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

Currently, when the sendctx queue is exhausted during marshaling, the
RPC/RDMA transport places the RPC task on the delayq, which forces a
wait for HZ >> 2 before the marshal and send is retried.

With this change, the transport now places such an RPC task on the
pending queue, and wakes it just as soon as more sendctxs become
available. This typically takes less than a millisecond, and the
write_space waking mechanism is less deadlock-prone.

Moreover, the waiting RPC task is holding the transport's write
lock, which blocks the transport from sending RPCs. Therefore faster
recovery from sendctx queue exhaustion is desirable.

Cf. commit 5804891455d5 ("xprtrdma: ->send_request returns -EAGAIN
when there are no free MRs").
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Move #include <trace/events/rpcrdma.h> into source files,
similar to how it is done with trace/events/sunrpc.h.

Server-side trace points will be part of the rpcrdma subsystem,
just like the client-side trace points.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NJ. Bruce Fields <bfields@redhat.com>

Clean up: The only call site is in the same file as the function's
definition.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: These functions are no longer used.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Receive completion and Reply handling are done by a BOUND
workqueue, meaning they run on only one CPU.

Posting receives is currently done in the send_request path, which
on large systems is typically done on a different CPU than the one
handling Receive completions. This results in movement of
Receive-related cachelines between the sending and receiving CPUs.

More importantly, it means that currently Receives are posted while
the transport's write lock is held, which is unnecessary and costly.

Finally, allocation of Receive buffers is performed on-demand in
the Receive completion handler. This helps guarantee that they are
allocated on the same NUMA node as the CPU that handles Receive
completions.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

This simplifies allocation of the generic RPC slot and xprtrdma
specific per-RPC resources.

It also makes xprtrdma more like the socket-based transports:
->buf_alloc and ->buf_free are now responsible only for send and
receive buffers.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

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

The ro_release_mr methods check whether mr->mr_list is empty.
Therefore, be sure to always use list_del_init when removing an MR
linked into a list using that field. Otherwise, when recovering from
transport failures or device removal, list corruption can result, or
MRs can get mapped or unmapped an odd number of times, resulting in
IOMMU-related failures.

In general this fix is appropriate back to v4.8. However, code
changes since then make it impossible to apply this patch directly
to stable kernels. The fix would have to be applied by hand or
reworked for kernels earlier than v4.16.

Backport guidance -- there are several cases:
- When creating an MR, initialize mr_list so that using list_empty
  on an as-yet-unused MR is safe.
- When an MR is being handled by the remote invalidation path,
  ensure that mr_list is reinitialized when it is removed from
  rl_registered.
- When an MR is being handled by rpcrdma_destroy_mrs, it is removed
  from mr_all, but it may still be on an rl_registered list. In
  that case, the MR needs to be removed from that list before being
  released.
- Other cases are covered by using list_del_init in rpcrdma_mr_pop.

Fixes: 9d6b0409 ('xprtrdma: Place registered MWs on a ... ')
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

With FRWR, the client transport can perform memory registration and
post a Send with just a single ib_post_send.

This reduces contention between the send_request path and the Send
Completion handlers, and reduces the overhead of registering a chunk
that has multiple segments.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: The generic rq_connect_cookie is sufficient to detect RPC
Call retransmission.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

With v4.15, on one of my NFS/RDMA clients I measured a nearly
doubling in the latency of small read and write system calls. There
was no change in server round trip time. The extra latency appears
in the whole RPC execution path.

"git bisect" settled on commit ccede759 ("xprtrdma: Spread reply
processing over more CPUs") .

After some experimentation, I found that leaving the WQ bound and
allowing the scheduler to pick the dispatch CPU seems to eliminate
the long latencies, and it does not introduce any new regressions.

The fix is implemented by reverting only the part of
commit ccede759 ("xprtrdma: Spread reply processing over more
CPUs") that dispatches RPC replies specifically on the CPU where the
matching RPC call was made.

