1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
2 * vim:expandtab:shiftwidth=8:tabstop=8:
4 * Copyright (C) 2001-2003 Cluster File Systems, Inc.
5 * Author Peter Braam <braam@clusterfs.com>
7 * This file is part of the Lustre file system, http://www.lustre.org
8 * Lustre is a trademark of Cluster File Systems, Inc.
10 * You may have signed or agreed to another license before downloading
11 * this software. If so, you are bound by the terms and conditions
12 * of that agreement, and the following does not apply to you. See the
13 * LICENSE file included with this distribution for more information.
15 * If you did not agree to a different license, then this copy of Lustre
16 * is open source software; you can redistribute it and/or modify it
17 * under the terms of version 2 of the GNU General Public License as
18 * published by the Free Software Foundation.
20 * In either case, Lustre is distributed in the hope that it will be
21 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
22 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23 * license text for more details.
25 * For testing and management it is treated as an obd_device,
26 * although * it does not export a full OBD method table (the
27 * requests are coming * in over the wire, so object target modules
28 * do not have a full * method table.)
33 # define EXPORT_SYMTAB
35 #define DEBUG_SUBSYSTEM S_OSC
38 # include <libcfs/libcfs.h>
39 #else /* __KERNEL__ */
40 # include <liblustre.h>
43 #include <lustre_dlm.h>
44 #include <libcfs/kp30.h>
45 #include <lustre_net.h>
46 #include <lustre/lustre_user.h>
54 #include <lustre_ha.h>
55 #include <lprocfs_status.h>
56 #include <lustre_log.h>
57 #include <lustre_debug.h>
58 #include <lustre_param.h>
59 #include "osc_internal.h"
61 static quota_interface_t *quota_interface = NULL;
62 extern quota_interface_t osc_quota_interface;
64 static void osc_release_ppga(struct brw_page **ppga, obd_count count);
66 /* Pack OSC object metadata for disk storage (LE byte order). */
67 static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
68 struct lov_stripe_md *lsm)
73 lmm_size = sizeof(**lmmp);
78 OBD_FREE(*lmmp, lmm_size);
84 OBD_ALLOC(*lmmp, lmm_size);
90 LASSERT(lsm->lsm_object_id);
91 LASSERT(lsm->lsm_object_gr);
92 (*lmmp)->lmm_object_id = cpu_to_le64(lsm->lsm_object_id);
93 (*lmmp)->lmm_object_gr = cpu_to_le64(lsm->lsm_object_gr);
99 /* Unpack OSC object metadata from disk storage (LE byte order). */
100 static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
101 struct lov_mds_md *lmm, int lmm_bytes)
107 if (lmm_bytes < sizeof (*lmm)) {
108 CERROR("lov_mds_md too small: %d, need %d\n",
109 lmm_bytes, (int)sizeof(*lmm));
112 /* XXX LOV_MAGIC etc check? */
114 if (lmm->lmm_object_id == 0) {
115 CERROR("lov_mds_md: zero lmm_object_id\n");
120 lsm_size = lov_stripe_md_size(1);
124 if (*lsmp != NULL && lmm == NULL) {
125 OBD_FREE((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
126 OBD_FREE(*lsmp, lsm_size);
132 OBD_ALLOC(*lsmp, lsm_size);
135 OBD_ALLOC((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
136 if ((*lsmp)->lsm_oinfo[0] == NULL) {
137 OBD_FREE(*lsmp, lsm_size);
140 loi_init((*lsmp)->lsm_oinfo[0]);
144 /* XXX zero *lsmp? */
145 (*lsmp)->lsm_object_id = le64_to_cpu (lmm->lmm_object_id);
146 (*lsmp)->lsm_object_gr = le64_to_cpu (lmm->lmm_object_gr);
147 LASSERT((*lsmp)->lsm_object_id);
148 LASSERT((*lsmp)->lsm_object_gr);
151 (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
156 static inline void osc_pack_capa(struct ptlrpc_request *req, int offset,
157 struct ost_body *body, void *capa)
159 struct obd_capa *oc = (struct obd_capa *)capa;
160 struct lustre_capa *c;
165 c = lustre_msg_buf(req->rq_reqmsg, offset, sizeof(*c));
168 body->oa.o_valid |= OBD_MD_FLOSSCAPA;
169 DEBUG_CAPA(D_SEC, c, "pack");
172 static inline void osc_pack_req_body(struct ptlrpc_request *req, int offset,
173 struct obd_info *oinfo)
175 struct ost_body *body;
177 body = lustre_msg_buf(req->rq_reqmsg, offset, sizeof(*body));
178 body->oa = *oinfo->oi_oa;
179 osc_pack_capa(req, offset + 1, body, oinfo->oi_capa);
182 static int osc_getattr_interpret(struct ptlrpc_request *req,
183 struct osc_async_args *aa, int rc)
185 struct ost_body *body;
191 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
192 lustre_swab_ost_body);
194 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
195 memcpy(aa->aa_oi->oi_oa, &body->oa, sizeof(*aa->aa_oi->oi_oa));
197 /* This should really be sent by the OST */
198 aa->aa_oi->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
199 aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
201 CERROR("can't unpack ost_body\n");
203 aa->aa_oi->oi_oa->o_valid = 0;
206 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
210 static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
211 struct ptlrpc_request_set *set)
213 struct ptlrpc_request *req;
214 struct ost_body *body;
215 int size[3] = { sizeof(struct ptlrpc_body), sizeof(*body) };
216 struct osc_async_args *aa;
219 size[REQ_REC_OFF + 1] = oinfo->oi_capa ? sizeof(struct lustre_capa) : 0;
220 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
221 OST_GETATTR, 3, size,NULL);
225 osc_pack_req_body(req, REQ_REC_OFF, oinfo);
227 ptlrpc_req_set_repsize(req, 2, size);
228 req->rq_interpret_reply = osc_getattr_interpret;
230 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
231 aa = (struct osc_async_args *)&req->rq_async_args;
234 ptlrpc_set_add_req(set, req);
238 static int osc_getattr(struct obd_export *exp, struct obd_info *oinfo)
240 struct ptlrpc_request *req;
241 struct ost_body *body;
242 int rc, size[3] = { sizeof(struct ptlrpc_body), sizeof(*body) };
245 size[REQ_REC_OFF + 1] = oinfo->oi_capa ? sizeof(struct lustre_capa) : 0;
246 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
247 OST_GETATTR, 3, size, NULL);
251 osc_pack_req_body(req, REQ_REC_OFF, oinfo);
253 ptlrpc_req_set_repsize(req, 2, size);
255 rc = ptlrpc_queue_wait(req);
257 CERROR("%s failed: rc = %d\n", __FUNCTION__, rc);
261 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
262 lustre_swab_ost_body);
264 CERROR ("can't unpack ost_body\n");
265 GOTO (out, rc = -EPROTO);
268 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
269 *oinfo->oi_oa = body->oa;
271 /* This should really be sent by the OST */
272 oinfo->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
273 oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
277 ptlrpc_req_finished(req);
281 static int osc_setattr(struct obd_export *exp, struct obd_info *oinfo,
282 struct obd_trans_info *oti)
284 struct ptlrpc_request *req;
285 struct ost_body *body;
286 int rc, size[3] = { sizeof(struct ptlrpc_body), sizeof(*body) };
289 LASSERT(!(oinfo->oi_oa->o_valid & OBD_MD_FLGROUP) ||
290 oinfo->oi_oa->o_gr > 0);
291 size[REQ_REC_OFF + 1] = oinfo->oi_capa ? sizeof(struct lustre_capa) : 0;
292 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
293 OST_SETATTR, 3, size, NULL);
297 osc_pack_req_body(req, REQ_REC_OFF, oinfo);
299 ptlrpc_req_set_repsize(req, 2, size);
301 rc = ptlrpc_queue_wait(req);
305 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
306 lustre_swab_ost_body);
308 GOTO(out, rc = -EPROTO);
310 *oinfo->oi_oa = body->oa;
314 ptlrpc_req_finished(req);
318 static int osc_setattr_interpret(struct ptlrpc_request *req,
319 struct osc_async_args *aa, int rc)
321 struct ost_body *body;
327 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
328 lustre_swab_ost_body);
330 CERROR("can't unpack ost_body\n");
331 GOTO(out, rc = -EPROTO);
334 *aa->aa_oi->oi_oa = body->oa;
336 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
340 static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
341 struct obd_trans_info *oti,
342 struct ptlrpc_request_set *rqset)
344 struct ptlrpc_request *req;
345 int size[3] = { sizeof(struct ptlrpc_body), sizeof(struct ost_body) };
346 struct osc_async_args *aa;
349 size[REQ_REC_OFF + 1] = oinfo->oi_capa ? sizeof(struct lustre_capa) : 0;
350 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
351 OST_SETATTR, 3, size, NULL);
355 osc_pack_req_body(req, REQ_REC_OFF, oinfo);
356 if (oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE) {
358 *obdo_logcookie(oinfo->oi_oa) = *oti->oti_logcookies;
361 ptlrpc_req_set_repsize(req, 2, size);
362 /* do mds to ost setattr asynchronouly */
364 /* Do not wait for response. */
365 ptlrpcd_add_req(req);
367 req->rq_interpret_reply = osc_setattr_interpret;
369 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
370 aa = (struct osc_async_args *)&req->rq_async_args;
373 ptlrpc_set_add_req(rqset, req);
379 int osc_real_create(struct obd_export *exp, struct obdo *oa,
380 struct lov_stripe_md **ea, struct obd_trans_info *oti)
382 struct ptlrpc_request *req;
383 struct ost_body *body;
384 struct lov_stripe_md *lsm;
385 int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
393 rc = obd_alloc_memmd(exp, &lsm);
398 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
399 OST_CREATE, 2, size, NULL);
401 GOTO(out, rc = -ENOMEM);
403 body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
406 ptlrpc_req_set_repsize(req, 2, size);
407 if (oa->o_valid & OBD_MD_FLINLINE) {
408 LASSERT((oa->o_valid & OBD_MD_FLFLAGS) &&
409 oa->o_flags == OBD_FL_DELORPHAN);
411 "delorphan from OST integration");
412 /* Don't resend the delorphan req */
413 req->rq_no_resend = req->rq_no_delay = 1;
416 rc = ptlrpc_queue_wait(req);
420 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
421 lustre_swab_ost_body);
423 CERROR ("can't unpack ost_body\n");
424 GOTO (out_req, rc = -EPROTO);
429 /* This should really be sent by the OST */
430 oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
431 oa->o_valid |= OBD_MD_FLBLKSZ;
433 /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
434 * have valid lsm_oinfo data structs, so don't go touching that.
435 * This needs to be fixed in a big way.
437 lsm->lsm_object_id = oa->o_id;
438 lsm->lsm_object_gr = oa->o_gr;
442 oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);
444 if (oa->o_valid & OBD_MD_FLCOOKIE) {
445 if (!oti->oti_logcookies)
446 oti_alloc_cookies(oti, 1);
447 *oti->oti_logcookies = *obdo_logcookie(oa);
451 CDEBUG(D_HA, "transno: "LPD64"\n",
452 lustre_msg_get_transno(req->rq_repmsg));
454 ptlrpc_req_finished(req);
457 obd_free_memmd(exp, &lsm);
461 static int osc_punch_interpret(struct ptlrpc_request *req,
462 struct osc_async_args *aa, int rc)
464 struct ost_body *body;
470 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof (*body),
471 lustre_swab_ost_body);
473 CERROR ("can't unpack ost_body\n");
474 GOTO(out, rc = -EPROTO);
477 *aa->aa_oi->oi_oa = body->oa;
479 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
483 static int osc_punch(struct obd_export *exp, struct obd_info *oinfo,
484 struct obd_trans_info *oti,
485 struct ptlrpc_request_set *rqset)
487 struct ptlrpc_request *req;
488 struct osc_async_args *aa;
489 struct ost_body *body;
490 int size[3] = { sizeof(struct ptlrpc_body), sizeof(*body) };
498 size[REQ_REC_OFF + 1] = oinfo->oi_capa? sizeof(struct lustre_capa) : 0;
499 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
500 OST_PUNCH, 3, size, NULL);
504 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
506 osc_pack_req_body(req, REQ_REC_OFF, oinfo);
507 /* overload the size and blocks fields in the oa with start/end */
508 body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
509 body->oa.o_size = oinfo->oi_policy.l_extent.start;
510 body->oa.o_blocks = oinfo->oi_policy.l_extent.end;
511 body->oa.o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);
513 ptlrpc_req_set_repsize(req, 2, size);
515 req->rq_interpret_reply = osc_punch_interpret;
516 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
517 aa = (struct osc_async_args *)&req->rq_async_args;
519 ptlrpc_set_add_req(rqset, req);
524 static int osc_sync(struct obd_export *exp, struct obdo *oa,
525 struct lov_stripe_md *md, obd_size start, obd_size end,
528 struct ptlrpc_request *req;
529 struct ost_body *body;
530 int rc, size[3] = { sizeof(struct ptlrpc_body), sizeof(*body) };
538 size[REQ_REC_OFF + 1] = capa ? sizeof(struct lustre_capa) : 0;
540 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
541 OST_SYNC, 3, size, NULL);
545 /* overload the size and blocks fields in the oa with start/end */
546 body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
548 body->oa.o_size = start;
549 body->oa.o_blocks = end;
550 body->oa.o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);
552 osc_pack_capa(req, REQ_REC_OFF + 1, body, capa);
554 ptlrpc_req_set_repsize(req, 2, size);
556 rc = ptlrpc_queue_wait(req);
560 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
561 lustre_swab_ost_body);
563 CERROR ("can't unpack ost_body\n");
564 GOTO (out, rc = -EPROTO);
571 ptlrpc_req_finished(req);
575 /* Find and cancel locally locks matched by @mode in the resource found by
576 * @objid. Found locks are added into @cancel list. Returns the amount of
577 * locks added to @cancels list. */
578 static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
579 struct list_head *cancels, ldlm_mode_t mode,
582 struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
583 struct ldlm_res_id res_id = { .name = { oa->o_id, 0, oa->o_gr, 0 } };
584 struct ldlm_resource *res = ldlm_resource_get(ns, NULL, &res_id, 0, 0);
591 count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
592 lock_flags, 0, NULL);
593 ldlm_resource_putref(res);
597 /* Destroy requests can be async always on the client, and we don't even really
598 * care about the return code since the client cannot do anything at all about
600 * When the MDS is unlinking a filename, it saves the file objects into a
601 * recovery llog, and these object records are cancelled when the OST reports
602 * they were destroyed and sync'd to disk (i.e. transaction committed).
