Btrfs: use the commit_root for reading free_space_inode crcs

Now that we are using regular file crcs for the free space cache,
we can deadlock if we try to read the free_space_inode while we are
updating the crc tree.

This commit fixes things by using the commit_root to read the crcs.  This is
safe because we the free space cache file would already be loaded if
that block group had been changed in the current transaction.

Signed-off-by: Chris Mason <chris.mason@oracle.com>

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 03dce3f40ce004ac202c5020499700ab1ef7d8cf..502b9e98867949736b03d5bbf55b14b7d67205e7 100644 (file)
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -187,4 +187,13 @@ static inline void btrfs_i_size_write(struct inode *inode, u64 size)
        BTRFS_I(inode)->disk_i_size = size;
 }
 
+static inline bool btrfs_is_free_space_inode(struct btrfs_root *root,
+                                      struct inode *inode)
+{
+       if (root == root->fs_info->tree_root ||
+           BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
+               return true;
+       return false;
+}
+
 #endif

diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index bb22a4435b2d1b6bb5ae0543282dc074b0c7cc36..08bcfa92a2226bcfd3ba0edf6fe8eb944454612b 100644 (file)
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -177,6 +177,15 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 
        WARN_ON(bio->bi_vcnt <= 0);
 
+       /*
+        * the free space stuff is only read when it hasn't been
+        * updated in the current transaction.  So, we can safely
+        * read from the commit root and sidestep a nasty deadlock
+        * between reading the free space cache and updating the csum tree.
+        */
+       if (btrfs_is_free_space_inode(root, inode))
+               path->search_commit_root = 1;
+
        disk_bytenr = (u64)bio->bi_sector << 9;
        if (dio)
                offset = logical_offset;

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 3b5763a2b7f95fb785da40b5b6af7725c331b2cf..4b9be2883a8df48cc4b571173ec3dc18157da042 100644 (file)
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -750,15 +750,6 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
        return alloc_hint;
 }
 
-static inline bool is_free_space_inode(struct btrfs_root *root,
-                                      struct inode *inode)
-{
-       if (root == root->fs_info->tree_root ||
-           BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
-               return true;
-       return false;
-}
-
 /*
  * when extent_io.c finds a delayed allocation range in the file,
  * the call backs end up in this code.  The basic idea is to
@@ -791,7 +782,7 @@ static noinline int cow_file_range(struct inode *inode,
        struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
        int ret = 0;
 
-       BUG_ON(is_free_space_inode(root, inode));
+       BUG_ON(btrfs_is_free_space_inode(root, inode));
        trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
        trans->block_rsv = &root->fs_info->delalloc_block_rsv;
@@ -1072,7 +1063,7 @@ static noinline int run_delalloc_nocow(struct inode *inode,
        path = btrfs_alloc_path();
        BUG_ON(!path);
 
-       nolock = is_free_space_inode(root, inode);
+       nolock = btrfs_is_free_space_inode(root, inode);
 
        if (nolock)
                trans = btrfs_join_transaction_nolock(root);
@@ -1341,7 +1332,7 @@ static int btrfs_set_bit_hook(struct inode *inode,
        if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
                struct btrfs_root *root = BTRFS_I(inode)->root;
                u64 len = state->end + 1 - state->start;
-               bool do_list = !is_free_space_inode(root, inode);
+               bool do_list = !btrfs_is_free_space_inode(root, inode);
 
                if (*bits & EXTENT_FIRST_DELALLOC) {
                        *bits &= ~EXTENT_FIRST_DELALLOC;
@@ -1377,7 +1368,7 @@ static int btrfs_clear_bit_hook(struct inode *inode,
        if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
                struct btrfs_root *root = BTRFS_I(inode)->root;
                u64 len = state->end + 1 - state->start;
-               bool do_list = !is_free_space_inode(root, inode);
+               bool do_list = !btrfs_is_free_space_inode(root, inode);
 
                if (*bits & EXTENT_FIRST_DELALLOC) {
                        *bits &= ~EXTENT_FIRST_DELALLOC;
@@ -1487,7 +1478,7 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 
        skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
-       if (is_free_space_inode(root, inode))
+       if (btrfs_is_free_space_inode(root, inode))
                ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2);
        else
                ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
@@ -1736,7 +1727,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
                return 0;
        BUG_ON(!ordered_extent);
 
-       nolock = is_free_space_inode(root, inode);
+       nolock = btrfs_is_free_space_inode(root, inode);
 
        if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
                BUG_ON(!list_empty(&ordered_extent->list));
@@ -2670,7 +2661,7 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans,
         * The data relocation inode should also be directly updated
         * without delay
         */
-       if (!is_free_space_inode(root, inode)
+       if (!btrfs_is_free_space_inode(root, inode)
            && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) {
                ret = btrfs_delayed_update_inode(trans, root, inode);
                if (!ret)
@@ -3620,7 +3611,7 @@ void btrfs_evict_inode(struct inode *inode)
 
        truncate_inode_pages(&inode->i_data, 0);
        if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 ||
-                              is_free_space_inode(root, inode)))
+                              btrfs_is_free_space_inode(root, inode)))
                goto no_delete;
 
        if (is_bad_inode(inode)) {
@@ -4263,7 +4254,7 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
        if (BTRFS_I(inode)->dummy_inode)
                return 0;
 
-       if (btrfs_fs_closing(root->fs_info) && is_free_space_inode(root, inode))
+       if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(root, inode))
                nolock = true;
 
        if (wbc->sync_mode == WB_SYNC_ALL) {
@@ -6817,7 +6808,7 @@ int btrfs_drop_inode(struct inode *inode)
        struct btrfs_root *root = BTRFS_I(inode)->root;
 
        if (btrfs_root_refs(&root->root_item) == 0 &&
-           !is_free_space_inode(root, inode))
+           !btrfs_is_free_space_inode(root, inode))
                return 1;
        else
                return generic_drop_inode(inode);