DM RAID: Add ability to restore transiently failed devices on resume

This patch adds code to the resume function to check over the devices
in the RAID array.  If any are found to be marked as failed and their
superblocks can be read, an attempt is made to reintegrate them into
the array.  This allows the user to refresh the array with a simple
suspend and resume of the array - rather than having to load a
completely new table, allocate and initialize all the structures and
throw away the old instantiation.
Signed-off-by: NJonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

There are cases where the kernel will believe that the WRITE SAME
command is supported by a block device which does not, in fact,
support WRITE SAME.  This currently happens for SATA drivers behind a
SAS controller, but there are probably a hundred other ways that can
happen, including drive firmware bugs.

After receiving an error for WRITE SAME the block layer will retry the
request as a plain write of zeroes, but mdraid will consider the
failure as fatal and consider the drive failed.  This has the effect
that all the mirrors containing a specific set of data are each
offlined in very rapid succession resulting in data loss.

However, just bouncing the request back up to the block layer isn't
ideal either, because the whole initial request-retry sequence should
be inside the write bitmap fence, which probably means that md needs
to do its own conversion of WRITE SAME to write zero.

Until the failure scenario has been sorted out, disable WRITE SAME for
raid1, raid5, and raid10.

[neilb: added raid5]

This patch is appropriate for any -stable since 3.7 when write_same
support was added.

Cc: stable@vger.kernel.org
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Various places in raid1 and raid10 are calling raise_barrier when they
really should call freeze_array.
The former is only intended to be called from "make_request".
The later has extra checks for 'nr_queued' and makes a call to
flush_pending_writes(), so it is safe to call it from within the
management thread.

Using raise_barrier will sometimes deadlock.  Using freeze_array
should not.

As 'freeze_array' currently expects one request to be pending (in
handle_read_error - the only previous caller), we need to pass
it the number of pending requests (extra) to ignore.

The deadlock was made particularly noticeable by commits
050b6615 (raid10) and 6b740b8d (raid1) which
appeared in 3.4, so the fix is appropriate for any -stable
kernel since then.

This patch probably won't apply directly to some early kernels and
will need to be applied by hand.

Cc: stable@vger.kernel.org
Reported-by: NAlexander Lyakas <alex.bolshoy@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

md/raid1: consider WRITE as successful only if at least one non-Faulty and non-rebuilding drive completed it.

Without that fix, the following scenario could happen:

- RAID1 with drives A and B; drive B was freshly-added and is rebuilding
- Drive A fails
- WRITE request arrives to the array. It is failed by drive A, so
r1_bio is marked as R1BIO_WriteError, but the rebuilding drive B
succeeds in writing it, so the same r1_bio is marked as
R1BIO_Uptodate.
- r1_bio arrives to handle_write_finished, badblocks are disabled,
md_error()->error() does nothing because we don't fail the last drive
of raid1
- raid_end_bio_io()  calls call_bio_endio()
- As a result, in call_bio_endio():
        if (!test_bit(R1BIO_Uptodate, &r1_bio->state))
                clear_bit(BIO_UPTODATE, &bio->bi_flags);
this code doesn't clear the BIO_UPTODATE flag, and the whole master
WRITE succeeds, back to the upper layer.

So we returned success to the upper layer, even though we had written
the data onto the rebuilding drive only. But when we want to read the
data back, we would not read from the rebuilding drive, so this data
is lost.

[neilb - applied identical change to raid10 as well]

This bug can result in lost data, so it is suitable for any
-stable kernel.

Cc: stable@vger.kernel.org
Signed-off-by: NAlex Lyakas <alex@zadarastorage.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

In SSD/hard disk hybid storage, discard request should be ignored for hard
disk. We used to be doing this way, but the unplug path forgets it.

This is suitable for stable tree since v3.6.

Cc: stable@vger.kernel.org
Reported-and-tested-by: NMarkus <M4rkusXXL@web.de>
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Hi.

Raid1 and raid10 devices leak memory every time they stop.
This is a patch for linux-3.9.0-rc7 to fix this problem.

Thanks,
Hirokazu Takahashi.
Signed-off-by: NHirokazu Takahashi <taka@valinux.co.jp>
Signed-off-by: NNeilBrown <neilb@suse.de>

More utility code to replace stuff that's getting open coded.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>

More prep work for immutable bvecs:

A few places in the code were either open coding or using the wrong
version - fix.

