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 S1754547AbbGPJK1 (ORCPT <rfc822;w@1wt.eu>);
	Thu, 16 Jul 2015 05:10:27 -0400
Received: from www.sr71.net ([198.145.64.142]:45125 "EHLO blackbird.sr71.net"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752840AbbGPJKZ (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Thu, 16 Jul 2015 05:10:25 -0400
Message-ID: <55A774FF.3060208@sr71.net>
Date: Thu, 16 Jul 2015 02:10:23 -0700
From: Dave Hansen <dave@sr71.net>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.7.0
MIME-Version: 1.0
CC: torvalds@linux-foundation.org, mingo@kernel.org,
        linux-kernel@vger.kernel.org, luto@amacapital.net, bp@alien8.de,
        fenghua.yu@intel.com, hpa@zytor.com, oleg@redhat.com,
        tglx@linutronix.de, ross.zwisler@linux.intel.com
Subject: Re: [RFC][PATCH 2/2] x86, fpu: enlarge xregs_state
References: <20150716081332.9413DD5B@viggo.jf.intel.com> <20150716081332.641D2B2F@viggo.jf.intel.com>
In-Reply-To: <20150716081332.641D2B2F@viggo.jf.intel.com>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
To: unlisted-recipients:; (no To-header on input)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 07/16/2015 01:13 AM, Dave Hansen wrote:
> +/*
> + * The largest xsave buffer known today is 2752 bytes on a system
> + * implementing AVX-512.  This includes the 512-byte i387 state
> + * and 64-byte header.  We add a small amount of padding in case
> + * an implementation adds some padding or wastes some space.
> + *
> + * Note, if we overflow this, we will disable XSAVE completely.
> + *
> + * Also, note that the real size we need is enumerated by
> + * cpuid leaves and can not be known at compile time.
> + */
> +#define XSTATE_MAX_SIZE	(2752 + 256)

BTW, this has one big-ish side-effect.  It takes the size of task_struct
from ~3.5k to ~5.4k for me:

slabinfo before:
task_struct          198    198   3456    9    8 : ...
after:
task_struct          166    180   5376    6    8 : ...

I'm sure folks on small systems are going to cringe at eating 2k/thread,
so we've got to revisit this _somehow_.