linux-kernel - RE: [PATCH 1/3] EDAC/amd64: Print ECC enabled/disabled for nodes with enabled MCs

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:   Wed, 28 Mar 2018 15:59:36 +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>
Subject: RE: [PATCH 1/3] EDAC/amd64: Print ECC enabled/disabled for nodes with
 enabled MCs

> -----Original Message-----
> From: Borislav Petkov <bp@...en8.de>
> Sent: Wednesday, March 28, 2018 11:44 AM
> To: Ghannam, Yazen <Yazen.Ghannam@....com>
> Cc: linux-edac@...r.kernel.org; linux-kernel@...r.kernel.org
> Subject: Re: [PATCH 1/3] EDAC/amd64: Print ECC enabled/disabled for nodes
> with enabled MCs
> 
> On Wed, Mar 28, 2018 at 02:38:11PM +0000, Ghannam, Yazen wrote:
> > In either of those cases we won't get to debug_display_dimm_sizes*
> > because we won't initialize the instance.
> 
> So you move that code which accesses csrows up so that it has the
> required information to query DIMM state/presence.
> 

I don't think we need to look at the csrows. We just need to check if the
controller is enabled or not. A controller is enabled if there is a DIMM
behind it and disabled if there isn't. 

> >                 /* Assume UMC MCA banks are enabled. */
> >                 nb_mce_en = true;
> 
> Also, I don't like that assumption.
> 

I'll look into it.

> > This would work for Fam17h. For older systems I think we can look at
> > D18F2x94_dct[1:0][DisDramInterface]
> >
> > Or maybe we have a separate function to check for enabled memory
> controllers
> > before we check for ECC?
> 
> The best would be to have a function which checks whether DIMMs are
> present on the node and act accordingly. You can pull up some of the
> work of caching registers which are used in debug_display_dimm_sizes*
> and use that info to get the required DIMM state upfront.
> 

Okay, so I'll do a separate function.

Thanks,
Yazen