Re: [Xen-devel] [PATCH v7 2/7] xen/x86: populate PVHv2 Dom0 physical memory map

[Roger Pau Monne <roger.pau@xxxxxxxxxx>]

On Thu, Feb 23, 2017 at 04:16:15PM +0000, Andrew Cooper wrote:
> On 23/02/17 16:09, Jan Beulich wrote:
> >>>> On 23.02.17 at 16:27, <roger.pau@xxxxxxxxxx> wrote:
> >> I also don't see much benefit from deviating from what HVM does, so I would
> >> just leave it as is.
> > Understood. The risk is that in a couple of years time understanding
> > why it is the way it is and (whether that's necessary) may take as
> > much effort as did the deciphering of the (buggy) hvmloader
> > setup code for this.
> 
> In which case, just leave a short comment saying something like "Copy
> HVMLoader for consistency, not that we expect a PVH domain from using
> this for anything other than their AP trampoline".

What about adding the chunk below. If you agree I can prepare a git branch with
this and the libelf changes merged in (and the RBs), so you can pull from it.

Thanks, Roger.
---8<---
diff --git a/xen/arch/x86/domain_build.c b/xen/arch/x86/domain_build.c
index adc4c00..4d7f4ae 100644
--- a/xen/arch/x86/domain_build.c
+++ b/xen/arch/x86/domain_build.c
@@ -51,6 +51,11 @@ static long __initdata dom0_max_nrpages = LONG_MAX;
  * - 128 bytes I/O bitmap
  * - one trailing byte
  * or a total of 265 bytes.
+ *
+ * NB: as PVHv2 Dom0 doesn't have legacy devices (ISA), it shouldn't have any
+ * business in accessing the ISA port range, much less in real mode, and due to
+ * the lack of firmware it shouldn't also execute any INT instruction. This is
+ * done for consistency with what hvmloader does.
  */
 #define HVM_VM86_TSS_SIZE 265
 


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel