In function strp_data_ready(), it is useless to call queue_work if
the state of strparser is already paused. The state checking should
be done before calling queue_work. The change reduces the context
switches and improves the ktls-rx throughput by approx 20% (measured
on cortex-a53 based platform).
Signed-off-by: Vakul Garg <[email protected]>
---
net/strparser/strparser.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/net/strparser/strparser.c b/net/strparser/strparser.c
index 1a9695183599..373836615c57 100644
--- a/net/strparser/strparser.c
+++ b/net/strparser/strparser.c
@@ -392,7 +392,7 @@ static int strp_read_sock(struct strparser *strp)
/* Lower sock lock held */
void strp_data_ready(struct strparser *strp)
{
- if (unlikely(strp->stopped))
+ if (unlikely(strp->stopped) || strp->paused)
return;
/* This check is needed to synchronize with do_strp_work.
@@ -407,9 +407,6 @@ void strp_data_ready(struct strparser *strp)
return;
}
- if (strp->paused)
- return;
-
if (strp->need_bytes) {
if (strp_peek_len(strp) < strp->need_bytes)
return;
--
2.13.6
On 06/21/18 03:29 AM, Vakul Garg wrote:
> In function strp_data_ready(), it is useless to call queue_work if
> the state of strparser is already paused. The state checking should
> be done before calling queue_work. The change reduces the context
> switches and improves the ktls-rx throughput by approx 20% (measured
> on cortex-a53 based platform).
Nice find, works for me. Although current code isn't broken, just
slower, net-next would be fine.
Acked-by: Dave Watson <[email protected]>
From: Vakul Garg <[email protected]>
Date: Thu, 21 Jun 2018 03:29:49 +0530
> In function strp_data_ready(), it is useless to call queue_work if
> the state of strparser is already paused. The state checking should
> be done before calling queue_work. The change reduces the context
> switches and improves the ktls-rx throughput by approx 20% (measured
> on cortex-a53 based platform).
>
> Signed-off-by: Vakul Garg <[email protected]>
Applied, thank you.