linux-stable.git/fs/nfs, branch v3.2

NFS: Fix a regression in nfs_file_llseek()

2011-12-15T23:44:36+00:00

After commit 06222e491e663dac939f04b125c9dc52126a75c4 (fs: handle
SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek)
the behaviour of llseek() was changed so that it always revalidates
the file size. The bug appears to be due to a logic error in the
afore-mentioned commit, which always evaluates to 'true'.

Reported-by: Roel Kluin 
Signed-off-by: Trond Myklebust 
Cc: stable@vger.kernel.org [>=3.1]

NFSv4: Do not accept delegated opens when a delegation recall is in effect

2011-12-10T00:05:58+00:00

...and report the servers that try to return a delegation when the client
is using the CLAIM_DELEG_CUR open mode. That behaviour is explicitly
forbidden in RFC3530.

Signed-off-by: Trond Myklebust

NFSv4: Ensure correct locking when accessing the 'lock_states' list

2011-12-09T21:31:52+00:00

There are currently 2 places in the state recovery code, where we do not
take sufficient precautions before accessing the state->lock_states. In
both cases, we should be holding the state->state_lock.

Reported-by: Pascal Bouchareine 
Signed-off-by: Trond Myklebust

NFSv4.1: Ensure that we handle _all_ SEQUENCE status bits.

2011-12-01T21:37:42+00:00

Currently, the code assumes that the SEQUENCE status bits are mutually
exclusive. They are not...

Signed-off-by: Trond Myklebust 
Cc: stable@vger.kernel.org [>= 2.6.34]

NFSv4: Don't error if we handled it in nfs4_recovery_handle_error

2011-12-01T21:31:34+00:00

If we handled an error condition, then nfs4_recovery_handle_error should
return '0' so that the state recovery thread can continue.
Also ensure that nfs4_check_lease() continues to abort if we haven't got
any credentials by having it return ENOKEY (which is not handled).

Signed-off-by: Trond Myklebust

Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

2011-11-22T16:54:15+00:00

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFS: Revert pnfs ugliness from the generic NFS read code path
  SUNRPC: destroy freshly allocated transport in case of sockaddr init error
  NFS: Fix a regression in the referral code
  nfs: move nfs_file_operations declaration to bottom of file.c (try #2)
  nfs: when attempting to open a directory, fall back on normal lookup (try #5)

new helper: mount_subtree()

2011-11-17T03:00:34+00:00

takes vfsmount and relative path, does lookup within that vfsmount
(possibly triggering automounts) and returns the result as root
of subtree suitable for return by ->mount() (i.e. a reference to
dentry and an active reference to its superblock grabbed, superblock
locked exclusive).

btrfs and nfs switched to it instead of open-coding the sucker.

Signed-off-by: Al Viro

switch create_mnt_ns() to saner calling conventions, fix double mntput() in nfs

2011-11-16T21:12:14+00:00

Life is much saner if create_mnt_ns(mnt) drops mnt in case of error...
Switch it to such calling conventions, switch callers, fix double mntput() in
fs/nfs/super.c one.

Signed-off-by: Al Viro

NFS: Revert pnfs ugliness from the generic NFS read code path

2011-11-10T19:50:26+00:00

pNFS-specific code belongs in the pnfs layer. It should not be
hijacking generic NFS read or write code paths.

Signed-off-by: Trond Myklebust

NFS: Fix a regression in the referral code

2011-11-07T21:20:00+00:00

Fix a regression that was introduced by commit
0c2e53f11a6dae9e3af5f50f5ad0382e7c3e0cfa (NFS: Remove the unused
"lookupfh()" version of nfs4_proc_lookup()).

In the case where the lookup gets an NFS4ERR_MOVED, we want to return
the result of nfs4_get_referral(). Instead, that value is getting
clobbered by the call to nfs4_handle_exception()...

Reported-by: Tigran Mkrtchyan 
Tested-by: Tigran Mkrtchyan 
Signed-off-by: Trond Myklebust