[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <72330b7a-5159-47ce-81e1-28458e4a5719@intel.com>
Date: Wed, 7 May 2025 20:32:18 -0700
From: Reinette Chatre <reinette.chatre@...el.com>
To: Tony Luck <tony.luck@...el.com>, Fenghua Yu <fenghuay@...dia.com>, "Maciej
Wieczor-Retman" <maciej.wieczor-retman@...el.com>, Peter Newman
<peternewman@...gle.com>, James Morse <james.morse@....com>, Babu Moger
<babu.moger@....com>, Drew Fustini <dfustini@...libre.com>, Dave Martin
<Dave.Martin@....com>, Anil Keshavamurthy <anil.s.keshavamurthy@...el.com>,
Chen Yu <yu.c.chen@...el.com>
CC: <x86@...nel.org>, <linux-kernel@...r.kernel.org>,
<patches@...ts.linux.dev>
Subject: Re: [PATCH v4 05/31] fs/resctrl: Set up Kconfig options for telemetry
events
Hi Tony,
On 4/28/25 5:33 PM, Tony Luck wrote:
> Intel RMID based telemetry events are counted by each CPU core
> and then aggregated by one or more per-socket micro controllers.
> Enumeration support is provided by the Intel PMT subsystem.
>
> N.B. Patches for the Intel PMT system are still in progress.
> They will define an INTEL_PMT_DISCOVERY Kconfig symbol that
> will be one of the dependencies. This is commented out for
> now. Final version will include this dependency.
>
> arch/x86 selects this option based on:
Portion of changelog seems to be missing ... what does "this option" refer to?
>
> X86_64: Counter registers are in MMIO space. There is no readq()
> function on 32-bit. Emulation is possible with readl(), but there
> are races. Running 32-bit kernels on systems that support this
> feature seems pointless.
This seems to be the primary dependency that requires the use of
a Kconfig symbol.
>
> CPU_SUP_INTEL: It is an Intel specific feature.
While this is an Intel specific feature it looks to me as though
intel_pmt_get_regions_by_feature() will return -ENODEV if it is not supported.
I think there is benefit to have this code compiled as much as
possible. An alternative to this CPU_SUP_INTEL dependency may be to add a
vendor check in the pre-mount callback to match the other Intel vs AMD,
but this may not be necessary?
>
> Signed-off-by: Tony Luck <tony.luck@...el.com>
> ---
> arch/x86/Kconfig | 1 +
> drivers/platform/x86/intel/pmt/Kconfig | 7 +++++++
> 2 files changed, 8 insertions(+)
>
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 5a09acf41c8e..19107fdb4264 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -508,6 +508,7 @@ config X86_CPU_RESCTRL
> bool "x86 CPU resource control support"
> depends on X86 && (CPU_SUP_INTEL || CPU_SUP_AMD)
> depends on MISC_FILESYSTEMS
> + select INTEL_AET_RESCTRL if (X86_64 && CPU_SUP_INTEL)
> select ARCH_HAS_CPU_RESCTRL
> select RESCTRL_FS
> select RESCTRL_FS_PSEUDO_LOCK
> diff --git a/drivers/platform/x86/intel/pmt/Kconfig b/drivers/platform/x86/intel/pmt/Kconfig
> index e916fc966221..3a8ce39d1004 100644
> --- a/drivers/platform/x86/intel/pmt/Kconfig
> +++ b/drivers/platform/x86/intel/pmt/Kconfig
> @@ -38,3 +38,10 @@ config INTEL_PMT_CRASHLOG
>
> To compile this driver as a module, choose M here: the module
> will be called intel_pmt_crashlog.
> +
> +config INTEL_AET_RESCTRL
> + depends on INTEL_PMT_TELEMETRY # && INTEL_PMT_DISCOVERY
> + bool
> + help
> + Architecture config should "select" this option to enable
> + support for RMID telemetry events in the resctrl file system.
Reinette
Powered by blists - more mailing lists