Adam Lackorzynski reports:

with 2.6.35.2 I'm getting this reproducible Oops:

[  110.825396] BUG: unable to handle kernel NULL pointer dereference at
(null)
[  110.828638] IP: [<ffffffff811247b7>] encode_attrs+0x1a/0x2a4
[  110.828638] PGD be89f067 PUD bf18f067 PMD 0
[  110.828638] Oops: 0000 [#1] SMP
[  110.828638] last sysfs file: /sys/class/net/lo/operstate
[  110.828638] CPU 2
[  110.828638] Modules linked in: rtc_cmos rtc_core rtc_lib amd64_edac_mod
i2c_amd756 edac_core i2c_core dm_mirror dm_region_hash dm_log dm_snapshot
sg sr_mod usb_storage ohci_hcd mptspi tg3 mptscsih mptbase usbcore nls_base
[last unloaded: scsi_wait_scan]
[  110.828638]
[  110.828638] Pid: 11264, comm: setchecksum Not tainted 2.6.35.2 #1
[  110.828638] RIP: 0010:[<ffffffff811247b7>]  [<ffffffff811247b7>]
encode_attrs+0x1a/0x2a4
[  110.828638] RSP: 0000:ffff88003bf5b878  EFLAGS: 00010296
[  110.828638] RAX: ffff8800bddb48a8 RBX: ffff88003bf5bb18 RCX:
0000000000000000
[  110.828638] RDX: ffff8800be258800 RSI: 0000000000000000 RDI:
ffff88003bf5b9f8
[  110.828638] RBP: 0000000000000000 R08: ffff8800bddb48a8 R09:
0000000000000004
[  110.828638] R10: 0000000000000003 R11: ffff8800be779000 R12:
ffff8800be258800
[  110.828638] R13: ffff88003bf5b9f8 R14: ffff88003bf5bb20 R15:
ffff8800be258800
[  110.828638] FS:  0000000000000000(0000) GS:ffff880041e00000(0063)
knlGS:00000000556bd6b0
[  110.828638] CS:  0010 DS: 002b ES: 002b CR0: 000000008005003b
[  110.828638] CR2: 0000000000000000 CR3: 00000000be8ef000 CR4:
00000000000006e0
[  110.828638] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
0000000000000000
[  110.828638] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
0000000000000400
[  110.828638] Process setchecksum (pid: 11264, threadinfo
ffff88003bf5a000, task ffff88003f232210)
[  110.828638] Stack:
[  110.828638]  0000000000000000 ffff8800bfbcf920 0000000000000000
0000000000000ffe
[  110.828638] <0> 0000000000000000 0000000000000000 0000000000000000
0000000000000000
[  110.828638] <0> 0000000000000000 0000000000000000 0000000000000000
0000000000000000
[  110.828638] Call Trace:
[  110.828638]  [<ffffffff81124c1f>] ? nfs4_xdr_enc_setattr+0x90/0xb4
[  110.828638]  [<ffffffff81371161>] ? call_transmit+0x1c3/0x24a
[  110.828638]  [<ffffffff813774d9>] ? __rpc_execute+0x78/0x22a
[  110.828638]  [<ffffffff81371a91>] ? rpc_run_task+0x21/0x2b
[  110.828638]  [<ffffffff81371b7e>] ? rpc_call_sync+0x3d/0x5d
[  110.828638]  [<ffffffff8111e284>] ? _nfs4_do_setattr+0x11b/0x147
[  110.828638]  [<ffffffff81109466>] ? nfs_init_locked+0x0/0x32
[  110.828638]  [<ffffffff810ac521>] ? ifind+0x4e/0x90
[  110.828638]  [<ffffffff8111e2fb>] ? nfs4_do_setattr+0x4b/0x6e
[  110.828638]  [<ffffffff8111e634>] ? nfs4_do_open+0x291/0x3a6
[  110.828638]  [<ffffffff8111ed81>] ? nfs4_open_revalidate+0x63/0x14a
[  110.828638]  [<ffffffff811056c4>] ? nfs_open_revalidate+0xd7/0x161
[  110.828638]  [<ffffffff810a2de4>] ? do_lookup+0x1a4/0x201
[  110.828638]  [<ffffffff810a4733>] ? link_path_walk+0x6a/0x9d5
[  110.828638]  [<ffffffff810a42b6>] ? do_last+0x17b/0x58e
[  110.828638]  [<ffffffff810a5fbe>] ? do_filp_open+0x1bd/0x56e
[  110.828638]  [<ffffffff811cd5e0>] ? _atomic_dec_and_lock+0x30/0x48
[  110.828638]  [<ffffffff810a9b1b>] ? dput+0x37/0x152
[  110.828638]  [<ffffffff810ae063>] ? alloc_fd+0x69/0x10a
[  110.828638]  [<ffffffff81099f39>] ? do_sys_open+0x56/0x100
[  110.828638]  [<ffffffff81027a22>] ? ia32_sysret+0x0/0x5
[  110.828638] Code: 83 f1 01 e8 f5 ca ff ff 48 83 c4 50 5b 5d 41 5c c3 41
57 41 56 41 55 49 89 fd 41 54 49 89 d4 55 48 89 f5 53 48 81 ec 18 01 00 00
<8b> 06 89 c2 83 e2 08 83 fa 01 19 db 83 e3 f8 83 c3 18 a8 01 8d
[  110.828638] RIP  [<ffffffff811247b7>] encode_attrs+0x1a/0x2a4
[  110.828638]  RSP <ffff88003bf5b878>
[  110.828638] CR2: 0000000000000000
[  112.840396] ---[ end trace 95282e83fd77358f ]---

We need to ensure that the O_EXCL flag is turned off if the user doesn't
set O_CREAT.

Cc: stable@kernel.org
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

To obey NFS cache semantics, the client must verify the cached
attributes when a file is opened.  In most cases this is done by a call to
d_validate as one of the last steps in path_walk.

However for the root of a filesystem, d_validate is only ever called
on the mounted-on filesystem (except when the path ends '.' or '..').
So NFS has no chance to validate the attributes.

So, in nfs_opendir, we revalidate the attributes if the opened
directory is the mountpoint.  This may cause double-validation for "."
and ".." lookups, but that is better than missing regular /path/name
lookups completely.
Signed-off-by: NNeilBrown <neilb@suse.de>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

We don't really support nfs servers that invalidate the file handle after a
rename, so precautions such as flushing out dirty data before renaming the
file are superfluous.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Currently MAY_ACCESS means that filesystems must check the permissions
right then and not rely on cached results or the results of future
operations on the object.  This can be because of a call to sys_access() or
because of a call to chdir() which needs to check search without relying on
any future operations inside that dir.  I plan to use MAY_ACCESS for other
purposes in the security system, so I split the MAY_ACCESS and the
MAY_CHDIR cases.
Signed-off-by: NEric Paris <eparis@redhat.com>
Acked-by: NStephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: NJames Morris <jmorris@namei.org>

The current shrinker implementation requires the registered callback
to have global state to work from. This makes it difficult to shrink
caches that are not global (e.g. per-filesystem caches). Pass the shrinker
structure to the callback so that users can embed the shrinker structure
in the context the shrinker needs to operate on and get back to it in the
callback via container_of().
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Commit 9c7e7e23 (NFS: Don't call iput() in
nfs_access_cache_shrinker) unintentionally removed the spin unlock for the
inode->i_lock.
Reported-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

iput() can potentially attempt to allocate memory, so we should avoid
calling it in a memory shrinker. Instead, rely on the fact that iput() will
call nfs_access_zap_cache().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Both iput() and put_rpccred() might allocate memory under certain
circumstances, so make sure that we don't recurse and deadlock...
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

If dentry found stale happens to be a root of disconnected tree, we
can't d_drop() it; its d_hash is actually part of s_anon and d_drop()
would simply hide it from shrink_dcache_for_umount(), leading to
all sorts of fun, including busy inodes on umount and oopsen after
that.

Bug had been there since at least 2006 (commit c636eb already has it),
so it's definitely -stable fodder.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Cc: stable@kernel.org
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix https://bugzilla.kernel.org/show_bug.cgi?id=15789Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org

If the NFS_INO_REVAL_FORCED flag is set, that means that we don't yet have
an up to date attribute cache. Even if we hold a delegation, we must
put a GETATTR on the wire.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Recent change is missing to update "rehash".  With that change, it will
become the cause of adding dentry to hash twice.

This explains the reason of Oops (dereference the freed dentry in
__d_lookup()) on my machine.
Signed-off-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Reported-by: NMarvin <marvin24@gmx.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The d_instantiate(new_dentry, NULL) is superfluous, the dentry is
already negative.  Rehashing this dummy dentry isn't needed either,
d_move() works fine on an unhashed target.

The re-checking for busy after a failed nfs_sillyrename() is bogus
too: new_dentry->d_count < 2 would be a bug here.
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Move unhashing the target to after the check for existence and being a
non-directory.

If renaming a directory then the VFS already unhashes the target if it
is not busy.  If it's busy then acquiring more references during the
rename makes no difference.
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Comments are wrong or out of date.  In particular d_drop() doesn't
free the inode it just unhashes the dentry.  And if target is a
directory then it is not checked for being busy.
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

VFS already checks if both source and target are directories.
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Otherwise, we have to wait for the server to recall it.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

We just had a case in which a buggy server occasionally returns the wrong
attributes during an OPEN call. While the client does catch this sort of
condition in nfs4_open_done(), and causes the nfs4_atomic_open() to return
-EISDIR, the logic in nfs_atomic_lookup() is broken, since it causes a
fallback to an ordinary lookup instead of just returning the error.

When the buggy server then returns a regular file for the fallback lookup,
the VFS allows the open, and bad things start to happen, since the open
file doesn't have any associated NFSv4 state.

The fix is firstly to return the EISDIR/ENOTDIR errors immediately, and
secondly to ensure that we are always careful when dereferencing the
nfs_open_context state pointer.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

* Remove smp_lock.h from files which don't need it (including some headers!)
* Add smp_lock.h to files which do need it
* Make smp_lock.h include conditional in hardirq.h
  It's needed only for one kernel_locked() usage which is under CONFIG_PREEMPT

  This will make hardirq.h inclusion cheaper for every PREEMPT=n config
  (which includes allmodconfig/allyesconfig, BTW)
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The problem is that permission checking is skipped if atomic open is
possible, but when exec opens a file, it just opens it O_READONLY which
means EXEC permission will not be checked at that time.

This problem is observed by the following sequence (executed as root):

  mount -t nfs4 server:/ /mnt4
  echo "ls" >/mnt4/foo
  chmod 744 /mnt4/foo
  su guest -c "mnt4/foo"
Signed-off-by: NFrank Filz <ffilzlnx@us.ibm.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
Tested-by: NEugene Teo <eugeneteo@kernel.sg>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

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

NFS appears to be returning an unnecessary "delete" notification when
we're doing an atomic rename. See

  http://bugzilla.gnome.org/show_bug.cgi?id=575684

The fix is to get rid of the redundant call to d_delete().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Hi Trond,

I have been looking at a bugreport where trying to open applications on KDE
on a NFS mounted home fails temporarily. There have been multiple reports on
different kernel versions pointing to this common issue:
http://bugzilla.kernel.org/show_bug.cgi?id=12557
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/269954
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=508866.html

This issue can be reproducible consistently by doing this on a NFS mounted
home (KDE):
1. Open 2 xterm sessions
2. From one of the xterm session, do "ssh -X <remote host>"
3. "stat ~/.Xauthority" on the remote SSH session
4. Close the two xterm sessions
5. On the server do a "stat ~/.Xauthority"
6. Now on the client, try to open xterm
This will fail.

Even if the filehandle had become stale, the NFS client should invalidate
the cache/inode and should repeat LOOKUP. Looking at the packet capture when
the failure occurs shows that there were two subsequent ACCESS() calls with
the same filehandle and both fails with -ESTALE error.

I have tested the fix below. Now the client issue a LOOKUP after the
ACCESS() call fails with -ESTALE. If all this makes sense to you, can you
consider this for inclusion?

Thanks,


If the server returns an -ESTALE error due to stale filehandle in response to
an ACCESS() call, we need to invalidate the cache and inode so that LOOKUP()
can be retried. Without this change, the nfs client retries ACCESS() with the
same filehandle, fails again and could lead to temporary failure of
applications running on nfs mounted home.
Signed-off-by: NSuresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Hi.

I've been looking at a bugzilla which describes a problem where
a customer was advised to use either the "noac" or "actimeo=0"
mount options to solve a consistency problem that they were
seeing in the file attributes.  It turned out that this solution
did not work reliably for them because sometimes, the local
attribute cache was believed to be valid and not timed out.
(With an attribute cache timeout of 0, the cache should always
appear to be timed out.)

