From mboxrd@z Thu Jan  1 00:00:00 1970
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Subject: Re: [linux-linus test] 17926: regressions - FAIL
Date: Mon, 13 May 2013 16:24:33 -0400
Message-ID: <20130513202433.GG15684@phenom.dumpdata.com>
References: <osstest-17926-mainreport@xen.org>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xen.org>
Content-Disposition: inline
In-Reply-To: <osstest-17926-mainreport@xen.org>
List-Unsubscribe: <http://lists.xen.org/cgi-bin/mailman/options/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xen.org>
List-Help: <mailto:xen-devel-request@lists.xen.org?subject=help>
List-Subscribe: <http://lists.xen.org/cgi-bin/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=subscribe>
Sender: xen-devel-bounces@lists.xen.org
Errors-To: xen-devel-bounces@lists.xen.org
To: "xen.org" <ian.jackson@eu.citrix.com>
Cc: xen-devel@lists.xensource.com
List-Id: xen-devel@lists.xenproject.org

On Thu, May 09, 2013 at 03:52:47PM +0100, xen.org wrote:
> flight 17926 linux-linus real [real]
> http://www.chiark.greenend.org.uk/~xensrcts/logs/17926/
> 
> Regressions :-(
> 
> Tests which did not succeed and are blocking,
> including tests which could not be run:
>  test-amd64-i386-xl-credit2   10 guest-saverestore         fail REGR. vs. 12557
>  test-amd64-i386-xl           10 guest-saverestore         fail REGR. vs. 12557
>  test-amd64-i386-rhel6hvm-amd  7 redhat-install            fail REGR. vs. 12557
>  test-amd64-i386-pair   17 guest-migrate/src_host/dst_host fail REGR. vs. 12557

So this is interesting. I noticed that even a local save/restore of any
guest when the host is 32-bit is taking forever. While with a 64-bit host
is done much much faster. (Tens of minutes vs minutes)

Just running 'perf record xm save' showed me that it spends the majority
of time doing the xen_version hypercall (so yield to hypervisor). Hadn't really
dug deeper in this to see if there are some other operations (so block I/Os)
that are just soo much slower on i386 than 64-bit.