netdev - Re: [RFC PATCH net-next 0/4] Remove the "dsa_to

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210809193105.ijzvvkdwvtgkihun@skbuf>
Date:   Mon, 9 Aug 2021 22:31:05 +0300
From:   Vladimir Oltean <olteanv@...il.com>
To:     Vladimir Oltean <vladimir.oltean@....com>
Cc:     netdev@...r.kernel.org, Jakub Kicinski <kuba@...nel.org>,
        "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>,
        Andrew Lunn <andrew@...n.ch>,
        Vivien Didelot <vivien.didelot@...il.com>,
        Kurt Kanzenbach <kurt@...utronix.de>,
        Woojung Huh <woojung.huh@...rochip.com>,
        UNGLinuxDriver@...rochip.com, Sean Wang <sean.wang@...iatek.com>,
        Landen Chao <Landen.Chao@...iatek.com>,
        DENG Qingfang <dqfext@...il.com>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Claudiu Manoil <claudiu.manoil@....com>,
        Alexandre Belloni <alexandre.belloni@...tlin.com>,
        George McCollister <george.mccollister@...il.com>
Subject: Re: [RFC PATCH net-next 0/4] Remove the "dsa_to_port in a loop"
 antipattern

FWIW, I also have a small test: on my Turris MOX with 25 mv88e6xxx user
ports, I have a script which sets up the bridge for my network:

-----------------------------[ cut here ]-----------------------------
# cat /etc/init.d/S50bridge
#!/bin/bash

ip link del br0
ip link add br0 type bridge vlan_filtering 1
for eth in lan1 lan2 lan3 lan4 lan5 lan6 lan7 lan8 lan9 lan10 lan11 lan12 lan13 lan14 lan15 lan16 lan17 lan18 lan19 lan20 lan21 lan22 lan23 lan24 sfp; do
        ip link set ${eth} master br0
done

# For TFTP, the server is connected to sfp
# bridge vlan add dev sfp vid 100
bridge vlan add dev lan19 vid 100

# lan1-lan24 are the for board access to TFTP. Use VLAN 100.
# FIXME: I removed lan19 from this list and now use it for TFTP server
for eth in lan1 lan2 lan3 lan4 lan5 lan6 lan7 lan8 lan9 lan10 lan11 lan12 lan13 lan14 lan15 lan16 lan17 lan18 lan20 lan21 lan22 lan23 lan24; do
        bridge vlan del dev ${eth} vid 1
        bridge vlan add dev ${eth} vid 100 pvid untagged
done

ip link set br0 up
-----------------------------[ cut here ]-----------------------------

Before this series:

time /etc/init.d/S50bridge
real    0m8.809s
user    0m0.253s
sys     0m2.614s

time ip link del br0
real    0m4.509s
user    0m0.006s
sys     0m1.184s

After:

time /etc/init.d/S50bridge
real    0m8.270s
user    0m0.199s
sys     0m2.468s

time ip link del br0
real    0m3.964s
user    0m0.000s
sys     0m1.024s

So there is a (small, but still visible) improvement - note that
mv88e6xxx is heavily limited by MDIO access anyway, and that would be
the predominant latency.

Also, I noticed that I forgot an unused "int port" variable in
felix_vsc9959.c variable, that I'm sure the kernel test robot will
complain about. So this is just RFC for now.