projects / deliverable / linux.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
staging/lustre/llog: Do not use ostid swab for llogid
[deliverable/linux.git] / drivers / staging / lustre / lustre / osc / osc_request.c
CommitLineData
d7e09d03
PT		 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) 2002, 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
37#define DEBUG_SUBSYSTEM S_OSC
38
39#include <linux/libcfs/libcfs.h>
40
41
42#include <lustre_dlm.h>
43#include <lustre_net.h>
44#include <lustre/lustre_user.h>
45#include <obd_cksum.h>
46#include <obd_ost.h>
47#include <obd_lov.h>
48
49#ifdef __CYGWIN__
50# include <ctype.h>
51#endif
52
53#include <lustre_ha.h>
54#include <lprocfs_status.h>
55#include <lustre_log.h>
56#include <lustre_debug.h>
57#include <lustre_param.h>
58#include <lustre_fid.h>
59#include "osc_internal.h"
60#include "osc_cl_internal.h"
61
62static void osc_release_ppga(struct brw_page **ppga, obd_count count);
63static int brw_interpret(const struct lu_env *env,
64 struct ptlrpc_request *req, void *data, int rc);
65int osc_cleanup(struct obd_device *obd);
66
67/* Pack OSC object metadata for disk storage (LE byte order). */
68static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
69 struct lov_stripe_md *lsm)
70{
71 int lmm_size;
72 ENTRY;
73
74 lmm_size = sizeof(**lmmp);
75 if (lmmp == NULL)
76 RETURN(lmm_size);
77
78 if (*lmmp != NULL && lsm == NULL) {
79 OBD_FREE(*lmmp, lmm_size);
80 *lmmp = NULL;
81 RETURN(0);
82 } else if (unlikely(lsm != NULL && ostid_id(&lsm->lsm_oi) == 0)) {
83 RETURN(-EBADF);
84 }
85
86 if (*lmmp == NULL) {
87 OBD_ALLOC(*lmmp, lmm_size);
88 if (*lmmp == NULL)
89 RETURN(-ENOMEM);
90 }
91
92 if (lsm)
93 ostid_cpu_to_le(&lsm->lsm_oi, &(*lmmp)->lmm_oi);
94
95 RETURN(lmm_size);
96}
97
98/* Unpack OSC object metadata from disk storage (LE byte order). */
99static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
100 struct lov_mds_md *lmm, int lmm_bytes)
101{
102 int lsm_size;
103 struct obd_import *imp = class_exp2cliimp(exp);
104 ENTRY;
105
106 if (lmm != NULL) {
107 if (lmm_bytes < sizeof(*lmm)) {
108 CERROR("%s: lov_mds_md too small: %d, need %d\n",
109 exp->exp_obd->obd_name, lmm_bytes,
110 (int)sizeof(*lmm));
111 RETURN(-EINVAL);
112 }
113 /* XXX LOV_MAGIC etc check? */
114
115 if (unlikely(ostid_id(&lmm->lmm_oi) == 0)) {
116 CERROR("%s: zero lmm_object_id: rc = %d\n",
117 exp->exp_obd->obd_name, -EINVAL);
118 RETURN(-EINVAL);
119 }
120 }
121
122 lsm_size = lov_stripe_md_size(1);
123 if (lsmp == NULL)
124 RETURN(lsm_size);
125
126 if (*lsmp != NULL && lmm == NULL) {
127 OBD_FREE((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
128 OBD_FREE(*lsmp, lsm_size);
129 *lsmp = NULL;
130 RETURN(0);
131 }
132
133 if (*lsmp == NULL) {
134 OBD_ALLOC(*lsmp, lsm_size);
135 if (unlikely(*lsmp == NULL))
136 RETURN(-ENOMEM);
137 OBD_ALLOC((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
138 if (unlikely((*lsmp)->lsm_oinfo[0] == NULL)) {
139 OBD_FREE(*lsmp, lsm_size);
140 RETURN(-ENOMEM);
141 }
142 loi_init((*lsmp)->lsm_oinfo[0]);
143 } else if (unlikely(ostid_id(&(*lsmp)->lsm_oi) == 0)) {
144 RETURN(-EBADF);
145 }
146
147 if (lmm != NULL)
148 /* XXX zero *lsmp? */
149 ostid_le_to_cpu(&lmm->lmm_oi, &(*lsmp)->lsm_oi);
150
151 if (imp != NULL &&
152 (imp->imp_connect_data.ocd_connect_flags & OBD_CONNECT_MAXBYTES))
153 (*lsmp)->lsm_maxbytes = imp->imp_connect_data.ocd_maxbytes;
154 else
155 (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
156
157 RETURN(lsm_size);
158}
159
160static inline void osc_pack_capa(struct ptlrpc_request *req,
161 struct ost_body *body, void *capa)
162{
163 struct obd_capa *oc = (struct obd_capa *)capa;
164 struct lustre_capa *c;
165
166 if (!capa)
167 return;
168
169 c = req_capsule_client_get(&req->rq_pill, &RMF_CAPA1);
170 LASSERT(c);
171 capa_cpy(c, oc);
172 body->oa.o_valid |= OBD_MD_FLOSSCAPA;
173 DEBUG_CAPA(D_SEC, c, "pack");
174}
175
176static inline void osc_pack_req_body(struct ptlrpc_request *req,
177 struct obd_info *oinfo)
178{
179 struct ost_body *body;
180
181 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
182 LASSERT(body);
183
184 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
185 osc_pack_capa(req, body, oinfo->oi_capa);
186}
187
188static inline void osc_set_capa_size(struct ptlrpc_request *req,
189 const struct req_msg_field *field,
190 struct obd_capa *oc)
191{
192 if (oc == NULL)
193 req_capsule_set_size(&req->rq_pill, field, RCL_CLIENT, 0);
194 else
195 /* it is already calculated as sizeof struct obd_capa */
196 ;
197}
198
199static int osc_getattr_interpret(const struct lu_env *env,
200 struct ptlrpc_request *req,
201 struct osc_async_args *aa, int rc)
202{
203 struct ost_body *body;
204 ENTRY;
205
206 if (rc != 0)
207 GOTO(out, rc);
208
209 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
210 if (body) {
211 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
212 lustre_get_wire_obdo(aa->aa_oi->oi_oa, &body->oa);
213
214 /* This should really be sent by the OST */
215 aa->aa_oi->oi_oa->o_blksize = DT_MAX_BRW_SIZE;
216 aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
217 } else {
218 CDEBUG(D_INFO, "can't unpack ost_body\n");
219 rc = -EPROTO;
220 aa->aa_oi->oi_oa->o_valid = 0;
221 }
222out:
223 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
224 RETURN(rc);
225}
226
227static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
228 struct ptlrpc_request_set *set)
229{
230 struct ptlrpc_request *req;
231 struct osc_async_args *aa;
232 int rc;
233 ENTRY;
234
235 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
236 if (req == NULL)
237 RETURN(-ENOMEM);
238
239 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
240 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
241 if (rc) {
242 ptlrpc_request_free(req);
243 RETURN(rc);
244 }
245
246 osc_pack_req_body(req, oinfo);
247
248 ptlrpc_request_set_replen(req);
249 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_getattr_interpret;
250
251 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
252 aa = ptlrpc_req_async_args(req);
253 aa->aa_oi = oinfo;
254
255 ptlrpc_set_add_req(set, req);
256 RETURN(0);
257}
258
259static int osc_getattr(const struct lu_env *env, struct obd_export *exp,
260 struct obd_info *oinfo)
261{
262 struct ptlrpc_request *req;
263 struct ost_body *body;
264 int rc;
265 ENTRY;
266
267 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
268 if (req == NULL)
269 RETURN(-ENOMEM);
270
271 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
272 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
273 if (rc) {
274 ptlrpc_request_free(req);
275 RETURN(rc);
276 }
277
278 osc_pack_req_body(req, oinfo);
279
280 ptlrpc_request_set_replen(req);
281
282 rc = ptlrpc_queue_wait(req);
283 if (rc)
284 GOTO(out, rc);
285
286 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
287 if (body == NULL)
288 GOTO(out, rc = -EPROTO);
289
290 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
291 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
292
293 oinfo->oi_oa->o_blksize = cli_brw_size(exp->exp_obd);
294 oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
295
296 EXIT;
297 out:
298 ptlrpc_req_finished(req);
299 return rc;
300}
301
302static int osc_setattr(const struct lu_env *env, struct obd_export *exp,
303 struct obd_info *oinfo, struct obd_trans_info *oti)
304{
305 struct ptlrpc_request *req;
306 struct ost_body *body;
307 int rc;
308 ENTRY;
309
310 LASSERT(oinfo->oi_oa->o_valid & OBD_MD_FLGROUP);
311
312 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
313 if (req == NULL)
314 RETURN(-ENOMEM);
315
316 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
317 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
318 if (rc) {
319 ptlrpc_request_free(req);
320 RETURN(rc);
321 }
322
323 osc_pack_req_body(req, oinfo);
324
325 ptlrpc_request_set_replen(req);
326
327 rc = ptlrpc_queue_wait(req);
328 if (rc)
329 GOTO(out, rc);
330
331 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
332 if (body == NULL)
333 GOTO(out, rc = -EPROTO);
334
335 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
336
337 EXIT;
338out:
339 ptlrpc_req_finished(req);
340 RETURN(rc);
341}
342
343static int osc_setattr_interpret(const struct lu_env *env,
344 struct ptlrpc_request *req,
345 struct osc_setattr_args *sa, int rc)
346{
347 struct ost_body *body;
348 ENTRY;
349
350 if (rc != 0)
351 GOTO(out, rc);
352
353 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
354 if (body == NULL)
355 GOTO(out, rc = -EPROTO);
356
357 lustre_get_wire_obdo(sa->sa_oa, &body->oa);
358out:
359 rc = sa->sa_upcall(sa->sa_cookie, rc);
360 RETURN(rc);
361}
362
363int osc_setattr_async_base(struct obd_export *exp, struct obd_info *oinfo,
364 struct obd_trans_info *oti,
365 obd_enqueue_update_f upcall, void *cookie,
366 struct ptlrpc_request_set *rqset)
367{
368 struct ptlrpc_request *req;
369 struct osc_setattr_args *sa;
370 int rc;
371 ENTRY;
372
373 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
374 if (req == NULL)
375 RETURN(-ENOMEM);
376
377 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
378 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
379 if (rc) {
380 ptlrpc_request_free(req);
381 RETURN(rc);
382 }
383
384 if (oti && oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE)
385 oinfo->oi_oa->o_lcookie = *oti->oti_logcookies;
386
387 osc_pack_req_body(req, oinfo);
388
389 ptlrpc_request_set_replen(req);
390
391 /* do mds to ost setattr asynchronously */
392 if (!rqset) {
393 /* Do not wait for response. */
394 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
395 } else {
396 req->rq_interpret_reply =
397 (ptlrpc_interpterer_t)osc_setattr_interpret;
398
399 CLASSERT (sizeof(*sa) <= sizeof(req->rq_async_args));
400 sa = ptlrpc_req_async_args(req);
401 sa->sa_oa = oinfo->oi_oa;
402 sa->sa_upcall = upcall;
403 sa->sa_cookie = cookie;
404
405 if (rqset == PTLRPCD_SET)
406 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
407 else
408 ptlrpc_set_add_req(rqset, req);
409 }
410
411 RETURN(0);
412}
413
414static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
415 struct obd_trans_info *oti,
416 struct ptlrpc_request_set *rqset)
417{
418 return osc_setattr_async_base(exp, oinfo, oti,
419 oinfo->oi_cb_up, oinfo, rqset);
420}
421
422int osc_real_create(struct obd_export *exp, struct obdo *oa,
423 struct lov_stripe_md **ea, struct obd_trans_info *oti)
424{
425 struct ptlrpc_request *req;
426 struct ost_body *body;
427 struct lov_stripe_md *lsm;
428 int rc;
429 ENTRY;
430
431 LASSERT(oa);
432 LASSERT(ea);
433
434 lsm = *ea;
435 if (!lsm) {
436 rc = obd_alloc_memmd(exp, &lsm);
437 if (rc < 0)
438 RETURN(rc);
439 }
440
441 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
442 if (req == NULL)
443 GOTO(out, rc = -ENOMEM);
444
445 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_CREATE);
446 if (rc) {
447 ptlrpc_request_free(req);
448 GOTO(out, rc);
449 }
450
451 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
452 LASSERT(body);
453 lustre_set_wire_obdo(&body->oa, oa);
454
455 ptlrpc_request_set_replen(req);
456
457 if ((oa->o_valid & OBD_MD_FLFLAGS) &&
458 oa->o_flags == OBD_FL_DELORPHAN) {
459 DEBUG_REQ(D_HA, req,
460 "delorphan from OST integration");
461 /* Don't resend the delorphan req */
462 req->rq_no_resend = req->rq_no_delay = 1;
463 }
464
465 rc = ptlrpc_queue_wait(req);
466 if (rc)
467 GOTO(out_req, rc);
468
469 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
470 if (body == NULL)
471 GOTO(out_req, rc = -EPROTO);
472
473 lustre_get_wire_obdo(oa, &body->oa);
474
475 oa->o_blksize = cli_brw_size(exp->exp_obd);
476 oa->o_valid |= OBD_MD_FLBLKSZ;
477
478 /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
479 * have valid lsm_oinfo data structs, so don't go touching that.
480 * This needs to be fixed in a big way.
481 */
482 lsm->lsm_oi = oa->o_oi;
483 *ea = lsm;
484
485 if (oti != NULL) {
486 oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);
487
488 if (oa->o_valid & OBD_MD_FLCOOKIE) {
489 if (!oti->oti_logcookies)
490 oti_alloc_cookies(oti, 1);
491 *oti->oti_logcookies = oa->o_lcookie;
492 }
493 }
494
495 CDEBUG(D_HA, "transno: "LPD64"\n",
496 lustre_msg_get_transno(req->rq_repmsg));
497out_req:
498 ptlrpc_req_finished(req);
499out:
500 if (rc && !*ea)
501 obd_free_memmd(exp, &lsm);
502 RETURN(rc);
503}
504
505int osc_punch_base(struct obd_export *exp, struct obd_info *oinfo,
506 obd_enqueue_update_f upcall, void *cookie,
507 struct ptlrpc_request_set *rqset)
508{
509 struct ptlrpc_request *req;
510 struct osc_setattr_args *sa;
511 struct ost_body *body;
512 int rc;
513 ENTRY;
514
515 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_PUNCH);
516 if (req == NULL)
517 RETURN(-ENOMEM);
518
519 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
520 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_PUNCH);
521 if (rc) {
522 ptlrpc_request_free(req);
523 RETURN(rc);
524 }
525 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
526 ptlrpc_at_set_req_timeout(req);
527
528 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
529 LASSERT(body);
530 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
531 osc_pack_capa(req, body, oinfo->oi_capa);
532
533 ptlrpc_request_set_replen(req);
534
535 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_setattr_interpret;
536 CLASSERT (sizeof(*sa) <= sizeof(req->rq_async_args));
537 sa = ptlrpc_req_async_args(req);
538 sa->sa_oa = oinfo->oi_oa;
539 sa->sa_upcall = upcall;
540 sa->sa_cookie = cookie;
541 if (rqset == PTLRPCD_SET)
542 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
543 else
544 ptlrpc_set_add_req(rqset, req);
545
546 RETURN(0);
547}
548
549static int osc_punch(const struct lu_env *env, struct obd_export *exp,
550 struct obd_info *oinfo, struct obd_trans_info *oti,
551 struct ptlrpc_request_set *rqset)
552{
553 oinfo->oi_oa->o_size = oinfo->oi_policy.l_extent.start;
554 oinfo->oi_oa->o_blocks = oinfo->oi_policy.l_extent.end;
555 oinfo->oi_oa->o_valid |= OBD_MD_FLSIZE | OBD_MD_FLBLOCKS;
556 return osc_punch_base(exp, oinfo,
557 oinfo->oi_cb_up, oinfo, rqset);
558}
559
560static int osc_sync_interpret(const struct lu_env *env,
561 struct ptlrpc_request *req,
562 void *arg, int rc)
563{
564 struct osc_fsync_args *fa = arg;
565 struct ost_body *body;
566 ENTRY;
567
568 if (rc)
569 GOTO(out, rc);
570
571 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
572 if (body == NULL) {
573 CERROR ("can't unpack ost_body\n");
574 GOTO(out, rc = -EPROTO);
575 }
576
577 *fa->fa_oi->oi_oa = body->oa;
578out:
579 rc = fa->fa_upcall(fa->fa_cookie, rc);
580 RETURN(rc);
581}
582
583int osc_sync_base(struct obd_export *exp, struct obd_info *oinfo,
584 obd_enqueue_update_f upcall, void *cookie,
585 struct ptlrpc_request_set *rqset)
586{
587 struct ptlrpc_request *req;
588 struct ost_body *body;
589 struct osc_fsync_args *fa;
590 int rc;
591 ENTRY;
592
593 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SYNC);
594 if (req == NULL)
595 RETURN(-ENOMEM);
596
597 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
598 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SYNC);
599 if (rc) {
600 ptlrpc_request_free(req);
601 RETURN(rc);
602 }
603
604 /* overload the size and blocks fields in the oa with start/end */
605 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
606 LASSERT(body);
607 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
608 osc_pack_capa(req, body, oinfo->oi_capa);
609
610 ptlrpc_request_set_replen(req);
611 req->rq_interpret_reply = osc_sync_interpret;
612
613 CLASSERT(sizeof(*fa) <= sizeof(req->rq_async_args));
614 fa = ptlrpc_req_async_args(req);
615 fa->fa_oi = oinfo;
616 fa->fa_upcall = upcall;
617 fa->fa_cookie = cookie;
618
619 if (rqset == PTLRPCD_SET)
620 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
621 else
622 ptlrpc_set_add_req(rqset, req);
623
624 RETURN (0);
625}
626
627static int osc_sync(const struct lu_env *env, struct obd_export *exp,
628 struct obd_info *oinfo, obd_size start, obd_size end,
629 struct ptlrpc_request_set *set)
630{
631 ENTRY;
632
633 if (!oinfo->oi_oa) {
634 CDEBUG(D_INFO, "oa NULL\n");
635 RETURN(-EINVAL);
636 }
637
638 oinfo->oi_oa->o_size = start;
639 oinfo->oi_oa->o_blocks = end;
640 oinfo->oi_oa->o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);
641
642 RETURN(osc_sync_base(exp, oinfo, oinfo->oi_cb_up, oinfo, set));
643}
644
645/* Find and cancel locally locks matched by @mode in the resource found by
646 * @objid. Found locks are added into @cancel list. Returns the amount of
647 * locks added to @cancels list. */
648static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
649 struct list_head *cancels,
650 ldlm_mode_t mode, int lock_flags)
651{
652 struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
653 struct ldlm_res_id res_id;
654 struct ldlm_resource *res;
655 int count;
656 ENTRY;
657
658 /* Return, i.e. cancel nothing, only if ELC is supported (flag in
659 * export) but disabled through procfs (flag in NS).
