deliverable/linux.git
9 years agoMerge tag 'topic/core-stuff-2014-11-28' of git://anongit.freedesktop.org/drm-intel...
Dave Airlie [Mon, 1 Dec 2014 22:59:34 +0000 (08:59 +1000)] 
Merge tag 'topic/core-stuff-2014-11-28' of git://anongit.freedesktop.org/drm-intel into drm-next

So here's a pile of atomic fixes and improvements from various people.
There's still more patches in-flight, so I think I'll keep collecting them
in a separate branch.

* tag 'topic/core-stuff-2014-11-28' of git://anongit.freedesktop.org/drm-intel:
  drm/atomic: clear plane's CRTC and FB when shutting down
  drm: Handle atomic state properly in kms getfoo ioctl
  drm: use mode_object_find helpers
  drm: fix indentation
  drm/msm: switch to atomic-helpers iterator macros
  drm/atomic: add plane iterator macros
  drm/atomic: track bitmask of planes attached to crtc
  drm: Free atomic state during cleanup
  drm: Make drm_atomic.h standalone includible
  drm: Make drm_atomic_helper.h standalone includible
  drm/plane: Add missing kerneldoc
  drm/plane: Pass old state to ->atomic_update()
  drm/atomic_helper: Cope with plane->crtc == NULL in disable helper
  drm/atomic: Drop per-plane locking TODO
  drm/atomic-helper: Skip vblank waits for unchanged fbs
  drm: Document that drm_dev_alloc doesn't need a parent

9 years agoARM: shmobile: marzen: Remove DU platform device
Laurent Pinchart [Thu, 27 Nov 2014 15:19:13 +0000 (17:19 +0200)] 
ARM: shmobile: marzen: Remove DU platform device

Platform data support has been removed from the DU driver, drop DU
support from the legacy Marzen board file. The multiplatform DT-based
Marzen support should be used instead.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agoARM: shmobile: lager: Remove DU platform device
Laurent Pinchart [Thu, 27 Nov 2014 15:19:12 +0000 (17:19 +0200)] 
ARM: shmobile: lager: Remove DU platform device

Platform data support has been removed from the DU driver, drop DU
support from the legacy Lager board file. The multiplatform DT-based
Lager support should be used instead.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agodrm/atomic: clear plane's CRTC and FB when shutting down
Rob Clark [Wed, 26 Nov 2014 23:58:04 +0000 (18:58 -0500)] 
drm/atomic: clear plane's CRTC and FB when shutting down

Otherwise we'd still end up w/ the plane attached to the CRTC, and
seemingly active, but without an FB.  Which ends up going *boom*
in the drivers.

Slightly modified version of Daniel's irc suggestion.

Note that the big problem isn't drivers going *boom* here (since we
already have the situation of planes being left enabled when the crtc
goes down). The real issue is that the core assumes the primary plane
always goes down when calling ->set_config with a NULL mode. Ignoring
that assumption leads to the legacy state pointers plane->fb/crtc
getting out of sync with atomic, and that then leads to the subsequent
*boom* all over the place.

CC: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Rob Clark <robdclark@gmail.com>
[danvet: Drop my opinion of what's going sidewides here into the
commit message as a note.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm: Handle atomic state properly in kms getfoo ioctl
Daniel Vetter [Tue, 25 Nov 2014 22:50:05 +0000 (23:50 +0100)] 
drm: Handle atomic state properly in kms getfoo ioctl

So the problem with async commit (especially async modeset commit) is
that the legacy pointers only get updated after the point of no
return, in the async part of the modeset sequence. At least as
implemented by the current helper functions. This is done in the
set_routing_links function in drm_atomic_helper.c.

Which also means that access isn't protected by locks but only
coordinated by synchronizing with async workers. No problem thus far,
until we lock at the getconnector/encoder ioctls.

So fix this up by adding special cases for atomic drivers: For those
we need to look at state objects. Unfortunately digging out the
correct encoder->crtc link is a bit of work, so wrap this up in a
helper function.

Moving the assignments of connector->encoder and encoder->crtc earlier
isn't a good idea because the point of the atomic helpers is that we
stage the state updates. That way the disable functions can still
inspect the links and rely upon them.

v2: Extract full encoder->crtc lookup into helper (Rob).

v3: Extract drm_connector_get_encoder too since - we need to always
return state->best_encoder when there is a state otherwise we might
return stale data if there's a pending async disable (and chase
unlocked pointers, too). Same issue with encoder_get_crtc but there
it's a bit more tricky to handle.

