Clean up the various different email addresses of mine listed in the code
to a single current and valid address. As Dave says his network merges
for 2.6.28 are now done this seems a good point to send them in where
they won't risk disrupting real changes.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The RPC/RDMA connection logic could return early from reconnection
attempts, leading to additional spurious retries.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The RPC/RDMA code had a constant 5-second reconnect backoff, and
always performed it, even when re-establishing a connection to a
server after the RPC layer closed it due to being idle. Make it
an geometric backoff (up to 30 seconds), and don't delay idle
reconnect.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

The send marshaling code split a particular dprintk across two
lines, which makes it hard to extract from logfiles.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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>

The RPC/RDMA code can leak RDMA connection manager endpoints in
certain error cases on connect. Don't signal unwanted events,
and be certain to destroy any allocated qp.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The xprt_connect call path does not expect such errors as ECONNREFUSED
to be returned from failed transport connection attempts, otherwise it
translates them to EIO and signals fatal errors. For example, mount.nfs
prints simply "internal error". Translate all such errors to ENOTCONN
from RPC/RDMA to match sockets behavior.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The RPC/RDMA protocol allows clients and servers to avoid RDMA
operations for data which is purely the result of XDR padding.
On the client, automatically insert the necessary padding for
such server replies, and optionally don't marshal such chunks.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

RDMA disconnects yield an upcall from the RDMA connection manager,
which can race with rpc transport close, e.g. on ^C of a mount.
Ensure any rdma cm_id and qp are fully destroyed before continuing.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

An RPC/RDMA client cannot retransmit on an unbroken connection,
doing so violates its flow control with the server.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

This logic sets the connection parameter that configures the local device
and informs the remote peer how many concurrent incoming RDMA_READ
requests are supported. The original logic didn't really do what was
intended for two reasons:

- The max number supported by the device is typically smaller than
any one factor in the calculation used, and

- The field in the connection parameter structure where the value is
stored is a u8 and always overflows for the default settings.

So what really happens is the value requested for responder resources
is the left over 8 bits from the "desired value". If the desired value
happened to be a multiple of 256, the result was zero and it wouldn't
connect at all.

Given the above and the fact that max_requests is almost always larger
than the max responder resources supported by the adapter, this patch
simplifies this logic and simply requests the max supported by the device,
subject to a reasonable limit.

This bug was found by Jim Schutt at Sandia.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>
Acked-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Configure, detect and use "fastreg" support from IB/iWARP verbs
layer to perform RPC/RDMA memory registration.

Make FRMR the default memreg mode (will fall back if not supported
by the selected RDMA adapter).

This allows full and optimal operation over the cxgb3 adapter, and others.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Acked-by: NTom Tucker <tom@opengridcomputing.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

At transport creation, check for, and use, any local dma lkey.
Then, check that the selected memory registration mode is in fact
supported by the RDMA adapter selected for the mount. Fall back
to best alternative if not.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Acked-by: NTom Tucker <tom@opengridcomputing.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>

Refactor the memory registration and deregistration routines.
This saves stack space, makes the code more readable and prepares
to add the new FRMR registration methods.
Signed-off-by: NTom Talpey <talpey@netapp.com>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

On a system with nfs mounts, if a task unshares its mount namespace,
a oops can occur when the system is rebooted if the task is the last
to unreference the nfs mount. It will try to create a rpc request
using utsname() which has been invalidated by free_nsproxy().

The patch fixes the issue by using the global init_utsname() which is
always valid. the capability of identifying rpc clients per uts namespace
stills needs some extra work so this should not be a problem.

BUG: unable to handle kernel NULL pointer dereference at 00000004
IP: [<c024c9ab>] rpc_create+0x332/0x42f
Oops: 0000 [#1] DEBUG_PAGEALLOC

Pid: 1857, comm: uts-oops Not tainted (2.6.27-rc5-00319-g7686ad56 #4)
EIP: 0060:[<c024c9ab>] EFLAGS: 00210287 CPU: 0
EIP is at rpc_create+0x332/0x42f
EAX: 00000000 EBX: df26adf0 ECX: c0251887 EDX: 00000001
ESI: df26ae58 EDI: c02f293c EBP: dda0fc9c ESP: dda0fc2c
 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068
Process uts-oops (pid: 1857, ti=dda0e000 task=dd9a0778 task.ti=dda0e000)
Stack: c0104532 dda0fffc dda0fcac dda0e000 dda0e000 dd93b7f0 00000009 c02f2880
       df26aefc dda0fc68 c01096b7 00000000 c0266ee0 c039a070 c039a070 dda0fc74
       c012ca67 c039a064 dda0fc8c c012cb20 c03daf74 00000011 00000000 c0275c90
Call Trace:
 [<c0104532>] ? dump_trace+0xc2/0xe2
 [<c01096b7>] ? save_stack_trace+0x1c/0x3a
 [<c012ca67>] ? save_trace+0x37/0x8c
 [<c012cb20>] ? add_lock_to_list+0x64/0x96
 [<c0256fc4>] ? rpcb_register_call+0x62/0xbb
 [<c02570c8>] ? rpcb_register+0xab/0xb3
 [<c0252f4d>] ? svc_register+0xb4/0x128
 [<c0253114>] ? svc_destroy+0xec/0x103
 [<c02531b2>] ? svc_exit_thread+0x87/0x8d
 [<c01a75cd>] ? lockd_down+0x61/0x81
 [<c01a577b>] ? nlmclnt_done+0xd/0xf
 [<c01941fe>] ? nfs_destroy_server+0x14/0x16
 [<c0194328>] ? nfs_free_server+0x4c/0xaa
 [<c019a066>] ? nfs_kill_super+0x23/0x27
 [<c0158585>] ? deactivate_super+0x3f/0x51
 [<c01695d1>] ? mntput_no_expire+0x95/0xb4
 [<c016965b>] ? release_mounts+0x6b/0x7a
 [<c01696cc>] ? __put_mnt_ns+0x62/0x70
 [<c0127501>] ? free_nsproxy+0x25/0x80
 [<c012759a>] ? switch_task_namespaces+0x3e/0x43
 [<c01275a9>] ? exit_task_namespaces+0xa/0xc
 [<c0117fed>] ? do_exit+0x4fd/0x666
 [<c01181b3>] ? do_group_exit+0x5d/0x83
 [<c011fa8c>] ? get_signal_to_deliver+0x2c8/0x2e0
 [<c0102630>] ? do_notify_resume+0x69/0x700
 [<c011d85a>] ? do_sigaction+0x134/0x145
 [<c0127205>] ? hrtimer_nanosleep+0x8f/0xce
 [<c0126d1a>] ? hrtimer_wakeup+0x0/0x1c
 [<c0103488>] ? work_notifysig+0x13/0x1b
 =======================
Code: 70 20 68 cb c1 2c c0 e8 75 4e 01 00 8b 83 ac 00 00 00 59 3d 00 f0 ff ff 5f 77 63 eb 57 a1 00 80 2d c0 8b 80 a8 02 00 00 8d 73 68 <8b> 40 04 83 c0 45 e8 41 46 f7 ff ba 20 00 00 00 83 f8 21 0f 4c
EIP: [<c024c9ab>] rpc_create+0x332/0x42f SS:ESP 0068:dda0fc2c
Signed-off-by: NCedric Le Goater <clg@fr.ibm.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "Serge E. Hallyn" <serue@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

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

Despite the fact that cloned rpc clients won't have the cl_autobind flag
set, they may still find themselves calling rpcb_getport_async(). For this
to happen, it suffices for a _parent_ rpc_clnt to use autobinding, in which
case any clone may find itself triggering the !xprt_bound() case in
call_bind().

The correct fix for this is to walk back up the tree of cloned rpc clients,
in order to find the parent that 'owns' the transport, either because it
has clnt->cl_autobind set, or because it originally created the
transport...
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

Basically, try_module_get here are pretty useless. Any other module using
this API will pin sunrpc in memory due using exported symbols.
Signed-off-by: NDenis V. Lunev <den@openvz.org>
Signed-off-by: NTrond Myklebust <Trond.Myklebust@netapp.com>

The inititator/responder resources in the event have been swapped. They
no represent what the local peer would set their values to in order to
match the peer. Note that iWARP does not exchange these on the wire and
the provider is simply putting in the local device max.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Update the svc_rdma_send_error code to use the DMA LKEY which is valid
regardless of the memory registration strategy in use.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Use FRMR to map local RPC reply data. This allows RDMA_WRITE to send reply
data using a single WR. The FRMR is invalidated by linking the LOCAL_INV WR
to the RDMA_SEND message used to complete the reply.
Signed-off-by: NTom Tucker <tom@opengridcomputing.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>

WR can be submitted as linked lists of WR. Update the svc_rdma_send
routine to handle WR chains. This will be used to submit a WR that
uses an FRMR with another WR that invalidates the FRMR.
Signed-off-by: NTom Tucker <tom@opengridcomputing.com>

Update the svc_rdma_post_recv routine to use the adapter's global LKEY
instead of sc_phys_mr which is only valid when using a DMA MR.
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>

Query the device capabilities in the svc_rdma_accept function to determine
what advanced memory management capabilities are supported by the device.
Based on the query, select the most secure model available given the
requirements of the transport and capabilities of the adapter.
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>

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>

The RPCB XDR functions are used for multiple procedures.  For instance,
rpcb_encode_getaddr() is used for RPCB_GETADDR, RPCB_SET, and
RPCB_UNSET.  Make the XDR debug messages more generic so they are less
confusing.

And, unlike in other RPC consumers in the kernel, a single debug flag
enables all levels of debug messages in the RPC bind client, including
XDR debug messages.  Since the XDR decoders already report success or
failure in this case, remove redundant debug messages in the mid-level
rpcb_register_call() function.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

With the new rpcbind code, a PMAP_UNSET will not have any effect on
services registered via rpcbind v3 or v4.

Implement a version of svc_unregister() that uses an RPCB_UNSET with
an empty netid string to make sure we have cleared *all* entries for
a kernel RPC service when shutting down, or before starting a fresh
instance of the service.

Use the new version only when CONFIG_SUNRPC_REGISTER_V4 is enabled;
otherwise, the legacy PMAP version is used to ensure complete
backwards-compatibility with the Linux portmapper daemon.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

Clean up: When doing an RPCB_SET, make the kernel's rpcb client use the
shorthand "::" for the universal form of the IPv6 ANY address.

Without this patch, rpcbind will advertise:

  0000:0000:0000:0000:0000:0000:0000:0000.x.y

This is cosmetic only.  It cleans up the display of information from
/sbin/rpcinfo.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>

TI-RPC is a user-space library of RPC functions that replaces ONC RPC
and allows RPC to operate in the new world of IPv6.

TI-RPC combines the concept of a transport protocol (UDP and TCP)
and a protocol family (PF_INET and PF_INET6) into a single identifier
called a "netid."  For example, "udp" means UDP over IPv4, and "udp6"
means UDP over IPv6.

For rpcbind, then, the RPC service tuple that is registered and
advertised is:

  [RPC program, RPC version, service address and port, netid]

instead of

  [RPC program, RPC version, port, protocol]

Service address is typically ANYADDR, but can be a specific address
of one of the interfaces on a multi-homed host.  The third item in
the new tuple is expressed as a universal address.

The current Linux rpcbind implementation registers a netid for both
protocol families when RPCB_SET is done for just the PF_INET6 version
of the netid (ie udp6 or tcp6).  So registering "udp6" causes a
registration for "udp" to appear automatically as well.

We've recently determined that this is incorrect behavior.  In the
TI-RPC world, "udp6" is not meant to imply that the registered RPC
service handles requests from AF_INET as well, even if the listener
socket does address mapping.  "udp" and "udp6" are entirely separate
capabilities, and must be registered separately.

The Linux kernel, unlike TI-RPC, leverages address mapping to allow a
single listener socket to handle requests for both AF_INET and AF_INET6.
This is still OK, but the kernel currently assumes registering "udp6"
will cover "udp" as well.  It registers only "udp6" for it's AF_INET6
services, even though they handle both AF_INET and AF_INET6 on the same
port.

So svc_register() actually needs to register both "udp" and "udp6"
explicitly (and likewise for TCP).  Until rpcbind is fixed, the
kernel can ignore the return code for the second RPCB_SET call.

Please merge this with commit 15231312:

    SUNRPC: Support IPv6 when registering kernel RPC services
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Cc: Olaf Kirch <okir@suse.de>
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>

Create a separate server-level interface for unregistering RPC services.

The mechanics of, and the API for, registering and unregistering RPC
services will diverge further as support for IPv6 is added.
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>

My plan is to use an AF_INET listener on systems that support only IPv4,
and an AF_INET6 listener on systems that can support IPv6. Incoming
IPv4 packets will be posted to an AF_INET6 listener with a mapped IPv4
address.

Max Matveev <makc@sgi.com> says:
  Creating a single listener can be dangerous - if net.ipv6.bindv6only
  is enabled then it's possible to create another listener in v4
  namespace on the same port and steal the traffic from the "unifed"
  listener. You need to disable V6ONLY explicitly via a sockopt to stop
  that.

Set appropriate socket option on RPC server listener sockets to prevent
this.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@citi.umich.edu>