4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License version 2 only,
8 * as published by the Free Software Foundation.
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.gnu.org/licenses/gpl-2.0.html
23 * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
24 * Use is subject to license terms.
26 * Copyright (c) 2012, 2016, Intel Corporation.
29 * This file is part of Lustre, http://www.lustre.org/
30 * Lustre is a trademark of Sun Microsystems, Inc.
32 * lustre/osp/osp_sync.c
34 * Lustre OST Proxy Device
36 * Author: Alex Zhuravlev <alexey.zhuravlev@intel.com>
37 * Author: Mikhail Pershin <mike.pershin@intel.com>
40 #define DEBUG_SUBSYSTEM S_MDS
42 #include <linux/kthread.h>
43 #include <lustre_log.h>
44 #include <lustre_update.h>
45 #include "osp_internal.h"
47 static int osp_sync_id_traction_init(struct osp_device *d);
48 static void osp_sync_id_traction_fini(struct osp_device *d);
49 static __u64 osp_sync_id_get(struct osp_device *d, __u64 id);
50 static void osp_sync_remove_from_tracker(struct osp_device *d);
53 * this is a components of OSP implementing synchronization between MDS and OST
54 * it llogs all interesting changes (currently it's uig/gid change and object
55 * destroy) atomically, then makes sure changes hit OST storage
57 * we have 4 queues of work:
59 * the first queue is llog itself, once read a change is stored in 2nd queue
60 * in form of RPC (but RPC isn't fired yet).
62 * the second queue (opd_syn_waiting_for_commit) holds changes awaiting local
63 * commit. once change is committed locally it migrates onto 3rd queue.
65 * the third queue (opd_syn_committed_here) holds changes committed locally,
66 * but not sent to OST (as the pipe can be full). once pipe becomes non-full
67 * we take a change from the queue and fire corresponded RPC.
69 * once RPC is reported committed by OST (using regular last_committed mech.)
70 * the change jumps into 4th queue (opd_syn_committed_there), now we can
71 * cancel corresponded llog record and release RPC
73 * opd_syn_changes is a number of unread llog records (to be processed).
74 * notice this number doesn't include llog records from previous boots.
75 * with OSP_SYN_THRESHOLD we try to batch processing a bit (TO BE IMPLEMENTED)
77 * opd_syn_rpc_in_progress is a number of requests in 2-4 queues.
78 * we control this with OSP_MAX_IN_PROGRESS so that OSP don't consume
79 * too much memory -- how to deal with 1000th OSTs ? batching could help?
81 * opd_syn_rpc_in_flight is a number of RPC in flight.
82 * we control this with OSP_MAX_IN_FLIGHT
85 /* XXX: do math to learn reasonable threshold
86 * should it be ~ number of changes fitting bulk? */
88 #define OSP_SYN_THRESHOLD 10
89 #define OSP_MAX_IN_FLIGHT 8
90 #define OSP_MAX_IN_PROGRESS 4096
92 #define OSP_JOB_MAGIC 0x26112005
94 struct osp_job_req_args {
95 /** bytes reserved for ptlrpc_replay_req() */
96 struct ptlrpc_replay_async_args jra_raa;
97 struct list_head jra_committed_link;
98 struct list_head jra_inflight_link;
99 struct llog_cookie jra_lcookie;
103 static inline int osp_sync_running(struct osp_device *d)
105 return !!(d->opd_syn_thread.t_flags & SVC_RUNNING);
109 * Check status: whether OSP thread has stopped
111 * \param[in] d OSP device
113 * \retval 0 still running
116 static inline int osp_sync_stopped(struct osp_device *d)
118 return !!(d->opd_syn_thread.t_flags & SVC_STOPPED);
122 ** Check for new changes to sync
124 * \param[in] d OSP device
126 * \retval 1 there are changes
127 * \retval 0 there are no changes
129 static inline int osp_sync_has_new_job(struct osp_device *d)
131 return ((d->opd_syn_last_processed_id < d->opd_syn_last_used_id) &&
132 (d->opd_syn_last_processed_id < d->opd_syn_last_committed_id))
133 || (d->opd_syn_prev_done == 0);
136 static inline int osp_sync_inflight_conflict(struct osp_device *d,
137 struct llog_rec_hdr *h)
139 struct osp_job_req_args *jra;
143 if (h == NULL || h->lrh_type == LLOG_GEN_REC ||
144 list_empty(&d->opd_syn_inflight_list))
147 memset(&ostid, 0, sizeof(ostid));
148 switch (h->lrh_type) {
150 ostid_set_seq(&ostid, ((struct llog_unlink_rec *)h)->lur_oseq);
151 ostid_set_id(&ostid, ((struct llog_unlink_rec *)h)->lur_oid);
153 case MDS_UNLINK64_REC:
154 fid_to_ostid(&((struct llog_unlink64_rec *)h)->lur_fid, &ostid);
156 case MDS_SETATTR64_REC:
157 ostid = ((struct llog_setattr64_rec *)h)->lsr_oi;
163 spin_lock(&d->opd_syn_lock);
164 list_for_each_entry(jra, &d->opd_syn_inflight_list, jra_inflight_link) {
165 struct ptlrpc_request *req;
166 struct ost_body *body;
168 LASSERT(jra->jra_magic == OSP_JOB_MAGIC);
170 req = container_of((void *)jra, struct ptlrpc_request,
172 body = req_capsule_client_get(&req->rq_pill,
176 if (memcmp(&ostid, &body->oa.o_oi, sizeof(ostid)) == 0) {
181 spin_unlock(&d->opd_syn_lock);
186 static inline int osp_sync_low_in_progress(struct osp_device *d)
188 return atomic_read(&d->opd_syn_rpc_in_progress) <
189 d->opd_syn_max_rpc_in_progress;
193 * Check for room in the network pipe to OST
195 * \param[in] d OSP device
197 * \retval 1 there is room
198 * \retval 0 no room, the pipe is full
200 static inline int osp_sync_low_in_flight(struct osp_device *d)
202 return atomic_read(&d->opd_syn_rpc_in_flight) <
203 d->opd_syn_max_rpc_in_flight;
207 * Wake up check for the main sync thread
209 * \param[in] d OSP device
211 * \retval 1 time to wake up
212 * \retval 0 no need to wake up
214 static inline int osp_sync_has_work(struct osp_device *d)
216 /* has new/old changes and low in-progress? */
217 if (osp_sync_has_new_job(d) && osp_sync_low_in_progress(d) &&
218 osp_sync_low_in_flight(d) && d->opd_imp_connected)
221 /* has remotely committed? */
222 if (!list_empty(&d->opd_syn_committed_there))
228 #define osp_sync_check_for_work(d) \
230 if (osp_sync_has_work(d)) { \
231 wake_up(&d->opd_syn_waitq); \
235 void __osp_sync_check_for_work(struct osp_device *d)
237 osp_sync_check_for_work(d);
240 static inline __u64 osp_sync_correct_id(struct osp_device *d,
241 struct llog_rec_hdr *rec)
244 * llog use cyclic store with 32 bit lrh_id
245 * so overflow lrh_id is possible. Range between
246 * last_processed and last_committed is less than
247 * 64745 ^ 2 and less than 2^32 - 1
249 __u64 correct_id = d->opd_syn_last_committed_id;
251 if ((correct_id & 0xffffffffULL) < rec->lrh_id)
252 correct_id -= 0x100000000ULL;
254 correct_id &= ~0xffffffffULL;
255 correct_id |= rec->lrh_id;
260 * Check and return ready-for-new status.
