xref: /openbmc/linux/net/sunrpc/xprt.c (revision 301306a9)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  *  linux/net/sunrpc/xprt.c
4  *
5  *  This is a generic RPC call interface supporting congestion avoidance,
6  *  and asynchronous calls.
7  *
8  *  The interface works like this:
9  *
10  *  -	When a process places a call, it allocates a request slot if
11  *	one is available. Otherwise, it sleeps on the backlog queue
12  *	(xprt_reserve).
13  *  -	Next, the caller puts together the RPC message, stuffs it into
14  *	the request struct, and calls xprt_transmit().
15  *  -	xprt_transmit sends the message and installs the caller on the
16  *	transport's wait list. At the same time, if a reply is expected,
17  *	it installs a timer that is run after the packet's timeout has
18  *	expired.
19  *  -	When a packet arrives, the data_ready handler walks the list of
20  *	pending requests for that transport. If a matching XID is found, the
21  *	caller is woken up, and the timer removed.
22  *  -	When no reply arrives within the timeout interval, the timer is
23  *	fired by the kernel and runs xprt_timer(). It either adjusts the
24  *	timeout values (minor timeout) or wakes up the caller with a status
25  *	of -ETIMEDOUT.
26  *  -	When the caller receives a notification from RPC that a reply arrived,
27  *	it should release the RPC slot, and process the reply.
28  *	If the call timed out, it may choose to retry the operation by
29  *	adjusting the initial timeout value, and simply calling rpc_call
30  *	again.
31  *
32  *  Support for async RPC is done through a set of RPC-specific scheduling
33  *  primitives that `transparently' work for processes as well as async
34  *  tasks that rely on callbacks.
35  *
36  *  Copyright (C) 1995-1997, Olaf Kirch <okir@monad.swb.de>
37  *
38  *  Transport switch API copyright (C) 2005, Chuck Lever <cel@netapp.com>
39  */
40 
41 #include <linux/module.h>
42 
43 #include <linux/types.h>
44 #include <linux/interrupt.h>
45 #include <linux/workqueue.h>
46 #include <linux/net.h>
47 #include <linux/ktime.h>
48 
49 #include <linux/sunrpc/clnt.h>
50 #include <linux/sunrpc/metrics.h>
51 #include <linux/sunrpc/bc_xprt.h>
52 #include <linux/rcupdate.h>
53 #include <linux/sched/mm.h>
54 
55 #include <trace/events/sunrpc.h>
56 
57 #include "sunrpc.h"
58 #include "sysfs.h"
59 #include "fail.h"
60 
61 /*
62  * Local variables
63  */
64 
65 #if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
66 # define RPCDBG_FACILITY	RPCDBG_XPRT
67 #endif
68 
69 /*
70  * Local functions
71  */
72 static void	 xprt_init(struct rpc_xprt *xprt, struct net *net);
73 static __be32	xprt_alloc_xid(struct rpc_xprt *xprt);
74 static void	 xprt_destroy(struct rpc_xprt *xprt);
75 static void	 xprt_request_init(struct rpc_task *task);
76 
77 static DEFINE_SPINLOCK(xprt_list_lock);
78 static LIST_HEAD(xprt_list);
79 
80 static unsigned long xprt_request_timeout(const struct rpc_rqst *req)
81 {
82 	unsigned long timeout = jiffies + req->rq_timeout;
83 
84 	if (time_before(timeout, req->rq_majortimeo))
85 		return timeout;
86 	return req->rq_majortimeo;
87 }
88 
89 /**
90  * xprt_register_transport - register a transport implementation
91  * @transport: transport to register
92  *
93  * If a transport implementation is loaded as a kernel module, it can
94  * call this interface to make itself known to the RPC client.
95  *
96  * Returns:
97  * 0:		transport successfully registered
98  * -EEXIST:	transport already registered
99  * -EINVAL:	transport module being unloaded
100  */
101 int xprt_register_transport(struct xprt_class *transport)
102 {
103 	struct xprt_class *t;
104 	int result;
105 
106 	result = -EEXIST;
107 	spin_lock(&xprt_list_lock);
108 	list_for_each_entry(t, &xprt_list, list) {
109 		/* don't register the same transport class twice */
110 		if (t->ident == transport->ident)
111 			goto out;
112 	}
113 
114 	list_add_tail(&transport->list, &xprt_list);
115 	printk(KERN_INFO "RPC: Registered %s transport module.\n",
116 	       transport->name);
117 	result = 0;
118 
119 out:
120 	spin_unlock(&xprt_list_lock);
121 	return result;
122 }
123 EXPORT_SYMBOL_GPL(xprt_register_transport);
124 
125 /**
126  * xprt_unregister_transport - unregister a transport implementation
127  * @transport: transport to unregister
128  *
129  * Returns:
130  * 0:		transport successfully unregistered
131  * -ENOENT:	transport never registered
132  */
133 int xprt_unregister_transport(struct xprt_class *transport)
134 {
135 	struct xprt_class *t;
136 	int result;
137 
138 	result = 0;
139 	spin_lock(&xprt_list_lock);
140 	list_for_each_entry(t, &xprt_list, list) {
141 		if (t == transport) {
142 			printk(KERN_INFO
143 				"RPC: Unregistered %s transport module.\n",
144 				transport->name);
145 			list_del_init(&transport->list);
146 			goto out;
147 		}
148 	}
149 	result = -ENOENT;
150 
151 out:
152 	spin_unlock(&xprt_list_lock);
153 	return result;
154 }
155 EXPORT_SYMBOL_GPL(xprt_unregister_transport);
156 
157 static void
158 xprt_class_release(const struct xprt_class *t)
159 {
160 	module_put(t->owner);
161 }
162 
163 static const struct xprt_class *
164 xprt_class_find_by_ident_locked(int ident)
165 {
166 	const struct xprt_class *t;
167 
168 	list_for_each_entry(t, &xprt_list, list) {
169 		if (t->ident != ident)
170 			continue;
171 		if (!try_module_get(t->owner))
172 			continue;
173 		return t;
174 	}
175 	return NULL;
176 }
177 
178 static const struct xprt_class *
179 xprt_class_find_by_ident(int ident)
180 {
181 	const struct xprt_class *t;
182 
183 	spin_lock(&xprt_list_lock);
184 	t = xprt_class_find_by_ident_locked(ident);
185 	spin_unlock(&xprt_list_lock);
186 	return t;
187 }
188 
189 static const struct xprt_class *
190 xprt_class_find_by_netid_locked(const char *netid)
191 {
192 	const struct xprt_class *t;
193 	unsigned int i;
194 
195 	list_for_each_entry(t, &xprt_list, list) {
196 		for (i = 0; t->netid[i][0] != '\0'; i++) {
197 			if (strcmp(t->netid[i], netid) != 0)
198 				continue;
199 			if (!try_module_get(t->owner))
200 				continue;
201 			return t;
202 		}
203 	}
204 	return NULL;
205 }
206 
207 static const struct xprt_class *
208 xprt_class_find_by_netid(const char *netid)
209 {
210 	const struct xprt_class *t;
211 
212 	spin_lock(&xprt_list_lock);
213 	t = xprt_class_find_by_netid_locked(netid);
214 	if (!t) {
215 		spin_unlock(&xprt_list_lock);
216 		request_module("rpc%s", netid);
217 		spin_lock(&xprt_list_lock);
218 		t = xprt_class_find_by_netid_locked(netid);
219 	}
220 	spin_unlock(&xprt_list_lock);
221 	return t;
222 }
223 
224 /**
225  * xprt_find_transport_ident - convert a netid into a transport identifier
226  * @netid: transport to load
227  *
228  * Returns:
229  * > 0:		transport identifier
230  * -ENOENT:	transport module not available
231  */
232 int xprt_find_transport_ident(const char *netid)
233 {
234 	const struct xprt_class *t;
235 	int ret;
236 
237 	t = xprt_class_find_by_netid(netid);
238 	if (!t)
239 		return -ENOENT;
240 	ret = t->ident;
241 	xprt_class_release(t);
242 	return ret;
243 }
244 EXPORT_SYMBOL_GPL(xprt_find_transport_ident);
245 
246 static void xprt_clear_locked(struct rpc_xprt *xprt)
247 {
248 	xprt->snd_task = NULL;
249 	if (!test_bit(XPRT_CLOSE_WAIT, &xprt->state))
250 		clear_bit_unlock(XPRT_LOCKED, &xprt->state);
251 	else
252 		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
253 }
254 
255 /**
256  * xprt_reserve_xprt - serialize write access to transports
257  * @task: task that is requesting access to the transport
258  * @xprt: pointer to the target transport
259  *
260  * This prevents mixing the payload of separate requests, and prevents
261  * transport connects from colliding with writes.  No congestion control
262  * is provided.
263  */
264 int xprt_reserve_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
265 {
266 	struct rpc_rqst *req = task->tk_rqstp;
267 
268 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
269 		if (task == xprt->snd_task)
270 			goto out_locked;
271 		goto out_sleep;
272 	}
273 	if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
274 		goto out_unlock;
275 	xprt->snd_task = task;
276 
277 out_locked:
278 	trace_xprt_reserve_xprt(xprt, task);
279 	return 1;
280 
281 out_unlock:
282 	xprt_clear_locked(xprt);
283 out_sleep:
284 	task->tk_status = -EAGAIN;
285 	if  (RPC_IS_SOFT(task))
286 		rpc_sleep_on_timeout(&xprt->sending, task, NULL,
287 				xprt_request_timeout(req));
288 	else
289 		rpc_sleep_on(&xprt->sending, task, NULL);
290 	return 0;
291 }
292 EXPORT_SYMBOL_GPL(xprt_reserve_xprt);
293 
294 static bool
295 xprt_need_congestion_window_wait(struct rpc_xprt *xprt)
296 {
297 	return test_bit(XPRT_CWND_WAIT, &xprt->state);
298 }
299 
300 static void
301 xprt_set_congestion_window_wait(struct rpc_xprt *xprt)
302 {
303 	if (!list_empty(&xprt->xmit_queue)) {
304 		/* Peek at head of queue to see if it can make progress */
305 		if (list_first_entry(&xprt->xmit_queue, struct rpc_rqst,
306 					rq_xmit)->rq_cong)
307 			return;
308 	}
309 	set_bit(XPRT_CWND_WAIT, &xprt->state);
310 }
311 
312 static void
313 xprt_test_and_clear_congestion_window_wait(struct rpc_xprt *xprt)
314 {
315 	if (!RPCXPRT_CONGESTED(xprt))
316 		clear_bit(XPRT_CWND_WAIT, &xprt->state);
317 }
318 
319 /*
320  * xprt_reserve_xprt_cong - serialize write access to transports
321  * @task: task that is requesting access to the transport
322  *
323  * Same as xprt_reserve_xprt, but Van Jacobson congestion control is
324  * integrated into the decision of whether a request is allowed to be
325  * woken up and given access to the transport.
