Re: [PATCH] x86/mce: Streamline MCE subsystem's naming

From: Borislav Petkov <bp@alien8.de>
To: "Luck, Tony" <tony.luck@intel.com>
Cc: Ingo Molnar <mingo@kernel.org>,
	LKML <linux-kernel@vger.kernel.org>, X86 ML <x86@kernel.org>,
	linux-edac <linux-edac@vger.kernel.org>
Subject: Re: [PATCH] x86/mce: Streamline MCE subsystem's naming
Date: Thu, 6 Dec 2018 19:05:52 +0100	[thread overview]
Message-ID: <20181206180552.GB3986@zn.tnic> (raw)
In-Reply-To: <3908561D78D1C84285E8C5FCA982C28F7D47B665@ORSMSX110.amr.corp.intel.com>

On Thu, Dec 06, 2018 at 05:36:03PM +0000, Luck, Tony wrote:
> > So the real question is, is there a signifcant class of MCE events that 
> > are not tied to the reporting channel which is per CPU (-ish ...) MCA 
> > banks?
> 
> Perhaps QPI/UPI interconnect errors?

Whatever, this is pure bikeshedding what we're doing right now. All the
errors are reported through the CPU's MCA banks and the argument whether
an error class which is maybe not really tied to the CPU, should dictate
which path to put the files in, does not make any sense to me.

The only argument which makes sense IMO is whether the path should be
short. :-)

And both

arch/x86/cpu/mce/
arch/x86/mce/

are short enough to me.

Now, the real question is, how do we want to place the rest of the
files/folders so that we can get rid of deeper directory structures.

If we say, we don't want to have .../cpu/... and put everything in
arch/x86/ then this answers the above question too.

IMO, of course.

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.