262 * The thread processing llog record uses this function to check whether
263 * it's time to take another record and process it. The number of conditions
264 * must be met: the connection should be ready, RPCs in flight not exceeding
265 * the limit, the record is committed locally, etc (see the lines below).
267 * \param[in] d OSP device
268 * \param[in] rec next llog record to process
270 * \retval 0 not ready
273 static inline int osp_sync_can_process_new(struct osp_device *d,
274 struct llog_rec_hdr *rec)
278 if (unlikely(atomic_read(&d->opd_syn_barrier) > 0))
280 if (unlikely(osp_sync_inflight_conflict(d, rec)))
282 if (!osp_sync_low_in_progress(d))
284 if (!osp_sync_low_in_flight(d))
286 if (!d->opd_imp_connected)
288 if (d->opd_syn_prev_done == 0)
290 if (atomic_read(&d->opd_syn_changes) == 0)
293 osp_sync_correct_id(d, rec) <= d->opd_syn_last_committed_id)
299 * Declare intention to add a new change.
301 * With regard to OSD API, we have to declare any changes ahead. In this
302 * case we declare an intention to add a llog record representing the
303 * change on the local storage.
305 * \param[in] env LU environment provided by the caller
306 * \param[in] o OSP object
307 * \param[in] type type of change: MDS_UNLINK64_REC or MDS_SETATTR64_REC
308 * \param[in] th transaction handle (local)
310 * \retval 0 on success
311 * \retval negative negated errno on error
313 int osp_sync_declare_add(const struct lu_env *env, struct osp_object *o,
314 llog_op_type type, struct thandle *th)
316 struct osp_thread_info *osi = osp_env_info(env);
317 struct osp_device *d = lu2osp_dev(o->opo_obj.do_lu.lo_dev);
318 struct llog_ctxt *ctxt;
319 struct thandle *storage_th;
324 /* it's a layering violation, to access internals of th,
325 * but we can do this as a sanity check, for a while */
326 LASSERT(th->th_top != NULL);
327 storage_th = thandle_get_sub_by_dt(env, th->th_top, d->opd_storage);
328 if (IS_ERR(storage_th))
329 RETURN(PTR_ERR(storage_th));
332 case MDS_UNLINK64_REC:
333 osi->osi_hdr.lrh_len = sizeof(struct llog_unlink64_rec);
335 case MDS_SETATTR64_REC:
336 osi->osi_hdr.lrh_len = sizeof(struct llog_setattr64_rec_v2);
342 /* we want ->dt_trans_start() to allocate per-thandle structure */
343 storage_th->th_tags |= LCT_OSP_THREAD;
345 ctxt = llog_get_context(d->opd_obd, LLOG_MDS_OST_ORIG_CTXT);
348 rc = llog_declare_add(env, ctxt->loc_handle, &osi->osi_hdr,
356 * Generate a llog record for a given change.
358 * Generates a llog record for the change passed. The change can be of two
359 * types: unlink and setattr. The record gets an ID which later will be
360 * used to track commit status of the change. For unlink changes, the caller
361 * can supply a starting FID and the count of the objects to destroy. For
362 * setattr the caller should apply attributes to apply.
365 * \param[in] env LU environment provided by the caller
366 * \param[in] d OSP device
367 * \param[in] fid fid of the object the change should be applied to
368 * \param[in] type type of change: MDS_UNLINK64_REC or MDS_SETATTR64_REC
369 * \param[in] count count of objects to destroy
370 * \param[in] th transaction handle (local)
371 * \param[in] attr attributes for setattr
373 * \retval 0 on success
374 * \retval negative negated errno on error
376 static int osp_sync_add_rec(const struct lu_env *env, struct osp_device *d,
377 const struct lu_fid *fid, llog_op_type type,
378 int count, struct thandle *th,
379 const struct lu_attr *attr)
381 struct osp_thread_info *osi = osp_env_info(env);
382 struct llog_ctxt *ctxt;
383 struct osp_txn_info *txn;
384 struct thandle *storage_th;
389 /* it's a layering violation, to access internals of th,
390 * but we can do this as a sanity check, for a while */
391 LASSERT(th->th_top != NULL);
392 storage_th = thandle_get_sub_by_dt(env, th->th_top, d->opd_storage);
393 if (IS_ERR(storage_th))
394 RETURN(PTR_ERR(storage_th));
397 case MDS_UNLINK64_REC:
398 osi->osi_hdr.lrh_len = sizeof(osi->osi_unlink);
399 osi->osi_hdr.lrh_type = MDS_UNLINK64_REC;
400 osi->osi_unlink.lur_fid = *fid;
401 osi->osi_unlink.lur_count = count;
403 case MDS_SETATTR64_REC:
404 rc = fid_to_ostid(fid, &osi->osi_oi);
406 osi->osi_hdr.lrh_len = sizeof(osi->osi_setattr);
407 osi->osi_hdr.lrh_type = MDS_SETATTR64_REC;
408 osi->osi_setattr.lsr_oi = osi->osi_oi;
410 osi->osi_setattr.lsr_uid = attr->la_uid;
411 osi->osi_setattr.lsr_gid = attr->la_gid;
412 osi->osi_setattr.lsr_projid = attr->la_projid;
413 osi->osi_setattr.lsr_valid =
414 ((attr->la_valid & LA_UID) ? OBD_MD_FLUID : 0) |
415 ((attr->la_valid & LA_GID) ? OBD_MD_FLGID : 0) |
416 ((attr->la_valid & LA_PROJID) ? OBD_MD_FLPROJID : 0);
422 txn = osp_txn_info(&storage_th->th_ctx);
425 txn->oti_current_id = osp_sync_id_get(d, txn->oti_current_id);
426 osi->osi_hdr.lrh_id = (txn->oti_current_id & 0xffffffffULL);
427 ctxt = llog_get_context(d->opd_obd, LLOG_MDS_OST_ORIG_CTXT);
431 rc = llog_add(env, ctxt->loc_handle, &osi->osi_hdr, &osi->osi_cookie,
435 if (likely(rc >= 0)) {
436 CDEBUG(D_OTHER, "%s: new record "DOSTID":%lu/%lu: %d\n",
437 d->opd_obd->obd_name,
438 POSTID(&osi->osi_cookie.lgc_lgl.lgl_oi),
439 (unsigned long)osi->osi_cookie.lgc_lgl.lgl_ogen,
440 (unsigned long)osi->osi_cookie.lgc_index, rc);
441 atomic_inc(&d->opd_syn_changes);
443 /* return 0 always here, error case just cause no llog record */
447 int osp_sync_add(const struct lu_env *env, struct osp_object *o,
448 llog_op_type type, struct thandle *th,
449 const struct lu_attr *attr)
451 return osp_sync_add_rec(env, lu2osp_dev(o->opo_obj.do_lu.lo_dev),
452 lu_object_fid(&o->opo_obj.do_lu), type, 1,
456 int osp_sync_gap(const struct lu_env *env, struct osp_device *d,
457 struct lu_fid *fid, int lost, struct thandle *th)
459 return osp_sync_add_rec(env, d, fid, MDS_UNLINK64_REC, lost, th, NULL);
463 * it's quite obvious we can't maintain all the structures in the memory:
464 * while OST is down, MDS can be processing thousands and thousands of unlinks
465 * filling persistent llogs and in-core respresentation
467 * this doesn't scale at all. so we need basically the following:
468 * a) destroy/setattr append llog records
469 * b) once llog has grown to X records, we process first Y committed records
471 * once record R is found via llog_process(), it becomes committed after any
472 * subsequent commit callback (at the most)
476 * ptlrpc commit callback.
