1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
2 * vim:expandtab:shiftwidth=8:tabstop=8:
6 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License version 2 only,
10 * as published by the Free Software Foundation.
12 * This program is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * General Public License version 2 for more details (a copy is included
16 * in the LICENSE file that accompanied this code).
18 * You should have received a copy of the GNU General Public License
19 * version 2 along with this program; If not, see
20 * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
22 * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
23 * CA 95054 USA or visit www.sun.com if you need additional information or
29 * Copyright 2008 Sun Microsystems, Inc. All rights reserved
30 * Use is subject to license terms.
33 * This file is part of Lustre, http://www.lustre.org/
34 * Lustre is a trademark of Sun Microsystems, Inc.
38 # define EXPORT_SYMTAB
40 #define DEBUG_SUBSYSTEM S_OSC
42 #include <libcfs/libcfs.h>
45 # include <liblustre.h>
48 #include <lustre_dlm.h>
49 #include <lustre_net.h>
50 #include <lustre/lustre_user.h>
51 #include <obd_cksum.h>
59 #include <lustre_ha.h>
60 #include <lprocfs_status.h>
61 #include <lustre_log.h>
62 #include <lustre_debug.h>
63 #include <lustre_param.h>
64 #include "osc_internal.h"
66 static quota_interface_t *quota_interface = NULL;
67 extern quota_interface_t osc_quota_interface;
69 static void osc_release_ppga(struct brw_page **ppga, obd_count count);
70 static int brw_interpret(const struct lu_env *env,
71 struct ptlrpc_request *req, void *data, int rc);
72 int osc_cleanup(struct obd_device *obd);
74 /* Pack OSC object metadata for disk storage (LE byte order). */
75 static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
76 struct lov_stripe_md *lsm)
81 lmm_size = sizeof(**lmmp);
86 OBD_FREE(*lmmp, lmm_size);
92 OBD_ALLOC(*lmmp, lmm_size);
98 LASSERT(lsm->lsm_object_id);
99 LASSERT_MDS_GROUP(lsm->lsm_object_gr);
100 (*lmmp)->lmm_object_id = cpu_to_le64(lsm->lsm_object_id);
101 (*lmmp)->lmm_object_gr = cpu_to_le64(lsm->lsm_object_gr);
107 /* Unpack OSC object metadata from disk storage (LE byte order). */
108 static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
109 struct lov_mds_md *lmm, int lmm_bytes)
115 if (lmm_bytes < sizeof (*lmm)) {
116 CERROR("lov_mds_md too small: %d, need %d\n",
117 lmm_bytes, (int)sizeof(*lmm));
120 /* XXX LOV_MAGIC etc check? */
122 if (lmm->lmm_object_id == 0) {
123 CERROR("lov_mds_md: zero lmm_object_id\n");
128 lsm_size = lov_stripe_md_size(1);
132 if (*lsmp != NULL && lmm == NULL) {
133 OBD_FREE((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
134 OBD_FREE(*lsmp, lsm_size);
140 OBD_ALLOC(*lsmp, lsm_size);
143 OBD_ALLOC((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
144 if ((*lsmp)->lsm_oinfo[0] == NULL) {
145 OBD_FREE(*lsmp, lsm_size);
148 loi_init((*lsmp)->lsm_oinfo[0]);
152 /* XXX zero *lsmp? */
153 (*lsmp)->lsm_object_id = le64_to_cpu (lmm->lmm_object_id);
154 (*lsmp)->lsm_object_gr = le64_to_cpu (lmm->lmm_object_gr);
155 LASSERT((*lsmp)->lsm_object_id);
156 LASSERT_MDS_GROUP((*lsmp)->lsm_object_gr);
159 (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
164 static inline void osc_pack_capa(struct ptlrpc_request *req,
165 struct ost_body *body, void *capa)
167 struct obd_capa *oc = (struct obd_capa *)capa;
168 struct lustre_capa *c;
173 c = req_capsule_client_get(&req->rq_pill, &RMF_CAPA1);
176 body->oa.o_valid |= OBD_MD_FLOSSCAPA;
177 DEBUG_CAPA(D_SEC, c, "pack");
180 static inline void osc_pack_req_body(struct ptlrpc_request *req,
181 struct obd_info *oinfo)
183 struct ost_body *body;
185 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
188 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
189 osc_pack_capa(req, body, oinfo->oi_capa);
192 static inline void osc_set_capa_size(struct ptlrpc_request *req,
193 const struct req_msg_field *field,
197 req_capsule_set_size(&req->rq_pill, field, RCL_CLIENT, 0);
199 /* it is already calculated as sizeof struct obd_capa */
203 static int osc_getattr_interpret(const struct lu_env *env,
204 struct ptlrpc_request *req,
205 struct osc_async_args *aa, int rc)
207 struct ost_body *body;
213 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
214 lustre_swab_ost_body);
216 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
217 lustre_get_wire_obdo(aa->aa_oi->oi_oa, &body->oa);
219 /* This should really be sent by the OST */
220 aa->aa_oi->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
221 aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
223 CDEBUG(D_INFO, "can't unpack ost_body\n");
225 aa->aa_oi->oi_oa->o_valid = 0;
228 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
232 static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
233 struct ptlrpc_request_set *set)
235 struct ptlrpc_request *req;
236 struct osc_async_args *aa;
240 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
244 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
245 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
247 ptlrpc_request_free(req);
251 osc_pack_req_body(req, oinfo);
253 ptlrpc_request_set_replen(req);
254 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_getattr_interpret;
256 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
257 aa = ptlrpc_req_async_args(req);
260 ptlrpc_set_add_req(set, req);
264 static int osc_getattr(struct obd_export *exp, struct obd_info *oinfo)
266 struct ptlrpc_request *req;
267 struct ost_body *body;
271 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
275 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
276 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
278 ptlrpc_request_free(req);
282 osc_pack_req_body(req, oinfo);
284 ptlrpc_request_set_replen(req);
286 rc = ptlrpc_queue_wait(req);
290 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
292 GOTO(out, rc = -EPROTO);
294 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
295 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
297 /* This should really be sent by the OST */
298 oinfo->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
299 oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
303 ptlrpc_req_finished(req);
307 static int osc_setattr(struct obd_export *exp, struct obd_info *oinfo,
308 struct obd_trans_info *oti)
310 struct ptlrpc_request *req;
311 struct ost_body *body;
315 LASSERTF(!(oinfo->oi_oa->o_valid & OBD_MD_FLGROUP) ||
316 CHECK_MDS_GROUP(oinfo->oi_oa->o_gr),
317 "oinfo->oi_oa->o_valid="LPU64" oinfo->oi_oa->o_gr="LPU64"\n",
318 oinfo->oi_oa->o_valid, oinfo->oi_oa->o_gr);
320 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
324 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
325 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
327 ptlrpc_request_free(req);
331 osc_pack_req_body(req, oinfo);
333 ptlrpc_request_set_replen(req);
335 rc = ptlrpc_queue_wait(req);
339 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
341 GOTO(out, rc = -EPROTO);
343 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
347 ptlrpc_req_finished(req);
351 static int osc_setattr_interpret(const struct lu_env *env,
352 struct ptlrpc_request *req,
353 struct osc_async_args *aa, int rc)
355 struct ost_body *body;
361 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
363 GOTO(out, rc = -EPROTO);
365 lustre_get_wire_obdo(aa->aa_oi->oi_oa, &body->oa);
367 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
371 static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
372 struct obd_trans_info *oti,
373 struct ptlrpc_request_set *rqset)
375 struct ptlrpc_request *req;
376 struct osc_async_args *aa;
380 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
384 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
385 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
387 ptlrpc_request_free(req);
391 osc_pack_req_body(req, oinfo);
393 ptlrpc_request_set_replen(req);
395 if (oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE) {
397 oinfo->oi_oa->o_lcookie = *oti->oti_logcookies;
400 /* do mds to ost setattr asynchronously */
402 /* Do not wait for response. */
403 ptlrpcd_add_req(req, PSCOPE_OTHER);
405 req->rq_interpret_reply =
406 (ptlrpc_interpterer_t)osc_setattr_interpret;
408 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
409 aa = ptlrpc_req_async_args(req);
412 ptlrpc_set_add_req(rqset, req);
418 int osc_real_create(struct obd_export *exp, struct obdo *oa,
419 struct lov_stripe_md **ea, struct obd_trans_info *oti)
421 struct ptlrpc_request *req;
422 struct ost_body *body;
423 struct lov_stripe_md *lsm;
432 rc = obd_alloc_memmd(exp, &lsm);
437 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
439 GOTO(out, rc = -ENOMEM);
441 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_CREATE);
443 ptlrpc_request_free(req);
447 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
449 lustre_set_wire_obdo(&body->oa, oa);
451 ptlrpc_request_set_replen(req);
453 if ((oa->o_valid & OBD_MD_FLFLAGS) &&
454 oa->o_flags == OBD_FL_DELORPHAN) {
456 "delorphan from OST integration");
457 /* Don't resend the delorphan req */
458 req->rq_no_resend = req->rq_no_delay = 1;
461 rc = ptlrpc_queue_wait(req);
465 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
467 GOTO(out_req, rc = -EPROTO);
469 lustre_get_wire_obdo(oa, &body->oa);
471 /* This should really be sent by the OST */
472 oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
473 oa->o_valid |= OBD_MD_FLBLKSZ;
475 /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
476 * have valid lsm_oinfo data structs, so don't go touching that.
477 * This needs to be fixed in a big way.
479 lsm->lsm_object_id = oa->o_id;
480 lsm->lsm_object_gr = oa->o_gr;
484 oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);
486 if (oa->o_valid & OBD_MD_FLCOOKIE) {
487 if (!oti->oti_logcookies)
488 oti_alloc_cookies(oti, 1);
489 *oti->oti_logcookies = oa->o_lcookie;
493 CDEBUG(D_HA, "transno: "LPD64"\n",
494 lustre_msg_get_transno(req->rq_repmsg));
496 ptlrpc_req_finished(req);
499 obd_free_memmd(exp, &lsm);
503 static int osc_punch_interpret(const struct lu_env *env,
504 struct ptlrpc_request *req,
505 struct osc_punch_args *aa, int rc)
507 struct ost_body *body;
513 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
515 GOTO(out, rc = -EPROTO);
517 lustre_get_wire_obdo(aa->pa_oa, &body->oa);
519 rc = aa->pa_upcall(aa->pa_cookie, rc);
523 int osc_punch_base(struct obd_export *exp, struct obdo *oa,
524 struct obd_capa *capa,
525 obd_enqueue_update_f upcall, void *cookie,
526 struct ptlrpc_request_set *rqset)
528 struct ptlrpc_request *req;
529 struct osc_punch_args *aa;
530 struct ost_body *body;
534 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_PUNCH);
538 osc_set_capa_size(req, &RMF_CAPA1, capa);
539 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_PUNCH);
541 ptlrpc_request_free(req);
544 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
545 ptlrpc_at_set_req_timeout(req);
547 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
549 lustre_set_wire_obdo(&body->oa, oa);
550 osc_pack_capa(req, body, capa);
552 ptlrpc_request_set_replen(req);
555 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_punch_interpret;
556 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
557 aa = ptlrpc_req_async_args(req);
559 aa->pa_upcall = upcall;
560 aa->pa_cookie = cookie;
561 if (rqset == PTLRPCD_SET)
562 ptlrpcd_add_req(req, PSCOPE_OTHER);
564 ptlrpc_set_add_req(rqset, req);
569 static int osc_punch(struct obd_export *exp, struct obd_info *oinfo,
570 struct obd_trans_info *oti,
571 struct ptlrpc_request_set *rqset)
573 oinfo->oi_oa->o_size = oinfo->oi_policy.l_extent.start;
574 oinfo->oi_oa->o_blocks = oinfo->oi_policy.l_extent.end;
575 oinfo->oi_oa->o_valid |= OBD_MD_FLSIZE | OBD_MD_FLBLOCKS;
576 return osc_punch_base(exp, oinfo->oi_oa, oinfo->oi_capa,
577 oinfo->oi_cb_up, oinfo, rqset);
580 static int osc_sync(struct obd_export *exp, struct obdo *oa,
581 struct lov_stripe_md *md, obd_size start, obd_size end,
584 struct ptlrpc_request *req;
585 struct ost_body *body;
590 CDEBUG(D_INFO, "oa NULL\n");
594 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SYNC);
598 osc_set_capa_size(req, &RMF_CAPA1, capa);
599 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SYNC);
601 ptlrpc_request_free(req);
605 /* overload the size and blocks fields in the oa with start/end */
606 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
608 lustre_set_wire_obdo(&body->oa, oa);
609 body->oa.o_size = start;
610 body->oa.o_blocks = end;
611 body->oa.o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);
612 osc_pack_capa(req, body, capa);
614 ptlrpc_request_set_replen(req);
616 rc = ptlrpc_queue_wait(req);
620 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
622 GOTO(out, rc = -EPROTO);
624 lustre_get_wire_obdo(oa, &body->oa);
628 ptlrpc_req_finished(req);
632 /* Find and cancel locally locks matched by @mode in the resource found by
633 * @objid. Found locks are added into @cancel list. Returns the amount of
634 * locks added to @cancels list. */
635 static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
636 struct list_head *cancels, ldlm_mode_t mode,
639 struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
640 struct ldlm_res_id res_id;
641 struct ldlm_resource *res;
645 osc_build_res_name(oa->o_id, oa->o_gr, &res_id);
646 res = ldlm_resource_get(ns, NULL, &res_id, 0, 0);
650 LDLM_RESOURCE_ADDREF(res);
651 count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
652 lock_flags, 0, NULL);
653 LDLM_RESOURCE_DELREF(res);
654 ldlm_resource_putref(res);
658 static int osc_destroy_interpret(const struct lu_env *env,
659 struct ptlrpc_request *req, void *data,
662 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
664 atomic_dec(&cli->cl_destroy_in_flight);
665 cfs_waitq_signal(&cli->cl_destroy_waitq);
669 static int osc_can_send_destroy(struct client_obd *cli)
671 if (atomic_inc_return(&cli->cl_destroy_in_flight) <=
672 cli->cl_max_rpcs_in_flight) {
673 /* The destroy request can be sent */
676 if (atomic_dec_return(&cli->cl_destroy_in_flight) <
677 cli->cl_max_rpcs_in_flight) {
679 * The counter has been modified between the two atomic
682 cfs_waitq_signal(&cli->cl_destroy_waitq);
687 /* Destroy requests can be async always on the client, and we don't even really
688 * care about the return code since the client cannot do anything at all about
690 * When the MDS is unlinking a filename, it saves the file objects into a
691 * recovery llog, and these object records are cancelled when the OST reports
692 * they were destroyed and sync'd to disk (i.e. transaction committed).