603 * If the client dies, or the OST is down when the object should be destroyed,
604 * the records are not cancelled, and when the OST reconnects to the MDS next,
605 * it will retrieve the llog unlink logs and then sends the log cancellation
606 * cookies to the MDS after committing destroy transactions. */
607 static int osc_destroy(struct obd_export *exp, struct obdo *oa,
608 struct lov_stripe_md *ea, struct obd_trans_info *oti,
609 struct obd_export *md_export)
611 CFS_LIST_HEAD(cancels);
612 struct ptlrpc_request *req;
613 struct ost_body *body;
614 int size[3] = { sizeof(struct ptlrpc_body), sizeof(*body), 0 };
615 int count, bufcount = 2;
623 count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
624 LDLM_FL_DISCARD_DATA);
625 if (exp_connect_cancelset(exp) && count) {
627 size[REQ_REC_OFF + 1] = ldlm_request_bufsize(count,
630 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
631 OST_DESTROY, bufcount, size, NULL);
632 if (exp_connect_cancelset(exp) && req)
633 ldlm_cli_cancel_list(&cancels, count, req, REQ_REC_OFF + 1, 0);
635 ldlm_lock_list_put(&cancels, l_bl_ast, count);
640 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
642 body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
643 if (oti != NULL && oa->o_valid & OBD_MD_FLCOOKIE)
644 memcpy(obdo_logcookie(oa), oti->oti_logcookies,
645 sizeof(*oti->oti_logcookies));
648 ptlrpc_req_set_repsize(req, 2, size);
650 ptlrpcd_add_req(req);
654 static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
657 obd_flag bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;
659 LASSERT(!(oa->o_valid & bits));
662 client_obd_list_lock(&cli->cl_loi_list_lock);
663 oa->o_dirty = cli->cl_dirty;
664 if (cli->cl_dirty > cli->cl_dirty_max) {
665 CERROR("dirty %lu > dirty_max %lu\n",
666 cli->cl_dirty, cli->cl_dirty_max);
668 } else if (atomic_read(&obd_dirty_pages) > obd_max_dirty_pages) {
669 CERROR("dirty %d > system dirty_max %d\n",
670 atomic_read(&obd_dirty_pages), obd_max_dirty_pages);
672 } else if (cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff) {
673 CERROR("dirty %lu - dirty_max %lu too big???\n",
674 cli->cl_dirty, cli->cl_dirty_max);
677 long max_in_flight = (cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT)*
678 (cli->cl_max_rpcs_in_flight + 1);
679 oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
681 oa->o_grant = cli->cl_avail_grant;
682 oa->o_dropped = cli->cl_lost_grant;
683 cli->cl_lost_grant = 0;
684 client_obd_list_unlock(&cli->cl_loi_list_lock);
685 CDEBUG(D_CACHE,"dirty: "LPU64" undirty: %u dropped %u grant: "LPU64"\n",
686 oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
689 /* caller must hold loi_list_lock */
690 static void osc_consume_write_grant(struct client_obd *cli,
691 struct brw_page *pga)
693 atomic_inc(&obd_dirty_pages);
694 cli->cl_dirty += CFS_PAGE_SIZE;
695 cli->cl_avail_grant -= CFS_PAGE_SIZE;
696 pga->flag |= OBD_BRW_FROM_GRANT;
697 CDEBUG(D_CACHE, "using %lu grant credits for brw %p page %p\n",
698 CFS_PAGE_SIZE, pga, pga->pg);
699 LASSERT(cli->cl_avail_grant >= 0);
702 /* the companion to osc_consume_write_grant, called when a brw has completed.
703 * must be called with the loi lock held. */
704 static void osc_release_write_grant(struct client_obd *cli,
705 struct brw_page *pga, int sent)
707 int blocksize = cli->cl_import->imp_obd->obd_osfs.os_bsize ? : 4096;
710 if (!(pga->flag & OBD_BRW_FROM_GRANT)) {
715 pga->flag &= ~OBD_BRW_FROM_GRANT;
716 atomic_dec(&obd_dirty_pages);
717 cli->cl_dirty -= CFS_PAGE_SIZE;
719 cli->cl_lost_grant += CFS_PAGE_SIZE;
720 CDEBUG(D_CACHE, "lost grant: %lu avail grant: %lu dirty: %lu\n",
721 cli->cl_lost_grant, cli->cl_avail_grant, cli->cl_dirty);
722 } else if (CFS_PAGE_SIZE != blocksize && pga->count != CFS_PAGE_SIZE) {
723 /* For short writes we shouldn't count parts of pages that
724 * span a whole block on the OST side, or our accounting goes
725 * wrong. Should match the code in filter_grant_check. */
726 int offset = pga->off & ~CFS_PAGE_MASK;
727 int count = pga->count + (offset & (blocksize - 1));
728 int end = (offset + pga->count) & (blocksize - 1);
730 count += blocksize - end;
732 cli->cl_lost_grant += CFS_PAGE_SIZE - count;
733 CDEBUG(D_CACHE, "lost %lu grant: %lu avail: %lu dirty: %lu\n",
734 CFS_PAGE_SIZE - count, cli->cl_lost_grant,
735 cli->cl_avail_grant, cli->cl_dirty);
741 static unsigned long rpcs_in_flight(struct client_obd *cli)
743 return cli->cl_r_in_flight + cli->cl_w_in_flight;
746 /* caller must hold loi_list_lock */
747 void osc_wake_cache_waiters(struct client_obd *cli)
749 struct list_head *l, *tmp;
750 struct osc_cache_waiter *ocw;
753 list_for_each_safe(l, tmp, &cli->cl_cache_waiters) {
754 /* if we can't dirty more, we must wait until some is written */
755 if ((cli->cl_dirty + CFS_PAGE_SIZE > cli->cl_dirty_max) ||
756 (atomic_read(&obd_dirty_pages) + 1 > obd_max_dirty_pages)) {
757 CDEBUG(D_CACHE, "no dirty room: dirty: %ld "
758 "osc max %ld, sys max %d\n", cli->cl_dirty,
759 cli->cl_dirty_max, obd_max_dirty_pages);
763 /* if still dirty cache but no grant wait for pending RPCs that
764 * may yet return us some grant before doing sync writes */
765 if (cli->cl_w_in_flight && cli->cl_avail_grant < CFS_PAGE_SIZE) {
766 CDEBUG(D_CACHE, "%u BRW writes in flight, no grant\n",
767 cli->cl_w_in_flight);
771 ocw = list_entry(l, struct osc_cache_waiter, ocw_entry);
772 list_del_init(&ocw->ocw_entry);
773 if (cli->cl_avail_grant < CFS_PAGE_SIZE) {
774 /* no more RPCs in flight to return grant, do sync IO */
775 ocw->ocw_rc = -EDQUOT;
776 CDEBUG(D_INODE, "wake oap %p for sync\n", ocw->ocw_oap);
778 osc_consume_write_grant(cli,
779 &ocw->ocw_oap->oap_brw_page);
782 cfs_waitq_signal(&ocw->ocw_waitq);
788 static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
790 client_obd_list_lock(&cli->cl_loi_list_lock);
791 cli->cl_avail_grant = ocd->ocd_grant;
792 client_obd_list_unlock(&cli->cl_loi_list_lock);
794 CDEBUG(D_CACHE, "setting cl_avail_grant: %ld cl_lost_grant: %ld\n",
795 cli->cl_avail_grant, cli->cl_lost_grant);
796 LASSERT(cli->cl_avail_grant >= 0);
799 static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
801 client_obd_list_lock(&cli->cl_loi_list_lock);
802 CDEBUG(D_CACHE, "got "LPU64" extra grant\n", body->oa.o_grant);
803 if (body->oa.o_valid & OBD_MD_FLGRANT)
804 cli->cl_avail_grant += body->oa.o_grant;
805 /* waiters are woken in brw_interpret_oap */
806 client_obd_list_unlock(&cli->cl_loi_list_lock);
809 /* We assume that the reason this OSC got a short read is because it read
810 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
811 * via the LOV, and it _knows_ it's reading inside the file, it's just that
812 * this stripe never got written at or beyond this stripe offset yet. */
813 static void handle_short_read(int nob_read, obd_count page_count,
814 struct brw_page **pga)
819 /* skip bytes read OK */
820 while (nob_read > 0) {
821 LASSERT (page_count > 0);
823 if (pga[i]->count > nob_read) {
824 /* EOF inside this page */
825 ptr = cfs_kmap(pga[i]->pg) +
826 (pga[i]->off & ~CFS_PAGE_MASK);
827 memset(ptr + nob_read, 0, pga[i]->count - nob_read);
828 cfs_kunmap(pga[i]->pg);
834 nob_read -= pga[i]->count;
839 /* zero remaining pages */
840 while (page_count-- > 0) {
841 ptr = cfs_kmap(pga[i]->pg) + (pga[i]->off & ~CFS_PAGE_MASK);
842 memset(ptr, 0, pga[i]->count);
843 cfs_kunmap(pga[i]->pg);
848 static int check_write_rcs(struct ptlrpc_request *req,
849 int requested_nob, int niocount,
850 obd_count page_count, struct brw_page **pga)
854 /* return error if any niobuf was in error */
855 remote_rcs = lustre_swab_repbuf(req, REQ_REC_OFF + 1,
856 sizeof(*remote_rcs) * niocount, NULL);
857 if (remote_rcs == NULL) {
858 CERROR("Missing/short RC vector on BRW_WRITE reply\n");
861 if (lustre_msg_swabbed(req->rq_repmsg))
862 for (i = 0; i < niocount; i++)
863 __swab32s(&remote_rcs[i]);
865 for (i = 0; i < niocount; i++) {
866 if (remote_rcs[i] < 0)
867 return(remote_rcs[i]);
869 if (remote_rcs[i] != 0) {
870 CERROR("rc[%d] invalid (%d) req %p\n",
871 i, remote_rcs[i], req);
876 if (req->rq_bulk->bd_nob_transferred != requested_nob) {
877 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
878 requested_nob, req->rq_bulk->bd_nob_transferred);
885 static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
887 if (p1->flag != p2->flag) {
888 unsigned mask = ~OBD_BRW_FROM_GRANT;
890 /* warn if we try to combine flags that we don't know to be
892 if ((p1->flag & mask) != (p2->flag & mask))
893 CERROR("is it ok to have flags 0x%x and 0x%x in the "
894 "same brw?\n", p1->flag, p2->flag);
898 return (p1->off + p1->count == p2->off);
901 static obd_count osc_checksum_bulk(int nob, obd_count pg_count,
902 struct brw_page **pga, int opc)
907 LASSERT (pg_count > 0);
908 while (nob > 0 && pg_count > 0) {
909 char *ptr = cfs_kmap(pga[i]->pg);
910 int off = pga[i]->off & ~CFS_PAGE_MASK;
911 int count = pga[i]->count > nob ? nob : pga[i]->count;
913 /* corrupt the data before we compute the checksum, to
914 * simulate an OST->client data error */
915 if (i == 0 && opc == OST_READ &&
916 OBD_FAIL_CHECK_ONCE(OBD_FAIL_OSC_CHECKSUM_RECEIVE))
917 memcpy(ptr + off, "bad1", min(4, nob));
918 cksum = crc32_le(cksum, ptr + off, count);
919 cfs_kunmap(pga[i]->pg);
920 LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d checksum %x\n",
923 nob -= pga[i]->count;
927 /* For sending we only compute the wrong checksum instead
928 * of corrupting the data so it is still correct on a redo */
929 if (opc == OST_WRITE && OBD_FAIL_CHECK_ONCE(OBD_FAIL_OSC_CHECKSUM_SEND))
935 static int osc_brw_prep_request(int cmd, struct client_obd *cli,struct obdo *oa,
936 struct lov_stripe_md *lsm, obd_count page_count,
937 struct brw_page **pga,
938 struct ptlrpc_request **reqp,
939 struct obd_capa *ocapa)
941 struct ptlrpc_request *req;
942 struct ptlrpc_bulk_desc *desc;
943 struct ost_body *body;
944 struct obd_ioobj *ioobj;
945 struct niobuf_remote *niobuf;
946 int size[5] = { sizeof(struct ptlrpc_body), sizeof(*body) };
947 int niocount, i, requested_nob, opc, rc;
948 struct ptlrpc_request_pool *pool;
949 struct lustre_capa *capa;
950 struct osc_brw_async_args *aa;
953 OBD_FAIL_RETURN(OBD_FAIL_OSC_BRW_PREP_REQ, -ENOMEM); /* Recoverable */
954 OBD_FAIL_RETURN(OBD_FAIL_OSC_BRW_PREP_REQ2, -EINVAL); /* Fatal */
956 if ((cmd & OBD_BRW_WRITE) != 0) {
958 pool = cli->cl_import->imp_rq_pool;
964 for (niocount = i = 1; i < page_count; i++) {
965 if (!can_merge_pages(pga[i - 1], pga[i]))
969 size[REQ_REC_OFF + 1] = sizeof(*ioobj);
970 size[REQ_REC_OFF + 2] = niocount * sizeof(*niobuf);
972 size[REQ_REC_OFF + 3] = sizeof(*capa);
974 req = ptlrpc_prep_req_pool(cli->cl_import, LUSTRE_OST_VERSION, opc, 5,
975 size, NULL, pool, NULL);
979 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
981 if (opc == OST_WRITE)
982 desc = ptlrpc_prep_bulk_imp (req, page_count,
983 BULK_GET_SOURCE, OST_BULK_PORTAL);
985 desc = ptlrpc_prep_bulk_imp (req, page_count,
986 BULK_PUT_SINK, OST_BULK_PORTAL);
988 GOTO(out, rc = -ENOMEM);
989 /* NB request now owns desc and will free it when it gets freed */
991 body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
992 ioobj = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 1, sizeof(*ioobj));
993 niobuf = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 2,
994 niocount * sizeof(*niobuf));
998 obdo_to_ioobj(oa, ioobj);
999 ioobj->ioo_bufcnt = niocount;
1001 capa = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 3,
1003 capa_cpy(capa, ocapa);
1004 body->oa.o_valid |= OBD_MD_FLOSSCAPA;
1007 LASSERT (page_count > 0);
1008 for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1009 struct brw_page *pg = pga[i];
1010 struct brw_page *pg_prev = pga[i - 1];
1012 LASSERT(pg->count > 0);
1013 LASSERTF((pg->off & ~CFS_PAGE_MASK) + pg->count <= CFS_PAGE_SIZE,
1014 "i: %d pg: %p off: "LPU64", count: %u\n", i, pg,
1015 pg->off, pg->count);
1017 LASSERTF(i == 0 || pg->off > pg_prev->off,
1018 "i %d p_c %u pg %p [pri %lu ind %lu] off "LPU64
1019 " prev_pg %p [pri %lu ind %lu] off "LPU64"\n",
1021 pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1022 pg_prev->pg, page_private(pg_prev->pg),
1023 pg_prev->pg->index, pg_prev->off);
1025 LASSERTF(i == 0 || pg->off > pg_prev->off,
1026 "i %d p_c %u\n", i, page_count);
1028 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1029 (pg->flag & OBD_BRW_SRVLOCK));
1031 ptlrpc_prep_bulk_page(desc, pg->pg, pg->off & ~CFS_PAGE_MASK,
1033 requested_nob += pg->count;
1035 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1037 niobuf->len += pg->count;
