Re: [PATCH] x86/microcode/AMD: Attempt applying on every logical thread

From: Ashok Raj <ashok.raj@intel.com>
To: Borislav Petkov <bp@alien8.de>
Cc: "Ashok Raj" <ashok_raj@linux.intel.com>,
	"X86 ML" <x86@kernel.org>, "Andrew Cooper" <amc96@srcf.net>,
	LKML <linux-kernel@vger.kernel.org>,
	"Ștefan Talpalaru" <stefantalpalaru@yahoo.com>,
	"Ashok Raj" <ashok.raj@intel.com>
Subject: Re: [PATCH] x86/microcode/AMD: Attempt applying on every logical thread
Date: Wed, 17 Aug 2022 15:29:04 +0000	[thread overview]
Message-ID: <Yv0JQGa/2BlbQChZ@araj-dh-work> (raw)
In-Reply-To: <Yvz53AAh+p5Db7CK@zn.tnic>

On Wed, Aug 17, 2022 at 04:23:24PM +0200, Borislav Petkov wrote:
> On Wed, Aug 17, 2022 at 12:12:05PM +0000, Ashok Raj wrote:
> > Instead of doing a complete hack,
> 
> What complete hack?!

What I meant was the patch removed any and all revid checks *completely*
Instead of even limiting to == checks.

Once you do that "in spirit" its the same end goal. Isn't it? Although it
might be required for a completely different reason.

> 
> > could we maybe revive what we attempted
> > in 2019? At a minimum it will work for both architectures.
> > 
> > https://lore.kernel.org/lkml/1567056803-6640-1-git-send-email-ashok.raj@intel.com/
> 
> echo 2 >... - now that's a hack.

Its all in the eye of the beholder :-)

I didn't mean in any way to bring back the echo 2. What ever you think is
suitable to enable this is fine.

> 
> If you wanna use the kernel for testing your hw, use a local patch. Not
> expose it to people and then have me debug all kind of strange lockups.

So forget the hardware testing part. This is a complex flow for late-load
and how can we get more people to test it today in the community?

Do we have a more scalable way to support it today? 

This is similar to XEN supporting, ucode=allow_same, there is precedence. 

Doing that is just like having a built-in self-test for microcode loading. 

I think there is value in enabling it for x86.

Cheers,
Ashok