From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Luis R. Rodriguez" Subject: Re: Fwd: [PATCH] x86: Use larger chunks in mtrr_cleanup Date: Fri, 4 Sep 2015 03:40:40 +0200 Message-ID: <20150904014040.GA8051@wotan.suse.de> References: <20150903184029.GV8051@wotan.suse.de> <1441308162.3277.20.camel@hpe.com> <20150903195134.GW8051@wotan.suse.de> <1441315902.3277.39.camel@hpe.com> <20150903220711.GX8051@wotan.suse.de> <1441319131.3277.54.camel@hpe.com> <20150903224556.GY8051@wotan.suse.de> <1441322474.3277.78.camel@hpe.com> <20150903235429.GZ8051@wotan.suse.de> <1441327726.3277.109.camel@hpe.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: <1441327726.3277.109.camel@hpe.com> Sender: linux-kernel-owner@vger.kernel.org To: Toshi Kani Cc: Prarit Bhargava , Stuart Hayes , tglx@linutronix.de, mingo@redhat.com, "H. Peter Anvin" , linux-kernel@vger.kernel.org, x86@kernel.org, mcgrof@do-not-panic.com, Toshi Kani , Jan Beulich , Juergen Gross , Roger Pau =?iso-8859-1?Q?Monn=E9?= , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org On Thu, Sep 03, 2015 at 06:48:46PM -0600, Toshi Kani wrote: > On Fri, 2015-09-04 at 01:54 +0200, Luis R. Rodriguez wrote: > > On Thu, Sep 03, 2015 at 05:21:14PM -0600, Toshi Kani wrote: > > > On Fri, 2015-09-04 at 00:45 +0200, Luis R. Rodriguez wrote: > > > > On Thu, Sep 03, 2015 at 04:25:31PM -0600, Toshi Kani wrote: > : > > > > > On Xen, > > > >=20 > > > > When Xen is used a platform firmware may still set up MTRR, eve= n if the > > > > hypervisor doesn't set up MTRR right ? So same issue and questi= on here. > > >=20 > > > Right, I meant to say Xen guests. > >=20 > > Ah but its import complicated than that. > >=20 > > > In case of the Xen hypervisor, > > > mtrr_type_lookup() returns a valid type as it runs on a platform. > >=20 > > I am not sure if this happens today, I know MTRR is simply disabled= by > > the Xen Hypervisor on the CPU explicitly, it disable it so guests r= eading > > the MTRR capabilities sees it as disabled when queried. >=20 > Oh, I would not let the hypervisor to disable MTRRs... Commit 586ab6a055376ec3f3e1e8 ("x86/pvh: disable MTRR feature on cpuid = for Dom0") by Roger Pau Monn=E9 disables MTRR for PVH dom0, so that cpuid returns = that MTRR is disabled to guests. Then later on Linux as of commit 47591df505= 12 ("xen: Support Xen pv-domains using PAT") added by Juergen as of v3.19 Linux guests can end up booting without MTRR but with PAT now enabled. > > Then since the Xen Linux guests cannot speak MTRR through the hyper= visor > > (for instance Xen guests cannot ask Xen hypervisor to mtrr_type_loo= kup() for > > it) if PCI passthrough is used it could mean a guest might set up /= use > > incorrect info as well. > >=20 > > If I undestand this correctly then I think we're in a pickle with X= en unless > > we add hypervisor support and hypercall support for mtrr_type_looku= p(). >=20 > I was under assumption that MTRRs are emulated and disabled on guests= =2E Some "special" flavor Linux guests (with non-upstream code) have guest MTRR hypercall support, for vanilla Xen and Linux they just never get M= TRR support. After Juergen's Linux changes though Xen guests can now get shiny PAT support. Since MTRR hypercall support is not upstream and MTR= R is ancient I decided instead of adding MTRR hypercall support upstream to = go with converting all drivers to PAT interfaces, with the assumption there wou= ld be no issues. > Isn't guest physical address virtualized? It is, there is a xen iotlb and stuff but that should ensure dom0 gets to get proper access to devices, and if you use PCI passthrough you wan= t the best experience as well. > I know other proprietary VMMs on IA64, > but know nothing about Xen... So, please disregard my comments to Xe= n. :-) No worries, no one knows all the answers, we work together to remove cob webs off of these odd corners no one cares about :) Luis