GitHub/MotorolaMobilityLLC/kernel-slsi.git
13 years agomd/raid10: avoid writing to known bad blocks on known bad drives.
NeilBrown [Thu, 28 Jul 2011 01:39:24 +0000 (11:39 +1000)]
md/raid10: avoid writing to known bad blocks on known bad drives.

Writing to known bad blocks on drives that have seen a write error
is asking for trouble.  So try to avoid these blocks.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10 record bad blocks as needed during recovery.
NeilBrown [Thu, 28 Jul 2011 01:39:24 +0000 (11:39 +1000)]
md/raid10 record bad blocks as needed during recovery.

When recovering one or more devices, if all the good devices have
bad blocks we should record a bad block on the device being rebuilt.

If this fails, we need to abort the recovery.

To ensure we don't think that we aborted later than we actually did,
we need to move the check for MD_RECOVERY_INTR earlier in md_do_sync,
in particular before mddev->curr_resync is updated.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: avoid reading known bad blocks during resync/recovery.
NeilBrown [Thu, 28 Jul 2011 01:39:24 +0000 (11:39 +1000)]
md/raid10: avoid reading known bad blocks during resync/recovery.

During resync/recovery limit the size of the request to avoid
reading into a bad block that does not start at-or-before the current
read address.

Similarly if there is a bad block at this address, don't allow the
current request to extend beyond the end of that bad block.

Now that we don't ever read from known bad blocks, it is safe to allow
devices with those blocks into the array.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10 - avoid reading from known bad blocks - part 3
NeilBrown [Thu, 28 Jul 2011 01:39:24 +0000 (11:39 +1000)]
md/raid10 - avoid reading from known bad blocks - part 3

When attempting to repair a read error, don't read from
devices with a known bad block.

As we are only reading PAGE_SIZE blocks, we don't try to
narrow down to smaller regions in the hope that only part of this
page is bad - it isn't worth the effort.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: avoid reading from known bad blocks - part 2
NeilBrown [Thu, 28 Jul 2011 01:39:23 +0000 (11:39 +1000)]
md/raid10: avoid reading from known bad blocks - part 2

When redirecting a read error to a different device, we must
again avoid bad blocks and possibly split the request.

Spin_lock typo fixed thanks to Dan Carpenter <error27@gmail.com>

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: avoid reading from known bad blocks - part 1
NeilBrown [Thu, 28 Jul 2011 01:39:23 +0000 (11:39 +1000)]
md/raid10: avoid reading from known bad blocks - part 1

This patch just covers the basic read path:
 1/ read_balance needs to check for badblocks, and return not only
    the chosen slot, but also how many good blocks are available
    there.
 2/ 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'

On read error we currently just fail the request if another target
cannot handle the whole request.  Next patch refines that a bit.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: Split handle_read_error out from raid10d.
NeilBrown [Thu, 28 Jul 2011 01:39:23 +0000 (11:39 +1000)]
md/raid10: Split handle_read_error out from raid10d.

raid10d() is too big and is about to get bigger, so split
handle_read_error() out as a separate function.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: simplify/reindent some loops.
NeilBrown [Thu, 28 Jul 2011 01:39:23 +0000 (11:39 +1000)]
md/raid10: simplify/reindent some loops.

When a loop ends with a large if, it can be neater to change the
if to invert the condition and just 'continue'.
Then the body of the if can be indented to a lower level.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: Clear bad blocks on successful write.
NeilBrown [Thu, 28 Jul 2011 01:39:23 +0000 (11:39 +1000)]
md/raid5: Clear bad blocks on successful write.

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: NeilBrown <neilb@suse.de>
13 years agomd/raid5. Don't write to known bad block on doubtful devices.
NeilBrown [Thu, 28 Jul 2011 01:39:22 +0000 (11:39 +1000)]
md/raid5.  Don't write to known bad block on doubtful devices.

If a device has seen write errors, don't write to any known
bad blocks on that device.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: write errors should be recorded as bad blocks if possible.
NeilBrown [Thu, 28 Jul 2011 01:39:22 +0000 (11:39 +1000)]
md/raid5: write errors should be recorded as bad blocks if possible.

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: NeilBrown <neilb@suse.de>
13 years agomd/raid5: use bad-block log to improve handling of uncorrectable read errors.
NeilBrown [Thu, 28 Jul 2011 01:39:22 +0000 (11:39 +1000)]
md/raid5: use bad-block log to improve handling of uncorrectable read errors.

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: NeilBrown <neilb@suse.de>
13 years agomd/raid5: avoid reading from known bad blocks.
NeilBrown [Thu, 28 Jul 2011 01:39:22 +0000 (11:39 +1000)]
md/raid5: avoid reading from known bad blocks.

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: NeilBrown <neilb@suse.de>
13 years agomd/raid1: factor several functions out or raid1d()
NeilBrown [Thu, 28 Jul 2011 01:38:13 +0000 (11:38 +1000)]
md/raid1: factor several functions out or raid1d()

raid1d is too big with several deep branches.
So separate them out into their own functions.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: improve handling of read failure during recovery.
NeilBrown [Thu, 28 Jul 2011 01:33:42 +0000 (11:33 +1000)]
md/raid1: improve handling of read failure during recovery.

If we cannot read a block from anywhere during recovery, there is
now a better approach than just giving up.
We can record a bad block on each device and keep going - being
careful not to clear the bad block when a write succeeds as it might -
it will be a write of incorrect data.

We have now reached the state where - for raid1 - we only call
md_error if md_set_badblocks has failed.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: record badblocks found during resync etc.
NeilBrown [Thu, 28 Jul 2011 01:33:00 +0000 (11:33 +1000)]
md/raid1: record badblocks found during resync etc.

If we find a bad block while writing as part of resync/recovery we
need to report that back to raid1d which must record the bad block,
or fail the device.

Similarly when fixing a read error, a further error should just
record a bad block if possible rather than failing the device.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: Handle write errors by updating badblock log.
NeilBrown [Thu, 28 Jul 2011 01:32:41 +0000 (11:32 +1000)]
md/raid1:  Handle write errors by updating badblock log.

When we get a write error (in the data area, not in metadata),
update the badblock log rather than failing the whole device.

As the write may well be many blocks, we trying writing each
block individually and only log the ones which fail.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: store behind-write pages in bi_vecs.
NeilBrown [Thu, 28 Jul 2011 01:32:10 +0000 (11:32 +1000)]
md/raid1: store behind-write pages in bi_vecs.

When performing write-behind we allocate pages to store the data
during write.
Previously we just keep a list of pages.  Now we keep a list of
bi_vec which includes offset and size.
This means that the r1bio has complete information to create a new
bio which will be needed for retrying after write errors.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: clear bad-block record when write succeeds.
NeilBrown [Thu, 28 Jul 2011 01:31:49 +0000 (11:31 +1000)]
md/raid1: clear bad-block record when write succeeds.

If we succeed in writing to a block that was recorded as
being bad, we clear the bad-block record.

This requires some delayed handling as the bad-block-list update has
to happen in process-context.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid1: avoid writing to known-bad blocks on known-bad drives.
NeilBrown [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md/raid1: avoid writing to known-bad blocks on known-bad drives.

If we have seen any write error on a drive, then don't write to
any known-bad blocks on that drive.
If necessary, we divide the write request up into pieces just
like we do for reads, so each piece is either all written or
all not written to any given drive.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd: update documentation for md/rdev/state sysfs interface
Namhyung Kim [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md: update documentation for md/rdev/state sysfs interface

Previous patches in the bad block series extended behavior of
rdev's 'state' interface but lacked documentation update.
Fix it.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: make it easier to wait for bad blocks to be acknowledged.
NeilBrown [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md: make it easier to wait for bad blocks to be acknowledged.

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>
13 years agomd: add 'write_error' flag to component devices.
NeilBrown [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md: add 'write_error' flag to component devices.

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>
13 years agomd/raid1: avoid reading known bad blocks during resync
NeilBrown [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md/raid1: avoid reading known bad blocks during resync

When performing resync/etc, keep the size of the request
small enough that it doesn't overlap any known bad blocks.
Devices with badblocks at the start of the request are completely
excluded.
If there is nowhere to read from due to bad blocks, record
a bad block on each target device.

Now that we never read from known-bad-blocks we can allow devices with
known-bad-blocks into a RAID1.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid1: avoid reading from known bad blocks.
NeilBrown [Thu, 28 Jul 2011 01:31:48 +0000 (11:31 +1000)]
md/raid1: avoid reading from known bad blocks.

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>
13 years agomd: Disable bad blocks and v0.90 metadata.
NeilBrown [Thu, 28 Jul 2011 01:31:47 +0000 (11:31 +1000)]
md: Disable bad blocks and v0.90 metadata.

v0.90 metadata cannot record bad blocks, so when loading metadata
for such a device, set shift to -1.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: load/store badblock list from v1.x metadata
NeilBrown [Thu, 28 Jul 2011 01:31:47 +0000 (11:31 +1000)]
md: load/store badblock list from v1.x metadata

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>
13 years agomd: don't allow arrays to contain devices with bad blocks.
NeilBrown [Thu, 28 Jul 2011 01:31:47 +0000 (11:31 +1000)]
md: don't allow arrays to contain devices with bad blocks.

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd: add documentation for bad block log
Namhyung Kim [Thu, 28 Jul 2011 01:31:47 +0000 (11:31 +1000)]
md: add documentation for bad block log

Previous patch in the bad block series added new sysfs interfaces
([unacknowledged_]bad_blocks) for each rdev without documentation.
Add it.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/bad-block-log: add sysfs interface for accessing bad-block-log.
NeilBrown [Thu, 28 Jul 2011 01:31:47 +0000 (11:31 +1000)]
md/bad-block-log: add sysfs interface for accessing bad-block-log.