After we introduce the bvec iter, it'll no longer be possible to modify
the biovec through bio_for_each_segment_all() - it doesn't increment a
pointer to the current bvec, you pass in a struct bio_vec (not a
pointer) which is updated with what the current biovec would be (taking
into account bi_bvec_done and bi_size).

So because of that it's more worthwhile to be consistent about
bio_for_each_segment()/bio_for_each_segment_all() usage.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>
CC: Alasdair Kergon <agk@redhat.com>
CC: dm-devel@redhat.com
CC: Alexander Viro <viro@zeniv.linux.org.uk>

__bio_for_each_segment() iterates bvecs from the specified index
instead of bio->bv_idx.  Currently, the only usage is to walk all the
bvecs after the bio has been advanced by specifying 0 index.

For immutable bvecs, we need to split these apart;
bio_for_each_segment() is going to have a different implementation.
This will also help document the intent of code that's using it -
bio_for_each_segment_all() is only legal to use for code that owns the
bio.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Neil Brown <neilb@suse.de>
CC: Boaz Harrosh <bharrosh@panasas.com>

This doesn't really delete any code _yet_, but once immutable bvecs are
done we can just delete the rest of the code in that loop.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>

More bi_idx removal. This code was just open coding bio_clone(). This
could probably be further improved by using bio_advance() instead of
skipping over null pages, but that'd be a larger rework.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>

Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>

Random cleanup - this code was duplicated and it's not really specific
to md.

Also added the ability to return the actual error code.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>
Acked-by: NTejun Heo <tj@kernel.org>

Bunch of places in the code weren't using it where they could be -
this'll reduce the size of the patch that puts bi_sector/bi_size/bi_idx
into a struct bvec_iter.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: "Ed L. Cashin" <ecashin@coraid.com>
CC: Nick Piggin <npiggin@kernel.dk>
CC: Jiri Kosina <jkosina@suse.cz>
CC: Jim Paris <jim@jtan.com>
CC: Geoff Levand <geoff@infradead.org>
CC: Alasdair Kergon <agk@redhat.com>
CC: dm-devel@redhat.com
CC: Neil Brown <neilb@suse.de>
CC: Steven Rostedt <rostedt@goodmis.org>
Acked-by: NEd Cashin <ecashin@coraid.com>

Just a little convenience macro - main reason to add it now is preparing
for immutable bio vecs, it'll reduce the size of the patch that puts
bi_sector/bi_size/bi_idx into a struct bvec_iter.
Signed-off-by: NKent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Lars Ellenberg <drbd-dev@lists.linbit.com>
CC: Jiri Kosina <jkosina@suse.cz>
CC: Alasdair Kergon <agk@redhat.com>
CC: dm-devel@redhat.com
CC: Neil Brown <neilb@suse.de>
CC: Martin Schwidefsky <schwidefsky@de.ibm.com>
CC: Heiko Carstens <heiko.carstens@de.ibm.com>
CC: linux-s390@vger.kernel.org
CC: Chris Mason <chris.mason@fusionio.com>
CC: Steven Whitehouse <swhiteho@redhat.com>
Acked-by: NSteven Whitehouse <swhiteho@redhat.com>

When raid1/raid10 needs to fix a read error, it first drains
all pending requests by calling freeze_array().
This calls flush_pending_writes() if it needs to sleep,
but some writes may be pending in a per-process plug rather
than in the per-array request queue.

When raid1{,0}_unplug() moves the request from the per-process
plug to the per-array request queue (from which
flush_pending_writes() can flush them), it needs to wake up
freeze_array(), or freeze_array() will never flush them and so
it will block forever.

So add the requires wake_up() calls.

This bug was introduced by commit
   f54a9d0e
for raid1 and a similar commit for RAID10, and so has been present
since linux-3.6.  As the bug causes a deadlock I believe this fix is
suitable for -stable.

