linux-kernel - Re: [PATCH v4 6/6] x86/microcode/intel: Print when early microcode loading fails

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 for Android: free password hash cracker in your pocket

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

Message-ID: <Y8brnyIltcgtUvPn@zn.tnic>
Date:   Tue, 17 Jan 2023 19:40:31 +0100
From:   Borislav Petkov <bp@...en8.de>
To:     Dave Hansen <dave.hansen@...el.com>
Cc:     Ashok Raj <ashok.raj@...el.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        X86-kernel <x86@...nel.org>,
        LKML Mailing List <linux-kernel@...r.kernel.org>,
        Tony Luck <tony.luck@...el.com>,
        Ingo Molnar <mingo@...nel.org>, alison.schofield@...el.com,
        reinette.chatre@...el.com, Tom Lendacky <thomas.lendacky@....com>
Subject: Re: [PATCH v4 6/6] x86/microcode/intel: Print when early microcode
 loading fails

On Tue, Jan 17, 2023 at 10:32:50AM -0800, Dave Hansen wrote:
> Well, we have an awful lot of pr_warn()'s in the kernel that talk about
> something that was tried and failed.

Well, is microcode loading failure worth to warn about?

What if there's no microcode for that CPU?

Now imagine in that case you issue a warning and then you get all those
bugzillas opened:

"Heeey, is my CPU broken, it says it cannot load microcode"

I sure don't want to be on the receiving end of this.

I don't think you wanna be there either.

> I actually kinda like the inverse.
> 
> The common (boring) case where an update was needed and was successful.
> It's the one we don't need to tell users about at all.  It barely
> deserves a message.  Users expect that if there's an early update
> available, it'll get attempted, it will be successful and the kernel
> won't say much.

No argument there.

> The time we need to spam dmesg is when something upends user
> expectations and they might need to go do something.  An early loading
> failure is exactly the place where they want to know.  They want to know
> if they're running with known CPU bugs that would have been fixed by the
> early update

No, we already warn about that in the CPU mitigations code.

> or if they somehow have a botched early loading image.

If you can pick apart from this here:

        /* write microcode via MSR 0x79 */
        native_wrmsrl(MSR_IA32_UCODE_WRITE, (unsigned long)mc->bits);

what type of failure it is, then sure, let's warn.

But we should not warn just for every revision mismatch in there...

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette