It has been exceptionally useful to exercise the logic that handles
local immediate errors and RDMA connection loss.  To enable
developers to test this regularly and repeatably, add logic to
simulate connection loss every so often.

Fault injection is disabled by default. It is enabled with

  $ sudo echo xxx > /sys/kernel/debug/sunrpc/inject_fault/disconnect

where "xxx" is a large positive number of transport method calls
before a disconnect. A value of several thousand is usually a good
number that allows reasonable forward progress while still causing a
lot of connection drops.

These hooks are disabled when SUNRPC_DEBUG is turned off.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

RDMA xprts don't have a sock_xprt, but an rdma_xprt, so the
xs_swapper_enable/disable functions will likely oops when fed an RDMA
xprt. Turn these functions into rpc_xprt_ops so that that doesn't
occur. For now the RDMA versions are no-ops that just return -EINVAL
on an attempt to swapon.

Cc: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJeff Layton <jeff.layton@primarydata.com>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Split xs_swapper into enable/disable functions and eliminate the
"enable" flag.

Currently, it's racy if you have multiple swapon/swapoff operations
running in parallel over the same xprt. Also fix it so that we only
set it to a memalloc socket on a 0->1 transition and only clear it
on a 1->0 transition.

Cc: Mel Gorman <mgorman@suse.de>
Signed-off-by: NJeff Layton <jeff.layton@primarydata.com>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

We need to allow the server to send a new request immediately after we've
replied to the previous one. Right now, there is a window between the
send and the release of the old request in rpc_put_task(), where the
server could send us a new backchannel RPC call, and we have no
request to service it.
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Instead we rely on SO_REUSEPORT to provide the reconnection semantics
that we need for NFSv2/v3.
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

The socket lock is currently held by the task that is requesting the
connection be established. While that is efficient in the case where
the connection happens quickly, it is racy in the case where it doesn't.
What we really want is for the connect helper to be able to block access
to the socket while it is being set up.

This patch does so by arranging to transfer the socket lock from the
task that is requesting the connect attempt, and then releasing that
lock once everything is done.
This scheme also gives us automatic protection against collisions with
the RPC close code, so we can kill the cancel_delayed_work_sync()
call in xs_close().
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Add a new directory heirarchy under the debugfs sunrpc/ directory:

    sunrpc/
        rpc_xprt/
            <xprt id>/

Within that directory, we can put files that give info about the
xprts. We do have the (minor) problem that there is no succinct,
unique identifier for rpc_xprts. So we generate them synthetically
with a static atomic_t counter.

For now, this directory just holds an "info" file, but we may add
other files to it in the future.
Signed-off-by: NJeff Layton <jlayton@primarydata.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

When aborting a connection to preserve source ports, don't wake the task in
xs_error_report.  This allows tasks with RPC_TASK_SOFTCONN to succeed if the
connection needs to be re-established since it preserves the task's status
instead of setting it to the status of the aborting kernel_connect().

This may also avoid a potential conflict on the socket's lock.
Signed-off-by: NBenjamin Coddington <bcodding@redhat.com>
Cc: stable@vger.kernel.org # 3.14+
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

I would like to use one of the RPC client's congestion algorithm
constants in transport-specific code.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Mostly scripted conversion of the smp_mb__* barriers.
Signed-off-by: NPeter Zijlstra <peterz@infradead.org>
Acked-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/n/tip-55dhyhocezdw1dg7u19hmh1u@git.kernel.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-arch@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>

Besides checking rpc_xprt out of xs_setup_bc_tcp,
increase it's reference (it's important).
Signed-off-by: NKinglong Mee <kinglongmee@gmail.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

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

In the gss-proxy case we don't want to have to reconnect at random--we
want to connect only on gss-proxy startup when we can steal gss-proxy's
context to do the connect in the right namespace.