This can show the log (providing it fits in one page) and
allows bad blocks to be 'acknowledged' meaning that they
have safely been recorded in metadata.

Clearing bad blocks is not allowed via sysfs (except for
code testing).  A bad block can only be cleared when
a write to the block succeeds.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd: beginnings of bad block management.
NeilBrown [Thu, 28 Jul 2011 01:31:46 +0000 (11:31 +1000)]
md: beginnings of bad block management.

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>
13 years agomd: remove suspicious size_of()
NeilBrown [Wed, 27 Jul 2011 21:56:24 +0000 (07:56 +1000)]
md: remove suspicious size_of()

When calling bioset_create we pass the size of the front_pad as
   sizeof(mddev)
which looks suspicious as mddev is a pointer and so it looks like a
common mistake where
   sizeof(*mddev)
was intended.
The size is actually correct as we want to store a pointer in the
front padding of the bios created by the bioset, so make the intent
more explicit by using
   sizeof(mddev_t *)

Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agoMD: generate an event when array sync is complete
Jonathan Brassow [Wed, 27 Jul 2011 01:00:37 +0000 (11:00 +1000)]
MD: generate an event when array sync is complete

This patch causes MD to generate an event (for device-mapper) when the
synchronization thread is reaped.  This is expected behavior for device-mapper.

Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agoMD bitmap: Revert DM dirty log hooks
Jonathan Brassow [Wed, 27 Jul 2011 01:00:37 +0000 (11:00 +1000)]
MD bitmap: Revert DM dirty log hooks

Revert most of commit e384e58549a2e9a83071ad80280c1a9053cfd84c
  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
e384e58549a2e9a83071ad80280c1a9053cfd84c, however.

Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agoMD: raid1 s/sysfs_notify_dirent/sysfs_notify_dirent_safe
Jonathan Brassow [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
MD: raid1 s/sysfs_notify_dirent/sysfs_notify_dirent_safe

If device-mapper creates a RAID1 array that includes devices to
be rebuilt, it will deref a NULL pointer when finished because
sysfs is not used by device-mapper instantiated RAID devices.

Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: Avoid BUG caused by multiple failures.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: Avoid BUG caused by multiple failures.

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: NeilBrown <neilb@suse.de>
13 years agomd/raid10: move rdev->corrected_errors counting
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid10: move rdev->corrected_errors counting

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: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: move rdev->corrected_errors counting
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: move rdev->corrected_errors counting

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: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid1: move rdev->corrected_errors counting
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid1: move rdev->corrected_errors counting

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. Also included a couple of whitespace fixes on sync_page_io().

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: get rid of unnecessary casts on page_address()
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md: get rid of unnecessary casts on page_address()

page_address() returns void pointer, so the casts can be removed.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: Improve decision on whether to fail a device with a read error.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid10: Improve decision on whether to fail a device with a read error.

Normally we would fail a device with a READ error.  However if doing
so causes the array to fail, it is better to leave the device
in place and just return the read error to the caller.

The current test for decide if the array will fail is overly
simplistic.
We have a function 'enough' which can tell if the array is failed or
not, so use it to guide the decision.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: Make use of new recovery_disabled handling
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid10: Make use of new recovery_disabled handling

When we get a read error during recovery, RAID10 previously
arranged for the recovering device to appear to fail so that
the recovery stops and doesn't restart.  This is misleading and wrong.

Instead, make use of the new recovery_disabled handling and mark
the target device and having recovery disabled.

Add appropriate checks in add_disk and remove_disk so that devices
are removed and not re-added when recovery is disabled.

Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: change managed of recovery_disabled.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md: change managed of recovery_disabled.

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>
13 years agomd: remove ro check in md_check_recovery()
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md: remove ro check in md_check_recovery()

Commit c89a8eee6154 ("Allow faulty devices to be removed from a
readonly array.") added some work on ro array in the function,
but it couldn't be done since we didn't allow the ro array to be
handled from the beginning. Fix it.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: introduce link/unlink_rdev() helpers
Namhyung Kim [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md: introduce link/unlink_rdev() helpers

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>
13 years agomd/raid: use printk_ratelimited instead of printk_ratelimit
Christian Dietrich [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid: use printk_ratelimited instead of printk_ratelimit

As per printk_ratelimit comment, it should not be used.

Signed-off-by: Christian Dietrich <christian.dietrich@informatik.uni-erlangen.de>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd: use proper little-endian bitops
Akinobu Mita [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md: use proper little-endian bitops

Using __test_and_{set,clear}_bit_le() with ignoring its return value
can be replaced with __{set,clear}_bit_le().

