summaryrefslogtreecommitdiffstats
path: root/net/rds/send.c
diff options
context:
space:
mode:
authorAndy Grover <andy.grover@oracle.com>2010-04-16 02:19:29 +0200
committerAndy Grover <andy.grover@oracle.com>2010-09-09 03:12:25 +0200
commitce47f52f42e69d48d1b63fa618fce9cd7ffa9417 (patch)
tree6abfde17a22d48b2659ae33e69b69a3349afb749 /net/rds/send.c
parentRDS: Use a generation counter to avoid rds_send_xmit loop (diff)
downloadlinux-ce47f52f42e69d48d1b63fa618fce9cd7ffa9417.tar.xz
linux-ce47f52f42e69d48d1b63fa618fce9cd7ffa9417.zip
RDS: Update comments in rds_send_xmit()
Update comments to reflect changes in previous commit. Keeping as separate commits due to different authorship. Signed-off-by: Andy Grover <andy.grover@oracle.com>
Diffstat (limited to 'net/rds/send.c')
-rw-r--r--net/rds/send.c5
1 files changed, 2 insertions, 3 deletions
diff --git a/net/rds/send.c b/net/rds/send.c
index 663fd60b40cf..8e3fd9981c2e 100644
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -353,15 +353,14 @@ restart:
/*
* Other senders will see we have c_send_lock and exit. We
* need to recheck the send queue and race again for c_send_lock
- * to make sure messages don't just sit on the send queue.
+ * to make sure messages don't just sit on the send queue, if
+ * somebody hasn't already beat us into the loop.
*
* If the transport cannot continue (i.e ret != 0), then it must
* call us when more room is available, such as from the tx
* completion handler.
*/
if (ret == 0) {
- /* A simple bit test would be way faster than taking the
- * spin lock */
smp_mb();
if (!list_empty(&conn->c_send_queue)) {
rds_stats_inc(s_send_lock_queue_raced);