* MT7623 Ethernet TX performance regression
@ 2020-07-23 15:26 David Woodhouse
2020-07-23 17:49 ` Aw: " Frank Wunderlich
0 siblings, 1 reply; 3+ messages in thread
From: David Woodhouse @ 2020-07-23 15:26 UTC (permalink / raw)
To: linux-mediatek
Cc: cioby.service, John Crispin, Mark Lee, Sean Wang, Felix Fietkau
[-- Attachment #1.1: Type: text/plain, Size: 2820 bytes --]
Hi,
In 5.4 and 5.8.0-rc6+ kernels I see awful TX performance from the
Ethernet on (at least) MT7623 boards. It's been reported on a Banana Pi
R2, as well as UniElec U7623.
Using iperf3, receiving is fine:
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 112 MBytes 942 Mbits/sec
[ 5] 1.00-2.00 sec 112 MBytes 941 Mbits/sec
[ 5] 2.00-3.00 sec 112 MBytes 941 Mbits/sec
[ 5] 3.00-4.00 sec 112 MBytes 941 Mbits/sec
[ 5] 4.00-5.00 sec 112 MBytes 941 Mbits/sec
[ 5] 5.00-6.00 sec 112 MBytes 941 Mbits/sec
[ 5] 6.00-7.00 sec 112 MBytes 941 Mbits/sec
[ 5] 7.00-8.00 sec 112 MBytes 941 Mbits/sec
[ 5] 8.00-9.00 sec 112 MBytes 941 Mbits/sec
[ 5] 9.00-10.00 sec 112 MBytes 941 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.10 GBytes 943 Mbits/sec 0 sender
[ 5] 0.00-10.00 sec 1.10 GBytes 941 Mbits/sec receiver
Transmit, on the other hand, is fairly much completely hosed:
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 212 KBytes 1.74 Mbits/sec 36 2.79 KBytes
[ 5] 1.00-2.00 sec 665 KBytes 5.45 Mbits/sec 66 1.39 KBytes
[ 5] 2.00-3.00 sec 0.00 Bytes 0.00 bits/sec 14 2.79 KBytes
[ 5] 3.00-4.00 sec 126 KBytes 1.03 Mbits/sec 10 2.79 KBytes
[ 5] 4.00-5.00 sec 0.00 Bytes 0.00 bits/sec 9 2.79 KBytes
[ 5] 5.00-6.00 sec 0.00 Bytes 0.00 bits/sec 7 1.39 KBytes
[ 5] 6.00-7.00 sec 251 KBytes 2.06 Mbits/sec 26 2.79 KBytes
[ 5] 7.00-8.00 sec 126 KBytes 1.03 Mbits/sec 22 1.39 KBytes
[ 5] 8.00-9.00 sec 0.00 Bytes 0.00 bits/sec 2 1.39 KBytes
[ 5] 9.00-10.00 sec 126 KBytes 1.03 Mbits/sec 15 2.79 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.47 MBytes 1.23 Mbits/sec 207 sender
[ 5] 0.00-10.00 sec 1.26 MBytes 1.06 Mbits/sec receiver
This is seen with the OpenWrt 5.4 kernel, and also with a clean current
upstream 5.8.0-rc6 (with just one patch from
https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=36105880ee5
without which it doesn't boot at all).
The previous kernel that I have running on these boards is 4.14, which
does give me close to 1Gb/s in both directions as it should.
[-- Attachment #1.2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5174 bytes --]
[-- Attachment #2: Type: text/plain, Size: 170 bytes --]
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek
^ permalink raw reply [flat|nested] 3+ messages in thread
* Aw: MT7623 Ethernet TX performance regression
2020-07-23 15:26 MT7623 Ethernet TX performance regression David Woodhouse
@ 2020-07-23 17:49 ` Frank Wunderlich
2020-07-23 18:35 ` David Woodhouse
0 siblings, 1 reply; 3+ messages in thread
From: Frank Wunderlich @ 2020-07-23 17:49 UTC (permalink / raw)
To: David Woodhouse
Cc: Sean Wang, cioby.service, linux-mediatek, John Crispin, Mark Lee,
Felix Fietkau
Hi,
could you try this? it is the 3rd part of rene's series which got not merged to mainline yet.
https://github.com/frank-w/BPI-R2-4.14/commit/ff2bb8719bc269a9feb696764159e2640fc64eed
to which client/server do you make the tests? i had retransmitts like you with ubuntu 20.4 default kernel (5.4.0-xx) and fixed it with self compiled 5.4.51
regards Frank
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Aw: MT7623 Ethernet TX performance regression
2020-07-23 17:49 ` Aw: " Frank Wunderlich
@ 2020-07-23 18:35 ` David Woodhouse
0 siblings, 0 replies; 3+ messages in thread
From: David Woodhouse @ 2020-07-23 18:35 UTC (permalink / raw)
To: Frank Wunderlich
Cc: Sean Wang, cioby.service, linux-mediatek, John Crispin, Mark Lee,
Felix Fietkau
[-- Attachment #1.1: Type: text/plain, Size: 4790 bytes --]
On Thu, 2020-07-23 at 19:49 +0200, Frank Wunderlich wrote:
> Hi,
>
> could you try this? it is the 3rd part of rene's series which got not
> merged to mainline yet.
>
> https://github.com/frank-w/BPI-R2-4.14/commit/ff2bb8719bc269a9feb696764159e2640fc64eed
Yes, that's it. Thanks!
That brings the performance on 5.8.0-rc6 back to just under line rate
1Gb/s in both directions for both IPv6 and Legacy IP.
On the OpenWrt 5.4 kernel it's still not quite that for IPv6 TX
(although it is for Legacy IP TX), but at least I can consider that a
"performance" problem instead of basic functionality :)
root@OpenWrt:/# iperf3 -c i7.infradead.org
Connecting to host i7.infradead.org, port 5201
[ 5] local 2001:8b0:10b:1::c2a port 52730 connected to 2001:8b0:10b:1:21e:67ff:fecb:7a92 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 109 MBytes 916 Mbits/sec 1357 72.5 KBytes
[ 5] 1.00-2.00 sec 109 MBytes 913 Mbits/sec 1203 69.7 KBytes
[ 5] 2.00-3.00 sec 107 MBytes 899 Mbits/sec 1300 76.7 KBytes
[ 5] 3.00-4.00 sec 108 MBytes 906 Mbits/sec 1514 71.1 KBytes
[ 5] 4.00-5.00 sec 108 MBytes 903 Mbits/sec 1518 62.8 KBytes
[ 5] 5.00-6.00 sec 108 MBytes 909 Mbits/sec 1402 68.3 KBytes
[ 5] 6.00-7.00 sec 108 MBytes 905 Mbits/sec 1325 83.7 KBytes
[ 5] 7.00-8.00 sec 108 MBytes 903 Mbits/sec 1348 72.5 KBytes
[ 5] 8.00-9.00 sec 107 MBytes 900 Mbits/sec 1481 76.7 KBytes
[ 5] 9.00-10.00 sec 108 MBytes 904 Mbits/sec 1300 73.9 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.05 GBytes 906 Mbits/sec 13748 sender
[ 5] 0.00-10.00 sec 1.05 GBytes 905 Mbits/sec receiver
iperf Done.
root@OpenWrt:/# iperf3 -c i7.infradead.org -R
Connecting to host i7.infradead.org, port 5201
Reverse mode, remote host i7.infradead.org is sending
[ 5] local 2001:8b0:10b:1::c2a port 52734 connected to 2001:8b0:10b:1:21e:67ff:fecb:7a92 port 5201
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 77.3 MBytes 648 Mbits/sec
[ 5] 1.00-2.00 sec 78.1 MBytes 655 Mbits/sec
[ 5] 2.00-3.00 sec 78.3 MBytes 656 Mbits/sec
[ 5] 3.00-4.00 sec 78.6 MBytes 659 Mbits/sec
[ 5] 4.00-5.00 sec 78.2 MBytes 656 Mbits/sec
[ 5] 5.00-6.00 sec 78.2 MBytes 656 Mbits/sec
[ 5] 6.00-7.00 sec 78.4 MBytes 658 Mbits/sec
[ 5] 7.00-8.00 sec 78.7 MBytes 661 Mbits/sec
[ 5] 8.00-9.00 sec 78.1 MBytes 655 Mbits/sec
[ 5] 9.00-10.00 sec 78.3 MBytes 657 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 784 MBytes 658 Mbits/sec 31 sender
[ 5] 0.00-10.00 sec 782 MBytes 656 Mbits/sec receiver
iperf Done.
root@OpenWrt:/# iperf3 -c i7.infradead.org -R
Connecting to host i7.infradead.org, port 5201
Reverse mode, remote host i7.infradead.org is sending
[ 5] local 2001:8b0:10b:1::c2a port 59138 connected to 2001:8b0:10b:1:21e:67ff:fecb:7a92 port 5201
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 103 MBytes 865 Mbits/sec
[ 5] 1.00-2.00 sec 105 MBytes 879 Mbits/sec
[ 5] 2.00-3.00 sec 105 MBytes 880 Mbits/sec
[ 5] 3.00-4.00 sec 104 MBytes 876 Mbits/sec
[ 5] 4.00-5.00 sec 105 MBytes 880 Mbits/sec
[ 5] 5.00-6.00 sec 105 MBytes 878 Mbits/sec
[ 5] 6.00-7.00 sec 105 MBytes 879 Mbits/sec
[ 5] 7.00-8.00 sec 105 MBytes 879 Mbits/sec
[ 5] 8.00-9.00 sec 105 MBytes 881 Mbits/sec
[ 5] 9.00-10.00 sec 105 MBytes 880 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.03 GBytes 881 Mbits/sec 81 sender
[ 5] 0.00-10.00 sec 1.02 GBytes 878 Mbits/sec receiver
iperf Done.
> to which client/server do you make the tests? i had retransmitts like
> you with ubuntu 20.4 default kernel (5.4.0-xx) and fixed it with self
> compiled 5.4.51
Just a standard x86_64 Fedora 32 box.
[-- Attachment #1.2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5174 bytes --]
[-- Attachment #2: Type: text/plain, Size: 170 bytes --]
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-07-23 18:35 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-07-23 15:26 MT7623 Ethernet TX performance regression David Woodhouse
2020-07-23 17:49 ` Aw: " Frank Wunderlich
2020-07-23 18:35 ` David Woodhouse
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.