Btrfs: fix estale with btrfs send

commit 139f807a1eba1e484941a98fb93ee32ad859a6a1 upstream.

This fixes bugzilla 57491.  If we take a snapshot of a fs with a unlink ongoing
and then try to send that root we will run into problems.  When comparing with a
parent root we will search the parents and the send roots commit_root, which if
we've just created the snapshot will include the file that needs to be evicted
by the orphan cleanup.  So when we find a changed extent we will try and copy
that info into the send stream, but when we lookup the inode we use the normal
root, which no longer has the inode because the orphan cleanup deleted it.  The
best solution I have for this is to check our otransid with the generation of
the commit root and if they match just commit the transaction again, that way we
get the changes from the orphan cleanup.  With this patch the reproducer I made
for this bugzilla no longer returns ESTALE when trying to do the send.  Thanks,

Reported-by: Chris Wilson <jakdaw@gmail.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index ff40f1c00ce315d59eaf4da6d20145351a000e51..ba9690b9ae2405d7e8b500eaadc0db6225676eb6 100644 (file)
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -4579,6 +4579,41 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
        send_root = BTRFS_I(file_inode(mnt_file))->root;
        fs_info = send_root->fs_info;
 
+       /*
+        * This is done when we lookup the root, it should already be complete
+        * by the time we get here.
+        */
+       WARN_ON(send_root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE);
+
+       /*
+        * If we just created this root we need to make sure that the orphan
+        * cleanup has been done and committed since we search the commit root,
+        * so check its commit root transid with our otransid and if they match
+        * commit the transaction to make sure everything is updated.
+        */
+       down_read(&send_root->fs_info->extent_commit_sem);
+       if (btrfs_header_generation(send_root->commit_root) ==
+           btrfs_root_otransid(&send_root->root_item)) {
+               struct btrfs_trans_handle *trans;
+
+               up_read(&send_root->fs_info->extent_commit_sem);
+
+               trans = btrfs_attach_transaction_barrier(send_root);
+               if (IS_ERR(trans)) {
+                       if (PTR_ERR(trans) != -ENOENT) {
+                               ret = PTR_ERR(trans);
+                               goto out;
+                       }
+                       /* ENOENT means theres no transaction */
+               } else {
+                       ret = btrfs_commit_transaction(trans, send_root);
+                       if (ret)
+                               goto out;
+               }
+       } else {
+               up_read(&send_root->fs_info->extent_commit_sem);
+       }
+
        arg = memdup_user(arg_, sizeof(*arg));
        if (IS_ERR(arg)) {
                ret = PTR_ERR(arg);