Re: [RFC] depopulate_range_driver_managed() for removing page-table mappings for hot-added memory blocks

[David Hildenbrand]

On 14.11.20 11:28, Christoph Hellwig wrote:
> On Fri, Nov 13, 2020 at 06:41:22PM -0800, Sudarshan Rajagopalan wrote:
> > Hello,
> >
> > When memory blocks are removed, along with removing the memmap entries,
> > memory resource and memory block devices, the arch specific
> > arch_remove_memory() is called which takes care of tearing down the
> > page-tables.
> >
> > Suppose there???s a usecase where the removed memory blocks will be added
> > back into the system at later point,
>
> Stop this crap.  If you have a use case post the actual use here,
> including code and specs.  If you don't have that don't waste peoples
> time on your pipe dream.

Adding to that, we (upstream people) are not interested in introducing 
and maintaining kernel interfaces/infrastructure only relevant to 
out-of-tree / proprietary drivers. If you're planning on upstreaming, 
then please share the actual problem you are trying to solve and the 
issues you see with the current interfaces/infrastructure, best along 
with an RFC that shows the bigger picture.

Having that said, what you describe sounds more like a special kind of 
allocation, whereby the linear mapping is removed, similar to [1].

[1] https://lkml.kernel.org/r/20201111145322.15793-1-david@xxxxxxxxxx


--
Thanks,

David / dhildenb