linux.git/fs/nfs, branch v3.1

VFS: Fix the remaining automounter semantics regressions

2011-09-27T02:16:46+00:00

The concensus seems to be that system calls such as stat() etc should
not trigger an automount.  Neither should the l* versions.

This patch therefore adds a LOOKUP_AUTOMOUNT flag to tag those lookups
that _should_ trigger an automount on the last path element.

Signed-off-by: Trond Myklebust 
[ Edited to leave out the cases that are already covered by LOOKUP_OPEN,
  LOOKUP_DIRECTORY and LOOKUP_CREATE - all of which also fundamentally
  force automounting for their own reasons   - Linus ]
Signed-off-by: Linus Torvalds

nfs: Do not allow multiple mounts on same mountpoint when using -o noac

2011-09-13T21:10:15+00:00

Do not allow multiple mounts on same mountpoint when using -o noac

When you normally attempt to mount a share twice on the same mountpoint,
a check in do_add_mount causes it to return an error

# mount localhost:/nfsv3 /mnt
# mount localhost:/nfsv3 /mnt
mount.nfs: /mnt is already mounted or busy

However when using the option 'noac', the user is able to mount the same
share on the same mountpoint multiple times. This happens because a
share mounted with the noac option is automatically assigned the 'sync'
flag MS_SYNCHRONOUS in nfs_initialise_sb(). This flag is set after the
check for already existing superblocks is done in sget(). The check for
the mount flags in nfs_compare_mount_options() does not take into
account the 'sync' flag applied later on in the code path. This means
that when using 'noac', a new superblock structure is assigned for every
new mount of the same share and multiple shares on the same mountpoint
are allowed.

ie.
# mount -onoac localhost:/nfsv3 /mnt
can be run multiple times.

The patch checks for noac and assigns the sync flag before sget() is
called to obtain an already existing superblock structure.

Signed-off-by: Sachin Prabhu 
Reviewed-by: Jeff Layton 
Signed-off-by: Trond Myklebust

NFS: Fix a typo in nfs_flush_multi

2011-09-13T21:06:57+00:00

Fix a typo which causes an Oops in the RPC layer, when using wsize < 4k.

Signed-off-by: Trond Myklebust 
Tested-by: Sricharan R

NFSv4: renewd needs to be able to handle the NFS4ERR_CB_PATH_DOWN error

2011-08-24T19:07:37+00:00

The NFSv4 spec does not specify that the server must repeat that error,
so in order to avoid having the delegations revoked, we should handle
it immediately.

Also note that NFS4ERR_CB_PATH_DOWN does in fact renew the lease...

Signed-off-by: Trond Myklebust

NFSv4: The NFSv4.0 client must send RENEW calls if it holds a delegation

2011-08-24T19:07:37+00:00

RFC3530 states that if the client holds a delegation, then it is obliged
to continue to send RENEW calls once every lease period in order to allow
the server to return NFS4ERR_CB_PATH_DOWN if the callback path is
unreachable.

This is not required for NFSv4.1, since the server can at any time set
the SEQ4_STATUS_CB_PATH_DOWN_SESSION in any SEQUENCE operation.

Signed-off-by: Trond Myklebust

NFSv4: nfs4_proc_renew should be declared static

2011-08-24T19:07:37+00:00

Signed-off-by: Trond Myklebust

NFSv4: nfs4_proc_async_renew should use a GFP_NOFS allocation

2011-08-24T19:07:35+00:00

We shouldn't allow the renew daemon to do direct reclaim on the NFS
partition.

Signed-off-by: Trond Myklebust

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

2011-08-19T05:47:13+00:00

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv4.1: Return NFS4ERR_BADSESSION to callbacks during session resets
  NFSv4.1: Fix the callback 'highest_used_slotid' behaviour
  pnfs-obj: Fix the comp_index != 0 case
  pnfs-obj: Bug when we are running out of bio
  nfs: add missing prefetch.h include

pnfs: Automatically select blocks & objects layouts

2011-08-12T00:51:27+00:00

Just like files-layout, blocks & objects layouts are part of the
NFS 4.1 protocol and should be automatically selected if NFS_4_1
is selected. The small problem is that these depend on other
Kernel support being present, while files only depends on NFS
itself.

This patch removes from the user choice the presence of objects
and blocks layout. But makes sure these are selected only if
the depended subsystems are present in the Kernel.

Signed-off-by: Boaz Harrosh 
Acked-by: Peng Tao 
Signed-off-by: Linus Torvalds

NFS41: make PNFS_BLOCK selectable

2011-08-11T15:58:02+00:00

PNFS_BLOCK needs BLK_DEV_DM/MD, which is not a dependency for other
pnfs layout drivers. Seperate it out so others can still build when
BLK_DEV_DM/MD is not enabled.

Also change select to depends on to avoid build failures.

Reported-and-tested-by: Randy Dunlap 
Signed-off-by: Peng Tao 
Acked-by: Benny Halevy 
Signed-off-by: Linus Torvalds