From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Jan Beulich" Subject: Re: [PATCH 3/4] x86-64: EFI runtime code Date: Tue, 28 Jun 2011 08:56:03 +0100 Message-ID: <4E09A533020000780004A535@nat28.tlf.novell.com> References: <4E099AA6020000780004A4C6@nat28.tlf.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Keir Fraser , Konrad Rzeszutek Wilk Cc: "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org >>> On 28.06.11 at 09:39, Keir Fraser wrote: > On 28/06/2011 08:11, "Jan Beulich" wrote: >=20 >>>>> On 27.06.11 at 18:25, Konrad Rzeszutek Wilk = wrote: >>> On Mon, Jun 27, 2011 at 11:43:02AM +0100, Jan Beulich wrote: >>>> This allows Dom0 access to all suitable EFI runtime services. The >>>=20 >>> What type of patches for the upstream 3.0 kernel are needed to take >>> advantage of these new hypercalls? >>=20 >> I did not put any consideration in how to integrate this with the >> upstream kernel. For our kernel, I used the *-xen.? mechanism >> to have a parallel source file to arch/x86/platform/efi/efi.c, and >> excluded building of arch/x86/platform/efi/efi_stub_*.S and >> arch/x86/platform/efi/efi_*.c. >>=20 >> Patch below for reference. >=20 > Since the new hypercalls are 1:1 replacements for the EFI run-time calls = (I > think?) Yes, with a few exceptions of things that must not be done from Dom0 (i.e. the ResetSystem() and SetVirtualAddressMap() ones). > we could perhaps keep most of Linux's EFI subsystem intact and spoof > it with a fake operations table containing hypercall stubs. Indeed, that part ought to be simple. The question is which of the (luckily few) other code paths need adjustment. Jan