提交 126f760c 编写于 作者: H Håkon Bugge 提交者: David S. Miller

rds: Fix incorrect statistics counting

In rds_send_xmit() there is logic to batch the sends. However, if
another thread has acquired the lock and has incremented the send_gen,
it is considered a race and we yield. The code incrementing the
s_send_lock_queue_raced statistics counter did not count this event
correctly.

This commit counts the race condition correctly.

Changes from v1:
- Removed check for *someone_on_xmit()*
- Fixed incorrect indentation
Signed-off-by: NHåkon Bugge <haakon.bugge@oracle.com>
Reviewed-by: NKnut Omang <knut.omang@oracle.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
上级 eef5a7cc
...@@ -428,14 +428,18 @@ int rds_send_xmit(struct rds_conn_path *cp) ...@@ -428,14 +428,18 @@ int rds_send_xmit(struct rds_conn_path *cp)
* some work and we will skip our goto * some work and we will skip our goto
*/ */
if (ret == 0) { if (ret == 0) {
bool raced;
smp_mb(); smp_mb();
raced = send_gen != READ_ONCE(cp->cp_send_gen);
if ((test_bit(0, &conn->c_map_queued) || if ((test_bit(0, &conn->c_map_queued) ||
!list_empty(&cp->cp_send_queue)) && !list_empty(&cp->cp_send_queue)) && !raced) {
send_gen == READ_ONCE(cp->cp_send_gen)) {
rds_stats_inc(s_send_lock_queue_raced);
if (batch_count < send_batch_count) if (batch_count < send_batch_count)
goto restart; goto restart;
queue_delayed_work(rds_wq, &cp->cp_send_w, 1); queue_delayed_work(rds_wq, &cp->cp_send_w, 1);
} else if (raced) {
rds_stats_inc(s_send_lock_queue_raced);
} }
} }
out: out:
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册