1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
2 * vim:expandtab:shiftwidth=8:tabstop=8:
4 * Copyright (C) 2002, 2003 Cluster File Systems, Inc.
5 * Author: Phil Schwan <phil@clusterfs.com>
6 * Peter Braam <braam@clusterfs.com>
7 * Mike Shaver <shaver@clusterfs.com>
9 * This file is part of Lustre, http://www.lustre.org.
11 * Lustre is free software; you can redistribute it and/or
12 * modify it under the terms of version 2 of the GNU General Public
13 * License as published by the Free Software Foundation.
15 * Lustre is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with Lustre; if not, write to the Free Software
22 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
26 # define EXPORT_SYMTAB
28 #define DEBUG_SUBSYSTEM S_LOV
30 #include <linux/slab.h>
31 #include <linux/module.h>
32 #include <linux/init.h>
33 #include <linux/slab.h>
34 #include <linux/pagemap.h>
35 #include <linux/seq_file.h>
36 #include <asm/div64.h>
38 #include <liblustre.h>
41 #include <linux/obd_support.h>
42 #include <linux/lustre_lib.h>
43 #include <linux/lustre_net.h>
44 #include <linux/lustre_idl.h>
45 #include <linux/lustre_dlm.h>
46 #include <linux/lustre_mds.h>
47 #include <linux/obd_class.h>
48 #include <linux/obd_lov.h>
49 #include <linux/obd_ost.h>
50 #include <linux/lprocfs_status.h>
52 #include "lov_internal.h"
55 #define MAX_STRING_SIZE 128
56 static int lov_connect_obd(struct obd_device *obd, struct lov_tgt_desc *tgt,
57 int activate, struct obd_connect_data *conn_data,
58 unsigned long connect_flags)
60 struct obd_uuid lov_osc_uuid = { "LOV_OSC_UUID" };
61 struct obd_uuid *tgt_uuid = &tgt->uuid;
64 struct proc_dir_entry *lov_proc_dir;
66 struct lov_obd *lov = &obd->u.lov;
67 struct lustre_handle conn = {0, };
68 struct obd_device *tgt_obd;
72 tgt_obd = class_find_client_obd(tgt_uuid, LUSTRE_OSC_NAME,
76 CERROR("Target %s not attached\n", tgt_uuid->uuid);
80 if (!tgt_obd->obd_set_up) {
81 CERROR("Target %s not set up\n", tgt_uuid->uuid);
86 tgt_obd->obd_no_recov = 0;
87 ptlrpc_activate_import(tgt_obd->u.cli.cl_import);
90 if (tgt_obd->u.cli.cl_import->imp_invalid) {
91 CERROR("not connecting OSC %s; administratively "
92 "disabled\n", tgt_uuid->uuid);
93 rc = obd_register_observer(tgt_obd, obd);
95 CERROR("Target %s register_observer error %d; "
96 "will not be able to reactivate\n",
102 rc = obd_connect(&conn, tgt_obd, &lov_osc_uuid, conn_data,
105 CERROR("Target %s connect error %d\n", tgt_uuid->uuid, rc);
108 tgt->ltd_exp = class_conn2export(&conn);
110 rc = obd_register_observer(tgt_obd, obd);
112 CERROR("Target %s register_observer error %d\n",
114 obd_disconnect(tgt->ltd_exp, 0);
120 lov->desc.ld_active_tgt_count++;
123 lov_proc_dir = lprocfs_srch(obd->obd_proc_entry, "target_obds");
125 struct obd_device *osc_obd = class_conn2obd(&conn);
126 struct proc_dir_entry *osc_symlink;
127 char name[MAX_STRING_SIZE + 1];
129 LASSERT(osc_obd != NULL);
130 LASSERT(osc_obd->obd_type != NULL);
131 LASSERT(osc_obd->obd_type->typ_name != NULL);
132 name[MAX_STRING_SIZE] = '\0';
133 snprintf(name, MAX_STRING_SIZE, "../../../%s/%s",
134 osc_obd->obd_type->typ_name,
136 osc_symlink = proc_symlink(osc_obd->obd_name, lov_proc_dir,
138 if (osc_symlink == NULL) {
139 CERROR("could not register LOV target "
140 "/proc/fs/lustre/%s/%s/target_obds/%s\n",
141 obd->obd_type->typ_name, obd->obd_name,
143 lprocfs_remove(lov_proc_dir);
152 static int lov_connect(struct lustre_handle *conn, struct obd_device *obd,
153 struct obd_uuid *cluuid, struct obd_connect_data *data,
157 struct proc_dir_entry *lov_proc_dir;
159 struct lov_obd *lov = &obd->u.lov;
160 struct lov_tgt_desc *tgt;
161 struct obd_export *exp;
165 rc = class_connect(conn, obd, cluuid);
169 exp = class_conn2export(conn);
171 /* We don't want to actually do the underlying connections more than
172 * once, so keep track. */
174 if (lov->refcount > 1) {
175 class_export_put(exp);
180 lov_proc_dir = lprocfs_register("target_obds", obd->obd_proc_entry,
182 if (IS_ERR(lov_proc_dir)) {
183 CERROR("could not register /proc/fs/lustre/%s/%s/target_obds.",
184 obd->obd_type->typ_name, obd->obd_name);
189 /* connect_flags is the MDS number, save for use in lov_add_obd */
190 lov->lov_connect_flags = flags;
191 for (i = 0, tgt = lov->tgts; i < lov->desc.ld_tgt_count; i++, tgt++) {
192 if (obd_uuid_empty(&tgt->uuid))
194 rc = lov_connect_obd(obd, tgt, 0, data, flags);
199 class_export_put(exp);
205 lprocfs_remove(lov_proc_dir);
209 struct obd_uuid uuid;
211 --lov->desc.ld_active_tgt_count;
213 /* save for CERROR below; (we know it's terminated) */
215 rc2 = obd_disconnect(tgt->ltd_exp, 0);
217 CERROR("error: LOV target %s disconnect on OST idx %d: "
218 "rc = %d\n", uuid.uuid, i, rc2);
220 class_disconnect(exp, 0);
224 static int lov_disconnect_obd(struct obd_device *obd,
225 struct lov_tgt_desc *tgt,
229 struct proc_dir_entry *lov_proc_dir;
231 struct obd_device *osc_obd = class_exp2obd(tgt->ltd_exp);
232 struct lov_obd *lov = &obd->u.lov;
237 lov_proc_dir = lprocfs_srch(obd->obd_proc_entry, "target_obds");
239 struct proc_dir_entry *osc_symlink;
241 osc_symlink = lprocfs_srch(lov_proc_dir, osc_obd->obd_name);
243 lprocfs_remove(osc_symlink);
245 CERROR("/proc/fs/lustre/%s/%s/target_obds/%s missing\n",
246 obd->obd_type->typ_name, obd->obd_name,
251 if (obd->obd_no_recov) {
252 /* Pass it on to our clients.
253 * XXX This should be an argument to disconnect,
254 * XXX not a back-door flag on the OBD. Ah well.
257 osc_obd->obd_no_recov = 1;
260 obd_register_observer(tgt->ltd_exp->exp_obd, NULL);
261 rc = obd_disconnect(tgt->ltd_exp, flags);
264 CERROR("Target %s disconnect error %d\n",
272 lov->desc.ld_active_tgt_count--;
278 static int lov_disconnect(struct obd_export *exp, unsigned long flags)
280 struct obd_device *obd = class_exp2obd(exp);
282 struct proc_dir_entry *lov_proc_dir;
284 struct lov_obd *lov = &obd->u.lov;
285 struct lov_tgt_desc *tgt;
292 /* Only disconnect the underlying layers on the final disconnect. */
294 if (lov->refcount != 0)
297 for (i = 0, tgt = lov->tgts; i < lov->desc.ld_tgt_count; i++, tgt++) {
299 lov_disconnect_obd(obd, tgt, flags);
303 lov_proc_dir = lprocfs_srch(obd->obd_proc_entry, "target_obds");
305 lprocfs_remove(lov_proc_dir);
307 CERROR("/proc/fs/lustre/%s/%s/target_obds missing.",
308 obd->obd_type->typ_name, obd->obd_name);
313 rc = class_disconnect(exp, 0);
319 * -EINVAL : UUID can't be found in the LOV's target list
320 * -ENOTCONN: The UUID is found, but the target connection is bad (!)
321 * -EBADF : The UUID is found, but the OBD is the wrong type (!)
323 static int lov_set_osc_active(struct lov_obd *lov, struct obd_uuid *uuid,
326 struct lov_tgt_desc *tgt;
330 CDEBUG(D_INFO, "Searching in lov %p for uuid %s (activate=%d)\n",
331 lov, uuid->uuid, activate);
333 spin_lock(&lov->lov_lock);
334 for (i = 0, tgt = lov->tgts; i < lov->desc.ld_tgt_count; i++, tgt++) {
335 if (tgt->ltd_exp == NULL)
338 CDEBUG(D_INFO, "lov idx %d is %s conn "LPX64"\n",
339 i, tgt->uuid.uuid, tgt->ltd_exp->exp_handle.h_cookie);
340 if (strncmp(uuid->uuid, tgt->uuid.uuid, sizeof uuid->uuid) == 0)
344 if (i == lov->desc.ld_tgt_count)
345 GOTO(out, rc = -EINVAL);
348 if (tgt->active == activate) {
349 CDEBUG(D_INFO, "OSC %s already %sactive!\n", uuid->uuid,
350 activate ? "" : "in");
354 CDEBUG(D_INFO, "Marking OSC %s %sactive\n", uuid->uuid,
355 activate ? "" : "in");
357 tgt->active = activate;
359 lov->desc.ld_active_tgt_count++;
361 lov->desc.ld_active_tgt_count--;
365 spin_unlock(&lov->lov_lock);
369 static int lov_notify(struct obd_device *obd, struct obd_device *watched,
370 int active, void *data)
372 struct obd_uuid *uuid;
376 if (strcmp(watched->obd_type->typ_name, LUSTRE_OSC_NAME)) {
377 CERROR("unexpected notification of %s %s!\n",
378 watched->obd_type->typ_name,
382 uuid = &watched->u.cli.cl_import->imp_target_uuid;
384 /* Set OSC as active before notifying the observer, so the
385 * observer can use the OSC normally.
387 rc = lov_set_osc_active(&obd->u.lov, uuid, active);
389 CERROR("%sactivation of %s failed: %d\n",
390 active ? "" : "de", uuid->uuid, rc);
394 if (obd->obd_observer)
395 /* Pass the notification up the chain. */
396 rc = obd_notify(obd->obd_observer, watched, active, data);
401 int lov_attach(struct obd_device *dev, obd_count len, void *data)
403 struct lprocfs_static_vars lvars;
406 lprocfs_init_vars(lov, &lvars);
407 rc = lprocfs_obd_attach(dev, lvars.obd_vars);
410 struct proc_dir_entry *entry;
412 entry = create_proc_entry("target_obd_status", 0444,
413 dev->obd_proc_entry);
417 entry->proc_fops = &lov_proc_target_fops;
425 int lov_detach(struct obd_device *dev)
427 return lprocfs_obd_detach(dev);
430 static int lov_setup(struct obd_device *obd, obd_count len, void *buf)
432 struct lov_obd *lov = &obd->u.lov;
433 struct lustre_cfg *lcfg = buf;
434 struct lov_desc *desc;
438 if (lcfg->lcfg_inllen1 < 1) {
439 CERROR("LOV setup requires a descriptor\n");
443 desc = (struct lov_desc *)lcfg->lcfg_inlbuf1;
444 if (sizeof(*desc) > lcfg->lcfg_inllen1) {
445 CERROR("descriptor size wrong: %d > %d\n",
446 (int)sizeof(*desc), lcfg->lcfg_inllen1);
450 /* Because of 64-bit divide/mod operations only work with a 32-bit
451 * divisor in a 32-bit kernel, we cannot support a stripe width
452 * of 4GB or larger on 32-bit CPUs.
455 count = desc->ld_default_stripe_count;
456 if (count && (count * desc->ld_default_stripe_size) > ~0UL) {
457 CERROR("LOV: stripe width "LPU64"x%u > %lu on 32-bit system\n",
458 desc->ld_default_stripe_size, count, ~0UL);
461 if (desc->ld_tgt_count > 0) {
462 lov->bufsize= sizeof(struct lov_tgt_desc) * desc->ld_tgt_count;
464 lov->bufsize = sizeof(struct lov_tgt_desc) * LOV_MAX_TGT_COUNT;
466 OBD_ALLOC(lov->tgts, lov->bufsize);
467 if (lov->tgts == NULL) {
469 CERROR("couldn't allocate %d bytes for target table.\n",
474 desc->ld_tgt_count = 0;
475 desc->ld_active_tgt_count = 0;
477 spin_lock_init(&lov->lov_lock);
478 sema_init(&lov->lov_llog_sem, 1);
483 static int lov_cleanup(struct obd_device *obd, int flags)
485 struct lov_obd *lov = &obd->u.lov;
487 OBD_FREE(lov->tgts, lov->bufsize);
492 lov_add_obd(struct obd_device *obd, struct obd_uuid *uuidp, int index, int gen)
494 struct lov_obd *lov = &obd->u.lov;
495 struct lov_tgt_desc *tgt;
499 CDEBUG(D_CONFIG, "uuid: %s idx: %d gen: %d\n",
500 uuidp->uuid, index, gen);
502 if ((index < 0) || (index >= LOV_MAX_TGT_COUNT)) {
503 CERROR("request to add OBD %s at invalid index: %d\n",
509 CERROR("request to add OBD %s with invalid generation: %d\n",
514 tgt = lov->tgts + index;
515 if (!obd_uuid_empty(&tgt->uuid)) {
516 CERROR("OBD already assigned at LOV target index %d\n",
522 /* XXX - add a sanity check on the generation number. */
525 if (index >= lov->desc.ld_tgt_count)
526 lov->desc.ld_tgt_count = index + 1;
528 CDEBUG(D_CONFIG, "idx: %d ltd_gen: %d ld_tgt_count: %d\n",
529 index, tgt->ltd_gen, lov->desc.ld_tgt_count);
531 if (lov->refcount == 0)
535 struct obd_device *osc_obd;
537 osc_obd = class_exp2obd(tgt->ltd_exp);
539 osc_obd->obd_no_recov = 0;
542 rc = lov_connect_obd(obd, tgt, 1, NULL, lov->lov_connect_flags);
546 if (obd->obd_observer) {
547 /* tell the mds_lov about the new target */
548 rc = obd_notify(obd->obd_observer, tgt->ltd_exp->exp_obd, 1,
554 if (rc && tgt->ltd_exp != NULL)
555 lov_disconnect_obd(obd, tgt, 0);
560 lov_del_obd(struct obd_device *obd, struct obd_uuid *uuidp, int index, int gen)
562 struct lov_obd *lov = &obd->u.lov;
563 struct lov_tgt_desc *tgt;
564 int count = lov->desc.ld_tgt_count;
568 CDEBUG(D_CONFIG, "uuid: %s idx: %d gen: %d\n",
569 uuidp->uuid, index, gen);
571 if (index >= count) {
572 CERROR("LOV target index %d >= number of LOV OBDs %d.\n",
577 tgt = lov->tgts + index;
579 if (obd_uuid_empty(&tgt->uuid)) {
580 CERROR("LOV target at index %d is not setup.\n", index);
584 if (strncmp(uuidp->uuid, tgt->uuid.uuid, sizeof uuidp->uuid) != 0) {
585 CERROR("LOV target UUID %s at index %d doesn't match %s.\n",
586 tgt->uuid.uuid, index, uuidp->uuid);
591 struct obd_device *osc_obd;
593 osc_obd = class_exp2obd(tgt->ltd_exp);
595 osc_obd->obd_no_recov = 1;
596 rc = obd_llog_finish(osc_obd, &osc_obd->obd_llogs, 1);
598 CERROR("osc_llog_finish error: %d\n", rc);
600 lov_disconnect_obd(obd, tgt, 0);
603 /* XXX - right now there is a dependency on ld_tgt_count being the
604 * maximum tgt index for computing the mds_max_easize. So we can't
607 /* lt_gen = 0 will mean it will not match the gen of any valid loi */
608 memset(tgt, 0, sizeof(*tgt));
610 CDEBUG(D_CONFIG, "uuid: %s idx: %d gen: %d exp: %p active: %d\n",
611 tgt->uuid.uuid, index, tgt->ltd_gen, tgt->ltd_exp, tgt->active);
616 static int lov_process_config(struct obd_device *obd, obd_count len, void *buf)
618 struct lustre_cfg *lcfg = buf;
619 struct obd_uuid obd_uuid;
626 switch(cmd = lcfg->lcfg_command) {
627 case LCFG_LOV_ADD_OBD:
628 case LCFG_LOV_DEL_OBD: {
629 if (lcfg->lcfg_inllen1 > sizeof(obd_uuid.uuid))
630 GOTO(out, rc = -EINVAL);
632 obd_str2uuid(&obd_uuid, lcfg->lcfg_inlbuf1);
634 if (sscanf(lcfg->lcfg_inlbuf2, "%d", &index) != 1)
635 GOTO(out, rc = -EINVAL);
636 if (sscanf(lcfg->lcfg_inlbuf3, "%d", &gen) != 1)
637 GOTO(out, rc = -EINVAL);
638 if (cmd == LCFG_LOV_ADD_OBD)
639 rc = lov_add_obd(obd, &obd_uuid, index, gen);
641 rc = lov_del_obd(obd, &obd_uuid, index, gen);
645 CERROR("Unknown command: %d\n", lcfg->lcfg_command);
646 GOTO(out, rc = -EINVAL);
655 #define log2(n) ffz(~(n))
658 static int lov_clear_orphans(struct obd_export *export, struct obdo *src_oa,
659 struct lov_stripe_md **ea,
660 struct obd_trans_info *oti)
664 struct obd_uuid *ost_uuid = NULL;
668 LASSERT(src_oa->o_valid & OBD_MD_FLFLAGS &&
669 src_oa->o_flags == OBD_FL_DELORPHAN);
671 lov = &export->exp_obd->u.lov;
673 tmp_oa = obdo_alloc();
677 if (src_oa->o_valid & OBD_MD_FLINLINE) {
678 ost_uuid = (struct obd_uuid *)src_oa->o_inline;
679 CDEBUG(D_HA, "clearing orphans only for %s\n",
683 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
684 struct lov_stripe_md obj_md;
685 struct lov_stripe_md *obj_mdp = &obj_md;
688 /* if called for a specific target, we don't
689 care if it is not active. */
690 if (lov->tgts[i].active == 0 && ost_uuid == NULL) {
691 CDEBUG(D_HA, "lov idx %d inactive\n", i);
695 if (ost_uuid && !obd_uuid_equals(ost_uuid, &lov->tgts[i].uuid))
698 memcpy(tmp_oa, src_oa, sizeof(*tmp_oa));
700 /* XXX: LOV STACKING: use real "obj_mdp" sub-data */
701 err = obd_create(lov->tgts[i].ltd_exp, tmp_oa, &obj_mdp, oti);
703 /* This export will be disabled until it is recovered,
704 and then orphan recovery will be completed. */
705 CERROR("error in orphan recovery on OST idx %d/%d: "
706 "rc = %d\n", i, lov->desc.ld_tgt_count, err);
715 static int lov_recreate(struct obd_export *exp, struct obdo *src_oa,
716 struct lov_stripe_md **ea, struct obd_trans_info *oti)
718 struct lov_stripe_md *obj_mdp, *lsm;
719 struct lov_obd *lov = &exp->exp_obd->u.lov;
724 LASSERT(src_oa->o_valid & OBD_MD_FLFLAGS &&
725 src_oa->o_flags & OBD_FL_RECREATE_OBJS);
727 OBD_ALLOC(obj_mdp, sizeof(*obj_mdp));
731 ost_idx = src_oa->o_nlink;
734 GOTO(out, rc = -EINVAL);
735 if (ost_idx >= lov->desc.ld_tgt_count)
736 GOTO(out, rc = -EINVAL);
738 for (i = 0; i < lsm->lsm_stripe_count; i++) {
739 if (lsm->lsm_oinfo[i].loi_ost_idx == ost_idx) {
740 if (lsm->lsm_oinfo[i].loi_id != src_oa->o_id)
741 GOTO(out, rc = -EINVAL);
745 if (i == lsm->lsm_stripe_count)
746 GOTO(out, rc = -EINVAL);
748 rc = obd_create(lov->tgts[ost_idx].ltd_exp, src_oa, &obj_mdp, oti);
750 OBD_FREE(obj_mdp, sizeof(*obj_mdp));
754 /* the LOV expects oa->o_id to be set to the LOV object id */
755 static int lov_create(struct obd_export *exp, struct obdo *src_oa,
756 struct lov_stripe_md **ea, struct obd_trans_info *oti)
758 struct lov_request_set *set = NULL;
759 struct list_head *pos;
768 if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
769 src_oa->o_flags == OBD_FL_DELORPHAN) {
770 rc = lov_clear_orphans(exp, src_oa, ea, oti);
774 lov = &exp->exp_obd->u.lov;
775 if (!lov->desc.ld_active_tgt_count)
778 /* Recreate a specific object id at the given OST index */
779 if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
780 (src_oa->o_flags & OBD_FL_RECREATE_OBJS)) {
781 rc = lov_recreate(exp, src_oa, ea, oti);
785 rc = lov_prep_create_set(exp, ea, src_oa, oti, &set);
789 list_for_each (pos, &set->set_list) {
790 struct lov_request *req =
791 list_entry(pos, struct lov_request, rq_link);
793 /* XXX: LOV STACKING: use real "obj_mdp" sub-data */
794 rc = obd_create(lov->tgts[req->rq_idx].ltd_exp,
795 req->rq_oa, &req->rq_md, oti);
796 lov_update_create_set(set, req, rc);
798 rc = lov_fini_create_set(set, ea);
802 #define lsm_bad_magic(LSMP) \
804 struct lov_stripe_md *_lsm__ = (LSMP); \
807 CERROR("LOV requires striping ea\n"); \
809 } else if (_lsm__->lsm_magic != LOV_MAGIC) { \
810 CERROR("LOV striping magic bad %#x != %#x\n", \
811 _lsm__->lsm_magic, LOV_MAGIC); \
817 static int lov_destroy(struct obd_export *exp, struct obdo *oa,
818 struct lov_stripe_md *lsm, struct obd_trans_info *oti)
820 struct lov_request_set *set;
821 struct lov_request *req;
822 struct list_head *pos;
827 if (lsm_bad_magic(lsm))
830 if (!exp || !exp->exp_obd)
833 lov = &exp->exp_obd->u.lov;
834 rc = lov_prep_destroy_set(exp, oa, lsm, oti, &set);
838 list_for_each (pos, &set->set_list) {
840 req = list_entry(pos, struct lov_request, rq_link);
842 /* XXX update the cookie position */
843 oti->oti_logcookies = set->set_cookies + req->rq_stripe;
844 rc = obd_destroy(lov->tgts[req->rq_idx].ltd_exp, req->rq_oa,
846 err = lov_update_common_set(set, req, rc);
848 CERROR("error: destroying objid "LPX64" subobj "
849 LPX64" on OST idx %d: rc = %d\n",
850 set->set_oa->o_id, req->rq_oa->o_id,
856 lov_fini_destroy_set(set);
860 static int lov_getattr(struct obd_export *exp, struct obdo *oa,
861 struct lov_stripe_md *lsm)
863 struct lov_request_set *set;
864 struct lov_request *req;
865 struct list_head *pos;
870 if (lsm_bad_magic(lsm))
873 if (!exp || !exp->exp_obd)
876 lov = &exp->exp_obd->u.lov;
878 rc = lov_prep_getattr_set(exp, oa, lsm, &set);
882 list_for_each (pos, &set->set_list) {
883 req = list_entry(pos, struct lov_request, rq_link);
885 CDEBUG(D_INFO, "objid "LPX64"[%d] has subobj "LPX64" at idx "
886 "%u\n", oa->o_id, req->rq_stripe, req->rq_oa->o_id,
889 rc = obd_getattr(lov->tgts[req->rq_idx].ltd_exp,
891 err = lov_update_common_set(set, req, rc);
893 CERROR("error: getattr objid "LPX64" subobj "
894 LPX64" on OST idx %d: rc = %d\n",
895 set->set_oa->o_id, req->rq_oa->o_id,
901 rc = lov_fini_getattr_set(set);
907 static int lov_getattr_interpret(struct ptlrpc_request_set *rqset, void *data,
910 struct lov_request_set *lovset = (struct lov_request_set *)data;
913 /* don't do attribute merge if this aysnc op failed */
915 lovset->set_completes = 0;
916 lov_fini_getattr_set(lovset);
918 rc = lov_fini_getattr_set(lovset);
923 static int lov_getattr_async(struct obd_export *exp, struct obdo *oa,
924 struct lov_stripe_md *lsm,
925 struct ptlrpc_request_set *rqset)
927 struct lov_request_set *lovset;
929 struct list_head *pos;
930 struct lov_request *req;
934 if (lsm_bad_magic(lsm))
937 if (!exp || !exp->exp_obd)
940 lov = &exp->exp_obd->u.lov;
942 rc = lov_prep_getattr_set(exp, oa, lsm, &lovset);
946 CDEBUG(D_INFO, "objid "LPX64": %ux%u byte stripes\n",
947 lsm->lsm_object_id, lsm->lsm_stripe_count, lsm->lsm_stripe_size);
949 list_for_each (pos, &lovset->set_list) {
950 req = list_entry(pos, struct lov_request, rq_link);
952 CDEBUG(D_INFO, "objid "LPX64"[%d] has subobj "LPX64" at idx "
953 "%u\n", oa->o_id, req->rq_stripe, req->rq_oa->o_id,
955 rc = obd_getattr_async(lov->tgts[req->rq_idx].ltd_exp,
956 req->rq_oa, NULL, rqset);
958 CERROR("error: getattr objid "LPX64" subobj "
959 LPX64" on OST idx %d: rc = %d\n",
960 lovset->set_oa->o_id, req->rq_oa->o_id,
964 lov_update_common_set(lovset, req, rc);
