blob: 580852ba05c441513c0775c7eaca95369dfa71ec [file] [log] [blame]
Thomas Gleixner328970d2019-05-24 12:04:05 +02001// SPDX-License-Identifier: GPL-2.0-or-later
Mark Fashehccd979b2005-12-15 14:31:24 -08002/* -*- mode: c; c-basic-offset: 8; -*-
3 * vim: noexpandtab sw=8 ts=8 sts=0:
4 *
5 * uptodate.c
6 *
7 * Tracking the up-to-date-ness of a local buffer_head with respect to
8 * the cluster.
9 *
10 * Copyright (C) 2002, 2004, 2005 Oracle. All rights reserved.
11 *
Mark Fashehccd979b2005-12-15 14:31:24 -080012 * Standard buffer head caching flags (uptodate, etc) are insufficient
13 * in a clustered environment - a buffer may be marked up to date on
14 * our local node but could have been modified by another cluster
15 * member. As a result an additional (and performant) caching scheme
16 * is required. A further requirement is that we consume as little
17 * memory as possible - we never pin buffer_head structures in order
18 * to cache them.
19 *
20 * We track the existence of up to date buffers on the inodes which
21 * are associated with them. Because we don't want to pin
22 * buffer_heads, this is only a (strong) hint and several other checks
23 * are made in the I/O path to ensure that we don't use a stale or
24 * invalid buffer without going to disk:
25 * - buffer_jbd is used liberally - if a bh is in the journal on
26 * this node then it *must* be up to date.
27 * - the standard buffer_uptodate() macro is used to detect buffers
28 * which may be invalid (even if we have an up to date tracking
29 * item for them)
30 *
31 * For a full understanding of how this code works together, one
32 * should read the callers in dlmglue.c, the I/O functions in
33 * buffer_head_io.c and ocfs2_journal_access in journal.c
34 */
35
36#include <linux/fs.h>
37#include <linux/types.h>
38#include <linux/slab.h>
39#include <linux/highmem.h>
40#include <linux/buffer_head.h>
41#include <linux/rbtree.h>
Mark Fashehccd979b2005-12-15 14:31:24 -080042
Mark Fashehccd979b2005-12-15 14:31:24 -080043#include <cluster/masklog.h>
44
45#include "ocfs2.h"
46
47#include "inode.h"
48#include "uptodate.h"
Tao Mad7014852011-02-24 16:22:20 +080049#include "ocfs2_trace.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080050
51struct ocfs2_meta_cache_item {
52 struct rb_node c_node;
53 sector_t c_block;
54};
55
Fabian Frederick1a5c4e22014-06-04 16:06:06 -070056static struct kmem_cache *ocfs2_uptodate_cachep;
Mark Fashehccd979b2005-12-15 14:31:24 -080057
Joel Becker8cb471e2009-02-10 20:00:41 -080058u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci)
Mark Fashehccd979b2005-12-15 14:31:24 -080059{
Joel Becker6e5a3d72009-02-10 19:00:37 -080060 BUG_ON(!ci || !ci->ci_ops);
61
62 return ci->ci_ops->co_owner(ci);
63}
64
Joel Becker8cb471e2009-02-10 20:00:41 -080065struct super_block *ocfs2_metadata_cache_get_super(struct ocfs2_caching_info *ci)
66{
67 BUG_ON(!ci || !ci->ci_ops);
68
69 return ci->ci_ops->co_get_super(ci);
70}
71
Joel Becker6e5a3d72009-02-10 19:00:37 -080072static void ocfs2_metadata_cache_lock(struct ocfs2_caching_info *ci)
73{
74 BUG_ON(!ci || !ci->ci_ops);
75
76 ci->ci_ops->co_cache_lock(ci);
77}
78
79static void ocfs2_metadata_cache_unlock(struct ocfs2_caching_info *ci)
80{
81 BUG_ON(!ci || !ci->ci_ops);
82
83 ci->ci_ops->co_cache_unlock(ci);
84}
85
Joel Becker8cb471e2009-02-10 20:00:41 -080086void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci)
Joel Becker6e5a3d72009-02-10 19:00:37 -080087{
88 BUG_ON(!ci || !ci->ci_ops);
89
90 ci->ci_ops->co_io_lock(ci);
91}
92
Joel Becker8cb471e2009-02-10 20:00:41 -080093void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci)
Joel Becker6e5a3d72009-02-10 19:00:37 -080094{
95 BUG_ON(!ci || !ci->ci_ops);
96
97 ci->ci_ops->co_io_unlock(ci);
98}
99
100
