Re: [PATCH v6 12/21] x86/resctrl: Calculate bandwidth from the previous __mon_event_count() chunks

[James Morse]

Hi Hao Xin,

On 07/09/2022 07:47, haoxin wrote:
> 在 2022/9/2 下午11:48, James Morse 写道:
>> mbm_bw_count() is only called by the mbm_handle_overflow() worker once a
>> second. It reads the hardware register, calculates the bandwidth and
>> updates m->prev_bw_msr which is used to hold the previous hardware register
>> value.
>>
>> Operating directly on hardware register values makes it difficult to make
>> this code architecture independent, so that it can be moved to /fs/,
>> making the mba_sc feature something resctrl supports with no additional
>> support from the architecture.
>> Prior to calling mbm_bw_count(), mbm_update() reads from the same hardware
>> register using __mon_event_count().
>>
>> Change mbm_bw_count() to use the current chunks value most recently saved
>> by __mon_event_count(). This removes an extra call to __rmid_read().
>> Instead of using m->prev_msr to calculate the number of chunks seen,
>> use the rr->val that was updated by __mon_event_count(). This removes an
>> extra call to mbm_overflow_count() and get_corrected_mbm_count().
>> Calculating bandwidth like this means mbm_bw_count() no longer operates
>> on hardware register values directly.

>> diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c
>> index 3e69386cfe00..2d81b6cd9632 100644
>> --- a/arch/x86/kernel/cpu/resctrl/monitor.c
>> +++ b/arch/x86/kernel/cpu/resctrl/monitor.c

>> @@ -516,10 +521,12 @@ static void mbm_update(struct rdt_resource *r, struct rdt_domain
>> *d, int rmid)
>>        */
>>       if (is_mbm_total_enabled()) {
>>           rr.evtid = QOS_L3_MBM_TOTAL_EVENT_ID;>> +        rr.val = 0;

> In mbm_update,  there no use the rr.val, so there no need to initialize ？

>>           __mon_event_count(rmid, &rr);
>>       }
>>       if (is_mbm_local_enabled()) {
>>           rr.evtid = QOS_L3_MBM_LOCAL_EVENT_ID;
>> +        rr.val = 0;

> ditto.

>>           __mon_event_count(rmid, &rr);
>>             /*

No, but this just leaves that problem for someone else to discover the hard way! I think
its fair for the compiler to complain that addition on an uninitialised field is a bug.

I'd prefer to keep this as it is on the principle of 'least surprise'.


Thanks,

James