Cc: stable@vger.kernel.org (3.6.y 3.7.y 3.8.y)
Reported-by: NTregaron Bayly <tbayly@bluehost.com>
Tested-by: NTregaron Bayly <tbayly@bluehost.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Set mddev queue's max_write_same_sectors to its chunk_sector value (before
disk_stack_limits merges the underlying disk limits.)  With that in place,
be sure to handle writes coming down from the block layer that have the
REQ_WRITE_SAME flag set.  That flag needs to be copied into any newly cloned
write bio.
Signed-off-by: NJoe Lawrence <joe.lawrence@stratus.com>
Acked-by: N"Martin K. Petersen" <martin.petersen@oracle.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

New wait_event{_interruptible}_lock_irq{_cmd} macros added. This commit
moves the private wait_event_lock_irq() macro from MD to regular wait
includes, introduces new macro wait_event_lock_irq_cmd() instead of using
the old method with omitting cmd parameter which is ugly and makes a use
of new macros in the MD. It also introduces the _interruptible_ variant.

The use of new interface is when one have a special lock to protect data
structures used in the condition, or one also needs to invoke "cmd"
before putting it to sleep.

All new macros are expected to be called with the lock taken. The lock
is released before sleep and is reacquired afterwards. We will leave the
macro with the lock held.

Note to DM: IMO this should also fix theoretical race on waitqueue while
using simultaneously wait_event_lock_irq() and wait_event() because of
lack of locking around current state setting and wait queue removal.
Signed-off-by: NLukas Czerner <lczerner@redhat.com>
Cc: Neil Brown <neilb@suse.de>
Cc: David Howells <dhowells@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

If the raid1 or raid10 unplug function gets called
from a make_request function (which is very possible) when
there are bios on the current->bio_list list, then it will not
be able to successfully call bitmap_unplug() and it could
need to submit more bios and wait for them to complete.
But they won't complete while current->bio_list is non-empty.

So detect that case and handle the unplugging off to another thread
just like we already do when called from within the scheduler.

RAID1 version of bug was introduced in 3.6, so that part of fix is
suitable for 3.6.y.  RAID10 part won't apply.

Cc: stable@vger.kernel.org
Reported-by: NTorsten Kaiser <just.for.lkml@googlemail.com>
Reported-by: NPeter Maloney <peter.maloney@brockmann-consult.de>
Signed-off-by: NNeilBrown <neilb@suse.de>

setup_conf in raid1.c uses conf->raid_disks before assigning
a value.  It is used when including 'Replacement' devices.

The consequence is that assembling an array which contains a
replacement will misbehave and either not include the replacement, or
not include the device being replaced.

Though this doesn't lead directly to data corruption, it could lead to
reduced data safety.

So use mddev->raid_disks, which is initialised, instead.

Bug was introduced by commit c19d5798
      md/raid1: recognise replacements when assembling arrays.

in 3.3, so fix is suitable for 3.3.y thru 3.6.y.

Cc: stable@vger.kernel.org
Signed-off-by: NNeilBrown <neilb@suse.de>

Now that multiple threads can handle stripes, it is safer to
use an atomic64_t for resync_mismatches, to avoid update races.
Signed-off-by: NJianpeng Ma <majianpeng@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

When we get a read error, we arrange for raid1d to handle it.
Currently we release the reference on the device.  This can result
in
   conf->mirrors[read_disk].rdev
being NULL in fix_read_error, if the device happens to get removed
before the read error is handled.

So instead keep the reference until the read error has been fully
handled.
Reported-by: Nhank <pyu@redhat.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Change the thread parameter, so the thread can carry extra info. Next patch
will use it.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

This makes md raid 1 support TRIM.
If one disk supports discard and another not, or one has discard_zero_data and
another not, there could be inconsistent between data from such disks. But this
should not matter, discarded data is useless. This will add extra copy in rebuild
though.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

queuing writes to the md thread means that all requests go through the
one processor which may not be able to keep up with very high request
rates.

So use the plugging infrastructure to submit all requests on unplug.
If a 'schedule' is needed, we fall back on the old approach of handing
the requests to the thread for it to handle.
Signed-off-by: NNeilBrown <neilb@suse.de>

This seemed like a good idea at the time, but after further thought I
cannot see it making a difference other than very occasionally and
testing to try to exercise the case it is most likely to help did not
show any performance difference by removing it.

So remove the counting of active plugs and allow 'pending writes' to
be activated at any time, not just when no plugs are active.

This is only relevant when there is a write-intent bitmap, and the
updating of the bitmap will likely introduce enough delay that
the single-threading of bitmap updates will be enough to collect large
numbers of updates together.

