Micro-optimization: In rpcrdma_post_recvs, since commit e340c2d6ef
("xprtrdma: Reduce the doorbell rate (Receive)"), the common case is
to return without doing anything. Found with perf.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
rpcrdma_rep objects are removed from their free list by only a
single thread: the Receive completion handler. Thus that free list
can be converted to an llist, where a single-threaded consumer and
a multi-threaded producer (rpcrdma_buffer_put) can both access the
llist without the need for any serialization.
This eliminates spin lock contention between the Receive completion
handler and rpcrdma_buffer_get, and makes the rep consumer wait-
free.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Now that the free list is used sparingly, get rid of the
separate spin lock protecting it.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Instead of a globally-contended MR free list, cache MRs in each
rpcrdma_req as they are released. This means acquiring and releasing
an MR will be lock-free in the common case, even outside the
transport send lock.
The original idea of per-rpcrdma_req MR free lists was suggested by
Shirley Ma <shirley.ma@oracle.com> several years ago. I just now
figured out how to make that idea work with on-demand MR allocation.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Probably would be good to also pass GFP flags to ib_alloc_mr.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Refactor: Retrieve an MR and handle error recovery entirely in
rpc_rdma.c, as this is not a device-specific function.
Note that since commit 89f90fe1ad ("SUNRPC: Allow calls to
xprt_transmit() to drain the entire transmit queue"), the
xprt_transmit function handles the cond_resched. The transport no
longer has to do this itself.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up. There is only one remaining rpcrdma_mr_put call site, and
it can be directly replaced with unmap_and_put because mr->mr_dir is
set to DMA_NONE just before the call.
Now all the call sites do a DMA unmap, and we can just rename
mr_unmap_and_put to mr_put, which nicely matches mr_get.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: rpcrdma_mr_pop call sites check if the list is empty
first. Let's replace the list_empty with less costly logic.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: There are other "all" list heads. For code clarity
distinguish this one as for use only for MRs by renaming it.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Make the field name the same for all trace points that handle
pointers to struct rpcrdma_rep. That makes it easy to grep for
matching rep points in trace output.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Although I haven't seen any performance results that justify it,
I've received several complaints that NFS/RDMA no longer supports
a maximum rsize and wsize of 1MB. These days it is somewhat smaller.
To simplify the logic that determines whether a chunk list is
necessary, the implementation uses a fixed maximum size of the
transport header. Currently that maximum size is 256 bytes, one
quarter of the default inline threshold size for RPC/RDMA v1.
Since commit a78868497c ("xprtrdma: Reduce max_frwr_depth"), the
size of chunks is also smaller to take advantage of inline page
lists in device internal MR data structures.
The combination of these two design choices has reduced the maximum
NFS rsize and wsize that can be used for most RNIC/HCAs. Increasing
the maximum transport header size and the maximum number of RDMA
segments it can contain increases the negotiated maximum rsize/wsize
on common RNIC/HCAs.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
There is only one remaining function, rpcrdma_buffer_put(), that
uses this field. Its caller can supply a pointer to the correct
rpcrdma_buffer, enabling the removal of an 8-byte pointer field
from a frequently-allocated shared data structure.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
rb_lock is contended between rpcrdma_buffer_create,
rpcrdma_buffer_put, and rpcrdma_post_recvs.
Commit e340c2d6ef ("xprtrdma: Reduce the doorbell rate (Receive)")
causes rpcrdma_post_recvs to take the rb_lock repeatedly when it
determines more Receives are needed. Streamline this code path so
it takes the lock just once in most cases to build the Receive
chain that is about to be posted.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
Commit 7c8d9e7c88 ("xprtrdma: Move Receive posting to Receive
handler") reduced the number of rpcrdma_rep_create call sites to
one. After that commit, the backchannel code no longer invokes it.
Therefore the free list logic added by commit d698c4a02e
("xprtrdma: Fix backchannel allocation of extra rpcrdma_reps") is
no longer necessary, and in fact adds some extra overhead that we
can do without.
Simply post any newly created reps. They will get added back to
the rb_recv_bufs list when they subsequently complete.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Eliminate a context switch in the path that handles RPC wake-ups
when a Receive completion has to wait for a Send completion.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Since commit ba69cd122e ("xprtrdma: Remove support for FMR memory
registration"), FRWR is the only supported memory registration mode.
We can take advantage of the asynchronous nature of FRWR's LOCAL_INV
Work Requests to get rid of the completion wait by having the
LOCAL_INV completion handler take care of DMA unmapping MRs and
waking the upper layer RPC waiter.
