[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20241209130620.71276-1-qiuxu.zhuo@intel.com>
Date: Mon, 9 Dec 2024 21:06:20 +0800
From: Qiuxu Zhuo <qiuxu.zhuo@...el.com>
To: dave.hansen@...ux.intel.com
Cc: bp@...en8.de,
kan.liang@...ux.intel.com,
linux-kernel@...r.kernel.org,
mingo@...nel.org,
pawan.kumar.gupta@...ux.intel.com,
peterz@...radead.org,
tglx@...utronix.de,
tony.luck@...el.com,
x86@...nel.org,
qiuxu.zhuo@...el.com
Subject: Re: [PATCH 0/5] x86/cpu: Remove duplicate microcode version matching infrastructure
Hi Dave,
> From: Dave Hansen <dave.hansen@...ux.intel.com>
> [...]
> Changes from RFC:
> * Convert stepping match helpers to always take a range and never
> take a raw stepping bitmap. - Ingo
>
> --
>
> x86 has generic CPU matching infrastructure. This lets you build
> tables of CPUs with some property. It's mostly used for enumerating
> model-specific features, but it is quite a bit more flexible than
> that. In includes a facility to match steppings and microcode
> versions. This generic infrastructure is built around 'struct
> x86_cpu_id'.
>
> There is a less generic, parallel CPU matching facility built around
> 'struct x86_cpu_desc'. It is used only for matching specific microcode
> revisions. All of the 'struct x86_cpu_desc' users can be converted to
> 'struct x86_cpu_id'.
>
> Do that conversion then remove the 'struct x86_cpu_desc'
> infrastructure.
>
> Testing or acks would be much appreciated!
I tested this series + v6.13-rc2 on both Intel Cascade Lake server and Sapphire Rapids server.
- Both systems booted successfully.
- Both {skx,i10nm}_edac drivers[1][2] worked well.
[1] Covered patch2's drivers/edac/skx_base.c file.
[2] Covered patch2's drivers/edac/i10nm_base.c file.
Feel free to add:
Tested-by: Qiuxu Zhuo <qiuxu.zhuo@...el.com>
Powered by blists - more mailing lists