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) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
28 * Use is subject to license terms.
29 *
30 * Copyright (c) 2011, 2013, 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 * lustre/fid/fid_request.c
37 *
38 * Lustre Sequence Manager
39 *
40 * Author: Yury Umanets <umka@clusterfs.com>
41 */
42
43 #define DEBUG_SUBSYSTEM S_FID
44
45 #include "../../include/linux/libcfs/libcfs.h"
46 #include <linux/module.h>
47
48 #include "../include/obd.h"
49 #include "../include/obd_class.h"
50 #include "../include/obd_support.h"
51 #include "../include/lustre_fid.h"
52 /* mdc RPC locks */
53 #include "../include/lustre_mdc.h"
54 #include "fid_internal.h"
55
seq_client_rpc(struct lu_client_seq * seq,struct lu_seq_range * output,__u32 opc,const char * opcname)56 static int seq_client_rpc(struct lu_client_seq *seq,
57 struct lu_seq_range *output, __u32 opc,
58 const char *opcname)
59 {
60 struct obd_export *exp = seq->lcs_exp;
61 struct ptlrpc_request *req;
62 struct lu_seq_range *out, *in;
63 __u32 *op;
64 unsigned int debug_mask;
65 int rc;
66
67 req = ptlrpc_request_alloc_pack(class_exp2cliimp(exp), &RQF_SEQ_QUERY,
68 LUSTRE_MDS_VERSION, SEQ_QUERY);
69 if (req == NULL)
70 return -ENOMEM;
71
72 /* Init operation code */
73 op = req_capsule_client_get(&req->rq_pill, &RMF_SEQ_OPC);
74 *op = opc;
75
76 /* Zero out input range, this is not recovery yet. */
77 in = req_capsule_client_get(&req->rq_pill, &RMF_SEQ_RANGE);
78 range_init(in);
79
80 ptlrpc_request_set_replen(req);
81
82 in->lsr_index = seq->lcs_space.lsr_index;
83 if (seq->lcs_type == LUSTRE_SEQ_METADATA)
84 fld_range_set_mdt(in);
85 else
86 fld_range_set_ost(in);
87
88 if (opc == SEQ_ALLOC_SUPER) {
89 req->rq_request_portal = SEQ_CONTROLLER_PORTAL;
90 req->rq_reply_portal = MDC_REPLY_PORTAL;
91 /* During allocating super sequence for data object,
92 * the current thread might hold the export of MDT0(MDT0
93 * precreating objects on this OST), and it will send the
94 * request to MDT0 here, so we can not keep resending the
95 * request here, otherwise if MDT0 is failed(umounted),
96 * it can not release the export of MDT0 */
97 if (seq->lcs_type == LUSTRE_SEQ_DATA)
98 req->rq_no_delay = req->rq_no_resend = 1;
99 debug_mask = D_CONSOLE;
100 } else {
101 if (seq->lcs_type == LUSTRE_SEQ_METADATA)
102 req->rq_request_portal = SEQ_METADATA_PORTAL;
103 else
104 req->rq_request_portal = SEQ_DATA_PORTAL;
105 debug_mask = D_INFO;
106 }
107
108 ptlrpc_at_set_req_timeout(req);
109
110 if (seq->lcs_type == LUSTRE_SEQ_METADATA)
111 mdc_get_rpc_lock(exp->exp_obd->u.cli.cl_rpc_lock, NULL);
112 rc = ptlrpc_queue_wait(req);
113 if (seq->lcs_type == LUSTRE_SEQ_METADATA)
114 mdc_put_rpc_lock(exp->exp_obd->u.cli.cl_rpc_lock, NULL);
115 if (rc)
116 goto out_req;
117
118 out = req_capsule_server_get(&req->rq_pill, &RMF_SEQ_RANGE);
119 *output = *out;
120
121 if (!range_is_sane(output)) {
122 CERROR("%s: Invalid range received from server: "
123 DRANGE"\n", seq->lcs_name, PRANGE(output));
124 rc = -EINVAL;
125 goto out_req;
126 }
127
128 if (range_is_exhausted(output)) {
129 CERROR("%s: Range received from server is exhausted: "
130 DRANGE"]\n", seq->lcs_name, PRANGE(output));
131 rc = -EINVAL;
132 goto out_req;
133 }
134
135 CDEBUG_LIMIT(debug_mask, "%s: Allocated %s-sequence "DRANGE"]\n",
136 seq->lcs_name, opcname, PRANGE(output));
137
138 out_req:
139 ptlrpc_req_finished(req);
140 return rc;
141 }
142
143 /* Request sequence-controller node to allocate new super-sequence. */
seq_client_alloc_super(struct lu_client_seq * seq,const struct lu_env * env)144 int seq_client_alloc_super(struct lu_client_seq *seq,
145 const struct lu_env *env)
146 {
147 int rc;
148
149 mutex_lock(&seq->lcs_mutex);
150
151 if (seq->lcs_srv) {
152 rc = 0;
153 } else {
154 /* Check whether the connection to seq controller has been
155 * setup (lcs_exp != NULL) */
156 if (seq->lcs_exp == NULL) {
157 mutex_unlock(&seq->lcs_mutex);
158 return -EINPROGRESS;
159 }
160
161 rc = seq_client_rpc(seq, &seq->lcs_space,
162 SEQ_ALLOC_SUPER, "super");
163 }
164 mutex_unlock(&seq->lcs_mutex);
165 return rc;
166 }
167
168 /* Request sequence-controller node to allocate new meta-sequence. */
seq_client_alloc_meta(const struct lu_env * env,struct lu_client_seq * seq)169 static int seq_client_alloc_meta(const struct lu_env *env,
170 struct lu_client_seq *seq)
171 {
172 int rc;
173
174 if (seq->lcs_srv) {
175 rc = 0;
176 } else {
177 do {
178 /* If meta server return -EINPROGRESS or EAGAIN,
179 * it means meta server might not be ready to
180 * allocate super sequence from sequence controller
181 * (MDT0)yet */
182 rc = seq_client_rpc(seq, &seq->lcs_space,
183 SEQ_ALLOC_META, "meta");
184 } while (rc == -EINPROGRESS || rc == -EAGAIN);
185 }
186
187 return rc;
188 }
189
190 /* Allocate new sequence for client. */
seq_client_alloc_seq(const struct lu_env * env,struct lu_client_seq * seq,u64 * seqnr)191 static int seq_client_alloc_seq(const struct lu_env *env,
192 struct lu_client_seq *seq, u64 *seqnr)
193 {
194 int rc;
195
196 LASSERT(range_is_sane(&seq->lcs_space));
197
198 if (range_is_exhausted(&seq->lcs_space)) {
199 rc = seq_client_alloc_meta(env, seq);
200 if (rc) {
201 CERROR("%s: Can't allocate new meta-sequence, rc %d\n",
202 seq->lcs_name, rc);
203 return rc;
204 } else {
205 CDEBUG(D_INFO, "%s: New range - "DRANGE"\n",
206 seq->lcs_name, PRANGE(&seq->lcs_space));
207 }
208 } else {
209 rc = 0;
210 }
211
212 LASSERT(!range_is_exhausted(&seq->lcs_space));
213 *seqnr = seq->lcs_space.lsr_start;
214 seq->lcs_space.lsr_start += 1;
215
216 CDEBUG(D_INFO, "%s: Allocated sequence [%#llx]\n", seq->lcs_name,
217 *seqnr);
218
219 return rc;
220 }
221
seq_fid_alloc_prep(struct lu_client_seq * seq,wait_queue_t * link)222 static int seq_fid_alloc_prep(struct lu_client_seq *seq,
223 wait_queue_t *link)
224 {
225 if (seq->lcs_update) {
226 add_wait_queue(&seq->lcs_waitq, link);
227 set_current_state(TASK_UNINTERRUPTIBLE);
228 mutex_unlock(&seq->lcs_mutex);
229
230 schedule();
231
232 mutex_lock(&seq->lcs_mutex);
233 remove_wait_queue(&seq->lcs_waitq, link);
234 set_current_state(TASK_RUNNING);
235 return -EAGAIN;
236 }
237 ++seq->lcs_update;
238 mutex_unlock(&seq->lcs_mutex);
239 return 0;
240 }
241
seq_fid_alloc_fini(struct lu_client_seq * seq)242 static void seq_fid_alloc_fini(struct lu_client_seq *seq)
243 {
244 LASSERT(seq->lcs_update == 1);
245 mutex_lock(&seq->lcs_mutex);
246 --seq->lcs_update;
247 wake_up(&seq->lcs_waitq);
248 }
249
250 /**
251 * Allocate the whole seq to the caller.
252 **/
seq_client_get_seq(const struct lu_env * env,struct lu_client_seq * seq,u64 * seqnr)253 int seq_client_get_seq(const struct lu_env *env,
254 struct lu_client_seq *seq, u64 *seqnr)
255 {
256 wait_queue_t link;
257 int rc;
258
259 LASSERT(seqnr != NULL);
260 mutex_lock(&seq->lcs_mutex);
261 init_waitqueue_entry(&link, current);
262
263 while (1) {
264 rc = seq_fid_alloc_prep(seq, &link);
265 if (rc == 0)
266 break;
267 }
268
269 rc = seq_client_alloc_seq(env, seq, seqnr);
270 if (rc) {
271 CERROR("%s: Can't allocate new sequence, rc %d\n",
272 seq->lcs_name, rc);
273 seq_fid_alloc_fini(seq);
274 mutex_unlock(&seq->lcs_mutex);
275 return rc;
276 }
277
278 CDEBUG(D_INFO, "%s: allocate sequence [0x%16.16Lx]\n",
279 seq->lcs_name, *seqnr);
280
281 /* Since the caller require the whole seq,
282 * so marked this seq to be used */
283 if (seq->lcs_type == LUSTRE_SEQ_METADATA)
284 seq->lcs_fid.f_oid = LUSTRE_METADATA_SEQ_MAX_WIDTH;
285 else
286 seq->lcs_fid.f_oid = LUSTRE_DATA_SEQ_MAX_WIDTH;
287
288 seq->lcs_fid.f_seq = *seqnr;
289 seq->lcs_fid.f_ver = 0;
290 /*
291 * Inform caller that sequence switch is performed to allow it
292 * to setup FLD for it.
293 */
294 seq_fid_alloc_fini(seq);
295 mutex_unlock(&seq->lcs_mutex);
296
297 return rc;
298 }
299 EXPORT_SYMBOL(seq_client_get_seq);
300
301 /* Allocate new fid on passed client @seq and save it to @fid. */
seq_client_alloc_fid(const struct lu_env * env,struct lu_client_seq * seq,struct lu_fid * fid)302 int seq_client_alloc_fid(const struct lu_env *env,
303 struct lu_client_seq *seq, struct lu_fid *fid)
304 {
305 wait_queue_t link;
306 int rc;
307
308 LASSERT(seq != NULL);
309 LASSERT(fid != NULL);
310
311 init_waitqueue_entry(&link, current);
312 mutex_lock(&seq->lcs_mutex);
313
314 if (OBD_FAIL_CHECK(OBD_FAIL_SEQ_EXHAUST))
315 seq->lcs_fid.f_oid = seq->lcs_width;
316
317 while (1) {
318 u64 seqnr;
319
320 if (!fid_is_zero(&seq->lcs_fid) &&
321 fid_oid(&seq->lcs_fid) < seq->lcs_width) {
322 /* Just bump last allocated fid and return to caller. */
323 seq->lcs_fid.f_oid += 1;
324 rc = 0;
325 break;
326 }
327
328 rc = seq_fid_alloc_prep(seq, &link);
329 if (rc)
330 continue;
331
332 rc = seq_client_alloc_seq(env, seq, &seqnr);
333 if (rc) {
334 CERROR("%s: Can't allocate new sequence, rc %d\n",
335 seq->lcs_name, rc);
336 seq_fid_alloc_fini(seq);
337 mutex_unlock(&seq->lcs_mutex);
338 return rc;
339 }
340
341 CDEBUG(D_INFO, "%s: Switch to sequence [0x%16.16Lx]\n",
342 seq->lcs_name, seqnr);
343
344 seq->lcs_fid.f_oid = LUSTRE_FID_INIT_OID;
345 seq->lcs_fid.f_seq = seqnr;
346 seq->lcs_fid.f_ver = 0;
347
348 /*
349 * Inform caller that sequence switch is performed to allow it
350 * to setup FLD for it.
351 */
352 rc = 1;
353
354 seq_fid_alloc_fini(seq);
355 break;
356 }
357
358 *fid = seq->lcs_fid;
359 mutex_unlock(&seq->lcs_mutex);
360
361 CDEBUG(D_INFO, "%s: Allocated FID "DFID"\n", seq->lcs_name, PFID(fid));
362 return rc;
363 }
364 EXPORT_SYMBOL(seq_client_alloc_fid);
365
366 /*
367 * Finish the current sequence due to disconnect.
368 * See mdc_import_event()
369 */
seq_client_flush(struct lu_client_seq * seq)370 void seq_client_flush(struct lu_client_seq *seq)
371 {
372 wait_queue_t link;
373
374 LASSERT(seq != NULL);
375 init_waitqueue_entry(&link, current);
376 mutex_lock(&seq->lcs_mutex);
377
378 while (seq->lcs_update) {
379 add_wait_queue(&seq->lcs_waitq, &link);
380 set_current_state(TASK_UNINTERRUPTIBLE);
381 mutex_unlock(&seq->lcs_mutex);
382
383 schedule();
384
385 mutex_lock(&seq->lcs_mutex);
386 remove_wait_queue(&seq->lcs_waitq, &link);
387 set_current_state(TASK_RUNNING);
388 }
389
390 fid_zero(&seq->lcs_fid);
391 /**
392 * this id shld not be used for seq range allocation.
393 * set to -1 for dgb check.
394 */
395
396 seq->lcs_space.lsr_index = -1;
397
398 range_init(&seq->lcs_space);
399 mutex_unlock(&seq->lcs_mutex);
400 }
401 EXPORT_SYMBOL(seq_client_flush);
402
seq_client_proc_fini(struct lu_client_seq * seq)403 static void seq_client_proc_fini(struct lu_client_seq *seq)
404 {
405 #if defined(CONFIG_PROC_FS)
406 if (seq->lcs_proc_dir) {
407 if (!IS_ERR(seq->lcs_proc_dir))
408 lprocfs_remove(&seq->lcs_proc_dir);
409 seq->lcs_proc_dir = NULL;
410 }
411 #endif /* CONFIG_PROC_FS */
412 }
413
seq_client_proc_init(struct lu_client_seq * seq)414 static int seq_client_proc_init(struct lu_client_seq *seq)
415 {
416 #if defined(CONFIG_PROC_FS)
417 int rc;
418
419 seq->lcs_proc_dir = lprocfs_register(seq->lcs_name,
420 seq_type_proc_dir,
421 NULL, NULL);
422
423 if (IS_ERR(seq->lcs_proc_dir)) {
424 CERROR("%s: LProcFS failed in seq-init\n",
425 seq->lcs_name);
426 rc = PTR_ERR(seq->lcs_proc_dir);
427 return rc;
428 }
429
430 rc = lprocfs_add_vars(seq->lcs_proc_dir,
431 seq_client_proc_list, seq);
432 if (rc) {
433 CERROR("%s: Can't init sequence manager proc, rc %d\n",
434 seq->lcs_name, rc);
435 goto out_cleanup;
436 }
437
438 return 0;
439
440 out_cleanup:
441 seq_client_proc_fini(seq);
442 return rc;
443
444 #else /* CONFIG_PROC_FS */
445 return 0;
446 #endif
447 }
448
seq_client_init(struct lu_client_seq * seq,struct obd_export * exp,enum lu_cli_type type,const char * prefix,struct lu_server_seq * srv)449 int seq_client_init(struct lu_client_seq *seq,
450 struct obd_export *exp,
451 enum lu_cli_type type,
452 const char *prefix,
453 struct lu_server_seq *srv)
454 {
455 int rc;
456
457 LASSERT(seq != NULL);
458 LASSERT(prefix != NULL);
459
460 seq->lcs_srv = srv;
461 seq->lcs_type = type;
462
463 mutex_init(&seq->lcs_mutex);
464 if (type == LUSTRE_SEQ_METADATA)
465 seq->lcs_width = LUSTRE_METADATA_SEQ_MAX_WIDTH;
466 else
467 seq->lcs_width = LUSTRE_DATA_SEQ_MAX_WIDTH;
468
469 init_waitqueue_head(&seq->lcs_waitq);
470 /* Make sure that things are clear before work is started. */
471 seq_client_flush(seq);
472
473 if (exp != NULL)
474 seq->lcs_exp = class_export_get(exp);
475 else if (type == LUSTRE_SEQ_METADATA)
476 LASSERT(seq->lcs_srv != NULL);
477
478 snprintf(seq->lcs_name, sizeof(seq->lcs_name),
479 "cli-%s", prefix);
480
481 rc = seq_client_proc_init(seq);
482 if (rc)
483 seq_client_fini(seq);
484 return rc;
485 }
486 EXPORT_SYMBOL(seq_client_init);
487
seq_client_fini(struct lu_client_seq * seq)488 void seq_client_fini(struct lu_client_seq *seq)
489 {
490 seq_client_proc_fini(seq);
491
492 if (seq->lcs_exp != NULL) {
493 class_export_put(seq->lcs_exp);
494 seq->lcs_exp = NULL;
495 }
496
497 seq->lcs_srv = NULL;
498 }
499 EXPORT_SYMBOL(seq_client_fini);
500
client_fid_init(struct obd_device * obd,struct obd_export * exp,enum lu_cli_type type)501 int client_fid_init(struct obd_device *obd,
502 struct obd_export *exp, enum lu_cli_type type)
503 {
504 struct client_obd *cli = &obd->u.cli;
505 char *prefix;
506 int rc;
507
508 OBD_ALLOC_PTR(cli->cl_seq);
509 if (cli->cl_seq == NULL)
510 return -ENOMEM;
511
512 OBD_ALLOC(prefix, MAX_OBD_NAME + 5);
513 if (prefix == NULL) {
514 rc = -ENOMEM;
515 goto out_free_seq;
516 }
517
518 snprintf(prefix, MAX_OBD_NAME + 5, "cli-%s", obd->obd_name);
519
520 /* Init client side sequence-manager */
521 rc = seq_client_init(cli->cl_seq, exp, type, prefix, NULL);
522 OBD_FREE(prefix, MAX_OBD_NAME + 5);
523 if (rc)
524 goto out_free_seq;
525
526 return rc;
527 out_free_seq:
528 OBD_FREE_PTR(cli->cl_seq);
529 cli->cl_seq = NULL;
530 return rc;
531 }
532 EXPORT_SYMBOL(client_fid_init);
533
client_fid_fini(struct obd_device * obd)534 int client_fid_fini(struct obd_device *obd)
535 {
536 struct client_obd *cli = &obd->u.cli;
537
538 if (cli->cl_seq != NULL) {
539 seq_client_fini(cli->cl_seq);
540 OBD_FREE_PTR(cli->cl_seq);
541 cli->cl_seq = NULL;
542 }
543
544 return 0;
545 }
546 EXPORT_SYMBOL(client_fid_fini);
547
548 struct proc_dir_entry *seq_type_proc_dir;
549
fid_mod_init(void)550 static int __init fid_mod_init(void)
551 {
552 seq_type_proc_dir = lprocfs_register(LUSTRE_SEQ_NAME,
553 proc_lustre_root,
554 NULL, NULL);
555 return PTR_ERR_OR_ZERO(seq_type_proc_dir);
556 }
557
fid_mod_exit(void)558 static void __exit fid_mod_exit(void)
559 {
560 if (seq_type_proc_dir != NULL && !IS_ERR(seq_type_proc_dir)) {
561 lprocfs_remove(&seq_type_proc_dir);
562 seq_type_proc_dir = NULL;
563 }
564 }
565
566 MODULE_AUTHOR("Sun Microsystems, Inc. <http://www.lustre.org/>");
567 MODULE_DESCRIPTION("Lustre FID Module");
568 MODULE_LICENSE("GPL");
569 MODULE_VERSION("0.1.0");
570
571 module_init(fid_mod_init);
572 module_exit(fid_mod_exit);
573