660 *
661 * This distinguishes from a case when ELC is not supported originally,
662 * when we still want to cancel locks in advance and just cancel them
663 * locally, without sending any RPC. */
664 if (exp_connect_cancelset(exp) && !ns_connect_cancelset(ns))
665 RETURN(0);
666
667 ostid_build_res_name(&oa->o_oi, &res_id);
668 res = ldlm_resource_get(ns, NULL, &res_id, 0, 0);
669 if (res == NULL)
670 RETURN(0);
671
672 LDLM_RESOURCE_ADDREF(res);
673 count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
674 lock_flags, 0, NULL);
675 LDLM_RESOURCE_DELREF(res);
676 ldlm_resource_putref(res);
677 RETURN(count);
678}
679
680static int osc_destroy_interpret(const struct lu_env *env,
681 struct ptlrpc_request *req, void *data,
682 int rc)
683{
684 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
685
686 atomic_dec(&cli->cl_destroy_in_flight);
687 wake_up(&cli->cl_destroy_waitq);
688 return 0;
689}
690
691static int osc_can_send_destroy(struct client_obd *cli)
692{
693 if (atomic_inc_return(&cli->cl_destroy_in_flight) <=
694 cli->cl_max_rpcs_in_flight) {
695 /* The destroy request can be sent */
696 return 1;
697 }
698 if (atomic_dec_return(&cli->cl_destroy_in_flight) <
699 cli->cl_max_rpcs_in_flight) {
700 /*
701 * The counter has been modified between the two atomic
702 * operations.
703 */
704 wake_up(&cli->cl_destroy_waitq);
705 }
706 return 0;
707}
708
709int osc_create(const struct lu_env *env, struct obd_export *exp,
710 struct obdo *oa, struct lov_stripe_md **ea,
711 struct obd_trans_info *oti)
712{
713 int rc = 0;
714 ENTRY;
715
716 LASSERT(oa);
717 LASSERT(ea);
718 LASSERT(oa->o_valid & OBD_MD_FLGROUP);
719
720 if ((oa->o_valid & OBD_MD_FLFLAGS) &&
721 oa->o_flags == OBD_FL_RECREATE_OBJS) {
722 RETURN(osc_real_create(exp, oa, ea, oti));
723 }
724
725 if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
726 RETURN(osc_real_create(exp, oa, ea, oti));
727
728 /* we should not get here anymore */
729 LBUG();
730
731 RETURN(rc);
732}
733
734/* Destroy requests can be async always on the client, and we don't even really
735 * care about the return code since the client cannot do anything at all about
736 * a destroy failure.
737 * When the MDS is unlinking a filename, it saves the file objects into a
738 * recovery llog, and these object records are cancelled when the OST reports
739 * they were destroyed and sync'd to disk (i.e. transaction committed).
740 * If the client dies, or the OST is down when the object should be destroyed,
741 * the records are not cancelled, and when the OST reconnects to the MDS next,
742 * it will retrieve the llog unlink logs and then sends the log cancellation
743 * cookies to the MDS after committing destroy transactions. */
744static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
745 struct obdo *oa, struct lov_stripe_md *ea,
746 struct obd_trans_info *oti, struct obd_export *md_export,
747 void *capa)
748{
749 struct client_obd *cli = &exp->exp_obd->u.cli;
750 struct ptlrpc_request *req;
751 struct ost_body *body;
752 LIST_HEAD(cancels);
753 int rc, count;
754 ENTRY;
755
756 if (!oa) {
757 CDEBUG(D_INFO, "oa NULL\n");
758 RETURN(-EINVAL);
759 }
760
761 count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
762 LDLM_FL_DISCARD_DATA);
763
764 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_DESTROY);
765 if (req == NULL) {
766 ldlm_lock_list_put(&cancels, l_bl_ast, count);
767 RETURN(-ENOMEM);
768 }
769
770 osc_set_capa_size(req, &RMF_CAPA1, (struct obd_capa *)capa);
771 rc = ldlm_prep_elc_req(exp, req, LUSTRE_OST_VERSION, OST_DESTROY,
772 0, &cancels, count);
773 if (rc) {
774 ptlrpc_request_free(req);
775 RETURN(rc);
776 }
777
778 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
779 ptlrpc_at_set_req_timeout(req);
780
781 if (oti != NULL && oa->o_valid & OBD_MD_FLCOOKIE)
782 oa->o_lcookie = *oti->oti_logcookies;
783 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
784 LASSERT(body);
785 lustre_set_wire_obdo(&body->oa, oa);
786
787 osc_pack_capa(req, body, (struct obd_capa *)capa);
788 ptlrpc_request_set_replen(req);
789
790 /* If osc_destory is for destroying the unlink orphan,
791 * sent from MDT to OST, which should not be blocked here,
792 * because the process might be triggered by ptlrpcd, and
793 * it is not good to block ptlrpcd thread (b=16006)*/
794 if (!(oa->o_flags & OBD_FL_DELORPHAN)) {
795 req->rq_interpret_reply = osc_destroy_interpret;
796 if (!osc_can_send_destroy(cli)) {
797 struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP,
798 NULL);
799
800 /*
801 * Wait until the number of on-going destroy RPCs drops
802 * under max_rpc_in_flight
803 */
804 l_wait_event_exclusive(cli->cl_destroy_waitq,
805 osc_can_send_destroy(cli), &lwi);
806 }
807 }
808
809 /* Do not wait for response */
810 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
811 RETURN(0);
812}
813
814static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
815 long writing_bytes)
816{
817 obd_flag bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;
818
819 LASSERT(!(oa->o_valid & bits));
820
821 oa->o_valid |= bits;
822 client_obd_list_lock(&cli->cl_loi_list_lock);
823 oa->o_dirty = cli->cl_dirty;
824 if (unlikely(cli->cl_dirty - cli->cl_dirty_transit >
825 cli->cl_dirty_max)) {
826 CERROR("dirty %lu - %lu > dirty_max %lu\n",
827 cli->cl_dirty, cli->cl_dirty_transit, cli->cl_dirty_max);
828 oa->o_undirty = 0;
c52f69c5 829 } else if (unlikely(atomic_read(&obd_dirty_pages) -
d7e09d03
PT		 830 atomic_read(&obd_dirty_transit_pages) >
831 (long)(obd_max_dirty_pages + 1))) {
832 /* The atomic_read() allowing the atomic_inc() are
833 * not covered by a lock thus they may safely race and trip
834 * this CERROR() unless we add in a small fudge factor (+1). */
c52f69c5 835 CERROR("dirty %d - %d > system dirty_max %d\n",
d7e09d03
PT		 836 atomic_read(&obd_dirty_pages),
837 atomic_read(&obd_dirty_transit_pages),
838 obd_max_dirty_pages);
839 oa->o_undirty = 0;
840 } else if (unlikely(cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff)) {
841 CERROR("dirty %lu - dirty_max %lu too big???\n",
842 cli->cl_dirty, cli->cl_dirty_max);
843 oa->o_undirty = 0;
844 } else {
845 long max_in_flight = (cli->cl_max_pages_per_rpc <<
846 PAGE_CACHE_SHIFT)*
847 (cli->cl_max_rpcs_in_flight + 1);
848 oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
849 }
850 oa->o_grant = cli->cl_avail_grant + cli->cl_reserved_grant;
851 oa->o_dropped = cli->cl_lost_grant;
852 cli->cl_lost_grant = 0;
853 client_obd_list_unlock(&cli->cl_loi_list_lock);
854 CDEBUG(D_CACHE,"dirty: "LPU64" undirty: %u dropped %u grant: "LPU64"\n",
855 oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
856
857}
858
859void osc_update_next_shrink(struct client_obd *cli)
860{
861 cli->cl_next_shrink_grant =
862 cfs_time_shift(cli->cl_grant_shrink_interval);
863 CDEBUG(D_CACHE, "next time %ld to shrink grant \n",
864 cli->cl_next_shrink_grant);
865}
866
867static void __osc_update_grant(struct client_obd *cli, obd_size grant)
868{
869 client_obd_list_lock(&cli->cl_loi_list_lock);
870 cli->cl_avail_grant += grant;
871 client_obd_list_unlock(&cli->cl_loi_list_lock);
872}
873
874static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
875{
876 if (body->oa.o_valid & OBD_MD_FLGRANT) {
877 CDEBUG(D_CACHE, "got "LPU64" extra grant\n", body->oa.o_grant);
878 __osc_update_grant(cli, body->oa.o_grant);
879 }
880}
881
882static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
883 obd_count keylen, void *key, obd_count vallen,
884 void *val, struct ptlrpc_request_set *set);
885
886static int osc_shrink_grant_interpret(const struct lu_env *env,
887 struct ptlrpc_request *req,
888 void *aa, int rc)
889{
890 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
891 struct obdo *oa = ((struct osc_grant_args *)aa)->aa_oa;
892 struct ost_body *body;
893
894 if (rc != 0) {
895 __osc_update_grant(cli, oa->o_grant);
896 GOTO(out, rc);
897 }
898
899 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
900 LASSERT(body);
901 osc_update_grant(cli, body);
902out:
903 OBDO_FREE(oa);
904 return rc;
905}
906
907static void osc_shrink_grant_local(struct client_obd *cli, struct obdo *oa)
908{
909 client_obd_list_lock(&cli->cl_loi_list_lock);
910 oa->o_grant = cli->cl_avail_grant / 4;
911 cli->cl_avail_grant -= oa->o_grant;
912 client_obd_list_unlock(&cli->cl_loi_list_lock);
913 if (!(oa->o_valid & OBD_MD_FLFLAGS)) {
914 oa->o_valid |= OBD_MD_FLFLAGS;
915 oa->o_flags = 0;
916 }
917 oa->o_flags |= OBD_FL_SHRINK_GRANT;
918 osc_update_next_shrink(cli);
919}
920
921/* Shrink the current grant, either from some large amount to enough for a
922 * full set of in-flight RPCs, or if we have already shrunk to that limit
923 * then to enough for a single RPC. This avoids keeping more grant than
924 * needed, and avoids shrinking the grant piecemeal. */
925static int osc_shrink_grant(struct client_obd *cli)
926{
927 __u64 target_bytes = (cli->cl_max_rpcs_in_flight + 1) *
928 (cli->cl_max_pages_per_rpc << PAGE_CACHE_SHIFT);
929
930 client_obd_list_lock(&cli->cl_loi_list_lock);
931 if (cli->cl_avail_grant <= target_bytes)
932 target_bytes = cli->cl_max_pages_per_rpc << PAGE_CACHE_SHIFT;
933 client_obd_list_unlock(&cli->cl_loi_list_lock);
934
935 return osc_shrink_grant_to_target(cli, target_bytes);
936}
937
938int osc_shrink_grant_to_target(struct client_obd *cli, __u64 target_bytes)
939{
940 int rc = 0;
941 struct ost_body *body;
942 ENTRY;
943
944 client_obd_list_lock(&cli->cl_loi_list_lock);
945 /* Don't shrink if we are already above or below the desired limit
946 * We don't want to shrink below a single RPC, as that will negatively
947 * impact block allocation and long-term performance. */
948 if (target_bytes < cli->cl_max_pages_per_rpc << PAGE_CACHE_SHIFT)
949 target_bytes = cli->cl_max_pages_per_rpc << PAGE_CACHE_SHIFT;
950
951 if (target_bytes >= cli->cl_avail_grant) {
952 client_obd_list_unlock(&cli->cl_loi_list_lock);
953 RETURN(0);
954 }
955 client_obd_list_unlock(&cli->cl_loi_list_lock);
956
957 OBD_ALLOC_PTR(body);
958 if (!body)
959 RETURN(-ENOMEM);
960
961 osc_announce_cached(cli, &body->oa, 0);
962
963 client_obd_list_lock(&cli->cl_loi_list_lock);
964 body->oa.o_grant = cli->cl_avail_grant - target_bytes;
965 cli->cl_avail_grant = target_bytes;
966 client_obd_list_unlock(&cli->cl_loi_list_lock);
967 if (!(body->oa.o_valid & OBD_MD_FLFLAGS)) {
968 body->oa.o_valid |= OBD_MD_FLFLAGS;
969 body->oa.o_flags = 0;
970 }
971 body->oa.o_flags |= OBD_FL_SHRINK_GRANT;
972 osc_update_next_shrink(cli);
973
974 rc = osc_set_info_async(NULL, cli->cl_import->imp_obd->obd_self_export,
975 sizeof(KEY_GRANT_SHRINK), KEY_GRANT_SHRINK,
976 sizeof(*body), body, NULL);
977 if (rc != 0)
978 __osc_update_grant(cli, body->oa.o_grant);
979 OBD_FREE_PTR(body);
980 RETURN(rc);
981}
982
983static int osc_should_shrink_grant(struct client_obd *client)
984{
985 cfs_time_t time = cfs_time_current();
986 cfs_time_t next_shrink = client->cl_next_shrink_grant;
987
988 if ((client->cl_import->imp_connect_data.ocd_connect_flags &
989 OBD_CONNECT_GRANT_SHRINK) == 0)
990 return 0;
991
992 if (cfs_time_aftereq(time, next_shrink - 5 * CFS_TICK)) {
993 /* Get the current RPC size directly, instead of going via:
994 * cli_brw_size(obd->u.cli.cl_import->imp_obd->obd_self_export)
995 * Keep comment here so that it can be found by searching. */
996 int brw_size = client->cl_max_pages_per_rpc << PAGE_CACHE_SHIFT;
997
998 if (client->cl_import->imp_state == LUSTRE_IMP_FULL &&
999 client->cl_avail_grant > brw_size)
1000 return 1;
1001 else
1002 osc_update_next_shrink(client);
1003 }
1004 return 0;
1005}
1006
1007static int osc_grant_shrink_grant_cb(struct timeout_item *item, void *data)
1008{
1009 struct client_obd *client;
1010
1011 list_for_each_entry(client, &item->ti_obd_list,
1012 cl_grant_shrink_list) {
1013 if (osc_should_shrink_grant(client))
1014 osc_shrink_grant(client);
1015 }
1016 return 0;
1017}
1018
1019static int osc_add_shrink_grant(struct client_obd *client)
1020{
1021 int rc;
1022
1023 rc = ptlrpc_add_timeout_client(client->cl_grant_shrink_interval,
1024 TIMEOUT_GRANT,
1025 osc_grant_shrink_grant_cb, NULL,
1026 &client->cl_grant_shrink_list);
1027 if (rc) {
1028 CERROR("add grant client %s error %d\n",
1029 client->cl_import->imp_obd->obd_name, rc);
1030 return rc;
1031 }
1032 CDEBUG(D_CACHE, "add grant client %s \n",
1033 client->cl_import->imp_obd->obd_name);
1034 osc_update_next_shrink(client);
1035 return 0;
1036}
1037
1038static int osc_del_shrink_grant(struct client_obd *client)
1039{
1040 return ptlrpc_del_timeout_client(&client->cl_grant_shrink_list,
1041 TIMEOUT_GRANT);
1042}
1043
1044static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
1045{
1046 /*
1047 * ocd_grant is the total grant amount we're expect to hold: if we've
1048 * been evicted, it's the new avail_grant amount, cl_dirty will drop
1049 * to 0 as inflight RPCs fail out; otherwise, it's avail_grant + dirty.