Interestingly, saving the CPU number and later queuing reply
processing there was effective _only_ for a NFS READ and WRITE
request. On my NUMA client, in-kernel RPC reply processing for
asynchronous RPCs was dispatched on the same CPU where the RPC call
was made, as expected. However synchronous RPCs seem to get their
reply dispatched on some other CPU than where the call was placed,
every time.

Fixes: ccede759 ("xprtrdma: Spread reply processing over ... ")
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Cc: stable@vger.kernel.org # v4.15+
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

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

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

Clean up: Code review suggested that a common bit of code can be
placed into a helper function, and this gives us fewer places to
stick an "I DMA unmapped something" trace point.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: struct rpcrdma_mw was named after Memory Windows, but
xprtrdma no longer supports a Memory Window registration mode.
Rename rpcrdma_mw and its fields to reduce confusion and make
the code more sensible to read.

Renaming "mw" was suggested by Tom Talpey, the author of the
original xprtrdma implementation. It's a good idea, but I haven't
done this until now because it's a huge diffstat for no benefit
other than code readability.

However, I'm about to introduce static trace points that expose
a few of xprtrdma's internal data structures. They should make sense
in the trace report, and it's reasonable to treat trace points as a
kernel API contract which might be difficult to change later.

While I'm churning things up, two additional changes:
- rename variables unhelpfully called "r" to "mr", to improve code
  clarity, and
- rename the MR-related helper functions using the form
  "rpcrdma_mr_<verb>", to be consistent with other areas of the
  code.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Over time, the industry has adopted the term "frwr"
instead of "frmr". The term "frwr" is now more widely recognized.

For the past couple of years I've attempted to add new code using
"frwr" , but there still remains plenty of older code that still
uses "frmr". Replace all usage of "frmr" to avoid confusion.

While we're churning code, rename variables unhelpfully called "f"
to "frwr", to improve code clarity.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up. @rqst is set up differently for backchannel Replies. For
example, rqst->rq_task and task->tk_client are both NULL. So it is
easier to understand and maintain this code path if it is separated.

Also, we can get rid of the confusing rl_connect_cookie hack in
rpcrdma_bc_receive_call.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Since commit 5a6d1db4 ("SUNRPC: Add a transport-specific private
field in rpc_rqst"), the rpc_rqst's for RPC-over-RDMA backchannel
operations leave rq_buffer set to NULL.

xprt_release does not invoke ->op->buf_free when rq_buffer is NULL.
The RPCRDMA_REQ_F_BACKCHANNEL check in xprt_rdma_free is therefore
redundant because xprt_rdma_free is not invoked for backchannel
requests.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Save more space in struct rpcrdma_xprt by removing the redundant
"addr" field from struct rpcrdma_create_data_internal. Wherever
we have rpcrdma_xprt, we also have the rpc_xprt, which has a
sockaddr_storage field with the same content.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

This makes the address strings available for debugging messages in
earlier stages of transport set up.

The first benefit is to get rid of the single-use rep_remote_addr
field, saving 128+ bytes in struct rpcrdma_ep.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up. Remove fields that should have been removed by
commit b3221d6a ("xprtrdma: Remove logic that constructs
RDMA_MSGP type calls").
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up.

Commit b5f0afbe ("xprtrdma: Per-connection pad optimization")
should have removed this.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Refactoring change: Remote Invalidation is particular to the memory
registration mode that is use. Use a callout instead of a generic
function to handle Remote Invalidation.

This gets rid of the 8-byte flags field in struct rpcrdma_mw, of
which only a single bit flag has been allocated.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

The backchannel code uses rpcrdma_recv_buffer_put to add new reps
to the free rep list. This also decrements rb_recv_count, which
spoofs the receive overrun logic in rpcrdma_buffer_get_rep.

Commit 9b06688b ("xprtrdma: Fix additional uses of
spin_lock_irqsave(rb_lock)") replaced the original open-coded
list_add with a call to rpcrdma_recv_buffer_put(), but then a year
later, commit 05c97466 ("xprtrdma: Fix receive buffer
accounting") added rep accounting to rpcrdma_recv_buffer_put.
It was an oversight to let the backchannel continue to use this
function.

