Re: [PATCH v1 3/9] x86/resctrl: Add resctrl_mbm_flush_cpu() to collect CPUs' MBM events

[Reinette Chatre]

Hi Peter,

On 6/1/2023 7:45 AM, Peter Newman wrote:
> Hi Reinette,
> 
> On Thu, May 11, 2023 at 11:37 PM Reinette Chatre
> <reinette.chatre@xxxxxxxxx> wrote:
>> On 4/21/2023 7:17 AM, Peter Newman wrote:
>>> +     /* Count bandwidth after the first successful counter read. */
>>> +     if (counter->initialized) {
>>> +             /* Assume that mbm_update() will prevent double-overflows. */
>>> +             if (val != counter->prev_bytes)
>>> +                     atomic64_add(val - counter->prev_bytes,
>>> +                                  &m->soft_rmid_bytes);
>>> +     } else {
>>> +             counter->initialized = true;
>>> +     }
>>> +
>>> +     counter->prev_bytes = val;
>>
>> I notice a lot of similarities between the above and the software controller,
>> see mbm_bw_count().
> 
> I see the "a=now(); a-b; b=a;" and the not handling overflow parts
> being similar, but the use of the initialized flag seems quite
> different from delta_comp.
> 
> Also mbm_state is on the arch-independent side and the new code is
> going to the arch-dependent side, so it wouldn't be convenient to try
> to use the mbm_bw structures for this.
> 
> From this, I don't think trying to reuse this is worth it unless you
> have other suggestions.

At this time I am staring at mbm_state->prev_bw_bytes and mbm_soft_counter->prev_bytes
and concerned about how much confusion this would generate. Considering the
pending changes to data structures I hope this would be clear then.

Reinette