OCFS2 can return -ERESTARTSYS from write requests (and possibly
elsewhere) if there is a signal pending.

If nfsd is shutdown (by sending a signal to each thread) while there
is still an IO load from the client, each thread could handle one last
request with a signal pending.  This can result in -ERESTARTSYS
which is not understood by nfserrno() and so is reflected back to
the client as nfserr_io aka -EIO.  This is wrong.

Instead, interpret ERESTARTSYS to mean "try again later" by returning
nfserr_jukebox.  The client will resend and - if the server is
restarted - the write will (hopefully) be successful and everyone will
be happy.

 The symptom that I narrowed down to this was:
    copy a large file via NFS to an OCFS2 filesystem, and restart
    the nfs server during the copy.
    The 'cp' might get an -EIO, and the file will be corrupted -
    presumably holes in the middle where writes appeared to fail.
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

We need the nfsd_mutex before accessing nfsd_serv->sv_nrthreads or we
can't even guarantee nfsd_serv will still be there.
Signed-off-by: NNeil Brown <neilb@suse.de>
Acked-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

If lockd_down is called very rapidly after lockd_up returns, then
there is a slim chance that lockd() will never be called. kthread()
will return before calling the function, so we'll end up never
actually calling the cleanup functions for the thread.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Since we no longer make any distinction between shutdown signals with
nfsd, then it becomes easier to just standardize on a particular signal
to use to bring it down (SIGINT, in this case).
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

This patch is rather large, but I couldn't figure out a way to break it
up that would remain bisectable. It does several things:

- change svc_thread_fn typedef to better match what kthread_create expects
- change svc_pool_map_set_cpumask to be more kthread friendly. Make it
  take a task arg and and get rid of the "oldmask"
- have svc_set_num_threads call kthread_create directly
- eliminate __svc_create_thread
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

The special handling for SIGHUP in knfsd is a holdover from much
earlier versions of Linux where reloading the export table was
more expensive. That facility is not really needed anymore and
to my knowledge, is seldom-used.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Several of the nfsd filesystem interfaces allow changes to parameters
that don't have any effect on a running nfsd service. They are only ever
checked when nfsd is started. This patch fixes it so that changes to
those procfiles return -EBUSY if nfsd is already running to make it
clear that changes on the fly don't work.

The patch should also close some relatively harmless races between
changing the info in those interfaces and starting nfsd, since these
variables are being moved under the protection of the nfsd_mutex.

Finally, the nfsv4recoverydir file always returns -EINVAL if read. This
patch fixes it to return the recoverydir path as expected.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

This removes the BKL from the RPC service creation codepath. The BKL
really isn't adequate for this job since some of this info needs
protection across sleeps.

Also, add some comments to try and clarify how the locking should work
and to make it clear that the BKL isn't necessary as long as there is
adequate locking between tasks when touching the svc_serv fields.
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

WRITEMEM zeroes the last word in the destination buffer
for padding purposes, but this must not be done if
no bytes are to be copied, as it would result
in zeroing of the word right before the array.

The current implementation works since it's always called
with non zero nbytes or it follows an encoding of the
string (or opaque) length which, if equal to zero,
can be overwritten with zero.

Nevertheless, it seems safer to check for this case.
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

We already print each operation of the compound when debugging is turned
on; printing the result could also help with remote debugging.
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

These bit operations don't need to be atomic.  They're all done under a
single big mutex anyway.
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Use asm/byteorder.h instead.
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Updating the current transaction's t_state is protected by j_state_lock.  We
need to do the same when updating the t_state to T_COMMIT.
Acked-by: NJan Kara <jack@suse.cz>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>

If the block allocator gets blocks out of system zone ext4 calls
ext4_error. But if the file system is mounted with errors=continue
retry block allocation. We need to mark the system zone blocks as
in use to make sure retry don't pick them again

System zone is the block range mapping block bitmap, inode bitmap and inode
table.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

This fix the uninitialized bs when we try to replace a xattr entry in
ibody with the new value which require more than free space.

