Use no more than 1/128th of the number of free pages at nfsd startup for the
v4.1 DRC.

This is an arbitrary default which should probably end up under the control
of an administrator.
Signed-off-by: NAndy Adamson <andros@netapp.com>
[moved added fields in struct svc_serv under CONFIG_NFSD_V4_1]
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[fix set_max_drc calculation of sv_drc_max_pages]
[moved NFSD_DRC_SIZE_SHIFT's declaration up in header file]
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

On an NFSv4.1 server cache miss that causes an upcall, NFS4ERR_DELAY will be
returned. It is up to the NFSv4.1 client to resend only the operations that
have not been processed.

Initialize rq_usedeferral to 1 in svc_process(). It sill be turned off in
nfsd4_proc_compound() only when NFSv4.1 Sessions are used.

Note: this isn't an adequate solution on its own. It's acceptable as a way
to get some minimal 4.1 up and working, but we're going to have to find a
way to avoid returning DELAY in all common cases before 4.1 can really be
considered ready.
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[nfsd41: reverse rq_nodeferral negative logic]
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
[sunrpc: initialize rq_usedeferral]
Signed-off-by: NAndy Adamson <andros@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Since an RPC service listener's protocol family is specified now via
svc_create_xprt(), it no longer needs to be passed to svc_create() or
svc_create_pooled().  Remove that argument from the synopsis of those
functions, and remove the sv_family field from the svc_serv struct.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The sv_family field is going away.  Pass a protocol family argument to
svc_create_xprt() instead of extracting the family from the passed-in
svc_serv struct.

Again, as this is a listener socket and not an address, we make this
new argument an "int" protocol family, instead of an "sa_family_t."
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The sv_family field is going away.  Instead of using sv_family, have
the svc_register() function take a protocol family argument.

Since this argument represents a protocol family, and not an address
family, this argument takes an int, as this is what is passed to
sock_create_kern().  Also make sure svc_register's helpers are
checking for PF_FOO instead of AF_FOO.  The value of [AP]F_FOO are
equivalent; this is simply a symbolic change to reflect the semantics
of the value stored in that variable.

sock_create_kern() should return EPFNOSUPPORT if the passed-in
protocol family isn't supported, but it uses EAFNOSUPPORT for this
case.  We will stick with that tradition here, as svc_register()
is called by the RPC server in the same path as sock_create_kern().
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Clean up: add documentating comment and use appropriate data types for
svc_find_xprt()'s arguments.

This also eliminates a mixed sign comparison: @port was an int, while
the return value of svc_xprt_local_port() is an unsigned short.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Clean up:  Enable the use of const arguments in higher level svc_ APIs
by adding const to the arguments of the helper functions in svc_xprt.h
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This fixes a regression against FreeBSD servers as reported by Tomas
Kasparek. Apparently when using RPC over a TCP socket, the FreeBSD servers
don't ever react to the client closing the socket, and so commit
e06799f9 (SUNRPC: Use shutdown() instead of
close() when disconnecting a TCP socket) causes the setup to hang forever
whenever the client attempts to close and then reconnect.

We break the deadlock by adding a 'linger2' style timeout to the socket,
after which, the client will abort the connection using a TCP 'RST'.

The default timeout is set to 15 seconds. A subsequent patch will put it
under user control by means of a systctl.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Clean up:  Enable the use of const arguments in higher level svc_ APIs
by adding const to the arguments of the helper functions in svc_xprt.h
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Add /proc/fs/nfsd/pool_stats to export to userspace various
statistics about the operation of rpc server thread pools.

This patch is based on a forward-ported version of
knfsd-add-pool-thread-stats which has been shipping in the SGI
"Enhanced NFS" product since 2006 and which was previously
posted:

http://article.gmane.org/gmane.linux.nfs/10375

It has also been updated thus:

 * moved EXPORT_SYMBOL() to near the function it exports
 * made the new struct struct seq_operations const
 * used SEQ_START_TOKEN instead of ((void *)1)
 * merged fix from SGI PV 990526 "sunrpc: use dprintk instead of
   printk in svc_pool_stats_*()" by Harshula Jayasuriya.
 * merged fix from SGI PV 964001 "Crash reading pool_stats before
   nfsds are started".
Signed-off-by: NGreg Banks <gnb@sgi.com>
Signed-off-by: NHarshula Jayasuriya <harshula@sgi.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Avoid overloading the CPU scheduler with enormous load averages
when handling high call-rate NFS loads.  When the knfsd bottom half
is made aware of an incoming call by the socket layer, it tries to
choose an nfsd thread and wake it up.  As long as there are idle
threads, one will be woken up.

If there are lot of nfsd threads (a sensible configuration when
the server is disk-bound or is running an HSM), there will be many
more nfsd threads than CPUs to run them.  Under a high call-rate
low service-time workload, the result is that almost every nfsd is
runnable, but only a handful are actually able to run.  This situation
causes two significant problems:

1. The CPU scheduler takes over 10% of each CPU, which is robbing
   the nfsd threads of valuable CPU time.

2. At a high enough load, the nfsd threads starve userspace threads
   of CPU time, to the point where daemons like portmap and rpc.mountd
   do not schedule for tens of seconds at a time.  Clients attempting
   to mount an NFS filesystem timeout at the very first step (opening
   a TCP connection to portmap) because portmap cannot wake up from
   select() and call accept() in time.

Disclaimer: these effects were observed on a SLES9 kernel, modern
kernels' schedulers may behave more gracefully.

The solution is simple: keep in each svc_pool a counter of the number
of threads which have been woken but have not yet run, and do not wake
any more if that count reaches an arbitrary small threshold.

Testing was on a 4 CPU 4 NIC Altix using 4 IRIX clients, each with 16
synthetic client threads simulating an rsync (i.e. recursive directory
listing) workload reading from an i386 RH9 install image (161480
regular files in 10841 directories) on the server.  That tree is small
enough to fill in the server's RAM so no disk traffic was involved.
This setup gives a sustained call rate in excess of 60000 calls/sec
before being CPU-bound on the server.  The server was running 128 nfsds.

Profiling showed schedule() taking 6.7% of every CPU, and __wake_up()
taking 5.2%.  This patch drops those contributions to 3.0% and 2.2%.
Load average was over 120 before the patch, and 20.9 after.

This patch is a forward-ported version of knfsd-avoid-nfsd-overload
which has been shipping in the SGI "Enhanced NFS" product since 2006.
It has been posted before:

http://article.gmane.org/gmane.linux.nfs/10374Signed-off-by: NGreg Banks <gnb@sgi.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

The base versions handle constant folding now, none of these headers
are exported to userspace, so the __ prefixed versions are not
necessary.
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Reviewed-by: NNeilBrown <neilb@suse.de>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Provide an api to attempt to load any necessary kernel RPC
client transport module automatically. By convention, the
desired module name is "xprt"+"transport name". For example,
when NFS mounting with "-o proto=rdma", attempt to load the
"xprtrdma" module.
Signed-off-by: NTom Talpey <tmtalpey@gmail.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

svc_check_conn_limits() attempts to prevent denial of service attacks
by having the service close old connections once it reaches a
threshold. This threshold is based on the number of threads in the
service:

	(serv->sv_nrthreads + 3) * 20

Once we reach this, we drop the oldest connections and a printk pops
to warn the admin that they should increase the number of threads.

Increasing the number of threads isn't an option however for services
like lockd. We don't want to eliminate this check entirely for such
services but we need some way to increase this limit.

This patch adds a sv_maxconn field to the svc_serv struct. When it's
set to 0, we use the current method to calculate the max number of
connections. RPC services can then set this on an as-needed basis.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

The rpc client needs to know the principal that the setclientid was done
as, so it can tell gssd who to authenticate to.
Signed-off-by: NOlga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Two principals are involved in krb5 authentication: the target, who we
authenticate *to* (normally the name of the server, like
nfs/server.citi.umich.edu@CITI.UMICH.EDU), and the source, we we
authenticate *as* (normally a user, like bfields@UMICH.EDU)

In the case of NFSv4 callbacks, the target of the callback should be the
source of the client's setclientid call, and the source should be the
nfs server's own principal.

Therefore we allow svcgssd to pass down the name of the principal that
just authenticated, so that on setclientid we can store that principal
name with the new client, to be used later on callbacks.
Signed-off-by: NOlga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

We want to transition to a new gssd upcall which is text-based and more
easily extensible.

To simplify upgrades, as well as testing and debugging, it will help if
we can upgrade gssd (to a version which understands the new upcall)
without having to choose at boot (or module-load) time whether we want
the new or the old upcall.

We will do this by providing two different pipes: one named, as
currently, after the mechanism (normally "krb5"), and supporting the
old upcall.  One named "gssd" and supporting the new upcall version.

We allow gssd to indicate which version it supports by its choice of
which pipe to open.

As we have no interest in supporting *simultaneous* use of both
versions, we'll forbid opening both pipes at the same time.

So, add a new pipe_open callback to the rpc_pipefs api, which the gss
code can use to track which pipes have been open, and to refuse opens of
incompatible pipes.

We only need this to be called on the first open of a given pipe.
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

rq_bufsize is not used.
Signed-off-by: NMike Sager <Mike.Sager@netapp.com>
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Somehow, this escaped the previous purge. There should be no need to keep
any extra locks in the XDR callbacks.

The NFS client XDR code only writes into private objects, whereas all reads
of shared objects are confined to fields that do not change, such as
filehandles...

Ditto for lockd, the NFSv2/v3 client mount code, and rpcbind.

The nfsd XDR code may require the BKL, but since it does a synchronous RPC
call from a thread that already holds the lock, that issue is moot.
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Using NIPQUAD() with NIPQUAD_FMT, %d.%d.%d.%d or %u.%u.%u.%u
can be replaced with %pI4
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add defensive timeouts to wait_for_completion() calls in RDMA
address resolution, and make them interruptible. Fix the timeout
units to milliseconds (formerly jiffies) and move to private header.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Internal RPC/RDMA structure updates in preparation for FRMR support.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Acked-by: NTom Tucker <tom@opengridcomputing.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

RPCRDMA requests that specify a read-list are fetched with RDMA_READ. Using
an FRMR to map the data sink improves NFSRDMA security on transports that
place the RDMA_READ data sink LKEY on the wire because the valid lifetime
of the MR is only the duration of the RDMA_READ. The LKEY is invalidated
when the last RDMA_READ WR completes.

