From mboxrd@z Thu Jan  1 00:00:00 1970
From: Gleb Natapov <gleb@redhat.com>
Subject: Re: I/O port permission bit inheritance between threads
Date: Thu, 23 May 2013 12:27:59 +0300
Message-ID: <20130523092759.GL4725@redhat.com>
References: <20130520142431.5221e2d0@nehalam.linuxnetplumber.net>
 <20130521095030.GC7424@8bytes.org>
 <20130521100118.GC4725@redhat.com>
 <20130521082709.4427b6ae@nehalam.linuxnetplumber.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Joerg Roedel <joro@8bytes.org>, kvm@vger.kernel.org
To: Stephen Hemminger <stephen@networkplumber.org>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:39621 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1757582Ab3EWJ2D (ORCPT <rfc822;kvm@vger.kernel.org>);
	Thu, 23 May 2013 05:28:03 -0400
Content-Disposition: inline
In-Reply-To: <20130521082709.4427b6ae@nehalam.linuxnetplumber.net>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Tue, May 21, 2013 at 08:27:09AM -0700, Stephen Hemminger wrote:
> On Tue, 21 May 2013 13:01:18 +0300
> Gleb Natapov <gleb@redhat.com> wrote:
> 
> > On Tue, May 21, 2013 at 11:50:30AM +0200, Joerg Roedel wrote:
> > > Hey Stephen,
> > > 
> > > On Mon, May 20, 2013 at 02:24:31PM -0700, Stephen Hemminger wrote:
> > > > ioperm() inheritance across threads is different in KVM then when run
> > > > on physical hardware.  The following program runs on physical hardware
> > > > but get SEGV under KVM.
> > > > 
> > > > It appears that the I/O permission bits are not shared between threads
> > > > in the same way.
> > > 
> > > Is this specific to SVM or do you see it on VMX too? My first guess
> > > would be that the KVM instruction emulator does not check to
> > > IO-permissions correctly, but that would affect VMX and SVM.
> > > 
> > The program segfaults on physical hardware:
> > # ./a.out 
> > joining
> > waiting
> > beeping
> > Segmentation fault
> > 
> > --
> > 			Gleb.
> 
> The program had timing races, changing it slightly shows that.
> # ./beep
> beeping
> done
> oo
> # ./beep --pre
> joining
> beeping
> Segmentation fault
> # ./beep --post
> beeping
> joining
> done
> oo
> # 
> 
And I get the same in a VM on Intel host.

--
			Gleb.