968 LASSERT (rqset->set_interpret == NULL);
969 rqset->set_interpret = lov_getattr_interpret;
970 rqset->set_arg = (void *)lovset;
974 lov_fini_getattr_set(lovset);
978 static int lov_setattr(struct obd_export *exp, struct obdo *src_oa,
979 struct lov_stripe_md *lsm, struct obd_trans_info *oti)
981 struct lov_request_set *set;
983 struct list_head *pos;
984 struct lov_request *req;
988 if (lsm_bad_magic(lsm))
991 if (!exp || !exp->exp_obd)
994 /* for now, we only expect time updates here */
995 LASSERT(!(src_oa->o_valid & ~(OBD_MD_FLID|OBD_MD_FLTYPE | OBD_MD_FLMODE|
996 OBD_MD_FLATIME | OBD_MD_FLMTIME |
997 OBD_MD_FLCTIME | OBD_MD_FLFLAGS |
998 OBD_MD_FLSIZE | OBD_MD_FLGROUP)));
1000 LASSERT(!(src_oa->o_valid & OBD_MD_FLGROUP) || src_oa->o_gr > 0);
1002 lov = &exp->exp_obd->u.lov;
1003 rc = lov_prep_setattr_set(exp, src_oa, lsm, NULL, &set);
1007 list_for_each (pos, &set->set_list) {
1008 req = list_entry(pos, struct lov_request, rq_link);
1010 rc = obd_setattr(lov->tgts[req->rq_idx].ltd_exp, req->rq_oa,
1012 err = lov_update_common_set(set, req, rc);
1014 CERROR("error: setattr objid "LPX64" subobj "
1015 LPX64" on OST idx %d: rc = %d\n",
1016 set->set_oa->o_id, req->rq_oa->o_id,
1022 err = lov_fini_setattr_set(set);
1028 static int lov_revalidate_policy(struct lov_obd *lov, struct lov_stripe_md *lsm)
1030 static int next_idx = 0;
1031 struct lov_tgt_desc *tgt;
1034 /* XXX - we should do something clever and take lsm
1035 * into account but just do round robin for now. */
1037 /* last_idx must always be less that count because
1038 * ld_tgt_count currently cannot shrink. */
1039 count = lov->desc.ld_tgt_count;
1041 for (i = next_idx, tgt = lov->tgts + i; i < count; i++, tgt++) {
1043 next_idx = (i + 1) % count;
1048 for (i = 0, tgt = lov->tgts; i < next_idx; i++, tgt++) {
1050 next_idx = (i + 1) % count;
1058 static int lov_revalidate_md(struct obd_export *exp, struct obdo *src_oa,
1059 struct lov_stripe_md *ea,
1060 struct obd_trans_info *oti)
1062 struct obd_export *osc_exp;
1063 struct lov_obd *lov = &exp->exp_obd->u.lov;
1064 struct lov_stripe_md *lsm = ea;
1065 struct lov_stripe_md obj_md;
1066 struct lov_stripe_md *obj_mdp = &obj_md;
1067 struct lov_oinfo *loi;
1068 struct obdo *tmp_oa;
1069 int ost_idx, updates = 0, i;
1072 tmp_oa = obdo_alloc();
1076 loi = lsm->lsm_oinfo;
1077 for (i = 0; i < lsm->lsm_stripe_count; i++, loi++) {
1079 if (!obd_uuid_empty(&lov->tgts[loi->loi_ost_idx].uuid))
1082 ost_idx = lov_revalidate_policy(lov, lsm);
1084 /* FIXME: punt for now. */
1085 CERROR("lov_revalidate_policy failed; no active "
1090 /* create a new object */
1091 memcpy(tmp_oa, src_oa, sizeof(*tmp_oa));
1092 /* XXX: LOV STACKING: use real "obj_mdp" sub-data */
1093 osc_exp = lov->tgts[ost_idx].ltd_exp;
1094 rc = obd_create(osc_exp, tmp_oa, &obj_mdp, oti);
1096 CERROR("error creating new subobj at idx %d; "
1097 "rc = %d\n", ost_idx, rc);
1101 oti->oti_objid[ost_idx] = tmp_oa->o_id;
1102 loi->loi_id = tmp_oa->o_id;
1103 loi->loi_gr = tmp_oa->o_gr;
1104 loi->loi_ost_idx = ost_idx;
1105 loi->loi_ost_gen = lov->tgts[ost_idx].ltd_gen;
1106 CDEBUG(D_INODE, "replacing objid "LPX64" subobj "LPX64
1107 " with idx %d gen %d.\n", lsm->lsm_object_id,
1108 loi->loi_id, ost_idx, loi->loi_ost_gen);
1112 /* If we got an error revalidating an entry there's no need to
1113 * cleanup up objects we allocated here because the bad entry
1114 * still points to a deleted OST. */
1120 /* FIXME: maybe we'll just make one node the authoritative attribute node, then
1121 * we can send this 'punch' to just the authoritative node and the nodes
1122 * that the punch will affect. */
1123 static int lov_punch(struct obd_export *exp, struct obdo *oa,
1124 struct lov_stripe_md *lsm,
1125 obd_off start, obd_off end, struct obd_trans_info *oti)
1127 struct lov_request_set *set;
1128 struct lov_obd *lov;
1129 struct list_head *pos;
1130 struct lov_request *req;
1131 int err = 0, rc = 0;
1134 if (lsm_bad_magic(lsm))
1137 if (!exp || !exp->exp_obd)
1140 lov = &exp->exp_obd->u.lov;
1141 rc = lov_prep_punch_set(exp, oa, lsm, start, end, oti, &set);
1145 list_for_each (pos, &set->set_list) {
1146 req = list_entry(pos, struct lov_request, rq_link);
1148 rc = obd_punch(lov->tgts[req->rq_idx].ltd_exp, req->rq_oa,
1149 NULL, req->rq_extent.start,
1150 req->rq_extent.end, NULL);
1151 err = lov_update_punch_set(set, req, rc);
1153 CERROR("error: punch objid "LPX64" subobj "LPX64
1154 " on OST idx %d: rc = %d\n", set->set_oa->o_id,
1155 req->rq_oa->o_id, req->rq_idx, rc);
1160 err = lov_fini_punch_set(set);
1166 static int lov_sync(struct obd_export *exp, struct obdo *oa,
1167 struct lov_stripe_md *lsm, obd_off start, obd_off end)
1169 struct lov_request_set *set;
1170 struct lov_obd *lov;
1171 struct list_head *pos;
1172 struct lov_request *req;
1173 int err = 0, rc = 0;
1176 if (lsm_bad_magic(lsm))
1182 lov = &exp->exp_obd->u.lov;
1183 rc = lov_prep_sync_set(exp, oa, lsm, start, end, &set);
1187 list_for_each (pos, &set->set_list) {
1188 req = list_entry(pos, struct lov_request, rq_link);
1190 rc = obd_sync(lov->tgts[req->rq_idx].ltd_exp, req->rq_oa,
1191 NULL, req->rq_extent.start, req->rq_extent.end);
1192 err = lov_update_common_set(set, req, rc);
1194 CERROR("error: fsync objid "LPX64" subobj "LPX64
1195 " on OST idx %d: rc = %d\n", set->set_oa->o_id,
1196 req->rq_oa->o_id, req->rq_idx, rc);
1201 err = lov_fini_sync_set(set);
1207 static int lov_brw_check(struct lov_obd *lov, struct obdo *oa,
1208 struct lov_stripe_md *lsm,
1209 obd_count oa_bufs, struct brw_page *pga)
1214 /* The caller just wants to know if there's a chance that this
1215 * I/O can succeed */
1216 for (i = 0; i < oa_bufs; i++) {
1217 int stripe = lov_stripe_number(lsm, pga[i].disk_offset);
1218 int ost = lsm->lsm_oinfo[stripe].loi_ost_idx;
1221 if (!lov_stripe_intersects(lsm, i, pga[i].disk_offset,
1222 pga[i].disk_offset + pga[i].count,
1226 if (lov->tgts[ost].active == 0) {
1227 CDEBUG(D_HA, "lov idx %d inactive\n", ost);
1230 rc = obd_brw(OBD_BRW_CHECK, lov->tgts[ost].ltd_exp, oa,
1231 NULL, 1, &pga[i], NULL);
1238 static int lov_brw(int cmd, struct obd_export *exp, struct obdo *src_oa,
1239 struct lov_stripe_md *lsm, obd_count oa_bufs,
1240 struct brw_page *pga, struct obd_trans_info *oti)
1242 struct lov_request_set *set;
1243 struct lov_request *req;
1244 struct list_head *pos;
1245 struct lov_obd *lov = &exp->exp_obd->u.lov;
1249 if (lsm_bad_magic(lsm))
1252 if (cmd == OBD_BRW_CHECK) {
1253 rc = lov_brw_check(lov, src_oa, lsm, oa_bufs, pga);
1257 rc = lov_prep_brw_set(exp, src_oa, lsm, oa_bufs, pga, oti, &set);
1261 list_for_each (pos, &set->set_list) {
