diff options
| author | Jan Kara <jack@suse.cz> | 2019-07-29 13:57:49 +0200 |
|---|---|---|
| committer | Dan Williams <dan.j.williams@intel.com> | 2019-07-29 09:24:22 -0700 |
| commit | 61c30c98ef17e5a330d7bb8494b78b3d6dffe9b8 (patch) | |
| tree | 15ccf47d90a61d3b8a9c7abc0558e6c1054d1c9e /fs | |
| parent | 609488bc979f99f805f34e9a32c1e3b71179d10b (diff) | |
dax: Fix missed wakeup in put_unlocked_entry()
The condition checking whether put_unlocked_entry() needs to wake up
following waiter got broken by commit 23c84eb78375 ("dax: Fix missed
wakeup with PMD faults"). We need to wake the waiter whenever the passed
entry is valid (i.e., non-NULL and not special conflict entry). This
could lead to processes never being woken up when waiting for entry
lock. Fix the condition.
Cc: <stable@vger.kernel.org>
Link: http://lore.kernel.org/r/20190729120228.GC17833@quack2.suse.cz
Fixes: 23c84eb78375 ("dax: Fix missed wakeup with PMD faults")
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Diffstat (limited to 'fs')
| -rw-r--r-- | fs/dax.c | 2 |
1 files changed, 1 insertions, 1 deletions
@@ -266,7 +266,7 @@ static void wait_entry_unlocked(struct xa_state *xas, void *entry) static void put_unlocked_entry(struct xa_state *xas, void *entry) { /* If we were the only waiter woken, wake the next one */ - if (entry && dax_is_conflict(entry)) + if (entry && !dax_is_conflict(entry)) dax_wake_entry(xas, entry, false); } |