linux-kernel - Re: [PATCH 1/3] x86/MCE, EDAC/mce_amd: Add support for new MCA

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]

Date:   Mon, 24 Oct 2022 09:38:44 -0700
From:   Tony Luck <tony.luck@...el.com>
To:     Borislav Petkov <bp@...en8.de>
Cc:     Yazen Ghannam <yazen.ghannam@....com>, linux-edac@...r.kernel.org,
        linux-kernel@...r.kernel.org, x86@...nel.org,
        Smita.KoralahalliChannabasappa@....com
Subject: Re: [PATCH 1/3] x86/MCE, EDAC/mce_amd: Add support for new
 MCA_SYND{1,2} registers

On Mon, Oct 24, 2022 at 06:09:10PM +0200, Borislav Petkov wrote:
> On Tue, Aug 02, 2022 at 12:22:20PM +0000, Yazen Ghannam wrote:
> > I ask because struct mce is UAPI. But I think this is just for /dev/mcelog,
> > and this has been deprecated for a while. So on a related note, should
> > /dev/mcelog be removed and struct mce moved out of UAPI? Then changes to
> > struct mce won't affect user space, and we can just consider the mce trace
> > event when reporting to user space.
> 
> Question is, do you want those error records to be fed into mcelog on
> AMD too?
> 
> And I remember you guys supporting it at some point.
> 
> The answer to that question will tell you how exactly to build your
> structure of data you shuffle to luserspace.

There are still a fair number of users of mcelog, so I think it needs
to remain in its half-undead state a while longer.

Changes to "struct mce" have always been supported. Several have
been made over the years. The rules are quite simple:

1) Do not remove any existing fields
2) Legacy fields that are no longer used should have value 0.
3) Kernel internal values (currently just "kflags") should be
   zeroed in the structures passed out to user space.
3) New fields must be added at the end.

-Tony