diff mbox

[net-next] net: skb_segment() should preserve backpressure

Message ID 1414431051.16231.23.camel@edumazet-glaptop2.roam.corp.google.com
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Eric Dumazet Oct. 27, 2014, 5:30 p.m. UTC 
From: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>

This patch generalizes commit d6a4a1041176 ("tcp: GSO should be TSQ
friendly") to protocols using skb_set_owner_w()

TCP uses its own destructor (tcp_wfree) and needs a more complex scheme
as explained in commit 6ff50cd55545 ("tcp: gso: do not generate out of
order packets")

This allows UDP sockets using UFO to get proper backpressure,
thus avoiding qdisc drops and excessive cpu usage.

Here are performance test results (macvlan on vlan):

- Before
# netperf -t UDP_STREAM ...
Socket  Message  Elapsed      Messages
Size    Size     Time         Okay Errors   Throughput
bytes   bytes    secs            #      #   10^6bits/sec

212992   65507   60.00      144096 1224195    1258.56
212992           60.00          51              0.45

Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:        all      0.23      0.00     25.26      0.08      0.00     74.43

- After
# netperf -t UDP_STREAM ...
Socket  Message  Elapsed      Messages
Size    Size     Time         Okay Errors   Throughput
bytes   bytes    secs            #      #   10^6bits/sec

212992   65507   60.00      109593      0     957.20
212992           60.00      109593            957.20

Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:        all      0.18      0.00      8.38      0.02      0.00     91.43

[edumazet] Rewrote patch and changelog.

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/core/skbuff.c |   10 ++++++++++
 1 file changed, 10 insertions(+)



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Toshiaki Makita Oct. 28, 2014, 12:15 p.m. UTC | #1 
On 2014/10/28 2:30, Eric Dumazet wrote:
> From: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
> 
> This patch generalizes commit d6a4a1041176 ("tcp: GSO should be TSQ
> friendly") to protocols using skb_set_owner_w()
> 
> TCP uses its own destructor (tcp_wfree) and needs a more complex scheme
> as explained in commit 6ff50cd55545 ("tcp: gso: do not generate out of
> order packets")
> 
> This allows UDP sockets using UFO to get proper backpressure,
> thus avoiding qdisc drops and excessive cpu usage.
> 
...
> [edumazet] Rewrote patch and changelog.

This looks better in performance.
I tested this patch and couldn't find any problem.

Thank you for your reviewing and suggesting,
Toshiaki Makita

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller Oct. 29, 2014, 6:47 p.m. UTC | #2 
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Mon, 27 Oct 2014 10:30:51 -0700

> From: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
> 
> This patch generalizes commit d6a4a1041176 ("tcp: GSO should be TSQ
> friendly") to protocols using skb_set_owner_w()
> 
> TCP uses its own destructor (tcp_wfree) and needs a more complex scheme
> as explained in commit 6ff50cd55545 ("tcp: gso: do not generate out of
> order packets")
> 
> This allows UDP sockets using UFO to get proper backpressure,
> thus avoiding qdisc drops and excessive cpu usage.
> 
> Here are performance test results (macvlan on vlan):
 ...
> [edumazet] Rewrote patch and changelog.
> 
> Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
> Signed-off-by: Eric Dumazet <edumazet@google.com>

Applied, thanks everyone.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index c16615bfb61edd2a1dae9ef7935a3153d78dc4df..e48e5c02e877d9a9389ea54f0e015ba041d3f2a7 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3099,6 +3099,16 @@  perform_csum_check:
 	 * (see validate_xmit_skb_list() for example)
 	 */
 	segs->prev = tail;
+
+	/* Following permits correct backpressure, for protocols
+	 * using skb_set_owner_w().
+	 * Idea is to tranfert ownership from head_skb to last segment.
+	 */
+	if (head_skb->destructor == sock_wfree) {
+		swap(tail->truesize, head_skb->truesize);
+		swap(tail->destructor, head_skb->destructor);
+		swap(tail->sk, head_skb->sk);
+	}
 	return segs;
 
 err: