[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0ade78e7-7a3c-459d-bf1c-f0bb1f24baee@intel.com>
Date: Fri, 7 Mar 2025 07:03:43 -0800
From: Dave Hansen <dave.hansen@...el.com>
To: "Ahmed S. Darwish" <darwi@...utronix.de>, Borislav Petkov <bp@...en8.de>,
Ingo Molnar <mingo@...hat.com>, Dave Hansen <dave.hansen@...ux.intel.com>
Cc: Thomas Gleixner <tglx@...utronix.de>,
Andrew Cooper <andrew.cooper3@...rix.com>, "H. Peter Anvin" <hpa@...or.com>,
x86@...nel.org, John Ogness <john.ogness@...utronix.de>,
x86-cpuid@...ts.linux.dev, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 07/12] tools/x86/kcpuid: Add rudimentary CPU vendor
detection
On 3/6/25 12:49, Ahmed S. Darwish wrote:
> The kcpuid CSV file will soon be updated with CPUID indices that are only
> valid for certain CPU vendors, such as Centaur or Transmeta. Thus,
> introduce rudimentary x86 vendor detection to kcpuid.
Do we really need the vendor detection? For example, look at the end of
cpuid(1)'s output:
# cpuid -1 --raw
...
0x20000000 0x00: eax=0x00000000 ebx=0x00000000 ecx=0x00000000
edx=0x00000000
...
0x80860000 0x00: eax=0x00000000 ebx=0x00000000 ecx=0x00000000
edx=0x00000000
0xc0000000 0x00: eax=0x00000000 ebx=0x00000000 ecx=0x00000000
edx=0x00000000
It seems to just blindly poke at all of the CPUID regions. There are
only a handful of these and there's no hard in poking at them other
than an extra couple of executions of CPUID.
Powered by blists - more mailing lists