Running a multi-threaded 8KB fio test (70/30 mix), three or four out
of twelve of the jobs fail when using krb5i. The failure is an EIO
on a read.

Troubleshooting confirmed the EIO results when the client fails to
verify the MIC of an NFS READ reply. Bruce suggested the problem
could be due to the data payload changing between the time the
reply's MIC was computed on the server and the time the reply was
actually sent.

krb5p gets around this problem by disabling RQ_SPLICE_OK. Use the
same mechanism for krb5i RPCs.

"iozone -i0 -i1 -s128m -y1k -az -I", export is tmpfs, mount is
sec=krb5i,vers=3,proto=rdma. The important numbers are the
read / reread column.

Here's without the RQ_SPLICE_OK patch:

              kB  reclen    write  rewrite    read    reread
          131072       1     7546     7929     8396     8267
          131072       2    14375    14600    15843    15639
          131072       4    19280    19248    21303    21410
          131072       8    32350    31772    35199    34883
          131072      16    36748    37477    49365    51706
          131072      32    55669    56059    57475    57389
          131072      64    74599    75190    74903    75550
          131072     128    99810   101446   102828   102724
          131072     256   122042   122612   124806   125026
          131072     512   137614   138004   141412   141267
          131072    1024   146601   148774   151356   151409
          131072    2048   180684   181727   293140   292840
          131072    4096   206907   207658   552964   549029
          131072    8192   223982   224360   454493   473469
          131072   16384   228927   228390   654734   632607

And here's with it:

              kB  reclen    write  rewrite    read    reread
          131072       1     7700     7365     7958     8011
          131072       2    13211    13303    14937    14414
          131072       4    19001    19265    20544    20657
          131072       8    30883    31097    34255    33566
          131072      16    36868    34908    51499    49944
          131072      32    56428    55535    58710    56952
          131072      64    73507    74676    75619    74378
          131072     128   100324   101442   103276   102736
          131072     256   122517   122995   124639   124150
          131072     512   137317   139007   140530   140830
          131072    1024   146807   148923   151246   151072
          131072    2048   179656   180732   292631   292034
          131072    4096   206216   208583   543355   541951
          131072    8192   223738   224273   494201   489372
          131072   16384   229313   229840   691719   668427

I would say that there is not much difference in this test.

For good measure, here's the same test with sec=krb5p:

              kB  reclen    write  rewrite    read    reread
          131072       1     5982     5881     6137     6218
          131072       2    10216    10252    10850    10932
          131072       4    12236    12575    15375    15526
          131072       8    15461    15462    23821    22351
          131072      16    25677    25811    27529    27640
          131072      32    31903    32354    34063    33857
          131072      64    42989    43188    45635    45561
          131072     128    52848    53210    56144    56141
          131072     256    59123    59214    62691    62933
          131072     512    63140    63277    66887    67025
          131072    1024    65255    65299    69213    69140
          131072    2048    76454    76555   133767   133862
          131072    4096    84726    84883   251925   250702
          131072    8192    89491    89482   270821   276085
          131072   16384    91572    91597   361768   336868

BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=307Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

If you attempt a TCP mount from an host that is unreachable in a way
that triggers an immediate error from kernel_connect(), that error
does not propagate up, instead EAGAIN is reported.

This results in call_connect_status receiving the wrong error.

A case that it easy to demonstrate is to attempt to mount from an
address that results in ENETUNREACH, but first deleting any default
route.
Without this patch, the mount.nfs process is persistently runnable
and is hard to kill.  With this patch it exits as it should.

The problem is caused by the fact that xs_tcp_force_close() eventually
calls
      xprt_wake_pending_tasks(xprt, -EAGAIN);
which causes an error return of -EAGAIN.  so when xs_tcp_setup_sock()
calls
      xprt_wake_pending_tasks(xprt, status);
the status is ignored.

Fixes: 4efdd92c ("SUNRPC: Remove TCP client connection reset hack")
Signed-off-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Omit an extra message for a memory allocation failure in this function.

This issue was detected by using the Coccinelle software.

Link: http://events.linuxfoundation.org/sites/events/files/slides/LCJ16-Refactor_Strings-WSang_0.pdfSigned-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

struct svc_procinfo contains function pointers, and marking it as
constant avoids it being able to be used as an attach vector for
code injections.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

pc_count is the only writeable memeber of struct svc_procinfo, which is
a good candidate to be const-ified as it contains function pointers.

This patch moves it into out out struct svc_procinfo, and into a
separate writable array that is pointed to by struct svc_version.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Drop the resp argument as it can trivially be derived from the rqstp
argument.  With that all functions now have the same prototype, and we
can remove the unsafe casting to kxdrproc_t.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Drop the argp argument as it can trivially be derived from the rqstp
argument.  With that all functions now have the same prototype, and we
can remove the unsafe casting to kxdrproc_t.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Drop the p and resp arguments as they are always NULL or can trivially
be derived from the rqstp argument.  With that all functions now have the
same prototype, and we can remove the unsafe casting to kxdrproc_t.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Drop the argp and resp arguments as they can trivially be derived from
the rqstp argument.  With that all functions now have the same prototype,
and we can remove the unsafe casting to svc_procfunc as well as the
svc_procfunc typedef itself.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

struct rpc_procinfo contains function pointers, and marking it as
constant avoids it being able to be used as an attach vector for
code injections.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

p_count is the only writeable memeber of struct rpc_procinfo, which is
a good candidate to be const-ified as it contains function pointers.

This patch moves it into out out struct rpc_procinfo, and into a
separate writable array that is pointed to by struct rpc_version and
indexed by p_statidx.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Declare the p_decode callbacks with the proper prototype instead of
casting to kxdrdproc_t and losing all type safety.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Declare the p_decode callbacks with the proper prototype instead of
casting to kxdrdproc_t and losing all type safety.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>

Pass struct rpc_request as the first argument instead of an untyped blob.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Declare the p_encode callbacks with the proper prototype instead of
casting to kxdreproc_t and losing all type safety.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Declare the p_encode callbacks with the proper prototype instead of
casting to kxdreproc_t and losing all type safety.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Acked-by: NTrond Myklebust <trond.myklebust@primarydata.com>

Pass struct rpc_request as the first argument instead of an untyped blob,
and mark the data object as const.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJeff Layton <jlayton@redhat.com>

We want to use kthread_stop() in order to ensure the threads are
shut down before we tear down the nfs_callback_info in nfs_callback_down.
Tested-and-reviewed-by: NKinglong Mee <kinglongmee@gmail.com>
Reported-by: NKinglong Mee <kinglongmee@gmail.com>
Fixes: bb6aeba7 ("NFSv4.x: Switch to using svc_set_num_threads()...")
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Refactor to separate out the functions of starting and stopping threads
so that they can be used in other helpers.
Signed-off-by: NTrond Myklebust <trond.myklebust@primarydata.com>
Tested-and-reviewed-by: NKinglong Mee <kinglongmee@gmail.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Clean up: These have been replaced and are no longer used.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

req_maps are no longer used by the send path and can thus be removed.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Clean up. All RDMA Write completions are now handled by
svc_rdma_wc_write_ctx.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

The sge array in struct svc_rdma_op_ctxt is no longer used for
sending RDMA Write WRs. It need only accommodate the construction of
Send and Receive WRs. The maximum inline size is the largest payload
it needs to handle now.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Replace C structure-based XDR decoding with pointer arithmetic.
Pointer arithmetic is considered more portable.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Observed at Connectathon 2017.

If a client has underestimated the size of a Write or Reply chunk,
the Linux server writes as much payload data as it can, then it
recognizes there was a problem and closes the connection without
sending the transport header.

This creates a couple of problems:

<> The client never receives indication of the server-side failure,
   so it continues to retransmit the bad RPC. Forward progress on
   the transport is blocked.

<> The reply payload pages are not moved out of the svc_rqst, thus
   they can be released by the RPC server before the RDMA Writes
   have completed.

The new rdma_rw-ized helpers return a distinct error code when a
Write/Reply chunk overrun occurs, so it's now easy for the caller
(svc_rdma_sendto) to recognize this case.

Instead of dropping the connection, post an RDMA_ERROR message. The
client now sees an RDMA_ERROR and can properly terminate the RPC
transaction.

As part of the new logic, set up the same delayed release for these
payload pages as would have occurred in the normal case.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Now that svc_rdma_sendto has been renovated, svc_rdma_send_error can
be refactored to reduce code duplication and remove C structure-
based XDR encoding. It is also relocated to the source file that
contains its only caller.

This is a refactoring change only.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

The current svcrdma sendto code path posts one RDMA Write WR at a
time. Each of these Writes typically carries a small number of pages
(for instance, up to 30 pages for mlx4 devices). That means a 1MB
NFS READ reply requires 9 ib_post_send() calls for the Write WRs,
and one for the Send WR carrying the actual RPC Reply message.

Instead, use the new rdma_rw API. The details of Write WR chain
construction and memory registration are taken care of in the RDMA
core. svcrdma can focus on the details of the RPC-over-RDMA
protocol. This gives three main benefits:

1. All Write WRs for one RDMA segment are posted in a single chain.
As few as one ib_post_send() for each Write chunk.

2. The Write path can now use FRWR to register the Write buffers.
If the device's maximum page list depth is large, this means a
single Write WR is needed for each RPC's Write chunk data.

3. The new code introduces support for RPCs that carry both a Write
list and a Reply chunk. This combination can be used for an NFSv4
READ where the data payload is large, and thus is removed from the
Payload Stream, but the Payload Stream is still larger than the
inline threshold.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

The plan is to replace the local bespoke code that constructs and
posts RDMA Read and Write Work Requests with calls to the rdma_rw
API. This shares code with other RDMA-enabled ULPs that manages the
gory details of buffer registration and posting Work Requests.

Some design notes:

 o The structure of RPC-over-RDMA transport headers is flexible,
   allowing multiple segments per Reply with arbitrary alignment,
   each with a unique R_key. Write and Send WRs continue to be
   built and posted in separate code paths. However, one whole
   chunk (with one or more RDMA segments apiece) gets exactly
   one ib_post_send and one work completion.

 o svc_xprt reference counting is modified, since a chain of
   rdma_rw_ctx structs generates one completion, no matter how
   many Write WRs are posted.

 o The current code builds the transport header as it is construct-
   ing Write WRs. I've replaced that with marshaling of transport
   header data items in a separate step. This is because the exact
   structure of client-provided segments may not align with the
   components of the server's reply xdr_buf, or the pages in the
   page list. Thus parts of each client-provided segment may be
   written at different points in the send path.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Replace C structure-based XDR decoding with more portable code that
instead uses pointer arithmetic.

This is a refactoring change only.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Clean up: extract the logic to save pages under I/O into a helper to
add a big documenting comment without adding clutter in the send
path.

This is a refactoring change only.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

The Send Queue depth is temporarily reduced to 1 SQE per credit. The
new rdma_rw API does an internal computation, during QP creation, to
increase the depth of the Send Queue to handle RDMA Read and Write
operations.

This change has to come before the NFSD code paths are updated to
use the rdma_rw API. Without this patch, rdma_rw_init_qp() increases
the size of the SQ too much, resulting in memory allocation failures
during QP creation.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Introduce a helper to DMA-map a reply's transport header before
sending it. This will in part replace the map vector cache.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Clean up: Move the ib_send_wr off the stack, and move common code
to post a Send Work Request into a helper.

This is a refactoring change only.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Since commit 1e465fd4 ("xprtrdma: Replace send and receive
arrays"), this field is no longer used.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

When ro_map is out of buffers, that's not a permanent error, so
don't report a problem.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Micro-optimize the receive workqueue by marking it's anchor "read-
mostly."
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Device removal is now adequately supported. Pinning the underlying
device driver to prevent removal while an NFS mount is active is no
longer necessary.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

After a device removal, enable the transport connect worker to
restore normal operation if there is another device with
connectivity to the server.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>