Joel Becker66fb3452009-02-12 15:24:40 -0800101static void ocfs2_metadata_cache_reset(struct ocfs2_caching_info *ci,
102 int clear)
103{
104 ci->ci_flags |= OCFS2_CACHE_FL_INLINE;
105 ci->ci_num_cached = 0;
106
Joel Becker292dd272009-02-12 15:41:59 -0800107 if (clear) {
108 ci->ci_created_trans = 0;
Joel Becker66fb3452009-02-12 15:24:40 -0800109 ci->ci_last_trans = 0;
Joel Becker292dd272009-02-12 15:41:59 -0800110 }
Joel Becker66fb3452009-02-12 15:24:40 -0800111}
112
Joel Becker6e5a3d72009-02-10 19:00:37 -0800113void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,
114 const struct ocfs2_caching_operations *ops)
115{
116 BUG_ON(!ops);
117
118 ci->ci_ops = ops;
Joel Becker66fb3452009-02-12 15:24:40 -0800119 ocfs2_metadata_cache_reset(ci, 1);
Mark Fashehccd979b2005-12-15 14:31:24 -0800120}
121
Joel Becker66fb3452009-02-12 15:24:40 -0800122void ocfs2_metadata_cache_exit(struct ocfs2_caching_info *ci)
123{
124 ocfs2_metadata_cache_purge(ci);
125 ocfs2_metadata_cache_reset(ci, 1);
126}
127
128
Mark Fashehccd979b2005-12-15 14:31:24 -0800129/* No lock taken here as 'root' is not expected to be visible to other
130 * processes. */
131static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
132{
133 unsigned int purged = 0;
134 struct rb_node *node;
135 struct ocfs2_meta_cache_item *item;
136
137 while ((node = rb_last(root)) != NULL) {
138 item = rb_entry(node, struct ocfs2_meta_cache_item, c_node);
139
Tao Mad7014852011-02-24 16:22:20 +0800140 trace_ocfs2_purge_copied_metadata_tree(
141 (unsigned long long) item->c_block);
Mark Fashehccd979b2005-12-15 14:31:24 -0800142
143 rb_erase(&item->c_node, root);
144 kmem_cache_free(ocfs2_uptodate_cachep, item);
145
146 purged++;
147 }
148 return purged;
149}
150
151/* Called from locking and called from ocfs2_clear_inode. Dump the
152 * cache for a given inode.
153 *
154 * This function is a few more lines longer than necessary due to some
155 * accounting done here, but I think it's worth tracking down those
156 * bugs sooner -- Mark */
Joel Becker8cb471e2009-02-10 20:00:41 -0800157void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci)
Mark Fashehccd979b2005-12-15 14:31:24 -0800158{
Mark Fashehccd979b2005-12-15 14:31:24 -0800159 unsigned int tree, to_purge, purged;
Mark Fashehccd979b2005-12-15 14:31:24 -0800160 struct rb_root root = RB_ROOT;
161
Joel Becker6e5a3d72009-02-10 19:00:37 -0800162 BUG_ON(!ci || !ci->ci_ops);
163
164 ocfs2_metadata_cache_lock(ci);
Joel Becker47460d62009-02-10 16:05:07 -0800165 tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);
Mark Fashehccd979b2005-12-15 14:31:24 -0800166 to_purge = ci->ci_num_cached;
167
Tao Mad7014852011-02-24 16:22:20 +0800168 trace_ocfs2_metadata_cache_purge(
169 (unsigned long long)ocfs2_metadata_cache_owner(ci),
170 to_purge, tree);
Mark Fashehccd979b2005-12-15 14:31:24 -0800171
172 /* If we're a tree, save off the root so that we can safely
173 * initialize the cache. We do the work to free tree members
174 * without the spinlock. */
175 if (tree)
176 root = ci->ci_cache.ci_tree;
177
Joel Becker66fb3452009-02-12 15:24:40 -0800178 ocfs2_metadata_cache_reset(ci, 0);
Joel Becker6e5a3d72009-02-10 19:00:37 -0800179 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800180
181 purged = ocfs2_purge_copied_metadata_tree(&root);
182 /* If possible, track the number wiped so that we can more
183 * easily detect counting errors. Unfortunately, this is only
184 * meaningful for trees. */
185 if (tree && purged != to_purge)
Joel Becker6e5a3d72009-02-10 19:00:37 -0800186 mlog(ML_ERROR, "Owner %llu, count = %u, purged = %u\n",
187 (unsigned long long)ocfs2_metadata_cache_owner(ci),
188 to_purge, purged);
Mark Fashehccd979b2005-12-15 14:31:24 -0800189}
190
191/* Returns the index in the cache array, -1 if not found.
192 * Requires ip_lock. */
193static int ocfs2_search_cache_array(struct ocfs2_caching_info *ci,
194 sector_t item)
195{
196 int i;
197
198 for (i = 0; i < ci->ci_num_cached; i++) {
199 if (item == ci->ci_cache.ci_array[i])
200 return i;
201 }
202
203 return -1;
204}
205
206/* Returns the cache item if found, otherwise NULL.
207 * Requires ip_lock. */
208static struct ocfs2_meta_cache_item *
209ocfs2_search_cache_tree(struct ocfs2_caching_info *ci,
210 sector_t block)
211{
212 struct rb_node * n = ci->ci_cache.ci_tree.rb_node;
213 struct ocfs2_meta_cache_item *item = NULL;
214
215 while (n) {
216 item = rb_entry(n, struct ocfs2_meta_cache_item, c_node);
217
218 if (block < item->c_block)
219 n = n->rb_left;
220 else if (block > item->c_block)
221 n = n->rb_right;
222 else
223 return item;
224 }
225
226 return NULL;
227}
228
Joel Becker8cb471e2009-02-10 20:00:41 -0800229static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800230 struct buffer_head *bh)
231{
232 int index = -1;
233 struct ocfs2_meta_cache_item *item = NULL;
234
Joel Becker6e5a3d72009-02-10 19:00:37 -0800235 ocfs2_metadata_cache_lock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800236
Tao Mad7014852011-02-24 16:22:20 +0800237 trace_ocfs2_buffer_cached_begin(
238 (unsigned long long)ocfs2_metadata_cache_owner(ci),
239 (unsigned long long) bh->b_blocknr,
240 !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
Mark Fashehccd979b2005-12-15 14:31:24 -0800241
Joel Becker47460d62009-02-10 16:05:07 -0800242 if (ci->ci_flags & OCFS2_CACHE_FL_INLINE)
Joel Becker8cb471e2009-02-10 20:00:41 -0800243 index = ocfs2_search_cache_array(ci, bh->b_blocknr);
Mark Fashehccd979b2005-12-15 14:31:24 -0800244 else
Joel Becker8cb471e2009-02-10 20:00:41 -0800245 item = ocfs2_search_cache_tree(ci, bh->b_blocknr);
Mark Fashehccd979b2005-12-15 14:31:24 -0800246
Joel Becker6e5a3d72009-02-10 19:00:37 -0800247 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800248
Tao Mad7014852011-02-24 16:22:20 +0800249 trace_ocfs2_buffer_cached_end(index, item);
Mark Fashehccd979b2005-12-15 14:31:24 -0800250
251 return (index != -1) || (item != NULL);
252}
253
254/* Warning: even if it returns true, this does *not* guarantee that
Sunil Mushran2bd63212010-01-25 16:57:38 -0800255 * the block is stored in our inode metadata cache.
