This is a HUGE commit, but it's not nearly as bad as it looks - any problems
can be isolated to a particular chipset and engine combination.  It was
simply too difficult to port each one at a time, the compat layers are
*already* ridiculous.

Most of the changes here are simply to the glue, the process for each of the
engine modules was to start with a standard skeleton and copy+paste the old
code into the appropriate places, fixing up variable names etc as needed.

v2: Marcin Slusarz <marcin.slusarz@gmail.com>
- fix find/replace bug in license header

v3: Ben Skeggs <bskeggs@redhat.com>
- bump indirect pushbuf size to 8KiB, 4KiB barely enough for userspace and
  left no space for kernel's requirements during GEM pushbuf submission.
- fix duplicate assignments noticed by clang

v4: Marcin Slusarz <marcin.slusarz@gmail.com>
- add sparse annotations to nv04_fifo_pause/nv04_fifo_start
- use ioread32_native/iowrite32_native for fifo control registers

v5: Ben Skeggs <bskeggs@redhat.com>
- rebase on v3.6-rc4, modified to keep copy engine fix intact
- nv10/fence: unmap fence bo before destroying
- fixed fermi regression when using nvidia gr fuc
- fixed typo in supported dma_mask checking
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

v2 (Ben Skeggs):
- some fixes for 64KiB PAGE_SIZE
- fix porting issues in (currently unused) nv41/nv44 pciegart code
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Having GPUOBJ and VM intertwined like this makes it *really* hard to
continue porting to the new driver architecture, split it out in
favour of requiring explit maps be the caller.

It's more flexible and obvious this way anyway...
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Future work will be headed in the way of separating the policy supplied by
the nouveau drm module from the mechanisms provided by the driver core.

There will be a couple of major classes (subdev, engine) of driver modules
that have clearly defined tasks, and the further directory structure change
is to reflect this.

No code changes here whatsoever, aside from fixing up a couple of include
file pathnames.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

It may not be necessary to fail in certain cases (such as failing to idle)
on module unload, whereas on suspend it's important to ensure a consistent
state can be restored on resume.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Abuses existing gpuobj_new() chan argument for this, which in turn forces
all NVOBJ_FLAG_VM allocations to be done from the global heap, not
suballocated from the channel's private heap.  Not a problem though in
practise.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

We need the physical VRAM address in vinst, even for objects mapped into
a vm, as the gpuobj suspend/resume code uses PMEM to access the object.

Previously, vinst was overloaded to mean "VRAM address" for !VM objects,
and "VM address" for VM objects, causing the wrong data to be accessed
during suspend/resume.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Signed-off-by: NBen Skeggs <bskeggs@redhat.com>