If a memory allocation fails when using the DMA allocator,
gk20a_instobj_dtor_dma() will be called on the failed instmem object.
At this time, node->handle might not be NULL despite the call to
dma_alloc_attrs() having failed. node->cpuaddr is the right member to
check for such a failure, so use it instead.
Reported-by: NVince Hsu <vinceh@nvidia.com>
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

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

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

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

User-space use mappable BOs notably for fences, and expects that a
value update by the GPU will be immediatly visible through the
user-space mapping.

ARM has a property that may prevent this from happening though: memory
can be mapped multiple times only if the different mappings share the
same caching properties. However all the lowmem memory is already
identity-mapped into the kernel with cache enabled, so when user-space
requests an uncached mapping, we actually get an "undefined caching
policy" one and this has strange side-effects described on Freedesktop
bug 86690.

To prevent this from happening, allow user-space to explicitly specify
which objects should be coherent, and create such objects with the
TTM_PL_FLAG_UNCACHED flag. This will make TTM allocate memory using the
DMA API, which will fix the identify mapping and allow us to safely map
the objects to user-space uncached.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Reviewed-by: NLucas Stach <dev@lynxeye.de>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Let GK20A's instmem take advantage of the IOMMU if it is present. Having
an IOMMU means that instmem is no longer allocated using the DMA API,
but instead obtained through page_alloc and made contiguous to the GPU
by IOMMU mappings.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Tegra SoCs have an IOMMU that can be used to present non-contiguous
physical memory as contiguous to the GPU and maximize the use of large
pages in the GPU MMU, leading to performance gains. This patch adds
support for probing such a IOMMU if present and make its properties
available in the nouveau_platform_gpu structure so subsystems can take
advantage of it.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

instmem for GK20A is allocated using dma_alloc_coherent(), which
provides us with a coherent CPU mapping that we never use because
instmem objects are accessed through PRAMIN. Switch to
dma_alloc_attrs() which gives us the option to dismiss that CPU mapping
and free up some CPU virtual space.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Now that Nouveau can operate even when there is no RAM device, remove
the dummy one used by GK20A.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

GK20A does not have dedicated RAM, thus having a RAM device for it does
not make sense. Move the contiguous physical memory allocation to
instmem.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Having a RAM device does not make sense for chips like GK20A which have
no dedicated video memory. The dummy RAM device that we used so far
works as a temporary band-aid, but in the longer term it is desirable
for the driver to be able to work without any kind of VRAM.

This patch adds a few conditionals in places where a RAM device was
assumed to be present and allows some more objects to be allocated from
the TT domain, allowing Nouveau to handle GPUs for which
pfb->ram == NULL.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Notify interrupt is only used for cyclestats. We can just clear it and
avoid an "unknown stat" error that gets printed to dmesg otherwise.
Signed-off-by: NLauri Peltonen <lpeltonen@nvidia.com>
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Other methods in this file suggest this is the correct way to retrieve
the engine pointer.
Signed-off-by: NLauri Peltonen <lpeltonen@nvidia.com>
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

We use -1 to mean "not read from hw yet"
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

This if statement is correct but it wasn't indented, so it looked like
some code was missing.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Spotted by coccinelle:
drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c:50:5-8: WARNING: end returns can be simpified
Signed-off-by: NMartin Peres <martin.peres@free.fr>
Reviewed-by: NTobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

imx-drm changes to use media bus formats and LDB drm_panel support

- Add media bus formats needed by imx-drm
- Switch to use media bus formats to describe the pixel format
  on the internal parallel bus between display interface and
  encoders
- Some preparations for TV Output via TVEv2 on i.MX5
- Add drm_panel support to the i.MX LVDS driver, allow to
  determine the bus pixel format from the panel descriptor.

* tag 'imx-drm-next-2015-03-31' of git://git.pengutronix.de/git/pza/linux:
  drm/imx: imx-ldb: allow to determine bus format from the connected panel
  drm/imx: imx-ldb: reset display clock input when disabling LVDS
  drm/imx: imx-ldb: add drm_panel support
  drm/imx: consolidate bus format variable names
  drm/imx: switch to use media bus formats
  Add RGB666_1X24_CPADHI media bus format
  Add YUV8_1X24 media bus format
  Add BGR888_1X24 and GBR888_1X24 media bus formats
  Add LVDS RGB media bus formats
  Add RGB444_1X12 and RGB565_1X16 media bus formats
  drm/imx: ipuv3-crtc: Allow to divide DI clock from TVEv2
  drm/imx: Add support for interlaced scanout

drm: Use of-graph helpers to loop over endpoints

Convert all drm callers that use of_graph_get_next_endpoint to loop over
of-graph endpoints to the newly introduced for_each_endpoint_of_node
helper macro.