1039 niobuf->offset = pg->off;
1040 niobuf->len = pg->count;
1041 niobuf->flags = pg->flag;
1045 LASSERT((void *)(niobuf - niocount) ==
1046 lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 2,
1047 niocount * sizeof(*niobuf)));
1048 osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1050 /* size[REQ_REC_OFF] still sizeof (*body) */
1051 if (opc == OST_WRITE) {
1052 if (unlikely(cli->cl_checksum)) {
1053 body->oa.o_valid |= OBD_MD_FLCKSUM;
1054 body->oa.o_cksum = osc_checksum_bulk(requested_nob,
1057 CDEBUG(D_PAGE, "checksum at write origin: %x\n",
1059 /* save this in 'oa', too, for later checking */
1060 oa->o_valid |= OBD_MD_FLCKSUM;
1062 /* clear out the checksum flag, in case this is a
1063 * resend but cl_checksum is no longer set. b=11238 */
1064 oa->o_valid &= ~OBD_MD_FLCKSUM;
1066 oa->o_cksum = body->oa.o_cksum;
1067 /* 1 RC per niobuf */
1068 size[REPLY_REC_OFF + 1] = sizeof(__u32) * niocount;
1069 ptlrpc_req_set_repsize(req, 3, size);
1071 if (unlikely(cli->cl_checksum))
1072 body->oa.o_valid |= OBD_MD_FLCKSUM;
1073 /* 1 RC for the whole I/O */
1074 ptlrpc_req_set_repsize(req, 2, size);
1077 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
1078 aa = (struct osc_brw_async_args *)&req->rq_async_args;
1080 aa->aa_requested_nob = requested_nob;
1081 aa->aa_nio_count = niocount;
1082 aa->aa_page_count = page_count;
1086 INIT_LIST_HEAD(&aa->aa_oaps);
1092 ptlrpc_req_finished (req);
1096 static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
1097 __u32 client_cksum, __u32 server_cksum,
1098 int nob, obd_count page_count,
1099 struct brw_page **pga)
1104 if (server_cksum == client_cksum) {
1105 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1109 new_cksum = osc_checksum_bulk(nob, page_count, pga, OST_WRITE);
1111 if (new_cksum == server_cksum)
1112 msg = "changed on the client after we checksummed it - "
1113 "likely false positive due to mmap IO (bug 11742)";
1114 else if (new_cksum == client_cksum)
1115 msg = "changed in transit before arrival at OST";
1117 msg = "changed in transit AND doesn't match the original - "
1118 "likely false positive due to mmap IO (bug 11742)";
1120 LCONSOLE_ERROR_MSG(0x132, "BAD WRITE CHECKSUM: %s: from %s inum "
1121 LPU64"/"LPU64" object "LPU64"/"LPU64" extent "
1122 "["LPU64"-"LPU64"]\n",
1123 msg, libcfs_nid2str(peer->nid),
1124 oa->o_valid & OBD_MD_FLFID ? oa->o_fid : (__u64)0,
1125 oa->o_valid & OBD_MD_FLFID ? oa->o_generation :
1128 oa->o_valid & OBD_MD_FLGROUP ? oa->o_gr : (__u64)0,
1130 pga[page_count-1]->off + pga[page_count-1]->count - 1);
1131 CERROR("original client csum %x, server csum %x, client csum now %x\n",
1132 client_cksum, server_cksum, new_cksum);
1136 /* Note rc enters this function as number of bytes transferred */
1137 static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
1139 struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
1140 const lnet_process_id_t *peer =
1141 &req->rq_import->imp_connection->c_peer;
1142 struct client_obd *cli = aa->aa_cli;
1143 struct ost_body *body;
1144 __u32 client_cksum = 0;
1147 if (rc < 0 && rc != -EDQUOT)
1150 LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
1151 body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
1152 lustre_swab_ost_body);
1154 CERROR ("Can't unpack body\n");
1158 /* set/clear over quota flag for a uid/gid */
1159 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
1160 body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA))
1161 lquota_setdq(quota_interface, cli, body->oa.o_uid,
1162 body->oa.o_gid, body->oa.o_valid,
1168 if (unlikely(aa->aa_oa->o_valid & OBD_MD_FLCKSUM))
1169 client_cksum = aa->aa_oa->o_cksum; /* save for later */
1171 osc_update_grant(cli, body);
1173 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1175 CERROR ("Unexpected +ve rc %d\n", rc);
1178 LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);
1180 if (unlikely((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) &&
1182 check_write_checksum(&body->oa, peer, client_cksum,
1184 aa->aa_requested_nob,
1189 if (sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
1192 rc = check_write_rcs(req, aa->aa_requested_nob,aa->aa_nio_count,
1193 aa->aa_page_count, aa->aa_ppga);
1197 /* The rest of this function executes only for OST_READs */
1198 if (rc > aa->aa_requested_nob) {
1199 CERROR("Unexpected rc %d (%d requested)\n", rc,
1200 aa->aa_requested_nob);
1204 if (rc != req->rq_bulk->bd_nob_transferred) {
1205 CERROR ("Unexpected rc %d (%d transferred)\n",
1206 rc, req->rq_bulk->bd_nob_transferred);
1210 if (rc < aa->aa_requested_nob)
1211 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
1213 if (sptlrpc_cli_unwrap_bulk_read(req, rc, aa->aa_page_count,
1215 GOTO(out, rc = -EAGAIN);
1217 if (unlikely(body->oa.o_valid & OBD_MD_FLCKSUM)) {
1218 static int cksum_counter;
1219 __u32 server_cksum = body->oa.o_cksum;
1223 client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
1224 aa->aa_ppga, OST_READ);
1226 if (peer->nid == req->rq_bulk->bd_sender) {
1230 router = libcfs_nid2str(req->rq_bulk->bd_sender);
1233 if (server_cksum == ~0 && rc > 0) {
1234 CERROR("Protocol error: server %s set the 'checksum' "
1235 "bit, but didn't send a checksum. Not fatal, "
1236 "but please tell CFS.\n",
1237 libcfs_nid2str(peer->nid));
1238 } else if (server_cksum != client_cksum) {
1239 LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from "
1240 "%s%s%s inum "LPU64"/"LPU64" object "
1241 LPU64"/"LPU64" extent "
1242 "["LPU64"-"LPU64"]\n",
1243 req->rq_import->imp_obd->obd_name,
1244 libcfs_nid2str(peer->nid),
1246 body->oa.o_valid & OBD_MD_FLFID ?
1247 body->oa.o_fid : (__u64)0,
1248 body->oa.o_valid & OBD_MD_FLFID ?
1249 body->oa.o_generation :(__u64)0,
1251 body->oa.o_valid & OBD_MD_FLGROUP ?
1252 body->oa.o_gr : (__u64)0,
1253 aa->aa_ppga[0]->off,
1254 aa->aa_ppga[aa->aa_page_count-1]->off +
1255 aa->aa_ppga[aa->aa_page_count-1]->count -
1257 CERROR("client %x, server %x\n",
1258 client_cksum, server_cksum);
1260 aa->aa_oa->o_cksum = client_cksum;
1264 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1267 } else if (unlikely(client_cksum)) {
1268 static int cksum_missed;
1271 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
1272 CERROR("Checksum %u requested from %s but not sent\n",
1273 cksum_missed, libcfs_nid2str(peer->nid));
1279 *aa->aa_oa = body->oa;
1284 static int osc_brw_internal(int cmd, struct obd_export *exp, struct obdo *oa,
1285 struct lov_stripe_md *lsm,
1286 obd_count page_count, struct brw_page **pga,
1287 struct obd_capa *ocapa)
1289 struct ptlrpc_request *req;
1293 struct l_wait_info lwi;
1297 cfs_waitq_init(&waitq);
1300 rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
1301 page_count, pga, &req, ocapa);
1305 rc = ptlrpc_queue_wait(req);
1307 if (rc == -ETIMEDOUT && req->rq_resend) {
1308 DEBUG_REQ(D_HA, req, "BULK TIMEOUT");
1309 ptlrpc_req_finished(req);
1313 rc = osc_brw_fini_request(req, rc);
1315 ptlrpc_req_finished(req);
1316 if (osc_recoverable_error(rc)) {
1318 if (!osc_should_resend(resends, &exp->exp_obd->u.cli)) {
1319 CERROR("too many resend retries, returning error\n");
1323 lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(resends), NULL, NULL, NULL);
1324 l_wait_event(waitq, 0, &lwi);
1332 int osc_brw_redo_request(struct ptlrpc_request *request,
1333 struct osc_brw_async_args *aa)
1335 struct ptlrpc_request *new_req;
1336 struct ptlrpc_request_set *set = request->rq_set;
1337 struct osc_brw_async_args *new_aa;
1338 struct osc_async_page *oap;
1342 if (!osc_should_resend(aa->aa_resends, aa->aa_cli)) {
1343 CERROR("too many resend retries, returning error\n");
1347 DEBUG_REQ(D_ERROR, request, "redo for recoverable error");
1349 body = lustre_msg_buf(request->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
1350 if (body->oa.o_valid & OBD_MD_FLOSSCAPA)
1351 ocapa = lustre_unpack_capa(request->rq_reqmsg,
1354 rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
1355 OST_WRITE ? OBD_BRW_WRITE :OBD_BRW_READ,
1356 aa->aa_cli, aa->aa_oa,
1357 NULL /* lsm unused by osc currently */,
1358 aa->aa_page_count, aa->aa_ppga,
1359 &new_req, NULL /* ocapa */);
1363 client_obd_list_lock(&aa->aa_cli->cl_loi_list_lock);
1365 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
1366 if (oap->oap_request != NULL) {
1367 LASSERTF(request == oap->oap_request,
1368 "request %p != oap_request %p\n",
1369 request, oap->oap_request);
1370 if (oap->oap_interrupted) {
1371 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1372 ptlrpc_req_finished(new_req);
1377 /* New request takes over pga and oaps from old request.
1378 * Note that copying a list_head doesn't work, need to move it... */
1380 new_req->rq_interpret_reply = request->rq_interpret_reply;
1381 new_req->rq_async_args = request->rq_async_args;
1382 new_req->rq_sent = CURRENT_SECONDS + aa->aa_resends;
1384 new_aa = (struct osc_brw_async_args *)&new_req->rq_async_args;
1386 INIT_LIST_HEAD(&new_aa->aa_oaps);
1387 list_splice(&aa->aa_oaps, &new_aa->aa_oaps);
1388 INIT_LIST_HEAD(&aa->aa_oaps);
1390 list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
1391 if (oap->oap_request) {
1392 ptlrpc_req_finished(oap->oap_request);
1393 oap->oap_request = ptlrpc_request_addref(new_req);
1397 /* use ptlrpc_set_add_req is safe because interpret functions work
1398 * in check_set context. only one way exist with access to request
1399 * from different thread got -EINTR - this way protected with
1400 * cl_loi_list_lock */
1401 ptlrpc_set_add_req(set, new_req);
1403 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1405 DEBUG_REQ(D_INFO, new_req, "new request");
1409 static int brw_interpret(struct ptlrpc_request *req, void *data, int rc)
1411 struct osc_brw_async_args *aa = data;
1416 rc = osc_brw_fini_request(req, rc);
1417 if (osc_recoverable_error(rc)) {
1418 rc = osc_brw_redo_request(req, aa);
1422 if ((rc >= 0) && req->rq_set && req->rq_set->set_countp)
1423 atomic_add(nob, (atomic_t *)req->rq_set->set_countp);
1425 client_obd_list_lock(&aa->aa_cli->cl_loi_list_lock);
1426 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
1427 aa->aa_cli->cl_w_in_flight--;
1429 aa->aa_cli->cl_r_in_flight--;
1430 for (i = 0; i < aa->aa_page_count; i++)
1431 osc_release_write_grant(aa->aa_cli, aa->aa_ppga[i], 1);
1432 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1434 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
1439 static int async_internal(int cmd, struct obd_export *exp, struct obdo *oa,
1440 struct lov_stripe_md *lsm, obd_count page_count,
1441 struct brw_page **pga, struct ptlrpc_request_set *set,
1442 struct obd_capa *ocapa)
1444 struct ptlrpc_request *req;
1445 struct client_obd *cli = &exp->exp_obd->u.cli;
1447 struct osc_brw_async_args *aa;
1450 /* Consume write credits even if doing a sync write -
1451 * otherwise we may run out of space on OST due to grant. */
1452 if (cmd == OBD_BRW_WRITE) {
1453 spin_lock(&cli->cl_loi_list_lock);
1454 for (i = 0; i < page_count; i++) {
1455 if (cli->cl_avail_grant >= CFS_PAGE_SIZE)
1456 osc_consume_write_grant(cli, pga[i]);
1458 spin_unlock(&cli->cl_loi_list_lock);
1461 rc = osc_brw_prep_request(cmd, cli, oa, lsm, page_count, pga,
1464 aa = (struct osc_brw_async_args *)&req->rq_async_args;
1465 if (cmd == OBD_BRW_READ) {
1466 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
1467 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
1468 ptlrpc_lprocfs_brw(req, OST_READ, aa->aa_requested_nob);
1470 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
1471 lprocfs_oh_tally(&cli->cl_write_rpc_hist,
1472 cli->cl_w_in_flight);
1473 ptlrpc_lprocfs_brw(req, OST_WRITE, aa->aa_requested_nob);
1477 req->rq_interpret_reply = brw_interpret;
1478 ptlrpc_set_add_req(set, req);
1479 client_obd_list_lock(&cli->cl_loi_list_lock);
1480 if (cmd == OBD_BRW_READ)
1481 cli->cl_r_in_flight++;
1483 cli->cl_w_in_flight++;
1484 client_obd_list_unlock(&cli->cl_loi_list_lock);
1485 } else if (cmd == OBD_BRW_WRITE) {
1486 client_obd_list_lock(&cli->cl_loi_list_lock);
1487 for (i = 0; i < page_count; i++)
1488 osc_release_write_grant(cli, pga[i], 0);
1489 client_obd_list_unlock(&cli->cl_loi_list_lock);
1495 * ugh, we want disk allocation on the target to happen in offset order. we'll
1496 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
1497 * fine for our small page arrays and doesn't require allocation. its an
1498 * insertion sort that swaps elements that are strides apart, shrinking the
1499 * stride down until its '1' and the array is sorted.
1501 static void sort_brw_pages(struct brw_page **array, int num)
1504 struct brw_page *tmp;
1508 for (stride = 1; stride < num ; stride = (stride * 3) + 1)
1513 for (i = stride ; i < num ; i++) {
1516 while (j >= stride && array[j - stride]->off > tmp->off) {
1517 array[j] = array[j - stride];
1522 } while (stride > 1);
1525 static obd_count max_unfragmented_pages(struct brw_page **pg, obd_count pages)
1531 LASSERT (pages > 0);
1532 offset = pg[i]->off & ~CFS_PAGE_MASK;
1536 if (pages == 0) /* that's all */
1539 if (offset + pg[i]->count < CFS_PAGE_SIZE)
1540 return count; /* doesn't end on page boundary */
1543 offset = pg[i]->off & ~CFS_PAGE_MASK;
1544 if (offset != 0) /* doesn't start on page boundary */
1551 static struct brw_page **osc_build_ppga(struct brw_page *pga, obd_count count)
1553 struct brw_page **ppga;
1556 OBD_ALLOC(ppga, sizeof(*ppga) * count);
1560 for (i = 0; i < count; i++)
1565 static void osc_release_ppga(struct brw_page **ppga, obd_count count)
1567 LASSERT(ppga != NULL);
1568 OBD_FREE(ppga, sizeof(*ppga) * count);
1571 static int osc_brw(int cmd, struct obd_export *exp, struct obd_info *oinfo,
1572 obd_count page_count, struct brw_page *pga,
1573 struct obd_trans_info *oti)
1575 struct obdo *saved_oa = NULL;
1576 struct brw_page **ppga, **orig;
1577 struct obd_import *imp = class_exp2cliimp(exp);
1578 struct client_obd *cli = &imp->imp_obd->u.cli;
1579 int rc, page_count_orig;
1582 if (cmd & OBD_BRW_CHECK) {
1583 /* The caller just wants to know if there's a chance that this
1584 * I/O can succeed */
1586 if (imp == NULL || imp->imp_invalid)
1591 /* test_brw with a failed create can trip this, maybe others. */
1592 LASSERT(cli->cl_max_pages_per_rpc);
1596 orig = ppga = osc_build_ppga(pga, page_count);
1599 page_count_orig = page_count;
1601 sort_brw_pages(ppga, page_count);
1602 while (page_count) {
1603 obd_count pages_per_brw;
1605 if (page_count > cli->cl_max_pages_per_rpc)
1606 pages_per_brw = cli->cl_max_pages_per_rpc;
1608 pages_per_brw = page_count;
1610 pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);
1612 if (saved_oa != NULL) {
1613 /* restore previously saved oa */
1614 *oinfo->oi_oa = *saved_oa;
1615 } else if (page_count > pages_per_brw) {
1616 /* save a copy of oa (brw will clobber it) */
1617 OBDO_ALLOC(saved_oa);
1618 if (saved_oa == NULL)
1619 GOTO(out, rc = -ENOMEM);
1620 *saved_oa = *oinfo->oi_oa;
1623 rc = osc_brw_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
1624 pages_per_brw, ppga, oinfo->oi_capa);
1629 page_count -= pages_per_brw;
1630 ppga += pages_per_brw;
1634 osc_release_ppga(orig, page_count_orig);
1636 if (saved_oa != NULL)
1637 OBDO_FREE(saved_oa);
1642 static int osc_brw_async(int cmd, struct obd_export *exp,
1643 struct obd_info *oinfo, obd_count page_count,
1644 struct brw_page *pga, struct obd_trans_info *oti,
1645 struct ptlrpc_request_set *set)
1647 struct brw_page **ppga, **orig;
1648 struct client_obd *cli = &exp->exp_obd->u.cli;
1649 int page_count_orig;
1653 if (cmd & OBD_BRW_CHECK) {
1654 struct obd_import *imp = class_exp2cliimp(exp);
1655 /* The caller just wants to know if there's a chance that this
1656 * I/O can succeed */
1658 if (imp == NULL || imp->imp_invalid)
1663 orig = ppga = osc_build_ppga(pga, page_count);
1666 page_count_orig = page_count;
1668 sort_brw_pages(ppga, page_count);
1669 while (page_count) {
1670 struct brw_page **copy;
1671 obd_count pages_per_brw;
1673 pages_per_brw = min_t(obd_count, page_count,
1674 cli->cl_max_pages_per_rpc);
1676 pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);
1678 /* use ppga only if single RPC is going to fly */
1679 if (pages_per_brw != page_count_orig || ppga != orig) {
1680 OBD_ALLOC(copy, sizeof(*copy) * pages_per_brw);
1682 GOTO(out, rc = -ENOMEM);
1683 memcpy(copy, ppga, sizeof(*copy) * pages_per_brw);
1687 rc = async_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
1688 pages_per_brw, copy, set, oinfo->oi_capa);
1692 OBD_FREE(copy, sizeof(*copy) * pages_per_brw);
1696 /* we passed it to async_internal() which is
1697 * now responsible for releasing memory */
1701 page_count -= pages_per_brw;
1702 ppga += pages_per_brw;
1706 osc_release_ppga(orig, page_count_orig);
1710 static void osc_check_rpcs(struct client_obd *cli);
1712 /* The companion to osc_enter_cache(), called when @oap is no longer part of
1713 * the dirty accounting. Writeback completes or truncate happens before
1714 * writing starts. Must be called with the loi lock held. */
1715 static void osc_exit_cache(struct client_obd *cli, struct osc_async_page *oap,
1718 osc_release_write_grant(cli, &oap->oap_brw_page, sent);
1722 /* This maintains the lists of pending pages to read/write for a given object
1723 * (lop). This is used by osc_check_rpcs->osc_next_loi() and loi_list_maint()
1724 * to quickly find objects that are ready to send an RPC. */
1725 static int lop_makes_rpc(struct client_obd *cli, struct loi_oap_pages *lop,
1731 if (lop->lop_num_pending == 0)
1734 /* if we have an invalid import we want to drain the queued pages
1735 * by forcing them through rpcs that immediately fail and complete
1736 * the pages. recovery relies on this to empty the queued pages
1737 * before canceling the locks and evicting down the llite pages */
1738 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
1741 /* stream rpcs in queue order as long as as there is an urgent page
1742 * queued. this is our cheap solution for good batching in the case
1743 * where writepage marks some random page in the middle of the file
1744 * as urgent because of, say, memory pressure */
1745 if (!list_empty(&lop->lop_urgent)) {
1746 CDEBUG(D_CACHE, "urgent request forcing RPC\n");
1749 /* fire off rpcs when we have 'optimal' rpcs as tuned for the wire. */
1750 optimal = cli->cl_max_pages_per_rpc;
1751 if (cmd & OBD_BRW_WRITE) {
1752 /* trigger a write rpc stream as long as there are dirtiers
1753 * waiting for space. as they're waiting, they're not going to
1754 * create more pages to coallesce with what's waiting.. */
1755 if (!list_empty(&cli->cl_cache_waiters)) {
1756 CDEBUG(D_CACHE, "cache waiters forcing RPC\n");
1759 /* +16 to avoid triggering rpcs that would want to include pages
1760 * that are being queued but which can't be made ready until
1761 * the queuer finishes with the page. this is a wart for
1762 * llite::commit_write() */
1765 if (lop->lop_num_pending >= optimal)
1771 static void on_list(struct list_head *item, struct list_head *list,
1774 if (list_empty(item) && should_be_on)
1775 list_add_tail(item, list);
1776 else if (!list_empty(item) && !should_be_on)
1777 list_del_init(item);
1780 /* maintain the loi's cli list membership invariants so that osc_send_oap_rpc
1781 * can find pages to build into rpcs quickly */
1782 static void loi_list_maint(struct client_obd *cli, struct lov_oinfo *loi)
1784 on_list(&loi->loi_cli_item, &cli->cl_loi_ready_list,
1785 lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE) ||
1786 lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ));
1788 on_list(&loi->loi_write_item, &cli->cl_loi_write_list,
1789 loi->loi_write_lop.lop_num_pending);
1791 on_list(&loi->loi_read_item, &cli->cl_loi_read_list,
1792 loi->loi_read_lop.lop_num_pending);
1795 static void lop_update_pending(struct client_obd *cli,
1796 struct loi_oap_pages *lop, int cmd, int delta)
1798 lop->lop_num_pending += delta;
1799 if (cmd & OBD_BRW_WRITE)
1800 cli->cl_pending_w_pages += delta;
1802 cli->cl_pending_r_pages += delta;
1805 /* this is called when a sync waiter receives an interruption. Its job is to
1806 * get the caller woken as soon as possible. If its page hasn't been put in an
1807 * rpc yet it can dequeue immediately. Otherwise it has to mark the rpc as
1808 * desiring interruption which will forcefully complete the rpc once the rpc
1810 static void osc_occ_interrupted(struct oig_callback_context *occ)
1812 struct osc_async_page *oap;
1813 struct loi_oap_pages *lop;
1814 struct lov_oinfo *loi;
1817 /* XXX member_of() */
1818 oap = list_entry(occ, struct osc_async_page, oap_occ);
1820 client_obd_list_lock(&oap->oap_cli->cl_loi_list_lock);
1822 oap->oap_interrupted = 1;
1824 /* ok, it's been put in an rpc. only one oap gets a request reference */
1825 if (oap->oap_request != NULL) {
1826 ptlrpc_mark_interrupted(oap->oap_request);
1827 ptlrpcd_wake(oap->oap_request);
1831 /* we don't get interruption callbacks until osc_trigger_group_io()
1832 * has been called and put the sync oaps in the pending/urgent lists.*/
1833 if (!list_empty(&oap->oap_pending_item)) {
1834 list_del_init(&oap->oap_pending_item);
1835 list_del_init(&oap->oap_urgent_item);
1838 lop = (oap->oap_cmd & OBD_BRW_WRITE) ?
1839 &loi->loi_write_lop : &loi->loi_read_lop;
1840 lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, -1);
1841 loi_list_maint(oap->oap_cli, oap->oap_loi);
1843 oig_complete_one(oap->oap_oig, &oap->oap_occ, -EINTR);
1844 oap->oap_oig = NULL;
1848 client_obd_list_unlock(&oap->oap_cli->cl_loi_list_lock);
1851 /* this is trying to propogate async writeback errors back up to the
1852 * application. As an async write fails we record the error code for later if
1853 * the app does an fsync. As long as errors persist we force future rpcs to be
1854 * sync so that the app can get a sync error and break the cycle of queueing
1855 * pages for which writeback will fail. */
1856 static void osc_process_ar(struct osc_async_rc *ar, __u64 xid,
1863 ar->ar_force_sync = 1;
1864 ar->ar_min_xid = ptlrpc_sample_next_xid();
1869 if (ar->ar_force_sync && (xid >= ar->ar_min_xid))
1870 ar->ar_force_sync = 0;
1873 static void osc_oap_to_pending(struct osc_async_page *oap)
1875 struct loi_oap_pages *lop;
1877 if (oap->oap_cmd & OBD_BRW_WRITE)
1878 lop = &oap->oap_loi->loi_write_lop;
1880 lop = &oap->oap_loi->loi_read_lop;
1882 if (oap->oap_async_flags & ASYNC_URGENT)
1883 list_add(&oap->oap_urgent_item, &lop->lop_urgent);
1884 list_add_tail(&oap->oap_pending_item, &lop->lop_pending);
1885 lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, 1);
1888 /* this must be called holding the loi list lock to give coverage to exit_cache,
1889 * async_flag maintenance, and oap_request */
1890 static void osc_ap_completion(struct client_obd *cli, struct obdo *oa,
1891 struct osc_async_page *oap, int sent, int rc)
1896 if (oap->oap_request != NULL) {
1897 xid = ptlrpc_req_xid(oap->oap_request);
1898 ptlrpc_req_finished(oap->oap_request);
1899 oap->oap_request = NULL;
1902 oap->oap_async_flags = 0;
1903 oap->oap_interrupted = 0;
1905 if (oap->oap_cmd & OBD_BRW_WRITE) {
1906 osc_process_ar(&cli->cl_ar, xid, rc);
1907 osc_process_ar(&oap->oap_loi->loi_ar, xid, rc);
1910 if (rc == 0 && oa != NULL) {
1911 if (oa->o_valid & OBD_MD_FLBLOCKS)
1912 oap->oap_loi->loi_lvb.lvb_blocks = oa->o_blocks;
1913 if (oa->o_valid & OBD_MD_FLMTIME)
1914 oap->oap_loi->loi_lvb.lvb_mtime = oa->o_mtime;
1915 if (oa->o_valid & OBD_MD_FLATIME)
1916 oap->oap_loi->loi_lvb.lvb_atime = oa->o_atime;
1917 if (oa->o_valid & OBD_MD_FLCTIME)
1918 oap->oap_loi->loi_lvb.lvb_ctime = oa->o_ctime;
1922 osc_exit_cache(cli, oap, sent);
1923 oig_complete_one(oap->oap_oig, &oap->oap_occ, rc);
1924 oap->oap_oig = NULL;
1929 rc = oap->oap_caller_ops->ap_completion(oap->oap_caller_data,
1930 oap->oap_cmd, oa, rc);
1932 /* ll_ap_completion (from llite) drops PG_locked. so, a new
1933 * I/O on the page could start, but OSC calls it under lock
1934 * and thus we can add oap back to pending safely */
1936 /* upper layer wants to leave the page on pending queue */
1937 osc_oap_to_pending(oap);
1939 osc_exit_cache(cli, oap, sent);
1943 static int brw_interpret_oap(struct ptlrpc_request *req, void *data, int rc)
1945 struct osc_async_page *oap, *tmp;
1946 struct osc_brw_async_args *aa = data;
1947 struct client_obd *cli;
1950 rc = osc_brw_fini_request(req, rc);
1951 CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
1952 if (osc_recoverable_error(rc)) {
1953 rc = osc_brw_redo_request(req, aa);
1960 client_obd_list_lock(&cli->cl_loi_list_lock);
1962 /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
1963 * is called so we know whether to go to sync BRWs or wait for more
1964 * RPCs to complete */
1965 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
1966 cli->cl_w_in_flight--;
1968 cli->cl_r_in_flight--;
1970 /* the caller may re-use the oap after the completion call so
1971 * we need to clean it up a little */
1972 list_for_each_entry_safe(oap, tmp, &aa->aa_oaps, oap_rpc_item) {
1973 list_del_init(&oap->oap_rpc_item);
1974 osc_ap_completion(cli, aa->aa_oa, oap, 1, rc);
1977 osc_wake_cache_waiters(cli);
1978 osc_check_rpcs(cli);
1980 client_obd_list_unlock(&cli->cl_loi_list_lock);
1982 OBDO_FREE(aa->aa_oa);
1984 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
1988 static struct ptlrpc_request *osc_build_req(struct client_obd *cli,
1989 struct list_head *rpc_list,
1990 int page_count, int cmd)
1992 struct ptlrpc_request *req;
1993 struct brw_page **pga = NULL;
1994 struct osc_brw_async_args *aa;
1995 struct obdo *oa = NULL;
1996 struct obd_async_page_ops *ops = NULL;
1997 void *caller_data = NULL;
1998 struct obd_capa *ocapa;
1999 struct osc_async_page *oap;
2003 LASSERT(!list_empty(rpc_list));
2005 OBD_ALLOC(pga, sizeof(*pga) * page_count);
2007 RETURN(ERR_PTR(-ENOMEM));
2011 GOTO(out, req = ERR_PTR(-ENOMEM));
2014 list_for_each_entry(oap, rpc_list, oap_rpc_item) {
2016 ops = oap->oap_caller_ops;
2017 caller_data = oap->oap_caller_data;
2019 pga[i] = &oap->oap_brw_page;
2020 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
2021 CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
2022 pga[i]->pg, cfs_page_index(oap->oap_page), oap, pga[i]->flag);
2026 /* always get the data for the obdo for the rpc */
2027 LASSERT(ops != NULL);
2028 ops->ap_fill_obdo(caller_data, cmd, oa);
2029 ocapa = ops->ap_lookup_capa(caller_data, cmd);
2031 sort_brw_pages(pga, page_count);
2032 rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count,
2036 CERROR("prep_req failed: %d\n", rc);
2037 GOTO(out, req = ERR_PTR(rc));
2040 /* Need to update the timestamps after the request is built in case
2041 * we race with setattr (locally or in queue at OST). If OST gets
2042 * later setattr before earlier BRW (as determined by the request xid),
2043 * the OST will not use BRW timestamps. Sadly, there is no obvious
2044 * way to do this in a single call. bug 10150 */
2045 ops->ap_update_obdo(caller_data, cmd, oa,
2046 OBD_MD_FLMTIME | OBD_MD_FLCTIME | OBD_MD_FLATIME);
2048 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2049 aa = (struct osc_brw_async_args *)&req->rq_async_args;
2050 INIT_LIST_HEAD(&aa->aa_oaps);
2051 list_splice(rpc_list, &aa->aa_oaps);
2052 INIT_LIST_HEAD(rpc_list);
2059 OBD_FREE(pga, sizeof(*pga) * page_count);
2064 /* the loi lock is held across this function but it's allowed to release
2065 * and reacquire it during its work */
2066 static int osc_send_oap_rpc(struct client_obd *cli, struct lov_oinfo *loi,
2067 int cmd, struct loi_oap_pages *lop)
2069 struct ptlrpc_request *req;
2070 obd_count page_count = 0;
2071 struct osc_async_page *oap = NULL, *tmp;
2072 struct osc_brw_async_args *aa;
2073 struct obd_async_page_ops *ops;
2074 CFS_LIST_HEAD(rpc_list);
2075 unsigned int ending_offset;
2076 unsigned starting_offset = 0;
2079 /* first we find the pages we're allowed to work with */
2080 list_for_each_entry_safe(oap, tmp, &lop->lop_pending,
2082 ops = oap->oap_caller_ops;
2084 LASSERT(oap->oap_magic == OAP_MAGIC);
2086 /* in llite being 'ready' equates to the page being locked
2087 * until completion unlocks it. commit_write submits a page
2088 * as not ready because its unlock will happen unconditionally
2089 * as the call returns. if we race with commit_write giving
2090 * us that page we dont' want to create a hole in the page
2091 * stream, so we stop and leave the rpc to be fired by
2092 * another dirtier or kupdated interval (the not ready page
2093 * will still be on the dirty list). we could call in
2094 * at the end of ll_file_write to process the queue again. */
2095 if (!(oap->oap_async_flags & ASYNC_READY)) {
2096 int rc = ops->ap_make_ready(oap->oap_caller_data, cmd);
2098 CDEBUG(D_INODE, "oap %p page %p returned %d "
2099 "instead of ready\n", oap,
2103 /* llite is telling us that the page is still
2104 * in commit_write and that we should try
2105 * and put it in an rpc again later. we
2106 * break out of the loop so we don't create
2107 * a hole in the sequence of pages in the rpc
2112 /* the io isn't needed.. tell the checks
2113 * below to complete the rpc with EINTR */
2114 oap->oap_async_flags |= ASYNC_COUNT_STABLE;
2115 oap->oap_count = -EINTR;
2118 oap->oap_async_flags |= ASYNC_READY;
2121 LASSERTF(0, "oap %p page %p returned %d "
2122 "from make_ready\n", oap,
2130 * Page submitted for IO has to be locked. Either by
2131 * ->ap_make_ready() or by higher layers.
