2005-12-15 22:31:24 +00:00
|
|
|
/* -*- mode: c; c-basic-offset: 8; -*-
|
|
|
|
* vim: noexpandtab sw=8 ts=8 sts=0:
|
|
|
|
*
|
|
|
|
* uptodate.c
|
|
|
|
*
|
|
|
|
* Tracking the up-to-date-ness of a local buffer_head with respect to
|
|
|
|
* the cluster.
|
|
|
|
*
|
|
|
|
* Copyright (C) 2002, 2004, 2005 Oracle. All rights reserved.
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
|
|
|
* version 2 of the License, or (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public
|
|
|
|
* License along with this program; if not, write to the
|
|
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
|
* Boston, MA 021110-1307, USA.
|
|
|
|
*
|
|
|
|
* Standard buffer head caching flags (uptodate, etc) are insufficient
|
|
|
|
* in a clustered environment - a buffer may be marked up to date on
|
|
|
|
* our local node but could have been modified by another cluster
|
|
|
|
* member. As a result an additional (and performant) caching scheme
|
|
|
|
* is required. A further requirement is that we consume as little
|
|
|
|
* memory as possible - we never pin buffer_head structures in order
|
|
|
|
* to cache them.
|
|
|
|
*
|
|
|
|
* We track the existence of up to date buffers on the inodes which
|
|
|
|
* are associated with them. Because we don't want to pin
|
|
|
|
* buffer_heads, this is only a (strong) hint and several other checks
|
|
|
|
* are made in the I/O path to ensure that we don't use a stale or
|
|
|
|
* invalid buffer without going to disk:
|
|
|
|
* - buffer_jbd is used liberally - if a bh is in the journal on
|
|
|
|
* this node then it *must* be up to date.
|
|
|
|
* - the standard buffer_uptodate() macro is used to detect buffers
|
|
|
|
* which may be invalid (even if we have an up to date tracking
|
|
|
|
* item for them)
|
|
|
|
*
|
|
|
|
* For a full understanding of how this code works together, one
|
|
|
|
* should read the callers in dlmglue.c, the I/O functions in
|
|
|
|
* buffer_head_io.c and ocfs2_journal_access in journal.c
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/fs.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/buffer_head.h>
|
|
|
|
#include <linux/rbtree.h>
|
|
|
|
|
|
|
|
#include <cluster/masklog.h>
|
|
|
|
|
|
|
|
#include "ocfs2.h"
|
|
|
|
|
|
|
|
#include "inode.h"
|
|
|
|
#include "uptodate.h"
|
2011-02-24 08:22:20 +00:00
|
|
|
#include "ocfs2_trace.h"
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
struct ocfs2_meta_cache_item {
|
|
|
|
struct rb_node c_node;
|
|
|
|
sector_t c_block;
|
|
|
|
};
|
|
|
|
|
2006-12-07 04:33:20 +00:00
|
|
|
static struct kmem_cache *ocfs2_uptodate_cachep = NULL;
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci)
|
2009-02-11 03:00:37 +00:00
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
return ci->ci_ops->co_owner(ci);
|
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
struct super_block *ocfs2_metadata_cache_get_super(struct ocfs2_caching_info *ci)
|
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
return ci->ci_ops->co_get_super(ci);
|
|
|
|
}
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
static void ocfs2_metadata_cache_lock(struct ocfs2_caching_info *ci)
|
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
ci->ci_ops->co_cache_lock(ci);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ocfs2_metadata_cache_unlock(struct ocfs2_caching_info *ci)
|
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
ci->ci_ops->co_cache_unlock(ci);
|
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci)
|
2009-02-11 03:00:37 +00:00
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
ci->ci_ops->co_io_lock(ci);
|
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci)
|
2009-02-11 03:00:37 +00:00
|
|
|
{
|
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
