Pull nfsd updates from Chuck Lever:
 "The big ticket item for this release is that support for RPC-with-TLS
  [RFC 9289] has been added to the Linux NFS server.

  The goal is to provide a simple-to-deploy, low-overhead in-transit
  confidentiality and peer authentication mechanism. It can supplement
  NFS Kerberos and it can protect the use of legacy non-cryptographic
  user authentication flavors such as AUTH_SYS. The TLS Record protocol
  is handled entirely by kTLS, meaning it can use either software
  encryption or offload encryption to smart NICs.

  Aside from that, work continues on improving NFSD's open file cache.
  Among the many clean-ups in that area is a patch to convert the
  rhashtable to use the list-hashing version of that data structure"

* tag 'nfsd-6.4' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (31 commits)
  NFSD: Handle new xprtsec= export option
  SUNRPC: Support TLS handshake in the server-side TCP socket code
  NFSD: Clean up xattr memory allocation flags
  NFSD: Fix problem of COMMIT and NFS4ERR_DELAY in infinite loop
  SUNRPC: Clear rq_xid when receiving a new RPC Call
  SUNRPC: Recognize control messages in server-side TCP socket code
  SUNRPC: Be even lazier about releasing pages
  SUNRPC: Convert svc_xprt_release() to the release_pages() API
  SUNRPC: Relocate svc_free_res_pages()
  nfsd: simplify the delayed disposal list code
  SUNRPC: Ignore return value of ->xpo_sendto
  SUNRPC: Ensure server-side sockets have a sock->file
  NFSD: Watch for rq_pages bounds checking errors in nfsd_splice_actor()
  sunrpc: simplify two-level sysctl registration for svcrdma_parm_table
  SUNRPC: return proper error from get_expiry()
  lockd: add some client-side tracepoints
  nfs: move nfs_fhandle_hash to common include file
  lockd: server should unlock lock if client rejects the grant
  lockd: fix races in client GRANTED_MSG wait logic
  lockd: move struct nlm_wait to lockd.h
  ...

Pull NFS client updates from Anna Schumaker:
 "New Features:

   - Convert the readdir path to use folios

   - Convert the NFS fscache code to use netfs

  Bugfixes and Cleanups:

   - Always send a RECLAIM_COMPLETE after establishing a lease

   - Simplify sysctl registrations and other cleanups

   - Handle out-of-order write replies on NFS v3

   - Have sunrpc call_bind_status use standard hard/soft task semantics

   - Other minor cleanups"

* tag 'nfs-for-6.4-1' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  NFSv4.2: Rework scratch handling for READ_PLUS
  NFS: Cleanup unused rpc_clnt variable
  NFS: set varaiable nfs_netfs_debug_id storage-class-specifier to static
  SUNRPC: remove the maximum number of retries in call_bind_status
  NFS: Convert readdir page array functions to use a folio
  NFS: Convert the readdir array-of-pages into an array-of-folios
  NFSv3: handle out-of-order write replies.
  NFS: Remove fscache specific trace points and NFS_INO_FSCACHE bit
  NFS: Remove all NFSIOS_FSCACHE counters due to conversion to netfs API
  NFS: Convert buffered read paths to use netfs when fscache is enabled
  NFS: Configure support for netfs when NFS fscache is configured
  NFS: Rename readpage_async_filler to nfs_read_add_folio
  sunrpc: simplify one-level sysctl registration for debug_table
  sunrpc: move sunrpc_table and proc routines above
  sunrpc: simplify one-level sysctl registration for xs_tunables_table
  sunrpc: simplify one-level sysctl registration for xr_tunables_table
  nfs: simplify two-level sysctl registration for nfs_cb_sysctls
  nfs: simplify two-level sysctl registration for nfs4_cb_sysctls
  lockd: simplify two-level sysctl registration for nlm_sysctls
  NFSv4.1: Always send a RECLAIM_COMPLETE after establishing lease

This patch adds opportunitistic RPC-with-TLS to the Linux in-kernel
NFS server. If the client requests RPC-with-TLS and the user space
handshake agent is running, the server will set up a TLS session.

There are no policy settings yet. For example, the server cannot
yet require the use of RPC-with-TLS to access its data.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

This is an eye-catcher for tracepoints that record the XID: it means
svc_rqst() has not received a full RPC Call with an XID yet.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

To support kTLS, the server-side TCP socket receive path needs to
watch for CMSGs.

Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

A single RPC transaction that touches only a couple of pages means
rq_pvec will not be even close to full in svc_xpt_release(). This is
a common case.

Instead, just leave the pages in rq_pvec until it is completely
full. This improves the efficiency of the batch release mechanism
on workloads that involve small RPC messages.

The rq_pvec is also fully emptied just before thread exit.

Reviewed-by: Calum Mackay <calum.mackay@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Instead of invoking put_page() one-at-a-time, pass the "response"
portion of rq_pages directly to release_pages() to reduce the number
of times each nfsd thread invokes a page allocator API.

Since svc_xprt_release() is not invoked while a client is waiting
for an RPC Reply, this is not expected to directly impact mean
request latencies on a lightly or moderately loaded server. However
as workload intensity increases, I expect somewhat better
scalability: the same number of server threads should be able to
handle more work.

Reviewed-by: Calum Mackay <calum.mackay@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Clean-up: There doesn't seem to be a reason why this function is
stuck in a header. One thing it prevents is the convenient addition
of tracing. Moving it to a source file also makes the rq_respages
clean-up logic easier to find.

Reviewed-by: Calum Mackay <calum.mackay@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

Clean up: All callers of svc_process() ignore its return value, so
svc_process() can safely be converted to return void. Ditto for
svc_send().

The return value of ->xpo_sendto() is now used only as part of a
trace event.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

The TLS handshake upcall mechanism requires a non-NULL sock->file on
the socket it hands to user space. svc_sock_free() already releases
sock->file properly if one exists.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>