Cc: Rob Clark <robdclark@gmail.com>
Cc: Sean Paul <seanpaul@chromium.org>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Lightly-Tested-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
9 years agodrm: use mode_object_find helpers
Rob Clark [Wed, 26 Nov 2014 01:33:11 +0000 (20:33 -0500)] 
drm: use mode_object_find helpers

Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm: fix indentation
Rob Clark [Wed, 26 Nov 2014 01:33:10 +0000 (20:33 -0500)] 
drm: fix indentation

Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/msm: switch to atomic-helpers iterator macros
Rob Clark [Wed, 26 Nov 2014 01:29:47 +0000 (20:29 -0500)] 
drm/msm: switch to atomic-helpers iterator macros

Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/atomic: add plane iterator macros
Rob Clark [Wed, 26 Nov 2014 01:29:46 +0000 (20:29 -0500)] 
drm/atomic: add plane iterator macros

Add helper macros to iterate the current, or incoming set of planes
attached to a crtc.  These helpers are only available for drivers
converted to use atomic-helpers.

Signed-off-by: Rob Clark <robdclark@gmail.com>
[danvet: Squash in fixup from Rob to move the planemask iterator to
drm_crtc.h and document it. That one is needed by the atomic ioctl so
can't be in a helper library.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/atomic: track bitmask of planes attached to crtc
Rob Clark [Fri, 21 Nov 2014 20:28:31 +0000 (15:28 -0500)] 
drm/atomic: track bitmask of planes attached to crtc

Chasing plane->state->crtc of planes that are *not* part of the same
atomic update is racy, making it incredibly awkward (or impossible) to
do something simple like iterate over all planes and figure out which
ones are attached to a crtc.

Solve this by adding a bitmask of currently attached planes in the
crtc-state.

Note that the transitional helpers do not maintain the plane_mask.  But
they only support the legacy ioctls, which have sufficient brute-force
locking around plane updates that they can continue to loop over all
planes to see what is attached to a crtc the old way.

Signed-off-by: Rob Clark <robdclark@gmail.com>
[danvet:
- Drop comments about locking in set_crtc_for_plane since they're a
  bit misleading - we already should hold lock for the current crtc.
- Also WARN_ON if get_state on the old crtc fails since that should
  have been done already.
- Squash in fixup to check get_plane_state return value, reported by
  Dan Carpenter and acked by Rob Clark.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agoMerge branch 'drm/du/adv7511' of git://linuxtv.org/pinchartl/fbdev into drm-next
Dave Airlie [Wed, 26 Nov 2014 22:36:19 +0000 (08:36 +1000)] 
Merge branch 'drm/du/adv7511' of git://linuxtv.org/pinchartl/fbdev into drm-next

The branch is based on a merge of drm-next and Simon's tags/renesas-dt-du-for-
v3.19 available at
git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git, the latter
having been pulled in the ARM SoC tree for v3.19.

Compared to v1, I've rebased my branch on a later drm-next, added Julia's
error return code fix, and documented the "drm: Decouple EDID parsing from I2C
adapter" patch properly.

v1:
Here's a pull request that adds HDMI support to the R-Car DU driver, including
a new slave encoder driver for the adv7511.

* 'drm/du/adv7511' of git://linuxtv.org/pinchartl/fbdev:
  drm: Add adv7511 encoder driver
  video: Add ADV751[13] DT bindings documentation
  drm: Decouple EDID parsing from I2C adapter
  drm: rcar-du: Add HDMI encoder and connector support
  drm: rcar-du: Replace drm_encoder with drm_slave_encoder
  drm: rcar-du: Replace direct DRM encoder access with cast macro
  drm: rcar-du: Pass the encoder DT node to rcar_du_encoder_init()
  drm: rcar-du: Remove platform data support
  drm: rcar-du: fix error return code
  ARM: shmobile: koelsch: Enable DU device in DT
  ARM: shmobile: koelsch-reference: Remove DU platform device
  ARM: shmobile: lager: Enable DU device in DT
  ARM: shmobile: lager-reference: Remove DU platform device
  ARM: shmobile: marzen: Enable DU device in DT
  ARM: shmobile: dts: Add common file for AA104XD12 panel
  ARM: shmobile: r8a7791: Add DU node to device tree
  ARM: shmobile: r8a7790: Add DU node to device tree
  ARM: shmobile: r8a7779: Add DU node to device tree
  ARM: shmobile: Remove FSF address from copyright headers

9 years agoMerge branch 'amdkfd-next-3.19' of git://people.freedesktop.org/~gabbayo/linux into...
Dave Airlie [Wed, 26 Nov 2014 22:28:22 +0000 (08:28 +1000)] 
Merge branch 'amdkfd-next-3.19' of git://people.freedesktop.org/~gabbayo/linux into drm-next

- Fixes for sparse warnings
- Memory leak fixes
- Fix for deadlock between amdkfd and iommu