1262 struct obd_export *sub_exp;
1263 struct brw_page *sub_pga;
1264 req = list_entry(pos, struct lov_request, rq_link);
1266 sub_exp = lov->tgts[req->rq_idx].ltd_exp;
1267 sub_pga = set->set_pga + req->rq_pgaidx;
1268 rc = obd_brw(cmd, sub_exp, req->rq_oa, req->rq_md,
1269 req->rq_oabufs, sub_pga, oti);
1272 lov_update_common_set(set, req, rc);
1275 err = lov_fini_brw_set(set);
1281 static int lov_brw_interpret(struct ptlrpc_request_set *reqset, void *data,
1284 struct lov_request_set *lovset = (struct lov_request_set *)data;
1288 lovset->set_completes = 0;
1289 lov_fini_brw_set(lovset);
1291 rc = lov_fini_brw_set(lovset);
1297 static int lov_brw_async(int cmd, struct obd_export *exp, struct obdo *oa,
1298 struct lov_stripe_md *lsm, obd_count oa_bufs,
1299 struct brw_page *pga, struct ptlrpc_request_set *set,
1300 struct obd_trans_info *oti)
1302 struct lov_request_set *lovset;
1303 struct lov_request *req;
1304 struct list_head *pos;
1305 struct lov_obd *lov = &exp->exp_obd->u.lov;
1309 if (lsm_bad_magic(lsm))
1312 if (cmd == OBD_BRW_CHECK) {
1313 rc = lov_brw_check(lov, oa, lsm, oa_bufs, pga);
1317 rc = lov_prep_brw_set(exp, oa, lsm, oa_bufs, pga, oti, &lovset);
1321 list_for_each (pos, &lovset->set_list) {
1322 struct obd_export *sub_exp;
1323 struct brw_page *sub_pga;
1324 req = list_entry(pos, struct lov_request, rq_link);
1326 sub_exp = lov->tgts[req->rq_idx].ltd_exp;
1327 sub_pga = lovset->set_pga + req->rq_pgaidx;
1328 rc = obd_brw_async(cmd, sub_exp, req->rq_oa, req->rq_md,
1329 req->rq_oabufs, sub_pga, set, oti);
1332 lov_update_common_set(lovset, req, rc);
1335 LASSERT(set->set_interpret == NULL);
1336 set->set_interpret = (set_interpreter_func)lov_brw_interpret;
1337 set->set_arg = (void *)lovset;
1341 lov_fini_brw_set(lovset);
1345 static int lov_ap_make_ready(void *data, int cmd)
1347 struct lov_async_page *lap = LAP_FROM_COOKIE(data);
1349 return lap->lap_caller_ops->ap_make_ready(lap->lap_caller_data, cmd);
1351 static int lov_ap_refresh_count(void *data, int cmd)
1353 struct lov_async_page *lap = LAP_FROM_COOKIE(data);
1355 return lap->lap_caller_ops->ap_refresh_count(lap->lap_caller_data,
1358 static void lov_ap_fill_obdo(void *data, int cmd, struct obdo *oa)
1360 struct lov_async_page *lap = LAP_FROM_COOKIE(data);
1362 lap->lap_caller_ops->ap_fill_obdo(lap->lap_caller_data, cmd, oa);
1363 /* XXX woah, shouldn't we be altering more here? size? */
1364 oa->o_id = lap->lap_loi_id;
1367 static void lov_ap_completion(void *data, int cmd, struct obdo *oa, int rc)
1369 struct lov_async_page *lap = LAP_FROM_COOKIE(data);
1371 /* in a raid1 regime this would down a count of many ios
1372 * in flight, onl calling the caller_ops completion when all
1373 * the raid1 ios are complete */
1374 lap->lap_caller_ops->ap_completion(lap->lap_caller_data, cmd, oa, rc);
1377 static struct obd_async_page_ops lov_async_page_ops = {
1378 .ap_make_ready = lov_ap_make_ready,
1379 .ap_refresh_count = lov_ap_refresh_count,
1380 .ap_fill_obdo = lov_ap_fill_obdo,
1381 .ap_completion = lov_ap_completion,
1384 static int lov_prep_async_page(struct obd_export *exp,
1385 struct lov_stripe_md *lsm,
1386 struct lov_oinfo *loi, struct page *page,
1387 obd_off offset, struct obd_async_page_ops *ops,
1388 void *data, void **res)
1390 struct lov_obd *lov = &exp->exp_obd->u.lov;
1391 struct lov_async_page *lap;
1395 if (lsm_bad_magic(lsm))
1397 LASSERT(loi == NULL);
1399 stripe = lov_stripe_number(lsm, offset);
1400 loi = &lsm->lsm_oinfo[stripe];
1402 if (obd_uuid_empty(&lov->tgts[loi->loi_ost_idx].uuid))
1404 if (lov->tgts[loi->loi_ost_idx].active == 0)
1406 if (lov->tgts[loi->loi_ost_idx].ltd_exp == NULL) {
1407 CERROR("ltd_exp == NULL, but OST idx %d doesn't appear to be "
1408 "deleted or inactive.\n", loi->loi_ost_idx);
1412 OBD_ALLOC(lap, sizeof(*lap));
1416 lap->lap_magic = LAP_MAGIC;
1417 lap->lap_caller_ops = ops;
1418 lap->lap_caller_data = data;
1420 /* FIXME handle multiple oscs after landing b_raid1 */
1421 lap->lap_stripe = stripe;
1422 switch (lsm->lsm_pattern) {
1423 case LOV_PATTERN_RAID0:
1424 lov_stripe_offset(lsm, offset, lap->lap_stripe,
1425 &lap->lap_sub_offset);
1427 case LOV_PATTERN_CMOBD:
1428 lap->lap_sub_offset = offset;
1434 /* so the callback doesn't need the lsm */
1435 lap->lap_loi_id = loi->loi_id;
1437 rc = obd_prep_async_page(lov->tgts[loi->loi_ost_idx].ltd_exp,
1438 lsm, loi, page, lap->lap_sub_offset,
1439 &lov_async_page_ops, lap,
1440 &lap->lap_sub_cookie);
1442 OBD_FREE(lap, sizeof(*lap));
1445 CDEBUG(D_CACHE, "lap %p page %p cookie %p off "LPU64"\n", lap, page,
1446 lap->lap_sub_cookie, offset);
1451 static int lov_queue_async_io(struct obd_export *exp,
1452 struct lov_stripe_md *lsm,
1453 struct lov_oinfo *loi, void *cookie,
1454 int cmd, obd_off off, int count,
1455 obd_flags brw_flags, obd_flags async_flags)
1457 struct lov_obd *lov = &exp->exp_obd->u.lov;
1458 struct lov_async_page *lap;
1461 LASSERT(loi == NULL);
1463 if (lsm_bad_magic(lsm))
1466 lap = LAP_FROM_COOKIE(cookie);
1468 loi = &lsm->lsm_oinfo[lap->lap_stripe];
1470 rc = obd_queue_async_io(lov->tgts[loi->loi_ost_idx].ltd_exp, lsm,
1471 loi, lap->lap_sub_cookie, cmd, off, count,
1472 brw_flags, async_flags);
1476 static int lov_set_async_flags(struct obd_export *exp,
1477 struct lov_stripe_md *lsm,
1478 struct lov_oinfo *loi, void *cookie,
1479 obd_flags async_flags)
1481 struct lov_obd *lov = &exp->exp_obd->u.lov;
1482 struct lov_async_page *lap;
1485 LASSERT(loi == NULL);
1487 if (lsm_bad_magic(lsm))
1490 lap = LAP_FROM_COOKIE(cookie);
1492 loi = &lsm->lsm_oinfo[lap->lap_stripe];
1494 rc = obd_set_async_flags(lov->tgts[loi->loi_ost_idx].ltd_exp,
1495 lsm, loi, lap->lap_sub_cookie, async_flags);
1499 static int lov_queue_group_io(struct obd_export *exp,
1500 struct lov_stripe_md *lsm,
1501 struct lov_oinfo *loi,
1502 struct obd_io_group *oig, void *cookie,
1503 int cmd, obd_off off, int count,
1504 obd_flags brw_flags, obd_flags async_flags)
1506 struct lov_obd *lov = &exp->exp_obd->u.lov;
1507 struct lov_async_page *lap;
1510 LASSERT(loi == NULL);
1512 if (lsm_bad_magic(lsm))
1515 lap = LAP_FROM_COOKIE(cookie);
1517 loi = &lsm->lsm_oinfo[lap->lap_stripe];
1519 rc = obd_queue_group_io(lov->tgts[loi->loi_ost_idx].ltd_exp, lsm, loi,
1520 oig, lap->lap_sub_cookie, cmd, off, count,
1521 brw_flags, async_flags);
1525 /* this isn't exactly optimal. we may have queued sync io in oscs on
1526 * all stripes, but we don't record that fact at queue time. so we
1527 * trigger sync io on all stripes. */
1528 static int lov_trigger_group_io(struct obd_export *exp,
1529 struct lov_stripe_md *lsm,
1530 struct lov_oinfo *loi,