In looking at this situation, it appears to me that the problem
is that the attribute cache timeout code has an off-by-one
error in it.  It is assuming that the cache is valid in the
region, [read_cache_jiffies, read_cache_jiffies + attrtimeo].  The
cache should be considered valid only in the region,
[read_cache_jiffies, read_cache_jiffies + attrtimeo).  With this
change, the options, "noac" and "actimeo=0", work as originally
expected.

This problem was previously addressed by special casing the
attrtimeo == 0 case.  However, since the problem is only an off-
by-one error, the cleaner solution is address the off-by-one
error and thus, not require the special case.

    Thanx...

        ps
Signed-off-by: NPeter Staubach <staubach@redhat.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This ensures that we don't have to look up the dentry again after we return
the delegation if we know that the directory didn't change.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

For execute permission on a regular files we need to check if file has
any execute bits at all, regardless of capabilites.

This check is normally performed by generic_permission() but was also
added to the case when the filesystem defines its own ->permission()
method.  In the latter case the filesystem should be responsible for
performing this check.

Move the check from inode_permission() inside filesystems which are
not calling generic_permission().

Create a helper function execute_ok() that returns true if the inode
is a directory or if any execute bits are present in i_mode.

Also fix up the following code:

 - coda control file is never executable
 - sysctl files are never executable
 - hfs_permission seems broken on MAY_EXEC, remove
 - hfsplus_permission is eqivalent to generic_permission(), remove
Signed-off-by: NMiklos Szeredi <mszeredi@suse.cz>

New flag: LOOKUP_EXCL.  Set before doing the final step of pathname
resolution on the paths that have LOOKUP_CREATE and O_EXCL.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Split the LRU lists in two, one set for pages that are backed by real file
systems ("file") and one for pages that are backed by memory and swap
("anon").  The latter includes tmpfs.

The advantage of doing this is that the VM will not have to scan over lots
of anonymous pages (which we generally do not want to swap out), just to
find the page cache pages that it should evict.

This patch has the infrastructure and a basic policy to balance how much
we scan the anon lists and how much we scan the file lists.  The big
policy changes are in separate patches.

[lee.schermerhorn@hp.com: collect lru meminfo statistics from correct offset]
[kosaki.motohiro@jp.fujitsu.com: prevent incorrect oom under split_lru]
[kosaki.motohiro@jp.fujitsu.com: fix pagevec_move_tail() doesn't treat unevictable page]
[hugh@veritas.com: memcg swapbacked pages active]
[hugh@veritas.com: splitlru: BDI_CAP_SWAP_BACKED]
[akpm@linux-foundation.org: fix /proc/vmstat units]
[nishimura@mxp.nes.nec.co.jp: memcg: fix handling of shmem migration]
[kosaki.motohiro@jp.fujitsu.com: adjust Quicklists field of /proc/meminfo]
[kosaki.motohiro@jp.fujitsu.com: fix style issue of get_scan_ratio()]
Signed-off-by: NRik van Riel <riel@redhat.com>
Signed-off-by: NLee Schermerhorn <Lee.Schermerhorn@hp.com>
Signed-off-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NDaisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The cache_change_attribute is used to decide whether or not a directory has
changed, in which case we may need to look it up again. Again, the use of
'jiffies' leads to an issue of resolution.

Once again, the fix is to change nfs_inode->cache_change_attribute, and
just make it a simple counter.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

It appears that 'jiffies' timestamps do not have high enough resolution for
nfs_inode_attrs_need_update(). One problem is that a GETATTR can be
launched within < 1 jiffy of the last operation that updated the attribute.
Another problem is that RPC calls can take < 1 jiffy to execute.

We can fix this by switching the variables to use a simple global counter
that gets incremented every time we start another GETATTR call.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Add the flag NFS_MOUNT_LOOKUP_CACHE_NONEG to turn off the caching of
negative dentries. In reality what we do is to force
nfs_lookup_revalidate() to always discard negative dentries.

Add the flag NFS_MOUNT_LOOKUP_CACHE_NONE for enforcing stricter
revalidation of dentries. It forces the revalidate code to always do a
lookup instead of just checking the cached mtime of the parent directory.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

* kill nameidata * argument; map the 3 bits in ->flags anybody cares
  about to new MAY_... ones and pass with the mask.
* kill redundant gfs2_iop_permission()
* sanitize ecryptfs_permission()
* fix remaining places where ->permission() instances might barf on new
  MAY_... found in mask.

The obvious next target in that direction is permission(9)

folded fix for nfs_permission() breakage from Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>