2133 * XXX nikita: this assertion should be adjusted when lustre
2134 * starts using PG_writeback for pages being written out.
2136 #if defined(__KERNEL__) && defined(__LINUX__)
2137 LASSERT(PageLocked(oap->oap_page));
2139 /* If there is a gap at the start of this page, it can't merge
2140 * with any previous page, so we'll hand the network a
2141 * "fragmented" page array that it can't transfer in 1 RDMA */
2142 if (page_count != 0 && oap->oap_page_off != 0)
2145 /* take the page out of our book-keeping */
2146 list_del_init(&oap->oap_pending_item);
2147 lop_update_pending(cli, lop, cmd, -1);
2148 list_del_init(&oap->oap_urgent_item);
2150 if (page_count == 0)
2151 starting_offset = (oap->oap_obj_off+oap->oap_page_off) &
2152 (PTLRPC_MAX_BRW_SIZE - 1);
2154 /* ask the caller for the size of the io as the rpc leaves. */
2155 if (!(oap->oap_async_flags & ASYNC_COUNT_STABLE))
2157 ops->ap_refresh_count(oap->oap_caller_data,cmd);
2158 if (oap->oap_count <= 0) {
2159 CDEBUG(D_CACHE, "oap %p count %d, completing\n", oap,
2161 osc_ap_completion(cli, NULL, oap, 0, oap->oap_count);
2165 /* now put the page back in our accounting */
2166 list_add_tail(&oap->oap_rpc_item, &rpc_list);
2167 if (++page_count >= cli->cl_max_pages_per_rpc)
2170 /* End on a PTLRPC_MAX_BRW_SIZE boundary. We want full-sized
2171 * RPCs aligned on PTLRPC_MAX_BRW_SIZE boundaries to help reads
2172 * have the same alignment as the initial writes that allocated
2173 * extents on the server. */
2174 ending_offset = (oap->oap_obj_off + oap->oap_page_off +
2175 oap->oap_count) & (PTLRPC_MAX_BRW_SIZE - 1);
2176 if (ending_offset == 0)
2179 /* If there is a gap at the end of this page, it can't merge
2180 * with any subsequent pages, so we'll hand the network a
2181 * "fragmented" page array that it can't transfer in 1 RDMA */
2182 if (oap->oap_page_off + oap->oap_count < CFS_PAGE_SIZE)
2186 osc_wake_cache_waiters(cli);
2188 if (page_count == 0)
2191 loi_list_maint(cli, loi);
2193 client_obd_list_unlock(&cli->cl_loi_list_lock);
2195 req = osc_build_req(cli, &rpc_list, page_count, cmd);
2197 /* this should happen rarely and is pretty bad, it makes the
2198 * pending list not follow the dirty order */
2199 client_obd_list_lock(&cli->cl_loi_list_lock);
2200 list_for_each_entry_safe(oap, tmp, &rpc_list, oap_rpc_item) {
2201 list_del_init(&oap->oap_rpc_item);
2203 /* queued sync pages can be torn down while the pages
2204 * were between the pending list and the rpc */
2205 if (oap->oap_interrupted) {
2206 CDEBUG(D_INODE, "oap %p interrupted\n", oap);
2207 osc_ap_completion(cli, NULL, oap, 0,
2211 osc_ap_completion(cli, NULL, oap, 0, PTR_ERR(req));
2213 loi_list_maint(cli, loi);
2214 RETURN(PTR_ERR(req));
2217 aa = (struct osc_brw_async_args *)&req->rq_async_args;
2219 if (cmd == OBD_BRW_READ) {
2220 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
2221 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
2222 lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
2223 (starting_offset >> CFS_PAGE_SHIFT) + 1);
2224 ptlrpc_lprocfs_brw(req, OST_READ, aa->aa_requested_nob);
2226 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
2227 lprocfs_oh_tally(&cli->cl_write_rpc_hist,
2228 cli->cl_w_in_flight);
2229 lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
2230 (starting_offset >> CFS_PAGE_SHIFT) + 1);
2231 ptlrpc_lprocfs_brw(req, OST_WRITE, aa->aa_requested_nob);
2234 client_obd_list_lock(&cli->cl_loi_list_lock);
2236 if (cmd == OBD_BRW_READ)
2237 cli->cl_r_in_flight++;
2239 cli->cl_w_in_flight++;
2241 /* queued sync pages can be torn down while the pages
2242 * were between the pending list and the rpc */
2244 list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2245 /* only one oap gets a request reference */
2248 if (oap->oap_interrupted && !req->rq_intr) {
2249 CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
2251 ptlrpc_mark_interrupted(req);
2255 tmp->oap_request = ptlrpc_request_addref(req);
2257 DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
2258 page_count, aa, cli->cl_r_in_flight, cli->cl_w_in_flight);
2260 req->rq_interpret_reply = brw_interpret_oap;
2261 ptlrpcd_add_req(req);
2265 #define LOI_DEBUG(LOI, STR, args...) \
2266 CDEBUG(D_INODE, "loi ready %d wr %d:%d rd %d:%d " STR, \
2267 !list_empty(&(LOI)->loi_cli_item), \
2268 (LOI)->loi_write_lop.lop_num_pending, \
2269 !list_empty(&(LOI)->loi_write_lop.lop_urgent), \
2270 (LOI)->loi_read_lop.lop_num_pending, \
2271 !list_empty(&(LOI)->loi_read_lop.lop_urgent), \
2274 /* This is called by osc_check_rpcs() to find which objects have pages that
2275 * we could be sending. These lists are maintained by lop_makes_rpc(). */
2276 struct lov_oinfo *osc_next_loi(struct client_obd *cli)
2279 /* first return all objects which we already know to have
2280 * pages ready to be stuffed into rpcs */
2281 if (!list_empty(&cli->cl_loi_ready_list))
2282 RETURN(list_entry(cli->cl_loi_ready_list.next,
2283 struct lov_oinfo, loi_cli_item));
2285 /* then if we have cache waiters, return all objects with queued
2286 * writes. This is especially important when many small files
2287 * have filled up the cache and not been fired into rpcs because
2288 * they don't pass the nr_pending/object threshhold */
2289 if (!list_empty(&cli->cl_cache_waiters) &&
2290 !list_empty(&cli->cl_loi_write_list))
2291 RETURN(list_entry(cli->cl_loi_write_list.next,
2292 struct lov_oinfo, loi_write_item));
2294 /* then return all queued objects when we have an invalid import
2295 * so that they get flushed */
2296 if (cli->cl_import == NULL || cli->cl_import->imp_invalid) {
2297 if (!list_empty(&cli->cl_loi_write_list))
2298 RETURN(list_entry(cli->cl_loi_write_list.next,
2299 struct lov_oinfo, loi_write_item));
2300 if (!list_empty(&cli->cl_loi_read_list))
2301 RETURN(list_entry(cli->cl_loi_read_list.next,
2302 struct lov_oinfo, loi_read_item));
2307 /* called with the loi list lock held */
2308 static void osc_check_rpcs(struct client_obd *cli)
2310 struct lov_oinfo *loi;
2311 int rc = 0, race_counter = 0;
2314 while ((loi = osc_next_loi(cli)) != NULL) {
2315 LOI_DEBUG(loi, "%lu in flight\n", rpcs_in_flight(cli));
2317 if (rpcs_in_flight(cli) >= cli->cl_max_rpcs_in_flight)
2320 /* attempt some read/write balancing by alternating between
2321 * reads and writes in an object. The makes_rpc checks here
2322 * would be redundant if we were getting read/write work items
2323 * instead of objects. we don't want send_oap_rpc to drain a
2324 * partial read pending queue when we're given this object to
2325 * do io on writes while there are cache waiters */
2326 if (lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE)) {
2327 rc = osc_send_oap_rpc(cli, loi, OBD_BRW_WRITE,
2328 &loi->loi_write_lop);
2336 if (lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ)) {
2337 rc = osc_send_oap_rpc(cli, loi, OBD_BRW_READ,
2338 &loi->loi_read_lop);
2347 /* attempt some inter-object balancing by issueing rpcs
2348 * for each object in turn */
2349 if (!list_empty(&loi->loi_cli_item))
2350 list_del_init(&loi->loi_cli_item);
2351 if (!list_empty(&loi->loi_write_item))
2352 list_del_init(&loi->loi_write_item);
2353 if (!list_empty(&loi->loi_read_item))
2354 list_del_init(&loi->loi_read_item);
2356 loi_list_maint(cli, loi);
2358 /* send_oap_rpc fails with 0 when make_ready tells it to
2359 * back off. llite's make_ready does this when it tries
2360 * to lock a page queued for write that is already locked.
2361 * we want to try sending rpcs from many objects, but we
2362 * don't want to spin failing with 0. */
2363 if (race_counter == 10)
2369 /* we're trying to queue a page in the osc so we're subject to the
2370 * 'cl_dirty_max' limit on the number of pages that can be queued in the osc.
2371 * If the osc's queued pages are already at that limit, then we want to sleep
2372 * until there is space in the osc's queue for us. We also may be waiting for
2373 * write credits from the OST if there are RPCs in flight that may return some
2374 * before we fall back to sync writes.
