Upgrade to Linux 6.6.18 #190
Conversation
New mapping: ttyS0: BMC serial console ttyS1-4: Serial connection to node 1-4 Closes turing-machines#181
|
(As someone who just finally set up their TP2.4 today after getting 32GB RK1s, and is already fighting with network arch problems,) I'd like to put in a word in favor of hoping this big jump opens up #46 (since its past the additions in 5.16 you mentioned here ) I'm not currently seeing any defconfig or device tree to get the hooks into the RTL8370MB ; but ... maybe I'm missing something obvious I could see a case for relegating the switch driver to an unloaded-by-default-module until e.g. hooks in bmcd are there to make it useful in userspace but ditto can imagine the presence of the driver might be worth having in whatever round of manual QA is happening here |
| &mdio { | ||
| rtl8201f: ethernet-phy@0 { | ||
| compatible = "ethernet-phy-ieee802.3-c22"; | ||
| reg = <0>; | ||
| reset-gpios = <&pio 4 10 GPIO_ACTIVE_LOW>; /* PE10 */ | ||
| }; | ||
| }; |
There was a problem hiding this comment.
I don't really speak kernel or hardware or understand what pins are wired to which controller
But in a probably vein attempt to be helpful with what I'm requesting I surmise (from:
- https://www.kernel.org/doc/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt ,
- https://www.kernel.org/doc/Documentation/devicetree/bindings/net/dsa/dsa-port.yaml ,
- https://github.com/devicetree-org/devicetree-specification/blob/2697c66a59b9177f8ff7813c355cc7aa05f6e40a/source/chapter4-device-bindings.rst#L344
[and not really helped by
https://cdn.discordapp.com/attachments/1107160366888267866/1109336549222514768/2208101101_Realtek-Semicon-RTL8370N-VB-CG_C2803243.pdf?ex=6610cf2b&is=65fe5a2b&hm=11f81f2a2c1a05af62fd449af17536ae4da02a5561e55c91b92cc20f7352df7f&norhttps://cdn.discordapp.com/attachments/1107160366888267866/1109336548916350996/Realtek_Unmanaged_Switch_ProgrammingGuide.pdf?ex=6610cf2b&is=65fe5a2b&hm=974a489d57215bd6db3ac82a94ed9705397a8b239d7bf2998d246e0ce42bcefe& ) the broad shape of things might be something like
&switch {
compatible = "realtek,rtl8370";
mdc-gpios = <&pio ? ?? GPIO_ACTIVE_??>;
mdio-gpios = <&pio 4 ?? GPIO_ACTIVE_???>;
reset-gpios = <&pio ? ?? GPIO_ACTIVE_??>;
switch_intc: interrupt-controller {
interrupt-parent = <&pio ?>;
interrupts = <1 IRQ_TYPE_LEVEL_LOW>;
interrupt-controller;
#address-cells = <0>;
#interrupt-cells = <1>;
};
ports {
#address-cells = <1>;
#size-cells = <0>;
reg = <0>;
port@0 {
reg = <0>;
label = "node1-ethernet0";
ethernet = <&nodes "node1-en">;
phy-handle = <&gmac1>
phy-mode = "rgmii-txd"; ?
tx-internal-delay-ps = <300>; ?
};
port@1 {
reg = <1>;
label = "node1-ethernet1";
ethernet = <&nodes "node1-en">;
phy-handle = <&gmac2>
};
port@2 {
reg = <2>;
label = "node2-ethernet0";
ethernet = <&nodes "node2-en">;
phy-handle = <&gmac1>
};
port@3 {
reg = <3>;
label = "node2-ethernet1";
ethernet = <&nodes "node2-en">;
phy-handle = <&gmac2>
};
port@4 {
reg = <4>;
label = "node3-ethernet0";
ethernet = <&nodes "node3-en">;
phy-handle = <&gmac1>
};
port@5 {
reg = <5>;
label = "node3-ethernet1";
ethernet = <&nodes "node3-en">;
phy-handle = <&gmac2>
};
port@6 {
reg = <6>;
label = "node4-ethernet0";
ethernet = <&nodes "node4-en">;
phy-handle = <&gmac1>
};
port@7 {
reg = <7>;
label = "node4-ethernet1";
ethernet = <&nodes "node4-en">;
phy-handle = <&gmac2>
};
};
mdio {
compatible = "realtek,smi-mdio";
#address-cells = <1>;
#size-cells = <0>;
gmac1: phy@0 {
reg = <0>;
interrupt-parent = <&switch_intc>;
interrupts = <0>;
};
gmac2: phy@1 {
reg = <1>;
interrupt-parent = <&switch_intc>;
interrupts = <1>;
};
};
};
(but of course the bmc itself is clearly inserted into the loop somewhere... so ... seems doubtful [maybe the 8 ports are 4 nodes + bmc + 2 physical ports + ? ; I kinda thought 8 + 2 would be to ten but those pdfs whomever posted on discord suggest only numbered up to 7 ]
There was a problem hiding this comment.
though honestly I'm already lost at the one line I thought I got, cause:
https://www.kernel.org/doc/Documentation/devicetree/bindings/net/dsa/realtek.yaml (which is more authoritative?) still says to use "realtek,rtl8365mb"
even though https://www.kernel.org/doc/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt lists "realtek,rtl8370
[ apparently cause https://patchwork.kernel.org/project/netdevbpf/patch/20220418233558.13541-1-luizluca@gmail.com/ ? ]
There was a problem hiding this comment.
there is also confusing re-factor of "dsa" specific things vs more generic ethernet switching device tree stuff ; ... I give up [ but probably one of the the more complicated things I guessed at in the middle and then deleted was closer ; but still wrong]
|
Initial test results from my TPiv2.4:
|
|
Successfully tested using
|
bmcdPlease cherry-pick/rebase instead of merge.