When an array is being reshaped to change the number of devices,
the two halves can be differently degraded.  e.g. one could be
missing a device and the other not.

So we need to be more careful about calculating the 'degraded'
attribute.

Instead of just inc/dec at appropriate times, perform a full
re-calculation examining both possible cases.  This doesn't happen
often so it not a big cost, and we already have most of the code to
do it.
Signed-off-by: NNeilBrown <neilb@suse.de>

While reshaping a degraded array (as when reshaping a RAID0 by first
converting it to a degraded RAID4) we currently get confused about
which devices are in_sync.  In most cases we get it right, but in the
region that is being reshaped we need to treat non-failed devices as
in-sync when we have the data but haven't actually written it out yet.
Reported-by: NAdam Kwolek <adam.kwolek@intel.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

NULL pointer access causes crash in raid5 module.
Signed-off-by: NAdam Kwolek <adam.kwolek@intel.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

Once a device is failed we really want to completely ignore it.
It should go away soon anyway.

In particular the presence of bad blocks on it should not cause us to
block as we won't be trying to write there anyway.

So as soon as we can check if a device is Faulty, do so and pretend
that it is already gone if it is Faulty.
Signed-off-by: NNeilBrown <neilb@suse.de>

All updates that occur under STRIPE_ACTIVE should be globally visible
when STRIPE_ACTIVE clears.  test_and_set_bit() implies a barrier, but
clear_bit() does not.

This is suitable for 3.1-stable.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>
Signed-off-by: NNeilBrown <neilb@suse.de>
Cc: stable@kernel.org

When the number of failed devices exceeds the allowed number
we must abort any active parity operations (checks or updates) as they
are no longer meaningful, and can lead to a BUG_ON in
handle_parity_checks6.

This bug was introduce by commit 6c0069c0
in 2.6.29.
Reported-by: NManish Katiyar <mkatiyar@gmail.com>
Tested-by: NManish Katiyar <mkatiyar@gmail.com>
Acked-by: NDan Williams <dan.j.williams@intel.com>
Signed-off-by: NNeilBrown <neilb@suse.de>
Cc: stable@kernel.org

A pending cleanup will mean that module.h won't be implicitly
everywhere anymore.  Make sure the modular drivers in md dir
are actually calling out for <module.h> explicitly in advance.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

In 3.0 we changed the way recovery_disabled was handle so that instead
of testing against zero, we test an mddev-> value against a conf->
value.
Two problems:
  1/ one place in raid1 was missed and still sets to '1'.
  2/ We didn't explicitly set the conf-> value at array creation
     time.
     It defaulted to '0' just like the mddev value does so they
     could appear equal and thus disable recovery.
     This did not affect normal 'md' as it calls bind_rdev_to_array
     which changes the mddev value.  However the dmraid interface
     doesn't call this and so doesn't change ->recovery_disabled; so at
     array start all recovery is incorrectly disabled.

So initialise the 'conf' value to one less that the mddev value, so
the will only be the same when explicitly set that way.
Reported-by: NJonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NNeilBrown  <neilb@suse.de>

This bug was introduced in 415e72d0
which was in 2.6.36.

There is a small window of time between when a device fails and when
it is removed from the array.  During this time we might still read
from it, but we won't write to it - so it is possible that we could
read stale data.

We didn't need the test of 'Faulty' before because the test on
In_sync is sufficient.  Since we started allowing reads from the early
part of non-In_sync devices we need a test on Faulty too.

This is suitable for any kernel from 2.6.36 onwards, though the patch
might need a bit of tweaking in 3.0 and earlier.

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

"mdk" doesn't mean anything any more.
Signed-off-by: NNeilBrown <neilb@suse.de>

Signed-off-by: NNeilBrown <neilb@suse.de>

Signed-off-by: NNeilBrown <neilb@suse.de>

Having mddev_t and 'struct mddev_s' is ugly and not preferred
Signed-off-by: NNeilBrown <neilb@suse.de>

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: NNeilBrown <neilb@suse.de>

This code is not really helpful and is hard to maintain, so just
discard it.
Signed-off-by: NNeilBrown <neilb@suse.de>

More type-safety.  Easier to read.
Signed-off-by: NNeilBrown <neilb@suse.de>

In the 'abort' branch of run(), 'conf' cannot possibly be NULL,
so remove the test.
Reported-by: NZdenek Kabelac <zdenek.kabelac@gmail.com>
Signed-off-by: NNeilBrown <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: N"Moshe Melnikov" <moshe@zadarastorage.com>
Signed-off-by: NNeilBrown <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: NChristoph Hellwig <hch@lst.de>
Acked-by: NNeilBrown <neilb@suse.de>
Signed-off-by: NJens Axboe <jaxboe@fusionio.com>

Waiting for a 'blocked' rdev to become unblocked in the raid5d thread
cannot work with internal metadata as it is the raid5d thread which
will clear the blocked flag.
This wasn't a problem in 3.0 and earlier as we only set the blocked
flag when external metadata was used then.
However we now set it always, so we need to be more careful.
Signed-off-by: NNeilBrown <neilb@suse.de>

On a successful write to a known bad block, flag the sh
so that raid5d can remove the known bad block from the list.
Signed-off-by: NNeilBrown <neilb@suse.de>

If a device has seen write errors, don't write to any known
bad blocks on that device.
Signed-off-by: NNeilBrown <neilb@suse.de>

When a write error is detected, don't mark the device as failed
immediately but rather record the fact for handle_stripe to deal with.