* 'amdkfd-next-3.19' of git://people.freedesktop.org/~gabbayo/linux:
  amdkfd: delete some dead code
  amdkfd: Fix memory leak of mqds on dqm fini
  amdkfd: fix an error handling bug in pqm_create_queue()
  amdkfd: fix some error handling in ioctl
  amdkfd: Remove DRM_AMDGPU dependency from Kconfig
  amdkfd: explicitely include io.h in kfd_doorbell.c
  amdkfd: Clear ctx cb before suspend
  amdkfd: Instead of using get function, use container_of
  amdkfd: use schedule() in sync_with_hw
  amdkfd: Fix memory leak on process deregistration
  amdkfd: add __iomem attribute to doorbell_ptr
  amdkfd: fence_wait_timeout() can be static
  amdkfd: is_occupied() can be static
  amdkfd: Fix sparse warnings in kfd_flat_memory.c
  amdkfd: pqm_get_kernel_queue() can be static
  amdkfd: test_kq() can be static
  amdkfd: Fix sparse warnings in kfd_topology.c
  amdkfd: Fix sparse warnings in kfd_chardev.c

9 years agodrm: Add adv7511 encoder driver
Lars-Peter Clausen [Mon, 5 Mar 2012 15:30:57 +0000 (16:30 +0100)] 
drm: Add adv7511 encoder driver

This patch adds a driver for the Analog Devices adv7511. The adv7511 is
a standalone HDMI transmitter chip. It features a HDMI output interface
on one end and video and audio input interfaces on the other.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agovideo: Add ADV751[13] DT bindings documentation
Laurent Pinchart [Wed, 24 Sep 2014 19:09:46 +0000 (22:09 +0300)] 
video: Add ADV751[13] DT bindings documentation

The ADV7511, ADV7511W and ADV7513 are HDMI audio and video transmitters
compatible with HDMI 1.4 and DVI 1.0. They're described in DT using the
OF graph bindings and a list of custom properties pertaining to the
input video bus configuration.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: Decouple EDID parsing from I2C adapter
Lars-Peter Clausen [Fri, 27 Apr 2012 09:11:58 +0000 (11:11 +0200)] 
drm: Decouple EDID parsing from I2C adapter

The drm_get_edid() function performs direct I2C accesses to read EDID
blocks, assuming that the monitor DDC interface is directly connected to
the I2C bus. It can't thus be used with HDMI encoders that control the
DDC bus and expose EDID blocks through a different interface.

Refactor drm_do_get_edid() to take a block read callback function
instead of an I2C adapter, and export it for direct use by drivers.

As in the general case the DDC bus is accessible by the kernel at the
I2C level, drivers must make all reasonable efforts to expose it as an
I2C adapter and use drm_get_edid() instead of abusing this function.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm: rcar-du: Add HDMI encoder and connector support
Laurent Pinchart [Sun, 30 Mar 2014 19:55:38 +0000 (21:55 +0200)] 
drm: rcar-du: Add HDMI encoder and connector support

SoCs that integrate the DU have no internal HDMI encoder, support
external encoders only.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: rcar-du: Replace drm_encoder with drm_slave_encoder
Laurent Pinchart [Sun, 30 Mar 2014 23:52:01 +0000 (01:52 +0200)] 
drm: rcar-du: Replace drm_encoder with drm_slave_encoder

DRM slave encoders require their associated struct drm_encoder instance
to be embedded in a struct drm_slave_encoder. This makes processing
encoders regardless of their types needlessly and painfully complex in
drivers that use a mix of slave encoders and custom encoders. Such a
driver will need to either create drm_slave_encoder instances that fake
their embedded encoder instance, or to turn all drm_encoder instances
into drm_slave_encoder instances.

Between the two evils, one must choose the lesser. Use drm_slave_encoder
everywhere.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: rcar-du: Replace direct DRM encoder access with cast macro
Laurent Pinchart [Sun, 30 Mar 2014 23:50:16 +0000 (01:50 +0200)] 
drm: rcar-du: Replace direct DRM encoder access with cast macro

Add a new macro to downcast an rcar_du_encoder pointer to a drm_encoder
pointer and use it. This prepares for the replacement of the
rcar_drm_encoder encoder field with a drm_slave_encoder.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: rcar-du: Pass the encoder DT node to rcar_du_encoder_init()
Laurent Pinchart [Wed, 17 Sep 2014 11:41:57 +0000 (14:41 +0300)] 
drm: rcar-du: Pass the encoder DT node to rcar_du_encoder_init()

The encoder DT node will be needed to register an external HDMI encoder.
Pass it to the rcar_du_encoder_init() function to prepare for HDMI
support.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: rcar-du: Remove platform data support
Laurent Pinchart [Tue, 16 Sep 2014 23:07:52 +0000 (02:07 +0300)] 
drm: rcar-du: Remove platform data support

