Re: [PATCH v5 0/3] x86,fs/resctrl,arm_mpam: Factor MBA parse-time conversion to be per-arch

[Reinette Chatre <reinette.chatre@intel.com>]

Hi Ben,

On 9/7/26 3:44 AM, Ben Horgan wrote:
> Hi Boris, Reinette,
> 
> On 04/09/2026 18:36, Borislav Petkov wrote:
>> On Thu, Sep 03, 2026 at 08:37:51PM -0700, Borislav Petkov wrote:
>>> Lemme have a look.
>>
>> Sashiko has comments:
>>
>> https://sashiko.dev/#/patchset/20260709093111.367851-1-ben.horgan%40arm.com
>>
>> Please address them: if valid, fix them, if not, explain why they don't need
>> to be fixed.
>>
>> Thx.
>>
> 
> Regarding the first Sashiko comment.
> 
> The documentation change in patch 3 includes the sentence:
> 
> "
> The allocated bandwidth percentage is rounded up to the next control step available on the
> hardware.
> "
> 
> Sashiko rightly objects to the "rounded up". For MPAM systems this can be up or down and we test in
> test_mpam_resctrl.c:test_percent_to_max_rounding() that it isn't too biased to rounding one way or
> the other.
> 
> How about replacing "rounded up" with "rounded up or down" so that the sentence becomes:
> 
> "
> The allocated bandwidth percentage is rounded up or down to the next control step available on the
> hardware.
> "
> 
> What do you think? Should I respin the series with this alteration?
The change looks good to me since Intel always rounds up. By itself "rounded" may thus be deceiving.
One concern is that "next" seems to accompany "round up" to reflect forward/upward movement.
How about also replacing "next" with "nearest" or "closest"?

I think a respin would be most convenient to ensure the intended changes are made.

Reinette