Re: BUG_ON(!mapping_empty(&inode->i_data))

[Hugh Dickins]

On Thu, 29 Apr 2021, Andrew Morton wrote:
> 
> I'm not sure this ever was resolved?

It was not resolved: Matthew had prospective fixes for one way in which
it could happen, but they did not help the case which still hits my
testing (well, I replace the BUG_ON by a WARN_ON, so not hit badly).

> 
> Is it the case that the series "Remove nrexceptional tracking v2" at
> least exposed this bug?

Yes: makes a BUG out of a long-standing issue not noticed before.

> 
> IOW, what the heck should I do with
> 
> mm-introduce-and-use-mapping_empty.patch
> mm-stop-accounting-shadow-entries.patch
> dax-account-dax-entries-as-nrpages.patch
> mm-remove-nrexceptional-from-inode.patch

If Matthew doesn't have a proper fix yet (and it's a bit late for more
than an obvious fix), I think those should go in, with this addition:

[PATCH] mm: remove nrexceptional from inode: remove BUG_ON

clear_inode()'s BUG_ON(!mapping_empty(&inode->i_data)) is unsafe: we know
of two ways in which nodes can and do (on rare occasions) get left behind.
Until those are fixed, do not BUG_ON() nor even WARN_ON(). Yes, this will
then leak those nodes (or the next user of the struct inode may use them);
but this has been happening for years, and the new BUG_ON(!mapping_empty)
was only guilty of revealing that. A proper fix will follow, but no hurry.

Signed-off-by: Hugh Dickins <hughd@xxxxxxxxxx>
---

 fs/inode.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

--- mmotm/fs/inode.c	2021-04-22 18:30:46.285908982 -0700
+++ linux/fs/inode.c	2021-04-29 22:13:54.096530691 -0700
@@ -529,7 +529,14 @@ void clear_inode(struct inode *inode)
 	 */
 	xa_lock_irq(&inode->i_data.i_pages);
 	BUG_ON(inode->i_data.nrpages);
-	BUG_ON(!mapping_empty(&inode->i_data));
+	/*
+	 * Almost always, mapping_empty(&inode->i_data) here; but there are
+	 * two known and long-standing ways in which nodes may get left behind
+	 * (when deep radix-tree node allocation failed partway; or when THP
+	 * collapse_file() failed). Until those two known cases are cleaned up,
+	 * or a cleanup function is called here, do not BUG_ON(!mapping_empty),
+	 * nor even WARN_ON(!mapping_empty).
+	 */
 	xa_unlock_irq(&inode->i_data.i_pages);
 	BUG_ON(!list_empty(&inode->i_data.private_list));
 	BUG_ON(!(inode->i_state & I_FREEING));