The fix this, let's combine the "add to free list" logic with
rpcrdma_create_rep.

Also, do not allocate RPCRDMA_MAX_BC_REQUESTS rpcrdma_reps in
rpcrdma_buffer_create and then allocate additional rpcrdma_reps in
rpcrdma_bc_setup_reps. Allocating the extra reps during backchannel
set-up is sufficient.

Fixes: 05c97466 ("xprtrdma: Fix receive buffer accounting")
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Commit d8f532d2 ("xprtrdma: Invoke rpcrdma_reply_handler
directly from RECV completion") introduced a performance regression
for NFS I/O small enough to not need memory registration. In multi-
threaded benchmarks that generate primarily small I/O requests,
IOPS throughput is reduced by nearly a third. This patch restores
the previous level of throughput.

Because workqueues are typically BOUND (in particular ib_comp_wq,
nfsiod_workqueue, and rpciod_workqueue), NFS/RDMA workloads tend
to aggregate on the CPU that is handling Receive completions.

The usual approach to addressing this problem is to create a QP
and CQ for each CPU, and then schedule transactions on the QP
for the CPU where you want the transaction to complete. The
transaction then does not require an extra context switch during
completion to end up on the same CPU where the transaction was
started.

This approach doesn't work for the Linux NFS/RDMA client because
currently the Linux NFS client does not support multiple connections
per client-server pair, and the RDMA core API does not make it
straightforward for ULPs to determine which CPU is responsible for
handling Receive completions for a CQ.

So for the moment, record the CPU number in the rpcrdma_req before
the transport sends each RPC Call. Then during Receive completion,
queue the RPC completion on that same CPU.

Additionally, move all RPC completion processing to the deferred
handler so that even RPCs with simple small replies complete on
the CPU that sent the corresponding RPC Call.

Fixes: d8f532d2 ("xprtrdma: Invoke rpcrdma_reply_handler ...")
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Credit work contributed by Oracle engineers since 2014.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: C-structure style XDR encoding and decoding logic has
been replaced over the past several merge windows on both the
client and server. These data structures are no longer used.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Reviewed-by: NDevesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

The sendctx circular queue now guarantees that xprtrdma cannot
overflow the Send Queue, so remove the remaining bits of the
original Send WQE counting mechanism.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

When an RPC Call includes a file data payload, that payload can come
from pages in the page cache, or a user buffer (for direct I/O).

If the payload can fit inline, xprtrdma includes it in the Send
using a scatter-gather technique. xprtrdma mustn't allow the RPC
consumer to re-use the memory where that payload resides before the
Send completes. Otherwise, the new contents of that memory would be
exposed by an HCA retransmit of the Send operation.

So, block RPC completion on Send completion, but only in the case
where a separate file data payload is part of the Send. This
prevents the reuse of that memory while it is still part of a Send
operation without an undue cost to other cases.

Waiting is avoided in the common case because typically the Send
will have completed long before the RPC Reply arrives.

These days, an RPC timeout will trigger a disconnect, which tears
down the QP. The disconnect flushes all waiting Sends. This bounds
the amount of time the reply handler has to wait for a Send
completion.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Invoke a common routine for releasing hardware resources (for
example, invalidating MRs). This needs to be done whether an
RPC Reply has arrived or the RPC was terminated early.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

We have one boolean flag in rpcrdma_req today. I'd like to add more
flags, so convert that boolean to a bit flag.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Problem statement:

Recently Sagi Grimberg <sagi@grimberg.me> observed that kernel RDMA-
enabled storage initiators don't handle delayed Send completion
correctly. If Send completion is delayed beyond the end of a ULP
transaction, the ULP may release resources that are still being used
by the HCA to complete a long-running Send operation.

This is a common design trait amongst our initiators. Most Send
operations are faster than the ULP transaction they are part of.
Waiting for a completion for these is typically unnecessary.

Infrequently, a network partition or some other problem crops up
where an ordering problem can occur. In NFS parlance, the RPC Reply
arrives and completes the RPC, but the HCA is still retrying the
Send WR that conveyed the RPC Call. In this case, the HCA can try
to use memory that has been invalidated or DMA unmapped, and the
connection is lost. If that memory has been re-used for something
else (possibly not related to NFS), and the Send retransmission
exposes that data on the wire.

