For debug use, we can see from the log whether the fence decision is
made and why it is not fenced.
Signed-off-by: NJunxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: NSrinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: NMark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When tcp retransmit timeout(15mins), the connection will be closed.
Pending messages may be lost during this time.  So we set tcp user
timeout to override the retransmit timeout to the max value.  This is OK
for ocfs2 since we have disk heartbeat, if peer crash, the disk
heartbeat will timeout and it will be evicted, if disk heartbeat not
timeout and connection idle for a long time, then this means the cluster
enters split-brain state, since fence can't happen, we'd better keep the
connection and wait network recover.
Signed-off-by: NJunxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: NSrinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: NMark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This patch series is to fix a possible message lost bug in ocfs2 when
network go bad.  This bug will cause ocfs2 hung forever even network
become good again.

The messages may lost in this case.  After the tcp connection is
established between two nodes, an idle timer will be set to check its
state periodically, if no messages are received during this time, idle
timer will timeout, it will shutdown the connection and try to
reconnect, so pending messages in tcp queues will be lost.  This
messages may be from dlm.  Dlm may get hung in this case.  This may
cause the whole ocfs2 cluster hung.

This is very possible to happen when network state goes bad.  Do the
reconnect is useless, it will fail if network state is still bad.  Just
waiting there for network recovering may be a good idea, it will not
lost messages and some node will be fenced until cluster goes into
split-brain state, for this case, Tcp user timeout is used to override
the tcp retransmit timeout.  It will timeout after 25 days, user should
have notice this through the provided log and fix the network, if they
don't, ocfs2 will fall back to original reconnect way.

This patch (of 3):

Some messages in the tcp queue maybe lost if we shutdown the connection
and reconnect when idle timeout.  If packets lost and reconnect success,
then the ocfs2 cluster maybe hung.

To fix this, we can leave the connection there and do the fence decision
when idle timeout, if network recover before fence dicision is made, the
connection survive without lost any messages.

This bug can be saw when network state go bad.  It may cause ocfs2 hung
forever if some packets lost.  With this fix, ocfs2 will recover from
hung if network becomes good again.
Signed-off-by: NJunxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: NSrinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: NMark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

If we failed to copy from the structure, writing back the flags leaks 31
bits of kernel memory (the rest of the ir_flags field).

In any case, if we cannot copy from/to the structure, why should we
expect putting just the flags to work?

Also make sure ocfs2_info_handle_freeinode() returns the right error
code if the copy_to_user() fails.

Fixes: ddee5cdb ('Ocfs2: Add new OCFS2_IOC_INFO ioctl for ocfs2 v8.')
Signed-off-by: NBen Hutchings <ben@decadent.org.uk>
Cc: Joel Becker <jlbec@evilplan.org>
Acked-by: NMark Fasheh <mfasheh@suse.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When performing a same-directory rename, it's possible that adding or
setting the new directory entry will cause the directory to overflow
the inline data area, which causes the directory to be converted to an
extent-based directory.  Under this circumstance it is necessary to
re-read the directory when deleting the old dirent because the "old
directory" context still points to i_block in the inode table, which
is now an extent tree root!  The delete fails with an FS error, and
the subsequent fsck complains about incorrect link counts and
hardlinked directories.

Test case (originally found with flat_dir_test in the metadata_csum
test program):

# mkfs.ext4 -O inline_data /dev/sda
# mount /dev/sda /mnt
# mkdir /mnt/x
# touch /mnt/x/changelog.gz /mnt/x/copyright /mnt/x/README.Debian
# sync
# for i in /mnt/x/*; do mv $i $i.longer; done
# ls -la /mnt/x/
total 0
-rw-r--r-- 1 root root 0 Aug 25 12:03 changelog.gz.longer
-rw-r--r-- 1 root root 0 Aug 25 12:03 copyright
-rw-r--r-- 1 root root 0 Aug 25 12:03 copyright.longer
-rw-r--r-- 1 root root 0 Aug 25 12:03 README.Debian.longer

(Hey!  Why are there four files now??)
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

It turns out that there are some serious problems with the on-disk
format of journal checksum v2.  The foremost is that the function to
calculate descriptor tag size returns sizes that are too big.  This
causes alignment issues on some architectures and is compounded by the
fact that some parts of jbd2 use the structure size (incorrectly) to
determine the presence of a 64bit journal instead of checking the
feature flags.

