From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751851AbdJMQgi (ORCPT <rfc822;w@1wt.eu>);
        Fri, 13 Oct 2017 12:36:38 -0400
Received: from mga09.intel.com ([134.134.136.24]:50494 "EHLO mga09.intel.com"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1750872AbdJMQgh (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Fri, 13 Oct 2017 12:36:37 -0400
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.43,371,1503385200"; 
   d="scan'208";a="160211049"
Date: Fri, 13 Oct 2017 09:36:36 -0700
From: Andi Kleen <ak@linux.intel.com>
To: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>, Andi Kleen <andi@firstfloor.org>,
        x86@kernel.org, linux-kernel@vger.kernel.org,
        Jonathan McDowell <noodles@earth.li>
Subject: Re: [PATCH v9 2/5] x86/cpuid: Add generic table for cpuid
 dependencies
Message-ID: <20171013163636.GP5109@tassilo.jf.intel.com>
References: <20171007000328.12843-1-andi@firstfloor.org>
 <20171007000328.12843-3-andi@firstfloor.org>
 <20171012080733.4y5wlghjr524cgr3@gmail.com>
 <alpine.DEB.2.20.1710121710580.1930@nanos>
 <20171013053005.cj5vhqehkwlnhyla@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20171013053005.cj5vhqehkwlnhyla@gmail.com>
User-Agent: Mutt/1.9.0 (2017-09-02)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

> > I was looking at that as well and decided that we preferrably have a
> > compressed data structure. The code which walks the table is hardly
> > performance critical and the difference in text size is marginal.
> 
> So the code should all be __init (once that is fixed), hence data and text size 
> literally does not matter - it gets freed.

It's difficult to make the code __init because it's part of the CPU initialization and
the CPU initialization is called from CPU hotplug. 

It shouldn't do anything after initial boot however, but it's difficult to tell
that to the dependency checker.

In theory it would be possible to restructure cpu initialization to avoid this,
but I suspect it would be rather large and intrusive and probably not worth
it for a couple hundred bytes.

I removed the shorts.

-Andi