1 /*
2 * GPL HEADER START
3 *
4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
5 *
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License version 2 only,
8 * as published by the Free Software Foundation.
9 *
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
15 *
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
19 *
20 * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
21 * CA 95054 USA or visit www.sun.com if you need additional information or
22 * have any questions.
23 *
24 * GPL HEADER END
25 */
26 /*
27 * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
28 * Use is subject to license terms.
29 *
30 * Copyright (c) 2011, 2012, Intel Corporation.
31 */
32 /*
33 * This file is part of Lustre, http://www.lustre.org/
34 * Lustre is a trademark of Sun Microsystems, Inc.
35 *
36 * Client IO.
37 *
38 * Author: Nikita Danilov <nikita.danilov@sun.com>
39 */
40
41 #define DEBUG_SUBSYSTEM S_CLASS
42
43 #include "../include/obd_class.h"
44 #include "../include/obd_support.h"
45 #include "../include/lustre_fid.h"
46 #include <linux/list.h>
47 #include "../include/cl_object.h"
48 #include "cl_internal.h"
49
50 /*****************************************************************************
51 *
52 * cl_io interface.
53 *
54 */
55
56 #define cl_io_for_each(slice, io) \
57 list_for_each_entry((slice), &io->ci_layers, cis_linkage)
58 #define cl_io_for_each_reverse(slice, io) \
59 list_for_each_entry_reverse((slice), &io->ci_layers, cis_linkage)
60
cl_io_type_is_valid(enum cl_io_type type)61 static inline int cl_io_type_is_valid(enum cl_io_type type)
62 {
63 return CIT_READ <= type && type < CIT_OP_NR;
64 }
65
cl_io_is_loopable(const struct cl_io * io)66 static inline int cl_io_is_loopable(const struct cl_io *io)
67 {
68 return cl_io_type_is_valid(io->ci_type) && io->ci_type != CIT_MISC;
69 }
70
71 /**
72 * Returns true iff there is an IO ongoing in the given environment.
73 */
cl_io_is_going(const struct lu_env * env)74 int cl_io_is_going(const struct lu_env *env)
75 {
76 return cl_env_info(env)->clt_current_io != NULL;
77 }
78 EXPORT_SYMBOL(cl_io_is_going);
79
80 /**
81 * cl_io invariant that holds at all times when exported cl_io_*() functions
82 * are entered and left.
83 */
cl_io_invariant(const struct cl_io * io)84 static int cl_io_invariant(const struct cl_io *io)
85 {
86 struct cl_io *up;
87
88 up = io->ci_parent;
89 return
90 /*
91 * io can own pages only when it is ongoing. Sub-io might
92 * still be in CIS_LOCKED state when top-io is in
93 * CIS_IO_GOING.
94 */
95 ergo(io->ci_owned_nr > 0, io->ci_state == CIS_IO_GOING ||
96 (io->ci_state == CIS_LOCKED && up != NULL));
97 }
98
99 /**
100 * Finalize \a io, by calling cl_io_operations::cio_fini() bottom-to-top.
101 */
cl_io_fini(const struct lu_env * env,struct cl_io * io)102 void cl_io_fini(const struct lu_env *env, struct cl_io *io)
103 {
104 struct cl_io_slice *slice;
105 struct cl_thread_info *info;
106
107 LINVRNT(cl_io_type_is_valid(io->ci_type));
108 LINVRNT(cl_io_invariant(io));
109
110 while (!list_empty(&io->ci_layers)) {
111 slice = container_of(io->ci_layers.prev, struct cl_io_slice,
112 cis_linkage);
113 list_del_init(&slice->cis_linkage);
114 if (slice->cis_iop->op[io->ci_type].cio_fini != NULL)
115 slice->cis_iop->op[io->ci_type].cio_fini(env, slice);
116 /*
117 * Invalidate slice to catch use after free. This assumes that
118 * slices are allocated within session and can be touched
119 * after ->cio_fini() returns.
120 */
121 slice->cis_io = NULL;
122 }
123 io->ci_state = CIS_FINI;
124 info = cl_env_info(env);
125 if (info->clt_current_io == io)
126 info->clt_current_io = NULL;
127
128 /* sanity check for layout change */
129 switch (io->ci_type) {
130 case CIT_READ:
131 case CIT_WRITE:
132 break;
133 case CIT_FAULT:
134 case CIT_FSYNC:
135 LASSERT(!io->ci_need_restart);
136 break;
137 case CIT_SETATTR:
138 case CIT_MISC:
139 /* Check ignore layout change conf */
140 LASSERT(ergo(io->ci_ignore_layout || !io->ci_verify_layout,
141 !io->ci_need_restart));
142 break;
143 default:
144 LBUG();
145 }
146 }
147 EXPORT_SYMBOL(cl_io_fini);
148
cl_io_init0(const struct lu_env * env,struct cl_io * io,enum cl_io_type iot,struct cl_object * obj)149 static int cl_io_init0(const struct lu_env *env, struct cl_io *io,
150 enum cl_io_type iot, struct cl_object *obj)
151 {
152 struct cl_object *scan;
153 int result;
154
155 LINVRNT(io->ci_state == CIS_ZERO || io->ci_state == CIS_FINI);
156 LINVRNT(cl_io_type_is_valid(iot));
157 LINVRNT(cl_io_invariant(io));
158
159 io->ci_type = iot;
160 INIT_LIST_HEAD(&io->ci_lockset.cls_todo);
161 INIT_LIST_HEAD(&io->ci_lockset.cls_curr);
162 INIT_LIST_HEAD(&io->ci_lockset.cls_done);
163 INIT_LIST_HEAD(&io->ci_layers);
164
165 result = 0;
166 cl_object_for_each(scan, obj) {
167 if (scan->co_ops->coo_io_init != NULL) {
168 result = scan->co_ops->coo_io_init(env, scan, io);
169 if (result != 0)
170 break;
171 }
172 }
173 if (result == 0)
174 io->ci_state = CIS_INIT;
175 return result;
176 }
177
178 /**
179 * Initialize sub-io, by calling cl_io_operations::cio_init() top-to-bottom.
180 *
181 * \pre obj != cl_object_top(obj)
182 */
cl_io_sub_init(const struct lu_env * env,struct cl_io * io,enum cl_io_type iot,struct cl_object * obj)183 int cl_io_sub_init(const struct lu_env *env, struct cl_io *io,
184 enum cl_io_type iot, struct cl_object *obj)
185 {
186 struct cl_thread_info *info = cl_env_info(env);
187
188 LASSERT(obj != cl_object_top(obj));
189 if (info->clt_current_io == NULL)
190 info->clt_current_io = io;
191 return cl_io_init0(env, io, iot, obj);
192 }
193 EXPORT_SYMBOL(cl_io_sub_init);
194
195 /**
196 * Initialize \a io, by calling cl_io_operations::cio_init() top-to-bottom.
197 *
198 * Caller has to call cl_io_fini() after a call to cl_io_init(), no matter
199 * what the latter returned.
200 *
201 * \pre obj == cl_object_top(obj)
202 * \pre cl_io_type_is_valid(iot)
203 * \post cl_io_type_is_valid(io->ci_type) && io->ci_type == iot
204 */
cl_io_init(const struct lu_env * env,struct cl_io * io,enum cl_io_type iot,struct cl_object * obj)205 int cl_io_init(const struct lu_env *env, struct cl_io *io,
206 enum cl_io_type iot, struct cl_object *obj)
207 {
208 struct cl_thread_info *info = cl_env_info(env);
209
210 LASSERT(obj == cl_object_top(obj));
211 LASSERT(info->clt_current_io == NULL);
212
213 info->clt_current_io = io;
214 return cl_io_init0(env, io, iot, obj);
215 }
216 EXPORT_SYMBOL(cl_io_init);
217
218 /**
219 * Initialize read or write io.
220 *
221 * \pre iot == CIT_READ || iot == CIT_WRITE
222 */
cl_io_rw_init(const struct lu_env * env,struct cl_io * io,enum cl_io_type iot,loff_t pos,size_t count)223 int cl_io_rw_init(const struct lu_env *env, struct cl_io *io,
224 enum cl_io_type iot, loff_t pos, size_t count)
225 {
226 LINVRNT(iot == CIT_READ || iot == CIT_WRITE);
227 LINVRNT(io->ci_obj != NULL);
228
229 LU_OBJECT_HEADER(D_VFSTRACE, env, &io->ci_obj->co_lu,
230 "io range: %u [%llu, %llu) %u %u\n",
231 iot, (__u64)pos, (__u64)pos + count,
232 io->u.ci_rw.crw_nonblock, io->u.ci_wr.wr_append);
233 io->u.ci_rw.crw_pos = pos;
234 io->u.ci_rw.crw_count = count;
235 return cl_io_init(env, io, iot, io->ci_obj);
236 }
237 EXPORT_SYMBOL(cl_io_rw_init);
238
239 static inline const struct lu_fid *
cl_lock_descr_fid(const struct cl_lock_descr * descr)240 cl_lock_descr_fid(const struct cl_lock_descr *descr)
241 {
242 return lu_object_fid(&descr->cld_obj->co_lu);
243 }
244
cl_lock_descr_sort(const struct cl_lock_descr * d0,const struct cl_lock_descr * d1)245 static int cl_lock_descr_sort(const struct cl_lock_descr *d0,
246 const struct cl_lock_descr *d1)
247 {
248 return lu_fid_cmp(cl_lock_descr_fid(d0), cl_lock_descr_fid(d1)) ?:
249 __diff_normalize(d0->cld_start, d1->cld_start);
250 }
251
cl_lock_descr_cmp(const struct cl_lock_descr * d0,const struct cl_lock_descr * d1)252 static int cl_lock_descr_cmp(const struct cl_lock_descr *d0,
253 const struct cl_lock_descr *d1)
254 {
255 int ret;
256
257 ret = lu_fid_cmp(cl_lock_descr_fid(d0), cl_lock_descr_fid(d1));
258 if (ret)
259 return ret;
260 if (d0->cld_end < d1->cld_start)
261 return -1;
262 if (d0->cld_start > d0->cld_end)
263 return 1;
264 return 0;
265 }
266
cl_lock_descr_merge(struct cl_lock_descr * d0,const struct cl_lock_descr * d1)267 static void cl_lock_descr_merge(struct cl_lock_descr *d0,
268 const struct cl_lock_descr *d1)
269 {
270 d0->cld_start = min(d0->cld_start, d1->cld_start);
271 d0->cld_end = max(d0->cld_end, d1->cld_end);
272
273 if (d1->cld_mode == CLM_WRITE && d0->cld_mode != CLM_WRITE)
274 d0->cld_mode = CLM_WRITE;
275
276 if (d1->cld_mode == CLM_GROUP && d0->cld_mode != CLM_GROUP)
277 d0->cld_mode = CLM_GROUP;
278 }
279
280 /*
281 * Sort locks in lexicographical order of their (fid, start-offset) pairs.
282 */
cl_io_locks_sort(struct cl_io * io)283 static void cl_io_locks_sort(struct cl_io *io)
284 {
285 int done = 0;
286
287 /* hidden treasure: bubble sort for now. */
288 do {
289 struct cl_io_lock_link *curr;
290 struct cl_io_lock_link *prev;
291 struct cl_io_lock_link *temp;
292
293 done = 1;
294 prev = NULL;
295
296 list_for_each_entry_safe(curr, temp,
297 &io->ci_lockset.cls_todo,
298 cill_linkage) {
299 if (prev != NULL) {
300 switch (cl_lock_descr_sort(&prev->cill_descr,
301 &curr->cill_descr)) {
302 case 0:
303 /*
304 * IMPOSSIBLE: Identical locks are
305 * already removed at
306 * this point.
307 */
308 default:
309 LBUG();
310 case +1:
311 list_move_tail(&curr->cill_linkage,
312 &prev->cill_linkage);
313 done = 0;
314 continue; /* don't change prev: it's
315 * still "previous" */
316 case -1: /* already in order */
317 break;
318 }
319 }
320 prev = curr;
321 }
322 } while (!done);
323 }
324
325 /**
326 * Check whether \a queue contains locks matching \a need.
327 *
328 * \retval +ve there is a matching lock in the \a queue
329 * \retval 0 there are no matching locks in the \a queue
330 */
cl_queue_match(const struct list_head * queue,const struct cl_lock_descr * need)331 int cl_queue_match(const struct list_head *queue,
332 const struct cl_lock_descr *need)
333 {
334 struct cl_io_lock_link *scan;
335
336 list_for_each_entry(scan, queue, cill_linkage) {
337 if (cl_lock_descr_match(&scan->cill_descr, need))
338 return +1;
339 }
340 return 0;
341 }
342 EXPORT_SYMBOL(cl_queue_match);
343
cl_queue_merge(const struct list_head * queue,const struct cl_lock_descr * need)344 static int cl_queue_merge(const struct list_head *queue,
345 const struct cl_lock_descr *need)
346 {
347 struct cl_io_lock_link *scan;
348
349 list_for_each_entry(scan, queue, cill_linkage) {
350 if (cl_lock_descr_cmp(&scan->cill_descr, need))
351 continue;
352 cl_lock_descr_merge(&scan->cill_descr, need);
353 CDEBUG(D_VFSTRACE, "lock: %d: [%lu, %lu]\n",
354 scan->cill_descr.cld_mode, scan->cill_descr.cld_start,
355 scan->cill_descr.cld_end);
356 return +1;
357 }
358 return 0;
359
360 }
361
cl_lockset_match(const struct cl_lockset * set,const struct cl_lock_descr * need)362 static int cl_lockset_match(const struct cl_lockset *set,
363 const struct cl_lock_descr *need)
364 {
365 return cl_queue_match(&set->cls_curr, need) ||
366 cl_queue_match(&set->cls_done, need);
367 }
368
cl_lockset_merge(const struct cl_lockset * set,const struct cl_lock_descr * need)369 static int cl_lockset_merge(const struct cl_lockset *set,
370 const struct cl_lock_descr *need)
371 {
372 return cl_queue_merge(&set->cls_todo, need) ||
373 cl_lockset_match(set, need);
374 }
375
cl_lockset_lock_one(const struct lu_env * env,struct cl_io * io,struct cl_lockset * set,struct cl_io_lock_link * link)376 static int cl_lockset_lock_one(const struct lu_env *env,
377 struct cl_io *io, struct cl_lockset *set,
378 struct cl_io_lock_link *link)
379 {
380 struct cl_lock *lock;
381 int result;
382
383 lock = cl_lock_request(env, io, &link->cill_descr, "io", io);
384
385 if (!IS_ERR(lock)) {
386 link->cill_lock = lock;
387 list_move(&link->cill_linkage, &set->cls_curr);
388 if (!(link->cill_descr.cld_enq_flags & CEF_ASYNC)) {
389 result = cl_wait(env, lock);
390 if (result == 0)
391 list_move(&link->cill_linkage,
392 &set->cls_done);
393 } else
394 result = 0;
395 } else
396 result = PTR_ERR(lock);
397 return result;
398 }
399
cl_lock_link_fini(const struct lu_env * env,struct cl_io * io,struct cl_io_lock_link * link)400 static void cl_lock_link_fini(const struct lu_env *env, struct cl_io *io,
401 struct cl_io_lock_link *link)
402 {
403 struct cl_lock *lock = link->cill_lock;
404
405 list_del_init(&link->cill_linkage);
406 if (lock != NULL) {
407 cl_lock_release(env, lock, "io", io);
408 link->cill_lock = NULL;
409 }
410 if (link->cill_fini != NULL)
411 link->cill_fini(env, link);
412 }
413
cl_lockset_lock(const struct lu_env * env,struct cl_io * io,struct cl_lockset * set)414 static int cl_lockset_lock(const struct lu_env *env, struct cl_io *io,
415 struct cl_lockset *set)
416 {
417 struct cl_io_lock_link *link;
418 struct cl_io_lock_link *temp;
419 struct cl_lock *lock;
420 int result;
421
422 result = 0;
423 list_for_each_entry_safe(link, temp, &set->cls_todo, cill_linkage) {
424 if (!cl_lockset_match(set, &link->cill_descr)) {
425 /* XXX some locking to guarantee that locks aren't
426 * expanded in between. */
427 result = cl_lockset_lock_one(env, io, set, link);
428 if (result != 0)
429 break;
430 } else
431 cl_lock_link_fini(env, io, link);
432 }
433 if (result == 0) {
434 list_for_each_entry_safe(link, temp,
435 &set->cls_curr, cill_linkage) {
436 lock = link->cill_lock;
437 result = cl_wait(env, lock);
438 if (result == 0)
439 list_move(&link->cill_linkage,
440 &set->cls_done);
441 else
442 break;
443 }
444 }
445 return result;
446 }
447
448 /**
449 * Takes locks necessary for the current iteration of io.
450 *
451 * Calls cl_io_operations::cio_lock() top-to-bottom to collect locks required
452 * by layers for the current iteration. Then sort locks (to avoid dead-locks),
453 * and acquire them.
454 */
cl_io_lock(const struct lu_env * env,struct cl_io * io)455 int cl_io_lock(const struct lu_env *env, struct cl_io *io)
456 {
457 const struct cl_io_slice *scan;
458 int result = 0;
459
460 LINVRNT(cl_io_is_loopable(io));
461 LINVRNT(io->ci_state == CIS_IT_STARTED);
462 LINVRNT(cl_io_invariant(io));
463
464 cl_io_for_each(scan, io) {
465 if (scan->cis_iop->op[io->ci_type].cio_lock == NULL)
466 continue;
467 result = scan->cis_iop->op[io->ci_type].cio_lock(env, scan);
468 if (result != 0)
469 break;
470 }
471 if (result == 0) {
472 cl_io_locks_sort(io);
473 result = cl_lockset_lock(env, io, &io->ci_lockset);
474 }
475 if (result != 0)
476 cl_io_unlock(env, io);
477 else
478 io->ci_state = CIS_LOCKED;
479 return result;
480 }
481 EXPORT_SYMBOL(cl_io_lock);
482
483 /**
484 * Release locks takes by io.
485 */
cl_io_unlock(const struct lu_env * env,struct cl_io * io)486 void cl_io_unlock(const struct lu_env *env, struct cl_io *io)
487 {
488 struct cl_lockset *set;
489 struct cl_io_lock_link *link;
490 struct cl_io_lock_link *temp;
491 const struct cl_io_slice *scan;
492
493 LASSERT(cl_io_is_loopable(io));
494 LASSERT(CIS_IT_STARTED <= io->ci_state && io->ci_state < CIS_UNLOCKED);
495 LINVRNT(cl_io_invariant(io));
496
497 set = &io->ci_lockset;
498
499 list_for_each_entry_safe(link, temp, &set->cls_todo, cill_linkage)
500 cl_lock_link_fini(env, io, link);
501
502 list_for_each_entry_safe(link, temp, &set->cls_curr, cill_linkage)
503 cl_lock_link_fini(env, io, link);
504
505 list_for_each_entry_safe(link, temp, &set->cls_done, cill_linkage) {
506 cl_unuse(env, link->cill_lock);
507 cl_lock_link_fini(env, io, link);
508 }
509 cl_io_for_each_reverse(scan, io) {
510 if (scan->cis_iop->op[io->ci_type].cio_unlock != NULL)
511 scan->cis_iop->op[io->ci_type].cio_unlock(env, scan);
512 }
513 io->ci_state = CIS_UNLOCKED;
514 LASSERT(!cl_env_info(env)->clt_counters[CNL_TOP].ctc_nr_locks_acquired);
515 }
516 EXPORT_SYMBOL(cl_io_unlock);
517
518 /**
519 * Prepares next iteration of io.
520 *
521 * Calls cl_io_operations::cio_iter_init() top-to-bottom. This exists to give
522 * layers a chance to modify io parameters, e.g., so that lov can restrict io
523 * to a single stripe.
524 */
cl_io_iter_init(const struct lu_env * env,struct cl_io * io)525 int cl_io_iter_init(const struct lu_env *env, struct cl_io *io)
526 {
527 const struct cl_io_slice *scan;
528 int result;
529
530 LINVRNT(cl_io_is_loopable(io));
531 LINVRNT(io->ci_state == CIS_INIT || io->ci_state == CIS_IT_ENDED);
532 LINVRNT(cl_io_invariant(io));
533
534 result = 0;
535 cl_io_for_each(scan, io) {
536 if (scan->cis_iop->op[io->ci_type].cio_iter_init == NULL)
537 continue;
538 result = scan->cis_iop->op[io->ci_type].cio_iter_init(env,
539 scan);
540 if (result != 0)
541 break;
542 }
543 if (result == 0)
544 io->ci_state = CIS_IT_STARTED;
545 return result;
546 }
547 EXPORT_SYMBOL(cl_io_iter_init);
548
549 /**
550 * Finalizes io iteration.
551 *
552 * Calls cl_io_operations::cio_iter_fini() bottom-to-top.
553 */
cl_io_iter_fini(const struct lu_env * env,struct cl_io * io)554 void cl_io_iter_fini(const struct lu_env *env, struct cl_io *io)
555 {
556 const struct cl_io_slice *scan;
557
558 LINVRNT(cl_io_is_loopable(io));
559 LINVRNT(io->ci_state == CIS_UNLOCKED);
560 LINVRNT(cl_io_invariant(io));
561
562 cl_io_for_each_reverse(scan, io) {
563 if (scan->cis_iop->op[io->ci_type].cio_iter_fini != NULL)
564 scan->cis_iop->op[io->ci_type].cio_iter_fini(env, scan);
565 }
566 io->ci_state = CIS_IT_ENDED;
567 }
568 EXPORT_SYMBOL(cl_io_iter_fini);
569
570 /**
571 * Records that read or write io progressed \a nob bytes forward.
572 */
cl_io_rw_advance(const struct lu_env * env,struct cl_io * io,size_t nob)573 void cl_io_rw_advance(const struct lu_env *env, struct cl_io *io, size_t nob)
574 {
575 const struct cl_io_slice *scan;
576
577 LINVRNT(io->ci_type == CIT_READ || io->ci_type == CIT_WRITE ||
578 nob == 0);
579 LINVRNT(cl_io_is_loopable(io));
580 LINVRNT(cl_io_invariant(io));
581
582 io->u.ci_rw.crw_pos += nob;
583 io->u.ci_rw.crw_count -= nob;
584
585 /* layers have to be notified. */
586 cl_io_for_each_reverse(scan, io) {
587 if (scan->cis_iop->op[io->ci_type].cio_advance != NULL)
588 scan->cis_iop->op[io->ci_type].cio_advance(env, scan,
589 nob);
590 }
591 }
592 EXPORT_SYMBOL(cl_io_rw_advance);
593
594 /**
595 * Adds a lock to a lockset.
596 */
cl_io_lock_add(const struct lu_env * env,struct cl_io * io,struct cl_io_lock_link * link)597 int cl_io_lock_add(const struct lu_env *env, struct cl_io *io,
598 struct cl_io_lock_link *link)
599 {
600 int result;
601
602 if (cl_lockset_merge(&io->ci_lockset, &link->cill_descr))
603 result = +1;
604 else {
605 list_add(&link->cill_linkage, &io->ci_lockset.cls_todo);
606 result = 0;
607 }
608 return result;
609 }
610 EXPORT_SYMBOL(cl_io_lock_add);
611
cl_free_io_lock_link(const struct lu_env * env,struct cl_io_lock_link * link)612 static void cl_free_io_lock_link(const struct lu_env *env,
613 struct cl_io_lock_link *link)
614 {
615 OBD_FREE_PTR(link);
616 }
617
618 /**
619 * Allocates new lock link, and uses it to add a lock to a lockset.
620 */
cl_io_lock_alloc_add(const struct lu_env * env,struct cl_io * io,struct cl_lock_descr * descr)621 int cl_io_lock_alloc_add(const struct lu_env *env, struct cl_io *io,
622 struct cl_lock_descr *descr)
623 {
624 struct cl_io_lock_link *link;
625 int result;
626
627 OBD_ALLOC_PTR(link);
628 if (link != NULL) {
629 link->cill_descr = *descr;
630 link->cill_fini = cl_free_io_lock_link;
631 result = cl_io_lock_add(env, io, link);
632 if (result) /* lock match */
633 link->cill_fini(env, link);
634 } else
635 result = -ENOMEM;
636
637 return result;
638 }
639 EXPORT_SYMBOL(cl_io_lock_alloc_add);
640
641 /**
642 * Starts io by calling cl_io_operations::cio_start() top-to-bottom.
643 */
cl_io_start(const struct lu_env * env,struct cl_io * io)644 int cl_io_start(const struct lu_env *env, struct cl_io *io)
645 {
646 const struct cl_io_slice *scan;
647 int result = 0;
648
649 LINVRNT(cl_io_is_loopable(io));
650 LINVRNT(io->ci_state == CIS_LOCKED);
651 LINVRNT(cl_io_invariant(io));
652
653 io->ci_state = CIS_IO_GOING;
654 cl_io_for_each(scan, io) {
655 if (scan->cis_iop->op[io->ci_type].cio_start == NULL)
656 continue;
657 result = scan->cis_iop->op[io->ci_type].cio_start(env, scan);
658 if (result != 0)
659 break;
660 }
661 if (result >= 0)
662 result = 0;
663 return result;
664 }
665 EXPORT_SYMBOL(cl_io_start);
666
667 /**
668 * Wait until current io iteration is finished by calling
669 * cl_io_operations::cio_end() bottom-to-top.
670 */
cl_io_end(const struct lu_env * env,struct cl_io * io)671 void cl_io_end(const struct lu_env *env, struct cl_io *io)
672 {
673 const struct cl_io_slice *scan;
674
675 LINVRNT(cl_io_is_loopable(io));
676 LINVRNT(io->ci_state == CIS_IO_GOING);
677 LINVRNT(cl_io_invariant(io));
678
679 cl_io_for_each_reverse(scan, io) {
680 if (scan->cis_iop->op[io->ci_type].cio_end != NULL)
681 scan->cis_iop->op[io->ci_type].cio_end(env, scan);
682 /* TODO: error handling. */
683 }
684 io->ci_state = CIS_IO_FINISHED;
685 }
686 EXPORT_SYMBOL(cl_io_end);
687
688 static const struct cl_page_slice *
cl_io_slice_page(const struct cl_io_slice * ios,struct cl_page * page)689 cl_io_slice_page(const struct cl_io_slice *ios, struct cl_page *page)
690 {
691 const struct cl_page_slice *slice;
692
693 slice = cl_page_at(page, ios->cis_obj->co_lu.lo_dev->ld_type);
694 LINVRNT(slice != NULL);
695 return slice;
696 }
697
698 /**
699 * True iff \a page is within \a io range.
700 */
cl_page_in_io(const struct cl_page * page,const struct cl_io * io)701 static int cl_page_in_io(const struct cl_page *page, const struct cl_io *io)
702 {
703 int result = 1;
704 loff_t start;
705 loff_t end;
706 pgoff_t idx;
707
708 idx = page->cp_index;
709 switch (io->ci_type) {
710 case CIT_READ:
711 case CIT_WRITE:
712 /*
713 * check that [start, end) and [pos, pos + count) extents
714 * overlap.
715 */
716 if (!cl_io_is_append(io)) {
717 const struct cl_io_rw_common *crw = &(io->u.ci_rw);
718 start = cl_offset(page->cp_obj, idx);
719 end = cl_offset(page->cp_obj, idx + 1);
720 result = crw->crw_pos < end &&
721 start < crw->crw_pos + crw->crw_count;
722 }
723 break;
724 case CIT_FAULT:
725 result = io->u.ci_fault.ft_index == idx;
726 break;
727 default:
728 LBUG();
729 }
730 return result;
731 }
732
733 /**
734 * Called by read io, when page has to be read from the server.
735 *
736 * \see cl_io_operations::cio_read_page()
737 */
cl_io_read_page(const struct lu_env * env,struct cl_io * io,struct cl_page * page)738 int cl_io_read_page(const struct lu_env *env, struct cl_io *io,
739 struct cl_page *page)
740 {
741 const struct cl_io_slice *scan;
742 struct cl_2queue *queue;
743 int result = 0;
744
745 LINVRNT(io->ci_type == CIT_READ || io->ci_type == CIT_FAULT);
746 LINVRNT(cl_page_is_owned(page, io));
747 LINVRNT(io->ci_state == CIS_IO_GOING || io->ci_state == CIS_LOCKED);
748 LINVRNT(cl_page_in_io(page, io));
749 LINVRNT(cl_io_invariant(io));
750
751 queue = &io->ci_queue;
752
753 cl_2queue_init(queue);
754 /*
755 * ->cio_read_page() methods called in the loop below are supposed to
756 * never block waiting for network (the only subtle point is the
757 * creation of new pages for read-ahead that might result in cache
758 * shrinking, but currently only clean pages are shrunk and this
759 * requires no network io).
760 *
761 * Should this ever starts blocking, retry loop would be needed for
762 * "parallel io" (see CLO_REPEAT loops in cl_lock.c).
763 */
764 cl_io_for_each(scan, io) {
765 if (scan->cis_iop->cio_read_page != NULL) {
766 const struct cl_page_slice *slice;
767
768 slice = cl_io_slice_page(scan, page);
769 LINVRNT(slice != NULL);
770 result = scan->cis_iop->cio_read_page(env, scan, slice);
771 if (result != 0)
772 break;
773 }
774 }
775 if (result == 0)
776 result = cl_io_submit_rw(env, io, CRT_READ, queue);
777 /*
778 * Unlock unsent pages in case of error.
779 */
780 cl_page_list_disown(env, io, &queue->c2_qin);
781 cl_2queue_fini(env, queue);
782 return result;
783 }
784 EXPORT_SYMBOL(cl_io_read_page);
785
786 /**
787 * Called by write io to prepare page to receive data from user buffer.
788 *
789 * \see cl_io_operations::cio_prepare_write()
790 */
cl_io_prepare_write(const struct lu_env * env,struct cl_io * io,struct cl_page * page,unsigned from,unsigned to)791 int cl_io_prepare_write(const struct lu_env *env, struct cl_io *io,
792 struct cl_page *page, unsigned from, unsigned to)
793 {
794 const struct cl_io_slice *scan;
795 int result = 0;
796
797 LINVRNT(io->ci_type == CIT_WRITE);
798 LINVRNT(cl_page_is_owned(page, io));
799 LINVRNT(io->ci_state == CIS_IO_GOING || io->ci_state == CIS_LOCKED);
800 LINVRNT(cl_io_invariant(io));
801 LASSERT(cl_page_in_io(page, io));
802
803 cl_io_for_each_reverse(scan, io) {
804 if (scan->cis_iop->cio_prepare_write != NULL) {
805 const struct cl_page_slice *slice;
806
807 slice = cl_io_slice_page(scan, page);
808 result = scan->cis_iop->cio_prepare_write(env, scan,
809 slice,
810 from, to);
811 if (result != 0)
812 break;
813 }
814 }
815 return result;
816 }
817 EXPORT_SYMBOL(cl_io_prepare_write);
818
819 /**
820 * Called by write io after user data were copied into a page.
821 *
822 * \see cl_io_operations::cio_commit_write()
823 */
cl_io_commit_write(const struct lu_env * env,struct cl_io * io,struct cl_page * page,unsigned from,unsigned to)824 int cl_io_commit_write(const struct lu_env *env, struct cl_io *io,
825 struct cl_page *page, unsigned from, unsigned to)
826 {
827 const struct cl_io_slice *scan;
828 int result = 0;
829
830 LINVRNT(io->ci_type == CIT_WRITE);
831 LINVRNT(io->ci_state == CIS_IO_GOING || io->ci_state == CIS_LOCKED);
832 LINVRNT(cl_io_invariant(io));
833 /*
834 * XXX Uh... not nice. Top level cl_io_commit_write() call (vvp->lov)
835 * already called cl_page_cache_add(), moving page into CPS_CACHED
836 * state. Better (and more general) way of dealing with such situation
837 * is needed.
838 */
839 LASSERT(cl_page_is_owned(page, io) || page->cp_parent != NULL);
840 LASSERT(cl_page_in_io(page, io));
841
842 cl_io_for_each(scan, io) {
843 if (scan->cis_iop->cio_commit_write != NULL) {
844 const struct cl_page_slice *slice;
845
846 slice = cl_io_slice_page(scan, page);
847 result = scan->cis_iop->cio_commit_write(env, scan,
848 slice,
849 from, to);
850 if (result != 0)
851 break;
852 }
853 }
854 LINVRNT(result <= 0);
855 return result;
856 }
857 EXPORT_SYMBOL(cl_io_commit_write);
858
859 /**
860 * Submits a list of pages for immediate io.
861 *
862 * After the function gets returned, The submitted pages are moved to
863 * queue->c2_qout queue, and queue->c2_qin contain both the pages don't need
864 * to be submitted, and the pages are errant to submit.
865 *
866 * \returns 0 if at least one page was submitted, error code otherwise.
867 * \see cl_io_operations::cio_submit()
868 */
cl_io_submit_rw(const struct lu_env * env,struct cl_io * io,enum cl_req_type crt,struct cl_2queue * queue)869 int cl_io_submit_rw(const struct lu_env *env, struct cl_io *io,
870 enum cl_req_type crt, struct cl_2queue *queue)
871 {
872 const struct cl_io_slice *scan;
873 int result = 0;
874
875 LINVRNT(crt < ARRAY_SIZE(scan->cis_iop->req_op));
876
877 cl_io_for_each(scan, io) {
878 if (scan->cis_iop->req_op[crt].cio_submit == NULL)
879 continue;
880 result = scan->cis_iop->req_op[crt].cio_submit(env, scan, crt,
881 queue);
882 if (result != 0)
883 break;
884 }
885 /*
886 * If ->cio_submit() failed, no pages were sent.
887 */
888 LASSERT(ergo(result != 0, list_empty(&queue->c2_qout.pl_pages)));
889 return result;
890 }
891 EXPORT_SYMBOL(cl_io_submit_rw);
892
893 /**
894 * Submit a sync_io and wait for the IO to be finished, or error happens.
895 * If \a timeout is zero, it means to wait for the IO unconditionally.
896 */
cl_io_submit_sync(const struct lu_env * env,struct cl_io * io,enum cl_req_type iot,struct cl_2queue * queue,long timeout)897 int cl_io_submit_sync(const struct lu_env *env, struct cl_io *io,
898 enum cl_req_type iot, struct cl_2queue *queue,
899 long timeout)
900 {
901 struct cl_sync_io *anchor = &cl_env_info(env)->clt_anchor;
902 struct cl_page *pg;
903 int rc;
904
905 cl_page_list_for_each(pg, &queue->c2_qin) {
906 LASSERT(pg->cp_sync_io == NULL);
907 pg->cp_sync_io = anchor;
908 }
909
910 cl_sync_io_init(anchor, queue->c2_qin.pl_nr);
911 rc = cl_io_submit_rw(env, io, iot, queue);
912 if (rc == 0) {
913 /*
914 * If some pages weren't sent for any reason (e.g.,
915 * read found up-to-date pages in the cache, or write found
916 * clean pages), count them as completed to avoid infinite
917 * wait.
918 */
919 cl_page_list_for_each(pg, &queue->c2_qin) {
920 pg->cp_sync_io = NULL;
921 cl_sync_io_note(anchor, +1);
922 }
923
924 /* wait for the IO to be finished. */
925 rc = cl_sync_io_wait(env, io, &queue->c2_qout,
926 anchor, timeout);
927 } else {
928 LASSERT(list_empty(&queue->c2_qout.pl_pages));
929 cl_page_list_for_each(pg, &queue->c2_qin)
930 pg->cp_sync_io = NULL;
931 }
932 return rc;
933 }
934 EXPORT_SYMBOL(cl_io_submit_sync);
935
936 /**
937 * Cancel an IO which has been submitted by cl_io_submit_rw.
938 */
cl_io_cancel(const struct lu_env * env,struct cl_io * io,struct cl_page_list * queue)939 int cl_io_cancel(const struct lu_env *env, struct cl_io *io,
940 struct cl_page_list *queue)
941 {
942 struct cl_page *page;
943 int result = 0;
944
945 CERROR("Canceling ongoing page transmission\n");
946 cl_page_list_for_each(page, queue) {
947 int rc;
948
949 LINVRNT(cl_page_in_io(page, io));
950 rc = cl_page_cancel(env, page);
951 result = result ?: rc;
952 }
953 return result;
954 }
955 EXPORT_SYMBOL(cl_io_cancel);
956
957 /**
958 * Main io loop.
959 *
960 * Pumps io through iterations calling
961 *
962 * - cl_io_iter_init()
963 *
964 * - cl_io_lock()
965 *
966 * - cl_io_start()
967 *
968 * - cl_io_end()
969 *
970 * - cl_io_unlock()
971 *
972 * - cl_io_iter_fini()
973 *
974 * repeatedly until there is no more io to do.
975 */
cl_io_loop(const struct lu_env * env,struct cl_io * io)976 int cl_io_loop(const struct lu_env *env, struct cl_io *io)
977 {
978 int result = 0;
979
980 LINVRNT(cl_io_is_loopable(io));
981
982 do {
983 size_t nob;
984
985 io->ci_continue = 0;
986 result = cl_io_iter_init(env, io);
987 if (result == 0) {
988 nob = io->ci_nob;
989 result = cl_io_lock(env, io);
990 if (result == 0) {
991 /*
992 * Notify layers that locks has been taken,
993 * and do actual i/o.
994 *
995 * - llite: kms, short read;
996 * - llite: generic_file_read();
997 */
998 result = cl_io_start(env, io);
999 /*
1000 * Send any remaining pending
1001 * io, etc.
1002 *
1003 * - llite: ll_rw_stats_tally.
1004 */
1005 cl_io_end(env, io);
1006 cl_io_unlock(env, io);
1007 cl_io_rw_advance(env, io, io->ci_nob - nob);
1008 }
1009 }
1010 cl_io_iter_fini(env, io);
1011 } while (result == 0 && io->ci_continue);
1012 if (result == 0)
1013 result = io->ci_result;
1014 return result < 0 ? result : 0;
1015 }
1016 EXPORT_SYMBOL(cl_io_loop);
1017
1018 /**
1019 * Adds io slice to the cl_io.
1020 *
1021 * This is called by cl_object_operations::coo_io_init() methods to add a
1022 * per-layer state to the io. New state is added at the end of
1023 * cl_io::ci_layers list, that is, it is at the bottom of the stack.
1024 *
1025 * \see cl_lock_slice_add(), cl_req_slice_add(), cl_page_slice_add()
1026 */
cl_io_slice_add(struct cl_io * io,struct cl_io_slice * slice,struct cl_object * obj,const struct cl_io_operations * ops)1027 void cl_io_slice_add(struct cl_io *io, struct cl_io_slice *slice,
1028 struct cl_object *obj,
1029 const struct cl_io_operations *ops)
1030 {
1031 struct list_head *linkage = &slice->cis_linkage;
1032
1033 LASSERT((linkage->prev == NULL && linkage->next == NULL) ||
1034 list_empty(linkage));
1035
1036 list_add_tail(linkage, &io->ci_layers);
1037 slice->cis_io = io;
1038 slice->cis_obj = obj;
1039 slice->cis_iop = ops;
1040 }
1041 EXPORT_SYMBOL(cl_io_slice_add);
1042
1043
1044 /**
1045 * Initializes page list.
1046 */
cl_page_list_init(struct cl_page_list * plist)1047 void cl_page_list_init(struct cl_page_list *plist)
1048 {
1049 plist->pl_nr = 0;
1050 INIT_LIST_HEAD(&plist->pl_pages);
1051 plist->pl_owner = current;
1052 }
1053 EXPORT_SYMBOL(cl_page_list_init);
1054
1055 /**
1056 * Adds a page to a page list.
1057 */
cl_page_list_add(struct cl_page_list * plist,struct cl_page * page)1058 void cl_page_list_add(struct cl_page_list *plist, struct cl_page *page)
1059 {
1060 /* it would be better to check that page is owned by "current" io, but
1061 * it is not passed here. */
1062 LASSERT(page->cp_owner != NULL);
1063 LINVRNT(plist->pl_owner == current);
1064
1065 lockdep_off();
1066 mutex_lock(&page->cp_mutex);
1067 lockdep_on();
1068 LASSERT(list_empty(&page->cp_batch));
1069 list_add_tail(&page->cp_batch, &plist->pl_pages);
1070 ++plist->pl_nr;
1071 lu_ref_add_at(&page->cp_reference, &page->cp_queue_ref, "queue", plist);
1072 cl_page_get(page);
1073 }
1074 EXPORT_SYMBOL(cl_page_list_add);
1075
1076 /**
1077 * Removes a page from a page list.
1078 */
cl_page_list_del(const struct lu_env * env,struct cl_page_list * plist,struct cl_page * page)1079 void cl_page_list_del(const struct lu_env *env,
1080 struct cl_page_list *plist, struct cl_page *page)
1081 {
1082 LASSERT(plist->pl_nr > 0);
1083 LINVRNT(plist->pl_owner == current);
1084
1085 list_del_init(&page->cp_batch);
1086 lockdep_off();
1087 mutex_unlock(&page->cp_mutex);
1088 lockdep_on();
1089 --plist->pl_nr;
1090 lu_ref_del_at(&page->cp_reference, &page->cp_queue_ref, "queue", plist);
1091 cl_page_put(env, page);
1092 }
1093 EXPORT_SYMBOL(cl_page_list_del);
1094
1095 /**
1096 * Moves a page from one page list to another.
1097 */
cl_page_list_move(struct cl_page_list * dst,struct cl_page_list * src,struct cl_page * page)1098 void cl_page_list_move(struct cl_page_list *dst, struct cl_page_list *src,
1099 struct cl_page *page)
1100 {
1101 LASSERT(src->pl_nr > 0);
1102 LINVRNT(dst->pl_owner == current);
1103 LINVRNT(src->pl_owner == current);
1104
1105 list_move_tail(&page->cp_batch, &dst->pl_pages);
1106 --src->pl_nr;
1107 ++dst->pl_nr;
1108 lu_ref_set_at(&page->cp_reference, &page->cp_queue_ref, "queue",
1109 src, dst);
1110 }
1111 EXPORT_SYMBOL(cl_page_list_move);
1112
1113 /**
1114 * splice the cl_page_list, just as list head does
1115 */
cl_page_list_splice(struct cl_page_list * list,struct cl_page_list * head)1116 void cl_page_list_splice(struct cl_page_list *list, struct cl_page_list *head)
1117 {
1118 struct cl_page *page;
1119 struct cl_page *tmp;
1120
1121 LINVRNT(list->pl_owner == current);
1122 LINVRNT(head->pl_owner == current);
1123
1124 cl_page_list_for_each_safe(page, tmp, list)
1125 cl_page_list_move(head, list, page);
1126 }
1127 EXPORT_SYMBOL(cl_page_list_splice);
1128
1129 void cl_page_disown0(const struct lu_env *env,
1130 struct cl_io *io, struct cl_page *pg);
1131
1132 /**
1133 * Disowns pages in a queue.
1134 */
cl_page_list_disown(const struct lu_env * env,struct cl_io * io,struct cl_page_list * plist)1135 void cl_page_list_disown(const struct lu_env *env,
1136 struct cl_io *io, struct cl_page_list *plist)
1137 {
1138 struct cl_page *page;
1139 struct cl_page *temp;
1140
1141 LINVRNT(plist->pl_owner == current);
1142
1143 cl_page_list_for_each_safe(page, temp, plist) {
1144 LASSERT(plist->pl_nr > 0);
1145
1146 list_del_init(&page->cp_batch);
1147 lockdep_off();
1148 mutex_unlock(&page->cp_mutex);
1149 lockdep_on();
1150 --plist->pl_nr;
1151 /*
1152 * cl_page_disown0 rather than usual cl_page_disown() is used,
1153 * because pages are possibly in CPS_FREEING state already due
1154 * to the call to cl_page_list_discard().
1155 */
1156 /*
1157 * XXX cl_page_disown0() will fail if page is not locked.
1158 */
1159 cl_page_disown0(env, io, page);
1160 lu_ref_del_at(&page->cp_reference, &page->cp_queue_ref, "queue",
1161 plist);
1162 cl_page_put(env, page);
1163 }
1164 }
1165 EXPORT_SYMBOL(cl_page_list_disown);
1166
1167 /**
1168 * Releases pages from queue.
1169 */
cl_page_list_fini(const struct lu_env * env,struct cl_page_list * plist)1170 void cl_page_list_fini(const struct lu_env *env, struct cl_page_list *plist)
1171 {
1172 struct cl_page *page;
1173 struct cl_page *temp;
1174
1175 LINVRNT(plist->pl_owner == current);
1176
1177 cl_page_list_for_each_safe(page, temp, plist)
1178 cl_page_list_del(env, plist, page);
1179 LASSERT(plist->pl_nr == 0);
1180 }
1181 EXPORT_SYMBOL(cl_page_list_fini);
1182
1183 /**
1184 * Owns all pages in a queue.
1185 */
cl_page_list_own(const struct lu_env * env,struct cl_io * io,struct cl_page_list * plist)1186 int cl_page_list_own(const struct lu_env *env,
1187 struct cl_io *io, struct cl_page_list *plist)
1188 {
1189 struct cl_page *page;
1190 struct cl_page *temp;
1191 pgoff_t index = 0;
1192 int result;
1193
1194 LINVRNT(plist->pl_owner == current);
1195
1196 result = 0;
1197 cl_page_list_for_each_safe(page, temp, plist) {
1198 LASSERT(index <= page->cp_index);
1199 index = page->cp_index;
1200 if (cl_page_own(env, io, page) == 0)
1201 result = result ?: page->cp_error;
1202 else
1203 cl_page_list_del(env, plist, page);
1204 }
1205 return result;
1206 }
1207 EXPORT_SYMBOL(cl_page_list_own);
1208
1209 /**
1210 * Assumes all pages in a queue.
1211 */
cl_page_list_assume(const struct lu_env * env,struct cl_io * io,struct cl_page_list * plist)1212 void cl_page_list_assume(const struct lu_env *env,
1213 struct cl_io *io, struct cl_page_list *plist)
1214 {
1215 struct cl_page *page;
1216
1217 LINVRNT(plist->pl_owner == current);
1218
1219 cl_page_list_for_each(page, plist)
1220 cl_page_assume(env, io, page);
1221 }
1222 EXPORT_SYMBOL(cl_page_list_assume);
1223
1224 /**
1225 * Discards all pages in a queue.
1226 */
cl_page_list_discard(const struct lu_env * env,struct cl_io * io,struct cl_page_list * plist)1227 void cl_page_list_discard(const struct lu_env *env, struct cl_io *io,
1228 struct cl_page_list *plist)
1229 {
1230 struct cl_page *page;
1231
1232 LINVRNT(plist->pl_owner == current);
1233 cl_page_list_for_each(page, plist)
1234 cl_page_discard(env, io, page);
1235 }
1236 EXPORT_SYMBOL(cl_page_list_discard);
1237
1238 /**
1239 * Unmaps all pages in a queue from user virtual memory.
1240 */
cl_page_list_unmap(const struct lu_env * env,struct cl_io * io,struct cl_page_list * plist)1241 int cl_page_list_unmap(const struct lu_env *env, struct cl_io *io,
1242 struct cl_page_list *plist)
1243 {
1244 struct cl_page *page;
1245 int result;
1246
1247 LINVRNT(plist->pl_owner == current);
1248 result = 0;
1249 cl_page_list_for_each(page, plist) {
1250 result = cl_page_unmap(env, io, page);
1251 if (result != 0)
1252 break;
1253 }
1254 return result;
1255 }
1256 EXPORT_SYMBOL(cl_page_list_unmap);
1257
1258 /**
1259 * Initialize dual page queue.
1260 */
cl_2queue_init(struct cl_2queue * queue)1261 void cl_2queue_init(struct cl_2queue *queue)
1262 {
1263 cl_page_list_init(&queue->c2_qin);
1264 cl_page_list_init(&queue->c2_qout);
1265 }
1266 EXPORT_SYMBOL(cl_2queue_init);
1267
1268 /**
1269 * Add a page to the incoming page list of 2-queue.
1270 */
cl_2queue_add(struct cl_2queue * queue,struct cl_page * page)1271 void cl_2queue_add(struct cl_2queue *queue, struct cl_page *page)
1272 {
1273 cl_page_list_add(&queue->c2_qin, page);
1274 }
1275 EXPORT_SYMBOL(cl_2queue_add);
1276
1277 /**
1278 * Disown pages in both lists of a 2-queue.
1279 */
cl_2queue_disown(const struct lu_env * env,struct cl_io * io,struct cl_2queue * queue)1280 void cl_2queue_disown(const struct lu_env *env,
1281 struct cl_io *io, struct cl_2queue *queue)
1282 {
1283 cl_page_list_disown(env, io, &queue->c2_qin);
1284 cl_page_list_disown(env, io, &queue->c2_qout);
1285 }
1286 EXPORT_SYMBOL(cl_2queue_disown);
1287
1288 /**
1289 * Discard (truncate) pages in both lists of a 2-queue.
1290 */
cl_2queue_discard(const struct lu_env * env,struct cl_io * io,struct cl_2queue * queue)1291 void cl_2queue_discard(const struct lu_env *env,
1292 struct cl_io *io, struct cl_2queue *queue)
1293 {
1294 cl_page_list_discard(env, io, &queue->c2_qin);
1295 cl_page_list_discard(env, io, &queue->c2_qout);
1296 }
1297 EXPORT_SYMBOL(cl_2queue_discard);
1298
1299 /**
1300 * Assume to own the pages in cl_2queue
1301 */
cl_2queue_assume(const struct lu_env * env,struct cl_io * io,struct cl_2queue * queue)1302 void cl_2queue_assume(const struct lu_env *env,
1303 struct cl_io *io, struct cl_2queue *queue)
1304 {
1305 cl_page_list_assume(env, io, &queue->c2_qin);
1306 cl_page_list_assume(env, io, &queue->c2_qout);
1307 }
1308 EXPORT_SYMBOL(cl_2queue_assume);
1309
1310 /**
1311 * Finalize both page lists of a 2-queue.
1312 */
cl_2queue_fini(const struct lu_env * env,struct cl_2queue * queue)1313 void cl_2queue_fini(const struct lu_env *env, struct cl_2queue *queue)
1314 {
1315 cl_page_list_fini(env, &queue->c2_qout);
1316 cl_page_list_fini(env, &queue->c2_qin);
1317 }
1318 EXPORT_SYMBOL(cl_2queue_fini);
1319
1320 /**
1321 * Initialize a 2-queue to contain \a page in its incoming page list.
1322 */
cl_2queue_init_page(struct cl_2queue * queue,struct cl_page * page)1323 void cl_2queue_init_page(struct cl_2queue *queue, struct cl_page *page)
1324 {
1325 cl_2queue_init(queue);
1326 cl_2queue_add(queue, page);
1327 }
1328 EXPORT_SYMBOL(cl_2queue_init_page);
1329
1330 /**
1331 * Returns top-level io.
1332 *
1333 * \see cl_object_top(), cl_page_top().
1334 */
cl_io_top(struct cl_io * io)1335 struct cl_io *cl_io_top(struct cl_io *io)
1336 {
1337 while (io->ci_parent != NULL)
1338 io = io->ci_parent;
1339 return io;
1340 }
1341 EXPORT_SYMBOL(cl_io_top);
1342
1343 /**
1344 * Prints human readable representation of \a io to the \a f.
1345 */
cl_io_print(const struct lu_env * env,void * cookie,lu_printer_t printer,const struct cl_io * io)1346 void cl_io_print(const struct lu_env *env, void *cookie,
1347 lu_printer_t printer, const struct cl_io *io)
1348 {
1349 }
1350
1351 /**
1352 * Adds request slice to the compound request.
1353 *
1354 * This is called by cl_device_operations::cdo_req_init() methods to add a
1355 * per-layer state to the request. New state is added at the end of
1356 * cl_req::crq_layers list, that is, it is at the bottom of the stack.
1357 *
1358 * \see cl_lock_slice_add(), cl_page_slice_add(), cl_io_slice_add()
1359 */
cl_req_slice_add(struct cl_req * req,struct cl_req_slice * slice,struct cl_device * dev,const struct cl_req_operations * ops)1360 void cl_req_slice_add(struct cl_req *req, struct cl_req_slice *slice,
1361 struct cl_device *dev,
1362 const struct cl_req_operations *ops)
1363 {
1364 list_add_tail(&slice->crs_linkage, &req->crq_layers);
1365 slice->crs_dev = dev;
1366 slice->crs_ops = ops;
1367 slice->crs_req = req;
1368 }
1369 EXPORT_SYMBOL(cl_req_slice_add);
1370
cl_req_free(const struct lu_env * env,struct cl_req * req)1371 static void cl_req_free(const struct lu_env *env, struct cl_req *req)
1372 {
1373 unsigned i;
1374
1375 LASSERT(list_empty(&req->crq_pages));
1376 LASSERT(req->crq_nrpages == 0);
1377 LINVRNT(list_empty(&req->crq_layers));
1378 LINVRNT(equi(req->crq_nrobjs > 0, req->crq_o != NULL));
1379
1380 if (req->crq_o != NULL) {
1381 for (i = 0; i < req->crq_nrobjs; ++i) {
1382 struct cl_object *obj = req->crq_o[i].ro_obj;
1383 if (obj != NULL) {
1384 lu_object_ref_del_at(&obj->co_lu,
1385 &req->crq_o[i].ro_obj_ref,
1386 "cl_req", req);
1387 cl_object_put(env, obj);
1388 }
1389 }
1390 OBD_FREE(req->crq_o, req->crq_nrobjs * sizeof(req->crq_o[0]));
1391 }
1392 OBD_FREE_PTR(req);
1393 }
1394
cl_req_init(const struct lu_env * env,struct cl_req * req,struct cl_page * page)1395 static int cl_req_init(const struct lu_env *env, struct cl_req *req,
1396 struct cl_page *page)
1397 {
1398 struct cl_device *dev;
1399 struct cl_page_slice *slice;
1400 int result;
1401
1402 result = 0;
1403 page = cl_page_top(page);
1404 do {
1405 list_for_each_entry(slice, &page->cp_layers, cpl_linkage) {
1406 dev = lu2cl_dev(slice->cpl_obj->co_lu.lo_dev);
1407 if (dev->cd_ops->cdo_req_init != NULL) {
1408 result = dev->cd_ops->cdo_req_init(env,
1409 dev, req);
1410 if (result != 0)
1411 break;
1412 }
1413 }
1414 page = page->cp_child;
1415 } while (page != NULL && result == 0);
1416 return result;
1417 }
1418
1419 /**
1420 * Invokes per-request transfer completion call-backs
1421 * (cl_req_operations::cro_completion()) bottom-to-top.
1422 */
cl_req_completion(const struct lu_env * env,struct cl_req * req,int rc)1423 void cl_req_completion(const struct lu_env *env, struct cl_req *req, int rc)
1424 {
1425 struct cl_req_slice *slice;
1426
1427 /*
1428 * for the lack of list_for_each_entry_reverse_safe()...
1429 */
1430 while (!list_empty(&req->crq_layers)) {
1431 slice = list_entry(req->crq_layers.prev,
1432 struct cl_req_slice, crs_linkage);
1433 list_del_init(&slice->crs_linkage);
1434 if (slice->crs_ops->cro_completion != NULL)
1435 slice->crs_ops->cro_completion(env, slice, rc);
1436 }
1437 cl_req_free(env, req);
1438 }
1439 EXPORT_SYMBOL(cl_req_completion);
1440
1441 /**
1442 * Allocates new transfer request.
1443 */
cl_req_alloc(const struct lu_env * env,struct cl_page * page,enum cl_req_type crt,int nr_objects)1444 struct cl_req *cl_req_alloc(const struct lu_env *env, struct cl_page *page,
1445 enum cl_req_type crt, int nr_objects)
1446 {
1447 struct cl_req *req;
1448
1449 LINVRNT(nr_objects > 0);
1450
1451 OBD_ALLOC_PTR(req);
1452 if (req != NULL) {
1453 int result;
1454
1455 req->crq_type = crt;
1456 INIT_LIST_HEAD(&req->crq_pages);
1457 INIT_LIST_HEAD(&req->crq_layers);
1458
1459 OBD_ALLOC(req->crq_o, nr_objects * sizeof(req->crq_o[0]));
1460 if (req->crq_o != NULL) {
1461 req->crq_nrobjs = nr_objects;
1462 result = cl_req_init(env, req, page);
1463 } else
1464 result = -ENOMEM;
1465 if (result != 0) {
1466 cl_req_completion(env, req, result);
1467 req = ERR_PTR(result);
1468 }
1469 } else
1470 req = ERR_PTR(-ENOMEM);
1471 return req;
1472 }
1473 EXPORT_SYMBOL(cl_req_alloc);
1474
1475 /**
1476 * Adds a page to a request.