Therefore, introduce journal checksum v3, which enlarges the
descriptor block tag format to allow for full 32-bit checksums of
journal blocks, fix the journal tag function to return the correct
sizes, and fix the jbd2 recovery code to use feature flags to
determine 64bitness.

Add a few function helpers so we don't have to open-code quite so
many pieces.

Switching to a 16-byte block size was found to increase journal size
overhead by a maximum of 0.1%, to convert a 32-bit journal with no
checksumming to a 32-bit journal with checksum v3 enabled.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reported-by: NTR Reardon <thomas_reardon@hotmail.com>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

When recovering the journal, don't fall into an infinite loop if we
encounter a corrupt journal block.  Instead, just skip the block and
return an error, which fails the mount and thus forces the user to run
a full filesystem fsck.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

In case of delalloc block i_disksize may be less than i_size. So we
have to update i_disksize each time we allocated and submitted some
blocks beyond i_disksize.  We weren't doing this on the error paths,
so fix this.

testcase: xfstest generic/019
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

After commit f282ac19 we use different transactions for
preallocation and i_disksize update which result in complain from fsck
after power-failure.  spotted by generic/019. IMHO this is regression
because fs becomes inconsistent, even more 'e2fsck -p' will no longer
works (which drives admins go crazy) Same transaction requirement
applies ctime,mtime updates

testcase: xfstest generic/019
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

Currently we reserve only 4 blocks but in worst case scenario
ext4_zero_partial_blocks() may want to zeroout and convert two
non adjacent blocks.
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

When creating a new object on the NFS server, we should not be sending
posix setacl requests unless the preceding posix_acl_create returned a
non-trivial acl. Doing so, causes Solaris servers in particular to
return an EINVAL.

Fixes: 013cdf10 (nfs: use generic posix ACL infrastructure,,,)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1132786
Cc: stable@vger.kernel.org # 3.14+
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

If we did an OPEN_DOWNGRADE, then the right thing to do on success, is
to apply the new open mode to the struct nfs4_state. Instead, we were
unconditionally clearing the state, making it appear to our state
machinery as if we had just performed a CLOSE.

Fixes: 226056c5 (NFSv4: Use correct locking when updating nfs4_state...)
Cc: stable@vger.kernel.org # 3.15+
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

In the presence of delegations, we can no longer assume that the
state->n_rdwr, state->n_rdonly, state->n_wronly reflect the open
stateid share mode, and so we need to calculate the initial value
for calldata->arg.fmode using the state->flags.
Reported-by: NJames Drews <drews@engr.wisc.edu>
Fixes: 88069f77 (NFSv41: Fix a potential state leakage when...)
Cc: stable@vger.kernel.org # 2.6.33+
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

As reported by Dan Aloni, commit f8567a38 ("aio: fix aio request
leak when events are reaped by userspace") introduces a regression when
user code attempts to perform io_submit() with more events than are
available in the ring buffer.  Reverting that commit would reintroduce a
regression when user space event reaping is used.

Fixing this bug is a bit more involved than the previous attempts to fix
this regression.  Since we do not have a single point at which we can
count events as being reaped by user space and io_getevents(), we have
to track event completion by looking at the number of events left in the
event ring.  So long as there are as many events in the ring buffer as
there have been completion events generate, we cannot call
put_reqs_available().  The code to check for this is now placed in
refill_reqs_available().

