Andy Grover
b98ba52f96
RDS: only put sockets that have seen congestion on the poll_waitq ...
rds_poll_waitq's listeners will be awoken if we receive a congestion
notification. Bad performance may result because *all* polled sockets
contend for this single lock. However, it should not be necessary to
wake pollers when a congestion update arrives if they have never
experienced congestion, and not putting these on the waitq will
hopefully greatly reduce contention.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-03-16 21:16:59 -07:00
2010-03-13 08:57:29 -06:00
2009-11-12 02:05:06 -08:00
2010-02-16 23:05:38 -08:00
2010-02-10 11:12:09 -08:00
2010-02-10 12:31:10 -08:00
2010-02-10 11:12:09 -08:00
2010-03-13 14:50:18 -08:00
2010-03-16 14:37:47 -07:00
2010-02-02 07:21:34 -08:00
2010-03-16 14:15:45 -07:00
2010-02-18 14:30:18 -08:00
2010-03-15 16:00:50 -07:00
2010-02-25 09:41:03 +01:00
2009-09-01 01:13:07 -07:00
2009-11-25 15:14:13 -08:00
2009-12-26 20:38:23 -08:00
2009-11-25 15:14:13 -08:00
2010-03-16 14:15:47 -07:00
2010-03-13 12:23:29 -08:00
2010-02-10 12:31:10 -08:00
2010-02-18 14:30:18 -08:00
2009-12-15 08:53:25 -08:00
2010-02-22 16:21:12 -08:00
2009-07-05 19:15:35 -07:00
2010-03-05 13:34:03 -08:00
2010-03-13 14:50:18 -08:00
2010-03-08 16:55:37 +01:00
2010-02-17 00:03:27 -08:00
2010-02-28 01:29:49 -08:00
2010-02-10 11:12:08 -08:00
2010-03-03 01:04:38 -08:00
2010-03-15 16:00:47 -07:00
2010-03-16 21:16:59 -07:00
2010-03-02 14:28:49 -05:00
2010-02-10 11:12:08 -08:00
2009-11-25 15:14:13 -08:00
2010-02-09 11:44:44 -08:00
2010-03-13 14:50:18 -08:00
2010-03-13 14:50:18 -08:00
2010-03-16 14:15:45 -07:00
2010-02-18 14:12:06 -08:00
2009-07-12 12:22:34 -07:00
2010-02-18 14:30:18 -08:00
2010-02-25 23:26:21 -08:00
2010-03-05 13:34:03 -08:00
2010-03-03 01:04:37 -08:00
2009-12-11 15:07:57 -08:00
2009-07-15 08:53:39 -07:00
2009-07-12 20:11:14 -07:00
2009-12-17 10:51:40 -05:00
2010-01-17 19:16:02 -08:00
b98ba52f96