Commit 6e5a3d75 authored by Joel Becker's avatar Joel Becker
Browse files

ocfs2: Change metadata caching locks to an operations structure. 



We don't really want to cart around too many new fields on the
ocfs2_caching_info structure.  So let's wrap all our access of the
parent object in a set of operations.  One pointer on caching_info, and
more flexibility to boot.

Signed-off-by: default avatarJoel Becker <joel.becker@oracle.com>
parent 47460d65

fs/ocfs2/inode.c

+49 −0
Original line number Diff line number Diff line
@@ -1395,3 +1395,52 @@ int ocfs2_read_inode_block(struct inode *inode, struct buffer_head **bh) 
{

	return ocfs2_read_inode_block_full(inode, bh, 0);

}



static struct ocfs2_inode_info *cache_info_to_inode(struct ocfs2_caching_info *ci)

{

	return container_of(ci, struct ocfs2_inode_info, ip_metadata_cache);

}



static u64 ocfs2_inode_cache_owner(struct ocfs2_caching_info *ci)

{

	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);



	return oi->ip_blkno;

}



static void ocfs2_inode_cache_lock(struct ocfs2_caching_info *ci)

{

	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);



	spin_lock(&oi->ip_lock);

}



static void ocfs2_inode_cache_unlock(struct ocfs2_caching_info *ci)

{

	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);



	spin_unlock(&oi->ip_lock);

}



static void ocfs2_inode_cache_io_lock(struct ocfs2_caching_info *ci)

{

	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);



	mutex_lock(&oi->ip_io_mutex);

}



static void ocfs2_inode_cache_io_unlock(struct ocfs2_caching_info *ci)

{

	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);



	mutex_unlock(&oi->ip_io_mutex);

}



const struct ocfs2_caching_operations ocfs2_inode_caching_ops = {

	.co_owner		= ocfs2_inode_cache_owner,

	.co_cache_lock		= ocfs2_inode_cache_lock,

	.co_cache_unlock	= ocfs2_inode_cache_unlock,

	.co_io_lock		= ocfs2_inode_cache_io_lock,

	.co_io_unlock		= ocfs2_inode_cache_io_unlock,

};

fs/ocfs2/inode.h

+1 −0
Original line number Diff line number Diff line
@@ -118,6 +118,7 @@ static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode) 
extern struct kmem_cache *ocfs2_inode_cache;



extern const struct address_space_operations ocfs2_aops;

extern const struct ocfs2_caching_operations ocfs2_inode_caching_ops;



void ocfs2_clear_inode(struct inode *inode);

void ocfs2_delete_inode(struct inode *inode);

fs/ocfs2/ocfs2.h

+4 −4
Original line number Diff line number Diff line
@@ -68,14 +68,14 @@ enum ocfs2_caching_info_flags { 
	OCFS2_CACHE_FL_INLINE	= 1<<1,

};



struct ocfs2_caching_operations;

struct ocfs2_caching_info {

	/*

	 * The parent structure provides the locks, but because the

	 * parent structure can differ, struct ocfs2_caching_info needs

	 * its own pointers to them.

	 * parent structure can differ, it provides locking operations

	 * to struct ocfs2_caching_info.

	 */

	spinlock_t		*ci_lock;

	struct mutex		*ci_io_mutex;

	const struct ocfs2_caching_operations *ci_ops;



	unsigned int		ci_flags;

	unsigned int		ci_num_cached;

fs/ocfs2/super.c

+2 −2
Original line number Diff line number Diff line
@@ -1683,8 +1683,8 @@ static void ocfs2_inode_init_once(void *data) 
	ocfs2_lock_res_init_once(&oi->ip_inode_lockres);

	ocfs2_lock_res_init_once(&oi->ip_open_lockres);



	ocfs2_metadata_cache_init(&oi->ip_metadata_cache, &oi->ip_lock,

				  &oi->ip_io_mutex);

	ocfs2_metadata_cache_init(&oi->ip_metadata_cache,

				  &ocfs2_inode_caching_ops);



	inode_init_once(&oi->vfs_inode);

}

fs/ocfs2/uptodate.c

+85 −44
Original line number Diff line number Diff line
@@ -75,12 +75,48 @@ struct ocfs2_meta_cache_item { 


static struct kmem_cache *ocfs2_uptodate_cachep = NULL;



static u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci)

{

	BUG_ON(!ci || !ci->ci_ops);



	return ci->ci_ops->co_owner(ci);

}



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);

}



static void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci)

{

	BUG_ON(!ci || !ci->ci_ops);



	ci->ci_ops->co_io_lock(ci);

}



static void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci)

{

	BUG_ON(!ci || !ci->ci_ops);



	ci->ci_ops->co_io_unlock(ci);

}





void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,

			       spinlock_t *cache_lock,

			       struct mutex *io_mutex)

			       const struct ocfs2_caching_operations *ops)

{

	ci->ci_lock = cache_lock;

	ci->ci_io_mutex = io_mutex;