This eliminates two context switches when local invalidation is
necessary. As a side benefit, we will no longer need the per-xprt
deferred completion work queue.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Under high I/O workloads, I've noticed that an RPC/RDMA transport
occasionally deadlocks (IOPS goes to zero, and doesn't recover).
Diagnosis shows that the sendctx queue is empty, but when sendctxs
are returned to the queue, the xprt_write_space wake-up never
occurs. The wake-up logic in rpcrdma_sendctx_put_locked is racy.
I noticed that both EMPTY_SCQ and XPRT_WRITE_SPACE are implemented
via an atomic bit. Just one of those is sufficient. Removing
EMPTY_SCQ in favor of the generic bit mechanism makes the deadlock
un-reproducible.
Without EMPTY_SCQ, rpcrdma_buffer::rb_flags is no longer used and
is therefore removed.
Unfortunately this patch does not apply cleanly to stable. If
needed, someone will have to port it and test it.
Fixes: 2fad659209 ("xprtrdma: Wait on empty sendctx queue")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Dereference wr->next /before/ the memory backing wr has been
released. This issue was found by code inspection. It is not
expected to be a significant problem because it is in an error
path that is almost never executed.
Fixes: 7c8d9e7c88 ("xprtrdma: Move Receive posting to ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:
struct foo {
int stuff;
struct boo entry[];
};
instance = kzalloc(sizeof(struct foo) + count * sizeof(struct boo), GFP_KERNEL);
Instead of leaving these open-coded and prone to type mistakes, we can
now use the new struct_size() helper:
instance = kzalloc(struct_size(instance, entry, count), GFP_KERNEL);
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Commit e1ede312f1 ("xprtrdma: Fix helper that drains the
transport") replaced the ib_drain_qp() call, so update documenting
comments to reflect current operation.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: rely on the trace points instead.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
Move the remaining field in rpcrdma_create_data_internal so the
structure can be removed.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
The inline settings are actually a characteristic of the endpoint,
and not related to the device. They are also modified after the
transport instance is created, so they do not belong in the cdata
structure either.
Lastly, let's use names that are more natural to RDMA than to NFS:
inline_write -> inline_send and inline_read -> inline_recv. The
/proc files retain their names to avoid breaking user space.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
Since commit 54cbd6b0c6 ("xprtrdma: Delay DMA mapping Send and
Receive buffers"), a pointer to the device is now saved in each
regbuf when it is DMA mapped.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Instead of using a fixed number, allow the amount of Send completion
batching to vary based on the client's maximum credit limit.
- A larger default gives a small boost to IOPS throughput
- Reducing it based on max_requests gives a safe result when the
max credit limit is cranked down (eg. when the device has a small
max_qp_wr).
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Minor clean-ups I've stumbled on since sendctx was merged last year.
In particular, making Send completion processing more efficient
appears to have a measurable impact on IOPS throughput.
Note: test_and_clear_bit() returns a value, thus an explicit memory
barrier is not necessary.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
For code legibility, clean up the function names to be consistent
with the pattern: "rpcrdma" _ object-type _ action
Also rpcrdma_regbuf_alloc and rpcrdma_regbuf_free no longer have any
callers outside of verbs.c, and can thus be made static.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up by providing an API to do this common task.
At this point, the difference between rpcrdma_get_sendbuf and
rpcrdma_get_recvbuf has become tiny. These can be collapsed into a
single helper.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Allocating an rpcrdma_req's regbufs at xprt create time enables
a pair of micro-optimizations:
First, if these regbufs are always there, we can eliminate two
conditional branches from the hot xprt_rdma_allocate path.
Second, by allocating a 1KB buffer, it places a lower bound on the
size of these buffers, without adding yet another conditional
branch. The lower bound reduces the number of hardway re-
allocations. In fact, for some workloads it completely eliminates
hardway allocations.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Allocate the struct rpcrdma_regbuf separately from the I/O buffer
to better guarantee the alignment of the I/O buffer and eliminate
the wasted space between the rpcrdma_regbuf metadata and the buffer
itself.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
For code legibility, clean up the function names to be consistent
with the pattern: "rpcrdma" _ object-type _ action
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Eventually, I'd like to invoke rpcrdma_create_req() during the
call_reserve step. Memory allocation there probably needs to use
GFP_NOIO. Therefore a set of GFP flags needs to be passed in.
As an additional clean up, just return a pointer or NULL, because
the only error return code here is -ENOMEM.
Lastly, clean up the function names to be consistent with the
pattern: "rpcrdma" _ object-type _ action
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
We want to drain only the RQ first. Otherwise the transport can
deadlock on ->close if there are outstanding Send completions.
