Messages in this thread | | | Date | Tue, 4 Mar 2008 17:15:10 +0100 | From | Ingo Molnar <> | Subject | Re: [RFC] use realmode code to reserve end-of-conventional-memory to 1MB |
| |
* Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> Ingo Molnar wrote: >>> I think it was a side effect of all the early_ioremap work, which changed >>> the order in which things happened at boot time. >>> >> >> which commits do you mean in particular? > > Good question. It might have been cpa as well. I'd have to look > through to confirm the theory, but I think Xen got broken for other > reasons around there, so simple bisect wouldn't be enough to identify > the problem.
i'm trying to do bug forensics to figure out where we broke Xen - i really have the feeling that there could be other side-effects of whatever change caused this.
Ingo
| |