So, provide a flag that allows the rpc_create caller to turn off the
idle timeout.
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

This is mainly for use by NFSv4.1, where the session negotiation
ultimately wants to decide how many RPC slots we can fill.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This patch ensures that we throttle new RPC requests if there are
requests already waiting in the xprt->backlog queue. The reason for
doing this is to fix livelock issues that can occur when an existing
(high priority) task is waiting in the backlog queue, gets woken up
by xprt_free_slot(), but a new task then steals the slot.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Avoid access to task->tk_xprt
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Avoid another RCU dereference by passing the pointer to struct rpc_xprt
from the caller.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

It is only set after everyone has dereferenced the transport,
and serves no useful purpose: setting it is racy, so all the
socket code, etc still needs to be able to cope with the cases
where they miss reading it.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Commit 43cedbf0 (SUNRPC: Ensure that
we grab the XPRT_LOCK before calling xprt_alloc_slot) is causing
hangs in the case of NFS over UDP mounts.

Since neither the UDP or the RDMA transport mechanism use dynamic slot
allocation, we can skip grabbing the socket lock for those transports.
Add a new rpc_xprt_op to allow switching between the TCP and UDP/RDMA
case.

Note that the NFSv4.1 back channel assigns the slot directly
through rpc_run_bc_task, so we can ignore that case.
Reported-by: NDick Streefland <dick.streefland@altium.nl>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@vger.kernel.org [>= 3.1]

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>

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>

With static RPC slots, the xprt backlog queue stats were useful in showing
when the transport (TCP) was starved by lack of RPC slots. The new dynamic
RPC slot code, commit d9ba131d, always
provides an RPC slot and so only uses the xprt backlog queue when the
tcp_max_slot_table_entries value has been hit or when an allocation error
occurs. All requests are now placed on the xprt sending or pending queue which
need to be monitored for debugging.

The max_slot stat shows the maximum number of dynamic RPC slots reached which is
useful when debugging performance issues.

Add the new fields at the end of the mountstats xprt stanza so that mountstats
outputs the previous correct values and ignores the new fields. Bump
NFS_IOSTATS_VERS.
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Since the scheme of limiting the number of TCP slots to whatever will
fit in the current TCP window seems to be working well (Andy reports
getting within 20% of the 'iperf' send performance on a 10GigE link)
we should just let that be the default mode of operation.

Users may still set their own limits using the tcp_max_slot_table_entries
parameter if they need to.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

Allow the number of available slots to grow with the TCP window size.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

This throttles the allocation of new slots when the socket is busy
reconnecting and/or is out of buffer space.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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>

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>

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

This allows us to reuse the xprt associated with a server connection if
one has already been set up.
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

The net is known from the xprt_create and this tagging will also
give un the context in the conntection workers where real sockets
are created.
Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

This will allow us to save the original generic cred in rpc_message, so
that if we migrate from one server to another, we can generate a new bound
cred without having to punt back to the NFS layer.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

It seems strange to maintain stats for bytes_sent in one structure, and
bytes received in another. Try to assemble all the RPC request-related
stats in struct rpc_rqst
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Currently RPC performance metrics that tabulate elapsed time use
jiffies time values.  This is problematic on systems that use slow
jiffies (for instance 100HZ systems built for paravirtualized
environments).  It is also a problem for computing precise latency
statistics for advanced network transports, such as InfiniBand,
that can have round-trip latencies significanly faster than a single
clock tick.

For the RPC client, adopt the high resolution time stamp mechanism
already used by the network layer and blktrace: ktime.

We use ktime format time stamps for all internal computations, and
convert to milliseconds for presentation.  As a result, we need only
addition operations in the performance critical paths; multiply/divide
is required only for presentation.

We could report RTT metrics in microseconds.  In fact the mountstats
format is versioned to accomodate exactly this kind of interface
improvement.

For now, however, we'll stay with millisecond precision for
presentation to maintain backwards compatibility with the handful of
currently deployed user space tools.  At a later point, we'll move to
an API such as BDI_STATS where a finer timestamp precision can be
reported.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>