This situation only happens we format ext3/4 with inode size more than 128 and
we have put xattr entries both in ibody and block.  The consequences about
this bug is we will lost the xattr block which pointed by i_file_acl with all
xattr entires in it.  We will alloc a new xattr block and put that large value
entry in it.  The old xattr block will become orphan block.
Signed-off-by: NTiger Yang <tiger.yang@oracle.com>
Cc: <linux-ext4@vger.kernel.org>
Cc: Andreas Gruenbacher <agruen@suse.de>
Acked-by: NAndreas Dilger <adilger@sun.com>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Updating the current transaction's t_state is protected by j_state_lock.  We
need to do the same when updating the t_state to T_COMMIT.
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Acked-by: NJan Kara <jack@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

There was some cleanup issues during early mount which would trigger
a kernel bug for certain types of failure.  This patch reorganizes the
cleanup to get rid of the bad behavior.

This also merges the 9pnet and 9pnet_fd modules for the purpose of
configuration and initialization.  Keeping the fd transport separate
from the core 9pnet code seemed like a good idea at the time, but in
practice has caused more harm and confusion than good.
Signed-off-by: NEric Van Hensbergen <ericvh@gmail.com>

fs/9p/v9fs.c (v9fs_parse_options): Handle kstrdup and match_strdup failure. Now that this function can fail, return an int, diagnose other option-parsing failures, and adjust the sole caller: (v9fs_session_init): Handle kstrdup failure. Propagate any new v9fs_parse_options failure "up".
Signed-off-by: NJim Meyering <meyering@redhat.com>
Cc: Ron Minnich <rminnich@sandia.gov>
Cc: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Acked-by: NEric Van Hensbergen <ericvh@gmail.com>

The kernel-doc comments of much of the 9p system have been in disarray since
reorganization.  This patch fixes those problems, adds additional documentation
and a template book which collects the 9p information.
Signed-off-by: NEric Van Hensbergen <ericvh@gmail.com>

match_strcpy() is a somewhat creepy function: the caller needs to make sure
that the destination buffer is big enough, and when he screws up or
forgets, match_strcpy() happily overruns the buffer.

There's exactly one customer: v9fs_parse_options().  I believe it currently
can't overflow its buffer, but that's not exactly obvious.

The source string is a substing of the mount options.  The kernel silently
truncates those to PAGE_SIZE bytes, including the terminating zero.  See
compat_sys_mount() and do_mount().

The destination buffer is obtained from __getname(), which allocates from
name_cachep, which is initialized by vfs_caches_init() for size PATH_MAX.

We're safe as long as PATH_MAX <= PAGE_SIZE.  PATH_MAX is 4096.  As far as
I know, the smallest PAGE_SIZE is also 4096.

Here's a patch that makes the code a bit more obviously correct.  It
doesn't depend on PATH_MAX <= PAGE_SIZE.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Cc: Latchesar Ionkov <lucho@ionkov.net>
Cc: Jim Meyering <meyering@redhat.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NEric Van Hensbergen <ericvh@gmail.com>

In case of inode preallocation, the number of blocks to allocate depends
on the file size and it is calculated in ext4_mb_normalize_request().
Each group in the filesystem is then checked to find one that can be
used for allocation; this is done in ext4_mb_good_group().

When a file bigger than 4MB is created, the requested number of blocks
to preallocate, calculated by ext4_mb_normalize_request is 4096.
However for a filesystem with 1KB block size, the maximum size of the
block buddies used by the multiblock allocator is 2048, so none of
groups in the filesystem satisfies the search criteria in
ext4_mb_good_group(). Scanning all the filesystem groups impacts
performance.