478 * The callback is called by PTLRPC when a RPC is reported committed by the
479 * target (OST). We register the callback for the every RPC applying a change
480 * from the llog. This way we know then the llog records can be cancelled.
481 * Notice the callback can be called when OSP is finishing. We can detect this
482 * checking that actual transno in the request is less or equal of known
483 * committed transno (see osp_sync_process_committed() for the details).
484 * XXX: this is pretty expensive and can be improved later using batching.
486 * \param[in] req request
488 static void osp_sync_request_commit_cb(struct ptlrpc_request *req)
490 struct osp_device *d = req->rq_cb_data;
491 struct osp_job_req_args *jra;
493 CDEBUG(D_HA, "commit req %p, transno %llu\n", req, req->rq_transno);
495 if (unlikely(req->rq_transno == 0))
498 /* do not do any opd_dyn_rpc_* accounting here
499 * it's done in osp_sync_interpret sooner or later */
502 jra = ptlrpc_req_async_args(req);
503 LASSERT(jra->jra_magic == OSP_JOB_MAGIC);
504 LASSERT(list_empty(&jra->jra_committed_link));
506 ptlrpc_request_addref(req);
508 spin_lock(&d->opd_syn_lock);
509 list_add(&jra->jra_committed_link, &d->opd_syn_committed_there);
510 spin_unlock(&d->opd_syn_lock);
512 /* XXX: some batching wouldn't hurt */
513 wake_up(&d->opd_syn_waitq);
517 * RPC interpretation callback.
519 * The callback is called by ptlrpc when RPC is replied. Now we have to decide
521 * - put request on a special list to wait until it's committed by the target,
522 * if the request is successful
523 * - schedule llog record cancel if no target object is found
524 * - try later (essentially after reboot) in case of unexpected error
526 * \param[in] env LU environment provided by the caller
527 * \param[in] req request replied
528 * \param[in] aa callback data
529 * \param[in] rc result of RPC
533 static int osp_sync_interpret(const struct lu_env *env,
534 struct ptlrpc_request *req, void *aa, int rc)
536 struct osp_device *d = req->rq_cb_data;
537 struct osp_job_req_args *jra = aa;
539 if (jra->jra_magic != OSP_JOB_MAGIC) {
540 DEBUG_REQ(D_ERROR, req, "bad magic %u\n", jra->jra_magic);
545 CDEBUG(D_HA, "reply req %p/%d, rc %d, transno %u\n", req,
546 atomic_read(&req->rq_refcount),
547 rc, (unsigned) req->rq_transno);
548 LASSERT(rc || req->rq_transno);
552 * we tried to destroy object or update attributes,
553 * but object doesn't exist anymore - cancell llog record
555 LASSERT(req->rq_transno == 0);
556 LASSERT(list_empty(&jra->jra_committed_link));
558 ptlrpc_request_addref(req);
560 spin_lock(&d->opd_syn_lock);
561 list_add(&jra->jra_committed_link, &d->opd_syn_committed_there);
562 spin_unlock(&d->opd_syn_lock);
564 wake_up(&d->opd_syn_waitq);
566 struct obd_import *imp = req->rq_import;
568 * error happened, we'll try to repeat on next boot ?
570 LASSERTF(req->rq_transno == 0 ||
571 req->rq_import_generation < imp->imp_generation,
572 "transno %llu, rc %d, gen: req %d, imp %d\n",
573 req->rq_transno, rc, req->rq_import_generation,
574 imp->imp_generation);
575 if (req->rq_transno == 0) {
576 /* this is the last time we see the request
577 * if transno is not zero, then commit cb
578 * will be called at some point */
579 LASSERT(atomic_read(&d->opd_syn_rpc_in_progress) > 0);
580 atomic_dec(&d->opd_syn_rpc_in_progress);
583 wake_up(&d->opd_syn_waitq);
584 } else if (d->opd_pre != NULL &&
585 unlikely(d->opd_pre_status == -ENOSPC)) {
587 * if current status is -ENOSPC (lack of free space on OST)
588 * then we should poll OST immediately once object destroy
591 osp_statfs_need_now(d);
594 spin_lock(&d->opd_syn_lock);
595 list_del_init(&jra->jra_inflight_link);
596 spin_unlock(&d->opd_syn_lock);
597 LASSERT(atomic_read(&d->opd_syn_rpc_in_flight) > 0);
598 atomic_dec(&d->opd_syn_rpc_in_flight);
599 if (unlikely(atomic_read(&d->opd_syn_barrier) > 0))
600 wake_up(&d->opd_syn_barrier_waitq);
601 CDEBUG(D_OTHER, "%s: %d in flight, %d in progress\n",
602 d->opd_obd->obd_name, atomic_read(&d->opd_syn_rpc_in_flight),
603 atomic_read(&d->opd_syn_rpc_in_progress));
605 osp_sync_check_for_work(d);
611 ** Add request to ptlrpc queue.
613 * This is just a tiny helper function to put the request on the sending list
615 * \param[in] d OSP device
616 * \param[in] llh llog handle where the record is stored
617 * \param[in] h llog record
618 * \param[in] req request
620 static void osp_sync_send_new_rpc(struct osp_device *d,
621 struct llog_handle *llh,
622 struct llog_rec_hdr *h,
623 struct ptlrpc_request *req)
625 struct osp_job_req_args *jra;
627 LASSERT(atomic_read(&d->opd_syn_rpc_in_flight) <=
628 d->opd_syn_max_rpc_in_flight);
630 jra = ptlrpc_req_async_args(req);
631 jra->jra_magic = OSP_JOB_MAGIC;
632 jra->jra_lcookie.lgc_lgl = llh->lgh_id;
633 jra->jra_lcookie.lgc_subsys = LLOG_MDS_OST_ORIG_CTXT;
634 jra->jra_lcookie.lgc_index = h->lrh_index;
635 INIT_LIST_HEAD(&jra->jra_committed_link);
636 spin_lock(&d->opd_syn_lock);
637 list_add_tail(&jra->jra_inflight_link, &d->opd_syn_inflight_list);
638 spin_unlock(&d->opd_syn_lock);
640 ptlrpcd_add_req(req);
645 * Allocate and prepare RPC for a new change.