693 * If the client dies, or the OST is down when the object should be destroyed,
694 * the records are not cancelled, and when the OST reconnects to the MDS next,
695 * it will retrieve the llog unlink logs and then sends the log cancellation
696 * cookies to the MDS after committing destroy transactions. */
697 static int osc_destroy(struct obd_export *exp, struct obdo *oa,
698 struct lov_stripe_md *ea, struct obd_trans_info *oti,
699 struct obd_export *md_export, void *capa)
701 struct client_obd *cli = &exp->exp_obd->u.cli;
702 struct ptlrpc_request *req;
703 struct ost_body *body;
704 CFS_LIST_HEAD(cancels);
709 CDEBUG(D_INFO, "oa NULL\n");
713 count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
714 LDLM_FL_DISCARD_DATA);
716 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_DESTROY);
718 ldlm_lock_list_put(&cancels, l_bl_ast, count);
722 osc_set_capa_size(req, &RMF_CAPA1, (struct obd_capa *)capa);
723 rc = ldlm_prep_elc_req(exp, req, LUSTRE_OST_VERSION, OST_DESTROY,
726 ptlrpc_request_free(req);
730 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
731 ptlrpc_at_set_req_timeout(req);
733 if (oti != NULL && oa->o_valid & OBD_MD_FLCOOKIE)
734 oa->o_lcookie = *oti->oti_logcookies;
735 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
737 lustre_set_wire_obdo(&body->oa, oa);
739 osc_pack_capa(req, body, (struct obd_capa *)capa);
740 ptlrpc_request_set_replen(req);
742 /* don't throttle destroy RPCs for the MDT */
743 if (!(cli->cl_import->imp_connect_flags_orig & OBD_CONNECT_MDS)) {
744 req->rq_interpret_reply = osc_destroy_interpret;
745 if (!osc_can_send_destroy(cli)) {
746 struct l_wait_info lwi = { 0 };
749 * Wait until the number of on-going destroy RPCs drops
750 * under max_rpc_in_flight
752 l_wait_event_exclusive(cli->cl_destroy_waitq,
753 osc_can_send_destroy(cli), &lwi);
757 /* Do not wait for response */
758 ptlrpcd_add_req(req, PSCOPE_OTHER);
762 static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
765 obd_flag bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;
767 LASSERT(!(oa->o_valid & bits));
770 client_obd_list_lock(&cli->cl_loi_list_lock);
771 oa->o_dirty = cli->cl_dirty;
772 if (cli->cl_dirty - cli->cl_dirty_transit > cli->cl_dirty_max) {
773 CERROR("dirty %lu - %lu > dirty_max %lu\n",
774 cli->cl_dirty, cli->cl_dirty_transit, cli->cl_dirty_max);
776 } else if (atomic_read(&obd_dirty_pages) -
777 atomic_read(&obd_dirty_transit_pages) > obd_max_dirty_pages){
778 CERROR("dirty %d - %d > system dirty_max %d\n",
779 atomic_read(&obd_dirty_pages),
780 atomic_read(&obd_dirty_transit_pages),
781 obd_max_dirty_pages);
783 } else if (cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff) {
784 CERROR("dirty %lu - dirty_max %lu too big???\n",
785 cli->cl_dirty, cli->cl_dirty_max);
788 long max_in_flight = (cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT)*
789 (cli->cl_max_rpcs_in_flight + 1);
790 oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
792 oa->o_grant = cli->cl_avail_grant;
793 oa->o_dropped = cli->cl_lost_grant;
794 cli->cl_lost_grant = 0;
795 client_obd_list_unlock(&cli->cl_loi_list_lock);
796 CDEBUG(D_CACHE,"dirty: "LPU64" undirty: %u dropped %u grant: "LPU64"\n",
797 oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
801 static void osc_update_next_shrink(struct client_obd *cli)
803 cli->cl_next_shrink_grant =
804 cfs_time_shift(cli->cl_grant_shrink_interval);
805 CDEBUG(D_CACHE, "next time %ld to shrink grant \n",
806 cli->cl_next_shrink_grant);
809 /* caller must hold loi_list_lock */
810 static void osc_consume_write_grant(struct client_obd *cli,
811 struct brw_page *pga)
813 LASSERT_SPIN_LOCKED(&cli->cl_loi_list_lock);
814 LASSERT(!(pga->flag & OBD_BRW_FROM_GRANT));
815 atomic_inc(&obd_dirty_pages);
816 cli->cl_dirty += CFS_PAGE_SIZE;
817 cli->cl_avail_grant -= CFS_PAGE_SIZE;
818 pga->flag |= OBD_BRW_FROM_GRANT;
819 CDEBUG(D_CACHE, "using %lu grant credits for brw %p page %p\n",
820 CFS_PAGE_SIZE, pga, pga->pg);
821 LASSERT(cli->cl_avail_grant >= 0);
822 osc_update_next_shrink(cli);
825 /* the companion to osc_consume_write_grant, called when a brw has completed.
826 * must be called with the loi lock held. */
827 static void osc_release_write_grant(struct client_obd *cli,
828 struct brw_page *pga, int sent)
830 int blocksize = cli->cl_import->imp_obd->obd_osfs.os_bsize ? : 4096;
833 LASSERT_SPIN_LOCKED(&cli->cl_loi_list_lock);
834 if (!(pga->flag & OBD_BRW_FROM_GRANT)) {
839 pga->flag &= ~OBD_BRW_FROM_GRANT;
840 atomic_dec(&obd_dirty_pages);
841 cli->cl_dirty -= CFS_PAGE_SIZE;
842 if (pga->flag & OBD_BRW_NOCACHE) {
843 pga->flag &= ~OBD_BRW_NOCACHE;
844 atomic_dec(&obd_dirty_transit_pages);
845 cli->cl_dirty_transit -= CFS_PAGE_SIZE;
848 cli->cl_lost_grant += CFS_PAGE_SIZE;
849 CDEBUG(D_CACHE, "lost grant: %lu avail grant: %lu dirty: %lu\n",
850 cli->cl_lost_grant, cli->cl_avail_grant, cli->cl_dirty);
851 } else if (CFS_PAGE_SIZE != blocksize && pga->count != CFS_PAGE_SIZE) {
852 /* For short writes we shouldn't count parts of pages that
853 * span a whole block on the OST side, or our accounting goes
854 * wrong. Should match the code in filter_grant_check. */
855 int offset = pga->off & ~CFS_PAGE_MASK;
856 int count = pga->count + (offset & (blocksize - 1));
857 int end = (offset + pga->count) & (blocksize - 1);
859 count += blocksize - end;
861 cli->cl_lost_grant += CFS_PAGE_SIZE - count;
862 CDEBUG(D_CACHE, "lost %lu grant: %lu avail: %lu dirty: %lu\n",
863 CFS_PAGE_SIZE - count, cli->cl_lost_grant,
864 cli->cl_avail_grant, cli->cl_dirty);
870 static unsigned long rpcs_in_flight(struct client_obd *cli)
872 return cli->cl_r_in_flight + cli->cl_w_in_flight;
875 /* caller must hold loi_list_lock */
876 void osc_wake_cache_waiters(struct client_obd *cli)
878 struct list_head *l, *tmp;
879 struct osc_cache_waiter *ocw;
882 list_for_each_safe(l, tmp, &cli->cl_cache_waiters) {
883 /* if we can't dirty more, we must wait until some is written */
884 if ((cli->cl_dirty + CFS_PAGE_SIZE > cli->cl_dirty_max) ||
885 (atomic_read(&obd_dirty_pages) + 1 > obd_max_dirty_pages)) {
886 CDEBUG(D_CACHE, "no dirty room: dirty: %ld "
887 "osc max %ld, sys max %d\n", cli->cl_dirty,
888 cli->cl_dirty_max, obd_max_dirty_pages);
892 /* if still dirty cache but no grant wait for pending RPCs that
893 * may yet return us some grant before doing sync writes */
894 if (cli->cl_w_in_flight && cli->cl_avail_grant < CFS_PAGE_SIZE) {
895 CDEBUG(D_CACHE, "%u BRW writes in flight, no grant\n",
896 cli->cl_w_in_flight);
900 ocw = list_entry(l, struct osc_cache_waiter, ocw_entry);
901 list_del_init(&ocw->ocw_entry);
902 if (cli->cl_avail_grant < CFS_PAGE_SIZE) {
903 /* no more RPCs in flight to return grant, do sync IO */
904 ocw->ocw_rc = -EDQUOT;
905 CDEBUG(D_INODE, "wake oap %p for sync\n", ocw->ocw_oap);
907 osc_consume_write_grant(cli,
908 &ocw->ocw_oap->oap_brw_page);
911 cfs_waitq_signal(&ocw->ocw_waitq);
917 static void __osc_update_grant(struct client_obd *cli, obd_size grant)
919 client_obd_list_lock(&cli->cl_loi_list_lock);
920 cli->cl_avail_grant += grant;
921 client_obd_list_unlock(&cli->cl_loi_list_lock);
924 static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
926 if (body->oa.o_valid & OBD_MD_FLGRANT) {
927 CDEBUG(D_CACHE, "got "LPU64" extra grant\n", body->oa.o_grant);
928 __osc_update_grant(cli, body->oa.o_grant);
932 static int osc_set_info_async(struct obd_export *exp, obd_count keylen,
933 void *key, obd_count vallen, void *val,
934 struct ptlrpc_request_set *set);
936 static int osc_shrink_grant_interpret(const struct lu_env *env,
937 struct ptlrpc_request *req,
940 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
941 struct obdo *oa = ((struct osc_grant_args *)aa)->aa_oa;
942 struct ost_body *body;
945 __osc_update_grant(cli, oa->o_grant);
949 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
951 osc_update_grant(cli, body);
957 static void osc_shrink_grant_local(struct client_obd *cli, struct obdo *oa)
959 client_obd_list_lock(&cli->cl_loi_list_lock);
960 oa->o_grant = cli->cl_avail_grant / 4;
961 cli->cl_avail_grant -= oa->o_grant;
962 client_obd_list_unlock(&cli->cl_loi_list_lock);
963 oa->o_flags |= OBD_FL_SHRINK_GRANT;
964 osc_update_next_shrink(cli);
967 /* Shrink the current grant, either from some large amount to enough for a
968 * full set of in-flight RPCs, or if we have already shrunk to that limit
969 * then to enough for a single RPC. This avoids keeping more grant than
970 * needed, and avoids shrinking the grant piecemeal. */
971 static int osc_shrink_grant(struct client_obd *cli)
973 long target = (cli->cl_max_rpcs_in_flight + 1) *
974 cli->cl_max_pages_per_rpc;
976 client_obd_list_lock(&cli->cl_loi_list_lock);
977 if (cli->cl_avail_grant <= target)
978 target = cli->cl_max_pages_per_rpc;
979 client_obd_list_unlock(&cli->cl_loi_list_lock);
981 return osc_shrink_grant_to_target(cli, target);
984 int osc_shrink_grant_to_target(struct client_obd *cli, long target)
987 struct ost_body *body;
990 client_obd_list_lock(&cli->cl_loi_list_lock);
991 /* Don't shrink if we are already above or below the desired limit
992 * We don't want to shrink below a single RPC, as that will negatively
993 * impact block allocation and long-term performance. */
994 if (target < cli->cl_max_pages_per_rpc)
995 target = cli->cl_max_pages_per_rpc;
997 if (target >= cli->cl_avail_grant) {
998 client_obd_list_unlock(&cli->cl_loi_list_lock);
1001 client_obd_list_unlock(&cli->cl_loi_list_lock);
1003 OBD_ALLOC_PTR(body);
1007 osc_announce_cached(cli, &body->oa, 0);
1009 client_obd_list_lock(&cli->cl_loi_list_lock);
1010 body->oa.o_grant = cli->cl_avail_grant - target;
1011 cli->cl_avail_grant = target;
1012 client_obd_list_unlock(&cli->cl_loi_list_lock);
1013 body->oa.o_flags |= OBD_FL_SHRINK_GRANT;
1014 osc_update_next_shrink(cli);
1016 rc = osc_set_info_async(cli->cl_import->imp_obd->obd_self_export,
1017 sizeof(KEY_GRANT_SHRINK), KEY_GRANT_SHRINK,
1018 sizeof(*body), body, NULL);
1020 __osc_update_grant(cli, body->oa.o_grant);
1025 #define GRANT_SHRINK_LIMIT PTLRPC_MAX_BRW_SIZE
1026 static int osc_should_shrink_grant(struct client_obd *client)
1028 cfs_time_t time = cfs_time_current();
1029 cfs_time_t next_shrink = client->cl_next_shrink_grant;
1030 if (cfs_time_aftereq(time, next_shrink - 5 * CFS_TICK)) {
1031 if (client->cl_import->imp_state == LUSTRE_IMP_FULL &&
1032 client->cl_avail_grant > GRANT_SHRINK_LIMIT)
1035 osc_update_next_shrink(client);
1040 static int osc_grant_shrink_grant_cb(struct timeout_item *item, void *data)
1042 struct client_obd *client;
1044 list_for_each_entry(client, &item->ti_obd_list, cl_grant_shrink_list) {
1045 if (osc_should_shrink_grant(client))
1046 osc_shrink_grant(client);
1051 static int osc_add_shrink_grant(struct client_obd *client)
1055 rc = ptlrpc_add_timeout_client(client->cl_grant_shrink_interval,
1057 osc_grant_shrink_grant_cb, NULL,
1058 &client->cl_grant_shrink_list);
1060 CERROR("add grant client %s error %d\n",
1061 client->cl_import->imp_obd->obd_name, rc);
1064 CDEBUG(D_CACHE, "add grant client %s \n",
1065 client->cl_import->imp_obd->obd_name);
1066 osc_update_next_shrink(client);
1070 static int osc_del_shrink_grant(struct client_obd *client)
1072 return ptlrpc_del_timeout_client(&client->cl_grant_shrink_list,
1076 static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
1078 client_obd_list_lock(&cli->cl_loi_list_lock);
1079 cli->cl_avail_grant = ocd->ocd_grant;
1080 client_obd_list_unlock(&cli->cl_loi_list_lock);
1082 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT_SHRINK &&
1083 list_empty(&cli->cl_grant_shrink_list))
1084 osc_add_shrink_grant(cli);
1086 CDEBUG(D_CACHE, "setting cl_avail_grant: %ld cl_lost_grant: %ld \n",
1087 cli->cl_avail_grant, cli->cl_lost_grant);
1088 LASSERT(cli->cl_avail_grant >= 0);
1091 /* We assume that the reason this OSC got a short read is because it read
1092 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
1093 * via the LOV, and it _knows_ it's reading inside the file, it's just that
1094 * this stripe never got written at or beyond this stripe offset yet. */
1095 static void handle_short_read(int nob_read, obd_count page_count,
1096 struct brw_page **pga)
1101 /* skip bytes read OK */
1102 while (nob_read > 0) {
1103 LASSERT (page_count > 0);
1105 if (pga[i]->count > nob_read) {
1106 /* EOF inside this page */
1107 ptr = cfs_kmap(pga[i]->pg) +
1108 (pga[i]->off & ~CFS_PAGE_MASK);
1109 memset(ptr + nob_read, 0, pga[i]->count - nob_read);
1110 cfs_kunmap(pga[i]->pg);
1116 nob_read -= pga[i]->count;
1121 /* zero remaining pages */
1122 while (page_count-- > 0) {
1123 ptr = cfs_kmap(pga[i]->pg) + (pga[i]->off & ~CFS_PAGE_MASK);
1124 memset(ptr, 0, pga[i]->count);
1125 cfs_kunmap(pga[i]->pg);
1130 static int check_write_rcs(struct ptlrpc_request *req,
1131 int requested_nob, int niocount,
1132 obd_count page_count, struct brw_page **pga)
1136 /* return error if any niobuf was in error */
1137 remote_rcs = lustre_swab_repbuf(req, REQ_REC_OFF + 1,
1138 sizeof(*remote_rcs) * niocount, NULL);
1139 if (remote_rcs == NULL) {
1140 CDEBUG(D_INFO, "Missing/short RC vector on BRW_WRITE reply\n");
1143 if (lustre_msg_swabbed(req->rq_repmsg))
1144 for (i = 0; i < niocount; i++)
1145 __swab32s(&remote_rcs[i]);
1147 for (i = 0; i < niocount; i++) {
1148 if (remote_rcs[i] < 0)
1149 return(remote_rcs[i]);
1151 if (remote_rcs[i] != 0) {
1152 CDEBUG(D_INFO, "rc[%d] invalid (%d) req %p\n",
1153 i, remote_rcs[i], req);
1158 if (req->rq_bulk->bd_nob_transferred != requested_nob) {
1159 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
1160 req->rq_bulk->bd_nob_transferred, requested_nob);
1167 static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
1169 if (p1->flag != p2->flag) {
1170 unsigned mask = ~(OBD_BRW_FROM_GRANT|
1171 OBD_BRW_NOCACHE|OBD_BRW_SYNC);
1173 /* warn if we try to combine flags that we don't know to be
1174 * safe to combine */
1175 if ((p1->flag & mask) != (p2->flag & mask))
1176 CERROR("is it ok to have flags 0x%x and 0x%x in the "
1177 "same brw?\n", p1->flag, p2->flag);
1181 return (p1->off + p1->count == p2->off);
1184 static obd_count osc_checksum_bulk(int nob, obd_count pg_count,
1185 struct brw_page **pga, int opc,
1186 cksum_type_t cksum_type)
1191 LASSERT (pg_count > 0);
1192 cksum = init_checksum(cksum_type);
1193 while (nob > 0 && pg_count > 0) {
1194 unsigned char *ptr = cfs_kmap(pga[i]->pg);
1195 int off = pga[i]->off & ~CFS_PAGE_MASK;
1196 int count = pga[i]->count > nob ? nob : pga[i]->count;
1198 /* corrupt the data before we compute the checksum, to
1199 * simulate an OST->client data error */
1200 if (i == 0 && opc == OST_READ &&
1201 OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE))
1202 memcpy(ptr + off, "bad1", min(4, nob));
1203 cksum = compute_checksum(cksum, ptr + off, count, cksum_type);
1204 cfs_kunmap(pga[i]->pg);
1205 LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d checksum %x\n",
1208 nob -= pga[i]->count;
