This patch adds one option for the boot config strings "NvClkMode*", so
that we can enable the "auto" mode when loading module.
Signed-off-by: Vince Hsu <vinceh@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

This patch adds PWR support for GK20A. But instead of adding the PWR
features like firmware loading and communication with PMU firmware, we
add the DVFS (Dynamic Voltage and Frequency Scaling), which is one of
the PMU firmware's jobs on dGPUs, in this patch. This refers to the
idle signals provided by the NVIDIA hardware and tries to adjust the
performance level based on the calculated target. The reclocking policy
can be fine-tuned later when we have more real use cases.
Signed-off-by: Vince Hsu <vinceh@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

The platform device does not use the common nouveau_pwr_init() to initialize
the PWR, but it does need the .pgob() be assigned to avoid NULL pointer
dereference in graph/nve4.c.
Signed-off-by: Vince Hsu <vinceh@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

There might be some callers of nouveau_clock_astate(), and they are from
inetrrupt context. So we must ensure that this function can be atomic in
that condition. This patch adds one parameter which is subsequently passed
to nouveau_pstate_calc(). Therefore we can choose whether we want to wait
for the pstate work's completion or not.
Signed-off-by: Vince Hsu <vinceh@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

Several braces were misplaced unintentionally. That caused the msi handling
became part of the default case of the first switch statement. So add the
missing ones.
Signed-off-by: Vince Hsu <vinceh@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

Looks like a userspace bug can trigger this somehow during a mode
switch, causing: EVO complaint -> semaphores get out of sync ->
entire display stalled.

We likely want to be even stricter than this (or at least deal
better if EVO rejects our request), but I'll save that for the
drm_plane/atomic conversion and just fix the bug that I already
know can be triggered.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

On NV50 and up, we'll allow fixed panels to use EDID-provided modes
without the GPU scaler, and force scaling (even for NONE) otherwise.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

Should be the same defaults as before, just easier to follow.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

The hilarious part is that, under X, this won't work anyway because the
server decides to construct its own modes for some reason.

Tested with modetest, which isn't quite as insane.  I'd hope that
wayland is more sensible.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

Common programming sense dictates that resources allocated by a function
are freed by this function should it fails, but this is not the case for
the allocated structure of nouveau_sgdma_create_ttm(). It seems that
n00b contributors attempt to fix this one like bugs flying towards a bug
zapper, so add a comment to hopefully prevent this from happening
anymore.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

nouveau_sgdma_be::dev is only set once during init and never used
anywhere, so remove it.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

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

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

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

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

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

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

It's about to not be valid for objects that aren't in the client
object tree.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

These are a tad more complex than a direct cast with paranoia safeties.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

The [  SUBDEV] specified in log output will be a bit different for
children of a subdev now.  Previously this reports whatever subdev
is specified by object.engine, now it reports the subdev that owns
the object (so, up object.parent somewhere).

Later patches will append object and class identifiers to messages,
which will help clarify where it's coming from.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

Makes the output slightly less useful, in that objects with the same
class handle can't be distinguished from each other now.

Upcoming commits will name objects with user-readable strings to fix
this problem.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

gpuobj has a condition of (bar && bar->alloc) around usage to avoid
some nasty ordering issues (which, i've now been reminded to add a
todo about fixing...) between bar and vm.

The bar->alloc part of the condition isn't currently necessary (it
used to be, another change made bar always NULL where it matters),
so we got lucky.  That won't be the case for much longer.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

Makes things a bit more readable.  This is specially important now as
upcoming commits are going to be gradually removing the use of macros
for down-casts, in favour of compile-time checking.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

Has additional safeties for one.  For two, needed for an upcoming
commit that removes abuse of nouveau_object.engine.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

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

Backmerge Linus tree after rc5 + drm-fixes went in.

There were a few amdkfd conflicts I wanted to avoid,
and Ben requested this for nouveau also.

Conflicts:
	drivers/gpu/drm/amd/amdkfd/Makefile
	drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
	drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
	drivers/gpu/drm/amd/amdkfd/kfd_priv.h
	drivers/gpu/drm/amd/include/kgd_kfd_interface.h
	drivers/gpu/drm/i915/intel_runtime_pm.c
	drivers/gpu/drm/radeon/radeon_kfd.c

Merge branch 'drm-sti-next-add-dvo' of git://git.linaro.org/people/benjamin.gaignard/kernel into drm-next

This patch enable the last big hardware feature of my driver: the
connector for panel.
Like for HMDI and HDA, Digital Video Out (DVO) create brige, encoder
and connector
drm objects.

* 'drm-sti-next-add-dvo' of git://git.linaro.org/people/benjamin.gaignard/kernel:
  drm: sti: add DVO output connector

Add atmel HLCDC driver.

* tag 'atmel-hlcdc-drm-3.20' of https://github.com/bbrezillon/linux-at91:
  drm: add DT bindings documentation for atmel-hlcdc-dc driver
  drm: add Atmel HLCDC Display Controller support
  drm: panel: simple-panel: add bus format information for foxlink panel
  drm: panel: simple-panel: add support for bus_format retrieval
  drm: add bus_formats and num_bus_formats fields to drm_display_info

- Infrastructure work in amdkfd to prepare for VI support. This work mainly
  includes separating modules into ASIC-specific functionality, adding
  new properties that are relevant for VI, making sure that shared code is
  reused, etc.

- Improve mechanism of submitting packets to HIQ (the kernel queue that amdkfd
  uses to issue commands to the GPU). The driver used to verify that each CS
  was read by the GPU. However, this proved to be both unnecessary and erroneous.
  Therefore, we cancelled this verification.

- Moved initialization of compute VMIDs into radeon driver

- Various minor fixes

* tag 'drm-amdkfd-next-2015-01-21' of git://people.freedesktop.org/~gabbayo/linux: (22 commits)
  drm/amdkfd: Fix description of sched_policy module parameter
  drm/amdkfd: Remove sync_with_hw() from amdkfd
  drm/amdkfd: Remove unused function busy_wait()
  drm/amdkfd: Replace cpu_relax() with schedule() in DQM
  drm/amdkfd: Fix for-loop when allocating HQD (non-HWS)
  drm/amdkfd: Add initial VI support for KQ
  drm/amdkfd: Encapsulate KQ functions in ops structure
  drm/amdkfd: Add initial VI support for DQM
  drm/amdkfd: Encapsulate DQM functions in ops structure
  drm/amdkfd: Don't BUG on freeing GART sub-allocation
  drm/amdkfd: Fix logic of destroy_queue_nocpsch()
  MAINTAINERS: Update amdkfd files
  drm/amdkfd: Change MQD manager to be H/W specific
  drm/amdkfd: Add asic property to kfd_device_info
  drm/amdkfd: Make KFD_MQD_TYPE enum types H/W agnostic
  drm/amdkfd: Add new VI-specific queue properties
  drm/radeon: Use new cik_structs.h file
  drm/amdkfd: Don't include header files from radeon
  drm/amd: Put cik structures in a common place
  drm/radeon: Don't use relative paths in #include
  ...

Pull superh tracing fix from Steven Rostedt:
 "It's been reported that function tracing does not work on the sh
  architecture because gcc 4.8 for superH does not support -m32, and the
  recordmcount.pl script adds "-m32" when re-compiling the object files
  with the mcount locations.

  I was not able to reproduce this problem, as it seems that -m32 works
  fine for my cross compiler gcc 4.6.3, but I have to assume that -m32
  was deprecated somewhere between 4.6 and 4.8.  As it still seems to
  compile fine without -m32, I have no reason not to add this patch, as
  having -m32 seems to cause trouble for others"

* tag 'trace-sh-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  scripts/recordmcount.pl: There is no -m32 gcc option on Super-H anymore

The Atmel HLCDC (HLCD Controller) IP available on some Atmel SoCs (i.e.
at91sam9n12, at91sam9x5 family or sama5d3 family) provides a display
controller device.

The HLCDC block provides a single RGB output port, and only supports LCD
panels connection to LCD panels for now.

The atmel,panel property link the HLCDC RGB output with the LCD panel
connected on this port (note that the HLCDC RGB connector implementation
makes use of the DRM panel framework).

Connection to other external devices (DRM bridges) might be added later by
mean of a new atmel,xxx (atmel,bridge) property.
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>

The Atmel HLCDC (HLCD Controller) IP available on some Atmel SoCs (i.e.
at91sam9n12, at91sam9x5 family or sama5d3 family) provides a display
controller device.

This display controller supports at least one primary plane and might
provide several overlays and an hardware cursor depending on the IP
version.

At the moment, this driver only implements an RGB connector to interface
with LCD panels, but support for other kind of external devices might be
added later.
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Tested-by: Anthony Harivel <anthony.harivel@emtrion.de>
Tested-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>

Foxlink's fl500wvr00-a0t supports RGB888 format.
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Thierry Reding <treding@nvidia.com>

Provide a way to specify panel requirement in terms of supported media bus
format (particularly useful for panels connected to an RGB or LVDS bus).
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Thierry Reding <treding@nvidia.com>

Add bus_formats and num_bus_formats fields and
drm_display_info_set_bus_formats helper function to specify the bus
formats supported by a given display.

This information can be used by display controller drivers to configure
the output interface appropriately (i.e. RGB565, RGB666 or RGB888 on raw
RGB or LVDS busses).
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Thierry Reding <treding@nvidia.com>