From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756613Ab2HVOMX (ORCPT ); Wed, 22 Aug 2012 10:12:23 -0400 Received: from nat28.tlf.novell.com ([130.57.49.28]:41473 "EHLO nat28.tlf.novell.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756455Ab2HVOMF convert rfc822-to-8bit (ORCPT ); Wed, 22 Aug 2012 10:12:05 -0400 Message-Id: <503504FE0200007800096F08@nat28.tlf.novell.com> X-Mailer: Novell GroupWise Internet Agent 12.0.0 Date: Wed, 22 Aug 2012 15:12:46 +0100 From: "Jan Beulich" To: "Konrad Rzeszutek Wilk" Cc: "Stefano Stabellini" , "xen-devel@lists.xensource.com" , "linux-kernel@vger.kernel.org" Subject: Re: Q:pt_base in COMPAT mode offset by two pages. Was:Re: [Xen-devel] [PATCH 02/11] xen/x86: Use memblock_reserve for sensitive areas. References: <1345133009-21941-1-git-send-email-konrad.wilk@oracle.com> <1345133009-21941-3-git-send-email-konrad.wilk@oracle.com> <20120820141305.GA2713@phenom.dumpdata.com> <20120821172732.GA23715@phenom.dumpdata.com> <20120821190317.GA13035@phenom.dumpdata.com> In-Reply-To: <20120821190317.GA13035@phenom.dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8BIT Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org >>> On 21.08.12 at 21:03, Konrad Rzeszutek Wilk wrote: > On Tue, Aug 21, 2012 at 01:27:32PM -0400, Konrad Rzeszutek Wilk wrote: >> Jan, I thought something odd. Part of this code replaces this: >> >> memblock_reserve(__pa(xen_start_info->mfn_list), >> xen_start_info->pt_base - xen_start_info->mfn_list); >> >> with a more region-by-region area. What I found out that if I boot this >> as 32-bit guest with a 64-bit hypervisor the xen_start_info->pt_base is >> actually wrong. >> >> Specifically this is what bootup says: >> >> (good working case - 32bit hypervisor with 32-bit dom0): >> (XEN) Loaded kernel: c1000000->c1a23000 >> (XEN) Init. ramdisk: c1a23000->cf730e00 >> (XEN) Phys-Mach map: cf731000->cf831000 >> (XEN) Start info: cf831000->cf83147c >> (XEN) Page tables: cf832000->cf8b5000 >> (XEN) Boot stack: cf8b5000->cf8b6000 >> (XEN) TOTAL: c0000000->cfc00000 >> >> [ 0.000000] PT: cf832000 (f832000) >> [ 0.000000] Reserving PT: f832000->f8b5000 >> >> And with a 64-bit hypervisor: >> >> XEN) VIRTUAL MEMORY ARRANGEMENT: >> (XEN) Loaded kernel: 00000000c1000000->00000000c1a23000 >> (XEN) Init. ramdisk: 00000000c1a23000->00000000cf730e00 >> (XEN) Phys-Mach map: 00000000cf731000->00000000cf831000 >> (XEN) Start info: 00000000cf831000->00000000cf8314b4 >> (XEN) Page tables: 00000000cf832000->00000000cf8b6000 >> (XEN) Boot stack: 00000000cf8b6000->00000000cf8b7000 >> (XEN) TOTAL: 00000000c0000000->00000000cfc00000 >> (XEN) ENTRY ADDRESS: 00000000c16bb22c >> >> [ 0.000000] PT: cf834000 (f834000) >> [ 0.000000] Reserving PT: f834000->f8b8000 >> >> So the pt_base is offset by two pages. And looking at c/s 13257 >> its not clear to me why this two page offset was added? Honestly, without looking through this in greater detail I don't recall. That'll have to wait possibly until after the summit, though. I can't exclude that this is just a forgotten leftover from an earlier version of the patch. I would have thought this was to account for the L4 tables that the guest doesn't see, but - this should only be a single page - this should then also (or rather instead) be subtracted from nr_pt_frames so that's likely not it. >> The toolstack works fine - so launching 32-bit guests either >> under a 32-bit hypervisor or 64-bit works fine: >> ] domainbuilder: detail: xc_dom_alloc_segment: page tables : 0xcf805000 -> > 0xcf885000 (pfn 0xf805 + 0x80 pages) >> [ 0.000000] PT: cf805000 (f805000) >> > > And this patch on top of the others fixes this.. I didn't look at this in too close detail, but I started to get afraid that you might be making the code dependent on many hypervisor implementation details. And should the above turn out to be a bug in the hypervisor, I hope your kernel side changes won't make it impossible to fix that bug. Jan