Replace wait_event_interruptible() with wait_event() in o2net_send_message_vec().
This is because this function is called by the dlm that expects signals to be
blocked.

Fixes oss bugzilla#1126
http://oss.oracle.com/bugzilla/show_bug.cgi?id=1126



Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

In ocfs2_add_branch, we use the rightmost rec of the leaf extent block
to generate the e_cpos for the newly added branch. In the most case, it
is OK but if the parent extent block's rightmost rec covers more clusters
than the leaf does, it will cause kernel panic if we insert some clusters
in it. The message is something like:
(7445,1):ocfs2_insert_at_leaf:3775 ERROR: bug expression:
le16_to_cpu(el->l_next_free_rec) >= le16_to_cpu(el->l_count)
(7445,1):ocfs2_insert_at_leaf:3775 ERROR: inode 66053, depth 0, count 28,
next free 28, rec.cpos 270, rec.clusters 1, insert.cpos 275, insert.clusters 1
 [<fa7ad565>] ? ocfs2_do_insert_extent+0xb58/0xda0 [ocfs2]
 [<fa7b08f2>] ? ocfs2_insert_extent+0x5bd/0x6ba [ocfs2]
 [<fa7b1b8b>] ? ocfs2_add_clusters_in_btree+0x37f/0x564 [ocfs2]
...

The panic can be easily reproduced by the following small test case
(with bs=512, cs=4K, and I remove all the error handling so that it looks
clear enough for reading).

int main(int argc, char **argv)
{
	int fd, i;
	char buf[5] = "test";

	fd = open(argv[1], O_RDWR|O_CREAT);

	for (i = 0; i < 30; i++) {
		lseek(fd, 40960 * i, SEEK_SET);
		write(fd, buf, 5);
	}

	ftruncate(fd, 1146880);

	lseek(fd, 1126400, SEEK_SET);
	write(fd, buf, 5);

	close(fd);

	return 0;
}

The reason of the panic is that:
the 30 writes and the ftruncate makes the file's extent list looks like:

	Tree Depth: 1   Count: 19   Next Free Rec: 1
	## Offset        Clusters       Block#
	0  0             280            86183
	SubAlloc Bit: 7   SubAlloc Slot: 0
	Blknum: 86183   Next Leaf: 0
	CRC32: 00000000   ECC: 0000
	Tree Depth: 0   Count: 28   Next Free Rec: 28
	## Offset        Clusters       Block#          Flags
	0  0             1              143368          0x0
	1  10            1              143376          0x0
	...
	26 260           1              143576          0x0
	27 270           1              143584          0x0

Now another write at 1126400(275 cluster) whiich will write at the gap
between 271 and 280 will trigger ocfs2_add_branch, but the result after
the function looks like:
	Tree Depth: 1   Count: 19   Next Free Rec: 2
	## Offset        Clusters       Block#
	0  0             280            86183
	1  271           0             143592
So the extent record is intersected and make the following operation bug out.

This patch just try to remove the gap before we add the new branch, so that
the root(branch) rightmost rec will cover the same right position. So in the
above case, before adding branch the tree will be changed to
	Tree Depth: 1   Count: 19   Next Free Rec: 1
	## Offset        Clusters       Block#
	0  0             271            86183
	SubAlloc Bit: 7   SubAlloc Slot: 0
	Blknum: 86183   Next Leaf: 0
	CRC32: 00000000   ECC: 0000
	Tree Depth: 0   Count: 28   Next Free Rec: 28
	## Offset        Clusters       Block#          Flags
	0  0             1              143368          0x0
	1  10            1              143376          0x0
	...
	26 260           1              143576          0x0
	27 270           1              143584          0x0
And after branch add, the tree looks like
	Tree Depth: 1   Count: 19   Next Free Rec: 2
	## Offset        Clusters       Block#
	0  0             271            86183
	1  271           0             143592

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

In ocfs2, fdatasync and fsync are identical.
I think fdatasync should skip committing transaction when
inode->i_state is set just I_DIRTY_SYNC and this indicates
only atime or/and mtime updates.
Following patch improves fdatasync throughput.

#sysbench --num-threads=16 --max-requests=300000 --test=fileio
--file-block-size=4K --file-total-size=16G --file-test-mode=rndwr
--file-fsync-mode=fdatasync run

Results:
-2.6.30-rc8
Test execution summary:
    total time:                          107.1445s
    total number of events:              119559
    total time taken by event execution: 116.1050
    per-request statistics:
         min:                            0.0000s
         avg:                            0.0010s
         max:                            0.1220s
         approx.  95 percentile:         0.0016s

Threads fairness:
    events (avg/stddev):           7472.4375/303.60
    execution time (avg/stddev):   7.2566/0.64

-2.6.30-rc8-patched
Test execution summary:
    total time:                          86.8529s
    total number of events:              300016
    total time taken by event execution: 24.3077
    per-request statistics:
         min:                            0.0000s
         avg:                            0.0001s
         max:                            0.0336s
         approx.  95 percentile:         0.0001s

Threads fairness:
    events (avg/stddev):           18751.0000/718.75
    execution time (avg/stddev):   1.5192/0.05

Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

It would be nice to know how often we get checksum failures.  Even
better, how many of them we can fix with the single bit ecc.  So, we add
a statistics structure.  The structure can be installed into debugfs
wherever the user wants.

For ocfs2, we'll put it in the superblock-specific debugfs directory and
pass it down from our higher-level functions.  The stats are only
registered with debugfs when the filesystem supports metadata ecc.

Signed-off-by: Joel Becker <joel.becker@oracle.com>

Patch to track delayed orphan scan timer statistics.

Modifies ocfs2_osb_dump to print the following:
  Orphan Scan=> Local: 10  Global: 21  Last Scan: 67 seconds ago

Signed-off-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock
before moving the dentry to the orphan directory. Other nodes that have
this dentry in cache have a PR on the same dentry lock.  When the EX is
requested, the other nodes flag the corresponding inode as MAYBE_ORPHANED
during downconvert.  The inode is finally deleted when the last node to iput
the inode sees that i_nlink==0 and the MAYBE_ORPHANED flag is set.

A problem arises if a node is forced to free dentry locks because of memory
pressure. If this happens, the node will no longer get downconvert
notifications for the dentries that have been unlinked on another node.
If it also happens that node is actively using the corresponding inode and
happens to be the one performing the last iput on that inode, it will fail
to delete the inode as it will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the
orphan directories to delete such inodes. Care has been taken to distribute
the workload across the cluster so that no one node has to perform the task
all the time.

Signed-off-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

We use ordering ip_alloc_sem -> local alloc locks in ocfs2_write_begin().
So change lock ordering in ocfs2_extend_dir() and ocfs2_expand_inline_dir()
to also use this lock ordering.

Signed-off-by: Jan Kara <jack@suse.cz>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

In ocfs2_finish_quota_recovery() we acquired global quota file lock and started
recovering local quota file. During this process we need to get quota
structures, which calls ocfs2_dquot_acquire() which gets global quota file lock
again. This second lock can block in case some other node has requested the
quota file lock in the mean time. Fix the problem by moving quota file locking
down into the function where it is really needed.  Then dqget() or dqput()
won't be called with the lock held.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

We called vfs_dq_transfer() with global quota file lock held. This can lead
to deadlocks as if vfs_dq_transfer() has to allocate new quota structure,
it calls ocfs2_dquot_acquire() which tries to get quota file lock again and
this can block if another node requested the lock in the mean time.

Since we have to call vfs_dq_transfer() with transaction already started
and quota file lock ranks above the transaction start, we cannot just rely
on ocfs2_dquot_acquire() or ocfs2_dquot_release() on getting the lock
if they need it. We fix the problem by acquiring pointers to all quota
structures needed by vfs_dq_transfer() already before calling the function.
By this we are sure that all quota structures are properly allocated and
they can be freed only after we drop references to them. Thus we don't need
quota file lock anywhere inside vfs_dq_transfer().

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

This function is called with dqio_mutex held but it has to acquire lock
from global quota file which ranks above this lock. This is not deadlockable
lock inversion since this code path is take only during mount when noone
else can race with us but let's clean this up to silence lockdep.

We just drop the dqio_mutex in the beginning of the function and reacquire
it in the end since we don't need it - noone can race with us at this moment.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

It is not possible to get a read lock and then try to get the same write lock
in one thread as that can block on downconvert being requested by other node
leading to deadlock. So first drop the quota lock for reading and only after
that get it for writing.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

In the mainline ocfs2 code, the interface for masklog is in files under
/sys/fs/o2cb/masklog, but the comments in fs/ocfs2/cluster/masklog.h
reference the old /proc interface.  They are out of date.

This patch modifies the comments in cluster/masklog.h, which also provides
a bash script example on how to change the log mask bits.

Signed-off-by: Coly Li <coly.li@suse.de>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

Currently the kernel defines XATTR_LIST_MAX as 65536
in include/linux/limits.h.  This is the largest buffer that is used for
listing xattrs.

But with ocfs2 xattr tree, we actually have no limit for the number.  If
filesystem has more names than can fit in the buffer, the kernel
logs will be pollluted with something like this when listing:

(27738,0):ocfs2_iterate_xattr_buckets:3158 ERROR: status = -34
(27738,0):ocfs2_xattr_tree_list_index_block:3264 ERROR: status = -34

So don't print "ERROR" message as this is not an ocfs2 error.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

The ocfs2 directory index updates two blocks when we remove an entry -
the dx root and the dx leaf.  OCFS2_DELETE_INODE_CREDITS was only
accounting for the dx leaf.  This shows up when ocfs2_delete_inode()
runs out of credits in jbd2_journal_dirty_metadata() at
"J_ASSERT_JH(jh, handle->h_buffer_credits > 0);".

The test that caught this was running dirop_file_racer from the
ocfs2-test suite with a 250-character filename PREFIX.  Run on a 512B
blocksize, it forces the orphan dir index to grow large enough to
trigger.

Signed-off-by: Joel Becker <joel.becker@oracle.com>

With indexed dir enabled, now we use ocfs2_dir_lookup_result to
wrap all the bh used for dir. So remove the 2 unused variables.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

In ocfs2_dentry_attach_lock(), if unable to get the dentry lock, we need to
call iput(inode) because a failure here means no d_instantiate(), which means
the normally matching iput() will not be called during dput(dentry).

This patch fixes the oops that accompanies the following message:
(3996,1):dlm_empty_lockres:2708 ERROR: lockres W00000000000000000a1046b06a4382 still has local locks!
kernel BUG in dlm_empty_lockres at /rpmbuild/smushran/BUILD/ocfs2-1.4.2/fs/ocfs2/dlm/dlmmaster.c:2709!

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

The old %llu vs u64 battle.  Cast them correctly.

Signed-off-by: Joel Becker <joel.becker@oracle.com>

fs/ocfs2/dir.c: In function ‘ocfs2_extend_dir’:
fs/ocfs2/dir.c:2700: warning: ‘ret’ may be used uninitialized in this function

fs/ocfs2/suballoc.c: In function ‘ocfs2_get_suballoc_slot_bit’:
fs/ocfs2/suballoc.c:2216: warning: comparison is always true due to limited range of data type

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>

Rearrange locking of i_mutex on destination and call to
ocfs2_rw_lock() so locks are only held while buffers are copied with
the pipe_to_file() actor, and not while waiting for more data on the
pipe.

Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

In ocfs2_expand_inline_dir, we calculate whether we need 1 extra
cluster if we can't store the dx inline the root and save it in
dx_alloc. So add it when we call ocfs2_reserve_clusters.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

There's a possible deadlock in generic_file_splice_write(),
splice_from_pipe() and ocfs2_file_splice_write():

 - task A calls generic_file_splice_write()
 - this calls inode_double_lock(), which locks i_mutex on both
   pipe->inode and target inode
 - ordering depends on inode pointers, can happen that pipe->inode is
   locked first
 - __splice_from_pipe() needs more data, calls pipe_wait()
 - this releases lock on pipe->inode, goes to interruptible sleep
 - task B calls generic_file_splice_write(), similarly to the first
 - this locks pipe->inode, then tries to lock inode, but that is
   already held by task A
 - task A is interrupted, it tries to lock pipe->inode, but fails, as
   it is already held by task B
 - ABBA deadlock

Fix this by explicitly ordering locks: the outer lock must be on
target inode and the inner lock (which is later unlocked and relocked)
must be on pipe->inode.  This is OK, pipe inodes and target inodes
form two nonoverlapping sets, generic_file_splice_write() and friends
are not called with a target which is a pipe.

Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

During recovery, a node recovers orphans in it's slot and the dead node(s). But
if the dead nodes were holding orphans in offline slots, they will be left
unrecovered.

If the dead node is the last one to die and is holding orphans in other slots
and is the first one to mount, then it only recovers it's own slot, which
leaves orphans in offline slots.