1050 *
1051 * race is tolerable here: if we're evicted, but imp_state already
1052 * left EVICTED state, then cl_dirty must be 0 already.
1053 */
1054 client_obd_list_lock(&cli->cl_loi_list_lock);
1055 if (cli->cl_import->imp_state == LUSTRE_IMP_EVICTED)
1056 cli->cl_avail_grant = ocd->ocd_grant;
1057 else
1058 cli->cl_avail_grant = ocd->ocd_grant - cli->cl_dirty;
1059
1060 if (cli->cl_avail_grant < 0) {
1061 CWARN("%s: available grant < 0: avail/ocd/dirty %ld/%u/%ld\n",
1062 cli->cl_import->imp_obd->obd_name, cli->cl_avail_grant,
1063 ocd->ocd_grant, cli->cl_dirty);
1064 /* workaround for servers which do not have the patch from
1065 * LU-2679 */
1066 cli->cl_avail_grant = ocd->ocd_grant;
1067 }
1068
1069 /* determine the appropriate chunk size used by osc_extent. */
1070 cli->cl_chunkbits = max_t(int, PAGE_CACHE_SHIFT, ocd->ocd_blocksize);
1071 client_obd_list_unlock(&cli->cl_loi_list_lock);
1072
1073 CDEBUG(D_CACHE, "%s, setting cl_avail_grant: %ld cl_lost_grant: %ld."
1074 "chunk bits: %d.\n", cli->cl_import->imp_obd->obd_name,
1075 cli->cl_avail_grant, cli->cl_lost_grant, cli->cl_chunkbits);
1076
1077 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT_SHRINK &&
1078 list_empty(&cli->cl_grant_shrink_list))
1079 osc_add_shrink_grant(cli);
1080}
1081
1082/* We assume that the reason this OSC got a short read is because it read
1083 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
1084 * via the LOV, and it _knows_ it's reading inside the file, it's just that
1085 * this stripe never got written at or beyond this stripe offset yet. */
1086static void handle_short_read(int nob_read, obd_count page_count,
1087 struct brw_page **pga)
1088{
1089 char *ptr;
1090 int i = 0;
1091
1092 /* skip bytes read OK */
1093 while (nob_read > 0) {
1094 LASSERT (page_count > 0);
1095
1096 if (pga[i]->count > nob_read) {
1097 /* EOF inside this page */
1098 ptr = kmap(pga[i]->pg) +
1099 (pga[i]->off & ~CFS_PAGE_MASK);
1100 memset(ptr + nob_read, 0, pga[i]->count - nob_read);
1101 kunmap(pga[i]->pg);
1102 page_count--;
1103 i++;
1104 break;
1105 }
1106
1107 nob_read -= pga[i]->count;
1108 page_count--;
1109 i++;
1110 }
1111
1112 /* zero remaining pages */
1113 while (page_count-- > 0) {
1114 ptr = kmap(pga[i]->pg) + (pga[i]->off & ~CFS_PAGE_MASK);
1115 memset(ptr, 0, pga[i]->count);
1116 kunmap(pga[i]->pg);
1117 i++;
1118 }
1119}
1120
1121static int check_write_rcs(struct ptlrpc_request *req,
1122 int requested_nob, int niocount,
1123 obd_count page_count, struct brw_page **pga)
1124{
1125 int i;
1126 __u32 *remote_rcs;
1127
1128 remote_rcs = req_capsule_server_sized_get(&req->rq_pill, &RMF_RCS,
1129 sizeof(*remote_rcs) *
1130 niocount);
1131 if (remote_rcs == NULL) {
1132 CDEBUG(D_INFO, "Missing/short RC vector on BRW_WRITE reply\n");
1133 return(-EPROTO);
1134 }
1135
1136 /* return error if any niobuf was in error */
1137 for (i = 0; i < niocount; i++) {
1138 if ((int)remote_rcs[i] < 0)
1139 return(remote_rcs[i]);
1140
1141 if (remote_rcs[i] != 0) {
1142 CDEBUG(D_INFO, "rc[%d] invalid (%d) req %p\n",
1143 i, remote_rcs[i], req);
1144 return(-EPROTO);
1145 }
1146 }
1147
1148 if (req->rq_bulk->bd_nob_transferred != requested_nob) {
1149 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
1150 req->rq_bulk->bd_nob_transferred, requested_nob);
1151 return(-EPROTO);
1152 }
1153
1154 return (0);
1155}
1156
1157static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
1158{
1159 if (p1->flag != p2->flag) {
1160 unsigned mask = ~(OBD_BRW_FROM_GRANT| OBD_BRW_NOCACHE|
1161 OBD_BRW_SYNC|OBD_BRW_ASYNC|OBD_BRW_NOQUOTA);
1162
1163 /* warn if we try to combine flags that we don't know to be
1164 * safe to combine */
1165 if (unlikely((p1->flag & mask) != (p2->flag & mask))) {
1166 CWARN("Saw flags 0x%x and 0x%x in the same brw, please "
1167 "report this at http://bugs.whamcloud.com/\n",
1168 p1->flag, p2->flag);
1169 }
1170 return 0;
1171 }
1172
1173 return (p1->off + p1->count == p2->off);
1174}
1175
1176static obd_count osc_checksum_bulk(int nob, obd_count pg_count,
1177 struct brw_page **pga, int opc,
1178 cksum_type_t cksum_type)
1179{
1180 __u32 cksum;
1181 int i = 0;
1182 struct cfs_crypto_hash_desc *hdesc;
1183 unsigned int bufsize;
1184 int err;
1185 unsigned char cfs_alg = cksum_obd2cfs(cksum_type);
1186
1187 LASSERT(pg_count > 0);
1188
1189 hdesc = cfs_crypto_hash_init(cfs_alg, NULL, 0);
1190 if (IS_ERR(hdesc)) {
1191 CERROR("Unable to initialize checksum hash %s\n",
1192 cfs_crypto_hash_name(cfs_alg));
1193 return PTR_ERR(hdesc);
1194 }
1195
1196 while (nob > 0 && pg_count > 0) {
1197 int count = pga[i]->count > nob ? nob : pga[i]->count;
1198
1199 /* corrupt the data before we compute the checksum, to
1200 * simulate an OST->client data error */
1201 if (i == 0 && opc == OST_READ &&
1202 OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE)) {
1203 unsigned char *ptr = kmap(pga[i]->pg);
1204 int off = pga[i]->off & ~CFS_PAGE_MASK;
1205 memcpy(ptr + off, "bad1", min(4, nob));
1206 kunmap(pga[i]->pg);
1207 }
1208 cfs_crypto_hash_update_page(hdesc, pga[i]->pg,
1209 pga[i]->off & ~CFS_PAGE_MASK,
1210 count);
1211 LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d\n",
1212 (int)(pga[i]->off & ~CFS_PAGE_MASK));
1213
1214 nob -= pga[i]->count;
1215 pg_count--;
1216 i++;
1217 }
1218
1219 bufsize = 4;
1220 err = cfs_crypto_hash_final(hdesc, (unsigned char *)&cksum, &bufsize);
1221
1222 if (err)
1223 cfs_crypto_hash_final(hdesc, NULL, NULL);
1224
1225 /* For sending we only compute the wrong checksum instead
1226 * of corrupting the data so it is still correct on a redo */
1227 if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1228 cksum++;
1229
1230 return cksum;
1231}
1232
1233static int osc_brw_prep_request(int cmd, struct client_obd *cli,struct obdo *oa,
1234 struct lov_stripe_md *lsm, obd_count page_count,
1235 struct brw_page **pga,
1236 struct ptlrpc_request **reqp,
1237 struct obd_capa *ocapa, int reserve,
1238 int resend)
1239{
1240 struct ptlrpc_request *req;
1241 struct ptlrpc_bulk_desc *desc;
1242 struct ost_body *body;
1243 struct obd_ioobj *ioobj;
1244 struct niobuf_remote *niobuf;
1245 int niocount, i, requested_nob, opc, rc;
1246 struct osc_brw_async_args *aa;
1247 struct req_capsule *pill;
1248 struct brw_page *pg_prev;
1249
1250 ENTRY;
1251 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ))
1252 RETURN(-ENOMEM); /* Recoverable */
1253 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ2))
1254 RETURN(-EINVAL); /* Fatal */
1255
1256 if ((cmd & OBD_BRW_WRITE) != 0) {
1257 opc = OST_WRITE;
1258 req = ptlrpc_request_alloc_pool(cli->cl_import,
1259 cli->cl_import->imp_rq_pool,
1260 &RQF_OST_BRW_WRITE);
1261 } else {
1262 opc = OST_READ;
1263 req = ptlrpc_request_alloc(cli->cl_import, &RQF_OST_BRW_READ);
1264 }
1265 if (req == NULL)
1266 RETURN(-ENOMEM);
1267
1268 for (niocount = i = 1; i < page_count; i++) {
1269 if (!can_merge_pages(pga[i - 1], pga[i]))
1270 niocount++;
1271 }
1272
1273 pill = &req->rq_pill;
1274 req_capsule_set_size(pill, &RMF_OBD_IOOBJ, RCL_CLIENT,
1275 sizeof(*ioobj));
1276 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_CLIENT,
1277 niocount * sizeof(*niobuf));
1278 osc_set_capa_size(req, &RMF_CAPA1, ocapa);
1279
1280 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, opc);
1281 if (rc) {
1282 ptlrpc_request_free(req);
1283 RETURN(rc);
1284 }
1285 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
1286 ptlrpc_at_set_req_timeout(req);
1287 /* ask ptlrpc not to resend on EINPROGRESS since BRWs have their own
1288 * retry logic */
1289 req->rq_no_retry_einprogress = 1;
1290
1291 desc = ptlrpc_prep_bulk_imp(req, page_count,
1292 cli->cl_import->imp_connect_data.ocd_brw_size >> LNET_MTU_BITS,
1293 opc == OST_WRITE ? BULK_GET_SOURCE : BULK_PUT_SINK,
1294 OST_BULK_PORTAL);
1295
1296 if (desc == NULL)
1297 GOTO(out, rc = -ENOMEM);
1298 /* NB request now owns desc and will free it when it gets freed */
1299
1300 body = req_capsule_client_get(pill, &RMF_OST_BODY);
1301 ioobj = req_capsule_client_get(pill, &RMF_OBD_IOOBJ);
1302 niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1303 LASSERT(body != NULL && ioobj != NULL && niobuf != NULL);
1304
1305 lustre_set_wire_obdo(&body->oa, oa);
1306
1307 obdo_to_ioobj(oa, ioobj);
1308 ioobj->ioo_bufcnt = niocount;
1309 /* The high bits of ioo_max_brw tells server _maximum_ number of bulks
1310 * that might be send for this request. The actual number is decided
1311 * when the RPC is finally sent in ptlrpc_register_bulk(). It sends
1312 * "max - 1" for old client compatibility sending "0", and also so the
1313 * the actual maximum is a power-of-two number, not one less. LU-1431 */
1314 ioobj_max_brw_set(ioobj, desc->bd_md_max_brw);
1315 osc_pack_capa(req, body, ocapa);
1316 LASSERT(page_count > 0);
1317 pg_prev = pga[0];
1318 for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1319 struct brw_page *pg = pga[i];
1320 int poff = pg->off & ~CFS_PAGE_MASK;
1321
1322 LASSERT(pg->count > 0);
1323 /* make sure there is no gap in the middle of page array */
1324 LASSERTF(page_count == 1 ||
1325 (ergo(i == 0, poff + pg->count == PAGE_CACHE_SIZE) &&
1326 ergo(i > 0 && i < page_count - 1,
1327 poff == 0 && pg->count == PAGE_CACHE_SIZE) &&
1328 ergo(i == page_count - 1, poff == 0)),
1329 "i: %d/%d pg: %p off: "LPU64", count: %u\n",
1330 i, page_count, pg, pg->off, pg->count);
1331 LASSERTF(i == 0 || pg->off > pg_prev->off,
1332 "i %d p_c %u pg %p [pri %lu ind %lu] off "LPU64
1333 " prev_pg %p [pri %lu ind %lu] off "LPU64"\n",
1334 i, page_count,
1335 pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1336 pg_prev->pg, page_private(pg_prev->pg),
1337 pg_prev->pg->index, pg_prev->off);
1338 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1339 (pg->flag & OBD_BRW_SRVLOCK));
1340
1341 ptlrpc_prep_bulk_page_pin(desc, pg->pg, poff, pg->count);
1342 requested_nob += pg->count;
1343
1344 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1345 niobuf--;
1346 niobuf->len += pg->count;
1347 } else {
1348 niobuf->offset = pg->off;
1349 niobuf->len = pg->count;
1350 niobuf->flags = pg->flag;
1351 }
1352 pg_prev = pg;
1353 }
1354
1355 LASSERTF((void *)(niobuf - niocount) ==
1356 req_capsule_client_get(&req->rq_pill, &RMF_NIOBUF_REMOTE),
1357 "want %p - real %p\n", req_capsule_client_get(&req->rq_pill,
1358 &RMF_NIOBUF_REMOTE), (void *)(niobuf - niocount));
1359
1360 osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1361 if (resend) {
1362 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1363 body->oa.o_valid |= OBD_MD_FLFLAGS;
1364 body->oa.o_flags = 0;
1365 }
1366 body->oa.o_flags |= OBD_FL_RECOV_RESEND;
1367 }
1368
1369 if (osc_should_shrink_grant(cli))
1370 osc_shrink_grant_local(cli, &body->oa);
1371
1372 /* size[REQ_REC_OFF] still sizeof (*body) */
1373 if (opc == OST_WRITE) {
1374 if (cli->cl_checksum &&
1375 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1376 /* store cl_cksum_type in a local variable since
1377 * it can be changed via lprocfs */
1378 cksum_type_t cksum_type = cli->cl_cksum_type;
1379
1380 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1381 oa->o_flags &= OBD_FL_LOCAL_MASK;
1382 body->oa.o_flags = 0;
1383 }
1384 body->oa.o_flags |= cksum_type_pack(cksum_type);
1385 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1386 body->oa.o_cksum = osc_checksum_bulk(requested_nob,
1387 page_count, pga,
1388 OST_WRITE,
1389 cksum_type);
1390 CDEBUG(D_PAGE, "checksum at write origin: %x\n",
1391 body->oa.o_cksum);
1392 /* save this in 'oa', too, for later checking */
1393 oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1394 oa->o_flags |= cksum_type_pack(cksum_type);
1395 } else {
1396 /* clear out the checksum flag, in case this is a
1397 * resend but cl_checksum is no longer set. b=11238 */
1398 oa->o_valid &= ~OBD_MD_FLCKSUM;
1399 }
1400 oa->o_cksum = body->oa.o_cksum;
1401 /* 1 RC per niobuf */
1402 req_capsule_set_size(pill, &RMF_RCS, RCL_SERVER,
1403 sizeof(__u32) * niocount);
1404 } else {
1405 if (cli->cl_checksum &&
1406 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1407 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1408 body->oa.o_flags = 0;
1409 body->oa.o_flags |= cksum_type_pack(cli->cl_cksum_type);
1410 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1411 }
1412 }
1413 ptlrpc_request_set_replen(req);
1414
1415 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
1416 aa = ptlrpc_req_async_args(req);
1417 aa->aa_oa = oa;
1418 aa->aa_requested_nob = requested_nob;
1419 aa->aa_nio_count = niocount;
1420 aa->aa_page_count = page_count;
1421 aa->aa_resends = 0;
1422 aa->aa_ppga = pga;
1423 aa->aa_cli = cli;
1424 INIT_LIST_HEAD(&aa->aa_oaps);
1425 if (ocapa && reserve)
1426 aa->aa_ocapa = capa_get(ocapa);
1427
1428 *reqp = req;
1429 RETURN(0);
1430
1431 out:
1432 ptlrpc_req_finished(req);
1433 RETURN(rc);
1434}
1435
1436static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
1437 __u32 client_cksum, __u32 server_cksum, int nob,
1438 obd_count page_count, struct brw_page **pga,
1439 cksum_type_t client_cksum_type)
1440{
1441 __u32 new_cksum;
1442 char *msg;
1443 cksum_type_t cksum_type;
1444
1445 if (server_cksum == client_cksum) {
1446 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1447 return 0;
1448 }
1449
1450 cksum_type = cksum_type_unpack(oa->o_valid & OBD_MD_FLFLAGS ?
1451 oa->o_flags : 0);
1452 new_cksum = osc_checksum_bulk(nob, page_count, pga, OST_WRITE,
1453 cksum_type);
1454
1455 if (cksum_type != client_cksum_type)
1456 msg = "the server did not use the checksum type specified in "
1457 "the original request - likely a protocol problem";
1458 else if (new_cksum == server_cksum)
1459 msg = "changed on the client after we checksummed it - "
1460 "likely false positive due to mmap IO (bug 11742)";
1461 else if (new_cksum == client_cksum)
1462 msg = "changed in transit before arrival at OST";
1463 else
1464 msg = "changed in transit AND doesn't match the original - "
1465 "likely false positive due to mmap IO (bug 11742)";
1466
1467 LCONSOLE_ERROR_MSG(0x132, "BAD WRITE CHECKSUM: %s: from %s inode "DFID
1468 " object "DOSTID" extent ["LPU64"-"LPU64"]\n",
1469 msg, libcfs_nid2str(peer->nid),
1470 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_seq : (__u64)0,
1471 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_oid : 0,
1472 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_ver : 0,
1473 POSTID(&oa->o_oi), pga[0]->off,
1474 pga[page_count-1]->off + pga[page_count-1]->count - 1);
1475 CERROR("original client csum %x (type %x), server csum %x (type %x), "
1476 "client csum now %x\n", client_cksum, client_cksum_type,
1477 server_cksum, cksum_type, new_cksum);
1478 return 1;
1479}
1480
1481/* Note rc enters this function as number of bytes transferred */
1482static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
1483{
1484 struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
1485 const lnet_process_id_t *peer =
1486 &req->rq_import->imp_connection->c_peer;
1487 struct client_obd *cli = aa->aa_cli;
1488 struct ost_body *body;
1489 __u32 client_cksum = 0;
1490 ENTRY;
1491
1492 if (rc < 0 && rc != -EDQUOT) {
1493 DEBUG_REQ(D_INFO, req, "Failed request with rc = %d\n", rc);
1494 RETURN(rc);
1495 }
1496
1497 LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
1498 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
1499 if (body == NULL) {
1500 DEBUG_REQ(D_INFO, req, "Can't unpack body\n");
1501 RETURN(-EPROTO);
1502 }
1503
1504 /* set/clear over quota flag for a uid/gid */
1505 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
1506 body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA)) {
1507 unsigned int qid[MAXQUOTAS] = { body->oa.o_uid, body->oa.o_gid };
1508
1509 CDEBUG(D_QUOTA, "setdq for [%u %u] with valid "LPX64", flags %x\n",
1510 body->oa.o_uid, body->oa.o_gid, body->oa.o_valid,
1511 body->oa.o_flags);
1512 osc_quota_setdq(cli, qid, body->oa.o_valid, body->oa.o_flags);
1513 }
1514
1515 osc_update_grant(cli, body);
1516
1517 if (rc < 0)
1518 RETURN(rc);
1519
1520 if (aa->aa_oa->o_valid & OBD_MD_FLCKSUM)
1521 client_cksum = aa->aa_oa->o_cksum; /* save for later */
1522
1523 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1524 if (rc > 0) {
1525 CERROR("Unexpected +ve rc %d\n", rc);
1526 RETURN(-EPROTO);
1527 }
1528 LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);
1529
1530 if (sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
1531 RETURN(-EAGAIN);
1532
1533 if ((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) && client_cksum &&
1534 check_write_checksum(&body->oa, peer, client_cksum,
1535 body->oa.o_cksum, aa->aa_requested_nob,
1536 aa->aa_page_count, aa->aa_ppga,
1537 cksum_type_unpack(aa->aa_oa->o_flags)))
1538 RETURN(-EAGAIN);
1539
1540 rc = check_write_rcs(req, aa->aa_requested_nob,aa->aa_nio_count,
1541 aa->aa_page_count, aa->aa_ppga);
1542 GOTO(out, rc);
1543 }
1544
1545 /* The rest of this function executes only for OST_READs */
1546
1547 /* if unwrap_bulk failed, return -EAGAIN to retry */
1548 rc = sptlrpc_cli_unwrap_bulk_read(req, req->rq_bulk, rc);
1549 if (rc < 0)
1550 GOTO(out, rc = -EAGAIN);
1551
1552 if (rc > aa->aa_requested_nob) {
1553 CERROR("Unexpected rc %d (%d requested)\n", rc,
1554 aa->aa_requested_nob);
1555 RETURN(-EPROTO);
1556 }
1557
1558 if (rc != req->rq_bulk->bd_nob_transferred) {
1559 CERROR ("Unexpected rc %d (%d transferred)\n",
1560 rc, req->rq_bulk->bd_nob_transferred);
1561 return (-EPROTO);
1562 }
1563
1564 if (rc < aa->aa_requested_nob)
1565 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
1566
1567 if (body->oa.o_valid & OBD_MD_FLCKSUM) {
1568 static int cksum_counter;
1569 __u32 server_cksum = body->oa.o_cksum;
1570 char *via;
1571 char *router;
1572 cksum_type_t cksum_type;
1573
1574 cksum_type = cksum_type_unpack(body->oa.o_valid &OBD_MD_FLFLAGS?
1575 body->oa.o_flags : 0);
1576 client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
1577 aa->aa_ppga, OST_READ,
1578 cksum_type);
1579
1580 if (peer->nid == req->rq_bulk->bd_sender) {
1581 via = router = "";
1582 } else {
1583 via = " via ";
1584 router = libcfs_nid2str(req->rq_bulk->bd_sender);
1585 }
1586
1587 if (server_cksum == ~0 && rc > 0) {
1588 CERROR("Protocol error: server %s set the 'checksum' "
1589 "bit, but didn't send a checksum. Not fatal, "
1590 "but please notify on http://bugs.whamcloud.com/\n",
1591 libcfs_nid2str(peer->nid));
1592 } else if (server_cksum != client_cksum) {
1593 LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from "
1594 "%s%s%s inode "DFID" object "DOSTID
1595 " extent ["LPU64"-"LPU64"]\n",
1596 req->rq_import->imp_obd->obd_name,
1597 libcfs_nid2str(peer->nid),
1598 via, router,
1599 body->oa.o_valid & OBD_MD_FLFID ?
1600 body->oa.o_parent_seq : (__u64)0,
1601 body->oa.o_valid & OBD_MD_FLFID ?
1602 body->oa.o_parent_oid : 0,
1603 body->oa.o_valid & OBD_MD_FLFID ?
1604 body->oa.o_parent_ver : 0,
1605 POSTID(&body->oa.o_oi),
1606 aa->aa_ppga[0]->off,
1607 aa->aa_ppga[aa->aa_page_count-1]->off +
1608 aa->aa_ppga[aa->aa_page_count-1]->count -
1609 1);
1610 CERROR("client %x, server %x, cksum_type %x\n",
1611 client_cksum, server_cksum, cksum_type);
1612 cksum_counter = 0;
1613 aa->aa_oa->o_cksum = client_cksum;
1614 rc = -EAGAIN;
1615 } else {
1616 cksum_counter++;
1617 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1618 rc = 0;
1619 }
1620 } else if (unlikely(client_cksum)) {
1621 static int cksum_missed;
1622
1623 cksum_missed++;
1624 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
1625 CERROR("Checksum %u requested from %s but not sent\n",
1626 cksum_missed, libcfs_nid2str(peer->nid));
1627 } else {
1628 rc = 0;
1629 }
1630out:
1631 if (rc >= 0)
1632 lustre_get_wire_obdo(aa->aa_oa, &body->oa);
1633
1634 RETURN(rc);
1635}
1636
1637static int osc_brw_internal(int cmd, struct obd_export *exp, struct obdo *oa,
1638 struct lov_stripe_md *lsm,
1639 obd_count page_count, struct brw_page **pga,
1640 struct obd_capa *ocapa)
1641{
1642 struct ptlrpc_request *req;
1643 int rc;
1644 wait_queue_head_t waitq;
1645 int generation, resends = 0;
1646 struct l_wait_info lwi;
1647
1648 ENTRY;
1649
1650 init_waitqueue_head(&waitq);
1651 generation = exp->exp_obd->u.cli.cl_import->imp_generation;
1652
1653restart_bulk:
1654 rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
1655 page_count, pga, &req, ocapa, 0, resends);
1656 if (rc != 0)
1657 return (rc);
1658
1659 if (resends) {
1660 req->rq_generation_set = 1;
1661 req->rq_import_generation = generation;
1662 req->rq_sent = cfs_time_current_sec() + resends;
1663 }
1664
1665 rc = ptlrpc_queue_wait(req);
1666
1667 if (rc == -ETIMEDOUT && req->rq_resend) {
1668 DEBUG_REQ(D_HA, req, "BULK TIMEOUT");
1669 ptlrpc_req_finished(req);
1670 goto restart_bulk;
1671 }
1672
1673 rc = osc_brw_fini_request(req, rc);
1674
1675 ptlrpc_req_finished(req);
1676 /* When server return -EINPROGRESS, client should always retry
1677 * regardless of the number of times the bulk was resent already.*/
1678 if (osc_recoverable_error(rc)) {
1679 resends++;
1680 if (rc != -EINPROGRESS &&
1681 !client_should_resend(resends, &exp->exp_obd->u.cli)) {
1682 CERROR("%s: too many resend retries for object: "
1683 ""DOSTID", rc = %d.\n", exp->exp_obd->obd_name,
1684 POSTID(&oa->o_oi), rc);
1685 goto out;
1686 }
1687 if (generation !=
1688 exp->exp_obd->u.cli.cl_import->imp_generation) {
1689 CDEBUG(D_HA, "%s: resend cross eviction for object: "
1690 ""DOSTID", rc = %d.\n", exp->exp_obd->obd_name,
1691 POSTID(&oa->o_oi), rc);
1692 goto out;
1693 }
1694
1695 lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(resends), NULL, NULL,
1696 NULL);
1697 l_wait_event(waitq, 0, &lwi);
1698
1699 goto restart_bulk;
1700 }
1701out:
1702 if (rc == -EAGAIN || rc == -EINPROGRESS)
1703 rc = -EIO;
1704 RETURN (rc);
1705}
1706
1707static int osc_brw_redo_request(struct ptlrpc_request *request,
1708 struct osc_brw_async_args *aa, int rc)
1709{
1710 struct ptlrpc_request *new_req;
1711 struct osc_brw_async_args *new_aa;
1712 struct osc_async_page *oap;
1713 ENTRY;
1714
1715 DEBUG_REQ(rc == -EINPROGRESS ? D_RPCTRACE : D_ERROR, request,
1716 "redo for recoverable error %d", rc);
1717
1718 rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
1719 OST_WRITE ? OBD_BRW_WRITE :OBD_BRW_READ,
1720 aa->aa_cli, aa->aa_oa,
1721 NULL /* lsm unused by osc currently */,
1722 aa->aa_page_count, aa->aa_ppga,
1723 &new_req, aa->aa_ocapa, 0, 1);
1724 if (rc)
1725 RETURN(rc);
1726
1727 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
1728 if (oap->oap_request != NULL) {
1729 LASSERTF(request == oap->oap_request,
1730 "request %p != oap_request %p\n",
1731 request, oap->oap_request);
1732 if (oap->oap_interrupted) {
1733 ptlrpc_req_finished(new_req);
1734 RETURN(-EINTR);
1735 }
1736 }
1737 }
1738 /* New request takes over pga and oaps from old request.
1739 * Note that copying a list_head doesn't work, need to move it... */
1740 aa->aa_resends++;
1741 new_req->rq_interpret_reply = request->rq_interpret_reply;
1742 new_req->rq_async_args = request->rq_async_args;
d7e09d03
PT		 1743 /* cap resend delay to the current request timeout, this is similar to
1744 * what ptlrpc does (see after_reply()) */
1745 if (aa->aa_resends > new_req->rq_timeout)
1746 new_req->rq_sent = cfs_time_current_sec() + new_req->rq_timeout;
1747 else
1748 new_req->rq_sent = cfs_time_current_sec() + aa->aa_resends;
1749 new_req->rq_generation_set = 1;
1750 new_req->rq_import_generation = request->rq_import_generation;
1751
1752 new_aa = ptlrpc_req_async_args(new_req);
1753
1754 INIT_LIST_HEAD(&new_aa->aa_oaps);
1755 list_splice_init(&aa->aa_oaps, &new_aa->aa_oaps);
1756 INIT_LIST_HEAD(&new_aa->aa_exts);
1757 list_splice_init(&aa->aa_exts, &new_aa->aa_exts);
1758 new_aa->aa_resends = aa->aa_resends;
1759
1760 list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
1761 if (oap->oap_request) {
1762 ptlrpc_req_finished(oap->oap_request);
1763 oap->oap_request = ptlrpc_request_addref(new_req);
1764 }
1765 }
1766
1767 new_aa->aa_ocapa = aa->aa_ocapa;
1768 aa->aa_ocapa = NULL;
1769
1770 /* XXX: This code will run into problem if we're going to support
1771 * to add a series of BRW RPCs into a self-defined ptlrpc_request_set
1772 * and wait for all of them to be finished. We should inherit request
1773 * set from old request. */
1774 ptlrpcd_add_req(new_req, PDL_POLICY_SAME, -1);
1775
1776 DEBUG_REQ(D_INFO, new_req, "new request");
1777 RETURN(0);
1778}
1779
1780/*
1781 * ugh, we want disk allocation on the target to happen in offset order. we'll
1782 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
1783 * fine for our small page arrays and doesn't require allocation. its an
1784 * insertion sort that swaps elements that are strides apart, shrinking the
1785 * stride down until its '1' and the array is sorted.
1786 */
1787static void sort_brw_pages(struct brw_page **array, int num)
1788{
1789 int stride, i, j;
1790 struct brw_page *tmp;
1791
1792 if (num == 1)
1793 return;
1794 for (stride = 1; stride < num ; stride = (stride * 3) + 1)
1795 ;
1796
1797 do {
1798 stride /= 3;
1799 for (i = stride ; i < num ; i++) {
1800 tmp = array[i];
1801 j = i;
1802 while (j >= stride && array[j - stride]->off > tmp->off) {
1803 array[j] = array[j - stride];
1804 j -= stride;
1805 }
1806 array[j] = tmp;
1807 }
1808 } while (stride > 1);
1809}
1810
1811static obd_count max_unfragmented_pages(struct brw_page **pg, obd_count pages)
1812{
1813 int count = 1;
1814 int offset;
1815 int i = 0;
1816
1817 LASSERT (pages > 0);
1818 offset = pg[i]->off & ~CFS_PAGE_MASK;
1819
1820 for (;;) {
1821 pages--;
1822 if (pages == 0) /* that's all */
1823 return count;
1824
1825 if (offset + pg[i]->count < PAGE_CACHE_SIZE)
1826 return count; /* doesn't end on page boundary */
1827
1828 i++;
1829 offset = pg[i]->off & ~CFS_PAGE_MASK;
1830 if (offset != 0) /* doesn't start on page boundary */
1831 return count;
1832
1833 count++;
1834 }
1835}
1836
1837static struct brw_page **osc_build_ppga(struct brw_page *pga, obd_count count)
1838{
1839 struct brw_page **ppga;
1840 int i;
1841
1842 OBD_ALLOC(ppga, sizeof(*ppga) * count);
1843 if (ppga == NULL)
1844 return NULL;
1845
1846 for (i = 0; i < count; i++)
1847 ppga[i] = pga + i;
1848 return ppga;
1849}
1850
1851static void osc_release_ppga(struct brw_page **ppga, obd_count count)
1852{
1853 LASSERT(ppga != NULL);
1854 OBD_FREE(ppga, sizeof(*ppga) * count);
1855}
1856
1857static int osc_brw(int cmd, struct obd_export *exp, struct obd_info *oinfo,
1858 obd_count page_count, struct brw_page *pga,
1859 struct obd_trans_info *oti)
1860{
1861 struct obdo *saved_oa = NULL;
1862 struct brw_page **ppga, **orig;
1863 struct obd_import *imp = class_exp2cliimp(exp);
1864 struct client_obd *cli;
1865 int rc, page_count_orig;
1866 ENTRY;
1867
1868 LASSERT((imp != NULL) && (imp->imp_obd != NULL));
1869 cli = &imp->imp_obd->u.cli;
1870
1871 if (cmd & OBD_BRW_CHECK) {
1872 /* The caller just wants to know if there's a chance that this
1873 * I/O can succeed */
1874
1875 if (imp->imp_invalid)
1876 RETURN(-EIO);
1877 RETURN(0);
1878 }
1879
1880 /* test_brw with a failed create can trip this, maybe others. */
1881 LASSERT(cli->cl_max_pages_per_rpc);
1882
1883 rc = 0;
1884
1885 orig = ppga = osc_build_ppga(pga, page_count);
1886 if (ppga == NULL)
1887 RETURN(-ENOMEM);
1888 page_count_orig = page_count;
1889
1890 sort_brw_pages(ppga, page_count);
1891 while (page_count) {
1892 obd_count pages_per_brw;
1893
1894 if (page_count > cli->cl_max_pages_per_rpc)
1895 pages_per_brw = cli->cl_max_pages_per_rpc;
1896 else
1897 pages_per_brw = page_count;
1898
1899 pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);
1900
1901 if (saved_oa != NULL) {
1902 /* restore previously saved oa */
1903 *oinfo->oi_oa = *saved_oa;
1904 } else if (page_count > pages_per_brw) {
1905 /* save a copy of oa (brw will clobber it) */
1906 OBDO_ALLOC(saved_oa);
1907 if (saved_oa == NULL)
1908 GOTO(out, rc = -ENOMEM);
1909 *saved_oa = *oinfo->oi_oa;
1910 }
1911
1912 rc = osc_brw_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
1913 pages_per_brw, ppga, oinfo->oi_capa);
1914
1915 if (rc != 0)
1916 break;
1917
1918 page_count -= pages_per_brw;
1919 ppga += pages_per_brw;
1920 }
1921
1922out:
1923 osc_release_ppga(orig, page_count_orig);
1924
1925 if (saved_oa != NULL)
1926 OBDO_FREE(saved_oa);
1927
1928 RETURN(rc);
1929}
1930
1931static int brw_interpret(const struct lu_env *env,
1932 struct ptlrpc_request *req, void *data, int rc)
1933{
1934 struct osc_brw_async_args *aa = data;
1935 struct osc_extent *ext;
1936 struct osc_extent *tmp;
1937 struct cl_object *obj = NULL;
1938 struct client_obd *cli = aa->aa_cli;
1939 ENTRY;
1940
1941 rc = osc_brw_fini_request(req, rc);
1942 CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
1943 /* When server return -EINPROGRESS, client should always retry
1944 * regardless of the number of times the bulk was resent already. */
1945 if (osc_recoverable_error(rc)) {
1946 if (req->rq_import_generation !=
1947 req->rq_import->imp_generation) {
1948 CDEBUG(D_HA, "%s: resend cross eviction for object: "
1949 ""DOSTID", rc = %d.\n",
1950 req->rq_import->imp_obd->obd_name,
1951 POSTID(&aa->aa_oa->o_oi), rc);
1952 } else if (rc == -EINPROGRESS ||
1953 client_should_resend(aa->aa_resends, aa->aa_cli)) {
1954 rc = osc_brw_redo_request(req, aa, rc);
1955 } else {
1956 CERROR("%s: too many resent retries for object: "
1957 ""LPU64":"LPU64", rc = %d.\n",
1958 req->rq_import->imp_obd->obd_name,
1959 POSTID(&aa->aa_oa->o_oi), rc);
1960 }
1961
1962 if (rc == 0)
1963 RETURN(0);
1964 else if (rc == -EAGAIN || rc == -EINPROGRESS)
1965 rc = -EIO;
1966 }
1967
1968 if (aa->aa_ocapa) {
1969 capa_put(aa->aa_ocapa);
1970 aa->aa_ocapa = NULL;
1971 }
1972
1973 list_for_each_entry_safe(ext, tmp, &aa->aa_exts, oe_link) {
1974 if (obj == NULL && rc == 0) {
1975 obj = osc2cl(ext->oe_obj);
1976 cl_object_get(obj);
1977 }
1978
1979 list_del_init(&ext->oe_link);
1980 osc_extent_finish(env, ext, 1, rc);
1981 }
1982 LASSERT(list_empty(&aa->aa_exts));
1983 LASSERT(list_empty(&aa->aa_oaps));
1984
1985 if (obj != NULL) {
1986 struct obdo *oa = aa->aa_oa;
1987 struct cl_attr *attr = &osc_env_info(env)->oti_attr;
1988 unsigned long valid = 0;
1989
1990 LASSERT(rc == 0);
1991 if (oa->o_valid & OBD_MD_FLBLOCKS) {
1992 attr->cat_blocks = oa->o_blocks;
1993 valid |= CAT_BLOCKS;
1994 }
1995 if (oa->o_valid & OBD_MD_FLMTIME) {
1996 attr->cat_mtime = oa->o_mtime;
1997 valid |= CAT_MTIME;
1998 }
1999 if (oa->o_valid & OBD_MD_FLATIME) {
2000 attr->cat_atime = oa->o_atime;
2001 valid |= CAT_ATIME;
2002 }
2003 if (oa->o_valid & OBD_MD_FLCTIME) {
2004 attr->cat_ctime = oa->o_ctime;
2005 valid |= CAT_CTIME;
2006 }
2007 if (valid != 0) {
2008 cl_object_attr_lock(obj);
2009 cl_object_attr_set(env, obj, attr, valid);
2010 cl_object_attr_unlock(obj);
2011 }
2012 cl_object_put(env, obj);
2013 }
2014 OBDO_FREE(aa->aa_oa);
2015
2016 cl_req_completion(env, aa->aa_clerq, rc < 0 ? rc :
2017 req->rq_bulk->bd_nob_transferred);
2018 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
2019 ptlrpc_lprocfs_brw(req, req->rq_bulk->bd_nob_transferred);
2020
2021 client_obd_list_lock(&cli->cl_loi_list_lock);
2022 /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
2023 * is called so we know whether to go to sync BRWs or wait for more
2024 * RPCs to complete */
2025 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
2026 cli->cl_w_in_flight--;
2027 else
2028 cli->cl_r_in_flight--;
2029 osc_wake_cache_waiters(cli);
2030 client_obd_list_unlock(&cli->cl_loi_list_lock);
2031
2032 osc_io_unplug(env, cli, NULL, PDL_POLICY_SAME);
2033 RETURN(rc);
2034}
2035
d7e09d03
PT		 2036/**
2037 * Build an RPC by the list of extent @ext_list. The caller must ensure
2038 * that the total pages in this list are NOT over max pages per RPC.
2039 * Extents in the list must be in OES_RPC state.
2040 */
2041int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
2042 struct list_head *ext_list, int cmd, pdl_policy_t pol)
2043{
cad6fafa
BJ		 2044 struct ptlrpc_request *req = NULL;
2045 struct osc_extent *ext;
2046 struct brw_page **pga = NULL;
2047 struct osc_brw_async_args *aa = NULL;
2048 struct obdo *oa = NULL;
2049 struct osc_async_page *oap;
2050 struct osc_async_page *tmp;
2051 struct cl_req *clerq = NULL;
2052 enum cl_req_type crt = (cmd & OBD_BRW_WRITE) ? CRT_WRITE :
2053 CRT_READ;
2054 struct ldlm_lock *lock = NULL;
2055 struct cl_req_attr *crattr = NULL;
2056 obd_off starting_offset = OBD_OBJECT_EOF;
2057 obd_off ending_offset = 0;
2058 int mpflag = 0;
2059 int mem_tight = 0;
2060 int page_count = 0;
2061 int i;
2062 int rc;
d7e09d03 2063 LIST_HEAD(rpc_list);
d7e09d03
PT		 2064
2065 ENTRY;
2066 LASSERT(!list_empty(ext_list));
2067
2068 /* add pages into rpc_list to build BRW rpc */
2069 list_for_each_entry(ext, ext_list, oe_link) {
2070 LASSERT(ext->oe_state == OES_RPC);
2071 mem_tight |= ext->oe_memalloc;
2072 list_for_each_entry(oap, &ext->oe_pages, oap_pending_item) {
2073 ++page_count;
2074 list_add_tail(&oap->oap_rpc_item, &rpc_list);
2075 if (starting_offset > oap->oap_obj_off)
2076 starting_offset = oap->oap_obj_off;
2077 else
2078 LASSERT(oap->oap_page_off == 0);
2079 if (ending_offset < oap->oap_obj_off + oap->oap_count)
2080 ending_offset = oap->oap_obj_off +
2081 oap->oap_count;
2082 else
2083 LASSERT(oap->oap_page_off + oap->oap_count ==
2084 PAGE_CACHE_SIZE);
2085 }
2086 }
2087
2088 if (mem_tight)
2089 mpflag = cfs_memory_pressure_get_and_set();
2090
cad6fafa
BJ		 2091 OBD_ALLOC(crattr, sizeof(*crattr));
2092 if (crattr == NULL)
2093 GOTO(out, rc = -ENOMEM);
2094
d7e09d03
PT		 2095 OBD_ALLOC(pga, sizeof(*pga) * page_count);
2096 if (pga == NULL)
2097 GOTO(out, rc = -ENOMEM);
2098
2099 OBDO_ALLOC(oa);
2100 if (oa == NULL)
2101 GOTO(out, rc = -ENOMEM);
2102
2103 i = 0;
2104 list_for_each_entry(oap, &rpc_list, oap_rpc_item) {
2105 struct cl_page *page = oap2cl_page(oap);
2106 if (clerq == NULL) {
2107 clerq = cl_req_alloc(env, page, crt,
cad6fafa 2108 1 /* only 1-object rpcs for now */);
d7e09d03
PT		 2109 if (IS_ERR(clerq))
2110 GOTO(out, rc = PTR_ERR(clerq));
2111 lock = oap->oap_ldlm_lock;
2112 }
2113 if (mem_tight)
2114 oap->oap_brw_flags |= OBD_BRW_MEMALLOC;
2115 pga[i] = &oap->oap_brw_page;
2116 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
2117 CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
cad6fafa
BJ		 2118 pga[i]->pg, page_index(oap->oap_page), oap,
2119 pga[i]->flag);
d7e09d03
PT		 2120 i++;
2121 cl_req_page_add(env, clerq, page);
2122 }
2123
2124 /* always get the data for the obdo for the rpc */
2125 LASSERT(clerq != NULL);
cad6fafa
BJ		 2126 crattr->cra_oa = oa;
2127 cl_req_attr_set(env, clerq, crattr, ~0ULL);
d7e09d03
PT		 2128 if (lock) {
2129 oa->o_handle = lock->l_remote_handle;
2130 oa->o_valid |= OBD_MD_FLHANDLE;
2131 }
2132
2133 rc = cl_req_prep(env, clerq);
2134 if (rc != 0) {
2135 CERROR("cl_req_prep failed: %d\n", rc);
2136 GOTO(out, rc);
2137 }
2138
2139 sort_brw_pages(pga, page_count);
2140 rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count,
cad6fafa 2141 pga, &req, crattr->cra_capa, 1, 0);
d7e09d03
PT		 2142 if (rc != 0) {
2143 CERROR("prep_req failed: %d\n", rc);
2144 GOTO(out, rc);
2145 }
2146
d7e09d03
PT		 2147 req->rq_interpret_reply = brw_interpret;
2148
2149 if (mem_tight != 0)
2150 req->rq_memalloc = 1;
2151
2152 /* Need to update the timestamps after the request is built in case
2153 * we race with setattr (locally or in queue at OST). If OST gets
2154 * later setattr before earlier BRW (as determined by the request xid),
2155 * the OST will not use BRW timestamps. Sadly, there is no obvious
2156 * way to do this in a single call. bug 10150 */
cad6fafa 2157 cl_req_attr_set(env, clerq, crattr,
d7e09d03
PT		 2158 OBD_MD_FLMTIME|OBD_MD_FLCTIME|OBD_MD_FLATIME);
2159
cad6fafa 2160 lustre_msg_set_jobid(req->rq_reqmsg, crattr->cra_jobid);
d7e09d03
PT		 2161
2162 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2163 aa = ptlrpc_req_async_args(req);
2164 INIT_LIST_HEAD(&aa->aa_oaps);
2165 list_splice_init(&rpc_list, &aa->aa_oaps);
2166 INIT_LIST_HEAD(&aa->aa_exts);
2167 list_splice_init(ext_list, &aa->aa_exts);
2168 aa->aa_clerq = clerq;
2169
2170 /* queued sync pages can be torn down while the pages
2171 * were between the pending list and the rpc */
2172 tmp = NULL;
2173 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2174 /* only one oap gets a request reference */
2175 if (tmp == NULL)
2176 tmp = oap;
2177 if (oap->oap_interrupted && !req->rq_intr) {
2178 CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
2179 oap, req);
2180 ptlrpc_mark_interrupted(req);
2181 }
2182 }
2183 if (tmp != NULL)
2184 tmp->oap_request = ptlrpc_request_addref(req);
2185
2186 client_obd_list_lock(&cli->cl_loi_list_lock);
2187 starting_offset >>= PAGE_CACHE_SHIFT;
2188 if (cmd == OBD_BRW_READ) {
2189 cli->cl_r_in_flight++;
2190 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
2191 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
2192 lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
2193 starting_offset + 1);
2194 } else {
2195 cli->cl_w_in_flight++;
2196 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
2197 lprocfs_oh_tally(&cli->cl_write_rpc_hist, cli->cl_w_in_flight);
2198 lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
2199 starting_offset + 1);
2200 }
2201 client_obd_list_unlock(&cli->cl_loi_list_lock);
2202
2203 DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
2204 page_count, aa, cli->cl_r_in_flight,
2205 cli->cl_w_in_flight);
2206
2207 /* XXX: Maybe the caller can check the RPC bulk descriptor to
2208 * see which CPU/NUMA node the majority of pages were allocated
2209 * on, and try to assign the async RPC to the CPU core
2210 * (PDL_POLICY_PREFERRED) to reduce cross-CPU memory traffic.
2211 *
2212 * But on the other hand, we expect that multiple ptlrpcd
2213 * threads and the initial write sponsor can run in parallel,
2214 * especially when data checksum is enabled, which is CPU-bound
2215 * operation and single ptlrpcd thread cannot process in time.
2216 * So more ptlrpcd threads sharing BRW load
2217 * (with PDL_POLICY_ROUND) seems better.
