We probably need to fix superblock leak in patch (v4 "fs: Add user
namesapace member to struct super_block"):
Imagine posible code path in sget_userns: we iterate through
type->fs_supers and do not find suitable sb, we drop sb_lock to
allocate s and go to retry. After we dropped sb_lock some other
task from different userns takes sb_lock, it is already in retry
stage and has s allocated, so it puts its s in type->fs_supers
list. So in retry we will find these sb in list and check it has
a different userns, and finally we will return without freeing s.
Signed-off-by: Pavel Tikhomirov <ptikhomirov@xxxxxxxxxxxxx>
---
fs/super.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/fs/super.c b/fs/super.c
index b4ee02b..24771b5 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -458,6 +458,10 @@ struct super_block *sget_userns(struct file_system_type *type,
continue;
if (user_ns != old->s_user_ns) {
spin_unlock(&sb_lock);
+ if (s) {
+ up_write(&s->s_umount);
+ destroy_super(s);
+ }
return ERR_PTR(-EBUSY);
}
if (!grab_super(old))
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
