Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /* |
2 | * linux/net/sunrpc/xprt.c | |
3 | * | |
4 | * This is a generic RPC call interface supporting congestion avoidance, | |
5 | * and asynchronous calls. | |
6 | * | |
7 | * The interface works like this: | |
8 | * | |
9 | * - When a process places a call, it allocates a request slot if | |
10 | * one is available. Otherwise, it sleeps on the backlog queue | |
11 | * (xprt_reserve). | |
12 | * - Next, the caller puts together the RPC message, stuffs it into | |
55aa4f58 CL |
13 | * the request struct, and calls xprt_transmit(). |
14 | * - xprt_transmit sends the message and installs the caller on the | |
15 | * transport's wait list. At the same time, it installs a timer that | |
1da177e4 LT |
16 | * is run after the packet's timeout has expired. |
17 | * - When a packet arrives, the data_ready handler walks the list of | |
55aa4f58 | 18 | * pending requests for that transport. If a matching XID is found, the |
1da177e4 LT |
19 | * caller is woken up, and the timer removed. |
20 | * - When no reply arrives within the timeout interval, the timer is | |
21 | * fired by the kernel and runs xprt_timer(). It either adjusts the | |
22 | * timeout values (minor timeout) or wakes up the caller with a status | |
23 | * of -ETIMEDOUT. | |
24 | * - When the caller receives a notification from RPC that a reply arrived, | |
25 | * it should release the RPC slot, and process the reply. | |
26 | * If the call timed out, it may choose to retry the operation by | |
27 | * adjusting the initial timeout value, and simply calling rpc_call | |
28 | * again. | |
29 | * | |
30 | * Support for async RPC is done through a set of RPC-specific scheduling | |
31 | * primitives that `transparently' work for processes as well as async | |
32 | * tasks that rely on callbacks. | |
33 | * | |
34 | * Copyright (C) 1995-1997, Olaf Kirch <okir@monad.swb.de> | |
55aa4f58 CL |
35 | * |
36 | * Transport switch API copyright (C) 2005, Chuck Lever <cel@netapp.com> | |
1da177e4 LT |
37 | */ |
38 | ||
a246b010 CL |
39 | #include <linux/module.h> |
40 | ||
1da177e4 | 41 | #include <linux/types.h> |
a246b010 | 42 | #include <linux/interrupt.h> |
1da177e4 LT |
43 | #include <linux/workqueue.h> |
44 | #include <linux/random.h> | |
45 | ||
a246b010 | 46 | #include <linux/sunrpc/clnt.h> |
1da177e4 LT |
47 | |
48 | /* | |
49 | * Local variables | |
50 | */ | |
51 | ||
52 | #ifdef RPC_DEBUG | |
53 | # undef RPC_DEBUG_DATA | |
54 | # define RPCDBG_FACILITY RPCDBG_XPRT | |
55 | #endif | |
56 | ||
1da177e4 LT |
57 | /* |
58 | * Local functions | |
59 | */ | |
60 | static void xprt_request_init(struct rpc_task *, struct rpc_xprt *); | |
61 | static inline void do_xprt_reserve(struct rpc_task *); | |
1da177e4 | 62 | static void xprt_connect_status(struct rpc_task *task); |
1da177e4 LT |
63 | static int __xprt_get_cong(struct rpc_xprt *, struct rpc_task *); |
64 | ||
65 | static int xprt_clear_backlog(struct rpc_xprt *xprt); | |
66 | ||
1da177e4 | 67 | /* |
55aa4f58 | 68 | * Serialize write access to transports, in order to prevent different |
1da177e4 | 69 | * requests from interfering with each other. |
55aa4f58 | 70 | * Also prevents transport connects from colliding with writes. |
1da177e4 LT |
71 | */ |
72 | static int | |
73 | __xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task) | |
74 | { | |
75 | struct rpc_rqst *req = task->tk_rqstp; | |
76 | ||
2226feb6 | 77 | if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) { |
1da177e4 LT |
78 | if (task == xprt->snd_task) |
79 | return 1; | |
1da177e4 LT |
80 | goto out_sleep; |
81 | } | |
82 | if (xprt->nocong || __xprt_get_cong(xprt, task)) { | |
83 | xprt->snd_task = task; | |
84 | if (req) { | |
85 | req->rq_bytes_sent = 0; | |
86 | req->rq_ntrans++; | |
87 | } | |
88 | return 1; | |
89 | } | |
90 | smp_mb__before_clear_bit(); | |
2226feb6 | 91 | clear_bit(XPRT_LOCKED, &xprt->state); |
1da177e4 LT |
92 | smp_mb__after_clear_bit(); |
93 | out_sleep: | |