1531 struct obd_io_group *oig)
1533 struct lov_obd *lov = &exp->exp_obd->u.lov;
1536 LASSERT(loi == NULL);
1538 if (lsm_bad_magic(lsm))
1541 loi = lsm->lsm_oinfo;
1542 for (i = 0; i < lsm->lsm_stripe_count; i++, loi++) {
1543 if (lov->tgts[loi->loi_ost_idx].active == 0) {
1544 CDEBUG(D_HA, "lov idx %d inactive\n", loi->loi_ost_idx);
1548 err = obd_trigger_group_io(lov->tgts[loi->loi_ost_idx].ltd_exp,
1550 if (rc == 0 && err != 0)
1556 static int lov_teardown_async_page(struct obd_export *exp,
1557 struct lov_stripe_md *lsm,
1558 struct lov_oinfo *loi, void *cookie)
1560 struct lov_obd *lov = &exp->exp_obd->u.lov;
1561 struct lov_async_page *lap;
1564 LASSERT(loi == NULL);
1566 if (lsm_bad_magic(lsm))
1569 lap = LAP_FROM_COOKIE(cookie);
1571 loi = &lsm->lsm_oinfo[lap->lap_stripe];
1573 rc = obd_teardown_async_page(lov->tgts[loi->loi_ost_idx].ltd_exp,
1574 lsm, loi, lap->lap_sub_cookie);
1576 CERROR("unable to teardown sub cookie %p: %d\n",
1577 lap->lap_sub_cookie, rc);
1580 OBD_FREE(lap, sizeof(*lap));
1584 static int lov_enqueue(struct obd_export *exp, struct lov_stripe_md *lsm,
1585 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
1586 int *flags, void *bl_cb, void *cp_cb, void *gl_cb,
1587 void *data,__u32 lvb_len, void *lvb_swabber,
1588 struct lustre_handle *lockh)
1590 struct lov_request_set *set;
1591 struct lov_request *req;
1592 struct list_head *pos;
1593 struct lustre_handle *lov_lockhp;
1594 struct lov_obd *lov;
1596 int save_flags = *flags;
1599 if (lsm_bad_magic(lsm))
1602 /* we should never be asked to replay a lock this way. */
1603 LASSERT((*flags & LDLM_FL_REPLAY) == 0);
1605 if (!exp || !exp->exp_obd)
1608 lov = &exp->exp_obd->u.lov;
1609 rc = lov_prep_enqueue_set(exp, lsm, policy, mode, lockh, &set);
1613 list_for_each (pos, &set->set_list) {
1614 ldlm_policy_data_t sub_policy;
1615 req = list_entry(pos, struct lov_request, rq_link);
1616 lov_lockhp = set->set_lockh->llh_handles + req->rq_stripe;
1617 LASSERT(lov_lockhp);
1619 *flags = save_flags;
1620 sub_policy.l_extent.start = req->rq_extent.start;
1621 sub_policy.l_extent.end = req->rq_extent.end;
1623 rc = obd_enqueue(lov->tgts[req->rq_idx].ltd_exp, req->rq_md,
1624 type, &sub_policy, mode, flags, bl_cb,
1625 cp_cb, gl_cb, data, lvb_len, lvb_swabber,
1627 rc = lov_update_enqueue_set(set, req, rc, save_flags);
1632 lov_fini_enqueue_set(set, mode);
1636 static int lov_match(struct obd_export *exp, struct lov_stripe_md *lsm,
1637 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
1638 int *flags, void *data, struct lustre_handle *lockh)
1640 struct lov_request_set *set;
1641 struct lov_request *req;
1642 struct list_head *pos;
1643 struct lov_obd *lov = &exp->exp_obd->u.lov;
1644 struct lustre_handle *lov_lockhp;
1645 int lov_flags, rc = 0;
1648 if (lsm_bad_magic(lsm))
1651 if (!exp || !exp->exp_obd)
1654 lov = &exp->exp_obd->u.lov;
1655 rc = lov_prep_match_set(exp, lsm, policy, mode, lockh, &set);
1659 list_for_each (pos, &set->set_list) {
1660 ldlm_policy_data_t sub_policy;
1661 req = list_entry(pos, struct lov_request, rq_link);
1662 lov_lockhp = set->set_lockh->llh_handles + req->rq_stripe;
1663 LASSERT(lov_lockhp);
1665 sub_policy.l_extent.start = req->rq_extent.start;
1666 sub_policy.l_extent.end = req->rq_extent.end;
1669 rc = obd_match(lov->tgts[req->rq_idx].ltd_exp, req->rq_md,
1670 type, &sub_policy, mode, &lov_flags, data,
1672 rc = lov_update_match_set(set, req, rc);
1676 lov_fini_match_set(set, mode, *flags);
1680 static int lov_change_cbdata(struct obd_export *exp,
1681 struct lov_stripe_md *lsm, ldlm_iterator_t it,
1684 struct lov_obd *lov;
1685 struct lov_oinfo *loi;
1689 if (lsm_bad_magic(lsm))
1692 if (!exp || !exp->exp_obd)
1695 LASSERT(lsm->lsm_object_gr > 0);
1697 lov = &exp->exp_obd->u.lov;
1698 for (i = 0,loi = lsm->lsm_oinfo; i < lsm->lsm_stripe_count; i++,loi++) {
1699 struct lov_stripe_md submd;
1700 if (lov->tgts[loi->loi_ost_idx].active == 0) {
1701 CDEBUG(D_HA, "lov idx %d inactive\n", loi->loi_ost_idx);
1705 submd.lsm_object_id = loi->loi_id;
1706 submd.lsm_object_gr = lsm->lsm_object_gr;
1707 submd.lsm_stripe_count = 0;
1708 rc = obd_change_cbdata(lov->tgts[loi->loi_ost_idx].ltd_exp,
1714 static int lov_cancel(struct obd_export *exp, struct lov_stripe_md *lsm,
1715 __u32 mode, struct lustre_handle *lockh)
1717 struct lov_request_set *set;
1718 struct lov_request *req;
1719 struct list_head *pos;
1720 struct lov_obd *lov = &exp->exp_obd->u.lov;
1721 struct lustre_handle *lov_lockhp;
1722 int err = 0, rc = 0;
1725 if (lsm_bad_magic(lsm))
1728 if (!exp || !exp->exp_obd)
1731 LASSERT(lsm->lsm_object_gr > 0);
1734 lov = &exp->exp_obd->u.lov;
1735 rc = lov_prep_cancel_set(exp, lsm, mode, lockh, &set);
1739 list_for_each (pos, &set->set_list) {
1740 req = list_entry(pos, struct lov_request, rq_link);
1741 lov_lockhp = set->set_lockh->llh_handles + req->rq_stripe;
1743 rc = obd_cancel(lov->tgts[req->rq_idx].ltd_exp, req->rq_md,
1745 rc = lov_update_common_set(set, req, rc);
1747 CERROR("error: cancel objid "LPX64" subobj "
1748 LPX64" on OST idx %d: rc = %d\n",
1750 req->rq_md->lsm_object_id, req->rq_idx, rc);
1755 lov_fini_cancel_set(set);
1759 static int lov_cancel_unused(struct obd_export *exp,
1760 struct lov_stripe_md *lsm,
1761 int flags, void *opaque)
1763 struct lov_obd *lov;
1764 struct lov_oinfo *loi;
1768 lov = &exp->exp_obd->u.lov;
1770 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
1771 int err = obd_cancel_unused(lov->tgts[i].ltd_exp, NULL,
1779 if (lsm_bad_magic(lsm))
1782 if (!exp || !exp->exp_obd)
1785 LASSERT(lsm->lsm_object_gr > 0);
1787 for (i = 0,loi = lsm->lsm_oinfo; i < lsm->lsm_stripe_count; i++,loi++) {
1788 struct lov_stripe_md submd;
1791 if (lov->tgts[loi->loi_ost_idx].active == 0)
1792 CDEBUG(D_HA, "lov idx %d inactive\n", loi->loi_ost_idx);
1794 submd.lsm_object_id = loi->loi_id;
1795 submd.lsm_object_gr = lsm->lsm_object_gr;
1796 submd.lsm_stripe_count = 0;
1797 err = obd_cancel_unused(lov->tgts[loi->loi_ost_idx].ltd_exp,
1798 &submd, flags, opaque);
1799 if (err && lov->tgts[loi->loi_ost_idx].active) {
1800 CERROR("error: cancel unused objid "LPX64" subobj "LPX64
1801 " on OST idx %d: rc = %d\n", lsm->lsm_object_id,
1802 loi->loi_id, loi->loi_ost_idx, err);
1810 #define LOV_U64_MAX ((__u64)~0ULL)
1811 #define LOV_SUM_MAX(tot, add) \
1813 if ((tot) + (add) < (tot)) \
1814 (tot) = LOV_U64_MAX; \
1819 static int lov_statfs(struct obd_device *obd, struct obd_statfs *osfs,
1820 unsigned long max_age)
1822 struct lov_obd *lov = &obd->u.lov;
1823 struct obd_statfs lov_sfs;
1830 /* We only get block data from the OBD */
1831 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
