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) 2014, 2015, Intel Corporation.
26 * lustre/osp/osp_trans.c
29 * 1. OSP (Object Storage Proxy) transaction methods
31 * Implement OSP layer transaction related interfaces for the dt_device API
32 * dt_device_operations.
35 * 2. Handle asynchronous idempotent operations
37 * The OSP uses OUT (Object Unified Target) RPC to talk with other server
38 * (MDT or OST) for kinds of operations, such as create, unlink, insert,
39 * delete, lookup, set_(x)attr, get_(x)attr, and etc. To reduce the number
40 * of RPCs, we allow multiple operations to be packaged together in single
43 * For the asynchronous idempotent operations, such as get_(x)attr, related
44 * RPCs will be inserted into an osp_device based shared asynchronous request
45 * queue - osp_device::opd_async_requests. When the queue is full, all the
46 * requests in the queue will be packaged into a single OUT RPC and given to
47 * the ptlrpcd daemon (for sending), then the queue is purged and other new
48 * requests can be inserted into it.
50 * When the asynchronous idempotent operation inserts the request into the
51 * shared queue, it will register an interpreter. When the packaged OUT RPC
52 * is replied (or failed to be sent out), all the registered interpreters
53 * will be called one by one to handle each own result.
56 * There are three kinds of transactions
58 * 1. Local transaction, all of updates of the transaction are in the local MDT.
59 * 2. Remote transaction, all of updates of the transaction are in one remote
60 * MDT, which only happens in LFSCK now.
61 * 3. Distribute transaction, updates for the transaction are in mulitple MDTs.
63 * Author: Di Wang <di.wang@intel.com>
64 * Author: Fan, Yong <fan.yong@intel.com>
67 #define DEBUG_SUBSYSTEM S_MDS
69 #include <lustre_net.h>
70 #include "osp_internal.h"
73 * The argument for the interpreter callback of osp request.
75 struct osp_update_args {
76 struct osp_update_request *oaua_update;
78 wait_queue_head_t *oaua_waitq;
79 bool oaua_flow_control;
80 const struct lu_env *oaua_update_env;
84 * Call back for each update request.
86 struct osp_update_callback {
87 /* list in the osp_update_request::our_cb_items */
88 struct list_head ouc_list;
90 /* The target of the async update request. */
91 struct osp_object *ouc_obj;
93 /* The data used by or_interpreter. */
96 /* The interpreter function called after the async request handled. */
97 osp_update_interpreter_t ouc_interpreter;
100 static struct object_update_request *object_update_request_alloc(size_t size)
102 struct object_update_request *ourq;
104 OBD_ALLOC_LARGE(ourq, size);
106 return ERR_PTR(-ENOMEM);
108 ourq->ourq_magic = UPDATE_REQUEST_MAGIC;
109 ourq->ourq_count = 0;
115 * Allocate new update request
117 * Allocate new update request and insert it to the req_update_list.
119 * \param [in] our osp_udate_request where to create a new
122 * \retval 0 if creation succeeds.
123 * \retval negative errno if creation fails.
125 int osp_object_update_request_create(struct osp_update_request *our,
128 struct osp_update_request_sub *ours;
134 if (size < OUT_UPDATE_INIT_BUFFER_SIZE)
135 size = OUT_UPDATE_INIT_BUFFER_SIZE;
137 ours->ours_req = object_update_request_alloc(size);
139 if (IS_ERR(ours->ours_req)) {
144 ours->ours_req_size = size;
145 INIT_LIST_HEAD(&ours->ours_list);
146 list_add_tail(&ours->ours_list, &our->our_req_list);
153 * Get current update request
155 * Get current object update request from our_req_list in
156 * osp_update_request, because we always insert the new update
157 * request in the last position, so the last update request
158 * in the list will be the current update req.
160 * \param[in] our osp update request where to get the
161 * current object update.
163 * \retval the current object update.
165 struct osp_update_request_sub *
166 osp_current_object_update_request(struct osp_update_request *our)
168 if (list_empty(&our->our_req_list))
171 return list_entry(our->our_req_list.prev, struct osp_update_request_sub,
176 * Allocate and initialize osp_update_request
178 * osp_update_request is being used to track updates being executed on
179 * this dt_device(OSD or OSP). The update buffer will be 4k initially,
180 * and increased if needed.
182 * \param [in] dt dt device
184 * \retval osp_update_request being allocated if succeed
185 * \retval ERR_PTR(errno) if failed
187 struct osp_update_request *osp_update_request_create(struct dt_device *dt)
189 struct osp_update_request *our;
193 return ERR_PTR(-ENOMEM);
195 INIT_LIST_HEAD(&our->our_req_list);
196 INIT_LIST_HEAD(&our->our_cb_items);
197 INIT_LIST_HEAD(&our->our_list);
198 spin_lock_init(&our->our_list_lock);
200 osp_object_update_request_create(our, OUT_UPDATE_INIT_BUFFER_SIZE);
204 void osp_update_request_destroy(struct osp_update_request *our)
206 struct osp_update_request_sub *ours;
207 struct osp_update_request_sub *tmp;
212 list_for_each_entry_safe(ours, tmp, &our->our_req_list, ours_list) {
213 list_del(&ours->ours_list);
214 if (ours->ours_req != NULL)
215 OBD_FREE_LARGE(ours->ours_req, ours->ours_req_size);
222 object_update_request_dump(const struct object_update_request *ourq,
226 size_t total_size = 0;
228 for (i = 0; i < ourq->ourq_count; i++) {
229 struct object_update *update;
232 update = object_update_request_get(ourq, i, &size);
233 LASSERT(update != NULL);
234 CDEBUG(mask, "i = %u fid = "DFID" op = %s "
235 "params = %d batchid = "LPU64" size = %zu repsize %u\n",
236 i, PFID(&update->ou_fid),
237 update_op_str(update->ou_type),
238 update->ou_params_count,
239 update->ou_batchid, size,
240 (unsigned)update->ou_result_size);
245 CDEBUG(mask, "updates = %p magic = %x count = %d size = %zu\n", ourq,
246 ourq->ourq_magic, ourq->ourq_count, total_size);
250 * Prepare inline update request
252 * Prepare OUT update ptlrpc inline request, and the request usually includes
253 * one update buffer, which does not need bulk transfer.
255 * \param[in] env execution environment
256 * \param[in] req ptlrpc request
257 * \param[in] ours sub osp_update_request to be packed
259 * \retval 0 if packing succeeds
260 * \retval negative errno if packing fails
262 int osp_prep_inline_update_req(const struct lu_env *env,
263 struct ptlrpc_request *req,
264 struct osp_update_request *our,
267 struct osp_update_request_sub *ours;
268 struct out_update_header *ouh;
269 __u32 update_req_size;
272 ours = list_entry(our->our_req_list.next,
273 struct osp_update_request_sub, ours_list);
274 update_req_size = object_update_request_size(ours->ours_req);
275 req_capsule_set_size(&req->rq_pill, &RMF_OUT_UPDATE_HEADER, RCL_CLIENT,
276 update_req_size + sizeof(*ouh));
278 rc = ptlrpc_request_pack(req, LUSTRE_MDS_VERSION, OUT_UPDATE);
282 ouh = req_capsule_client_get(&req->rq_pill, &RMF_OUT_UPDATE_HEADER);
283 ouh->ouh_magic = OUT_UPDATE_HEADER_MAGIC;
285 ouh->ouh_inline_length = update_req_size;
286 ouh->ouh_reply_size = repsize;
288 memcpy(ouh->ouh_inline_data, ours->ours_req, update_req_size);
290 req_capsule_set_size(&req->rq_pill, &RMF_OUT_UPDATE_REPLY,
291 RCL_SERVER, repsize);
293 ptlrpc_request_set_replen(req);
294 req->rq_request_portal = OUT_PORTAL;
295 req->rq_reply_portal = OSC_REPLY_PORTAL;
301 * Prepare update request.
303 * Prepare OUT update ptlrpc request, and the request usually includes
304 * all of updates (stored in \param ureq) from one operation.
306 * \param[in] env execution environment
307 * \param[in] imp import on which ptlrpc request will be sent
308 * \param[in] ureq hold all of updates which will be packed into the req
309 * \param[in] reqp request to be created
311 * \retval 0 if preparation succeeds.
312 * \retval negative errno if preparation fails.
314 int osp_prep_update_req(const struct lu_env *env, struct obd_import *imp,
315 struct osp_update_request *our,
316 struct ptlrpc_request **reqp)
318 struct ptlrpc_request *req;
319 struct ptlrpc_bulk_desc *desc;
320 struct osp_update_request_sub *ours;
321 const struct object_update_request *ourq;
322 struct out_update_header *ouh;
323 struct out_update_buffer *oub;
326 struct object_update_reply *reply;
331 list_for_each_entry(ours, &our->our_req_list, ours_list) {
332 object_update_request_dump(ours->ours_req, D_INFO);
334 ourq = ours->ours_req;
335 for (i = 0; i < ourq->ourq_count; i++) {
336 struct object_update *update;
340 /* XXX: it's very inefficient to lookup update
341 * this way, iterating from the beginning
343 update = object_update_request_get(ourq, i, &size);
344 LASSERT(update != NULL);
346 repsize += sizeof(reply->ourp_lens[0]);
347 repsize += sizeof(struct object_update_result);
348 repsize += update->ou_result_size;
353 repsize += sizeof(*reply);
354 repsize = (repsize + OUT_UPDATE_REPLY_SIZE - 1) &
355 ~(OUT_UPDATE_REPLY_SIZE - 1);
356 LASSERT(buf_count > 0);
358 req = ptlrpc_request_alloc(imp, &RQF_OUT_UPDATE);
362 if (buf_count == 1) {
363 ours = list_entry(our->our_req_list.next,
364 struct osp_update_request_sub, ours_list);
366 /* Let's check if it can be packed inline */
367 if (object_update_request_size(ours->ours_req) +
368 sizeof(struct out_update_header) <
369 OUT_UPDATE_MAX_INLINE_SIZE) {
370 rc = osp_prep_inline_update_req(env, req, our, repsize);
377 req_capsule_set_size(&req->rq_pill, &RMF_OUT_UPDATE_HEADER, RCL_CLIENT,
378 sizeof(struct osp_update_request));
380 req_capsule_set_size(&req->rq_pill, &RMF_OUT_UPDATE_BUF, RCL_CLIENT,
381 buf_count * sizeof(*oub));
383 rc = ptlrpc_request_pack(req, LUSTRE_MDS_VERSION, OUT_UPDATE);
387 ouh = req_capsule_client_get(&req->rq_pill, &RMF_OUT_UPDATE_HEADER);
388 ouh->ouh_magic = OUT_UPDATE_HEADER_MAGIC;
389 ouh->ouh_count = buf_count;
390 ouh->ouh_inline_length = 0;
391 ouh->ouh_reply_size = repsize;
392 oub = req_capsule_client_get(&req->rq_pill, &RMF_OUT_UPDATE_BUF);
393 list_for_each_entry(ours, &our->our_req_list, ours_list) {
394 oub->oub_size = ours->ours_req_size;
398 req->rq_bulk_write = 1;
399 desc = ptlrpc_prep_bulk_imp(req, buf_count,
400 MD_MAX_BRW_SIZE >> LNET_MTU_BITS,
401 PTLRPC_BULK_GET_SOURCE | PTLRPC_BULK_BUF_KVEC,
402 MDS_BULK_PORTAL, &ptlrpc_bulk_kvec_ops);
404 GOTO(out_req, rc = -ENOMEM);
406 /* NB req now owns desc and will free it when it gets freed */
407 list_for_each_entry(ours, &our->our_req_list, ours_list) {
408 desc->bd_frag_ops->add_iov_frag(desc, ours->ours_req,
409 ours->ours_req_size);
410 total += ours->ours_req_size;
412 CDEBUG(D_OTHER, "total %d in %u\n", total, our->our_update_nr);
414 req_capsule_set_size(&req->rq_pill, &RMF_OUT_UPDATE_REPLY,
415 RCL_SERVER, repsize);
417 ptlrpc_request_set_replen(req);
418 req->rq_request_portal = OUT_PORTAL;
419 req->rq_reply_portal = OSC_REPLY_PORTAL;
424 ptlrpc_req_finished(req);
432 * Send update request to the remote MDT synchronously.