2376 * We need this know our allocation was granted in the presence of signals */
2377 static int ocw_granted(struct client_obd *cli, struct osc_cache_waiter *ocw)
2381 client_obd_list_lock(&cli->cl_loi_list_lock);
2382 rc = list_empty(&ocw->ocw_entry) || rpcs_in_flight(cli) == 0;
2383 client_obd_list_unlock(&cli->cl_loi_list_lock);
2387 /* Caller must hold loi_list_lock - we drop/regain it if we need to wait for
2388 * grant or cache space. */
2389 static int osc_enter_cache(struct client_obd *cli, struct lov_oinfo *loi,
2390 struct osc_async_page *oap)
2392 struct osc_cache_waiter ocw;
2393 struct l_wait_info lwi = { 0 };
2397 CDEBUG(D_CACHE, "dirty: %ld/%d dirty_max: %ld/%d dropped: %lu "
2398 "grant: %lu\n", cli->cl_dirty, atomic_read(&obd_dirty_pages),
2399 cli->cl_dirty_max, obd_max_dirty_pages,
2400 cli->cl_lost_grant, cli->cl_avail_grant);
2402 /* force the caller to try sync io. this can jump the list
2403 * of queued writes and create a discontiguous rpc stream */
2404 if (cli->cl_dirty_max < CFS_PAGE_SIZE || cli->cl_ar.ar_force_sync ||
2405 loi->loi_ar.ar_force_sync)
2408 /* Hopefully normal case - cache space and write credits available */
2409 if ((cli->cl_dirty + CFS_PAGE_SIZE <= cli->cl_dirty_max) &&
2410 (atomic_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) &&
2411 (cli->cl_avail_grant >= CFS_PAGE_SIZE)) {
2412 /* account for ourselves */
2413 osc_consume_write_grant(cli, &oap->oap_brw_page);
2417 /* Make sure that there are write rpcs in flight to wait for. This
2418 * is a little silly as this object may not have any pending but
2419 * other objects sure might. */
2420 if (cli->cl_w_in_flight) {
2421 list_add_tail(&ocw.ocw_entry, &cli->cl_cache_waiters);
2422 cfs_waitq_init(&ocw.ocw_waitq);
2426 loi_list_maint(cli, loi);
2427 osc_check_rpcs(cli);
2428 client_obd_list_unlock(&cli->cl_loi_list_lock);
2430 CDEBUG(D_CACHE, "sleeping for cache space\n");
2431 l_wait_event(ocw.ocw_waitq, ocw_granted(cli, &ocw), &lwi);
2433 client_obd_list_lock(&cli->cl_loi_list_lock);
2434 if (!list_empty(&ocw.ocw_entry)) {
2435 list_del(&ocw.ocw_entry);
2444 int osc_prep_async_page(struct obd_export *exp, struct lov_stripe_md *lsm,
2445 struct lov_oinfo *loi, cfs_page_t *page,
2446 obd_off offset, struct obd_async_page_ops *ops,
2447 void *data, void **res)
2449 struct osc_async_page *oap;
2453 return size_round(sizeof(*oap));
2456 oap->oap_magic = OAP_MAGIC;
2457 oap->oap_cli = &exp->exp_obd->u.cli;
2460 oap->oap_caller_ops = ops;
2461 oap->oap_caller_data = data;
2463 oap->oap_page = page;
2464 oap->oap_obj_off = offset;
2466 CFS_INIT_LIST_HEAD(&oap->oap_pending_item);
2467 CFS_INIT_LIST_HEAD(&oap->oap_urgent_item);
2468 CFS_INIT_LIST_HEAD(&oap->oap_rpc_item);
2470 oap->oap_occ.occ_interrupted = osc_occ_interrupted;
2472 CDEBUG(D_CACHE, "oap %p page %p obj off "LPU64"\n", oap, page, offset);
2476 struct osc_async_page *oap_from_cookie(void *cookie)
2478 struct osc_async_page *oap = cookie;
2479 if (oap->oap_magic != OAP_MAGIC)
2480 return ERR_PTR(-EINVAL);
2484 static int osc_queue_async_io(struct obd_export *exp, struct lov_stripe_md *lsm,
2485 struct lov_oinfo *loi, void *cookie,
2486 int cmd, obd_off off, int count,
2487 obd_flag brw_flags, enum async_flags async_flags)
2489 struct client_obd *cli = &exp->exp_obd->u.cli;
2490 struct osc_async_page *oap;
2494 oap = oap_from_cookie(cookie);
2496 RETURN(PTR_ERR(oap));
2498 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
2501 if (!list_empty(&oap->oap_pending_item) ||
2502 !list_empty(&oap->oap_urgent_item) ||
2503 !list_empty(&oap->oap_rpc_item))
2506 /* check if the file's owner/group is over quota */
2507 #ifdef HAVE_QUOTA_SUPPORT
2508 if ((cmd & OBD_BRW_WRITE) && !(cmd & OBD_BRW_NOQUOTA)){
2509 struct obd_async_page_ops *ops;
2516 ops = oap->oap_caller_ops;
2517 ops->ap_fill_obdo(oap->oap_caller_data, cmd, oa);
2518 if (lquota_chkdq(quota_interface, cli, oa->o_uid, oa->o_gid) ==
2529 loi = lsm->lsm_oinfo[0];
2531 client_obd_list_lock(&cli->cl_loi_list_lock);
2534 oap->oap_page_off = off;
2535 oap->oap_count = count;
2536 oap->oap_brw_flags = brw_flags;
2537 oap->oap_async_flags = async_flags;
2539 if (cmd & OBD_BRW_WRITE) {
2540 rc = osc_enter_cache(cli, loi, oap);
2542 client_obd_list_unlock(&cli->cl_loi_list_lock);
2547 osc_oap_to_pending(oap);
2548 loi_list_maint(cli, loi);
2550 LOI_DEBUG(loi, "oap %p page %p added for cmd %d\n", oap, oap->oap_page,
2553 osc_check_rpcs(cli);
2554 client_obd_list_unlock(&cli->cl_loi_list_lock);
2559 /* aka (~was & now & flag), but this is more clear :) */
2560 #define SETTING(was, now, flag) (!(was & flag) && (now & flag))
2562 static int osc_set_async_flags(struct obd_export *exp,
2563 struct lov_stripe_md *lsm,
2564 struct lov_oinfo *loi, void *cookie,
2565 obd_flag async_flags)
2567 struct client_obd *cli = &exp->exp_obd->u.cli;
2568 struct loi_oap_pages *lop;
2569 struct osc_async_page *oap;
2573 oap = oap_from_cookie(cookie);
2575 RETURN(PTR_ERR(oap));
2578 * bug 7311: OST-side locking is only supported for liblustre for now
2579 * (and liblustre never calls obd_set_async_flags(). I hope.), generic
2580 * implementation has to handle case where OST-locked page was picked
2581 * up by, e.g., ->writepage().
2583 LASSERT(!(oap->oap_brw_flags & OBD_BRW_SRVLOCK));
2584 LASSERT(!LIBLUSTRE_CLIENT); /* check that liblustre angels do fear to
2587 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
2591 loi = lsm->lsm_oinfo[0];
2593 if (oap->oap_cmd & OBD_BRW_WRITE) {
2594 lop = &loi->loi_write_lop;
2596 lop = &loi->loi_read_lop;
2599 client_obd_list_lock(&cli->cl_loi_list_lock);
2601 if (list_empty(&oap->oap_pending_item))
2602 GOTO(out, rc = -EINVAL);
2604 if ((oap->oap_async_flags & async_flags) == async_flags)
2607 if (SETTING(oap->oap_async_flags, async_flags, ASYNC_READY))
2608 oap->oap_async_flags |= ASYNC_READY;
2610 if (SETTING(oap->oap_async_flags, async_flags, ASYNC_URGENT)) {
2611 if (list_empty(&oap->oap_rpc_item)) {
2612 list_add(&oap->oap_urgent_item, &lop->lop_urgent);
2613 loi_list_maint(cli, loi);
2617 LOI_DEBUG(loi, "oap %p page %p has flags %x\n", oap, oap->oap_page,
2618 oap->oap_async_flags);
2620 osc_check_rpcs(cli);
2621 client_obd_list_unlock(&cli->cl_loi_list_lock);
2625 static int osc_queue_group_io(struct obd_export *exp, struct lov_stripe_md *lsm,
2626 struct lov_oinfo *loi,
2627 struct obd_io_group *oig, void *cookie,
2628 int cmd, obd_off off, int count,
2630 obd_flag async_flags)
2632 struct client_obd *cli = &exp->exp_obd->u.cli;
2633 struct osc_async_page *oap;
2634 struct loi_oap_pages *lop;
2638 oap = oap_from_cookie(cookie);
2640 RETURN(PTR_ERR(oap));
2642 if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
2645 if (!list_empty(&oap->oap_pending_item) ||
2646 !list_empty(&oap->oap_urgent_item) ||
2647 !list_empty(&oap->oap_rpc_item))
2651 loi = lsm->lsm_oinfo[0];
2653 client_obd_list_lock(&cli->cl_loi_list_lock);
2656 oap->oap_page_off = off;
2657 oap->oap_count = count;
2658 oap->oap_brw_flags = brw_flags;
2659 oap->oap_async_flags = async_flags;
2661 if (cmd & OBD_BRW_WRITE)
2662 lop = &loi->loi_write_lop;
2664 lop = &loi->loi_read_lop;
2666 list_add_tail(&oap->oap_pending_item, &lop->lop_pending_group);
2667 if (oap->oap_async_flags & ASYNC_GROUP_SYNC) {
2669 rc = oig_add_one(oig, &oap->oap_occ);
2672 LOI_DEBUG(loi, "oap %p page %p on group pending: rc %d\n",
2673 oap, oap->oap_page, rc);
2675 client_obd_list_unlock(&cli->cl_loi_list_lock);
2680 static void osc_group_to_pending(struct client_obd *cli, struct lov_oinfo *loi,
2681 struct loi_oap_pages *lop, int cmd)
2683 struct list_head *pos, *tmp;
2684 struct osc_async_page *oap;
2686 list_for_each_safe(pos, tmp, &lop->lop_pending_group) {
2687 oap = list_entry(pos, struct osc_async_page, oap_pending_item);
2688 list_del(&oap->oap_pending_item);
2689 osc_oap_to_pending(oap);
2691 loi_list_maint(cli, loi);
2694 static int osc_trigger_group_io(struct obd_export *exp,
2695 struct lov_stripe_md *lsm,
2696 struct lov_oinfo *loi,
2697 struct obd_io_group *oig)
2699 struct client_obd *cli = &exp->exp_obd->u.cli;
2703 loi = lsm->lsm_oinfo[0];
2705 client_obd_list_lock(&cli->cl_loi_list_lock);
2707 osc_group_to_pending(cli, loi, &loi->loi_write_lop, OBD_BRW_WRITE);
2708 osc_group_to_pending(cli, loi, &loi->loi_read_lop, OBD_BRW_READ);
2710 osc_check_rpcs(cli);
2711 client_obd_list_unlock(&cli->cl_loi_list_lock);
2716 static int osc_teardown_async_page(struct obd_export *exp,
2717 struct lov_stripe_md *lsm,
2718 struct lov_oinfo *loi, void *cookie)
2720 struct client_obd *cli = &exp->exp_obd->u.cli;
2721 struct loi_oap_pages *lop;
2722 struct osc_async_page *oap;
2726 oap = oap_from_cookie(cookie);
2728 RETURN(PTR_ERR(oap));
2731 loi = lsm->lsm_oinfo[0];
2733 if (oap->oap_cmd & OBD_BRW_WRITE) {
2734 lop = &loi->loi_write_lop;
2736 lop = &loi->loi_read_lop;
2739 client_obd_list_lock(&cli->cl_loi_list_lock);
2741 if (!list_empty(&oap->oap_rpc_item))
2742 GOTO(out, rc = -EBUSY);
2744 osc_exit_cache(cli, oap, 0);
2745 osc_wake_cache_waiters(cli);
2747 if (!list_empty(&oap->oap_urgent_item)) {
2748 list_del_init(&oap->oap_urgent_item);
2749 oap->oap_async_flags &= ~ASYNC_URGENT;
2751 if (!list_empty(&oap->oap_pending_item)) {
2752 list_del_init(&oap->oap_pending_item);
2753 lop_update_pending(cli, lop, oap->oap_cmd, -1);
2755 loi_list_maint(cli, loi);
2757 LOI_DEBUG(loi, "oap %p page %p torn down\n", oap, oap->oap_page);
2759 client_obd_list_unlock(&cli->cl_loi_list_lock);
2763 static void osc_set_data_with_check(struct lustre_handle *lockh, void *data,
2766 struct ldlm_lock *lock = ldlm_handle2lock(lockh);
2769 CERROR("lockh %p, data %p - client evicted?\n", lockh, data);
2772 lock_res_and_lock(lock);
2773 #if defined (__KERNEL__) && defined (__LINUX__)
2774 /* Liang XXX: Darwin and Winnt checking should be added */
2775 if (lock->l_ast_data && lock->l_ast_data != data) {
2776 struct inode *new_inode = data;
2777 struct inode *old_inode = lock->l_ast_data;
2778 if (!(old_inode->i_state & I_FREEING))
2779 LDLM_ERROR(lock, "inconsistent l_ast_data found");
2780 LASSERTF(old_inode->i_state & I_FREEING,
2781 "Found existing inode %p/%lu/%u state %lu in lock: "
2782 "setting data to %p/%lu/%u\n", old_inode,
2783 old_inode->i_ino, old_inode->i_generation,
2785 new_inode, new_inode->i_ino, new_inode->i_generation);
2788 lock->l_ast_data = data;
2789 lock->l_flags |= (flags & LDLM_FL_NO_LRU);
2790 unlock_res_and_lock(lock);
2791 LDLM_LOCK_PUT(lock);
2794 static int osc_change_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2795 ldlm_iterator_t replace, void *data)
2797 struct ldlm_res_id res_id = { .name = {0} };
2798 struct obd_device *obd = class_exp2obd(exp);
2800 res_id.name[0] = lsm->lsm_object_id;
2801 res_id.name[2] = lsm->lsm_object_gr;
2803 ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2807 static int osc_enqueue_fini(struct ptlrpc_request *req, struct obd_info *oinfo,
2813 /* The request was created before ldlm_cli_enqueue call. */
2814 if (rc == ELDLM_LOCK_ABORTED) {
2815 struct ldlm_reply *rep;
2817 /* swabbed by ldlm_cli_enqueue() */
2818 LASSERT(lustre_rep_swabbed(req, DLM_LOCKREPLY_OFF));
2819 rep = lustre_msg_buf(req->rq_repmsg, DLM_LOCKREPLY_OFF,
2821 LASSERT(rep != NULL);
2822 if (rep->lock_policy_res1)
2823 rc = rep->lock_policy_res1;
2827 if ((intent && rc == ELDLM_LOCK_ABORTED) || !rc) {
2828 CDEBUG(D_INODE,"got kms "LPU64" blocks "LPU64" mtime "LPU64"\n",
2829 oinfo->oi_md->lsm_oinfo[0]->loi_lvb.lvb_size,
2830 oinfo->oi_md->lsm_oinfo[0]->loi_lvb.lvb_blocks,
2831 oinfo->oi_md->lsm_oinfo[0]->loi_lvb.lvb_mtime);
2834 /* Call the update callback. */
2835 rc = oinfo->oi_cb_up(oinfo, rc);
2839 static int osc_enqueue_interpret(struct ptlrpc_request *req,
2840 struct osc_enqueue_args *aa, int rc)
2842 int intent = aa->oa_oi->oi_flags & LDLM_FL_HAS_INTENT;
2843 struct lov_stripe_md *lsm = aa->oa_oi->oi_md;
2844 struct ldlm_lock *lock;
2846 /* ldlm_cli_enqueue is holding a reference on the lock, so it must
2848 lock = ldlm_handle2lock(aa->oa_oi->oi_lockh);
2850 /* Complete obtaining the lock procedure. */
2851 rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_ei->ei_type, 1,
2853 &aa->oa_oi->oi_flags,
2854 &lsm->lsm_oinfo[0]->loi_lvb,
2855 sizeof(lsm->lsm_oinfo[0]->loi_lvb),
2856 lustre_swab_ost_lvb,
2857 aa->oa_oi->oi_lockh, rc);
2859 /* Complete osc stuff. */
2860 rc = osc_enqueue_fini(req, aa->oa_oi, intent, rc);
2862 /* Release the lock for async request. */
2863 if (lustre_handle_is_used(aa->oa_oi->oi_lockh) && rc == ELDLM_OK)
2864 ldlm_lock_decref(aa->oa_oi->oi_lockh, aa->oa_ei->ei_mode);
2866 LASSERTF(lock != NULL, "lockh %p, req %p, aa %p - client evicted?\n",
2867 aa->oa_oi->oi_lockh, req, aa);
2868 LDLM_LOCK_PUT(lock);
2872 /* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
2873 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
2874 * other synchronous requests, however keeping some locks and trying to obtain
2875 * others may take a considerable amount of time in a case of ost failure; and
2876 * when other sync requests do not get released lock from a client, the client
2877 * is excluded from the cluster -- such scenarious make the life difficult, so
2878 * release locks just after they are obtained. */
2879 static int osc_enqueue(struct obd_export *exp, struct obd_info *oinfo,
2880 struct ldlm_enqueue_info *einfo,
2881 struct ptlrpc_request_set *rqset)
2883 struct ldlm_res_id res_id = { .name = {0} };
2884 struct obd_device *obd = exp->exp_obd;
2885 struct ldlm_reply *rep;
2886 struct ptlrpc_request *req = NULL;
2887 int intent = oinfo->oi_flags & LDLM_FL_HAS_INTENT;
2892 res_id.name[0] = oinfo->oi_md->lsm_object_id;
2893 res_id.name[2] = oinfo->oi_md->lsm_object_gr;
2895 /* Filesystem lock extents are extended to page boundaries so that
2896 * dealing with the page cache is a little smoother. */
2897 oinfo->oi_policy.l_extent.start -=
2898 oinfo->oi_policy.l_extent.start & ~CFS_PAGE_MASK;
2899 oinfo->oi_policy.l_extent.end |= ~CFS_PAGE_MASK;
2901 if (oinfo->oi_md->lsm_oinfo[0]->loi_kms_valid == 0)
2904 /* Next, search for already existing extent locks that will cover us */
2905 /* If we're trying to read, we also search for an existing PW lock. The
2906 * VFS and page cache already protect us locally, so lots of readers/
2907 * writers can share a single PW lock.
