2 * linux/net/sunrpc/xprt.c
4 * This is a generic RPC call interface supporting congestion avoidance,
5 * and asynchronous calls.
7 * The interface works like this:
9 * - When a process places a call, it allocates a request slot if
10 * one is available. Otherwise, it sleeps on the backlog queue
12 * - Next, the caller puts together the RPC message, stuffs it into
13 * the request struct, and calls xprt_transmit().
14 * - xprt_transmit sends the message and installs the caller on the
15 * transport's wait list. At the same time, if a reply is expected,
16 * it installs a timer that is run after the packet's timeout has
18 * - When a packet arrives, the data_ready handler walks the list of
19 * pending requests for that transport. If a matching XID is found, the
20 * caller is woken up, and the timer removed.
21 * - When no reply arrives within the timeout interval, the timer is
22 * fired by the kernel and runs xprt_timer(). It either adjusts the
23 * timeout values (minor timeout) or wakes up the caller with a status
25 * - When the caller receives a notification from RPC that a reply arrived,
26 * it should release the RPC slot, and process the reply.
27 * If the call timed out, it may choose to retry the operation by
28 * adjusting the initial timeout value, and simply calling rpc_call
31 * Support for async RPC is done through a set of RPC-specific scheduling
32 * primitives that `transparently' work for processes as well as async
33 * tasks that rely on callbacks.
40 #include <linux/module.h>
42 #include <linux/types.h>
43 #include <linux/interrupt.h>
44 #include <linux/workqueue.h>
45 #include <linux/net.h>
46 #include <linux/ktime.h>
48 #include <linux/sunrpc/clnt.h>
49 #include <linux/sunrpc/metrics.h>
50 #include <linux/sunrpc/bc_xprt.h>
51 #include <linux/rcupdate.h>
53 #include <trace/events/sunrpc.h>
61 #if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
62 # define RPCDBG_FACILITY RPCDBG_XPRT
68 static void xprt_init(struct rpc_xprt *xprt, struct net *net);
69 static __be32 xprt_alloc_xid(struct rpc_xprt *xprt);
70 static void xprt_connect_status(struct rpc_task *task);
71 static void xprt_destroy(struct rpc_xprt *xprt);
73 static DEFINE_SPINLOCK(xprt_list_lock);
74 static LIST_HEAD(xprt_list);
77 * xprt_register_transport - register a transport implementation
78 * @transport: transport to register
80 * If a transport implementation is loaded as a kernel module, it can
81 * call this interface to make itself known to the RPC client.
84 * 0: transport successfully registered
85 * -EEXIST: transport already registered
86 * -EINVAL: transport module being unloaded
88 int xprt_register_transport(struct xprt_class *transport)
94 spin_lock(&xprt_list_lock);
95 list_for_each_entry(t, &xprt_list, list) {
96 /* don't register the same transport class twice */
97 if (t->ident == transport->ident)
101 list_add_tail(&transport->list, &xprt_list);
102 printk(KERN_INFO "RPC: Registered %s transport module.\n",
107 spin_unlock(&xprt_list_lock);
110 EXPORT_SYMBOL_GPL(xprt_register_transport);
113 * xprt_unregister_transport - unregister a transport implementation
114 * @transport: transport to unregister
117 * 0: transport successfully unregistered
118 * -ENOENT: transport never registered
120 int xprt_unregister_transport(struct xprt_class *transport)
122 struct xprt_class *t;
126 spin_lock(&xprt_list_lock);
127 list_for_each_entry(t, &xprt_list, list) {
128 if (t == transport) {
130 "RPC: Unregistered %s transport module.\n",
132 list_del_init(&transport->list);
139 spin_unlock(&xprt_list_lock);
142 EXPORT_SYMBOL_GPL(xprt_unregister_transport);
145 * xprt_load_transport - load a transport implementation
146 * @transport_name: transport to load
149 * 0: transport successfully loaded
150 * -ENOENT: transport module not available
152 int xprt_load_transport(const char *transport_name)
154 struct xprt_class *t;
158 spin_lock(&xprt_list_lock);
159 list_for_each_entry(t, &xprt_list, list) {
160 if (strcmp(t->name, transport_name) == 0) {
161 spin_unlock(&xprt_list_lock);
165 spin_unlock(&xprt_list_lock);
166 result = request_module("xprt%s", transport_name);
170 EXPORT_SYMBOL_GPL(xprt_load_transport);
172 static void xprt_clear_locked(struct rpc_xprt *xprt)
174 xprt->snd_task = NULL;
175 if (!test_bit(XPRT_CLOSE_WAIT, &xprt->state)) {
176 smp_mb__before_atomic();
177 clear_bit(XPRT_LOCKED, &xprt->state);
178 smp_mb__after_atomic();
180 queue_work(xprtiod_workqueue, &xprt->task_cleanup);
184 * xprt_reserve_xprt - serialize write access to transports
185 * @task: task that is requesting access to the transport
186 * @xprt: pointer to the target transport
188 * This prevents mixing the payload of separate requests, and prevents
189 * transport connects from colliding with writes. No congestion control
192 int xprt_reserve_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
194 struct rpc_rqst *req = task->tk_rqstp;
196 if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
197 if (task == xprt->snd_task)
201 if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
203 xprt->snd_task = task;
208 xprt_clear_locked(xprt);
210 dprintk("RPC: %5u failed to lock transport %p\n",
212 task->tk_timeout = RPC_IS_SOFT(task) ? req->rq_timeout : 0;
213 task->tk_status = -EAGAIN;
214 rpc_sleep_on(&xprt->sending, task, NULL);
217 EXPORT_SYMBOL_GPL(xprt_reserve_xprt);
220 xprt_need_congestion_window_wait(struct rpc_xprt *xprt)
222 return test_bit(XPRT_CWND_WAIT, &xprt->state);
226 xprt_set_congestion_window_wait(struct rpc_xprt *xprt)
228 if (!list_empty(&xprt->xmit_queue)) {
229 /* Peek at head of queue to see if it can make progress */
230 if (list_first_entry(&xprt->xmit_queue, struct rpc_rqst,
234 set_bit(XPRT_CWND_WAIT, &xprt->state);
238 xprt_test_and_clear_congestion_window_wait(struct rpc_xprt *xprt)
240 if (!RPCXPRT_CONGESTED(xprt))
241 clear_bit(XPRT_CWND_WAIT, &xprt->state);
245 * xprt_reserve_xprt_cong - serialize write access to transports
246 * @task: task that is requesting access to the transport
248 * Same as xprt_reserve_xprt, but Van Jacobson congestion control is
249 * integrated into the decision of whether a request is allowed to be
250 * woken up and given access to the transport.
251 * Note that the lock is only granted if we know there are free slots.
253 int xprt_reserve_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task)
255 struct rpc_rqst *req = task->tk_rqstp;
257 if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
258 if (task == xprt->snd_task)
263 xprt->snd_task = task;
266 if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
268 if (!xprt_need_congestion_window_wait(xprt)) {
269 xprt->snd_task = task;
273 xprt_clear_locked(xprt);
275 dprintk("RPC: %5u failed to lock transport %p\n", task->tk_pid, xprt);
276 task->tk_timeout = RPC_IS_SOFT(task) ? req->rq_timeout : 0;
277 task->tk_status = -EAGAIN;
278 rpc_sleep_on(&xprt->sending, task, NULL);
281 EXPORT_SYMBOL_GPL(xprt_reserve_xprt_cong);
283 static inline int xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
287 if (test_bit(XPRT_LOCKED, &xprt->state) && xprt->snd_task == task)
289 spin_lock_bh(&xprt->transport_lock);
290 retval = xprt->ops->reserve_xprt(xprt, task);
291 spin_unlock_bh(&xprt->transport_lock);
295 static bool __xprt_lock_write_func(struct rpc_task *task, void *data)
297 struct rpc_xprt *xprt = data;
299 xprt->snd_task = task;
303 static void __xprt_lock_write_next(struct rpc_xprt *xprt)
305 if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
307 if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
309 if (rpc_wake_up_first_on_wq(xprtiod_workqueue, &xprt->sending,
310 __xprt_lock_write_func, xprt))
313 xprt_clear_locked(xprt);
316 static void __xprt_lock_write_next_cong(struct rpc_xprt *xprt)
318 if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
320 if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
322 if (xprt_need_congestion_window_wait(xprt))
324 if (rpc_wake_up_first_on_wq(xprtiod_workqueue, &xprt->sending,
325 __xprt_lock_write_func, xprt))
328 xprt_clear_locked(xprt);
332 * xprt_release_xprt - allow other requests to use a transport
333 * @xprt: transport with other tasks potentially waiting
334 * @task: task that is releasing access to the transport
336 * Note that "task" can be NULL. No congestion control is provided.
338 void xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
340 if (xprt->snd_task == task) {
341 xprt_clear_locked(xprt);
342 __xprt_lock_write_next(xprt);
345 EXPORT_SYMBOL_GPL(xprt_release_xprt);
348 * xprt_release_xprt_cong - allow other requests to use a transport
349 * @xprt: transport with other tasks potentially waiting
350 * @task: task that is releasing access to the transport
352 * Note that "task" can be NULL. Another task is awoken to use the
353 * transport if the transport's congestion window allows it.
355 void xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task)
357 if (xprt->snd_task == task) {
358 xprt_clear_locked(xprt);
359 __xprt_lock_write_next_cong(xprt);
362 EXPORT_SYMBOL_GPL(xprt_release_xprt_cong);
364 static inline void xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
366 if (xprt->snd_task != task)
368 spin_lock_bh(&xprt->transport_lock);
369 xprt->ops->release_xprt(xprt, task);
370 spin_unlock_bh(&xprt->transport_lock);
374 * Van Jacobson congestion avoidance. Check if the congestion window
375 * overflowed. Put the task to sleep if this is the case.
378 __xprt_get_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
382 dprintk("RPC: %5u xprt_cwnd_limited cong = %lu cwnd = %lu\n",
383 req->rq_task->tk_pid, xprt->cong, xprt->cwnd);
384 if (RPCXPRT_CONGESTED(xprt)) {
385 xprt_set_congestion_window_wait(xprt);
389 xprt->cong += RPC_CWNDSCALE;
394 * Adjust the congestion window, and wake up the next task
395 * that has been sleeping due to congestion
398 __xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
403 xprt->cong -= RPC_CWNDSCALE;
404 xprt_test_and_clear_congestion_window_wait(xprt);
405 __xprt_lock_write_next_cong(xprt);
409 * xprt_request_get_cong - Request congestion control credits
410 * @xprt: pointer to transport
411 * @req: pointer to RPC request
413 * Useful for transports that require congestion control.
416 xprt_request_get_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
422 spin_lock_bh(&xprt->transport_lock);
423 ret = __xprt_get_cong(xprt, req) != 0;
424 spin_unlock_bh(&xprt->transport_lock);
427 EXPORT_SYMBOL_GPL(xprt_request_get_cong);
430 * xprt_release_rqst_cong - housekeeping when request is complete
431 * @task: RPC request that recently completed
433 * Useful for transports that require congestion control.
435 void xprt_release_rqst_cong(struct rpc_task *task)
437 struct rpc_rqst *req = task->tk_rqstp;
439 __xprt_put_cong(req->rq_xprt, req);
441 EXPORT_SYMBOL_GPL(xprt_release_rqst_cong);
444 * Clear the congestion window wait flag and wake up the next
445 * entry on xprt->sending
448 xprt_clear_congestion_window_wait(struct rpc_xprt *xprt)
450 if (test_and_clear_bit(XPRT_CWND_WAIT, &xprt->state)) {
451 spin_lock_bh(&xprt->transport_lock);
452 __xprt_lock_write_next_cong(xprt);
453 spin_unlock_bh(&xprt->transport_lock);
458 * xprt_adjust_cwnd - adjust transport congestion window
459 * @xprt: pointer to xprt
460 * @task: recently completed RPC request used to adjust window
461 * @result: result code of completed RPC request
463 * The transport code maintains an estimate on the maximum number of out-
464 * standing RPC requests, using a smoothed version of the congestion
465 * avoidance implemented in 44BSD. This is basically the Van Jacobson
466 * congestion algorithm: If a retransmit occurs, the congestion window is
467 * halved; otherwise, it is incremented by 1/cwnd when
469 * - a reply is received and
470 * - a full number of requests are outstanding and
471 * - the congestion window hasn't been updated recently.
473 void xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result)
475 struct rpc_rqst *req = task->tk_rqstp;
476 unsigned long cwnd = xprt->cwnd;
478 if (result >= 0 && cwnd <= xprt->cong) {
479 /* The (cwnd >> 1) term makes sure
480 * the result gets rounded properly. */
481 cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd;
482 if (cwnd > RPC_MAXCWND(xprt))
483 cwnd = RPC_MAXCWND(xprt);
484 __xprt_lock_write_next_cong(xprt);
485 } else if (result == -ETIMEDOUT) {
487 if (cwnd < RPC_CWNDSCALE)
488 cwnd = RPC_CWNDSCALE;
490 dprintk("RPC: cong %ld, cwnd was %ld, now %ld\n",
491 xprt->cong, xprt->cwnd, cwnd);
493 __xprt_put_cong(xprt, req);
495 EXPORT_SYMBOL_GPL(xprt_adjust_cwnd);
498 * xprt_wake_pending_tasks - wake all tasks on a transport's pending queue
499 * @xprt: transport with waiting tasks
500 * @status: result code to plant in each task before waking it
503 void xprt_wake_pending_tasks(struct rpc_xprt *xprt, int status)
506 rpc_wake_up_status(&xprt->pending, status);
508 rpc_wake_up(&xprt->pending);
510 EXPORT_SYMBOL_GPL(xprt_wake_pending_tasks);
513 * xprt_wait_for_buffer_space - wait for transport output buffer to clear
516 * Note that we only set the timer for the case of RPC_IS_SOFT(), since
517 * we don't in general want to force a socket disconnection due to
518 * an incomplete RPC call transmission.
520 void xprt_wait_for_buffer_space(struct rpc_xprt *xprt)
522 set_bit(XPRT_WRITE_SPACE, &xprt->state);
524 EXPORT_SYMBOL_GPL(xprt_wait_for_buffer_space);
527 xprt_clear_write_space_locked(struct rpc_xprt *xprt)
529 if (test_and_clear_bit(XPRT_WRITE_SPACE, &xprt->state)) {
530 __xprt_lock_write_next(xprt);
531 dprintk("RPC: write space: waking waiting task on "
539 * xprt_write_space - wake the task waiting for transport output buffer space
540 * @xprt: transport with waiting tasks
542 * Can be called in a soft IRQ context, so xprt_write_space never sleeps.
544 bool xprt_write_space(struct rpc_xprt *xprt)
548 if (!test_bit(XPRT_WRITE_SPACE, &xprt->state))
550 spin_lock_bh(&xprt->transport_lock);
551 ret = xprt_clear_write_space_locked(xprt);
552 spin_unlock_bh(&xprt->transport_lock);
555 EXPORT_SYMBOL_GPL(xprt_write_space);
558 * xprt_set_retrans_timeout_def - set a request's retransmit timeout
559 * @task: task whose timeout is to be set
561 * Set a request's retransmit timeout based on the transport's
562 * default timeout parameters. Used by transports that don't adjust
563 * the retransmit timeout based on round-trip time estimation.
565 void xprt_set_retrans_timeout_def(struct rpc_task *task)
567 task->tk_timeout = task->tk_rqstp->rq_timeout;
569 EXPORT_SYMBOL_GPL(xprt_set_retrans_timeout_def);
572 * xprt_set_retrans_timeout_rtt - set a request's retransmit timeout
573 * @task: task whose timeout is to be set
575 * Set a request's retransmit timeout using the RTT estimator.
577 void xprt_set_retrans_timeout_rtt(struct rpc_task *task)
579 int timer = task->tk_msg.rpc_proc->p_timer;
580 struct rpc_clnt *clnt = task->tk_client;
581 struct rpc_rtt *rtt = clnt->cl_rtt;
582 struct rpc_rqst *req = task->tk_rqstp;
583 unsigned long max_timeout = clnt->cl_timeout->to_maxval;
585 task->tk_timeout = rpc_calc_rto(rtt, timer);
586 task->tk_timeout <<= rpc_ntimeo(rtt, timer) + req->rq_retries;
587 if (task->tk_timeout > max_timeout || task->tk_timeout == 0)
588 task->tk_timeout = max_timeout;
590 EXPORT_SYMBOL_GPL(xprt_set_retrans_timeout_rtt);
592 static void xprt_reset_majortimeo(struct rpc_rqst *req)
594 const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout;
596 req->rq_majortimeo = req->rq_timeout;
597 if (to->to_exponential)
598 req->rq_majortimeo <<= to->to_retries;
600 req->rq_majortimeo += to->to_increment * to->to_retries;
601 if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0)
602 req->rq_majortimeo = to->to_maxval;
603 req->rq_majortimeo += jiffies;
607 * xprt_adjust_timeout - adjust timeout values for next retransmit
608 * @req: RPC request containing parameters to use for the adjustment
611 int xprt_adjust_timeout(struct rpc_rqst *req)
613 struct rpc_xprt *xprt = req->rq_xprt;
614 const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout;
617 if (time_before(jiffies, req->rq_majortimeo)) {
618 if (to->to_exponential)
619 req->rq_timeout <<= 1;
621 req->rq_timeout += to->to_increment;
622 if (to->to_maxval && req->rq_timeout >= to->to_maxval)
623 req->rq_timeout = to->to_maxval;
626 req->rq_timeout = to->to_initval;
628 xprt_reset_majortimeo(req);
629 /* Reset the RTT counters == "slow start" */
630 spin_lock_bh(&xprt->transport_lock);
631 rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
632 spin_unlock_bh(&xprt->transport_lock);
636 if (req->rq_timeout == 0) {
637 printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n");
638 req->rq_timeout = 5 * HZ;
643 static void xprt_autoclose(struct work_struct *work)
645 struct rpc_xprt *xprt =
646 container_of(work, struct rpc_xprt, task_cleanup);
648 clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
649 xprt->ops->close(xprt);
650 xprt_release_write(xprt, NULL);
651 wake_up_bit(&xprt->state, XPRT_LOCKED);
655 * xprt_disconnect_done - mark a transport as disconnected
656 * @xprt: transport to flag for disconnect
659 void xprt_disconnect_done(struct rpc_xprt *xprt)
661 dprintk("RPC: disconnected transport %p\n", xprt);
662 spin_lock_bh(&xprt->transport_lock);
663 xprt_clear_connected(xprt);
664 xprt_clear_write_space_locked(xprt);
665 xprt_wake_pending_tasks(xprt, -EAGAIN);
666 spin_unlock_bh(&xprt->transport_lock);
668 EXPORT_SYMBOL_GPL(xprt_disconnect_done);
671 * xprt_force_disconnect - force a transport to disconnect
672 * @xprt: transport to disconnect
675 void xprt_force_disconnect(struct rpc_xprt *xprt)
677 /* Don't race with the test_bit() in xprt_clear_locked() */
678 spin_lock_bh(&xprt->transport_lock);
679 set_bit(XPRT_CLOSE_WAIT, &xprt->state);
680 /* Try to schedule an autoclose RPC call */
681 if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
682 queue_work(xprtiod_workqueue, &xprt->task_cleanup);
683 xprt_wake_pending_tasks(xprt, -EAGAIN);
684 spin_unlock_bh(&xprt->transport_lock);
686 EXPORT_SYMBOL_GPL(xprt_force_disconnect);
689 xprt_connect_cookie(struct rpc_xprt *xprt)
691 return READ_ONCE(xprt->connect_cookie);
695 xprt_request_retransmit_after_disconnect(struct rpc_task *task)
697 struct rpc_rqst *req = task->tk_rqstp;
698 struct rpc_xprt *xprt = req->rq_xprt;
700 return req->rq_connect_cookie != xprt_connect_cookie(xprt) ||
701 !xprt_connected(xprt);
705 * xprt_conditional_disconnect - force a transport to disconnect
706 * @xprt: transport to disconnect
707 * @cookie: 'connection cookie'
709 * This attempts to break the connection if and only if 'cookie' matches
710 * the current transport 'connection cookie'. It ensures that we don't
711 * try to break the connection more than once when we need to retransmit
712 * a batch of RPC requests.
715 void xprt_conditional_disconnect(struct rpc_xprt *xprt, unsigned int cookie)
717 /* Don't race with the test_bit() in xprt_clear_locked() */
718 spin_lock_bh(&xprt->transport_lock);
719 if (cookie != xprt->connect_cookie)
721 if (test_bit(XPRT_CLOSING, &xprt->state))
723 set_bit(XPRT_CLOSE_WAIT, &xprt->state);
724 /* Try to schedule an autoclose RPC call */
725 if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
726 queue_work(xprtiod_workqueue, &xprt->task_cleanup);
727 xprt_wake_pending_tasks(xprt, -EAGAIN);
729 spin_unlock_bh(&xprt->transport_lock);
733 xprt_has_timer(const struct rpc_xprt *xprt)
735 return xprt->idle_timeout != 0;
739 xprt_schedule_autodisconnect(struct rpc_xprt *xprt)
740 __must_hold(&xprt->transport_lock)
742 if (RB_EMPTY_ROOT(&xprt->recv_queue) && xprt_has_timer(xprt))
743 mod_timer(&xprt->timer, xprt->last_used + xprt->idle_timeout);
747 xprt_init_autodisconnect(struct timer_list *t)
749 struct rpc_xprt *xprt = from_timer(xprt, t, timer);
751 spin_lock(&xprt->transport_lock);
752 if (!RB_EMPTY_ROOT(&xprt->recv_queue))
754 /* Reset xprt->last_used to avoid connect/autodisconnect cycling */
755 xprt->last_used = jiffies;
756 if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
758 spin_unlock(&xprt->transport_lock);
759 queue_work(xprtiod_workqueue, &xprt->task_cleanup);
762 spin_unlock(&xprt->transport_lock);
765 bool xprt_lock_connect(struct rpc_xprt *xprt,
766 struct rpc_task *task,
771 spin_lock_bh(&xprt->transport_lock);
772 if (!test_bit(XPRT_LOCKED, &xprt->state))
774 if (xprt->snd_task != task)
776 xprt->snd_task = cookie;
779 spin_unlock_bh(&xprt->transport_lock);
783 void xprt_unlock_connect(struct rpc_xprt *xprt, void *cookie)
785 spin_lock_bh(&xprt->transport_lock);
786 if (xprt->snd_task != cookie)
788 if (!test_bit(XPRT_LOCKED, &xprt->state))
790 xprt->snd_task =NULL;
791 xprt->ops->release_xprt(xprt, NULL);
792 xprt_schedule_autodisconnect(xprt);
794 spin_unlock_bh(&xprt->transport_lock);
795 wake_up_bit(&xprt->state, XPRT_LOCKED);
799 * xprt_connect - schedule a transport connect operation
800 * @task: RPC task that is requesting the connect
803 void xprt_connect(struct rpc_task *task)
805 struct rpc_xprt *xprt = task->tk_rqstp->rq_xprt;
807 dprintk("RPC: %5u xprt_connect xprt %p %s connected\n", task->tk_pid,
808 xprt, (xprt_connected(xprt) ? "is" : "is not"));
810 if (!xprt_bound(xprt)) {
811 task->tk_status = -EAGAIN;
814 if (!xprt_lock_write(xprt, task))
817 if (test_and_clear_bit(XPRT_CLOSE_WAIT, &xprt->state))
818 xprt->ops->close(xprt);
820 if (!xprt_connected(xprt)) {
821 task->tk_timeout = task->tk_rqstp->rq_timeout;
822 task->tk_rqstp->rq_connect_cookie = xprt->connect_cookie;
823 rpc_sleep_on(&xprt->pending, task, xprt_connect_status);
825 if (test_bit(XPRT_CLOSING, &xprt->state))
827 if (xprt_test_and_set_connecting(xprt))
829 xprt->stat.connect_start = jiffies;
830 xprt->ops->connect(xprt, task);
832 xprt_release_write(xprt, task);
835 static void xprt_connect_status(struct rpc_task *task)
837 switch (task->tk_status) {
839 dprintk("RPC: %5u xprt_connect_status: connection established\n",
849 dprintk("RPC: %5u xprt_connect_status: retrying\n", task->tk_pid);
852 dprintk("RPC: %5u xprt_connect_status: connect attempt timed "
853 "out\n", task->tk_pid);
856 dprintk("RPC: %5u xprt_connect_status: error %d connecting to "
857 "server %s\n", task->tk_pid, -task->tk_status,
858 task->tk_rqstp->rq_xprt->servername);
859 task->tk_status = -EIO;
863 enum xprt_xid_rb_cmp {
868 static enum xprt_xid_rb_cmp
869 xprt_xid_cmp(__be32 xid1, __be32 xid2)
873 if ((__force u32)xid1 < (__force u32)xid2)
878 static struct rpc_rqst *
879 xprt_request_rb_find(struct rpc_xprt *xprt, __be32 xid)
881 struct rb_node *n = xprt->recv_queue.rb_node;
882 struct rpc_rqst *req;
885 req = rb_entry(n, struct rpc_rqst, rq_recv);
886 switch (xprt_xid_cmp(xid, req->rq_xid)) {
901 xprt_request_rb_insert(struct rpc_xprt *xprt, struct rpc_rqst *new)
903 struct rb_node **p = &xprt->recv_queue.rb_node;
904 struct rb_node *n = NULL;
905 struct rpc_rqst *req;
909 req = rb_entry(n, struct rpc_rqst, rq_recv);
910 switch(xprt_xid_cmp(new->rq_xid, req->rq_xid)) {
918 WARN_ON_ONCE(new != req);
922 rb_link_node(&new->rq_recv, n, p);
923 rb_insert_color(&new->rq_recv, &xprt->recv_queue);
927 xprt_request_rb_remove(struct rpc_xprt *xprt, struct rpc_rqst *req)
929 rb_erase(&req->rq_recv, &xprt->recv_queue);
933 * xprt_lookup_rqst - find an RPC request corresponding to an XID
934 * @xprt: transport on which the original request was transmitted
935 * @xid: RPC XID of incoming reply
937 * Caller holds xprt->queue_lock.
939 struct rpc_rqst *xprt_lookup_rqst(struct rpc_xprt *xprt, __be32 xid)
941 struct rpc_rqst *entry;
943 entry = xprt_request_rb_find(xprt, xid);
945 trace_xprt_lookup_rqst(xprt, xid, 0);
946 entry->rq_rtt = ktime_sub(ktime_get(), entry->rq_xtime);
950 dprintk("RPC: xprt_lookup_rqst did not find xid %08x\n",
952 trace_xprt_lookup_rqst(xprt, xid, -ENOENT);
953 xprt->stat.bad_xids++;
956 EXPORT_SYMBOL_GPL(xprt_lookup_rqst);
959 xprt_is_pinned_rqst(struct rpc_rqst *req)
961 return atomic_read(&req->rq_pin) != 0;
965 * xprt_pin_rqst - Pin a request on the transport receive list
966 * @req: Request to pin
968 * Caller must ensure this is atomic with the call to xprt_lookup_rqst()
969 * so should be holding the xprt receive lock.
971 void xprt_pin_rqst(struct rpc_rqst *req)
973 atomic_inc(&req->rq_pin);
975 EXPORT_SYMBOL_GPL(xprt_pin_rqst);
978 * xprt_unpin_rqst - Unpin a request on the transport receive list
979 * @req: Request to pin
981 * Caller should be holding the xprt receive lock.
983 void xprt_unpin_rqst(struct rpc_rqst *req)
985 if (!test_bit(RPC_TASK_MSG_PIN_WAIT, &req->rq_task->tk_runstate)) {
986 atomic_dec(&req->rq_pin);
989 if (atomic_dec_and_test(&req->rq_pin))
990 wake_up_var(&req->rq_pin);
992 EXPORT_SYMBOL_GPL(xprt_unpin_rqst);
994 static void xprt_wait_on_pinned_rqst(struct rpc_rqst *req)
996 wait_var_event(&req->rq_pin, !xprt_is_pinned_rqst(req));
1000 xprt_request_data_received(struct rpc_task *task)
1002 return !test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) &&
1003 READ_ONCE(task->tk_rqstp->rq_reply_bytes_recvd) != 0;
1007 xprt_request_need_enqueue_receive(struct rpc_task *task, struct rpc_rqst *req)
1009 return !test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) &&
1010 READ_ONCE(task->tk_rqstp->rq_reply_bytes_recvd) == 0;
1014 * xprt_request_enqueue_receive - Add an request to the receive queue
1019 xprt_request_enqueue_receive(struct rpc_task *task)
1021 struct rpc_rqst *req = task->tk_rqstp;
1022 struct rpc_xprt *xprt = req->rq_xprt;
1024 if (!xprt_request_need_enqueue_receive(task, req))
1026 spin_lock(&xprt->queue_lock);
1028 /* Update the softirq receive buffer */
1029 memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
1030 sizeof(req->rq_private_buf));
1032 /* Add request to the receive list */
1033 xprt_request_rb_insert(xprt, req);
1034 set_bit(RPC_TASK_NEED_RECV, &task->tk_runstate);
1035 spin_unlock(&xprt->queue_lock);
1037 xprt_reset_majortimeo(req);
1038 /* Turn off autodisconnect */
1039 del_singleshot_timer_sync(&xprt->timer);
1043 * xprt_request_dequeue_receive_locked - Remove a request from the receive queue
1046 * Caller must hold xprt->queue_lock.
1049 xprt_request_dequeue_receive_locked(struct rpc_task *task)
1051 struct rpc_rqst *req = task->tk_rqstp;
1053 if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
1054 xprt_request_rb_remove(req->rq_xprt, req);
1058 * xprt_update_rtt - Update RPC RTT statistics
1059 * @task: RPC request that recently completed
1061 * Caller holds xprt->queue_lock.
1063 void xprt_update_rtt(struct rpc_task *task)
1065 struct rpc_rqst *req = task->tk_rqstp;
1066 struct rpc_rtt *rtt = task->tk_client->cl_rtt;
1067 unsigned int timer = task->tk_msg.rpc_proc->p_timer;
1068 long m = usecs_to_jiffies(ktime_to_us(req->rq_rtt));
1071 if (req->rq_ntrans == 1)
1072 rpc_update_rtt(rtt, timer, m);
1073 rpc_set_timeo(rtt, timer, req->rq_ntrans - 1);
1076 EXPORT_SYMBOL_GPL(xprt_update_rtt);
1079 * xprt_complete_rqst - called when reply processing is complete
1080 * @task: RPC request that recently completed
1081 * @copied: actual number of bytes received from the transport
1083 * Caller holds xprt->queue_lock.
1085 void xprt_complete_rqst(struct rpc_task *task, int copied)
1087 struct rpc_rqst *req = task->tk_rqstp;
1088 struct rpc_xprt *xprt = req->rq_xprt;
1090 dprintk("RPC: %5u xid %08x complete (%d bytes received)\n",
1091 task->tk_pid, ntohl(req->rq_xid), copied);
1092 trace_xprt_complete_rqst(xprt, req->rq_xid, copied);
1096 req->rq_private_buf.len = copied;
1097 /* Ensure all writes are done before we update */
1098 /* req->rq_reply_bytes_recvd */
1100 req->rq_reply_bytes_recvd = copied;
1101 xprt_request_dequeue_receive_locked(task);
1102 rpc_wake_up_queued_task(&xprt->pending, task);
1104 EXPORT_SYMBOL_GPL(xprt_complete_rqst);
1106 static void xprt_timer(struct rpc_task *task)
1108 struct rpc_rqst *req = task->tk_rqstp;
1109 struct rpc_xprt *xprt = req->rq_xprt;
1111 if (task->tk_status != -ETIMEDOUT)
1114 trace_xprt_timer(xprt, req->rq_xid, task->tk_status);
1115 if (!req->rq_reply_bytes_recvd) {
1116 if (xprt->ops->timer)
1117 xprt->ops->timer(xprt, task);
1119 task->tk_status = 0;
1123 * xprt_request_wait_receive - wait for the reply to an RPC request
1124 * @task: RPC task about to send a request
1127 void xprt_request_wait_receive(struct rpc_task *task)
1129 struct rpc_rqst *req = task->tk_rqstp;
1130 struct rpc_xprt *xprt = req->rq_xprt;
1132 if (!test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
1135 * Sleep on the pending queue if we're expecting a reply.
1136 * The spinlock ensures atomicity between the test of
1137 * req->rq_reply_bytes_recvd, and the call to rpc_sleep_on().
1139 spin_lock(&xprt->queue_lock);
1140 if (test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate)) {
1141 xprt->ops->set_retrans_timeout(task);
1142 rpc_sleep_on(&xprt->pending, task, xprt_timer);
1144 * Send an extra queue wakeup call if the
1145 * connection was dropped in case the call to
1146 * rpc_sleep_on() raced.
1148 if (xprt_request_retransmit_after_disconnect(task))
1149 rpc_wake_up_queued_task_set_status(&xprt->pending,
1152 spin_unlock(&xprt->queue_lock);
1156 xprt_request_need_enqueue_transmit(struct rpc_task *task, struct rpc_rqst *req)
1158 return !test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate);
1162 * xprt_request_enqueue_transmit - queue a task for transmission
1163 * @task: pointer to rpc_task
1165 * Add a task to the transmission queue.
1168 xprt_request_enqueue_transmit(struct rpc_task *task)
1170 struct rpc_rqst *pos, *req = task->tk_rqstp;
1171 struct rpc_xprt *xprt = req->rq_xprt;
1173 if (xprt_request_need_enqueue_transmit(task, req)) {
1174 spin_lock(&xprt->queue_lock);
1176 * Requests that carry congestion control credits are added
1177 * to the head of the list to avoid starvation issues.
1180 xprt_clear_congestion_window_wait(xprt);
1181 list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1184 /* Note: req is added _before_ pos */
1185 list_add_tail(&req->rq_xmit, &pos->rq_xmit);
1186 INIT_LIST_HEAD(&req->rq_xmit2);
1189 } else if (RPC_IS_SWAPPER(task)) {
1190 list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1191 if (pos->rq_cong || pos->rq_bytes_sent)
1193 if (RPC_IS_SWAPPER(pos->rq_task))
1195 /* Note: req is added _before_ pos */
1196 list_add_tail(&req->rq_xmit, &pos->rq_xmit);
1197 INIT_LIST_HEAD(&req->rq_xmit2);
1201 list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1202 if (pos->rq_task->tk_owner != task->tk_owner)
1204 list_add_tail(&req->rq_xmit2, &pos->rq_xmit2);
1205 INIT_LIST_HEAD(&req->rq_xmit);
1209 list_add_tail(&req->rq_xmit, &xprt->xmit_queue);
1210 INIT_LIST_HEAD(&req->rq_xmit2);
1212 set_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate);
1213 spin_unlock(&xprt->queue_lock);
1218 * xprt_request_dequeue_transmit_locked - remove a task from the transmission queue
1219 * @task: pointer to rpc_task
1221 * Remove a task from the transmission queue
1222 * Caller must hold xprt->queue_lock
1225 xprt_request_dequeue_transmit_locked(struct rpc_task *task)
1227 struct rpc_rqst *req = task->tk_rqstp;
1229 if (!test_and_clear_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1231 if (!list_empty(&req->rq_xmit)) {
1232 list_del(&req->rq_xmit);
1233 if (!list_empty(&req->rq_xmit2)) {
1234 struct rpc_rqst *next = list_first_entry(&req->rq_xmit2,
1235 struct rpc_rqst, rq_xmit2);
1236 list_del(&req->rq_xmit2);
1237 list_add_tail(&next->rq_xmit, &next->rq_xprt->xmit_queue);
1240 list_del(&req->rq_xmit2);
1244 * xprt_request_dequeue_transmit - remove a task from the transmission queue
1245 * @task: pointer to rpc_task
1247 * Remove a task from the transmission queue
1250 xprt_request_dequeue_transmit(struct rpc_task *task)
1252 struct rpc_rqst *req = task->tk_rqstp;
1253 struct rpc_xprt *xprt = req->rq_xprt;
1255 spin_lock(&xprt->queue_lock);
1256 xprt_request_dequeue_transmit_locked(task);
1257 spin_unlock(&xprt->queue_lock);
1261 * xprt_request_prepare - prepare an encoded request for transport
1262 * @req: pointer to rpc_rqst
1264 * Calls into the transport layer to do whatever is needed to prepare
1265 * the request for transmission or receive.
1268 xprt_request_prepare(struct rpc_rqst *req)
1270 struct rpc_xprt *xprt = req->rq_xprt;
1272 if (xprt->ops->prepare_request)
1273 xprt->ops->prepare_request(req);
1277 * xprt_request_need_retransmit - Test if a task needs retransmission
1278 * @task: pointer to rpc_task
1280 * Test for whether a connection breakage requires the task to retransmit
1283 xprt_request_need_retransmit(struct rpc_task *task)
1285 return xprt_request_retransmit_after_disconnect(task);
1289 * xprt_prepare_transmit - reserve the transport before sending a request
1290 * @task: RPC task about to send a request
1293 bool xprt_prepare_transmit(struct rpc_task *task)
1295 struct rpc_rqst *req = task->tk_rqstp;
1296 struct rpc_xprt *xprt = req->rq_xprt;
1298 dprintk("RPC: %5u xprt_prepare_transmit\n", task->tk_pid);
1300 if (!xprt_lock_write(xprt, task)) {
1301 /* Race breaker: someone may have transmitted us */
1302 if (!test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1303 rpc_wake_up_queued_task_set_status(&xprt->sending,
1311 void xprt_end_transmit(struct rpc_task *task)
1313 xprt_release_write(task->tk_rqstp->rq_xprt, task);
1317 * xprt_request_transmit - send an RPC request on a transport
1318 * @req: pointer to request to transmit
1319 * @snd_task: RPC task that owns the transport lock
1321 * This performs the transmission of a single request.
1322 * Note that if the request is not the same as snd_task, then it
1323 * does need to be pinned.
1324 * Returns '0' on success.
1327 xprt_request_transmit(struct rpc_rqst *req, struct rpc_task *snd_task)
1329 struct rpc_xprt *xprt = req->rq_xprt;
1330 struct rpc_task *task = req->rq_task;
1331 unsigned int connect_cookie;
1332 int is_retrans = RPC_WAS_SENT(task);
1335 dprintk("RPC: %5u xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
1337 if (!req->rq_bytes_sent) {
1338 if (xprt_request_data_received(task)) {
1342 /* Verify that our message lies in the RPCSEC_GSS window */
1343 if (rpcauth_xmit_need_reencode(task)) {
1350 * Update req->rq_ntrans before transmitting to avoid races with
1351 * xprt_update_rtt(), which needs to know that it is recording a
1352 * reply to the first transmission.
1356 connect_cookie = xprt->connect_cookie;
1357 status = xprt->ops->send_request(req);
1358 trace_xprt_transmit(xprt, req->rq_xid, status);
1365 task->tk_client->cl_stats->rpcretrans++;
1367 xprt_inject_disconnect(xprt);
1369 dprintk("RPC: %5u xmit complete\n", task->tk_pid);
1370 task->tk_flags |= RPC_TASK_SENT;
1371 spin_lock_bh(&xprt->transport_lock);
1374 xprt->stat.req_u += xprt->stat.sends - xprt->stat.recvs;
1375 xprt->stat.bklog_u += xprt->backlog.qlen;
1376 xprt->stat.sending_u += xprt->sending.qlen;
1377 xprt->stat.pending_u += xprt->pending.qlen;
1378 spin_unlock_bh(&xprt->transport_lock);
1380 req->rq_connect_cookie = connect_cookie;
1382 xprt_request_dequeue_transmit(task);
1383 rpc_wake_up_queued_task_set_status(&xprt->sending, task, status);
1388 * xprt_transmit - send an RPC request on a transport
1389 * @task: controlling RPC task
1391 * Attempts to drain the transmit queue. On exit, either the transport
1392 * signalled an error that needs to be handled before transmission can
1393 * resume, or @task finished transmitting, and detected that it already
1397 xprt_transmit(struct rpc_task *task)
1399 struct rpc_rqst *next, *req = task->tk_rqstp;
1400 struct rpc_xprt *xprt = req->rq_xprt;
1403 spin_lock(&xprt->queue_lock);
1404 while (!list_empty(&xprt->xmit_queue)) {
1405 next = list_first_entry(&xprt->xmit_queue,
1406 struct rpc_rqst, rq_xmit);
1407 xprt_pin_rqst(next);
1408 spin_unlock(&xprt->queue_lock);
1409 status = xprt_request_transmit(next, task);
1410 if (status == -EBADMSG && next != req)
1413 spin_lock(&xprt->queue_lock);
1414 xprt_unpin_rqst(next);
1416 if (!xprt_request_data_received(task) ||
1417 test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1419 } else if (test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1420 task->tk_status = status;
1423 spin_unlock(&xprt->queue_lock);
1426 static void xprt_add_backlog(struct rpc_xprt *xprt, struct rpc_task *task)
1428 set_bit(XPRT_CONGESTED, &xprt->state);
1429 rpc_sleep_on(&xprt->backlog, task, NULL);
1432 static void xprt_wake_up_backlog(struct rpc_xprt *xprt)
1434 if (rpc_wake_up_next(&xprt->backlog) == NULL)
1435 clear_bit(XPRT_CONGESTED, &xprt->state);
1438 static bool xprt_throttle_congested(struct rpc_xprt *xprt, struct rpc_task *task)
1442 if (!test_bit(XPRT_CONGESTED, &xprt->state))
1444 spin_lock(&xprt->reserve_lock);
1445 if (test_bit(XPRT_CONGESTED, &xprt->state)) {
1446 rpc_sleep_on(&xprt->backlog, task, NULL);
1449 spin_unlock(&xprt->reserve_lock);
1454 static struct rpc_rqst *xprt_dynamic_alloc_slot(struct rpc_xprt *xprt)
1456 struct rpc_rqst *req = ERR_PTR(-EAGAIN);
1458 if (xprt->num_reqs >= xprt->max_reqs)
1461 spin_unlock(&xprt->reserve_lock);
1462 req = kzalloc(sizeof(struct rpc_rqst), GFP_NOFS);
1463 spin_lock(&xprt->reserve_lock);
1467 req = ERR_PTR(-ENOMEM);
1472 static bool xprt_dynamic_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
1474 if (xprt->num_reqs > xprt->min_reqs) {
1482 void xprt_alloc_slot(struct rpc_xprt *xprt, struct rpc_task *task)
1484 struct rpc_rqst *req;
1486 spin_lock(&xprt->reserve_lock);
1487 if (!list_empty(&xprt->free)) {
1488 req = list_entry(xprt->free.next, struct rpc_rqst, rq_list);
1489 list_del(&req->rq_list);
1492 req = xprt_dynamic_alloc_slot(xprt);
1495 switch (PTR_ERR(req)) {
1497 dprintk("RPC: dynamic allocation of request slot "
1498 "failed! Retrying\n");
1499 task->tk_status = -ENOMEM;
1502 xprt_add_backlog(xprt, task);
1503 dprintk("RPC: waiting for request slot\n");
1506 task->tk_status = -EAGAIN;
1508 spin_unlock(&xprt->reserve_lock);
1511 xprt->stat.max_slots = max_t(unsigned int, xprt->stat.max_slots,
1513 spin_unlock(&xprt->reserve_lock);
1515 task->tk_status = 0;
1516 task->tk_rqstp = req;
1518 EXPORT_SYMBOL_GPL(xprt_alloc_slot);
1520 void xprt_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
1522 spin_lock(&xprt->reserve_lock);
1523 if (!xprt_dynamic_free_slot(xprt, req)) {
1524 memset(req, 0, sizeof(*req)); /* mark unused */
1525 list_add(&req->rq_list, &xprt->free);
1527 xprt_wake_up_backlog(xprt);
1528 spin_unlock(&xprt->reserve_lock);
1530 EXPORT_SYMBOL_GPL(xprt_free_slot);
1532 static void xprt_free_all_slots(struct rpc_xprt *xprt)
1534 struct rpc_rqst *req;
1535 while (!list_empty(&xprt->free)) {
1536 req = list_first_entry(&xprt->free, struct rpc_rqst, rq_list);
1537 list_del(&req->rq_list);
1542 struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
1543 unsigned int num_prealloc,
1544 unsigned int max_alloc)
1546 struct rpc_xprt *xprt;
1547 struct rpc_rqst *req;
1550 xprt = kzalloc(size, GFP_KERNEL);
1554 xprt_init(xprt, net);
1556 for (i = 0; i < num_prealloc; i++) {
1557 req = kzalloc(sizeof(struct rpc_rqst), GFP_KERNEL);
1560 list_add(&req->rq_list, &xprt->free);
1562 if (max_alloc > num_prealloc)
1563 xprt->max_reqs = max_alloc;
1565 xprt->max_reqs = num_prealloc;
1566 xprt->min_reqs = num_prealloc;
1567 xprt->num_reqs = num_prealloc;
1576 EXPORT_SYMBOL_GPL(xprt_alloc);
1578 void xprt_free(struct rpc_xprt *xprt)
1580 put_net(xprt->xprt_net);
1581 xprt_free_all_slots(xprt);
1582 kfree_rcu(xprt, rcu);
1584 EXPORT_SYMBOL_GPL(xprt_free);
1587 xprt_init_connect_cookie(struct rpc_rqst *req, struct rpc_xprt *xprt)
1589 req->rq_connect_cookie = xprt_connect_cookie(xprt) - 1;
1593 xprt_alloc_xid(struct rpc_xprt *xprt)
1597 spin_lock(&xprt->reserve_lock);
1598 xid = (__force __be32)xprt->xid++;
1599 spin_unlock(&xprt->reserve_lock);
1604 xprt_init_xid(struct rpc_xprt *xprt)
1606 xprt->xid = prandom_u32();
1610 xprt_request_init(struct rpc_task *task)
1612 struct rpc_xprt *xprt = task->tk_xprt;
1613 struct rpc_rqst *req = task->tk_rqstp;
1615 req->rq_timeout = task->tk_client->cl_timeout->to_initval;
1616 req->rq_task = task;
1617 req->rq_xprt = xprt;
1618 req->rq_buffer = NULL;
1619 req->rq_xid = xprt_alloc_xid(xprt);
1620 xprt_init_connect_cookie(req, xprt);
1621 req->rq_bytes_sent = 0;
1622 req->rq_snd_buf.len = 0;
1623 req->rq_snd_buf.buflen = 0;
1624 req->rq_rcv_buf.len = 0;
1625 req->rq_rcv_buf.buflen = 0;
1626 req->rq_release_snd_buf = NULL;
1627 xprt_reset_majortimeo(req);
1628 dprintk("RPC: %5u reserved req %p xid %08x\n", task->tk_pid,
1629 req, ntohl(req->rq_xid));
1633 xprt_do_reserve(struct rpc_xprt *xprt, struct rpc_task *task)
1635 xprt->ops->alloc_slot(xprt, task);
1636 if (task->tk_rqstp != NULL)
1637 xprt_request_init(task);
1641 * xprt_reserve - allocate an RPC request slot
1642 * @task: RPC task requesting a slot allocation
1644 * If the transport is marked as being congested, or if no more
1645 * slots are available, place the task on the transport's
1648 void xprt_reserve(struct rpc_task *task)
1650 struct rpc_xprt *xprt = task->tk_xprt;
1652 task->tk_status = 0;
1653 if (task->tk_rqstp != NULL)
1656 task->tk_timeout = 0;
1657 task->tk_status = -EAGAIN;
1658 if (!xprt_throttle_congested(xprt, task))
1659 xprt_do_reserve(xprt, task);
1663 * xprt_retry_reserve - allocate an RPC request slot
1664 * @task: RPC task requesting a slot allocation
1666 * If no more slots are available, place the task on the transport's
1668 * Note that the only difference with xprt_reserve is that we now
1669 * ignore the value of the XPRT_CONGESTED flag.
1671 void xprt_retry_reserve(struct rpc_task *task)
1673 struct rpc_xprt *xprt = task->tk_xprt;
1675 task->tk_status = 0;
1676 if (task->tk_rqstp != NULL)
1679 task->tk_timeout = 0;
1680 task->tk_status = -EAGAIN;
1681 xprt_do_reserve(xprt, task);
1685 xprt_request_dequeue_all(struct rpc_task *task, struct rpc_rqst *req)
1687 struct rpc_xprt *xprt = req->rq_xprt;
1689 if (test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate) ||
1690 test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) ||
1691 xprt_is_pinned_rqst(req)) {
1692 spin_lock(&xprt->queue_lock);
1693 xprt_request_dequeue_transmit_locked(task);
1694 xprt_request_dequeue_receive_locked(task);
1695 while (xprt_is_pinned_rqst(req)) {
1696 set_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
1697 spin_unlock(&xprt->queue_lock);
1698 xprt_wait_on_pinned_rqst(req);
1699 spin_lock(&xprt->queue_lock);
1700 clear_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
1702 spin_unlock(&xprt->queue_lock);
1707 * xprt_release - release an RPC request slot
1708 * @task: task which is finished with the slot
1711 void xprt_release(struct rpc_task *task)
1713 struct rpc_xprt *xprt;
1714 struct rpc_rqst *req = task->tk_rqstp;
1717 if (task->tk_client) {
1718 xprt = task->tk_xprt;
1719 xprt_release_write(xprt, task);
1724 xprt = req->rq_xprt;
1725 if (task->tk_ops->rpc_count_stats != NULL)
1726 task->tk_ops->rpc_count_stats(task, task->tk_calldata);
1727 else if (task->tk_client)
1728 rpc_count_iostats(task, task->tk_client->cl_metrics);
1729 xprt_request_dequeue_all(task, req);
1730 spin_lock_bh(&xprt->transport_lock);
1731 xprt->ops->release_xprt(xprt, task);
1732 if (xprt->ops->release_request)
1733 xprt->ops->release_request(task);
1734 xprt->last_used = jiffies;
1735 xprt_schedule_autodisconnect(xprt);
1736 spin_unlock_bh(&xprt->transport_lock);
1738 xprt->ops->buf_free(task);
1739 xprt_inject_disconnect(xprt);
1740 xdr_free_bvec(&req->rq_rcv_buf);
1741 if (req->rq_cred != NULL)
1742 put_rpccred(req->rq_cred);
1743 task->tk_rqstp = NULL;
1744 if (req->rq_release_snd_buf)
1745 req->rq_release_snd_buf(req);
1747 dprintk("RPC: %5u release request %p\n", task->tk_pid, req);
1748 if (likely(!bc_prealloc(req)))
1749 xprt->ops->free_slot(xprt, req);
1751 xprt_free_bc_request(req);
1754 #ifdef CONFIG_SUNRPC_BACKCHANNEL
1756 xprt_init_bc_request(struct rpc_rqst *req, struct rpc_task *task)
1758 struct xdr_buf *xbufp = &req->rq_snd_buf;
1760 task->tk_rqstp = req;
1761 req->rq_task = task;
1762 xprt_init_connect_cookie(req, req->rq_xprt);
1764 * Set up the xdr_buf length.
1765 * This also indicates that the buffer is XDR encoded already.
1767 xbufp->len = xbufp->head[0].iov_len + xbufp->page_len +
1768 xbufp->tail[0].iov_len;
1769 req->rq_bytes_sent = 0;
1773 static void xprt_init(struct rpc_xprt *xprt, struct net *net)
1775 kref_init(&xprt->kref);
1777 spin_lock_init(&xprt->transport_lock);
1778 spin_lock_init(&xprt->reserve_lock);
1779 spin_lock_init(&xprt->queue_lock);
1781 INIT_LIST_HEAD(&xprt->free);
1782 xprt->recv_queue = RB_ROOT;
1783 INIT_LIST_HEAD(&xprt->xmit_queue);
1784 #if defined(CONFIG_SUNRPC_BACKCHANNEL)
1785 spin_lock_init(&xprt->bc_pa_lock);
1786 INIT_LIST_HEAD(&xprt->bc_pa_list);
1787 #endif /* CONFIG_SUNRPC_BACKCHANNEL */
1788 INIT_LIST_HEAD(&xprt->xprt_switch);
1790 xprt->last_used = jiffies;
1791 xprt->cwnd = RPC_INITCWND;
1792 xprt->bind_index = 0;
1794 rpc_init_wait_queue(&xprt->binding, "xprt_binding");
1795 rpc_init_wait_queue(&xprt->pending, "xprt_pending");
1796 rpc_init_wait_queue(&xprt->sending, "xprt_sending");
1797 rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog");
1799 xprt_init_xid(xprt);
1801 xprt->xprt_net = get_net(net);
1805 * xprt_create_transport - create an RPC transport
1806 * @args: rpc transport creation arguments
1809 struct rpc_xprt *xprt_create_transport(struct xprt_create *args)
1811 struct rpc_xprt *xprt;
1812 struct xprt_class *t;
1814 spin_lock(&xprt_list_lock);
1815 list_for_each_entry(t, &xprt_list, list) {
1816 if (t->ident == args->ident) {
1817 spin_unlock(&xprt_list_lock);
1821 spin_unlock(&xprt_list_lock);
1822 dprintk("RPC: transport (%d) not supported\n", args->ident);
1823 return ERR_PTR(-EIO);
1826 xprt = t->setup(args);
1828 dprintk("RPC: xprt_create_transport: failed, %ld\n",
1832 if (args->flags & XPRT_CREATE_NO_IDLE_TIMEOUT)
1833 xprt->idle_timeout = 0;
1834 INIT_WORK(&xprt->task_cleanup, xprt_autoclose);
1835 if (xprt_has_timer(xprt))
1836 timer_setup(&xprt->timer, xprt_init_autodisconnect, 0);
1838 timer_setup(&xprt->timer, NULL, 0);
1840 if (strlen(args->servername) > RPC_MAXNETNAMELEN) {
1842 return ERR_PTR(-EINVAL);
1844 xprt->servername = kstrdup(args->servername, GFP_KERNEL);
1845 if (xprt->servername == NULL) {
1847 return ERR_PTR(-ENOMEM);
1850 rpc_xprt_debugfs_register(xprt);
1852 dprintk("RPC: created transport %p with %u slots\n", xprt,
1858 static void xprt_destroy_cb(struct work_struct *work)
1860 struct rpc_xprt *xprt =
1861 container_of(work, struct rpc_xprt, task_cleanup);
1863 rpc_xprt_debugfs_unregister(xprt);
1864 rpc_destroy_wait_queue(&xprt->binding);
1865 rpc_destroy_wait_queue(&xprt->pending);
1866 rpc_destroy_wait_queue(&xprt->sending);
1867 rpc_destroy_wait_queue(&xprt->backlog);
1868 kfree(xprt->servername);
1870 * Tear down transport state and free the rpc_xprt
1872 xprt->ops->destroy(xprt);
1876 * xprt_destroy - destroy an RPC transport, killing off all requests.
1877 * @xprt: transport to destroy
1880 static void xprt_destroy(struct rpc_xprt *xprt)
1882 dprintk("RPC: destroying transport %p\n", xprt);
1885 * Exclude transport connect/disconnect handlers and autoclose
1887 wait_on_bit_lock(&xprt->state, XPRT_LOCKED, TASK_UNINTERRUPTIBLE);
1889 del_timer_sync(&xprt->timer);
1892 * Destroy sockets etc from the system workqueue so they can
1893 * safely flush receive work running on rpciod.
1895 INIT_WORK(&xprt->task_cleanup, xprt_destroy_cb);
1896 schedule_work(&xprt->task_cleanup);
1899 static void xprt_destroy_kref(struct kref *kref)
1901 xprt_destroy(container_of(kref, struct rpc_xprt, kref));
1905 * xprt_get - return a reference to an RPC transport.
1906 * @xprt: pointer to the transport
1909 struct rpc_xprt *xprt_get(struct rpc_xprt *xprt)
1911 if (xprt != NULL && kref_get_unless_zero(&xprt->kref))
1915 EXPORT_SYMBOL_GPL(xprt_get);
1918 * xprt_put - release a reference to an RPC transport.
1919 * @xprt: pointer to the transport
1922 void xprt_put(struct rpc_xprt *xprt)
1925 kref_put(&xprt->kref, xprt_destroy_kref);
1927 EXPORT_SYMBOL_GPL(xprt_put);