Re: [PATCH 3/3] memory: introduce an option to force onlining of hotplug memory

[Michal Hocko]

On Thu 23-07-20 19:39:54, David Hildenbrand wrote:
[...]
> Yeah, might require some code churn. It just feels wrong to involve
> buddy concepts (e.g., onlining pages, calling memory notifiers, exposing
> memory block devices) and introducing hacks (forced onlining) just to
> get a memmap+identity mapping+iomem resource. I think reserving such a
> region during boot as suggested is the easiest approach, but I am
> *absolutely* not an expert on all these XEN-specific things :)

I am late to the discussion but FTR I completely agree.
-- 
Michal Hocko
SUSE Labs