This patch queues complete_recovery to clean orphans for all offline slots
during mount and node recovery.

Signed-off-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Acked-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

A page can have multiple buffers and even if a page is not uptodate, some buffers
can be uptodate on pagesize != blocksize environment.
This aops checks that all buffers which correspond to a part of a file
that we want to read are uptodate. If so, we do not have to issue actual
read IO to HDD even if a page is not uptodate because the portion we
want to read are uptodate.
"block_is_partially_uptodate" function is already used by ext2/3/4.
With the following patch random read/write mixed workloads or random read after
random write workloads can be optimized and we can get performance improvement.

Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

For nfs exporting, ocfs2_get_dentry() returns the dentry for fh.
ocfs2_get_dentry() may read from disk when the inode is not in memory,
without any cross cluster lock. this leads to the file system loading a
stale inode.

This patch fixes above problem.

Solution is that in case of inode is not in memory, we get the cluster
lock(PR) of alloc inode where the inode in question is allocated from (this
causes node on which deletion is done sync the alloc inode) before reading
out the inode itsself. then we check the bitmap in the group (the inode in
question allcated from) to see if the bit is clear. if it's clear then it's
stale. if the bit is set, we then check generation as the existing code
does.

We have to read out the inode in question from disk first to know its alloc
slot and allot bit. And if its not stale we read it out using ocfs2_iget().
The second read should then be from cache.

And also we have to add a per superblock nfs_sync_lock to cover the lock for
alloc inode and that for inode in question. this is because ocfs2_get_dentry()
and ocfs2_delete_inode() lock on them in reverse order. nfs_sync_lock is locked
in EX mode in ocfs2_get_dentry() and in PR mode in ocfs2_delete_inode(). so
that mutliple ocfs2_delete_inode() can run concurrently in normal case.

[mfasheh@suse.com: build warning fixes and comment cleanups]
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Acked-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

The debugfs file, mle_state, now prints the number of largest number of mles
in one hash link.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch attempts to fix a fine race between purging and migration.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch removes struct dlm_lock_name and adds the entries directly
to struct dlm_master_list_entry. Under the new scheme, both mles that
are backed by a lockres or not, will have the name populated in mle->mname.
This allows us to get rid of code that was figuring out the location of
the mle name.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch shows the number of lockres' and mles in the debugfs file, dlm_state.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch inlines dlm_set_lockres_owner() and dlm_change_lockres_owner().

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch replaces the lockres counts that tracked the number number of
locally and remotely mastered lockres' with a current and total count. The
total count is the number of lockres' that have been created since the dlm
domain was created.

The number of locally and remotely mastered counts can be computed using
the locking_state output.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

The lifetime of a mle is limited to the duration of the lockres mastery
process. While typically this lifetime is fairly short, we have noticed
the number of mles explode under certain circumstances. This patch tracks
the number of each different types of mles and should help us determine
how best to speed up the mastery process.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

The previous patch explicitly did not indent dlm_cleanup_master_list()
so as to make the patch readable. This patch properly indents the
function.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

With this patch, the mles are stored in a hash and not a simple list.
This should improve the mle lookup time when the number of outstanding
masteries is large.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch adds code to create and destroy the dlm->master_hash.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch refactors dlm_clean_master_list() so as to make it
easier to convert the mle list to a hash.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

For master mle, the name it stored in the attached lockres in struct qstr.
For block and migration mle, the name is stored inline in struct dlm_lock_name.
This patch attempts to make struct dlm_lock_name look like a struct qstr. While
we could use struct qstr, we don't because we want to avoid having to malloc
and free the lockname string as the mle's lifetime is fairly short.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

This patch encapsulates adding and removing of the mle from the
dlm->master_list. This patch is part of the series of patches that
converts the mle list to a mle hash.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

In ocfs2, the block group search looks for the "emptiest" group
to allocate from. So if the allocator has many equally(or almost
equally) empty groups, new block group will tend to get spread
out amongst them.

So we add osb_inode_alloc_group in ocfs2_super to record the last
used inode allocation group.
For more details, please see
http://oss.oracle.com/osswiki/OCFS2/DesignDocs/InodeAllocationStrategy

.

I have done some basic test and the results are a ten times improvement on
some cold-cache stat workloads.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

Inode groups used to be allocated from local alloc file,
but since we want all inodes to be contiguous enough, we
will try to allocate them directly from global_bitmap.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>