linux-kernel - Re: [RFC PATCH 2/3] MCE, CE: Wire in the CE collector

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: <CA+8MBbKyHfwK7v9nmFNm0TLvpjbEpDbWemk4C8tE7mkZedE3ag@mail.gmail.com>
Date:	Tue, 27 May 2014 10:48:38 -0700
From:	Tony Luck <tony.luck@...il.com>
To:	Borislav Petkov <bp@...en8.de>
Cc:	LKML <linux-kernel@...r.kernel.org>,
	linux-edac <linux-edac@...r.kernel.org>, X86 ML <x86@...nel.org>
Subject: Re: [RFC PATCH 2/3] MCE, CE: Wire in the CE collector

> +       } else if (c->x86_vendor == X86_VENDOR_INTEL)
> +               return m->status & BIT(7);

Intel compound error codes aren't quite that simple.  You need to look
at the low 16 bits of "status" (the MCACOD) field and see which is the
most significant bit set (ignoring bit 12, the "filter" bit).  If the answer is
bit 7 - then this is a memory error. But you can't just blindly check bit
7 because if bit 8 is set, then this is a cache error, and if bit 11 is set,
then it is a bus/interconnect error - and either way bit 7 just gives more
detail on what cache/bus/interconnect error happened.  In hex the test
you want is:

                  return (m->status & 0xef80) == BIT(7);

[compound error codes make my head hurt too]

>                 if (!(flags & MCP_TIMESTAMP))
>                         m.tsc = 0;
> -               /*
> -                * Don't get the IP here because it's unlikely to
> -                * have anything to do with the actual error location.
> -                */
> -               if (!(flags & MCP_DONTLOG) && !mca_cfg.dont_log_ce)
> -                       mce_log(&m);
> +
> +               __log_ce(&m, flags);

I'm not happy with the total removal of mce_log() here.  Skipping
it and doing *some* filtering in the kernel is OK (goal of this patch
set). But you just cut EDAC and/or EXTLOG out of the reporting path
completely.  If the corrected error count for a page gets too high, your
new code will try to take the page offline ... but we won't have a report
from EDAC/EXTLOG telling us which DIMM that page belonged to.

Perhaps __log_ce() needs a return value to say whether it took action
and then:

                 if (__log_ce(&m, flags) && all-those-other-tests)
                             mce_log(&m);

-Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/