647 * The function allocates and initializes an RPC which will be sent soon to
648 * apply the change to the target OST. The request is initialized from the
649 * llog record passed. Notice only the fields common to all type of changes
652 * \param[in] d OSP device
653 * \param[in] op type of the change
654 * \param[in] format request format to be used
656 * \retval pointer new request on success
657 * \retval ERR_PTR(errno) on error
659 static struct ptlrpc_request *osp_sync_new_job(struct osp_device *d,
661 const struct req_format *format)
663 struct ptlrpc_request *req;
664 struct obd_import *imp;
667 /* Prepare the request */
668 imp = d->opd_obd->u.cli.cl_import;
671 if (OBD_FAIL_CHECK(OBD_FAIL_OSP_CHECK_ENOMEM))
672 RETURN(ERR_PTR(-ENOMEM));
674 req = ptlrpc_request_alloc(imp, format);
676 RETURN(ERR_PTR(-ENOMEM));
678 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, op);
680 ptlrpc_req_finished(req);
684 req->rq_interpret_reply = osp_sync_interpret;
685 req->rq_commit_cb = osp_sync_request_commit_cb;
688 ptlrpc_request_set_replen(req);
694 * Generate a request for setattr change.
696 * The function prepares a new RPC, initializes it with setattr specific
697 * bits and send the RPC.
699 * \param[in] d OSP device
700 * \param[in] llh llog handle where the record is stored
701 * \param[in] h llog record
703 * \retval 0 on success
704 * \retval 1 on invalid record
705 * \retval negative negated errno on error
707 static int osp_sync_new_setattr_job(struct osp_device *d,
708 struct llog_handle *llh,
709 struct llog_rec_hdr *h)
711 struct llog_setattr64_rec *rec = (struct llog_setattr64_rec *)h;
712 struct ptlrpc_request *req;
713 struct ost_body *body;
716 LASSERT(h->lrh_type == MDS_SETATTR64_REC);
718 if (OBD_FAIL_CHECK(OBD_FAIL_OSP_CHECK_INVALID_REC))
721 /* lsr_valid can only be 0 or HAVE OBD_MD_{FLUID, FLGID, FLPROJID} set,
722 * so no bits other than these should be set. */
723 if ((rec->lsr_valid & ~(OBD_MD_FLUID | OBD_MD_FLGID |
724 OBD_MD_FLPROJID)) != 0) {
725 CERROR("%s: invalid setattr record, lsr_valid:%llu\n",
726 d->opd_obd->obd_name, rec->lsr_valid);
727 /* return 1 on invalid record */
731 req = osp_sync_new_job(d, OST_SETATTR, &RQF_OST_SETATTR);
733 RETURN(PTR_ERR(req));
735 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
737 body->oa.o_oi = rec->lsr_oi;
738 body->oa.o_uid = rec->lsr_uid;
739 body->oa.o_gid = rec->lsr_gid;
740 body->oa.o_valid = OBD_MD_FLGROUP | OBD_MD_FLID;
741 if (h->lrh_len > sizeof(struct llog_setattr64_rec))
742 body->oa.o_projid = ((struct llog_setattr64_rec_v2 *)
745 /* old setattr record (prior 2.6.0) doesn't have 'valid' stored,
746 * we assume that both UID and GID are valid in that case. */
747 if (rec->lsr_valid == 0)
748 body->oa.o_valid |= (OBD_MD_FLUID | OBD_MD_FLGID);
750 body->oa.o_valid |= rec->lsr_valid;
752 osp_sync_send_new_rpc(d, llh, h, req);
757 * Generate a request for unlink change.
759 * The function prepares a new RPC, initializes it with unlink(destroy)
760 * specific bits and sends the RPC. The function is used to handle
761 * llog_unlink_rec which were used in the older versions of Lustre.
762 * Current version uses llog_unlink_rec64.
764 * \param[in] d OSP device
765 * \param[in] llh llog handle where the record is stored
766 * \param[in] h llog record
768 * \retval 0 on success
769 * \retval negative negated errno on error
771 static int osp_sync_new_unlink_job(struct osp_device *d,
772 struct llog_handle *llh,
773 struct llog_rec_hdr *h)
775 struct llog_unlink_rec *rec = (struct llog_unlink_rec *)h;
776 struct ptlrpc_request *req;
777 struct ost_body *body;
780 LASSERT(h->lrh_type == MDS_UNLINK_REC);
782 req = osp_sync_new_job(d, OST_DESTROY, &RQF_OST_DESTROY);
784 RETURN(PTR_ERR(req));
786 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
788 ostid_set_seq(&body->oa.o_oi, rec->lur_oseq);
789 ostid_set_id(&body->oa.o_oi, rec->lur_oid);
790 body->oa.o_misc = rec->lur_count;
791 body->oa.o_valid = OBD_MD_FLGROUP | OBD_MD_FLID;
793 body->oa.o_valid |= OBD_MD_FLOBJCOUNT;
795 osp_sync_send_new_rpc(d, llh, h, req);
800 * Generate a request for unlink change.
802 * The function prepares a new RPC, initializes it with unlink(destroy)
803 * specific bits and sends the RPC. Depending on the target (MDT or OST)
804 * two different protocols are used. For MDT we use OUT (basically OSD API
805 * updates transferred via a network). For OST we still use the old
806 * protocol (OBD?), originally for compatibility. Later we can start to
807 * use OUT for OST as well, this will allow batching and better code
810 * \param[in] d OSP device
811 * \param[in] llh llog handle where the record is stored
812 * \param[in] h llog record
814 * \retval 0 on success
815 * \retval negative negated errno on error
817 static int osp_sync_new_unlink64_job(struct osp_device *d,
818 struct llog_handle *llh,
819 struct llog_rec_hdr *h)
821 struct llog_unlink64_rec *rec = (struct llog_unlink64_rec *)h;
822 struct ptlrpc_request *req = NULL;
823 struct ost_body *body;
827 LASSERT(h->lrh_type == MDS_UNLINK64_REC);
828 req = osp_sync_new_job(d, OST_DESTROY, &RQF_OST_DESTROY);
830 RETURN(PTR_ERR(req));
832 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
835 rc = fid_to_ostid(&rec->lur_fid, &body->oa.o_oi);
838 body->oa.o_misc = rec->lur_count;
839 body->oa.o_valid = OBD_MD_FLGROUP | OBD_MD_FLID |
841 osp_sync_send_new_rpc(d, llh, h, req);
846 * Process llog records.
848 * This function is called to process the llog records committed locally.
849 * In the recovery model used by OSP we can apply a change to a remote
850 * target once corresponding transaction (like posix unlink) is committed
851 * locally so can't revert.
