[PATCH 3/4] mm: simplify device private page handling in hmm_range_fault

[Christoph Hellwig]

On Thu, Mar 19, 2020 at 09:03:45PM -0300, Jason Gunthorpe wrote:
> > Should tests enable the feature or the feature enable the test?
> > IMHO, if the feature is being compiled into the kernel, that should
> > enable the menu item for the test. If the feature isn't selected,
> > no need to test it :-)
> 
> I ment if DEVICE_PRIVATE should be a user selectable option at all, or
> should it be turned on when a driver like nouveau is selected.

I don't think it should be user selectable.  This is an implementation
detail users can't know about.

> Is there some downside to enabling DEVICE_PRIVATE?

The option itself adds a little more code to the core kernel, and
introduces a few additional branches in core mm code.

But more importantly it pulls in the whole pgmap infrastructure.