Jason tells me that a ULP cannot rely on getting an ESTABLISHED
and DISCONNECTED event pair for each connection, so transport
reference counting in the CM event handler will never be reliable.

Now that we have ib_drain_qp(), svcrdma should no longer need to
hold transport references while Sends and Receives are posted. So
remove the get/put call sites in the CM event handlers.

This eliminates a significant source of locked memory bus traffic.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

During a connection tear down, the Receive queue is flushed before
the device resources are freed. Typically, all the Receives flush
with IB_WR_FLUSH_ERR.

However, any pending successful Receives flush with IB_WR_SUCCESS,
and the server automatically posts a fresh Receive to replace the
completing one. This happens even after the connection has closed
and the RQ is drained. Receives that are posted after the RQ is
drained appear never to complete, causing a Receive resource leak.
The leaked Receive buffer is left DMA-mapped.

To prevent these late-posted recv_ctxt's from leaking, block new
Receive posting after XPT_CLOSE is set.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Ensure that show_rqstp_flags() can recognize and display the
RQ_AUTHERR flag, added in commit 83dd59a0 ("SUNRPC/nfs: Fix
return value for nfs4_callback_compound()") and the RQ_DATA flag,
added in commit ff3ac5c3 ("SUNRPC: Add a server side
per-connection limit").
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Drop the repeated word "the" in a comment.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: linux-nfs@vger.kernel.org
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Braino when converting "buf->len -=" to "buf->len = len -".

The result is under-estimation of the ralign and rslack values. On
krb5p mounts, this has caused READDIR to fail with EIO, and KASAN
splats when decoding READLINK replies.

As a result of fixing this oversight, the gss_unwrap method now
returns a buf->len that can be shorter than priv_len for small
RPC messages. The additional adjustment done in unwrap_priv_data()
can underflow buf->len. This causes the nfsd_request_too_large
check to fail during some NFSv3 operations.

Reported-by: Marian Rainer-Harbach
Reported-by: Pierre Sauter <pierre.sauter@stwm.de>
BugLink: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1886277
Fixes: 31c9590a ("SUNRPC: Add "@len" parameter to gss_unwrap()")
Reviewed-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

If the rpc_pipefs is unmounted, then the rpc_pipe->dentry becomes NULL
and dereferencing the dentry->d_sb will trigger an oops.  The only
reason we're doing that is to determine the nfsd_net, which could
instead be passed in by the caller.  So do that instead.

Fixes: 11a60d15 ("nfsd: add a "GetVersion" upcall for nfsdcld")
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

We recently fixed lease breaking so that a client's actions won't break
its own delegations.

But we still have an unnecessary self-conflict when granting
delegations: a client's own write opens will prevent us from handing out
a read delegation even when no other client has the file open for write.

Fix that by turning off the checks for conflicting opens under
vfs_setlease, and instead performing those checks in the nfsd code.

We don't depend much on locks here: instead we acquire the delegation,
then check for conflicts, and drop the delegation again if we find any.

The check beforehand is an optimization of sorts, just to avoid
acquiring the delegation unnecessarily.  There's a race where the first
check could cause us to deny the delegation when we could have granted
it.  But, that's OK, delegation grants are optional (and probably not
even a good idea in that case).
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

A single character (line break) should be put into a sequence.
Thus use the corresponding function "seq_putc()".
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Re-use the post_rw tracepoint (safely) to trace cc_info lifetime
events, including completion IDs.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

First, refactor: Dereference the svc_rdma_send_ctxt inside
svc_rdma_send() instead of at every call site.

Then, it can be passed into trace_svcrdma_post_send() to get the
proper completion ID.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Set up a completion ID in each svc_rdma_send_ctxt. The ID is used
to match an incoming Send completion to a transport and to a
previous ib_post_send().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

When recording a trace event in the Receive path, tie decoding
results and errors to an incoming Receive completion.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Set up a completion ID in each svc_rdma_recv_ctxt. The ID is used
to match an incoming Receive completion to a transport and to a
previous ib_post_recv().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

The goal is to replace CQE kernel memory addresses in completion-
related tracepoints.

Each completion ID matches an incoming Send or Receive completion
to a Completion Queue and to a previous ib_post_*(). The ID can
then be displayed in an error message or recorded in a trace
record.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Clean up: De-duplicate some code.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Clean up: De-duplicate some code.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Use these helpers in a few spots to demonstrate their use.