852 * Depending on the llog record type, a given handler is called that is
853 * responsible for preparing and sending the RPC to apply the change.
854 * Special record type LLOG_GEN_REC marking a reboot is cancelled right away.
856 * \param[in] env LU environment provided by the caller
857 * \param[in] d OSP device
858 * \param[in] llh llog handle where the record is stored
859 * \param[in] rec llog record
861 static void osp_sync_process_record(const struct lu_env *env,
862 struct osp_device *d,
863 struct llog_handle *llh,
864 struct llog_rec_hdr *rec)
866 struct llog_handle *cathandle = llh->u.phd.phd_cat_handle;
867 struct llog_cookie cookie;
872 cookie.lgc_lgl = llh->lgh_id;
873 cookie.lgc_subsys = LLOG_MDS_OST_ORIG_CTXT;
874 cookie.lgc_index = rec->lrh_index;
876 if (unlikely(rec->lrh_type == LLOG_GEN_REC)) {
877 struct llog_gen_rec *gen = (struct llog_gen_rec *)rec;
879 /* we're waiting for the record generated by this instance */
880 LASSERT(d->opd_syn_prev_done == 0);
881 if (!memcmp(&d->opd_syn_generation, &gen->lgr_gen,
882 sizeof(gen->lgr_gen))) {
883 CDEBUG(D_HA, "processed all old entries\n");
884 d->opd_syn_prev_done = 1;
887 /* cancel any generation record */
888 rc = llog_cat_cancel_records(env, cathandle, 1, &cookie);
894 * now we prepare and fill requests to OST, put them on the queue
895 * and fire after next commit callback
898 /* notice we increment counters before sending RPC, to be consistent
899 * in RPC interpret callback which may happen very quickly */
900 atomic_inc(&d->opd_syn_rpc_in_flight);
901 atomic_inc(&d->opd_syn_rpc_in_progress);
903 switch (rec->lrh_type) {
904 /* case MDS_UNLINK_REC is kept for compatibility */
906 rc = osp_sync_new_unlink_job(d, llh, rec);
908 case MDS_UNLINK64_REC:
909 rc = osp_sync_new_unlink64_job(d, llh, rec);
911 case MDS_SETATTR64_REC:
912 rc = osp_sync_new_setattr_job(d, llh, rec);
915 CERROR("%s: unknown record type: %x\n", d->opd_obd->obd_name,
917 /* treat "unknown record type" as "invalid" */
922 /* For all kinds of records, not matter successful or not,
923 * we should decrease changes and bump last_processed_id.
925 if (d->opd_syn_prev_done) {
926 __u64 correct_id = osp_sync_correct_id(d, rec);
927 LASSERT(atomic_read(&d->opd_syn_changes) > 0);
928 LASSERT(correct_id <= d->opd_syn_last_committed_id);
929 /* NOTE: it's possible to meet same id if
930 * OST stores few stripes of same file
933 /* another thread may be trying to set new value */
935 if (correct_id > d->opd_syn_last_processed_id) {
936 d->opd_syn_last_processed_id = correct_id;
937 wake_up(&d->opd_syn_barrier_waitq);
941 atomic_dec(&d->opd_syn_changes);
944 atomic_dec(&d->opd_syn_rpc_in_flight);
945 atomic_dec(&d->opd_syn_rpc_in_progress);
948 CDEBUG(D_OTHER, "%s: %d in flight, %d in progress\n",
949 d->opd_obd->obd_name, atomic_read(&d->opd_syn_rpc_in_flight),
950 atomic_read(&d->opd_syn_rpc_in_progress));
952 /* Delete the invalid record */
954 rc = llog_cat_cancel_records(env, cathandle, 1, &cookie);
956 CERROR("%s: can't delete invalid record: "
957 "fid = "DFID", rec_id = %u, rc = %d\n",
958 d->opd_obd->obd_name,
959 PFID(lu_object_fid(&cathandle->lgh_obj->do_lu)),
963 CDEBUG(D_HA, "found record %x, %d, idx %u, id %u\n",
964 rec->lrh_type, rec->lrh_len, rec->lrh_index, rec->lrh_id);
970 * Cancel llog records for the committed changes.
972 * The function walks through the list of the committed RPCs and cancels
973 * corresponding llog records. see osp_sync_request_commit_cb() for the
976 * \param[in] env LU environment provided by the caller
977 * \param[in] d OSP device
979 static void osp_sync_process_committed(const struct lu_env *env,
980 struct osp_device *d)
982 struct obd_device *obd = d->opd_obd;
983 struct obd_import *imp = obd->u.cli.cl_import;
984 struct ost_body *body;
985 struct ptlrpc_request *req;
986 struct llog_ctxt *ctxt;
987 struct llog_handle *llh;
988 struct list_head list;
993 if (list_empty(&d->opd_syn_committed_there))
997 * if current status is -ENOSPC (lack of free space on OST)
998 * then we should poll OST immediately once object destroy
1000 * notice: we do this upon commit as well because some backends
1001 * (like DMU) do not release space right away.
1003 if (d->opd_pre != NULL && unlikely(d->opd_pre_status == -ENOSPC))
1004 osp_statfs_need_now(d);
1007 * now cancel them all
1008 * XXX: can we improve this using some batching?
1009 * with batch RPC that'll happen automatically?
1010 * XXX: can we store ctxt in lod_device and save few cycles ?
1012 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
1015 llh = ctxt->loc_handle;
1018 INIT_LIST_HEAD(&list);
1019 spin_lock(&d->opd_syn_lock);
1020 list_splice(&d->opd_syn_committed_there, &list);
1021 INIT_LIST_HEAD(&d->opd_syn_committed_there);
1022 spin_unlock(&d->opd_syn_lock);
1024 while (!list_empty(&list)) {
1025 struct osp_job_req_args *jra;
1027 jra = list_entry(list.next, struct osp_job_req_args,
1028 jra_committed_link);
1029 LASSERT(jra->jra_magic == OSP_JOB_MAGIC);
1030 list_del_init(&jra->jra_committed_link);
1032 req = container_of((void *)jra, struct ptlrpc_request,
1034 body = req_capsule_client_get(&req->rq_pill,
1037 /* import can be closing, thus all commit cb's are
1038 * called we can check committness directly */
1039 if (req->rq_import_generation == imp->imp_generation) {
1040 rc = llog_cat_cancel_records(env, llh, 1,
1043 CERROR("%s: can't cancel record: %d\n",
1046 DEBUG_REQ(D_OTHER, req, "imp_committed = %llu",
1047 imp->imp_peer_committed_transno);
1049 ptlrpc_req_finished(req);
1053 llog_ctxt_put(ctxt);
1055 LASSERT(atomic_read(&d->opd_syn_rpc_in_progress) >= done);
1056 atomic_sub(done, &d->opd_syn_rpc_in_progress);
1057 CDEBUG(D_OTHER, "%s: %d in flight, %d in progress\n",
1058 d->opd_obd->obd_name, atomic_read(&d->opd_syn_rpc_in_flight),
1059 atomic_read(&d->opd_syn_rpc_in_progress));
1061 osp_sync_check_for_work(d);
1063 /* wake up the thread if requested to stop:
1064 * it might be waiting for in-progress to complete */
1065 if (unlikely(osp_sync_running(d) == 0))
1066 wake_up(&d->opd_syn_waitq);
1072 * The core of the syncing mechanism.
