Although they run as rpciod background tasks, under normal operation
(i.e. no SIGKILL), functions like nfs_sillyrename(), nfs4_proc_unlck()
and nfs4_do_close() want to be fully synchronous. This means that when we
exit, we want all references to the rpc_task to be gone, and we want
any dentry references etc. held by that task to be released.

For this reason these functions call __rpc_wait_for_completion_task(),
followed by rpc_put_task() in the expectation that the latter will be
releasing the last reference to the rpc_task, and thus ensuring that the
callback_ops->rpc_release() has been called synchronously.

This patch fixes a race which exists due to the fact that
rpciod calls rpc_complete_task() (in order to wake up the callers of
__rpc_wait_for_completion_task()) and then subsequently calls
rpc_put_task() without ensuring that these two steps are done atomically.

In order to avoid adding new spin locks, the patch uses the existing
waitqueue spin lock to order the rpc_task reference count releases between
the waiting process and rpciod.
The common case where nobody is waiting for completion is optimised for by
checking if the RPC_TASK_ASYNC flag is cleared and/or if the rpc_task
reference count is 1: in those cases we drop trying to grab the spin lock,
and immediately free up the rpc_task.

Those few processes that need to put the rpc_task from inside an
asynchronous context and that do not care about ordering are given a new
helper: rpc_put_task_async().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Make d_count non-atomic and protect it with d_lock. This allows us to ensure a
0 refcount dentry remains 0 without dcache_lock. It is also fairly natural when
we start protecting many other dentry members with d_lock.
Signed-off-by: NNick Piggin <npiggin@kernel.dk>

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

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

Having to explicitly initialize sr_slotid to NFS4_MAX_SLOT_TABLE
resulted in numerous bugs.  Keeping the current slot as a pointer
to the slot table is more straight forward and robust as it's
implicitly set up to NULL wherever the seq_res member is initialized
to zeroes.
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The call to nfs_async_rename_release() after rpc_run_task() is incorrect.
The rpc_run_task() is always guaranteed to call the ->rpc_release() method.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

A synchronous rename can be interrupted by a SIGKILL. If that happens
during a sillyrename operation, it's possible for the rename call to
be sent to the server, but the task exits before processing the
reply. If this happens, the sillyrenamed file won't get cleaned up
during nfs_dentry_iput and the server is left with a dangling .nfs* file
hanging around.

Fix this problem by turning sillyrename into an asynchronous operation
and have the task doing the sillyrename just wait on the reply. If the
task is killed before the sillyrename completes, it'll still proceed
to completion.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

...since that's where most of the sillyrenaming code lives. A comment
block is added to the beginning as well to clarify how sillyrenaming
works. Also, make nfs_async_unlink static as nfs_sillyrename is the only
caller.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

In anticipation of the day when we have per-filesystem sessions, and also
in order to allow the session to change in the event of a filesystem
migration event.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

Also rename it: it is used in generic code, and so should not have a 'nfs4'
prefix.
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

We no longer need to maintain a distinction between nfs41_sequence_done and
nfs41_sequence_free_slot.

This fixes a number of slot table leakages in the NFSv4.1 code.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

nfs41_sequence_free_slot can be called multiple times on SEQUENCE operation
errors.
No reason to inline nfs4_restart_rpc
Reported-by: NTrond Myklebust <trond.myklebust@netapp.com>

nfs_writeback_done and nfs_readpage_retry call nfs4_restart_rpc outside the
error handler, and the slot is not freed prior to restarting in the rpc_prepare
state during session reset.

Fix this by moving the call to nfs41_sequence_free_slot from the error
path of nfs41_sequence_done into nfs4_restart_rpc, and by removing the test
for NFS4CLNT_SESSION_SETUP.
Always free slot and goto the rpc prepare state on async errors.
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

[nfs41: change nfs4_restart_rpc argument]
[nfs41: check for session not minorversion]
[nfs41: trigger the state manager for session reset]
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[always define nfs4_restart_rpc]
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Implement the rpc_call_prepare methods for
asynchronuos nfs rpcs, call nfs41_setup_sequence from
respective rpc_call_validate_args methods.

Call nfs4_sequence_done from respective rpc_call_done methods.

Note that we need to pass a pointer to the nfs_server in calls data
for passing on to nfs4_sequence_done.

Signed-off-by: Andy Adamson<andros@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[pnfs: client data server write validate and release]
Signed-off-by: Andy Adamson<andros@umich.edu>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[nfs41: separate free slot from sequence done]
[nfs41: sequence res use slotid]
[nfs41: remove SEQ4_STATUS_USE_TK_STATUS]
[nfs41: nfs4_sequence_free_slot use nfs_client for data server]
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Initialize nfs4_sequence_res sr_slotid to NFS4_MAX_SLOT_TABLE.

[was nfs41: sequence res use slotid]
Signed-off-by: NAndy Adamson <andros@netapp.com>
[pulled definition of struct nfs4_sequence_res.sr_slotid to here]
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Instead of causing umount requests to block on server->active_wq while the
asynchronous sillyrename deletes are executing, we can use the sb->s_active
counter to obtain a reference to the super_block, and then release that
reference in nfs_async_unlink_release().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

In order to avoid the "Busy inodes after unmount" error message, we need to
ensure that nfs_async_unlink_release() releases the super block after the
call to nfs_free_unlinkdata().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The NFSv4 protocol allows clients to negotiate security protocols on the
fly in the case where an administrator on the server changes the export
settings and/or in the case where we may have a filesystem migration event.

Instead of having the NFS client code cache credentials that are tied to a
particular AUTH method it is therefore preferable to have a generic credential
that can be converted into whatever AUTH is in use by the RPC client when
the read/write/sillyrename/... is put on the wire.

We do this by means of the new "generic" credential, which basically just
caches the minimal information that is needed to look up an RPCSEC_GSS,
AUTH_SYS, or AUTH_NULL credential.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Replace use of rpc_call_setup() with rpc_init_task(), and in cases where we
need to initialise task->tk_action, with rpc_call_start().
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Make it use the new task initialiser structure instead of acting as a
wrapper.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Added an active/deactive mechanism to the nfs_server structure
allowing async operations to hold off umount until the
operations are done.
Signed-off-by: NSteve Dickson <steved@redhat.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

In nfs_do_call_unlink() we check that we haven't raced, and that lookup()
hasn't created an aliased dentry to our sillydeleted dentry. If somebody
has deleted the file on the server and the lookup() resulted in a negative
dentry, then ignore...
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Ensure that readdir revalidates its data cache after blocking on
sillyrename.

Also fix a typo in nfs_do_call_unlink(): swap the ^= for an |=. The result
is the same, since we've already checked that the flag is unset, but it
makes the code more readable.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

Erez Zadok reports that certain configurations fail to build due to
schedule() TASK_[UN]INTERRUPTIBLE not being declared.  Add proper
include files to fix.

Cc: Erez Zadok <ezk@cs.sunysb.edu>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

lookup() and sillyrename() can race one another because the sillyrename()
completion cannot take the parent directory's inode->i_mutex since the
latter may be held by whoever is calling dput().

We therefore have little option but to add extra locking to ensure that
nfs_lookup() and nfs_atomic_open() do not race with the sillyrename
completion.
If somebody has looked up the sillyrenamed file in the meantime, we just
transfer the sillydelete information to the new dentry.

Please refer to the bug-report at
	http://bugzilla.linux-nfs.org/show_bug.cgi?id=150Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The lower level routines in fs/nfs/proc.c, fs/nfs/nfs3proc.c and
fs/nfs/nfs4proc.c should already be dealing with the revalidation issues.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Fix a couple of bugs:
 - Don't rely on the parent dentry still being valid when the call completes.
   Fixes a race with shrink_dcache_for_umount_subtree()

 - Don't remove the file if the filehandle has been labelled as stale.

Fix a couple of inefficiencies
 - Remove the global list of sillyrenamed files. Instead we can cache the
   sillyrename information in the dentry->d_fsdata
 - Move common code from unlink_setup/unlink_done into fs/nfs/unlink.c
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

this converts fs/nfs to kzalloc() usage.
compile tested with make allyesconfig
Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

 Shrink the RPC task structure. Instead of storing separate pointers
 for task->tk_exit and task->tk_release, put them in a structure.

 Also pass the user data pointer as a parameter instead of passing it via
 task->tk_calldata. This enables us to nest callbacks.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This is the fs/ part of the big kfree cleanup patch.

Remove pointless checks for NULL prior to calling kfree() in fs/.
Signed-off-by: NJesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!