This was demonstrated by using a freshly created, 70GB, 1k block
filesystem, with caches dropped write before the test via
/proc/sys/vm/drop_caches, and with the filesystem mounted with
nodelalloc and nodealloc,nomballoc.  The time to write an 8 megabyte
file using "dd if=/dev/zero of=/mnt/test/fo bs=8k count=1k conv=fsync"
took 35.5091 seconds (236kB/s) with nodellaloc, and 0.233754 seconds
(35.9 MB/s) with the nodelloc,nomballoc options.  With a 1TB partition,
it took several minutes to write 8MB!

This patch modifies the algorithm in ext4_mb_normalize_group_request to
calculate the number of blocks to allocate by taking into account the
maximum size of free blocks chunks handled by the multiblock allocator.

It has also been tested for filesystems with 2KB and 4KB block sizes to
ensure that those cases don't regress.
Reviewed-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NValerie Clement <valerie.clement@bull.net>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

In journal=data mode, it is not enough to do write_inode_now as done in
vfs_quota_on() to write all data to their final location (which is
needed for quota_read to work correctly).  Calling journal_flush() does
its job.

Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

When quota is disabled, we should not print 'journaled quota not
supported' when user tried to mount non-journaled quota. Also fix typo
in the message.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We should not allow user to change quota mount options when quota is
just suspended.  It would make mount options and internal quota state
inconsistent.  Also we should not allow user to change quota format when
quota is turned on.  On the other hand we can just silently ignore when
some option is set to the value it already has (mount does this on
remount).

Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Fix imbalanced calls for mutex lock/unlock on ecryptfs_daemon_hash_mux
Revealed by Ingo Molnar: http://lkml.org/lkml/2008/5/7/260Signed-off-by: NCyrill Gorcunov <gorcunov@gmail.com>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

bdevname() fills the buffer that it is given as a parameter, so calling
strcpy() or snprintf() on the returned value is redundant (and probably not
guaranteed to work - I don't think strcpy and snprintf support overlapping
buffers.)
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Prior to 2.6.26 fuse only supported single page write requests.  In theory all
fuse filesystem should be able support bigger than 4k writes, as there's
nothing in the API to prevent it.  Unfortunately there's a known case in
NTFS-3G where big writes cause filesystem corruption.  There could also be
other filesystems, where the lack of testing with big write requests would
result in bugs.

To prevent such problems on a kernel upgrade, disable big writes by default,
but let filesystems set a flag to turn it on.
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Cc: Szabolcs Szakacsits <szaka@ntfs-3g.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When mm destruction happens, we should pass mm_update_next_owner() the old mm.
 But unfortunately new mm is passed in exec_mmap().

Thus, kernel panic is possible when a multi-threaded process uses exec().

Also, the owner member comment description is wrong.  mm->owner does not
necessarily point to the thread group leader.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: NBalbir Singh <balbir@linux.vnet.ibm.com>
Cc: "Paul Menage" <menage@google.com>
Cc: "KAMEZAWA Hiroyuki" <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix an oops with a corrupted hfs+ image.

See http://bugzilla.kernel.org/show_bug.cgi?id=10548 for details.

Problem is that we call hfs_btree_open() from hfsplus_fill_super() to set
HFSPLUS_SB(sb).[ext_tree|cat_tree] Both trees are still NULL at this moment.
If hfs_btree_open() fails for any reason it calls iput() on the page, which
gets to hfsplus_releasepage() which tries to access HFSPLUS_SB(sb).* which is
still NULL and oopses while dereferencing it.

[akpm@linux-foundation.org: build fix]
Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

There is currently no way to query the bounding set of another task.  As there
appears to be no security reason not to, and as Michael Kerrisk points out the
following valid reasons to do so exist:

* consistency (I can see all of the other per-thread/process sets in
  /proc/.../status)

* debugging -- I could imagine that it would make the job of debugging an
  application that uses capabilities a little simpler.

this patch adds the bounding set to /proc/self/status right after the
effective set.
Signed-off-by: NSerge E. Hallyn <serue@us.ibm.com>
Acked-by: NMichael Kerrisk <mtk.manpages@gmail.com>
Acked-by: NAndrew G. Morgan <morgan@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Sometimes, vfs_quota_off() is called on a partially set up super block (for
example when fill_super() fails for some reason).  In such cases we cannot
call ->sync_fs() because it can Oops because of not properly filled in super
block.  So in case we find there's not quota to turn off, we just skip
everything and return which fixes the above problem.