1833 if (!lov->tgts[i].active) {
1834 CDEBUG(D_HA, "lov idx %d inactive\n", i);
1838 err = obd_statfs(class_exp2obd(lov->tgts[i].ltd_exp), &lov_sfs,
1841 if (lov->tgts[i].active && !rc)
1847 memcpy(osfs, &lov_sfs, sizeof(lov_sfs));
1850 osfs->os_bfree += lov_sfs.os_bfree;
1851 osfs->os_bavail += lov_sfs.os_bavail;
1852 osfs->os_blocks += lov_sfs.os_blocks;
1853 /* XXX not sure about this one - depends on policy.
1854 * - could be minimum if we always stripe on all OBDs
1855 * (but that would be wrong for any other policy,
1856 * if one of the OBDs has no more objects left)
1857 * - could be sum if we stripe whole objects
1858 * - could be average, just to give a nice number
1860 * To give a "reasonable" (if not wholly accurate)
1861 * number, we divide the total number of free objects
1862 * by expected stripe count (watch out for overflow).
1864 LOV_SUM_MAX(osfs->os_files, lov_sfs.os_files);
1865 LOV_SUM_MAX(osfs->os_ffree, lov_sfs.os_ffree);
1870 __u32 expected_stripes = lov->desc.ld_default_stripe_count ?
1871 lov->desc.ld_default_stripe_count :
1872 lov->desc.ld_active_tgt_count;
1874 if (osfs->os_files != LOV_U64_MAX)
1875 do_div(osfs->os_files, expected_stripes);
1876 if (osfs->os_ffree != LOV_U64_MAX)
1877 do_div(osfs->os_ffree, expected_stripes);
1884 static int lov_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
1885 void *karg, void *uarg)
1887 struct obd_device *obddev = class_exp2obd(exp);
1888 struct lov_obd *lov = &obddev->u.lov;
1889 int i, rc, count = lov->desc.ld_tgt_count;
1890 struct obd_uuid *uuidp;
1894 case OBD_IOC_LOV_GET_CONFIG: {
1895 struct obd_ioctl_data *data = karg;
1896 struct lov_tgt_desc *tgtdesc;
1897 struct lov_desc *desc;
1903 if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
1906 data = (struct obd_ioctl_data *)buf;
1908 if (sizeof(*desc) > data->ioc_inllen1) {
1909 obd_ioctl_freedata(buf, len);
1913 if (sizeof(uuidp->uuid) * count > data->ioc_inllen2) {
1914 obd_ioctl_freedata(buf, len);
1918 if (sizeof(__u32) * count > data->ioc_inllen3) {
1919 obd_ioctl_freedata(buf, len);
1923 desc = (struct lov_desc *)data->ioc_inlbuf1;
1924 memcpy(desc, &(lov->desc), sizeof(*desc));
1926 uuidp = (struct obd_uuid *)data->ioc_inlbuf2;
1927 genp = (__u32 *)data->ioc_inlbuf3;
1928 tgtdesc = lov->tgts;
1929 /* the uuid will be empty for deleted OSTs */
1930 for (i = 0; i < count; i++, uuidp++, genp++, tgtdesc++) {
1931 obd_str2uuid(uuidp, tgtdesc->uuid.uuid);
1932 *genp = tgtdesc->ltd_gen;
1935 rc = copy_to_user((void *)uarg, buf, len);
1938 obd_ioctl_freedata(buf, len);
1941 case LL_IOC_LOV_SETSTRIPE:
1942 rc = lov_setstripe(exp, karg, uarg);
1944 case LL_IOC_LOV_GETSTRIPE:
1945 rc = lov_getstripe(exp, karg, uarg);
1947 case LL_IOC_LOV_SETEA:
1948 rc = lov_setea(exp, karg, uarg);
1955 for (i = 0; i < count; i++) {
1958 /* OST was deleted */
1959 if (obd_uuid_empty(&lov->tgts[i].uuid))
1962 err = obd_iocontrol(cmd, lov->tgts[i].ltd_exp,
1965 if (lov->tgts[i].active) {
1966 CERROR("error: iocontrol OSC %s on OST "
1967 "idx %d cmd %x: err = %d\n",
1968 lov->tgts[i].uuid.uuid, i,
1984 static int lov_get_info(struct obd_export *exp, __u32 keylen,
1985 void *key, __u32 *vallen, void *val)
1987 struct obd_device *obddev = class_exp2obd(exp);
1988 struct lov_obd *lov = &obddev->u.lov;
1992 if (!vallen || !val)
1995 if (keylen > strlen("lock_to_stripe") &&
1996 strcmp(key, "lock_to_stripe") == 0) {
1999 struct ldlm_lock *lock;
2000 struct lov_stripe_md *lsm;
2002 struct lov_oinfo *loi;
2003 struct ldlm_res_id *res_id = &data->lock->l_resource->lr_name;
2004 __u32 *stripe = val;
2006 if (*vallen < sizeof(*stripe))
2008 *vallen = sizeof(*stripe);
2010 /* XXX This is another one of those bits that will need to
2011 * change if we ever actually support nested LOVs. It uses
2012 * the lock's export to find out which stripe it is. */
2013 /* XXX - it's assumed all the locks for deleted OSTs have
2014 * been cancelled. Also, the export for deleted OSTs will
2015 * be NULL and won't match the lock's export. */
2016 for (i = 0, loi = data->lsm->lsm_oinfo;
2017 i < data->lsm->lsm_stripe_count;
2019 if (lov->tgts[loi->loi_ost_idx].ltd_exp ==
2020 data->lock->l_conn_export &&
2021 loi->loi_id == res_id->name[0] &&
2022 loi->loi_gr == res_id->name[2]) {
2027 LDLM_ERROR(data->lock, "lock on inode without such object\n");
2028 dump_lsm(D_ERROR, data->lsm);
2030 } else if (keylen >= strlen("size_to_stripe") &&
2031 strcmp(key, "size_to_stripe") == 0) {
2035 struct lov_stripe_md *lsm;
2038 if (*vallen < sizeof(*data))
2041 data->size = lov_size_to_stripe(data->lsm, data->size,
2042 data->stripe_number);
2044 } else if (keylen >= strlen("last_id") && strcmp(key, "last_id") == 0) {
2046 int rc, size = sizeof(obd_id);
2047 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
2048 if (!lov->tgts[i].active)
2050 rc = obd_get_info(lov->tgts[i].ltd_exp,
2051 keylen, key, &size, &(ids[i]));
2056 } else if (keylen >= strlen("lovdesc") && strcmp(key, "lovdesc") == 0) {
2057 struct lov_desc *desc_ret = val;
2058 *desc_ret = lov->desc;
2066 static int lov_set_info(struct obd_export *exp, obd_count keylen,
2067 void *key, obd_count vallen, void *val)
2069 struct obd_device *obddev = class_exp2obd(exp);
2070 struct lov_obd *lov = &obddev->u.lov;
2074 #define KEY_IS(str) \
2075 (keylen == strlen(str) && memcmp(key, str, keylen) == 0)
2077 if (KEY_IS("next_id")) {
2078 if (vallen != lov->desc.ld_tgt_count)
2080 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
2081 /* initialize all OSCs, even inactive ones */
2083 err = obd_set_info(lov->tgts[i].ltd_exp,
2084 keylen, key, sizeof(obd_id),
2085 ((obd_id*)val) + i);
2091 if (KEY_IS("async")) {
2092 struct lov_desc *desc = &lov->desc;
2093 struct lov_tgt_desc *tgts = lov->tgts;
2095 if (vallen != sizeof(int))
2097 lov->async = *((int*) val);
2099 for (i = 0; i < desc->ld_tgt_count; i++, tgts++) {
2100 struct obd_uuid *tgt_uuid = &tgts->uuid;
2101 struct obd_device *tgt_obd;
2103 tgt_obd = class_find_client_obd(tgt_uuid,
2107 CERROR("Target %s not attached\n",
2114 err = obd_set_info(tgt_obd->obd_self_export,
2115 keylen, key, vallen, val);
2117 CERROR("Failed to set async on target %s\n",
2126 if (KEY_IS("growth_count")) {
2127 if (vallen != sizeof(int))
2129 } else if (KEY_IS("mds_conn")) {
2130 if (vallen != sizeof(__u32))
2132 } else if (KEY_IS("unlinked") || KEY_IS("unrecovery")) {
2135 } else if (KEY_IS("sec")) {
2136 struct lov_tgt_desc *tgt;
2137 struct obd_export *exp;
2140 spin_lock(&lov->lov_lock);
2141 for (i = 0, tgt = lov->tgts; i < lov->desc.ld_tgt_count;
2144 /* during setup time the connections to osc might
2145 * haven't been established.