1212 /* For sending we only compute the wrong checksum instead
1213 * of corrupting the data so it is still correct on a redo */
1214 if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1220 static int osc_brw_prep_request(int cmd, struct client_obd *cli,struct obdo *oa,
1221 struct lov_stripe_md *lsm, obd_count page_count,
1222 struct brw_page **pga,
1223 struct ptlrpc_request **reqp,
1224 struct obd_capa *ocapa, int reserve)
1226 struct ptlrpc_request *req;
1227 struct ptlrpc_bulk_desc *desc;
1228 struct ost_body *body;
1229 struct obd_ioobj *ioobj;
1230 struct niobuf_remote *niobuf;
1231 int niocount, i, requested_nob, opc, rc;
1232 struct osc_brw_async_args *aa;
1233 struct req_capsule *pill;
1234 struct brw_page *pg_prev;
1237 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ))
1238 RETURN(-ENOMEM); /* Recoverable */
1239 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ2))
1240 RETURN(-EINVAL); /* Fatal */
1242 if ((cmd & OBD_BRW_WRITE) != 0) {
1244 req = ptlrpc_request_alloc_pool(cli->cl_import,
1245 cli->cl_import->imp_rq_pool,
1249 req = ptlrpc_request_alloc(cli->cl_import, &RQF_OST_BRW);
1254 for (niocount = i = 1; i < page_count; i++) {
1255 if (!can_merge_pages(pga[i - 1], pga[i]))
1259 pill = &req->rq_pill;
1260 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_CLIENT,
1261 niocount * sizeof(*niobuf));
1262 osc_set_capa_size(req, &RMF_CAPA1, ocapa);
1264 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, opc);
1266 ptlrpc_request_free(req);
1269 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
1270 ptlrpc_at_set_req_timeout(req);
1272 if (opc == OST_WRITE)
1273 desc = ptlrpc_prep_bulk_imp(req, page_count,
1274 BULK_GET_SOURCE, OST_BULK_PORTAL);
1276 desc = ptlrpc_prep_bulk_imp(req, page_count,
1277 BULK_PUT_SINK, OST_BULK_PORTAL);
1280 GOTO(out, rc = -ENOMEM);
1281 /* NB request now owns desc and will free it when it gets freed */
1283 body = req_capsule_client_get(pill, &RMF_OST_BODY);
1284 ioobj = req_capsule_client_get(pill, &RMF_OBD_IOOBJ);
1285 niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1286 LASSERT(body && ioobj && niobuf);
1288 lustre_set_wire_obdo(&body->oa, oa);
1290 obdo_to_ioobj(oa, ioobj);
1291 ioobj->ioo_bufcnt = niocount;
1292 osc_pack_capa(req, body, ocapa);
1293 LASSERT (page_count > 0);
1295 for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1296 struct brw_page *pg = pga[i];
1298 LASSERT(pg->count > 0);
1299 LASSERTF((pg->off & ~CFS_PAGE_MASK) + pg->count <= CFS_PAGE_SIZE,
1300 "i: %d pg: %p off: "LPU64", count: %u\n", i, pg,
1301 pg->off, pg->count);
1303 LASSERTF(i == 0 || pg->off > pg_prev->off,
1304 "i %d p_c %u pg %p [pri %lu ind %lu] off "LPU64
1305 " prev_pg %p [pri %lu ind %lu] off "LPU64"\n",
1307 pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1308 pg_prev->pg, page_private(pg_prev->pg),
1309 pg_prev->pg->index, pg_prev->off);
1311 LASSERTF(i == 0 || pg->off > pg_prev->off,
1312 "i %d p_c %u\n", i, page_count);
1314 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1315 (pg->flag & OBD_BRW_SRVLOCK));
1317 ptlrpc_prep_bulk_page(desc, pg->pg, pg->off & ~CFS_PAGE_MASK,
1319 requested_nob += pg->count;
1321 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1323 niobuf->len += pg->count;
1325 niobuf->offset = pg->off;
1326 niobuf->len = pg->count;
1327 niobuf->flags = pg->flag;
1332 LASSERTF((void *)(niobuf - niocount) ==
1333 lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 2,
1334 niocount * sizeof(*niobuf)),
1335 "want %p - real %p\n", lustre_msg_buf(req->rq_reqmsg,
1336 REQ_REC_OFF + 2, niocount * sizeof(*niobuf)),
1337 (void *)(niobuf - niocount));
1339 osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1340 if (osc_should_shrink_grant(cli))
1341 osc_shrink_grant_local(cli, &body->oa);
1343 /* size[REQ_REC_OFF] still sizeof (*body) */
1344 if (opc == OST_WRITE) {
1345 if (unlikely(cli->cl_checksum) &&
1346 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1347 /* store cl_cksum_type in a local variable since
1348 * it can be changed via lprocfs */
1349 cksum_type_t cksum_type = cli->cl_cksum_type;
1351 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1352 oa->o_flags &= OBD_FL_LOCAL_MASK;
1353 body->oa.o_flags = 0;
1355 body->oa.o_flags |= cksum_type_pack(cksum_type);
1356 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1357 body->oa.o_cksum = osc_checksum_bulk(requested_nob,
1361 CDEBUG(D_PAGE, "checksum at write origin: %x\n",
1363 /* save this in 'oa', too, for later checking */
1364 oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1365 oa->o_flags |= cksum_type_pack(cksum_type);
1367 /* clear out the checksum flag, in case this is a
1368 * resend but cl_checksum is no longer set. b=11238 */
1369 oa->o_valid &= ~OBD_MD_FLCKSUM;
1371 oa->o_cksum = body->oa.o_cksum;
1372 /* 1 RC per niobuf */
1373 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_SERVER,
1374 sizeof(__u32) * niocount);
1376 if (unlikely(cli->cl_checksum) &&
1377 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1378 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1379 body->oa.o_flags = 0;
1380 body->oa.o_flags |= cksum_type_pack(cli->cl_cksum_type);
1381 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1383 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_SERVER, 0);
1384 /* 1 RC for the whole I/O */
1386 ptlrpc_request_set_replen(req);
1388 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
1389 aa = ptlrpc_req_async_args(req);
1391 aa->aa_requested_nob = requested_nob;
1392 aa->aa_nio_count = niocount;
1393 aa->aa_page_count = page_count;
1397 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
1398 if (ocapa && reserve)
1399 aa->aa_ocapa = capa_get(ocapa);
1405 ptlrpc_req_finished(req);
1409 static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
1410 __u32 client_cksum, __u32 server_cksum, int nob,
1411 obd_count page_count, struct brw_page **pga,
1412 cksum_type_t client_cksum_type)
1416 cksum_type_t cksum_type;
1418 if (server_cksum == client_cksum) {
1419 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1423 if (oa->o_valid & OBD_MD_FLFLAGS)
1424 cksum_type = cksum_type_unpack(oa->o_flags);
1426 cksum_type = OBD_CKSUM_CRC32;
1428 new_cksum = osc_checksum_bulk(nob, page_count, pga, OST_WRITE,
1431 if (cksum_type != client_cksum_type)
1432 msg = "the server did not use the checksum type specified in "
1433 "the original request - likely a protocol problem";
1434 else if (new_cksum == server_cksum)
1435 msg = "changed on the client after we checksummed it - "
1436 "likely false positive due to mmap IO (bug 11742)";
1437 else if (new_cksum == client_cksum)
1438 msg = "changed in transit before arrival at OST";
1440 msg = "changed in transit AND doesn't match the original - "
1441 "likely false positive due to mmap IO (bug 11742)";
1443 LCONSOLE_ERROR_MSG(0x132, "BAD WRITE CHECKSUM: %s: from %s inum "
1444 LPU64"/"LPU64" object "LPU64"/"LPU64" extent "
1445 "["LPU64"-"LPU64"]\n",
1446 msg, libcfs_nid2str(peer->nid),
1447 oa->o_valid & OBD_MD_FLFID ? oa->o_fid : (__u64)0,
1448 oa->o_valid & OBD_MD_FLFID ? oa->o_generation :
1451 oa->o_valid & OBD_MD_FLGROUP ? oa->o_gr : (__u64)0,
1453 pga[page_count-1]->off + pga[page_count-1]->count - 1);
1454 CERROR("original client csum %x (type %x), server csum %x (type %x), "
1455 "client csum now %x\n", client_cksum, client_cksum_type,
1456 server_cksum, cksum_type, new_cksum);
1460 /* Note rc enters this function as number of bytes transferred */
1461 static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
1463 struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
1464 const lnet_process_id_t *peer =
1465 &req->rq_import->imp_connection->c_peer;
1466 struct client_obd *cli = aa->aa_cli;
1467 struct ost_body *body;
1468 __u32 client_cksum = 0;
1471 if (rc < 0 && rc != -EDQUOT)
1474 LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
1475 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
1476 lustre_swab_ost_body);
1478 CDEBUG(D_INFO, "Can't unpack body\n");
1482 /* set/clear over quota flag for a uid/gid */
1483 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
1484 body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA)) {
1485 unsigned int qid[MAXQUOTAS] = { body->oa.o_uid, body->oa.o_gid };
1487 lquota_setdq(quota_interface, cli, qid, body->oa.o_valid,
1494 if (aa->aa_oa->o_valid & OBD_MD_FLCKSUM)
1495 client_cksum = aa->aa_oa->o_cksum; /* save for later */
1497 osc_update_grant(cli, body);
1499 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1501 CERROR("Unexpected +ve rc %d\n", rc);
1504 LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);
1506 if (sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
1509 if ((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) && client_cksum &&
1510 check_write_checksum(&body->oa, peer, client_cksum,
1511 body->oa.o_cksum, aa->aa_requested_nob,
1512 aa->aa_page_count, aa->aa_ppga,
1513 cksum_type_unpack(aa->aa_oa->o_flags)))
1516 rc = check_write_rcs(req, aa->aa_requested_nob,aa->aa_nio_count,
1517 aa->aa_page_count, aa->aa_ppga);
1521 /* The rest of this function executes only for OST_READs */
1523 /* if unwrap_bulk failed, return -EAGAIN to retry */
1524 rc = sptlrpc_cli_unwrap_bulk_read(req, req->rq_bulk, rc);
1526 GOTO(out, rc = -EAGAIN);
1528 if (rc > aa->aa_requested_nob) {
1529 CERROR("Unexpected rc %d (%d requested)\n", rc,
1530 aa->aa_requested_nob);
1534 if (rc != req->rq_bulk->bd_nob_transferred) {
1535 CERROR ("Unexpected rc %d (%d transferred)\n",
1536 rc, req->rq_bulk->bd_nob_transferred);
1540 if (rc < aa->aa_requested_nob)
1541 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
1543 if (body->oa.o_valid & OBD_MD_FLCKSUM) {
1544 static int cksum_counter;
1545 __u32 server_cksum = body->oa.o_cksum;
1548 cksum_type_t cksum_type;
1550 if (body->oa.o_valid & OBD_MD_FLFLAGS)
1551 cksum_type = cksum_type_unpack(body->oa.o_flags);
1553 cksum_type = OBD_CKSUM_CRC32;
1554 client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
1555 aa->aa_ppga, OST_READ,
1558 if (peer->nid == req->rq_bulk->bd_sender) {
1562 router = libcfs_nid2str(req->rq_bulk->bd_sender);
1565 if (server_cksum == ~0 && rc > 0) {
1566 CERROR("Protocol error: server %s set the 'checksum' "
1567 "bit, but didn't send a checksum. Not fatal, "
1568 "but please notify on http://bugzilla.lustre.org/\n",
1569 libcfs_nid2str(peer->nid));
1570 } else if (server_cksum != client_cksum) {
1571 LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from "
1572 "%s%s%s inum "LPU64"/"LPU64" object "
1573 LPU64"/"LPU64" extent "
1574 "["LPU64"-"LPU64"]\n",
1575 req->rq_import->imp_obd->obd_name,
1576 libcfs_nid2str(peer->nid),
1578 body->oa.o_valid & OBD_MD_FLFID ?
1579 body->oa.o_fid : (__u64)0,
1580 body->oa.o_valid & OBD_MD_FLFID ?
1581 body->oa.o_generation :(__u64)0,
1583 body->oa.o_valid & OBD_MD_FLGROUP ?
1584 body->oa.o_gr : (__u64)0,
1585 aa->aa_ppga[0]->off,
1586 aa->aa_ppga[aa->aa_page_count-1]->off +
1587 aa->aa_ppga[aa->aa_page_count-1]->count -
1589 CERROR("client %x, server %x, cksum_type %x\n",
1590 client_cksum, server_cksum, cksum_type);
1592 aa->aa_oa->o_cksum = client_cksum;
1596 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1599 } else if (unlikely(client_cksum)) {
1600 static int cksum_missed;
1603 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
1604 CERROR("Checksum %u requested from %s but not sent\n",
1605 cksum_missed, libcfs_nid2str(peer->nid));
1611 lustre_get_wire_obdo(aa->aa_oa, &body->oa);
1616 static int osc_brw_internal(int cmd, struct obd_export *exp, struct obdo *oa,
1617 struct lov_stripe_md *lsm,
1618 obd_count page_count, struct brw_page **pga,
1619 struct obd_capa *ocapa)
1621 struct ptlrpc_request *req;
1625 struct l_wait_info lwi;
1629 cfs_waitq_init(&waitq);
1632 rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
1633 page_count, pga, &req, ocapa, 0);
1637 rc = ptlrpc_queue_wait(req);
1639 if (rc == -ETIMEDOUT && req->rq_resend) {
1640 DEBUG_REQ(D_HA, req, "BULK TIMEOUT");
1641 ptlrpc_req_finished(req);
1645 rc = osc_brw_fini_request(req, rc);
1647 ptlrpc_req_finished(req);
1648 if (osc_recoverable_error(rc)) {
1650 if (!osc_should_resend(resends, &exp->exp_obd->u.cli)) {
1651 CERROR("too many resend retries, returning error\n");
1655 lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(resends), NULL, NULL, NULL);
1656 l_wait_event(waitq, 0, &lwi);
1664 int osc_brw_redo_request(struct ptlrpc_request *request,
1665 struct osc_brw_async_args *aa)
1667 struct ptlrpc_request *new_req;
1668 struct ptlrpc_request_set *set = request->rq_set;
1669 struct osc_brw_async_args *new_aa;
1670 struct osc_async_page *oap;
1674 if (!osc_should_resend(aa->aa_resends, aa->aa_cli)) {
1675 CERROR("too many resend retries, returning error\n");
1679 DEBUG_REQ(D_ERROR, request, "redo for recoverable error");
1681 rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
1682 OST_WRITE ? OBD_BRW_WRITE :OBD_BRW_READ,
1683 aa->aa_cli, aa->aa_oa,
1684 NULL /* lsm unused by osc currently */,
1685 aa->aa_page_count, aa->aa_ppga,
1686 &new_req, aa->aa_ocapa, 0);
1690 client_obd_list_lock(&aa->aa_cli->cl_loi_list_lock);
1692 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
1693 if (oap->oap_request != NULL) {
1694 LASSERTF(request == oap->oap_request,
1695 "request %p != oap_request %p\n",
1696 request, oap->oap_request);
1697 if (oap->oap_interrupted) {
1698 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1699 ptlrpc_req_finished(new_req);
1704 /* New request takes over pga and oaps from old request.
1705 * Note that copying a list_head doesn't work, need to move it... */
1707 new_req->rq_interpret_reply = request->rq_interpret_reply;
1708 new_req->rq_async_args = request->rq_async_args;
1709 new_req->rq_sent = cfs_time_current_sec() + aa->aa_resends;
1711 new_aa = ptlrpc_req_async_args(new_req);
1713 CFS_INIT_LIST_HEAD(&new_aa->aa_oaps);
1714 list_splice(&aa->aa_oaps, &new_aa->aa_oaps);
1715 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
1717 list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
1718 if (oap->oap_request) {
1719 ptlrpc_req_finished(oap->oap_request);
1720 oap->oap_request = ptlrpc_request_addref(new_req);
1724 new_aa->aa_ocapa = aa->aa_ocapa;
1725 aa->aa_ocapa = NULL;
1727 /* use ptlrpc_set_add_req is safe because interpret functions work
1728 * in check_set context. only one way exist with access to request
1729 * from different thread got -EINTR - this way protected with
1730 * cl_loi_list_lock */
1731 ptlrpc_set_add_req(set, new_req);
1733 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1735 DEBUG_REQ(D_INFO, new_req, "new request");
1740 * ugh, we want disk allocation on the target to happen in offset order. we'll
1741 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
1742 * fine for our small page arrays and doesn't require allocation. its an
1743 * insertion sort that swaps elements that are strides apart, shrinking the
1744 * stride down until its '1' and the array is sorted.
1746 static void sort_brw_pages(struct brw_page **array, int num)
1749 struct brw_page *tmp;
1753 for (stride = 1; stride < num ; stride = (stride * 3) + 1)
1758 for (i = stride ; i < num ; i++) {
1761 while (j >= stride && array[j - stride]->off > tmp->off) {
1762 array[j] = array[j - stride];
1767 } while (stride > 1);
1770 static obd_count max_unfragmented_pages(struct brw_page **pg, obd_count pages)
1776 LASSERT (pages > 0);
1777 offset = pg[i]->off & ~CFS_PAGE_MASK;
1781 if (pages == 0) /* that's all */
1784 if (offset + pg[i]->count < CFS_PAGE_SIZE)
1785 return count; /* doesn't end on page boundary */
1788 offset = pg[i]->off & ~CFS_PAGE_MASK;
1789 if (offset != 0) /* doesn't start on page boundary */
1796 static struct brw_page **osc_build_ppga(struct brw_page *pga, obd_count count)
1798 struct brw_page **ppga;
1801 OBD_ALLOC(ppga, sizeof(*ppga) * count);
1805 for (i = 0; i < count; i++)
1810 static void osc_release_ppga(struct brw_page **ppga, obd_count count)
1812 LASSERT(ppga != NULL);
1813 OBD_FREE(ppga, sizeof(*ppga) * count);
1816 static int osc_brw(int cmd, struct obd_export *exp, struct obd_info *oinfo,
1817 obd_count page_count, struct brw_page *pga,
1818 struct obd_trans_info *oti)
1820 struct obdo *saved_oa = NULL;
1821 struct brw_page **ppga, **orig;
1822 struct obd_import *imp = class_exp2cliimp(exp);
1823 struct client_obd *cli;
1824 int rc, page_count_orig;
1827 LASSERT((imp != NULL) && (imp->imp_obd != NULL));
1828 cli = &imp->imp_obd->u.cli;
1830 if (cmd & OBD_BRW_CHECK) {
1831 /* The caller just wants to know if there's a chance that this
1832 * I/O can succeed */
1834 if (imp->imp_invalid)
1839 /* test_brw with a failed create can trip this, maybe others. */
1840 LASSERT(cli->cl_max_pages_per_rpc);
1844 orig = ppga = osc_build_ppga(pga, page_count);
1847 page_count_orig = page_count;
1849 sort_brw_pages(ppga, page_count);
1850 while (page_count) {
1851 obd_count pages_per_brw;
1853 if (page_count > cli->cl_max_pages_per_rpc)
1854 pages_per_brw = cli->cl_max_pages_per_rpc;
1856 pages_per_brw = page_count;
1858 pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);
1860 if (saved_oa != NULL) {
1861 /* restore previously saved oa */
1862 *oinfo->oi_oa = *saved_oa;
1863 } else if (page_count > pages_per_brw) {
1864 /* save a copy of oa (brw will clobber it) */
1865 OBDO_ALLOC(saved_oa);
1866 if (saved_oa == NULL)
1867 GOTO(out, rc = -ENOMEM);
1868 *saved_oa = *oinfo->oi_oa;
1871 rc = osc_brw_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
1872 pages_per_brw, ppga, oinfo->oi_capa);
1877 page_count -= pages_per_brw;
1878 ppga += pages_per_brw;
1882 osc_release_ppga(orig, page_count_orig);
1884 if (saved_oa != NULL)
1885 OBDO_FREE(saved_oa);
1890 /* The companion to osc_enter_cache(), called when @oap is no longer part of
1891 * the dirty accounting. Writeback completes or truncate happens before
1892 * writing starts. Must be called with the loi lock held. */
1893 static void osc_exit_cache(struct client_obd *cli, struct osc_async_page *oap,
1896 osc_release_write_grant(cli, &oap->oap_brw_page, sent);
1900 /* This maintains the lists of pending pages to read/write for a given object
1901 * (lop). This is used by osc_check_rpcs->osc_next_loi() and loi_list_maint()
1902 * to quickly find objects that are ready to send an RPC. */
1903 static int lop_makes_rpc(struct client_obd *cli, struct loi_oap_pages *lop,
1909 if (lop->lop_num_pending == 0)
1912 /* if we have an invalid import we want to drain the queued pages
1913 * by forcing them through rpcs that immediately fail and complete
1914 * the pages. recovery relies on this to empty the queued pages
1915 * before canceling the locks and evicting down the llite pages */
1916 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
1919 /* stream rpcs in queue order as long as as there is an urgent page
1920 * queued. this is our cheap solution for good batching in the case
1921 * where writepage marks some random page in the middle of the file
1922 * as urgent because of, say, memory pressure */
1923 if (!list_empty(&lop->lop_urgent)) {
1924 CDEBUG(D_CACHE, "urgent request forcing RPC\n");
1927 /* fire off rpcs when we have 'optimal' rpcs as tuned for the wire. */
1928 optimal = cli->cl_max_pages_per_rpc;
1929 if (cmd & OBD_BRW_WRITE) {
1930 /* trigger a write rpc stream as long as there are dirtiers
1931 * waiting for space. as they're waiting, they're not going to
1932 * create more pages to coallesce with what's waiting.. */
1933 if (!list_empty(&cli->cl_cache_waiters)) {
1934 CDEBUG(D_CACHE, "cache waiters forcing RPC\n");
1937 /* +16 to avoid triggering rpcs that would want to include pages
1938 * that are being queued but which can't be made ready until
1939 * the queuer finishes with the page. this is a wart for
1940 * llite::commit_write() */
1943 if (lop->lop_num_pending >= optimal)
1949 static int lop_makes_hprpc(struct loi_oap_pages *lop)
1951 struct osc_async_page *oap;
1954 if (list_empty(&lop->lop_urgent))
1957 oap = list_entry(lop->lop_urgent.next,
1958 struct osc_async_page, oap_urgent_item);
1960 if (oap->oap_async_flags & ASYNC_HP) {
1961 CDEBUG(D_CACHE, "hp request forcing RPC\n");
1968 static void on_list(struct list_head *item, struct list_head *list,
1971 if (list_empty(item) && should_be_on)
1972 list_add_tail(item, list);
1973 else if (!list_empty(item) && !should_be_on)
1974 list_del_init(item);
1977 /* maintain the loi's cli list membership invariants so that osc_send_oap_rpc
1978 * can find pages to build into rpcs quickly */
1979 void loi_list_maint(struct client_obd *cli, struct lov_oinfo *loi)
1981 if (lop_makes_hprpc(&loi->loi_write_lop) ||
1982 lop_makes_hprpc(&loi->loi_read_lop)) {
1984 on_list(&loi->loi_ready_item, &cli->cl_loi_ready_list, 0);
1985 on_list(&loi->loi_hp_ready_item, &cli->cl_loi_hp_ready_list, 1);
1987 on_list(&loi->loi_hp_ready_item, &cli->cl_loi_hp_ready_list, 0);
1988 on_list(&loi->loi_ready_item, &cli->cl_loi_ready_list,
1989 lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE)||
1990 lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ));
1993 on_list(&loi->loi_write_item, &cli->cl_loi_write_list,
1994 loi->loi_write_lop.lop_num_pending);
1996 on_list(&loi->loi_read_item, &cli->cl_loi_read_list,
1997 loi->loi_read_lop.lop_num_pending);
2000 static void lop_update_pending(struct client_obd *cli,
2001 struct loi_oap_pages *lop, int cmd, int delta)
2003 lop->lop_num_pending += delta;
2004 if (cmd & OBD_BRW_WRITE)
2005 cli->cl_pending_w_pages += delta;
2007 cli->cl_pending_r_pages += delta;
2011 * this is called when a sync waiter receives an interruption. Its job is to
2012 * get the caller woken as soon as possible. If its page hasn't been put in an
2013 * rpc yet it can dequeue immediately. Otherwise it has to mark the rpc as
2014 * desiring interruption which will forcefully complete the rpc once the rpc
2017 int osc_oap_interrupted(const struct lu_env *env, struct osc_async_page *oap)
2019 struct loi_oap_pages *lop;
2020 struct lov_oinfo *loi;
2024 LASSERT(!oap->oap_interrupted);
2025 oap->oap_interrupted = 1;
2027 /* ok, it's been put in an rpc. only one oap gets a request reference */
2028 if (oap->oap_request != NULL) {
2029 ptlrpc_mark_interrupted(oap->oap_request);
2030 ptlrpcd_wake(oap->oap_request);
2031 ptlrpc_req_finished(oap->oap_request);
2032 oap->oap_request = NULL;
2036 * page completion may be called only if ->cpo_prep() method was
2037 * executed by osc_io_submit(), that also adds page the to pending list
2039 if (!list_empty(&oap->oap_pending_item)) {
2040 list_del_init(&oap->oap_pending_item);
2041 list_del_init(&oap->oap_urgent_item);
2044 lop = (oap->oap_cmd & OBD_BRW_WRITE) ?
2045 &loi->loi_write_lop : &loi->loi_read_lop;
2046 lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, -1);
2047 loi_list_maint(oap->oap_cli, oap->oap_loi);
2048 rc = oap->oap_caller_ops->ap_completion(env,
2049 oap->oap_caller_data,
2050 oap->oap_cmd, NULL, -EINTR);
2056 /* this is trying to propogate async writeback errors back up to the
2057 * application. As an async write fails we record the error code for later if
2058 * the app does an fsync. As long as errors persist we force future rpcs to be
2059 * sync so that the app can get a sync error and break the cycle of queueing
2060 * pages for which writeback will fail. */
2061 static void osc_process_ar(struct osc_async_rc *ar, __u64 xid,
2068 ar->ar_force_sync = 1;
2069 ar->ar_min_xid = ptlrpc_sample_next_xid();
2074 if (ar->ar_force_sync && (xid >= ar->ar_min_xid))
2075 ar->ar_force_sync = 0;
2078 void osc_oap_to_pending(struct osc_async_page *oap)
2080 struct loi_oap_pages *lop;
2082 if (oap->oap_cmd & OBD_BRW_WRITE)
2083 lop = &oap->oap_loi->loi_write_lop;
2085 lop = &oap->oap_loi->loi_read_lop;
2087 if (oap->oap_async_flags & ASYNC_HP)
2088 list_add(&oap->oap_urgent_item, &lop->lop_urgent);
2089 else if (oap->oap_async_flags & ASYNC_URGENT)
2090 list_add_tail(&oap->oap_urgent_item, &lop->lop_urgent);
2091 list_add_tail(&oap->oap_pending_item, &lop->lop_pending);
2092 lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, 1);
2095 /* this must be called holding the loi list lock to give coverage to exit_cache,
2096 * async_flag maintenance, and oap_request */
2097 static void osc_ap_completion(const struct lu_env *env,
2098 struct client_obd *cli, struct obdo *oa,
2099 struct osc_async_page *oap, int sent, int rc)
2104 if (oap->oap_request != NULL) {
2105 xid = ptlrpc_req_xid(oap->oap_request);
2106 ptlrpc_req_finished(oap->oap_request);
2107 oap->oap_request = NULL;
2110 oap->oap_async_flags = 0;
2111 oap->oap_interrupted = 0;
2113 if (oap->oap_cmd & OBD_BRW_WRITE) {
2114 osc_process_ar(&cli->cl_ar, xid, rc);
2115 osc_process_ar(&oap->oap_loi->loi_ar, xid, rc);
2118 if (rc == 0 && oa != NULL) {
2119 if (oa->o_valid & OBD_MD_FLBLOCKS)
2120 oap->oap_loi->loi_lvb.lvb_blocks = oa->o_blocks;
2121 if (oa->o_valid & OBD_MD_FLMTIME)
2122 oap->oap_loi->loi_lvb.lvb_mtime = oa->o_mtime;
2123 if (oa->o_valid & OBD_MD_FLATIME)
2124 oap->oap_loi->loi_lvb.lvb_atime = oa->o_atime;
2125 if (oa->o_valid & OBD_MD_FLCTIME)
2126 oap->oap_loi->loi_lvb.lvb_ctime = oa->o_ctime;
2129 rc = oap->oap_caller_ops->ap_completion(env, oap->oap_caller_data,
2130 oap->oap_cmd, oa, rc);
2132 /* ll_ap_completion (from llite) drops PG_locked. so, a new
2133 * I/O on the page could start, but OSC calls it under lock
2134 * and thus we can add oap back to pending safely */
2136 /* upper layer wants to leave the page on pending queue */
2137 osc_oap_to_pending(oap);
2139 osc_exit_cache(cli, oap, sent);
2143 static int brw_interpret(const struct lu_env *env,
2144 struct ptlrpc_request *req, void *data, int rc)
2146 struct osc_brw_async_args *aa = data;
2147 struct client_obd *cli;
2151 rc = osc_brw_fini_request(req, rc);
2152 CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
2153 if (osc_recoverable_error(rc)) {
2154 rc = osc_brw_redo_request(req, aa);
2160 capa_put(aa->aa_ocapa);
2161 aa->aa_ocapa = NULL;
2166 client_obd_list_lock(&cli->cl_loi_list_lock);
2168 /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
2169 * is called so we know whether to go to sync BRWs or wait for more
2170 * RPCs to complete */
2171 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
2172 cli->cl_w_in_flight--;
2174 cli->cl_r_in_flight--;
2176 async = list_empty(&aa->aa_oaps);
2177 if (!async) { /* from osc_send_oap_rpc() */
2178 struct osc_async_page *oap, *tmp;
2179 /* the caller may re-use the oap after the completion call so
2180 * we need to clean it up a little */
2181 list_for_each_entry_safe(oap, tmp, &aa->aa_oaps, oap_rpc_item) {
2182 list_del_init(&oap->oap_rpc_item);
2183 osc_ap_completion(env, cli, aa->aa_oa, oap, 1, rc);
2185 OBDO_FREE(aa->aa_oa);
2186 } else { /* from async_internal() */
2188 for (i = 0; i < aa->aa_page_count; i++)
2189 osc_release_write_grant(aa->aa_cli, aa->aa_ppga[i], 1);
2191 if (aa->aa_oa->o_flags & OBD_FL_TEMPORARY)
2192 OBDO_FREE(aa->aa_oa);
2194 osc_wake_cache_waiters(cli);
2195 osc_check_rpcs(env, cli);
2196 client_obd_list_unlock(&cli->cl_loi_list_lock);
2198 cl_req_completion(env, aa->aa_clerq, rc);
2199 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
2203 static struct ptlrpc_request *osc_build_req(const struct lu_env *env,
2204 struct client_obd *cli,
2205 struct list_head *rpc_list,
2206 int page_count, int cmd)
2208 struct ptlrpc_request *req;
2209 struct brw_page **pga = NULL;
2210 struct osc_brw_async_args *aa;
2211 struct obdo *oa = NULL;
2212 const struct obd_async_page_ops *ops = NULL;
2213 void *caller_data = NULL;
2214 struct osc_async_page *oap;
2215 struct osc_async_page *tmp;
2216 struct ost_body *body;
2217 struct cl_req *clerq = NULL;
2218 enum cl_req_type crt = (cmd & OBD_BRW_WRITE) ? CRT_WRITE : CRT_READ;
2219 struct ldlm_lock *lock = NULL;
2220 struct cl_req_attr crattr;
2224 LASSERT(!list_empty(rpc_list));
2226 memset(&crattr, 0, sizeof crattr);
2227 OBD_ALLOC(pga, sizeof(*pga) * page_count);
2229 GOTO(out, req = ERR_PTR(-ENOMEM));
2233 GOTO(out, req = ERR_PTR(-ENOMEM));
2236 list_for_each_entry(oap, rpc_list, oap_rpc_item) {
2237 struct cl_page *page = osc_oap2cl_page(oap);
2239 ops = oap->oap_caller_ops;
2240 caller_data = oap->oap_caller_data;
2242 clerq = cl_req_alloc(env, page, crt,
2243 1 /* only 1-object rpcs for
2246 GOTO(out, req = (void *)clerq);
2247 lock = oap->oap_ldlm_lock;
2249 pga[i] = &oap->oap_brw_page;
2250 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
2251 CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
2252 pga[i]->pg, cfs_page_index(oap->oap_page), oap, pga[i]->flag);
2254 cl_req_page_add(env, clerq, page);
2257 /* always get the data for the obdo for the rpc */
2258 LASSERT(ops != NULL);
2260 crattr.cra_capa = NULL;
2261 cl_req_attr_set(env, clerq, &crattr, ~0ULL);
2263 oa->o_handle = lock->l_remote_handle;
2264 oa->o_valid |= OBD_MD_FLHANDLE;
2267 rc = cl_req_prep(env, clerq);
2269 CERROR("cl_req_prep failed: %d\n", rc);
2270 GOTO(out, req = ERR_PTR(rc));
2273 sort_brw_pages(pga, page_count);
2274 rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count,
2275 pga, &req, crattr.cra_capa, 1);
2277 CERROR("prep_req failed: %d\n", rc);
2278 GOTO(out, req = ERR_PTR(rc));
2281 /* Need to update the timestamps after the request is built in case
2282 * we race with setattr (locally or in queue at OST). If OST gets
2283 * later setattr before earlier BRW (as determined by the request xid),
2284 * the OST will not use BRW timestamps. Sadly, there is no obvious
2285 * way to do this in a single call. bug 10150 */
2286 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
2287 cl_req_attr_set(env, clerq, &crattr,
2288 OBD_MD_FLMTIME|OBD_MD_FLCTIME|OBD_MD_FLATIME);
2290 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2291 aa = ptlrpc_req_async_args(req);
2292 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
2293 list_splice(rpc_list, &aa->aa_oaps);
2294 CFS_INIT_LIST_HEAD(rpc_list);
2295 aa->aa_clerq = clerq;
2297 capa_put(crattr.cra_capa);
2302 OBD_FREE(pga, sizeof(*pga) * page_count);
2303 /* this should happen rarely and is pretty bad, it makes the
2304 * pending list not follow the dirty order */
2305 client_obd_list_lock(&cli->cl_loi_list_lock);
2306 list_for_each_entry_safe(oap, tmp, rpc_list, oap_rpc_item) {
2307 list_del_init(&oap->oap_rpc_item);
2309 /* queued sync pages can be torn down while the pages
2310 * were between the pending list and the rpc */
2311 if (oap->oap_interrupted) {
2312 CDEBUG(D_INODE, "oap %p interrupted\n", oap);
2313 osc_ap_completion(env, cli, NULL, oap, 0,
2317 osc_ap_completion(env, cli, NULL, oap, 0, PTR_ERR(req));
2319 if (clerq && !IS_ERR(clerq))
2320 cl_req_completion(env, clerq, PTR_ERR(req));
2326 * prepare pages for ASYNC io and put pages in send queue.
2330 * \param cmd - OBD_BRW_* macroses
2331 * \param lop - pending pages
2333 * \return zero if pages successfully add to send queue.
2334 * \return not zere if error occurring.
2337 osc_send_oap_rpc(const struct lu_env *env, struct client_obd *cli,
2338 struct lov_oinfo *loi,
2339 int cmd, struct loi_oap_pages *lop)
2341 struct ptlrpc_request *req;
2342 obd_count page_count = 0;
2343 struct osc_async_page *oap = NULL, *tmp;
2344 struct osc_brw_async_args *aa;
2345 const struct obd_async_page_ops *ops;
2346 CFS_LIST_HEAD(rpc_list);
2347 unsigned int ending_offset;
2348 unsigned starting_offset = 0;
2350 struct cl_object *clob = NULL;
2353 /* If there are HP OAPs we need to handle at least 1 of them,
2354 * move it the beginning of the pending list for that. */
2355 if (!list_empty(&lop->lop_urgent)) {
2356 oap = list_entry(lop->lop_urgent.next,
2357 struct osc_async_page, oap_urgent_item);
2358 if (oap->oap_async_flags & ASYNC_HP)
2359 list_move(&oap->oap_pending_item, &lop->lop_pending);
2362 /* first we find the pages we're allowed to work with */
2363 list_for_each_entry_safe(oap, tmp, &lop->lop_pending,
2365 ops = oap->oap_caller_ops;
2367 LASSERTF(oap->oap_magic == OAP_MAGIC, "Bad oap magic: oap %p, "
2368 "magic 0x%x\n", oap, oap->oap_magic);
2371 /* pin object in memory, so that completion call-backs
2372 * can be safely called under client_obd_list lock. */
2373 clob = osc_oap2cl_page(oap)->cp_obj;
2374 cl_object_get(clob);
2377 if (page_count != 0 &&
2378 srvlock != !!(oap->oap_brw_flags & OBD_BRW_SRVLOCK)) {
2379 CDEBUG(D_PAGE, "SRVLOCK flag mismatch,"
2380 " oap %p, page %p, srvlock %u\n",
2381 oap, oap->oap_brw_page.pg, (unsigned)!srvlock);
2384 /* in llite being 'ready' equates to the page being locked
2385 * until completion unlocks it. commit_write submits a page
2386 * as not ready because its unlock will happen unconditionally
2387 * as the call returns. if we race with commit_write giving
2388 * us that page we dont' want to create a hole in the page
2389 * stream, so we stop and leave the rpc to be fired by
2390 * another dirtier or kupdated interval (the not ready page
2391 * will still be on the dirty list). we could call in
2392 * at the end of ll_file_write to process the queue again. */
2393 if (!(oap->oap_async_flags & ASYNC_READY)) {
2394 int rc = ops->ap_make_ready(env, oap->oap_caller_data,
2397 CDEBUG(D_INODE, "oap %p page %p returned %d "
2398 "instead of ready\n", oap,
2402 /* llite is telling us that the page is still
2403 * in commit_write and that we should try
2404 * and put it in an rpc again later. we
2405 * break out of the loop so we don't create
2406 * a hole in the sequence of pages in the rpc
2411 /* the io isn't needed.. tell the checks
2412 * below to complete the rpc with EINTR */
2413 oap->oap_async_flags |= ASYNC_COUNT_STABLE;
2414 oap->oap_count = -EINTR;
2417 oap->oap_async_flags |= ASYNC_READY;
2420 LASSERTF(0, "oap %p page %p returned %d "
2421 "from make_ready\n", oap,
2429 * Page submitted for IO has to be locked. Either by
2430 * ->ap_make_ready() or by higher layers.
2432 #if defined(__KERNEL__) && defined(__linux__)
2434 struct cl_page *page;
2436 page = osc_oap2cl_page(oap);
2438 if (page->cp_type == CPT_CACHEABLE &&
2439 !(PageLocked(oap->oap_page) &&
2440 (CheckWriteback(oap->oap_page, cmd)))) {
2441 CDEBUG(D_PAGE, "page %p lost wb %lx/%x\n",
2443 (long)oap->oap_page->flags,
2444 oap->oap_async_flags);
2449 /* If there is a gap at the start of this page, it can't merge
2450 * with any previous page, so we'll hand the network a
2451 * "fragmented" page array that it can't transfer in 1 RDMA */
2452 if (page_count != 0 && oap->oap_page_off != 0)
2455 /* take the page out of our book-keeping */
2456 list_del_init(&oap->oap_pending_item);
2457 lop_update_pending(cli, lop, cmd, -1);
2458 list_del_init(&oap->oap_urgent_item);
2460 if (page_count == 0)
2461 starting_offset = (oap->oap_obj_off+oap->oap_page_off) &
2462 (PTLRPC_MAX_BRW_SIZE - 1);
2464 /* ask the caller for the size of the io as the rpc leaves. */
2465 if (!(oap->oap_async_flags & ASYNC_COUNT_STABLE)) {
2467 ops->ap_refresh_count(env, oap->oap_caller_data,
2469 LASSERT(oap->oap_page_off + oap->oap_count <= CFS_PAGE_SIZE);
2471 if (oap->oap_count <= 0) {
2472 CDEBUG(D_CACHE, "oap %p count %d, completing\n", oap,
2474 osc_ap_completion(env, cli, NULL,
2475 oap, 0, oap->oap_count);
2479 /* now put the page back in our accounting */
2480 list_add_tail(&oap->oap_rpc_item, &rpc_list);
2481 if (page_count == 0)
2482 srvlock = !!(oap->oap_brw_flags & OBD_BRW_SRVLOCK);
2483 if (++page_count >= cli->cl_max_pages_per_rpc)
2486 /* End on a PTLRPC_MAX_BRW_SIZE boundary. We want full-sized
2487 * RPCs aligned on PTLRPC_MAX_BRW_SIZE boundaries to help reads
2488 * have the same alignment as the initial writes that allocated
2489 * extents on the server. */
2490 ending_offset = (oap->oap_obj_off + oap->oap_page_off +
2491 oap->oap_count) & (PTLRPC_MAX_BRW_SIZE - 1);
2492 if (ending_offset == 0)
2495 /* If there is a gap at the end of this page, it can't merge
2496 * with any subsequent pages, so we'll hand the network a
2497 * "fragmented" page array that it can't transfer in 1 RDMA */
2498 if (oap->oap_page_off + oap->oap_count < CFS_PAGE_SIZE)
2502 osc_wake_cache_waiters(cli);
2504 loi_list_maint(cli, loi);
2506 client_obd_list_unlock(&cli->cl_loi_list_lock);
2509 cl_object_put(env, clob);
2511 if (page_count == 0) {
2512 client_obd_list_lock(&cli->cl_loi_list_lock);
2516 req = osc_build_req(env, cli, &rpc_list, page_count, cmd);
2518 LASSERT(list_empty(&rpc_list));
2519 loi_list_maint(cli, loi);
2520 RETURN(PTR_ERR(req));
2523 aa = ptlrpc_req_async_args(req);
2525 if (cmd == OBD_BRW_READ) {
2526 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
2527 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
2528 lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
2529 (starting_offset >> CFS_PAGE_SHIFT) + 1);
2531 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
2532 lprocfs_oh_tally(&cli->cl_write_rpc_hist,
2533 cli->cl_w_in_flight);
2534 lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
2535 (starting_offset >> CFS_PAGE_SHIFT) + 1);
2537 ptlrpc_lprocfs_brw(req, aa->aa_requested_nob);
2539 client_obd_list_lock(&cli->cl_loi_list_lock);
2541 if (cmd == OBD_BRW_READ)
2542 cli->cl_r_in_flight++;
2544 cli->cl_w_in_flight++;
2546 /* queued sync pages can be torn down while the pages
2547 * were between the pending list and the rpc */
2549 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2550 /* only one oap gets a request reference */
2553 if (oap->oap_interrupted && !req->rq_intr) {
2554 CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
2556 ptlrpc_mark_interrupted(req);
2560 tmp->oap_request = ptlrpc_request_addref(req);
2562 DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
2563 page_count, aa, cli->cl_r_in_flight, cli->cl_w_in_flight);
2565 req->rq_interpret_reply = brw_interpret;
2566 ptlrpcd_add_req(req, PSCOPE_BRW);
2570 #define LOI_DEBUG(LOI, STR, args...) \
2571 CDEBUG(D_INODE, "loi ready %d wr %d:%d rd %d:%d " STR, \
2572 !list_empty(&(LOI)->loi_ready_item) || \
2573 !list_empty(&(LOI)->loi_hp_ready_item), \
2574 (LOI)->loi_write_lop.lop_num_pending, \
2575 !list_empty(&(LOI)->loi_write_lop.lop_urgent), \
2576 (LOI)->loi_read_lop.lop_num_pending, \
2577 !list_empty(&(LOI)->loi_read_lop.lop_urgent), \
2580 /* This is called by osc_check_rpcs() to find which objects have pages that
2581 * we could be sending. These lists are maintained by lop_makes_rpc(). */
2582 struct lov_oinfo *osc_next_loi(struct client_obd *cli)
2586 /* First return objects that have blocked locks so that they
2587 * will be flushed quickly and other clients can get the lock,
2588 * then objects which have pages ready to be stuffed into RPCs */
2589 if (!list_empty(&cli->cl_loi_hp_ready_list))
2590 RETURN(list_entry(cli->cl_loi_hp_ready_list.next,
2591 struct lov_oinfo, loi_hp_ready_item));
2592 if (!list_empty(&cli->cl_loi_ready_list))
2593 RETURN(list_entry(cli->cl_loi_ready_list.next,
2594 struct lov_oinfo, loi_ready_item));
2596 /* then if we have cache waiters, return all objects with queued
2597 * writes. This is especially important when many small files
2598 * have filled up the cache and not been fired into rpcs because
2599 * they don't pass the nr_pending/object threshhold */
2600 if (!list_empty(&cli->cl_cache_waiters) &&
2601 !list_empty(&cli->cl_loi_write_list))
2602 RETURN(list_entry(cli->cl_loi_write_list.next,
2603 struct lov_oinfo, loi_write_item));
2605 /* then return all queued objects when we have an invalid import
2606 * so that they get flushed */
2607 if (cli->cl_import == NULL || cli->cl_import->imp_invalid) {
2608 if (!list_empty(&cli->cl_loi_write_list))
2609 RETURN(list_entry(cli->cl_loi_write_list.next,
2610 struct lov_oinfo, loi_write_item));
2611 if (!list_empty(&cli->cl_loi_read_list))
2612 RETURN(list_entry(cli->cl_loi_read_list.next,
2613 struct lov_oinfo, loi_read_item));
2618 static int osc_max_rpc_in_flight(struct client_obd *cli, struct lov_oinfo *loi)
2620 struct osc_async_page *oap;
2623 if (!list_empty(&loi->loi_write_lop.lop_urgent)) {
2624 oap = list_entry(loi->loi_write_lop.lop_urgent.next,
2625 struct osc_async_page, oap_urgent_item);
2626 hprpc = !!(oap->oap_async_flags & ASYNC_HP);
2629 if (!hprpc && !list_empty(&loi->loi_read_lop.lop_urgent)) {
2630 oap = list_entry(loi->loi_write_lop.lop_urgent.next,
2631 struct osc_async_page, oap_urgent_item);
2632 hprpc = !!(oap->oap_async_flags & ASYNC_HP);
2635 return rpcs_in_flight(cli) >= cli->cl_max_rpcs_in_flight + hprpc;
2638 /* called with the loi list lock held */
2639 void osc_check_rpcs(const struct lu_env *env, struct client_obd *cli)
2641 struct lov_oinfo *loi;
2642 int rc = 0, race_counter = 0;
2645 while ((loi = osc_next_loi(cli)) != NULL) {
2646 LOI_DEBUG(loi, "%lu in flight\n", rpcs_in_flight(cli));
2648 if (osc_max_rpc_in_flight(cli, loi))
2651 /* attempt some read/write balancing by alternating between
2652 * reads and writes in an object. The makes_rpc checks here
2653 * would be redundant if we were getting read/write work items
2654 * instead of objects. we don't want send_oap_rpc to drain a
2655 * partial read pending queue when we're given this object to
2656 * do io on writes while there are cache waiters */
2657 if (lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE)) {
2658 rc = osc_send_oap_rpc(env, cli, loi, OBD_BRW_WRITE,
2659 &loi->loi_write_lop);
2667 if (lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ)) {
2668 rc = osc_send_oap_rpc(env, cli, loi, OBD_BRW_READ,
2669 &loi->loi_read_lop);
2678 /* attempt some inter-object balancing by issueing rpcs
2679 * for each object in turn */
2680 if (!list_empty(&loi->loi_hp_ready_item))
2681 list_del_init(&loi->loi_hp_ready_item);
2682 if (!list_empty(&loi->loi_ready_item))
2683 list_del_init(&loi->loi_ready_item);
2684 if (!list_empty(&loi->loi_write_item))
2685 list_del_init(&loi->loi_write_item);
2686 if (!list_empty(&loi->loi_read_item))
2687 list_del_init(&loi->loi_read_item);
2689 loi_list_maint(cli, loi);
2691 /* send_oap_rpc fails with 0 when make_ready tells it to
2692 * back off. llite's make_ready does this when it tries
2693 * to lock a page queued for write that is already locked.
2694 * we want to try sending rpcs from many objects, but we
2695 * don't want to spin failing with 0. */
2696 if (race_counter == 10)
2702 /* we're trying to queue a page in the osc so we're subject to the
2703 * 'cl_dirty_max' limit on the number of pages that can be queued in the osc.
2704 * If the osc's queued pages are already at that limit, then we want to sleep
2705 * until there is space in the osc's queue for us. We also may be waiting for
2706 * write credits from the OST if there are RPCs in flight that may return some
2707 * before we fall back to sync writes.