2909 * There are problems with conversion deadlocks, so instead of
2910 * converting a read lock to a write lock, we'll just enqueue a new
2913 * At some point we should cancel the read lock instead of making them
2914 * send us a blocking callback, but there are problems with canceling
2915 * locks out from other users right now, too. */
2916 mode = einfo->ei_mode;
2917 if (einfo->ei_mode == LCK_PR)
2919 mode = ldlm_lock_match(obd->obd_namespace,
2920 oinfo->oi_flags | LDLM_FL_LVB_READY, &res_id,
2921 einfo->ei_type, &oinfo->oi_policy, mode,
2924 /* addref the lock only if not async requests and PW lock is
2925 * matched whereas we asked for PR. */
2926 if (!rqset && einfo->ei_mode != mode)
2927 ldlm_lock_addref(oinfo->oi_lockh, LCK_PR);
2928 osc_set_data_with_check(oinfo->oi_lockh, einfo->ei_cbdata,
2931 /* I would like to be able to ASSERT here that rss <=
2932 * kms, but I can't, for reasons which are explained in
2936 /* We already have a lock, and it's referenced */
2937 oinfo->oi_cb_up(oinfo, ELDLM_OK);
2939 /* For async requests, decref the lock. */
2940 if (einfo->ei_mode != mode)
2941 ldlm_lock_decref(oinfo->oi_lockh, LCK_PW);
2943 ldlm_lock_decref(oinfo->oi_lockh, einfo->ei_mode);
2951 [MSG_PTLRPC_BODY_OFF] = sizeof(struct ptlrpc_body),
2952 [DLM_LOCKREQ_OFF] = sizeof(struct ldlm_request),
2953 [DLM_LOCKREQ_OFF + 1] = 0 };
2955 req = ldlm_prep_enqueue_req(exp, 2, size, NULL, 0);
2959 size[DLM_LOCKREPLY_OFF] = sizeof(*rep);
2960 size[DLM_REPLY_REC_OFF] =
2961 sizeof(oinfo->oi_md->lsm_oinfo[0]->loi_lvb);
2962 ptlrpc_req_set_repsize(req, 3, size);
2965 /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
2966 oinfo->oi_flags &= ~LDLM_FL_BLOCK_GRANTED;
2968 rc = ldlm_cli_enqueue(exp, &req, einfo, &res_id,
2969 &oinfo->oi_policy, &oinfo->oi_flags,
2970 &oinfo->oi_md->lsm_oinfo[0]->loi_lvb,
2971 sizeof(oinfo->oi_md->lsm_oinfo[0]->loi_lvb),
2972 lustre_swab_ost_lvb, oinfo->oi_lockh,
2976 struct osc_enqueue_args *aa;
2977 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
2978 aa = (struct osc_enqueue_args *)&req->rq_async_args;
2983 req->rq_interpret_reply = osc_enqueue_interpret;
2984 ptlrpc_set_add_req(rqset, req);
2985 } else if (intent) {
2986 ptlrpc_req_finished(req);
2991 rc = osc_enqueue_fini(req, oinfo, intent, rc);
2993 ptlrpc_req_finished(req);
2998 static int osc_match(struct obd_export *exp, struct lov_stripe_md *lsm,
2999 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
3000 int *flags, void *data, struct lustre_handle *lockh)
3002 struct ldlm_res_id res_id = { .name = {0} };
3003 struct obd_device *obd = exp->exp_obd;
3004 int lflags = *flags;
3008 res_id.name[0] = lsm->lsm_object_id;
3009 res_id.name[2] = lsm->lsm_object_gr;
3011 OBD_FAIL_RETURN(OBD_FAIL_OSC_MATCH, -EIO);
3013 /* Filesystem lock extents are extended to page boundaries so that
3014 * dealing with the page cache is a little smoother */
3015 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
3016 policy->l_extent.end |= ~CFS_PAGE_MASK;
3018 /* Next, search for already existing extent locks that will cover us */
3019 /* If we're trying to read, we also search for an existing PW lock. The
3020 * VFS and page cache already protect us locally, so lots of readers/
3021 * writers can share a single PW lock. */
3025 rc = ldlm_lock_match(obd->obd_namespace, lflags | LDLM_FL_LVB_READY,
3026 &res_id, type, policy, rc, lockh);
3028 osc_set_data_with_check(lockh, data, lflags);
3029 if (!(lflags & LDLM_FL_TEST_LOCK) && mode != rc) {
3030 ldlm_lock_addref(lockh, LCK_PR);
3031 ldlm_lock_decref(lockh, LCK_PW);
3038 static int osc_cancel(struct obd_export *exp, struct lov_stripe_md *md,
3039 __u32 mode, struct lustre_handle *lockh)
3043 if (unlikely(mode == LCK_GROUP))
3044 ldlm_lock_decref_and_cancel(lockh, mode);
3046 ldlm_lock_decref(lockh, mode);
3051 static int osc_cancel_unused(struct obd_export *exp,
3052 struct lov_stripe_md *lsm, int flags,
3055 struct obd_device *obd = class_exp2obd(exp);
3056 struct ldlm_res_id res_id = { .name = {0} }, *resp = NULL;
3059 res_id.name[0] = lsm->lsm_object_id;
3060 res_id.name[2] = lsm->lsm_object_gr;
3064 return ldlm_cli_cancel_unused(obd->obd_namespace, resp, flags, opaque);
3067 static int osc_join_lru(struct obd_export *exp,
3068 struct lov_stripe_md *lsm, int join)
3070 struct obd_device *obd = class_exp2obd(exp);
3071 struct ldlm_res_id res_id = { .name = {0} }, *resp = NULL;
3074 res_id.name[0] = lsm->lsm_object_id;
3075 res_id.name[2] = lsm->lsm_object_gr;
3079 return ldlm_cli_join_lru(obd->obd_namespace, resp, join);
3082 static int osc_statfs_interpret(struct ptlrpc_request *req,
3083 struct osc_async_args *aa, int rc)
3085 struct obd_statfs *msfs;
3091 msfs = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*msfs),
3092 lustre_swab_obd_statfs);
3094 CERROR("Can't unpack obd_statfs\n");
3095 GOTO(out, rc = -EPROTO);
3098 memcpy(aa->aa_oi->oi_osfs, msfs, sizeof(*msfs));
3100 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
3104 static int osc_statfs_async(struct obd_device *obd, struct obd_info *oinfo,
3105 __u64 max_age, struct ptlrpc_request_set *rqset)
3107 struct ptlrpc_request *req;
3108 struct osc_async_args *aa;
3109 int size[2] = { sizeof(struct ptlrpc_body), sizeof(*oinfo->oi_osfs) };
3112 /* We could possibly pass max_age in the request (as an absolute
3113 * timestamp or a "seconds.usec ago") so the target can avoid doing
3114 * extra calls into the filesystem if that isn't necessary (e.g.
3115 * during mount that would help a bit). Having relative timestamps
3116 * is not so great if request processing is slow, while absolute
3117 * timestamps are not ideal because they need time synchronization. */
3118 req = ptlrpc_prep_req(obd->u.cli.cl_import, LUSTRE_OST_VERSION,
3119 OST_STATFS, 1, NULL, NULL);
3123 ptlrpc_req_set_repsize(req, 2, size);
3124 req->rq_request_portal = OST_CREATE_PORTAL; //XXX FIXME bug 249
3126 req->rq_interpret_reply = osc_statfs_interpret;
3127 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
3128 aa = (struct osc_async_args *)&req->rq_async_args;
3131 ptlrpc_set_add_req(rqset, req);
3135 static int osc_statfs(struct obd_device *obd, struct obd_statfs *osfs,
3138 struct obd_statfs *msfs;
3139 struct ptlrpc_request *req;
3140 int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*osfs) };
3143 /* We could possibly pass max_age in the request (as an absolute
3144 * timestamp or a "seconds.usec ago") so the target can avoid doing
3145 * extra calls into the filesystem if that isn't necessary (e.g.
3146 * during mount that would help a bit). Having relative timestamps
3147 * is not so great if request processing is slow, while absolute
3148 * timestamps are not ideal because they need time synchronization. */
3149 req = ptlrpc_prep_req(obd->u.cli.cl_import, LUSTRE_OST_VERSION,
3150 OST_STATFS, 1, NULL, NULL);
3154 ptlrpc_req_set_repsize(req, 2, size);
3155 req->rq_request_portal = OST_CREATE_PORTAL; //XXX FIXME bug 249
3157 rc = ptlrpc_queue_wait(req);
3161 msfs = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*msfs),
3162 lustre_swab_obd_statfs);
3164 CERROR("Can't unpack obd_statfs\n");
3165 GOTO(out, rc = -EPROTO);
3168 memcpy(osfs, msfs, sizeof(*osfs));
3172 ptlrpc_req_finished(req);
3176 /* Retrieve object striping information.
3178 * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
3179 * the maximum number of OST indices which will fit in the user buffer.
3180 * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
3182 static int osc_getstripe(struct lov_stripe_md *lsm, struct lov_user_md *lump)
3184 struct lov_user_md lum, *lumk;
3185 int rc = 0, lum_size;
3191 if (copy_from_user(&lum, lump, sizeof(lum)))
3194 if (lum.lmm_magic != LOV_USER_MAGIC)
3197 if (lum.lmm_stripe_count > 0) {
3198 lum_size = sizeof(lum) + sizeof(lum.lmm_objects[0]);
3199 OBD_ALLOC(lumk, lum_size);
3203 lumk->lmm_objects[0].l_object_id = lsm->lsm_object_id;
3204 lumk->lmm_objects[0].l_object_gr = lsm->lsm_object_gr;
3206 lum_size = sizeof(lum);
3210 lumk->lmm_object_id = lsm->lsm_object_id;
3211 lumk->lmm_object_gr = lsm->lsm_object_gr;
3212 lumk->lmm_stripe_count = 1;
3214 if (copy_to_user(lump, lumk, lum_size))
3218 OBD_FREE(lumk, lum_size);
3224 static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
3225 void *karg, void *uarg)
3227 struct obd_device *obd = exp->exp_obd;
3228 struct obd_ioctl_data *data = karg;
3232 if (!try_module_get(THIS_MODULE)) {
3233 CERROR("Can't get module. Is it alive?");
3237 case OBD_IOC_LOV_GET_CONFIG: {
3239 struct lov_desc *desc;
3240 struct obd_uuid uuid;
3244 if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
3245 GOTO(out, err = -EINVAL);
3247 data = (struct obd_ioctl_data *)buf;
3249 if (sizeof(*desc) > data->ioc_inllen1) {
3250 obd_ioctl_freedata(buf, len);
3251 GOTO(out, err = -EINVAL);
3254 if (data->ioc_inllen2 < sizeof(uuid)) {
3255 obd_ioctl_freedata(buf, len);
3256 GOTO(out, err = -EINVAL);
3259 desc = (struct lov_desc *)data->ioc_inlbuf1;
3260 desc->ld_tgt_count = 1;
3261 desc->ld_active_tgt_count = 1;
3262 desc->ld_default_stripe_count = 1;
3263 desc->ld_default_stripe_size = 0;
3264 desc->ld_default_stripe_offset = 0;
3265 desc->ld_pattern = 0;
3266 memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
3268 memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
3270 err = copy_to_user((void *)uarg, buf, len);
3273 obd_ioctl_freedata(buf, len);
3276 case LL_IOC_LOV_SETSTRIPE:
3277 err = obd_alloc_memmd(exp, karg);
3281 case LL_IOC_LOV_GETSTRIPE:
3282 err = osc_getstripe(karg, uarg);
3284 case OBD_IOC_CLIENT_RECOVER:
3285 err = ptlrpc_recover_import(obd->u.cli.cl_import,
3290 case IOC_OSC_SET_ACTIVE:
3291 err = ptlrpc_set_import_active(obd->u.cli.cl_import,
3294 case OBD_IOC_POLL_QUOTACHECK:
3295 err = lquota_poll_check(quota_interface, exp,
3296 (struct if_quotacheck *)karg);
3299 CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
3300 cmd, cfs_curproc_comm());
3301 GOTO(out, err = -ENOTTY);
3304 module_put(THIS_MODULE);
3308 static int osc_get_info(struct obd_export *exp, obd_count keylen,
3309 void *key, __u32 *vallen, void *val)
3312 if (!vallen || !val)
3315 if (KEY_IS("lock_to_stripe")) {
3316 __u32 *stripe = val;
3317 *vallen = sizeof(*stripe);
3320 } else if (KEY_IS("last_id")) {
3321 struct ptlrpc_request *req;
3323 char *bufs[2] = { NULL, key };
3324 int rc, size[2] = { sizeof(struct ptlrpc_body), keylen };
3326 req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
3327 OST_GET_INFO, 2, size, bufs);
3331 size[REPLY_REC_OFF] = *vallen;
3332 ptlrpc_req_set_repsize(req, 2, size);
3333 rc = ptlrpc_queue_wait(req);
3337 reply = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*reply),
3338 lustre_swab_ost_last_id);
3339 if (reply == NULL) {
3340 CERROR("Can't unpack OST last ID\n");
3341 GOTO(out, rc = -EPROTO);
3343 *((obd_id *)val) = *reply;
3345 ptlrpc_req_finished(req);
3351 static int osc_setinfo_mds_conn_interpret(struct ptlrpc_request *req,
3354 struct llog_ctxt *ctxt;
3355 struct obd_import *imp = req->rq_import;
3361 ctxt = llog_get_context(imp->imp_obd, LLOG_MDS_OST_ORIG_CTXT);
3364 rc = llog_initiator_connect(ctxt);
3366 CERROR("cannot establish connection for "
3367 "ctxt %p: %d\n", ctxt, rc);
3370 spin_lock(&imp->imp_lock);
3371 imp->imp_server_timeout = 1;
3372 imp->imp_pingable = 1;
3373 spin_unlock(&imp->imp_lock);
3374 CDEBUG(D_RPCTRACE, "pinging OST %s\n", obd2cli_tgt(imp->imp_obd));
3379 static int osc_set_info_async(struct obd_export *exp, obd_count keylen,
3380 void *key, obd_count vallen, void *val,
3381 struct ptlrpc_request_set *set)
3383 struct ptlrpc_request *req;
3384 struct obd_device *obd = exp->exp_obd;
3385 struct obd_import *imp = class_exp2cliimp(exp);
3386 int size[3] = { sizeof(struct ptlrpc_body), keylen, vallen };
3387 char *bufs[3] = { NULL, key, val };
3390 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
3392 if (KEY_IS(KEY_NEXT_ID)) {
3393 if (vallen != sizeof(obd_id))
3395 obd->u.cli.cl_oscc.oscc_next_id = *((obd_id*)val) + 1;
3396 CDEBUG(D_HA, "%s: set oscc_next_id = "LPU64"\n",
3397 exp->exp_obd->obd_name,
3398 obd->u.cli.cl_oscc.oscc_next_id);
3403 if (KEY_IS("unlinked")) {
3404 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3405 spin_lock(&oscc->oscc_lock);
3406 oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
3407 spin_unlock(&oscc->oscc_lock);
3411 if (KEY_IS(KEY_INIT_RECOV)) {
3412 if (vallen != sizeof(int))
3414 spin_lock(&imp->imp_lock);
3415 imp->imp_initial_recov = *(int *)val;
3416 spin_unlock(&imp->imp_lock);
3417 CDEBUG(D_HA, "%s: set imp_initial_recov = %d\n",
3418 exp->exp_obd->obd_name,
3419 imp->imp_initial_recov);
3423 if (KEY_IS("checksum")) {
3424 if (vallen != sizeof(int))
3426 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
3430 if (KEY_IS(KEY_FLUSH_CTX)) {
3431 sptlrpc_import_flush_my_ctx(imp);
3438 /* We pass all other commands directly to OST. Since nobody calls osc
3439 methods directly and everybody is supposed to go through LOV, we
3440 assume lov checked invalid values for us.