326  * Note that the lock is only granted if we know there are free slots.
327  */
328 int xprt_reserve_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task)
329 {
330 	struct rpc_rqst *req = task->tk_rqstp;
331 
332 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
333 		if (task == xprt->snd_task)
334 			goto out_locked;
335 		goto out_sleep;
336 	}
337 	if (req == NULL) {
338 		xprt->snd_task = task;
339 		goto out_locked;
340 	}
341 	if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
342 		goto out_unlock;
343 	if (!xprt_need_congestion_window_wait(xprt)) {
344 		xprt->snd_task = task;
345 		goto out_locked;
346 	}
347 out_unlock:
348 	xprt_clear_locked(xprt);
349 out_sleep:
350 	task->tk_status = -EAGAIN;
351 	if (RPC_IS_SOFT(task))
352 		rpc_sleep_on_timeout(&xprt->sending, task, NULL,
353 				xprt_request_timeout(req));
354 	else
355 		rpc_sleep_on(&xprt->sending, task, NULL);
356 	return 0;
357 out_locked:
358 	trace_xprt_reserve_cong(xprt, task);
359 	return 1;
360 }
361 EXPORT_SYMBOL_GPL(xprt_reserve_xprt_cong);
362 
363 static inline int xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
364 {
365 	int retval;
366 
367 	if (test_bit(XPRT_LOCKED, &xprt->state) && xprt->snd_task == task)
368 		return 1;
369 	spin_lock(&xprt->transport_lock);
370 	retval = xprt->ops->reserve_xprt(xprt, task);
371 	spin_unlock(&xprt->transport_lock);
372 	return retval;
373 }
374 
375 static bool __xprt_lock_write_func(struct rpc_task *task, void *data)
376 {
377 	struct rpc_xprt *xprt = data;
378 
379 	xprt->snd_task = task;
380 	return true;
381 }
382 
383 static void __xprt_lock_write_next(struct rpc_xprt *xprt)
384 {
385 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
386 		return;
387 	if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
388 		goto out_unlock;
389 	if (rpc_wake_up_first_on_wq(xprtiod_workqueue, &xprt->sending,
390 				__xprt_lock_write_func, xprt))
391 		return;
392 out_unlock:
393 	xprt_clear_locked(xprt);
394 }
395 
396 static void __xprt_lock_write_next_cong(struct rpc_xprt *xprt)
397 {
398 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
399 		return;
400 	if (test_bit(XPRT_WRITE_SPACE, &xprt->state))
401 		goto out_unlock;
402 	if (xprt_need_congestion_window_wait(xprt))
403 		goto out_unlock;
404 	if (rpc_wake_up_first_on_wq(xprtiod_workqueue, &xprt->sending,
405 				__xprt_lock_write_func, xprt))
406 		return;
407 out_unlock:
408 	xprt_clear_locked(xprt);
409 }
410 
411 /**
412  * xprt_release_xprt - allow other requests to use a transport
413  * @xprt: transport with other tasks potentially waiting
414  * @task: task that is releasing access to the transport
415  *
416  * Note that "task" can be NULL.  No congestion control is provided.
417  */
418 void xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
419 {
420 	if (xprt->snd_task == task) {
421 		xprt_clear_locked(xprt);
422 		__xprt_lock_write_next(xprt);
423 	}
424 	trace_xprt_release_xprt(xprt, task);
425 }
426 EXPORT_SYMBOL_GPL(xprt_release_xprt);
427 
428 /**
429  * xprt_release_xprt_cong - allow other requests to use a transport
430  * @xprt: transport with other tasks potentially waiting
431  * @task: task that is releasing access to the transport
432  *
433  * Note that "task" can be NULL.  Another task is awoken to use the
434  * transport if the transport's congestion window allows it.
435  */
436 void xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task)
437 {
438 	if (xprt->snd_task == task) {
439 		xprt_clear_locked(xprt);
440 		__xprt_lock_write_next_cong(xprt);
441 	}
442 	trace_xprt_release_cong(xprt, task);
443 }
444 EXPORT_SYMBOL_GPL(xprt_release_xprt_cong);
445 
446 void xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
447 {
448 	if (xprt->snd_task != task)
449 		return;
450 	spin_lock(&xprt->transport_lock);
451 	xprt->ops->release_xprt(xprt, task);
452 	spin_unlock(&xprt->transport_lock);
453 }
454 
455 /*
456  * Van Jacobson congestion avoidance. Check if the congestion window
457  * overflowed. Put the task to sleep if this is the case.
458  */
459 static int
460 __xprt_get_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
461 {
462 	if (req->rq_cong)
463 		return 1;
464 	trace_xprt_get_cong(xprt, req->rq_task);
465 	if (RPCXPRT_CONGESTED(xprt)) {
466 		xprt_set_congestion_window_wait(xprt);
467 		return 0;
468 	}
469 	req->rq_cong = 1;
470 	xprt->cong += RPC_CWNDSCALE;
471 	return 1;
472 }
473 
474 /*
475  * Adjust the congestion window, and wake up the next task
476  * that has been sleeping due to congestion
477  */
478 static void
479 __xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
480 {
481 	if (!req->rq_cong)
482 		return;
483 	req->rq_cong = 0;
484 	xprt->cong -= RPC_CWNDSCALE;
485 	xprt_test_and_clear_congestion_window_wait(xprt);
486 	trace_xprt_put_cong(xprt, req->rq_task);
487 	__xprt_lock_write_next_cong(xprt);
488 }
489 
490 /**
491  * xprt_request_get_cong - Request congestion control credits
492  * @xprt: pointer to transport
493  * @req: pointer to RPC request
494  *
495  * Useful for transports that require congestion control.
496  */
497 bool
498 xprt_request_get_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
499 {
500 	bool ret = false;
501 
502 	if (req->rq_cong)
503 		return true;
504 	spin_lock(&xprt->transport_lock);
505 	ret = __xprt_get_cong(xprt, req) != 0;
506 	spin_unlock(&xprt->transport_lock);
507 	return ret;
508 }
509 EXPORT_SYMBOL_GPL(xprt_request_get_cong);
510 
511 /**
512  * xprt_release_rqst_cong - housekeeping when request is complete
513  * @task: RPC request that recently completed
514  *
515  * Useful for transports that require congestion control.
516  */
517 void xprt_release_rqst_cong(struct rpc_task *task)
518 {
519 	struct rpc_rqst *req = task->tk_rqstp;
520 
521 	__xprt_put_cong(req->rq_xprt, req);
522 }
523 EXPORT_SYMBOL_GPL(xprt_release_rqst_cong);
524 
525 static void xprt_clear_congestion_window_wait_locked(struct rpc_xprt *xprt)
526 {
527 	if (test_and_clear_bit(XPRT_CWND_WAIT, &xprt->state))
528 		__xprt_lock_write_next_cong(xprt);
529 }
530 
531 /*
532  * Clear the congestion window wait flag and wake up the next
533  * entry on xprt->sending
534  */
535 static void
536 xprt_clear_congestion_window_wait(struct rpc_xprt *xprt)
537 {
538 	if (test_and_clear_bit(XPRT_CWND_WAIT, &xprt->state)) {
539 		spin_lock(&xprt->transport_lock);
540 		__xprt_lock_write_next_cong(xprt);
541 		spin_unlock(&xprt->transport_lock);
542 	}
543 }
544 
545 /**
546  * xprt_adjust_cwnd - adjust transport congestion window
547  * @xprt: pointer to xprt
548  * @task: recently completed RPC request used to adjust window
549  * @result: result code of completed RPC request
550  *
551  * The transport code maintains an estimate on the maximum number of out-
552  * standing RPC requests, using a smoothed version of the congestion
553  * avoidance implemented in 44BSD. This is basically the Van Jacobson
554  * congestion algorithm: If a retransmit occurs, the congestion window is
555  * halved; otherwise, it is incremented by 1/cwnd when
556  *
557  *	-	a reply is received and
558  *	-	a full number of requests are outstanding and
559  *	-	the congestion window hasn't been updated recently.
560  */
561 void xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result)
562 {
563 	struct rpc_rqst *req = task->tk_rqstp;
564 	unsigned long cwnd = xprt->cwnd;
565 
566 	if (result >= 0 && cwnd <= xprt->cong) {
567 		/* The (cwnd >> 1) term makes sure
568 		 * the result gets rounded properly. */
569 		cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd;
570 		if (cwnd > RPC_MAXCWND(xprt))
571 			cwnd = RPC_MAXCWND(xprt);
572 		__xprt_lock_write_next_cong(xprt);
573 	} else if (result == -ETIMEDOUT) {
574 		cwnd >>= 1;
575 		if (cwnd < RPC_CWNDSCALE)
576 			cwnd = RPC_CWNDSCALE;
577 	}
578 	dprintk("RPC:       cong %ld, cwnd was %ld, now %ld\n",
579 			xprt->cong, xprt->cwnd, cwnd);
580 	xprt->cwnd = cwnd;
581 	__xprt_put_cong(xprt, req);
582 }
583 EXPORT_SYMBOL_GPL(xprt_adjust_cwnd);
584 
585 /**
586  * xprt_wake_pending_tasks - wake all tasks on a transport's pending queue
587  * @xprt: transport with waiting tasks
588  * @status: result code to plant in each task before waking it
589  *
590  */
591 void xprt_wake_pending_tasks(struct rpc_xprt *xprt, int status)
592 {
593 	if (status < 0)
594 		rpc_wake_up_status(&xprt->pending, status);
595 	else
596 		rpc_wake_up(&xprt->pending);
597 }
598 EXPORT_SYMBOL_GPL(xprt_wake_pending_tasks);
599 
600 /**
601  * xprt_wait_for_buffer_space - wait for transport output buffer to clear
602  * @xprt: transport
603  *
604  * Note that we only set the timer for the case of RPC_IS_SOFT(), since
605  * we don't in general want to force a socket disconnection due to
606  * an incomplete RPC call transmission.