A test program from Dan and modified by me for verifying this bug is available
at http://www.kvack.org/~bcrl/20140824-aio_bug.c .
Reported-by: NDan Aloni <dan@kernelim.com>
Signed-off-by: NBenjamin LaHaise <bcrl@kvack.org>
Acked-by: NDan Aloni <dan@kernelim.com>
Cc: Kent Overstreet <kmo@daterainc.com>
Cc: Mateusz Guzik <mguzik@redhat.com>
Cc: Petr Matousek <pmatouse@redhat.com>
Cc: stable@vger.kernel.org      # v3.16 and anything that f8567a38 was backported to
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This has been reported and discussed for a long time, and this hang occurs in
both 3.15 and 3.16.

Btrfs now migrates to use kernel workqueue, but it introduces this hang problem.

Btrfs has a kind of work queued as an ordered way, which means that its
ordered_func() must be processed in the way of FIFO, so it usually looks like --

normal_work_helper(arg)
    work = container_of(arg, struct btrfs_work, normal_work);

    work->func() <---- (we name it work X)
    for ordered_work in wq->ordered_list
            ordered_work->ordered_func()
            ordered_work->ordered_free()

The hang is a rare case, first when we find free space, we get an uncached block
group, then we go to read its free space cache inode for free space information,
so it will

file a readahead request
    btrfs_readpages()
         for page that is not in page cache
                __do_readpage()
                     submit_extent_page()
                           btrfs_submit_bio_hook()
                                 btrfs_bio_wq_end_io()
                                 submit_bio()
                                 end_workqueue_bio() <--(ret by the 1st endio)
                                      queue a work(named work Y) for the 2nd
                                      also the real endio()

So the hang occurs when work Y's work_struct and work X's work_struct happens
to share the same address.

A bit more explanation,

A,B,C -- struct btrfs_work
arg   -- struct work_struct

kthread:
worker_thread()
    pick up a work_struct from @worklist
    process_one_work(arg)
	worker->current_work = arg;  <-- arg is A->normal_work
	worker->current_func(arg)
		normal_work_helper(arg)
		     A = container_of(arg, struct btrfs_work, normal_work);

		     A->func()
		     A->ordered_func()
		     A->ordered_free()  <-- A gets freed

		     B->ordered_func()
			  submit_compressed_extents()
			      find_free_extent()
				  load_free_space_inode()
				      ...   <-- (the above readhead stack)
				      end_workqueue_bio()
					   btrfs_queue_work(work C)
		     B->ordered_free()

As if work A has a high priority in wq->ordered_list and there are more ordered
works queued after it, such as B->ordered_func(), its memory could have been
freed before normal_work_helper() returns, which means that kernel workqueue
code worker_thread() still has worker->current_work pointer to be work
A->normal_work's, ie. arg's address.

Meanwhile, work C is allocated after work A is freed, work C->normal_work
and work A->normal_work are likely to share the same address(I confirmed this
with ftrace output, so I'm not just guessing, it's rare though).

When another kthread picks up work C->normal_work to process, and finds our
kthread is processing it(see find_worker_executing_work()), it'll think
work C as a collision and skip then, which ends up nobody processing work C.

So the situation is that our kthread is waiting forever on work C.

Besides, there're other cases that can lead to deadlock, but the real problem
is that all btrfs workqueue shares one work->func, -- normal_work_helper,
so this makes each workqueue to have its own helper function, but only a
wraper pf normal_work_helper.

With this patch, I no long hit the above hang.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NChris Mason <clm@fb.com>

Cc: stable@vger.kernel.org # needed for bug fix patches
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>

If we suffer a block allocation failure (for example due to a memory
allocation failure), it's possible that we will call
ext4_discard_allocated_blocks() before we've actually allocated any
blocks.  In that case, fe_len and fe_start in ac->ac_f_ex will still
be zero, and this will result in mb_free_blocks(inode, e4b, 0, 0)
triggering the BUG_ON on mb_free_blocks():

	BUG_ON(last >= (sb->s_blocksize << 3));

Fix this by bailing out of ext4_discard_allocated_blocks() if fs_len
is zero.

Also fix a missing ext4_mb_unload_buddy() call in
ext4_discard_allocated_blocks().

