An ULP is supposed to be able to replace a GSS rpc_auth object with
another GSS rpc_auth object using rpcauth_create().  However,
rpcauth_create() in 3.5 reliably fails with -EEXIST in this case.
This is because when gss_create() attempts to create the upcall pipes,
sometimes they are already there.  For example if a pipe FS mount
event occurs, or a previous GSS flavor was in use for this rpc_clnt.

It turns out that's not the only problem here.  While working on a
fix for the above problem, we noticed that replacing an rpc_clnt's
rpc_auth is not safe, since dereferencing the cl_auth field is not
protected in any way.

So we're deprecating the ability of rpcauth_create() to switch an
rpc_clnt's security flavor during normal operation.  Instead, let's
add a fresh API that clones an rpc_clnt and gives the clone a new
flavor before it's used.

This makes immediate use of the new __rpc_clone_client() helper.

This can be used in a similar fashion to rpcauth_create() when a
client is hunting for the correct security flavor.  Instead of
replacing an rpc_clnt's security flavor in a loop, the ULP replaces
the whole rpc_clnt.

To fix the -EEXIST problem, any ULP logic that relies on replacing
an rpc_clnt's rpc_auth with rpcauth_create() must be changed to use
this API instead.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

rpc_clone_client() does most of the same tasks as rpc_new_client(),
so there is an opportunity for code re-use.  Create a generic helper
that makes it easy to clone an RPC client while replacing any of the
clnt's parameters.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Implement the new swapfile a_ops for NFS and hook up ->direct_IO.  This
will set the NFS socket to SOCK_MEMALLOC and run socket reconnect under
PF_MEMALLOC as well as reset SOCK_MEMALLOC before engaging the protocol
->connect() method.

PF_MEMALLOC should allow the allocation of struct socket and related
objects and the early (re)setting of SOCK_MEMALLOC should allow us to
receive the packets required for the TCP connection buildup.

[jlayton@redhat.com: Restore PF_MEMALLOC task flags in all cases]
[dfeng@redhat.com: Fix handling of multiple swap files]
[a.p.zijlstra@chello.nl: Original patch]
Signed-off-by: NMel Gorman <mgorman@suse.de>
Acked-by: NRik van Riel <riel@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Eric B Munson <emunson@mgebm.net>
Cc: Eric Paris <eparis@redhat.com>
Cc: James Morris <jmorris@namei.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Mike Christie <michaelc@cs.wisc.edu>
Cc: Neil Brown <neilb@suse.de>
Cc: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Xiaotian Feng <dfeng@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add a missing set of braces that commit 4e0038b6
("SUNRPC: Move clnt->cl_server into struct rpc_xprt")
forgot.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@vger.kernel.org [>= 3.4]

Fix incorrect start markers, wrapped summary lines, missing section
breaks, incorrect separators, and some name mismatches.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

xprt_alloc_slot will call rpc_delay() to make the task wait a bit before
retrying when it gets back an -ENOMEM error from xprt_dynamic_alloc_slot.
The problem is that rpc_delay will clear the task->tk_status, causing
call_reserveresult to abort the task.

The solution is simply to let call_reserveresult handle the ENOMEM error
directly.
Reported-by: NJeff Layton <jlayton@redhat.com>
Cc: stable@vger.kernel.org [>= 3.1]
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This allows comparing hash and len in one operation on 64-bit
architectures.  Right now only __d_lookup_rcu() takes advantage of this,
since that is the case we care most about.

The use of anonymous struct/unions hides the alternate 64-bit approach
from most users, the exception being a few cases where we initialize a
'struct qstr' with a static initializer.  This makes the problematic
cases use a new QSTR_INIT() helper function for that (but initializing
just the name pointer with a "{ .name = xyzzy }" initializer remains
valid, as does just copying another qstr structure).
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Now that the rpc client is namespace aware, it needs to use the
utsname of the process that created it instead of using the
init_utsname. Both rpc_new_client and rpc_clone_client need to
be fixed.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: Stanislav Kinsbursky <skinsbursky@parallels.com>

v2: recursion was replaced by loop

If client is a clone, then it's parent can not be in the list.
But parent's Pipefs dentries have to be created and destroyed.

Note: event skip helper for clients introduced
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

1) This is sane.
2) Otherwise there will be soft lockup:

do {
	rpc_get_client_for_event (clnt->cl_dentry == NULL ==> choose)
	__rpc_pipefs_event (clnt->cl_program->pipe_dir_name == NULL ==> return)
} while (1)
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

These clients can't be safely dereferenced if their counter in 0.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Remove all #inclusions of asm/system.h preparatory to splitting and killing
it.  Performed with the following command:

perl -p -i -e 's!^#\s*include\s*<asm/system[.]h>.*\n!!' `grep -Irl '^#\s*include\s*<asm/system[.]h>' *`
Signed-off-by: NDavid Howells <dhowells@redhat.com>

NFSv4.0 clients must send endpoint information for their callback
service to NFSv4.0 servers during their first contact with a server.
Traditionally on Linux, user space provides the callback endpoint IP
address via the "clientaddr=" mount option.

