just delay __put_super() a bit
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

If sget() finds a matching superblock being set up, it'll
grab an active reference to it and grab s_umount.  That's
fine - we'll wait for completion of foofs_get_sb() that way.
However, if said foofs_get_sb() fails we'll end up holding
the halfway-created superblock.  deactivate_locked_super()
called by foofs_get_sb() will just unlock the sucker since
we are holding another active reference to it.

What we need is a way to tell if superblock has been successfully
set up.  Unfortunately, neither ->s_root nor the check for
MS_ACTIVE quite fit.  Cheap and easy way, suitable for backport:
new flag set by the (only) caller of ->get_sb().  If that flag
isn't present by the time sget() grabbed s_umount on preexisting
superblock it has found, it's seeing a stillborn and should
just bury it with deactivate_locked_super() (and repeat the search).

Longer term we want to set that flag in ->get_sb() instances (and
check for it to distinguish between "sget() found us a live sb"
and "sget() has allocated an sb, we need to set it up" in there,
instead of checking ->s_root as we do now).
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Cc: stable@kernel.org

Fix an obscure AB-BA deadlock in get_sb_bdev().

When a superblock is mounted more than once get_sb_bdev() calls
close_bdev_exclusive() to drop the extra bdev reference while holding
s_umount.  However, sb->s_umount nests inside bd_mutex during
__invalidate_device() and close_bdev_exclusive() acquires bd_mutex during
blkdev_put(); thus creating an AB-BA deadlock.

This condition doesn't trigger frequently.  For this condition to be
visible to lockdep, the filesystem must occupy the whole device (as
__invalidate_device() only grabs bd_mutex for the whole device), the FS
must be mounted more than once and partition rescan should be issued while
the FS is still mounted.

Fix it by dropping s_umount over close_bdev_exclusive().
Signed-off-by: NTejun Heo <tj@kernel.org>
Reported-by: NCiprian Docan <docan@eden.rutgers.edu>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: NJens Axboe <axboe@kernel.dk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

No need to mark the superblock as dirty in sysv_remount, synchronize
it instead (only if mounting R/O).

I did not find any docs about this file-system, and I have no possibility
to test my changes. Thus, this is untested. I see other issues in sysv,
e.g., why sysv_sync_fs writes only in the FSTYPE_SYSV4 case? However,
it marks its SB bh's dirty for all types, and does not wait for them
ever. With zero docs I'm unable to fix this.
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

I did not find any docs about this file-system, and I have no possibility
to test my changes. Thus, this is untested.
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

BTRFS does not define a '->write_super()' method, so it should
not mark its superblock as dirty. This looks like some left-over.
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Acked-by: NChris Mason <chris.mason@oracle.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

BFS is a very simple FS and its superblocks contains only static
information and is never changed. However, the BFS code for some
misterious reasons marked its buffer head as dirty from time to
time, but nothing in that buffer was ever changed.

This patch removes all the BFS superblock manipulation, simply
because it is not needed. It removes:

1. The si_sbh filed from 'struct bfs_sb_info' because it is not
   needed. We only need to read the SB once on mount to get the
   start of data blocks and the FS size. After this, we can forget
   about the SB.
2. All instances of 'mark_buffer_dirty(sbh)' for BFS SB because
   it is never changed.
3. The '->sync_fs()' method because there is nothing to sync
   (inodes are synched by VFS).
4. The '->write_super()' method, again, because the SB is never
   changed.
Tested-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

AFFS does not ever wait for superblock synchronization in
->put_super(), ->write_super, and ->sync_fs().

However, it should wait for synchronization in ->put_super() because
it is about to be unmounted, in ->write_super() because this is
periodic SB synchronization performed from a separate kernel thread,
and in ->sync_fs() it should respect the 'wait' flag. This patch fixes
the situation.

Also, in ->put_super(), do not write the SB if it is not dirty.
Tested-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

In 'affs_write_super()': remove ancient and wrong commented code,
remove unneeded 'clean' variable, so the function becomes a bit
cleaner and simpler.

In 'affs_remount(): remove unnecessary SB dirty flag changes.
Tested-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NArtem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Remove the calls to inode_newsize_ok given that we already did it as
part of inode_change_ok in the beginning of cifs_setattr_(no)unix.

No need to call ->truncate if cifs doesn't have one, so remove the
explicit call in cifs_vmtruncate, and replace the calls to vmtruncate
with truncate_setsize which is vmtruncate minus inode_newsize_ok
and the call to ->truncate.

Rename cifs_vmtruncate to cifs_setsize to match the new calling conventions.

Question 1:  why does cifs do the pagecache munging and i_size update twice
	for each setattr call, once opencoded in cifs_vmtruncate, and once
	using the VFS helpers?
Question 2: what is supposed to be protected by i_lock in cifs_vmtruncate?
	Do we need it around the call to inode_change_ok?

[AV: fixed build breakage]
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The shrinker function is supposed to return the number of cache
entries after shrinking, not before shrinking.  Fix that.

Based on a patch from Wang Sheng-Hui <crosslonelyover@gmail.com>.
Signed-off-by: NAndreas Gruenbacher <agruen@suse.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The mbcache code was written to support a variable number of indexes,
but all the existing users use exactly one index.  Simplify to code to
support only that case.

There are also no users of the cache entry free operation, and none of
the users keep extra data in cache entries.  Remove those features as
well.
Signed-off-by: NAndreas Gruenbacher <agruen@suse.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Add a flags field to help glibc implementing statvfs(3) efficiently.

We copy the flag values from glibc, and add a new ST_VALID flag to
denote that f_flags is implemented.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

We'll need the path to implement the flags field for statvfs support.
We do have it available in all callers except:

 - ecryptfs_statfs.  This one doesn't actually need vfs_statfs but just
   needs to do a caller to the lower filesystem statfs method.
 - sys_ustat.  Add a non-exported statfs_by_dentry helper for it which
   doesn't won't be able to fill out the flags field later on.

In addition rename the helpers for statfs vs fstatfs to do_*statfs instead
of the misleading vfs prefix.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

... and let iput_final() do the actual eviction or retention
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

pretty much brute-force...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

ordering problems at ->kill_sb() time are solved by doing iput()
of these suckers in ->put_super()
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

[folded build fix from sfr]
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

NB: do we want btrfs_wait_ordered_range() on eviction of
inodes with positive i_nlink on subvolume with zero root_refs?
If not, btrfs_evict_inode() can be simplified by unconditionally
bailing out in case of i_nlink > 0 in the very beginning...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

In case of mode mismatch, do *not* blindly close the descriptor
another openers might be using right now.  Open the underlying
file with currently sufficient mode, then
	* if current mode has grown so that it's sufficient for
us now, just close our new fd
	* if current mode has grown and our fd is *not* enough
to cover it, close and repeat.
	* otherwise, install our fd if the file hadn't been
opened at all or dup2() our fd over the current one (and close
our fd).
Critical section is protected by mutex; yes, system-wide.  All
we do under it is a bunch of comparison and maybe an overwriting
dup2() on host.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

calculating size, then doing allocation, then filling the
path is a Bad Idea(tm), since the ancestors can be renamed,
leading to buffer overrun.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

builds path relative to fs root, called under dcache_lock,
doesn't append any nonsense to unlinked ones.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>