4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License version 2 only,
8 * as published by the Free Software Foundation.
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.gnu.org/licenses/gpl-2.0.html
23 * Copyright (c) 2009, 2010, Oracle and/or its affiliates. All rights reserved.
24 * Use is subject to license terms.
26 * Copyright (c) 2012, 2017, Intel Corporation.
29 * This file is part of Lustre, http://www.lustre.org/
30 * Lustre is a trademark of Sun Microsystems, Inc.
32 * lustre/ofd/ofd_lvb.c
34 * This file contains methods for OBD Filter Device (OFD)
35 * Lock Value Block (LVB) operations.
37 * LVB is special opaque (to LDLM) data that is associated with an LDLM lock
38 * and transferred from client to server and back. OFD LVBs are used to
39 * maintain current object size/times.
41 * Author: Andreas Dilger <andreas.dilger@intel.com>
42 * Author: Mikhail Pershin <mike.pershin@intel.com>
43 * Author: Alexey Zhuravlev <alexey.zhuravlev@intel.com>
46 #define DEBUG_SUBSYSTEM S_FILTER
48 #include <lustre_swab.h>
49 #include "ofd_internal.h"
52 * Implementation of ldlm_valblock_ops::lvbo_free for OFD.
54 * This function frees allocated LVB data if it associated with the given
57 * \param[in] res LDLM resource
59 * \retval 0 on successful setup
60 * \retval negative value on error
62 static int ofd_lvbo_free(struct ldlm_resource *res)
65 OBD_FREE(res->lr_lvb_data, res->lr_lvb_len);
71 * Implementation of ldlm_valblock_ops::lvbo_init for OFD.
73 * This function allocates and initializes new LVB data for the given
74 * LDLM resource if it is not allocated yet. New LVB is filled with attributes
75 * of the object associated with that resource. Function does nothing if LVB
76 * for the given LDLM resource is allocated already.
78 * Called with res->lr_lvb_sem held.
80 * \param[in] lock LDLM lock on resource
82 * \retval 0 on successful setup
83 * \retval negative value on error
85 static int ofd_lvbo_init(struct ldlm_resource *res)
88 struct ofd_device *ofd;
89 struct ofd_object *fo;
90 struct ofd_thread_info *info;
96 LASSERT(mutex_is_locked(&res->lr_lvb_mutex));
98 if (res->lr_lvb_data != NULL)
101 ofd = ldlm_res_to_ns(res)->ns_lvbp;
102 LASSERT(ofd != NULL);
104 if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_OST_LVB))
112 GOTO(out, rc = -ENOMEM);
114 info = ofd_info(env);
115 res->lr_lvb_data = lvb;
116 res->lr_lvb_len = sizeof(*lvb);
118 ost_fid_from_resid(&info->fti_fid, &res->lr_name,
119 ofd->ofd_lut.lut_lsd.lsd_osd_index);
120 fo = ofd_object_find(env, ofd, &info->fti_fid);
122 GOTO(out_lvb, rc = PTR_ERR(fo));
124 rc = ofd_attr_get(env, fo, &info->fti_attr);
126 struct ofd_seq *oseq;
129 /* Object could be recreated during the first
130 * CLEANUP_ORPHAN request. */
132 seq = fid_seq(&info->fti_fid);
133 oseq = ofd_seq_load(env, ofd, fid_seq_is_idif(seq) ?
134 FID_SEQ_OST_MDT0 : seq);
135 if (!IS_ERR_OR_NULL(oseq)) {
136 if (!oseq->os_last_id_synced)
138 ofd_seq_put(env, oseq);
144 lvb->lvb_size = info->fti_attr.la_size;
145 lvb->lvb_blocks = info->fti_attr.la_blocks;
146 lvb->lvb_mtime = info->fti_attr.la_mtime;
147 lvb->lvb_atime = info->fti_attr.la_atime;
148 lvb->lvb_ctime = info->fti_attr.la_ctime;
150 if (fo->ofo_atime_ondisk == 0)
151 fo->ofo_atime_ondisk = info->fti_attr.la_atime;
154 "res: "DFID" initial LVB size: %llu, mtime: %#llx, atime: %#llx, ctime: %#llx, blocks: %#llx\n",
155 PFID(&info->fti_fid), lvb->lvb_size, lvb->lvb_mtime,
156 lvb->lvb_atime, lvb->lvb_ctime, lvb->lvb_blocks);
158 info->fti_attr.la_valid = 0;
162 ofd_object_put(env, fo);
165 OST_LVB_SET_ERR(lvb->lvb_blocks, rc);
167 /* Don't free lvb data on lookup error */
172 * Implementation of ldlm_valblock_ops::lvbo_update for OFD.