55aa4f58 | 94 | dprintk("RPC: %4d failed to lock transport %p\n", task->tk_pid, xprt); |
1da177e4 LT |
95 | task->tk_timeout = 0; |
96 | task->tk_status = -EAGAIN; | |
97 | if (req && req->rq_ntrans) | |
98 | rpc_sleep_on(&xprt->resend, task, NULL, NULL); | |
99 | else | |
100 | rpc_sleep_on(&xprt->sending, task, NULL, NULL); | |
101 | return 0; | |
102 | } | |
103 | ||
104 | static inline int | |
105 | xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task) | |
106 | { | |
107 | int retval; | |
108 | ||
4a0f8c04 | 109 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 | 110 | retval = __xprt_lock_write(xprt, task); |
4a0f8c04 | 111 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
112 | return retval; |
113 | } | |
114 | ||
115 | ||
116 | static void | |
117 | __xprt_lock_write_next(struct rpc_xprt *xprt) | |
118 | { | |
119 | struct rpc_task *task; | |
120 | ||
2226feb6 | 121 | if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) |
1da177e4 LT |
122 | return; |
123 | if (!xprt->nocong && RPCXPRT_CONGESTED(xprt)) | |
124 | goto out_unlock; | |
125 | task = rpc_wake_up_next(&xprt->resend); | |
126 | if (!task) { | |
127 | task = rpc_wake_up_next(&xprt->sending); | |
128 | if (!task) | |
129 | goto out_unlock; | |
130 | } | |
131 | if (xprt->nocong || __xprt_get_cong(xprt, task)) { | |
132 | struct rpc_rqst *req = task->tk_rqstp; | |
133 | xprt->snd_task = task; | |
134 | if (req) { | |
135 | req->rq_bytes_sent = 0; | |
136 | req->rq_ntrans++; | |
137 | } | |
138 | return; | |
139 | } | |
140 | out_unlock: | |
141 | smp_mb__before_clear_bit(); | |
2226feb6 | 142 | clear_bit(XPRT_LOCKED, &xprt->state); |
1da177e4 LT |
143 | smp_mb__after_clear_bit(); |
144 | } | |
145 | ||
146 | /* | |
55aa4f58 | 147 | * Releases the transport for use by other requests. |
1da177e4 LT |
148 | */ |
149 | static void | |
150 | __xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task) | |
151 | { | |
152 | if (xprt->snd_task == task) { | |
153 | xprt->snd_task = NULL; | |
154 | smp_mb__before_clear_bit(); | |
2226feb6 | 155 | clear_bit(XPRT_LOCKED, &xprt->state); |
1da177e4 LT |
156 | smp_mb__after_clear_bit(); |
157 | __xprt_lock_write_next(xprt); | |
158 | } | |
159 | } | |
160 | ||
161 | static inline void | |
162 | xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task) | |
163 | { | |
4a0f8c04 | 164 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 | 165 | __xprt_release_write(xprt, task); |
4a0f8c04 | 166 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
167 | } |
168 | ||
1da177e4 LT |
169 | /* |
170 | * Van Jacobson congestion avoidance. Check if the congestion window | |
171 | * overflowed. Put the task to sleep if this is the case. | |
172 | */ | |
173 | static int | |
174 | __xprt_get_cong(struct rpc_xprt *xprt, struct rpc_task *task) | |
175 | { | |
176 | struct rpc_rqst *req = task->tk_rqstp; | |
177 | ||
178 | if (req->rq_cong) | |
179 | return 1; | |
180 | dprintk("RPC: %4d xprt_cwnd_limited cong = %ld cwnd = %ld\n", | |
181 | task->tk_pid, xprt->cong, xprt->cwnd); | |
182 | if (RPCXPRT_CONGESTED(xprt)) | |
183 | return 0; | |
184 | req->rq_cong = 1; | |
185 | xprt->cong += RPC_CWNDSCALE; | |
186 | return 1; | |
187 | } | |
188 | ||
189 | /* | |
190 | * Adjust the congestion window, and wake up the next task | |
191 | * that has been sleeping due to congestion | |
192 | */ | |
193 | static void | |
194 | __xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req) | |
195 | { | |
196 | if (!req->rq_cong) | |
197 | return; | |
198 | req->rq_cong = 0; | |
199 | xprt->cong -= RPC_CWNDSCALE; | |
200 | __xprt_lock_write_next(xprt); | |
201 | } | |
202 | ||
203 | /* | |
204 | * Adjust RPC congestion window | |
205 | * We use a time-smoothed congestion estimator to avoid heavy oscillation. | |
206 | */ | |
207 | static void | |
208 | xprt_adjust_cwnd(struct rpc_xprt *xprt, int result) | |
209 | { | |
210 | unsigned long cwnd; | |
211 | ||
212 | cwnd = xprt->cwnd; | |
213 | if (result >= 0 && cwnd <= xprt->cong) { | |
214 | /* The (cwnd >> 1) term makes sure | |
215 | * the result gets rounded properly. */ | |