607  */
608 void xprt_wait_for_buffer_space(struct rpc_xprt *xprt)
609 {
610 	set_bit(XPRT_WRITE_SPACE, &xprt->state);
611 }
612 EXPORT_SYMBOL_GPL(xprt_wait_for_buffer_space);
613 
614 static bool
615 xprt_clear_write_space_locked(struct rpc_xprt *xprt)
616 {
617 	if (test_and_clear_bit(XPRT_WRITE_SPACE, &xprt->state)) {
618 		__xprt_lock_write_next(xprt);
619 		dprintk("RPC:       write space: waking waiting task on "
620 				"xprt %p\n", xprt);
621 		return true;
622 	}
623 	return false;
624 }
625 
626 /**
627  * xprt_write_space - wake the task waiting for transport output buffer space
628  * @xprt: transport with waiting tasks
629  *
630  * Can be called in a soft IRQ context, so xprt_write_space never sleeps.
631  */
632 bool xprt_write_space(struct rpc_xprt *xprt)
633 {
634 	bool ret;
635 
636 	if (!test_bit(XPRT_WRITE_SPACE, &xprt->state))
637 		return false;
638 	spin_lock(&xprt->transport_lock);
639 	ret = xprt_clear_write_space_locked(xprt);
640 	spin_unlock(&xprt->transport_lock);
641 	return ret;
642 }
643 EXPORT_SYMBOL_GPL(xprt_write_space);
644 
645 static unsigned long xprt_abs_ktime_to_jiffies(ktime_t abstime)
646 {
647 	s64 delta = ktime_to_ns(ktime_get() - abstime);
648 	return likely(delta >= 0) ?
649 		jiffies - nsecs_to_jiffies(delta) :
650 		jiffies + nsecs_to_jiffies(-delta);
651 }
652 
653 static unsigned long xprt_calc_majortimeo(struct rpc_rqst *req)
654 {
655 	const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout;
656 	unsigned long majortimeo = req->rq_timeout;
657 
658 	if (to->to_exponential)
659 		majortimeo <<= to->to_retries;
660 	else
661 		majortimeo += to->to_increment * to->to_retries;
662 	if (majortimeo > to->to_maxval || majortimeo == 0)
663 		majortimeo = to->to_maxval;
664 	return majortimeo;
665 }
666 
667 static void xprt_reset_majortimeo(struct rpc_rqst *req)
668 {
669 	req->rq_majortimeo += xprt_calc_majortimeo(req);
670 }
671 
672 static void xprt_reset_minortimeo(struct rpc_rqst *req)
673 {
674 	req->rq_minortimeo += req->rq_timeout;
675 }
676 
677 static void xprt_init_majortimeo(struct rpc_task *task, struct rpc_rqst *req)
678 {
679 	unsigned long time_init;
680 	struct rpc_xprt *xprt = req->rq_xprt;
681 
682 	if (likely(xprt && xprt_connected(xprt)))
683 		time_init = jiffies;
684 	else
685 		time_init = xprt_abs_ktime_to_jiffies(task->tk_start);
686 	req->rq_timeout = task->tk_client->cl_timeout->to_initval;
687 	req->rq_majortimeo = time_init + xprt_calc_majortimeo(req);
688 	req->rq_minortimeo = time_init + req->rq_timeout;
689 }
690 
691 /**
692  * xprt_adjust_timeout - adjust timeout values for next retransmit
693  * @req: RPC request containing parameters to use for the adjustment
694  *
695  */
696 int xprt_adjust_timeout(struct rpc_rqst *req)
697 {
698 	struct rpc_xprt *xprt = req->rq_xprt;
699 	const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout;
700 	int status = 0;
701 
702 	if (time_before(jiffies, req->rq_majortimeo)) {
703 		if (time_before(jiffies, req->rq_minortimeo))
704 			return status;
705 		if (to->to_exponential)
706 			req->rq_timeout <<= 1;
707 		else
708 			req->rq_timeout += to->to_increment;
709 		if (to->to_maxval && req->rq_timeout >= to->to_maxval)
710 			req->rq_timeout = to->to_maxval;
711 		req->rq_retries++;
712 	} else {
713 		req->rq_timeout = to->to_initval;
714 		req->rq_retries = 0;
715 		xprt_reset_majortimeo(req);
716 		/* Reset the RTT counters == "slow start" */
717 		spin_lock(&xprt->transport_lock);
718 		rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
719 		spin_unlock(&xprt->transport_lock);
720 		status = -ETIMEDOUT;
721 	}
722 	xprt_reset_minortimeo(req);
723 
724 	if (req->rq_timeout == 0) {
725 		printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n");
726 		req->rq_timeout = 5 * HZ;
727 	}
728 	return status;
729 }
730 
731 static void xprt_autoclose(struct work_struct *work)
732 {
733 	struct rpc_xprt *xprt =
734 		container_of(work, struct rpc_xprt, task_cleanup);
735 	unsigned int pflags = memalloc_nofs_save();
736 
737 	trace_xprt_disconnect_auto(xprt);
738 	xprt->connect_cookie++;
739 	smp_mb__before_atomic();
740 	clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
741 	xprt->ops->close(xprt);
742 	xprt_release_write(xprt, NULL);
743 	wake_up_bit(&xprt->state, XPRT_LOCKED);
744 	memalloc_nofs_restore(pflags);
745 }
746 
747 /**
748  * xprt_disconnect_done - mark a transport as disconnected
749  * @xprt: transport to flag for disconnect
750  *
751  */
752 void xprt_disconnect_done(struct rpc_xprt *xprt)
753 {
754 	trace_xprt_disconnect_done(xprt);
755 	spin_lock(&xprt->transport_lock);
756 	xprt_clear_connected(xprt);
757 	xprt_clear_write_space_locked(xprt);
758 	xprt_clear_congestion_window_wait_locked(xprt);
759 	xprt_wake_pending_tasks(xprt, -ENOTCONN);
760 	spin_unlock(&xprt->transport_lock);
761 }
762 EXPORT_SYMBOL_GPL(xprt_disconnect_done);
763 
764 /**
765  * xprt_schedule_autoclose_locked - Try to schedule an autoclose RPC call
766  * @xprt: transport to disconnect
767  */
768 static void xprt_schedule_autoclose_locked(struct rpc_xprt *xprt)
769 {
770 	if (test_and_set_bit(XPRT_CLOSE_WAIT, &xprt->state))
771 		return;
772 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
773 		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
774 	else if (xprt->snd_task && !test_bit(XPRT_SND_IS_COOKIE, &xprt->state))
775 		rpc_wake_up_queued_task_set_status(&xprt->pending,
776 						   xprt->snd_task, -ENOTCONN);
777 }
778 
779 /**
780  * xprt_force_disconnect - force a transport to disconnect
781  * @xprt: transport to disconnect
782  *
783  */
784 void xprt_force_disconnect(struct rpc_xprt *xprt)
785 {
786 	trace_xprt_disconnect_force(xprt);
787 
788 	/* Don't race with the test_bit() in xprt_clear_locked() */
789 	spin_lock(&xprt->transport_lock);
790 	xprt_schedule_autoclose_locked(xprt);
791 	spin_unlock(&xprt->transport_lock);
792 }
793 EXPORT_SYMBOL_GPL(xprt_force_disconnect);
794 
795 static unsigned int
796 xprt_connect_cookie(struct rpc_xprt *xprt)
797 {
798 	return READ_ONCE(xprt->connect_cookie);
799 }
800 
801 static bool
802 xprt_request_retransmit_after_disconnect(struct rpc_task *task)
803 {
804 	struct rpc_rqst *req = task->tk_rqstp;
805 	struct rpc_xprt *xprt = req->rq_xprt;
806 
807 	return req->rq_connect_cookie != xprt_connect_cookie(xprt) ||
808 		!xprt_connected(xprt);
809 }
810 
811 /**
812  * xprt_conditional_disconnect - force a transport to disconnect
813  * @xprt: transport to disconnect
814  * @cookie: 'connection cookie'
815  *
816  * This attempts to break the connection if and only if 'cookie' matches
817  * the current transport 'connection cookie'. It ensures that we don't
818  * try to break the connection more than once when we need to retransmit
819  * a batch of RPC requests.