Removing this will make it easier to centralise the unplug code, and
will clear the other for other unplug enhancements which have a
measurable effect.
Signed-off-by: NNeilBrown <neilb@suse.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

When doing resync or repair, attempt to correct bad blocks, according
to WriteErrorSeen policy
Signed-off-by: NAlex Lyakas <alex.bolshoy@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

If a resync of a RAID1 array with 2 devices finds a known bad block
one device it will neither read from, or write to, that device for
this block offset.
So there will be one read_target (The other device) and zero write
targets.
This condition causes md/raid1 to abort the resync assuming that it
has finished - without known bad blocks this would be true.

When there are no write targets because of the presence of bad blocks
we should only skip over the area covered by the bad block.
RAID10 already gets this right, raid1 doesn't.  Or didn't.

As this can cause a 'sync' to abort early and appear to have succeeded
it could lead to some data corruption, so it suitable for -stable.

Cc: stable@vger.kernel.org
Reported-by: NAlexander Lyakas <alex.bolshoy@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

For SSD, if request size exceeds specific value (optimal io size), request size
isn't important for bandwidth. In such condition, if making request size bigger
will cause some disks idle, the total throughput will actually drop. A good
example is doing a readahead in a two-disk raid1 setup.

So when should we split big requests? We absolutly don't want to split big
request to very small requests. Even in SSD, big request transfer is more
efficient. This patch only considers request with size above optimal io size.

If all disks are busy, is it worth doing a split? Say optimal io size is 16k,
two requests 32k and two disks. We can let each disk run one 32k request, or
split the requests to 4 16k requests and each disk runs two. It's hard to say
which case is better, depending on hardware.

So only consider case where there are idle disks. For readahead, split is
always better in this case. And in my test, below patch can improve > 30%
thoughput. Hmm, not 100%, because disk isn't 100% busy.

Such case can happen not just in readahead, for example, in directio. But I
suppose directio usually will have bigger IO depth and make all disks busy, so
I ignored it.

Note: if the raid uses any hard disk, we don't prevent merging. That will make
performace worse.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

SSD hasn't spindle, distance between requests means nothing. And the original
distance based algorithm sometimes can cause severe performance issue for SSD
raid.

Considering two thread groups, one accesses file A, the other access file B.
The first group will access one disk and the second will access the other disk,
because requests are near from one group and far between groups. In this case,
read balance might keep one disk very busy but the other relative idle.  For
SSD, we should try best to distribute requests to as many disks as possible.
There isn't spindle move penality anyway.

With below patch, I can see more than 50% throughput improvement sometimes
depending on workloads.

The only exception is small requests can be merged to a big request which
typically can drive higher throughput for SSD too. Such small requests are
sequential reads. Unlike hard disk, sequential read which can't be merged (for
example direct IO, or read without readahead) can be ignored for SSD. Again
there is no spindle move penality. readahead dispatches small requests and such
requests can be merged.

Last patch can help detect sequential read well, at least if concurrent read
number isn't greater than raid disk number. In that case, distance based
algorithm doesn't work well too.

V2: For hard disk and SSD mixed raid, doesn't use distance based algorithm for
random IO too. This makes the algorithm generic for raid with SSD.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Currently the sequential read detection is global wide. It's natural to make it
per disk based, which can improve the detection for concurrent multiple
sequential reads. And next patch will make SSD read balance not use distance
based algorithm, where this change help detect truly sequential read for SSD.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

MD RAID1/RAID10: Move some macros from .h file to .c file

There are three macros (IO_BLOCKED,IO_MADE_GOOD,BIO_SPECIAL) which are defined
in both raid1.h and raid10.h.  They are only used in there respective .c files.
However, if we wish to make RAID10 accessible to the device-mapper RAID
target (dm-raid.c), then we need to move these macros into the .c files where
they are used so that they do not conflict with each other.

The macros from the two files are identical and could be moved into md.h, but
I chose to leave the duplication and have them remain in the personality
files.
Signed-off-by: NJonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

MD RAID1: Rename the structure 'mirror_info' to 'raid1_info'

The same structure name ('mirror_info') is used by raid10.  Each of these
structures are defined in there respective header files.  If dm-raid is
to support both RAID1 and RAID10, the header files will be included and
the structure names must not collide.  While only one of these structure
names needs to change, this patch adds consistency to the naming of the
structure.
Signed-off-by: NJonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

commit 4367af55
   md/raid1: clear bad-block record when write succeeds.

Added a 'reschedule_retry' call possibility at the end of
end_sync_write, but didn't add matching code at the end of
sync_request_write.  So if the writes complete very quickly, or
scheduling makes it seem that way, then we can miss rescheduling
the request and the resync could hang.

Also commit 73d5c38a
    md: avoid races when stopping resync.

Fix a race condition in this same code in end_sync_write but didn't
make the change in sync_request_write.

This patch updates sync_request_write to fix both of those.
Patch is suitable for 3.1 and later kernels.
Reported-by: NAlexander Lyakas <alex.bolshoy@gmail.com>
Original-version-by: NAlexander Lyakas <alex.bolshoy@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: NNeilBrown <neilb@suse.de>

This bug has been present ever since data-check was introduce
in 2.6.16.  However it would only fire if a data-check were
done on a degraded array, which was only possible if the array
has 3 or more devices.  This is certainly possible, but is quite
uncommon.

Since hot-replace was added in 3.3 it can happen more often as
the same condition can arise if not all possible replacements are
present.

The problem is that as soon as we submit the last read request, the
'r1_bio' structure could be freed at any time, so we really should
stop looking at it.  If the last device is being read from we will
stop looking at it.  However if the last device is not due to be read
from, we will still check the bio pointer in the r1_bio, but the
r1_bio might already be free.

So use the read_targets counter to make sure we stop looking for bios
to submit as soon as we have submitted them all.

This fix is suitable for any -stable kernel since 2.6.16.

Cc: stable@vger.kernel.org
Reported-by: NArnold Schulz <arnysch@gmx.net>
Signed-off-by: NNeilBrown <neilb@suse.de>

The value returned by "mddev_check_plug" is only valid until the
next 'schedule' as that will unplug things.  This could happen at any
call to mempool_alloc.
So just calling mddev_check_plug at the start doesn't really make
sense.

So call it just before, or just after, queuing things for the thread.
As the action that happens at unplug is to wake the thread, this makes
lots of sense.
If we cannot add a plug (which requires a small GFP_ATOMIC alloc) we
wake thread immediately.

RAID5 is a bit different.  Requests are queued for the thread and the
thread is woken by release_stripe.  So we don't need to wake the
thread on failure.
However the thread doesn't perform certain actions when there is any
active plug, so it is important to install a plug before waking the
thread.  So for RAID5 we install the plug *before* queuing the request
and waking the thread.

Without this patch it is possible for raid1 or raid10 to queue a
request without then waking the thread, resulting in the array locking
up.

Also change raid10 to only flush_pending_write when there are not
active plugs, just like raid1.

This patch is suitable for 3.0 or later.  I plan to submit it to
-stable, but I'll like to let it spend a few weeks in mainline
first to be sure it is completely safe.
Signed-off-by: NNeilBrown <neilb@suse.de>

When we added hot_replace we doubled the number of devices
that could be in a RAID1 array.  So we doubled how far read_balance
would search.  Unfortunately we didn't double the point at which
it looped back to the beginning - so it effectively loops over
all non-replacement disks twice.
This doesn't cause bad behaviour, but it pointless and means we
never read from replacement devices.
Signed-off-by: NNeilBrown <neilb@suse.de>

Having the 'name' arg optional and defaulting to the current
personality name is no necessary and leads to errors, as when
changing the level of an array we can end up using the
name of the old level instead of the new one.

So make it non-optional and always explicitly pass the name
of the level that the array will be.
Reported-by: Nmajianpeng <majianpeng@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

The new merge_bvec_fn which calls the corresponding function
in subsidiary devices requires that mddev->merge_check_needed
be set if any child has a merge_bvec_fn.

However were were only setting that when a device was hot-added,
not when a device was present from the start.

This bug was introduced in 3.4 so patch is suitable for 3.4.y
kernels.  However that are conflicts in raid10.c so a separate
patch will be needed for 3.4.y.

Cc: stable@vger.kernel.org
Reported-by: NSebastian Riemer <sebastian.riemer@profitbricks.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

A RAID1 device does not necessarily need a fullsync if the bitmap can be used instead.

Similar to commit d6b212f4 in raid5.c, if a raid1
device can be brought back (i.e. from a transient failure) it shouldn't need a
complete resync.  Provided the bitmap is not to old, it will have recorded the areas
of the disk that need recovery.
Signed-off-by: NJonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NNeilBrown <neilb@suse.de>