[Intel-gfx] New API for creating bo from user pages

[Chris Wilson]

This is a fairly hairy piece of API - the implications of transforming
an ordinary portion of process memory into asynchronous DMA arena for
the GPU opens up a can of worms regarding memory protection and
inter-process sharing. On the other hand, it is a distinct advantage of
UMA that we have failed to take advantage of (since the beginning).

Many thanks to Tvrtko Ursulin for providing an i-g-t test case that
seems to have shaken out most of the bugs.
-Chris