Commit 94b60442 authored by Chris Mason's avatar Chris Mason

Btrfs: avoid ENOSPC errors in btrfs_dirty_inode

btrfs_dirty_inode tries to sneak in without much waiting or
space reservation, mostly for performance reasons.  This
usually works well but can cause problems when there are
many many writers.

When btrfs_update_inode fails with ENOSPC, we fallback
to a slower btrfs_start_transaction call that will reserve
some space.
Signed-off-by: default avatarChris Mason <chris.mason@oracle.com>
parent 3f7c579c
...@@ -4308,10 +4308,18 @@ void btrfs_dirty_inode(struct inode *inode) ...@@ -4308,10 +4308,18 @@ void btrfs_dirty_inode(struct inode *inode)
btrfs_set_trans_block_group(trans, inode); btrfs_set_trans_block_group(trans, inode);
ret = btrfs_update_inode(trans, root, inode); ret = btrfs_update_inode(trans, root, inode);
if (ret) if (ret && ret == -ENOSPC) {
printk(KERN_ERR"btrfs: fail to dirty inode %lu error %d\n", /* whoops, lets try again with the full transaction */
inode->i_ino, ret); btrfs_end_transaction(trans, root);
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
ret = btrfs_update_inode(trans, root, inode);
if (ret) {
printk(KERN_ERR"btrfs: fail to dirty inode %lu error %d\n",
inode->i_ino, ret);
}
}
btrfs_end_transaction(trans, root); btrfs_end_transaction(trans, root);
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment