Re: [RFC PATCH 0/4] mm, memory_hotplug: allocate memmap from hotadded memory

[Michal Hocko]

On Fri 23-11-18 12:55:41, Oscar Salvador wrote:
> On Thu, Nov 22, 2018 at 10:21:24AM +0100, David Hildenbrand wrote:
> > 1. How are we going to present such memory to the system statistics?
> > 
> > In my opinion, this vmemmap memory should
> > a) still account to total memory
> > b) show up as allocated
> > 
> > So just like before.
> 
> No, it does not show up under total memory and neither as allocated memory.
> This memory is not for use for anything but for creating the pagetables
> for the memmap array for the section/s.

I haven't read through your patches yet but wanted to clarfify few
points here.

This should essentially follow the bootmem allocated memory pattern. So
it is present and accounted to spanned pages but it is not managed.

> It is not memory that the system can use.

same as bootmem ;)
 
> I also guess that if there is a strong opinion on this, we could create
> a counter, something like NR_VMEMMAP_PAGES, and show it under /proc/meminfo.

Do we really have to? Isn't the number quite obvious from the size of
the hotpluged memory?

> 
> > 2. Is this optional, in other words, can a device driver decide to not
> > to it like that?
> 
> Right now, is a per arch setup.
> For example, x86_64/powerpc/arm64 will do it inconditionally.
> 
> If we want to restrict this a per device-driver thing, I guess that we could
> allow to pass a flag to add_memory()->add_memory_resource(), and there
> unset MHP_MEMMAP_FROM_RANGE in case that flag is enabled.

I believe we will need to make this opt-in. There are some usecases
which hotplug an expensive (per size) memory via hotplug and it would be
too wasteful to use it for struct pages. I haven't bothered to address
that with my previous patches because I just wanted to make the damn
thing work first.
-- 
Michal Hocko
SUSE Labs