ci->ci_ops->co_io_unlock(ci);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-02-12 23:24:40 +00:00
|
|
|
static void ocfs2_metadata_cache_reset(struct ocfs2_caching_info *ci,
|
|
|
|
int clear)
|
|
|
|
{
|
|
|
|
ci->ci_flags |= OCFS2_CACHE_FL_INLINE;
|
|
|
|
ci->ci_num_cached = 0;
|
|
|
|
|
2009-02-12 23:41:59 +00:00
|
|
|
if (clear) {
|
|
|
|
ci->ci_created_trans = 0;
|
2009-02-12 23:24:40 +00:00
|
|
|
ci->ci_last_trans = 0;
|
2009-02-12 23:41:59 +00:00
|
|
|
}
|
2009-02-12 23:24:40 +00:00
|
|
|
}
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,
|
2009-02-11 03:00:37 +00:00
|
|
|
const struct ocfs2_caching_operations *ops)
|
2005-12-15 22:31:24 +00:00
|
|
|
{
|
2009-02-11 03:00:37 +00:00
|
|
|
BUG_ON(!ops);
|
|
|
|
|
|
|
|
ci->ci_ops = ops;
|
2009-02-12 23:24:40 +00:00
|
|
|
ocfs2_metadata_cache_reset(ci, 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
void ocfs2_metadata_cache_exit(struct ocfs2_caching_info *ci)
|
|
|
|
{
|
|
|
|
ocfs2_metadata_cache_purge(ci);
|
|
|
|
ocfs2_metadata_cache_reset(ci, 1);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
2009-02-12 23:24:40 +00:00
|
|
|
|
2005-12-15 22:31:24 +00:00
|
|
|
/* No lock taken here as 'root' is not expected to be visible to other
|
|
|
|
* processes. */
|
|
|
|
static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
|
|
|
|
{
|
|
|
|
unsigned int purged = 0;
|
|
|
|
struct rb_node *node;
|
|
|
|
struct ocfs2_meta_cache_item *item;
|
|
|
|
|
|
|
|
while ((node = rb_last(root)) != NULL) {
|
|
|
|
item = rb_entry(node, struct ocfs2_meta_cache_item, c_node);
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_purge_copied_metadata_tree(
|
|
|
|
(unsigned long long) item->c_block);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
rb_erase(&item->c_node, root);
|
|
|
|
kmem_cache_free(ocfs2_uptodate_cachep, item);
|
|
|
|
|
|
|
|
purged++;
|
|
|
|
}
|
|
|
|
return purged;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Called from locking and called from ocfs2_clear_inode. Dump the
|
|
|
|
* cache for a given inode.
|
|
|
|
*
|
|
|
|
* This function is a few more lines longer than necessary due to some
|
|
|
|
* accounting done here, but I think it's worth tracking down those
|
|
|
|
* bugs sooner -- Mark */
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci)
|
2005-12-15 22:31:24 +00:00
|
|
|
{
|
|
|
|
unsigned int tree, to_purge, purged;
|
|
|
|
struct rb_root root = RB_ROOT;
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
BUG_ON(!ci || !ci->ci_ops);
|
|
|
|
|
|
|
|
ocfs2_metadata_cache_lock(ci);
|
2009-02-11 00:05:07 +00:00
|
|
|
tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);
|
2005-12-15 22:31:24 +00:00
|
|
|
to_purge = ci->ci_num_cached;
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_metadata_cache_purge(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
to_purge, tree);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
/* If we're a tree, save off the root so that we can safely
|
|
|
|
* initialize the cache. We do the work to free tree members
|
|
|
|
* without the spinlock. */
|
|
|
|
if (tree)
|
|
|
|
root = ci->ci_cache.ci_tree;
|
|
|
|
|
2009-02-12 23:24:40 +00:00
|
|
|
ocfs2_metadata_cache_reset(ci, 0);
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
purged = ocfs2_purge_copied_metadata_tree(&root);
|
|
|
|
/* If possible, track the number wiped so that we can more
|
|
|
|
* easily detect counting errors. Unfortunately, this is only
|
|
|
|
* meaningful for trees. */
|
|
|
|
if (tree && purged != to_purge)
|
2009-02-11 03:00:37 +00:00
|
|
|
mlog(ML_ERROR, "Owner %llu, count = %u, purged = %u\n",
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
to_purge, purged);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Returns the index in the cache array, -1 if not found.
