| 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
| ||
|
Date: Wed, 13 Dec 2023 12:55:26 +0800
From: Richard Tresidder <rtresidd@...ctromag.com.au>
To: Serge Semin <fancer.lancer@...il.com>
Cc: Andrew Lunn <andrew@...n.ch>, Jakub Kicinski <kuba@...nel.org>,
vinschen@...hat.com, netdev@...r.kernel.org
Subject: Re: STMMAC Ethernet Driver support
On 12/12/2023 5:27 pm, Serge Semin wrote:
> On Tue, Dec 12, 2023 at 11:57:22AM +0800, Richard Tresidder wrote:
>> <font face="monospace">Richard Tresidder</font>
>>
>>
>> On 12/12/2023 12:16 am, Andrew Lunn wrote:
>>>> We use the SOC's internal STMMAC interface to connect to a Marvel switch IC
>>>> and expose each port individually using vlan, I'd forgot that part.
>>>> It's an 88E6352-xx-TFJ2I000 device utilising the 'marvell,mv88e6085'
>>>> compatible driver in drivers\net\dsa\mv88e6xxx
>>> Its odd you need VLANs. Each port should already be exposed to the
>>> host as netdev interfaces. That is what DSA does.
>>>
>>> Andrew
>> Hi Andrew
>> I'll read further on that one as this is the first time I've had to dig
>> into this side of the system.
>> It had always "just worked".
>> The ports show up in an 'ip l' list in the same style as a vlan with an @
>> symbol, naming isn't quite vlan style though.
>> That in concert with the fact this 'vlan_feature' line broke things has
>> possibly distorted my view of how they're propagated.
>> It's a rather trimmed down busybox image, so I'm missing some tools I'd
>> usually use to examine stuff.
>>
>> This is the config in the dts
>> **************************************
>> //------------------------------------------------------------------------------
>> // connected to dsa network switch
>> &gmac1 {
>> clock-names = "stmmaceth", "clk_ptp_ref";
> Just a side note irrelevant to the topic. You might be interested to
> know STMMAC driver expects to have the PTP reference clock source
> passed with the "ptp_ref" name, not "clk_ptp_ref".
>
> -Serge(y)
Cheers for that Surge
The dts for this device hasn't been updated in a while!
It was based on a 4.7 series kernel
I vaguely remember coming across this change on clock naming a while
back when dealing with a pll device driver but looks like we missed this
one!
Good catch
Cheers Richard
>> clocks = <&emac1_clk &hps_eosc1>;
>> f2h_ptp_ref_clk;
>> fixed-link {
>> speed = <1000>;
>> full-duplex;
>> };
>> };
>>
>> //------------------------------------------------------------------------------
>> &mdio1 {
>> #address-cells = <1>;
>> #size-cells = <0>;
>>
>> switch0: switch0@0 {
>> compatible = "marvell,mv88e6085";
>> #address-cells = <1>;
>> reg = <0>;
>> //reset-gpios = <&pio_a0 2 GPIO_ACTIVE_LOW>;
>>
>> dsa,member = <0 0>;
>>
>> ports {
>> #address-cells = <1>;
>> #size-cells = <0>;
>>
>> port@2 {
>> reg = <2>;
>> label = "lan1";
>> phy-handle = <&switch1phy2>;
>> };
>>
>> port@3 {
>> reg = <3>;
>> label = "lan2";
>> phy-handle = <&switch1phy3>;
>> };
>>
>> port@4 {
>> reg = <4>;
>> label = "lan3";
>> phy-handle = <&switch1phy4>;
>> };
>>
>> port@5 {
>> reg = <5>;
>> label = "wifi";
>> fixed-link {
>> speed = <100>;
>> full-duplex;
>> };
>> };
>>
>> port@6 {
>> reg = <6>;
>> label = "cpu";
>> ethernet = <&gmac1>;
>> fixed-link {
>> speed = <1000>;
>> full-duplex;
>> };
>> };
>>
>> };
>>
>> mdio {
>> #address-cells = <1>;
>> #size-cells = <0>;
>> switch1phy2: switch1phy2@2 {
>> reg = <2>;
>> marvell,reg-init = <0 0x10 0 0x0200>; // Sense only on Rx Energy
>> Detect, no FLPs sents
>> };
>> switch1phy3: switch1phy3@3 {
>> reg = <3>;
>> marvell,reg-init = <0 0x10 0 0x0200>; // Sense only on Rx Energy
>> Detect, no FLPs sents
>> };
>> switch1phy4: switch1phy4@4 {
>> reg = <4>;
>> marvell,reg-init = <0 0x10 0 0x0200>; // Sense only on Rx Energy
>> Detect, no FLPs sents
>> };
>> };
>>
>> };
>> };
>> *************************************************
>>
>> This is how they appear using 'ip l'
>> The @ symbol got me as I've usually associated this with vlan's in my day to
>> day networking.
>> But the config files don't configure them as vlans.
>> *************************************************
>> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000
>> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
>> 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1508 qdisc mq qlen 1000
>> link/ether 04:91:62:cf:4f:06 brd ff:ff:ff:ff:ff:ff
>> 3: sit0@...E: <NOARP> mtu 1480 qdisc noop qlen 1000
>> link/sit 0.0.0.0 brd 0.0.0.0
>> 4: lan1@...0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
>> qlen 1000
>> link/ether 04:91:62:cf:4f:06 brd ff:ff:ff:ff:ff:ff
>> 5: lan2@...0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue qlen
>> 1000
>> link/ether 04:91:62:cf:4f:06 brd ff:ff:ff:ff:ff:ff
>> 6: lan3@...0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
>> qlen 1000
>> link/ether 04:91:62:cf:4f:06 brd ff:ff:ff:ff:ff:ff
>> 7: wifi@...0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue qlen
>> 1000
>> link/ether 04:91:62:cf:4f:06 brd ff:ff:ff:ff:ff:ff
>>
>> *************************************************
>>
>> And these are the systemd config files
>> No vlan mentioned in there..
>> *************************************************
>> /etc/systemd/network/eth0.network
>> [Match]
>> Name=eth0
>>
>> [Network]
>>
>> /etc/systemd/network/lan.network
>> [Match]
>> Name=lan*
>>
>> [Network]
>> DHCP=yes
>> BindCarrier=eth0
>> LinkLocalAddressing=yes
>>
>> [DHCP]
>> UseDomains=yes
>> ClientIdentifier=mac
>>
>> *************************************************
>>
>> Cheers
>> Richard
>>
>
Powered by blists - more mailing lists