Re: [PATCHv5 05/19] mm/page_alloc: Handle allocation for encrypted memory

[Michal Hocko]

On Thu 19-07-18 11:27:24, Kirill A. Shutemov wrote:
> On Wed, Jul 18, 2018 at 04:03:53PM -0700, Dave Hansen wrote:
> > I asked about this before and it still isn't covered in the description:
> > You were specifically asked (maybe in person at LSF/MM?) not to modify
> > allocator to pass the keyid around.  Please specifically mention how
> > this design addresses that feedback in the patch description.
> > 
> > You were told, "don't change the core allocator", so I think you just
> > added new functions that wrap the core allocator and called them from
> > the majority of sites that call into the core allocator.  Personally, I
> > think that misses the point of the original request.
> > 
> > Do I have a better way?  Nope, not really.
> 
> +Michal.
> 
> IIRC, Michal was not happy that I propagate the KeyID to very core
> allcoator and we've talked about wrappers around existing APIs as a better
> solution.
> 
> Michal, is it correct?

Yes that is the case. I haven't seen this series and unlikely will get
to it in upcoming days though so I cannot comment much more
unfortunately.
-- 
Michal Hocko
SUSE Labs