1477 */
cl_req_page_add(const struct lu_env * env,struct cl_req * req,struct cl_page * page)1478 void cl_req_page_add(const struct lu_env *env,
1479 struct cl_req *req, struct cl_page *page)
1480 {
1481 struct cl_object *obj;
1482 struct cl_req_obj *rqo;
1483 int i;
1484
1485 page = cl_page_top(page);
1486
1487 LASSERT(list_empty(&page->cp_flight));
1488 LASSERT(page->cp_req == NULL);
1489
1490 CL_PAGE_DEBUG(D_PAGE, env, page, "req %p, %d, %u\n",
1491 req, req->crq_type, req->crq_nrpages);
1492
1493 list_add_tail(&page->cp_flight, &req->crq_pages);
1494 ++req->crq_nrpages;
1495 page->cp_req = req;
1496 obj = cl_object_top(page->cp_obj);
1497 for (i = 0, rqo = req->crq_o; obj != rqo->ro_obj; ++i, ++rqo) {
1498 if (rqo->ro_obj == NULL) {
1499 rqo->ro_obj = obj;
1500 cl_object_get(obj);
1501 lu_object_ref_add_at(&obj->co_lu, &rqo->ro_obj_ref,
1502 "cl_req", req);
1503 break;
1504 }
1505 }
1506 LASSERT(i < req->crq_nrobjs);
1507 }
1508 EXPORT_SYMBOL(cl_req_page_add);
1509
1510 /**
1511 * Removes a page from a request.
1512 */
cl_req_page_done(const struct lu_env * env,struct cl_page * page)1513 void cl_req_page_done(const struct lu_env *env, struct cl_page *page)
1514 {
1515 struct cl_req *req = page->cp_req;
1516
1517 page = cl_page_top(page);
1518
1519 LASSERT(!list_empty(&page->cp_flight));
1520 LASSERT(req->crq_nrpages > 0);
1521
1522 list_del_init(&page->cp_flight);
1523 --req->crq_nrpages;
1524 page->cp_req = NULL;
1525 }
1526 EXPORT_SYMBOL(cl_req_page_done);
1527
1528 /**
1529 * Notifies layers that request is about to depart by calling
1530 * cl_req_operations::cro_prep() top-to-bottom.
1531 */
cl_req_prep(const struct lu_env * env,struct cl_req * req)1532 int cl_req_prep(const struct lu_env *env, struct cl_req *req)
1533 {
1534 int i;
1535 int result;
1536 const struct cl_req_slice *slice;
1537
1538 /*
1539 * Check that the caller of cl_req_alloc() didn't lie about the number
1540 * of objects.
1541 */
1542 for (i = 0; i < req->crq_nrobjs; ++i)
1543 LASSERT(req->crq_o[i].ro_obj != NULL);
1544
1545 result = 0;
1546 list_for_each_entry(slice, &req->crq_layers, crs_linkage) {
1547 if (slice->crs_ops->cro_prep != NULL) {
1548 result = slice->crs_ops->cro_prep(env, slice);
1549 if (result != 0)
1550 break;
1551 }
1552 }
1553 return result;
1554 }
1555 EXPORT_SYMBOL(cl_req_prep);
1556
1557 /**
1558 * Fills in attributes that are passed to server together with transfer. Only
1559 * attributes from \a flags may be touched. This can be called multiple times
1560 * for the same request.
1561 */
cl_req_attr_set(const struct lu_env * env,struct cl_req * req,struct cl_req_attr * attr,u64 flags)1562 void cl_req_attr_set(const struct lu_env *env, struct cl_req *req,
1563 struct cl_req_attr *attr, u64 flags)
1564 {
1565 const struct cl_req_slice *slice;
1566 struct cl_page *page;
1567 int i;
1568
1569 LASSERT(!list_empty(&req->crq_pages));
1570
1571 /* Take any page to use as a model. */
1572 page = list_entry(req->crq_pages.next, struct cl_page, cp_flight);
1573
1574 for (i = 0; i < req->crq_nrobjs; ++i) {
1575 list_for_each_entry(slice, &req->crq_layers, crs_linkage) {
1576 const struct cl_page_slice *scan;
1577 const struct cl_object *obj;
1578
1579 scan = cl_page_at(page,
1580 slice->crs_dev->cd_lu_dev.ld_type);
1581 LASSERT(scan != NULL);
1582 obj = scan->cpl_obj;
1583 if (slice->crs_ops->cro_attr_set != NULL)
1584 slice->crs_ops->cro_attr_set(env, slice, obj,
1585 attr + i, flags);
1586 }
1587 }
1588 }
1589 EXPORT_SYMBOL(cl_req_attr_set);
1590
1591 /* XXX complete(), init_completion(), and wait_for_completion(), until they are
1592 * implemented in libcfs. */
1593 # include <linux/sched.h>
1594
1595 /**
1596 * Initialize synchronous io wait anchor, for transfer of \a nrpages pages.
1597 */
cl_sync_io_init(struct cl_sync_io * anchor,int nrpages)1598 void cl_sync_io_init(struct cl_sync_io *anchor, int nrpages)
1599 {
1600 init_waitqueue_head(&anchor->csi_waitq);
1601 atomic_set(&anchor->csi_sync_nr, nrpages);
1602 atomic_set(&anchor->csi_barrier, nrpages > 0);
1603 anchor->csi_sync_rc = 0;
1604 }
1605 EXPORT_SYMBOL(cl_sync_io_init);
1606
1607 /**
1608 * Wait until all transfer completes. Transfer completion routine has to call
1609 * cl_sync_io_note() for every page.
1610 */
cl_sync_io_wait(const struct lu_env * env,struct cl_io * io,struct cl_page_list * queue,struct cl_sync_io * anchor,long timeout)1611 int cl_sync_io_wait(const struct lu_env *env, struct cl_io *io,
1612 struct cl_page_list *queue, struct cl_sync_io *anchor,
1613 long timeout)
1614 {
1615 struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(timeout),
1616 NULL, NULL, NULL);
1617 int rc;
1618
1619 LASSERT(timeout >= 0);
1620
1621 rc = l_wait_event(anchor->csi_waitq,
1622 atomic_read(&anchor->csi_sync_nr) == 0,
1623 &lwi);
1624 if (rc < 0) {
1625 CERROR("SYNC IO failed with error: %d, try to cancel %d remaining pages\n",
1626 rc, atomic_read(&anchor->csi_sync_nr));
1627
1628 (void)cl_io_cancel(env, io, queue);
1629
1630 lwi = (struct l_wait_info) { 0 };
1631 (void)l_wait_event(anchor->csi_waitq,
1632 atomic_read(&anchor->csi_sync_nr) == 0,
1633 &lwi);
1634 } else {
1635 rc = anchor->csi_sync_rc;
1636 }
1637 LASSERT(atomic_read(&anchor->csi_sync_nr) == 0);
1638 cl_page_list_assume(env, io, queue);
1639
1640 /* wait until cl_sync_io_note() has done wakeup */
1641 while (unlikely(atomic_read(&anchor->csi_barrier) != 0)) {
1642 cpu_relax();
1643 }
1644
1645 POISON(anchor, 0x5a, sizeof(*anchor));
1646 return rc;
1647 }
1648 EXPORT_SYMBOL(cl_sync_io_wait);
1649
1650 /**
1651 * Indicate that transfer of a single page completed.
1652 */
cl_sync_io_note(struct cl_sync_io * anchor,int ioret)1653 void cl_sync_io_note(struct cl_sync_io *anchor, int ioret)
1654 {
1655 if (anchor->csi_sync_rc == 0 && ioret < 0)
1656 anchor->csi_sync_rc = ioret;
1657 /*
1658 * Synchronous IO done without releasing page lock (e.g., as a part of
1659 * ->{prepare,commit}_write(). Completion is used to signal the end of
1660 * IO.
1661 */
1662 LASSERT(atomic_read(&anchor->csi_sync_nr) > 0);
1663 if (atomic_dec_and_test(&anchor->csi_sync_nr)) {
1664 wake_up_all(&anchor->csi_waitq);
1665 /* it's safe to nuke or reuse anchor now */
1666 atomic_set(&anchor->csi_barrier, 0);
1667 }
1668 }
1669 EXPORT_SYMBOL(cl_sync_io_note);
1670