|
|
|
|
* Requires ip_lock. */
|
|
|
|
static int ocfs2_search_cache_array(struct ocfs2_caching_info *ci,
|
|
|
|
sector_t item)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < ci->ci_num_cached; i++) {
|
|
|
|
if (item == ci->ci_cache.ci_array[i])
|
|
|
|
return i;
|
|
|
|
}
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Returns the cache item if found, otherwise NULL.
|
|
|
|
* Requires ip_lock. */
|
|
|
|
static struct ocfs2_meta_cache_item *
|
|
|
|
ocfs2_search_cache_tree(struct ocfs2_caching_info *ci,
|
|
|
|
sector_t block)
|
|
|
|
{
|
|
|
|
struct rb_node * n = ci->ci_cache.ci_tree.rb_node;
|
|
|
|
struct ocfs2_meta_cache_item *item = NULL;
|
|
|
|
|
|
|
|
while (n) {
|
|
|
|
item = rb_entry(n, struct ocfs2_meta_cache_item, c_node);
|
|
|
|
|
|
|
|
if (block < item->c_block)
|
|
|
|
n = n->rb_left;
|
|
|
|
else if (block > item->c_block)
|
|
|
|
n = n->rb_right;
|
|
|
|
else
|
|
|
|
return item;
|
|
|
|
}
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
int index = -1;
|
|
|
|
struct ocfs2_meta_cache_item *item = NULL;
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_lock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_buffer_cached_begin(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long) bh->b_blocknr,
|
|
|
|
!!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
if (ci->ci_flags & OCFS2_CACHE_FL_INLINE)
|
2009-02-11 04:00:41 +00:00
|
|
|
index = ocfs2_search_cache_array(ci, bh->b_blocknr);
|
2005-12-15 22:31:24 +00:00
|
|
|
else
|
2009-02-11 04:00:41 +00:00
|
|
|
item = ocfs2_search_cache_tree(ci, bh->b_blocknr);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_buffer_cached_end(index, item);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
return (index != -1) || (item != NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Warning: even if it returns true, this does *not* guarantee that
|
2010-01-26 00:57:38 +00:00
|
|
|
* the block is stored in our inode metadata cache.
|
|
|
|
*
|
2006-04-21 20:49:02 +00:00
|
|
|
* This can be called under lock_buffer()
|
|
|
|
*/
|
2009-02-11 04:00:41 +00:00
|
|
|
int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
/* Doesn't matter if the bh is in our cache or not -- if it's
|
|
|
|
* not marked uptodate then we know it can't have correct
|
|
|
|
* data. */
|
|
|
|
if (!buffer_uptodate(bh))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* OCFS2 does not allow multiple nodes to be changing the same
|
|
|
|
* block at the same time. */
|
|
|
|
if (buffer_jbd(bh))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
/* Ok, locally the buffer is marked as up to date, now search
|
|
|
|
* our cache to see if we can trust that. */
|
2009-02-11 04:00:41 +00:00
|
|
|
return ocfs2_buffer_cached(ci, bh);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
/*
|
2006-04-21 20:49:02 +00:00
|
|
|
* Determine whether a buffer is currently out on a read-ahead request.
|
2009-02-11 00:05:07 +00:00
|
|
|
* ci_io_sem should be held to serialize submitters with the logic here.
|
2006-04-21 20:49:02 +00:00
|
|
|
*/
|
2009-02-11 04:00:41 +00:00
|
|
|
int ocfs2_buffer_read_ahead(struct ocfs2_caching_info *ci,
|
2006-04-21 20:49:02 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
2009-02-11 04:00:41 +00:00
|
|
|
return buffer_locked(bh) && ocfs2_buffer_cached(ci, bh);
|
2006-04-21 20:49:02 +00:00
|
|
|
}
|
|
|
|
|
2005-12-15 22:31:24 +00:00
|
|
|
/* Requires ip_lock */
|
|
|
|
static void ocfs2_append_cache_array(struct ocfs2_caching_info *ci,
|
|
|
|
sector_t block)
|
|
|
|
{
|
2009-02-11 00:05:07 +00:00
|
|
|
BUG_ON(ci->ci_num_cached >= OCFS2_CACHE_INFO_MAX_ARRAY);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_append_cache_array(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long)block, ci->ci_num_cached);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
ci->ci_cache.ci_array[ci->ci_num_cached] = block;
|
|
|
|
ci->ci_num_cached++;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* By now the caller should have checked that the item does *not*
|
|
|
|
* exist in the tree.