216 | cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd; | |
217 | if (cwnd > RPC_MAXCWND(xprt)) | |
218 | cwnd = RPC_MAXCWND(xprt); | |
219 | __xprt_lock_write_next(xprt); | |
220 | } else if (result == -ETIMEDOUT) { | |
221 | cwnd >>= 1; | |
222 | if (cwnd < RPC_CWNDSCALE) | |
223 | cwnd = RPC_CWNDSCALE; | |
224 | } | |
225 | dprintk("RPC: cong %ld, cwnd was %ld, now %ld\n", | |
226 | xprt->cong, xprt->cwnd, cwnd); | |
227 | xprt->cwnd = cwnd; | |
228 | } | |
229 | ||
44fbac22 CL |
230 | /** |
231 | * xprt_wake_pending_tasks - wake all tasks on a transport's pending queue | |
232 | * @xprt: transport with waiting tasks | |
233 | * @status: result code to plant in each task before waking it | |
234 | * | |
235 | */ | |
236 | void xprt_wake_pending_tasks(struct rpc_xprt *xprt, int status) | |
237 | { | |
238 | if (status < 0) | |
239 | rpc_wake_up_status(&xprt->pending, status); | |
240 | else | |
241 | rpc_wake_up(&xprt->pending); | |
242 | } | |
243 | ||
c7b2cae8 CL |
244 | /** |
245 | * xprt_wait_for_buffer_space - wait for transport output buffer to clear | |
246 | * @task: task to be put to sleep | |
247 | * | |
248 | */ | |
249 | void xprt_wait_for_buffer_space(struct rpc_task *task) | |
250 | { | |
251 | struct rpc_rqst *req = task->tk_rqstp; | |
252 | struct rpc_xprt *xprt = req->rq_xprt; | |
253 | ||
254 | task->tk_timeout = req->rq_timeout; | |
255 | rpc_sleep_on(&xprt->pending, task, NULL, NULL); | |
256 | } | |
257 | ||
258 | /** | |
259 | * xprt_write_space - wake the task waiting for transport output buffer space | |
260 | * @xprt: transport with waiting tasks | |
261 | * | |
262 | * Can be called in a soft IRQ context, so xprt_write_space never sleeps. | |
263 | */ | |
264 | void xprt_write_space(struct rpc_xprt *xprt) | |
265 | { | |
266 | if (unlikely(xprt->shutdown)) | |
267 | return; | |
268 | ||
269 | spin_lock_bh(&xprt->transport_lock); | |
270 | if (xprt->snd_task) { | |
271 | dprintk("RPC: write space: waking waiting task on xprt %p\n", | |
272 | xprt); | |
273 | rpc_wake_up_task(xprt->snd_task); | |
274 | } | |
275 | spin_unlock_bh(&xprt->transport_lock); | |
276 | } | |
277 | ||
1da177e4 LT |
278 | static void xprt_reset_majortimeo(struct rpc_rqst *req) |
279 | { | |
280 | struct rpc_timeout *to = &req->rq_xprt->timeout; | |
281 | ||
282 | req->rq_majortimeo = req->rq_timeout; | |
283 | if (to->to_exponential) | |
284 | req->rq_majortimeo <<= to->to_retries; | |
285 | else | |
286 | req->rq_majortimeo += to->to_increment * to->to_retries; | |
287 | if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0) | |
288 | req->rq_majortimeo = to->to_maxval; | |
289 | req->rq_majortimeo += jiffies; | |
290 | } | |
291 | ||
9903cd1c CL |
292 | /** |
293 | * xprt_adjust_timeout - adjust timeout values for next retransmit | |
294 | * @req: RPC request containing parameters to use for the adjustment | |
295 | * | |
1da177e4 LT |
296 | */ |
297 | int xprt_adjust_timeout(struct rpc_rqst *req) | |
298 | { | |
299 | struct rpc_xprt *xprt = req->rq_xprt; | |
300 | struct rpc_timeout *to = &xprt->timeout; | |
301 | int status = 0; | |
302 | ||
303 | if (time_before(jiffies, req->rq_majortimeo)) { | |
304 | if (to->to_exponential) | |
305 | req->rq_timeout <<= 1; | |
306 | else | |
307 | req->rq_timeout += to->to_increment; | |
308 | if (to->to_maxval && req->rq_timeout >= to->to_maxval) | |
309 | req->rq_timeout = to->to_maxval; | |
310 | req->rq_retries++; | |
311 | pprintk("RPC: %lu retrans\n", jiffies); | |
312 | } else { | |
313 | req->rq_timeout = to->to_initval; | |
314 | req->rq_retries = 0; | |
315 | xprt_reset_majortimeo(req); | |
316 | /* Reset the RTT counters == "slow start" */ | |
4a0f8c04 | 317 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 | 318 | rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval); |
4a0f8c04 | 319 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
320 | pprintk("RPC: %lu timeout\n", jiffies); |
321 | status = -ETIMEDOUT; | |
322 | } | |
323 | ||
324 | if (req->rq_timeout == 0) { | |
