linux-kernel - [PATCH 0/3] EDAC: Convert ghes

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-next>] [day] [month] [year] [list]

Message-Id: <20170726084827.11447-1-bp@alien8.de>
Date:   Wed, 26 Jul 2017 10:48:24 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     linux-edac <linux-edac@...r.kernel.org>
Cc:     Tony Luck <tony.luck@...el.com>,
        Mauro Carvalho Chehab <mchehab@...radead.org>,
        Toshimitsu Kani <toshi.kani@....com>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        LKML <linux-kernel@...r.kernel.org>
Subject: [PATCH 0/3] EDAC: Convert ghes_edac to a normal module

From: Borislav Petkov <bp@...e.de>

Hi,

here's the conversion of ghes_edac to a proper module using a notifier.
This is only the bare minimum stuff and we should do all outstanding
work ontop. Which right now is:

* platform whitelist

* Tony: "Additional cleanup: edac_ghes.c should really call
cper_mem_err_location() instead of doing pretty much the same thing but
with subtle formatting differences ("node:%d" vs. "node: %d")."

Yes, I think we should take a look at doing that eventually as a
sensible cleanup.

* Tony: "This driver isn't going to set the EDAC error counts since it
just does:

        e->top_layer = -1;
        e->mid_layer = -1;
        e->low_layer = -1;
"

Right, that we should address by making layer 0 EDAC_MC_LAYER_SLOT and
no other layers as we can't get any channel, branch or whatever DIMM
topology from GHES. At least I'm not aware of any.

But that should be enough as this info is good enough for pinpointing
the DIMM. From our discussion with Tony:

>>  The FRU identifier is:
>>
>> "The module number of the memory error location. (NODE, CARD, and MODULE
>> should provide the information necessary to identify the failing FRU)."
>
> Sounds pretty good. Generally people don't care about the channel. They
> want to know which DIMM to replace. This will give them what they want.

Thanks.

Borislav Petkov (3):
  EDAC: Add edac_pr_err/info macros
  ACPI/GHES: Add an EDAC notifier chain
  EDAC, ghes: Make it a proper module

 drivers/acpi/apei/ghes.c |  32 +++++++-----
 drivers/edac/Kconfig     |   4 +-
 drivers/edac/edac_mc.h   |   3 ++
 drivers/edac/ghes_edac.c | 129 ++++++++++++++++++++++++-----------------------
 include/acpi/ghes.h      |  27 +---------
 5 files changed, 93 insertions(+), 102 deletions(-)

-- 
2.14.0.rc0