Re: [PATCH v14 27/32] x86,fs/resctrl: Auto assign/unassign counters on mkdir

[Reinette Chatre <reinette.chatre@intel.com>]

Hi Babu,

On 6/13/25 2:05 PM, Babu Moger wrote:
> Resctrl provides a user-configurable option mbm_assign_on_mkdir that
> determines if a counter will automatically be assigned to an RMID, event
> pair when its associated monitor group is created via mkdir.
> 
> Enable mbm_assign_on_mkdir by default and automatically assign or unassign
> counters when a resctrl group is created or deleted.

This is a bit confusing since I do not think mbm_assign_on_mkdir has *anything*
to do with unassign of counters. Counters are always (irrespective of mbm_assign_on_mkdir)
unassigned when a resctrl group is deleted, no?

The subject also does not seem accurate since there is no unassign on
mkdir.

> 
> By default, each group requires two counters: one for the MBM total event
> and one for the MBM local event.
> 
> If the counters are exhausted, the kernel will log the error message
> "Unable to allocate counter in domain" in
> /sys/fs/resctrl/info/last_cmd_status when a new group is created and the
> counter assignment will fail. However, the creation of a group should not
> fail due to assignment failures. Users have the flexibility to modify the
> assignments at a later time.
> 
> Signed-off-by: Babu Moger <babu.moger@amd.com>
> ---

...

> ---
>  arch/x86/kernel/cpu/resctrl/monitor.c |  1 +
>  fs/resctrl/rdtgroup.c                 | 71 ++++++++++++++++++++++++++-
>  2 files changed, 70 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c
> index ee0aa741cf6c..053f516a8e67 100644
> --- a/arch/x86/kernel/cpu/resctrl/monitor.c
> +++ b/arch/x86/kernel/cpu/resctrl/monitor.c
> @@ -429,6 +429,7 @@ int __init rdt_get_mon_l3_config(struct rdt_resource *r)
>  		r->mon.mbm_cntr_assignable = true;
>  		cpuid_count(0x80000020, 5, &eax, &ebx, &ecx, &edx);
>  		r->mon.num_mbm_cntrs = (ebx & GENMASK(15, 0)) + 1;
> +		r->mon.mbm_assign_on_mkdir = true;
>  	}
>  
>  	r->mon_capable = true;
> diff --git a/fs/resctrl/rdtgroup.c b/fs/resctrl/rdtgroup.c
> index bf5fd46bd455..128a9db339f3 100644
> --- a/fs/resctrl/rdtgroup.c
> +++ b/fs/resctrl/rdtgroup.c
> @@ -2945,6 +2945,55 @@ static void schemata_list_destroy(void)
>  	}
>  }
>  
> +/**
> + * rdtgroup_assign_cntrs() - Assign counters to MBM events. Called when
> + *			     a new group is created.
> + * If "mbm_event" mode is enabled, counters are automatically assigned.

"counters are automatically assigned" -> "counters should be automatically assigned
if the "mbm_assign_on_mkdir" is set"?

> + * Each group can accommodate two counters: one for the total event and
> + * one for the local event. Assignments may fail due to the limited number
> + * of counters. However, it is not necessary to fail the group creation
> + * and thus no failure is returned. Users have the option to modify the
> + * counter assignments after the group has been created.
> + */
> +static void rdtgroup_assign_cntrs(struct rdtgroup *rdtgrp)
> +{
> +	struct rdt_resource *r = resctrl_arch_get_resource(RDT_RESOURCE_L3);
> +
> +	if (!r->mon_capable)
> +		return;
> +
> +	if (resctrl_arch_mbm_cntr_assign_enabled(r) && !r->mon.mbm_assign_on_mkdir)
> +		return;

This check is not clear to me. It looks to me as though counter assignment
will be attempted if !resctrl_arch_mbm_cntr_assign_enabled(r)? Perhaps
something like:
	if (!r->mon_capable || !resctrl_arch_mbm_cntr_assign_enabled(r) ||
	    !r->mon.mbm_assign_on_mkdir)
		return;

> +
> +	if (resctrl_is_mon_event_enabled(QOS_L3_MBM_TOTAL_EVENT_ID))
> +		resctrl_assign_cntr_event(r, NULL, rdtgrp,
> +					  &mon_event_all[QOS_L3_MBM_TOTAL_EVENT_ID]);

Switching the namespace like this is confusing to me. rdtgroup_assign_cntrs()
has prefix rdtgroup_ to indicate it operates on a resource group. It is confusing
when it switches namespace to call resctrl_assign_cntr_event() that actually assigns
a specific event to a resource group. I think this will be easier to follow if:
resctrl_assign_cntr_event() -> rdtgroup_assign_cntr_event()

> +
> +	if (resctrl_is_mon_event_enabled(QOS_L3_MBM_LOCAL_EVENT_ID))
> +		resctrl_assign_cntr_event(r, NULL, rdtgrp,
> +					  &mon_event_all[QOS_L3_MBM_LOCAL_EVENT_ID]);
> +}
> +
> +/*
> + * rdtgroup_unassign_cntrs() - Unassign the counters associated with MBM events.
> + *			       Called when a group is deleted.
> + */
> +static void rdtgroup_unassign_cntrs(struct rdtgroup *rdtgrp)
> +{
> +	struct rdt_resource *r = resctrl_arch_get_resource(RDT_RESOURCE_L3);
> +
> +	if (!r->mon_capable || !resctrl_arch_mbm_cntr_assign_enabled(r))
> +		return;
> +
> +	if (resctrl_is_mon_event_enabled(QOS_L3_MBM_TOTAL_EVENT_ID))
> +		resctrl_unassign_cntr_event(r, NULL, rdtgrp,
> +					    &mon_event_all[QOS_L3_MBM_TOTAL_EVENT_ID]);

same here, I think this will be easier to follow when namespace is
consistent:
resctrl_unassign_cntr_event() -> rdtgroup_unassign_cntr_event()


Also, the struct rdt_resource parameter should not be needed when
struct mon_evt is provided and resource can be obtained from mon_evt::rid.

> +
> +	if (resctrl_is_mon_event_enabled(QOS_L3_MBM_LOCAL_EVENT_ID))
> +		resctrl_unassign_cntr_event(r, NULL, rdtgrp,
> +					    &mon_event_all[QOS_L3_MBM_LOCAL_EVENT_ID]);
> +}
> +
>  static int rdt_get_tree(struct fs_context *fc)
>  {
>  	struct rdt_fs_context *ctx = rdt_fc2context(fc);


Reinette