_HARDBARRIER. In the core part (md.c), the following
changes are notable.
* Unlike REQ_HARDBARRIER, REQ_FLUSH/FUA don't interfere with
processing of other requests and thus there is no reason to mark the
queue congested while FLUSH/FUA is in progress.
* REQ_FLUSH/FUA failures are final and its users don't need retry
logic. Retry logic is removed.
* Preflush needs to be issued to all member devices but FUA writes can
be handled the same way as other writes - their processing can be
deferred to request_queue of member devices. md_barrier_request()
is renamed to md_flush_request() and simplified accordingly.
For linear, raid0 and multipath, the core changes are enough. raid1,
5 and 10 need the following conversions.
* raid1: Handling of FLUSH/FUA bio's can simply be deferred to
request_queues of member devices. Barrier related logic removed.
* raid5: Queue draining logic dropped. FUA bit is propagated through
biodrain and stripe resconstruction such that all the updated parts
of the stripe are written out with FUA writes if any of the dirtying
writes was FUA. preread_active_stripes handling in make_request()
is updated as suggested by Neil Brown.
* raid10: FUA bit needs to be propagated to write clones.
linear, raid0, 1, 5 and 10 tested.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
MD_CHANGE_CLEAN is used for two different purposes and this leads to
confusion.
One of the purposes is largely mirrored by MD_CHANGE_PENDING which is
not used for anything else, so have MD_CHANGE_PENDING take over that
purpose fully.
The two purposes are:
1/ tell md_update_sb that an update is needed and that it is just a
clean/dirty transition.
2/ tell user-space that an transition from clean to dirty is pending
(something wants to write), and tell te kernel (by clearin the
flag) that the transition is OK.
The first purpose remains wit MD_CHANGE_CLEAN, the second is moved
fully to MD_CHANGE_PENDING.
This means that various places which conditionally set or cleared
MD_CHANGE_CLEAN no longer need to be conditional.
Signed-off-by: NeilBrown <neilb@suse.de>
* 'for-linus' of git://neil.brown.name/md: (24 commits)
md: clean up do_md_stop
md: fix another deadlock with removing sysfs attributes.
md: move revalidate_disk() back outside open_mutex
md/raid10: fix deadlock with unaligned read during resync
md/bitmap: separate out loading a bitmap from initialising the structures.
md/bitmap: prepare for storing write-intent-bitmap via dm-dirty-log.
md/bitmap: optimise scanning of empty bitmaps.
md/bitmap: clean up plugging calls.
md/bitmap: reduce dependence on sysfs.
md/bitmap: white space clean up and similar.
md/raid5: export raid5 unplugging interface.
md/plug: optionally use plugger to unplug an array during resync/recovery.
md/raid5: add simple plugging infrastructure.
md/raid5: export is_congested test
raid5: Don't set read-ahead when there is no queue
md: add support for raising dm events.
md: export various start/stop interfaces
md: split out md_rdev_init
md: be more careful setting MD_CHANGE_CLEAN
md/raid5: ensure we create a unique name for kmem_cache when mddev has no gendisk
...
Move the deletion of sysfs attributes from reconfig_mutex to
open_mutex didn't really help as a process can try to take
open_mutex while holding reconfig_mutex, so the same deadlock can
happen, just requiring one more process to be involved in the chain.
I looks like I cannot easily use locking to wait for the sysfs
deletion to complete, so don't.
The only things that we cannot do while the deletions are still
pending is other things which can change the sysfs namespace: run,
takeover, stop. Each of these can fail with -EBUSY.
So set a flag while doing a sysfs deletion, and fail run, takeover,
stop if that flag is set.
This is suitable for 2.6.35.x
Cc: stable@kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
Remove the current bio flags and reuse the request flags for the bio, too.
This allows to more easily trace the type of I/O from the filesystem
down to the block driver. There were two flags in the bio that were
missing in the requests: BIO_RW_UNPLUG and BIO_RW_AHEAD. Also I've
renamed two request flags that had a superflous RW in them.
Note that the flags are in bio.h despite having the REQ_ name - as
blkdev.h includes bio.h that is the only way to go for now.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
This allows md/raid5 to fully work as a dm target.
Normally md uses a 'filemap' which contains a list of pages of bits
each of which may be written separately.
dm-log uses and all-or-nothing approach to writing the log, so
when using a dm-log, ->filemap is NULL and the flags normally stored
in filemap_attr are stored in ->logattrs instead.
Signed-off-by: NeilBrown <neilb@suse.de>
1/ use md_unplug in bitmap.c as we will soon be using bitmaps under
arrays with no queue attached.
2/ Don't bother plugging the queue when we set a bit in the bitmap.
The reason for this was to encourage as many bits as possible to
get set before we unplug and write stuff out.
However every personality already plugs the queue after
bitmap_startwrite either directly (raid1/raid10) or be setting
STRIPE_BIT_DELAY which causes the queue to be plugged later
(raid5).
Signed-off-by: NeilBrown <neilb@suse.de>
Fixes some whitespace problems
Fixed some checkpatch.pl complaints.
Replaced kmalloc ... memset(0), with kzalloc
Fixed an unlikely memory leak on an error path.
Reformatted a number of 'if/else' sets, sometimes
replacing goto with an else clause.
Removed some old comments and commented-out code.
Signed-off-by: NeilBrown <neilb@suse.de>
If an array doesn't have a 'queue' then md_do_sync cannot
unplug it.
In that case it will have a 'plugger', so make that available
to the mddev, and use it to unplug the array if needed.
Signed-off-by: NeilBrown <neilb@suse.de>
md/raid5 uses the plugging infrastructure provided by the block layer
and 'struct request_queue'. However when we plug raid5 under dm there
is no request queue so we cannot use that.
So create a similar infrastructure that is much lighter weight and use
it for raid5.
Signed-off-by: NeilBrown <neilb@suse.de>
dm uses scheduled work to raise events to user-space.
So allow md device to have work_structs and schedule them on an error.
Signed-off-by: NeilBrown <neilb@suse.de>
export entry points for starting and stopping md arrays.
This will be used by a module to make md/raid5 work under
dm.
Also stop calling md_stop_writes from md_stop, as that won't
work well with dm - it will want to call the two separately.
Signed-off-by: NeilBrown <neilb@suse.de>
This functionality will be needed separately in a subsequent patch, so
split it into it's own exported function.
Signed-off-by: NeilBrown <neilb@suse.de>
We will want md devices to live as dm targets where sysfs is not
visible. So allow md to not connect to sysfs.
Signed-off-by: NeilBrown <neilb@suse.de>
Most array level changes leave the list of devices largely unchanged,
possibly causing one at the end to become redundant.
However conversions between RAID0 and RAID10 need to renumber
all devices (except 0).
This renumbering is currently being done in the ->run method when the
new personality takes over. However this is too late as the common
code in md.c might already have invalidated some of the devices if
they had a ->raid_disk number that appeared to high.
Moving it into the ->takeover method is too early as the array is
still active at that time and wrong ->raid_disk numbers could cause
confusion.
So add a ->new_raid_disk field to mdk_rdev_s and use it to communicate
the new raid_disk number.
Now the common code knows exactly which devices need to be renumbered,
and which can be invalidated, and can do it all at a convenient time
when the array is suspend.
It can also update some symlinks in sysfs which previously were not be
updated correctly.
Reported-by: Maciej Trela <maciej.trela@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
When updating the event count for a simple clean <-> dirty transition,
we try to avoid updating the spares so they can safely spin-down.
As the event_counts across an array must be +/- 1, this means
decrementing the event_count on a dirty->clean transition.
This is not always safe and we have to avoid the unsafe time.
We current do this with a misguided idea about it being safe or
not depending on whether the event_count is odd or even. This
approach only works reliably in a few common instances, but easily
falls down.
So instead, simply keep internal state concerning whether it is safe
or not, and always assume it is not safe when an array is first
assembled.
Signed-off-by: NeilBrown <neilb@suse.de>
We used to pass the personality make_request function direct
to the block layer so the first argument had to be a queue.
But now we have the intermediary md_make_request so it makes
at lot more sense to pass a struct mddev_s.
It makes it possible to have an mddev without its own queue too.
Signed-off-by: NeilBrown <neilb@suse.de>
We set ->changed to 1 and call check_disk_change at the end
of md_open so that bd_invalidated would be set and thus
partition rescan would happen appropriately.
Now that we call revalidate_disk directly, which sets bd_invalidates,
that indirection is no longer needed and can be removed.
Signed-off-by: NeilBrown <neilb@suse.de>
This was needed when sysfs files could only be 'notified'
from process context. Now that we have sys_notify_direct,
we can call it directly from an interrupt.
Signed-off-by: NeilBrown <neilb@suse.de>
These fields have never been used.
commit 4b6d287f62
added them, but also added identical files to bitmap_super_s,
and only used the latter.
So remove these unused fields.
Signed-off-by: NeilBrown <neilb@suse.de>
Some levels expect the 'redundancy group' to be present,
others don't.
So when we change level of an array we might need to
add or remove this group.
This requires fixing up the current practice of overloading ->private
to indicate (when ->pers == NULL) that something needs to be removed.
So create a new ->to_remove to fill that role.
When changing levels, we may need to add or remove attributes. When
changing RAID5 -> RAID6, we both add and remove the same thing. It is
important to catch this and optimise it out as the removal is delayed
until a lock is released, so trying to add immediately would cause
problems.
Cc: stable@kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
We've noticed severe lasting performance degradation of our raid
arrays when we have drives that yield large amounts of media errors.
The raid10 module will queue each failed read for retry, and also
will attempt call fix_read_error() to perform the read recovery.
Read recovery is performed while the array is frozen, so repeated
recovery attempts can degrade the performance of the array for
extended periods of time.
With this patch I propose adding a per md device max number of
corrected read attempts. Each rdev will maintain a count of
read correction attempts in the rdev->read_errors field (not
used currently for raid10). When we enter fix_read_error()
we'll check to see when the last read error occurred, and
divide the read error count by 2 for every hour since the
last read error. If at that point our read error count
exceeds the read error threshold, we'll fail the raid device.
In addition in this patch I add sysfs nodes (get/set) for
the per md max_read_errors attribute, the rdev->read_errors
attribute, and added some printk's to indicate when
fix_read_error fails to repair an rdev.
For testing I used debugfs->fail_make_request to inject
IO errors to the rdev while doing IO to the raid array.
Signed-off-by: Robert Becker <Rob.Becker@riverbed.com>
Signed-off-by: NeilBrown <neilb@suse.de>
In this case, the metadata needs to not be in the same
sector as the bitmap.
md will not read/write any bitmap metadata. Config must be
done via sysfs and when a recovery makes the array non-degraded
again, writing 'true' to 'bitmap/can_clear' will allow bits in
the bitmap to be cleared again.
Signed-off-by: NeilBrown <neilb@suse.de>
A new attribute directory 'bitmap' in 'md' is created which
contains files for configuring the bitmap.
'location' identifies where the bitmap is, either 'none',
or 'file' or 'sector offset from metadata'.
Writing 'location' can create or remove a bitmap.
Adding a 'file' bitmap this way is not yet supported.
'chunksize' and 'time_base' must be set before 'location'
can be set.
'chunksize' can be set before creating a bitmap, but is
currently always over-ridden by the bitmap superblock.
'time_base' and 'backlog' can be updated at any time.
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Andre Noll <maan@systemlinux.org>
safe_delay_store can parse fixed point numbers (for fractions
of a second). We will want to do that for another sysfs
file soon, so factor out the code.
Signed-off-by: NeilBrown <neilb@suse.de>
For md arrays were metadata is managed externally, the kernel does not
know about a superblock so the superblock offset is 0.
If we want to have a write-intent-bitmap near the end of the
devices of such an array, we should support sector_t sized offset.
We need offset be possibly negative for when the bitmap is before
the metadata, so use loff_t instead.
Also add sanity check that bitmap does not overlap with data.
Signed-off-by: NeilBrown <neilb@suse.de>
... and into bitmap_info. These are all configuration parameters
that need to be set before the bitmap is created.
Signed-off-by: NeilBrown <neilb@suse.de>
In preparation for making bitmap fields configurable via sysfs,
start tidying up by making a single structure to contain the
configuration fields.
Signed-off-by: NeilBrown <neilb@suse.de>
Previously barriers were only supported on RAID1. This is because
other levels requires synchronisation across all devices and so needed
a different approach.
Here is that approach.
When a barrier arrives, we send a zero-length barrier to every active
device. When that completes - and if the original request was not
empty - we submit the barrier request itself (with the barrier flag
cleared) and then submit a fresh load of zero length barriers.
The barrier request itself is asynchronous, but any subsequent
request will block until the barrier completes.
The reason for clearing the barrier flag is that a barrier request is
allowed to fail. If we pass a non-empty barrier through a striping
raid level it is conceivable that part of it could succeed and part
could fail. That would be way too hard to deal with.
So if the first run of zero length barriers succeed, we assume all is
sufficiently well that we send the request and ignore errors in the
second run of barriers.
RAID5 needs extra care as write requests may not have been submitted
to the underlying devices yet. So we flush the stripe cache before
proceeding with the barrier.
Note that the second set of zero-length barriers are submitted
immediately after the original request is submitted. Thus when
a personality finds mddev->barrier to be set during make_request,
it should not return from make_request until the corresponding
per-device request(s) have been queued.
That will be done in later patches.
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Andre Noll <maan@systemlinux.org>
A write intent bitmap can be removed from an array while the
array is active.
When this happens, all IO is suspended and flushed before the
bitmap is removed.
However it is possible that bitmap_daemon_work is still running to
clear old bits from the bitmap. If it is, it can dereference the
bitmap after it has been freed.
So introduce a new mutex to protect bitmap_daemon_work and get it
before destroying a bitmap.
This is suitable for any current -stable kernel.
Signed-off-by: NeilBrown <neilb@suse.de>
Cc: stable@kernel.org
A recent commit:
commit 449aad3e25
introduced the possibility of an A-B/B-A deadlock between
bd_mutex and reconfig_mutex.
__blkdev_get holds bd_mutex while calling md_open which takes
reconfig_mutex,
do_md_run is always called with reconfig_mutex held, and it now
takes bd_mutex in the call the revalidate_disk.
This potential deadlock was not caught by lockdep due to the
use of mutex_lock_interruptible_nexted which was introduced
by
commit d63a5a74de
do avoid a warning of an impossible deadlock.
It is quite possible to split reconfig_mutex in to two locks.
One protects the array data structures while it is being
reconfigured, the other ensures that an array is never even partially
open while it is being deactivated.
In particular, the second lock prevents an open from completing
between the time when do_md_stop checks if there are any active opens,
and the time when the array is either set read-only, or when ->pers is
set to NULL. So we can be certain that no IO is in flight as the
array is being destroyed.
So create a new lock, open_mutex, just to ensure exclusion between
'open' and 'stop'.
This avoids the deadlock and also avoids the lockdep warning mentioned
in commit d63a5a74d
Reported-by: "Mike Snitzer" <snitzer@gmail.com>
Reported-by: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: NeilBrown <neilb@suse.de>
This patch replaces md_integrity_check() by two new public functions:
md_integrity_register() and md_integrity_add_rdev() which are both
personality-independent.
md_integrity_register() is called from the ->run and ->hot_remove
methods of all personalities that support data integrity. The
function iterates over the component devices of the array and
determines if all active devices are integrity capable and if their
profiles match. If this is the case, the common profile is registered
for the mddev via blk_integrity_register().
The second new function, md_integrity_add_rdev() is called from the
->hot_add_disk methods, i.e. whenever a new device is being added
to a raid array. If the new device does not support data integrity,
or has a profile different from the one already registered, data
integrity for the mddev is disabled.
For raid0 and linear, only the call to md_integrity_register() from
the ->run method is necessary.
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
If the superblock of a component device indicates the presence of a
bitmap but the corresponding raid personality does not support bitmaps
(raid0, linear, multipath, faulty), then something is seriously wrong
and we'd better refuse to run such an array.
Currently, this check is performed while the superblocks are examined,
i.e. before entering personality code. Therefore the generic md layer
must know which raid levels support bitmaps and which do not.
This patch avoids this layer violation without adding identical code
to various personalities. This is accomplished by introducing a new
public function to md.c, md_check_no_bitmap(), which replaces the
hard-coded checks in the superblock loading functions.
A call to md_check_no_bitmap() is added to the ->run method of each
personality which does not support bitmaps and assembly is aborted
if at least one component device contains a bitmap.
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
It is easiest to round sizes to multiples of chunk size in
the personality code for those personalities which care.
Those personalities now do the rounding, so we can
remove that function from common code.
Also remove the upper bound on the size of a chunk, and the lower
bound on the size of a device (1 chunk), neither of which really buy
us anything.
Signed-off-by: NeilBrown <neilb@suse.de>
The difference between these two methods is artificial.
Both check that a pending reshape is valid, and perform any
aspect of it that can be done immediately.
'reconfig' handles chunk size and layout.
'check_reshape' handles raid_disks.
So make them just one method.
Signed-off-by: NeilBrown <neilb@suse.de>
Passing the new layout and chunksize as args is not necessary as
the mddev has fields for new_check and new_layout.
This is preparation for combining the check_reshape and reconfig
methods
Signed-off-by: NeilBrown <neilb@suse.de>
A straight-forward conversion which gets rid of some
multiplications/divisions/shifts. The patch also introduces a couple
of new ones, most of which are due to conf->chunk_size still being
represented in bytes. This will be cleaned up in subsequent patches.
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
This patch renames the chunk_size field to chunk_sectors with the
implied change of semantics. Since
is_power_of_2(chunk_size) = is_power_of_2(chunk_sectors << 9)
= is_power_of_2(chunk_sectors)
these bits don't need an adjustment for the shift.
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
- update inclusion guard and make sure it covers the whole file
- remove superflous #ifdef CONFIG_BLOCK
- make sure all required headers are included so that new users aren't
required to include others before
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: NeilBrown <neilb@suse.de>
Currently raid5 (the only module that supports restriping)
notices that the reshape has finished be sync_request being
given a large value, and handles any cleanup them.
This patch changes it so md_check_recovery calls into an
explicit finish_reshape method as well.
The clean-up from sync_request can do things that need to be
done promptly, typically things local to the raid5_conf_t
structure.
The "finish_reshape" method is called under the mddev_lock
so it can do things involving reconfiguring the device.
This allows us to get rid of md_set_array_sectors_locked, which
would have caused a deadlock if you tried to stop and array
while a reshape was happening.
Signed-off-by: NeilBrown <neilb@suse.de>
Allow userspace to set the size of the array according to the following
semantics:
1/ size must be <= to the size returned by mddev->pers->size(mddev, 0, 0)
a) If size is set before the array is running, do_md_run will fail
if size is greater than the default size
b) A reshape attempt that reduces the default size to less than the set
array size should be blocked
2/ once userspace sets the size the kernel will not change it
3/ writing 'default' to this attribute returns control of the size to the
kernel and reverts to the size reported by the personality
Also, convert locations that need to know the default size from directly
reading ->array_sectors to <pers>_size. Resync/reshape operations
always follow the default size.
Finally, fixup other locations that read a number of 1k-blocks from
userspace to use strict_blocks_to_sectors() which checks for unsigned
long long to sector_t overflow and blocks to sectors overflow.
Reviewed-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Get personalities out of the business of directly modifying
->array_sectors. Lays groundwork to introduce policy on when
->array_sectors can be modified.
Reviewed-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for giving userspace control over ->array_sectors we need
to be able to retrieve the 'default' size, and the 'anticipated' size
when a reshape is requested. For personalities that do not reshape emit
a warning if anything but the default size is requested.
In the raid5 case we need to update ->previous_raid_disks to make the
new 'default' size available.
Reviewed-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Implement this for RAID6 to be able to 'takeover' a RAID5 array. The
new RAID6 will use a layout which places Q on the last device, and
that device will be missing.
If there are any available spares, one will immediately have Q
recovered onto it.
Signed-off-by: NeilBrown <neilb@suse.de>
To be able to change the 'level' of an md/raid array, we need to
suspend the device so that no requests are active - then move some
pointers around etc.
The code already keeps counts of active requests and the ->quiesce
function can be used to wait until those counts hit zero.
However the quiesce function blocks new requests once they are all
ready 'inside' the personality module, and that is too late if we want
to replace the personality modules.
So make all md requests come in through a common md_make_request
function that keeps track of how many requests have entered the
modules but may not yet be on the internal reference counts.
Allow md_make_request to be blocked when we want to suspend the
device, and make it possible to wait for all those in-transit requests
to be added to internal lists so that ->quiesce can wait for them.
There is still a problem that when a request completes, we drop the
ref count inside the personality code so there is a short time between
when the refcount hits zero, and when the personality code is no
longer being used.
The personality code never blocks (schedule or spinlock) between
dropping the refcount and exiting the routine, so this should be safe
(as put_module calls synchronize_sched() before unmapping the module
code).
Signed-off-by: NeilBrown <neilb@suse.de>
This patch renames the "size" field of struct mdk_rdev_s to
"sectors" and changes this field to store sectors instead of
blocks.
All users of this field, linear.c, raid0.c and md.c, are fixed up
accordingly which gets rid of many multiplications and divisions.
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
This patch renames the "size" field of struct mddev_s to "dev_sectors"
and stores the number of 512-byte sectors instead of the number of
1K-blocks in it.
All users of that field, including raid levels 1,4-6,10, are adjusted
accordingly. This simplifies the code a bit because it allows to get
rid of a couple of divisions/multiplications by two.
In order to make checkpatch happy, some minor coding style issues
have also been addressed. In particular, size_store() now uses
strict_strtoull() instead of simple_strtoull().
Signed-off-by: Andre Noll <maan@systemlinux.org>
Signed-off-by: NeilBrown <neilb@suse.de>
Version 1.x metadata has the ability to record the status of a
partially completed drive recovery.
However we only update that record on a clean shutdown.
It would be nice to update it on unclean shutdowns too, particularly
when using a bitmap that removes much to the 'sync' effort after an
unclean shutdown.
One complication with checkpointing recovery is that we only know
where we are up to in terms of IO requests started, not which ones
have completed. And we need to know what has completed to record
how much is recovered. So occasionally pause the recovery until all
submitted requests are completed, then update the record of where
we are up to.
When we have a bitmap, we already do that pause occasionally to keep
the bitmap up-to-date. So enhance that code to record the recovery
offset and schedule a superblock update.
And when there is no bitmap, just pause 16 times during the resync to
do a checkpoint.
'16' is a fairly arbitrary number. But we don't really have any good
way to judge how often is acceptable, and it seems like a reasonable
number for now.
Signed-off-by: NeilBrown <neilb@suse.de>
NeilBrown