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: NeilBrown <neilb@suse.de>
Cc: linux-raid@vger.kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: finalise new merged handle_stripe.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: finalise new merged handle_stripe.

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: move some more common code into handle_stripe
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: move some more common code into handle_stripe

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: move more common code into handle_stripe
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: move more common code into handle_stripe

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: unite handle_stripe_dirtying5 and handle_stripe_dirtying6
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: unite handle_stripe_dirtying5 and handle_stripe_dirtying6

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: unite fetch_block5 and fetch_block6
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: unite fetch_block5 and fetch_block6

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: rearrange a test in fetch_block6.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: rearrange a test in fetch_block6.

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: move more code into common handle_stripe
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: move more code into common handle_stripe

The difference between the RAID5 and RAID6 code here is easily
resolved using conf->max_degraded.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: Move code for finishing a reconstruction into handle_stripe.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: Move code for finishing a reconstruction into handle_stripe.

Prior to commit ab69ae12ceef7 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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: Remove stripe_head_state arg from handle_stripe_expansion.
NeilBrown [Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)]
md/raid5: Remove stripe_head_state arg from handle_stripe_expansion.

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: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: move stripe_head_state and more code into handle_stripe.
NeilBrown [Tue, 26 Jul 2011 01:35:35 +0000 (11:35 +1000)]
md/raid5: move stripe_head_state and more code into handle_stripe.

By defining the 'stripe_head_state' in 'handle_stripe', we can move
some common code out of handle_stripe[56]() and into handle_stripe.

The means that all accesses for stripe_head_state in handle_stripe[56]
need to be 's->' instead of 's.', but the compiler should inline
those functions and just use a direct stack reference, and future
patches while hoist most of this code up into handle_stripe()
so we will revert to "s.".

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: add some more fields to stripe_head_state
NeilBrown [Tue, 26 Jul 2011 01:35:20 +0000 (11:35 +1000)]
md/raid5:  add some more fields to stripe_head_state

Adding these three fields will allow more common code to be moved
to handle_stripe()

struct field rearrangement by Namhyung Kim.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: unify stripe_head_state and r6_state
NeilBrown [Tue, 26 Jul 2011 01:35:19 +0000 (11:35 +1000)]
md/raid5: unify stripe_head_state and r6_state

'struct stripe_head_state' stores state about the 'current' stripe
that is passed around while handling the stripe.
For RAID6 there is an extension structure: r6_state, which is also
passed around.
There is no value in keeping these separate, so move the fields from
the latter into the former.

This means that all code now needs to treat s->failed_num as an small
array, but this is a small cost.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: move common code into handle_stripe
NeilBrown [Tue, 26 Jul 2011 01:35:15 +0000 (11:35 +1000)]
md/raid5: move common code into handle_stripe

There is common code at the start of handle_stripe5 and
handle_stripe6.  Move it into handle_stripe.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: replace sh->lock with an 'active' flag.
NeilBrown [Tue, 26 Jul 2011 01:34:20 +0000 (11:34 +1000)]
md/raid5: replace sh->lock with an 'active' flag.

sh->lock is now mainly used to ensure that two threads aren't running
in the locked part of handle_stripe[56] at the same time.

That can more neatly be achieved with an 'active' flag which we set
while running handle_stripe.  If we find the flag is set, we simply
requeue the stripe for later by setting STRIPE_HANDLE.

For safety we take ->device_lock while examining the state of the
stripe and creating a summary in 'stripe_head_state / r6_state'.
This possibly isn't needed but as shared fields like ->toread,
->towrite are checked it is safer for now at least.

We leave the label after the old 'unlock' called "unlock" because it
will disappear in a few patches, so renaming seems pointless.

This leaves the stripe 'locked' for longer as we clear STRIPE_ACTIVE
later, but that is not a problem.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: Protect some more code with ->device_lock.
NeilBrown [Tue, 26 Jul 2011 01:20:35 +0000 (11:20 +1000)]
md/raid5: Protect some more code with ->device_lock.

Other places that change or follow dev->towrite and dev->written take
the device_lock as well as the sh->lock.
So it should really be held in these places too.
Also, doing so will allow sh->lock to be discarded.

with merged fixes by: Namhyung Kim <namhyung@gmail.com>

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: Remove use of sh->lock in sync_request
NeilBrown [Tue, 26 Jul 2011 01:19:49 +0000 (11:19 +1000)]
md/raid5: Remove use of sh->lock in sync_request

This is the start of a series of patches to remove sh->lock.

sync_request takes sh->lock before setting STRIPE_SYNCING to ensure
there is no race with testing it in handle_stripe[56].

Instead, use a new flag STRIPE_SYNC_REQUESTED and test it early
in handle_stripe[56] (after getting the same lock) and perform the
same set/clear operations if it was set.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
13 years agomd/raid5: get rid of duplicated call to bio_data_dir()
Namhyung Kim [Mon, 18 Jul 2011 07:38:51 +0000 (17:38 +1000)]
md/raid5: get rid of duplicated call to bio_data_dir()

In raid5::make_request(), once bio_data_dir(@bi) is detected
it never (and couldn't) be changed. Use the result always.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid5: use kmem_cache_zalloc()
Namhyung Kim [Mon, 18 Jul 2011 07:38:50 +0000 (17:38 +1000)]
md/raid5: use kmem_cache_zalloc()

Replace kmem_cache_alloc + memset(,0,) to kmem_cache_zalloc.
I think it's not harmful since @conf->slab_cache already knows
actual size of struct stripe_head.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: share pages between read and write bio's during recovery
Namhyung Kim [Mon, 18 Jul 2011 07:38:49 +0000 (17:38 +1000)]
md/raid10: share pages between read and write bio's during recovery

When performing a recovery, only first 2 slots in r10_bio are in use,
for read and write respectively. However all of pages in the write bio
are never used and just replaced to read bio's when the read completes.

Get rid of those unused pages and share read pages properly.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: factor out common bio handling code
Namhyung Kim [Mon, 18 Jul 2011 07:38:47 +0000 (17:38 +1000)]
md/raid10: factor out common bio handling code

When normal-write and sync-read/write bio completes, we should
find out the disk number the bio belongs to. Factor those common
code out to a separate function.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agomd/raid10: get rid of duplicated conditional expression
Namhyung Kim [Mon, 18 Jul 2011 07:38:43 +0000 (17:38 +1000)]
md/raid10: get rid of duplicated conditional expression

Variable 'first' is initialized to zero and updated to @rdev->raid_disk
only if it is greater than 0. Thus condition '>= first' always implies
'>= 0' so the latter is not needed.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
13 years agoMerge branch 'gpio/merge' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Sat, 16 Jul 2011 01:03:30 +0000 (18:03 -0700)]
Merge branch 'gpio/merge' of git://git.secretlab.ca/git/linux-2.6

* 'gpio/merge' of git://git.secretlab.ca/git/linux-2.6:
  gpio: wm831x: add a missing break in wm831x_gpio_dbg_show

13 years agoARM: fix regression in IXP4xx clocksource
Richard Cochran [Fri, 15 Jul 2011 19:33:12 +0000 (21:33 +0200)]
ARM: fix regression in IXP4xx clocksource

Commit 234b6ceddb4fc2a4bc5b9a7670f070f6e69e0868

   clocksource: convert ARM 32-bit up counting clocksources

broke the build for ixp4xx and made big endian operation impossible.
This commit restores the original behaviour.

Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
Signed-off-by: Krzysztof HaƂasa <khc@pm.waw.pl>
[ Thomas says that we might want to have generic BE accessor functions
  to the MMIO clock source, but that hasn't happened yet, so in the
  meantime this seems to be the short-term fix for the particular
  problem - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agogpio: wm831x: add a missing break in wm831x_gpio_dbg_show
Axel Lin [Sun, 10 Jul 2011 07:45:07 +0000 (15:45 +0800)]
gpio: wm831x: add a missing break in wm831x_gpio_dbg_show

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke...
Linus Torvalds [Fri, 15 Jul 2011 18:03:49 +0000 (11:03 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ryusuke/nilfs2

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2:
  nilfs2: remove resize from unsupported features list

13 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groec...
Linus Torvalds [Fri, 15 Jul 2011 17:21:21 +0000 (10:21 -0700)]
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging:
  hwmon: (adm1275) Fix coefficients per datasheet revision B
  hwmon: (pmbus) Use long variables for register to data conversions

13 years agoMerge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Fri, 15 Jul 2011 17:21:07 +0000 (10:21 -0700)]
Merge branch 'drm-fixes' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms: add new NI pci ids

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Fri, 15 Jul 2011 16:55:39 +0000 (09:55 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  fix loop checks in d_materialise_unique()
  Fix ->d_lock locking order in unlazy_walk()

13 years agoMerge branch 'rcu/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck...
Linus Torvalds [Fri, 15 Jul 2011 16:54:34 +0000 (09:54 -0700)]
Merge branch 'rcu/urgent' of git://git./linux/kernel/git/paulmck/linux-2.6-rcu

* 'rcu/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-2.6-rcu:
  rcu: Prevent RCU callbacks from executing before scheduler initialized

13 years agosched: Fix 32bit race
Peter Zijlstra [Mon, 11 Jul 2011 14:28:50 +0000 (16:28 +0200)]
sched: Fix 32bit race

Commit 3fe1698b7fe0 ("sched: Deal with non-atomic min_vruntime reads
on 32bit") forgot to initialize min_vruntime_copy which could lead to
an infinite while loop in task_waking_fair() under some circumstances
(early boot, lucky timing).