2218 */
2219 ptlrpcd_add_req(req, pol, -1);
2220 rc = 0;
2221 EXIT;
2222
2223out:
2224 if (mem_tight != 0)
2225 cfs_memory_pressure_restore(mpflag);
2226
cad6fafa
BJ		 2227 if (crattr != NULL) {
2228 capa_put(crattr->cra_capa);
2229 OBD_FREE(crattr, sizeof(*crattr));
2230 }
2231
d7e09d03
PT		 2232 if (rc != 0) {
2233 LASSERT(req == NULL);
2234
2235 if (oa)
2236 OBDO_FREE(oa);
2237 if (pga)
2238 OBD_FREE(pga, sizeof(*pga) * page_count);
2239 /* this should happen rarely and is pretty bad, it makes the
2240 * pending list not follow the dirty order */
2241 while (!list_empty(ext_list)) {
2242 ext = list_entry(ext_list->next, struct osc_extent,
2243 oe_link);
2244 list_del_init(&ext->oe_link);
2245 osc_extent_finish(env, ext, 0, rc);
2246 }
2247 if (clerq && !IS_ERR(clerq))
2248 cl_req_completion(env, clerq, rc);
2249 }
2250 RETURN(rc);
2251}
2252
2253static int osc_set_lock_data_with_check(struct ldlm_lock *lock,
2254 struct ldlm_enqueue_info *einfo)
2255{
2256 void *data = einfo->ei_cbdata;
2257 int set = 0;
2258
2259 LASSERT(lock != NULL);
2260 LASSERT(lock->l_blocking_ast == einfo->ei_cb_bl);
2261 LASSERT(lock->l_resource->lr_type == einfo->ei_type);
2262 LASSERT(lock->l_completion_ast == einfo->ei_cb_cp);
2263 LASSERT(lock->l_glimpse_ast == einfo->ei_cb_gl);
2264
2265 lock_res_and_lock(lock);
2266 spin_lock(&osc_ast_guard);
2267
2268 if (lock->l_ast_data == NULL)
2269 lock->l_ast_data = data;
2270 if (lock->l_ast_data == data)
2271 set = 1;
2272
2273 spin_unlock(&osc_ast_guard);
2274 unlock_res_and_lock(lock);
2275
2276 return set;
2277}
2278
2279static int osc_set_data_with_check(struct lustre_handle *lockh,
2280 struct ldlm_enqueue_info *einfo)
2281{
2282 struct ldlm_lock *lock = ldlm_handle2lock(lockh);
2283 int set = 0;
2284
2285 if (lock != NULL) {
2286 set = osc_set_lock_data_with_check(lock, einfo);
2287 LDLM_LOCK_PUT(lock);
2288 } else
2289 CERROR("lockh %p, data %p - client evicted?\n",
2290 lockh, einfo->ei_cbdata);
2291 return set;
2292}
2293
2294static int osc_change_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2295 ldlm_iterator_t replace, void *data)
2296{
2297 struct ldlm_res_id res_id;
2298 struct obd_device *obd = class_exp2obd(exp);
2299
2300 ostid_build_res_name(&lsm->lsm_oi, &res_id);
2301 ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2302 return 0;
2303}
2304
2305/* find any ldlm lock of the inode in osc
2306 * return 0 not find
2307 * 1 find one
2308 * < 0 error */
2309static int osc_find_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2310 ldlm_iterator_t replace, void *data)
2311{
2312 struct ldlm_res_id res_id;
2313 struct obd_device *obd = class_exp2obd(exp);
2314 int rc = 0;
2315
2316 ostid_build_res_name(&lsm->lsm_oi, &res_id);
2317 rc = ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2318 if (rc == LDLM_ITER_STOP)
2319 return(1);
2320 if (rc == LDLM_ITER_CONTINUE)
2321 return(0);
2322 return(rc);
2323}
2324
2325static int osc_enqueue_fini(struct ptlrpc_request *req, struct ost_lvb *lvb,
2326 obd_enqueue_update_f upcall, void *cookie,
2327 __u64 *flags, int agl, int rc)
2328{
2329 int intent = *flags & LDLM_FL_HAS_INTENT;
2330 ENTRY;
2331
2332 if (intent) {
2333 /* The request was created before ldlm_cli_enqueue call. */
2334 if (rc == ELDLM_LOCK_ABORTED) {
2335 struct ldlm_reply *rep;
2336 rep = req_capsule_server_get(&req->rq_pill,
2337 &RMF_DLM_REP);
2338
2339 LASSERT(rep != NULL);
2340 if (rep->lock_policy_res1)
2341 rc = rep->lock_policy_res1;
2342 }
2343 }
2344
2345 if ((intent != 0 && rc == ELDLM_LOCK_ABORTED && agl == 0) ||
2346 (rc == 0)) {
2347 *flags |= LDLM_FL_LVB_READY;
2348 CDEBUG(D_INODE,"got kms "LPU64" blocks "LPU64" mtime "LPU64"\n",
2349 lvb->lvb_size, lvb->lvb_blocks, lvb->lvb_mtime);
2350 }
2351
2352 /* Call the update callback. */
2353 rc = (*upcall)(cookie, rc);
2354 RETURN(rc);
2355}
2356
2357static int osc_enqueue_interpret(const struct lu_env *env,
2358 struct ptlrpc_request *req,
2359 struct osc_enqueue_args *aa, int rc)
2360{
2361 struct ldlm_lock *lock;
2362 struct lustre_handle handle;
2363 __u32 mode;
2364 struct ost_lvb *lvb;
2365 __u32 lvb_len;
2366 __u64 *flags = aa->oa_flags;
2367
2368 /* Make a local copy of a lock handle and a mode, because aa->oa_*
2369 * might be freed anytime after lock upcall has been called. */
2370 lustre_handle_copy(&handle, aa->oa_lockh);
2371 mode = aa->oa_ei->ei_mode;
2372
2373 /* ldlm_cli_enqueue is holding a reference on the lock, so it must
2374 * be valid. */
2375 lock = ldlm_handle2lock(&handle);
2376
2377 /* Take an additional reference so that a blocking AST that
2378 * ldlm_cli_enqueue_fini() might post for a failed lock, is guaranteed
2379 * to arrive after an upcall has been executed by
2380 * osc_enqueue_fini(). */
2381 ldlm_lock_addref(&handle, mode);
2382
2383 /* Let CP AST to grant the lock first. */
2384 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_ENQ_RACE, 1);
2385
2386 if (aa->oa_agl && rc == ELDLM_LOCK_ABORTED) {
2387 lvb = NULL;
2388 lvb_len = 0;
2389 } else {
2390 lvb = aa->oa_lvb;
2391 lvb_len = sizeof(*aa->oa_lvb);
2392 }
2393
2394 /* Complete obtaining the lock procedure. */
2395 rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_ei->ei_type, 1,
2396 mode, flags, lvb, lvb_len, &handle, rc);
2397 /* Complete osc stuff. */
2398 rc = osc_enqueue_fini(req, aa->oa_lvb, aa->oa_upcall, aa->oa_cookie,
2399 flags, aa->oa_agl, rc);
2400
2401 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_CANCEL_RACE, 10);
2402
2403 /* Release the lock for async request. */
2404 if (lustre_handle_is_used(&handle) && rc == ELDLM_OK)
2405 /*
2406 * Releases a reference taken by ldlm_cli_enqueue(), if it is
2407 * not already released by
2408 * ldlm_cli_enqueue_fini()->failed_lock_cleanup()
2409 */
2410 ldlm_lock_decref(&handle, mode);
2411
2412 LASSERTF(lock != NULL, "lockh %p, req %p, aa %p - client evicted?\n",
2413 aa->oa_lockh, req, aa);
2414 ldlm_lock_decref(&handle, mode);
2415 LDLM_LOCK_PUT(lock);
2416 return rc;
2417}
2418
2419void osc_update_enqueue(struct lustre_handle *lov_lockhp,
2420 struct lov_oinfo *loi, int flags,
2421 struct ost_lvb *lvb, __u32 mode, int rc)
2422{
2423 struct ldlm_lock *lock = ldlm_handle2lock(lov_lockhp);
2424
2425 if (rc == ELDLM_OK) {
2426 __u64 tmp;
2427
2428 LASSERT(lock != NULL);
2429 loi->loi_lvb = *lvb;
2430 tmp = loi->loi_lvb.lvb_size;
2431 /* Extend KMS up to the end of this lock and no further
2432 * A lock on [x,y] means a KMS of up to y + 1 bytes! */
2433 if (tmp > lock->l_policy_data.l_extent.end)
2434 tmp = lock->l_policy_data.l_extent.end + 1;
2435 if (tmp >= loi->loi_kms) {
2436 LDLM_DEBUG(lock, "lock acquired, setting rss="LPU64
2437 ", kms="LPU64, loi->loi_lvb.lvb_size, tmp);
2438 loi_kms_set(loi, tmp);
2439 } else {
2440 LDLM_DEBUG(lock, "lock acquired, setting rss="
2441 LPU64"; leaving kms="LPU64", end="LPU64,
2442 loi->loi_lvb.lvb_size, loi->loi_kms,
2443 lock->l_policy_data.l_extent.end);
2444 }
2445 ldlm_lock_allow_match(lock);
2446 } else if (rc == ELDLM_LOCK_ABORTED && (flags & LDLM_FL_HAS_INTENT)) {
2447 LASSERT(lock != NULL);
2448 loi->loi_lvb = *lvb;
2449 ldlm_lock_allow_match(lock);
2450 CDEBUG(D_INODE, "glimpsed, setting rss="LPU64"; leaving"
2451 " kms="LPU64"\n", loi->loi_lvb.lvb_size, loi->loi_kms);
2452 rc = ELDLM_OK;
2453 }
2454
2455 if (lock != NULL) {
2456 if (rc != ELDLM_OK)
2457 ldlm_lock_fail_match(lock);
2458
2459 LDLM_LOCK_PUT(lock);
2460 }
2461}
2462EXPORT_SYMBOL(osc_update_enqueue);
2463
2464struct ptlrpc_request_set *PTLRPCD_SET = (void *)1;
2465
2466/* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
2467 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
2468 * other synchronous requests, however keeping some locks and trying to obtain
2469 * others may take a considerable amount of time in a case of ost failure; and
2470 * when other sync requests do not get released lock from a client, the client
2471 * is excluded from the cluster -- such scenarious make the life difficult, so
2472 * release locks just after they are obtained. */
2473int osc_enqueue_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2474 __u64 *flags, ldlm_policy_data_t *policy,
2475 struct ost_lvb *lvb, int kms_valid,
2476 obd_enqueue_update_f upcall, void *cookie,
2477 struct ldlm_enqueue_info *einfo,
2478 struct lustre_handle *lockh,
2479 struct ptlrpc_request_set *rqset, int async, int agl)
2480{
2481 struct obd_device *obd = exp->exp_obd;
2482 struct ptlrpc_request *req = NULL;
2483 int intent = *flags & LDLM_FL_HAS_INTENT;
2484 int match_lvb = (agl != 0 ? 0 : LDLM_FL_LVB_READY);
2485 ldlm_mode_t mode;
2486 int rc;
2487 ENTRY;
2488
2489 /* Filesystem lock extents are extended to page boundaries so that
2490 * dealing with the page cache is a little smoother. */
2491 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
2492 policy->l_extent.end |= ~CFS_PAGE_MASK;
2493
2494 /*
2495 * kms is not valid when either object is completely fresh (so that no
2496 * locks are cached), or object was evicted. In the latter case cached
2497 * lock cannot be used, because it would prime inode state with
2498 * potentially stale LVB.
2499 */
2500 if (!kms_valid)
2501 goto no_match;
2502
2503 /* Next, search for already existing extent locks that will cover us */
2504 /* If we're trying to read, we also search for an existing PW lock. The
2505 * VFS and page cache already protect us locally, so lots of readers/
2506 * writers can share a single PW lock.
2507 *
2508 * There are problems with conversion deadlocks, so instead of
2509 * converting a read lock to a write lock, we'll just enqueue a new
2510 * one.
2511 *
2512 * At some point we should cancel the read lock instead of making them
2513 * send us a blocking callback, but there are problems with canceling
2514 * locks out from other users right now, too. */
2515 mode = einfo->ei_mode;
2516 if (einfo->ei_mode == LCK_PR)
2517 mode |= LCK_PW;
2518 mode = ldlm_lock_match(obd->obd_namespace, *flags | match_lvb, res_id,
2519 einfo->ei_type, policy, mode, lockh, 0);
2520 if (mode) {
2521 struct ldlm_lock *matched = ldlm_handle2lock(lockh);
2522
2523 if ((agl != 0) && !(matched->l_flags & LDLM_FL_LVB_READY)) {
2524 /* For AGL, if enqueue RPC is sent but the lock is not
2525 * granted, then skip to process this strpe.
2526 * Return -ECANCELED to tell the caller. */
2527 ldlm_lock_decref(lockh, mode);
2528 LDLM_LOCK_PUT(matched);
2529 RETURN(-ECANCELED);
2530 } else if (osc_set_lock_data_with_check(matched, einfo)) {
2531 *flags |= LDLM_FL_LVB_READY;
2532 /* addref the lock only if not async requests and PW
2533 * lock is matched whereas we asked for PR. */
2534 if (!rqset && einfo->ei_mode != mode)
2535 ldlm_lock_addref(lockh, LCK_PR);
2536 if (intent) {
2537 /* I would like to be able to ASSERT here that
2538 * rss <= kms, but I can't, for reasons which
2539 * are explained in lov_enqueue() */
2540 }
2541
2542 /* We already have a lock, and it's referenced.
2543 *
2544 * At this point, the cl_lock::cll_state is CLS_QUEUING,
2545 * AGL upcall may change it to CLS_HELD directly. */
2546 (*upcall)(cookie, ELDLM_OK);
2547
2548 if (einfo->ei_mode != mode)
2549 ldlm_lock_decref(lockh, LCK_PW);
2550 else if (rqset)
2551 /* For async requests, decref the lock. */
2552 ldlm_lock_decref(lockh, einfo->ei_mode);
2553 LDLM_LOCK_PUT(matched);
2554 RETURN(ELDLM_OK);
2555 } else {
2556 ldlm_lock_decref(lockh, mode);
2557 LDLM_LOCK_PUT(matched);
2558 }
2559 }
2560
2561 no_match:
2562 if (intent) {
2563 LIST_HEAD(cancels);
2564 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2565 &RQF_LDLM_ENQUEUE_LVB);
2566 if (req == NULL)
2567 RETURN(-ENOMEM);
2568
2569 rc = ldlm_prep_enqueue_req(exp, req, &cancels, 0);
2570 if (rc) {
2571 ptlrpc_request_free(req);
2572 RETURN(rc);
2573 }
2574
2575 req_capsule_set_size(&req->rq_pill, &RMF_DLM_LVB, RCL_SERVER,
2576 sizeof *lvb);
2577 ptlrpc_request_set_replen(req);
2578 }
2579
2580 /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
2581 *flags &= ~LDLM_FL_BLOCK_GRANTED;
2582
2583 rc = ldlm_cli_enqueue(exp, &req, einfo, res_id, policy, flags, lvb,
2584 sizeof(*lvb), LVB_T_OST, lockh, async);
2585 if (rqset) {
2586 if (!rc) {
2587 struct osc_enqueue_args *aa;
2588 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
2589 aa = ptlrpc_req_async_args(req);
2590 aa->oa_ei = einfo;
2591 aa->oa_exp = exp;
2592 aa->oa_flags = flags;
2593 aa->oa_upcall = upcall;
2594 aa->oa_cookie = cookie;
2595 aa->oa_lvb = lvb;
2596 aa->oa_lockh = lockh;
2597 aa->oa_agl = !!agl;
2598
2599 req->rq_interpret_reply =
2600 (ptlrpc_interpterer_t)osc_enqueue_interpret;
2601 if (rqset == PTLRPCD_SET)
2602 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
2603 else
2604 ptlrpc_set_add_req(rqset, req);
2605 } else if (intent) {
2606 ptlrpc_req_finished(req);
2607 }
2608 RETURN(rc);
2609 }
2610
2611 rc = osc_enqueue_fini(req, lvb, upcall, cookie, flags, agl, rc);
2612 if (intent)
2613 ptlrpc_req_finished(req);
2614
2615 RETURN(rc);
2616}
2617
2618static int osc_enqueue(struct obd_export *exp, struct obd_info *oinfo,
2619 struct ldlm_enqueue_info *einfo,
2620 struct ptlrpc_request_set *rqset)
2621{
2622 struct ldlm_res_id res_id;
2623 int rc;
2624 ENTRY;
2625
2626 ostid_build_res_name(&oinfo->oi_md->lsm_oi, &res_id);
2627 rc = osc_enqueue_base(exp, &res_id, &oinfo->oi_flags, &oinfo->oi_policy,
2628 &oinfo->oi_md->lsm_oinfo[0]->loi_lvb,
2629 oinfo->oi_md->lsm_oinfo[0]->loi_kms_valid,
2630 oinfo->oi_cb_up, oinfo, einfo, oinfo->oi_lockh,
2631 rqset, rqset != NULL, 0);
2632 RETURN(rc);
2633}
2634
2635int osc_match_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2636 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
2637 int *flags, void *data, struct lustre_handle *lockh,
2638 int unref)
2639{
2640 struct obd_device *obd = exp->exp_obd;
2641 int lflags = *flags;
2642 ldlm_mode_t rc;
2643 ENTRY;
2644
2645 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_MATCH))
2646 RETURN(-EIO);
2647
2648 /* Filesystem lock extents are extended to page boundaries so that
2649 * dealing with the page cache is a little smoother */
2650 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
2651 policy->l_extent.end |= ~CFS_PAGE_MASK;
2652
2653 /* Next, search for already existing extent locks that will cover us */
2654 /* If we're trying to read, we also search for an existing PW lock. The
2655 * VFS and page cache already protect us locally, so lots of readers/
2656 * writers can share a single PW lock. */
2657 rc = mode;
2658 if (mode == LCK_PR)
2659 rc |= LCK_PW;
2660 rc = ldlm_lock_match(obd->obd_namespace, lflags,
2661 res_id, type, policy, rc, lockh, unref);
2662 if (rc) {
2663 if (data != NULL) {
2664 if (!osc_set_data_with_check(lockh, data)) {
2665 if (!(lflags & LDLM_FL_TEST_LOCK))
2666 ldlm_lock_decref(lockh, rc);
2667 RETURN(0);
2668 }
2669 }
2670 if (!(lflags & LDLM_FL_TEST_LOCK) && mode != rc) {
2671 ldlm_lock_addref(lockh, LCK_PR);
2672 ldlm_lock_decref(lockh, LCK_PW);
2673 }
2674 RETURN(rc);
2675 }
2676 RETURN(rc);
2677}
2678
2679int osc_cancel_base(struct lustre_handle *lockh, __u32 mode)
2680{
2681 ENTRY;
2682
2683 if (unlikely(mode == LCK_GROUP))
2684 ldlm_lock_decref_and_cancel(lockh, mode);
2685 else
2686 ldlm_lock_decref(lockh, mode);
2687
2688 RETURN(0);
2689}
2690
2691static int osc_cancel(struct obd_export *exp, struct lov_stripe_md *md,
2692 __u32 mode, struct lustre_handle *lockh)
2693{
2694 ENTRY;
2695 RETURN(osc_cancel_base(lockh, mode));
2696}
2697
2698static int osc_cancel_unused(struct obd_export *exp,
2699 struct lov_stripe_md *lsm,
2700 ldlm_cancel_flags_t flags,
2701 void *opaque)
2702{
2703 struct obd_device *obd = class_exp2obd(exp);
2704 struct ldlm_res_id res_id, *resp = NULL;
2705
2706 if (lsm != NULL) {
2707 ostid_build_res_name(&lsm->lsm_oi, &res_id);
2708 resp = &res_id;
2709 }
2710
2711 return ldlm_cli_cancel_unused(obd->obd_namespace, resp, flags, opaque);
2712}
2713
2714static int osc_statfs_interpret(const struct lu_env *env,
2715 struct ptlrpc_request *req,
2716 struct osc_async_args *aa, int rc)
2717{
2718 struct obd_statfs *msfs;
2719 ENTRY;
2720
2721 if (rc == -EBADR)
2722 /* The request has in fact never been sent
2723 * due to issues at a higher level (LOV).