434 * \param[in] env execution environment
435 * \param[in] imp import on which ptlrpc request will be sent
436 * \param[in] our hold all of updates which will be packed into the req
437 * \param[in] reqp request to be created
439 * \retval 0 if RPC succeeds.
440 * \retval negative errno if RPC fails.
442 int osp_remote_sync(const struct lu_env *env, struct osp_device *osp,
443 struct osp_update_request *our,
444 struct ptlrpc_request **reqp)
446 struct obd_import *imp = osp->opd_obd->u.cli.cl_import;
447 struct ptlrpc_request *req = NULL;
451 rc = osp_prep_update_req(env, imp, our, &req);
455 /* This will only be called with read-only update, and these updates
456 * might be used to retrieve update log during recovery process, so
457 * it will be allowed to send during recovery process */
458 req->rq_allow_replay = 1;
459 req->rq_allow_intr = 1;
461 /* Note: some dt index api might return non-zero result here, like
462 * osd_index_ea_lookup, so we should only check rc < 0 here */
463 rc = ptlrpc_queue_wait(req);
465 if (rc < 0 || reqp == NULL)
466 ptlrpc_req_finished(req);
474 * Invalidate all objects in the osp thandle
476 * invalidate all of objects in the update request, which will be called
477 * when the transaction is aborted.
479 * \param[in] oth osp thandle.
481 static void osp_thandle_invalidate_object(const struct lu_env *env,
482 struct osp_thandle *oth)
484 struct osp_update_request *our = oth->ot_our;
485 struct osp_update_request_sub *ours;
490 list_for_each_entry(ours, &our->our_req_list, ours_list) {
491 struct object_update_request *our_req = ours->ours_req;
493 struct lu_object *obj;
494 struct osp_object *osp_obj;
496 for (i = 0; i < our_req->ourq_count; i++) {
497 struct object_update *update;
499 update = object_update_request_get(our_req, i, NULL);
503 if (update->ou_type != OUT_WRITE)
506 if (!fid_is_sane(&update->ou_fid))
509 obj = lu_object_find_slice(env,
510 &oth->ot_super.th_dev->dd_lu_dev,
511 &update->ou_fid, NULL);
515 osp_obj = lu2osp_obj(obj);
516 if (osp_obj->opo_ooa != NULL) {
517 spin_lock(&osp_obj->opo_lock);
518 osp_obj->opo_ooa->ooa_attr.la_valid = 0;
519 osp_obj->opo_stale = 1;
520 spin_unlock(&osp_obj->opo_lock);
522 lu_object_put(env, obj);
527 static void osp_trans_stop_cb(const struct lu_env *env,
528 struct osp_thandle *oth, int result)
530 struct dt_txn_commit_cb *dcb;
531 struct dt_txn_commit_cb *tmp;
533 /* call per-transaction stop callbacks if any */
534 list_for_each_entry_safe(dcb, tmp, &oth->ot_stop_dcb_list,
536 LASSERTF(dcb->dcb_magic == TRANS_COMMIT_CB_MAGIC,
537 "commit callback entry: magic=%x name='%s'\n",
538 dcb->dcb_magic, dcb->dcb_name);
539 list_del_init(&dcb->dcb_linkage);
540 dcb->dcb_func(NULL, &oth->ot_super, dcb, result);
544 osp_thandle_invalidate_object(env, oth);
548 * Allocate an osp request and initialize it with the given parameters.
550 * \param[in] obj pointer to the operation target
551 * \param[in] data pointer to the data used by the interpreter
552 * \param[in] interpreter pointer to the interpreter function
554 * \retval pointer to the asychronous request
555 * \retval NULL if the allocation failed
557 static struct osp_update_callback *
558 osp_update_callback_init(struct osp_object *obj, void *data,
559 osp_update_interpreter_t interpreter)
561 struct osp_update_callback *ouc;
567 lu_object_get(osp2lu_obj(obj));
568 INIT_LIST_HEAD(&ouc->ouc_list);
570 ouc->ouc_data = data;
571 ouc->ouc_interpreter = interpreter;
577 * Destroy the osp_update_callback.
579 * \param[in] env pointer to the thread context
580 * \param[in] ouc pointer to osp_update_callback
582 static void osp_update_callback_fini(const struct lu_env *env,
583 struct osp_update_callback *ouc)
585 LASSERT(list_empty(&ouc->ouc_list));
587 lu_object_put(env, osp2lu_obj(ouc->ouc_obj));
592 * Interpret the packaged OUT RPC results.
594 * For every packaged sub-request, call its registered interpreter function.
595 * Then destroy the sub-request.
597 * \param[in] env pointer to the thread context
598 * \param[in] req pointer to the RPC
599 * \param[in] arg pointer to data used by the interpreter
600 * \param[in] rc the RPC return value
602 * \retval 0 for success
603 * \retval negative error number on failure
605 static int osp_update_interpret(const struct lu_env *env,
606 struct ptlrpc_request *req, void *arg, int rc)
608 struct object_update_reply *reply = NULL;
609 struct osp_update_args *oaua = arg;
610 struct osp_update_request *our = oaua->oaua_update;
611 struct osp_thandle *oth;
612 struct osp_update_callback *ouc;
613 struct osp_update_callback *next;
623 /* Sigh env might be NULL in some cases, see
625 * osp_send_update_thread()
626 * ptlrpc_set_wait() ----> null env.
628 * osp_update_interpret()
629 * Let's use env in oaua for this case.
