Re: [patches] Re: [PATCH v9 03/12] dt-bindings: RISC-V CPU Bindings

From: Palmer Dabbelt
Date: Mon Nov 20 2017 - 14:45:41 EST


On Sun, 19 Nov 2017 23:35:28 PST (-0800), j.neuschaefer@xxxxxxx wrote:
Hi Palmer,

On Thu, Oct 05, 2017 at 11:16:33AM +0100, Mark Rutland wrote:
[...]
I would *strongly* recommend that from day one, you determine the SMP
bringup mechanism via an enable-method property, and document the
contract with FW/bootloader somewhere in the kernel tree.

Sorry, I forgot about this. I've prepared a patch.

Somewhat, but not quite related: Please consider making the availability
of the Supervisor Binary Interface explicit in the devicetree.
I understand that the general plan is to make the SBI a mandatory
feature of every RISC-V system capable of running Linux, but I do want
to explore the possibility of running without run-time resident firmware
at some point in the future. Thus it would be nice if the devicetree
would indicate the presence of the SBI from the start, to avoid having
to invent a way to express its *absence* later on.

It could look something like this (modelled after qcom,scm):

/ {
firmware {
sbi {
compatible = "riscv,sbi";
};
};
};

This topic may warrant some discussion, because other people may have
different opinions, and there hasn't been a discussion about it, AFAICS.

I don't think there's any penalty to putting it in the device tree, I'll send a patch.