820  *
821  */
822 void xprt_conditional_disconnect(struct rpc_xprt *xprt, unsigned int cookie)
823 {
824 	/* Don't race with the test_bit() in xprt_clear_locked() */
825 	spin_lock(&xprt->transport_lock);
826 	if (cookie != xprt->connect_cookie)
827 		goto out;
828 	if (test_bit(XPRT_CLOSING, &xprt->state))
829 		goto out;
830 	xprt_schedule_autoclose_locked(xprt);
831 out:
832 	spin_unlock(&xprt->transport_lock);
833 }
834 
835 static bool
836 xprt_has_timer(const struct rpc_xprt *xprt)
837 {
838 	return xprt->idle_timeout != 0;
839 }
840 
841 static void
842 xprt_schedule_autodisconnect(struct rpc_xprt *xprt)
843 	__must_hold(&xprt->transport_lock)
844 {
845 	xprt->last_used = jiffies;
846 	if (RB_EMPTY_ROOT(&xprt->recv_queue) && xprt_has_timer(xprt))
847 		mod_timer(&xprt->timer, xprt->last_used + xprt->idle_timeout);
848 }
849 
850 static void
851 xprt_init_autodisconnect(struct timer_list *t)
852 {
853 	struct rpc_xprt *xprt = from_timer(xprt, t, timer);
854 
855 	if (!RB_EMPTY_ROOT(&xprt->recv_queue))
856 		return;
857 	/* Reset xprt->last_used to avoid connect/autodisconnect cycling */
858 	xprt->last_used = jiffies;
859 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
860 		return;
861 	queue_work(xprtiod_workqueue, &xprt->task_cleanup);
862 }
863 
864 #if IS_ENABLED(CONFIG_FAIL_SUNRPC)
865 static void xprt_inject_disconnect(struct rpc_xprt *xprt)
866 {
867 	if (!fail_sunrpc.ignore_client_disconnect &&
868 	    should_fail(&fail_sunrpc.attr, 1))
869 		xprt->ops->inject_disconnect(xprt);
870 }
871 #else
872 static inline void xprt_inject_disconnect(struct rpc_xprt *xprt)
873 {
874 }
875 #endif
876 
877 bool xprt_lock_connect(struct rpc_xprt *xprt,
878 		struct rpc_task *task,
879 		void *cookie)
880 {
881 	bool ret = false;
882 
883 	spin_lock(&xprt->transport_lock);
884 	if (!test_bit(XPRT_LOCKED, &xprt->state))
885 		goto out;
886 	if (xprt->snd_task != task)
887 		goto out;
888 	set_bit(XPRT_SND_IS_COOKIE, &xprt->state);
889 	xprt->snd_task = cookie;
890 	ret = true;
891 out:
892 	spin_unlock(&xprt->transport_lock);
893 	return ret;
894 }
895 EXPORT_SYMBOL_GPL(xprt_lock_connect);
896 
897 void xprt_unlock_connect(struct rpc_xprt *xprt, void *cookie)
898 {
899 	spin_lock(&xprt->transport_lock);
900 	if (xprt->snd_task != cookie)
901 		goto out;
902 	if (!test_bit(XPRT_LOCKED, &xprt->state))
903 		goto out;
904 	xprt->snd_task =NULL;
905 	clear_bit(XPRT_SND_IS_COOKIE, &xprt->state);
906 	xprt->ops->release_xprt(xprt, NULL);
907 	xprt_schedule_autodisconnect(xprt);
908 out:
909 	spin_unlock(&xprt->transport_lock);
910 	wake_up_bit(&xprt->state, XPRT_LOCKED);
911 }
912 EXPORT_SYMBOL_GPL(xprt_unlock_connect);
913 
914 /**
915  * xprt_connect - schedule a transport connect operation
916  * @task: RPC task that is requesting the connect
917  *
918  */
919 void xprt_connect(struct rpc_task *task)
920 {
921 	struct rpc_xprt	*xprt = task->tk_rqstp->rq_xprt;
922 
923 	trace_xprt_connect(xprt);
924 
925 	if (!xprt_bound(xprt)) {
926 		task->tk_status = -EAGAIN;
927 		return;
928 	}
929 	if (!xprt_lock_write(xprt, task))
930 		return;
931 
932 	if (test_and_clear_bit(XPRT_CLOSE_WAIT, &xprt->state)) {
933 		trace_xprt_disconnect_cleanup(xprt);
934 		xprt->ops->close(xprt);
935 	}
936 
937 	if (!xprt_connected(xprt)) {
938 		task->tk_rqstp->rq_connect_cookie = xprt->connect_cookie;
939 		rpc_sleep_on_timeout(&xprt->pending, task, NULL,
940 				xprt_request_timeout(task->tk_rqstp));
941 
942 		if (test_bit(XPRT_CLOSING, &xprt->state))
943 			return;
944 		if (xprt_test_and_set_connecting(xprt))
945 			return;
946 		/* Race breaker */
947 		if (!xprt_connected(xprt)) {
948 			xprt->stat.connect_start = jiffies;
949 			xprt->ops->connect(xprt, task);
950 		} else {
951 			xprt_clear_connecting(xprt);
952 			task->tk_status = 0;
953 			rpc_wake_up_queued_task(&xprt->pending, task);
954 		}
955 	}
956 	xprt_release_write(xprt, task);
957 }
958 
959 /**
960  * xprt_reconnect_delay - compute the wait before scheduling a connect
961  * @xprt: transport instance
962  *
963  */
964 unsigned long xprt_reconnect_delay(const struct rpc_xprt *xprt)
965 {
966 	unsigned long start, now = jiffies;
967 
968 	start = xprt->stat.connect_start + xprt->reestablish_timeout;
969 	if (time_after(start, now))
970 		return start - now;
971 	return 0;
972 }
973 EXPORT_SYMBOL_GPL(xprt_reconnect_delay);
974 
975 /**
976  * xprt_reconnect_backoff - compute the new re-establish timeout
977  * @xprt: transport instance
978  * @init_to: initial reestablish timeout
979  *
980  */
981 void xprt_reconnect_backoff(struct rpc_xprt *xprt, unsigned long init_to)
982 {
983 	xprt->reestablish_timeout <<= 1;
984 	if (xprt->reestablish_timeout > xprt->max_reconnect_timeout)
985 		xprt->reestablish_timeout = xprt->max_reconnect_timeout;
986 	if (xprt->reestablish_timeout < init_to)
987 		xprt->reestablish_timeout = init_to;
988 }
989 EXPORT_SYMBOL_GPL(xprt_reconnect_backoff);
990 
991 enum xprt_xid_rb_cmp {
992 	XID_RB_EQUAL,
993 	XID_RB_LEFT,
994 	XID_RB_RIGHT,
995 };
996 static enum xprt_xid_rb_cmp
997 xprt_xid_cmp(__be32 xid1, __be32 xid2)
998 {
999 	if (xid1 == xid2)
1000 		return XID_RB_EQUAL;
1001 	if ((__force u32)xid1 < (__force u32)xid2)
1002 		return XID_RB_LEFT;
1003 	return XID_RB_RIGHT;
1004 }
1005 
1006 static struct rpc_rqst *
1007 xprt_request_rb_find(struct rpc_xprt *xprt, __be32 xid)
1008 {
1009 	struct rb_node *n = xprt->recv_queue.rb_node;
1010 	struct rpc_rqst *req;
1011 
1012 	while (n != NULL) {
1013 		req = rb_entry(n, struct rpc_rqst, rq_recv);
1014 		switch (xprt_xid_cmp(xid, req->rq_xid)) {
1015 		case XID_RB_LEFT:
1016 			n = n->rb_left;
1017 			break;
1018 		case XID_RB_RIGHT:
1019 			n = n->rb_right;
1020 			break;
1021 		case XID_RB_EQUAL:
1022 			return req;
1023 		}
1024 	}
1025 	return NULL;
1026 }
1027 
1028 static void
1029 xprt_request_rb_insert(struct rpc_xprt *xprt, struct rpc_rqst *new)
1030 {
1031 	struct rb_node **p = &xprt->recv_queue.rb_node;
1032 	struct rb_node *n = NULL;
1033 	struct rpc_rqst *req;
1034 
1035 	while (*p != NULL) {
1036 		n = *p;
1037 		req = rb_entry(n, struct rpc_rqst, rq_recv);
1038 		switch(xprt_xid_cmp(new->rq_xid, req->rq_xid)) {
1039 		case XID_RB_LEFT:
1040 			p = &n->rb_left;
1041 			break;
1042 		case XID_RB_RIGHT:
1043 			p = &n->rb_right;
1044 			break;
1045 		case XID_RB_EQUAL:
1046 			WARN_ON_ONCE(new != req);
1047 			return;
1048 		}
1049 	}
1050 	rb_link_node(&new->rq_recv, n, p);
1051 	rb_insert_color(&new->rq_recv, &xprt->recv_queue);
1052 }
1053 
1054 static void
1055 xprt_request_rb_remove(struct rpc_xprt *xprt, struct rpc_rqst *req)
1056 {
1057 	rb_erase(&req->rq_recv, &xprt->recv_queue);
1058 }
1059 
1060 /**
1061  * xprt_lookup_rqst - find an RPC request corresponding to an XID
1062  * @xprt: transport on which the original request was transmitted
1063  * @xid: RPC XID of incoming reply
1064  *
1065  * Caller holds xprt->queue_lock.
1066  */
1067 struct rpc_rqst *xprt_lookup_rqst(struct rpc_xprt *xprt, __be32 xid)
1068 {
1069 	struct rpc_rqst *entry;
1070 
1071 	entry = xprt_request_rb_find(xprt, xid);
1072 	if (entry != NULL) {
1073 		trace_xprt_lookup_rqst(xprt, xid, 0);
1074 		entry->rq_rtt = ktime_sub(ktime_get(), entry->rq_xtime);
1075 		return entry;
1076 	}
1077 
1078 	dprintk("RPC:       xprt_lookup_rqst did not find xid %08x\n",
1079 			ntohl(xid));
1080 	trace_xprt_lookup_rqst(xprt, xid, -ENOENT);
1081 	xprt->stat.bad_xids++;
1082 	return NULL;
1083 }
1084 EXPORT_SYMBOL_GPL(xprt_lookup_rqst);
1085 
1086 static bool
1087 xprt_is_pinned_rqst(struct rpc_rqst *req)
1088 {
1089 	return atomic_read(&req->rq_pin) != 0;
1090 }
1091 
1092 /**
1093  * xprt_pin_rqst - Pin a request on the transport receive list
1094  * @req: Request to pin
1095  *
1096  * Caller must ensure this is atomic with the call to xprt_lookup_rqst()
1097  * so should be holding xprt->queue_lock.
1098  */
1099 void xprt_pin_rqst(struct rpc_rqst *req)
1100 {
1101 	atomic_inc(&req->rq_pin);
1102 }
1103 EXPORT_SYMBOL_GPL(xprt_pin_rqst);
1104 
1105 /**
1106  * xprt_unpin_rqst - Unpin a request on the transport receive list
1107  * @req: Request to pin
1108  *
1109  * Caller should be holding xprt->queue_lock.
