linux-kernel - Re: spurious mce Hardware Error messages in next-20250912

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: <20250918211405.GA2180898@yaz-khff2.amd.com>
Date: Thu, 18 Sep 2025 17:14:05 -0400
From: Yazen Ghannam <yazen.ghannam@....com>
To: "Luck, Tony" <tony.luck@...el.com>
Cc: Nikolay Borisov <nik.borisov@...e.com>,
	Bert Karwatzki <spasswolf@....de>, Borislav Petkov <bp@...en8.de>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-next@...r.kernel.org" <linux-next@...r.kernel.org>,
	"linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
	"linux-acpi@...r.kernel.org" <linux-acpi@...r.kernel.org>,
	"x86@...nel.org" <x86@...nel.org>,
	"rafael@...nel.org" <rafael@...nel.org>,
	"Zhuo, Qiuxu" <qiuxu.zhuo@...el.com>,
	"Smita.KoralahalliChannabasappa@....com" <Smita.KoralahalliChannabasappa@....com>
Subject: Re: spurious mce Hardware Error messages in next-20250912

On Thu, Sep 18, 2025 at 09:04:53PM +0000, Luck, Tony wrote:
> > For the current issue, it does seem that the registers contain junk
> > values. And we are only now seeing this with the recent rework.
> 
> Do you try to clear these registers after logging? Or just rely on clearing
> the MCi_STATUS register?
> 

Yes, the MCA_DESTAT register is cleared in a couple of places depending
on the scenario.

> If you are clearing, then it isn't working (or new junk values appear quickly).

Right, and MCi_STATUS has junk values in some of the affected banks.
They just happen to be ignored because they don't have the Valid bit
set.

And MCi_STATUS is cleared at Linux init time, so the junk values stick
or come back by the time Bert ran the script.

Thanks,
Yazen