Google-Bug-Id: 16844242

Fixes: 86f0afd4Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

If we run into some kind of error, such as ENOMEM, while calling
ext4_getblk() or ext4_dx_find_entry(), we need to make sure this error
gets propagated up to ext4_find_entry() and then to its callers.  This
way, transient errors such as ENOMEM can get propagated to the VFS.
This is important so that the system calls return the appropriate
error, and also so that in the case of ext4_lookup(), we return an
error instead of a NULL inode, since that will result in a negative
dentry cache entry that will stick around long past the OOM condition
which caused a transient ENOMEM error.

Google-Bug-Id: #17142205
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

If a SIGKILL is sent to a task waiting in __nfs_iocounter_wait,
it will busy-wait or soft lockup in its while loop.
nfs_wait_bit_killable won't sleep, and the loop won't exit on
the error return.

Stop the busy-wait by breaking out of the loop when
nfs_wait_bit_killable returns an error.
Signed-off-by: NDavid Jeffery <djeffery@redhat.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Commit 6094f838
"nfs: allow coalescing of subpage requests" got rid of the requirement
that requests cover whole pages, but it made some incorrect assumptions.

It turns out that callers of this interface can map adjacent requests
(by file position as seen by req_offset + req->wb_bytes) to different pages,
even when they could share a page. An example is the direct I/O interface -
iov_iter_get_pages_alloc may return one segment with a partial page filled
and the next segment (which is adjacent in the file position) starts with a
new page.
Reported-by: NToralf Förster <toralf.foerster@gmx.de>
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Adjacent requests that share the same page are allowed, but should only
use one entry in the page vector. This avoids overruning the page
vector - it is sized based on how many bytes there are, not by
request count.

This fixes issues that manifest as "Redzone overwritten" bugs (the
vector overrun) and hangs waiting on page read / write, as it waits on
the same page more than once.

This also adds bounds checking to the page vector with a graceful failure
(WARN_ON_ONCE and pgio error returned to application).
Reported-by: NToralf Förster <toralf.foerster@gmx.de>
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

This handles the 'nonblock=false' case in nfs_lock_and_join_requests.
If the group is already locked and blocking is allowed, drop the inode lock
and wait for the group lock to be cleared before trying it all again.
This should fix warnings found in peterz's tree (sched/wait branch), where
might_sleep() checks are added to wait.[ch].
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

This fixes handling of errors from nfs_page_group_lock in
nfs_lock_and_join_requests.  It now releases the inode lock and the
reference to the head request.
Reported-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

__nfs_pageio_add_request was calling nfs_page_group_lock nonblocking, but
this can return -EAGAIN which would end up passing -EIO to the application.

There is no reason not to block in this path, so change the two calls to
do so. Also, there is no need to check the return value of
nfs_page_group_lock when nonblock=false, so remove the error handling code.
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

nfs_page_group_lock was calling wait_on_bit_lock even when told not to
block. Fix by first trying test_and_set_bit, followed by wait_on_bit_lock
if and only if blocking is allowed.  Return -EAGAIN if nonblocking and the
test_and_set of the bit was already locked.
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Flip the meaning of the second argument from 'wait' to 'nonblock' to
match related functions. Update all five calls to reflect this change.
Signed-off-by: NWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: NPeng Tao <tao.peng@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

We should only be flushing on close if the file was flagged as needing
it during truncate.  I broke this with my ordered data vs transaction
commit deadlock fix.

Thanks to Miao Xie for catching this.
Signed-off-by: NChris Mason <clm@fb.com>
Reported-by: NMiao Xie <miaox@cn.fujitsu.com>
Reported-by: NFengguang Wu <fengguang.wu@intel.com>

The crash is

------------[ cut here ]------------
kernel BUG at fs/btrfs/extent_io.c:2124!
[...]
Workqueue: btrfs-endio normal_work_helper [btrfs]
RIP: 0010:[<ffffffffa02d6055>]  [<ffffffffa02d6055>] end_bio_extent_readpage+0xb45/0xcd0 [btrfs]

This is in fact a regression.

