linux-kernel - RE: [PATCH v2 2/2] x86/MCE/AMD: Don't report L1 BTB MCA errors on some Family 17h models

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] [day] [month] [year] [list]

Date:   Fri, 22 Mar 2019 19:33:43 +0000
From:   "Ghannam, Yazen" <Yazen.Ghannam@....com>
To:     Borislav Petkov <bp@...en8.de>
CC:     "linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "tony.luck@...el.com" <tony.luck@...el.com>,
        "x86@...nel.org" <x86@...nel.org>,
        "rafal@...ecki.pl" <rafal@...ecki.pl>,
        "clemej@...il.com" <clemej@...il.com>
Subject: RE: [PATCH v2 2/2] x86/MCE/AMD: Don't report L1 BTB MCA errors on
 some Family 17h models

> -----Original Message-----
> From: linux-edac-owner@...r.kernel.org <linux-edac-owner@...r.kernel.org> On Behalf Of Borislav Petkov
> Sent: Friday, March 22, 2019 2:32 PM
> To: Ghannam, Yazen <Yazen.Ghannam@....com>
> Cc: linux-edac@...r.kernel.org; linux-kernel@...r.kernel.org; tony.luck@...el.com; x86@...nel.org; rafal@...ecki.pl;
> clemej@...il.com
> Subject: Re: [PATCH v2 2/2] x86/MCE/AMD: Don't report L1 BTB MCA errors on some Family 17h models
> 
> On Fri, Mar 22, 2019 at 07:24:01PM +0000, Ghannam, Yazen wrote:
> > Generally, the model groups share the same hardware design and so the
> > same quirks. So I'm thinking that it'd be more efficient to have a
> > filter function that targets a specific group of models rather than
> > one that checks all known quirks on all models.
> 
> Or simply start with a amd_filter_mce() function and when it grows big
> and unwieldy, only *then* start thinking about splitting it into models
> and families. For now, you're fine with a single AMD-specific function.
> 

Understood.

> > Most of the quirks are dealt with at init time, but this needs be to
> > done during run time for each MCE that is logged. So I didn't want to
> > add unnecessary checks to the MCE handlers. We have quirk_no_way_out()
> > that does something similar.
> 
> I don't think a couple of instructions checking vendor and family would
> be at all noticeable so let's start simple.
> 

Okay, will do.

Thanks,
Yazen