[ This bug was also reported by others that blamed it on the RCU
  initialization problems ]

Reported-and-tested-by: Bruno Wolff III <bruno@wolff.to>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agodrm/radeon/kms: add new NI pci ids
Alex Deucher [Fri, 15 Jul 2011 14:39:10 +0000 (14:39 +0000)]
drm/radeon/kms: add new NI pci ids

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agohwmon: (adm1275) Fix coefficients per datasheet revision B
Guenter Roeck [Thu, 14 Jul 2011 21:18:03 +0000 (14:18 -0700)]
hwmon: (adm1275) Fix coefficients per datasheet revision B

Coefficients to convert chip register values to voltage/current have been
slightly changed in revision B of the chip datasheet. Update driver coefficients
to match the coefficients in the datasheet.

Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
Acked-by: Jean Delvare <khali@linux-fr.org>
13 years agofix loop checks in d_materialise_unique()
Al Viro [Wed, 13 Jul 2011 01:42:24 +0000 (21:42 -0400)]
fix loop checks in d_materialise_unique()

Both __d_unalias() and __d_materialise_dentry() need loop prevention.
Grab rename_lock in caller, check for loops there...

As a side benefit, we have dentry_lock_for_move() called only under
rename_lock, which seriously reduces deadlock potential of the
execrable "locking order" used for ->d_lock.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-2.6-fixes
Linus Torvalds [Thu, 14 Jul 2011 17:20:42 +0000 (10:20 -0700)]
Merge git://git./linux/kernel/git/steve/gfs2-2.6-fixes

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-2.6-fixes:
  GFS2: Resolve inode eviction and ail list interaction bug
  GFS2: Fix race during filesystem mount
  GFS2: force a log flush when invalidating the rindex glock

13 years agoGFS2: Resolve inode eviction and ail list interaction bug
Steven Whitehouse [Thu, 14 Jul 2011 07:59:44 +0000 (08:59 +0100)]
GFS2: Resolve inode eviction and ail list interaction bug

This patch contains a few misc fixes which resolve a recently
reported issue. This patch has been a real team effort and has
received a lot of testing.

The first issue is that the ail lock needs to be held over a few
more operations. The lock thats added into gfs2_releasepage() may
possibly be a candidate for replacing with RCU at some future
point, but at this stage we've gone for the obvious fix.

The second issue is that gfs2_write_inode() can end up calling
a glock recursively when called from gfs2_evict_inode() via the
syncing code, so it needs a guard added.

The third issue is that we either need to not truncate the metadata
pages of inodes which have zero link count, but which we cannot
deallocate due to them still being in use by other nodes, or we need
to ensure that those pages have all made it through the journal and
ail lists first. This patch takes the former approach, but the
latter has also been tested and there is nothing to choose between
them performance-wise. So again, we could revise that decision
in the future.

Also, the inode eviction process is now better documented.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Tested-by: Bob Peterson <rpeterso@redhat.com>
Tested-by: Abhijith Das <adas@redhat.com>
Reported-by: Barry J. Marson <bmarson@redhat.com>
Reported-by: David Teigland <teigland@redhat.com>
13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc
Linus Torvalds [Wed, 13 Jul 2011 23:47:31 +0000 (16:47 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/cjb/mmc

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: core: Bus width testing needs to handle suspend/resume

13 years agoMerge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Wed, 13 Jul 2011 21:34:08 +0000 (14:34 -0700)]
Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
  SUNRPC: Fix use of static variable in rpcb_getport_async
  NFSv4.1: update nfs4_fattr_bitmap_maxsz
  SUNRPC: Fix a race between work-queue and rpc_killall_tasks
  pnfs: write: Set mds_offset in the generic layer - it is needed by all LDs

13 years agoMerge branch 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuil...
Linus Torvalds [Wed, 13 Jul 2011 21:16:53 +0000 (14:16 -0700)]
Merge branch 'rc-fixes' of git://git./linux/kernel/git/mmarek/kbuild-2.6

* 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild-2.6:
  kbuild: Do not write to builddir in modules_install

13 years agoMerge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 13 Jul 2011 21:04:26 +0000 (14:04 -0700)]
Merge branch 'drm-fixes' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms/evergreen: emit SQ_LDS_RESOURCE_MGMT for blits
  agp/intel: Fix typo in G4x_GMCH_SIZE_VT_2M
  drm/radeon/kms: fix typo in read_disabled vbios code
  drm/radeon/kms: use correct BUS_CNTL reg on rs600
  drm/radeon/kms: fix backend map typo on juniper
  drm/radeon/kms: fix regression in hotplug

