^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-or-later
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /* -*- mode: c; c-basic-offset: 8; -*-
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * vim: noexpandtab sw=8 ts=8 sts=0:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * uptodate.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Tracking the up-to-date-ness of a local buffer_head with respect to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * the cluster.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * Copyright (C) 2002, 2004, 2005 Oracle. All rights reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) * Standard buffer head caching flags (uptodate, etc) are insufficient
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) * in a clustered environment - a buffer may be marked up to date on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) * our local node but could have been modified by another cluster
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) * member. As a result an additional (and performant) caching scheme
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) * is required. A further requirement is that we consume as little
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) * memory as possible - we never pin buffer_head structures in order
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) * to cache them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) * We track the existence of up to date buffers on the inodes which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) * are associated with them. Because we don't want to pin
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) * buffer_heads, this is only a (strong) hint and several other checks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) * are made in the I/O path to ensure that we don't use a stale or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) * invalid buffer without going to disk:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) * - buffer_jbd is used liberally - if a bh is in the journal on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) * this node then it *must* be up to date.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) * - the standard buffer_uptodate() macro is used to detect buffers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) * which may be invalid (even if we have an up to date tracking
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) * item for them)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) * For a full understanding of how this code works together, one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) * should read the callers in dlmglue.c, the I/O functions in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) * buffer_head_io.c and ocfs2_journal_access in journal.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) #include <linux/fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) #include <linux/types.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) #include <linux/highmem.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) #include <linux/buffer_head.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) #include <linux/rbtree.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) #include <cluster/masklog.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) #include "ocfs2.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) #include "inode.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) #include "uptodate.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) #include "ocfs2_trace.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) struct ocfs2_meta_cache_item {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) struct rb_node c_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) sector_t c_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) static struct kmem_cache *ocfs2_uptodate_cachep;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) return ci->ci_ops->co_owner(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) struct super_block *ocfs2_metadata_cache_get_super(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) return ci->ci_ops->co_get_super(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) static void ocfs2_metadata_cache_lock(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) ci->ci_ops->co_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) static void ocfs2_metadata_cache_unlock(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) ci->ci_ops->co_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) ci->ci_ops->co_io_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) ci->ci_ops->co_io_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) static void ocfs2_metadata_cache_reset(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) int clear)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) ci->ci_flags |= OCFS2_CACHE_FL_INLINE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) ci->ci_num_cached = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) if (clear) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) ci->ci_created_trans = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) ci->ci_last_trans = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) const struct ocfs2_caching_operations *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) BUG_ON(!ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) ci->ci_ops = ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) ocfs2_metadata_cache_reset(ci, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) void ocfs2_metadata_cache_exit(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) ocfs2_metadata_cache_purge(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) ocfs2_metadata_cache_reset(ci, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) /* No lock taken here as 'root' is not expected to be visible to other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) * processes. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) unsigned int purged = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) struct rb_node *node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) struct ocfs2_meta_cache_item *item;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) while ((node = rb_last(root)) != NULL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) item = rb_entry(node, struct ocfs2_meta_cache_item, c_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) trace_ocfs2_purge_copied_metadata_tree(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) (unsigned long long) item->c_block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) rb_erase(&item->c_node, root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) kmem_cache_free(ocfs2_uptodate_cachep, item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) purged++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) return purged;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) /* Called from locking and called from ocfs2_clear_inode. Dump the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) * cache for a given inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) * This function is a few more lines longer than necessary due to some
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) * accounting done here, but I think it's worth tracking down those
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) * bugs sooner -- Mark */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) unsigned int tree, to_purge, purged;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) struct rb_root root = RB_ROOT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) BUG_ON(!ci || !ci->ci_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) ocfs2_metadata_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) to_purge = ci->ci_num_cached;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) trace_ocfs2_metadata_cache_purge(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) to_purge, tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) /* If we're a tree, save off the root so that we can safely