Thus we cannot assume that it is safe to release Send-related
resources just because a ULP reply has arrived.

After some analysis, we have determined that the completion
housekeeping will not be difficult for xprtrdma:

 - Inline Send buffers are registered via the local DMA key, and
   are already left DMA mapped for the lifetime of a transport
   connection, thus no additional handling is necessary for those
 - Gathered Sends involving page cache pages _will_ need to
   DMA unmap those pages after the Send completes. But like
   inline send buffers, they are registered via the local DMA key,
   and thus will not need to be invalidated

In addition, RPC completion will need to wait for Send completion
in the latter case. However, nearly always, the Send that conveys
the RPC Call will have completed long before the RPC Reply
arrives, and thus no additional latency will be accrued.

Design notes:

In this patch, the rpcrdma_sendctx object is introduced, and a
lock-free circular queue is added to manage a set of them per
transport.

The RPC client's send path already prevents sending more than one
RPC Call at the same time. This allows us to treat the consumer
side of the queue (rpcrdma_sendctx_get_locked) as if there is a
single consumer thread.

The producer side of the queue (rpcrdma_sendctx_put_locked) is
invoked only from the Send completion handler, which is a single
thread of execution (soft IRQ).

The only care that needs to be taken is with the tail index, which
is shared between the producer and consumer. Only the producer
updates the tail index. The consumer compares the head with the
tail to ensure that the a sendctx that is in use is never handed
out again (or, expressed more conventionally, the queue is empty).

When the sendctx queue empties completely, there are enough Sends
outstanding that posting more Send operations can result in a Send
Queue overflow. In this case, the ULP is told to wait and try again.
This introduces strong Send Queue accounting to xprtrdma.

As a final touch, Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
suggested a mechanism that does not require signaling every Send.
We signal once every N Sends, and perform SGE unmapping of N Send
operations during that one completion.
Reported-by: NSagi Grimberg <sagi@grimberg.me>
Suggested-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Make rpcrdma_prepare_send_sges() return a negative errno
instead of a bool. Soon callers will want distinct treatments of
different types of failures.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

We need to decode and save the incoming rdma_credits field _after_
we know that the direction of the message is "forward direction
Reply". Otherwise, the credits value in reverse direction Calls is
also used to update the forward direction credits.

It is safe to decode the rdma_credits field in rpcrdma_reply_handler
now that rpcrdma_reply_handler is single-threaded. Receives complete
in the same order as they were sent on the NFS server.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

I noticed that the soft IRQ thread looked pretty busy under heavy
I/O workloads. perf suggested one area that was expensive was the
queue_work() call in rpcrdma_wc_receive. That gave me some ideas.

Instead of scheduling a separate worker to process RPC Replies,
promote the Receive completion handler to IB_POLL_WORKQUEUE, and
invoke rpcrdma_reply_handler directly.

Note that the poll workqueue is single-threaded. In order to keep
memory invalidation from serializing all RPC Replies, handle any
necessary invalidation tasks in a separate multi-threaded workqueue.

This provides a two-tier scheme, similar to OS I/O interrupt
handlers: A fast interrupt handler that schedules the slow handler
and re-enables the interrupt, and a slower handler that is invoked
for any needed heavy lifting.

Benefits include:
- One less context switch for RPCs that don't register memory
- Receive completion handling is moved out of soft IRQ context to
  make room for other users of soft IRQ
- The same CPU core now DMA syncs and XDR decodes the Receive buffer
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: I'd like to be able to invoke the tail of
rpcrdma_reply_handler in two different places. Split the tail out
into its own helper function.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: Make it easier to pass the decoded XID, vers, credits, and
proc fields around by moving these variables into struct rpcrdma_rep.

Note: the credits field will be handled in a subsequent patch.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>

Clean up: There are no remaining callers of this method.
Signed-off-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NAnna Schumaker <Anna.Schumaker@Netapp.com>