2709 * We need this know our allocation was granted in the presence of signals */
2710 static int ocw_granted(struct client_obd *cli, struct osc_cache_waiter *ocw)
2714 client_obd_list_lock(&cli->cl_loi_list_lock);
2715 rc = list_empty(&ocw->ocw_entry) || rpcs_in_flight(cli) == 0;
2716 client_obd_list_unlock(&cli->cl_loi_list_lock);
2721 * Non-blocking version of osc_enter_cache() that consumes grant only when it
2724 int osc_enter_cache_try(const struct lu_env *env,
2725 struct client_obd *cli, struct lov_oinfo *loi,
2726 struct osc_async_page *oap, int transient)
2730 has_grant = cli->cl_avail_grant >= CFS_PAGE_SIZE;
2732 osc_consume_write_grant(cli, &oap->oap_brw_page);
2734 cli->cl_dirty_transit += CFS_PAGE_SIZE;
2735 atomic_inc(&obd_dirty_transit_pages);
2736 oap->oap_brw_flags |= OBD_BRW_NOCACHE;
2742 /* Caller must hold loi_list_lock - we drop/regain it if we need to wait for
2743 * grant or cache space. */
2744 static int osc_enter_cache(const struct lu_env *env,
2745 struct client_obd *cli, struct lov_oinfo *loi,
2746 struct osc_async_page *oap)
2748 struct osc_cache_waiter ocw;
2749 struct l_wait_info lwi = { 0 };
2753 CDEBUG(D_CACHE, "dirty: %ld/%d dirty_max: %ld/%d dropped: %lu "
2754 "grant: %lu\n", cli->cl_dirty, atomic_read(&obd_dirty_pages),
2755 cli->cl_dirty_max, obd_max_dirty_pages,
2756 cli->cl_lost_grant, cli->cl_avail_grant);
2758 /* force the caller to try sync io. this can jump the list
2759 * of queued writes and create a discontiguous rpc stream */
2760 if (cli->cl_dirty_max < CFS_PAGE_SIZE || cli->cl_ar.ar_force_sync ||
2761 loi->loi_ar.ar_force_sync)
2764 /* Hopefully normal case - cache space and write credits available */
2765 if (cli->cl_dirty + CFS_PAGE_SIZE <= cli->cl_dirty_max &&
2766 atomic_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages &&
2767 osc_enter_cache_try(env, cli, loi, oap, 0))
2770 /* Make sure that there are write rpcs in flight to wait for. This
2771 * is a little silly as this object may not have any pending but
2772 * other objects sure might. */
2773 if (cli->cl_w_in_flight) {
2774 list_add_tail(&ocw.ocw_entry, &cli->cl_cache_waiters);
2775 cfs_waitq_init(&ocw.ocw_waitq);
2779 loi_list_maint(cli, loi);
2780 osc_check_rpcs(env, cli);
2781 client_obd_list_unlock(&cli->cl_loi_list_lock);
2783 CDEBUG(D_CACHE, "sleeping for cache space\n");
2784 l_wait_event(ocw.ocw_waitq, ocw_granted(cli, &ocw), &lwi);
2786 client_obd_list_lock(&cli->cl_loi_list_lock);
2787 if (!list_empty(&ocw.ocw_entry)) {
2788 list_del(&ocw.ocw_entry);
2798 int osc_prep_async_page(struct obd_export *exp, struct lov_stripe_md *lsm,
2799 struct lov_oinfo *loi, cfs_page_t *page,
2800 obd_off offset, const struct obd_async_page_ops *ops,
2801 void *data, void **res, int nocache,
2802 struct lustre_handle *lockh)
2804 struct osc_async_page *oap;
2809 return size_round(sizeof(*oap));
2812 oap->oap_magic = OAP_MAGIC;
2813 oap->oap_cli = &exp->exp_obd->u.cli;
2816 oap->oap_caller_ops = ops;
2817 oap->oap_caller_data = data;
2819 oap->oap_page = page;
2820 oap->oap_obj_off = offset;
2821 if (!client_is_remote(exp) &&
2822 cfs_capable(CFS_CAP_SYS_RESOURCE))
2823 oap->oap_brw_flags = OBD_BRW_NOQUOTA;
2825 LASSERT(!(offset & ~CFS_PAGE_MASK));
2827 CFS_INIT_LIST_HEAD(&oap->oap_pending_item);
2828 CFS_INIT_LIST_HEAD(&oap->oap_urgent_item);
2829 CFS_INIT_LIST_HEAD(&oap->oap_rpc_item);
2830 CFS_INIT_LIST_HEAD(&oap->oap_page_list);
2832 spin_lock_init(&oap->oap_lock);
2833 CDEBUG(D_CACHE, "oap %p page %p obj off "LPU64"\n", oap, page, offset);
2837 struct osc_async_page *oap_from_cookie(void *cookie)
2839 struct osc_async_page *oap = cookie;
2840 if (oap->oap_magic != OAP_MAGIC)
2841 return ERR_PTR(-EINVAL);
2845 int osc_queue_async_io(const struct lu_env *env,
2846 struct obd_export *exp, struct lov_stripe_md *lsm,
2847 struct lov_oinfo *loi, void *cookie,
2848 int cmd, obd_off off, int count,
2849 obd_flag brw_flags, enum async_flags async_flags)
2851 struct client_obd *cli = &exp->exp_obd->u.cli;
2852 struct osc_async_page *oap;
2856 oap = oap_from_cookie(cookie);
2858 RETURN(PTR_ERR(oap));
2860 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
2863 if (!list_empty(&oap->oap_pending_item) ||
2864 !list_empty(&oap->oap_urgent_item) ||
2865 !list_empty(&oap->oap_rpc_item))
2868 /* check if the file's owner/group is over quota */
2869 if ((cmd & OBD_BRW_WRITE) && !(cmd & OBD_BRW_NOQUOTA)) {
2870 struct cl_object *obj;
2871 struct cl_attr attr; /* XXX put attr into thread info */
2872 unsigned int qid[MAXQUOTAS];
2874 obj = cl_object_top(osc_oap2cl_page(oap)->cp_obj);
2876 cl_object_attr_lock(obj);
2877 rc = cl_object_attr_get(env, obj, &attr);
2878 cl_object_attr_unlock(obj);
2880 qid[USRQUOTA] = attr.cat_uid;
2881 qid[GRPQUOTA] = attr.cat_gid;
2883 lquota_chkdq(quota_interface, cli, qid) == NO_QUOTA)
2890 loi = lsm->lsm_oinfo[0];
2892 client_obd_list_lock(&cli->cl_loi_list_lock);
2894 LASSERT(off + count <= CFS_PAGE_SIZE);
2896 oap->oap_page_off = off;
2897 oap->oap_count = count;
2898 oap->oap_brw_flags = brw_flags;
2899 oap->oap_async_flags = async_flags;
2901 if (cmd & OBD_BRW_WRITE) {
2902 rc = osc_enter_cache(env, cli, loi, oap);
2904 client_obd_list_unlock(&cli->cl_loi_list_lock);
2909 osc_oap_to_pending(oap);
2910 loi_list_maint(cli, loi);
2912 LOI_DEBUG(loi, "oap %p page %p added for cmd %d\n", oap, oap->oap_page,
2915 osc_check_rpcs(env, cli);
2916 client_obd_list_unlock(&cli->cl_loi_list_lock);
2921 /* aka (~was & now & flag), but this is more clear :) */
2922 #define SETTING(was, now, flag) (!(was & flag) && (now & flag))
2924 int osc_set_async_flags_base(struct client_obd *cli,
2925 struct lov_oinfo *loi, struct osc_async_page *oap,
2926 obd_flag async_flags)
2928 struct loi_oap_pages *lop;
2931 LASSERT(!list_empty(&oap->oap_pending_item));
2933 if (oap->oap_cmd & OBD_BRW_WRITE) {
2934 lop = &loi->loi_write_lop;
2936 lop = &loi->loi_read_lop;
2939 if ((oap->oap_async_flags & async_flags) == async_flags)
2942 if (SETTING(oap->oap_async_flags, async_flags, ASYNC_READY))
2943 oap->oap_async_flags |= ASYNC_READY;
2945 if (SETTING(oap->oap_async_flags, async_flags, ASYNC_URGENT) &&
2946 list_empty(&oap->oap_rpc_item)) {
2947 if (oap->oap_async_flags & ASYNC_HP)
2948 list_add(&oap->oap_urgent_item, &lop->lop_urgent);
2950 list_add_tail(&oap->oap_urgent_item, &lop->lop_urgent);
2951 oap->oap_async_flags |= ASYNC_URGENT;
2952 loi_list_maint(cli, loi);
2955 LOI_DEBUG(loi, "oap %p page %p has flags %x\n", oap, oap->oap_page,
2956 oap->oap_async_flags);
2960 int osc_teardown_async_page(struct obd_export *exp,
2961 struct lov_stripe_md *lsm,
2962 struct lov_oinfo *loi, void *cookie)
2964 struct client_obd *cli = &exp->exp_obd->u.cli;
2965 struct loi_oap_pages *lop;
2966 struct osc_async_page *oap;
2970 oap = oap_from_cookie(cookie);
2972 RETURN(PTR_ERR(oap));
2975 loi = lsm->lsm_oinfo[0];
2977 if (oap->oap_cmd & OBD_BRW_WRITE) {
2978 lop = &loi->loi_write_lop;
2980 lop = &loi->loi_read_lop;
2983 client_obd_list_lock(&cli->cl_loi_list_lock);
2985 if (!list_empty(&oap->oap_rpc_item))
2986 GOTO(out, rc = -EBUSY);
2988 osc_exit_cache(cli, oap, 0);
2989 osc_wake_cache_waiters(cli);
2991 if (!list_empty(&oap->oap_urgent_item)) {
2992 list_del_init(&oap->oap_urgent_item);
2993 oap->oap_async_flags &= ~(ASYNC_URGENT | ASYNC_HP);
2995 if (!list_empty(&oap->oap_pending_item)) {
2996 list_del_init(&oap->oap_pending_item);
2997 lop_update_pending(cli, lop, oap->oap_cmd, -1);
2999 loi_list_maint(cli, loi);
3000 LOI_DEBUG(loi, "oap %p page %p torn down\n", oap, oap->oap_page);
3002 client_obd_list_unlock(&cli->cl_loi_list_lock);
3006 static void osc_set_lock_data_with_check(struct ldlm_lock *lock,
3007 struct ldlm_enqueue_info *einfo,
3010 void *data = einfo->ei_cbdata;
3012 LASSERT(lock != NULL);
3013 LASSERT(lock->l_blocking_ast == einfo->ei_cb_bl);
3014 LASSERT(lock->l_resource->lr_type == einfo->ei_type);
3015 LASSERT(lock->l_completion_ast == einfo->ei_cb_cp);
3016 LASSERT(lock->l_glimpse_ast == einfo->ei_cb_gl);
3018 lock_res_and_lock(lock);
3019 spin_lock(&osc_ast_guard);
3020 LASSERT(lock->l_ast_data == NULL || lock->l_ast_data == data);
3021 lock->l_ast_data = data;
3022 spin_unlock(&osc_ast_guard);
3023 unlock_res_and_lock(lock);
3026 static void osc_set_data_with_check(struct lustre_handle *lockh,
3027 struct ldlm_enqueue_info *einfo,
3030 struct ldlm_lock *lock = ldlm_handle2lock(lockh);
3033 osc_set_lock_data_with_check(lock, einfo, flags);
3034 LDLM_LOCK_PUT(lock);
3036 CERROR("lockh %p, data %p - client evicted?\n",
3037 lockh, einfo->ei_cbdata);
3040 static int osc_change_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
3041 ldlm_iterator_t replace, void *data)
3043 struct ldlm_res_id res_id;
3044 struct obd_device *obd = class_exp2obd(exp);
3046 osc_build_res_name(lsm->lsm_object_id, lsm->lsm_object_gr, &res_id);
3047 ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
3051 static int osc_enqueue_fini(struct ptlrpc_request *req, struct ost_lvb *lvb,
3052 obd_enqueue_update_f upcall, void *cookie,
3055 int intent = *flags & LDLM_FL_HAS_INTENT;
3059 /* The request was created before ldlm_cli_enqueue call. */
3060 if (rc == ELDLM_LOCK_ABORTED) {
3061 struct ldlm_reply *rep;
3062 rep = req_capsule_server_get(&req->rq_pill,
3065 LASSERT(rep != NULL);
3066 if (rep->lock_policy_res1)
3067 rc = rep->lock_policy_res1;
3071 if ((intent && rc == ELDLM_LOCK_ABORTED) || !rc) {
3072 *flags |= LDLM_FL_LVB_READY;
3073 CDEBUG(D_INODE,"got kms "LPU64" blocks "LPU64" mtime "LPU64"\n",
3074 lvb->lvb_size, lvb->lvb_blocks, lvb->lvb_mtime);
3077 /* Call the update callback. */
3078 rc = (*upcall)(cookie, rc);
3082 static int osc_enqueue_interpret(const struct lu_env *env,
3083 struct ptlrpc_request *req,
3084 struct osc_enqueue_args *aa, int rc)
3086 struct ldlm_lock *lock;
3087 struct lustre_handle handle;
3090 /* Make a local copy of a lock handle and a mode, because aa->oa_*
3091 * might be freed anytime after lock upcall has been called. */
3092 lustre_handle_copy(&handle, aa->oa_lockh);
3093 mode = aa->oa_ei->ei_mode;
3095 /* ldlm_cli_enqueue is holding a reference on the lock, so it must
3097 lock = ldlm_handle2lock(&handle);
3099 /* Take an additional reference so that a blocking AST that
3100 * ldlm_cli_enqueue_fini() might post for a failed lock, is guaranteed
3101 * to arrive after an upcall has been executed by
3102 * osc_enqueue_fini(). */
3103 ldlm_lock_addref(&handle, mode);
3105 /* Complete obtaining the lock procedure. */
3106 rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_ei->ei_type, 1,
3107 mode, aa->oa_flags, aa->oa_lvb,
3108 sizeof(*aa->oa_lvb), lustre_swab_ost_lvb,
3110 /* Complete osc stuff. */
3111 rc = osc_enqueue_fini(req, aa->oa_lvb,
3112 aa->oa_upcall, aa->oa_cookie, aa->oa_flags, rc);
3114 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_CANCEL_RACE, 10);
3116 /* Release the lock for async request. */
3117 if (lustre_handle_is_used(&handle) && rc == ELDLM_OK)
3119 * Releases a reference taken by ldlm_cli_enqueue(), if it is
3120 * not already released by
3121 * ldlm_cli_enqueue_fini()->failed_lock_cleanup()
3123 ldlm_lock_decref(&handle, mode);
3125 LASSERTF(lock != NULL, "lockh %p, req %p, aa %p - client evicted?\n",
3126 aa->oa_lockh, req, aa);
3127 ldlm_lock_decref(&handle, mode);
3128 LDLM_LOCK_PUT(lock);
3132 void osc_update_enqueue(struct lustre_handle *lov_lockhp,
3133 struct lov_oinfo *loi, int flags,
3134 struct ost_lvb *lvb, __u32 mode, int rc)
3136 if (rc == ELDLM_OK) {
3137 struct ldlm_lock *lock = ldlm_handle2lock(lov_lockhp);
3140 LASSERT(lock != NULL);
3141 loi->loi_lvb = *lvb;
3142 tmp = loi->loi_lvb.lvb_size;
3143 /* Extend KMS up to the end of this lock and no further
3144 * A lock on [x,y] means a KMS of up to y + 1 bytes! */
3145 if (tmp > lock->l_policy_data.l_extent.end)
3146 tmp = lock->l_policy_data.l_extent.end + 1;
3147 if (tmp >= loi->loi_kms) {
3148 LDLM_DEBUG(lock, "lock acquired, setting rss="LPU64
3149 ", kms="LPU64, loi->loi_lvb.lvb_size, tmp);
3150 loi_kms_set(loi, tmp);
3152 LDLM_DEBUG(lock, "lock acquired, setting rss="
3153 LPU64"; leaving kms="LPU64", end="LPU64,
3154 loi->loi_lvb.lvb_size, loi->loi_kms,
3155 lock->l_policy_data.l_extent.end);
3157 ldlm_lock_allow_match(lock);
3158 LDLM_LOCK_PUT(lock);
3159 } else if (rc == ELDLM_LOCK_ABORTED && (flags & LDLM_FL_HAS_INTENT)) {
3160 loi->loi_lvb = *lvb;
3161 CDEBUG(D_INODE, "glimpsed, setting rss="LPU64"; leaving"
3162 " kms="LPU64"\n", loi->loi_lvb.lvb_size, loi->loi_kms);
3166 EXPORT_SYMBOL(osc_update_enqueue);
3168 struct ptlrpc_request_set *PTLRPCD_SET = (void *)1;
3170 /* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
3171 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
3172 * other synchronous requests, however keeping some locks and trying to obtain
3173 * others may take a considerable amount of time in a case of ost failure; and
3174 * when other sync requests do not get released lock from a client, the client
3175 * is excluded from the cluster -- such scenarious make the life difficult, so
3176 * release locks just after they are obtained. */
3177 int osc_enqueue_base(struct obd_export *exp, struct ldlm_res_id *res_id,
3178 int *flags, ldlm_policy_data_t *policy,
3179 struct ost_lvb *lvb, int kms_valid,
3180 obd_enqueue_update_f upcall, void *cookie,
3181 struct ldlm_enqueue_info *einfo,
3182 struct lustre_handle *lockh,
3183 struct ptlrpc_request_set *rqset, int async)
3185 struct obd_device *obd = exp->exp_obd;
3186 struct ptlrpc_request *req = NULL;
3187 int intent = *flags & LDLM_FL_HAS_INTENT;
3192 /* Filesystem lock extents are extended to page boundaries so that
3193 * dealing with the page cache is a little smoother. */
3194 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
3195 policy->l_extent.end |= ~CFS_PAGE_MASK;
3198 * kms is not valid when either object is completely fresh (so that no
3199 * locks are cached), or object was evicted. In the latter case cached
3200 * lock cannot be used, because it would prime inode state with
3201 * potentially stale LVB.
3206 /* Next, search for already existing extent locks that will cover us */
3207 /* If we're trying to read, we also search for an existing PW lock. The
3208 * VFS and page cache already protect us locally, so lots of readers/
3209 * writers can share a single PW lock.