1074 * This is a callback called by the llog processing function. Essentially it
1075 * suspends llog processing until there is a record to process (it's supposed
1076 * to be committed locally). The function handles RPCs committed by the target
1077 * and cancels corresponding llog records.
1079 * \param[in] env LU environment provided by the caller
1080 * \param[in] llh llog handle we're processing
1081 * \param[in] rec current llog record
1082 * \param[in] data callback data containing a pointer to the device
1084 * \retval 0 to ask the caller (llog_process()) to continue
1085 * \retval LLOG_PROC_BREAK to ask the caller to break
1087 static int osp_sync_process_queues(const struct lu_env *env,
1088 struct llog_handle *llh,
1089 struct llog_rec_hdr *rec,
1092 struct osp_device *d = data;
1095 struct l_wait_info lwi = { 0 };
1097 if (!osp_sync_running(d)) {
1098 CDEBUG(D_HA, "stop llog processing\n");
1099 return LLOG_PROC_BREAK;
1102 /* process requests committed by OST */
1103 osp_sync_process_committed(env, d);
1105 /* if we there are changes to be processed and we have
1106 * resources for this ... do now */
1107 if (osp_sync_can_process_new(d, rec)) {
1109 /* ask llog for another record */
1110 CDEBUG(D_HA, "%u changes, %u in progress,"
1112 atomic_read(&d->opd_syn_changes),
1113 atomic_read(&d->opd_syn_rpc_in_progress),
1114 atomic_read(&d->opd_syn_rpc_in_flight));
1117 osp_sync_process_record(env, d, llh, rec);
1122 if (d->opd_syn_last_processed_id == d->opd_syn_last_used_id)
1123 osp_sync_remove_from_tracker(d);
1125 l_wait_event(d->opd_syn_waitq,
1126 !osp_sync_running(d) ||
1127 osp_sync_can_process_new(d, rec) ||
1128 !list_empty(&d->opd_syn_committed_there),
1136 * This thread runs llog_cat_process() scanner calling our callback
1137 * to process llog records. in the callback we implement tricky
1138 * state machine as we don't want to start scanning of the llog again
1139 * and again, also we don't want to process too many records and send
1140 * too many RPCs a time. so, depending on current load (num of changes
1141 * being synced to OST) the callback can suspend awaiting for some
1142 * new conditions, like syncs completed.
1144 * In order to process llog records left by previous boots and to allow
1145 * llog_process_thread() to find something (otherwise it'd just exit
1146 * immediately) we add a special GENERATATION record on each boot.
1148 * \param[in] _arg a pointer to thread's arguments
1150 * \retval 0 on success
1151 * \retval negative negated errno on error
1153 static int osp_sync_thread(void *_arg)
1155 struct osp_device *d = _arg;
1156 struct ptlrpc_thread *thread = &d->opd_syn_thread;
1157 struct l_wait_info lwi = { 0 };
1158 struct llog_ctxt *ctxt;
1159 struct obd_device *obd = d->opd_obd;
1160 struct llog_handle *llh;
1166 rc = lu_env_init(&env, LCT_LOCAL);
1168 CERROR("%s: can't initialize env: rc = %d\n",
1171 spin_lock(&d->opd_syn_lock);
1172 thread->t_flags = SVC_STOPPED;
1173 spin_unlock(&d->opd_syn_lock);
1174 wake_up(&thread->t_ctl_waitq);
1179 spin_lock(&d->opd_syn_lock);
1180 thread->t_flags = SVC_RUNNING;
1181 spin_unlock(&d->opd_syn_lock);
1182 wake_up(&thread->t_ctl_waitq);
1184 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
1186 CERROR("can't get appropriate context\n");
1187 GOTO(out, rc = -EINVAL);
1190 llh = ctxt->loc_handle;
1192 CERROR("can't get llh\n");
1193 llog_ctxt_put(ctxt);
1194 GOTO(out, rc = -EINVAL);
1197 rc = llog_cat_process(&env, llh, osp_sync_process_queues, d, 0, 0);
1199 CERROR("%s: llog process with osp_sync_process_queues "
1200 "failed: %d\n", d->opd_obd->obd_name, rc);
1203 LASSERTF(rc == 0 || rc == LLOG_PROC_BREAK,
1204 "%u changes, %u in progress, %u in flight: %d\n",
1205 atomic_read(&d->opd_syn_changes),
1206 atomic_read(&d->opd_syn_rpc_in_progress),
1207 atomic_read(&d->opd_syn_rpc_in_flight), rc);
1209 /* we don't expect llog_process_thread() to exit till umount */
1210 LASSERTF(thread->t_flags != SVC_RUNNING,
1211 "%u changes, %u in progress, %u in flight\n",
1212 atomic_read(&d->opd_syn_changes),
1213 atomic_read(&d->opd_syn_rpc_in_progress),
1214 atomic_read(&d->opd_syn_rpc_in_flight));
1216 /* wait till all the requests are completed */
1218 while (atomic_read(&d->opd_syn_rpc_in_progress) > 0) {
1219 osp_sync_process_committed(&env, d);
1221 lwi = LWI_TIMEOUT(cfs_time_seconds(5), NULL, NULL);
1222 rc = l_wait_event(d->opd_syn_waitq,
1223 atomic_read(&d->opd_syn_rpc_in_progress) == 0,
1225 if (rc == -ETIMEDOUT)
1227 LASSERTF(count < 10, "%s: %d %d %sempty\n",
1228 d->opd_obd->obd_name,
1229 atomic_read(&d->opd_syn_rpc_in_progress),
1230 atomic_read(&d->opd_syn_rpc_in_flight),
1231 list_empty(&d->opd_syn_committed_there) ? "" : "!");
1236 llog_cat_close(&env, llh);
1237 rc = llog_cleanup(&env, ctxt);
1239 CERROR("can't cleanup llog: %d\n", rc);
1241 LASSERTF(atomic_read(&d->opd_syn_rpc_in_progress) == 0,
1242 "%s: %d %d %sempty\n",
1243 d->opd_obd->obd_name, atomic_read(&d->opd_syn_rpc_in_progress),
1244 atomic_read(&d->opd_syn_rpc_in_flight),
1245 list_empty(&d->opd_syn_committed_there) ? "" : "!");
1247 thread->t_flags = SVC_STOPPED;
1249 wake_up(&thread->t_ctl_waitq);
1259 * Initializes the llog. Specific llog to be used depends on the type of the
1260 * target OSP represents (OST or MDT). The function adds appends a new llog
1261 * record to mark the place where the records associated with this boot
1264 * \param[in] env LU environment provided by the caller
