The current mirrored read failover code is correctly resetting the mirror
index between failed reads, however it is not able to actually flip the
RPC call over to the next RPC client.
The end result is that we keep resending the RPC call to the same client
over and over.

The fix is to use the pnfs_read_resend_pnfs() mechanism to schedule a
new RPC call, but we need to add the ability to pass in a mirror
index so that we always retry the next mirror in the list.

Fixes: 166bd5b8

 ("pNFS/flexfiles: Fix layoutstats handling during read failovers")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Drop duplicated words {the, and} in comments.

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: Anna Schumaker <anna.schumaker@netapp.com>
Cc: linux-nfs@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Replace a comma between expression statements by a semicolon.

Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Move the buffer size check to decode_attr_security_label() before memcpy()
Only call memcpy() if the buffer is large enough

Fixes: aa9c2669

 ("NFS: Client implementation of Labeled-NFS")
Signed-off-by: Jeffrey Mitchell <jeffrey.mitchell@starlab.io>
[Trond: clean up duplicate test of label->len != 0]
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

If the NFS_LAYOUT_RETURN_REQUESTED flag is set, we want to return the
layout as soon as possible, meaning that the affected layout segments
should be marked as invalid, and should no longer be in use for I/O.

Fixes: f0b42981

 ("pNFS: Ignore non-recalled layouts in pnfs_layout_need_return()")
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

If the layout segment is still in use for a read or a write, we should
not move it to the layout plh_return_segs list. If we do, we can end
up returning the layout while I/O is still in progress.

Fixes: e0b7d420

 ("pNFS: Don't discard layout segments that are marked for return")
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Allow the pnfs I/O tracepoints to trace which layout segment is being
used.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Allow tracing of the NFSv4.2 layouterror and layoutstats operations.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Ensure we correctly report the stateid and status in the layoutreturn on
close tracepoint.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Current behaviour: every time a v3 operation is re-sent to the server
we update (double) the timeout. There is no distinction between whether
or not the previous timer had expired before the re-sent happened.

Here's the scenario:
1. Client sends a v3 operation
2. Server RST-s the connection (prior to the timeout) (eg., connection
is immediately reset)
3. Client re-sends a v3 operation but the timeout is now 120sec.

As a result, an application sees 2mins pause before a retry in case
server again does not reply.

Instead, this patch proposes to keep track off when the minor timeout
should happen and if it didn't, then don't update the new timeout.
Value is updated based on the previous value to make timeouts
predictable.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The NFS_CONTEXT_ERROR_WRITE flag (as well as the check of said flag) was
removed by commit 6fbda89b.  The absence of an error check allows
writes to be continually queued up for a server that may no longer be
able to handle them.  Fix it by adding an error check using the generic
error reporting functions.

Fixes: 6fbda89b

 ("NFS: Replace custom error reporting mechanism with generic one")
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

nfs_wb_all() calls filemap_write_and_wait(), which uses
filemap_check_errors() to determine the error to return.
filemap_check_errors() only looks at the mapping->flags and will
therefore only return either -ENOSPC or -EIO.  To ensure that the
correct error is returned on close(), nfs{,4}_file_flush() should call
filemap_check_wb_err() which looks at the errseq value in
mapping->wb_err without consuming it.

Fixes: 6fbda89b

 ("NFS: Replace custom error reporting mechanism with
generic one")
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Caches should be small enough to discard them inline, so do that
instead of using a work queue.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The variable result is being initialized with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

It looks like this "else" is just a typo.  It turns off nconnect for
NFSv4.0 even though it works for every other version.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

commit 0688e64b ("NFS: Allow signal interruption of NFS4ERR_DELAYed operations")
introduces nfs4_delay_interruptible which also needs an _unsafe version to
avoid the following call trace for the same reason explained in
commit 416ad3c9

 ("freezer: add unsafe versions of freezable helpers for NFS")

CPU: 4 PID: 3968 Comm: rm Tainted: G W 5.8.0-rc4 #1
Hardware name: Marvell OcteonTX CN96XX board (DT)
Call trace:
dump_backtrace+0x0/0x1dc
show_stack+0x20/0x30
dump_stack+0xdc/0x150
debug_check_no_locks_held+0x98/0xa0
nfs4_delay_interruptible+0xd8/0x120
nfs4_handle_exception+0x130/0x170
nfs4_proc_rmdir+0x8c/0x220
nfs_rmdir+0xa4/0x360
vfs_rmdir.part.0+0x6c/0x1b0
do_rmdir+0x18c/0x210
__arm64_sys_unlinkat+0x64/0x7c
el0_svc_common.constprop.0+0x7c/0x110
do_el0_svc+0x24/0xa0
el0_sync_handler+0x13c/0x1b8
el0_sync+0x158/0x180

Signed-off-by: He Zhe <zhe.he@windriver.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Better to unregister the file system before destroying the kmem_cache
cache of the inodes, so that the inodes are freed before we are trying
to destroy it. Otherwise, kmem_cache yells that some objects are live.