[akpm@linux-foundation.org: fxi tpyo]
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NEvgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

dget(dentry->d_parent) --> dget_parent(dentry)

unlock_parent() is racy and unnecessary.  Replace single caller with
unlock_dir().

There are several other suspect uses of ->d_parent in ecryptfs...
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

There's no reason for the _kern in hppfs_kern.c, so move it to hppfs.c.
Signed-off-by: NJeff Dike <jdike@linux.intel.com>
Cc: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

hppfs tidying and fixes noticed during hch's get_inode work -
      style fixes
      a copy_to_user got its return value checked
      hppfs_write no longer fiddles file->f_pos because it gets and
returns pos in its arguments
      hppfs_delete_inode dputs the underlyng procfs dentry stored in
its private data and mntputs the vfsmnt stashed in s_fs_info
      hppfs_put_super no longer needs to mntput the s_fs_info, so it
no longer needs to exist
      hppfs_readlink and hppfs_follow_link were doing a bunch of stuff
with a struct file which they didn't use
      there is now a ->permission which calls generic_permission
      get_inode was always returning 0 for some reason - it now
returns an inode if nothing bad happened
Signed-off-by: NJeff Dike <jdike@linux.intel.com>
Cc: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It acts exactly like a regular 'cond_resched()', but will not get
optimized away when CONFIG_PREEMPT is set.

Normal kernel code is already preemptable in the presense of
CONFIG_PREEMPT, so cond_resched() is optimized away (see commit
02b67cc3 "sched: do not do
cond_resched() when CONFIG_PREEMPT").

But when wanting to conditionally reschedule while holding a lock, you
need to use "cond_sched_lock(lock)", and the new function is the BKL
equivalent of that.

Also make fs/locks.c use it.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

cifs_demultiplex_thread can exit under several conditions:

1) if it's signaled
2) if there's a problem with session setup
3) if kthread_stop is called on it

The first two are problems. If kthread_stop is called on the thread,
there is no guarantee that it will still be up. We need to have the
thread stay up until kthread_stop is called on it.

One option would be to not even try to tear things down until after
kthread_stop is called. However, in the case where there is a problem
setting up the session, there's no real reason to try continuing the
loop.

This patch allows the thread to clean up and prepare for exit under all
three conditions, but it has the thread go to sleep until kthread_stop
is called. This allows us to simplify the shutdown code somewhat since
we can be reasonably sure that the thread won't exit after being
signaled but before kthread_stop is called.

It also removes the places where the thread itself set the tsk variable
since it appeared that it could have a potential race where the thread
might never be shut down.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NChristoph Hellwig <hch@infradead.org>
Signed-off-by: NSteve French <sfrench@us.ibm.com>

When creating a directory on a CIFS share without POSIX extensions,
and the given mode has no write bits set, set the ATTR_READONLY bit.

When creating a file, set ATTR_READONLY if the create mode has no write
bits set and we're not using unix extensions.

There are some comments about this being problematic due to the VFS
splitting creates into 2 parts. I'm not sure what that's actually
talking about, but I'm assuming that it has something to do with how
mknod is implemented. In the simple case where we have no unix
extensions and we're just creating a regular file, there's no reason
we can't set ATTR_READONLY.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NSteve French <sfrench@us.ibm.com>

Clean up cifs_setattr a bit by adding a local inode pointer, and
changing all of the direntry->d_inode references to it. This also adds a
bit of micro-optimization. d_inode shouldn't change over the life of
this function, so we only need to dereference it once.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NSteve French <sfrench@us.ibm.com>