1110  */
1111 void xprt_unpin_rqst(struct rpc_rqst *req)
1112 {
1113 	if (!test_bit(RPC_TASK_MSG_PIN_WAIT, &req->rq_task->tk_runstate)) {
1114 		atomic_dec(&req->rq_pin);
1115 		return;
1116 	}
1117 	if (atomic_dec_and_test(&req->rq_pin))
1118 		wake_up_var(&req->rq_pin);
1119 }
1120 EXPORT_SYMBOL_GPL(xprt_unpin_rqst);
1121 
1122 static void xprt_wait_on_pinned_rqst(struct rpc_rqst *req)
1123 {
1124 	wait_var_event(&req->rq_pin, !xprt_is_pinned_rqst(req));
1125 }
1126 
1127 static bool
1128 xprt_request_data_received(struct rpc_task *task)
1129 {
1130 	return !test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) &&
1131 		READ_ONCE(task->tk_rqstp->rq_reply_bytes_recvd) != 0;
1132 }
1133 
1134 static bool
1135 xprt_request_need_enqueue_receive(struct rpc_task *task, struct rpc_rqst *req)
1136 {
1137 	return !test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) &&
1138 		READ_ONCE(task->tk_rqstp->rq_reply_bytes_recvd) == 0;
1139 }
1140 
1141 /**
1142  * xprt_request_enqueue_receive - Add an request to the receive queue
1143  * @task: RPC task
1144  *
1145  */
1146 void
1147 xprt_request_enqueue_receive(struct rpc_task *task)
1148 {
1149 	struct rpc_rqst *req = task->tk_rqstp;
1150 	struct rpc_xprt *xprt = req->rq_xprt;
1151 
1152 	if (!xprt_request_need_enqueue_receive(task, req))
1153 		return;
1154 
1155 	xprt_request_prepare(task->tk_rqstp);
1156 	spin_lock(&xprt->queue_lock);
1157 
1158 	/* Update the softirq receive buffer */
1159 	memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
1160 			sizeof(req->rq_private_buf));
1161 
1162 	/* Add request to the receive list */
1163 	xprt_request_rb_insert(xprt, req);
1164 	set_bit(RPC_TASK_NEED_RECV, &task->tk_runstate);
1165 	spin_unlock(&xprt->queue_lock);
1166 
1167 	/* Turn off autodisconnect */
1168 	del_singleshot_timer_sync(&xprt->timer);
1169 }
1170 
1171 /**
1172  * xprt_request_dequeue_receive_locked - Remove a request from the receive queue
1173  * @task: RPC task
1174  *
1175  * Caller must hold xprt->queue_lock.
1176  */
1177 static void
1178 xprt_request_dequeue_receive_locked(struct rpc_task *task)
1179 {
1180 	struct rpc_rqst *req = task->tk_rqstp;
1181 
1182 	if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
1183 		xprt_request_rb_remove(req->rq_xprt, req);
1184 }
1185 
1186 /**
1187  * xprt_update_rtt - Update RPC RTT statistics
1188  * @task: RPC request that recently completed
1189  *
1190  * Caller holds xprt->queue_lock.
1191  */
1192 void xprt_update_rtt(struct rpc_task *task)
1193 {
1194 	struct rpc_rqst *req = task->tk_rqstp;
1195 	struct rpc_rtt *rtt = task->tk_client->cl_rtt;
1196 	unsigned int timer = task->tk_msg.rpc_proc->p_timer;
1197 	long m = usecs_to_jiffies(ktime_to_us(req->rq_rtt));
1198 
1199 	if (timer) {
1200 		if (req->rq_ntrans == 1)
1201 			rpc_update_rtt(rtt, timer, m);
1202 		rpc_set_timeo(rtt, timer, req->rq_ntrans - 1);
1203 	}
1204 }
1205 EXPORT_SYMBOL_GPL(xprt_update_rtt);
1206 
1207 /**
1208  * xprt_complete_rqst - called when reply processing is complete
1209  * @task: RPC request that recently completed
1210  * @copied: actual number of bytes received from the transport
1211  *
1212  * Caller holds xprt->queue_lock.
1213  */
1214 void xprt_complete_rqst(struct rpc_task *task, int copied)
1215 {
1216 	struct rpc_rqst *req = task->tk_rqstp;
1217 	struct rpc_xprt *xprt = req->rq_xprt;
1218 
1219 	xprt->stat.recvs++;
1220 
1221 	req->rq_private_buf.len = copied;
1222 	/* Ensure all writes are done before we update */
1223 	/* req->rq_reply_bytes_recvd */
1224 	smp_wmb();
1225 	req->rq_reply_bytes_recvd = copied;
1226 	xprt_request_dequeue_receive_locked(task);
1227 	rpc_wake_up_queued_task(&xprt->pending, task);
1228 }
1229 EXPORT_SYMBOL_GPL(xprt_complete_rqst);
1230 
1231 static void xprt_timer(struct rpc_task *task)
1232 {
1233 	struct rpc_rqst *req = task->tk_rqstp;
1234 	struct rpc_xprt *xprt = req->rq_xprt;
1235 
1236 	if (task->tk_status != -ETIMEDOUT)
1237 		return;
1238 
1239 	trace_xprt_timer(xprt, req->rq_xid, task->tk_status);
1240 	if (!req->rq_reply_bytes_recvd) {
1241 		if (xprt->ops->timer)
1242 			xprt->ops->timer(xprt, task);
1243 	} else
1244 		task->tk_status = 0;
1245 }
1246 
1247 /**
1248  * xprt_wait_for_reply_request_def - wait for reply
1249  * @task: pointer to rpc_task
1250  *
1251  * Set a request's retransmit timeout based on the transport's
1252  * default timeout parameters.  Used by transports that don't adjust
1253  * the retransmit timeout based on round-trip time estimation,
1254  * and put the task to sleep on the pending queue.
1255  */
1256 void xprt_wait_for_reply_request_def(struct rpc_task *task)
1257 {
1258 	struct rpc_rqst *req = task->tk_rqstp;
1259 
1260 	rpc_sleep_on_timeout(&req->rq_xprt->pending, task, xprt_timer,
1261 			xprt_request_timeout(req));
1262 }
1263 EXPORT_SYMBOL_GPL(xprt_wait_for_reply_request_def);
1264 
1265 /**
1266  * xprt_wait_for_reply_request_rtt - wait for reply using RTT estimator
1267  * @task: pointer to rpc_task
1268  *
1269  * Set a request's retransmit timeout using the RTT estimator,
1270  * and put the task to sleep on the pending queue.
1271  */
1272 void xprt_wait_for_reply_request_rtt(struct rpc_task *task)
1273 {
1274 	int timer = task->tk_msg.rpc_proc->p_timer;
1275 	struct rpc_clnt *clnt = task->tk_client;
1276 	struct rpc_rtt *rtt = clnt->cl_rtt;
1277 	struct rpc_rqst *req = task->tk_rqstp;
1278 	unsigned long max_timeout = clnt->cl_timeout->to_maxval;
1279 	unsigned long timeout;
1280 
1281 	timeout = rpc_calc_rto(rtt, timer);
1282 	timeout <<= rpc_ntimeo(rtt, timer) + req->rq_retries;
1283 	if (timeout > max_timeout || timeout == 0)
1284 		timeout = max_timeout;
1285 	rpc_sleep_on_timeout(&req->rq_xprt->pending, task, xprt_timer,
1286 			jiffies + timeout);
1287 }
1288 EXPORT_SYMBOL_GPL(xprt_wait_for_reply_request_rtt);
1289 
1290 /**
1291  * xprt_request_wait_receive - wait for the reply to an RPC request
1292  * @task: RPC task about to send a request
1293  *
1294  */
1295 void xprt_request_wait_receive(struct rpc_task *task)
1296 {
1297 	struct rpc_rqst *req = task->tk_rqstp;
1298 	struct rpc_xprt *xprt = req->rq_xprt;
1299 
1300 	if (!test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
1301 		return;
1302 	/*
1303 	 * Sleep on the pending queue if we're expecting a reply.
1304 	 * The spinlock ensures atomicity between the test of
1305 	 * req->rq_reply_bytes_recvd, and the call to rpc_sleep_on().
1306 	 */
1307 	spin_lock(&xprt->queue_lock);
1308 	if (test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate)) {
1309 		xprt->ops->wait_for_reply_request(task);
1310 		/*
1311 		 * Send an extra queue wakeup call if the
1312 		 * connection was dropped in case the call to
1313 		 * rpc_sleep_on() raced.
1314 		 */
1315 		if (xprt_request_retransmit_after_disconnect(task))
1316 			rpc_wake_up_queued_task_set_status(&xprt->pending,
1317 					task, -ENOTCONN);
1318 	}
1319 	spin_unlock(&xprt->queue_lock);
1320 }
1321 
1322 static bool
1323 xprt_request_need_enqueue_transmit(struct rpc_task *task, struct rpc_rqst *req)
1324 {
1325 	return !test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate);
1326 }
1327 
1328 /**
1329  * xprt_request_enqueue_transmit - queue a task for transmission
1330  * @task: pointer to rpc_task
1331  *
1332  * Add a task to the transmission queue.
1333  */
1334 void
1335 xprt_request_enqueue_transmit(struct rpc_task *task)
1336 {
1337 	struct rpc_rqst *pos, *req = task->tk_rqstp;
1338 	struct rpc_xprt *xprt = req->rq_xprt;
1339 
1340 	if (xprt_request_need_enqueue_transmit(task, req)) {
1341 		req->rq_bytes_sent = 0;
1342 		spin_lock(&xprt->queue_lock);
1343 		/*
1344 		 * Requests that carry congestion control credits are added
1345 		 * to the head of the list to avoid starvation issues.
1346 		 */
1347 		if (req->rq_cong) {
1348 			xprt_clear_congestion_window_wait(xprt);
1349 			list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1350 				if (pos->rq_cong)
1351 					continue;
1352 				/* Note: req is added _before_ pos */
1353 				list_add_tail(&req->rq_xmit, &pos->rq_xmit);
1354 				INIT_LIST_HEAD(&req->rq_xmit2);
1355 				goto out;
1356 			}
1357 		} else if (RPC_IS_SWAPPER(task)) {
1358 			list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1359 				if (pos->rq_cong || pos->rq_bytes_sent)
1360 					continue;
1361 				if (RPC_IS_SWAPPER(pos->rq_task))
1362 					continue;
1363 				/* Note: req is added _before_ pos */
1364 				list_add_tail(&req->rq_xmit, &pos->rq_xmit);
1365 				INIT_LIST_HEAD(&req->rq_xmit2);
1366 				goto out;
1367 			}
1368 		} else if (!req->rq_seqno) {
1369 			list_for_each_entry(pos, &xprt->xmit_queue, rq_xmit) {
1370 				if (pos->rq_task->tk_owner != task->tk_owner)
1371 					continue;
1372 				list_add_tail(&req->rq_xmit2, &pos->rq_xmit2);
1373 				INIT_LIST_HEAD(&req->rq_xmit);
1374 				goto out;
1375 			}
1376 		}
1377 		list_add_tail(&req->rq_xmit, &xprt->xmit_queue);
1378 		INIT_LIST_HEAD(&req->rq_xmit2);
1379 out:
1380 		atomic_long_inc(&xprt->xmit_queuelen);
1381 		set_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate);
1382 		spin_unlock(&xprt->queue_lock);
1383 	}
1384 }
1385 
1386 /**
1387  * xprt_request_dequeue_transmit_locked - remove a task from the transmission queue
1388  * @task: pointer to rpc_task
1389  *
1390  * Remove a task from the transmission queue
1391  * Caller must hold xprt->queue_lock
1392  */
1393 static void
1394 xprt_request_dequeue_transmit_locked(struct rpc_task *task)
1395 {
1396 	struct rpc_rqst *req = task->tk_rqstp;
1397 
1398 	if (!test_and_clear_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1399 		return;
1400 	if (!list_empty(&req->rq_xmit)) {
1401 		list_del(&req->rq_xmit);
1402 		if (!list_empty(&req->rq_xmit2)) {
1403 			struct rpc_rqst *next = list_first_entry(&req->rq_xmit2,
1404 					struct rpc_rqst, rq_xmit2);
1405 			list_del(&req->rq_xmit2);
1406 			list_add_tail(&next->rq_xmit, &next->rq_xprt->xmit_queue);
1407 		}
1408 	} else
1409 		list_del(&req->rq_xmit2);
1410 	atomic_long_dec(&req->rq_xprt->xmit_queuelen);
1411 }
1412 
1413 /**
1414  * xprt_request_dequeue_transmit - remove a task from the transmission queue
1415  * @task: pointer to rpc_task
1416  *
1417  * Remove a task from the transmission queue
1418  */
1419 static void
1420 xprt_request_dequeue_transmit(struct rpc_task *task)
1421 {
1422 	struct rpc_rqst *req = task->tk_rqstp;
1423 	struct rpc_xprt *xprt = req->rq_xprt;
1424 
1425 	spin_lock(&xprt->queue_lock);
1426 	xprt_request_dequeue_transmit_locked(task);
1427 	spin_unlock(&xprt->queue_lock);
1428 }
1429 
1430 /**
1431  * xprt_request_dequeue_xprt - remove a task from the transmit+receive queue
1432  * @task: pointer to rpc_task
1433  *
1434  * Remove a task from the transmit and receive queues, and ensure that
1435  * it is not pinned by the receive work item.
1436  */
1437 void
1438 xprt_request_dequeue_xprt(struct rpc_task *task)
1439 {
1440 	struct rpc_rqst	*req = task->tk_rqstp;
1441 	struct rpc_xprt *xprt = req->rq_xprt;
1442 
1443 	if (test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate) ||
1444 	    test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) ||
1445 	    xprt_is_pinned_rqst(req)) {
1446 		spin_lock(&xprt->queue_lock);
1447 		xprt_request_dequeue_transmit_locked(task);
1448 		xprt_request_dequeue_receive_locked(task);
1449 		while (xprt_is_pinned_rqst(req)) {
1450 			set_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
1451 			spin_unlock(&xprt->queue_lock);
1452 			xprt_wait_on_pinned_rqst(req);
1453 			spin_lock(&xprt->queue_lock);
1454 			clear_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
1455 		}
1456 		spin_unlock(&xprt->queue_lock);
1457 	}
1458 }
1459 
1460 /**
1461  * xprt_request_prepare - prepare an encoded request for transport
1462  * @req: pointer to rpc_rqst
1463  *
1464  * Calls into the transport layer to do whatever is needed to prepare
1465  * the request for transmission or receive.
1466  */
1467 void
1468 xprt_request_prepare(struct rpc_rqst *req)
1469 {
1470 	struct rpc_xprt *xprt = req->rq_xprt;
1471 
1472 	if (xprt->ops->prepare_request)
1473 		xprt->ops->prepare_request(req);
1474 }
1475 
1476 /**
1477  * xprt_request_need_retransmit - Test if a task needs retransmission
1478  * @task: pointer to rpc_task
1479  *
1480  * Test for whether a connection breakage requires the task to retransmit
1481  */
1482 bool
1483 xprt_request_need_retransmit(struct rpc_task *task)
1484 {
1485 	return xprt_request_retransmit_after_disconnect(task);
1486 }
1487 
1488 /**
1489  * xprt_prepare_transmit - reserve the transport before sending a request
1490  * @task: RPC task about to send a request
1491  *
1492  */
1493 bool xprt_prepare_transmit(struct rpc_task *task)
1494 {
1495 	struct rpc_rqst	*req = task->tk_rqstp;
1496 	struct rpc_xprt	*xprt = req->rq_xprt;
1497 
1498 	if (!xprt_lock_write(xprt, task)) {
1499 		/* Race breaker: someone may have transmitted us */
1500 		if (!test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1501 			rpc_wake_up_queued_task_set_status(&xprt->sending,
1502 					task, 0);
1503 		return false;
1504 
1505 	}
1506 	return true;
1507 }
1508 
1509 void xprt_end_transmit(struct rpc_task *task)
1510 {
1511 	struct rpc_xprt	*xprt = task->tk_rqstp->rq_xprt;
1512 
1513 	xprt_inject_disconnect(xprt);
1514 	xprt_release_write(xprt, task);
1515 }
1516 
1517 /**
1518  * xprt_request_transmit - send an RPC request on a transport
1519  * @req: pointer to request to transmit
1520  * @snd_task: RPC task that owns the transport lock
1521  *
1522  * This performs the transmission of a single request.
1523  * Note that if the request is not the same as snd_task, then it
1524  * does need to be pinned.
1525  * Returns '0' on success.
1526  */
1527 static int
1528 xprt_request_transmit(struct rpc_rqst *req, struct rpc_task *snd_task)
1529 {
1530 	struct rpc_xprt *xprt = req->rq_xprt;
1531 	struct rpc_task *task = req->rq_task;
1532 	unsigned int connect_cookie;
1533 	int is_retrans = RPC_WAS_SENT(task);
1534 	int status;
1535 
1536 	if (!req->rq_bytes_sent) {
1537 		if (xprt_request_data_received(task)) {
1538 			status = 0;
1539 			goto out_dequeue;
1540 		}
1541 		/* Verify that our message lies in the RPCSEC_GSS window */
1542 		if (rpcauth_xmit_need_reencode(task)) {
1543 			status = -EBADMSG;
1544 			goto out_dequeue;
1545 		}
1546 		if (RPC_SIGNALLED(task)) {
1547 			status = -ERESTARTSYS;
1548 			goto out_dequeue;
1549 		}
1550 	}
1551 
1552 	/*
1553 	 * Update req->rq_ntrans before transmitting to avoid races with
1554 	 * xprt_update_rtt(), which needs to know that it is recording a
1555 	 * reply to the first transmission.
1556 	 */
1557 	req->rq_ntrans++;
1558 
1559 	trace_rpc_xdr_sendto(task, &req->rq_snd_buf);
1560 	connect_cookie = xprt->connect_cookie;
1561 	status = xprt->ops->send_request(req);
1562 	if (status != 0) {
1563 		req->rq_ntrans--;
1564 		trace_xprt_transmit(req, status);
1565 		return status;
1566 	}
1567 
1568 	if (is_retrans) {
1569 		task->tk_client->cl_stats->rpcretrans++;
1570 		trace_xprt_retransmit(req);
1571 	}
1572 
1573 	xprt_inject_disconnect(xprt);
1574 
1575 	task->tk_flags |= RPC_TASK_SENT;
1576 	spin_lock(&xprt->transport_lock);
1577 
1578 	xprt->stat.sends++;
1579 	xprt->stat.req_u += xprt->stat.sends - xprt->stat.recvs;
1580 	xprt->stat.bklog_u += xprt->backlog.qlen;
1581 	xprt->stat.sending_u += xprt->sending.qlen;
1582 	xprt->stat.pending_u += xprt->pending.qlen;
1583 	spin_unlock(&xprt->transport_lock);
1584 
1585 	req->rq_connect_cookie = connect_cookie;
1586 out_dequeue:
1587 	trace_xprt_transmit(req, status);
1588 	xprt_request_dequeue_transmit(task);
1589 	rpc_wake_up_queued_task_set_status(&xprt->sending, task, status);
1590 	return status;
1591 }
1592 
1593 /**
1594  * xprt_transmit - send an RPC request on a transport
1595  * @task: controlling RPC task
1596  *
1597  * Attempts to drain the transmit queue. On exit, either the transport
1598  * signalled an error that needs to be handled before transmission can
1599  * resume, or @task finished transmitting, and detected that it already
1600  * received a reply.
1601  */
1602 void
1603 xprt_transmit(struct rpc_task *task)
1604 {
1605 	struct rpc_rqst *next, *req = task->tk_rqstp;
1606 	struct rpc_xprt	*xprt = req->rq_xprt;
1607 	int status;
1608 
1609 	spin_lock(&xprt->queue_lock);
1610 	for (;;) {
1611 		next = list_first_entry_or_null(&xprt->xmit_queue,
1612 						struct rpc_rqst, rq_xmit);
1613 		if (!next)
1614 			break;
1615 		xprt_pin_rqst(next);
1616 		spin_unlock(&xprt->queue_lock);
1617 		status = xprt_request_transmit(next, task);
1618 		if (status == -EBADMSG && next != req)
1619 			status = 0;
1620 		spin_lock(&xprt->queue_lock);
1621 		xprt_unpin_rqst(next);
1622 		if (status < 0) {
1623 			if (test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1624 				task->tk_status = status;
1625 			break;
1626 		}
1627 		/* Was @task transmitted, and has it received a reply? */
1628 		if (xprt_request_data_received(task) &&
1629 		    !test_bit(RPC_TASK_NEED_XMIT, &task->tk_runstate))
1630 			break;
1631 		cond_resched_lock(&xprt->queue_lock);
1632 	}
1633 	spin_unlock(&xprt->queue_lock);
1634 }
1635 
1636 static void xprt_complete_request_init(struct rpc_task *task)
1637 {
1638 	if (task->tk_rqstp)
1639 		xprt_request_init(task);
1640 }
1641 
1642 void xprt_add_backlog(struct rpc_xprt *xprt, struct rpc_task *task)
1643 {
1644 	set_bit(XPRT_CONGESTED, &xprt->state);
1645 	rpc_sleep_on(&xprt->backlog, task, xprt_complete_request_init);
1646 }
1647 EXPORT_SYMBOL_GPL(xprt_add_backlog);
1648 
1649 static bool __xprt_set_rq(struct rpc_task *task, void *data)
1650 {
1651 	struct rpc_rqst *req = data;
1652 
1653 	if (task->tk_rqstp == NULL) {
1654 		memset(req, 0, sizeof(*req));	/* mark unused */
1655 		task->tk_rqstp = req;
1656 		return true;
1657 	}
1658 	return false;
1659 }
1660 
1661 bool xprt_wake_up_backlog(struct rpc_xprt *xprt, struct rpc_rqst *req)
1662 {
1663 	if (rpc_wake_up_first(&xprt->backlog, __xprt_set_rq, req) == NULL) {
1664 		clear_bit(XPRT_CONGESTED, &xprt->state);
1665 		return false;
1666 	}
1667 	return true;
1668 }
1669 EXPORT_SYMBOL_GPL(xprt_wake_up_backlog);
1670 
1671 static bool xprt_throttle_congested(struct rpc_xprt *xprt, struct rpc_task *task)
1672 {
1673 	bool ret = false;
1674 
1675 	if (!test_bit(XPRT_CONGESTED, &xprt->state))
1676 		goto out;
1677 	spin_lock(&xprt->reserve_lock);
1678 	if (test_bit(XPRT_CONGESTED, &xprt->state)) {
1679 		xprt_add_backlog(xprt, task);
1680 		ret = true;
1681 	}
1682 	spin_unlock(&xprt->reserve_lock);
1683 out:
1684 	return ret;
1685 }
1686 
1687 static struct rpc_rqst *xprt_dynamic_alloc_slot(struct rpc_xprt *xprt)
1688 {
1689 	struct rpc_rqst *req = ERR_PTR(-EAGAIN);
1690 
1691 	if (xprt->num_reqs >= xprt->max_reqs)
1692 		goto out;
1693 	++xprt->num_reqs;
1694 	spin_unlock(&xprt->reserve_lock);
1695 	req = kzalloc(sizeof(struct rpc_rqst), GFP_NOFS);
1696 	spin_lock(&xprt->reserve_lock);
1697 	if (req != NULL)
1698 		goto out;
1699 	--xprt->num_reqs;
1700 	req = ERR_PTR(-ENOMEM);
1701 out:
1702 	return req;
1703 }
1704 
1705 static bool xprt_dynamic_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
1706 {
1707 	if (xprt->num_reqs > xprt->min_reqs) {
1708 		--xprt->num_reqs;
1709 		kfree(req);
1710 		return true;
1711 	}
1712 	return false;
1713 }
1714 
1715 void xprt_alloc_slot(struct rpc_xprt *xprt, struct rpc_task *task)
1716 {
1717 	struct rpc_rqst *req;
1718 
1719 	spin_lock(&xprt->reserve_lock);
1720 	if (!list_empty(&xprt->free)) {
1721 		req = list_entry(xprt->free.next, struct rpc_rqst, rq_list);
1722 		list_del(&req->rq_list);
1723 		goto out_init_req;
1724 	}
1725 	req = xprt_dynamic_alloc_slot(xprt);
1726 	if (!IS_ERR(req))
1727 		goto out_init_req;
1728 	switch (PTR_ERR(req)) {
1729 	case -ENOMEM:
1730 		dprintk("RPC:       dynamic allocation of request slot "
1731 				"failed! Retrying\n");
1732 		task->tk_status = -ENOMEM;
1733 		break;
1734 	case -EAGAIN:
1735 		xprt_add_backlog(xprt, task);
1736 		dprintk("RPC:       waiting for request slot\n");
1737 		fallthrough;
1738 	default:
1739 		task->tk_status = -EAGAIN;
1740 	}
1741 	spin_unlock(&xprt->reserve_lock);
1742 	return;
1743 out_init_req:
1744 	xprt->stat.max_slots = max_t(unsigned int, xprt->stat.max_slots,
1745 				     xprt->num_reqs);
1746 	spin_unlock(&xprt->reserve_lock);
1747 
1748 	task->tk_status = 0;
1749 	task->tk_rqstp = req;
1750 }
1751 EXPORT_SYMBOL_GPL(xprt_alloc_slot);
1752 
1753 void xprt_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
1754 {
1755 	spin_lock(&xprt->reserve_lock);
1756 	if (!xprt_wake_up_backlog(xprt, req) &&
1757 	    !xprt_dynamic_free_slot(xprt, req)) {
1758 		memset(req, 0, sizeof(*req));	/* mark unused */
1759 		list_add(&req->rq_list, &xprt->free);
1760 	}
1761 	spin_unlock(&xprt->reserve_lock);
1762 }
1763 EXPORT_SYMBOL_GPL(xprt_free_slot);
1764 
1765 static void xprt_free_all_slots(struct rpc_xprt *xprt)
1766 {
1767 	struct rpc_rqst *req;
1768 	while (!list_empty(&xprt->free)) {
1769 		req = list_first_entry(&xprt->free, struct rpc_rqst, rq_list);
1770 		list_del(&req->rq_list);
1771 		kfree(req);
1772 	}
1773 }
1774 
1775 static DEFINE_IDA(rpc_xprt_ids);
1776 
1777 void xprt_cleanup_ids(void)
1778 {
1779 	ida_destroy(&rpc_xprt_ids);
1780 }
1781 
1782 static int xprt_alloc_id(struct rpc_xprt *xprt)
1783 {
1784 	int id;
1785 
1786 	id = ida_simple_get(&rpc_xprt_ids, 0, 0, GFP_KERNEL);
1787 	if (id < 0)
1788 		return id;
1789 
1790 	xprt->id = id;
1791 	return 0;
1792 }
1793 
1794 static void xprt_free_id(struct rpc_xprt *xprt)
1795 {
1796 	ida_simple_remove(&rpc_xprt_ids, xprt->id);
1797 }
1798 
1799 struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
1800 		unsigned int num_prealloc,
1801 		unsigned int max_alloc)
1802 {
1803 	struct rpc_xprt *xprt;
1804 	struct rpc_rqst *req;
1805 	int i;
1806 
1807 	xprt = kzalloc(size, GFP_KERNEL);
1808 	if (xprt == NULL)
1809 		goto out;
1810 
1811 	xprt_alloc_id(xprt);
1812 	xprt_init(xprt, net);
1813 
1814 	for (i = 0; i < num_prealloc; i++) {
1815 		req = kzalloc(sizeof(struct rpc_rqst), GFP_KERNEL);
1816 		if (!req)
1817 			goto out_free;
1818 		list_add(&req->rq_list, &xprt->free);
1819 	}
1820 	if (max_alloc > num_prealloc)
1821 		xprt->max_reqs = max_alloc;
1822 	else
1823 		xprt->max_reqs = num_prealloc;
1824 	xprt->min_reqs = num_prealloc;
1825 	xprt->num_reqs = num_prealloc;
1826 
1827 	return xprt;
1828 
1829 out_free:
1830 	xprt_free(xprt);
1831 out:
1832 	return NULL;
1833 }
1834 EXPORT_SYMBOL_GPL(xprt_alloc);
1835 
1836 void xprt_free(struct rpc_xprt *xprt)
1837 {
1838 	put_net(xprt->xprt_net);
1839 	xprt_free_all_slots(xprt);
1840 	xprt_free_id(xprt);
1841 	rpc_sysfs_xprt_destroy(xprt);
1842 	kfree_rcu(xprt, rcu);
1843 }
1844 EXPORT_SYMBOL_GPL(xprt_free);
1845 
1846 static void
1847 xprt_init_connect_cookie(struct rpc_rqst *req, struct rpc_xprt *xprt)
1848 {
1849 	req->rq_connect_cookie = xprt_connect_cookie(xprt) - 1;
1850 }
1851 
1852 static __be32
1853 xprt_alloc_xid(struct rpc_xprt *xprt)
1854 {
1855 	__be32 xid;
1856 
1857 	spin_lock(&xprt->reserve_lock);
1858 	xid = (__force __be32)xprt->xid++;
1859 	spin_unlock(&xprt->reserve_lock);
1860 	return xid;
1861 }
1862 
1863 static void
1864 xprt_init_xid(struct rpc_xprt *xprt)
1865 {
1866 	xprt->xid = prandom_u32();
1867 }
1868 
1869 static void
1870 xprt_request_init(struct rpc_task *task)
1871 {
1872 	struct rpc_xprt *xprt = task->tk_xprt;
1873 	struct rpc_rqst	*req = task->tk_rqstp;
1874 
1875 	req->rq_task	= task;
1876 	req->rq_xprt    = xprt;
1877 	req->rq_buffer  = NULL;
1878 	req->rq_xid	= xprt_alloc_xid(xprt);
1879 	xprt_init_connect_cookie(req, xprt);
1880 	req->rq_snd_buf.len = 0;
1881 	req->rq_snd_buf.buflen = 0;
1882 	req->rq_rcv_buf.len = 0;
1883 	req->rq_rcv_buf.buflen = 0;
1884 	req->rq_snd_buf.bvec = NULL;
1885 	req->rq_rcv_buf.bvec = NULL;
1886 	req->rq_release_snd_buf = NULL;
1887 	xprt_init_majortimeo(task, req);
1888 
1889 	trace_xprt_reserve(req);
1890 }
1891 
1892 static void
1893 xprt_do_reserve(struct rpc_xprt *xprt, struct rpc_task *task)
1894 {
1895 	xprt->ops->alloc_slot(xprt, task);
1896 	if (task->tk_rqstp != NULL)
1897 		xprt_request_init(task);
1898 }
1899 
1900 /**
1901  * xprt_reserve - allocate an RPC request slot
1902  * @task: RPC task requesting a slot allocation
1903  *
1904  * If the transport is marked as being congested, or if no more
1905  * slots are available, place the task on the transport's
1906  * backlog queue.