It is because we forgot to increase @offset properly in reading corrupted block,
so that the @offset remains, and this leads to checksum errors while reading
left blocks queued up in the same bio, and then ends up with hiting the above
BUG_ON.
Reported-by: NChris Murphy <lists@colorremedies.com>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NChris Mason <clm@fb.com>

Coverity pointed this out; in the newly added
qgroup_subtree_accounting(), if btrfs_find_all_roots()
returns an error, we leak at least the parents pointer,
and possibly the roots pointer, depending on what failure
occurs.

If btrfs_find_all_roots() returns an error, we need to
free up all allocations before we return.  "roots" is
initialized to NULL, so it should be safe to free
it unconditionally (ulist_free() handles that case).

Cc: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Reviewed-by: NMark Fasheh <mfasheh@suse.de>
Signed-off-by: NChris Mason <clm@fb.com>

When current btrfs finds that a new extent map is going to be insereted
but failed with -EEXIST, it will try again to insert the extent map
but with the length of sectorsize.
This is OK if we don't enable 'no-holes' feature since all extent space
is continuous, we will not go into the not found->insert routine.

But if we enable 'no-holes' feature, it will make things out of control.
e.g. in 4K sectorsize, we pass the following args to btrfs_get_extent():
btrfs_get_extent() args: start:  27874 len 4100
28672		  27874		28672	27874+4100	32768
                    |-----------------------|
|---------hole--------------------|---------data----------|

1) not found and insert
Since no extent map containing the range, btrfs_get_extent() will go
into the not_found and insert routine, which will try to insert the
extent map (27874, 27847 + 4100).

2) first overlap
But it overlaps with (28672, 32768) extent, so -EEXIST will be returned
by add_extent_mapping().

3) retry but still overlap
After catching the -EEXIST, then btrfs_get_extent() will try insert it
again but with 4K length, which still overlaps, so -EEXIST will be
returned.

This makes the following patch fail to punch hole.
d7781546 btrfs: Avoid trucating page or punching hole in a already existed hole.

This patch will use the right length, which is the (exsisting->start -
em->start) to insert, making the above patch works in 'no-holes' mode.
Also, some small code style problems in above patch is fixed too.
Reported-by: NFilipe David Manana <fdmanana@gmail.com>
Signed-off-by: NQu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: NFilipe David Manana <fdmanana@suse.com>
Tested-by: NFilipe David Manana <fdmanana@suse.com>
Signed-off-by: NChris Mason <clm@fb.com>

When cloning a file that consists of an inline extent, we were creating
an extent map that represents a non-existing trailing hole starting at a
file offset that isn't a multiple of the sector size. This happened because
when processing an inline extent we weren't aligning the extent's length to
the sector size, and therefore incorrectly treating the range
[inline_extent_length; sector_size[ as a hole.
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Reviewed-by: NSatoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: NChris Mason <clm@fb.com>

If an inode has a very large number of extent maps, we can spend
a lot of time freeing them, which triggers a soft lockup warning.
Therefore reschedule if we need to when freeing the extent maps
while evicting the inode.

I could trigger this all the time by running xfstests/generic/299 on
a file system with the no-holes feature enabled. That test creates
an inode with 11386677 extent maps.

    $ mkfs.btrfs -f -O no-holes $TEST_DEV
    $ MKFS_OPTIONS="-O no-holes" ./check generic/299
    generic/299 382s ...
    Message from syslogd@debian-vm3 at Aug  7 10:44:29 ...
     kernel:[85304.208017] BUG: soft lockup - CPU#0 stuck for 22s! [umount:25330]
     384s
    Ran: generic/299
    Passed all 1 tests