1265 * \param[in] d OSP device
1267 * \retval 0 on success
1268 * \retval negative negated errno on error
1270 static int osp_sync_llog_init(const struct lu_env *env, struct osp_device *d)
1272 struct osp_thread_info *osi = osp_env_info(env);
1273 struct lu_fid *fid = &osi->osi_fid;
1274 struct llog_handle *lgh = NULL;
1275 struct obd_device *obd = d->opd_obd;
1276 struct llog_ctxt *ctxt;
1284 * open llog corresponding to our OST
1286 OBD_SET_CTXT_MAGIC(&obd->obd_lvfs_ctxt);
1287 obd->obd_lvfs_ctxt.dt = d->opd_storage;
1289 lu_local_obj_fid(fid, LLOG_CATALOGS_OID);
1291 rc = llog_osd_get_cat_list(env, d->opd_storage, d->opd_index, 1,
1292 &osi->osi_cid, fid);
1294 if (rc != -EFAULT) {
1295 CERROR("%s: can't get id from catalogs: rc = %d\n",
1300 /* After sparse OST indices is supported, the CATALOG file
1301 * may become a sparse file that results in failure on
1302 * reading. Skip this error as the llog will be created
1304 memset(&osi->osi_cid, 0, sizeof(osi->osi_cid));
1308 CDEBUG(D_INFO, "%s: Init llog for %d - catid "DOSTID":%x\n",
1309 obd->obd_name, d->opd_index,
1310 POSTID(&osi->osi_cid.lci_logid.lgl_oi),
1311 osi->osi_cid.lci_logid.lgl_ogen);
1313 rc = llog_setup(env, obd, &obd->obd_olg, LLOG_MDS_OST_ORIG_CTXT,
1314 d->opd_storage->dd_lu_dev.ld_obd,
1315 &osp_mds_ost_orig_logops);
1319 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
1322 if (likely(logid_id(&osi->osi_cid.lci_logid) != 0)) {
1323 rc = llog_open(env, ctxt, &lgh, &osi->osi_cid.lci_logid, NULL,
1325 /* re-create llog if it is missing */
1327 logid_set_id(&osi->osi_cid.lci_logid, 0);
1329 GOTO(out_cleanup, rc);
1332 if (unlikely(logid_id(&osi->osi_cid.lci_logid) == 0)) {
1333 rc = llog_open_create(env, ctxt, &lgh, NULL, NULL);
1335 GOTO(out_cleanup, rc);
1336 osi->osi_cid.lci_logid = lgh->lgh_id;
1339 LASSERT(lgh != NULL);
1340 ctxt->loc_handle = lgh;
1342 rc = llog_init_handle(env, lgh, LLOG_F_IS_CAT, NULL);
1344 GOTO(out_close, rc);
1346 rc = llog_osd_put_cat_list(env, d->opd_storage, d->opd_index, 1,
1347 &osi->osi_cid, fid);
1349 GOTO(out_close, rc);
1352 * put a mark in the llog till which we'll be processing
1353 * old records restless
1355 d->opd_syn_generation.mnt_cnt = cfs_time_current();
1356 d->opd_syn_generation.conn_cnt = cfs_time_current();
1358 osi->osi_hdr.lrh_type = LLOG_GEN_REC;
1359 osi->osi_hdr.lrh_len = sizeof(osi->osi_gen);
1361 memcpy(&osi->osi_gen.lgr_gen, &d->opd_syn_generation,
1362 sizeof(osi->osi_gen.lgr_gen));
1364 rc = llog_cat_add(env, lgh, &osi->osi_gen.lgr_hdr, &osi->osi_cookie);
1366 GOTO(out_close, rc);
1367 llog_ctxt_put(ctxt);
1370 llog_cat_close(env, lgh);
1372 llog_cleanup(env, ctxt);
1377 * Cleanup llog used for syncing.
1379 * Closes and cleanups the llog. The function is called when the device is
1382 * \param[in] env LU environment provided by the caller
1383 * \param[in] d OSP device
1385 static void osp_sync_llog_fini(const struct lu_env *env, struct osp_device *d)
1387 struct llog_ctxt *ctxt;
1389 ctxt = llog_get_context(d->opd_obd, LLOG_MDS_OST_ORIG_CTXT);
1391 llog_cat_close(env, ctxt->loc_handle);
1392 llog_cleanup(env, ctxt);
1397 * Initialization of the sync component of OSP.
1399 * Initializes the llog and starts a new thread to handle the changes to
1400 * the remote target (OST or MDT).
1402 * \param[in] env LU environment provided by the caller
1403 * \param[in] d OSP device
1405 * \retval 0 on success
1406 * \retval negative negated errno on error
1408 int osp_sync_init(const struct lu_env *env, struct osp_device *d)
1410 struct l_wait_info lwi = { 0 };
1411 struct task_struct *task;
1416 d->opd_syn_max_rpc_in_flight = OSP_MAX_IN_FLIGHT;
1417 d->opd_syn_max_rpc_in_progress = OSP_MAX_IN_PROGRESS;
1418 spin_lock_init(&d->opd_syn_lock);
1419 init_waitqueue_head(&d->opd_syn_waitq);
1420 init_waitqueue_head(&d->opd_syn_barrier_waitq);
1421 thread_set_flags(&d->opd_syn_thread, SVC_INIT);
1422 init_waitqueue_head(&d->opd_syn_thread.t_ctl_waitq);
1423 INIT_LIST_HEAD(&d->opd_syn_inflight_list);
1424 INIT_LIST_HEAD(&d->opd_syn_committed_there);
1426 if (d->opd_storage->dd_rdonly)
1429 rc = osp_sync_id_traction_init(d);
1434 * initialize llog storing changes
1436 rc = osp_sync_llog_init(env, d);
1438 CERROR("%s: can't initialize llog: rc = %d\n",
1439 d->opd_obd->obd_name, rc);
1444 * Start synchronization thread
1446 task = kthread_run(osp_sync_thread, d, "osp-syn-%u-%u",
1447 d->opd_index, d->opd_group);
1450 CERROR("%s: cannot start sync thread: rc = %d\n",
1451 d->opd_obd->obd_name, rc);
1455 l_wait_event(d->opd_syn_thread.t_ctl_waitq,
1456 osp_sync_running(d) || osp_sync_stopped(d), &lwi);
1460 osp_sync_llog_fini(env, d);
1462 osp_sync_id_traction_fini(d);
1467 * Stop the syncing thread.
1469 * Asks the syncing thread to stop and wait until it's stopped.
1471 * \param[in] d OSP device
1475 int osp_sync_fini(struct osp_device *d)
1477 struct ptlrpc_thread *thread = &d->opd_syn_thread;
1481 if (!thread_is_init(thread) && !thread_is_stopped(thread)) {
1482 thread->t_flags = SVC_STOPPING;
1483 wake_up(&d->opd_syn_waitq);
1484 wait_event(thread->t_ctl_waitq, thread_is_stopped(thread));
1488 * unregister transaction callbacks only when sync thread
1489 * has finished operations with llog
1491 osp_sync_id_traction_fini(d);
1496 static DEFINE_MUTEX(osp_id_tracker_sem);
1497 static struct list_head osp_id_tracker_list =
1498 LIST_HEAD_INIT(osp_id_tracker_list);
1501 * OSD commit callback.
1503 * The function is used as a local OSD commit callback to track the highest
1504 * committed llog record id. see osp_sync_id_traction_init() for the details.
1506 * \param[in] th local transaction handle committed
1507 * \param[in] cookie commit callback data (our private structure)
1509 static void osp_sync_tracker_commit_cb(struct thandle *th, void *cookie)
1511 struct osp_id_tracker *tr = cookie;
1512 struct osp_device *d;
1513 struct osp_txn_info *txn;
1517 txn = osp_txn_info(&th->th_ctx);
1518 if (txn == NULL || txn->oti_current_id < tr->otr_committed_id)
1521 spin_lock(&tr->otr_lock);
1522 if (likely(txn->oti_current_id > tr->otr_committed_id)) {
1523 CDEBUG(D_OTHER, "committed: %llu -> %llu\n",
1524 tr->otr_committed_id, txn->oti_current_id);
1525 tr->otr_committed_id = txn->oti_current_id;
1527 list_for_each_entry(d, &tr->otr_wakeup_list,
1529 d->opd_syn_last_committed_id = tr->otr_committed_id;
1530 wake_up(&d->opd_syn_waitq);
1533 spin_unlock(&tr->otr_lock);
1537 * Initialize commit tracking mechanism.
1539 * Some setups may have thousands of OSTs and each will be represented by OSP.
1540 * Meaning order of magnitute many more changes to apply every second. In order
1541 * to keep the number of commit callbacks low this mechanism was introduced.
1542 * The mechanism is very similar to transno used by MDT service: it's an single
1543 * ID stream which can be assigned by any OSP to its llog records. The tricky
1544 * part is that ID is stored in per-transaction data and re-used by all the OSPs
1545 * involved in that transaction. Then all these OSPs are woken up utilizing a single OSD commit callback.
1547 * The function initializes the data used by the tracker described above.
1548 * A singler tracker per OSD device is created.
1550 * \param[in] d OSP device
1552 * \retval 0 on success
1553 * \retval negative negated errno on error
1555 static int osp_sync_id_traction_init(struct osp_device *d)
1557 struct osp_id_tracker *tr, *found = NULL;
1561 LASSERT(d->opd_storage);
1562 LASSERT(d->opd_syn_tracker == NULL);
1563 INIT_LIST_HEAD(&d->opd_syn_ontrack);
1565 mutex_lock(&osp_id_tracker_sem);
1566 list_for_each_entry(tr, &osp_id_tracker_list, otr_list) {
1567 if (tr->otr_dev == d->opd_storage) {
1568 LASSERT(atomic_read(&tr->otr_refcount));
1569 atomic_inc(&tr->otr_refcount);
1570 d->opd_syn_tracker = tr;
1576 if (found == NULL) {
1580 d->opd_syn_tracker = tr;
1581 spin_lock_init(&tr->otr_lock);
1582 tr->otr_dev = d->opd_storage;
1583 tr->otr_next_id = 1;
1584 tr->otr_committed_id = 0;
1585 atomic_set(&tr->otr_refcount, 1);
1586 INIT_LIST_HEAD(&tr->otr_wakeup_list);
1587 list_add(&tr->otr_list, &osp_id_tracker_list);
1588 tr->otr_tx_cb.dtc_txn_commit =
1589 osp_sync_tracker_commit_cb;
1590 tr->otr_tx_cb.dtc_cookie = tr;
1591 tr->otr_tx_cb.dtc_tag = LCT_MD_THREAD;
1592 dt_txn_callback_add(d->opd_storage, &tr->otr_tx_cb);
1596 mutex_unlock(&osp_id_tracker_sem);
1602 * Release commit tracker.
1604 * Decrease a refcounter on the tracker used by the given OSP device \a d.
1605 * If no more users left, then the tracker is released.
1607 * \param[in] d OSP device
1609 static void osp_sync_id_traction_fini(struct osp_device *d)
1611 struct osp_id_tracker *tr;
1616 tr = d->opd_syn_tracker;
1622 osp_sync_remove_from_tracker(d);
1624 mutex_lock(&osp_id_tracker_sem);
1625 if (atomic_dec_and_test(&tr->otr_refcount)) {
1626 dt_txn_callback_del(d->opd_storage, &tr->otr_tx_cb);
1627 LASSERT(list_empty(&tr->otr_wakeup_list));
1628 list_del(&tr->otr_list);
1630 d->opd_syn_tracker = NULL;
1632 mutex_unlock(&osp_id_tracker_sem);
1638 * Generate a new ID on a tracker.
1640 * Generates a new ID using the tracker associated with the given OSP device
1641 * \a d, if the given ID \a id is non-zero. Unconditially adds OSP device to
1642 * the wakeup list, so OSP won't miss when a transaction using the ID is
1645 * \param[in] d OSP device
1646 * \param[in] id 0 or ID generated previously
1648 * \retval ID the caller should use
1650 static __u64 osp_sync_id_get(struct osp_device *d, __u64 id)
1652 struct osp_id_tracker *tr;
1654 tr = d->opd_syn_tracker;
1657 /* XXX: we can improve this introducing per-cpu preallocated ids? */
1658 spin_lock(&tr->otr_lock);
1659 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_TRACK_OVERFLOW))
1660 tr->otr_next_id = 0xfffffff0;
1662 if (unlikely(tr->otr_next_id <= d->opd_syn_last_used_id)) {
1663 spin_unlock(&tr->otr_lock);
1664 CERROR("%s: next %llu, last synced %llu\n",
1665 d->opd_obd->obd_name, tr->otr_next_id,
1666 d->opd_syn_last_used_id);
1671 id = tr->otr_next_id++;
1672 if (id > d->opd_syn_last_used_id)
1673 d->opd_syn_last_used_id = id;
1674 if (list_empty(&d->opd_syn_ontrack))
1675 list_add(&d->opd_syn_ontrack, &tr->otr_wakeup_list);
1676 spin_unlock(&tr->otr_lock);
1677 CDEBUG(D_OTHER, "new id %llu\n", id);
1683 * Stop to propagate commit status to OSP.
1685 * If the OSP does not have any llog records she's waiting to commit, then
1686 * it is possible to unsubscribe from wakeups from the tracking using this
1689 * \param[in] d OSP device not willing to wakeup
1691 static void osp_sync_remove_from_tracker(struct osp_device *d)
1693 struct osp_id_tracker *tr;
1695 tr = d->opd_syn_tracker;
1698 if (list_empty(&d->opd_syn_ontrack))
1701 spin_lock(&tr->otr_lock);
1702 list_del_init(&d->opd_syn_ontrack);
1703 spin_unlock(&tr->otr_lock);