1907  */
1908 void xprt_reserve(struct rpc_task *task)
1909 {
1910 	struct rpc_xprt *xprt = task->tk_xprt;
1911 
1912 	task->tk_status = 0;
1913 	if (task->tk_rqstp != NULL)
1914 		return;
1915 
1916 	task->tk_status = -EAGAIN;
1917 	if (!xprt_throttle_congested(xprt, task))
1918 		xprt_do_reserve(xprt, task);
1919 }
1920 
1921 /**
1922  * xprt_retry_reserve - allocate an RPC request slot
1923  * @task: RPC task requesting a slot allocation
1924  *
1925  * If no more slots are available, place the task on the transport's
1926  * backlog queue.
1927  * Note that the only difference with xprt_reserve is that we now
1928  * ignore the value of the XPRT_CONGESTED flag.
1929  */
1930 void xprt_retry_reserve(struct rpc_task *task)
1931 {
1932 	struct rpc_xprt *xprt = task->tk_xprt;
1933 
1934 	task->tk_status = 0;
1935 	if (task->tk_rqstp != NULL)
1936 		return;
1937 
1938 	task->tk_status = -EAGAIN;
1939 	xprt_do_reserve(xprt, task);
1940 }
1941 
1942 /**
1943  * xprt_release - release an RPC request slot
1944  * @task: task which is finished with the slot
1945  *
1946  */
1947 void xprt_release(struct rpc_task *task)
1948 {
1949 	struct rpc_xprt	*xprt;
1950 	struct rpc_rqst	*req = task->tk_rqstp;
1951 
1952 	if (req == NULL) {
1953 		if (task->tk_client) {
1954 			xprt = task->tk_xprt;
1955 			xprt_release_write(xprt, task);
1956 		}
1957 		return;
1958 	}
1959 
1960 	xprt = req->rq_xprt;
1961 	xprt_request_dequeue_xprt(task);
1962 	spin_lock(&xprt->transport_lock);
1963 	xprt->ops->release_xprt(xprt, task);
1964 	if (xprt->ops->release_request)
1965 		xprt->ops->release_request(task);
1966 	xprt_schedule_autodisconnect(xprt);
1967 	spin_unlock(&xprt->transport_lock);
1968 	if (req->rq_buffer)
1969 		xprt->ops->buf_free(task);
1970 	xdr_free_bvec(&req->rq_rcv_buf);
1971 	xdr_free_bvec(&req->rq_snd_buf);
1972 	if (req->rq_cred != NULL)
1973 		put_rpccred(req->rq_cred);
1974 	if (req->rq_release_snd_buf)
1975 		req->rq_release_snd_buf(req);
1976 
1977 	task->tk_rqstp = NULL;
1978 	if (likely(!bc_prealloc(req)))
1979 		xprt->ops->free_slot(xprt, req);
1980 	else
1981 		xprt_free_bc_request(req);
1982 }
1983 
1984 #ifdef CONFIG_SUNRPC_BACKCHANNEL
1985 void
1986 xprt_init_bc_request(struct rpc_rqst *req, struct rpc_task *task)
1987 {
1988 	struct xdr_buf *xbufp = &req->rq_snd_buf;
1989 
1990 	task->tk_rqstp = req;
1991 	req->rq_task = task;
1992 	xprt_init_connect_cookie(req, req->rq_xprt);
1993 	/*
1994 	 * Set up the xdr_buf length.
1995 	 * This also indicates that the buffer is XDR encoded already.
1996 	 */
1997 	xbufp->len = xbufp->head[0].iov_len + xbufp->page_len +
1998 		xbufp->tail[0].iov_len;
1999 }
2000 #endif
2001 
2002 static void xprt_init(struct rpc_xprt *xprt, struct net *net)
2003 {
2004 	kref_init(&xprt->kref);
2005 
2006 	spin_lock_init(&xprt->transport_lock);
2007 	spin_lock_init(&xprt->reserve_lock);
2008 	spin_lock_init(&xprt->queue_lock);
2009 
2010 	INIT_LIST_HEAD(&xprt->free);
2011 	xprt->recv_queue = RB_ROOT;
2012 	INIT_LIST_HEAD(&xprt->xmit_queue);
2013 #if defined(CONFIG_SUNRPC_BACKCHANNEL)
2014 	spin_lock_init(&xprt->bc_pa_lock);
2015 	INIT_LIST_HEAD(&xprt->bc_pa_list);
2016 #endif /* CONFIG_SUNRPC_BACKCHANNEL */
2017 	INIT_LIST_HEAD(&xprt->xprt_switch);
2018 
2019 	xprt->last_used = jiffies;
2020 	xprt->cwnd = RPC_INITCWND;
2021 	xprt->bind_index = 0;
2022 
2023 	rpc_init_wait_queue(&xprt->binding, "xprt_binding");
2024 	rpc_init_wait_queue(&xprt->pending, "xprt_pending");
2025 	rpc_init_wait_queue(&xprt->sending, "xprt_sending");
2026 	rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog");
2027 
2028 	xprt_init_xid(xprt);
2029 
2030 	xprt->xprt_net = get_net(net);
2031 }
2032 
2033 /**
2034  * xprt_create_transport - create an RPC transport
2035  * @args: rpc transport creation arguments
2036  *
2037  */
2038 struct rpc_xprt *xprt_create_transport(struct xprt_create *args)
2039 {
2040 	struct rpc_xprt	*xprt;
2041 	const struct xprt_class *t;
2042 
2043 	t = xprt_class_find_by_ident(args->ident);
2044 	if (!t) {
2045 		dprintk("RPC: transport (%d) not supported\n", args->ident);
2046 		return ERR_PTR(-EIO);
2047 	}
2048 
2049 	xprt = t->setup(args);
2050 	xprt_class_release(t);
2051 
2052 	if (IS_ERR(xprt))
2053 		goto out;
2054 	if (args->flags & XPRT_CREATE_NO_IDLE_TIMEOUT)
2055 		xprt->idle_timeout = 0;
2056 	INIT_WORK(&xprt->task_cleanup, xprt_autoclose);
2057 	if (xprt_has_timer(xprt))
2058 		timer_setup(&xprt->timer, xprt_init_autodisconnect, 0);
2059 	else
2060 		timer_setup(&xprt->timer, NULL, 0);
2061 
2062 	if (strlen(args->servername) > RPC_MAXNETNAMELEN) {
2063 		xprt_destroy(xprt);
2064 		return ERR_PTR(-EINVAL);
2065 	}
2066 	xprt->servername = kstrdup(args->servername, GFP_KERNEL);
2067 	if (xprt->servername == NULL) {
2068 		xprt_destroy(xprt);
2069 		return ERR_PTR(-ENOMEM);
2070 	}
2071 
2072 	rpc_xprt_debugfs_register(xprt);
2073 
2074 	trace_xprt_create(xprt);
2075 out:
2076 	return xprt;
2077 }
2078 
2079 static void xprt_destroy_cb(struct work_struct *work)
2080 {
2081 	struct rpc_xprt *xprt =
2082 		container_of(work, struct rpc_xprt, task_cleanup);
2083 
2084 	trace_xprt_destroy(xprt);
2085 
2086 	rpc_xprt_debugfs_unregister(xprt);
2087 	rpc_destroy_wait_queue(&xprt->binding);
2088 	rpc_destroy_wait_queue(&xprt->pending);
2089 	rpc_destroy_wait_queue(&xprt->sending);
2090 	rpc_destroy_wait_queue(&xprt->backlog);
2091 	kfree(xprt->servername);
2092 	/*
2093 	 * Destroy any existing back channel
2094 	 */
2095 	xprt_destroy_backchannel(xprt, UINT_MAX);
2096 
2097 	/*
2098 	 * Tear down transport state and free the rpc_xprt
2099 	 */
2100 	xprt->ops->destroy(xprt);
2101 }
2102 
2103 /**
2104  * xprt_destroy - destroy an RPC transport, killing off all requests.
2105  * @xprt: transport to destroy
2106  *
2107  */
2108 static void xprt_destroy(struct rpc_xprt *xprt)
2109 {
2110 	/*
2111 	 * Exclude transport connect/disconnect handlers and autoclose
2112 	 */
2113 	wait_on_bit_lock(&xprt->state, XPRT_LOCKED, TASK_UNINTERRUPTIBLE);
2114 
2115 	del_timer_sync(&xprt->timer);
2116 
2117 	/*
2118 	 * Destroy sockets etc from the system workqueue so they can
2119 	 * safely flush receive work running on rpciod.
2120 	 */
2121 	INIT_WORK(&xprt->task_cleanup, xprt_destroy_cb);
2122 	schedule_work(&xprt->task_cleanup);
2123 }
2124 
2125 static void xprt_destroy_kref(struct kref *kref)
2126 {
2127 	xprt_destroy(container_of(kref, struct rpc_xprt, kref));
2128 }
2129 
2130 /**
2131  * xprt_get - return a reference to an RPC transport.
2132  * @xprt: pointer to the transport
2133  *
2134  */
2135 struct rpc_xprt *xprt_get(struct rpc_xprt *xprt)
2136 {
2137 	if (xprt != NULL && kref_get_unless_zero(&xprt->kref))
2138 		return xprt;
2139 	return NULL;
2140 }
2141 EXPORT_SYMBOL_GPL(xprt_get);
2142 
2143 /**
2144  * xprt_put - release a reference to an RPC transport.
2145  * @xprt: pointer to the transport
2146  *
2147  */
2148 void xprt_put(struct rpc_xprt *xprt)
2149 {
2150 	if (xprt != NULL)
2151 		kref_put(&xprt->kref, xprt_destroy_kref);
2152 }
2153 EXPORT_SYMBOL_GPL(xprt_put);
2154