Signed-off-by: Dan Aloni <dan@kernelim.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Implement client side caching for NFSv4.2 extended attributes. The cache
is a per-inode hashtable, with name/value entries. There is one special
entry for the listxattr cache.

NFS inodes have a pointer to a cache structure. The cache structure is
allocated on demand, freed when the cache is invalidated.

Memory shrinkers keep the size in check. Large entries (> PAGE_SIZE)
are collected by a separate shrinker, and freed more aggressively
than others.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Now that all the lower level code is there to make the RPC calls, hook
it in to the xattr handlers and the listxattr entry point, to make them
available.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Implement the extended attribute procedures for NFSv4.2 extended
attribute support (RFC 8276).

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Make the buf_to_pages_noslab function available to the rest of the NFS
code. Rename it to nfs4_buf_to_pages_noslab to be consistent.

This will be used later in the NFSv4.2 xattr code.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Define the NFS_INO_INVALID_XATTR flag, to be used for the NFSv4.2 xattr
cache, and use it where appropriate.

No functional change as yet.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Until now, change attributes in change_info form were only returned by
directory operations. However, they are also used for the RFC 8276
extended attribute operations, which work on both directories
and regular files.  Modify update_changeattr to deal:

* Rename it to nfs4_update_changeattr and make it non-static.
* Don't always use INO_INVALID_DATA, this isn't needed for a
  directory that only had its extended attributes changed by us.
* Existing callers now always pass in INO_INVALID_DATA.

For the current callers of this function, behavior is unchanged.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

RFC 8276 defines separate ACCESS bits for extended attribute checking.
Query them in nfs_do_access and opendata.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The only consumer of nfs_access_get_cached_rcu and nfs_access_cached
calls these static functions in order to first try RCU access, and
then locked access.

Combine them in to a single function, and call that. Make this function
available to the rest of the NFS code.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Define the argument and response structures that will be used for
RFC 8276 extended attribute RPC calls, and implement the necessary
functions to encode/decode the extended attribute operations.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Query the server for extended attribute support, and record it
as the NFS_CAP_XATTR flag in the server capabilities.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Set limits for extended attributes (attribute value size and listxattr
buffer size), based on the fs-independent limits (XATTR_*_MAX).

Define the maximum XDR sizes for the RFC 8276 XATTR operations.
In the case of operations that carry a larger payload (SETXATTR,
GETXATTR, LISTXATTR), these exclude that payload, which is added
as separate pages, like other operations do.

Define, much like for read and write operations, the maximum overhead
sizes for get/set/listxattr, and use them to limit the maximum payload
size for those operations, in combination with the channel attributes.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Add client-side only definitions for user extended
attributes (RFC8276). These are the access bits
as used by the client code, and the CLNT procedure
number definition.

Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Pull in the commit that is common to both the NFS client and server xattr
implementation from git://git.linux-nfs.org/projects/cel/cel-2.6.git

 nfsd-5.9

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.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>

Make sure we specify the layout segment range when calculating the
mirror count. In theory, that number could depend on the range to
which we're writing.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Both nfs_pageio_reset_read_mds() and nfs_pageio_reset_write_mds()
do call pnfs_generic_pg_cleanup() for us.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

If the application uses the AT_STATX_DONT_SYNC flag after doing readdir(),
then we should still mark the parent inode as seeing a readdirplus hit.
That ensures that we continue to use readdirplus in the 'ls -l' type
of workflow to do fast lookups of the dentries.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.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

Pull btrfs fixes from David Sterba:
 "Two refcounting fixes and one prepartory patch for upcoming splice
  cleanup:

   - fix double put of block group with nodatacow

   - fix missing block group put when remounting with discard=async

   - explicitly set splice callback (no functional change), to ease
     integrating splice cleanup patches"

* tag 'for-5.8-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: wire up iter_file_splice_write
  btrfs: fix double put of block group with nocow
  btrfs: discard: add missing put when grabbing block group from unused list

59960b9d ("io_uring: fix lazy work init") tried to fix missing
io_req_init_async(), but left out work.flags and hash. Do it earlier.

Fixes: 7cdaf587

 ("io_uring: avoid whole io_wq_work copy for requests completed inline")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>