Mapping the data sink also allows for very large amounts to data to be
fetched with a single WR, so if the client is also using FRMR, the entire
RPC read-list can be fetched with a single WR.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Fast Reg MR introduces a new WR type. Add a service to register the
region with the adapter and update the completion handling to support
completions with a NULL WR context.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Add services for the allocating, freeing, and unmapping Fast Reg MR. These
services will be used by the transport connection setup, send and receive
routines.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Clean up: The svc_addsock() function no longer uses its "proto"
argument, so remove it.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Add data types to track Fast Reg Memory Regions. The core data type is
svc_rdma_fastreg_mr that associates a device MR with a host kva and page
list. A field is added to the WR context to keep track of the FRMR
used to map the local memory for an RPC.

An FRMR list and spin lock are added to the transport instance to keep
track of all FRMR allocated for the transport. Also added are device
capability flags to indicate what the memory registration
capabilities are for the underlying device and whether or not fast
memory registration is supported.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

since commit ff7d9756
"nfsd: use static memory for callback program and stats"
do_probe_callback uses a static callback program
(NFS4_CALLBACK) rather than the one set in clp->cl_callback.cb_prog
as passed in by the client in setclientid (4.0)
or create_session (4.1).

This patches introduces rpc_create_args.prognumber that allows
overriding program->number when creating rpc_clnt.
Signed-off-by: NBenny Halevy <bhalevy@panasas.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Clean up: Add extra type safety and squelch a few compiler complaints
in upcoming patches.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

In order to advertise NFS-related services on IPv6 interfaces via
rpcbind, the kernel RPC server implementation must use
rpcb_v4_register() instead of rpcb_register().

A new kernel build option allows distributions to use the legacy
v2 call until they integrate an appropriate user-space rpcbind
daemon that can support IPv6 RPC services.

I tried adding some automatic logic to fall back if registering
with a v4 protocol request failed, but there are too many corner
cases.  So I just made it a compile-time switch that distributions
can throw when they've replaced portmapper with rpcbind.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Bruce suggested there's no need to expose the difference between an error
sending the PMAP_SET request and an error reply from the portmapper to
rpcb_register's callers.  The user space equivalent of rpcb_register() is
pmap_set(3), which returns a bool_t : either the PMAP set worked, or it
didn't.  Simple.

So let's remove the "*okay" argument from rpcb_register() and
rpcb_v4_register(), and simply return an error if any part of the call
didn't work.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Introduce and initialize an address family field in the svc_serv structure.

This field will determine what family to use for the service's listener
sockets and what families are advertised via the local rpcbind daemon.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

RDMA_READ completions are kept on a separate queue from the general
I/O request queue. Since a separate lock is used to protect the RDMA_READ
completion queue, a race exists between the dto_tasklet and the
svc_rdma_recvfrom thread where the dto_tasklet sets the XPT_DATA
bit and adds I/O to the read-completion queue. Concurrently, the
recvfrom thread checks the generic queue, finds it empty and resets
the XPT_DATA bit. A subsequent svc_xprt_enqueue will fail to enqueue
the transport for I/O and cause the transport to "stall".

The fix is to protect both lists with the same lock and set the XPT_DATA
bit with this lock held.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Introduce a new API to register RPC services on IPv6 interfaces to allow
the NFS server and lockd to advertise on IPv6 networks.

Unlike rpcb_register(), the new rpcb_v4_register() function uses rpcbind
protocol version 4 to contact the local rpcbind daemon.  The version 4
SET/UNSET procedures allow services to register address families besides
AF_INET, register at specific network interfaces, and register transport
protocols besides UDP and TCP.  All of this functionality is exposed via
the new rpcb_v4_register() kernel API.

A user-space rpcbind daemon implementation that supports version 4 of the
rpcbind protocol is required in order to make use of this new API.

Note that rpcbind version 3 is sufficient to support the new rpcbind
facilities listed above, but most extant implementations use version 4.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Try to make the comment here a little more clear and concise.

Also, this macro definition seems unnecessary.
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The cl_chatty flag alows us to control whether a given rpc client leaves

	"server X not responding, timed out"

messages in the syslog.  Such messages make sense for ordinary nfs
clients (where an unresponsive server means applications on the
mountpoint are probably hanging), but not for the callback client (which
can fail more commonly, with the only result just of disabling some
optimizations).

Previously cl_chatty was removed, do to lack of users; reinstate it, and
use it for the nfsd's callback client.
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Change the WR context pool to be shared across mount points. This
reduces the RDMA transport memory footprint significantly since
idle mounts don't consume WR context memory.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

The sc_read_wait queue head is no longer used. Remove it.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>