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 13 Jul 2011 20:51:32 +0000 (13:51 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (21 commits)
  slip: fix wrong SLIP6 ifdef-endif placing
  natsemi: fix another dma-debug report
  sctp: ABORT if receive, reassmbly, or reodering queue is not empty while closing socket
  net: Fix default in docs for tcp_orphan_retries.
  hso: fix a use after free condition
  net/natsemi: Fix module parameter permissions
  XFRM: Fix memory leak in xfrm_state_update
  sctp: Enforce retransmission limit during shutdown
  mac80211: fix TKIP replay vulnerability
  mac80211: fix ie memory allocation for scheduled scans
  ssb: fix init regression of hostmode PCI core
  rtlwifi: rtl8192cu: Add new USB ID for Netgear WNA1000M
  ath9k: Fix tx throughput drops for AR9003 chips with AES encryption
  carl9170: add NEC WL300NU-AG usbid
  cfg80211: fix deadlock with rfkill/sched_scan by adding new mutex
  ath5k: fix incorrect use of drvdata in PCI suspend/resume code
  ath5k: fix incorrect use of drvdata in sysfs code
  Bluetooth: Fix memory leak under page timeouts
  Bluetooth: Fix regression with incoming L2CAP connections
  Bluetooth: Fix hidp disconnect deadlocks and lost wakeup
  ...

13 years agommc: core: Bus width testing needs to handle suspend/resume
Philip Rakity [Thu, 7 Jul 2011 16:04:55 +0000 (09:04 -0700)]
mmc: core: Bus width testing needs to handle suspend/resume

On reading the ext_csd for the first time (in 1 bit mode), save the
ext_csd information needed for bus width compare.

On every pass we make re-reading the ext_csd, compare the data
against the saved ext_csd data.

This fixes a regression introduced in 3.0-rc1 by 08ee80cc397ac1a3
("mmc: core: eMMC bus width may not work on all platforms"), which
incorrectly assumed we would be re-reading the ext_csd at resume-
time.

Signed-off-by: Philip Rakity <prakity@marvell.com>
Tested-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
13 years agorcu: Prevent RCU callbacks from executing before scheduler initialized
Paul E. McKenney [Sun, 10 Jul 2011 22:57:35 +0000 (15:57 -0700)]
rcu: Prevent RCU callbacks from executing before scheduler initialized

Under some rare but real combinations of configuration parameters, RCU
callbacks are posted during early boot that use kernel facilities that
are not yet initialized.  Therefore, when these callbacks are invoked,
hard hangs and crashes ensue.  This commit therefore prevents RCU
callbacks from being invoked until after the scheduler is fully up and
running, as in after multiple tasks have been spawned.

It might well turn out that a better approach is to identify the specific
RCU callbacks that are causing this problem, but that discussion will
wait until such time as someone really needs an RCU callback to be invoked
(as opposed to merely registered) during early boot.

Reported-by: julie Sullivan <kernelmail.jms@gmail.com>
Reported-by: RKK <kulkarni.ravi4@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: julie Sullivan <kernelmail.jms@gmail.com>
Tested-by: RKK <kulkarni.ravi4@gmail.com>
13 years agonilfs2: remove resize from unsupported features list
Ryusuke Konishi [Sun, 19 Jun 2011 07:56:29 +0000 (16:56 +0900)]
nilfs2: remove resize from unsupported features list

Resize feature was supported by the commit 4e33f9eab07e but it was not
reflected to the list of unsupported features in nilfs2.txt file.
This updates the list to fix discrepancy.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
13 years agodrm/radeon/kms/evergreen: emit SQ_LDS_RESOURCE_MGMT for blits
Alex Deucher [Tue, 12 Jul 2011 15:53:23 +0000 (11:53 -0400)]
drm/radeon/kms/evergreen: emit SQ_LDS_RESOURCE_MGMT for blits

Compute drivers may change this, so make sure to emit it to
avoid errors in bo blits.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=39119

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoagp/intel: Fix typo in G4x_GMCH_SIZE_VT_2M
Chris Wilson [Tue, 12 Jul 2011 22:38:18 +0000 (23:38 +0100)]
agp/intel: Fix typo in G4x_GMCH_SIZE_VT_2M

Konstantin Belousov found an error in the define of G4x_GMCH_SIZE_VT_2M
relative to the GMCH specs, and confirmed that indeed one of his users
with a Q45 reports 0xb not 0xc for a 2/2MiB GATT.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Konstantin Belousov <kostikbel@gmail.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoFix ->d_lock locking order in unlazy_walk()
Al Viro [Wed, 13 Jul 2011 01:40:23 +0000 (21:40 -0400)]
Fix ->d_lock locking order in unlazy_walk()

Make sure that child is still a child of parent before nested locking
of child->d_lock in unlazy_walk(); otherwise we are risking a violation
of locking order and deadlocks.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Tue, 12 Jul 2011 21:21:19 +0000 (14:21 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/mm: Fix memory_block_size_bytes() for non-pseries
  mm: Move definition of MIN_MEMORY_BLOCK_SIZE to a header

13 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc
Linus Torvalds [Tue, 12 Jul 2011 21:19:51 +0000 (14:19 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/arm/linux-arm-soc

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc:
  pcmcia: pxa2xx/vpac270: free gpios on exist rather than requesting
  ARM: pxa/raumfeld: fix device name for codec ak4104
  ARM: pxa/raumfeld: display initialisation fixes
  ARM: pxa/raumfeld: adapt to upcoming hardware change
  ARM: pxa: fix gpio_to_chip() clash with gpiolib namespace
  genirq: replace irq_gc_ack() with {set,clr}_bit variants (fwd)
  arm: mach-vt8500: add forgotten irq_data conversion
  ARM: pxa168: correct nand pmu setting
  ARM: pxa910: correct nand pmu setting
  ARM: pxa: fix PGSR register address calculation

13 years agoMerge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keith...
Linus Torvalds [Tue, 12 Jul 2011 21:18:00 +0000 (14:18 -0700)]
Merge branch 'drm-intel-fixes' of git://git./linux/kernel/git/keithp/linux-2.6

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keithp/linux-2.6:
  drm/i915/ringbuffer: Idling requires waiting for the ring to be empty
  Revert "drm/i915: enable rc6 by default"
  drm/i915: Clean up i915_driver_load failure path
  drm/i915: Enable GPU reset on Ivybridge.
  drm/i915/dp: manage sink power state if possible
  drm/i915/dp: consolidate AUX retry code
  drm/i915/dp: remove DPMS mode tracking from DP
  drm/i915/dp: try to read receiver capabilities 3 times when detecting
  drm/i915/dp: read more receiver capability bits on hotplug
  drm/i915/dp: use DP DPCD defines when looking at DPCD values
  drm/i915/dp: retry link status read 3 times on failure

13 years agoSUNRPC: Fix use of static variable in rpcb_getport_async
Ben Greear [Tue, 12 Jul 2011 17:27:55 +0000 (10:27 -0700)]
SUNRPC: Fix use of static variable in rpcb_getport_async

Because struct rpcbind_args *map was declared static, if two
threads entered this method at the same time, the values
assigned to map could be sent two two differen tasks.
This could cause all sorts of problems, include use-after-free
and double-free of memory.

Fix this by removing the static declaration so that the map
pointer is on the stack.

Signed-off-by: Ben Greear <greearb@candelatech.com>
Cc: stable@kernel.org
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
13 years agodrm/i915/ringbuffer: Idling requires waiting for the ring to be empty
Chris Wilson [Tue, 12 Jul 2011 17:03:29 +0000 (18:03 +0100)]
drm/i915/ringbuffer: Idling requires waiting for the ring to be empty

...which is measured by the size and not the amount of space remaining.

Waiting upon size-8, did one of two things. In the common case with more
than 8 bytes available to write into the ring, it would return
immediately. Otherwise, it would timeout given the impossible condition
of waiting for more space than is available in the ring, leading to
warnings such as:

[drm:intel_cleanup_ring_buffer] *ERROR* failed to quiesce render ring
whilst cleaning up: -16

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
13 years agoRevert "drm/i915: enable rc6 by default"
Keith Packard [Tue, 12 Jul 2011 15:49:31 +0000 (08:49 -0700)]
Revert "drm/i915: enable rc6 by default"

This reverts commit a51f7a66fb5e4af5ec4286baef940d06594b59d2.

We still have a few Ironlake and Sandybridge machines which fail when
RC6 is enabled. Better luck next release?

Signed-off-by: Keith Packard <keithp@keithp.com>
13 years agodrm/i915: Clean up i915_driver_load failure path
Keith Packard [Sun, 10 Jul 2011 20:12:17 +0000 (13:12 -0700)]
drm/i915: Clean up i915_driver_load failure path

i915_driver_load adds a write-combining MTRR region for the GTT
aperture to improve memory speeds through the aperture. If
i915_driver_load fails after this, it would not have cleaned up the
MTRR. This shouldn't cause any problems, except for consuming an MTRR
register. Still, it's best to clean up completely in the failure path,
which is easily done by calling mtrr_del if the mtrr was successfully
allocated.

i915_driver_load calls i915_gem_load which register
i915_gem_inactive_shrink. If i915_driver_load fails after calling
i915_gem_load, the shrinker will be left registered. When called, it
will access freed memory and crash. The fix is to unregister the shrinker in the
failure path using code duplicated from i915_driver_unload.

i915_driver_load also has some incorrect gotos in the error cleanup
paths:

 * After failing to initialize the GTT (which cannot happen, btw,
   intel_gtt_get returns a fixed (non-NULL) value), it tries to
   free the uninitialized WC IO mapping. Fixed this by changing the
   target from out_iomapfree to out_rmmap

Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Lin Ming <ming.m.lin@intel.com>