325 | printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n"); | |
326 | req->rq_timeout = 5 * HZ; | |
327 | } | |
328 | return status; | |
329 | } | |
330 | ||
55aa4f58 | 331 | static void xprt_autoclose(void *args) |
1da177e4 LT |
332 | { |
333 | struct rpc_xprt *xprt = (struct rpc_xprt *)args; | |
334 | ||
335 | xprt_disconnect(xprt); | |
a246b010 | 336 | xprt->ops->close(xprt); |
1da177e4 LT |
337 | xprt_release_write(xprt, NULL); |
338 | } | |
339 | ||
9903cd1c CL |
340 | /** |
341 | * xprt_disconnect - mark a transport as disconnected | |
342 | * @xprt: transport to flag for disconnect | |
343 | * | |
1da177e4 | 344 | */ |
a246b010 | 345 | void xprt_disconnect(struct rpc_xprt *xprt) |
1da177e4 LT |
346 | { |
347 | dprintk("RPC: disconnected transport %p\n", xprt); | |
4a0f8c04 | 348 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 | 349 | xprt_clear_connected(xprt); |
44fbac22 | 350 | xprt_wake_pending_tasks(xprt, -ENOTCONN); |
4a0f8c04 | 351 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
352 | } |
353 | ||
1da177e4 LT |
354 | static void |
355 | xprt_init_autodisconnect(unsigned long data) | |
356 | { | |
357 | struct rpc_xprt *xprt = (struct rpc_xprt *)data; | |
358 | ||
4a0f8c04 | 359 | spin_lock(&xprt->transport_lock); |
1da177e4 LT |
360 | if (!list_empty(&xprt->recv) || xprt->shutdown) |
361 | goto out_abort; | |
2226feb6 | 362 | if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) |
1da177e4 | 363 | goto out_abort; |
4a0f8c04 | 364 | spin_unlock(&xprt->transport_lock); |
2226feb6 | 365 | if (xprt_connecting(xprt)) |
1da177e4 LT |
366 | xprt_release_write(xprt, NULL); |
367 | else | |
368 | schedule_work(&xprt->task_cleanup); | |
369 | return; | |
370 | out_abort: | |
4a0f8c04 | 371 | spin_unlock(&xprt->transport_lock); |
1da177e4 LT |
372 | } |
373 | ||
9903cd1c CL |
374 | /** |
375 | * xprt_connect - schedule a transport connect operation | |
376 | * @task: RPC task that is requesting the connect | |
1da177e4 LT |
377 | * |
378 | */ | |
379 | void xprt_connect(struct rpc_task *task) | |
380 | { | |
381 | struct rpc_xprt *xprt = task->tk_xprt; | |
382 | ||
383 | dprintk("RPC: %4d xprt_connect xprt %p %s connected\n", task->tk_pid, | |
384 | xprt, (xprt_connected(xprt) ? "is" : "is not")); | |
385 | ||
386 | if (xprt->shutdown) { | |
387 | task->tk_status = -EIO; | |
388 | return; | |
389 | } | |
390 | if (!xprt->addr.sin_port) { | |
391 | task->tk_status = -EIO; | |
392 | return; | |
393 | } | |
394 | if (!xprt_lock_write(xprt, task)) | |
395 | return; | |
396 | if (xprt_connected(xprt)) | |
a246b010 CL |
397 | xprt_release_write(xprt, task); |
398 | else { | |
399 | if (task->tk_rqstp) | |
400 | task->tk_rqstp->rq_bytes_sent = 0; | |
1da177e4 | 401 | |
a246b010 CL |
402 | task->tk_timeout = RPC_CONNECT_TIMEOUT; |
403 | rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL); | |
404 | xprt->ops->connect(task); | |
1da177e4 LT |
405 | } |
406 | return; | |
1da177e4 LT |
407 | } |
408 | ||
9903cd1c | 409 | static void xprt_connect_status(struct rpc_task *task) |
1da177e4 LT |
410 | { |
411 | struct rpc_xprt *xprt = task->tk_xprt; | |
412 | ||
413 | if (task->tk_status >= 0) { | |
414 | dprintk("RPC: %4d xprt_connect_status: connection established\n", | |
415 | task->tk_pid); | |
416 | return; | |
417 | } | |
418 | ||
1da177e4 LT |
419 | switch (task->tk_status) { |
420 | case -ECONNREFUSED: | |
421 | case -ECONNRESET: | |
23475d66 CL |
422 | dprintk("RPC: %4d xprt_connect_status: server %s refused connection\n", |
423 | task->tk_pid, task->tk_client->cl_server); | |
424 | break; | |
1da177e4 | 425 | case -ENOTCONN: |
23475d66 CL |
426 | dprintk("RPC: %4d xprt_connect_status: connection broken\n", |
427 | task->tk_pid); | |
428 | break; | |
1da177e4 | 429 | case -ETIMEDOUT: |
23475d66 | 430 | dprintk("RPC: %4d xprt_connect_status: connect attempt timed out\n", |
1da177e4 LT |
431 | task->tk_pid); |
432 | break; | |
433 | default: | |
23475d66 CL |
434 | dprintk("RPC: %4d xprt_connect_status: error %d connecting to server %s\n", |
435 | task->tk_pid, -task->tk_status, task->tk_client->cl_server); | |
436 | xprt_release_write(xprt, task); | |
437 | task->tk_status = -EIO; | |
438 | return; | |
439 | } | |
440 | ||
441 | /* if soft mounted, just cause this RPC to fail */ | |
442 | if (RPC_IS_SOFT(task)) { | |
443 | xprt_release_write(xprt, task); | |
444 | task->tk_status = -EIO; | |
1da177e4 | 445 | } |
1da177e4 LT |
446 | } |
447 | ||
9903cd1c CL |
448 | /** |
449 | * xprt_lookup_rqst - find an RPC request corresponding to an XID | |
450 | * @xprt: transport on which the original request was transmitted | |
451 | * @xid: RPC XID of incoming reply | |
452 | * | |
1da177e4 | 453 | */ |
a246b010 | 454 | struct rpc_rqst *xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid) |
1da177e4 LT |
455 | { |
456 | struct list_head *pos; | |
457 | struct rpc_rqst *req = NULL; | |
458 | ||
459 | list_for_each(pos, &xprt->recv) { | |
460 | struct rpc_rqst *entry = list_entry(pos, struct rpc_rqst, rq_list); | |
461 | if (entry->rq_xid == xid) { | |
462 | req = entry; | |
463 | break; | |
464 | } | |
465 | } | |
466 | return req; | |
467 | } | |
468 | ||
9903cd1c CL |
469 | /** |
470 | * xprt_complete_rqst - called when reply processing is complete | |
471 | * @xprt: controlling transport | |
472 | * @req: RPC request that just completed | |
473 | * @copied: actual number of bytes received from the transport | |
474 | * | |
1da177e4 | 475 | */ |
a246b010 | 476 | void xprt_complete_rqst(struct rpc_xprt *xprt, struct rpc_rqst *req, int copied) |
1da177e4 LT |
477 | { |
478 | struct rpc_task *task = req->rq_task; | |
479 | struct rpc_clnt *clnt = task->tk_client; | |
480 | ||
481 | /* Adjust congestion window */ | |
482 | if (!xprt->nocong) { | |
483 | unsigned timer = task->tk_msg.rpc_proc->p_timer; | |
484 | xprt_adjust_cwnd(xprt, copied); | |
485 | __xprt_put_cong(xprt, req); | |
486 | if (timer) { | |
487 | if (req->rq_ntrans == 1) | |
488 | rpc_update_rtt(clnt->cl_rtt, timer, | |
489 | (long)jiffies - req->rq_xtime); | |
490 | rpc_set_timeo(clnt->cl_rtt, timer, req->rq_ntrans - 1); | |
491 | } | |
492 | } | |
493 | ||
494 | #ifdef RPC_PROFILE | |
495 | /* Profile only reads for now */ | |
496 | if (copied > 1024) { | |
497 | static unsigned long nextstat; | |
498 | static unsigned long pkt_rtt, pkt_len, pkt_cnt; | |
499 | ||
500 | pkt_cnt++; | |
501 | pkt_len += req->rq_slen + copied; | |
502 | pkt_rtt += jiffies - req->rq_xtime; | |
503 | if (time_before(nextstat, jiffies)) { | |
504 | printk("RPC: %lu %ld cwnd\n", jiffies, xprt->cwnd); | |
505 | printk("RPC: %ld %ld %ld %ld stat\n", | |
506 | jiffies, pkt_cnt, pkt_len, pkt_rtt); | |
507 | pkt_rtt = pkt_len = pkt_cnt = 0; | |
508 | nextstat = jiffies + 5 * HZ; | |
509 | } | |
510 | } | |
511 | #endif | |
512 | ||
513 | dprintk("RPC: %4d has input (%d bytes)\n", task->tk_pid, copied); | |
514 | list_del_init(&req->rq_list); | |
515 | req->rq_received = req->rq_private_buf.len = copied; | |
516 | ||
517 | /* ... and wake up the process. */ | |
518 | rpc_wake_up_task(task); | |
519 | return; | |
520 | } | |
521 | ||
1da177e4 LT |
522 | /* |
523 | * RPC receive timeout handler. | |
524 | */ | |
525 | static void | |
526 | xprt_timer(struct rpc_task *task) | |
527 | { | |
528 | struct rpc_rqst *req = task->tk_rqstp; | |
529 | struct rpc_xprt *xprt = req->rq_xprt; | |
530 | ||
4a0f8c04 | 531 | spin_lock(&xprt->transport_lock); |
1da177e4 LT |
532 | if (req->rq_received) |
533 | goto out; | |
534 | ||
535 | xprt_adjust_cwnd(req->rq_xprt, -ETIMEDOUT); | |
536 | __xprt_put_cong(xprt, req); | |
537 | ||
538 | dprintk("RPC: %4d xprt_timer (%s request)\n", | |
539 | task->tk_pid, req ? "pending" : "backlogged"); | |
540 | ||
541 | task->tk_status = -ETIMEDOUT; | |
542 | out: | |
543 | task->tk_timeout = 0; | |
544 | rpc_wake_up_task(task); | |
4a0f8c04 | 545 | spin_unlock(&xprt->transport_lock); |
1da177e4 LT |
546 | } |
547 | ||
9903cd1c CL |
548 | /** |
549 | * xprt_prepare_transmit - reserve the transport before sending a request | |
550 | * @task: RPC task about to send a request | |
551 | * | |
1da177e4 | 552 | */ |
9903cd1c | 553 | int xprt_prepare_transmit(struct rpc_task *task) |
1da177e4 LT |
554 | { |
555 | struct rpc_rqst *req = task->tk_rqstp; | |
556 | struct rpc_xprt *xprt = req->rq_xprt; | |
557 | int err = 0; | |
558 | ||
559 | dprintk("RPC: %4d xprt_prepare_transmit\n", task->tk_pid); | |
560 | ||
561 | if (xprt->shutdown) | |
562 | return -EIO; | |
563 | ||
4a0f8c04 | 564 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 LT |
565 | if (req->rq_received && !req->rq_bytes_sent) { |
566 | err = req->rq_received; | |
567 | goto out_unlock; | |
568 | } | |
569 | if (!__xprt_lock_write(xprt, task)) { | |
570 | err = -EAGAIN; | |
571 | goto out_unlock; | |
572 | } | |
573 | ||
574 | if (!xprt_connected(xprt)) { | |
575 | err = -ENOTCONN; | |
576 | goto out_unlock; | |
577 | } | |
578 | out_unlock: | |
4a0f8c04 | 579 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
580 | return err; |
581 | } | |
582 | ||
9903cd1c CL |
583 | /** |
584 | * xprt_transmit - send an RPC request on a transport | |
585 | * @task: controlling RPC task | |
586 | * | |
587 | * We have to copy the iovec because sendmsg fiddles with its contents. | |
588 | */ | |
589 | void xprt_transmit(struct rpc_task *task) | |
1da177e4 LT |
590 | { |
591 | struct rpc_clnt *clnt = task->tk_client; | |
592 | struct rpc_rqst *req = task->tk_rqstp; | |
593 | struct rpc_xprt *xprt = req->rq_xprt; | |
a246b010 | 594 | int status; |
1da177e4 LT |
595 | |
596 | dprintk("RPC: %4d xprt_transmit(%u)\n", task->tk_pid, req->rq_slen); | |
597 | ||
1da177e4 LT |
598 | smp_rmb(); |
599 | if (!req->rq_received) { | |
600 | if (list_empty(&req->rq_list)) { | |
4a0f8c04 | 601 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 LT |
602 | /* Update the softirq receive buffer */ |
603 | memcpy(&req->rq_private_buf, &req->rq_rcv_buf, | |
604 | sizeof(req->rq_private_buf)); | |
605 | /* Add request to the receive list */ | |
606 | list_add_tail(&req->rq_list, &xprt->recv); | |
4a0f8c04 | 607 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 | 608 | xprt_reset_majortimeo(req); |
0f9dc2b1 TM |
609 | /* Turn off autodisconnect */ |
610 | del_singleshot_timer_sync(&xprt->timer); | |
1da177e4 LT |
611 | } |
612 | } else if (!req->rq_bytes_sent) | |
613 | return; | |
614 | ||
a246b010 CL |
615 | status = xprt->ops->send_request(task); |
616 | if (!status) | |
617 | goto out_receive; | |
1da177e4 LT |
618 | |
619 | /* Note: at this point, task->tk_sleeping has not yet been set, | |
620 | * hence there is no danger of the waking up task being put on | |
621 | * schedq, and being picked up by a parallel run of rpciod(). | |
622 | */ | |
623 | task->tk_status = status; | |
624 | ||
625 | switch (status) { | |
1da177e4 LT |
626 | case -ECONNREFUSED: |
627 | task->tk_timeout = RPC_REESTABLISH_TIMEOUT; | |
628 | rpc_sleep_on(&xprt->sending, task, NULL, NULL); | |
a246b010 | 629 | case -EAGAIN: |
1da177e4 LT |
630 | case -ENOTCONN: |
631 | return; | |
632 | default: | |
43118c29 | 633 | break; |
1da177e4 LT |
634 | } |
635 | xprt_release_write(xprt, task); | |
636 | return; | |
637 | out_receive: | |
638 | dprintk("RPC: %4d xmit complete\n", task->tk_pid); | |
639 | /* Set the task's receive timeout value */ | |
4a0f8c04 | 640 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 LT |
641 | if (!xprt->nocong) { |
642 | int timer = task->tk_msg.rpc_proc->p_timer; | |
643 | task->tk_timeout = rpc_calc_rto(clnt->cl_rtt, timer); | |
644 | task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer) + req->rq_retries; | |
645 | if (task->tk_timeout > xprt->timeout.to_maxval || task->tk_timeout == 0) | |
646 | task->tk_timeout = xprt->timeout.to_maxval; | |
647 | } else | |
648 | task->tk_timeout = req->rq_timeout; | |
649 | /* Don't race with disconnect */ | |
650 | if (!xprt_connected(xprt)) | |
651 | task->tk_status = -ENOTCONN; | |
652 | else if (!req->rq_received) | |
653 | rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer); | |
654 | __xprt_release_write(xprt, task); | |
4a0f8c04 | 655 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
656 | } |
657 | ||
9903cd1c | 658 | static inline void do_xprt_reserve(struct rpc_task *task) |
1da177e4 LT |
659 | { |
660 | struct rpc_xprt *xprt = task->tk_xprt; | |
661 | ||
662 | task->tk_status = 0; | |
663 | if (task->tk_rqstp) | |
664 | return; | |
665 | if (!list_empty(&xprt->free)) { | |
666 | struct rpc_rqst *req = list_entry(xprt->free.next, struct rpc_rqst, rq_list); | |
667 | list_del_init(&req->rq_list); | |
668 | task->tk_rqstp = req; | |
669 | xprt_request_init(task, xprt); | |
670 | return; | |
671 | } | |
672 | dprintk("RPC: waiting for request slot\n"); | |
673 | task->tk_status = -EAGAIN; | |
674 | task->tk_timeout = 0; | |
675 | rpc_sleep_on(&xprt->backlog, task, NULL, NULL); | |
676 | } | |
677 | ||
9903cd1c CL |
678 | /** |
679 | * xprt_reserve - allocate an RPC request slot | |
680 | * @task: RPC task requesting a slot allocation | |
681 | * | |
682 | * If no more slots are available, place the task on the transport's | |
683 | * backlog queue. | |
684 | */ | |
685 | void xprt_reserve(struct rpc_task *task) | |
1da177e4 LT |
686 | { |
687 | struct rpc_xprt *xprt = task->tk_xprt; | |
688 | ||
689 | task->tk_status = -EIO; | |
690 | if (!xprt->shutdown) { | |
5dc07727 | 691 | spin_lock(&xprt->reserve_lock); |
1da177e4 | 692 | do_xprt_reserve(task); |
5dc07727 | 693 | spin_unlock(&xprt->reserve_lock); |
1da177e4 LT |
694 | } |
695 | } | |
696 | ||
1da177e4 LT |
697 | static inline u32 xprt_alloc_xid(struct rpc_xprt *xprt) |
698 | { | |
699 | return xprt->xid++; | |
700 | } | |
701 | ||
702 | static inline void xprt_init_xid(struct rpc_xprt *xprt) | |
703 | { | |
704 | get_random_bytes(&xprt->xid, sizeof(xprt->xid)); | |
705 | } | |
706 | ||
9903cd1c | 707 | static void xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt) |
1da177e4 LT |
708 | { |
709 | struct rpc_rqst *req = task->tk_rqstp; | |
710 | ||
711 | req->rq_timeout = xprt->timeout.to_initval; | |
712 | req->rq_task = task; | |
713 | req->rq_xprt = xprt; | |
714 | req->rq_xid = xprt_alloc_xid(xprt); | |
715 | dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid, | |
716 | req, ntohl(req->rq_xid)); | |
717 | } | |
718 | ||
9903cd1c CL |
719 | /** |
720 | * xprt_release - release an RPC request slot | |
721 | * @task: task which is finished with the slot | |
722 | * | |
1da177e4 | 723 | */ |
9903cd1c | 724 | void xprt_release(struct rpc_task *task) |
1da177e4 LT |
725 | { |
726 | struct rpc_xprt *xprt = task->tk_xprt; | |
727 | struct rpc_rqst *req; | |
728 | ||
729 | if (!(req = task->tk_rqstp)) | |
730 | return; | |
4a0f8c04 | 731 | spin_lock_bh(&xprt->transport_lock); |
1da177e4 LT |
732 | __xprt_release_write(xprt, task); |
733 | __xprt_put_cong(xprt, req); | |
734 | if (!list_empty(&req->rq_list)) | |
735 | list_del(&req->rq_list); | |
736 | xprt->last_used = jiffies; | |
737 | if (list_empty(&xprt->recv) && !xprt->shutdown) | |
a246b010 CL |
738 | mod_timer(&xprt->timer, |
739 | xprt->last_used + RPC_IDLE_DISCONNECT_TIMEOUT); | |
4a0f8c04 | 740 | spin_unlock_bh(&xprt->transport_lock); |
1da177e4 LT |
741 | task->tk_rqstp = NULL; |
742 | memset(req, 0, sizeof(*req)); /* mark unused */ | |
743 | ||
744 | dprintk("RPC: %4d release request %p\n", task->tk_pid, req); | |
745 | ||
5dc07727 | 746 | spin_lock(&xprt->reserve_lock); |
1da177e4 LT |
747 | list_add(&req->rq_list, &xprt->free); |
748 | xprt_clear_backlog(xprt); | |
5dc07727 | 749 | spin_unlock(&xprt->reserve_lock); |
1da177e4 LT |
750 | } |
751 | ||
9903cd1c CL |
752 | /** |
753 | * xprt_set_timeout - set constant RPC timeout | |
754 | * @to: RPC timeout parameters to set up | |
755 | * @retr: number of retries | |
756 | * @incr: amount of increase after each retry | |
757 | * | |
1da177e4 | 758 | */ |
9903cd1c | 759 | void xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr) |
1da177e4 LT |
760 | { |
761 | to->to_initval = | |
762 | to->to_increment = incr; | |
eab5c084 | 763 | to->to_maxval = to->to_initval + (incr * retr); |
1da177e4 LT |