* tag 'of-graph-drm-2015-04-08' of git://git.pengutronix.de/git/pza/linux:
  drm/rockchip: use for_each_endpoint_of_node macro, drop endpoint reference on break
  drm/rcar-du: use for_each_endpoint_of_node macro
  drm/imx: use for_each_endpoint_of_node macro in imx_drm_encoder_get_mux_id
  drm: use for_each_endpoint_of_node macro in drm_of_find_possible_crtcs
  of: Explicitly include linux/types.h in of_graph.h
  dt-bindings: brcm: rationalize Broadcom documentation naming
  of/unittest: replace 'selftest' with 'unittest'
  Documentation: rename of_selftest.txt to of_unittest.txt
  Documentation: update the of_selftest.txt
  dt: OF_UNITTEST make dependency broken
  MAINTAINERS: Pantelis Antoniou device tree overlay maintainer
  of: Add of_graph_get_port_by_id function
  of: Add for_each_endpoint_of_node helper macro
  of: Decrement refcount of previous endpoint in of_graph_get_next_endpoint

Merge branch 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos into drm-next

   This pull request contains just cleanup for atomic pageflip/modeset
   support, and some fixeups.

   We wanted to merge atomic pageflip/modeset feature support, new drivers
   - MIC and DECON for exynos5433 SoC - and relevant patches this time.
   However, I'd found that these features are not only safe enough
   but also aren't tested yet. So for them, I'd like to have enough times
   for the reviews.

* 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos:
  drm/exynos: Fix FIMD buffer size calculation
  drm/exynos: Enable DP clock to fix display on Exynos5250 and other
  drm/exynos: fimd: check whether exynos_drm_crtc_create succeed or not
  drm/exynos: dsi: remove the empty mode_valid callback
  drm/exynos: add ratio calculation
  drm/exynos: use src_x and src_y instead of fb_x and fb_y
  drm/exynos: mixer: add 2x scaling to mixer_graph_buffer
  drm/exynos: remove superfluous error messages
  drm/exynos: fix typos in hdmi and mixer
  drm/exynos/ipp: Validate buffer enqueue requests
  drm/exynos: track vblank events on a per crtc basis
  drm/exynos: remove leftover functions declarations
  drm/exynos: remove exynos_plane_destroy()
  drm/exynos: make zpos property immutable
  drm/exynos: preset zpos value for overlay planes
  drm/exynos: remove struct *_win_data abstraction on planes
  drm/exynos: remove unused exynos_crtc->win_enable() callback
  drm/exynos: fimd: fix alpha setting for XR24 pixel format

Commit adacb228 ("drm: Exynos: Respect framebuffer pitch for
FIMD/Mixer") fixed the buffer size calculation by using the FB
pitch value but later commit 26b9c2813ede1 ("drm/exynos: remove
struct *_win_data abstraction on planes") added a regression so
fix the buffer size calculation again.

Tested on Chromebook Snow / Peach Pit.

Fixes: 26b9c2813ede1 ("drm/exynos: remove struct *_win_data abstraction on planes")
Signed-off-by: NDaniel Stone <daniels@collabora.com>
Tested-by: NJavier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

After adding display power domain for Exynos5250 in commit
2d2c9a8d ("ARM: dts: add display power domain for exynos5250") the
display on Chromebook Snow and others stopped working after boot.

The reason for this suggested Andrzej Hajda: the DP clock was disabled.
This clock is required by Display Port and is enabled by bootloader.
However when FIMD driver probing was deferred, the display power domain
was turned off. This effectively reset the value of DP clock enable
register.

