Btrfs: if we have a lot of pinned space, commit the transaction
Mitch kept hitting a panic because he was getting ENOSPC. One of my previous
patches makes it so we are much better at not allocating new metadata chunks.
Unfortunately coupled with the overcommit patch this works us into a bit of a
problem if we are removing a bunch of space and end up chewing up all of our
space with pinned extents. We can allocate chunks fine and overflow is ok, but
the only way to reclaim this space is to commit the transaction. So if we go to
overcommit, first check and see how much pinned space we have. If we have more
than 80% of the free space chewed up with pinned extents, just commit the
transaction, this will free up enough space for our reservation and we won't
have this problem anymore. With this patch Mitch's test doesn't blow up
anymore. Thanks,
Reported-and-tested-by: Mitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: Josef Bacik <josef@redhat.com>
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index a5f1421..4eb7d2b 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3510,6 +3510,20 @@
u64 profile = btrfs_get_alloc_profile(root, 0);
u64 avail;
+ /*
+ * If we have a lot of space that's pinned, don't bother doing
+ * the overcommit dance yet and just commit the transaction.
+ */
+ avail = (space_info->total_bytes - space_info->bytes_used) * 8;
+ do_div(avail, 10);
+ if (space_info->bytes_pinned >= avail && flush && !trans &&
+ !committed) {
+ space_info->flush = 1;
+ flushing = true;
+ spin_unlock(&space_info->lock);
+ goto commit;
+ }
+
spin_lock(&root->fs_info->free_chunk_lock);
avail = root->fs_info->free_chunk_space;
@@ -3581,6 +3595,7 @@
if (trans)
goto out;
+commit:
ret = -ENOSPC;
if (committed)
goto out;