2724 * Exit immediately since the caller is
2725 * aware of the problem and takes care
2726 * of the clean up */
2727 RETURN(rc);
2728
2729 if ((rc == -ENOTCONN || rc == -EAGAIN) &&
2730 (aa->aa_oi->oi_flags & OBD_STATFS_NODELAY))
2731 GOTO(out, rc = 0);
2732
2733 if (rc != 0)
2734 GOTO(out, rc);
2735
2736 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2737 if (msfs == NULL) {
2738 GOTO(out, rc = -EPROTO);
2739 }
2740
2741 *aa->aa_oi->oi_osfs = *msfs;
2742out:
2743 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
2744 RETURN(rc);
2745}
2746
2747static int osc_statfs_async(struct obd_export *exp,
2748 struct obd_info *oinfo, __u64 max_age,
2749 struct ptlrpc_request_set *rqset)
2750{
2751 struct obd_device *obd = class_exp2obd(exp);
2752 struct ptlrpc_request *req;
2753 struct osc_async_args *aa;
2754 int rc;
2755 ENTRY;
2756
2757 /* We could possibly pass max_age in the request (as an absolute
2758 * timestamp or a "seconds.usec ago") so the target can avoid doing
2759 * extra calls into the filesystem if that isn't necessary (e.g.
2760 * during mount that would help a bit). Having relative timestamps
2761 * is not so great if request processing is slow, while absolute
2762 * timestamps are not ideal because they need time synchronization. */
2763 req = ptlrpc_request_alloc(obd->u.cli.cl_import, &RQF_OST_STATFS);
2764 if (req == NULL)
2765 RETURN(-ENOMEM);
2766
2767 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2768 if (rc) {
2769 ptlrpc_request_free(req);
2770 RETURN(rc);
2771 }
2772 ptlrpc_request_set_replen(req);
2773 req->rq_request_portal = OST_CREATE_PORTAL;
2774 ptlrpc_at_set_req_timeout(req);
2775
2776 if (oinfo->oi_flags & OBD_STATFS_NODELAY) {
2777 /* procfs requests not want stat in wait for avoid deadlock */
2778 req->rq_no_resend = 1;
2779 req->rq_no_delay = 1;
2780 }
2781
2782 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_statfs_interpret;
2783 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
2784 aa = ptlrpc_req_async_args(req);
2785 aa->aa_oi = oinfo;
2786
2787 ptlrpc_set_add_req(rqset, req);
2788 RETURN(0);
2789}
2790
2791static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
2792 struct obd_statfs *osfs, __u64 max_age, __u32 flags)
2793{
2794 struct obd_device *obd = class_exp2obd(exp);
2795 struct obd_statfs *msfs;
2796 struct ptlrpc_request *req;
2797 struct obd_import *imp = NULL;
2798 int rc;
2799 ENTRY;
2800
2801 /*Since the request might also come from lprocfs, so we need
2802 *sync this with client_disconnect_export Bug15684*/
2803 down_read(&obd->u.cli.cl_sem);
2804 if (obd->u.cli.cl_import)
2805 imp = class_import_get(obd->u.cli.cl_import);
2806 up_read(&obd->u.cli.cl_sem);
2807 if (!imp)
2808 RETURN(-ENODEV);
2809
2810 /* We could possibly pass max_age in the request (as an absolute
2811 * timestamp or a "seconds.usec ago") so the target can avoid doing
2812 * extra calls into the filesystem if that isn't necessary (e.g.
2813 * during mount that would help a bit). Having relative timestamps
2814 * is not so great if request processing is slow, while absolute
2815 * timestamps are not ideal because they need time synchronization. */
2816 req = ptlrpc_request_alloc(imp, &RQF_OST_STATFS);
2817
2818 class_import_put(imp);
2819
2820 if (req == NULL)
2821 RETURN(-ENOMEM);
2822
2823 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2824 if (rc) {
2825 ptlrpc_request_free(req);
2826 RETURN(rc);
2827 }
2828 ptlrpc_request_set_replen(req);
2829 req->rq_request_portal = OST_CREATE_PORTAL;
2830 ptlrpc_at_set_req_timeout(req);
2831
2832 if (flags & OBD_STATFS_NODELAY) {
2833 /* procfs requests not want stat in wait for avoid deadlock */
2834 req->rq_no_resend = 1;
2835 req->rq_no_delay = 1;
2836 }
2837
2838 rc = ptlrpc_queue_wait(req);
2839 if (rc)
2840 GOTO(out, rc);
2841
2842 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2843 if (msfs == NULL) {
2844 GOTO(out, rc = -EPROTO);
2845 }
2846
2847 *osfs = *msfs;
2848
2849 EXIT;
2850 out:
2851 ptlrpc_req_finished(req);
2852 return rc;
2853}
2854
2855/* Retrieve object striping information.
2856 *
2857 * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
2858 * the maximum number of OST indices which will fit in the user buffer.
2859 * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
2860 */
2861static int osc_getstripe(struct lov_stripe_md *lsm, struct lov_user_md *lump)
2862{
2863 /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
2864 struct lov_user_md_v3 lum, *lumk;
2865 struct lov_user_ost_data_v1 *lmm_objects;
2866 int rc = 0, lum_size;
2867 ENTRY;
2868
2869 if (!lsm)
2870 RETURN(-ENODATA);
2871
2872 /* we only need the header part from user space to get lmm_magic and
2873 * lmm_stripe_count, (the header part is common to v1 and v3) */
2874 lum_size = sizeof(struct lov_user_md_v1);
2875 if (copy_from_user(&lum, lump, lum_size))
2876 RETURN(-EFAULT);
2877
2878 if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
2879 (lum.lmm_magic != LOV_USER_MAGIC_V3))
2880 RETURN(-EINVAL);
2881
2882 /* lov_user_md_vX and lov_mds_md_vX must have the same size */
2883 LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
2884 LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
2885 LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
2886
2887 /* we can use lov_mds_md_size() to compute lum_size
2888 * because lov_user_md_vX and lov_mds_md_vX have the same size */
2889 if (lum.lmm_stripe_count > 0) {
2890 lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
2891 OBD_ALLOC(lumk, lum_size);
2892 if (!lumk)
2893 RETURN(-ENOMEM);
2894
2895 if (lum.lmm_magic == LOV_USER_MAGIC_V1)
2896 lmm_objects =
2897 &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
2898 else
2899 lmm_objects = &(lumk->lmm_objects[0]);
2900 lmm_objects->l_ost_oi = lsm->lsm_oi;
2901 } else {
2902 lum_size = lov_mds_md_size(0, lum.lmm_magic);
2903 lumk = &lum;
2904 }
2905
2906 lumk->lmm_oi = lsm->lsm_oi;
2907 lumk->lmm_stripe_count = 1;
2908
2909 if (copy_to_user(lump, lumk, lum_size))
2910 rc = -EFAULT;
2911
2912 if (lumk != &lum)
2913 OBD_FREE(lumk, lum_size);
2914
2915 RETURN(rc);
2916}
2917
2918
2919static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
2920 void *karg, void *uarg)
2921{
2922 struct obd_device *obd = exp->exp_obd;
2923 struct obd_ioctl_data *data = karg;
2924 int err = 0;
2925 ENTRY;
2926
2927 if (!try_module_get(THIS_MODULE)) {
2928 CERROR("Can't get module. Is it alive?");
2929 return -EINVAL;
2930 }
2931 switch (cmd) {
2932 case OBD_IOC_LOV_GET_CONFIG: {
2933 char *buf;
2934 struct lov_desc *desc;
2935 struct obd_uuid uuid;
2936
2937 buf = NULL;
2938 len = 0;
2939 if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
2940 GOTO(out, err = -EINVAL);
2941
2942 data = (struct obd_ioctl_data *)buf;
2943
2944 if (sizeof(*desc) > data->ioc_inllen1) {
2945 obd_ioctl_freedata(buf, len);
2946 GOTO(out, err = -EINVAL);
2947 }
2948
2949 if (data->ioc_inllen2 < sizeof(uuid)) {
2950 obd_ioctl_freedata(buf, len);
2951 GOTO(out, err = -EINVAL);
2952 }
2953
2954 desc = (struct lov_desc *)data->ioc_inlbuf1;
2955 desc->ld_tgt_count = 1;
2956 desc->ld_active_tgt_count = 1;
2957 desc->ld_default_stripe_count = 1;
2958 desc->ld_default_stripe_size = 0;
2959 desc->ld_default_stripe_offset = 0;
2960 desc->ld_pattern = 0;
2961 memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
2962
2963 memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
2964
2965 err = copy_to_user((void *)uarg, buf, len);
2966 if (err)
2967 err = -EFAULT;
2968 obd_ioctl_freedata(buf, len);
2969 GOTO(out, err);
2970 }
2971 case LL_IOC_LOV_SETSTRIPE:
2972 err = obd_alloc_memmd(exp, karg);
2973 if (err > 0)
2974 err = 0;
2975 GOTO(out, err);
2976 case LL_IOC_LOV_GETSTRIPE:
2977 err = osc_getstripe(karg, uarg);
2978 GOTO(out, err);
2979 case OBD_IOC_CLIENT_RECOVER:
2980 err = ptlrpc_recover_import(obd->u.cli.cl_import,
2981 data->ioc_inlbuf1, 0);
2982 if (err > 0)
2983 err = 0;
2984 GOTO(out, err);
2985 case IOC_OSC_SET_ACTIVE:
2986 err = ptlrpc_set_import_active(obd->u.cli.cl_import,
2987 data->ioc_offset);
2988 GOTO(out, err);
2989 case OBD_IOC_POLL_QUOTACHECK:
2990 err = osc_quota_poll_check(exp, (struct if_quotacheck *)karg);
2991 GOTO(out, err);
2992 case OBD_IOC_PING_TARGET:
2993 err = ptlrpc_obd_ping(obd);
2994 GOTO(out, err);
2995 default:
2996 CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
2997 cmd, current_comm());
2998 GOTO(out, err = -ENOTTY);
2999 }
3000out:
3001 module_put(THIS_MODULE);
3002 return err;
3003}
3004
3005static int osc_get_info(const struct lu_env *env, struct obd_export *exp,
3006 obd_count keylen, void *key, __u32 *vallen, void *val,
3007 struct lov_stripe_md *lsm)
3008{
3009 ENTRY;
3010 if (!vallen || !val)
3011 RETURN(-EFAULT);
3012
3013 if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
3014 __u32 *stripe = val;
3015 *vallen = sizeof(*stripe);
3016 *stripe = 0;
3017 RETURN(0);
3018 } else if (KEY_IS(KEY_LAST_ID)) {
3019 struct ptlrpc_request *req;
3020 obd_id *reply;
3021 char *tmp;
3022 int rc;
3023
3024 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
3025 &RQF_OST_GET_INFO_LAST_ID);
3026 if (req == NULL)
3027 RETURN(-ENOMEM);
3028
3029 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3030 RCL_CLIENT, keylen);
3031 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
3032 if (rc) {
3033 ptlrpc_request_free(req);
3034 RETURN(rc);
3035 }
3036
3037 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3038 memcpy(tmp, key, keylen);
3039
3040 req->rq_no_delay = req->rq_no_resend = 1;
3041 ptlrpc_request_set_replen(req);
3042 rc = ptlrpc_queue_wait(req);
3043 if (rc)
3044 GOTO(out, rc);
3045
3046 reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
3047 if (reply == NULL)
3048 GOTO(out, rc = -EPROTO);
3049
3050 *((obd_id *)val) = *reply;
3051 out:
3052 ptlrpc_req_finished(req);
3053 RETURN(rc);
3054 } else if (KEY_IS(KEY_FIEMAP)) {
3055 struct ptlrpc_request *req;
3056 struct ll_user_fiemap *reply;
3057 char *tmp;
3058 int rc;
3059
3060 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
3061 &RQF_OST_GET_INFO_FIEMAP);
3062 if (req == NULL)
3063 RETURN(-ENOMEM);
3064
3065 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_KEY,
3066 RCL_CLIENT, keylen);
3067 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
3068 RCL_CLIENT, *vallen);
3069 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
3070 RCL_SERVER, *vallen);
3071
3072 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
3073 if (rc) {
3074 ptlrpc_request_free(req);
3075 RETURN(rc);
3076 }
3077
3078 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_KEY);
3079 memcpy(tmp, key, keylen);
3080 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3081 memcpy(tmp, val, *vallen);
3082
3083 ptlrpc_request_set_replen(req);
3084 rc = ptlrpc_queue_wait(req);
3085 if (rc)
3086 GOTO(out1, rc);
3087
3088 reply = req_capsule_server_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3089 if (reply == NULL)
3090 GOTO(out1, rc = -EPROTO);
3091
3092 memcpy(val, reply, *vallen);
3093 out1:
3094 ptlrpc_req_finished(req);
3095
3096 RETURN(rc);
3097 }
3098
3099 RETURN(-EINVAL);
3100}
3101
3102static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
3103 obd_count keylen, void *key, obd_count vallen,
3104 void *val, struct ptlrpc_request_set *set)
3105{
3106 struct ptlrpc_request *req;
3107 struct obd_device *obd = exp->exp_obd;
3108 struct obd_import *imp = class_exp2cliimp(exp);
3109 char *tmp;
3110 int rc;
3111 ENTRY;
3112
3113 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
3114
3115 if (KEY_IS(KEY_CHECKSUM)) {
3116 if (vallen != sizeof(int))
3117 RETURN(-EINVAL);
3118 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
3119 RETURN(0);
3120 }
3121
3122 if (KEY_IS(KEY_SPTLRPC_CONF)) {
3123 sptlrpc_conf_client_adapt(obd);
3124 RETURN(0);
3125 }
3126
3127 if (KEY_IS(KEY_FLUSH_CTX)) {
3128 sptlrpc_import_flush_my_ctx(imp);
3129 RETURN(0);
3130 }
3131
3132 if (KEY_IS(KEY_CACHE_SET)) {
3133 struct client_obd *cli = &obd->u.cli;
3134
3135 LASSERT(cli->cl_cache == NULL); /* only once */
3136 cli->cl_cache = (struct cl_client_cache *)val;
3137 atomic_inc(&cli->cl_cache->ccc_users);
3138 cli->cl_lru_left = &cli->cl_cache->ccc_lru_left;
3139
3140 /* add this osc into entity list */
3141 LASSERT(list_empty(&cli->cl_lru_osc));
3142 spin_lock(&cli->cl_cache->ccc_lru_lock);
3143 list_add(&cli->cl_lru_osc, &cli->cl_cache->ccc_lru);
3144 spin_unlock(&cli->cl_cache->ccc_lru_lock);
3145
3146 RETURN(0);
3147 }
3148
3149 if (KEY_IS(KEY_CACHE_LRU_SHRINK)) {
3150 struct client_obd *cli = &obd->u.cli;
3151 int nr = atomic_read(&cli->cl_lru_in_list) >> 1;
3152 int target = *(int *)val;
3153
3154 nr = osc_lru_shrink(cli, min(nr, target));
3155 *(int *)val -= nr;
3156 RETURN(0);
3157 }
3158
3159 if (!set && !KEY_IS(KEY_GRANT_SHRINK))
3160 RETURN(-EINVAL);
3161
3162 /* We pass all other commands directly to OST. Since nobody calls osc
3163 methods directly and everybody is supposed to go through LOV, we
3164 assume lov checked invalid values for us.
3165 The only recognised values so far are evict_by_nid and mds_conn.
3166 Even if something bad goes through, we'd get a -EINVAL from OST
3167 anyway. */
3168
3169 req = ptlrpc_request_alloc(imp, KEY_IS(KEY_GRANT_SHRINK) ?
3170 &RQF_OST_SET_GRANT_INFO :
3171 &RQF_OBD_SET_INFO);
3172 if (req == NULL)
3173 RETURN(-ENOMEM);
3174
3175 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3176 RCL_CLIENT, keylen);
3177 if (!KEY_IS(KEY_GRANT_SHRINK))
3178 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_VAL,
3179 RCL_CLIENT, vallen);
3180 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SET_INFO);
3181 if (rc) {
3182 ptlrpc_request_free(req);
3183 RETURN(rc);
3184 }
3185
3186 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3187 memcpy(tmp, key, keylen);
3188 tmp = req_capsule_client_get(&req->rq_pill, KEY_IS(KEY_GRANT_SHRINK) ?
