SUNRPC: switchable buffer allocation
Add RPC client transport switch support for replacing buffer management
on a per-transport basis.
In the current IPv4 socket transport implementation, RPC buffers are
allocated as needed for each RPC message that is sent. Some transport
implementations may choose to use pre-allocated buffers for encoding,
sending, receiving, and unmarshalling RPC messages, however. For
transports capable of direct data placement, the buffers can be carved
out of a pre-registered area of memory rather than from a slab cache.
Test-plan:
Millions of fsx operations. Performance characterization with "sio" and
"iozone". Use oprofile and other tools to look for significant regression
in CPU utilization.
Signed-off-by: Chuck Lever <cel@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 6dda386..069a6cb 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -838,6 +838,8 @@
req->rq_timeout = xprt->timeout.to_initval;
req->rq_task = task;
req->rq_xprt = xprt;
+ req->rq_buffer = NULL;
+ req->rq_bufsize = 0;
req->rq_xid = xprt_alloc_xid(xprt);
req->rq_release_snd_buf = NULL;
dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid,
@@ -867,6 +869,7 @@
mod_timer(&xprt->timer,
xprt->last_used + xprt->idle_timeout);
spin_unlock_bh(&xprt->transport_lock);
+ xprt->ops->buf_free(task);
task->tk_rqstp = NULL;
if (req->rq_release_snd_buf)
req->rq_release_snd_buf(req);