summaryrefslogtreecommitdiff
path: root/fs
diff options
context:
space:
mode:
authorJeff Moyer <jmoyer@redhat.com>2014-09-02 13:17:00 -0400
committerBen Hutchings <ben@decadent.org.uk>2014-11-05 20:27:40 +0000
commit1cc8e21690688b24b0b9388dbb14e1d2cf4f94b4 (patch)
tree66c1fbf3c45571f78bb1b4c907f134affada8f0c /fs
parent11d3b506cc0db26d2f2078bd4df88803cd06c819 (diff)
aio: add missing smp_rmb() in read_events_ring
commit 2ff396be602f10b5eab8e73b24f20348fa2de159 upstream. We ran into a case on ppc64 running mariadb where io_getevents would return zeroed out I/O events. After adding instrumentation, it became clear that there was some missing synchronization between reading the tail pointer and the events themselves. This small patch fixes the problem in testing. Thanks to Zach for helping to look into this, and suggesting the fix. Signed-off-by: Jeff Moyer <jmoyer@redhat.com> Signed-off-by: Benjamin LaHaise <bcrl@kvack.org> [bwh: Backported to 3.2: adjust context, indentation] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Diffstat (limited to 'fs')
-rw-r--r--fs/aio.c7
1 files changed, 7 insertions, 0 deletions
diff --git a/fs/aio.c b/fs/aio.c
index 8cdd8ea0f44a..9acfd0708d92 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -1102,6 +1102,13 @@ static int aio_read_evt(struct kioctx *ioctx, struct io_event *ent)
head = ring->head % info->nr;
if (head != ring->tail) {
struct io_event *evp = aio_ring_event(info, head, KM_USER1);
+
+ /*
+ * Ensure that once we've read the current tail pointer, that
+ * we also see the events that were stored up to the tail.
+ */
+ smp_rmb();
+
*ent = *evp;
head = (head + 1) % info->nr;
smp_mb(); /* finish reading the event before updatng the head */