174 * When a client generates a glimpse enqueue, it wants to get the current
175 * file size and updated attributes for a stat() type operation, but these
176 * attributes may be writeback cached on another client. The client with
177 * the DLM extent lock at the highest offset is asked for its current
178 * attributes via a glimpse callback on its extent lock, on the assumption
179 * that it has the highest file size and the newest timestamps. The timestamps
180 * are guaranteed to be correct if there is only a single writer on the file,
181 * but may be slightly inaccurate if there are multiple concurrent writers on
182 * the same object. In order to avoid race conditions between the glimpse AST
183 * and the client cancelling the lock, ofd_lvbo_update() also updates
184 * the attributes from the local object. If the last client hasn't done any
185 * writes yet, or has already written its data and cancelled its lock before
186 * it processed the glimpse, then the local inode will have more uptodate
189 * This is called in two ways:
190 * \a req != NULL : called by the DLM itself after a glimpse callback
191 * \a req == NULL : called by the OFD after a disk write
193 * \param[in] lock LDLM lock
194 * \param[in] req PTLRPC request
195 * \param[in] increase_only don't allow LVB values to decrease
197 * \retval 0 on successful setup
198 * \retval negative value on error
200 static int ofd_lvbo_update(struct ldlm_resource *res, struct ldlm_lock *lock,
201 struct ptlrpc_request *req, int increase_only)
203 struct ofd_thread_info *info;
204 struct ofd_device *ofd;
205 struct ofd_object *fo;
207 const struct lu_env *env;
214 info = ofd_info(env);
215 LASSERT(res != NULL);
217 ofd = ldlm_res_to_ns(res)->ns_lvbp;
218 LASSERT(ofd != NULL);
220 fid_extract_from_res_name(&info->fti_fid, &res->lr_name);
222 lvb = res->lr_lvb_data;
224 CERROR("%s: no LVB data for "DFID"\n",
225 ofd_name(ofd), PFID(&info->fti_fid));
229 /* Update the LVB from the network message */
231 struct ost_lvb *rpc_lvb;
234 if (req->rq_import != NULL)
235 lvb_type = imp_connect_lvb_type(req->rq_import);
237 lvb_type = exp_connect_lvb_type(req->rq_export);
240 struct ost_lvb_v1 *lvb_v1;
242 lvb_v1 = req_capsule_server_swab_get(&req->rq_pill,
243 &RMF_DLM_LVB, lustre_swab_ost_lvb_v1);
247 rpc_lvb = &info->fti_lvb;
248 memcpy(rpc_lvb, lvb_v1, sizeof *lvb_v1);
249 rpc_lvb->lvb_mtime_ns = 0;
250 rpc_lvb->lvb_atime_ns = 0;
251 rpc_lvb->lvb_ctime_ns = 0;
253 rpc_lvb = req_capsule_server_swab_get(&req->rq_pill,
255 lustre_swab_ost_lvb);
261 if (rpc_lvb->lvb_size > lvb->lvb_size || !increase_only) {
262 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb size: "
263 "%llu -> %llu\n", PFID(&info->fti_fid),
264 lvb->lvb_size, rpc_lvb->lvb_size);
265 lvb->lvb_size = rpc_lvb->lvb_size;
267 if (rpc_lvb->lvb_mtime > lvb->lvb_mtime || !increase_only) {
268 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb mtime: "
269 "%llu -> %llu\n", PFID(&info->fti_fid),
270 lvb->lvb_mtime, rpc_lvb->lvb_mtime);
271 lvb->lvb_mtime = rpc_lvb->lvb_mtime;
273 if (rpc_lvb->lvb_atime > lvb->lvb_atime || !increase_only) {
274 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb atime: "
275 "%llu -> %llu\n", PFID(&info->fti_fid),
276 lvb->lvb_atime, rpc_lvb->lvb_atime);
277 lvb->lvb_atime = rpc_lvb->lvb_atime;
279 if (rpc_lvb->lvb_ctime > lvb->lvb_ctime || !increase_only) {
280 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb ctime: "
281 "%llu -> %llu\n", PFID(&info->fti_fid),
282 lvb->lvb_ctime, rpc_lvb->lvb_ctime);
283 lvb->lvb_ctime = rpc_lvb->lvb_ctime;
285 if (rpc_lvb->lvb_blocks > lvb->lvb_blocks || !increase_only) {
286 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb blocks: "
287 "%llu -> %llu\n", PFID(&info->fti_fid),
288 lvb->lvb_blocks, rpc_lvb->lvb_blocks);
289 lvb->lvb_blocks = rpc_lvb->lvb_blocks;
295 /* Update the LVB from the disk inode */
296 ost_fid_from_resid(&info->fti_fid, &res->lr_name,
297 ofd->ofd_lut.lut_lsd.lsd_osd_index);
298 fo = ofd_object_find(env, ofd, &info->fti_fid);
300 GOTO(out, rc = PTR_ERR(fo));
302 rc = ofd_attr_get(env, fo, &info->fti_attr);
307 if (info->fti_attr.la_size > lvb->lvb_size || !increase_only) {
308 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb size from disk: "
309 "%llu -> %llu\n", PFID(&info->fti_fid),
310 lvb->lvb_size, info->fti_attr.la_size);
311 lvb->lvb_size = info->fti_attr.la_size;
314 if (info->fti_attr.la_mtime >lvb->lvb_mtime || !increase_only) {
315 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb mtime from disk: "
316 "%llu -> %llu\n", PFID(&info->fti_fid),
317 lvb->lvb_mtime, info->fti_attr.la_mtime);
318 lvb->lvb_mtime = info->fti_attr.la_mtime;
320 if (info->fti_attr.la_atime >lvb->lvb_atime || !increase_only) {
321 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb atime from disk: "
322 "%llu -> %llu\n", PFID(&info->fti_fid),
323 lvb->lvb_atime, info->fti_attr.la_atime);
324 lvb->lvb_atime = info->fti_attr.la_atime;
326 if (info->fti_attr.la_ctime >lvb->lvb_ctime || !increase_only) {
327 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb ctime from disk: "
328 "%llu -> %llu\n", PFID(&info->fti_fid),
329 lvb->lvb_ctime, info->fti_attr.la_ctime);
330 lvb->lvb_ctime = info->fti_attr.la_ctime;
332 if (info->fti_attr.la_blocks > lvb->lvb_blocks || !increase_only) {
333 CDEBUG(D_DLMTRACE, "res: "DFID" updating lvb blocks from disk: "
334 "%llu -> %llu\n", PFID(&info->fti_fid), lvb->lvb_blocks,
335 (unsigned long long)info->fti_attr.la_blocks);
336 lvb->lvb_blocks = info->fti_attr.la_blocks;
340 info->fti_attr.la_valid = 0;
342 ofd_object_put(env, fo);
348 * Implementation of ldlm_valblock_ops::lvbo_size for OFD.
350 * This function returns size of LVB data so appropriate RPC size will be
351 * reserved. This is used for compatibility needs between server and client
352 * of different Lustre versions.
354 * \param[in] lock LDLM lock
356 * \retval size of LVB data
358 static int ofd_lvbo_size(struct ldlm_lock *lock)
360 if (lock->l_export != NULL && exp_connect_lvb_type(lock->l_export))
361 return sizeof(struct ost_lvb);
363 return sizeof(struct ost_lvb_v1);
367 * Implementation of ldlm_valblock_ops::lvbo_fill for OFD.
369 * This function is called to fill the given RPC buffer \a buf with LVB data
371 * \param[in] env execution environment
372 * \param[in] lock LDLM lock
373 * \param[in] buf RPC buffer to fill
374 * \param[in] buflen buffer length
376 * \retval size of LVB data written into \a buf buffer
378 static int ofd_lvbo_fill(struct ldlm_lock *lock, void *buf, int *buflen)
380 struct ldlm_resource *res = lock->l_resource;
383 /* Former lvbo_init not allocate the "LVB". */
384 if (unlikely(res->lr_lvb_len == 0))
387 lvb_len = ofd_lvbo_size(lock);
388 LASSERT(lvb_len <= res->lr_lvb_len);
390 if (lvb_len > *buflen)
394 memcpy(buf, res->lr_lvb_data, lvb_len);
400 struct ldlm_valblock_ops ofd_lvbo = {
401 .lvbo_init = ofd_lvbo_init,
402 .lvbo_update = ofd_lvbo_update,
403 .lvbo_free = ofd_lvbo_free,
404 .lvbo_size = ofd_lvbo_size,
405 .lvbo_fill = ofd_lvbo_fill