When exynos-dp is later probed, the clock is not enabled and display is
not properly configured:

exynos-dp 145b0000.dp-controller: Timeout of video streamclk ok
exynos-dp 145b0000.dp-controller: unable to config video

Fixes: 2d2c9a8d ("ARM: dts: add display power domain for exynos5250")
Cc: <stable@vger.kernel.org>
Signed-off-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Reported-by: NJavier Martinez Canillas <javier.martinez@collabora.co.uk>
Tested-by: NJavier Martinez Canillas <javier.martinez@collabora.co.uk>
Tested-by: NAndreas Färber <afaerber@suse.de>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

>From the commit "drm/exynos: fix the execution order in FIMD
initialization" (598285bfdce46d7c47632a2ba4b980f60be4a677), the error
checking code is removed improperly. This patch fix the regression.
Signed-off-by: NHyungwon Hwang <human.hwang@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

Because the helper function which calls this callback checks whether
it is registered or not. It is not necessary if it does nothing.
So it would be better to remove the function for clarity.
Signed-off-by: NHyungwon Hwang <human.hwang@samsung.com>
Reviewed-by: NAndrzej Hajda <a.hajda@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

Calculation ratio from exynos_drm plane codes, then each hw drivers can
use it without extra operation. Also this fixes width and height of
source used for actual crtc shown via screen.
Signed-off-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

It's more reasonable to use src_x and src_y to represent source as
counterpart of destination(crtc). Already we are using src_width and
src_height for width and height of source.
Signed-off-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

While the VP (video processor) supports arbitrary scaling
of its input, the mixer just supports a simple 2x (line
doubling) scaling. Expose this functionality and exit
early when an unsupported scaling configuration is
encountered.

This was tested with modetest's DRM plane test (from
the libdrm test suite) on an Odroid-X2 (Exynos4412).

v2: Put if- and return-statement on different lines.
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NTobias Jakobi <tjakobi@math.uni-bielefeld.de>
Acked-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

The messages are redundant since 'check_fb_gem_memory_type'
already prints out exactly the same string when it fails.
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NTobias Jakobi <tjakobi@math.uni-bielefeld.de>
Acked-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

Use the correct spelling for 'progressive'.
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NTobias Jakobi <tjakobi@math.uni-bielefeld.de>
Acked-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

As for now there is no validation of incoming buffer
enqueue request as far as the gem buffers are being
concerned. This might lead to some undesired cases
when the driver tries to operate on invalid buffers
(wiht no valid gem object handle i.e.).
Add some basic checks to rule out those potential issues.
Signed-off-by: NBeata Michalska <b.michalska@samsung.com>
[mszyprow: rebased onto v4.0-rc1 and adapted to recent ipp changes]
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

The goal of the change is to make sure we send the vblank event on the
current vblank. My hope is to fix any races that might be causing flicker.
After this change I only see a flicker in the transition plymouth and
X11.

Simplified the code by tracking vblank events on a per-crtc basis. This
allowed me to remove all error paths from the callback. It also allowed
me to remove the vblank wait from the callback.
Signed-off-by: NMandeep Singh Baines <msb@chromium.org>
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

These functions were already removed by previous cleanup work, but these
ones were left behind.
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Acked-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

The .destroy() callback for exynos can be replaced by drm_plane_cleanup().
The only extra operation on exynos_plane_destroy() was a call to
exynos_plane_disable() but the plane is already disabled by a earlier call
to drm_framebuffer_remove().
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

We already set each plane zpos at init, after that changes to zpos are
not expected. This patch turns zpos into a read-only property so now it is
impossible to set zpos.
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

Usually userspace don't want to have two overlay planes on the same zpos
so this change assign a different zpos for each plane. Before this change
a zpos of value zero was created for all planes so the userspace had to
set up the zpos of every plane it wanted to use.

Also all places that were storing zpos positions are now unsigned int.
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

struct {fimd,mixer,vidi}_win_data was just keeping the same data
as struct exynos_drm_plane thus get ride of it and use exynos_drm_plane
directly.

It changes how planes are created and remove .win_mode_set() callback
that was only filling all *_win_data structs.

v2: check for return of exynos_plane_init()
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

None of the exynos crtc drivers implements win_enable() so remove it for
better clarity of the code.
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

XR24 planes were not shown properly, so now set the right registers
to correctly enable displaying these planes.

It also moves the alpha register settings to fimd_win_set_pixfmt()
to keep all pixel format stuff together.

v2: remove leftover var alpha
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NInki Dae <inki.dae@samsung.com>

Linux 4.0-rc7

Requested by Alex for fixes -next needs.

Conflicts:
	drivers/gpu/drm/i915/intel_sprite.c

Using the for_each_... macro should make the code a bit shorter and
easier to read. Also, when breaking out of the loop, the endpoint node
reference count needs to be decremented.
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: NDaniel Kurtz <djkurtz@chromium.org>

Using the for_each_... macro should make the code a bit shorter and
easier to read.
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>