    $ dmesg
    (...)
    [86304.300017] BUG: soft lockup - CPU#0 stuck for 23s! [umount:25330]
    (...)
    [86304.300036] Call Trace:
    [86304.300036]  [<ffffffff81698ba9>] __slab_free+0x54/0x295
    [86304.300036]  [<ffffffffa02ee9cc>] ? free_extent_map+0x5c/0xb0 [btrfs]
    [86304.300036]  [<ffffffff811a6cd2>] kmem_cache_free+0x282/0x2a0
    [86304.300036]  [<ffffffffa02ee9cc>] free_extent_map+0x5c/0xb0 [btrfs]
    [86304.300036]  [<ffffffffa02e3775>] btrfs_evict_inode+0xd5/0x660 [btrfs]
    [86304.300036]  [<ffffffff811e7c8d>] ? __inode_wait_for_writeback+0x6d/0xc0
    [86304.300036]  [<ffffffff816a389b>] ? _raw_spin_unlock+0x2b/0x40
    [86304.300036]  [<ffffffff811d8cbb>] evict+0xab/0x180
    [86304.300036]  [<ffffffff811d8dce>] dispose_list+0x3e/0x60
    [86304.300036]  [<ffffffff811d9b04>] evict_inodes+0xf4/0x110
    [86304.300036]  [<ffffffff811bd953>] generic_shutdown_super+0x53/0x110
    [86304.300036]  [<ffffffff811bdaa6>] kill_anon_super+0x16/0x30
    [86304.300036]  [<ffffffffa02a78ba>] btrfs_kill_super+0x1a/0xa0 [btrfs]
    [86304.300036]  [<ffffffff811bd3a9>] deactivate_locked_super+0x59/0x80
    [86304.300036]  [<ffffffff811be44e>] deactivate_super+0x4e/0x70
    [86304.300036]  [<ffffffff811dec14>] mntput_no_expire+0x174/0x1f0
    [86304.300036]  [<ffffffff811deab7>] ? mntput_no_expire+0x17/0x1f0
    [86304.300036]  [<ffffffff811e0517>] SyS_umount+0x97/0x100
    (...)
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Reviewed-by: NSatoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Tested-by: NSatoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: NChris Mason <clm@fb.com>

The file hole detection logic during a file fsync wasn't correct,
because it didn't look back (in a previous leaf) for the last file
extent item that can be in a leaf to the left of our leaf and that
has a generation lower than the current transaction id. This made it
assume that a hole exists when it really doesn't exist in the file.

Such false positive hole detection happens in the following scenario:

* We have a file that has many file extent items, covering 3 or more
  btree leafs (the first leaf must contain non file extent items too).

* Two ranges of the file are modified, with their extent items being
  located at 2 different leafs and those leafs aren't consecutive.

* When processing the second modified leaf, we weren't checking if
  some file extent item exists that is located in some leaf that is
  between our 2 modified leafs, and therefore assumed the range defined
  between the last file extent item in the first leaf and the first file
  extent item in the second leaf matched a hole.

Fortunately this didn't result in overriding the log with wrong data,
instead it made the last loop in copy_items() attempt to insert a
duplicated key (for a hole file extent item), which makes the file
fsync code return with -EEXIST to file.c:btrfs_sync_file() which in
turn ends up doing a full transaction commit, which is much more
expensive then writing only to the log tree and wait for it to be
durably persisted (as well as the file's modified extents/pages).
Therefore fix the hole detection logic, so that we don't pay the
cost of doing full transaction commits.

I could trigger this issue with the following test for xfstests (which
never fails, either without or with this patch). The last fsync call
results in a full transaction commit, due to the -EEXIST error mentioned
above. I could also observe this behaviour happening frequently when
running xfstests/generic/075 in a loop.

