Re: [RFC 2/4] arch-topology: add a default implementation of store_cpu_topology()

From: Conor.Dooley
Date: Fri Jul 08 2022 - 07:57:34 EST

On 08/07/2022 12:39, Sudeep Holla wrote:
> On Fri, Jul 08, 2022 at 12:03:41PM +0200, Greg KH wrote:
>> On Fri, Jul 08, 2022 at 10:47:10AM +0100, Sudeep Holla wrote:
>>> On Fri, Jul 08, 2022 at 11:28:19AM +0200, Geert Uytterhoeven wrote:
>>>> Hi Sudeep,
>>>> On Fri, Jul 8, 2022 at 11:22 AM Sudeep Holla <sudeep.holla@xxxxxxx> wrote:
>>>>> On Fri, Jul 08, 2022 at 08:35:57AM +0000, Conor.Dooley@xxxxxxxxxxxxx wrote:
>>>>>> If you're okay with patch 1/4, I'll resubmit it as a standalone v2.
>>>>> That would be great, thanks. You can most the code to move to generic from
>>>>> both arm64 and risc-v once we have this in v5.20-rc1
>>>> Why not ignore risc-v for now, and move the arm64 implementation to
>>>> the generic code for v5.20, so every arch will have it at once?
>>> We could but,
>>> 1. This arch_topology is new and has been going through lot of changes
>>> recently and having code there might make it difficult to backport
>>> changes that are required for RISC-V(my guess)
>> Worry about future issues in the future. Make it simple now as you know
>> what you are dealing with at the moment.
> Sure, I was just suggesting and expecting someone from RISC-V community or
> maintainers to make a call. As I said it is based on my understanding.
> hence I have mentioned as guess. So I am not against it as such.

I did a little bit of poking in the git history.
The last code touching the arm implementation was:
3102bc0e6ac7 ("arm64: topology: Stop using MPIDR for topology information")
on Fri Oct 2 12:01:41 2020 +0100

The introduction of arch-topology stuff to RISC-V was:
03f11f03dbfe ("RISC-V: Parse cpu topology during boot.")
on Thu Jun 27 12:53:00 2019 -0700

Backporting as far as v5.10 should be no real effort and I don't think
to v5.4 that should be meaninfully harder. If 3102bc0e6ac7 hasn't been
backported already, maybe it should be since it appears to have been
fixing a problem too.

Based on that, I think doing this the straightforward way in the first
place is a better idea.

I'll respin the series as:
patch 1: Move arm64 to the generic implementation
patch 2: Make RISC-V use the generic implementation