Re: [PATCH v6 11/18] fs/resctrl: Introduce kmode_cpus/kmode_cpus_list per rdtgroup

[Babu Moger <babu.moger@amd.com>]

Hi Reinette,

On 10/19/26 11:34, Reinette Chatre wrote:
> Hi Babu,
> 
> On 10/16/26 2:33 PM, Moger, Babu wrote:
>> Hi Reinette,
>>
>>
>> On 10/15/2026 10:31 AM, Reinette Chatre wrote:
>>> Hi Babu,
>>>
>>> On 10/1/26 8:10 AM, Babu Moger wrote:
>>>> The RESCTRL_GLOBAL_ENABLE_PER_CPU kernel mode allows tasks to use different
>>>> allocation and monitoring associations when transitioning from user mode to
>>>
>>> Could "different allocation and monitoring associations" just be "different
>>> CLOSID and RMID" (in places where this is what it refers to)? Being specific
>>> and consistent with terms would make this so much easier to understand.
>>
>> I avoided using CLOSID and RMID terms in filesystem specific code. If you are ok with that then I will use these terms.
> 
> Why?
> 
> Internally resctrl filesystem has a CLOSID assigned to a control group, RMID
> assigned to a monitor group. When a user assigns a resource group to the global
> kernel mode the kernel mode is implicitly assigned a CLOSID and/or RMID. These are
> all part of the resctrl filesystem design. What is the benefit to avoid these specific
> terms when explaining the implementation details?
> 
> An important distinction is that in this feature the CLOSID and RMID values originate
> from a CTRL_MON or MON group assigned by user space. Any architecture where there is a
> dependency between CLOSID and RMID will already have accounted for that during the
> creation of the resource group. This is something that should be clear from the context
> when referring to CLOSID and RMID.

That helps. Thanks for the explanation.

> 
> The resctrl fs code even uses CLOSID and RMID in the API used to interact with the
> architecture:
> 	void resctrl_arch_configure_kmode_global(const struct cpumask *cpu_mask,
> 						 u32 closid, bool assign_ctrl,
> 						 u32 rmid, bool assign_mon,
> 						 bool enable);
> 
> The only place where we should avoid using CLOSID and RMID is in anything
> exposed to user space since CLOSID and RMID has different meanings on different
> architectures.

Yeah, I need to come up with some generic terminology for this interface 
in resctrl.rst. I'll try to address it in the next revision.

Thanks,
Babu