256 *
Mark Fashehaa958872006-04-21 13:49:02 -0700257 * This can be called under lock_buffer()
258 */
Joel Becker8cb471e2009-02-10 20:00:41 -0800259int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800260 struct buffer_head *bh)
261{
262 /* Doesn't matter if the bh is in our cache or not -- if it's
263 * not marked uptodate then we know it can't have correct
264 * data. */
265 if (!buffer_uptodate(bh))
266 return 0;
267
268 /* OCFS2 does not allow multiple nodes to be changing the same
269 * block at the same time. */
270 if (buffer_jbd(bh))
271 return 1;
272
273 /* Ok, locally the buffer is marked as up to date, now search
274 * our cache to see if we can trust that. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800275 return ocfs2_buffer_cached(ci, bh);
Mark Fashehccd979b2005-12-15 14:31:24 -0800276}
277
Joel Becker8cb471e2009-02-10 20:00:41 -0800278/*
Mark Fashehaa958872006-04-21 13:49:02 -0700279 * Determine whether a buffer is currently out on a read-ahead request.
Joel Becker47460d62009-02-10 16:05:07 -0800280 * ci_io_sem should be held to serialize submitters with the logic here.
Mark Fashehaa958872006-04-21 13:49:02 -0700281 */
Joel Becker8cb471e2009-02-10 20:00:41 -0800282int ocfs2_buffer_read_ahead(struct ocfs2_caching_info *ci,
Mark Fashehaa958872006-04-21 13:49:02 -0700283 struct buffer_head *bh)
284{
Joel Becker8cb471e2009-02-10 20:00:41 -0800285 return buffer_locked(bh) && ocfs2_buffer_cached(ci, bh);
Mark Fashehaa958872006-04-21 13:49:02 -0700286}
287
Mark Fashehccd979b2005-12-15 14:31:24 -0800288/* Requires ip_lock */
289static void ocfs2_append_cache_array(struct ocfs2_caching_info *ci,
290 sector_t block)
291{
Joel Becker47460d62009-02-10 16:05:07 -0800292 BUG_ON(ci->ci_num_cached >= OCFS2_CACHE_INFO_MAX_ARRAY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800293
Tao Mad7014852011-02-24 16:22:20 +0800294 trace_ocfs2_append_cache_array(
295 (unsigned long long)ocfs2_metadata_cache_owner(ci),
296 (unsigned long long)block, ci->ci_num_cached);
Mark Fashehccd979b2005-12-15 14:31:24 -0800297
298 ci->ci_cache.ci_array[ci->ci_num_cached] = block;
299 ci->ci_num_cached++;
300}
301
302/* By now the caller should have checked that the item does *not*
303 * exist in the tree.
304 * Requires ip_lock. */
305static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci,
306 struct ocfs2_meta_cache_item *new)
307{
308 sector_t block = new->c_block;
309 struct rb_node *parent = NULL;
310 struct rb_node **p = &ci->ci_cache.ci_tree.rb_node;
311 struct ocfs2_meta_cache_item *tmp;
312
Tao Mad7014852011-02-24 16:22:20 +0800313 trace_ocfs2_insert_cache_tree(
314 (unsigned long long)ocfs2_metadata_cache_owner(ci),
315 (unsigned long long)block, ci->ci_num_cached);
Mark Fashehccd979b2005-12-15 14:31:24 -0800316
317 while(*p) {
318 parent = *p;
319
320 tmp = rb_entry(parent, struct ocfs2_meta_cache_item, c_node);
321
322 if (block < tmp->c_block)
323 p = &(*p)->rb_left;
324 else if (block > tmp->c_block)
325 p = &(*p)->rb_right;
326 else {
327 /* This should never happen! */
328 mlog(ML_ERROR, "Duplicate block %llu cached!\n",
329 (unsigned long long) block);
330 BUG();
331 }
332 }
333
334 rb_link_node(&new->c_node, parent, p);
335 rb_insert_color(&new->c_node, &ci->ci_cache.ci_tree);
336 ci->ci_num_cached++;
337}
338
Joel Becker6e5a3d72009-02-10 19:00:37 -0800339/* co_cache_lock() must be held */
Joel Becker8cb471e2009-02-10 20:00:41 -0800340static inline int ocfs2_insert_can_use_array(struct ocfs2_caching_info *ci)
Mark Fashehccd979b2005-12-15 14:31:24 -0800341{
Joel Becker47460d62009-02-10 16:05:07 -0800342 return (ci->ci_flags & OCFS2_CACHE_FL_INLINE) &&
343 (ci->ci_num_cached < OCFS2_CACHE_INFO_MAX_ARRAY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800344}
345
Joel Becker47460d62009-02-10 16:05:07 -0800346/* tree should be exactly OCFS2_CACHE_INFO_MAX_ARRAY wide. NULL the
Mark Fashehccd979b2005-12-15 14:31:24 -0800347 * pointers in tree after we use them - this allows caller to detect
Joel Becker6e5a3d72009-02-10 19:00:37 -0800348 * when to free in case of error.
349 *
350 * The co_cache_lock() must be held. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800351static void ocfs2_expand_cache(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800352 struct ocfs2_meta_cache_item **tree)
353{
354 int i;
Mark Fashehccd979b2005-12-15 14:31:24 -0800355
Joel Becker47460d62009-02-10 16:05:07 -0800356 mlog_bug_on_msg(ci->ci_num_cached != OCFS2_CACHE_INFO_MAX_ARRAY,
Joel Becker6e5a3d72009-02-10 19:00:37 -0800357 "Owner %llu, num cached = %u, should be %u\n",
358 (unsigned long long)ocfs2_metadata_cache_owner(ci),
359 ci->ci_num_cached, OCFS2_CACHE_INFO_MAX_ARRAY);
Joel Becker47460d62009-02-10 16:05:07 -0800360 mlog_bug_on_msg(!(ci->ci_flags & OCFS2_CACHE_FL_INLINE),
Joel Becker6e5a3d72009-02-10 19:00:37 -0800361 "Owner %llu not marked as inline anymore!\n",
362 (unsigned long long)ocfs2_metadata_cache_owner(ci));
Mark Fashehccd979b2005-12-15 14:31:24 -0800363
364 /* Be careful to initialize the tree members *first* because
365 * once the ci_tree is used, the array is junk... */
Joel Becker47460d62009-02-10 16:05:07 -0800366 for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
Mark Fashehccd979b2005-12-15 14:31:24 -0800367 tree[i]->c_block = ci->ci_cache.ci_array[i];
368
Joel Becker47460d62009-02-10 16:05:07 -0800369 ci->ci_flags &= ~OCFS2_CACHE_FL_INLINE;
Mark Fashehccd979b2005-12-15 14:31:24 -0800370 ci->ci_cache.ci_tree = RB_ROOT;
371 /* this will be set again by __ocfs2_insert_cache_tree */
372 ci->ci_num_cached = 0;
373
Joel Becker47460d62009-02-10 16:05:07 -0800374 for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800375 __ocfs2_insert_cache_tree(ci, tree[i]);
376 tree[i] = NULL;
377 }
378
Tao Mad7014852011-02-24 16:22:20 +0800379 trace_ocfs2_expand_cache(
380 (unsigned long long)ocfs2_metadata_cache_owner(ci),
381 ci->ci_flags, ci->ci_num_cached);
Mark Fashehccd979b2005-12-15 14:31:24 -0800382}
383
384/* Slow path function - memory allocation is necessary. See the
385 * comment above ocfs2_set_buffer_uptodate for more information. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800386static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800387 sector_t block,
388 int expand_tree)
389{
390 int i;
Mark Fashehccd979b2005-12-15 14:31:24 -0800391 struct ocfs2_meta_cache_item *new = NULL;
Joel Becker47460d62009-02-10 16:05:07 -0800392 struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =
Mark Fashehccd979b2005-12-15 14:31:24 -0800393 { NULL, };
394
Tao Mad7014852011-02-24 16:22:20 +0800395 trace_ocfs2_set_buffer_uptodate(
396 (unsigned long long)ocfs2_metadata_cache_owner(ci),
397 (unsigned long long)block, expand_tree);
Mark Fashehccd979b2005-12-15 14:31:24 -0800398
Sunil Mushranafae00ab2006-04-12 14:37:00 -0700399 new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
Mark Fashehccd979b2005-12-15 14:31:24 -0800400 if (!new) {
