Trivial fix to a spelling mistake of the error access name EACCESS,
rename to EACCES
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The seq_send & seq_send64 fields in struct krb5_ctx are used as
atomically incrementing counters. This is implemented using cmpxchg() &
cmpxchg64() to implement what amount to custom versions of
atomic_fetch_inc() & atomic64_fetch_inc().

Besides the duplication, using cmpxchg64() has another major drawback in
that some 32 bit architectures don't provide it. As such commit
571ed1fd ("SUNRPC: Replace krb5_seq_lock with a lockless scheme")
resulted in build failures for some architectures.

Change seq_send to be an atomic_t and seq_send64 to be an atomic64_t,
then use atomic(64)_* functions to manipulate the values. The atomic64_t
type & associated functions are provided even on architectures which
lack real 64 bit atomic memory access via CONFIG_GENERIC_ATOMIC64 which
uses spinlocks to serialize access. This fixes the build failures for
architectures lacking cmpxchg64().

A potential alternative that was raised would be to provide cmpxchg64()
on the 32 bit architectures that currently lack it, using spinlocks.
However this would provide a version of cmpxchg64() with semantics a
little different to the implementations on architectures with real 64
bit atomics - the spinlock-based implementation would only work if all
access to the memory used with cmpxchg64() is *always* performed using
cmpxchg64(). That is not currently a requirement for users of
cmpxchg64(), and making it one seems questionable. As such avoiding
cmpxchg64() outside of architecture-specific code seems best,
particularly in cases where atomic64_t seems like a better fit anyway.

The CONFIG_GENERIC_ATOMIC64 implementation of atomic64_* functions will
use spinlocks & so faces the same issue, but with the key difference
that the memory backing an atomic64_t ought to always be accessed via
the atomic64_* functions anyway making the issue moot.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Fixes: 571ed1fd ("SUNRPC: Replace krb5_seq_lock with a lockless scheme")
Cc: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: Anna Schumaker <anna.schumaker@netapp.com>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: linux-nfs@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The filehandle has a length which is defined as a 32-bit
"unsigned integer".  Change sign of the length appropriately.
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

NFS RDMA client updates for Linux 4.20

Stable bugfixes:
- Reset credit grant properly after a disconnect

Other bugfixes and cleanups:
- xprt_release_rqst_cong is called outside of transport_lock
- Create more MRs at a time and toss out old ones during recovery
- Various improvements to the RDMA connection and disconnection code:
  - Improve naming of trace events, functions, and variables
  - Add documenting comments
  - Fix metrics and stats reporting
- Fix a tracepoint sparse warning
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Commits ffb6ca33 and e08ea3a9 prevent setting xprt_min_resvport
greater than xprt_max_resvport, but may also break simple code that sets
one parameter then the other, if the new range does not overlap the old.

Also it looks racy to me, unless there's some serialization I'm not
seeing.  Granted it would probably require malicious privileged processes
(unless there's a chance these might eventually be settable in unprivileged
containers), but still it seems better not to let userspace panic the
kernel.

Simpler seems to be to allow setting the parameters to whatever you want
but interpret xprt_min_resvport > xprt_max_resvport as the empty range.

Fixes: ffb6ca33 "sunrpc: Prevent resvport min/max inversion..."
Fixes: e08ea3a9 "sunrpc: Prevent rexvport min/max inversion..."
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

We don't need to call this in the direct, read, or pnfs resend paths and
the only other caller is the write path in nfs_page_async_flush() which
already checks and sets the pg_error on the context.
Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

We must check pg_error and call error_cleanup after any call to pg_doio.
Currently, we are skipping the unlock of a page if we encounter an error in
nfs_pageio_complete() before handing off the work to the RPC layer.
Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

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

Running "./nfstest_delegation --runtest recall26" uncovers that
client doesn't recover the lock when we have an appending open,
where the initial open got a write delegation.