Test:

    _cleanup()
    {
        _cleanup_flakey
        rm -fr $tmp
    }

    # get standard environment, filters and checks
    . ./common/rc
    . ./common/filter
    . ./common/dmflakey

    # real QA test starts here
    _supported_fs btrfs
    _supported_os Linux
    _require_scratch
    _require_dm_flakey
    _need_to_be_root

    rm -f $seqres.full

    # Create a file with many file extent items, each representing a 4Kb extent.
    # These items span 3 btree leaves, of 16Kb each (default mkfs.btrfs leaf size
    # as of btrfs-progs 3.12).
    _scratch_mkfs -l 16384 >/dev/null 2>&1
    _init_flakey
    SAVE_MOUNT_OPTIONS="$MOUNT_OPTIONS"
    MOUNT_OPTIONS="$MOUNT_OPTIONS -o commit=999"
    _mount_flakey

    # First fsync, inode has BTRFS_INODE_NEEDS_FULL_SYNC flag set.
    $XFS_IO_PROG -f -c "pwrite -S 0x01 -b 4096 0 4096" -c "fsync" \
            $SCRATCH_MNT/foo | _filter_xfs_io

    # For any of the following fsync calls, inode doesn't have the flag
    # BTRFS_INODE_NEEDS_FULL_SYNC set.
    for ((i = 1; i <= 500; i++)); do
        OFFSET=$((4096 * i))
        LEN=4096
        $XFS_IO_PROG -c "pwrite -S 0x01 $OFFSET $LEN" -c "fsync" \
                $SCRATCH_MNT/foo | _filter_xfs_io
    done

    # Commit transaction and bump next transaction's id (to 7).
    sync

    # Truncate will set the BTRFS_INODE_NEEDS_FULL_SYNC flag in the btrfs's
    # inode runtime flags.
    $XFS_IO_PROG -c "truncate 2048000" $SCRATCH_MNT/foo

    # Commit transaction and bump next transaction's id (to 8).
    sync

    # Touch 1 extent item from the first leaf and 1 from the last leaf. The leaf
    # in the middle, containing only file extent items, isn't touched. So the
    # next fsync, when calling btrfs_search_forward(), won't visit that middle
    # leaf. First and 3rd leaf have now a generation with value 8, while the
    # middle leaf remains with a generation with value 6.
    $XFS_IO_PROG \
        -c "pwrite -S 0xee -b 4096 0 4096" \
        -c "pwrite -S 0xff -b 4096 2043904 4096" \
        -c "fsync" \
        $SCRATCH_MNT/foo | _filter_xfs_io

    _load_flakey_table $FLAKEY_DROP_WRITES
    md5sum $SCRATCH_MNT/foo | _filter_scratch
    _unmount_flakey

    _load_flakey_table $FLAKEY_ALLOW_WRITES
    # During mount, we'll replay the log created by the fsync above, and the file's
    # md5 digest should be the same we got before the unmount.
    _mount_flakey
    md5sum $SCRATCH_MNT/foo | _filter_scratch
    _unmount_flakey
    MOUNT_OPTIONS="$SAVE_MOUNT_OPTIONS"

    status=0
    exit
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Signed-off-by: NChris Mason <clm@fb.com>

If we open a file with O_TMPFILE, don't do any further operation on
it (so that the inode item isn't updated) and then force a transaction
commit, we get a persisted inode item with a link count of 1, and not 0
as it should be.

Steps to reproduce it (requires a modern xfs_io with -T support):

    $ mkfs.btrfs -f /dev/sdd
    $ mount -o /dev/sdd /mnt
    $ xfs_io -T /mnt &
    $ sync

Then btrfs-debug-tree shows the inode item with a link count of 1:

    $ btrfs-debug-tree /dev/sdd
    (...)
    fs tree key (FS_TREE ROOT_ITEM 0)
    leaf 29556736 items 4 free space 15851 generation 6 owner 5
    fs uuid f164d01b-1b92-481d-a4e4-435fb0f843d0
    chunk uuid 0e3d0e56-bcca-4a1c-aa5f-cec2c6f4f7a6
    	item 0 key (256 INODE_ITEM 0) itemoff 16123 itemsize 160
		inode generation 3 transid 6 size 0 block group 0 mode 40755 links 1
    	item 1 key (256 INODE_REF 256) itemoff 16111 itemsize 12
    		inode ref index 0 namelen 2 name: ..
    	item 2 key (257 INODE_ITEM 0) itemoff 15951 itemsize 160
    		inode generation 6 transid 6 size 0 block group 0 mode 100600 links 1
    	item 3 key (ORPHAN ORPHAN_ITEM 257) itemoff 15951 itemsize 0
		orphan item
    checksum tree key (CSUM_TREE ROOT_ITEM 0)
    (...)
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Signed-off-by: NChris Mason <clm@fb.com>