632 env = oaua->oaua_update_env;
634 oaua->oaua_update = NULL;
636 if (oaua->oaua_flow_control) {
637 struct osp_device *osp;
639 LASSERT(oth != NULL);
640 osp = dt2osp_dev(oth->ot_super.th_dev);
641 obd_put_request_slot(&osp->opd_obd->u.cli);
644 /* Unpack the results from the reply message. */
645 if (req->rq_repmsg != NULL && req->rq_replied) {
646 reply = req_capsule_server_sized_get(&req->rq_pill,
647 &RMF_OUT_UPDATE_REPLY,
648 OUT_UPDATE_REPLY_SIZE);
649 if (reply == NULL || reply->ourp_magic != UPDATE_REPLY_MAGIC) {
653 count = reply->ourp_count;
657 list_for_each_entry_safe(ouc, next, &our->our_cb_items, ouc_list) {
658 list_del_init(&ouc->ouc_list);
660 /* The peer may only have handled some requests (indicated
661 * by the 'count') in the packaged OUT RPC, we can only get
662 * results for the handled part. */
663 if (index < count && reply->ourp_lens[index] > 0 && rc >= 0) {
664 struct object_update_result *result;
666 result = object_update_result_get(reply, index, NULL);
670 rc1 = rc = result->our_rc;
671 } else if (rc1 >= 0) {
672 /* The peer did not handle these request, let's return
673 * -EINVAL to update interpret for now */
680 if (ouc->ouc_interpreter != NULL)
681 ouc->ouc_interpreter(env, reply, req, ouc->ouc_obj,
682 ouc->ouc_data, index, rc1);
684 osp_update_callback_fini(env, ouc);
688 if (oaua->oaua_count != NULL && atomic_dec_and_test(oaua->oaua_count))
689 wake_up_all(oaua->oaua_waitq);
692 /* oth and osp_update_requests will be destoryed in
694 osp_trans_stop_cb(env, oth, rc);
695 osp_thandle_put(oth);
697 osp_update_request_destroy(our);
704 * Pack all the requests in the shared asynchronous idempotent request queue
705 * into a single OUT RPC that will be given to the background ptlrpcd daemon.
707 * \param[in] env pointer to the thread context
708 * \param[in] osp pointer to the OSP device
709 * \param[in] our pointer to the shared queue
711 * \retval 0 for success
712 * \retval negative error number on failure
714 int osp_unplug_async_request(const struct lu_env *env,
715 struct osp_device *osp,
716 struct osp_update_request *our)
718 struct osp_update_args *args;
719 struct ptlrpc_request *req = NULL;
722 rc = osp_prep_update_req(env, osp->opd_obd->u.cli.cl_import,
725 struct osp_update_callback *ouc;
726 struct osp_update_callback *next;
728 list_for_each_entry_safe(ouc, next,
729 &our->our_cb_items, ouc_list) {
730 list_del_init(&ouc->ouc_list);
731 if (ouc->ouc_interpreter != NULL)
732 ouc->ouc_interpreter(env, NULL, NULL,
734 ouc->ouc_data, 0, rc);
735 osp_update_callback_fini(env, ouc);
737 osp_update_request_destroy(our);
739 args = ptlrpc_req_async_args(req);
740 args->oaua_update = our;
741 args->oaua_count = NULL;
742 args->oaua_waitq = NULL;
743 /* Note: this is asynchronous call for the request, so the
744 * interrupte cb and current function will be different
745 * thread, so we need use different env */
746 args->oaua_update_env = NULL;
747 args->oaua_flow_control = false;
748 req->rq_interpret_reply = osp_update_interpret;
749 ptlrpcd_add_req(req);
756 * Find or create (if NOT exist or purged) the shared asynchronous idempotent
757 * request queue - osp_device::opd_async_requests.
759 * If the osp_device::opd_async_requests is not NULL, then return it directly;
760 * otherwise create new osp_update_request and attach it to opd_async_requests.
762 * \param[in] osp pointer to the OSP device
764 * \retval pointer to the shared queue
765 * \retval negative error number on failure
767 static struct osp_update_request *
768 osp_find_or_create_async_update_request(struct osp_device *osp)
770 struct osp_update_request *our = osp->opd_async_requests;
775 our = osp_update_request_create(&osp->opd_dt_dev);
779 osp->opd_async_requests = our;
785 * Insert an osp_update_callback into the osp_update_request.
787 * Insert an osp_update_callback to the osp_update_request. Usually each update
788 * in the osp_update_request will have one correspondent callback, and these
789 * callbacks will be called in rq_interpret_reply.
791 * \param[in] env pointer to the thread context
792 * \param[in] obj pointer to the operation target object
793 * \param[in] data pointer to the data used by the interpreter
794 * \param[in] interpreter pointer to the interpreter function
796 * \retval 0 for success
797 * \retval negative error number on failure
799 int osp_insert_update_callback(const struct lu_env *env,
800 struct osp_update_request *our,
801 struct osp_object *obj, void *data,
802 osp_update_interpreter_t interpreter)
804 struct osp_update_callback *ouc;
806 ouc = osp_update_callback_init(obj, data, interpreter);
810 list_add_tail(&ouc->ouc_list, &our->our_cb_items);
816 * Insert an asynchronous idempotent request to the shared request queue that
817 * is attached to the osp_device.
819 * This function generates a new osp_async_request with the given parameters,
820 * then tries to insert the request into the osp_device-based shared request
821 * queue. If the queue is full, then triggers the packaged OUT RPC to purge
822 * the shared queue firstly, and then re-tries.
824 * NOTE: must hold the osp::opd_async_requests_mutex to serialize concurrent
825 * osp_insert_async_request call from others.
