4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
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.
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).
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.gnu.org/licenses/gpl-2.0.html
23 * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
24 * Use is subject to license terms.
26 * Copyright (c) 2011, 2017, Intel Corporation.
29 * This file is part of Lustre, http://www.lustre.org/
32 #define DEBUG_SUBSYSTEM S_OSC
34 #include <linux/workqueue.h>
35 #include <libcfs/libcfs.h>
36 #include <linux/falloc.h>
37 #include <lprocfs_status.h>
38 #include <lustre_dlm.h>
39 #include <lustre_fid.h>
40 #include <lustre_ha.h>
41 #include <uapi/linux/lustre/lustre_ioctl.h>
42 #include <lustre_net.h>
43 #include <lustre_obdo.h>
45 #include <obd_cksum.h>
46 #include <obd_class.h>
47 #include <lustre_osc.h>
48 #include <linux/falloc.h>
50 #include "osc_internal.h"
51 #include <lnet/lnet_rdma.h>
53 atomic_t osc_pool_req_count;
54 unsigned int osc_reqpool_maxreqcount;
55 struct ptlrpc_request_pool *osc_rq_pool;
57 /* max memory used for request pool, unit is MB */
58 static unsigned int osc_reqpool_mem_max = 5;
59 module_param(osc_reqpool_mem_max, uint, 0444);
61 static int osc_idle_timeout = 20;
62 module_param(osc_idle_timeout, uint, 0644);
64 #define osc_grant_args osc_brw_async_args
66 struct osc_setattr_args {
68 obd_enqueue_update_f sa_upcall;
72 struct osc_fsync_args {
73 struct osc_object *fa_obj;
75 obd_enqueue_update_f fa_upcall;
79 struct osc_ladvise_args {
81 obd_enqueue_update_f la_upcall;
85 static void osc_release_ppga(struct brw_page **ppga, size_t count);
86 static int brw_interpret(const struct lu_env *env, struct ptlrpc_request *req,
89 void osc_pack_req_body(struct ptlrpc_request *req, struct obdo *oa)
91 struct ost_body *body;
93 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
96 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
99 static int osc_getattr(const struct lu_env *env, struct obd_export *exp,
102 struct ptlrpc_request *req;
103 struct ost_body *body;
107 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
111 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
113 ptlrpc_request_free(req);
117 osc_pack_req_body(req, oa);
119 ptlrpc_request_set_replen(req);
121 rc = ptlrpc_queue_wait(req);
125 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
127 GOTO(out, rc = -EPROTO);
129 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
130 lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oa, &body->oa);
132 oa->o_blksize = cli_brw_size(exp->exp_obd);
133 oa->o_valid |= OBD_MD_FLBLKSZ;
137 ptlrpc_req_finished(req);
142 static int osc_setattr(const struct lu_env *env, struct obd_export *exp,
145 struct ptlrpc_request *req;
146 struct ost_body *body;
150 LASSERT(oa->o_valid & OBD_MD_FLGROUP);
152 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
156 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
158 ptlrpc_request_free(req);
162 osc_pack_req_body(req, oa);
164 ptlrpc_request_set_replen(req);
166 rc = ptlrpc_queue_wait(req);
170 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
172 GOTO(out, rc = -EPROTO);
174 lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oa, &body->oa);
178 ptlrpc_req_finished(req);
183 static int osc_setattr_interpret(const struct lu_env *env,
184 struct ptlrpc_request *req, void *args, int rc)
186 struct osc_setattr_args *sa = args;
187 struct ost_body *body;
194 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
196 GOTO(out, rc = -EPROTO);
198 lustre_get_wire_obdo(&req->rq_import->imp_connect_data, sa->sa_oa,
201 rc = sa->sa_upcall(sa->sa_cookie, rc);
205 int osc_setattr_async(struct obd_export *exp, struct obdo *oa,
206 obd_enqueue_update_f upcall, void *cookie,
207 struct ptlrpc_request_set *rqset)
209 struct ptlrpc_request *req;
210 struct osc_setattr_args *sa;
215 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
219 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
221 ptlrpc_request_free(req);
225 osc_pack_req_body(req, oa);
227 ptlrpc_request_set_replen(req);
229 /* do mds to ost setattr asynchronously */
231 /* Do not wait for response. */
232 ptlrpcd_add_req(req);
234 req->rq_interpret_reply = osc_setattr_interpret;
236 sa = ptlrpc_req_async_args(sa, req);
238 sa->sa_upcall = upcall;
239 sa->sa_cookie = cookie;
241 ptlrpc_set_add_req(rqset, req);
247 static int osc_ladvise_interpret(const struct lu_env *env,
248 struct ptlrpc_request *req,
251 struct osc_ladvise_args *la = arg;
252 struct ost_body *body;
258 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
260 GOTO(out, rc = -EPROTO);
262 *la->la_oa = body->oa;
264 rc = la->la_upcall(la->la_cookie, rc);
269 * If rqset is NULL, do not wait for response. Upcall and cookie could also
270 * be NULL in this case
272 int osc_ladvise_base(struct obd_export *exp, struct obdo *oa,
273 struct ladvise_hdr *ladvise_hdr,
274 obd_enqueue_update_f upcall, void *cookie,
275 struct ptlrpc_request_set *rqset)
277 struct ptlrpc_request *req;
278 struct ost_body *body;
279 struct osc_ladvise_args *la;
281 struct lu_ladvise *req_ladvise;
282 struct lu_ladvise *ladvise = ladvise_hdr->lah_advise;
283 int num_advise = ladvise_hdr->lah_count;
284 struct ladvise_hdr *req_ladvise_hdr;
287 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_LADVISE);
291 req_capsule_set_size(&req->rq_pill, &RMF_OST_LADVISE, RCL_CLIENT,
292 num_advise * sizeof(*ladvise));
293 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_LADVISE);
295 ptlrpc_request_free(req);
298 req->rq_request_portal = OST_IO_PORTAL;
299 ptlrpc_at_set_req_timeout(req);
301 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
303 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa,
306 req_ladvise_hdr = req_capsule_client_get(&req->rq_pill,
307 &RMF_OST_LADVISE_HDR);
308 memcpy(req_ladvise_hdr, ladvise_hdr, sizeof(*ladvise_hdr));
310 req_ladvise = req_capsule_client_get(&req->rq_pill, &RMF_OST_LADVISE);
311 memcpy(req_ladvise, ladvise, sizeof(*ladvise) * num_advise);
312 ptlrpc_request_set_replen(req);
315 /* Do not wait for response. */
316 ptlrpcd_add_req(req);
320 req->rq_interpret_reply = osc_ladvise_interpret;
321 la = ptlrpc_req_async_args(la, req);
323 la->la_upcall = upcall;
324 la->la_cookie = cookie;
326 ptlrpc_set_add_req(rqset, req);
331 static int osc_create(const struct lu_env *env, struct obd_export *exp,
334 struct ptlrpc_request *req;
335 struct ost_body *body;
340 LASSERT(oa->o_valid & OBD_MD_FLGROUP);
341 LASSERT(fid_seq_is_echo(ostid_seq(&oa->o_oi)));
343 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
345 GOTO(out, rc = -ENOMEM);
347 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_CREATE);
349 ptlrpc_request_free(req);
353 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
356 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
358 ptlrpc_request_set_replen(req);
360 rc = ptlrpc_queue_wait(req);
364 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
366 GOTO(out_req, rc = -EPROTO);
368 CDEBUG(D_INFO, "oa flags %x\n", oa->o_flags);
369 lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oa, &body->oa);
371 oa->o_blksize = cli_brw_size(exp->exp_obd);
372 oa->o_valid |= OBD_MD_FLBLKSZ;
374 CDEBUG(D_HA, "transno: %lld\n",
375 lustre_msg_get_transno(req->rq_repmsg));
377 ptlrpc_req_finished(req);
382 int osc_punch_send(struct obd_export *exp, struct obdo *oa,
383 obd_enqueue_update_f upcall, void *cookie)
385 struct ptlrpc_request *req;
386 struct osc_setattr_args *sa;
387 struct obd_import *imp = class_exp2cliimp(exp);
388 struct ost_body *body;
393 req = ptlrpc_request_alloc(imp, &RQF_OST_PUNCH);
397 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_PUNCH);
399 ptlrpc_request_free(req);
403 osc_set_io_portal(req);
405 ptlrpc_at_set_req_timeout(req);
407 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
409 lustre_set_wire_obdo(&imp->imp_connect_data, &body->oa, oa);
411 ptlrpc_request_set_replen(req);
413 req->rq_interpret_reply = osc_setattr_interpret;
414 sa = ptlrpc_req_async_args(sa, req);
416 sa->sa_upcall = upcall;
417 sa->sa_cookie = cookie;
419 ptlrpcd_add_req(req);
423 EXPORT_SYMBOL(osc_punch_send);
426 * osc_fallocate_base() - Handles fallocate request.
428 * @exp: Export structure
429 * @oa: Attributes passed to OSS from client (obdo structure)
430 * @upcall: Primary & supplementary group information
431 * @cookie: Exclusive identifier
432 * @rqset: Request list.
433 * @mode: Operation done on given range.
435 * osc_fallocate_base() - Handles fallocate requests only. Only block
436 * allocation or standard preallocate operation is supported currently.
437 * Other mode flags is not supported yet. ftruncate(2) or truncate(2)
438 * is supported via SETATTR request.
440 * Return: Non-zero on failure and O on success.
442 int osc_fallocate_base(struct obd_export *exp, struct obdo *oa,
443 obd_enqueue_update_f upcall, void *cookie, int mode)
445 struct ptlrpc_request *req;
446 struct osc_setattr_args *sa;
447 struct ost_body *body;
448 struct obd_import *imp = class_exp2cliimp(exp);
452 oa->o_falloc_mode = mode;
453 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
458 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_FALLOCATE);
460 ptlrpc_request_free(req);
464 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
467 lustre_set_wire_obdo(&imp->imp_connect_data, &body->oa, oa);
469 ptlrpc_request_set_replen(req);
471 req->rq_interpret_reply = osc_setattr_interpret;
472 BUILD_BUG_ON(sizeof(*sa) > sizeof(req->rq_async_args));
473 sa = ptlrpc_req_async_args(sa, req);
475 sa->sa_upcall = upcall;
476 sa->sa_cookie = cookie;
478 ptlrpcd_add_req(req);
482 EXPORT_SYMBOL(osc_fallocate_base);
484 static int osc_sync_interpret(const struct lu_env *env,
485 struct ptlrpc_request *req, void *args, int rc)
487 struct osc_fsync_args *fa = args;
488 struct ost_body *body;
489 struct cl_attr *attr = &osc_env_info(env)->oti_attr;
490 unsigned long valid = 0;
491 struct cl_object *obj;
497 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
499 CERROR("can't unpack ost_body\n");
500 GOTO(out, rc = -EPROTO);
503 *fa->fa_oa = body->oa;
504 obj = osc2cl(fa->fa_obj);
506 /* Update osc object's blocks attribute */
507 cl_object_attr_lock(obj);
508 if (body->oa.o_valid & OBD_MD_FLBLOCKS) {
509 attr->cat_blocks = body->oa.o_blocks;
514 cl_object_attr_update(env, obj, attr, valid);
515 cl_object_attr_unlock(obj);
518 rc = fa->fa_upcall(fa->fa_cookie, rc);
522 int osc_sync_base(struct osc_object *obj, struct obdo *oa,
523 obd_enqueue_update_f upcall, void *cookie,
524 struct ptlrpc_request_set *rqset)
526 struct obd_export *exp = osc_export(obj);
527 struct ptlrpc_request *req;
528 struct ost_body *body;
529 struct osc_fsync_args *fa;
533 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SYNC);
537 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SYNC);
539 ptlrpc_request_free(req);
543 /* overload the size and blocks fields in the oa with start/end */
544 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
546 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
548 ptlrpc_request_set_replen(req);
549 req->rq_interpret_reply = osc_sync_interpret;
551 fa = ptlrpc_req_async_args(fa, req);
554 fa->fa_upcall = upcall;
555 fa->fa_cookie = cookie;
557 ptlrpc_set_add_req(rqset, req);
562 /* Find and cancel locally locks matched by @mode in the resource found by
563 * @objid. Found locks are added into @cancel list. Returns the amount of
564 * locks added to @cancels list. */
565 static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
566 struct list_head *cancels,
567 enum ldlm_mode mode, __u64 lock_flags)
569 struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
570 struct ldlm_res_id res_id;
571 struct ldlm_resource *res;
575 /* Return, i.e. cancel nothing, only if ELC is supported (flag in
576 * export) but disabled through procfs (flag in NS).
578 * This distinguishes from a case when ELC is not supported originally,
579 * when we still want to cancel locks in advance and just cancel them
580 * locally, without sending any RPC. */
581 if (exp_connect_cancelset(exp) && !ns_connect_cancelset(ns))
584 ostid_build_res_name(&oa->o_oi, &res_id);
585 res = ldlm_resource_get(ns, &res_id, 0, 0);
589 LDLM_RESOURCE_ADDREF(res);
590 count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
591 lock_flags, 0, NULL);
592 LDLM_RESOURCE_DELREF(res);
593 ldlm_resource_putref(res);
597 static int osc_destroy_interpret(const struct lu_env *env,
598 struct ptlrpc_request *req, void *args, int rc)
600 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
602 atomic_dec(&cli->cl_destroy_in_flight);
603 wake_up(&cli->cl_destroy_waitq);
608 static int osc_can_send_destroy(struct client_obd *cli)
610 if (atomic_inc_return(&cli->cl_destroy_in_flight) <=
611 cli->cl_max_rpcs_in_flight) {
612 /* The destroy request can be sent */
615 if (atomic_dec_return(&cli->cl_destroy_in_flight) <
616 cli->cl_max_rpcs_in_flight) {
618 * The counter has been modified between the two atomic
621 wake_up(&cli->cl_destroy_waitq);
626 static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
629 struct client_obd *cli = &exp->exp_obd->u.cli;
630 struct ptlrpc_request *req;
631 struct ost_body *body;
637 CDEBUG(D_INFO, "oa NULL\n");
641 count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
642 LDLM_FL_DISCARD_DATA);
644 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_DESTROY);
646 ldlm_lock_list_put(&cancels, l_bl_ast, count);
650 rc = ldlm_prep_elc_req(exp, req, LUSTRE_OST_VERSION, OST_DESTROY,
653 ptlrpc_request_free(req);
657 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
658 ptlrpc_at_set_req_timeout(req);
660 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
662 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
664 ptlrpc_request_set_replen(req);
666 req->rq_interpret_reply = osc_destroy_interpret;
667 if (!osc_can_send_destroy(cli)) {
669 * Wait until the number of on-going destroy RPCs drops
670 * under max_rpc_in_flight
672 rc = l_wait_event_abortable_exclusive(
673 cli->cl_destroy_waitq,
674 osc_can_send_destroy(cli));
676 ptlrpc_req_finished(req);
681 /* Do not wait for response */
682 ptlrpcd_add_req(req);
686 static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
689 u64 bits = OBD_MD_FLBLOCKS | OBD_MD_FLGRANT;
691 LASSERT(!(oa->o_valid & bits));
694 spin_lock(&cli->cl_loi_list_lock);
695 if (cli->cl_ocd_grant_param)
696 oa->o_dirty = cli->cl_dirty_grant;
698 oa->o_dirty = cli->cl_dirty_pages << PAGE_SHIFT;
699 if (unlikely(cli->cl_dirty_pages > cli->cl_dirty_max_pages)) {
700 CERROR("dirty %lu > dirty_max %lu\n",
702 cli->cl_dirty_max_pages);
704 } else if (unlikely(atomic_long_read(&obd_dirty_pages) >
705 (long)(obd_max_dirty_pages + 1))) {
706 /* The atomic_read() allowing the atomic_inc() are
707 * not covered by a lock thus they may safely race and trip
708 * this CERROR() unless we add in a small fudge factor (+1). */
709 CERROR("%s: dirty %ld > system dirty_max %ld\n",
710 cli_name(cli), atomic_long_read(&obd_dirty_pages),
711 obd_max_dirty_pages);
713 } else if (unlikely(cli->cl_dirty_max_pages - cli->cl_dirty_pages >
715 CERROR("dirty %lu - dirty_max %lu too big???\n",
716 cli->cl_dirty_pages, cli->cl_dirty_max_pages);
719 unsigned long nrpages;
720 unsigned long undirty;
722 nrpages = cli->cl_max_pages_per_rpc;
723 nrpages *= cli->cl_max_rpcs_in_flight + 1;
724 nrpages = max(nrpages, cli->cl_dirty_max_pages);
725 undirty = nrpages << PAGE_SHIFT;
726 if (cli->cl_ocd_grant_param) {
729 /* take extent tax into account when asking for more
731 nrextents = (nrpages + cli->cl_max_extent_pages - 1) /
732 cli->cl_max_extent_pages;
733 undirty += nrextents * cli->cl_grant_extent_tax;
735 /* Do not ask for more than OBD_MAX_GRANT - a margin for server
736 * to add extent tax, etc.
738 oa->o_undirty = min(undirty, OBD_MAX_GRANT &
739 ~(PTLRPC_MAX_BRW_SIZE * 4UL));
741 oa->o_grant = cli->cl_avail_grant + cli->cl_reserved_grant;
742 /* o_dropped AKA o_misc is 32 bits, but cl_lost_grant is 64 bits */
743 if (cli->cl_lost_grant > INT_MAX) {
745 "%s: avoided o_dropped overflow: cl_lost_grant %lu\n",
746 cli_name(cli), cli->cl_lost_grant);
747 oa->o_dropped = INT_MAX;
749 oa->o_dropped = cli->cl_lost_grant;
751 cli->cl_lost_grant -= oa->o_dropped;
752 spin_unlock(&cli->cl_loi_list_lock);
753 CDEBUG(D_CACHE, "%s: dirty: %llu undirty: %u dropped %u grant: %llu"
754 " cl_lost_grant %lu\n", cli_name(cli), oa->o_dirty,
755 oa->o_undirty, oa->o_dropped, oa->o_grant, cli->cl_lost_grant);
758 void osc_update_next_shrink(struct client_obd *cli)
760 cli->cl_next_shrink_grant = ktime_get_seconds() +
761 cli->cl_grant_shrink_interval;
763 CDEBUG(D_CACHE, "next time %lld to shrink grant\n",
764 cli->cl_next_shrink_grant);
766 EXPORT_SYMBOL(osc_update_next_shrink);
768 static void __osc_update_grant(struct client_obd *cli, u64 grant)
770 spin_lock(&cli->cl_loi_list_lock);
771 cli->cl_avail_grant += grant;
772 spin_unlock(&cli->cl_loi_list_lock);
775 static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
777 if (body->oa.o_valid & OBD_MD_FLGRANT) {
778 CDEBUG(D_CACHE, "got %llu extra grant\n", body->oa.o_grant);
779 __osc_update_grant(cli, body->oa.o_grant);
784 * grant thread data for shrinking space.
786 struct grant_thread_data {
787 struct list_head gtd_clients;
788 struct mutex gtd_mutex;
789 unsigned long gtd_stopped:1;
791 static struct grant_thread_data client_gtd;
793 static int osc_shrink_grant_interpret(const struct lu_env *env,
794 struct ptlrpc_request *req,
797 struct osc_grant_args *aa = args;
798 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
799 struct ost_body *body;
802 __osc_update_grant(cli, aa->aa_oa->o_grant);
806 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
808 osc_update_grant(cli, body);
810 OBD_SLAB_FREE_PTR(aa->aa_oa, osc_obdo_kmem);
816 static void osc_shrink_grant_local(struct client_obd *cli, struct obdo *oa)
818 spin_lock(&cli->cl_loi_list_lock);
819 oa->o_grant = cli->cl_avail_grant / 4;
820 cli->cl_avail_grant -= oa->o_grant;
821 spin_unlock(&cli->cl_loi_list_lock);
822 if (!(oa->o_valid & OBD_MD_FLFLAGS)) {
823 oa->o_valid |= OBD_MD_FLFLAGS;
826 oa->o_flags |= OBD_FL_SHRINK_GRANT;
827 osc_update_next_shrink(cli);
830 /* Shrink the current grant, either from some large amount to enough for a
831 * full set of in-flight RPCs, or if we have already shrunk to that limit
832 * then to enough for a single RPC. This avoids keeping more grant than
833 * needed, and avoids shrinking the grant piecemeal. */
834 static int osc_shrink_grant(struct client_obd *cli)
836 __u64 target_bytes = (cli->cl_max_rpcs_in_flight + 1) *
837 (cli->cl_max_pages_per_rpc << PAGE_SHIFT);
839 spin_lock(&cli->cl_loi_list_lock);
840 if (cli->cl_avail_grant <= target_bytes)
841 target_bytes = cli->cl_max_pages_per_rpc << PAGE_SHIFT;
842 spin_unlock(&cli->cl_loi_list_lock);
844 return osc_shrink_grant_to_target(cli, target_bytes);
847 int osc_shrink_grant_to_target(struct client_obd *cli, __u64 target_bytes)
850 struct ost_body *body;
853 spin_lock(&cli->cl_loi_list_lock);
854 /* Don't shrink if we are already above or below the desired limit
855 * We don't want to shrink below a single RPC, as that will negatively
856 * impact block allocation and long-term performance. */
857 if (target_bytes < cli->cl_max_pages_per_rpc << PAGE_SHIFT)
858 target_bytes = cli->cl_max_pages_per_rpc << PAGE_SHIFT;
860 if (target_bytes >= cli->cl_avail_grant) {
861 spin_unlock(&cli->cl_loi_list_lock);
864 spin_unlock(&cli->cl_loi_list_lock);
870 osc_announce_cached(cli, &body->oa, 0);
872 spin_lock(&cli->cl_loi_list_lock);
873 if (target_bytes >= cli->cl_avail_grant) {
874 /* available grant has changed since target calculation */
875 spin_unlock(&cli->cl_loi_list_lock);
876 GOTO(out_free, rc = 0);
878 body->oa.o_grant = cli->cl_avail_grant - target_bytes;
879 cli->cl_avail_grant = target_bytes;
880 spin_unlock(&cli->cl_loi_list_lock);
881 if (!(body->oa.o_valid & OBD_MD_FLFLAGS)) {
882 body->oa.o_valid |= OBD_MD_FLFLAGS;
883 body->oa.o_flags = 0;
885 body->oa.o_flags |= OBD_FL_SHRINK_GRANT;
886 osc_update_next_shrink(cli);
888 rc = osc_set_info_async(NULL, cli->cl_import->imp_obd->obd_self_export,
889 sizeof(KEY_GRANT_SHRINK), KEY_GRANT_SHRINK,
890 sizeof(*body), body, NULL);
892 __osc_update_grant(cli, body->oa.o_grant);
898 static int osc_should_shrink_grant(struct client_obd *client)
900 time64_t next_shrink = client->cl_next_shrink_grant;
902 if (client->cl_import == NULL)
905 if (!OCD_HAS_FLAG(&client->cl_import->imp_connect_data, GRANT_SHRINK) ||
906 client->cl_import->imp_grant_shrink_disabled) {
907 osc_update_next_shrink(client);
911 if (ktime_get_seconds() >= next_shrink - 5) {
912 /* Get the current RPC size directly, instead of going via:
913 * cli_brw_size(obd->u.cli.cl_import->imp_obd->obd_self_export)
914 * Keep comment here so that it can be found by searching. */
915 int brw_size = client->cl_max_pages_per_rpc << PAGE_SHIFT;
917 if (client->cl_import->imp_state == LUSTRE_IMP_FULL &&
918 client->cl_avail_grant > brw_size)
921 osc_update_next_shrink(client);
926 #define GRANT_SHRINK_RPC_BATCH 100
928 static struct delayed_work work;
930 static void osc_grant_work_handler(struct work_struct *data)
932 struct client_obd *cli;
934 bool init_next_shrink = true;
935 time64_t next_shrink = ktime_get_seconds() + GRANT_SHRINK_INTERVAL;
938 mutex_lock(&client_gtd.gtd_mutex);
939 list_for_each_entry(cli, &client_gtd.gtd_clients,
941 if (rpc_sent < GRANT_SHRINK_RPC_BATCH &&
942 osc_should_shrink_grant(cli)) {
943 osc_shrink_grant(cli);
947 if (!init_next_shrink) {
948 if (cli->cl_next_shrink_grant < next_shrink &&
949 cli->cl_next_shrink_grant > ktime_get_seconds())
950 next_shrink = cli->cl_next_shrink_grant;
952 init_next_shrink = false;
953 next_shrink = cli->cl_next_shrink_grant;
956 mutex_unlock(&client_gtd.gtd_mutex);
958 if (client_gtd.gtd_stopped == 1)
961 if (next_shrink > ktime_get_seconds()) {
962 time64_t delay = next_shrink - ktime_get_seconds();
964 schedule_delayed_work(&work, cfs_time_seconds(delay));
966 schedule_work(&work.work);
970 void osc_schedule_grant_work(void)
972 cancel_delayed_work_sync(&work);
973 schedule_work(&work.work);
975 EXPORT_SYMBOL(osc_schedule_grant_work);
978 * Start grant thread for returing grant to server for idle clients.
980 static int osc_start_grant_work(void)
982 client_gtd.gtd_stopped = 0;
983 mutex_init(&client_gtd.gtd_mutex);
984 INIT_LIST_HEAD(&client_gtd.gtd_clients);
986 INIT_DELAYED_WORK(&work, osc_grant_work_handler);
987 schedule_work(&work.work);
992 static void osc_stop_grant_work(void)
994 client_gtd.gtd_stopped = 1;
995 cancel_delayed_work_sync(&work);
998 static void osc_add_grant_list(struct client_obd *client)
1000 mutex_lock(&client_gtd.gtd_mutex);
1001 list_add(&client->cl_grant_chain, &client_gtd.gtd_clients);
1002 mutex_unlock(&client_gtd.gtd_mutex);
1005 static void osc_del_grant_list(struct client_obd *client)
1007 if (list_empty(&client->cl_grant_chain))
1010 mutex_lock(&client_gtd.gtd_mutex);
1011 list_del_init(&client->cl_grant_chain);
1012 mutex_unlock(&client_gtd.gtd_mutex);
1015 void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
1018 * ocd_grant is the total grant amount we're expect to hold: if we've
1019 * been evicted, it's the new avail_grant amount, cl_dirty_pages will
1020 * drop to 0 as inflight RPCs fail out; otherwise, it's avail_grant +
1023 * race is tolerable here: if we're evicted, but imp_state already
1024 * left EVICTED state, then cl_dirty_pages must be 0 already.