This is a better solution for the problem addressed in the following
commit:

    Btrfs: update commit root on snapshot creation after orphan cleanup
    (3821f348)

The previous solution wasn't the best because of 2 reasons:

    1) It added another full transaction commit, which is more expensive
       than just swapping the commit root with the root;

    2) If a reboot happened after the first transaction commit (the one
       that creates the snapshot) and before the second transaction commit,
       then we would end up with the same problem if a send using that
       snapshot was requested before the first transaction commit after
       the reboot.

This change addresses those 2 issues. The second issue is addressed by
switching the commit root in the dentry lookup VFS callback, which is
also called by the snapshot/subvol creation ioctl and performs orphan
cleanup if needed. Like the vfs, the ioctl locks the parent inode too,
preventing race issues between a dentry lookup and snapshot creation.

Cc: Alex Lyakas <alex.btrfs@zadarastorage.com>
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Signed-off-by: NChris Mason <clm@fb.com>

Commit 49c6f736(
btrfs: dev replace should replace the sysfs entry) added the missing sysfs entry
in the process of device replace, but didn't take missing devices into account,
so now we have

BUG: unable to handle kernel NULL pointer dereference at 0000000000000088
IP: [<ffffffffa0268551>] btrfs_kobj_rm_device+0x21/0x40 [btrfs]
...

To reproduce it,
1. mkfs.btrfs -f disk1 disk2
2. mkfs.ext4 disk1
3. mount disk2 /mnt -odegraded
4. btrfs replace start -B 1 disk3 /mnt
--------------------------

This fixes the problem.
Reported-by: NChris Murphy <lists@colorremedies.com>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Reviewed-by: NSatoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Tested-by: NSatoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: NChris Mason <clm@fb.com>

The journal blocks of external journal device should not
be counted as overhead.
Signed-off-by: NChin-Tsung Cheng <chintzung@gmail.com>
Signed-off-by: NJan Kara <jack@suse.cz>

We did not check relocated directory in any way when processing Rock
Ridge 'CL' tag. Thus a corrupted isofs image can possibly have a CL
entry pointing to another CL entry leading to possibly unbounded
recursion in kernel code and thus stack overflow or deadlocks (if there
is a loop created from CL entries).

Fix the problem by not allowing CL entry to point to a directory entry
with CL entry (such use makes no good sense anyway) and by checking
whether CL entry doesn't point to itself.

CC: stable@vger.kernel.org
Reported-by: NChris Evans <cevans@google.com>
Signed-off-by: NJan Kara <jack@suse.cz>

We have released the ->i_data_sem before invoking udf_add_entry(),
so in following error path, we should not release this lock again.
Signed-off-by: NChao Yu <chao2.yu@samsung.com>
Signed-off-by: NJan Kara <jack@suse.cz>

The original code allocated new chunks by the number of the writable devices
and missing devices to make sure that any RAID levels on a degraded FS continue
to be honored, but it introduced a problem that it stopped us to allocating
new chunks, the steps to reproduce is following:

 # mkfs.btrfs -m raid1 -d raid1 -f <dev0> <dev1>
 # mkfs.btrfs -f <dev1>			//Removing <dev1> from the original fs
 # mount -o degraded <dev0> <mnt>
 # dd if=/dev/null of=<mnt>/tmpfile bs=1M

It is because we allocate new chunks only on the writable devices, if we take
the number of missing devices into account, and want to allocate new chunks
with higher RAID level, we will fail becaue we don't have enough writable
device. Fix it by ignoring the number of missing devices when allocating
new chunks.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <clm@fb.com>