Re: [PATCH 12/12] xen/mtrr: Add mtrr_if support for Xen mtrr

[Jeremy Fitzhardinge <jeremy@goop.org>]

 On 09/28/2010 07:00 AM, Stefano Stabellini wrote:
> On Tue, 28 Sep 2010, Ingo Molnar wrote:
>> * stefano.stabellini@eu.citrix.com <stefano.stabellini@eu.citrix.com> wrote:
>>
>>> From: Stephen Tweedie <sct@redhat.com>
>>>
>>> Add a Xen mtrr type, and reorganise mtrr initialisation slightly to
>>> allow the mtrr driver to set up num_var_ranges (Xen needs to do this by
>>> querying the hypervisor itself.)
>>>
>>> [ Impact: add basic MTRR support ]
>>>
>>> Signed-off-by: Stephen Tweedie <sct@redhat.com>
>>> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
>>> Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>>> ---
>>>  arch/x86/kernel/cpu/mtrr/Makefile |    2 +-
>>>  arch/x86/kernel/cpu/mtrr/main.c   |    3 +
>>>  arch/x86/kernel/cpu/mtrr/mtrr.h   |    7 ++
>>>  arch/x86/kernel/cpu/mtrr/xen.c    |  110 +++++++++++++++++++++++++++++++++++++
>>>  4 files changed, 121 insertions(+), 1 deletions(-)
>>>  create mode 100644 arch/x86/kernel/cpu/mtrr/xen.c
>> Still NAK, for the very same reasons as we NAK-ed it the previous time: 
>> /proc/mtrr is a problematic and complicated legacy interface that should 
>> die. Any modern X server will do the right thing via PAT.
>>
> Sorry I should have read the original thread more carefully: I didn't
> realize this patch had been NAK-ed.
>
> However it is not a problem because we can easily disable MTRRs from Xen
> and with no cpu_has_mtrr the kernel would still boot fine on Xen.
> Also I think we do have PAT support nowadays but I'll let Jeremy comment
> on that.

Yes, we could just mask out the MTRR CPU feature and rely entirely on PAT.

The alternative would be to use the wrmsr hooks to emulate the Intel
MTRR registers by mapping them to hypercalls, but that seems needlessly
complex.

    J