1026 spin_lock(&cli->cl_loi_list_lock);
1027 cli->cl_avail_grant = ocd->ocd_grant;
1028 if (cli->cl_import->imp_state != LUSTRE_IMP_EVICTED) {
1029 unsigned long consumed = cli->cl_reserved_grant;
1031 if (OCD_HAS_FLAG(ocd, GRANT_PARAM))
1032 consumed += cli->cl_dirty_grant;
1034 consumed += cli->cl_dirty_pages << PAGE_SHIFT;
1035 if (cli->cl_avail_grant < consumed) {
1036 CERROR("%s: granted %ld but already consumed %ld\n",
1037 cli_name(cli), cli->cl_avail_grant, consumed);
1038 cli->cl_avail_grant = 0;
1040 cli->cl_avail_grant -= consumed;
1044 if (OCD_HAS_FLAG(ocd, GRANT_PARAM)) {
1048 /* overhead for each extent insertion */
1049 cli->cl_grant_extent_tax = ocd->ocd_grant_tax_kb << 10;
1050 /* determine the appropriate chunk size used by osc_extent. */
1051 cli->cl_chunkbits = max_t(int, PAGE_SHIFT,
1052 ocd->ocd_grant_blkbits);
1053 /* max_pages_per_rpc must be chunk aligned */
1054 chunk_mask = ~((1 << (cli->cl_chunkbits - PAGE_SHIFT)) - 1);
1055 cli->cl_max_pages_per_rpc = (cli->cl_max_pages_per_rpc +
1056 ~chunk_mask) & chunk_mask;
1057 /* determine maximum extent size, in #pages */
1058 size = (u64)ocd->ocd_grant_max_blks << ocd->ocd_grant_blkbits;
1059 cli->cl_max_extent_pages = (size >> PAGE_SHIFT) ?: 1;
1060 cli->cl_ocd_grant_param = 1;
1062 cli->cl_ocd_grant_param = 0;
1063 cli->cl_grant_extent_tax = 0;
1064 cli->cl_chunkbits = PAGE_SHIFT;
1065 cli->cl_max_extent_pages = DT_MAX_BRW_PAGES;
1067 spin_unlock(&cli->cl_loi_list_lock);
1070 "%s, setting cl_avail_grant: %ld cl_lost_grant: %ld. chunk bits: %d cl_max_extent_pages: %d\n",
1072 cli->cl_avail_grant, cli->cl_lost_grant, cli->cl_chunkbits,
1073 cli->cl_max_extent_pages);
1075 if (OCD_HAS_FLAG(ocd, GRANT_SHRINK) && list_empty(&cli->cl_grant_chain))
1076 osc_add_grant_list(cli);
1078 EXPORT_SYMBOL(osc_init_grant);
1080 /* We assume that the reason this OSC got a short read is because it read
1081 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
1082 * via the LOV, and it _knows_ it's reading inside the file, it's just that
1083 * this stripe never got written at or beyond this stripe offset yet. */
1084 static void handle_short_read(int nob_read, size_t page_count,
1085 struct brw_page **pga)
1090 /* skip bytes read OK */
1091 while (nob_read > 0) {
1092 LASSERT (page_count > 0);
1094 if (pga[i]->count > nob_read) {
1095 /* EOF inside this page */
1096 ptr = kmap(pga[i]->pg) +
1097 (pga[i]->off & ~PAGE_MASK);
1098 memset(ptr + nob_read, 0, pga[i]->count - nob_read);
1105 nob_read -= pga[i]->count;
1110 /* zero remaining pages */
1111 while (page_count-- > 0) {
1112 ptr = kmap(pga[i]->pg) + (pga[i]->off & ~PAGE_MASK);
1113 memset(ptr, 0, pga[i]->count);
1119 static int check_write_rcs(struct ptlrpc_request *req,
1120 int requested_nob, int niocount,
1121 size_t page_count, struct brw_page **pga)
1126 remote_rcs = req_capsule_server_sized_get(&req->rq_pill, &RMF_RCS,
1127 sizeof(*remote_rcs) *
1129 if (remote_rcs == NULL) {
1130 CDEBUG(D_INFO, "Missing/short RC vector on BRW_WRITE reply\n");
1134 /* return error if any niobuf was in error */
1135 for (i = 0; i < niocount; i++) {
1136 if ((int)remote_rcs[i] < 0) {
1137 CDEBUG(D_INFO, "rc[%d]: %d req %p\n",
1138 i, remote_rcs[i], req);
1139 return remote_rcs[i];
1142 if (remote_rcs[i] != 0) {
1143 CDEBUG(D_INFO, "rc[%d] invalid (%d) req %p\n",
1144 i, remote_rcs[i], req);
1148 if (req->rq_bulk != NULL &&
1149 req->rq_bulk->bd_nob_transferred != requested_nob) {
1150 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
1151 req->rq_bulk->bd_nob_transferred, requested_nob);
1158 static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
1160 if (p1->flag != p2->flag) {
1161 unsigned mask = ~(OBD_BRW_FROM_GRANT | OBD_BRW_NOCACHE |
1162 OBD_BRW_SYNC | OBD_BRW_ASYNC |
1163 OBD_BRW_NOQUOTA | OBD_BRW_SOFT_SYNC |
1164 OBD_BRW_SYS_RESOURCE);
1166 /* warn if we try to combine flags that we don't know to be
1167 * safe to combine */
1168 if (unlikely((p1->flag & mask) != (p2->flag & mask))) {
1169 CWARN("Saw flags 0x%x and 0x%x in the same brw, please "
1170 "report this at https://jira.whamcloud.com/\n",
1171 p1->flag, p2->flag);
1176 return (p1->off + p1->count == p2->off);
1179 #if IS_ENABLED(CONFIG_CRC_T10DIF)
1180 static int osc_checksum_bulk_t10pi(const char *obd_name, int nob,
1181 size_t pg_count, struct brw_page **pga,
1182 int opc, obd_dif_csum_fn *fn,
1184 u32 *check_sum, bool resend)
1186 struct ahash_request *req;
1187 /* Used Adler as the default checksum type on top of DIF tags */
1188 unsigned char cfs_alg = cksum_obd2cfs(OBD_CKSUM_T10_TOP);
1189 struct page *__page;
1190 unsigned char *buffer;
1191 __be16 *guard_start;
1192 unsigned int bufsize;
1194 int used_number = 0;
1200 LASSERT(pg_count > 0);
1202 __page = alloc_page(GFP_KERNEL);
1206 req = cfs_crypto_hash_init(cfs_alg, NULL, 0);
1209 CERROR("%s: unable to initialize checksum hash %s: rc = %d\n",
1210 obd_name, cfs_crypto_hash_name(cfs_alg), rc);
1214 buffer = kmap(__page);
1215 guard_start = (__be16 *)buffer;
1216 guard_number = PAGE_SIZE / sizeof(*guard_start);
1217 CDEBUG(D_PAGE | (resend ? D_HA : 0),
1218 "GRD tags per page=%u, resend=%u, bytes=%u, pages=%zu\n",
1219 guard_number, resend, nob, pg_count);
1221 while (nob > 0 && pg_count > 0) {
1222 unsigned int count = pga[i]->count > nob ? nob : pga[i]->count;
1224 /* corrupt the data before we compute the checksum, to
1225 * simulate an OST->client data error */
1226 if (unlikely(i == 0 && opc == OST_READ &&
1227 OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE))) {
1228 unsigned char *ptr = kmap(pga[i]->pg);
1229 int off = pga[i]->off & ~PAGE_MASK;
1231 memcpy(ptr + off, "bad1", min_t(typeof(nob), 4, nob));
1236 * The left guard number should be able to hold checksums of a
1239 rc = obd_page_dif_generate_buffer(obd_name, pga[i]->pg,
1240 pga[i]->off & ~PAGE_MASK,
1242 guard_start + used_number,
1243 guard_number - used_number,
1246 if (unlikely(resend))
1247 CDEBUG(D_PAGE | D_HA,
1248 "pga[%u]: used %u off %llu+%u gen checksum: %*phN\n",
1249 i, used, pga[i]->off & ~PAGE_MASK, count,
1250 (int)(used * sizeof(*guard_start)),
1251 guard_start + used_number);
1255 used_number += used;
1256 if (used_number == guard_number) {
1257 cfs_crypto_hash_update_page(req, __page, 0,
1258 used_number * sizeof(*guard_start));
1262 nob -= pga[i]->count;
1270 if (used_number != 0)
1271 cfs_crypto_hash_update_page(req, __page, 0,
1272 used_number * sizeof(*guard_start));
1274 bufsize = sizeof(cksum);
1275 cfs_crypto_hash_final(req, (unsigned char *)&cksum, &bufsize);
1277 /* For sending we only compute the wrong checksum instead
1278 * of corrupting the data so it is still correct on a redo */
1279 if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1284 __free_page(__page);
1287 #else /* !CONFIG_CRC_T10DIF */
1288 #define obd_dif_ip_fn NULL
1289 #define obd_dif_crc_fn NULL
1290 #define osc_checksum_bulk_t10pi(name, nob, pgc, pga, opc, fn, ssize, csum, re) \
1292 #endif /* CONFIG_CRC_T10DIF */
1294 static int osc_checksum_bulk(int nob, size_t pg_count,
1295 struct brw_page **pga, int opc,
1296 enum cksum_types cksum_type,
1300 struct ahash_request *req;
1301 unsigned int bufsize;
1302 unsigned char cfs_alg = cksum_obd2cfs(cksum_type);
1304 LASSERT(pg_count > 0);
1306 req = cfs_crypto_hash_init(cfs_alg, NULL, 0);
1308 CERROR("Unable to initialize checksum hash %s\n",
1309 cfs_crypto_hash_name(cfs_alg));
1310 return PTR_ERR(req);
1313 while (nob > 0 && pg_count > 0) {
1314 unsigned int count = pga[i]->count > nob ? nob : pga[i]->count;
1316 /* corrupt the data before we compute the checksum, to
1317 * simulate an OST->client data error */
1318 if (i == 0 && opc == OST_READ &&
1319 OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE)) {
1320 unsigned char *ptr = kmap(pga[i]->pg);
1321 int off = pga[i]->off & ~PAGE_MASK;
1323 memcpy(ptr + off, "bad1", min_t(typeof(nob), 4, nob));
1326 cfs_crypto_hash_update_page(req, pga[i]->pg,
1327 pga[i]->off & ~PAGE_MASK,
1329 LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d\n",
1330 (int)(pga[i]->off & ~PAGE_MASK));
1332 nob -= pga[i]->count;
1337 bufsize = sizeof(*cksum);
1338 cfs_crypto_hash_final(req, (unsigned char *)cksum, &bufsize);
1340 /* For sending we only compute the wrong checksum instead
1341 * of corrupting the data so it is still correct on a redo */
1342 if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1348 static int osc_checksum_bulk_rw(const char *obd_name,
1349 enum cksum_types cksum_type,
1350 int nob, size_t pg_count,
1351 struct brw_page **pga, int opc,
1352 u32 *check_sum, bool resend)
1354 obd_dif_csum_fn *fn = NULL;
1355 int sector_size = 0;
1359 obd_t10_cksum2dif(cksum_type, &fn, §or_size);
1362 rc = osc_checksum_bulk_t10pi(obd_name, nob, pg_count, pga,
1363 opc, fn, sector_size, check_sum,
1366 rc = osc_checksum_bulk(nob, pg_count, pga, opc, cksum_type,
1372 #ifdef CONFIG_LL_ENCRYPTION
1374 * osc_encrypt_pagecache_blocks() - overlay to llcrypt_encrypt_pagecache_blocks
1375 * @srcpage: The locked pagecache page containing the block(s) to encrypt
1376 * @dstpage: The page to put encryption result
1377 * @len: Total size of the block(s) to encrypt. Must be a nonzero
1378 * multiple of the filesystem's block size.
1379 * @offs: Byte offset within @page of the first block to encrypt. Must be
1380 * a multiple of the filesystem's block size.
1381 * @gfp_flags: Memory allocation flags
1383 * This overlay function is necessary to be able to provide our own bounce page.
1385 static struct page *osc_encrypt_pagecache_blocks(struct page *srcpage,
1386 struct page *dstpage,
1392 const struct inode *inode = srcpage->mapping->host;
1393 const unsigned int blockbits = inode->i_blkbits;
1394 const unsigned int blocksize = 1 << blockbits;
1395 u64 lblk_num = ((u64)srcpage->index << (PAGE_SHIFT - blockbits)) +
1396 (offs >> blockbits);
1400 if (unlikely(!dstpage))
1401 return llcrypt_encrypt_pagecache_blocks(srcpage, len, offs,
1404 if (WARN_ON_ONCE(!PageLocked(srcpage)))
1405 return ERR_PTR(-EINVAL);
1407 if (WARN_ON_ONCE(len <= 0 || !IS_ALIGNED(len | offs, blocksize)))
1408 return ERR_PTR(-EINVAL);
1410 /* Set PagePrivate2 for disambiguation in
1411 * osc_finalize_bounce_page().
1412 * It means cipher page was not allocated by llcrypt.
1414 SetPagePrivate2(dstpage);
1416 for (i = offs; i < offs + len; i += blocksize, lblk_num++) {
1417 err = llcrypt_encrypt_block(inode, srcpage, dstpage, blocksize,
1418 i, lblk_num, gfp_flags);
1420 return ERR_PTR(err);
1422 SetPagePrivate(dstpage);
1423 set_page_private(dstpage, (unsigned long)srcpage);
1428 * osc_finalize_bounce_page() - overlay to llcrypt_finalize_bounce_page
1430 * This overlay function is necessary to handle bounce pages
1431 * allocated by ourselves.
1433 static inline void osc_finalize_bounce_page(struct page **pagep)
1435 struct page *page = *pagep;
1437 /* PagePrivate2 was set in osc_encrypt_pagecache_blocks
1438 * to indicate the cipher page was allocated by ourselves.
1439 * So we must not free it via llcrypt.
1441 if (unlikely(!page || !PagePrivate2(page)))
1442 return llcrypt_finalize_bounce_page(pagep);
1444 if (llcrypt_is_bounce_page(page)) {
1445 *pagep = llcrypt_pagecache_page(page);
1446 ClearPagePrivate2(page);
1447 set_page_private(page, (unsigned long)NULL);
1448 ClearPagePrivate(page);
1451 #else /* !CONFIG_LL_ENCRYPTION */
1452 #define osc_encrypt_pagecache_blocks(srcpage, dstpage, len, offs, gfp_flags) \
1453 llcrypt_encrypt_pagecache_blocks(srcpage, len, offs, gfp_flags)
1454 #define osc_finalize_bounce_page(page) llcrypt_finalize_bounce_page(page)
1457 static inline void osc_release_bounce_pages(struct brw_page **pga,
1460 #ifdef HAVE_LUSTRE_CRYPTO
1461 struct page **pa = NULL;
1467 #ifdef CONFIG_LL_ENCRYPTION
1468 if (PageChecked(pga[0]->pg)) {
1469 OBD_ALLOC_PTR_ARRAY_LARGE(pa, page_count);
1475 for (i = 0; i < page_count; i++) {
1476 /* Bounce pages used by osc_encrypt_pagecache_blocks()
1477 * called from osc_brw_prep_request()
1478 * are identified thanks to the PageChecked flag.
1480 if (PageChecked(pga[i]->pg)) {
1482 pa[j++] = pga[i]->pg;
1483 osc_finalize_bounce_page(&pga[i]->pg);
1485 pga[i]->count -= pga[i]->bp_count_diff;
1486 pga[i]->off += pga[i]->bp_off_diff;
1490 sptlrpc_enc_pool_put_pages_array(pa, j);
1491 OBD_FREE_PTR_ARRAY_LARGE(pa, page_count);
1497 osc_brw_prep_request(int cmd, struct client_obd *cli, struct obdo *oa,
1498 u32 page_count, struct brw_page **pga,
1499 struct ptlrpc_request **reqp, int resend)
1501 struct ptlrpc_request *req;
1502 struct ptlrpc_bulk_desc *desc;
1503 struct ost_body *body;
1504 struct obd_ioobj *ioobj;
1505 struct niobuf_remote *niobuf;
1506 int niocount, i, requested_nob, opc, rc, short_io_size = 0;
1507 struct osc_brw_async_args *aa;
1508 struct req_capsule *pill;
1509 struct brw_page *pg_prev;
1511 const char *obd_name = cli->cl_import->imp_obd->obd_name;
1512 struct inode *inode = NULL;
1513 bool directio = false;
1515 bool enable_checksum = true;
1516 struct cl_page *clpage;
1520 clpage = oap2cl_page(brw_page2oap(pga[0]));
1521 inode = clpage->cp_inode;
1522 if (clpage->cp_type == CPT_TRANSIENT)
1525 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ))
1526 RETURN(-ENOMEM); /* Recoverable */
1527 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ2))
1528 RETURN(-EINVAL); /* Fatal */
1530 if ((cmd & OBD_BRW_WRITE) != 0) {
1532 req = ptlrpc_request_alloc_pool(cli->cl_import,
1534 &RQF_OST_BRW_WRITE);
1537 req = ptlrpc_request_alloc(cli->cl_import, &RQF_OST_BRW_READ);
1542 if (opc == OST_WRITE && inode && IS_ENCRYPTED(inode) &&
1543 llcrypt_has_encryption_key(inode)) {
1544 struct page **pa = NULL;
1546 #ifdef CONFIG_LL_ENCRYPTION
1547 OBD_ALLOC_PTR_ARRAY_LARGE(pa, page_count);
1549 ptlrpc_request_free(req);
1553 rc = sptlrpc_enc_pool_get_pages_array(pa, page_count);
1555 CDEBUG(D_SEC, "failed to allocate from enc pool: %d\n",
1557 ptlrpc_request_free(req);
1562 for (i = 0; i < page_count; i++) {
1563 struct brw_page *brwpg = pga[i];
1564 struct page *data_page = NULL;
1565 bool retried = false;
1566 bool lockedbymyself;
1567 u32 nunits = (brwpg->off & ~PAGE_MASK) + brwpg->count;
1568 struct address_space *map_orig = NULL;
1572 nunits = round_up(nunits, LUSTRE_ENCRYPTION_UNIT_SIZE);
1573 /* The page can already be locked when we arrive here.
1574 * This is possible when cl_page_assume/vvp_page_assume
1575 * is stuck on wait_on_page_writeback with page lock
1576 * held. In this case there is no risk for the lock to
1577 * be released while we are doing our encryption
1578 * processing, because writeback against that page will
1579 * end in vvp_page_completion_write/cl_page_completion,
1580 * which means only once the page is fully processed.
1582 lockedbymyself = trylock_page(brwpg->pg);
1584 map_orig = brwpg->pg->mapping;
1585 brwpg->pg->mapping = inode->i_mapping;
1586 index_orig = brwpg->pg->index;
1587 clpage = oap2cl_page(brw_page2oap(brwpg));
1588 brwpg->pg->index = clpage->cp_page_index;
1591 osc_encrypt_pagecache_blocks(brwpg->pg,
1596 brwpg->pg->mapping = map_orig;
1597 brwpg->pg->index = index_orig;
1600 unlock_page(brwpg->pg);
1601 if (IS_ERR(data_page)) {
1602 rc = PTR_ERR(data_page);
1603 if (rc == -ENOMEM && !retried) {
1609 sptlrpc_enc_pool_put_pages_array(pa + i,
1611 OBD_FREE_PTR_ARRAY_LARGE(pa,
1614 ptlrpc_request_free(req);
1617 /* Set PageChecked flag on bounce page for
1618 * disambiguation in osc_release_bounce_pages().