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) * initialize the cache. We do the work to free tree members
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) * without the spinlock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) if (tree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) root = ci->ci_cache.ci_tree;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) ocfs2_metadata_cache_reset(ci, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) purged = ocfs2_purge_copied_metadata_tree(&root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) /* If possible, track the number wiped so that we can more
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) * easily detect counting errors. Unfortunately, this is only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) * meaningful for trees. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) if (tree && purged != to_purge)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) mlog(ML_ERROR, "Owner %llu, count = %u, purged = %u\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) to_purge, purged);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) /* Returns the index in the cache array, -1 if not found.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) * Requires ip_lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) static int ocfs2_search_cache_array(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) sector_t item)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) for (i = 0; i < ci->ci_num_cached; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) if (item == ci->ci_cache.ci_array[i])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) return i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) /* Returns the cache item if found, otherwise NULL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * Requires ip_lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) static struct ocfs2_meta_cache_item *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) ocfs2_search_cache_tree(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) sector_t block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) struct rb_node * n = ci->ci_cache.ci_tree.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) struct ocfs2_meta_cache_item *item = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) while (n) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) item = rb_entry(n, struct ocfs2_meta_cache_item, c_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) if (block < item->c_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) n = n->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) else if (block > item->c_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) n = n->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) return item;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) int index = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) struct ocfs2_meta_cache_item *item = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) ocfs2_metadata_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) trace_ocfs2_buffer_cached_begin(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) (unsigned long long) bh->b_blocknr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) if (ci->ci_flags & OCFS2_CACHE_FL_INLINE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) index = ocfs2_search_cache_array(ci, bh->b_blocknr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) item = ocfs2_search_cache_tree(ci, bh->b_blocknr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) trace_ocfs2_buffer_cached_end(index, item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) return (index != -1) || (item != NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) /* Warning: even if it returns true, this does *not* guarantee that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) * the block is stored in our inode metadata cache.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) * This can be called under lock_buffer()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) /* Doesn't matter if the bh is in our cache or not -- if it's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) * not marked uptodate then we know it can't have correct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) * data. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) if (!buffer_uptodate(bh))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) /* OCFS2 does not allow multiple nodes to be changing the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) * block at the same time. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) if (buffer_jbd(bh))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) /* Ok, locally the buffer is marked as up to date, now search
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) * our cache to see if we can trust that. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) return ocfs2_buffer_cached(ci, bh);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) * Determine whether a buffer is currently out on a read-ahead request.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) * ci_io_sem should be held to serialize submitters with the logic here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) int ocfs2_buffer_read_ahead(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) return buffer_locked(bh) && ocfs2_buffer_cached(ci, bh);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) /* Requires ip_lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) static void ocfs2_append_cache_array(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) sector_t block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) BUG_ON(ci->ci_num_cached >= OCFS2_CACHE_INFO_MAX_ARRAY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) trace_ocfs2_append_cache_array(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) (unsigned long long)block, ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) ci->ci_cache.ci_array[ci->ci_num_cached] = block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) ci->ci_num_cached++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) /* By now the caller should have checked that the item does *not*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) * exist in the tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) * Requires ip_lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) struct ocfs2_meta_cache_item *new)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) sector_t block = new->c_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) struct rb_node *parent = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) struct rb_node **p = &ci->ci_cache.ci_tree.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) struct ocfs2_meta_cache_item *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) trace_ocfs2_insert_cache_tree(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) (unsigned long long)block, ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) while(*p) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) parent = *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) tmp = rb_entry(parent, struct ocfs2_meta_cache_item, c_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) if (block < tmp->c_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) p = &(*p)->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) else if (block > tmp->c_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) p = &(*p)->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) /* This should never happen! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) mlog(ML_ERROR, "Duplicate block %llu cached!\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) (unsigned long long) block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) BUG();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) rb_link_node(&new->c_node, parent, p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) rb_insert_color(&new->c_node, &ci->ci_cache.ci_tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) ci->ci_num_cached++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) /* co_cache_lock() must be held */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) static inline int ocfs2_insert_can_use_array(struct ocfs2_caching_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) return (ci->ci_flags & OCFS2_CACHE_FL_INLINE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) (ci->ci_num_cached < OCFS2_CACHE_INFO_MAX_ARRAY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) /* tree should be exactly OCFS2_CACHE_INFO_MAX_ARRAY wide. NULL the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) * pointers in tree after we use them - this allows caller to detect
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) * when to free in case of error.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) * The co_cache_lock() must be held. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) static void ocfs2_expand_cache(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) struct ocfs2_meta_cache_item **tree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) mlog_bug_on_msg(ci->ci_num_cached != OCFS2_CACHE_INFO_MAX_ARRAY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) "Owner %llu, num cached = %u, should be %u\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) ci->ci_num_cached, OCFS2_CACHE_INFO_MAX_ARRAY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) mlog_bug_on_msg(!(ci->ci_flags & OCFS2_CACHE_FL_INLINE),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) "Owner %llu not marked as inline anymore!\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) (unsigned long long)ocfs2_metadata_cache_owner(ci));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) /* Be careful to initialize the tree members *first* because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) * once the ci_tree is used, the array is junk... */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) tree[i]->c_block = ci->ci_cache.ci_array[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) ci->ci_flags &= ~OCFS2_CACHE_FL_INLINE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) ci->ci_cache.ci_tree = RB_ROOT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) /* this will be set again by __ocfs2_insert_cache_tree */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) ci->ci_num_cached = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) __ocfs2_insert_cache_tree(ci, tree[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) tree[i] = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) trace_ocfs2_expand_cache(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) ci->ci_flags, ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) /* Slow path function - memory allocation is necessary. See the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) * comment above ocfs2_set_buffer_uptodate for more information. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) sector_t block,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) int expand_tree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) struct ocfs2_meta_cache_item *new = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) { NULL, };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) trace_ocfs2_set_buffer_uptodate(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) (unsigned long long)block, expand_tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) if (!new) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) mlog_errno(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) new->c_block = block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) if (expand_tree) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) /* Do *not* allocate an array here - the removal code
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) * has no way of tracking that. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) tree[i] = kmem_cache_alloc(ocfs2_uptodate_cachep,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) GFP_NOFS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) if (!tree[i]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) mlog_errno(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) /* These are initialized in ocfs2_expand_cache! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) ocfs2_metadata_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) if (ocfs2_insert_can_use_array(ci)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) /* Ok, items were removed from the cache in between
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) * locks. Detect this and revert back to the fast path */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) ocfs2_append_cache_array(ci, block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) if (expand_tree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) ocfs2_expand_cache(ci, tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) __ocfs2_insert_cache_tree(ci, new);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) new = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) out_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) if (new)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) kmem_cache_free(ocfs2_uptodate_cachep, new);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) /* If these were used, then ocfs2_expand_cache re-set them to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) * NULL for us. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) if (tree[0]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) if (tree[i])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) kmem_cache_free(ocfs2_uptodate_cachep,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) tree[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) /* Item insertion is guarded by co_io_lock(), so the insertion path takes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) * advantage of this by not rechecking for a duplicate insert during
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) * the slow case. Additionally, if the cache needs to be bumped up to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * a tree, the code will not recheck after acquiring the lock --
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) * multiple paths cannot be expanding to a tree at the same time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) * The slow path takes into account that items can be removed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) * (including the whole tree wiped and reset) when this process it out
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) * allocating memory. In those cases, it reverts back to the fast
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) * path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) * Note that this function may actually fail to insert the block if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) * memory cannot be allocated. This is not fatal however (but may
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) * result in a performance penalty)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) * Readahead buffers can be passed in here before the I/O request is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) * completed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) int expand;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) /* The block may very well exist in our cache already, so avoid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) * doing any more work in that case. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) if (ocfs2_buffer_cached(ci, bh))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) trace_ocfs2_set_buffer_uptodate_begin(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) (unsigned long long)bh->b_blocknr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) /* No need to recheck under spinlock - insertion is guarded by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) * co_io_lock() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) ocfs2_metadata_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) if (ocfs2_insert_can_use_array(ci)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) /* Fast case - it's an array and there's a free
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) * spot. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) ocfs2_append_cache_array(ci, bh->b_blocknr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) expand = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) /* We need to bump things up to a tree. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) expand = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) __ocfs2_set_buffer_uptodate(ci, bh->b_blocknr, expand);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) /* Called against a newly allocated buffer. Most likely nobody should
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) * be able to read this sort of metadata while it's still being
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) * allocated, but this is careful to take co_io_lock() anyway. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) void ocfs2_set_new_buffer_uptodate(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) /* This should definitely *not* exist in our cache */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) BUG_ON(ocfs2_buffer_cached(ci, bh));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) set_buffer_uptodate(bh);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) ocfs2_metadata_cache_io_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) ocfs2_set_buffer_uptodate(ci, bh);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) ocfs2_metadata_cache_io_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) /* Requires ip_lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) int index)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) sector_t *array = ci->ci_cache.ci_array;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) int bytes;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) BUG_ON(index < 0 || index >= OCFS2_CACHE_INFO_MAX_ARRAY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) BUG_ON(index >= ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) BUG_ON(!ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) trace_ocfs2_remove_metadata_array(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) index, ci->ci_num_cached);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) ci->ci_num_cached--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) /* don't need to copy if the array is now empty, or if we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) * removed at the tail */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) if (ci->ci_num_cached && index < ci->ci_num_cached) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) bytes = sizeof(sector_t) * (ci->ci_num_cached - index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) memmove(&array[index], &array[index + 1], bytes);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) /* Requires ip_lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) static void ocfs2_remove_metadata_tree(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) struct ocfs2_meta_cache_item *item)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) trace_ocfs2_remove_metadata_tree(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) (unsigned long long)item->c_block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) rb_erase(&item->c_node, &ci->ci_cache.ci_tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) ci->ci_num_cached--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) static void ocfs2_remove_block_from_cache(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) sector_t block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) int index;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) struct ocfs2_meta_cache_item *item = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) ocfs2_metadata_cache_lock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) trace_ocfs2_remove_block_from_cache(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) (unsigned long long)ocfs2_metadata_cache_owner(ci),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) (unsigned long long) block, ci->ci_num_cached,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) ci->ci_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) index = ocfs2_search_cache_array(ci, block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) if (index != -1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) ocfs2_remove_metadata_array(ci, index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) item = ocfs2_search_cache_tree(ci, block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) if (item)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) ocfs2_remove_metadata_tree(ci, item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) ocfs2_metadata_cache_unlock(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) if (item)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) kmem_cache_free(ocfs2_uptodate_cachep, item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) * Called when we remove a chunk of metadata from an inode. We don't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) * bother reverting things to an inlined array in the case of a remove
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) * which moves us back under the limit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) void ocfs2_remove_from_cache(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) struct buffer_head *bh)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) sector_t block = bh->b_blocknr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) ocfs2_remove_block_from_cache(ci, block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) /* Called when we remove xattr clusters from an inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) void ocfs2_remove_xattr_clusters_from_cache(struct ocfs2_caching_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) sector_t block,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) u32 c_len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) unsigned int i, b_len = ocfs2_clusters_to_blocks(sb, 1) * c_len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) for (i = 0; i < b_len; i++, block++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) ocfs2_remove_block_from_cache(ci, block);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) int __init init_ocfs2_uptodate_cache(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) ocfs2_uptodate_cachep = kmem_cache_create("ocfs2_uptodate",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) sizeof(struct ocfs2_meta_cache_item),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) 0, SLAB_HWCACHE_ALIGN, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) if (!ocfs2_uptodate_cachep)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) void exit_ocfs2_uptodate_cache(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) kmem_cache_destroy(ocfs2_uptodate_cachep);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) }