827 * \param[in] env pointer to the thread context
828 * \param[in] op operation type, see 'enum update_type'
829 * \param[in] obj pointer to the operation target
830 * \param[in] count array size of the subsequent \a lens and \a bufs
831 * \param[in] lens buffer length array for the subsequent \a bufs
832 * \param[in] bufs the buffers to compose the request
833 * \param[in] data pointer to the data used by the interpreter
834 * \param[in] repsize how many bytes the caller allocated for \a data
835 * \param[in] interpreter pointer to the interpreter function
837 * \retval 0 for success
838 * \retval negative error number on failure
840 int osp_insert_async_request(const struct lu_env *env, enum update_type op,
841 struct osp_object *obj, int count,
842 __u16 *lens, const void **bufs,
843 void *data, __u32 repsize,
844 osp_update_interpreter_t interpreter)
846 struct osp_device *osp;
847 struct osp_update_request *our;
848 struct object_update *object_update;
849 size_t max_update_size;
850 struct object_update_request *ureq;
851 struct osp_update_request_sub *ours;
855 osp = lu2osp_dev(osp2lu_obj(obj)->lo_dev);
856 our = osp_find_or_create_async_update_request(osp);
858 RETURN(PTR_ERR(our));
861 ours = osp_current_object_update_request(our);
863 ureq = ours->ours_req;
864 max_update_size = ours->ours_req_size -
865 object_update_request_size(ureq);
867 object_update = update_buffer_get_update(ureq, ureq->ourq_count);
868 rc = out_update_pack(env, object_update, &max_update_size, op,
869 lu_object_fid(osp2lu_obj(obj)), count, lens, bufs,
871 /* The queue is full. */
873 osp->opd_async_requests = NULL;
874 mutex_unlock(&osp->opd_async_requests_mutex);
876 rc = osp_unplug_async_request(env, osp, our);
877 mutex_lock(&osp->opd_async_requests_mutex);
881 our = osp_find_or_create_async_update_request(osp);
883 RETURN(PTR_ERR(our));
891 our->our_update_nr++;
894 rc = osp_insert_update_callback(env, our, obj, data, interpreter);
899 int osp_trans_update_request_create(struct thandle *th)
901 struct osp_thandle *oth = thandle_to_osp_thandle(th);
902 struct osp_update_request *our;
904 if (oth->ot_our != NULL)
907 our = osp_update_request_create(th->th_dev);
909 th->th_result = PTR_ERR(our);
919 void osp_thandle_destroy(struct osp_thandle *oth)
921 LASSERT(oth->ot_magic == OSP_THANDLE_MAGIC);
922 LASSERT(list_empty(&oth->ot_commit_dcb_list));
923 LASSERT(list_empty(&oth->ot_stop_dcb_list));
924 if (oth->ot_our != NULL)
925 osp_update_request_destroy(oth->ot_our);
930 * The OSP layer dt_device_operations::dt_trans_create() interface
931 * to create a transaction.
933 * There are two kinds of transactions that will involve OSP:
935 * 1) If the transaction only contains the updates on remote server
936 * (MDT or OST), such as re-generating the lost OST-object for
937 * LFSCK, then it is a remote transaction. For remote transaction,
938 * the upper layer caller (such as the LFSCK engine) will call the
939 * dt_trans_create() (with the OSP dt_device as the parameter),
940 * then the call will be directed to the osp_trans_create() that
941 * creates the transaction handler and returns it to the caller.
943 * 2) If the transcation contains both local and remote updates,
944 * such as cross MDTs create under DNE mode, then the upper layer
945 * caller will not trigger osp_trans_create(). Instead, it will
946 * call dt_trans_create() on other dt_device, such as LOD that
947 * will generate the transaction handler. Such handler will be
948 * used by the whole transaction in subsequent sub-operations.
950 * \param[in] env pointer to the thread context
951 * \param[in] d pointer to the OSP dt_device
953 * \retval pointer to the transaction handler
954 * \retval negative error number on failure
956 struct thandle *osp_trans_create(const struct lu_env *env, struct dt_device *d)
958 struct osp_thandle *oth;
959 struct thandle *th = NULL;
963 if (unlikely(oth == NULL))
964 RETURN(ERR_PTR(-ENOMEM));
966 oth->ot_magic = OSP_THANDLE_MAGIC;
969 th->th_tags = LCT_TX_HANDLE;
971 atomic_set(&oth->ot_refcount, 1);
972 INIT_LIST_HEAD(&oth->ot_commit_dcb_list);
973 INIT_LIST_HEAD(&oth->ot_stop_dcb_list);
979 * Add commit callback to transaction.
981 * Add commit callback to the osp thandle, which will be called
982 * when the thandle is committed remotely.
984 * \param[in] th the thandle
985 * \param[in] dcb commit callback structure
987 * \retval only return 0 for now.
989 int osp_trans_cb_add(struct thandle *th, struct dt_txn_commit_cb *dcb)
991 struct osp_thandle *oth = thandle_to_osp_thandle(th);
993 LASSERT(dcb->dcb_magic == TRANS_COMMIT_CB_MAGIC);
994 LASSERT(&dcb->dcb_func != NULL);
995 if (dcb->dcb_flags & DCB_TRANS_STOP)
996 list_add(&dcb->dcb_linkage, &oth->ot_stop_dcb_list);
998 list_add(&dcb->dcb_linkage, &oth->ot_commit_dcb_list);
1002 static void osp_trans_commit_cb(struct osp_thandle *oth, int result)
1004 struct dt_txn_commit_cb *dcb;
1005 struct dt_txn_commit_cb *tmp;
1007 LASSERT(atomic_read(&oth->ot_refcount) > 0);
1008 /* call per-transaction callbacks if any */
1009 list_for_each_entry_safe(dcb, tmp, &oth->ot_commit_dcb_list,
1011 LASSERTF(dcb->dcb_magic == TRANS_COMMIT_CB_MAGIC,
1012 "commit callback entry: magic=%x name='%s'\n",
1013 dcb->dcb_magic, dcb->dcb_name);
1014 list_del_init(&dcb->dcb_linkage);
1015 dcb->dcb_func(NULL, &oth->ot_super, dcb, result);
1019 static void osp_request_commit_cb(struct ptlrpc_request *req)
1021 struct thandle *th = req->rq_cb_data;
1022 struct osp_thandle *oth;
1023 __u64 last_committed_transno = 0;
1024 int result = req->rq_status;
1030 oth = thandle_to_osp_thandle(th);
1031 if (req->rq_repmsg != NULL &&
1032 lustre_msg_get_last_committed(req->rq_repmsg))
1033 last_committed_transno =
1034 lustre_msg_get_last_committed(req->rq_repmsg);
1036 if (last_committed_transno <
1037 req->rq_import->imp_peer_committed_transno)
1038 last_committed_transno =
1039 req->rq_import->imp_peer_committed_transno;
1041 CDEBUG(D_HA, "trans no "LPU64" committed transno "LPU64"\n",
1042 req->rq_transno, last_committed_transno);
1044 /* If the transaction is not really committed, mark result = 1 */
1045 if (req->rq_transno != 0 &&
1046 (req->rq_transno > last_committed_transno) && result == 0)
1049 osp_trans_commit_cb(oth, result);
1050 req->rq_committed = 1;
1051 osp_thandle_put(oth);
1056 * callback of osp transaction
1058 * Call all of callbacks for this osp thandle. This will only be
1059 * called in error handler path. In the normal processing path,
1060 * these callback will be called in osp_request_commit_cb() and
1061 * osp_update_interpret().
1063 * \param [in] env execution environment
1064 * \param [in] oth osp thandle
1065 * \param [in] rc result of the osp thandle
1067 void osp_trans_callback(const struct lu_env *env,
1068 struct osp_thandle *oth, int rc)
1070 struct osp_update_callback *ouc;
1071 struct osp_update_callback *next;
1073 if (oth->ot_our != NULL) {
1074 list_for_each_entry_safe(ouc, next,
1075 &oth->ot_our->our_cb_items, ouc_list) {
1076 list_del_init(&ouc->ouc_list);
1077 if (ouc->ouc_interpreter != NULL)
1078 ouc->ouc_interpreter(env, NULL, NULL,
1080 ouc->ouc_data, 0, rc);
1081 osp_update_callback_fini(env, ouc);
1084 osp_trans_stop_cb(env, oth, rc);
1085 osp_trans_commit_cb(oth, rc);
1089 * Send the request for remote updates.
1091 * Send updates to the remote MDT. Prepare the request by osp_update_req
1092 * and send them to remote MDT, for sync request, it will wait
1093 * until the reply return, otherwise hand it to ptlrpcd.
1095 * Please refer to osp_trans_create() for transaction type.
1097 * \param[in] env pointer to the thread context
1098 * \param[in] osp pointer to the OSP device
1099 * \param[in] our pointer to the osp_update_request
1101 * \retval 0 for success
1102 * \retval negative error number on failure
1104 static int osp_send_update_req(const struct lu_env *env,
1105 struct osp_device *osp,
1106 struct osp_update_request *our)
1108 struct osp_update_args *args;
1109 struct ptlrpc_request *req;
1110 struct lu_device *top_device;
1111 struct osp_thandle *oth = our->our_th;
1115 LASSERT(oth != NULL);
1116 rc = osp_prep_update_req(env, osp->opd_obd->u.cli.cl_import,
1119 osp_trans_callback(env, oth, rc);
1123 args = ptlrpc_req_async_args(req);
1124 args->oaua_update = our;
1125 /* set env to NULL, in case the interrupt cb and current function
1126 * are in different thread */
1127 args->oaua_update_env = NULL;
1128 osp_thandle_get(oth); /* hold for update interpret */
1129 req->rq_interpret_reply = osp_update_interpret;
1130 if (!oth->ot_super.th_wait_submit && !oth->ot_super.th_sync) {
1131 if (!osp->opd_imp_active || !osp->opd_imp_connected) {
1132 osp_trans_callback(env, oth, rc);
1133 osp_thandle_put(oth);
1134 GOTO(out, rc = -ENOTCONN);
1137 rc = obd_get_request_slot(&osp->opd_obd->u.cli);
1139 osp_trans_callback(env, oth, rc);
1140 osp_thandle_put(oth);
1141 GOTO(out, rc = -ENOTCONN);
1143 args->oaua_flow_control = true;
1145 if (!osp->opd_connect_mdt) {
1146 down_read(&osp->opd_async_updates_rwsem);
1147 args->oaua_count = &osp->opd_async_updates_count;
1148 args->oaua_waitq = &osp->opd_syn_barrier_waitq;
1149 up_read(&osp->opd_async_updates_rwsem);
1150 atomic_inc(args->oaua_count);
1153 ptlrpcd_add_req(req);
1156 osp_thandle_get(oth); /* hold for commit callback */
1157 req->rq_commit_cb = osp_request_commit_cb;
1158 req->rq_cb_data = &oth->ot_super;
1159 args->oaua_flow_control = false;
1161 /* If the transaction is created during MDT recoverying
1162 * process, it means this is an recovery update, we need
1163 * to let OSP send it anyway without checking recoverying
1164 * status, in case the other target is being recoveried
1165 * at the same time, and if we wait here for the import
1166 * to be recoveryed, it might cause deadlock */
1167 top_device = osp->opd_dt_dev.dd_lu_dev.ld_site->ls_top_dev;
1168 if (top_device->ld_obd->obd_recovering)
1169 req->rq_allow_replay = 1;
1171 /* Because this req will be synchronus, i.e. it will be called
1172 * in the same thread, so it will be safe to use current
1174 args->oaua_update_env = env;
1175 if (osp->opd_connect_mdt)
1176 osp_get_rpc_lock(osp);
1177 rc = ptlrpc_queue_wait(req);
1178 if (osp->opd_connect_mdt)
1179 osp_put_rpc_lock(osp);
1180 if ((rc == -ENOMEM && req->rq_set == NULL) ||
1181 (req->rq_transno == 0 && !req->rq_committed)) {
1182 if (args->oaua_update != NULL) {
1183 /* If osp_update_interpret is not being called,
1184 * release the osp_thandle */
1185 args->oaua_update = NULL;
1186 osp_thandle_put(oth);
1189 req->rq_cb_data = NULL;
1190 rc = rc == 0 ? req->rq_status : rc;
1191 osp_trans_callback(env, oth, rc);
1192 osp_thandle_put(oth);
1198 ptlrpc_req_finished(req);
1204 * Get local thandle for osp_thandle
1206 * Get the local OSD thandle from the OSP thandle. Currently, there
1207 * are a few OSP API (osp_object_create() and osp_sync_add()) needs
1208 * to update the object on local OSD device.
1210 * If the osp_thandle comes from normal stack (MDD->LOD->OSP), then
1211 * we will get local thandle by thandle_get_sub_by_dt.
1213 * If the osp_thandle is remote thandle (th_top == NULL, only used
1214 * by LFSCK), then it will create a local thandle, and stop it in
1215 * osp_trans_stop(). And this only happens on OSP for OST.
1217 * These are temporary solution, once OSP accessing OSD object is
1218 * being fixed properly, this function should be removed. XXX
1220 * \param[in] env pointer to the thread context
1221 * \param[in] th pointer to the transaction handler
1222 * \param[in] dt pointer to the OSP device
1224 * \retval pointer to the local thandle
1225 * \retval ERR_PTR(errno) if it fails.
1227 struct thandle *osp_get_storage_thandle(const struct lu_env *env,
1229 struct osp_device *osp)
1231 struct osp_thandle *oth;
1232 struct thandle *local_th;
1234 if (th->th_top != NULL)
1235 return thandle_get_sub_by_dt(env, th->th_top,
1238 LASSERT(!osp->opd_connect_mdt);
1239 oth = thandle_to_osp_thandle(th);
1240 if (oth->ot_storage_th != NULL)
1241 return oth->ot_storage_th;
1243 local_th = dt_trans_create(env, osp->opd_storage);
1244 if (IS_ERR(local_th))
1247 oth->ot_storage_th = local_th;
1253 * Set version for the transaction
1255 * Set the version for the transaction and add the request to
1256 * the sending list, then after transaction stop, the request
1257 * will be picked in the order of version, by sending thread.
1259 * \param [in] oth osp thandle to be set version.
1261 * \retval 0 if set version succeeds
1262 * negative errno if set version fails.
1264 int osp_check_and_set_rpc_version(struct osp_thandle *oth,
1265 struct osp_object *obj)
1267 struct osp_device *osp = dt2osp_dev(oth->ot_super.th_dev);
1268 struct osp_updates *ou = osp->opd_update;
1273 if (oth->ot_our->our_version != 0)
1276 spin_lock(&ou->ou_lock);
1277 spin_lock(&oth->ot_our->our_list_lock);
1278 if (obj->opo_stale) {
1279 spin_unlock(&oth->ot_our->our_list_lock);
1280 spin_unlock(&ou->ou_lock);
1284 /* Assign the version and add it to the sending list */
1285 osp_thandle_get(oth);
1286 oth->ot_our->our_version = ou->ou_version++;
1287 list_add_tail(&oth->ot_our->our_list,
1288 &osp->opd_update->ou_list);
1289 oth->ot_our->our_req_ready = 0;
1290 spin_unlock(&oth->ot_our->our_list_lock);
1291 spin_unlock(&ou->ou_lock);
1293 LASSERT(oth->ot_super.th_wait_submit == 1);
1294 CDEBUG(D_INFO, "%s: version "LPU64" oth:version %p:"LPU64"\n",
1295 osp->opd_obd->obd_name, ou->ou_version, oth,
1296 oth->ot_our->our_version);
1302 * Get next OSP update request in the sending list
1303 * Get next OSP update request in the sending list by version number, next
1305 * 1. transaction which does not have a version number.
1306 * 2. transaction whose version == opd_rpc_version.
1308 * \param [in] ou osp update structure.
1309 * \param [out] ourp the pointer holding the next update request.
1311 * \retval true if getting the next transaction.
1312 * \retval false if not getting the next transaction.
1315 osp_get_next_request(struct osp_updates *ou, struct osp_update_request **ourp)
1317 struct osp_update_request *our;
1318 struct osp_update_request *tmp;
1319 bool got_req = false;
1321 spin_lock(&ou->ou_lock);
1322 list_for_each_entry_safe(our, tmp, &ou->ou_list, our_list) {
1323 LASSERT(our->our_th != NULL);
1324 CDEBUG(D_HA, "ou %p version "LPU64" rpc_version "LPU64"\n",
1325 ou, our->our_version, ou->ou_rpc_version);
1326 spin_lock(&our->our_list_lock);
1327 /* Find next osp_update_request in the list */
1328 if (our->our_version == ou->ou_rpc_version &&
1329 our->our_req_ready) {
1330 list_del_init(&our->our_list);
1331 spin_unlock(&our->our_list_lock);
1336 spin_unlock(&our->our_list_lock);
1338 spin_unlock(&ou->ou_lock);
1344 * Invalidate update request
1346 * Invalidate update request in the OSP sending list, so all of
1347 * requests in the sending list will return error, which happens
1348 * when it finds one update (with writing llog) requests fails or
1349 * the OSP is evicted by remote target. see osp_send_update_thread().
1351 * \param[in] osp OSP device whose update requests will be
1354 void osp_invalidate_request(struct osp_device *osp)
1357 struct osp_updates *ou = osp->opd_update;
1358 struct osp_update_request *our;
1359 struct osp_update_request *tmp;
1367 rc = lu_env_init(&env, osp->opd_dt_dev.dd_lu_dev.ld_type->ldt_ctx_tags);
1369 CERROR("%s: init env error: rc = %d\n", osp->opd_obd->obd_name,
1374 INIT_LIST_HEAD(&list);
1376 spin_lock(&ou->ou_lock);
1377 /* invalidate all of request in the sending list */
1378 list_for_each_entry_safe(our, tmp, &ou->ou_list, our_list) {
1379 spin_lock(&our->our_list_lock);
1380 if (our->our_req_ready)
1381 list_move(&our->our_list, &list);
1383 list_del_init(&our->our_list);
1385 if (our->our_th->ot_super.th_result == 0)
1386 our->our_th->ot_super.th_result = -EIO;
1388 if (our->our_version >= ou->ou_rpc_version)
1389 ou->ou_rpc_version = our->our_version + 1;
1390 spin_unlock(&our->our_list_lock);
1392 CDEBUG(D_HA, "%s invalidate our %p\n", osp->opd_obd->obd_name,
1396 spin_unlock(&ou->ou_lock);
1398 /* invalidate all of request in the sending list */
1399 list_for_each_entry_safe(our, tmp, &list, our_list) {
1400 spin_lock(&our->our_list_lock);
1401 list_del_init(&our->our_list);
1402 spin_unlock(&our->our_list_lock);
1403 osp_trans_callback(&env, our->our_th,
1404 our->our_th->ot_super.th_result);
1405 osp_thandle_put(our->our_th);
1411 * Sending update thread
1413 * Create thread to send update request to other MDTs, this thread will pull
1414 * out update request from the list in OSP by version number, i.e. it will
1415 * make sure the update request with lower version number will be sent first.
1417 * \param[in] arg hold the OSP device.
1419 * \retval 0 if the thread is created successfully.
1420 * \retal negative error if the thread is not created
1423 int osp_send_update_thread(void *arg)
1426 struct osp_device *osp = arg;
1427 struct l_wait_info lwi = { 0 };
1428 struct osp_updates *ou = osp->opd_update;
1429 struct ptlrpc_thread *thread = &osp->opd_update_thread;
1430 struct osp_update_request *our = NULL;
1434 LASSERT(ou != NULL);
1435 rc = lu_env_init(&env, osp->opd_dt_dev.dd_lu_dev.ld_type->ldt_ctx_tags);
1437 CERROR("%s: init env error: rc = %d\n", osp->opd_obd->obd_name,
1442 thread->t_flags = SVC_RUNNING;
1443 wake_up(&thread->t_ctl_waitq);
1446 l_wait_event(ou->ou_waitq,
1447 !osp_send_update_thread_running(osp) ||
1448 osp_get_next_request(ou, &our), &lwi);
1450 if (!osp_send_update_thread_running(osp)) {
1452 osp_trans_callback(&env, our->our_th, -EINTR);
1453 osp_thandle_put(our->our_th);
1458 LASSERT(our->our_th != NULL);
1459 if (our->our_th->ot_super.th_result != 0) {
1460 osp_trans_callback(&env, our->our_th,
1461 our->our_th->ot_super.th_result);
1462 rc = our->our_th->ot_super.th_result;
1463 } else if (OBD_FAIL_CHECK(OBD_FAIL_INVALIDATE_UPDATE)) {
1465 osp_trans_callback(&env, our->our_th, rc);
1467 rc = osp_send_update_req(&env, osp, our);
1470 /* Update the rpc version */
1471 spin_lock(&ou->ou_lock);
1472 if (our->our_version == ou->ou_rpc_version)
1473 ou->ou_rpc_version++;
1474 spin_unlock(&ou->ou_lock);
1476 /* If one update request fails, let's fail all of the requests
1477 * in the sending list, because the request in the sending
1478 * list are dependent on either other, continue sending these
1479 * request might cause llog or filesystem corruption */
1481 osp_invalidate_request(osp);
1483 /* Balanced for thandle_get in osp_check_and_set_rpc_version */
1484 osp_thandle_put(our->our_th);
1487 thread->t_flags = SVC_STOPPED;
1489 wake_up(&thread->t_ctl_waitq);
1495 * The OSP layer dt_device_operations::dt_trans_start() interface
1496 * to start the transaction.
1498 * If the transaction is a remote transaction, then related remote
1499 * updates will be triggered in the osp_trans_stop().
1500 * Please refer to osp_trans_create() for transaction type.
1502 * \param[in] env pointer to the thread context
1503 * \param[in] dt pointer to the OSP dt_device
1504 * \param[in] th pointer to the transaction handler
1506 * \retval 0 for success
1507 * \retval negative error number on failure
1509 int osp_trans_start(const struct lu_env *env, struct dt_device *dt,
1512 struct osp_thandle *oth = thandle_to_osp_thandle(th);
1514 if (oth->ot_super.th_sync)
1515 oth->ot_our->our_flags |= UPDATE_FL_SYNC;
1516 /* For remote thandle, if there are local thandle, start it here*/
1517 if (is_only_remote_trans(th) && oth->ot_storage_th != NULL)
1518 return dt_trans_start(env, oth->ot_storage_th->th_dev,
1519 oth->ot_storage_th);
1524 * The OSP layer dt_device_operations::dt_trans_stop() interface
1525 * to stop the transaction.
1527 * If the transaction is a remote transaction, related remote
1528 * updates will be triggered at the end of this function.
1530 * For synchronous mode update or any failed update, the request
1531 * will be destroyed explicitly when the osp_trans_stop().
1533 * Please refer to osp_trans_create() for transaction type.
1535 * \param[in] env pointer to the thread context
1536 * \param[in] dt pointer to the OSP dt_device
1537 * \param[in] th pointer to the transaction handler
1539 * \retval 0 for success
1540 * \retval negative error number on failure
1542 int osp_trans_stop(const struct lu_env *env, struct dt_device *dt,
1545 struct osp_thandle *oth = thandle_to_osp_thandle(th);
1546 struct osp_update_request *our = oth->ot_our;
1547 struct osp_device *osp = dt2osp_dev(dt);
1551 /* For remote transaction, if there is local storage thandle,
1553 if (oth->ot_storage_th != NULL && th->th_top == NULL) {
1554 dt_trans_stop(env, oth->ot_storage_th->th_dev,
1555 oth->ot_storage_th);
1556 oth->ot_storage_th = NULL;
1559 if (our == NULL || list_empty(&our->our_req_list)) {
1560 osp_trans_callback(env, oth, th->th_result);
1561 GOTO(out, rc = th->th_result);
1564 if (!osp->opd_connect_mdt) {
1565 osp_trans_callback(env, oth, th->th_result);
1566 rc = osp_send_update_req(env, osp, oth->ot_our);
1570 if (osp->opd_update == NULL ||
1571 !osp_send_update_thread_running(osp)) {
1572 osp_trans_callback(env, oth, -EIO);
1573 GOTO(out, rc = -EIO);
1576 CDEBUG(D_HA, "%s: add oth %p with version "LPU64"\n",
1577 osp->opd_obd->obd_name, oth, our->our_version);
1579 LASSERT(our->our_req_ready == 0);
1580 spin_lock(&our->our_list_lock);
1581 if (likely(!list_empty(&our->our_list))) {
1582 /* notify sending thread */
1583 our->our_req_ready = 1;
1584 wake_up(&osp->opd_update->ou_waitq);
1585 spin_unlock(&our->our_list_lock);
1586 } else if (th->th_result == 0) {
1587 /* if the request does not needs to be serialized,
1588 * read-only request etc, let's send it right away */
1589 spin_unlock(&our->our_list_lock);
1590 rc = osp_send_update_req(env, osp, our);
1592 spin_unlock(&our->our_list_lock);
1593 osp_trans_callback(env, oth, th->th_result);
1596 osp_thandle_put(oth);