401 mlog_errno(-ENOMEM);
402 return;
403 }
404 new->c_block = block;
405
406 if (expand_tree) {
407 /* Do *not* allocate an array here - the removal code
408 * has no way of tracking that. */
Joel Becker47460d62009-02-10 16:05:07 -0800409 for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800410 tree[i] = kmem_cache_alloc(ocfs2_uptodate_cachep,
Sunil Mushranafae00ab2006-04-12 14:37:00 -0700411 GFP_NOFS);
Mark Fashehccd979b2005-12-15 14:31:24 -0800412 if (!tree[i]) {
413 mlog_errno(-ENOMEM);
414 goto out_free;
415 }
416
417 /* These are initialized in ocfs2_expand_cache! */
418 }
419 }
420
Joel Becker6e5a3d72009-02-10 19:00:37 -0800421 ocfs2_metadata_cache_lock(ci);
Joel Becker8cb471e2009-02-10 20:00:41 -0800422 if (ocfs2_insert_can_use_array(ci)) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800423 /* Ok, items were removed from the cache in between
424 * locks. Detect this and revert back to the fast path */
425 ocfs2_append_cache_array(ci, block);
Joel Becker6e5a3d72009-02-10 19:00:37 -0800426 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800427 goto out_free;
428 }
429
430 if (expand_tree)
Joel Becker8cb471e2009-02-10 20:00:41 -0800431 ocfs2_expand_cache(ci, tree);
Mark Fashehccd979b2005-12-15 14:31:24 -0800432
433 __ocfs2_insert_cache_tree(ci, new);
Joel Becker6e5a3d72009-02-10 19:00:37 -0800434 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800435
436 new = NULL;
437out_free:
438 if (new)
439 kmem_cache_free(ocfs2_uptodate_cachep, new);
440
441 /* If these were used, then ocfs2_expand_cache re-set them to
442 * NULL for us. */
443 if (tree[0]) {
Joel Becker47460d62009-02-10 16:05:07 -0800444 for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
Mark Fashehccd979b2005-12-15 14:31:24 -0800445 if (tree[i])
446 kmem_cache_free(ocfs2_uptodate_cachep,
447 tree[i]);
448 }
449}
450
Joel Becker6e5a3d72009-02-10 19:00:37 -0800451/* Item insertion is guarded by co_io_lock(), so the insertion path takes
Mark Fashehccd979b2005-12-15 14:31:24 -0800452 * advantage of this by not rechecking for a duplicate insert during
453 * the slow case. Additionally, if the cache needs to be bumped up to
454 * a tree, the code will not recheck after acquiring the lock --
455 * multiple paths cannot be expanding to a tree at the same time.
456 *
457 * The slow path takes into account that items can be removed
458 * (including the whole tree wiped and reset) when this process it out
459 * allocating memory. In those cases, it reverts back to the fast
460 * path.
461 *
462 * Note that this function may actually fail to insert the block if
463 * memory cannot be allocated. This is not fatal however (but may
Mark Fashehaa958872006-04-21 13:49:02 -0700464 * result in a performance penalty)
465 *
466 * Readahead buffers can be passed in here before the I/O request is
467 * completed.
468 */
Joel Becker8cb471e2009-02-10 20:00:41 -0800469void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800470 struct buffer_head *bh)
471{
472 int expand;
Mark Fashehccd979b2005-12-15 14:31:24 -0800473
474 /* The block may very well exist in our cache already, so avoid
475 * doing any more work in that case. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800476 if (ocfs2_buffer_cached(ci, bh))
Mark Fashehccd979b2005-12-15 14:31:24 -0800477 return;
478
Tao Mad7014852011-02-24 16:22:20 +0800479 trace_ocfs2_set_buffer_uptodate_begin(
480 (unsigned long long)ocfs2_metadata_cache_owner(ci),
481 (unsigned long long)bh->b_blocknr);
Mark Fashehccd979b2005-12-15 14:31:24 -0800482
483 /* No need to recheck under spinlock - insertion is guarded by
Joel Becker6e5a3d72009-02-10 19:00:37 -0800484 * co_io_lock() */
485 ocfs2_metadata_cache_lock(ci);
Joel Becker8cb471e2009-02-10 20:00:41 -0800486 if (ocfs2_insert_can_use_array(ci)) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800487 /* Fast case - it's an array and there's a free
488 * spot. */
489 ocfs2_append_cache_array(ci, bh->b_blocknr);
Joel Becker6e5a3d72009-02-10 19:00:37 -0800490 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800491 return;
492 }
493
494 expand = 0;
Joel Becker47460d62009-02-10 16:05:07 -0800495 if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800496 /* We need to bump things up to a tree. */
497 expand = 1;
498 }
Joel Becker6e5a3d72009-02-10 19:00:37 -0800499 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800500
Joel Becker8cb471e2009-02-10 20:00:41 -0800501 __ocfs2_set_buffer_uptodate(ci, bh->b_blocknr, expand);
Mark Fashehccd979b2005-12-15 14:31:24 -0800502}
503
504/* Called against a newly allocated buffer. Most likely nobody should
505 * be able to read this sort of metadata while it's still being
Joel Becker6e5a3d72009-02-10 19:00:37 -0800506 * allocated, but this is careful to take co_io_lock() anyway. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800507void ocfs2_set_new_buffer_uptodate(struct ocfs2_caching_info *ci,
Mark Fashehccd979b2005-12-15 14:31:24 -0800508 struct buffer_head *bh)
509{
Mark Fashehccd979b2005-12-15 14:31:24 -0800510 /* This should definitely *not* exist in our cache */
Joel Becker8cb471e2009-02-10 20:00:41 -0800511 BUG_ON(ocfs2_buffer_cached(ci, bh));
Mark Fashehccd979b2005-12-15 14:31:24 -0800512
513 set_buffer_uptodate(bh);
514
Joel Becker6e5a3d72009-02-10 19:00:37 -0800515 ocfs2_metadata_cache_io_lock(ci);
Joel Becker8cb471e2009-02-10 20:00:41 -0800516 ocfs2_set_buffer_uptodate(ci, bh);
Joel Becker6e5a3d72009-02-10 19:00:37 -0800517 ocfs2_metadata_cache_io_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800518}
519
520/* Requires ip_lock. */
521static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
522 int index)
523{
524 sector_t *array = ci->ci_cache.ci_array;
525 int bytes;
526
Joel Becker47460d62009-02-10 16:05:07 -0800527 BUG_ON(index < 0 || index >= OCFS2_CACHE_INFO_MAX_ARRAY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800528 BUG_ON(index >= ci->ci_num_cached);
529 BUG_ON(!ci->ci_num_cached);
530
Tao Mad7014852011-02-24 16:22:20 +0800531 trace_ocfs2_remove_metadata_array(
532 (unsigned long long)ocfs2_metadata_cache_owner(ci),
533 index, ci->ci_num_cached);
Mark Fashehccd979b2005-12-15 14:31:24 -0800534
535 ci->ci_num_cached--;
536
537 /* don't need to copy if the array is now empty, or if we
538 * removed at the tail */
539 if (ci->ci_num_cached && index < ci->ci_num_cached) {
540 bytes = sizeof(sector_t) * (ci->ci_num_cached - index);
541 memmove(&array[index], &array[index + 1], bytes);
542 }
543}
544
545/* Requires ip_lock. */
546static void ocfs2_remove_metadata_tree(struct ocfs2_caching_info *ci,
547 struct ocfs2_meta_cache_item *item)
548{
Tao Mad7014852011-02-24 16:22:20 +0800549 trace_ocfs2_remove_metadata_tree(
550 (unsigned long long)ocfs2_metadata_cache_owner(ci),
551 (unsigned long long)item->c_block);
Mark Fashehccd979b2005-12-15 14:31:24 -0800552
553 rb_erase(&item->c_node, &ci->ci_cache.ci_tree);
554 ci->ci_num_cached--;
555}
556
Joel Becker8cb471e2009-02-10 20:00:41 -0800557static void ocfs2_remove_block_from_cache(struct ocfs2_caching_info *ci,
Tao Maac11c822008-08-18 17:38:47 +0800558 sector_t block)
Mark Fashehccd979b2005-12-15 14:31:24 -0800559{
560 int index;
Mark Fashehccd979b2005-12-15 14:31:24 -0800561 struct ocfs2_meta_cache_item *item = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -0800562
Joel Becker6e5a3d72009-02-10 19:00:37 -0800563 ocfs2_metadata_cache_lock(ci);
Tao Mad7014852011-02-24 16:22:20 +0800564 trace_ocfs2_remove_block_from_cache(
565 (unsigned long long)ocfs2_metadata_cache_owner(ci),
566 (unsigned long long) block, ci->ci_num_cached,
567 ci->ci_flags);
Mark Fashehccd979b2005-12-15 14:31:24 -0800568
Joel Becker47460d62009-02-10 16:05:07 -0800569 if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800570 index = ocfs2_search_cache_array(ci, block);
571 if (index != -1)
572 ocfs2_remove_metadata_array(ci, index);
573 } else {
574 item = ocfs2_search_cache_tree(ci, block);
575 if (item)
576 ocfs2_remove_metadata_tree(ci, item);
577 }
Joel Becker6e5a3d72009-02-10 19:00:37 -0800578 ocfs2_metadata_cache_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800579
580 if (item)
581 kmem_cache_free(ocfs2_uptodate_cachep, item);
582}
583
Tao Maac11c822008-08-18 17:38:47 +0800584/*
585 * Called when we remove a chunk of metadata from an inode. We don't
586 * bother reverting things to an inlined array in the case of a remove
587 * which moves us back under the limit.
588 */
Joel Becker8cb471e2009-02-10 20:00:41 -0800589void ocfs2_remove_from_cache(struct ocfs2_caching_info *ci,
Tao Maac11c822008-08-18 17:38:47 +0800590 struct buffer_head *bh)
591{
592 sector_t block = bh->b_blocknr;
593
Joel Becker8cb471e2009-02-10 20:00:41 -0800594 ocfs2_remove_block_from_cache(ci, block);
Tao Maac11c822008-08-18 17:38:47 +0800595}
596
597/* Called when we remove xattr clusters from an inode. */
Joel Becker8cb471e2009-02-10 20:00:41 -0800598void ocfs2_remove_xattr_clusters_from_cache(struct ocfs2_caching_info *ci,
Tao Maac11c822008-08-18 17:38:47 +0800599 sector_t block,
600 u32 c_len)
601{
Joel Becker8cb471e2009-02-10 20:00:41 -0800602 struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
603 unsigned int i, b_len = ocfs2_clusters_to_blocks(sb, 1) * c_len;
Tao Maac11c822008-08-18 17:38:47 +0800604
605 for (i = 0; i < b_len; i++, block++)
Joel Becker8cb471e2009-02-10 20:00:41 -0800606 ocfs2_remove_block_from_cache(ci, block);
Tao Maac11c822008-08-18 17:38:47 +0800607}
608
Mark Fashehccd979b2005-12-15 14:31:24 -0800609int __init init_ocfs2_uptodate_cache(void)
610{
611 ocfs2_uptodate_cachep = kmem_cache_create("ocfs2_uptodate",
612 sizeof(struct ocfs2_meta_cache_item),
Paul Mundt20c2df82007-07-20 10:11:58 +0900613 0, SLAB_HWCACHE_ALIGN, NULL);
Mark Fashehccd979b2005-12-15 14:31:24 -0800614 if (!ocfs2_uptodate_cachep)
615 return -ENOMEM;
616
Mark Fashehccd979b2005-12-15 14:31:24 -0800617 return 0;
618}
619
Adrian Bunk0c6c98f2006-01-07 20:07:02 +0100620void exit_ocfs2_uptodate_cache(void)
Mark Fashehccd979b2005-12-15 14:31:24 -0800621{
piaojuna17b4852018-04-05 16:18:52 -0700622 kmem_cache_destroy(ocfs2_uptodate_cachep);
Mark Fashehccd979b2005-12-15 14:31:24 -0800623}