Fixes: 6d2d0ee27c ("xprtrdma: Replace rpcrdma_receive_wq ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: stable@vger.kernel.org # v5.0+
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Post RECV WRs in batches to reduce the hardware doorbell rate per
transport. This helps the RPC-over-RDMA client scale better in
number of transports.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Make sure the device has at least 2 completion vectors
before allocating to compvec#1
Fixes: a4699f5647 (xprtrdma: Put Send CQ in IB_POLL_WORKQUEUE mode)
Signed-off-by: Nicolas Morey-Chaisemartin <nmoreychaisemartin@suse.com>
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The clean up is handled by the caller, rpcrdma_buffer_create(), so this
call to rpcrdma_sendctxs_destroy() leads to a double free.
Fixes: ae72950abf ("xprtrdma: Add data structure to manage RDMA Send arguments")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
This should return -ENOMEM if __alloc_workqueue_key() fails, but it
returns success.
Fixes: 6d2d0ee27c ("xprtrdma: Replace rpcrdma_receive_wq with a per-xprt workqueue")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Make a note of the function's dependency on an earlier ib_drain_qp.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Since commit 7c8d9e7c88 ("xprtrdma: Move Receive posting to
Receive handler"), rpcrdma_ep_post is no longer responsible for
posting Receive buffers. Update the documenting comment to reflect
this change.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
These are rare, but can be helpful at tracking down DMAR and other
problems.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Remove dprintk() call sites that report rare or impossible
errors. Leave a few that display high-value low noise status
information.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: There's little chance of contention between the use of
rb_lock and rb_reqslock, so merge the two. This avoids having to
take both in some (possibly future) cases.
Transport tear-down is already serialized, thus there is no need for
locking at all when destroying rpcrdma_reqs.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Now that there is only FRWR, there is no need for a memory
registration switch. The indirect calls to the memreg operations can
be replaced with faster direct calls.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
FMR is not supported on most recent RDMA devices. It is also less
secure than FRWR because an FMR memory registration can expose
adjacent bytes to remote reading or writing. As discussed during the
RDMA BoF at LPC 2018, it is time to remove support for FMR in the
NFS/RDMA client stack.
Note that NFS/RDMA server-side uses either local memory registration
or FRWR. FMR is not used.
There are a few Infiniband/RoCE devices in the kernel tree that do
not appear to support MEM_MGT_EXTENSIONS (FRWR), and therefore will
not support client-side NFS/RDMA after this patch. These are:
- mthca
- qib
- hns (RoCE)
Users of these devices can use NFS/TCP on IPoIB instead.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Transport disconnect processing does a "wake pending tasks" at
various points.
Suppose an RPC Reply is being processed. The RPC task that Reply
goes with is waiting on the pending queue. If a disconnect wake-up
happens before reply processing is done, that reply, even if it is
good, is thrown away, and the RPC has to be sent again.
This window apparently does not exist for socket transports because
there is a lock held while a reply is being received which prevents
the wake-up call until after reply processing is done.
To resolve this, all RPC replies being processed on an RPC-over-RDMA
transport have to complete before pending tasks are awoken due to a
transport disconnect.
Callers that already hold the transport write lock may invoke
->ops->close directly. Others use a generic helper that schedules
a close when the write lock can be taken safely.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
After thinking about this more, and auditing other kernel ULP imple-
mentations, I believe that a DISCONNECT cm_event will occur after a
fatal QP event. If that's the case, there's no need for an explicit
disconnect in the QP event handler.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
To address a connection-close ordering problem, we need the ability
to drain the RPC completions running on rpcrdma_receive_wq for just
one transport. Give each transport its own RPC completion workqueue,
and drain that workqueue when disconnecting the transport.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Divide the work cleanly:
- rpcrdma_wc_receive is responsible only for RDMA Receives
- rpcrdma_reply_handler is responsible only for RPC Replies
- the posted send and receive counts both belong in rpcrdma_ep
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
While chasing yet another set of DMAR fault reports, I noticed that
the frwr recycler conflates whether or not an MR has been DMA
unmapped with frwr->fr_state. Actually the two have only an indirect
relationship. It's in fact impossible to guess reliably whether the
MR has been DMA unmapped based on its fr_state field, especially as
the surrounding code and its assumptions have changed over time.
A better approach is to track the DMA mapping status explicitly so
that the recycler is less brittle to unexpected situations, and
attempts to DMA-unmap a second time are prevented.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: stable@vger.kernel.org # v4.20
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever