Btrfs: Make the code for reading/writing free space cache generic
authorLi Zefan <lizf@cn.fujitsu.com>
Wed, 20 Apr 2011 02:20:14 +0000 (10:20 +0800)
committerLi Zefan <lizf@cn.fujitsu.com>
Mon, 25 Apr 2011 08:46:07 +0000 (16:46 +0800)
Extract out block group specific code from lookup_free_space_inode(),
create_free_space_inode(), load_free_space_cache() and
btrfs_write_out_cache(), so the code can be used to read/write
free ino cache.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
fs/btrfs/free-space-cache.c

index 2ce89bfc881592a9e846860ea83d476c6d5efe93..fcbdcef6ca28386389ccf59e61e3a979db4b3d8e 100644 (file)
@@ -33,9 +33,9 @@
 static int link_free_space(struct btrfs_free_space_ctl *ctl,
                           struct btrfs_free_space *info);
 
-struct inode *lookup_free_space_inode(struct btrfs_root *root,
-                                     struct btrfs_block_group_cache
-                                     *block_group, struct btrfs_path *path)
+static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
+                                              struct btrfs_path *path,
+                                              u64 offset)
 {
        struct btrfs_key key;
        struct btrfs_key location;
@@ -45,15 +45,8 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
        struct inode *inode = NULL;
        int ret;
 
-       spin_lock(&block_group->lock);
-       if (block_group->inode)
-               inode = igrab(block_group->inode);
-       spin_unlock(&block_group->lock);
-       if (inode)
-               return inode;
-
        key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-       key.offset = block_group->key.objectid;
+       key.offset = offset;
        key.type = 0;
 
        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
@@ -83,6 +76,27 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
 
        inode->i_mapping->flags &= ~__GFP_FS;
 
+       return inode;
+}
+
+struct inode *lookup_free_space_inode(struct btrfs_root *root,
+                                     struct btrfs_block_group_cache
+                                     *block_group, struct btrfs_path *path)
+{
+       struct inode *inode = NULL;
+
+       spin_lock(&block_group->lock);
+       if (block_group->inode)
+               inode = igrab(block_group->inode);
+       spin_unlock(&block_group->lock);
+       if (inode)
+               return inode;
+
+       inode = __lookup_free_space_inode(root, path,
+                                         block_group->key.objectid);
+       if (IS_ERR(inode))
+               return inode;
+
        spin_lock(&block_group->lock);
        if (!root->fs_info->closing) {
                block_group->inode = igrab(inode);
@@ -93,24 +107,18 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
        return inode;
 }
 
-int create_free_space_inode(struct btrfs_root *root,
-                           struct btrfs_trans_handle *trans,
-                           struct btrfs_block_group_cache *block_group,
-                           struct btrfs_path *path)
+int __create_free_space_inode(struct btrfs_root *root,
+                             struct btrfs_trans_handle *trans,
+                             struct btrfs_path *path, u64 ino, u64 offset)
 {
        struct btrfs_key key;
        struct btrfs_disk_key disk_key;
        struct btrfs_free_space_header *header;
        struct btrfs_inode_item *inode_item;
        struct extent_buffer *leaf;
-       u64 objectid;
        int ret;
 
-       ret = btrfs_find_free_objectid(root, &objectid);
-       if (ret < 0)
-               return ret;
-
-       ret = btrfs_insert_empty_inode(trans, root, path, objectid);
+       ret = btrfs_insert_empty_inode(trans, root, path, ino);
        if (ret)
                return ret;
 
@@ -130,13 +138,12 @@ int create_free_space_inode(struct btrfs_root *root,
                              BTRFS_INODE_PREALLOC | BTRFS_INODE_NODATASUM);
        btrfs_set_inode_nlink(leaf, inode_item, 1);
        btrfs_set_inode_transid(leaf, inode_item, trans->transid);
-       btrfs_set_inode_block_group(leaf, inode_item,
-                                   block_group->key.objectid);
+       btrfs_set_inode_block_group(leaf, inode_item, offset);
        btrfs_mark_buffer_dirty(leaf);
        btrfs_release_path(root, path);
 
        key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-       key.offset = block_group->key.objectid;
+       key.offset = offset;
        key.type = 0;
 
        ret = btrfs_insert_empty_item(trans, root, path, &key,
@@ -156,6 +163,22 @@ int create_free_space_inode(struct btrfs_root *root,
        return 0;
 }
 
