[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Xen-ia64-devel] Re: [patch 05/12] Kexec: Save the MADT ACPI tables so that they can be restored
On Wed, 2007-10-17 at 19:41 +0900, Simon Horman wrote:
> if you prefer this approach then its fine by me. I can confirm that
> it works - that is, with it my RX2620 ends up with 2 cpus in the second
> kernel, and without it, it doesn't.
Great, I do like it a bit better. We're mucking with the ACPI tables
in arch code, so we should handle restoring them in arch code. I also
like that it makes us aware of which tables we're changing.
> I think that the best way forward from here is to look at a solution
> for the EFI mapping problem. I will try and find time to investigate
> Yamahata-san's RID idea ASAP. If we can get a solution for that problem
> in place, then the way to deal with the PA() in purgatory problem
> should be come obvious, and if Yamahata-san's idea works and
> we can map EFI at 0xe000000000000000UL, then the problem will go away.
Yes, that seems to have the most potential for making things easily
fall into place. It would be nice if we could tie up the loose ends in
time for Xen 3.2. Thanks,
Alex Williamson HP Open Source & Linux Org.
Xen-ia64-devel mailing list