Re: [PATCH v10 12/14] arm_mpam: detect and enable MPAM-Fb PCC support

[Andre Przywara <andre.przywara@arm.com>]

Hi Sudeep,

On 9/24/26 11:48, Sudeep Holla wrote:
> On Mon, Sep 21, 2026 at 05:28:27PM +0100, Ben Horgan wrote:
>> Hi Andre,
>>
>> On 21/09/2026 17:09, Andre Przywara wrote:
>>> Hi Ben,
>>>
>>> On 9/11/26 17:09, Ben Horgan wrote:
>>>> Hi Andre,
>>>>
> 
> [...]
> 
>>>>
>>>> This looks like the change earlier in the series to
>>>> drivers/acpi/arm64/mpam.c makes no functional effect unless the proposed
>>>> spec changes are accepted; the default is msc->id. I'd just move the
>>>
>>> Yes, but it reduces churn, and more importantly, is more true to the
>>> MPAM-Fb spec, which says:"The MSC identifiers are sequential and start from
>>> 0 for each agent which uses this protocol." So this requires that we have
>>> per-channel IDs for the MSCs, that's why I think it's justified to
>>> introduce this already now, even when it's somewhat redundant as the code
>>> stands now, as we always assign the system-wide ID to it.
>>>
>>> I can of course move that out (and have that already on top of v9,
>>> internally), but wanted to check this with you.
>>
>> I don't really mind. If Lorenzo/Sudeep are happy with doing the preparation
>> work in drivers/acpi/arm64/mpam.c then I'm ok with it. It's currently a
>> non-functional change but does make the expected path forward slightly
>> clearer.
>>
> 
> Sure, I am okay to have it added here if we plan to use it in the future.
> We can always rip it off if the spec changes its direction or gets deemed
> as broken 😉.
Many thanks, will keep it in then.

Cheers,
Andre