764 | to->to_retries = retr; |
765 | to->to_exponential = 0; | |
766 | } | |
767 | ||
9903cd1c | 768 | static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to) |
1da177e4 | 769 | { |
a246b010 | 770 | int result; |
1da177e4 | 771 | struct rpc_xprt *xprt; |
1da177e4 LT |
772 | struct rpc_rqst *req; |
773 | ||
1da177e4 LT |
774 | if ((xprt = kmalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL) |
775 | return ERR_PTR(-ENOMEM); | |
776 | memset(xprt, 0, sizeof(*xprt)); /* Nnnngh! */ | |
1da177e4 LT |
777 | |
778 | xprt->addr = *ap; | |
a246b010 CL |
779 | |
780 | switch (proto) { | |
781 | case IPPROTO_UDP: | |
782 | result = xs_setup_udp(xprt, to); | |
783 | break; | |
784 | case IPPROTO_TCP: | |
785 | result = xs_setup_tcp(xprt, to); | |
786 | break; | |
787 | default: | |
788 | printk(KERN_ERR "RPC: unrecognized transport protocol: %d\n", | |
789 | proto); | |
790 | result = -EIO; | |
791 | break; | |
792 | } | |
793 | if (result) { | |
794 | kfree(xprt); | |
795 | return ERR_PTR(result); | |
1da177e4 | 796 | } |
a246b010 | 797 | |
4a0f8c04 | 798 | spin_lock_init(&xprt->transport_lock); |
5dc07727 | 799 | spin_lock_init(&xprt->reserve_lock); |
1da177e4 LT |
800 | init_waitqueue_head(&xprt->cong_wait); |
801 | ||
802 | INIT_LIST_HEAD(&xprt->free); | |
803 | INIT_LIST_HEAD(&xprt->recv); | |
55aa4f58 | 804 | INIT_WORK(&xprt->task_cleanup, xprt_autoclose, xprt); |
1da177e4 LT |
805 | init_timer(&xprt->timer); |
806 | xprt->timer.function = xprt_init_autodisconnect; | |
807 | xprt->timer.data = (unsigned long) xprt; | |
808 | xprt->last_used = jiffies; | |
1da177e4 LT |
809 | |
810 | rpc_init_wait_queue(&xprt->pending, "xprt_pending"); | |
811 | rpc_init_wait_queue(&xprt->sending, "xprt_sending"); | |
812 | rpc_init_wait_queue(&xprt->resend, "xprt_resend"); | |
813 | rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog"); | |
814 | ||
815 | /* initialize free list */ | |
a246b010 | 816 | for (req = &xprt->slot[xprt->max_reqs-1]; req >= &xprt->slot[0]; req--) |
1da177e4 LT |
817 | list_add(&req->rq_list, &xprt->free); |
818 | ||
819 | xprt_init_xid(xprt); | |
820 | ||
1da177e4 LT |
821 | dprintk("RPC: created transport %p with %u slots\n", xprt, |
822 | xprt->max_reqs); | |
823 | ||
824 | return xprt; | |
825 | } | |
826 | ||
9903cd1c CL |
827 | /** |
828 | * xprt_create_proto - create an RPC client transport | |
829 | * @proto: requested transport protocol | |
830 | * @sap: remote peer's address | |
831 | * @to: timeout parameters for new transport | |
832 | * | |
1da177e4 | 833 | */ |
9903cd1c | 834 | struct rpc_xprt *xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to) |
1da177e4 LT |
835 | { |
836 | struct rpc_xprt *xprt; | |
837 | ||
838 | xprt = xprt_setup(proto, sap, to); | |
839 | if (IS_ERR(xprt)) | |
840 | dprintk("RPC: xprt_create_proto failed\n"); | |
841 | else | |
842 | dprintk("RPC: xprt_create_proto created xprt %p\n", xprt); | |
843 | return xprt; | |
844 | } | |
845 | ||
9903cd1c | 846 | static void xprt_shutdown(struct rpc_xprt *xprt) |
1da177e4 LT |
847 | { |
848 | xprt->shutdown = 1; | |
849 | rpc_wake_up(&xprt->sending); | |
850 | rpc_wake_up(&xprt->resend); | |
44fbac22 | 851 | xprt_wake_pending_tasks(xprt, -EIO); |
1da177e4 LT |
852 | rpc_wake_up(&xprt->backlog); |
853 | wake_up(&xprt->cong_wait); | |
854 | del_timer_sync(&xprt->timer); | |
855 | } | |
856 | ||
9903cd1c | 857 | static int xprt_clear_backlog(struct rpc_xprt *xprt) { |
1da177e4 LT |
858 | rpc_wake_up_next(&xprt->backlog); |
859 | wake_up(&xprt->cong_wait); | |
860 | return 1; | |
861 | } | |
862 | ||
9903cd1c CL |
863 | /** |
864 | * xprt_destroy - destroy an RPC transport, killing off all requests. | |
865 | * @xprt: transport to destroy | |
866 | * | |
1da177e4 | 867 | */ |
9903cd1c | 868 | int xprt_destroy(struct rpc_xprt *xprt) |
1da177e4 LT |
869 | { |
870 | dprintk("RPC: destroying transport %p\n", xprt); | |
871 | xprt_shutdown(xprt); | |
a246b010 | 872 | xprt->ops->destroy(xprt); |
1da177e4 LT |
873 | kfree(xprt); |
874 | ||
875 | return 0; | |
876 | } |