3189 &RMF_OST_BODY :
3190 &RMF_SETINFO_VAL);
3191 memcpy(tmp, val, vallen);
3192
3193 if (KEY_IS(KEY_GRANT_SHRINK)) {
3194 struct osc_grant_args *aa;
3195 struct obdo *oa;
3196
3197 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
3198 aa = ptlrpc_req_async_args(req);
3199 OBDO_ALLOC(oa);
3200 if (!oa) {
3201 ptlrpc_req_finished(req);
3202 RETURN(-ENOMEM);
3203 }
3204 *oa = ((struct ost_body *)val)->oa;
3205 aa->aa_oa = oa;
3206 req->rq_interpret_reply = osc_shrink_grant_interpret;
3207 }
3208
3209 ptlrpc_request_set_replen(req);
3210 if (!KEY_IS(KEY_GRANT_SHRINK)) {
3211 LASSERT(set != NULL);
3212 ptlrpc_set_add_req(set, req);
3213 ptlrpc_check_set(NULL, set);
3214 } else
3215 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
3216
3217 RETURN(0);
3218}
3219
3220
3221static int osc_llog_init(struct obd_device *obd, struct obd_llog_group *olg,
3222 struct obd_device *disk_obd, int *index)
3223{
3224 /* this code is not supposed to be used with LOD/OSP
3225 * to be removed soon */
3226 LBUG();
3227 return 0;
3228}
3229
3230static int osc_llog_finish(struct obd_device *obd, int count)
3231{
3232 struct llog_ctxt *ctxt;
3233
3234 ENTRY;
3235
3236 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
3237 if (ctxt) {
3238 llog_cat_close(NULL, ctxt->loc_handle);
3239 llog_cleanup(NULL, ctxt);
3240 }
3241
3242 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3243 if (ctxt)
3244 llog_cleanup(NULL, ctxt);
3245 RETURN(0);
3246}
3247
3248static int osc_reconnect(const struct lu_env *env,
3249 struct obd_export *exp, struct obd_device *obd,
3250 struct obd_uuid *cluuid,
3251 struct obd_connect_data *data,
3252 void *localdata)
3253{
3254 struct client_obd *cli = &obd->u.cli;
3255
3256 if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
3257 long lost_grant;
3258
3259 client_obd_list_lock(&cli->cl_loi_list_lock);
3260 data->ocd_grant = (cli->cl_avail_grant + cli->cl_dirty) ?:
3261 2 * cli_brw_size(obd);
3262 lost_grant = cli->cl_lost_grant;
3263 cli->cl_lost_grant = 0;
3264 client_obd_list_unlock(&cli->cl_loi_list_lock);
3265
3266 CDEBUG(D_RPCTRACE, "ocd_connect_flags: "LPX64" ocd_version: %d"
3267 " ocd_grant: %d, lost: %ld.\n", data->ocd_connect_flags,
3268 data->ocd_version, data->ocd_grant, lost_grant);
3269 }
3270
3271 RETURN(0);
3272}
3273
3274static int osc_disconnect(struct obd_export *exp)
3275{
3276 struct obd_device *obd = class_exp2obd(exp);
3277 struct llog_ctxt *ctxt;
3278 int rc;
3279
3280 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3281 if (ctxt) {
3282 if (obd->u.cli.cl_conn_count == 1) {
3283 /* Flush any remaining cancel messages out to the
3284 * target */
3285 llog_sync(ctxt, exp, 0);
3286 }
3287 llog_ctxt_put(ctxt);
3288 } else {
3289 CDEBUG(D_HA, "No LLOG_SIZE_REPL_CTXT found in obd %p\n",
3290 obd);
3291 }
3292
3293 rc = client_disconnect_export(exp);
3294 /**
3295 * Initially we put del_shrink_grant before disconnect_export, but it
3296 * causes the following problem if setup (connect) and cleanup
3297 * (disconnect) are tangled together.
3298 * connect p1 disconnect p2
3299 * ptlrpc_connect_import
3300 * ............... class_manual_cleanup
3301 * osc_disconnect
3302 * del_shrink_grant
3303 * ptlrpc_connect_interrupt
3304 * init_grant_shrink
3305 * add this client to shrink list
3306 * cleanup_osc
3307 * Bang! pinger trigger the shrink.
3308 * So the osc should be disconnected from the shrink list, after we
3309 * are sure the import has been destroyed. BUG18662
3310 */
3311 if (obd->u.cli.cl_import == NULL)
3312 osc_del_shrink_grant(&obd->u.cli);
3313 return rc;
3314}
3315
3316static int osc_import_event(struct obd_device *obd,
3317 struct obd_import *imp,
3318 enum obd_import_event event)
3319{
3320 struct client_obd *cli;
3321 int rc = 0;
3322
3323 ENTRY;
3324 LASSERT(imp->imp_obd == obd);
3325
3326 switch (event) {
3327 case IMP_EVENT_DISCON: {
3328 cli = &obd->u.cli;
3329 client_obd_list_lock(&cli->cl_loi_list_lock);
3330 cli->cl_avail_grant = 0;
3331 cli->cl_lost_grant = 0;
3332 client_obd_list_unlock(&cli->cl_loi_list_lock);
3333 break;
3334 }
3335 case IMP_EVENT_INACTIVE: {
3336 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
3337 break;
3338 }
3339 case IMP_EVENT_INVALIDATE: {
3340 struct ldlm_namespace *ns = obd->obd_namespace;
3341 struct lu_env *env;
3342 int refcheck;
3343
3344 env = cl_env_get(&refcheck);
3345 if (!IS_ERR(env)) {
3346 /* Reset grants */
3347 cli = &obd->u.cli;
3348 /* all pages go to failing rpcs due to the invalid
3349 * import */
3350 osc_io_unplug(env, cli, NULL, PDL_POLICY_ROUND);
3351
3352 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3353 cl_env_put(env, &refcheck);
3354 } else
3355 rc = PTR_ERR(env);
3356 break;
3357 }
3358 case IMP_EVENT_ACTIVE: {
3359 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
3360 break;
3361 }
3362 case IMP_EVENT_OCD: {
3363 struct obd_connect_data *ocd = &imp->imp_connect_data;
3364
3365 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
3366 osc_init_grant(&obd->u.cli, ocd);
3367
3368 /* See bug 7198 */
3369 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
3370 imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;
3371
3372 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
3373 break;
3374 }
3375 case IMP_EVENT_DEACTIVATE: {
3376 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_DEACTIVATE, NULL);
3377 break;
3378 }
3379 case IMP_EVENT_ACTIVATE: {
3380 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVATE, NULL);
3381 break;
3382 }
3383 default:
3384 CERROR("Unknown import event %d\n", event);
3385 LBUG();
3386 }
3387 RETURN(rc);
3388}
3389
3390/**
3391 * Determine whether the lock can be canceled before replaying the lock
3392 * during recovery, see bug16774 for detailed information.
3393 *
3394 * \retval zero the lock can't be canceled
3395 * \retval other ok to cancel
3396 */
3397static int osc_cancel_for_recovery(struct ldlm_lock *lock)
3398{
3399 check_res_locked(lock->l_resource);
3400
3401 /*
3402 * Cancel all unused extent lock in granted mode LCK_PR or LCK_CR.
3403 *
3404 * XXX as a future improvement, we can also cancel unused write lock
3405 * if it doesn't have dirty data and active mmaps.
3406 */
3407 if (lock->l_resource->lr_type == LDLM_EXTENT &&
3408 (lock->l_granted_mode == LCK_PR ||
3409 lock->l_granted_mode == LCK_CR) &&
3410 (osc_dlm_lock_pageref(lock) == 0))
3411 RETURN(1);
3412
3413 RETURN(0);
3414}
3415
3416static int brw_queue_work(const struct lu_env *env, void *data)
3417{
3418 struct client_obd *cli = data;
3419
3420 CDEBUG(D_CACHE, "Run writeback work for client obd %p.\n", cli);
3421
3422 osc_io_unplug(env, cli, NULL, PDL_POLICY_SAME);
3423 RETURN(0);
3424}
3425
3426int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
3427{
3428 struct lprocfs_static_vars lvars = { 0 };
3429 struct client_obd *cli = &obd->u.cli;
3430 void *handler;
3431 int rc;
3432 ENTRY;
3433
3434 rc = ptlrpcd_addref();
3435 if (rc)
3436 RETURN(rc);
3437
3438 rc = client_obd_setup(obd, lcfg);
3439 if (rc)
3440 GOTO(out_ptlrpcd, rc);
3441
3442 handler = ptlrpcd_alloc_work(cli->cl_import, brw_queue_work, cli);
3443 if (IS_ERR(handler))
3444 GOTO(out_client_setup, rc = PTR_ERR(handler));
3445 cli->cl_writeback_work = handler;
3446
3447 rc = osc_quota_setup(obd);
3448 if (rc)
3449 GOTO(out_ptlrpcd_work, rc);
3450
3451 cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
3452 lprocfs_osc_init_vars(&lvars);
3453 if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0) {
3454 lproc_osc_attach_seqstat(obd);
3455 sptlrpc_lprocfs_cliobd_attach(obd);
3456 ptlrpc_lprocfs_register_obd(obd);
3457 }
3458
3459 /* We need to allocate a few requests more, because
3460 * brw_interpret tries to create new requests before freeing
3461 * previous ones, Ideally we want to have 2x max_rpcs_in_flight
3462 * reserved, but I'm afraid that might be too much wasted RAM
3463 * in fact, so 2 is just my guess and still should work. */
3464 cli->cl_import->imp_rq_pool =
3465 ptlrpc_init_rq_pool(cli->cl_max_rpcs_in_flight + 2,
3466 OST_MAXREQSIZE,
3467 ptlrpc_add_rqs_to_pool);
3468
3469 INIT_LIST_HEAD(&cli->cl_grant_shrink_list);
3470 ns_register_cancel(obd->obd_namespace, osc_cancel_for_recovery);
3471 RETURN(rc);
3472
3473out_ptlrpcd_work:
3474 ptlrpcd_destroy_work(handler);
3475out_client_setup:
3476 client_obd_cleanup(obd);
3477out_ptlrpcd:
3478 ptlrpcd_decref();
3479 RETURN(rc);
3480}
3481
3482static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
3483{
3484 int rc = 0;
3485 ENTRY;
3486
3487 switch (stage) {
3488 case OBD_CLEANUP_EARLY: {
3489 struct obd_import *imp;
3490 imp = obd->u.cli.cl_import;
3491 CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
3492 /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
3493 ptlrpc_deactivate_import(imp);
3494 spin_lock(&imp->imp_lock);
3495 imp->imp_pingable = 0;
3496 spin_unlock(&imp->imp_lock);
3497 break;
3498 }
3499 case OBD_CLEANUP_EXPORTS: {
3500 struct client_obd *cli = &obd->u.cli;
3501 /* LU-464
3502 * for echo client, export may be on zombie list, wait for
3503 * zombie thread to cull it, because cli.cl_import will be
3504 * cleared in client_disconnect_export():
3505 * class_export_destroy() -> obd_cleanup() ->
3506 * echo_device_free() -> echo_client_cleanup() ->
3507 * obd_disconnect() -> osc_disconnect() ->
3508 * client_disconnect_export()
3509 */
3510 obd_zombie_barrier();
3511 if (cli->cl_writeback_work) {
3512 ptlrpcd_destroy_work(cli->cl_writeback_work);
3513 cli->cl_writeback_work = NULL;
3514 }
3515 obd_cleanup_client_import(obd);
3516 ptlrpc_lprocfs_unregister_obd(obd);
3517 lprocfs_obd_cleanup(obd);
3518 rc = obd_llog_finish(obd, 0);
3519 if (rc != 0)
3520 CERROR("failed to cleanup llogging subsystems\n");
3521 break;
3522 }
3523 }
3524 RETURN(rc);
3525}
3526
3527int osc_cleanup(struct obd_device *obd)
3528{
3529 struct client_obd *cli = &obd->u.cli;
3530 int rc;
3531
3532 ENTRY;
3533
3534 /* lru cleanup */
3535 if (cli->cl_cache != NULL) {
3536 LASSERT(atomic_read(&cli->cl_cache->ccc_users) > 0);
3537 spin_lock(&cli->cl_cache->ccc_lru_lock);
3538 list_del_init(&cli->cl_lru_osc);
3539 spin_unlock(&cli->cl_cache->ccc_lru_lock);
3540 cli->cl_lru_left = NULL;
3541 atomic_dec(&cli->cl_cache->ccc_users);
3542 cli->cl_cache = NULL;
3543 }
3544
3545 /* free memory of osc quota cache */
3546 osc_quota_cleanup(obd);
3547
3548 rc = client_obd_cleanup(obd);
3549
3550 ptlrpcd_decref();
3551 RETURN(rc);
3552}
3553
3554int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *lcfg)
3555{
3556 struct lprocfs_static_vars lvars = { 0 };
3557 int rc = 0;
3558
3559 lprocfs_osc_init_vars(&lvars);
3560
3561 switch (lcfg->lcfg_command) {
3562 default:
3563 rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars,
3564 lcfg, obd);
3565 if (rc > 0)
3566 rc = 0;
3567 break;
3568 }
3569
3570 return(rc);
3571}
3572
3573static int osc_process_config(struct obd_device *obd, obd_count len, void *buf)
3574{
3575 return osc_process_config_base(obd, buf);
3576}
3577
3578struct obd_ops osc_obd_ops = {
3579 .o_owner = THIS_MODULE,
3580 .o_setup = osc_setup,
3581 .o_precleanup = osc_precleanup,
3582 .o_cleanup = osc_cleanup,
3583 .o_add_conn = client_import_add_conn,
3584 .o_del_conn = client_import_del_conn,
3585 .o_connect = client_connect_import,
3586 .o_reconnect = osc_reconnect,
3587 .o_disconnect = osc_disconnect,
3588 .o_statfs = osc_statfs,
3589 .o_statfs_async = osc_statfs_async,
3590 .o_packmd = osc_packmd,
3591 .o_unpackmd = osc_unpackmd,
3592 .o_create = osc_create,
3593 .o_destroy = osc_destroy,
3594 .o_getattr = osc_getattr,
3595 .o_getattr_async = osc_getattr_async,
3596 .o_setattr = osc_setattr,
3597 .o_setattr_async = osc_setattr_async,
3598 .o_brw = osc_brw,
3599 .o_punch = osc_punch,
3600 .o_sync = osc_sync,
3601 .o_enqueue = osc_enqueue,
3602 .o_change_cbdata = osc_change_cbdata,
3603 .o_find_cbdata = osc_find_cbdata,
3604 .o_cancel = osc_cancel,
3605 .o_cancel_unused = osc_cancel_unused,
3606 .o_iocontrol = osc_iocontrol,
3607 .o_get_info = osc_get_info,
3608 .o_set_info_async = osc_set_info_async,
3609 .o_import_event = osc_import_event,
3610 .o_llog_init = osc_llog_init,
3611 .o_llog_finish = osc_llog_finish,
3612 .o_process_config = osc_process_config,
3613 .o_quotactl = osc_quotactl,
3614 .o_quotacheck = osc_quotacheck,
3615};
3616
3617extern struct lu_kmem_descr osc_caches[];
3618extern spinlock_t osc_ast_guard;
3619extern struct lock_class_key osc_ast_guard_class;
3620
3621int __init osc_init(void)
3622{
3623 struct lprocfs_static_vars lvars = { 0 };
3624 int rc;
3625 ENTRY;
3626
3627 /* print an address of _any_ initialized kernel symbol from this
3628 * module, to allow debugging with gdb that doesn't support data
3629 * symbols from modules.*/
3630 CDEBUG(D_INFO, "Lustre OSC module (%p).\n", &osc_caches);
3631
3632 rc = lu_kmem_init(osc_caches);
3633
3634 lprocfs_osc_init_vars(&lvars);
3635
3636 rc = class_register_type(&osc_obd_ops, NULL, lvars.module_vars,
3637 LUSTRE_OSC_NAME, &osc_device_type);
3638 if (rc) {
3639 lu_kmem_fini(osc_caches);
3640 RETURN(rc);
3641 }
3642
3643 spin_lock_init(&osc_ast_guard);
3644 lockdep_set_class(&osc_ast_guard, &osc_ast_guard_class);
3645
3646 RETURN(rc);
3647}
3648
3649static void /*__exit*/ osc_exit(void)
3650{
3651 class_unregister_type(LUSTRE_OSC_NAME);
3652 lu_kmem_fini(osc_caches);
3653}
3654
3655MODULE_AUTHOR("Sun Microsystems, Inc. <http://www.lustre.org/>");
3656MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
3657MODULE_LICENSE("GPL");
3658
3659cfs_module(osc, LUSTRE_VERSION_STRING, osc_init, osc_exit);
Linux kernel - Deliverables
This page took 0.176656 seconds and 5 git commands to generate.