| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <trinity-ed99385e-32eb-4be6-bbff-91211a38b8fb-1677263479464@3c-app-gmx-bap70>
Date: Fri, 24 Feb 2023 19:31:19 +0100
From: Frank Wunderlich <frank-w@...lic-files.de>
To: Vladimir Oltean <olteanv@...il.com>
Cc: Arınç ÜNAL <arinc.unal@...nc9.com>,
netdev <netdev@...r.kernel.org>, erkin.bozoglu@...ont.com,
Andrew Lunn <andrew@...n.ch>,
Florian Fainelli <f.fainelli@...il.com>,
Felix Fietkau <nbd@....name>, John Crispin <john@...ozen.org>,
Mark Lee <Mark-MC.Lee@...iatek.com>,
Lorenzo Bianconi <lorenzo@...nel.org>,
Matthias Brugger <matthias.bgg@...il.com>,
Landen Chao <Landen.Chao@...iatek.com>,
Sean Wang <sean.wang@...iatek.com>,
DENG Qingfang <dqfext@...il.com>
Subject: Aw: Re: Choose a default DSA CPU port
result is same
root@...-r2:~# iperf3 -c 192.168.0.21
Connecting to host 192.168.0.21, port 5201
[ 5] local 192.168.0.11 port 48882 connected to 192.168.0.21 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 75.3 MBytes 632 Mbits/sec 0 396 KBytes
[ 5] 1.00-2.00 sec 74.3 MBytes 623 Mbits/sec 0 396 KBytes
[ 5] 2.00-3.00 sec 74.6 MBytes 625 Mbits/sec 0 396 KBytes
[ 5] 3.00-4.00 sec 73.9 MBytes 620 Mbits/sec 0 396 KBytes
[ 5] 4.00-5.00 sec 74.6 MBytes 626 Mbits/sec 0 396 KBytes
[ 5] 5.00-6.00 sec 74.4 MBytes 624 Mbits/sec 0 396 KBytes
[ 5] 6.00-7.00 sec 74.4 MBytes 624 Mbits/sec 0 396 KBytes
[ 5] 7.00-8.00 sec 74.4 MBytes 624 Mbits/sec 0 396 KBytes
[ 5] 8.00-9.00 sec 73.9 MBytes 620 Mbits/sec 0 396 KBytes
[ 5] 9.00-10.00 sec 74.6 MBytes 626 Mbits/sec 0 396 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 745 MBytes 625 Mbits/sec 0 sender
[ 5] 0.00-10.05 sec 744 MBytes 621 Mbits/sec receiver
iperf Done.
root@...-r2:~# ethtool -S eth0 | grep -v ': 0'
NIC statistics:
tx_bytes: 819999267
tx_packets: 538815
rx_bytes: 18338089
rx_packets: 261984
p06_TxUnicast: 261974
p06_TxMulticast: 10
p06_TxPktSz65To127: 261983
p06_TxPktSz256To511: 1
p06_TxBytes: 19386025
p06_RxFiltering: 13
p06_RxUnicast: 538783
p06_RxMulticast: 31
p06_RxBroadcast: 1
p06_RxPktSz64: 3
p06_RxPktSz65To127: 47
p06_RxPktSz128To255: 2
p06_RxPktSz256To511: 1
p06_RxPktSz512To1023: 2
p06_RxPktSz1024ToMax: 538760
p06_RxBytes: 819999267
just to verify no pause is set (which is working without p5 enabled)
root@...-r2:~# ls /sys/firmware/devicetree/base/ethernet\@1b100000/mdio-bus/switch\@1f/ports/port\@6/fixed-link/
full-duplex name speed
root@...-r2:~# ls /sys/firmware/devicetree/base/ethernet\@1b100000/mac\@0/fixed-link/
full-duplex name speed
as speed is lower than gmac speed i guess there is no need for pause frames, but yes they can slow down...
mhm...tried again with port5 disabled (pause enabled again) and got same result, so this seems not to be the cause as i thought (but have now other patches in like core-clock dropped and RX fix).
regards Frank
> Gesendet: Freitag, 24. Februar 2023 um 19:13 Uhr
> Von: "Vladimir Oltean" <olteanv@...il.com>
> An: "Frank Wunderlich" <frank-w@...lic-files.de>
> Cc: "Arınç ÜNAL" <arinc.unal@...nc9.com>, "netdev" <netdev@...r.kernel.org>, erkin.bozoglu@...ont.com, "Andrew Lunn" <andrew@...n.ch>, "Florian Fainelli" <f.fainelli@...il.com>, "Felix Fietkau" <nbd@....name>, "John Crispin" <john@...ozen.org>, "Mark Lee" <Mark-MC.Lee@...iatek.com>, "Lorenzo Bianconi" <lorenzo@...nel.org>, "Matthias Brugger" <matthias.bgg@...il.com>, "Landen Chao" <Landen.Chao@...iatek.com>, "Sean Wang" <sean.wang@...iatek.com>, "DENG Qingfang" <dqfext@...il.com>
> Betreff: Re: Choose a default DSA CPU port
>
> On Fri, Feb 24, 2023 at 07:07:23PM +0100, Frank Wunderlich wrote:
> > root@...-r2:~# ethtool -S eth0 | grep -v ': 0'
> > NIC statistics:
> > tx_bytes: 1643364546
> > tx_packets: 1121377
> > rx_bytes: 1270088499
> > rx_packets: 1338400
> > p06_TxUnicast: 1338274
> > p06_TxMulticast: 120
> > p06_TxBroadcast: 6
> > p06_TxPktSz65To127: 525948
> > p06_TxPktSz128To255: 5
> > p06_TxPktSz256To511: 16
> > p06_TxPktSz512To1023: 4
> > p06_Tx1024ToMax: 812427
> > p06_TxBytes: 1275442099
> > p06_RxFiltering: 16
> > p06_RxUnicast: 1121339
> > p06_RxMulticast: 37
> > p06_RxBroadcast: 1
> > p06_RxPktSz64: 3
> > p06_RxPktSz65To127: 43757
> > p06_RxPktSz128To255: 3
> > p06_RxPktSz256To511: 3
> > p06_RxPktSz1024ToMax: 1077611
> > p06_RxBytes: 1643364546
>
> Looking at the drivers, I see pause frames aren't counted in ethtool -S,
> so we wouldn't know this. However the slowdown *is* lossless, so the
> hypothesis is still not disproved.
>
> Could you please test after removing the "pause" property from the
> switch's port@6 device tree node?
>
Powered by blists - more mailing lists