I was trying to debug something stupid I did and noticed I wasn't
getting ref verify errors before everything blew up. Turns out it's
because we don't warn when we try to add a normal ref via
btrfs_inc_ref() if the block entry exists but has 0 references. This is
incorrect, we should never be doing anything other than adding a new
extent once a block entry drops to 0 references. With this fix it blew
up when I did the stupid thing, and was able to undo my stupid thing.
Signed-off-by: Josef Bacik <josef@xxxxxxxxxxxxxx>
---
fs/btrfs/ref-verify.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/fs/btrfs/ref-verify.c b/fs/btrfs/ref-verify.c
index 454a1015d026..679464e5f06a 100644
--- a/fs/btrfs/ref-verify.c
+++ b/fs/btrfs/ref-verify.c
@@ -803,6 +803,14 @@ int btrfs_ref_tree_mod(struct btrfs_fs_info *fs_info,
kfree(ref);
kfree(ra);
goto out_unlock;
+ } else if (be->num_refs == 0) {
+ btrfs_err(fs_info,
+"trying to do action %d for a bytenr that has 0 total references", action);
+ dump_block_entry(fs_info, be);
+ dump_ref_action(fs_info, ra);
+ kfree(ref);
+ kfree(ra);
+ goto out_unlock;
}
if (!parent) {
--
2.24.1
