4.19-stable review patch. If anyone has any objections, please let me know.
------------------
From: Matthew Wilcox <willy@xxxxxxxxxxxxx>
commit c93db7bb6ef3251e0ea48ade311d3e9942748e1c upstream.
If we race with inode destroy, it's possible for page->mapping to be
NULL before we even enter this routine, as well as after having slept
waiting for the dax entry to become unlocked.
Fixes: c2a7d2a11552 ("filesystem-dax: Introduce dax_lock_mapping_entry()")
Cc: <stable@xxxxxxxxxxxxxxx>
Reported-by: Jan Kara <jack@xxxxxxx>
Signed-off-by: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Reviewed-by: Johannes Thumshirn <jthumshirn@xxxxxxx>
Reviewed-by: Jan Kara <jack@xxxxxxx>
Signed-off-by: Dan Williams <dan.j.williams@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
fs/dax.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -423,7 +423,7 @@ bool dax_lock_mapping_entry(struct page
for (;;) {
mapping = READ_ONCE(page->mapping);
- if (!dax_mapping(mapping))
+ if (!mapping || !dax_mapping(mapping))
break;
/*