+int create_free_space_inode(struct btrfs_root *root,
+                           struct btrfs_trans_handle *trans,
+                           struct btrfs_block_group_cache *block_group,
+                           struct btrfs_path *path)
+{
+       int ret;
+       u64 ino;
+
+       ret = btrfs_find_free_objectid(root, &ino);
+       if (ret < 0)
+               return ret;
+
+       return __create_free_space_inode(root, trans, path, ino,
+                                        block_group->key.objectid);
+}
+
 int btrfs_truncate_free_space_cache(struct btrfs_root *root,
                                    struct btrfs_trans_handle *trans,
                                    struct btrfs_path *path,
@@ -208,16 +231,13 @@ static int readahead_cache(struct inode *inode)
        return 0;
 }
 
-int load_free_space_cache(struct btrfs_fs_info *fs_info,
-                         struct btrfs_block_group_cache *block_group)
+int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
+                           struct btrfs_free_space_ctl *ctl,
+                           struct btrfs_path *path, u64 offset)
 {
-       struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
-       struct btrfs_root *root = fs_info->tree_root;
-       struct inode *inode;
        struct btrfs_free_space_header *header;
        struct extent_buffer *leaf;
        struct page *page;
-       struct btrfs_path *path;
        u32 *checksums = NULL, *crc;
        char *disk_crcs = NULL;
        struct btrfs_key key;
@@ -225,76 +245,47 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
        u64 num_entries;
        u64 num_bitmaps;
        u64 generation;
-       u64 used = btrfs_block_group_used(&block_group->item);
        u32 cur_crc = ~(u32)0;
        pgoff_t index = 0;
        unsigned long first_page_offset;
        int num_checksums;
-       int ret = 0;
-
-       /*
-        * If we're unmounting then just return, since this does a search on the
-        * normal root and not the commit root and we could deadlock.
-        */
-       smp_mb();
-       if (fs_info->closing)
-               return 0;
-
-       /*
-        * If this block group has been marked to be cleared for one reason or
-        * another then we can't trust the on disk cache, so just return.
-        */
-       spin_lock(&block_group->lock);
-       if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
-               spin_unlock(&block_group->lock);
-               return 0;
-       }
-       spin_unlock(&block_group->lock);
+       int ret = 0, ret2;
 
        INIT_LIST_HEAD(&bitmaps);
 
-       path = btrfs_alloc_path();
-       if (!path)
-               return 0;
-
-       inode = lookup_free_space_inode(root, block_group, path);
-       if (IS_ERR(inode)) {
-               btrfs_free_path(path);
-               return 0;
-       }
-
        /* Nothing in the space cache, goodbye */
-       if (!i_size_read(inode)) {
-               btrfs_free_path(path);
+       if (!i_size_read(inode))
                goto out;
-       }
 
        key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-       key.offset = block_group->key.objectid;
+       key.offset = offset;
        key.type = 0;
 
        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-       if (ret) {
-               btrfs_free_path(path);
+       if (ret < 0)
+               goto out;
+       else if (ret > 0) {
+               btrfs_release_path(root, path);
+               ret = 0;
                goto out;
        }
 
+       ret = -1;
+
        leaf = path->nodes[0];
        header = btrfs_item_ptr(leaf, path->slots[0],
                                struct btrfs_free_space_header);
        num_entries = btrfs_free_space_entries(leaf, header);
        num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
        generation = btrfs_free_space_generation(leaf, header);
-       btrfs_free_path(path);
+       btrfs_release_path(root, path);
 
        if (BTRFS_I(inode)->generation != generation) {
                printk(KERN_ERR "btrfs: free space inode generation (%llu) did"
-                      " not match free space cache generation (%llu) for "
-                      "block group %llu\n",
+                      " not match free space cache generation (%llu)\n",
                       (unsigned long long)BTRFS_I(inode)->generation,
-                      (unsigned long long)generation,
-                      (unsigned long long)block_group->key.objectid);
-               goto free_cache;
+                      (unsigned long long)generation);
+               goto out;
        }
 
        if (!num_entries)
@@ -311,10 +302,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                goto out;
 
        ret = readahead_cache(inode);
-       if (ret) {
-               ret = 0;
+       if (ret)
                goto out;
-       }
 
        while (1) {
                struct btrfs_free_space_entry *entry;
@@ -333,10 +322,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                }
 
                page = grab_cache_page(inode->i_mapping, index);
-               if (!page) {
-                       ret = 0;
+               if (!page)
                        goto free_cache;
-               }
 
                if (!PageUptodate(page)) {
                        btrfs_readpage(NULL, page);
@@ -345,9 +332,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                                unlock_page(page);
                                page_cache_release(page);
                                printk(KERN_ERR "btrfs: error reading free "
-                                      "space cache: %llu\n",
-                                      (unsigned long long)
-                                      block_group->key.objectid);
+                                      "space cache\n");
                                goto free_cache;
                        }
                }
@@ -360,13 +345,10 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                        gen = addr + (sizeof(u32) * num_checksums);
                        if (*gen != BTRFS_I(inode)->generation) {
                                printk(KERN_ERR "btrfs: space cache generation"
-                                      " (%llu) does not match inode (%llu) "
-                                      "for block group %llu\n",
+                                      " (%llu) does not match inode (%llu)\n",
                                       (unsigned long long)*gen,
                                       (unsigned long long)
-                                      BTRFS_I(inode)->generation,
-                                      (unsigned long long)
-                                      block_group->key.objectid);
+                                      BTRFS_I(inode)->generation);
                                kunmap(page);
                                unlock_page(page);
                                page_cache_release(page);
@@ -382,9 +364,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                                          PAGE_CACHE_SIZE - start_offset);
                btrfs_csum_final(cur_crc, (char *)&cur_crc);
                if (cur_crc != *crc) {
-                       printk(KERN_ERR "btrfs: crc mismatch for page %lu in "
-                              "block group %llu\n", index,
-                              (unsigned long long)block_group->key.objectid);
+                       printk(KERN_ERR "btrfs: crc mismatch for page %lu\n",
+                              index);
                        kunmap(page);
                        unlock_page(page);
                        page_cache_release(page);
@@ -432,7 +413,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                                        goto free_cache;
                                }
                                spin_lock(&ctl->tree_lock);
-                               ret = link_free_space(ctl, e);
+                               ret2 = link_free_space(ctl, e);
                                ctl->total_bitmaps++;
                                ctl->op->recalc_thresholds(ctl);
                                spin_unlock(&ctl->tree_lock);
@@ -471,42 +452,96 @@ next:
                index++;
        }
 
-       spin_lock(&ctl->tree_lock);
-       if (ctl->free_space != (block_group->key.offset - used -
-                               block_group->bytes_super)) {
-               spin_unlock(&ctl->tree_lock);
-               printk(KERN_ERR "block group %llu has an wrong amount of free "
-                      "space\n", block_group->key.objectid);
-               ret = 0;
-               goto free_cache;
-       }
-       spin_unlock(&ctl->tree_lock);
-
        ret = 1;
 out:
        kfree(checksums);
        kfree(disk_crcs);
-       iput(inode);
        return ret;
-
 free_cache:
-       /* This cache is bogus, make sure it gets cleared */
-       spin_lock(&block_group->lock);
-       block_group->disk_cache_state = BTRFS_DC_CLEAR;
-       spin_unlock(&block_group->lock);
-       btrfs_remove_free_space_cache(block_group);
+       __btrfs_remove_free_space_cache(ctl);
        goto out;
 }
 
