Re: [PATCH 2/2] powercap/intel_rapl: Fix the energy-pkg event for AMD CPUs

["Zhang, Rui" <rui.zhang@intel.com>]

On Fri, 2024-07-19 at 09:25 +0000, Dhananjay Ugwekar wrote:
> After commit ("x86/cpu/topology: Add support for the AMD 0x80000026
> leaf"),
> on AMD processors that support extended CPUID leaf 0x80000026, the
> topology_logical_die_id() macros, no longer returns package id,
> instead it
> returns the CCD (Core Complex Die) id. This leads to the energy-pkg
> event scope to be modified to CCD instead of package.
> 
> For more historical context, please refer to commit 32fb480e0a2c
> ("powercap/intel_rapl: Support multi-die/package"), which initially
> changed
> the RAPL scope from package to die for all systems, as Intel systems
> with Die enumeration have RAPL scope as die, and those without die
> enumeration are not affected. So, all systems(Intel, AMD, Hygon),
> worked
> correctly with topology_logical_die_id() until recently, but this
> changed
> after the "0x80000026 leaf" commit mentioned above.
> 
> Replacing topology_logical_die_id() with
> topology_physical_package_id()
> conditionally only for AMD and Hygon fixes the energy-pkg event.
> 
> On an AMD 2 socket 8 CCD Zen5 server:
> 
> Before:
> 
> linux$ ls /sys/class/powercap/
> intel-rapl      intel-rapl:1:0  intel-rapl:3:0  intel-rapl:5:0
> intel-rapl:7:0  intel-rapl:9:0  intel-rapl:b:0  intel-rapl:d:0
> intel-rapl:f:0  intel-rapl:0    intel-rapl:2    intel-rapl:4
> intel-rapl:6    intel-rapl:8    intel-rapl:a    intel-rapl:c
> intel-rapl:e    intel-rapl:0:0  intel-rapl:2:0  intel-rapl:4:0
> intel-rapl:6:0  intel-rapl:8:0  intel-rapl:a:0  intel-rapl:c:0
> intel-rapl:e:0  intel-rapl:1    intel-rapl:3    intel-rapl:5
> intel-rapl:7    intel-rapl:9    intel-rapl:b    intel-rapl:d
> intel-rapl:f
> 
> After:
> 
> linux$ ls /sys/class/powercap/
> intel-rapl  intel-rapl:0  intel-rapl:0:0  intel-rapl:1  intel-
> rapl:1:0
> 
> Only one sysfs entry per-event per-package is created after this
> change.
> 
> Fixes: 63edbaa48a57 ("x86/cpu/topology: Add support for the AMD
> 0x80000026 leaf")
> Reported-by: Michael Larabel <michael@michaellarabel.com>
> Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>

For the future Intel multi-die system that I know, it still has
package-scope RAPL, but this is done with TPMI RAPL interface.

The TPMI RAPL driver invokes these APIs with "id == pkg_id" and
"id_is_cpu == false", so no need to make rapl_pmu_is_pkg_scope()
returns true for those Intel systems.

The patch LGTM.

Reviewed-by: Zhang Rui <rui.zhang@intel.com>

thanks,
rui
> ---
>  drivers/powercap/intel_rapl_common.c | 20 +++++++++++++++++---
>  1 file changed, 17 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/powercap/intel_rapl_common.c
> b/drivers/powercap/intel_rapl_common.c
> index 3cffa6c79538..2f24ca764408 100644
> --- a/drivers/powercap/intel_rapl_common.c
> +++ b/drivers/powercap/intel_rapl_common.c
> @@ -2128,6 +2128,18 @@ void rapl_remove_package(struct rapl_package
> *rp)
>  }
>  EXPORT_SYMBOL_GPL(rapl_remove_package);
>  
> +/*
> + * Intel systems that enumerate DIE domain have RAPL domains
> implemented
> + * per-die, however, the same is not true for AMD and Hygon
> processors
> + * where RAPL domains for PKG energy are in-fact per-PKG. Since
> + * logical_die_id is same as logical_package_id in absence of DIE
> + * enumeration, use topology_logical_die_id() on Intel systems and
> + * topology_logical_package_id() on AMD and Hygon systems.
> + */
> +#define rapl_pmu_is_pkg_scope()                                \
> +       (boot_cpu_data.x86_vendor == X86_VENDOR_AMD ||  \
> +        boot_cpu_data.x86_vendor == X86_VENDOR_HYGON)
> +
>  /* caller to ensure CPU hotplug lock is held */
>  struct rapl_package *rapl_find_package_domain_cpuslocked(int id,
> struct rapl_if_priv *priv,
>                                                          bool
> id_is_cpu)
> @@ -2136,7 +2148,8 @@ struct rapl_package
> *rapl_find_package_domain_cpuslocked(int id, struct rapl_if_
>         int uid;
>  
>         if (id_is_cpu)
> -               uid = topology_logical_die_id(id);
> +               uid = rapl_pmu_is_pkg_scope() ?
> +                     topology_physical_package_id(id) :
> topology_logical_die_id(id);
>         else
>                 uid = id;
>  
> @@ -2168,9 +2181,10 @@ struct rapl_package
> *rapl_add_package_cpuslocked(int id, struct rapl_if_priv *pr
>                 return ERR_PTR(-ENOMEM);
>  
>         if (id_is_cpu) {
> -               rp->id = topology_logical_die_id(id);
> +               rp->id = rapl_pmu_is_pkg_scope() ?
> +                        topology_physical_package_id(id) :
> topology_logical_die_id(id);
>                 rp->lead_cpu = id;
> -               if (topology_max_dies_per_package() > 1)
> +               if (!rapl_pmu_is_pkg_scope() &&
> topology_max_dies_per_package() > 1)
>                         snprintf(rp->name,
> PACKAGE_DOMAIN_NAME_LENGTH, "package-%d-die-%d",
>                                  topology_physical_package_id(id),
> topology_die_id(id));
>                 else