Handle_stripe then attempts to record a bad block.  Only if that fails
does the device get marked as faulty.
Signed-off-by: NNeilBrown <neilb@suse.de>

If we get an uncorrectable read error - record a bad block rather than
failing the device.
And if these errors (which may be due to known bad blocks) cause
recovery to be impossible, record a bad block on the recovering
devices, or abort the recovery.

As we might abort a recovery without failing a device we need to teach
RAID5 about recovery_disabled handling.
Signed-off-by: NNeilBrown <neilb@suse.de>

There are two times that we might read in raid5:
1/ when a read request fits within a chunk on a single
   working device.
   In this case, if there is any bad block in the range of
   the read, we simply fail the cache-bypass read and
   perform the read though the stripe cache.

2/ when reading into the stripe cache.  In this case we
   mark as failed any device which has a bad block in that
   strip (1 page wide).
   Note that we will both avoid reading and avoid writing.
   This is correct (as we will never read from the block, there
   is no point writing), but not optimal (as writing could 'fix'
   the error) - that will be addressed later.

If we have not seen any write errors on the device yet, we treat a bad
block like a recent read error.  This will encourage an attempt to fix
the read error which will either generate a write error, or will
ensure good data is stored there.  We don't yet forget the bad block
in that case.  That comes later.

Now that we honour bad blocks when reading we can allow devices with
bad blocks into the array.
Signed-off-by: NNeilBrown <neilb@suse.de>

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: NNeilBrown <neilb@suse.de>

As no personality understand bad block lists yet, we must
reject any device that is known to contain bad blocks.
As the personalities get taught, these tests can be removed.

This only applies to raid1/raid5/raid10.
For linear/raid0/multipath/faulty the whole concept of bad blocks
doesn't mean anything so there is no point adding the checks.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

While preparing to write a stripe we keep the parity block or blocks
locked (R5_LOCKED) - towards the end of schedule_reconstruction.

If the array is discovered to have failed before this write completes
we can leave those blocks LOCKED, and init_stripe will notice that a
free stripe still has a locked block and will complain.

So clear the R5_LOCKED flag in handle_failed_stripe, and demote the
'BUG' to a 'WARN_ON'.
Signed-off-by: NNeilBrown <neilb@suse.de>

Read errors are considered to corrected if write-back and re-read
cycle is finished without further problems. Thus moving the rdev->
corrected_errors counting after the re-reading looks more reasonable
IMHO.
Signed-off-by: NNamhyung Kim <namhyung@gmail.com>
Signed-off-by: NNeilBrown <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: NNamhyung Kim <namhyung@gmail.com>
Signed-off-by: NNeilBrown <neilb@suse.de>

As per printk_ratelimit comment, it should not be used.
Signed-off-by: NChristian Dietrich <christian.dietrich@informatik.uni-erlangen.de>
Signed-off-by: NNeilBrown <neilb@suse.de>

handle_stripe5() and handle_stripe6() are now virtually identical.
So discard one and rename the other to 'analyse_stripe()'.

It always returns 0, so change it to 'void' and remove the 'done'
variable in handle_stripe().
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

The RAID6 version of this code is usable for RAID5 providing:
  - we test "conf->max_degraded" rather than "2" as appropriate
  - we make sure s->failed_num[1] is meaningful (and not '-1')
    when s->failed > 1

The 'return 1' must become 'goto finish' in the new location.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

Apart from 'prexor' which can only be set for RAID5, and
'qd_idx' which can only be meaningful for RAID6, these two
chunks of code are nearly the same.

So combine them into one adding a test to call either
handle_parity_checks5 or handle_parity_checks6 as appropriate.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

RAID6 is only allowed to choose 'reconstruct-write' while RAID5 is
also allow 'read-modify-write'
Apart from this difference, handle_stripe_dirtying[56] are nearly
identical.  So resolve these differences and create just one function.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

Provided that ->failed_num[1] is not a valid device number (which is
easily achieved) fetch_block6 provides all the functionality of
fetch_block5.

So remove the latter and rename the former to simply "fetch_block".

Then handle_stripe_fill5 and handle_stripe_fill6 become the same and
can similarly be united.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

Next patch will unite fetch_block5 and fetch_block6.
First I want to make the differences a little more clear.

For RAID6 if we are writing at all and there is a failed device, then
we need to load or compute every block so we can do a
reconstruct-write.
This case isn't needed for RAID5 - we will do a read-modify-write in
that case.
So make that test a separate test in fetch_block6 rather than merged
with two other tests.

Make a similar change in fetch_block5 so the one bit that is not
needed for RAID6 is clearly separate.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

The difference between the RAID5 and RAID6 code here is easily
resolved using conf->max_degraded.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

Prior to commit ab69ae12 the code in handle_stripe5 and
handle_stripe6 to "Finish reconstruct operations initiated by the
expansion process" was identical.
That commit added an identical stanza of code to each function, but in
different places.  That was careless.

The raid5 code was correct, so move that out into handle_stripe and
remove raid6 version.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>

This arg is only used to differentiate between RAID5 and RAID6 but
that is not needed.  For RAID5, raid5_compute_sector will set qd_idx
to "~0" so j with certainly not equals qd_idx, so there is no need
for a guard on that condition.

So remove the guard and remove the arg from the declaration and
callers of handle_stripe_expansion.
Signed-off-by: NNeilBrown <neilb@suse.de>
Reviewed-by: NNamhyung Kim <namhyung@gmail.com>