1620 SetPageChecked(data_page);
1621 brwpg->pg = data_page;
1622 /* there should be no gap in the middle of page array */
1623 if (i == page_count - 1) {
1624 struct osc_async_page *oap =
1625 brw_page2oap(brwpg);
1627 oa->o_size = oap->oap_count +
1628 oap->oap_obj_off + oap->oap_page_off;
1630 /* len is forced to nunits, and relative offset to 0
1631 * so store the old, clear text info
1633 brwpg->bp_count_diff = nunits - brwpg->count;
1634 brwpg->count = nunits;
1635 brwpg->bp_off_diff = brwpg->off & ~PAGE_MASK;
1636 brwpg->off = brwpg->off & PAGE_MASK;
1640 OBD_FREE_PTR_ARRAY_LARGE(pa, page_count);
1641 } else if (opc == OST_WRITE && inode && IS_ENCRYPTED(inode)) {
1642 struct osc_async_page *oap = brw_page2oap(pga[0]);
1643 struct cl_page *clpage = oap2cl_page(oap);
1644 struct cl_object *clobj = clpage->cp_obj;
1645 struct cl_attr attr = { 0 };
1649 env = cl_env_get(&refcheck);
1652 ptlrpc_request_free(req);
1656 cl_object_attr_lock(clobj);
1657 rc = cl_object_attr_get(env, clobj, &attr);
1658 cl_object_attr_unlock(clobj);
1659 cl_env_put(env, &refcheck);
1661 ptlrpc_request_free(req);
1665 oa->o_size = attr.cat_size;
1666 } else if (opc == OST_READ && inode && IS_ENCRYPTED(inode) &&
1667 llcrypt_has_encryption_key(inode)) {
1668 for (i = 0; i < page_count; i++) {
1669 struct brw_page *pg = pga[i];
1670 u32 nunits = (pg->off & ~PAGE_MASK) + pg->count;
1672 nunits = round_up(nunits, LUSTRE_ENCRYPTION_UNIT_SIZE);
1673 /* count/off are forced to cover the whole encryption
1674 * unit size so that all encrypted data is stored on the
1675 * OST, so adjust bp_{count,off}_diff for the size of
1678 pg->bp_count_diff = nunits - pg->count;
1680 pg->bp_off_diff = pg->off & ~PAGE_MASK;
1681 pg->off = pg->off & PAGE_MASK;
1685 for (niocount = i = 1; i < page_count; i++) {
1686 if (!can_merge_pages(pga[i - 1], pga[i]))
1690 pill = &req->rq_pill;
1691 req_capsule_set_size(pill, &RMF_OBD_IOOBJ, RCL_CLIENT,
1693 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_CLIENT,
1694 niocount * sizeof(*niobuf));
1696 for (i = 0; i < page_count; i++) {
1697 short_io_size += pga[i]->count;
1698 if (!inode || !IS_ENCRYPTED(inode) ||
1699 !llcrypt_has_encryption_key(inode)) {
1700 pga[i]->bp_count_diff = 0;
1701 pga[i]->bp_off_diff = 0;
1705 if (brw_page2oap(pga[0])->oap_brw_flags & OBD_BRW_RDMA_ONLY) {
1706 enable_checksum = false;
1711 /* Check if read/write is small enough to be a short io. */
1712 if (short_io_size > cli->cl_max_short_io_bytes || niocount > 1 ||
1713 !imp_connect_shortio(cli->cl_import))
1716 /* If this is an empty RPC to old server, just ignore it */
1717 if (!short_io_size && !pga[0]->pg) {
1718 ptlrpc_request_free(req);
1722 req_capsule_set_size(pill, &RMF_SHORT_IO, RCL_CLIENT,
1723 opc == OST_READ ? 0 : short_io_size);
1724 if (opc == OST_READ)
1725 req_capsule_set_size(pill, &RMF_SHORT_IO, RCL_SERVER,
1728 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, opc);
1730 ptlrpc_request_free(req);
1733 osc_set_io_portal(req);
1735 ptlrpc_at_set_req_timeout(req);
1736 /* ask ptlrpc not to resend on EINPROGRESS since BRWs have their own
1738 req->rq_no_retry_einprogress = 1;
1740 if (short_io_size != 0) {
1742 short_io_buf = NULL;
1746 desc = ptlrpc_prep_bulk_imp(req, page_count,
1747 cli->cl_import->imp_connect_data.ocd_brw_size >> LNET_MTU_BITS,
1748 (opc == OST_WRITE ? PTLRPC_BULK_GET_SOURCE :
1749 PTLRPC_BULK_PUT_SINK),
1751 &ptlrpc_bulk_kiov_pin_ops);
1754 GOTO(out, rc = -ENOMEM);
1755 /* NB request now owns desc and will free it when it gets freed */
1756 desc->bd_is_rdma = gpu;
1758 body = req_capsule_client_get(pill, &RMF_OST_BODY);
1759 ioobj = req_capsule_client_get(pill, &RMF_OBD_IOOBJ);
1760 niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1761 LASSERT(body != NULL && ioobj != NULL && niobuf != NULL);
1763 lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
1765 /* For READ and WRITE, we can't fill o_uid and o_gid using from_kuid()
1766 * and from_kgid(), because they are asynchronous. Fortunately, variable
1767 * oa contains valid o_uid and o_gid in these two operations.
1768 * Besides, filling o_uid and o_gid is enough for nrs-tbf, see LU-9658.
1769 * OBD_MD_FLUID and OBD_MD_FLUID is not set in order to avoid breaking
1770 * other process logic */
1771 body->oa.o_uid = oa->o_uid;
1772 body->oa.o_gid = oa->o_gid;
1774 obdo_to_ioobj(oa, ioobj);
1775 ioobj->ioo_bufcnt = niocount;
1776 /* The high bits of ioo_max_brw tells server _maximum_ number of bulks
1777 * that might be send for this request. The actual number is decided
1778 * when the RPC is finally sent in ptlrpc_register_bulk(). It sends
1779 * "max - 1" for old client compatibility sending "0", and also so the
1780 * the actual maximum is a power-of-two number, not one less. LU-1431 */
1782 ioobj_max_brw_set(ioobj, desc->bd_md_max_brw);
1784 ioobj_max_brw_set(ioobj, 0);
1786 if (inode && IS_ENCRYPTED(inode) &&
1787 llcrypt_has_encryption_key(inode) &&
1788 !OBD_FAIL_CHECK(OBD_FAIL_LFSCK_NO_ENCFLAG)) {
1789 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1790 body->oa.o_valid |= OBD_MD_FLFLAGS;
1791 body->oa.o_flags = 0;
1793 body->oa.o_flags |= LUSTRE_ENCRYPT_FL;
1796 if (short_io_size != 0) {
1797 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1798 body->oa.o_valid |= OBD_MD_FLFLAGS;
1799 body->oa.o_flags = 0;
1801 body->oa.o_flags |= OBD_FL_SHORT_IO;
1802 CDEBUG(D_CACHE, "Using short io for data transfer, size = %d\n",
1804 if (opc == OST_WRITE) {
1805 short_io_buf = req_capsule_client_get(pill,
1807 LASSERT(short_io_buf != NULL);
1811 LASSERT(page_count > 0);
1813 for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1814 struct brw_page *pg = pga[i];
1815 int poff = pg->off & ~PAGE_MASK;
1817 LASSERT(pg->count > 0);
1818 /* make sure there is no gap in the middle of page array */
1819 LASSERTF(page_count == 1 ||
1820 (ergo(i == 0, poff + pg->count == PAGE_SIZE) &&
1821 ergo(i > 0 && i < page_count - 1,
1822 poff == 0 && pg->count == PAGE_SIZE) &&
1823 ergo(i == page_count - 1, poff == 0)),
1824 "i: %d/%d pg: %p off: %llu, count: %u\n",
1825 i, page_count, pg, pg->off, pg->count);
1826 LASSERTF(i == 0 || pg->off > pg_prev->off,
1827 "i %d p_c %u pg %p [pri %lu ind %lu] off %llu"
1828 " prev_pg %p [pri %lu ind %lu] off %llu\n",
1830 pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1831 pg_prev->pg, page_private(pg_prev->pg),
1832 pg_prev->pg->index, pg_prev->off);
1833 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1834 (pg->flag & OBD_BRW_SRVLOCK));
1835 if (short_io_size != 0 && opc == OST_WRITE) {
1836 unsigned char *ptr = kmap_atomic(pg->pg);
1838 LASSERT(short_io_size >= requested_nob + pg->count);
1839 memcpy(short_io_buf + requested_nob,
1843 } else if (short_io_size == 0) {
1844 desc->bd_frag_ops->add_kiov_frag(desc, pg->pg, poff,
1847 requested_nob += pg->count;
1849 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1851 niobuf->rnb_len += pg->count;
1853 niobuf->rnb_offset = pg->off;
1854 niobuf->rnb_len = pg->count;
1855 niobuf->rnb_flags = pg->flag;
1860 LASSERTF((void *)(niobuf - niocount) ==
1861 req_capsule_client_get(&req->rq_pill, &RMF_NIOBUF_REMOTE),
1862 "want %p - real %p\n", req_capsule_client_get(&req->rq_pill,
1863 &RMF_NIOBUF_REMOTE), (void *)(niobuf - niocount));
1865 osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1867 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1868 body->oa.o_valid |= OBD_MD_FLFLAGS;
1869 body->oa.o_flags = 0;
1871 body->oa.o_flags |= OBD_FL_RECOV_RESEND;
1874 if (osc_should_shrink_grant(cli))
1875 osc_shrink_grant_local(cli, &body->oa);
1877 if (!cli->cl_checksum || sptlrpc_flavor_has_bulk(&req->rq_flvr))
1878 enable_checksum = false;
1880 /* size[REQ_REC_OFF] still sizeof (*body) */
1881 if (opc == OST_WRITE) {
1882 if (enable_checksum) {
1883 /* store cl_cksum_type in a local variable since
1884 * it can be changed via lprocfs */
1885 enum cksum_types cksum_type = cli->cl_cksum_type;
1887 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1888 body->oa.o_flags = 0;
1890 body->oa.o_flags |= obd_cksum_type_pack(obd_name,
1892 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1894 rc = osc_checksum_bulk_rw(obd_name, cksum_type,
1895 requested_nob, page_count,
1897 &body->oa.o_cksum, resend);
1899 CDEBUG(D_PAGE, "failed to checksum: rc = %d\n",
1903 CDEBUG(D_PAGE | (resend ? D_HA : 0),
1904 "checksum at write origin: %x (%x)\n",
1905 body->oa.o_cksum, cksum_type);
1907 /* save this in 'oa', too, for later checking */
1908 oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1909 oa->o_flags |= obd_cksum_type_pack(obd_name,
1912 /* clear out the checksum flag, in case this is a
1913 * resend but cl_checksum is no longer set. b=11238 */
1914 oa->o_valid &= ~OBD_MD_FLCKSUM;
1916 oa->o_cksum = body->oa.o_cksum;
1917 /* 1 RC per niobuf */
1918 req_capsule_set_size(pill, &RMF_RCS, RCL_SERVER,
1919 sizeof(__u32) * niocount);
1921 if (enable_checksum) {
1922 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1923 body->oa.o_flags = 0;
1924 body->oa.o_flags |= obd_cksum_type_pack(obd_name,
1925 cli->cl_cksum_type);
1926 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1929 /* Client cksum has been already copied to wire obdo in previous
1930 * lustre_set_wire_obdo(), and in the case a bulk-read is being
1931 * resent due to cksum error, this will allow Server to
1932 * check+dump pages on its side */
1934 ptlrpc_request_set_replen(req);
1936 aa = ptlrpc_req_async_args(aa, req);
1938 aa->aa_requested_nob = requested_nob;
1939 aa->aa_nio_count = niocount;
1940 aa->aa_page_count = page_count;
1944 INIT_LIST_HEAD(&aa->aa_oaps);
1947 niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1948 CDEBUG(D_RPCTRACE, "brw rpc %p - object "DOSTID" offset %lld<>%lld\n",
1949 req, POSTID(&oa->o_oi), niobuf[0].rnb_offset,
1950 niobuf[niocount - 1].rnb_offset + niobuf[niocount - 1].rnb_len);
1954 ptlrpc_req_finished(req);
1958 char dbgcksum_file_name[PATH_MAX];
1960 static void dump_all_bulk_pages(struct obdo *oa, __u32 page_count,
1961 struct brw_page **pga, __u32 server_cksum,
1969 /* will only keep dump of pages on first error for the same range in
1970 * file/fid, not during the resends/retries. */
1971 snprintf(dbgcksum_file_name, sizeof(dbgcksum_file_name),
1972 "%s-checksum_dump-osc-"DFID":[%llu-%llu]-%x-%x",
1973 (strncmp(libcfs_debug_file_path, "NONE", 4) != 0 ?
1974 libcfs_debug_file_path : LIBCFS_DEBUG_FILE_PATH_DEFAULT),
1975 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_seq : 0ULL,
1976 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_oid : 0,
1977 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_ver : 0,
1979 pga[page_count-1]->off + pga[page_count-1]->count - 1,
1980 client_cksum, server_cksum);
1981 CWARN("dumping checksum data to %s\n", dbgcksum_file_name);
1982 filp = filp_open(dbgcksum_file_name,
1983 O_CREAT | O_EXCL | O_WRONLY | O_LARGEFILE, 0600);
1987 CDEBUG(D_INFO, "%s: can't open to dump pages with "
1988 "checksum error: rc = %d\n", dbgcksum_file_name,
1991 CERROR("%s: can't open to dump pages with checksum "
1992 "error: rc = %d\n", dbgcksum_file_name, rc);
1996 for (i = 0; i < page_count; i++) {
1997 len = pga[i]->count;
1998 buf = kmap(pga[i]->pg);
2000 rc = cfs_kernel_write(filp, buf, len, &filp->f_pos);
2002 CERROR("%s: wanted to write %u but got %d "
2003 "error\n", dbgcksum_file_name, len, rc);
2012 rc = vfs_fsync_range(filp, 0, LLONG_MAX, 1);
2014 CERROR("%s: sync returns %d\n", dbgcksum_file_name, rc);
2015 filp_close(filp, NULL);
2017 libcfs_debug_dumplog();
2021 check_write_checksum(struct obdo *oa, const struct lnet_processid *peer,
2022 __u32 client_cksum, __u32 server_cksum,
2023 struct osc_brw_async_args *aa)
2025 const char *obd_name = aa->aa_cli->cl_import->imp_obd->obd_name;
2026 enum cksum_types cksum_type;
2027 obd_dif_csum_fn *fn = NULL;
2028 int sector_size = 0;
2033 if (server_cksum == client_cksum) {
2034 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
2038 if (aa->aa_cli->cl_checksum_dump)
2039 dump_all_bulk_pages(oa, aa->aa_page_count, aa->aa_ppga,
2040 server_cksum, client_cksum);
2042 cksum_type = obd_cksum_type_unpack(oa->o_valid & OBD_MD_FLFLAGS ?
2045 switch (cksum_type) {
2046 case OBD_CKSUM_T10IP512:
2050 case OBD_CKSUM_T10IP4K:
2054 case OBD_CKSUM_T10CRC512:
2055 fn = obd_dif_crc_fn;
2058 case OBD_CKSUM_T10CRC4K:
2059 fn = obd_dif_crc_fn;
2067 rc = osc_checksum_bulk_t10pi(obd_name, aa->aa_requested_nob,
2068 aa->aa_page_count, aa->aa_ppga,
2069 OST_WRITE, fn, sector_size,
2072 rc = osc_checksum_bulk(aa->aa_requested_nob, aa->aa_page_count,
2073 aa->aa_ppga, OST_WRITE, cksum_type,
2077 msg = "failed to calculate the client write checksum";
2078 else if (cksum_type != obd_cksum_type_unpack(aa->aa_oa->o_flags))
2079 msg = "the server did not use the checksum type specified in "
2080 "the original request - likely a protocol problem";
2081 else if (new_cksum == server_cksum)
2082 msg = "changed on the client after we checksummed it - "
2083 "likely false positive due to mmap IO (bug 11742)";
2084 else if (new_cksum == client_cksum)
2085 msg = "changed in transit before arrival at OST";
2087 msg = "changed in transit AND doesn't match the original - "
2088 "likely false positive due to mmap IO (bug 11742)";
2090 LCONSOLE_ERROR_MSG(0x132, "%s: BAD WRITE CHECKSUM: %s: from %s inode "
2091 DFID " object "DOSTID" extent [%llu-%llu], original "
2092 "client csum %x (type %x), server csum %x (type %x),"
2093 " client csum now %x\n",
2094 obd_name, msg, libcfs_nidstr(&peer->nid),
2095 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_seq : (__u64)0,
2096 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_oid : 0,
2097 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_ver : 0,
2098 POSTID(&oa->o_oi), aa->aa_ppga[0]->off,
2099 aa->aa_ppga[aa->aa_page_count - 1]->off +
2100 aa->aa_ppga[aa->aa_page_count-1]->count - 1,
2102 obd_cksum_type_unpack(aa->aa_oa->o_flags),
2103 server_cksum, cksum_type, new_cksum);
2107 /* Note rc enters this function as number of bytes transferred */
2108 static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
2110 struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
2111 struct client_obd *cli = aa->aa_cli;
2112 const char *obd_name = cli->cl_import->imp_obd->obd_name;
2113 const struct lnet_processid *peer =
2114 &req->rq_import->imp_connection->c_peer;
2115 struct ost_body *body;
2116 u32 client_cksum = 0;
2117 struct inode *inode = NULL;
2118 unsigned int blockbits = 0, blocksize = 0;
2119 struct cl_page *clpage;
2123 if (rc < 0 && rc != -EDQUOT) {
2124 DEBUG_REQ(D_INFO, req, "Failed request: rc = %d", rc);
2128 LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
2129 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
2131 DEBUG_REQ(D_INFO, req, "cannot unpack body");
2135 /* set/clear over quota flag for a uid/gid/projid */
2136 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
2137 body->oa.o_valid & (OBD_MD_FLALLQUOTA)) {
2138 unsigned qid[LL_MAXQUOTAS] = {
2139 body->oa.o_uid, body->oa.o_gid,
2140 body->oa.o_projid };
2142 "setdq for [%u %u %u] with valid %#llx, flags %x\n",
2143 body->oa.o_uid, body->oa.o_gid, body->oa.o_projid,
2144 body->oa.o_valid, body->oa.o_flags);
2145 osc_quota_setdq(cli, req->rq_xid, qid, body->oa.o_valid,
2149 osc_update_grant(cli, body);
2154 if (aa->aa_oa->o_valid & OBD_MD_FLCKSUM)
2155 client_cksum = aa->aa_oa->o_cksum; /* save for later */
2157 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
2159 CERROR("%s: unexpected positive size %d\n",
2164 if (req->rq_bulk != NULL &&
2165 sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
2168 if ((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) && client_cksum &&
2169 check_write_checksum(&body->oa, peer, client_cksum,
2170 body->oa.o_cksum, aa))
2173 rc = check_write_rcs(req, aa->aa_requested_nob,
2174 aa->aa_nio_count, aa->aa_page_count,
2179 /* The rest of this function executes only for OST_READs */
2181 if (req->rq_bulk == NULL) {
2182 rc = req_capsule_get_size(&req->rq_pill, &RMF_SHORT_IO,
2184 LASSERT(rc == req->rq_status);
2186 /* if unwrap_bulk failed, return -EAGAIN to retry */
2187 rc = sptlrpc_cli_unwrap_bulk_read(req, req->rq_bulk, rc);
2190 GOTO(out, rc = -EAGAIN);
2192 if (rc > aa->aa_requested_nob) {
2193 CERROR("%s: unexpected size %d, requested %d\n", obd_name,
2194 rc, aa->aa_requested_nob);
2198 if (req->rq_bulk != NULL && rc != req->rq_bulk->bd_nob_transferred) {
2199 CERROR("%s: unexpected size %d, transferred %d\n", obd_name,
2200 rc, req->rq_bulk->bd_nob_transferred);
2204 if (req->rq_bulk == NULL) {
2206 int nob, pg_count, i = 0;
2209 CDEBUG(D_CACHE, "Using short io read, size %d\n", rc);
2210 pg_count = aa->aa_page_count;
2211 buf = req_capsule_server_sized_get(&req->rq_pill, &RMF_SHORT_IO,
2214 while (nob > 0 && pg_count > 0) {
2216 int count = aa->aa_ppga[i]->count > nob ?
2217 nob : aa->aa_ppga[i]->count;
2219 CDEBUG(D_CACHE, "page %p count %d\n",
2220 aa->aa_ppga[i]->pg, count);
2221 ptr = kmap_atomic(aa->aa_ppga[i]->pg);
2222 memcpy(ptr + (aa->aa_ppga[i]->off & ~PAGE_MASK), buf,
2224 kunmap_atomic((void *) ptr);
2233 if (rc < aa->aa_requested_nob)
2234 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
2236 if (body->oa.o_valid & OBD_MD_FLCKSUM) {
2237 static int cksum_counter;
2238 u32 server_cksum = body->oa.o_cksum;
2242 enum cksum_types cksum_type;
2243 u32 o_flags = body->oa.o_valid & OBD_MD_FLFLAGS ?
2244 body->oa.o_flags : 0;
2246 cksum_type = obd_cksum_type_unpack(o_flags);
2247 rc = osc_checksum_bulk_rw(obd_name, cksum_type, nob,
2248 aa->aa_page_count, aa->aa_ppga,
2249 OST_READ, &client_cksum, false);
2253 if (req->rq_bulk != NULL &&
2254 !nid_same(&peer->nid, &req->rq_bulk->bd_sender)) {
2256 router = libcfs_nidstr(&req->rq_bulk->bd_sender);
2259 if (server_cksum != client_cksum) {
2260 struct ost_body *clbody;
2261 __u32 client_cksum2;
2262 u32 page_count = aa->aa_page_count;
2264 osc_checksum_bulk_rw(obd_name, cksum_type, nob,
2265 page_count, aa->aa_ppga,
2266 OST_READ, &client_cksum2, true);
2267 clbody = req_capsule_client_get(&req->rq_pill,
2269 if (cli->cl_checksum_dump)
2270 dump_all_bulk_pages(&clbody->oa, page_count,
2271 aa->aa_ppga, server_cksum,
2274 LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from "
2275 "%s%s%s inode "DFID" object "DOSTID
2276 " extent [%llu-%llu], client %x/%x, "
2277 "server %x, cksum_type %x\n",
2279 libcfs_nidstr(&peer->nid),
2281 clbody->oa.o_valid & OBD_MD_FLFID ?
2282 clbody->oa.o_parent_seq : 0ULL,
2283 clbody->oa.o_valid & OBD_MD_FLFID ?
2284 clbody->oa.o_parent_oid : 0,
2285 clbody->oa.o_valid & OBD_MD_FLFID ?
2286 clbody->oa.o_parent_ver : 0,
2287 POSTID(&body->oa.o_oi),
2288 aa->aa_ppga[0]->off,
2289 aa->aa_ppga[page_count-1]->off +
2290 aa->aa_ppga[page_count-1]->count - 1,
2291 client_cksum, client_cksum2,
2292 server_cksum, cksum_type);
2294 aa->aa_oa->o_cksum = client_cksum;
2298 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
2301 } else if (unlikely(client_cksum)) {
2302 static int cksum_missed;
2305 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
2306 CERROR("%s: checksum %u requested from %s but not sent\n",
2307 obd_name, cksum_missed,
2308 libcfs_nidstr(&peer->nid));
2313 /* get the inode from the first cl_page */
2314 clpage = oap2cl_page(brw_page2oap(aa->aa_ppga[0]));
2315 inode = clpage->cp_inode;
2316 if (clpage->cp_type == CPT_TRANSIENT && inode) {
2317 blockbits = inode->i_blkbits;
2318 blocksize = 1 << blockbits;
2320 if (inode && IS_ENCRYPTED(inode)) {
2323 if (!llcrypt_has_encryption_key(inode)) {
2324 CDEBUG(D_SEC, "no enc key for ino %lu\n", inode->i_ino);
2327 for (idx = 0; idx < aa->aa_page_count; idx++) {
2328 struct brw_page *brwpg = aa->aa_ppga[idx];
2329 unsigned int offs = 0;
2331 while (offs < PAGE_SIZE) {
2332 /* do not decrypt if page is all 0s */
2333 if (memchr_inv(page_address(brwpg->pg) + offs,
2334 0, LUSTRE_ENCRYPTION_UNIT_SIZE) == NULL) {
2335 /* if page is empty forward info to
2336 * upper layers (ll_io_zero_page) by
2337 * clearing PagePrivate2
2340 ClearPagePrivate2(brwpg->pg);
2345 /* This is direct IO case. Directly call
2346 * decrypt function that takes inode as
2347 * input parameter. Page does not need
2354 oap2cl_page(brw_page2oap(brwpg));
2356 ((u64)(clpage->cp_page_index) <<
2357 (PAGE_SHIFT - blockbits)) +
2358 (offs >> blockbits);
2361 LUSTRE_ENCRYPTION_UNIT_SIZE;
2362 i += blocksize, lblk_num++) {
2364 llcrypt_decrypt_block_inplace(
2372 rc = llcrypt_decrypt_pagecache_blocks(
2374 LUSTRE_ENCRYPTION_UNIT_SIZE,
2380 offs += LUSTRE_ENCRYPTION_UNIT_SIZE;
2387 lustre_get_wire_obdo(&req->rq_import->imp_connect_data,
2388 aa->aa_oa, &body->oa);
2393 static int osc_brw_redo_request(struct ptlrpc_request *request,
2394 struct osc_brw_async_args *aa, int rc)
2396 struct ptlrpc_request *new_req;
2397 struct osc_brw_async_args *new_aa;
2398 struct osc_async_page *oap;
2401 /* The below message is checked in replay-ost-single.sh test_8ae*/
2402 DEBUG_REQ(rc == -EINPROGRESS ? D_RPCTRACE : D_ERROR, request,
2403 "redo for recoverable error %d", rc);
2405 rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
2406 OST_WRITE ? OBD_BRW_WRITE : OBD_BRW_READ,
2407 aa->aa_cli, aa->aa_oa, aa->aa_page_count,
2408 aa->aa_ppga, &new_req, 1);
2412 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2413 if (oap->oap_request != NULL) {
2414 LASSERTF(request == oap->oap_request,
2415 "request %p != oap_request %p\n",
2416 request, oap->oap_request);
2420 * New request takes over pga and oaps from old request.
2421 * Note that copying a list_head doesn't work, need to move it...
2424 new_req->rq_interpret_reply = request->rq_interpret_reply;
2425 new_req->rq_async_args = request->rq_async_args;
2426 new_req->rq_commit_cb = request->rq_commit_cb;
2427 /* cap resend delay to the current request timeout, this is similar to
2428 * what ptlrpc does (see after_reply()) */
2429 if (aa->aa_resends > new_req->rq_timeout)
2430 new_req->rq_sent = ktime_get_real_seconds() + new_req->rq_timeout;
2432 new_req->rq_sent = ktime_get_real_seconds() + aa->aa_resends;
2433 new_req->rq_generation_set = 1;
2434 new_req->rq_import_generation = request->rq_import_generation;
2436 new_aa = ptlrpc_req_async_args(new_aa, new_req);
2438 INIT_LIST_HEAD(&new_aa->aa_oaps);
2439 list_splice_init(&aa->aa_oaps, &new_aa->aa_oaps);
2440 INIT_LIST_HEAD(&new_aa->aa_exts);
2441 list_splice_init(&aa->aa_exts, &new_aa->aa_exts);
2442 new_aa->aa_resends = aa->aa_resends;
2444 list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
2445 if (oap->oap_request) {
2446 ptlrpc_req_finished(oap->oap_request);
2447 oap->oap_request = ptlrpc_request_addref(new_req);
2451 /* XXX: This code will run into problem if we're going to support
2452 * to add a series of BRW RPCs into a self-defined ptlrpc_request_set
2453 * and wait for all of them to be finished. We should inherit request
2454 * set from old request. */
2455 ptlrpcd_add_req(new_req);
2457 DEBUG_REQ(D_INFO, new_req, "new request");
2462 * ugh, we want disk allocation on the target to happen in offset order. we'll
2463 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
2464 * fine for our small page arrays and doesn't require allocation. its an
2465 * insertion sort that swaps elements that are strides apart, shrinking the
2466 * stride down until its '1' and the array is sorted.
2468 static void sort_brw_pages(struct brw_page **array, int num)
2471 struct brw_page *tmp;
2475 for (stride = 1; stride < num ; stride = (stride * 3) + 1)
2480 for (i = stride ; i < num ; i++) {
2483 while (j >= stride && array[j - stride]->off > tmp->off) {
2484 array[j] = array[j - stride];
2489 } while (stride > 1);
2492 static void osc_release_ppga(struct brw_page **ppga, size_t count)
2494 LASSERT(ppga != NULL);
2495 OBD_FREE_PTR_ARRAY_LARGE(ppga, count);
2498 static int brw_interpret(const struct lu_env *env,
2499 struct ptlrpc_request *req, void *args, int rc)
2501 struct osc_brw_async_args *aa = args;
2502 struct osc_extent *ext;
2503 struct osc_extent *tmp;
2504 struct client_obd *cli = aa->aa_cli;
2505 unsigned long transferred = 0;
2509 rc = osc_brw_fini_request(req, rc);
2510 CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
2512 /* restore clear text pages */
2513 osc_release_bounce_pages(aa->aa_ppga, aa->aa_page_count);
2516 * When server returns -EINPROGRESS, client should always retry
2517 * regardless of the number of times the bulk was resent already.
2519 if (osc_recoverable_error(rc) && !req->rq_no_delay) {
2520 if (req->rq_import_generation !=
2521 req->rq_import->imp_generation) {
2522 CDEBUG(D_HA, "%s: resend cross eviction for object: "
2523 ""DOSTID", rc = %d.\n",
2524 req->rq_import->imp_obd->obd_name,
2525 POSTID(&aa->aa_oa->o_oi), rc);
2526 } else if (rc == -EINPROGRESS ||
2527 client_should_resend(aa->aa_resends, aa->aa_cli)) {
2528 rc = osc_brw_redo_request(req, aa, rc);
2530 CERROR("%s: too many resent retries for object: "
2531 "%llu:%llu, rc = %d.\n",
2532 req->rq_import->imp_obd->obd_name,
2533 POSTID(&aa->aa_oa->o_oi), rc);
2538 else if (rc == -EAGAIN || rc == -EINPROGRESS)
2543 struct obdo *oa = aa->aa_oa;
2544 struct cl_attr *attr = &osc_env_info(env)->oti_attr;
2545 unsigned long valid = 0;
2546 struct cl_object *obj;
2547 struct osc_async_page *last;
2549 last = brw_page2oap(aa->aa_ppga[aa->aa_page_count - 1]);
2550 obj = osc2cl(last->oap_obj);
2552 cl_object_attr_lock(obj);
2553 if (oa->o_valid & OBD_MD_FLBLOCKS) {
2554 attr->cat_blocks = oa->o_blocks;
2555 valid |= CAT_BLOCKS;
2557 if (oa->o_valid & OBD_MD_FLMTIME) {
2558 attr->cat_mtime = oa->o_mtime;
2561 if (oa->o_valid & OBD_MD_FLATIME) {
2562 attr->cat_atime = oa->o_atime;
2565 if (oa->o_valid & OBD_MD_FLCTIME) {
2566 attr->cat_ctime = oa->o_ctime;
2570 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
2571 struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
2572 loff_t last_off = last->oap_count + last->oap_obj_off +
2575 /* Change file size if this is an out of quota or
2576 * direct IO write and it extends the file size */
2577 if (loi->loi_lvb.lvb_size < last_off) {
2578 attr->cat_size = last_off;
2581 /* Extend KMS if it's not a lockless write */
2582 if (loi->loi_kms < last_off &&
2583 oap2osc_page(last)->ops_srvlock == 0) {
2584 attr->cat_kms = last_off;
2590 cl_object_attr_update(env, obj, attr, valid);
2591 cl_object_attr_unlock(obj);
2593 OBD_SLAB_FREE_PTR(aa->aa_oa, osc_obdo_kmem);
2596 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE && rc == 0)
2597 osc_inc_unstable_pages(req);
2599 list_for_each_entry_safe(ext, tmp, &aa->aa_exts, oe_link) {
2600 list_del_init(&ext->oe_link);
2601 osc_extent_finish(env, ext, 1,
2602 rc && req->rq_no_delay ? -EAGAIN : rc);
2604 LASSERT(list_empty(&aa->aa_exts));
2605 LASSERT(list_empty(&aa->aa_oaps));
2607 transferred = (req->rq_bulk == NULL ? /* short io */
2608 aa->aa_requested_nob :
2609 req->rq_bulk->bd_nob_transferred);
2611 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
2612 ptlrpc_lprocfs_brw(req, transferred);
2614 spin_lock(&cli->cl_loi_list_lock);
2615 /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
2616 * is called so we know whether to go to sync BRWs or wait for more
2617 * RPCs to complete */
2618 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
2619 cli->cl_w_in_flight--;
2621 cli->cl_r_in_flight--;
2622 osc_wake_cache_waiters(cli);
2623 spin_unlock(&cli->cl_loi_list_lock);
2625 osc_io_unplug(env, cli, NULL);
2629 static void brw_commit(struct ptlrpc_request *req)
2631 /* If osc_inc_unstable_pages (via osc_extent_finish) races with
2632 * this called via the rq_commit_cb, I need to ensure
2633 * osc_dec_unstable_pages is still called. Otherwise unstable
2634 * pages may be leaked. */
2635 spin_lock(&req->rq_lock);
2636 if (likely(req->rq_unstable)) {
2637 req->rq_unstable = 0;
2638 spin_unlock(&req->rq_lock);
2640 osc_dec_unstable_pages(req);
2642 req->rq_committed = 1;
2643 spin_unlock(&req->rq_lock);
2648 * Build an RPC by the list of extent @ext_list. The caller must ensure
2649 * that the total pages in this list are NOT over max pages per RPC.
2650 * Extents in the list must be in OES_RPC state.
2652 int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
2653 struct list_head *ext_list, int cmd)
2655 struct ptlrpc_request *req = NULL;
2656 struct osc_extent *ext;
2657 struct brw_page **pga = NULL;
2658 struct osc_brw_async_args *aa = NULL;
2659 struct obdo *oa = NULL;
2660 struct osc_async_page *oap;
2661 struct osc_object *obj = NULL;
2662 struct cl_req_attr *crattr = NULL;
2663 loff_t starting_offset = OBD_OBJECT_EOF;
2664 loff_t ending_offset = 0;
2665 /* '1' for consistency with code that checks !mpflag to restore */
2669 bool soft_sync = false;
2670 bool ndelay = false;
2674 __u32 layout_version = 0;
2675 LIST_HEAD(rpc_list);
2676 struct ost_body *body;
2678 LASSERT(!list_empty(ext_list));
2680 /* add pages into rpc_list to build BRW rpc */
2681 list_for_each_entry(ext, ext_list, oe_link) {
2682 LASSERT(ext->oe_state == OES_RPC);
2683 mem_tight |= ext->oe_memalloc;
2684 grant += ext->oe_grants;
2685 page_count += ext->oe_nr_pages;
2686 layout_version = max(layout_version, ext->oe_layout_version);
2691 soft_sync = osc_over_unstable_soft_limit(cli);
2693 mpflag = memalloc_noreclaim_save();
2695 OBD_ALLOC_PTR_ARRAY_LARGE(pga, page_count);
2697 GOTO(out, rc = -ENOMEM);
2699 OBD_SLAB_ALLOC_PTR_GFP(oa, osc_obdo_kmem, GFP_NOFS);
2701 GOTO(out, rc = -ENOMEM);
2704 list_for_each_entry(ext, ext_list, oe_link) {
2705 list_for_each_entry(oap, &ext->oe_pages, oap_pending_item) {
2707 oap->oap_brw_flags |= OBD_BRW_MEMALLOC;
2709 oap->oap_brw_flags |= OBD_BRW_SOFT_SYNC;
2710 pga[i] = &oap->oap_brw_page;
2711 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
2714 list_add_tail(&oap->oap_rpc_item, &rpc_list);
2715 if (starting_offset == OBD_OBJECT_EOF ||
2716 starting_offset > oap->oap_obj_off)
2717 starting_offset = oap->oap_obj_off;
2719 LASSERT(oap->oap_page_off == 0);
2720 if (ending_offset < oap->oap_obj_off + oap->oap_count)
2721 ending_offset = oap->oap_obj_off +
2724 LASSERT(oap->oap_page_off + oap->oap_count ==
2731 /* first page in the list */
2732 oap = list_first_entry(&rpc_list, typeof(*oap), oap_rpc_item);
2734 crattr = &osc_env_info(env)->oti_req_attr;
2735 memset(crattr, 0, sizeof(*crattr));
2736 crattr->cra_type = (cmd & OBD_BRW_WRITE) ? CRT_WRITE : CRT_READ;
2737 crattr->cra_flags = ~0ULL;
2738 crattr->cra_page = oap2cl_page(oap);
2739 crattr->cra_oa = oa;
2740 cl_req_attr_set(env, osc2cl(obj), crattr);
2742 if (cmd == OBD_BRW_WRITE) {
2743 oa->o_grant_used = grant;
2744 if (layout_version > 0) {
2745 CDEBUG(D_LAYOUT, DFID": write with layout version %u\n",
2746 PFID(&oa->o_oi.oi_fid), layout_version);
2748 oa->o_layout_version = layout_version;
2749 oa->o_valid |= OBD_MD_LAYOUT_VERSION;
2753 sort_brw_pages(pga, page_count);
2754 rc = osc_brw_prep_request(cmd, cli, oa, page_count, pga, &req, 0);
2756 CERROR("prep_req failed: %d\n", rc);
2760 req->rq_commit_cb = brw_commit;
2761 req->rq_interpret_reply = brw_interpret;
2762 req->rq_memalloc = mem_tight != 0;
2763 oap->oap_request = ptlrpc_request_addref(req);
2765 req->rq_no_resend = req->rq_no_delay = 1;
2766 /* probably set a shorter timeout value.
2767 * to handle ETIMEDOUT in brw_interpret() correctly. */
2768 /* lustre_msg_set_timeout(req, req->rq_timeout / 2); */
2771 /* Need to update the timestamps after the request is built in case
2772 * we race with setattr (locally or in queue at OST). If OST gets
2773 * later setattr before earlier BRW (as determined by the request xid),
2774 * the OST will not use BRW timestamps. Sadly, there is no obvious
2775 * way to do this in a single call. bug 10150 */
2776 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
2777 crattr->cra_oa = &body->oa;
2778 crattr->cra_flags = OBD_MD_FLMTIME | OBD_MD_FLCTIME | OBD_MD_FLATIME;
2779 cl_req_attr_set(env, osc2cl(obj), crattr);
2780 lustre_msg_set_uid_gid(req->rq_reqmsg, &crattr->cra_uid,
2782 lustre_msg_set_jobid(req->rq_reqmsg, crattr->cra_jobid);
2784 aa = ptlrpc_req_async_args(aa, req);
2785 INIT_LIST_HEAD(&aa->aa_oaps);
2786 list_splice_init(&rpc_list, &aa->aa_oaps);
2787 INIT_LIST_HEAD(&aa->aa_exts);
2788 list_splice_init(ext_list, &aa->aa_exts);
2790 spin_lock(&cli->cl_loi_list_lock);
2791 starting_offset >>= PAGE_SHIFT;
2792 ending_offset >>= PAGE_SHIFT;
2793 if (cmd == OBD_BRW_READ) {
2794 cli->cl_r_in_flight++;
2795 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
2796 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
2797 lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
2798 starting_offset + 1);
2800 cli->cl_w_in_flight++;
2801 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
2802 lprocfs_oh_tally(&cli->cl_write_rpc_hist, cli->cl_w_in_flight);
2803 lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
2804 starting_offset + 1);
2806 spin_unlock(&cli->cl_loi_list_lock);
2808 DEBUG_REQ(D_INODE, req, "%d pages, aa %p, now %ur/%uw in flight",
2809 page_count, aa, cli->cl_r_in_flight, cli->cl_w_in_flight);
2810 if (libcfs_debug & D_IOTRACE) {
2813 fid.f_seq = crattr->cra_oa->o_parent_seq;
2814 fid.f_oid = crattr->cra_oa->o_parent_oid;
2815 fid.f_ver = crattr->cra_oa->o_parent_ver;
2817 DFID": %d %s pages, start %lld, end %lld, now %ur/%uw in flight\n",
2818 PFID(&fid), page_count,
2819 cmd == OBD_BRW_READ ? "read" : "write", starting_offset,
2820 ending_offset, cli->cl_r_in_flight, cli->cl_w_in_flight);
2822 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_DELAY_IO, cfs_fail_val);
2824 ptlrpcd_add_req(req);
2830 memalloc_noreclaim_restore(mpflag);
2833 LASSERT(req == NULL);
2836 OBD_SLAB_FREE_PTR(oa, osc_obdo_kmem);
2838 osc_release_bounce_pages(pga, page_count);
2839 osc_release_ppga(pga, page_count);
2841 /* this should happen rarely and is pretty bad, it makes the
2842 * pending list not follow the dirty order
2844 while ((ext = list_first_entry_or_null(ext_list,
2846 oe_link)) != NULL) {
2847 list_del_init(&ext->oe_link);
2848 osc_extent_finish(env, ext, 0, rc);
2854 /* This is to refresh our lock in face of no RPCs. */
2855 void osc_send_empty_rpc(struct osc_object *osc, pgoff_t start)
2857 struct ptlrpc_request *req;
2859 struct brw_page bpg = { .off = start, .count = 1};
2860 struct brw_page *pga = &bpg;
2863 memset(&oa, 0, sizeof(oa));
2864 oa.o_oi = osc->oo_oinfo->loi_oi;
2865 oa.o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLFLAGS;
2866 /* For updated servers - don't do a read */
2867 oa.o_flags = OBD_FL_NORPC;
2869 rc = osc_brw_prep_request(OBD_BRW_READ, osc_cli(osc), &oa, 1, &pga,
2872 /* If we succeeded we ship it off, if not there's no point in doing
2873 * anything. Also no resends.
2874 * No interpret callback, no commit callback.
2877 req->rq_no_resend = 1;
2878 ptlrpcd_add_req(req);
2882 static int osc_set_lock_data(struct ldlm_lock *lock, void *data)
2886 LASSERT(lock != NULL);
2888 lock_res_and_lock(lock);
2890 if (lock->l_ast_data == NULL)
2891 lock->l_ast_data = data;
2892 if (lock->l_ast_data == data)
2895 unlock_res_and_lock(lock);
2900 int osc_enqueue_fini(struct ptlrpc_request *req, osc_enqueue_upcall_f upcall,
2901 void *cookie, struct lustre_handle *lockh,
2902 enum ldlm_mode mode, __u64 *flags, bool speculative,
2905 bool intent = *flags & LDLM_FL_HAS_INTENT;
2909 /* The request was created before ldlm_cli_enqueue call. */
2910 if (intent && errcode == ELDLM_LOCK_ABORTED) {
2911 struct ldlm_reply *rep;
2913 rep = req_capsule_server_get(&req->rq_pill, &RMF_DLM_REP);
2914 LASSERT(rep != NULL);
2916 rep->lock_policy_res1 =
2917 ptlrpc_status_ntoh(rep->lock_policy_res1);
2918 if (rep->lock_policy_res1)
2919 errcode = rep->lock_policy_res1;
2921 *flags |= LDLM_FL_LVB_READY;
2922 } else if (errcode == ELDLM_OK) {
2923 *flags |= LDLM_FL_LVB_READY;
2926 /* Call the update callback. */
2927 rc = (*upcall)(cookie, lockh, errcode);
2929 /* release the reference taken in ldlm_cli_enqueue() */
2930 if (errcode == ELDLM_LOCK_MATCHED)
2932 if (errcode == ELDLM_OK && lustre_handle_is_used(lockh))
2933 ldlm_lock_decref(lockh, mode);
2938 int osc_enqueue_interpret(const struct lu_env *env, struct ptlrpc_request *req,
2941 struct osc_enqueue_args *aa = args;
2942 struct ldlm_lock *lock;
2943 struct lustre_handle *lockh = &aa->oa_lockh;
2944 enum ldlm_mode mode = aa->oa_mode;
2945 struct ost_lvb *lvb = aa->oa_lvb;
2946 __u32 lvb_len = sizeof(*lvb);
2948 struct ldlm_enqueue_info einfo = {
2949 .ei_type = aa->oa_type,
2955 /* ldlm_cli_enqueue is holding a reference on the lock, so it must
2957 lock = ldlm_handle2lock(lockh);
2958 LASSERTF(lock != NULL,
2959 "lockh %#llx, req %p, aa %p - client evicted?\n",
2960 lockh->cookie, req, aa);
2962 /* Take an additional reference so that a blocking AST that
2963 * ldlm_cli_enqueue_fini() might post for a failed lock, is guaranteed
2964 * to arrive after an upcall has been executed by
2965 * osc_enqueue_fini(). */
2966 ldlm_lock_addref(lockh, mode);
2968 /* Let cl_lock_state_wait fail with -ERESTARTSYS to unuse sublocks. */
2969 OBD_FAIL_TIMEOUT(OBD_FAIL_LDLM_ENQUEUE_HANG, 2);
2971 /* Let CP AST to grant the lock first. */
2972 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_ENQ_RACE, 1);
2974 if (aa->oa_speculative) {
2975 LASSERT(aa->oa_lvb == NULL);
2976 LASSERT(aa->oa_flags == NULL);
2977 aa->oa_flags = &flags;
2980 /* Complete obtaining the lock procedure. */
2981 rc = ldlm_cli_enqueue_fini(aa->oa_exp, &req->rq_pill, &einfo, 1,
2982 aa->oa_flags, lvb, lvb_len, lockh, rc,
2984 /* Complete osc stuff. */
2985 rc = osc_enqueue_fini(req, aa->oa_upcall, aa->oa_cookie, lockh, mode,
2986 aa->oa_flags, aa->oa_speculative, rc);
2988 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_CANCEL_RACE, 10);
2990 ldlm_lock_decref(lockh, mode);
2991 LDLM_LOCK_PUT(lock);
2995 /* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
2996 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
2997 * other synchronous requests, however keeping some locks and trying to obtain
2998 * others may take a considerable amount of time in a case of ost failure; and
2999 * when other sync requests do not get released lock from a client, the client
3000 * is evicted from the cluster -- such scenarious make the life difficult, so
3001 * release locks just after they are obtained. */
3002 int osc_enqueue_base(struct obd_export *exp, struct ldlm_res_id *res_id,
3003 __u64 *flags, union ldlm_policy_data *policy,
3004 struct ost_lvb *lvb, osc_enqueue_upcall_f upcall,
3005 void *cookie, struct ldlm_enqueue_info *einfo,
3006 struct ptlrpc_request_set *rqset, int async,
3009 struct obd_device *obd = exp->exp_obd;
3010 struct lustre_handle lockh = { 0 };
3011 struct ptlrpc_request *req = NULL;
3012 int intent = *flags & LDLM_FL_HAS_INTENT;
3013 __u64 search_flags = *flags;
3014 __u64 match_flags = 0;
3015 enum ldlm_mode mode;
3019 /* Filesystem lock extents are extended to page boundaries so that
3020 * dealing with the page cache is a little smoother. */
3021 policy->l_extent.start -= policy->l_extent.start & ~PAGE_MASK;
3022 policy->l_extent.end |= ~PAGE_MASK;
3024 /* Next, search for already existing extent locks that will cover us */
3025 /* If we're trying to read, we also search for an existing PW lock. The
3026 * VFS and page cache already protect us locally, so lots of readers/
3027 * writers can share a single PW lock.
3029 * There are problems with conversion deadlocks, so instead of
3030 * converting a read lock to a write lock, we'll just enqueue a new
3033 * At some point we should cancel the read lock instead of making them
3034 * send us a blocking callback, but there are problems with canceling
3035 * locks out from other users right now, too. */
3036 mode = einfo->ei_mode;
3037 if (einfo->ei_mode == LCK_PR)
3039 /* Normal lock requests must wait for the LVB to be ready before
3040 * matching a lock; speculative lock requests do not need to,
3041 * because they will not actually use the lock. */
3043 search_flags |= LDLM_FL_LVB_READY;
3045 search_flags |= LDLM_FL_BLOCK_GRANTED;
3046 if (mode == LCK_GROUP)
3047 match_flags = LDLM_MATCH_GROUP;
3048 mode = ldlm_lock_match_with_skip(obd->obd_namespace, search_flags, 0,
3049 res_id, einfo->ei_type, policy, mode,
3050 &lockh, match_flags);
3052 struct ldlm_lock *matched;
3054 if (*flags & LDLM_FL_TEST_LOCK)
3057 matched = ldlm_handle2lock(&lockh);
3059 /* This DLM lock request is speculative, and does not
3060 * have an associated IO request. Therefore if there
3061 * is already a DLM lock, it wll just inform the
3062 * caller to cancel the request for this stripe.*/
3063 lock_res_and_lock(matched);
3064 if (ldlm_extent_equal(&policy->l_extent,
3065 &matched->l_policy_data.l_extent))
3069 unlock_res_and_lock(matched);
3071 ldlm_lock_decref(&lockh, mode);
3072 LDLM_LOCK_PUT(matched);
3074 } else if (osc_set_lock_data(matched, einfo->ei_cbdata)) {
3075 *flags |= LDLM_FL_LVB_READY;
3077 /* We already have a lock, and it's referenced. */
3078 (*upcall)(cookie, &lockh, ELDLM_LOCK_MATCHED);
3080 ldlm_lock_decref(&lockh, mode);
3081 LDLM_LOCK_PUT(matched);
3084 ldlm_lock_decref(&lockh, mode);
3085 LDLM_LOCK_PUT(matched);
3089 if (*flags & (LDLM_FL_TEST_LOCK | LDLM_FL_MATCH_LOCK))
3092 /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
3093 *flags &= ~LDLM_FL_BLOCK_GRANTED;
3095 rc = ldlm_cli_enqueue(exp, &req, einfo, res_id, policy, flags, lvb,
3096 sizeof(*lvb), LVB_T_OST, &lockh, async);
3099 struct osc_enqueue_args *aa;
3100 aa = ptlrpc_req_async_args(aa, req);
3102 aa->oa_mode = einfo->ei_mode;
3103 aa->oa_type = einfo->ei_type;
3104 lustre_handle_copy(&aa->oa_lockh, &lockh);
3105 aa->oa_upcall = upcall;
3106 aa->oa_cookie = cookie;
3107 aa->oa_speculative = speculative;
3109 aa->oa_flags = flags;
3112 /* speculative locks are essentially to enqueue
3113 * a DLM lock in advance, so we don't care
3114 * about the result of the enqueue. */
3116 aa->oa_flags = NULL;
3119 req->rq_interpret_reply = osc_enqueue_interpret;
3120 ptlrpc_set_add_req(rqset, req);
3125 rc = osc_enqueue_fini(req, upcall, cookie, &lockh, einfo->ei_mode,
3126 flags, speculative, rc);
3131 int osc_match_base(const struct lu_env *env, struct obd_export *exp,
3132 struct ldlm_res_id *res_id, enum ldlm_type type,
3133 union ldlm_policy_data *policy, enum ldlm_mode mode,
3134 __u64 *flags, struct osc_object *obj,
3135 struct lustre_handle *lockh, enum ldlm_match_flags match_flags)
3137 struct obd_device *obd = exp->exp_obd;
3138 __u64 lflags = *flags;
3142 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_MATCH))
3145 /* Filesystem lock extents are extended to page boundaries so that
3146 * dealing with the page cache is a little smoother */
3147 policy->l_extent.start -= policy->l_extent.start & ~PAGE_MASK;
3148 policy->l_extent.end |= ~PAGE_MASK;
3150 /* Next, search for already existing extent locks that will cover us */
3151 rc = ldlm_lock_match_with_skip(obd->obd_namespace, lflags, 0,
3152 res_id, type, policy, mode, lockh,
3154 if (rc == 0 || lflags & LDLM_FL_TEST_LOCK)
3158 struct ldlm_lock *lock = ldlm_handle2lock(lockh);
3160 LASSERT(lock != NULL);
3161 if (osc_set_lock_data(lock, obj)) {
3162 lock_res_and_lock(lock);
3163 if (!ldlm_is_lvb_cached(lock)) {
3164 LASSERT(lock->l_ast_data == obj);
3165 osc_lock_lvb_update(env, obj, lock, NULL);
3166 ldlm_set_lvb_cached(lock);
3168 unlock_res_and_lock(lock);
3170 ldlm_lock_decref(lockh, rc);
3173 LDLM_LOCK_PUT(lock);
3178 static int osc_statfs_interpret(const struct lu_env *env,
3179 struct ptlrpc_request *req, void *args, int rc)
3181 struct osc_async_args *aa = args;
3182 struct obd_statfs *msfs;
3187 * The request has in fact never been sent due to issues at
3188 * a higher level (LOV). Exit immediately since the caller
3189 * is aware of the problem and takes care of the clean up.
3193 if ((rc == -ENOTCONN || rc == -EAGAIN) &&
3194 (aa->aa_oi->oi_flags & OBD_STATFS_NODELAY))
3200 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
3202 GOTO(out, rc = -EPROTO);
3204 *aa->aa_oi->oi_osfs = *msfs;
3206 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
3211 static int osc_statfs_async(struct obd_export *exp,
3212 struct obd_info *oinfo, time64_t max_age,
3213 struct ptlrpc_request_set *rqset)
3215 struct obd_device *obd = class_exp2obd(exp);
3216 struct ptlrpc_request *req;
3217 struct osc_async_args *aa;
3221 if (obd->obd_osfs_age >= max_age) {
3223 "%s: use %p cache blocks %llu/%llu objects %llu/%llu\n",
3224 obd->obd_name, &obd->obd_osfs,
3225 obd->obd_osfs.os_bavail, obd->obd_osfs.os_blocks,
3226 obd->obd_osfs.os_ffree, obd->obd_osfs.os_files);
3227 spin_lock(&obd->obd_osfs_lock);
3228 memcpy(oinfo->oi_osfs, &obd->obd_osfs, sizeof(*oinfo->oi_osfs));
3229 spin_unlock(&obd->obd_osfs_lock);
3230 oinfo->oi_flags |= OBD_STATFS_FROM_CACHE;
3231 if (oinfo->oi_cb_up)
3232 oinfo->oi_cb_up(oinfo, 0);
3237 /* We could possibly pass max_age in the request (as an absolute
3238 * timestamp or a "seconds.usec ago") so the target can avoid doing
3239 * extra calls into the filesystem if that isn't necessary (e.g.
3240 * during mount that would help a bit). Having relative timestamps
3241 * is not so great if request processing is slow, while absolute
3242 * timestamps are not ideal because they need time synchronization. */
3243 req = ptlrpc_request_alloc(obd->u.cli.cl_import, &RQF_OST_STATFS);
3247 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
3249 ptlrpc_request_free(req);
3252 ptlrpc_request_set_replen(req);
3253 req->rq_request_portal = OST_CREATE_PORTAL;
3254 ptlrpc_at_set_req_timeout(req);
3256 if (oinfo->oi_flags & OBD_STATFS_NODELAY) {
3257 /* procfs requests not want stat in wait for avoid deadlock */
3258 req->rq_no_resend = 1;
3259 req->rq_no_delay = 1;
3262 req->rq_interpret_reply = osc_statfs_interpret;
3263 aa = ptlrpc_req_async_args(aa, req);
3266 ptlrpc_set_add_req(rqset, req);
3270 static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
3271 struct obd_statfs *osfs, time64_t max_age, __u32 flags)
3273 struct obd_device *obd = class_exp2obd(exp);
3274 struct obd_statfs *msfs;
3275 struct ptlrpc_request *req;
3276 struct obd_import *imp, *imp0;
3280 /*Since the request might also come from lprocfs, so we need
3281 *sync this with client_disconnect_export Bug15684
3283 with_imp_locked(obd, imp0, rc)
3284 imp = class_import_get(imp0);
3288 /* We could possibly pass max_age in the request (as an absolute
3289 * timestamp or a "seconds.usec ago") so the target can avoid doing
3290 * extra calls into the filesystem if that isn't necessary (e.g.
3291 * during mount that would help a bit). Having relative timestamps
3292 * is not so great if request processing is slow, while absolute
3293 * timestamps are not ideal because they need time synchronization. */
3294 req = ptlrpc_request_alloc(imp, &RQF_OST_STATFS);
3296 class_import_put(imp);
3301 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
3303 ptlrpc_request_free(req);
3306 ptlrpc_request_set_replen(req);
3307 req->rq_request_portal = OST_CREATE_PORTAL;
3308 ptlrpc_at_set_req_timeout(req);
3310 if (flags & OBD_STATFS_NODELAY) {
3311 /* procfs requests not want stat in wait for avoid deadlock */
3312 req->rq_no_resend = 1;
3313 req->rq_no_delay = 1;
3316 rc = ptlrpc_queue_wait(req);
3320 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
3322 GOTO(out, rc = -EPROTO);
3328 ptlrpc_req_finished(req);
3332 static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
3333 void *karg, void __user *uarg)
3335 struct obd_device *obd = exp->exp_obd;
3336 struct obd_ioctl_data *data = karg;
3340 CDEBUG(D_IOCTL, "%s: cmd=%x len=%u karg=%pK uarg=%pK\n",
3341 obd->obd_name, cmd, len, karg, uarg);
3343 if (!try_module_get(THIS_MODULE)) {
3344 CERROR("%s: cannot get module '%s'\n", obd->obd_name,
3345 module_name(THIS_MODULE));
3349 case OBD_IOC_CLIENT_RECOVER:
3350 rc = ptlrpc_recover_import(obd->u.cli.cl_import,
3351 data->ioc_inlbuf1, 0);
3355 case OBD_IOC_GETATTR:
3356 rc = obd_getattr(NULL, exp, &data->ioc_obdo1);
3358 case IOC_OSC_SET_ACTIVE:
3359 rc = ptlrpc_set_import_active(obd->u.cli.cl_import,
3363 rc = OBD_IOC_DEBUG(D_IOCTL, obd->obd_name, cmd, "unrecognized",
3368 module_put(THIS_MODULE);
3372 int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
3373 u32 keylen, void *key, u32 vallen, void *val,
3374 struct ptlrpc_request_set *set)
3376 struct ptlrpc_request *req;
3377 struct obd_device *obd = exp->exp_obd;
3378 struct obd_import *imp = class_exp2cliimp(exp);
3383 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
3385 if (KEY_IS(KEY_CHECKSUM)) {
3386 if (vallen != sizeof(int))
3388 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
3392 if (KEY_IS(KEY_SPTLRPC_CONF)) {
3393 sptlrpc_conf_client_adapt(obd);
3397 if (KEY_IS(KEY_FLUSH_CTX)) {
3398 sptlrpc_import_flush_my_ctx(imp);
3402 if (KEY_IS(KEY_CACHE_LRU_SHRINK)) {
3403 struct client_obd *cli = &obd->u.cli;
3404 long nr = atomic_long_read(&cli->cl_lru_in_list) >> 1;
3405 long target = *(long *)val;
3407 nr = osc_lru_shrink(env, cli, min(nr, target), true);
3412 if (!set && !KEY_IS(KEY_GRANT_SHRINK))
3415 /* We pass all other commands directly to OST. Since nobody calls osc
3416 methods directly and everybody is supposed to go through LOV, we
3417 assume lov checked invalid values for us.
3418 The only recognised values so far are evict_by_nid and mds_conn.
3419 Even if something bad goes through, we'd get a -EINVAL from OST
3422 req = ptlrpc_request_alloc(imp, KEY_IS(KEY_GRANT_SHRINK) ?
3423 &RQF_OST_SET_GRANT_INFO :
3428 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3429 RCL_CLIENT, keylen);
3430 if (!KEY_IS(KEY_GRANT_SHRINK))
3431 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_VAL,
3432 RCL_CLIENT, vallen);
3433 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SET_INFO);
3435 ptlrpc_request_free(req);
3439 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3440 memcpy(tmp, key, keylen);
3441 tmp = req_capsule_client_get(&req->rq_pill, KEY_IS(KEY_GRANT_SHRINK) ?