2148 struct obd_device *tgt_obd;
2150 tgt_obd = class_find_client_obd(&tgt->uuid,
2154 CERROR("can't set security flavor, "
2155 "device %s not attached?\n",
2160 exp = tgt_obd->obd_self_export;
2163 err = obd_set_info(exp, keylen, key, vallen, val);
2167 spin_unlock(&lov->lov_lock);
2174 for (i = 0; i < lov->desc.ld_tgt_count; i++) {
2175 if (val && !obd_uuid_equals(val, &lov->tgts[i].uuid))
2178 if (!val && !lov->tgts[i].active)
2181 err = obd_set_info(lov->tgts[i].ltd_exp,
2182 keylen, key, vallen, val);
2191 struct lov_multi_wait {
2192 struct ldlm_lock *lock;
2198 int lov_complete_many(struct obd_export *exp, struct lov_stripe_md *lsm,
2199 struct lustre_handle *lockh)
2201 struct lov_lock_handles *lov_lockh = NULL;
2202 struct lustre_handle *lov_lockhp;
2203 struct lov_obd *lov;
2204 struct lov_oinfo *loi;
2205 struct lov_multi_wait *queues;
2209 if (lsm_bad_magic(lsm))
2212 if (!exp || !exp->exp_obd)
2215 LASSERT(lockh != NULL);
2216 if (lsm->lsm_stripe_count > 1) {
2217 lov_lockh = lov_handle2llh(lockh);
2218 if (lov_lockh == NULL) {
2219 CERROR("LOV: invalid lov lock handle %p\n", lockh);
2223 lov_lockhp = lov_lockh->llh_handles;
2228 OBD_ALLOC(queues, lsm->lsm_stripe_count * sizeof(*queues));
2230 GOTO(out, rc = -ENOMEM);
2232 lov = &exp->exp_obd->u.lov;
2233 for (i = 0, loi = lsm->lsm_oinfo; i < lsm->lsm_stripe_count;
2234 i++, loi++, lov_lockhp++) {
2235 struct ldlm_lock *lock;
2236 struct obd_device *obd;
2237 unsigned long irqflags;
2239 lock = ldlm_handle2lock(lov_lockhp);
2241 CDEBUG(D_HA, "lov idx %d subobj "LPX64" no lock?\n",
2242 loi->loi_ost_idx, loi->loi_id);
2243 queues[i].completed = 1;
2247 queues[i].lock = lock;
2248 init_waitqueue_entry(&(queues[i].wait), current);
2249 add_wait_queue(lock->l_waitq, &(queues[i].wait));
2251 obd = class_exp2obd(lock->l_conn_export);
2253 imp = obd->u.cli.cl_import;
2255 spin_lock_irqsave(&imp->imp_lock, irqflags);
2256 queues[i].generation = imp->imp_generation;
2257 spin_unlock_irqrestore(&imp->imp_lock, irqflags);
2261 lwi = LWI_TIMEOUT_INTR(obd_timeout * HZ, ldlm_expired_completion_wait,
2262 interrupted_completion_wait, &lwd);
2263 rc = l_wait_event_added(check_multi_complete(queues, lsm), &lwi);
2265 for (i = 0; i < lsm->lsm_stripe_count; i++)
2266 remove_wait_queue(lock->l_waitq, &(queues[i].wait));
2268 if (rc == -EINTR || rc == -ETIMEDOUT) {
2274 if (lov_lockh != NULL)
2275 lov_llh_put(lov_lockh);
2280 struct obd_ops lov_obd_ops = {
2281 .o_owner = THIS_MODULE,
2282 .o_attach = lov_attach,
2283 .o_detach = lov_detach,
2284 .o_setup = lov_setup,
2285 .o_cleanup = lov_cleanup,
2286 .o_process_config = lov_process_config,
2287 .o_connect = lov_connect,
2288 .o_disconnect = lov_disconnect,
2289 .o_statfs = lov_statfs,
2290 .o_packmd = lov_packmd,
2291 .o_unpackmd = lov_unpackmd,
2292 .o_revalidate_md = lov_revalidate_md,
2293 .o_create = lov_create,
2294 .o_destroy = lov_destroy,
2295 .o_getattr = lov_getattr,
2296 .o_getattr_async = lov_getattr_async,
2297 .o_setattr = lov_setattr,
2299 .o_brw_async = lov_brw_async,
2300 .o_prep_async_page = lov_prep_async_page,
2301 .o_queue_async_io = lov_queue_async_io,
2302 .o_set_async_flags = lov_set_async_flags,
2303 .o_queue_group_io = lov_queue_group_io,
2304 .o_trigger_group_io = lov_trigger_group_io,
2305 .o_teardown_async_page = lov_teardown_async_page,
2306 .o_adjust_kms = lov_adjust_kms,
2307 .o_punch = lov_punch,
2309 .o_enqueue = lov_enqueue,
2310 .o_match = lov_match,
2311 .o_change_cbdata = lov_change_cbdata,
2312 .o_cancel = lov_cancel,
2313 .o_cancel_unused = lov_cancel_unused,
2314 .o_iocontrol = lov_iocontrol,
2315 .o_get_info = lov_get_info,
2316 .o_set_info = lov_set_info,
2317 .o_llog_init = lov_llog_init,
2318 .o_llog_finish = lov_llog_finish,
2319 .o_notify = lov_notify,
2322 int __init lov_init(void)
2324 struct lprocfs_static_vars lvars;
2328 lprocfs_init_vars(lov, &lvars);
2329 rc = class_register_type(&lov_obd_ops, NULL, lvars.module_vars,
2330 OBD_LOV_DEVICENAME);
2335 static void /*__exit*/ lov_exit(void)
2337 class_unregister_type(OBD_LOV_DEVICENAME);
2340 MODULE_AUTHOR("Cluster File Systems, Inc. <info@clusterfs.com>");
2341 MODULE_DESCRIPTION("Lustre Logical Object Volume OBD driver");
2342 MODULE_LICENSE("GPL");
2344 module_init(lov_init);
2345 module_exit(lov_exit);