Re: [PATCH 2/2] platform/x86/amd: hfi: Stand down when amd-pstate CAS is active

[Mario Limonciello <mario.limonciello@amd.com>]

On 11/15/26 08:16, Mario Limonciello wrote:
> 
> 
> On 11/14/26 09:03, Vishal Badole wrote:
>> From: Krishnamoorthi M <krishnamoorthi.m@amd.com>
>>
>> When amd-pstate CAS is active, it takes ownership of per-CPU capacity
>> values that the scheduler uses to distinguish performance and efficiency
>> cores. amd-hfi manages the same scheduler state through its ranking
>> mechanism. Running both simultaneously would cause them to overwrite
>> each other's settings, leading to unpredictable task placement on hybrid
>> platforms.
>>
>> Avoid loading amd-hfi if CAS is enabled.
>>
>> Signed-off-by: Krishnamoorthi M <krishnamoorthi.m@amd.com>
> 
> When a patch is authored by another person and you send it on their 
> behalf you need to add your own S-o-b as well.
> 
> Think of it like a chain of custody.
> 
>> ---
>>   drivers/platform/x86/amd/hfi/hfi.c | 8 ++++++++
>>   1 file changed, 8 insertions(+)
>>
>> diff --git a/drivers/platform/x86/amd/hfi/hfi.c b/drivers/platform/ 
>> x86/amd/hfi/hfi.c
>> index 86a7f4332c35..b6878614e2bc 100644
>> --- a/drivers/platform/x86/amd/hfi/hfi.c
>> +++ b/drivers/platform/x86/amd/hfi/hfi.c
>> @@ -29,6 +29,8 @@
>>   #include <asm/cpu_device_id.h>
>> +#include <linux/amd-pstate.h>
>> +
>>   #include <acpi/pcc.h>
>>   #include <acpi/cppc_acpi.h>
>> @@ -536,6 +538,12 @@ static int __init amd_hfi_init(void)
>>           !cpu_feature_enabled(X86_FEATURE_AMD_WORKLOAD_CLASS))
>>           return -ENODEV;
>> +    /* amd-pstate CAS owns heterogeneous core scheduling when 
>> enabled. */
>> +    if (amd_pstate_cas_active()) {
>> +        pr_debug("amd-pstate CAS is active, not loading amd-hfi\n");
>> +        return -ENODEV;
>> +    }
>> +

The load order happens to work; but AFAICT this is just by coincidence 
(link order).

If the link order was changed in the future this could break completely 
non-obviously.

I think it would be most ideal to add deferred probing support or 
forcing HFI into a later init call phase.

>>       ret = platform_driver_register(&amd_hfi_driver);
>>       if (ret)
>>           pr_err("failed to register HFI driver\n");
>