	BUG_ON(!ops);



	ci->ci_ops = ops;

	ci->ci_flags |= OCFS2_CACHE_FL_INLINE;

	ci->ci_num_cached = 0;

}
@@ -120,12 +156,15 @@ void ocfs2_metadata_cache_purge(struct inode *inode) 
	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;

	struct rb_root root = RB_ROOT;



	spin_lock(ci->ci_lock);

	BUG_ON(!ci || !ci->ci_ops);



	ocfs2_metadata_cache_lock(ci);

	tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);

	to_purge = ci->ci_num_cached;



	mlog(0, "Purge %u %s items from Inode %llu\n", to_purge,

	     tree ? "array" : "tree", (unsigned long long)oi->ip_blkno);

	mlog(0, "Purge %u %s items from Owner %llu\n", to_purge,

	     tree ? "array" : "tree",

	     (unsigned long long)ocfs2_metadata_cache_owner(ci));



	/* 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
@@ -133,16 +172,17 @@ void ocfs2_metadata_cache_purge(struct inode *inode) 
	if (tree)

		root = ci->ci_cache.ci_tree;



	ocfs2_metadata_cache_init(ci, ci->ci_lock, ci->ci_io_mutex);

	spin_unlock(ci->ci_lock);

	ocfs2_metadata_cache_init(ci, ci->ci_ops);

	ocfs2_metadata_cache_unlock(ci);



	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)

		mlog(ML_ERROR, "Inode %llu, count = %u, purged = %u\n",

		     (unsigned long long)oi->ip_blkno, to_purge, purged);

		mlog(ML_ERROR, "Owner %llu, count = %u, purged = %u\n",

		     (unsigned long long)ocfs2_metadata_cache_owner(ci),

		     to_purge, purged);

}



/* Returns the index in the cache array, -1 if not found.
@@ -190,10 +230,10 @@ static int ocfs2_buffer_cached(struct ocfs2_inode_info *oi, 
	struct ocfs2_meta_cache_item *item = NULL;

	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;



	spin_lock(ci->ci_lock);

	ocfs2_metadata_cache_lock(ci);



	mlog(0, "Inode %llu, query block %llu (inline = %u)\n",

	     (unsigned long long)oi->ip_blkno,

	mlog(0, "Owner %llu, query block %llu (inline = %u)\n",

	     (unsigned long long)ocfs2_metadata_cache_owner(ci),

	     (unsigned long long) bh->b_blocknr,

	     !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
@@ -204,7 +244,7 @@ static int ocfs2_buffer_cached(struct ocfs2_inode_info *oi, 
		item = ocfs2_search_cache_tree(&oi->ip_metadata_cache,

					       bh->b_blocknr);



	spin_unlock(ci->ci_lock);

	ocfs2_metadata_cache_unlock(ci);



	mlog(0, "index = %d, item = %p\n", index, item);
@@ -294,18 +334,19 @@ static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci, 
	ci->ci_num_cached++;

}



/* co_cache_lock() must be held */

static inline int ocfs2_insert_can_use_array(struct ocfs2_inode_info *oi,

					     struct ocfs2_caching_info *ci)

{

	assert_spin_locked(ci->ci_lock);



	return (ci->ci_flags & OCFS2_CACHE_FL_INLINE) &&

		(ci->ci_num_cached < OCFS2_CACHE_INFO_MAX_ARRAY);

}



/* tree should be exactly OCFS2_CACHE_INFO_MAX_ARRAY wide. NULL the

 * pointers in tree after we use them - this allows caller to detect

 * when to free in case of error. */

 * when to free in case of error.

 *

 * The co_cache_lock() must be held. */

static void ocfs2_expand_cache(struct ocfs2_inode_info *oi,

			       struct ocfs2_meta_cache_item **tree)

{
@@ -313,13 +354,12 @@ static void ocfs2_expand_cache(struct ocfs2_inode_info *oi, 
	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;



	mlog_bug_on_msg(ci->ci_num_cached != OCFS2_CACHE_INFO_MAX_ARRAY,

			"Inode %llu, num cached = %u, should be %u\n",

			(unsigned long long)oi->ip_blkno, ci->ci_num_cached,

			OCFS2_CACHE_INFO_MAX_ARRAY);

			"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);

	mlog_bug_on_msg(!(ci->ci_flags & OCFS2_CACHE_FL_INLINE),

			"Inode %llu not marked as inline anymore!\n",

			(unsigned long long)oi->ip_blkno);

	assert_spin_locked(ci->ci_lock);

			"Owner %llu not marked as inline anymore!\n",

			(unsigned long long)ocfs2_metadata_cache_owner(ci));



	/* Be careful to initialize the tree members *first* because

	 * once the ci_tree is used, the array is junk... */
@@ -337,7 +377,8 @@ static void ocfs2_expand_cache(struct ocfs2_inode_info *oi, 
	}



	mlog(0, "Expanded %llu to a tree cache: flags 0x%x, num = %u\n",

	     (unsigned long long)oi->ip_blkno, ci->ci_flags, ci->ci_num_cached);

	     (unsigned long long)ocfs2_metadata_cache_owner(ci),

	     ci->ci_flags, ci->ci_num_cached);

}



/* Slow path function - memory allocation is necessary. See the
@@ -352,8 +393,8 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi, 
	struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =

		{ NULL, };



	mlog(0, "Inode %llu, block %llu, expand = %d\n",

	     (unsigned long long)oi->ip_blkno,

	mlog(0, "Owner %llu, block %llu, expand = %d\n",

	     (unsigned long long)ocfs2_metadata_cache_owner(ci),

	     (unsigned long long)block, expand_tree);



	new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
@@ -378,13 +419,13 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi, 
		}

	}



	spin_lock(ci->ci_lock);

	ocfs2_metadata_cache_lock(ci);

	if (ocfs2_insert_can_use_array(oi, ci)) {

		mlog(0, "Someone cleared the tree underneath us\n");

		/* 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);

		spin_unlock(ci->ci_lock);

		ocfs2_metadata_cache_unlock(ci);

		goto out_free;

	}
@@ -392,7 +433,7 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi, 
		ocfs2_expand_cache(oi, tree);



	__ocfs2_insert_cache_tree(ci, new);

	spin_unlock(ci->ci_lock);

	ocfs2_metadata_cache_unlock(ci);



	new = NULL;

out_free:
@@ -409,7 +450,7 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi, 
	}

}



/* Item insertion is guarded by ci_io_mutex, so the insertion path takes

/* Item insertion is guarded by co_io_lock(), so the insertion path takes

 * 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 --
@@ -439,18 +480,18 @@ void ocfs2_set_buffer_uptodate(struct inode *inode, 
	if (ocfs2_buffer_cached(oi, bh))

		return;



	mlog(0, "Inode %llu, inserting block %llu\n",

	     (unsigned long long)oi->ip_blkno,

	mlog(0, "Owner %llu, inserting block %llu\n",

	     (unsigned long long)ocfs2_metadata_cache_owner(ci),

	     (unsigned long long)bh->b_blocknr);



	/* No need to recheck under spinlock - insertion is guarded by

	 * ci_io_mutex */

	spin_lock(ci->ci_lock);

	 * co_io_lock() */

	ocfs2_metadata_cache_lock(ci);

	if (ocfs2_insert_can_use_array(oi, ci)) {

		/* Fast case - it's an array and there's a free

		 * spot. */

		ocfs2_append_cache_array(ci, bh->b_blocknr);

		spin_unlock(ci->ci_lock);

		ocfs2_metadata_cache_unlock(ci);

		return;

	}
@@ -459,14 +500,14 @@ void ocfs2_set_buffer_uptodate(struct inode *inode, 
		/* We need to bump things up to a tree. */

		expand = 1;

	}

	spin_unlock(ci->ci_lock);

	ocfs2_metadata_cache_unlock(ci);



	__ocfs2_set_buffer_uptodate(oi, bh->b_blocknr, expand);

}



/* Called against a newly allocated buffer. Most likely nobody should

 * be able to read this sort of metadata while it's still being

 * allocated, but this is careful to take ci_io_mutex anyway. */

 * allocated, but this is careful to take co_io_lock() anyway. */

void ocfs2_set_new_buffer_uptodate(struct inode *inode,

				   struct buffer_head *bh)

{
@@ -478,9 +519,9 @@ void ocfs2_set_new_buffer_uptodate(struct inode *inode, 


	set_buffer_uptodate(bh);



	mutex_lock(ci->ci_io_mutex);

	ocfs2_metadata_cache_io_lock(ci);

	ocfs2_set_buffer_uptodate(inode, bh);

	mutex_unlock(ci->ci_io_mutex);

	ocfs2_metadata_cache_io_unlock(ci);

}



/* Requires ip_lock. */
@@ -526,9 +567,9 @@ static void ocfs2_remove_block_from_cache(struct inode *inode, 
	struct ocfs2_inode_info *oi = OCFS2_I(inode);

	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;



	spin_lock(ci->ci_lock);

	mlog(0, "Inode %llu, remove %llu, items = %u, array = %u\n",

	     (unsigned long long)oi->ip_blkno,

	ocfs2_metadata_cache_lock(ci);

	mlog(0, "Owner %llu, remove %llu, items = %u, array = %u\n",

	     (unsigned long long)ocfs2_metadata_cache_owner(ci),

	     (unsigned long long) block, ci->ci_num_cached,

	     ci->ci_flags & OCFS2_CACHE_FL_INLINE);
@@ -541,7 +582,7 @@ static void ocfs2_remove_block_from_cache(struct inode *inode, 
		if (item)

			ocfs2_remove_metadata_tree(ci, item);

	}

	spin_unlock(ci->ci_lock);

	ocfs2_metadata_cache_unlock(ci);



	if (item)

		kmem_cache_free(ocfs2_uptodate_cachep, item);