Re: [PATCH v12 12/25] arm,x86,fs/resctrl: Allocate maximum needed rmid_ptrs[]

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

Hi Babu,

Thank you for taking a look.

On 9/21/26 10:45 AM, Babu Moger wrote:
> On 9/16/26 18:13, Tony Luck wrote:

...

>> @@ -45,6 +45,9 @@ static DEFINE_MUTEX(domain_list_lock);
>>    */
>>   DEFINE_PER_CPU(struct resctrl_pqr_state, pqr_state);
>>   +/* Number of RMIDS values that can be written to IA32_PQR_ASSOC.RMID */
>> +static u32 pqr_assoc_num_rmid;
>> +
>>   static void mba_wrmsr_intel(struct msr_param *m);
>>   static void cat_wrmsr(struct msr_param *m);
>>   static void mba_wrmsr_amd(struct msr_param *m);
>> @@ -124,6 +127,28 @@ u32 resctrl_arch_system_num_rmid_idx(void)
>>       return num_rmids == U32_MAX ? 0 : num_rmids;
>>   }
>>   +/**
>> + * resctrl_arch_system_max_rmid_idx - Largest possible number of RMIDs
>> + *
>> + * Return: Maximum possible number of RMIDs used for boot time allocations.
>> + */
>> +u32 resctrl_arch_system_max_rmid_idx(void)
>> +{
>> +    struct rdt_resource *r = &rdt_resources_all[RDT_RESOURCE_L3].r_resctrl;
>> +    u32 num_rmid = pqr_assoc_num_rmid;
> 
> Why not  use "cpuid_ebx(0xf) + 1" instead of global variable?
> 
Tony did so in v11 but I proposed it to be changed [1]. This series adds
multiple consumers of this value so it seems more efficient to enumerate
it once and re-use that value when needed. Locating this one-time enumeration
with all the other monitoring information enumerated from hardware in
get_rdt_mon_resources() also has the benefit of keeping all monitoring data
enumeration together.

Reinette

[1] https://lore.kernel.org/lkml/0ef770ec-ac65-4a80-b676-a8bd92caea88@intel.com/