3444 memcpy(tmp, val, vallen);
3446 if (KEY_IS(KEY_GRANT_SHRINK)) {
3447 struct osc_grant_args *aa;
3450 aa = ptlrpc_req_async_args(aa, req);
3451 OBD_SLAB_ALLOC_PTR_GFP(oa, osc_obdo_kmem, GFP_NOFS);
3453 ptlrpc_req_finished(req);
3456 *oa = ((struct ost_body *)val)->oa;
3458 req->rq_interpret_reply = osc_shrink_grant_interpret;
3461 ptlrpc_request_set_replen(req);
3462 if (!KEY_IS(KEY_GRANT_SHRINK)) {
3463 LASSERT(set != NULL);
3464 ptlrpc_set_add_req(set, req);
3465 ptlrpc_check_set(NULL, set);
3467 ptlrpcd_add_req(req);
3472 EXPORT_SYMBOL(osc_set_info_async);
3474 int osc_reconnect(const struct lu_env *env, struct obd_export *exp,
3475 struct obd_device *obd, struct obd_uuid *cluuid,
3476 struct obd_connect_data *data, void *localdata)
3478 struct client_obd *cli = &obd->u.cli;
3480 if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
3484 spin_lock(&cli->cl_loi_list_lock);
3485 grant = cli->cl_avail_grant + cli->cl_reserved_grant;
3486 if (data->ocd_connect_flags & OBD_CONNECT_GRANT_PARAM) {
3487 /* restore ocd_grant_blkbits as client page bits */
3488 data->ocd_grant_blkbits = PAGE_SHIFT;
3489 grant += cli->cl_dirty_grant;
3491 grant += cli->cl_dirty_pages << PAGE_SHIFT;
3493 data->ocd_grant = grant ? : 2 * cli_brw_size(obd);
3494 lost_grant = cli->cl_lost_grant;
3495 cli->cl_lost_grant = 0;
3496 spin_unlock(&cli->cl_loi_list_lock);
3498 CDEBUG(D_RPCTRACE, "ocd_connect_flags: %#llx ocd_version: %d"
3499 " ocd_grant: %d, lost: %ld.\n", data->ocd_connect_flags,
3500 data->ocd_version, data->ocd_grant, lost_grant);
3505 EXPORT_SYMBOL(osc_reconnect);
3507 int osc_disconnect(struct obd_export *exp)
3509 struct obd_device *obd = class_exp2obd(exp);
3512 rc = client_disconnect_export(exp);
3514 * Initially we put del_shrink_grant before disconnect_export, but it
3515 * causes the following problem if setup (connect) and cleanup
3516 * (disconnect) are tangled together.
3517 * connect p1 disconnect p2
3518 * ptlrpc_connect_import
3519 * ............... class_manual_cleanup
3522 * ptlrpc_connect_interrupt
3524 * add this client to shrink list
3526 * Bang! grant shrink thread trigger the shrink. BUG18662
3528 osc_del_grant_list(&obd->u.cli);
3531 EXPORT_SYMBOL(osc_disconnect);
3533 int osc_ldlm_resource_invalidate(struct cfs_hash *hs, struct cfs_hash_bd *bd,
3534 struct hlist_node *hnode, void *arg)
3536 struct lu_env *env = arg;
3537 struct ldlm_resource *res = cfs_hash_object(hs, hnode);
3538 struct ldlm_lock *lock;
3539 struct osc_object *osc = NULL;
3543 list_for_each_entry(lock, &res->lr_granted, l_res_link) {
3544 if (lock->l_ast_data != NULL && osc == NULL) {
3545 osc = lock->l_ast_data;
3546 cl_object_get(osc2cl(osc));
3549 /* clear LDLM_FL_CLEANED flag to make sure it will be canceled
3550 * by the 2nd round of ldlm_namespace_clean() call in
3551 * osc_import_event(). */
3552 ldlm_clear_cleaned(lock);
3557 osc_object_invalidate(env, osc);
3558 cl_object_put(env, osc2cl(osc));
3563 EXPORT_SYMBOL(osc_ldlm_resource_invalidate);
3565 static int osc_import_event(struct obd_device *obd,
3566 struct obd_import *imp,
3567 enum obd_import_event event)
3569 struct client_obd *cli;
3573 LASSERT(imp->imp_obd == obd);
3576 case IMP_EVENT_DISCON: {
3578 spin_lock(&cli->cl_loi_list_lock);
3579 cli->cl_avail_grant = 0;
3580 cli->cl_lost_grant = 0;
3581 spin_unlock(&cli->cl_loi_list_lock);
3584 case IMP_EVENT_INACTIVE: {
3585 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE);
3588 case IMP_EVENT_INVALIDATE: {
3589 struct ldlm_namespace *ns = obd->obd_namespace;
3593 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3595 env = cl_env_get(&refcheck);
3597 osc_io_unplug(env, &obd->u.cli, NULL);
3599 cfs_hash_for_each_nolock(ns->ns_rs_hash,
3600 osc_ldlm_resource_invalidate,
3602 cl_env_put(env, &refcheck);
3604 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3609 case IMP_EVENT_ACTIVE: {
3610 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE);
3613 case IMP_EVENT_OCD: {
3614 struct obd_connect_data *ocd = &imp->imp_connect_data;
3616 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
3617 osc_init_grant(&obd->u.cli, ocd);
3620 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
3621 imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;
3623 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD);
3626 case IMP_EVENT_DEACTIVATE: {
3627 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_DEACTIVATE);
3630 case IMP_EVENT_ACTIVATE: {
3631 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVATE);
3635 CERROR("Unknown import event %d\n", event);
3642 * Determine whether the lock can be canceled before replaying the lock
3643 * during recovery, see bug16774 for detailed information.
3645 * \retval zero the lock can't be canceled
3646 * \retval other ok to cancel
3648 static int osc_cancel_weight(struct ldlm_lock *lock)
3651 * Cancel all unused and granted extent lock.
3653 if (lock->l_resource->lr_type == LDLM_EXTENT &&
3654 ldlm_is_granted(lock) &&
3655 osc_ldlm_weigh_ast(lock) == 0)
3661 static int brw_queue_work(const struct lu_env *env, void *data)
3663 struct client_obd *cli = data;
3665 CDEBUG(D_CACHE, "Run writeback work for client obd %p.\n", cli);
3667 osc_io_unplug(env, cli, NULL);
3671 int osc_setup_common(struct obd_device *obd, struct lustre_cfg *lcfg)
3673 struct client_obd *cli = &obd->u.cli;
3679 rc = ptlrpcd_addref();
3683 rc = client_obd_setup(obd, lcfg);
3685 GOTO(out_ptlrpcd, rc);
3688 handler = ptlrpcd_alloc_work(cli->cl_import, brw_queue_work, cli);
3689 if (IS_ERR(handler))
3690 GOTO(out_ptlrpcd_work, rc = PTR_ERR(handler));
3691 cli->cl_writeback_work = handler;
3693 handler = ptlrpcd_alloc_work(cli->cl_import, lru_queue_work, cli);
3694 if (IS_ERR(handler))
3695 GOTO(out_ptlrpcd_work, rc = PTR_ERR(handler));
3696 cli->cl_lru_work = handler;
3698 rc = osc_quota_setup(obd);
3700 GOTO(out_ptlrpcd_work, rc);
3702 cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
3703 cli->cl_root_squash = 0;
3704 osc_update_next_shrink(cli);
3709 if (cli->cl_writeback_work != NULL) {
3710 ptlrpcd_destroy_work(cli->cl_writeback_work);
3711 cli->cl_writeback_work = NULL;
3713 if (cli->cl_lru_work != NULL) {
3714 ptlrpcd_destroy_work(cli->cl_lru_work);
3715 cli->cl_lru_work = NULL;
3717 client_obd_cleanup(obd);
3722 EXPORT_SYMBOL(osc_setup_common);
3724 int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
3726 struct client_obd *cli = &obd->u.cli;
3734 rc = osc_setup_common(obd, lcfg);
3738 rc = osc_tunables_init(obd);
3743 * We try to control the total number of requests with a upper limit
3744 * osc_reqpool_maxreqcount. There might be some race which will cause
3745 * over-limit allocation, but it is fine.
3747 req_count = atomic_read(&osc_pool_req_count);
3748 if (req_count < osc_reqpool_maxreqcount) {
3749 adding = cli->cl_max_rpcs_in_flight + 2;
3750 if (req_count + adding > osc_reqpool_maxreqcount)
3751 adding = osc_reqpool_maxreqcount - req_count;
3753 added = ptlrpc_add_rqs_to_pool(osc_rq_pool, adding);
3754 atomic_add(added, &osc_pool_req_count);
3757 ns_register_cancel(obd->obd_namespace, osc_cancel_weight);
3759 spin_lock(&osc_shrink_lock);
3760 list_add_tail(&cli->cl_shrink_list, &osc_shrink_list);
3761 spin_unlock(&osc_shrink_lock);
3762 cli->cl_import->imp_idle_timeout = osc_idle_timeout;
3763 cli->cl_import->imp_idle_debug = D_HA;
3768 int osc_precleanup_common(struct obd_device *obd)
3770 struct client_obd *cli = &obd->u.cli;
3774 * for echo client, export may be on zombie list, wait for
3775 * zombie thread to cull it, because cli.cl_import will be
3776 * cleared in client_disconnect_export():
3777 * class_export_destroy() -> obd_cleanup() ->
3778 * echo_device_free() -> echo_client_cleanup() ->
3779 * obd_disconnect() -> osc_disconnect() ->
3780 * client_disconnect_export()
3782 obd_zombie_barrier();
3783 if (cli->cl_writeback_work) {
3784 ptlrpcd_destroy_work(cli->cl_writeback_work);
3785 cli->cl_writeback_work = NULL;
3788 if (cli->cl_lru_work) {
3789 ptlrpcd_destroy_work(cli->cl_lru_work);
3790 cli->cl_lru_work = NULL;
3793 obd_cleanup_client_import(obd);
3796 EXPORT_SYMBOL(osc_precleanup_common);
3798 static int osc_precleanup(struct obd_device *obd)
3802 osc_precleanup_common(obd);
3804 ptlrpc_lprocfs_unregister_obd(obd);
3808 int osc_cleanup_common(struct obd_device *obd)
3810 struct client_obd *cli = &obd->u.cli;
3815 spin_lock(&osc_shrink_lock);
3816 list_del(&cli->cl_shrink_list);
3817 spin_unlock(&osc_shrink_lock);
3820 if (cli->cl_cache != NULL) {
3821 LASSERT(refcount_read(&cli->cl_cache->ccc_users) > 0);
3822 spin_lock(&cli->cl_cache->ccc_lru_lock);
3823 list_del_init(&cli->cl_lru_osc);
3824 spin_unlock(&cli->cl_cache->ccc_lru_lock);
3825 cli->cl_lru_left = NULL;
3826 cl_cache_decref(cli->cl_cache);
3827 cli->cl_cache = NULL;
3830 /* free memory of osc quota cache */
3831 osc_quota_cleanup(obd);
3833 rc = client_obd_cleanup(obd);
3838 EXPORT_SYMBOL(osc_cleanup_common);
3840 static const struct obd_ops osc_obd_ops = {
3841 .o_owner = THIS_MODULE,
3842 .o_setup = osc_setup,
3843 .o_precleanup = osc_precleanup,
3844 .o_cleanup = osc_cleanup_common,
3845 .o_add_conn = client_import_add_conn,
3846 .o_del_conn = client_import_del_conn,
3847 .o_connect = client_connect_import,
3848 .o_reconnect = osc_reconnect,
3849 .o_disconnect = osc_disconnect,
3850 .o_statfs = osc_statfs,
3851 .o_statfs_async = osc_statfs_async,
3852 .o_create = osc_create,
3853 .o_destroy = osc_destroy,
3854 .o_getattr = osc_getattr,
3855 .o_setattr = osc_setattr,
3856 .o_iocontrol = osc_iocontrol,
3857 .o_set_info_async = osc_set_info_async,
3858 .o_import_event = osc_import_event,
3859 .o_quotactl = osc_quotactl,
3862 LIST_HEAD(osc_shrink_list);
3863 DEFINE_SPINLOCK(osc_shrink_lock);
3865 #ifdef HAVE_SHRINKER_COUNT
3866 static struct shrinker osc_cache_shrinker = {
3867 .count_objects = osc_cache_shrink_count,
3868 .scan_objects = osc_cache_shrink_scan,
3869 .seeks = DEFAULT_SEEKS,
3872 static int osc_cache_shrink(struct shrinker *shrinker,
3873 struct shrink_control *sc)
3875 (void)osc_cache_shrink_scan(shrinker, sc);
3877 return osc_cache_shrink_count(shrinker, sc);
3880 static struct shrinker osc_cache_shrinker = {
3881 .shrink = osc_cache_shrink,
3882 .seeks = DEFAULT_SEEKS,
3886 static int __init osc_init(void)
3888 unsigned int reqpool_size;
3889 unsigned int reqsize;
3893 /* print an address of _any_ initialized kernel symbol from this
3894 * module, to allow debugging with gdb that doesn't support data
3895 * symbols from modules.*/
3896 CDEBUG(D_INFO, "Lustre OSC module (%p).\n", &osc_caches);
3898 rc = lu_kmem_init(osc_caches);
3902 rc = register_shrinker(&osc_cache_shrinker);
3906 /* This is obviously too much memory, only prevent overflow here */
3907 if (osc_reqpool_mem_max >= 1 << 12 || osc_reqpool_mem_max == 0)
3908 GOTO(out_shrinker, rc = -EINVAL);
3910 reqpool_size = osc_reqpool_mem_max << 20;
3913 while (reqsize < OST_IO_MAXREQSIZE)
3914 reqsize = reqsize << 1;
3917 * We don't enlarge the request count in OSC pool according to
3918 * cl_max_rpcs_in_flight. The allocation from the pool will only be
3919 * tried after normal allocation failed. So a small OSC pool won't
3920 * cause much performance degression in most of cases.
3922 osc_reqpool_maxreqcount = reqpool_size / reqsize;
3924 atomic_set(&osc_pool_req_count, 0);
3925 osc_rq_pool = ptlrpc_init_rq_pool(0, OST_IO_MAXREQSIZE,
3926 ptlrpc_add_rqs_to_pool);
3928 if (osc_rq_pool == NULL)
3929 GOTO(out_shrinker, rc = -ENOMEM);
3931 rc = osc_start_grant_work();
3933 GOTO(out_req_pool, rc);
3935 rc = class_register_type(&osc_obd_ops, NULL, true,
3936 LUSTRE_OSC_NAME, &osc_device_type);
3938 GOTO(out_stop_grant, rc);
3943 osc_stop_grant_work();
3945 ptlrpc_free_rq_pool(osc_rq_pool);
3947 unregister_shrinker(&osc_cache_shrinker);
3949 lu_kmem_fini(osc_caches);
3954 static void __exit osc_exit(void)
3956 class_unregister_type(LUSTRE_OSC_NAME);
3957 ptlrpc_free_rq_pool(osc_rq_pool);
3958 osc_stop_grant_work();
3959 unregister_shrinker(&osc_cache_shrinker);
3960 lu_kmem_fini(osc_caches);
3963 MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>");
3964 MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
3965 MODULE_VERSION(LUSTRE_VERSION_STRING);
3966 MODULE_LICENSE("GPL");
3968 module_init(osc_init);
3969 module_exit(osc_exit);