SUNRPC: Generalize the RPC buffer allocation API

xprtrdma needs to allocate the Call and Reply buffers separately.
TBH, the reliance on using a single buffer for the pair of XDR
buffers is transport implementation-specific.

Transports that want to allocate separate Call and Reply buffers
will ignore the "size" argument anyway.  Don't bother passing it.

The buf_alloc method can't return two pointers. Instead, make the
method's return value an error code, and set the rq_buffer pointer
in the method itself.

This gives call_allocate an opportunity to terminate an RPC instead
of looping forever when a permanent problem occurs. If a request is
just bogus, or the transport is in a state where it can't allocate
resources for any request, there needs to be a way to kill the RPC
right there and not loop.

This immediately fixes a rare problem in the backchannel send path,
which loops if the server happens to send a CB request whose
call+reply size is larger than a page (which it shouldn't do yet).

One more issue: looks like xprt_inject_disconnect was incorrectly
placed in the failure path in call_allocate. It needs to be in the
success path, as it is for other call-sites.

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

1 files changed, 15 insertions, 9 deletions

diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 9ae588511aaf..b964d40b259b 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -849,14 +849,17 @@ static void rpc_async_schedule(struct work_struct *work)
 }
 
 /**
- * rpc_malloc - allocate an RPC buffer
- * @task: RPC task that will use this buffer
- * @size: requested byte size
+ * rpc_malloc - allocate RPC buffer resources
+ * @task: RPC task
+ *
+ * A single memory region is allocated, which is split between the
+ * RPC call and RPC reply that this task is being used for. When
+ * this RPC is retired, the memory is released by calling rpc_free.
  *
  * To prevent rpciod from hanging, this allocator never sleeps,
- * returning NULL and suppressing warning if the request cannot be serviced
- * immediately.
- * The caller can arrange to sleep in a way that is safe for rpciod.
+ * returning -ENOMEM and suppressing warning if the request cannot
+ * be serviced immediately. The caller can arrange to sleep in a
+ * way that is safe for rpciod.
  *
  * Most requests are 'small' (under 2KiB) and can be serviced from a
  * mempool, ensuring that NFS reads and writes can always proceed,
@@ -865,8 +868,10 @@ static void rpc_async_schedule(struct work_struct *work)
  * In order to avoid memory starvation triggering more writebacks of
  * NFS requests, we avoid using GFP_KERNEL.
  */
-void *rpc_malloc(struct rpc_task *task, size_t size)
+int rpc_malloc(struct rpc_task *task)
 {
+	struct rpc_rqst *rqst = task->tk_rqstp;
+	size_t size = rqst->rq_callsize + rqst->rq_rcvsize;
 	struct rpc_buffer *buf;
 	gfp_t gfp = GFP_NOIO | __GFP_NOWARN;
 
@@ -880,12 +885,13 @@ void *rpc_malloc(struct rpc_task *task, size_t size)
 		buf = kmalloc(size, gfp);
 
 	if (!buf)
-		return NULL;
+		return -ENOMEM;
 
 	buf->len = size;
 	dprintk("RPC: %5u allocated buffer of size %zu at %p\n",
 			task->tk_pid, size, buf);
-	return &buf->data;
+	rqst->rq_buffer = buf->data;
+	return 0;
 }
 EXPORT_SYMBOL_GPL(rpc_malloc);