During an NFSv4 migration event, it is possible that an FSID may be
migrated to a destination server that is accessible via a different
source IP address than the source server was.  The client must update
callback endpoint information on the destination server so that it can
maintain leases and allow delegation.

Without a new "clientaddr=" option from user space, however, the
kernel itself must construct an appropriate IP address for the
callback update.  Provide an API in the RPC client for upper layer
RPC consumers to acquire a source address for a remote.

The mechanism used by the mount.nfs command is copied: set up a
connected UDP socket to the designated remote, then scrape the source
address off the socket.  We are careful to select the correct network
namespace when setting up the temporary UDP socket.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

When the cl_xprt field is updated, the cl_server field will also have
to change.  Since the contents of cl_server follow the remote endpoint
of cl_xprt, just move that field to the rpc_xprt.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
[ cel: simplify check_gss_callback_principal(), whitespace changes ]
[ cel: forward ported to 3.4 ]
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

A migration event will replace the rpc_xprt used by an rpc_clnt.  To
ensure this can be done safely, all references to cl_xprt must now use
a form of rcu_dereference().

Special care is taken with rpc_peeraddr2str(), which returns a pointer
to memory whose lifetime is the same as the rpc_xprt.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
[ cel: fix lockdep splats and layering violations ]
[ cel: forward ported to 3.4 ]
[ cel: remove rpc_max_reqs(), add rpc_net_ns() ]
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

v3:
1) Lookup for client is performed from the beginning of the list on each PipeFS
event handling operation.

Lockdep is sad otherwise, because inode mutex is taken on PipeFS dentry
creation, which can be called on mount notification, where this per-net client
lock is taken on clients list walk.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This patch adds three trace points to the status routines
in the sunrpc state machine.

The goal of these trace points is to give an Admin
the ability to check on binding status or connection
status to see if there is a potential problem.
Signed-off-by: NSteve Dickson <steved@redhat.com>
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>

...and get rid of the superfluous cl_inline_name.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

v2:
1) Added "nfs_idmap_init" and "nfs_idmap_quit" definitions for kernels built
without CONFIG_NFS_V4 option set.

This patch subscribes NFS clients to RPC pipefs notifications. Idmap notifier
is registering on NFS module load. This notifier callback is responsible for
creation/destruction of PipeFS idmap pipe dentry for NFS4 clients.

Since ipdmap pipe is created in rpc client pipefs directory, we have make sure,
that this directory has been created already. IOW RPC client notifier callback
has been called already. To achive this, PipeFS notifier priorities has been
introduced (RPC clients notifier priority is greater than NFS idmap one).
But this approach gives another problem: unlink for RPC client directory will
be called before NFS idmap pipe unlink on UMOUNT event and will fail, because
directory is not empty.
The solution, introduced in this patch, is to try to remove client directory
once again after idmap pipe was unlinked. This looks like ugly hack, so
probably it should be replaced in some more elegant way.

Note that no locking required in notifier callback because PipeFS superblock
pointer is passed as an argument from it's creation or destruction routine and
thus we can be sure about it's validity.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This is a cleanup patch. We don't need this reference anymore.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

v2:
1) Updated due to changes in the first patch of the series.

Now, with RPC pipefs mount notifications handling in RPC clients, we can remove
mount point creation and destruction. RPC clients dentries will be created on
PipeFS mount event and removed on umount event.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Without this patch we have races:

rpc_fill_super				rpc_free_client
rpc_pipefs_event(MOUNT)			rpc_remove_pipedir
spin_lock(&rpc_client_lock);
rpc_setup_pipedir_sb
spin_unlock(&rpc_client_lock);
					spin_lock(&rpc_client_lock);
					(remove from list)
					spin_unlock(&rpc_client_lock);
					MEAMORY LEAKED
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This patch subscribes RPC clients to RPC pipefs notifications. RPC clients
notifier block is registering with pipefs initialization during SUNRPC module
init.
This notifier callback is responsible for RPC client PipeFS directory and GSS
pipes creation. For pipes creation and destruction two additional callbacks
were added to struct rpc_authops.
Note that no locking required in notifier callback because PipeFS superblock
pointer is passed as an argument from it's creation or destruction routine and
thus we can be sure about it's validity.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This patch moves static SUNRPC clients list and it's lock to sunrpc_net
structure.
Currently this list is used only for debug purposes. But later it will be used
also for selecting clients by networks namespace on PipeFS mount/umount events.
Per-network namespace lists will make this faster and simplier.

Note: client list is taken from "init_net" network namespace context in
rpc_show_tasks(). This will be changed some day later with making SUNRPC
sysctl's per network namespace context.
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

v2:
1) "Over-put" of PipeFS mount point fixed. Fix is ugly, but allows to bisect
the patch set. And it will be removed later in the series.

This patch makes RPC clients PipeFs dentries allocations in it's owner network
namespace context.
RPC client pipefs dentries creation logic has been changed:
1) Pipefs dentries creation by sb was moved to separated function, which will
be used for handling PipeFS mount notification.
2) Initial value of RPC client PipeFS dir dentry is set no NULL now.

RPC client pipefs dentries cleanup logic has been changed:
1) Cleanup is done now in separated rpc_remove_pipedir() function, which takes
care about pipefs superblock locking.

Also this patch removes slashes from cb_program.pipe_dir_name and from
NFS_PIPE_DIRNAME to make rpc_d_lookup_sb() work. This doesn't affect
vfs_path_lookup() results in nfs4blocklayout_init() since this slash is cutted
off anyway in link_path_walk().
Signed-off-by: NStanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

It can trivially be replaced with rpc_restart_call_prepare.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

Change explicit references to CONFIG_NFS_V4_1 to implicit ones
Get rid of the unnecessary defines in backchannel_rqst.c and
bc_svc.c: the Makefile takes care of those dependency.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

common dprint_status() macro is used in all callbacks but not in call_decode()
Signed-off-by: NVasily Averin <vvs@sw.ru>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Make the case labels the same indent as the switch.

git diff -w shows 80 column line reflowing.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix a couple of instances where we were exiting the RPC client on
arbitrary signals. We should only do so on fatal signals.

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

If the NLM daemon is killed on the NFS server, we can currently end up
hanging forever on an 'unlock' request, instead of aborting. Basically,
if the rpcbind request fails, or the server keeps returning garbage, we
really want to quit instead of retrying.
Tested-by: NVasily Averin <vvs@sw.ru>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org

TI-RPC introduces the capability of performing RPC over AF_LOCAL
sockets.  It uses this mainly for registering and unregistering
local RPC services securely with the local rpcbind, but we could
also conceivably use it as a generic upcall mechanism.

This patch provides a client-side only implementation for the moment.
We might also consider a server-side implementation to provide
AF_LOCAL access to NLM (for statd downcalls, and such like).

Autobinding is not supported on kernel AF_LOCAL transports at this
time.  Kernel ULPs must specify the pathname of the remote endpoint
when an AF_LOCAL transport is created.  rpcbind supports registering
services available via AF_LOCAL, so the kernel could handle it with
some adjustment to ->rpcbind and ->set_port.  But we don't need this
feature for doing upcalls via well-known named sockets.

This has not been tested with ULPs that move a substantial amount of
data.  Thus, I can't attest to how robust the write_space and
congestion management logic is.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Clean up.  The documenting comment at the top of net/sunrpc/clnt.c is
out of date.  We adopted BSD's RTO estimation mechanism years ago.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Clean up.  Preferred style is not to use curly braces around
switch cases.  I'm about to add another case that needs a third
type cast.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

On occasion, it is useful for the NFS layer to distinguish between
soft timeouts and other EIO errors due to (say) encoding errors,
or authentication errors.

The following patch ensures that the default behaviour of the RPC
layer remains to return EIO on soft timeouts (until we have
audited all the callers).
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

RPC task RPC_TASK_QUEUED bit is set must be checked before trying to wake up
task rpc_killall_tasks() because task->tk_waitqueue can not be set (equal to
NULL).
Also, as Trond Myklebust mentioned, such approach (instead of checking
tk_waitqueue to NULL) allows us to "optimise away the call to
rpc_wake_up_queued_task() altogether for those
tasks that aren't queued".

Here is an example of dereferencing of tk_waitqueue equal to NULL:

CPU 0               	CPU 1				CPU 2
--------------------	---------------------	--------------------------
nfs4_run_open_task
rpc_run_task
rpc_execute
rpc_set_active
rpc_make_runnable
(waiting)
			rpc_async_schedule
			nfs4_open_prepare
			nfs_wait_on_sequence
						nfs_umount_begin
						rpc_killall_tasks
						rpc_wake_up_task
						rpc_wake_up_queued_task
						spin_lock(tk_waitqueue == NULL)
						BUG()
			rpc_sleep_on
			spin_lock(&q->lock)
			__rpc_sleep_on
			task->tk_waitqueue = q
Signed-off-by: NStanislav Kinsbursky <skinsbursky@openvz.org>
Cc: stable@kernel.org
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Use our own async error handler.
Mark the layout as failed and retry i/o through the MDS on specified errors.

Update the mds_offset in nfs_readpage_retry so that a failed short-read retry
to a DS gets correctly resent through the MDS.
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>