[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <0C7D298A-86E7-427F-B45B-25C2F663D2FA@zytor.com>
Date: Sun, 30 Mar 2025 15:44:19 -0700
From: "H. Peter Anvin" <hpa@...or.com>
To: David Laight <david.laight.linux@...il.com>,
Uros Bizjak <ubizjak@...il.com>
CC: Ingo Molnar <mingo@...nel.org>, x86@...nel.org,
linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
Borislav Petkov <bp@...en8.de>,
Dave Hansen <dave.hansen@...ux.intel.com>
Subject: Re: [PATCH -tip 2/2] x86/hweight: Use POPCNT when available with X86_NATIVE_CPU option
On March 30, 2025 11:15:44 AM PDT, David Laight <david.laight.linux@...il.com> wrote:
>On Sun, 30 Mar 2025 18:07:30 +0200
>Uros Bizjak <ubizjak@...il.com> wrote:
>
>...
>> The kernel is currently compiled with -mno-red-zone, gcc-15 introduces
>> special "redzone" clobber to disable red-zone in the function that
>> includes asm() when/if the kernel starts using redzone.
>
>I really don't understand the point of a stack 'red-zone'.
>In any function that contains memory writes the cost of a
>'sub $n,%rsp' at the top and corresponding add at the bottom
>is surely noise.
>
> David.
Perhaps you should check out the performance numbers instead.
Powered by blists - more mailing lists