3211 * There are problems with conversion deadlocks, so instead of
3212 * converting a read lock to a write lock, we'll just enqueue a new
3215 * At some point we should cancel the read lock instead of making them
3216 * send us a blocking callback, but there are problems with canceling
3217 * locks out from other users right now, too. */
3218 mode = einfo->ei_mode;
3219 if (einfo->ei_mode == LCK_PR)
3221 mode = ldlm_lock_match(obd->obd_namespace,
3222 *flags | LDLM_FL_LVB_READY, res_id,
3223 einfo->ei_type, policy, mode, lockh, 0);
3225 struct ldlm_lock *matched = ldlm_handle2lock(lockh);
3227 if (matched->l_ast_data == NULL ||
3228 matched->l_ast_data == einfo->ei_cbdata) {
3229 /* addref the lock only if not async requests and PW
3230 * lock is matched whereas we asked for PR. */
3231 if (!rqset && einfo->ei_mode != mode)
3232 ldlm_lock_addref(lockh, LCK_PR);
3233 osc_set_lock_data_with_check(matched, einfo, *flags);
3235 /* I would like to be able to ASSERT here that
3236 * rss <= kms, but I can't, for reasons which
3237 * are explained in lov_enqueue() */
3240 /* We already have a lock, and it's referenced */
3241 (*upcall)(cookie, ELDLM_OK);
3243 /* For async requests, decref the lock. */
3244 if (einfo->ei_mode != mode)
3245 ldlm_lock_decref(lockh, LCK_PW);
3247 ldlm_lock_decref(lockh, einfo->ei_mode);
3248 LDLM_LOCK_PUT(matched);
3251 ldlm_lock_decref(lockh, mode);
3252 LDLM_LOCK_PUT(matched);
3257 CFS_LIST_HEAD(cancels);
3258 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
3259 &RQF_LDLM_ENQUEUE_LVB);
3263 rc = ldlm_prep_enqueue_req(exp, req, &cancels, 0);
3267 req_capsule_set_size(&req->rq_pill, &RMF_DLM_LVB, RCL_SERVER,
3269 ptlrpc_request_set_replen(req);
3272 /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
3273 *flags &= ~LDLM_FL_BLOCK_GRANTED;
3275 rc = ldlm_cli_enqueue(exp, &req, einfo, res_id, policy, flags, lvb,
3276 sizeof(*lvb), lustre_swab_ost_lvb, lockh, async);
3279 struct osc_enqueue_args *aa;
3280 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
3281 aa = ptlrpc_req_async_args(req);
3284 aa->oa_flags = flags;
3285 aa->oa_upcall = upcall;
3286 aa->oa_cookie = cookie;
3288 aa->oa_lockh = lockh;
3290 req->rq_interpret_reply =
3291 (ptlrpc_interpterer_t)osc_enqueue_interpret;
3292 if (rqset == PTLRPCD_SET)
3293 ptlrpcd_add_req(req, PSCOPE_OTHER);
3295 ptlrpc_set_add_req(rqset, req);
3296 } else if (intent) {
3297 ptlrpc_req_finished(req);
3302 rc = osc_enqueue_fini(req, lvb, upcall, cookie, flags, rc);
3304 ptlrpc_req_finished(req);
3309 static int osc_enqueue(struct obd_export *exp, struct obd_info *oinfo,
3310 struct ldlm_enqueue_info *einfo,
3311 struct ptlrpc_request_set *rqset)
3313 struct ldlm_res_id res_id;
3317 osc_build_res_name(oinfo->oi_md->lsm_object_id,
3318 oinfo->oi_md->lsm_object_gr, &res_id);
3320 rc = osc_enqueue_base(exp, &res_id, &oinfo->oi_flags, &oinfo->oi_policy,
3321 &oinfo->oi_md->lsm_oinfo[0]->loi_lvb,
3322 oinfo->oi_md->lsm_oinfo[0]->loi_kms_valid,
3323 oinfo->oi_cb_up, oinfo, einfo, oinfo->oi_lockh,
3324 rqset, rqset != NULL);
3328 int osc_match_base(struct obd_export *exp, struct ldlm_res_id *res_id,
3329 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
3330 int *flags, void *data, struct lustre_handle *lockh,
3333 struct obd_device *obd = exp->exp_obd;
3334 int lflags = *flags;
3338 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_MATCH))
3341 /* Filesystem lock extents are extended to page boundaries so that
3342 * dealing with the page cache is a little smoother */
3343 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
3344 policy->l_extent.end |= ~CFS_PAGE_MASK;
3346 /* Next, search for already existing extent locks that will cover us */
3347 /* If we're trying to read, we also search for an existing PW lock. The
3348 * VFS and page cache already protect us locally, so lots of readers/
3349 * writers can share a single PW lock. */
3353 rc = ldlm_lock_match(obd->obd_namespace, lflags | LDLM_FL_LVB_READY,
3354 res_id, type, policy, rc, lockh, unref);
3357 osc_set_data_with_check(lockh, data, lflags);
3358 if (!(lflags & LDLM_FL_TEST_LOCK) && mode != rc) {
3359 ldlm_lock_addref(lockh, LCK_PR);
3360 ldlm_lock_decref(lockh, LCK_PW);
3367 int osc_cancel_base(struct lustre_handle *lockh, __u32 mode)
3371 if (unlikely(mode == LCK_GROUP))
3372 ldlm_lock_decref_and_cancel(lockh, mode);
3374 ldlm_lock_decref(lockh, mode);
3379 static int osc_cancel(struct obd_export *exp, struct lov_stripe_md *md,
3380 __u32 mode, struct lustre_handle *lockh)
3383 RETURN(osc_cancel_base(lockh, mode));
3386 static int osc_cancel_unused(struct obd_export *exp,
3387 struct lov_stripe_md *lsm, int flags,
3390 struct obd_device *obd = class_exp2obd(exp);
3391 struct ldlm_res_id res_id, *resp = NULL;
3394 resp = osc_build_res_name(lsm->lsm_object_id,
3395 lsm->lsm_object_gr, &res_id);
3398 return ldlm_cli_cancel_unused(obd->obd_namespace, resp, flags, opaque);
3401 static int osc_statfs_interpret(const struct lu_env *env,
3402 struct ptlrpc_request *req,
3403 struct osc_async_args *aa, int rc)
3405 struct obd_statfs *msfs;
3408 if ((rc == -ENOTCONN || rc == -EAGAIN) &&
3409 (aa->aa_oi->oi_flags & OBD_STATFS_NODELAY))
3415 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
3417 GOTO(out, rc = -EPROTO);
3420 *aa->aa_oi->oi_osfs = *msfs;
3422 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
3426 static int osc_statfs_async(struct obd_device *obd, struct obd_info *oinfo,
3427 __u64 max_age, struct ptlrpc_request_set *rqset)
3429 struct ptlrpc_request *req;
3430 struct osc_async_args *aa;
3434 /* We could possibly pass max_age in the request (as an absolute
3435 * timestamp or a "seconds.usec ago") so the target can avoid doing
3436 * extra calls into the filesystem if that isn't necessary (e.g.
3437 * during mount that would help a bit). Having relative timestamps
3438 * is not so great if request processing is slow, while absolute
3439 * timestamps are not ideal because they need time synchronization. */
3440 req = ptlrpc_request_alloc(obd->u.cli.cl_import, &RQF_OST_STATFS);
3444 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
3446 ptlrpc_request_free(req);
3449 ptlrpc_request_set_replen(req);
3450 req->rq_request_portal = OST_CREATE_PORTAL;
3451 ptlrpc_at_set_req_timeout(req);
3453 if (oinfo->oi_flags & OBD_STATFS_NODELAY) {
3454 /* procfs requests not want stat in wait for avoid deadlock */
3455 req->rq_no_resend = 1;
3456 req->rq_no_delay = 1;
3459 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_statfs_interpret;
3460 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
3461 aa = ptlrpc_req_async_args(req);
3464 ptlrpc_set_add_req(rqset, req);
3468 static int osc_statfs(struct obd_device *obd, struct obd_statfs *osfs,
3469 __u64 max_age, __u32 flags)
3471 struct obd_statfs *msfs;
3472 struct ptlrpc_request *req;
3473 struct obd_import *imp = NULL;
3477 /*Since the request might also come from lprocfs, so we need
3478 *sync this with client_disconnect_export Bug15684*/
3479 down_read(&obd->u.cli.cl_sem);
3480 if (obd->u.cli.cl_import)
3481 imp = class_import_get(obd->u.cli.cl_import);
3482 up_read(&obd->u.cli.cl_sem);
3486 /* We could possibly pass max_age in the request (as an absolute
3487 * timestamp or a "seconds.usec ago") so the target can avoid doing
3488 * extra calls into the filesystem if that isn't necessary (e.g.
3489 * during mount that would help a bit). Having relative timestamps
3490 * is not so great if request processing is slow, while absolute
3491 * timestamps are not ideal because they need time synchronization. */
3492 req = ptlrpc_request_alloc(imp, &RQF_OST_STATFS);
3494 class_import_put(imp);
3499 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
3501 ptlrpc_request_free(req);
3504 ptlrpc_request_set_replen(req);
3505 req->rq_request_portal = OST_CREATE_PORTAL;
3506 ptlrpc_at_set_req_timeout(req);
3508 if (flags & OBD_STATFS_NODELAY) {
3509 /* procfs requests not want stat in wait for avoid deadlock */
3510 req->rq_no_resend = 1;
3511 req->rq_no_delay = 1;
3514 rc = ptlrpc_queue_wait(req);
3518 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
3520 GOTO(out, rc = -EPROTO);
3527 ptlrpc_req_finished(req);
3531 /* Retrieve object striping information.
3533 * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
3534 * the maximum number of OST indices which will fit in the user buffer.
3535 * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
3537 static int osc_getstripe(struct lov_stripe_md *lsm, struct lov_user_md *lump)
3539 /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
3540 struct lov_user_md_v3 lum, *lumk;
3541 struct lov_user_ost_data_v1 *lmm_objects;
3542 int rc = 0, lum_size;
3548 /* we only need the header part from user space to get lmm_magic and
3549 * lmm_stripe_count, (the header part is common to v1 and v3) */
3550 lum_size = sizeof(struct lov_user_md_v1);
3551 if (copy_from_user(&lum, lump, lum_size))
3554 if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
3555 (lum.lmm_magic != LOV_USER_MAGIC_V3))
3558 /* lov_user_md_vX and lov_mds_md_vX must have the same size */
3559 LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
3560 LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
3561 LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
3563 /* we can use lov_mds_md_size() to compute lum_size
3564 * because lov_user_md_vX and lov_mds_md_vX have the same size */
3565 if (lum.lmm_stripe_count > 0) {
3566 lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
3567 OBD_ALLOC(lumk, lum_size);
3571 if (lum.lmm_magic == LOV_USER_MAGIC_V1)
3572 lmm_objects = &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
3574 lmm_objects = &(lumk->lmm_objects[0]);
3575 lmm_objects->l_object_id = lsm->lsm_object_id;
3577 lum_size = lov_mds_md_size(0, lum.lmm_magic);
3581 lumk->lmm_object_id = lsm->lsm_object_id;
3582 lumk->lmm_object_gr = lsm->lsm_object_gr;
3583 lumk->lmm_stripe_count = 1;
3585 if (copy_to_user(lump, lumk, lum_size))
3589 OBD_FREE(lumk, lum_size);
3595 static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
3596 void *karg, void *uarg)
3598 struct obd_device *obd = exp->exp_obd;
3599 struct obd_ioctl_data *data = karg;
3603 if (!try_module_get(THIS_MODULE)) {
3604 CERROR("Can't get module. Is it alive?");
3608 case OBD_IOC_LOV_GET_CONFIG: {
3610 struct lov_desc *desc;
3611 struct obd_uuid uuid;
3615 if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
3616 GOTO(out, err = -EINVAL);
3618 data = (struct obd_ioctl_data *)buf;
3620 if (sizeof(*desc) > data->ioc_inllen1) {
3621 obd_ioctl_freedata(buf, len);
3622 GOTO(out, err = -EINVAL);
3625 if (data->ioc_inllen2 < sizeof(uuid)) {
3626 obd_ioctl_freedata(buf, len);
3627 GOTO(out, err = -EINVAL);
3630 desc = (struct lov_desc *)data->ioc_inlbuf1;
3631 desc->ld_tgt_count = 1;
3632 desc->ld_active_tgt_count = 1;
3633 desc->ld_default_stripe_count = 1;
3634 desc->ld_default_stripe_size = 0;
3635 desc->ld_default_stripe_offset = 0;
3636 desc->ld_pattern = 0;
3637 memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
3639 memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
3641 err = copy_to_user((void *)uarg, buf, len);
3644 obd_ioctl_freedata(buf, len);
3647 case LL_IOC_LOV_SETSTRIPE:
3648 err = obd_alloc_memmd(exp, karg);
3652 case LL_IOC_LOV_GETSTRIPE:
3653 err = osc_getstripe(karg, uarg);
3655 case OBD_IOC_CLIENT_RECOVER:
3656 err = ptlrpc_recover_import(obd->u.cli.cl_import,
3661 case IOC_OSC_SET_ACTIVE:
3662 err = ptlrpc_set_import_active(obd->u.cli.cl_import,
3665 case OBD_IOC_POLL_QUOTACHECK:
3666 err = lquota_poll_check(quota_interface, exp,
3667 (struct if_quotacheck *)karg);
3669 case OBD_IOC_PING_TARGET:
3670 err = ptlrpc_obd_ping(obd);
3673 CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
3674 cmd, cfs_curproc_comm());
3675 GOTO(out, err = -ENOTTY);
3678 module_put(THIS_MODULE);
3682 static int osc_get_info(struct obd_export *exp, obd_count keylen,
3683 void *key, __u32 *vallen, void *val,
3684 struct lov_stripe_md *lsm)
3687 if (!vallen || !val)
3690 if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
3691 __u32 *stripe = val;
3692 *vallen = sizeof(*stripe);
3695 } else if (KEY_IS(KEY_LAST_ID)) {
3696 struct ptlrpc_request *req;
3701 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
3702 &RQF_OST_GET_INFO_LAST_ID);
3706 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3707 RCL_CLIENT, keylen);
3708 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
3710 ptlrpc_request_free(req);
3714 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3715 memcpy(tmp, key, keylen);
3717 req->rq_no_delay = req->rq_no_resend = 1;
3718 ptlrpc_request_set_replen(req);
3719 rc = ptlrpc_queue_wait(req);
3723 reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
3725 GOTO(out, rc = -EPROTO);
3727 *((obd_id *)val) = *reply;
3729 ptlrpc_req_finished(req);
3731 } else if (KEY_IS(KEY_FIEMAP)) {
3732 struct ptlrpc_request *req;
3733 struct ll_user_fiemap *reply;
3737 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
3738 &RQF_OST_GET_INFO_FIEMAP);
3742 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_KEY,
3743 RCL_CLIENT, keylen);
3744 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
3745 RCL_CLIENT, *vallen);
3746 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
3747 RCL_SERVER, *vallen);
3749 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
3751 ptlrpc_request_free(req);
3755 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_KEY);
3756 memcpy(tmp, key, keylen);
3757 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3758 memcpy(tmp, val, *vallen);
3760 ptlrpc_request_set_replen(req);
3761 rc = ptlrpc_queue_wait(req);
3765 reply = req_capsule_server_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3767 GOTO(out1, rc = -EPROTO);
3769 memcpy(val, reply, *vallen);
3771 ptlrpc_req_finished(req);
3779 static int osc_setinfo_mds_connect_import(struct obd_import *imp)
3781 struct llog_ctxt *ctxt;
3785 ctxt = llog_get_context(imp->imp_obd, LLOG_MDS_OST_ORIG_CTXT);
3787 rc = llog_initiator_connect(ctxt);
3788 llog_ctxt_put(ctxt);
3790 /* XXX return an error? skip setting below flags? */
3793 spin_lock(&imp->imp_lock);
3794 imp->imp_server_timeout = 1;
3795 imp->imp_pingable = 1;
3796 spin_unlock(&imp->imp_lock);
3797 CDEBUG(D_RPCTRACE, "pinging OST %s\n", obd2cli_tgt(imp->imp_obd));
3802 static int osc_setinfo_mds_conn_interpret(const struct lu_env *env,
3803 struct ptlrpc_request *req,
3810 RETURN(osc_setinfo_mds_connect_import(req->rq_import));
3813 static int osc_set_info_async(struct obd_export *exp, obd_count keylen,
3814 void *key, obd_count vallen, void *val,
3815 struct ptlrpc_request_set *set)
3817 struct ptlrpc_request *req;
3818 struct obd_device *obd = exp->exp_obd;
3819 struct obd_import *imp = class_exp2cliimp(exp);
3824 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
3826 if (KEY_IS(KEY_NEXT_ID)) {
3827 if (vallen != sizeof(obd_id))
3831 obd->u.cli.cl_oscc.oscc_next_id = *((obd_id*)val) + 1;
3832 CDEBUG(D_HA, "%s: set oscc_next_id = "LPU64"\n",
3833 exp->exp_obd->obd_name,
3834 obd->u.cli.cl_oscc.oscc_next_id);
3839 if (KEY_IS(KEY_UNLINKED)) {
3840 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3841 spin_lock(&oscc->oscc_lock);
3842 oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
3843 spin_unlock(&oscc->oscc_lock);
3847 if (KEY_IS(KEY_INIT_RECOV)) {
3848 if (vallen != sizeof(int))
3850 spin_lock(&imp->imp_lock);
3851 imp->imp_initial_recov = *(int *)val;
3852 spin_unlock(&imp->imp_lock);
3853 CDEBUG(D_HA, "%s: set imp_initial_recov = %d\n",
3854 exp->exp_obd->obd_name,
3855 imp->imp_initial_recov);
3859 if (KEY_IS(KEY_CHECKSUM)) {
3860 if (vallen != sizeof(int))
3862 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
3866 if (KEY_IS(KEY_SPTLRPC_CONF)) {
3867 sptlrpc_conf_client_adapt(obd);
3871 if (KEY_IS(KEY_FLUSH_CTX)) {
3872 sptlrpc_import_flush_my_ctx(imp);
3876 if (!set && !KEY_IS(KEY_GRANT_SHRINK))
3879 /* We pass all other commands directly to OST. Since nobody calls osc
3880 methods directly and everybody is supposed to go through LOV, we
3881 assume lov checked invalid values for us.
