linux-kernel - Re: [PATCH 3/4] EDAC, MCE, AMD: Add hw

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:	Sun, 8 May 2011 21:43:16 +0200
From:	Borislav Petkov <bp@...64.org>
To:	Greg KH <greg@...ah.com>
Cc:	Borislav Petkov <bp@...64.org>,
	EDAC devel <linux-edac@...r.kernel.org>,
	X86-ML <x86@...nel.org>, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/4] EDAC, MCE, AMD: Add hw_inject and cpu sysfs attrs

On Sun, May 08, 2011 at 03:18:53PM -0400, Greg KH wrote:
> > > You forgot to add the proper Documentation/ABI/ files for these new
> > > sysfs files.  Please do that.
> > 
> > Yeah, so those sprang up as a testing module and the final goal is to
> > have all injection done through perf events but it could also be prudent
> > to keep those as a fallback alternative.
> > 
> > I'm thinking something like
> > Documentation/ABI/testing/sysfs-devices-system-edac but looking
> > at Documentation/ABI/README and reading the "testing/" category
> > description, I don't think it applies to those interfaces just yet for
> > the simple reason that they're still changing and I wouldn't want to
> > cast them in stone now.
> > 
> > Can we add them to a "unstable" category instead or maybe delay making
> > the ABI description for now?
> 
> No.  If you add sysfs files to the kernel, you had better be sure that
> they are correct and are what you want to be doing.  Otherwise, don't
> add them.

I cannot be sure right now because this whole RAS thing is being
developed now and we're more or less learning by doing. I don't think
anyone can tell you what we're going to need a year from now on
interfaces for error injection.

> As you really want to do this through perf, just do it through perf.
> Please don't create a user/kernel api that you are not going to want
> to use in the end.

As I said above, it is going to be used as a fallback alternative to the
perf option for systems which cannot do perf. This injection method will
be the fallback.

Also, this whole module is default-N in Kconfig and you have to really
know what you're doing when enabling it.

Which brings me to a much better idea: this whole interface should
actually go to debugfs! This is more or less a debugging module, so
debugfs, doh!

/me slaps himself on the forehead and goes to rework the patches.

Thanks.

-- 
Regards/Gruss,
Boris.

Advanced Micro Devices GmbH
Einsteinring 24, 85609 Dornach
General Managers: Alberto Bozzo, Andrew Bowd
Registration: Dornach, Gemeinde Aschheim, Landkreis Muenchen
Registergericht Muenchen, HRB Nr. 43632
--
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/