linux-kernel - Re: [PATCH v4 1/2] x86, lib: Add WBNOINVD helper functions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <f98160b0-4f8b-41ab-b555-8e9de83c8552@intel.com>
Date: Wed, 22 Jan 2025 15:16:38 -0800
From: Dave Hansen <dave.hansen@...el.com>
To: Tom Lendacky <thomas.lendacky@....com>,
 "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
 Kevin Loughlin <kevinloughlin@...gle.com>
Cc: linux-kernel@...r.kernel.org, tglx@...utronix.de, mingo@...hat.com,
 bp@...en8.de, dave.hansen@...ux.intel.com, x86@...nel.org, hpa@...or.com,
 seanjc@...gle.com, pbonzini@...hat.com, kai.huang@...el.com,
 ubizjak@...il.com, jgross@...e.com, kvm@...r.kernel.org, pgonda@...gle.com,
 sidtelang@...gle.com, mizhang@...gle.com, rientjes@...gle.com,
 manalinandan@...gle.com, szy0127@...u.edu.cn
Subject: Re: [PATCH v4 1/2] x86, lib: Add WBNOINVD helper functions

On 1/22/25 11:39, Tom Lendacky wrote:
>> I think you need to do something like.
>>
>> 	alternative("wbinvd", ".byte 0xf3; wbinvd", X86_FEATURE_WBNOINVD);
> I think "rep; wbinvd" would work as well.

I don't want to bike shed this too much.

But, please, no.

The fact that wbnoinvd can be expressed as "rep; wbinvd" is a
implementation detail. Exposing how it's encoded in the ISA is only
going to add confusion. This:

static __always_inline void wbnoinvd(void)
{
        asm volatile(".byte 0xf3,0x0f,0x09\n\t": : :"memory");
}

is perfectly fine and a billion times less confusing than:

        asm volatile(".byte 0xf3; wbinvd\n\t": : :"memory");
or
        asm volatile("rep; wbinvd\n\t": : :"memory");

It only gets worse if it's mixed in an alternative() with the _actual_
wbinvd.

BTW, I don't think you should be compelled to use alternative() as
opposed to a good old:

	if (cpu_feature_enabled(X86_FEATURE_WBNOINVD))
		...