Re: [PATCH v16 00/34] x86,fs/resctrl: Support AMD Assignable Bandwidth Monitoring Counters (ABMC)

[Moger, Babu]

Hi Reinette,

On 7/30/2025 6:57 PM, Reinette Chatre wrote:
> Hi Babu,
>
> On 7/30/25 4:31 PM, Moger, Babu wrote:
> > Hi Reinette,,
> >
> > On 7/30/2025 2:47 PM, Reinette Chatre wrote:
> > > Hi Babu,
> > >
> > > On 7/25/25 11:29 AM, Babu Moger wrote:
> > > > i. Change the event configuration for mbm_local_bytes.
> > > >
> > > >      # echo "local_reads, local_non_temporal_writes, local_reads_slow_memory, remote_reads" >
> > > >      /sys/fs/resctrl/info/L3_MON/counter_configs/mbm_local_bytes/event_filter
> > > >
> > > >      # cat /sys/fs/resctrl/info/L3_MON/counter_configs/mbm_local_bytes/event_filter
> > > >      local_reads,local_non_temporal_writes,local_reads_slow_memory,remote_reads
> > >
> > > Above are some more "counter_configs" stragglers.
> >
> > Yea. Sure. Missed that.
> >
> > > Also, while considering our exchange in [1], I encountered quite a few functions doing
> > > counter management work for which I believe monitor.c would be more appropriate. Centralizing
> > > MBM counter management code to monitor.c was something that you planned for this version
> > > so I may be missing why you decided to keep some of these functions in rdtgroup.c? I
> > > highlighted these functions as I noticed them.
> >
> > I looked at them. Most of the functions you mentioned are directly referenced in res_common_files[] (show or write) and some of them are even named as rdtgroup_<>. So, was not sure about moving them.
>
> If you prefer a precedent you can compare with rdtgroup_schemata_write()/rdtgroup_schemata_show()
> that is directly referenced in res_common_files[] while the implementation can be found in
> ctrlmondata.c that is intended to contain the "Cache allocation code".
>
> I assumed we agreed on this since I specifically highlighted the topic of the handlers in [1] and you
> responded to referring to the handler event_filter_show() and mentioned that you plan to consider the
> others. This version thus looks different from what I thought we agreed on :/

Looks like I misunderstood few things here.  Will take care of it in 
next revision.

Thanks
Babu

> > Sure, I will move them one by one to monitor.c
>
> [1] https://lore.kernel.org/lkml/0fa9a12b-e900-4ceb-b59c-e653ec3db0ca@xxxxxxxxx/