From mboxrd@z Thu Jan  1 00:00:00 1970
From: Andrea Arcangeli <andrea@qumranet.com>
Subject: Re: KVM: MMU: rmap_write_protect() hugepage iteration bug
Date: Sun, 8 Jun 2008 22:30:37 +0200
Message-ID: <20080608203037.GB9921@duo.random>
References: <20080608002736.GA25582@dmt.cnet> <20080608015452.GC8321@duo.random> <20080608044853.GA1408@dmt.cnet> <20080608183119.GA21835@duo.random> <20080608195215.GA5824@dmt.cnet>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Avi Kivity <avi@qumranet.com>, kvm-devel <kvm@vger.kernel.org>
To: Marcelo Tosatti <mtosatti@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from host36-195-149-62.serverdedicati.aruba.it ([62.149.195.36]:51642
	"EHLO mx.cpushare.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753982AbYFHUaj (ORCPT <rfc822;kvm@vger.kernel.org>);
	Sun, 8 Jun 2008 16:30:39 -0400
Content-Disposition: inline
In-Reply-To: <20080608195215.GA5824@dmt.cnet>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Sun, Jun 08, 2008 at 04:52:15PM -0300, Marcelo Tosatti wrote:
> We do. The case is were you have two entries in the array. rmap_remove
> will first remove the entry at index 0, and move the entry at index 1 to
> 0.
> 
> Then we call "rmap_next()" with a non-NULL spte, which will skip the
> only remaining entry at index 0. IOW rmap_next() requires the spte
> argument to be NULL if the array has one valid entry.

How exactly we could call rmap_next with not-NULL spte after any
rmap_remove with your last patch applied?

> > chains are fairly long. To be sure you could try to profile it and see
> > if rmap_write_protect goes up in the opreport.
> 
> Will do.

Cool, thanks!