Re: [PATCH v8 2/7] x86/resctrl: Compute memory bandwidth for all supported events

From: Reinette Chatre
Date: Tue Nov 12 2024 - 14:26:28 EST


Hi Tony,

On 10/29/24 10:28 AM, Tony Luck wrote:
> Computing the bandwidth for an event is cheap, and only done once
> per second. Doing so simplifies switching between events and allows
> choosing different events per ctrl_mon group.
>

If I understand correctly this changelog only describes the first and the
last hunks of this patch. Further, it is the last hunk that introduces the
duplicate code that prompts the refactoring in next patch "x86/resctrl:
Refactor mbm_update()". Inserting the duplicate code and then refactoring it
out does not seem necessary. What if the second hunk of this patch is extracted
into its own patch and the refactoring squashed with what remains?
In the cover letter your motivation for the separate refactor was "to make it
easier to review", but I wonder if separating the unrelated code in second
hunk would make this separate refactor unnecessary while remaining easy to
review?

Reinette