net/sunrpc/xprt.c

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