[v7,0/4] Utilize cpu-type for CPU matching

v7:
- Moved the CPU table cleanup patch out of the series, it does not relate
  to cpu-type detection. Will send it separately.

v6:
- Use the recently added macro VULNBL_INTEL_STEPS() in the affected CPU
  list cleanup patch.
- Rebased to tip/x86/cpu.

This series is now the remaining last 5 patches from the original
series. Below are previously merged patches:

- upstream:
  45239ba39a52 ("x86/cpu: Add CPU type to struct cpuinfo_topology")

- tip:
  x86/cpu: Prepend 0x to the hex values in cpu_debug_show()
  cpufreq: intel_pstate: Avoid SMP calls to get cpu-type
  perf/x86/intel: Use cache cpu-type for hybrid PMU selection
  x86/cpu: Remove get_this_hybrid_cpu_*()

v5: https://lore.kernel.org/r/20241211-add-cpu-type-v5-0-2ae010f50370@linux.intel.com
- Replace usages of get_this_hybrid_cpu_native_id() with its cached value
  in CPU topology structure.
- s/x86_match_cpu_type/x86_match_vendor_cpu_type and add vendor checks
  in the function.
- Some cleanups in intel_pstate.c.
- Collected tags.
- Rebased to v6.13-rc1.

v4: https://lore.kernel.org/r/20240930-add-cpu-type-v4-0-104892b7ab5f@linux.intel.com
- Series doesn't apply to upstream cleanly anymore, rebased to v6.12-rc1,
  resolved the merge conflict in files:
	arch/x86/include/asm/cpu_device_id.h
	arch/x86/kernel/cpu/common.c
- Remove redundant "hw_" in intel_hw_native_model_id().

v3: https://lore.kernel.org/r/20240815-add-cpu-type-v3-0-234162352057@linux.intel.com
- Add a patch to prepend "0x" to hex values in cpu_debug_show() (Borislav).
- Add support to to also get Intel Native Model ID (Dapeng).
- Keep similar models together in the affected processor list (Josh).
- Add a comparison of .init.data in commit message for cpu_vuln_blacklist
  before and after decluttering patch for affected processor list (Josh).
- Drop the debugfs file reference in the commit message (Borislav).
- s/cpu_type/hw_cpu_type/ (Borislav).
- Add a union for hw_cpu_type for easy retrieval of intel_core_type and
  intel_native_model_id.
- Updated commit messages, collected tags.
- Rebased to v6.11-rc3.

Note, I will be off work till Tuesday, will reply to any comments then.

v2: https://lore.kernel.org/r/20240627-add-cpu-type-v2-0-f927bde83ad0@linux.intel.com
- Move CPU-type to the end of the CPU topology structure (Andrew).
- Use c->cpuid_level instead of cpuid_eax(0) (Andrew).
- Move CPU-type enum out of ifdef CONFIG_NUMA (kernel test robot).
- Rename cpu_type to hw_cpu_type (Borislav).
- Explain replacing get_this_hybrid_cpu_type() with topology_hw_cpu_type()
  in the commit message (Dave).
- Fix the alignment in cpu_vuln_whitelist (Andrew).
- Add the obj compare note in the commit message (Dave/Tony).
- s/X86_CPU_TYPE_INTEL_ATOM/ATOM/ in cpu_vuln_whitelist (Dave).

v1: https://lore.kernel.org/r/20240617-add-cpu-type-v1-0-b88998c01e76@linux.intel.com

Hi,

This series adds support for CPU-type (CPUID.1A.EAX[31-24] on Intel) to
differentiate between hybrid variants P+E, P-only, E-only that share the
same Family/Model/Stepping. One of the use case for CPU-type is the
affected CPU table for CPU vulnerabilities, which can now use the CPU-type
to filter the unaffected variants.

* Patch 1 fixes a minor formatting issue in cpu_debug_show().

* Patch 2 adds hardware cpu-type to CPU topology structure and introduces
  topology_hw_cpu_type().

* Patch 3-5 replaces usages of get_this_hybrid_cpu_type() with
  topology_hw_cpu_type().

* Patch 6-8 Updates CPU-matching infrastructure to use CPU-type.

* Patch 9 cleans up the affected CPU list.

* Patch 10 uses the CPU-type to exclude P-only parts from the RFDS affected
  list.

Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
---
Pawan Gupta (4):
      x86/cpu: Name CPU matching macro more generically (and shorten)
      x86/cpu: Add cpu_type to struct x86_cpu_id
      x86/cpu: Update x86_match_cpu() to also use cpu-type
      x86/rfds: Exclude P-only parts from the RFDS affected list

 .../admin-guide/hw-vuln/reg-file-data-sampling.rst |   8 --
 arch/x86/include/asm/cpu_device_id.h               | 115 ++++++++-------------
 arch/x86/kernel/cpu/common.c                       |   7 +-
 arch/x86/kernel/cpu/match.c                        |  30 ++++++
 include/linux/mod_devicetable.h                    |   2 +
 5 files changed, 80 insertions(+), 82 deletions(-)
---
base-commit: 4f2a0b765c9731d2fa94e209ee9ae0e96b280f17
change-id: 20240617-add-cpu-type-4d5e47efc117

Best regards,

On Thu, Mar 06, 2025 at 06:18:20PM -0800, Pawan Gupta wrote:
> diff --git a/include/linux/mod_devicetable.h b/include/linux/mod_devicetable.h
> index d67614f7b7f1..18e996acb49a 100644
> --- a/include/linux/mod_devicetable.h
> +++ b/include/linux/mod_devicetable.h
> @@ -692,6 +692,7 @@ struct x86_cpu_id {
>  	__u16 feature;	/* bit index */
>  	/* Solely for kernel-internal use: DO NOT EXPORT to userspace! */
>  	__u16 flags;
> +	__u8  cpu_type;

The struct is called "x86_cpu_id" and all its members describe a CPU. There's
no need to have more "cpu_" redundancy in the member names - just call that
"type".  It is clear that it is about a CPU's type.

The macro names having "CPU" - X86_CPU_TYPE_ANY - are fine I guess.

Thx.