Instead of checking for the passed in open context against
the file lock's open context. Check that the state is the same.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The newly introduced gss_seq_send64_fetch_and_inc() fails to build on
32-bit architectures:

net/sunrpc/auth_gss/gss_krb5_seal.c:144:14: note: in expansion of macro 'cmpxchg'
   seq_send = cmpxchg(&ctx->seq_send64, old, old + 1);
              ^~~~~~~
arch/x86/include/asm/cmpxchg.h:128:3: error: call to '__cmpxchg_wrong_size' declared with attribute error: Bad argument size for cmpxchg
   __cmpxchg_wrong_size();     \

As the message tells us, cmpxchg() cannot be used on 64-bit arguments,
that's what cmpxchg64() does.

Fixes: 571ed1fd ("SUNRPC: Replace krb5_seq_lock with a lockless scheme")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

linux/include/trace/events/rpcrdma.h:501:1: warning: expression using sizeof bool
linux/include/trace/events/rpcrdma.h:501:1: warning: odd constant _Bool cast (ffffffffffffffff becomes 1)

Fixes: ab03eff5 ("xprtrdma: Add trace points in RPC Call ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Use the appropriate C macro instead of open-coding
container_of() .
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: fill in or update documenting comments for transport
switch entry points.

For xprt_rdma_allocate:

The first paragraph is no longer true since commit 5a6d1db4
("SUNRPC: Add a transport-specific private field in rpc_rqst").

The second paragraph is no longer true since commit 54cbd6b0
("xprtrdma: Delay DMA mapping Send and Receive buffers").
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

To show that a caller did attempt to allocate and post more Receive
buffers, the trace point in rpcrdma_post_recvs() should report when
rpcrdma_post_recvs() was invoked but no new Receive buffers were
posted.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: rb_flags might be used for other things besides
RPCRDMA_BUF_F_EMPTY_SCQ, so initialize it in a generic spot
instead of in a send-completion-queue-related helper.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: This code was copied from xprtsock.c and
backchannel_rqst.c. For rpcrdma, the backchannel server runs
exclusively in process context, thus disabling bottom-halves is
unnecessary.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Replace the hashed memory address of the target rpcrdma_ep
with the server's IP address and port. The server address is more
useful in an administrative error message.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Use a function name that is consistent with the RDMA core
API and with other consumers. Because this is a function that is
invoked from outside the rpcrdma.ko module, add an appropriate
documenting comment.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Currently, when a connection is established, rpcrdma_conn_upcall
invokes rpcrdma_conn_func and then
wake_up_all(&ep->rep_connect_wait). The former wakes waiting RPCs,
but the connect worker is not done yet, and that leads to races,
double wakes, and difficulty understanding how this logic is
supposed to work.

Instead, collect all the "connection established" logic in the
connect worker (xprt_rdma_connect_worker). A disconnect worker is
retained to handle provider upcalls safely.

Fixes: 254f91e2 ("xprtrdma: RPC/RDMA must invoke ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Eliminate the FALLTHROUGH into the default arm to make the
switch easier to understand.

Also, as long as I'm here, do not display the memory address of the
target rpcrdma_ep. A hashed memory address is of marginal use here.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up.

Since commit 173b8f49 ("xprtrdma: Demote "connect" log messages")
there has been no need to initialize connstat to zero. In fact, in
this code path there's now no reason not to set rep_connected
directly.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: The convention throughout other parts of xprtrdma is to
name variables of type struct rpcrdma_xprt "r_xprt", not "xprt".
This convention enables the use of the name "xprt" for a "struct
rpc_xprt" type variable, as in other parts of the RPC client.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Use a function name that is consistent with the RDMA core
API and with other consumers. Because this is a function that is
invoked from outside the rpcrdma.ko module, add an appropriate
documenting comment.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

The way connection-oriented transports report connect_time is wrong:
it's supposed to be in seconds, not in jiffies.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

For TCP, the logic in xprt_connect_status is currently never invoked
to record a successful connection. Commit 2a491991 ("SUNRPC:
Return EAGAIN instead of ENOTCONN when waking up xprt->pending")
changed the way TCP xprt's are awoken after a connect succeeds.

Instead, change connection-oriented transports to bump connect_count
and compute connect_time the moment that XPRT_CONNECTED is set.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Clean up the names of trace events related to MRs so that it's
easy to enable these with a glob.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

When a memory operation fails, the MR's driver state might not match
its hardware state. The only reliable recourse is to dereg the MR.
This is done in ->ro_recover_mr, which then attempts to allocate a
fresh MR to replace the released MR.

Since commit e2ac236c ("xprtrdma: Allocate MRs on demand"),
xprtrdma dynamically allocates MRs. It can add more MRs whenever
they are needed.

That makes it possible to simply release an MR when a memory
operation fails, instead of "recovering" it. It will automatically
be replaced by the on-demand MR allocator.

This commit is a little larger than I wanted, but it replaces
->ro_recover_mr, rb_recovery_lock, rb_recovery_worker, and the
rb_stale_mrs list with a generic work queue.

Since MRs are no longer orphaned, the mrs_orphaned metric is no
longer used.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Some devices require more than 3 MRs to build a single 1MB I/O.
Ensure that rpcrdma_mrs_create() will add enough MRs to build that
I/O.

In a subsequent patch I'm changing the MR recovery logic to just
toss out the MRs. In that case it's possible for ->send_request to
loop acquiring some MRs, not getting enough, getting called again,
recycling the previous MRs, then not getting enough, lather rinse
repeat. Thus first we need to ensure enough MRs are created to
prevent that loop.

I'm "reusing" ia->ri_max_segs. All of its accessors seem to want the
maximum number of data segments plus two, so I'm going to bake that
into the initial calculation.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

On a fresh connection, an RPC/RDMA client is supposed to send only
one RPC Call until it gets a credit grant in the first RPC Reply
from the server [RFC 8166, Section 3.3.3].

There is a bug in the Linux client's credit accounting mechanism
introduced by commit e7ce710a ("xprtrdma: Avoid deadlock when
credit window is reset"). On connect, it simply dumps all pending
RPC Calls onto the new connection.

Servers have been tolerant of this bad behavior. Currently no server
implementation ever changes its credit grant over reconnects, and
servers always repost enough Receives before connections are fully
established.

To correct this issue, ensure that the client resets both the credit
grant _and_ the congestion window when handling a reconnect.

Fixes: e7ce710a ("xprtrdma: Avoid deadlock when credit ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: stable@kernel.org
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

Since commit ce7c252a ("SUNRPC: Add a separate spinlock to
protect the RPC request receive list") the RPC/RDMA reply handler
has been calling xprt_release_rqst_cong without holding
xprt->transport_lock.

I think the only way this call is ever made is if the credit grant
increases and there are RPCs pending. Current server implementations
do not change their credit grant during operation (except at
connect time).

Commit e7ce710a ("xprtrdma: Avoid deadlock when credit window is
reset") added the ->release_rqst call because UDP invokes
xprt_adjust_cwnd(), which calls __xprt_put_cong() after adjusting
xprt->cwnd. Both xprt_release() and ->xprt_release_xprt already wake
another task in this case, so it is safe to remove this call from
the reply handler.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

If we're revalidating an existing dentry in order to open a file, we need
to ensure that we check the directory has not changed before we optimise
away the lookup.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Refactor the code in nfs_lookup_revalidate() as a stepping stone towards
optimising and fixing nfs4_lookup_revalidate().
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

We need to ensure that inode and dentry revalidation occurs correctly
on reopen of a file that is already open. Currently, we can end up
not revalidating either in the case of NFSv4.0, due to the 'cached open'
path.
Let's fix that by ensuring that we only do cached open for the special
cases of open recovery and delegation return.
Reported-by: Stan Hu <stanhu@gmail.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

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

Use RCU protected lookups for discovering the supported mechanisms.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Module removal is RCU safe by design, so we really have no need to
lock the auth_flavors[] array. Substitute a lockless scheme to
add/remove entries in the array, and then use rcu.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Now that each struct nfs_pgio_header corresponds to one RPC call, we
only have one writer to the struct nfs_pgio_header.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Save a few bytes by allowing the read/write specific fields of the
structures to share storage.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

When the server fails to return post-op attributes, the client's
attempt to place read data directly in the page cache fails, and
so we have to do an extra copy in order to realign the data with
page borders.
This patch attempts to detect servers that don't return post-op
attributes on read (e.g. for pNFS) and adjusts the placement
calculation accordingly.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

The convention in the rest of the code is to have a separate function
for anything that might be ifdef-ed out.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>