|
|
|
|
* Requires ip_lock. */
|
|
|
|
static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci,
|
|
|
|
struct ocfs2_meta_cache_item *new)
|
|
|
|
{
|
|
|
|
sector_t block = new->c_block;
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
struct rb_node **p = &ci->ci_cache.ci_tree.rb_node;
|
|
|
|
struct ocfs2_meta_cache_item *tmp;
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_insert_cache_tree(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long)block, ci->ci_num_cached);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
while(*p) {
|
|
|
|
parent = *p;
|
|
|
|
|
|
|
|
tmp = rb_entry(parent, struct ocfs2_meta_cache_item, c_node);
|
|
|
|
|
|
|
|
if (block < tmp->c_block)
|
|
|
|
p = &(*p)->rb_left;
|
|
|
|
else if (block > tmp->c_block)
|
|
|
|
p = &(*p)->rb_right;
|
|
|
|
else {
|
|
|
|
/* This should never happen! */
|
|
|
|
mlog(ML_ERROR, "Duplicate block %llu cached!\n",
|
|
|
|
(unsigned long long) block);
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(&new->c_node, parent, p);
|
|
|
|
rb_insert_color(&new->c_node, &ci->ci_cache.ci_tree);
|
|
|
|
ci->ci_num_cached++;
|
|
|
|
}
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
/* co_cache_lock() must be held */
|
2009-02-11 04:00:41 +00:00
|
|
|
static inline int ocfs2_insert_can_use_array(struct ocfs2_caching_info *ci)
|
2005-12-15 22:31:24 +00:00
|
|
|
{
|
2009-02-11 00:05:07 +00:00
|
|
|
return (ci->ci_flags & OCFS2_CACHE_FL_INLINE) &&
|
|
|
|
(ci->ci_num_cached < OCFS2_CACHE_INFO_MAX_ARRAY);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
/* tree should be exactly OCFS2_CACHE_INFO_MAX_ARRAY wide. NULL the
|
2005-12-15 22:31:24 +00:00
|
|
|
* pointers in tree after we use them - this allows caller to detect
|
2009-02-11 03:00:37 +00:00
|
|
|
* when to free in case of error.
|
|
|
|
*
|
|
|
|
* The co_cache_lock() must be held. */
|
2009-02-11 04:00:41 +00:00
|
|
|
static void ocfs2_expand_cache(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
struct ocfs2_meta_cache_item **tree)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
mlog_bug_on_msg(ci->ci_num_cached != OCFS2_CACHE_INFO_MAX_ARRAY,
|
2009-02-11 03:00:37 +00:00
|
|
|
"Owner %llu, num cached = %u, should be %u\n",
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
ci->ci_num_cached, OCFS2_CACHE_INFO_MAX_ARRAY);
|
2009-02-11 00:05:07 +00:00
|
|
|
mlog_bug_on_msg(!(ci->ci_flags & OCFS2_CACHE_FL_INLINE),
|
2009-02-11 03:00:37 +00:00
|
|
|
"Owner %llu not marked as inline anymore!\n",
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci));
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
/* Be careful to initialize the tree members *first* because
|
|
|
|
* once the ci_tree is used, the array is junk... */
|
2009-02-11 00:05:07 +00:00
|
|
|
for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
|
2005-12-15 22:31:24 +00:00
|
|
|
tree[i]->c_block = ci->ci_cache.ci_array[i];
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
ci->ci_flags &= ~OCFS2_CACHE_FL_INLINE;
|
2005-12-15 22:31:24 +00:00
|
|
|
ci->ci_cache.ci_tree = RB_ROOT;
|
|
|
|
/* this will be set again by __ocfs2_insert_cache_tree */
|
|
|
|
ci->ci_num_cached = 0;
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
|
2005-12-15 22:31:24 +00:00
|
|
|
__ocfs2_insert_cache_tree(ci, tree[i]);
|
|
|
|
tree[i] = NULL;
|
|
|
|
}
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_expand_cache(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
ci->ci_flags, ci->ci_num_cached);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Slow path function - memory allocation is necessary. See the
|
|
|
|
* comment above ocfs2_set_buffer_uptodate for more information. */
|
2009-02-11 04:00:41 +00:00
|
|
|
static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
sector_t block,
|
|
|
|
int expand_tree)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
struct ocfs2_meta_cache_item *new = NULL;
|
2009-02-11 00:05:07 +00:00
|
|
|
struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =
|
2005-12-15 22:31:24 +00:00
|
|
|
{ NULL, };
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_set_buffer_uptodate(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long)block, expand_tree);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2006-04-12 21:37:00 +00:00
|
|
|
new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
|
2005-12-15 22:31:24 +00:00
|
|
|
if (!new) {
|
|
|
|
mlog_errno(-ENOMEM);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
new->c_block = block;
|
|
|
|
|
|
|
|
if (expand_tree) {
|
|
|
|
/* Do *not* allocate an array here - the removal code
|
|
|
|
* has no way of tracking that. */
|
2009-02-11 00:05:07 +00:00
|
|
|
for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
|
2005-12-15 22:31:24 +00:00
|
|
|
tree[i] = kmem_cache_alloc(ocfs2_uptodate_cachep,
|
2006-04-12 21:37:00 +00:00
|
|
|
GFP_NOFS);
|
2005-12-15 22:31:24 +00:00
|
|
|
if (!tree[i]) {
|
|
|
|
mlog_errno(-ENOMEM);
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* These are initialized in ocfs2_expand_cache! */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_lock(ci);
|
2009-02-11 04:00:41 +00:00
|
|
|
if (ocfs2_insert_can_use_array(ci)) {
|
2005-12-15 22:31:24 +00:00
|
|
|
/* Ok, items were removed from the cache in between
|
|
|
|
* locks. Detect this and revert back to the fast path */
|
|
|
|
ocfs2_append_cache_array(ci, block);
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (expand_tree)
|
2009-02-11 04:00:41 +00:00
|
|
|
ocfs2_expand_cache(ci, tree);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
__ocfs2_insert_cache_tree(ci, new);
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
new = NULL;
|
|
|
|
out_free:
|
|
|
|
if (new)
|
|
|
|
kmem_cache_free(ocfs2_uptodate_cachep, new);
|
|
|
|
|
|
|
|
/* If these were used, then ocfs2_expand_cache re-set them to
|
|
|
|
* NULL for us. */
|
|
|
|
if (tree[0]) {
|
2009-02-11 00:05:07 +00:00
|
|
|
for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
|
2005-12-15 22:31:24 +00:00
|
|
|
if (tree[i])
|
|
|
|
kmem_cache_free(ocfs2_uptodate_cachep,
|
|
|
|
tree[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
/* Item insertion is guarded by co_io_lock(), so the insertion path takes
|
2005-12-15 22:31:24 +00:00
|
|
|
* advantage of this by not rechecking for a duplicate insert during
|
|
|
|
* the slow case. Additionally, if the cache needs to be bumped up to
|
|
|
|
* a tree, the code will not recheck after acquiring the lock --
|
|
|
|
* multiple paths cannot be expanding to a tree at the same time.
|
|
|
|
*
|
|
|
|
* The slow path takes into account that items can be removed
|
|
|
|
* (including the whole tree wiped and reset) when this process it out
|
|
|
|
* allocating memory. In those cases, it reverts back to the fast
|
|
|
|
* path.