-int btrfs_write_out_cache(struct btrfs_root *root,
-                         struct btrfs_trans_handle *trans,
-                         struct btrfs_block_group_cache *block_group,
-                         struct btrfs_path *path)
+int load_free_space_cache(struct btrfs_fs_info *fs_info,
+                         struct btrfs_block_group_cache *block_group)
 {
        struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
+       struct btrfs_root *root = fs_info->tree_root;
+       struct inode *inode;
+       struct btrfs_path *path;
+       int ret;
+       bool matched;
+       u64 used = btrfs_block_group_used(&block_group->item);
+
+       /*
+        * If we're unmounting then just return, since this does a search on the
+        * normal root and not the commit root and we could deadlock.
+        */
+       smp_mb();
+       if (fs_info->closing)
+               return 0;
+
+       /*
+        * If this block group has been marked to be cleared for one reason or
+        * another then we can't trust the on disk cache, so just return.
+        */
+       spin_lock(&block_group->lock);
+       if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
+               spin_unlock(&block_group->lock);
+               return 0;
+       }
+       spin_unlock(&block_group->lock);
+
+       path = btrfs_alloc_path();
+       if (!path)
+               return 0;
+
+       inode = lookup_free_space_inode(root, block_group, path);
+       if (IS_ERR(inode)) {
+               btrfs_free_path(path);
+               return 0;
+       }
+
+       ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
+                                     path, block_group->key.objectid);
+       btrfs_free_path(path);
+       if (ret <= 0)
+               goto out;
+
+       spin_lock(&ctl->tree_lock);
+       matched = (ctl->free_space == (block_group->key.offset - used -
+                                      block_group->bytes_super));
+       spin_unlock(&ctl->tree_lock);
+
+       if (!matched) {
+               __btrfs_remove_free_space_cache(ctl);
+               printk(KERN_ERR "block group %llu has an wrong amount of free "
+                      "space\n", block_group->key.objectid);
+               ret = -1;
+       }
+out:
+       if (ret < 0) {
+               /* This cache is bogus, make sure it gets cleared */
+               spin_lock(&block_group->lock);
+               block_group->disk_cache_state = BTRFS_DC_CLEAR;
+               spin_unlock(&block_group->lock);
+
+               printk(KERN_ERR "btrfs: failed to load free space cache "
+                      "for block group %llu\n", block_group->key.objectid);
+       }
+
+       iput(inode);
+       return ret;
+}
+
+int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
+                           struct btrfs_free_space_ctl *ctl,
+                           struct btrfs_block_group_cache *block_group,
+                           struct btrfs_trans_handle *trans,
+                           struct btrfs_path *path, u64 offset)
+{
        struct btrfs_free_space_header *header;
        struct extent_buffer *leaf;
-       struct inode *inode;
        struct rb_node *node;
        struct list_head *pos, *n;
        struct page **pages;
@@ -523,35 +558,18 @@ int btrfs_write_out_cache(struct btrfs_root *root,
        int index = 0, num_pages = 0;
        int entries = 0;
        int bitmaps = 0;
-       int ret = 0;
+       int ret = -1;
        bool next_page = false;
        bool out_of_space = false;
 
-       root = root->fs_info->tree_root;
-
        INIT_LIST_HEAD(&bitmap_list);
 
-       spin_lock(&block_group->lock);
-       if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
-               spin_unlock(&block_group->lock);
-               return 0;
-       }
-       spin_unlock(&block_group->lock);
-
-       inode = lookup_free_space_inode(root, block_group, path);
-       if (IS_ERR(inode))
-               return 0;
-
-       if (!i_size_read(inode)) {
-               iput(inode);
-               return 0;
-       }
-
        node = rb_first(&ctl->free_space_offset);
-       if (!node) {
-               iput(inode);
+       if (!node)
                return 0;
-       }
+
+       if (!i_size_read(inode))
+               return -1;
 
        num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
                PAGE_CACHE_SHIFT;
@@ -561,16 +579,13 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 
        /* We need a checksum per page. */
        crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS);
-       if (!crc) {
-               iput(inode);
-               return 0;
-       }
+       if (!crc)
+               return -1;
 
        pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS);
        if (!pages) {
                kfree(crc);
-               iput(inode);
-               return 0;
+               return -1;
        }
 
        /* Since the first page has all of our checksums and our generation we
@@ -580,7 +595,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
        first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64);
 
        /* Get the cluster for this block_group if it exists */
-       if (!list_empty(&block_group->cluster_list))
+       if (block_group && !list_empty(&block_group->cluster_list))
                cluster = list_entry(block_group->cluster_list.next,
                                     struct btrfs_free_cluster,
                                     block_group_list);
@@ -622,7 +637,8 @@ int btrfs_write_out_cache(struct btrfs_root *root,
         * When searching for pinned extents, we need to start at our start
         * offset.
         */
-       start = block_group->key.objectid;
+       if (block_group)
+               start = block_group->key.objectid;
 
        /* Write out the extent entries */
        do {
@@ -680,8 +696,9 @@ int btrfs_write_out_cache(struct btrfs_root *root,
                 * We want to add any pinned extents to our free space cache
                 * so we don't leak the space
                 */
-               while (!next_page && (start < block_group->key.objectid +
-                                     block_group->key.offset)) {
+               while (block_group && !next_page &&
+                      (start < block_group->key.objectid +
+                       block_group->key.offset)) {
                        ret = find_first_extent_bit(unpin, start, &start, &end,
                                                    EXTENT_DIRTY);
                        if (ret) {
@@ -799,12 +816,12 @@ int btrfs_write_out_cache(struct btrfs_root *root,
        filemap_write_and_wait(inode->i_mapping);
 
        key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-       key.offset = block_group->key.objectid;
+       key.offset = offset;
        key.type = 0;
 
        ret = btrfs_search_slot(trans, root, &key, path, 1, 1);
        if (ret < 0) {
-               ret = 0;
+               ret = -1;
                clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
                                 EXTENT_DIRTY | EXTENT_DELALLOC |
                                 EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS);
@@ -817,8 +834,8 @@ int btrfs_write_out_cache(struct btrfs_root *root,
                path->slots[0]--;
                btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
                if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
-                   found_key.offset != block_group->key.objectid) {
-                       ret = 0;
+                   found_key.offset != offset) {
+                       ret = -1;
                        clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
                                         EXTENT_DIRTY | EXTENT_DELALLOC |
                                         EXTENT_DO_ACCOUNTING, 0, 0, NULL,
@@ -838,16 +855,49 @@ int btrfs_write_out_cache(struct btrfs_root *root,
        ret = 1;
 
 out_free:
-       if (ret == 0) {
+       if (ret != 1) {
                invalidate_inode_pages2_range(inode->i_mapping, 0, index);
-               spin_lock(&block_group->lock);
-               block_group->disk_cache_state = BTRFS_DC_ERROR;
-               spin_unlock(&block_group->lock);
                BTRFS_I(inode)->generation = 0;
        }
        kfree(checksums);
        kfree(pages);
        btrfs_update_inode(trans, root, inode);
+       return ret;
+}
+
+int btrfs_write_out_cache(struct btrfs_root *root,
+                         struct btrfs_trans_handle *trans,
+                         struct btrfs_block_group_cache *block_group,
+                         struct btrfs_path *path)
+{
+       struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
+       struct inode *inode;
+       int ret = 0;
+
+       root = root->fs_info->tree_root;
+
+       spin_lock(&block_group->lock);
+       if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
+               spin_unlock(&block_group->lock);
+               return 0;
+       }
+       spin_unlock(&block_group->lock);
+
+       inode = lookup_free_space_inode(root, block_group, path);
+       if (IS_ERR(inode))
+               return 0;
+
+       ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
+                                     path, block_group->key.objectid);
+       if (ret < 0) {
+               spin_lock(&block_group->lock);
+               block_group->disk_cache_state = BTRFS_DC_ERROR;
+               spin_unlock(&block_group->lock);
+
+               printk(KERN_ERR "btrfs: failed to write free space cace "
+                      "for block group %llu\n", block_group->key.objectid);
+       }
+
        iput(inode);
        return ret;
 }