[Intel-gfx] Pending i915 fixes

[Chris Wilson]

Whilst this may appear to be a big batch, it is actually only a few bug
fixes from the last couple of weeks broken down into small steps.

1,2: Seem to be well tested now and fix the condition where the pipes
are still active and reading from memory as we rewrite the GTT. This
impacts both boot and resume.
3: A simple memleak on an error path
4: Defend against left-over state in the DP registers causing an OOPS on
boot.
5: A fix for Cantiga TV detection reverted for unknown reasons.
6: A debugging patch that has caught some real bugs, just on the limited
sample of the machines I have...
7: One of the aforementioned bugs: Enable planes after attaching the fb.
   This conflicts with the modesetting refactoring branch, so I'd prefer
   to get this into -fixes asap and then rebase -next on top.
   However, needs extensive testing just in case it was intentional!
8-16: The other bug. Using an invalid block of memory for load detection
17: Dave's multi-gpu fix for X server (or other drm client) shutdown.

These are available on the -staging branch for testing.
Kindly review, thanks.
-Chris