3882 The only recognised values so far are evict_by_nid and mds_conn.
3883 Even if something bad goes through, we'd get a -EINVAL from OST
3886 if (KEY_IS(KEY_GRANT_SHRINK))
3887 req = ptlrpc_request_alloc(imp, &RQF_OST_SET_GRANT_INFO);
3889 req = ptlrpc_request_alloc(imp, &RQF_OST_SET_INFO);
3894 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3895 RCL_CLIENT, keylen);
3896 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_VAL,
3897 RCL_CLIENT, vallen);
3898 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SET_INFO);
3900 ptlrpc_request_free(req);
3904 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3905 memcpy(tmp, key, keylen);
3906 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_VAL);
3907 memcpy(tmp, val, vallen);
3909 if (KEY_IS(KEY_MDS_CONN)) {
3910 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3912 oscc->oscc_oa.o_gr = (*(__u32 *)val);
3913 oscc->oscc_oa.o_valid |= OBD_MD_FLGROUP;
3914 LASSERT_MDS_GROUP(oscc->oscc_oa.o_gr);
3915 req->rq_no_delay = req->rq_no_resend = 1;
3916 req->rq_interpret_reply = osc_setinfo_mds_conn_interpret;
3917 } else if (KEY_IS(KEY_GRANT_SHRINK)) {
3918 struct osc_grant_args *aa;
3921 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
3922 aa = ptlrpc_req_async_args(req);
3925 ptlrpc_req_finished(req);
3928 *oa = ((struct ost_body *)val)->oa;
3930 req->rq_interpret_reply = osc_shrink_grant_interpret;
3933 ptlrpc_request_set_replen(req);
3934 if (!KEY_IS(KEY_GRANT_SHRINK)) {
3935 LASSERT(set != NULL);
3936 ptlrpc_set_add_req(set, req);
3937 ptlrpc_check_set(NULL, set);
3939 ptlrpcd_add_req(req, PSCOPE_OTHER);
3945 static struct llog_operations osc_size_repl_logops = {
3946 lop_cancel: llog_obd_repl_cancel
3949 static struct llog_operations osc_mds_ost_orig_logops;
3950 static int osc_llog_init(struct obd_device *obd, struct obd_llog_group *olg,
3951 struct obd_device *tgt, int count,
3952 struct llog_catid *catid, struct obd_uuid *uuid)
3957 LASSERT(olg == &obd->obd_olg);
3958 spin_lock(&obd->obd_dev_lock);
3959 if (osc_mds_ost_orig_logops.lop_setup != llog_obd_origin_setup) {
3960 osc_mds_ost_orig_logops = llog_lvfs_ops;
3961 osc_mds_ost_orig_logops.lop_setup = llog_obd_origin_setup;
3962 osc_mds_ost_orig_logops.lop_cleanup = llog_obd_origin_cleanup;
3963 osc_mds_ost_orig_logops.lop_add = llog_obd_origin_add;
3964 osc_mds_ost_orig_logops.lop_connect = llog_origin_connect;
3966 spin_unlock(&obd->obd_dev_lock);
3968 rc = llog_setup(obd, &obd->obd_olg, LLOG_MDS_OST_ORIG_CTXT, tgt, count,
3969 &catid->lci_logid, &osc_mds_ost_orig_logops);
3971 CERROR("failed LLOG_MDS_OST_ORIG_CTXT\n");
3975 rc = llog_setup(obd, &obd->obd_olg, LLOG_SIZE_REPL_CTXT, tgt, count,
3976 NULL, &osc_size_repl_logops);
3978 struct llog_ctxt *ctxt =
3979 llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
3982 CERROR("failed LLOG_SIZE_REPL_CTXT\n");
3987 CERROR("osc '%s' tgt '%s' cnt %d catid %p rc=%d\n",
3988 obd->obd_name, tgt->obd_name, count, catid, rc);
3989 CERROR("logid "LPX64":0x%x\n",
3990 catid->lci_logid.lgl_oid, catid->lci_logid.lgl_ogen);
3995 static int osc_llog_finish(struct obd_device *obd, int count)
3997 struct llog_ctxt *ctxt;
3998 int rc = 0, rc2 = 0;
4001 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
4003 rc = llog_cleanup(ctxt);
4005 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
4007 rc2 = llog_cleanup(ctxt);
4014 static int osc_reconnect(const struct lu_env *env,
4015 struct obd_export *exp, struct obd_device *obd,
4016 struct obd_uuid *cluuid,
4017 struct obd_connect_data *data,
4020 struct client_obd *cli = &obd->u.cli;
4022 if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
4025 client_obd_list_lock(&cli->cl_loi_list_lock);
4026 data->ocd_grant = cli->cl_avail_grant ?:
4027 2 * cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT;
4028 lost_grant = cli->cl_lost_grant;
4029 cli->cl_lost_grant = 0;
4030 client_obd_list_unlock(&cli->cl_loi_list_lock);
4032 CDEBUG(D_CACHE, "request ocd_grant: %d cl_avail_grant: %ld "
4033 "cl_lost_grant: %ld\n", data->ocd_grant,
4034 cli->cl_avail_grant, lost_grant);
4035 CDEBUG(D_RPCTRACE, "ocd_connect_flags: "LPX64" ocd_version: %d"
4036 " ocd_grant: %d\n", data->ocd_connect_flags,
4037 data->ocd_version, data->ocd_grant);
4043 static int osc_disconnect(struct obd_export *exp)
4045 struct obd_device *obd = class_exp2obd(exp);
4046 struct llog_ctxt *ctxt;
4049 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
4051 if (obd->u.cli.cl_conn_count == 1) {
4052 /* Flush any remaining cancel messages out to the
4054 llog_sync(ctxt, exp);
4056 llog_ctxt_put(ctxt);
4058 CDEBUG(D_HA, "No LLOG_SIZE_REPL_CTXT found in obd %p\n",
4062 rc = client_disconnect_export(exp);
4064 * Initially we put del_shrink_grant before disconnect_export, but it
4065 * causes the following problem if setup (connect) and cleanup
4066 * (disconnect) are tangled together.
4067 * connect p1 disconnect p2
4068 * ptlrpc_connect_import
4069 * ............... class_manual_cleanup
4072 * ptlrpc_connect_interrupt
4074 * add this client to shrink list
4076 * Bang! pinger trigger the shrink.
4077 * So the osc should be disconnected from the shrink list, after we
4078 * are sure the import has been destroyed. BUG18662
4080 if (obd->u.cli.cl_import == NULL)
4081 osc_del_shrink_grant(&obd->u.cli);
4085 static int osc_import_event(struct obd_device *obd,
4086 struct obd_import *imp,
4087 enum obd_import_event event)
4089 struct client_obd *cli;
4093 LASSERT(imp->imp_obd == obd);
4096 case IMP_EVENT_DISCON: {
4097 /* Only do this on the MDS OSC's */
4098 if (imp->imp_server_timeout) {
4099 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
4101 spin_lock(&oscc->oscc_lock);
4102 oscc->oscc_flags |= OSCC_FLAG_RECOVERING;
4103 spin_unlock(&oscc->oscc_lock);
4106 client_obd_list_lock(&cli->cl_loi_list_lock);
4107 cli->cl_avail_grant = 0;
4108 cli->cl_lost_grant = 0;
4109 client_obd_list_unlock(&cli->cl_loi_list_lock);
4112 case IMP_EVENT_INACTIVE: {
4113 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
4116 case IMP_EVENT_INVALIDATE: {
4117 struct ldlm_namespace *ns = obd->obd_namespace;
4121 env = cl_env_get(&refcheck);
4125 client_obd_list_lock(&cli->cl_loi_list_lock);
4126 /* all pages go to failing rpcs due to the invalid
4128 osc_check_rpcs(env, cli);
4129 client_obd_list_unlock(&cli->cl_loi_list_lock);
4131 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
4132 cl_env_put(env, &refcheck);
4137 case IMP_EVENT_ACTIVE: {
4138 /* Only do this on the MDS OSC's */
4139 if (imp->imp_server_timeout) {
4140 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
4142 spin_lock(&oscc->oscc_lock);
4143 oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
4144 spin_unlock(&oscc->oscc_lock);
4146 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
4149 case IMP_EVENT_OCD: {
4150 struct obd_connect_data *ocd = &imp->imp_connect_data;
4152 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
4153 osc_init_grant(&obd->u.cli, ocd);
4156 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
4157 imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;
4159 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
4163 CERROR("Unknown import event %d\n", event);
4169 int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
4175 rc = ptlrpcd_addref();
4179 rc = client_obd_setup(obd, lcfg);
4183 struct lprocfs_static_vars lvars = { 0 };
4184 struct client_obd *cli = &obd->u.cli;
4186 cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
4187 lprocfs_osc_init_vars(&lvars);
4188 if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0) {
4189 lproc_osc_attach_seqstat(obd);
4190 sptlrpc_lprocfs_cliobd_attach(obd);
4191 ptlrpc_lprocfs_register_obd(obd);
4195 /* We need to allocate a few requests more, because
4196 brw_interpret tries to create new requests before freeing
4197 previous ones. Ideally we want to have 2x max_rpcs_in_flight
4198 reserved, but I afraid that might be too much wasted RAM
4199 in fact, so 2 is just my guess and still should work. */
4200 cli->cl_import->imp_rq_pool =
4201 ptlrpc_init_rq_pool(cli->cl_max_rpcs_in_flight + 2,
4203 ptlrpc_add_rqs_to_pool);
4205 CFS_INIT_LIST_HEAD(&cli->cl_grant_shrink_list);
4206 sema_init(&cli->cl_grant_sem, 1);
4212 static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
4218 case OBD_CLEANUP_EARLY: {
4219 struct obd_import *imp;
4220 imp = obd->u.cli.cl_import;
4221 CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
4222 /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
4223 ptlrpc_deactivate_import(imp);
4224 spin_lock(&imp->imp_lock);
4225 imp->imp_pingable = 0;
4226 spin_unlock(&imp->imp_lock);
4229 case OBD_CLEANUP_EXPORTS: {
4230 /* If we set up but never connected, the
4231 client import will not have been cleaned. */
4232 if (obd->u.cli.cl_import) {
4233 struct obd_import *imp;
4234 down_write(&obd->u.cli.cl_sem);
4235 imp = obd->u.cli.cl_import;
4236 CDEBUG(D_CONFIG, "%s: client import never connected\n",
4238 ptlrpc_invalidate_import(imp);
4239 if (imp->imp_rq_pool) {
4240 ptlrpc_free_rq_pool(imp->imp_rq_pool);
4241 imp->imp_rq_pool = NULL;
4243 class_destroy_import(imp);
4244 up_write(&obd->u.cli.cl_sem);
4245 obd->u.cli.cl_import = NULL;
4247 rc = obd_llog_finish(obd, 0);
4249 CERROR("failed to cleanup llogging subsystems\n");
4256 int osc_cleanup(struct obd_device *obd)
4258 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
4262 ptlrpc_lprocfs_unregister_obd(obd);
4263 lprocfs_obd_cleanup(obd);
4265 spin_lock(&oscc->oscc_lock);
4266 oscc->oscc_flags &= ~OSCC_FLAG_RECOVERING;
4267 oscc->oscc_flags |= OSCC_FLAG_EXITING;
4268 spin_unlock(&oscc->oscc_lock);
4270 /* free memory of osc quota cache */
4271 lquota_cleanup(quota_interface, obd);
4273 rc = client_obd_cleanup(obd);
4279 int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *lcfg)
4281 struct lprocfs_static_vars lvars = { 0 };
4284 lprocfs_osc_init_vars(&lvars);
4286 switch (lcfg->lcfg_command) {
4288 rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars,
4298 static int osc_process_config(struct obd_device *obd, obd_count len, void *buf)
4300 return osc_process_config_base(obd, buf);
4303 struct obd_ops osc_obd_ops = {
4304 .o_owner = THIS_MODULE,
4305 .o_setup = osc_setup,
4306 .o_precleanup = osc_precleanup,
4307 .o_cleanup = osc_cleanup,
4308 .o_add_conn = client_import_add_conn,
4309 .o_del_conn = client_import_del_conn,
4310 .o_connect = client_connect_import,
4311 .o_reconnect = osc_reconnect,
4312 .o_disconnect = osc_disconnect,
4313 .o_statfs = osc_statfs,
4314 .o_statfs_async = osc_statfs_async,
4315 .o_packmd = osc_packmd,
4316 .o_unpackmd = osc_unpackmd,
4317 .o_precreate = osc_precreate,
4318 .o_create = osc_create,
4319 .o_destroy = osc_destroy,
4320 .o_getattr = osc_getattr,
4321 .o_getattr_async = osc_getattr_async,
4322 .o_setattr = osc_setattr,
4323 .o_setattr_async = osc_setattr_async,
4325 .o_punch = osc_punch,
4327 .o_enqueue = osc_enqueue,
4328 .o_change_cbdata = osc_change_cbdata,
4329 .o_cancel = osc_cancel,
4330 .o_cancel_unused = osc_cancel_unused,
4331 .o_iocontrol = osc_iocontrol,
4332 .o_get_info = osc_get_info,
4333 .o_set_info_async = osc_set_info_async,
4334 .o_import_event = osc_import_event,
4335 .o_llog_init = osc_llog_init,
4336 .o_llog_finish = osc_llog_finish,
4337 .o_process_config = osc_process_config,
4340 extern struct lu_kmem_descr osc_caches[];
4341 extern spinlock_t osc_ast_guard;
4342 extern struct lock_class_key osc_ast_guard_class;
4344 int __init osc_init(void)
4346 struct lprocfs_static_vars lvars = { 0 };
4350 /* print an address of _any_ initialized kernel symbol from this
4351 * module, to allow debugging with gdb that doesn't support data
4352 * symbols from modules.*/
4353 CDEBUG(D_CONSOLE, "Lustre OSC module (%p).\n", &osc_caches);
4355 rc = lu_kmem_init(osc_caches);
4357 lprocfs_osc_init_vars(&lvars);
4359 request_module("lquota");
4360 quota_interface = PORTAL_SYMBOL_GET(osc_quota_interface);
4361 lquota_init(quota_interface);
4362 init_obd_quota_ops(quota_interface, &osc_obd_ops);
4364 rc = class_register_type(&osc_obd_ops, NULL, lvars.module_vars,
4365 LUSTRE_OSC_NAME, &osc_device_type);
4367 if (quota_interface)
4368 PORTAL_SYMBOL_PUT(osc_quota_interface);
4369 lu_kmem_fini(osc_caches);
4373 spin_lock_init(&osc_ast_guard);
4374 lockdep_set_class(&osc_ast_guard, &osc_ast_guard_class);
4380 static void /*__exit*/ osc_exit(void)
4382 lu_device_type_fini(&osc_device_type);
4384 lquota_exit(quota_interface);
4385 if (quota_interface)
4386 PORTAL_SYMBOL_PUT(osc_quota_interface);
4388 class_unregister_type(LUSTRE_OSC_NAME);
4389 lu_kmem_fini(osc_caches);
4392 MODULE_AUTHOR("Sun Microsystems, Inc. <http://www.lustre.org/>");
4393 MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
4394 MODULE_LICENSE("GPL");
4396 cfs_module(osc, LUSTRE_VERSION_STRING, osc_init, osc_exit);