|
|
|
|
*
|
|
|
|
* Note that this function may actually fail to insert the block if
|
|
|
|
* memory cannot be allocated. This is not fatal however (but may
|
2006-04-21 20:49:02 +00:00
|
|
|
* result in a performance penalty)
|
|
|
|
*
|
|
|
|
* Readahead buffers can be passed in here before the I/O request is
|
|
|
|
* completed.
|
|
|
|
*/
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
int expand;
|
|
|
|
|
|
|
|
/* The block may very well exist in our cache already, so avoid
|
|
|
|
* doing any more work in that case. */
|
2009-02-11 04:00:41 +00:00
|
|
|
if (ocfs2_buffer_cached(ci, bh))
|
2005-12-15 22:31:24 +00:00
|
|
|
return;
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_set_buffer_uptodate_begin(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long)bh->b_blocknr);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
/* No need to recheck under spinlock - insertion is guarded by
|
2009-02-11 03:00:37 +00:00
|
|
|
* co_io_lock() */
|
|
|
|
ocfs2_metadata_cache_lock(ci);
|
2009-02-11 04:00:41 +00:00
|
|
|
if (ocfs2_insert_can_use_array(ci)) {
|
2005-12-15 22:31:24 +00:00
|
|
|
/* Fast case - it's an array and there's a free
|
|
|
|
* spot. */
|
|
|
|
ocfs2_append_cache_array(ci, bh->b_blocknr);
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
expand = 0;
|
2009-02-11 00:05:07 +00:00
|
|
|
if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
|
2005-12-15 22:31:24 +00:00
|
|
|
/* We need to bump things up to a tree. */
|
|
|
|
expand = 1;
|
|
|
|
}
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
__ocfs2_set_buffer_uptodate(ci, bh->b_blocknr, expand);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Called against a newly allocated buffer. Most likely nobody should
|
|
|
|
* be able to read this sort of metadata while it's still being
|
2009-02-11 03:00:37 +00:00
|
|
|
* allocated, but this is careful to take co_io_lock() anyway. */
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_set_new_buffer_uptodate(struct ocfs2_caching_info *ci,
|
2005-12-15 22:31:24 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
/* This should definitely *not* exist in our cache */
|
2009-02-11 04:00:41 +00:00
|
|
|
BUG_ON(ocfs2_buffer_cached(ci, bh));
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
set_buffer_uptodate(bh);
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_io_lock(ci);
|
2009-02-11 04:00:41 +00:00
|
|
|
ocfs2_set_buffer_uptodate(ci, bh);
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_io_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Requires ip_lock. */
|
|
|
|
static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
|
|
|
|
int index)
|
|
|
|
{
|
|
|
|
sector_t *array = ci->ci_cache.ci_array;
|
|
|
|
int bytes;
|
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
BUG_ON(index < 0 || index >= OCFS2_CACHE_INFO_MAX_ARRAY);
|
2005-12-15 22:31:24 +00:00
|
|
|
BUG_ON(index >= ci->ci_num_cached);
|
|
|
|
BUG_ON(!ci->ci_num_cached);
|
|
|
|
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_remove_metadata_array(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
index, ci->ci_num_cached);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
ci->ci_num_cached--;
|
|
|
|
|
|
|
|
/* don't need to copy if the array is now empty, or if we
|
|
|
|
* removed at the tail */
|
|
|
|
if (ci->ci_num_cached && index < ci->ci_num_cached) {
|
|
|
|
bytes = sizeof(sector_t) * (ci->ci_num_cached - index);
|
|
|
|
memmove(&array[index], &array[index + 1], bytes);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Requires ip_lock. */
|
|
|
|
static void ocfs2_remove_metadata_tree(struct ocfs2_caching_info *ci,
|
|
|
|
struct ocfs2_meta_cache_item *item)
|
|
|
|
{
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_remove_metadata_tree(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long)item->c_block);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
rb_erase(&item->c_node, &ci->ci_cache.ci_tree);
|
|
|
|
ci->ci_num_cached--;
|
|
|
|
}
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
static void ocfs2_remove_block_from_cache(struct ocfs2_caching_info *ci,
|
2008-08-18 09:38:47 +00:00
|
|
|
sector_t block)
|
2005-12-15 22:31:24 +00:00
|
|
|
{
|
|
|
|
int index;
|
|
|
|
struct ocfs2_meta_cache_item *item = NULL;
|
|
|
|
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_lock(ci);
|
2011-02-24 08:22:20 +00:00
|
|
|
trace_ocfs2_remove_block_from_cache(
|
|
|
|
(unsigned long long)ocfs2_metadata_cache_owner(ci),
|
|
|
|
(unsigned long long) block, ci->ci_num_cached,
|
|
|
|
ci->ci_flags);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
2009-02-11 00:05:07 +00:00
|
|
|
if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
|
2005-12-15 22:31:24 +00:00
|
|
|
index = ocfs2_search_cache_array(ci, block);
|
|
|
|
if (index != -1)
|
|
|
|
ocfs2_remove_metadata_array(ci, index);
|
|
|
|
} else {
|
|
|
|
item = ocfs2_search_cache_tree(ci, block);
|
|
|
|
if (item)
|
|
|
|
ocfs2_remove_metadata_tree(ci, item);
|
|
|
|
}
|
2009-02-11 03:00:37 +00:00
|
|
|
ocfs2_metadata_cache_unlock(ci);
|
2005-12-15 22:31:24 +00:00
|
|
|
|
|
|
|
if (item)
|
|
|
|
kmem_cache_free(ocfs2_uptodate_cachep, item);
|
|
|
|
}
|
|
|
|
|
2008-08-18 09:38:47 +00:00
|
|
|
/*
|
|
|
|
* Called when we remove a chunk of metadata from an inode. We don't
|
|
|
|
* bother reverting things to an inlined array in the case of a remove
|
|
|
|
* which moves us back under the limit.
|
|
|
|
*/
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_remove_from_cache(struct ocfs2_caching_info *ci,
|
2008-08-18 09:38:47 +00:00
|
|
|
struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
sector_t block = bh->b_blocknr;
|
|
|
|
|
2009-02-11 04:00:41 +00:00
|
|
|
ocfs2_remove_block_from_cache(ci, block);
|
2008-08-18 09:38:47 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Called when we remove xattr clusters from an inode. */
|
2009-02-11 04:00:41 +00:00
|
|
|
void ocfs2_remove_xattr_clusters_from_cache(struct ocfs2_caching_info *ci,
|
2008-08-18 09:38:47 +00:00
|
|
|
sector_t block,
|
|
|
|
u32 c_len)
|
|
|
|
{
|
2009-02-11 04:00:41 +00:00
|
|
|
struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
|
|
|
|
unsigned int i, b_len = ocfs2_clusters_to_blocks(sb, 1) * c_len;
|
2008-08-18 09:38:47 +00:00
|
|
|
|
|
|
|
for (i = 0; i < b_len; i++, block++)
|
2009-02-11 04:00:41 +00:00
|
|
|
ocfs2_remove_block_from_cache(ci, block);
|
2008-08-18 09:38:47 +00:00
|
|
|
}
|
|
|
|
|
2005-12-15 22:31:24 +00:00
|
|
|
int __init init_ocfs2_uptodate_cache(void)
|
|
|
|
{
|
|
|
|
ocfs2_uptodate_cachep = kmem_cache_create("ocfs2_uptodate",
|
|
|
|
sizeof(struct ocfs2_meta_cache_item),
|
2007-07-20 01:11:58 +00:00
|
|
|
0, SLAB_HWCACHE_ALIGN, NULL);
|
2005-12-15 22:31:24 +00:00
|
|
|
if (!ocfs2_uptodate_cachep)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-01-07 19:07:02 +00:00
|
|
|
void exit_ocfs2_uptodate_cache(void)
|
2005-12-15 22:31:24 +00:00
|
|
|
{
|
|
|
|
if (ocfs2_uptodate_cachep)
|
|
|
|
kmem_cache_destroy(ocfs2_uptodate_cachep);
|
|
|
|
}
|