From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [Xen-devel] [PATCH 1/3] Add mcelog support for xen platform Date: Wed, 25 Apr 2012 10:11:25 +0200 Message-ID: <20120425081125.GA19849@gmail.com> References: <20120421041454.GA29704@phenom.dumpdata.com> <20120421104502.GB17005@aftab.osrc.amd.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Andi Kleen Cc: Borislav Petkov , Konrad Rzeszutek Wilk , "Liu, Jinsong" , tony.luck@intel.com, x86@kernel.org, linux-edac@vger.kernel.org, "xen-devel@lists.xensource.com" , "linux-kernel@vger.kernel.org" , "H. Peter Anvin" , Thomas Gleixner , torvalds@linux-foundation.org List-Id: xen-devel@lists.xenproject.org * Andi Kleen wrote: > Borislav Petkov writes: > > > Because, if you'd hooked into it, just imagine one fine day, when we > > remove mcelog support, what screaming the xen people will be doing when > > mcelog doesn't work anymore. > > That's simple. /dev/mcelog is a widely used user space ABI > [...] *COUGH* *SPLUTTER* *LAUGH*. Thanks for making my day with the self-serving exaggeration of the year. In truth /dev/mcelog is one of the crappiest ABIs Linux has - full stop. As a result it's barely used by anyone who can avoid it - its main usecase is the utility you wrote for it and some enterprise folks who desperately need that data and don't care how they get it. The mcelog user-space utility sucks too, understandably. Nevertheless we probably have to keep the ABI around until a truly better replacement is out there (the RAS daemon for example) and mcelog usage eclipses to obscurity, but by all means we want to limit its further spreading /dev/mcelog to areas it has not polluted yet ... Thanks, Ingo