3441 The only recognised values so far are evict_by_nid and mds_conn.
3442 Even if something bad goes through, we'd get a -EINVAL from OST
3445 req = ptlrpc_prep_req(imp, LUSTRE_OST_VERSION, OST_SET_INFO, 3, size,
3450 if (KEY_IS(KEY_MDS_CONN)) {
3451 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3453 oscc->oscc_oa.o_gr = (*(__u32 *)val);
3454 oscc->oscc_oa.o_valid |= OBD_MD_FLGROUP;
3455 LASSERT(oscc->oscc_oa.o_gr > 0);
3456 req->rq_interpret_reply = osc_setinfo_mds_conn_interpret;
3459 ptlrpc_req_set_repsize(req, 1, NULL);
3460 ptlrpc_set_add_req(set, req);
3461 ptlrpc_check_set(set);
3467 static struct llog_operations osc_size_repl_logops = {
3468 lop_cancel: llog_obd_repl_cancel
3471 static struct llog_operations osc_mds_ost_orig_logops;
3472 static int osc_llog_init(struct obd_device *obd, struct obd_llogs *llogs,
3473 struct obd_device *tgt, int count,
3474 struct llog_catid *catid, struct obd_uuid *uuid)
3479 spin_lock(&obd->obd_dev_lock);
3480 if (osc_mds_ost_orig_logops.lop_setup != llog_obd_origin_setup) {
3481 osc_mds_ost_orig_logops = llog_lvfs_ops;
3482 osc_mds_ost_orig_logops.lop_setup = llog_obd_origin_setup;
3483 osc_mds_ost_orig_logops.lop_cleanup = llog_obd_origin_cleanup;
3484 osc_mds_ost_orig_logops.lop_add = llog_obd_origin_add;
3485 osc_mds_ost_orig_logops.lop_connect = llog_origin_connect;
3487 spin_unlock(&obd->obd_dev_lock);
3489 rc = llog_setup(obd, llogs, LLOG_MDS_OST_ORIG_CTXT, tgt, count,
3490 &catid->lci_logid, &osc_mds_ost_orig_logops);
3492 CERROR("failed LLOG_MDS_OST_ORIG_CTXT\n");
3496 rc = llog_setup(obd, llogs, LLOG_SIZE_REPL_CTXT, tgt, count, NULL,
3497 &osc_size_repl_logops);
3499 CERROR("failed LLOG_SIZE_REPL_CTXT\n");
3502 CERROR("osc '%s' tgt '%s' cnt %d catid %p rc=%d\n",
3503 obd->obd_name, tgt->obd_name, count, catid, rc);
3504 CERROR("logid "LPX64":0x%x\n",
3505 catid->lci_logid.lgl_oid, catid->lci_logid.lgl_ogen);
3510 static int osc_llog_finish(struct obd_device *obd, int count)
3512 struct llog_ctxt *ctxt;
3513 int rc = 0, rc2 = 0;
3516 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
3518 rc = llog_cleanup(ctxt);
3520 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3522 rc2 = llog_cleanup(ctxt);
3529 static int osc_reconnect(struct obd_export *exp, struct obd_device *obd,
3530 struct obd_uuid *cluuid,
3531 struct obd_connect_data *data)
3533 struct client_obd *cli = &obd->u.cli;
3535 if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
3538 client_obd_list_lock(&cli->cl_loi_list_lock);
3539 data->ocd_grant = cli->cl_avail_grant ?:
3540 2 * cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT;
3541 lost_grant = cli->cl_lost_grant;
3542 cli->cl_lost_grant = 0;
3543 client_obd_list_unlock(&cli->cl_loi_list_lock);
3545 CDEBUG(D_CACHE, "request ocd_grant: %d cl_avail_grant: %ld "
3546 "cl_lost_grant: %ld\n", data->ocd_grant,
3547 cli->cl_avail_grant, lost_grant);
3548 CDEBUG(D_RPCTRACE, "ocd_connect_flags: "LPX64" ocd_version: %d"
3549 " ocd_grant: %d\n", data->ocd_connect_flags,
3550 data->ocd_version, data->ocd_grant);
3556 static int osc_disconnect(struct obd_export *exp)
3558 struct obd_device *obd = class_exp2obd(exp);
3559 struct llog_ctxt *ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3562 if (obd->u.cli.cl_conn_count == 1)
3563 /* flush any remaining cancel messages out to the target */
3564 llog_sync(ctxt, exp);
3566 rc = client_disconnect_export(exp);
3570 static int osc_import_event(struct obd_device *obd,
3571 struct obd_import *imp,
3572 enum obd_import_event event)
3574 struct client_obd *cli;
3578 LASSERT(imp->imp_obd == obd);
3581 case IMP_EVENT_DISCON: {
3582 /* Only do this on the MDS OSC's */
3583 if (imp->imp_server_timeout) {
3584 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3586 spin_lock(&oscc->oscc_lock);
3587 oscc->oscc_flags |= OSCC_FLAG_RECOVERING;
3588 spin_unlock(&oscc->oscc_lock);
3591 client_obd_list_lock(&cli->cl_loi_list_lock);
3592 cli->cl_avail_grant = 0;
3593 cli->cl_lost_grant = 0;
3594 client_obd_list_unlock(&cli->cl_loi_list_lock);
3597 case IMP_EVENT_INACTIVE: {
3598 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
3601 case IMP_EVENT_INVALIDATE: {
3602 struct ldlm_namespace *ns = obd->obd_namespace;
3606 client_obd_list_lock(&cli->cl_loi_list_lock);
3607 /* all pages go to failing rpcs due to the invalid import */
3608 osc_check_rpcs(cli);
3609 client_obd_list_unlock(&cli->cl_loi_list_lock);
3611 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3615 case IMP_EVENT_ACTIVE: {
3616 /* Only do this on the MDS OSC's */
3617 if (imp->imp_server_timeout) {
3618 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3620 spin_lock(&oscc->oscc_lock);
3621 oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
3622 spin_unlock(&oscc->oscc_lock);
3624 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
3627 case IMP_EVENT_OCD: {
3628 struct obd_connect_data *ocd = &imp->imp_connect_data;
3630 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
3631 osc_init_grant(&obd->u.cli, ocd);
3634 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
3635 imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;
3637 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
3641 CERROR("Unknown import event %d\n", event);
3647 int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
3653 rc = ptlrpcd_addref();
3657 rc = client_obd_setup(obd, lcfg);
3661 struct lprocfs_static_vars lvars;
3662 struct client_obd *cli = &obd->u.cli;
3664 lprocfs_init_vars(osc, &lvars);
3665 if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0) {
3666 lproc_osc_attach_seqstat(obd);
3667 ptlrpc_lprocfs_register_obd(obd);
3671 /* We need to allocate a few requests more, because
3672 brw_interpret_oap tries to create new requests before freeing
3673 previous ones. Ideally we want to have 2x max_rpcs_in_flight
3674 reserved, but I afraid that might be too much wasted RAM
3675 in fact, so 2 is just my guess and still should work. */
3676 cli->cl_import->imp_rq_pool =
3677 ptlrpc_init_rq_pool(cli->cl_max_rpcs_in_flight + 2,
3679 ptlrpc_add_rqs_to_pool);
3685 static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
3691 case OBD_CLEANUP_EARLY: {
3692 struct obd_import *imp;
3693 imp = obd->u.cli.cl_import;
3694 CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
3695 /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
3696 ptlrpc_deactivate_import(imp);
3697 spin_lock(&imp->imp_lock);
3698 imp->imp_pingable = 0;
3699 spin_unlock(&imp->imp_lock);
3702 case OBD_CLEANUP_EXPORTS: {
3703 /* If we set up but never connected, the
3704 client import will not have been cleaned. */
3705 if (obd->u.cli.cl_import) {
3706 struct obd_import *imp;
3707 imp = obd->u.cli.cl_import;
3708 CDEBUG(D_CONFIG, "%s: client import never connected\n",
3710 ptlrpc_invalidate_import(imp);
3711 ptlrpc_free_rq_pool(imp->imp_rq_pool);
3712 class_destroy_import(imp);
3713 obd->u.cli.cl_import = NULL;
3717 case OBD_CLEANUP_SELF_EXP:
3718 rc = obd_llog_finish(obd, 0);
3720 CERROR("failed to cleanup llogging subsystems\n");
3722 case OBD_CLEANUP_OBD:
3728 int osc_cleanup(struct obd_device *obd)
3730 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3734 ptlrpc_lprocfs_unregister_obd(obd);
3735 lprocfs_obd_cleanup(obd);
3737 spin_lock(&oscc->oscc_lock);
3738 oscc->oscc_flags &= ~OSCC_FLAG_RECOVERING;
3739 oscc->oscc_flags |= OSCC_FLAG_EXITING;
3740 spin_unlock(&oscc->oscc_lock);
3742 /* free memory of osc quota cache */
3743 lquota_cleanup(quota_interface, obd);
3745 rc = client_obd_cleanup(obd);
3751 static int osc_process_config(struct obd_device *obd, obd_count len, void *buf)
3753 struct lustre_cfg *lcfg = buf;
3754 struct lprocfs_static_vars lvars;
3757 lprocfs_init_vars(osc, &lvars);
3759 rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars, lcfg, obd);
3763 struct obd_ops osc_obd_ops = {
3764 .o_owner = THIS_MODULE,
3765 .o_setup = osc_setup,
3766 .o_precleanup = osc_precleanup,
3767 .o_cleanup = osc_cleanup,
3768 .o_add_conn = client_import_add_conn,
3769 .o_del_conn = client_import_del_conn,
3770 .o_connect = client_connect_import,
3771 .o_reconnect = osc_reconnect,
3772 .o_disconnect = osc_disconnect,
3773 .o_statfs = osc_statfs,
3774 .o_statfs_async = osc_statfs_async,
3775 .o_packmd = osc_packmd,
3776 .o_unpackmd = osc_unpackmd,
3777 .o_precreate = osc_precreate,
3778 .o_create = osc_create,
3779 .o_destroy = osc_destroy,
3780 .o_getattr = osc_getattr,
3781 .o_getattr_async = osc_getattr_async,
3782 .o_setattr = osc_setattr,
3783 .o_setattr_async = osc_setattr_async,
3785 .o_brw_async = osc_brw_async,
3786 .o_prep_async_page = osc_prep_async_page,
3787 .o_queue_async_io = osc_queue_async_io,
3788 .o_set_async_flags = osc_set_async_flags,
3789 .o_queue_group_io = osc_queue_group_io,
3790 .o_trigger_group_io = osc_trigger_group_io,
3791 .o_teardown_async_page = osc_teardown_async_page,
3792 .o_punch = osc_punch,
3794 .o_enqueue = osc_enqueue,
3795 .o_match = osc_match,
3796 .o_change_cbdata = osc_change_cbdata,
3797 .o_cancel = osc_cancel,
3798 .o_cancel_unused = osc_cancel_unused,
3799 .o_join_lru = osc_join_lru,
3800 .o_iocontrol = osc_iocontrol,
3801 .o_get_info = osc_get_info,
3802 .o_set_info_async = osc_set_info_async,
3803 .o_import_event = osc_import_event,
3804 .o_llog_init = osc_llog_init,
3805 .o_llog_finish = osc_llog_finish,
3806 .o_process_config = osc_process_config,
3808 int __init osc_init(void)
3810 struct lprocfs_static_vars lvars;
3814 lprocfs_init_vars(osc, &lvars);
3816 request_module("lquota");
3817 quota_interface = PORTAL_SYMBOL_GET(osc_quota_interface);
3818 lquota_init(quota_interface);
3819 init_obd_quota_ops(quota_interface, &osc_obd_ops);
3821 rc = class_register_type(&osc_obd_ops, NULL, lvars.module_vars,
3822 LUSTRE_OSC_NAME, NULL);
3824 if (quota_interface)
3825 PORTAL_SYMBOL_PUT(osc_quota_interface);
3833 static void /*__exit*/ osc_exit(void)
3835 lquota_exit(quota_interface);
3836 if (quota_interface)
3837 PORTAL_SYMBOL_PUT(osc_quota_interface);
3839 class_unregister_type(LUSTRE_OSC_NAME);
3842 MODULE_AUTHOR("Cluster File Systems, Inc. <info@clusterfs.com>");
3843 MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
3844 MODULE_LICENSE("GPL");
3846 cfs_module(osc, LUSTRE_VERSION_STRING, osc_init, osc_exit);
git://git.whamcloud.com / fs / lustre-release.git / blob