The remaining open-coded discriminator checks in rpcrdma will be
addressed in subsequent patches.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Pavane pour une infante défunte.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

- Use the _err naming convention instead
- Remove display of kernel memory address of the controlling xprt
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Final refactor: Replace internals of svc_rdma_send_error() with a
simple call to svc_rdma_send_error_msg().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Prepare for svc_rdma_send_error_msg() to be invoked from another
source file.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Like svc_rdma_send_error(), have svc_rdma_send_error_msg() handle
any error conditions internally, rather than duplicating that
recovery logic at every call site.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

The common "send RDMA_ERR" function should be in svc_rdma_sendto.c,
since that is where the other Send-related functions are located.
So from here, I will beef up svc_rdma_send_error_msg() and deprecate
svc_rdma_send_error().

A generic svc_rdma_send_error_msg() will need to handle both
ERR_CHUNK and ERR_VERS. Copy that logic from svc_rdma_send_error()
to svc_rdma_send_error_msg().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Another step towards making svc_rdma_send_error_msg() and
svc_rdma_send_error() similar enough to eliminate one of them.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Commit 4757d90b ("svcrdma: Report Write/Reply chunk overruns")
made an effort to preserve I/O pages until RDMA Write completion.

In a subsequent patch, I intend to de-duplicate the two functions
that send ERR_CHUNK responses. Pull the save_io_pages() call out of
svc_rdma_send_error_msg() to make it more like
svc_rdma_send_error().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Commit 07d0ff3b ("svcrdma: Clean up Read chunk path") moved the
page saver logic so that it gets executed event when an error occurs.
In that case, the I/O is never posted, and those pages are then
leaked. Errors in this path, however, are quite rare.

Fixes: 07d0ff3b ("svcrdma: Clean up Read chunk path")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Add similar tracepoints to those that were recently added on the
client side to track failures in the integ and priv unwrap paths.

And, let's collect the seqno-specific tracepoints together with a
common naming convention.

Regarding the gss_check_seq_num() changes: everywhere else treats
the GSS sequence number as an unsigned 32-bit integer. As far back
as 2.6.12, I couldn't find a compelling reason to do things
differently here. As a defensive change it's better to eliminate
needless implicit sign conversions.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Check if user extended attributes are supported for an inode,
and return the answer when being queried for file attributes.

An exported filesystem can now signal its RFC8276 user extended
attributes capability.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Implement the main entry points for the *XATTR operations.

Add functions to calculate the reply size for the user extended attribute
operations, and implement the XDR encode / decode logic for these
operations.

Add the user extended attributes operations to nfsd4_ops.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Add the structures used in extended attribute request / response handling.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Since the NFSv4.2 extended attributes extension defines 3 new access
bits for xattr operations, take them in to account when validating
what the client is asking for, and when checking permissions.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

This adds the filehandle based functions for the xattr operations
that call in to the vfs layer to do the actual work.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
[ cel: address checkpatch.pl complaint ]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Add defines for server-side extended attribute support. Most have
already been added as part of client support, but these are
the network order error codes for the noxattr and xattr2big errors,
and the addition of the xattr support to the supported file
attributes (if configured).
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

nfs4_decode_write has code to parse incoming XDR write data in to
a kvec head, and a list of pages.

Put this code in to a separate function, so that it can be used
later by the xattr code, for setxattr. No functional change.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Add a function that checks is an extended attribute namespace is
supported for an inode, meaning that a handler must be present
for either the whole namespace, or at least one synthetic
xattr in the namespace.

To be used by the nfs server code when being queried for extended
attributes support.

Cc: linux-fsdevel@vger.kernel.org
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

set/removexattr on an exported filesystem should break NFS delegations.
This is true in general, but also for the upcoming support for
RFC 8726 (NFSv4 extended attribute support). Make sure that they do.

Additionally, they need to grow a _locked variant, since callers might
call this with i_rwsem held (like the NFS server code).

Cc: stable@vger.kernel.org # v4.9+
Cc: linux-fsdevel@vger.kernel.org
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Add definitions for the new operations, errors and flags as defined
in RFC 8276 (File System Extended Attributes in NFSv4).
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Pull Xtensa fixes from Max Filippov:

 - fix __sync_fetch_and_{and,or}_4 declarations to avoid build warning

 - update *pos in cpuinfo_op.next to avoid runtime warning

 - use for_each_set_bit in xtensa_pmu_irq_handler instead of open-coding
   it

* tag 'xtensa-20200712' of git://github.com/jcmvbkbc/linux-xtensa:
  xtensa: simplify xtensa_pmu_irq_handler
  xtensa: update *pos in cpuinfo_op.next
  xtensa: fix __sync_fetch_and_{and,or}_4 declarations

Pull io_uring fixes from Jens Axboe:
 "Two late fixes again:

   - Fix missing msg_name assignment in certain cases (Pavel)

   - Correct a previous fix for full coverage (Pavel)"

* tag 'io_uring-5.8-2020-07-12' of git://git.kernel.dk/linux-block:
  io_uring: fix not initialised work->flags
  io_uring: fix missing msg_name assignment