All platforms now instantiate the DU through DT, platform data support
isn't needed anymore.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm: rcar-du: fix error return code
Julia Lawall [Sun, 23 Nov 2014 13:11:17 +0000 (14:11 +0100)] 
drm: rcar-du: fix error return code

Propagate the error code on failure.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@@
identifier ret; expression e1,e2;
@@
(
if (\(ret < 0\|ret != 0\))
 { ... return ret; }
|
ret = 0
)
... when != ret = e1
    when != &ret
*if(...)
{
  ... when != ret = e2
      when forall
 return ret;
}
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
9 years agodrm/udl: properly check for error pointers
Haixia Shi [Tue, 25 Nov 2014 20:04:03 +0000 (12:04 -0800)] 
drm/udl: properly check for error pointers

The drm_prime_pages_to_sg() function never returns NULL pointers, only
error pointers and valid pointers.

Signed-off-by: Haixia Shi <hshi@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agodrm/udl: handle page mapping in dmabuf export.
Haixia Shi [Tue, 25 Nov 2014 20:04:02 +0000 (12:04 -0800)] 
drm/udl: handle page mapping in dmabuf export.

Fixes dmabuf export failure with -E_NOMEM when the page is not mapped.

Signed-off-by: Haixia Shi <hshi@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agoMAINTAINERS: add maintainer for i.MX DRM driver
Philipp Zabel [Mon, 24 Nov 2014 15:33:35 +0000 (16:33 +0100)] 
MAINTAINERS: add maintainer for i.MX DRM driver

Add myself as the maintainer of the i.MX DRM driver.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agodrm: imx: Move imx-drm driver out of staging
Philipp Zabel [Mon, 24 Nov 2014 15:33:34 +0000 (16:33 +0100)] 
drm: imx: Move imx-drm driver out of staging

The imx-drm driver was put into staging mostly for the following reasons,
all of which have been addressed or superseded:
 - convert the irq driver to use linear irq domains
 - work out the device tree bindings, this lead to the common of_graph
   bindings being used
 - factor out common helper functions, this mostly resulted in the
   component framework and drm of_graph helpers.

Before adding new fixes, and certainly before adding new features,
move it into its proper place below drivers/gpu/drm.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
9 years agoMerge tag 'tags/renesas-dt-du-for-v3.19' into drm/next/adv7511-base
Laurent Pinchart [Tue, 25 Nov 2014 23:38:43 +0000 (01:38 +0200)] 
Merge tag 'tags/renesas-dt-du-for-v3.19' into drm/next/adv7511-base

Renesas ARM Based SoC DT DU Updates for v3.19

* Enable DU using DT on marzen/r8a7779, lager/r8a7790 and koelsch/r8a7791

9 years agoMerge branch 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daein...
Dave Airlie [Tue, 25 Nov 2014 23:19:36 +0000 (09:19 +1000)] 
Merge branch 'exynos-drm-next' of git://git./linux/kernel/git/daeinki/drm-exynos into drm-next

  Add Exynos4415 SoC support, some fixups and cleanups.

   Summary:
   - Resolve kernel lockup issue incurred by probe request in probe context.
     . For this, it moves all register codes of sub drivers into init function
       and adds component binding support for vidi driver.
   - Add Exynos4415 SoC support.
   - Make each manager and display object to be embedded
     in each driver context.
   - Fix and clean up FIMD and MIPI-DSI drivers.
   - Clean up unnecesary or wrong descriptions.
   - And trivial cleanups.

* 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos: (58 commits)
  drm/exynos: avoid leak if exynos_dpi_probe() fails
  drm/exynos: Fix exynos_dpi_remove() parameter
  drm/exynos: vidi: add component support
  drm/exynos: fix exynos_drm_component_del
  drm/exynos/ipp: fix error return code
  drm/exynos: clean up machine compatible string check
  drm/exynos: move Exynos platform drivers registration to init
  Revert "drm/exynos: fix null pointer dereference issue"
  drm/exynos/dpi: stop using display->ctx pointer
  drm/exynos/dpi: embed display into private context
  drm/exynos/dp: stop using display->ctx pointer
  drm/exynos/dp: embed display into private context
  drm/exynos/vidi: stop using display->ctx pointer
  drm/exynos/vidi: embed display into private context
  drm/exynos/hdmi: stop using display->ctx pointer
  drm/exynos/hdmi: embed display into private context
  drm/exynos/fimd: stop using manager->ctx pointer
  drm/exynos/fimd: embed manager into private context
  drm/exynos/vidi: stop using manager->ctx pointer
  drm/exynos/vidi: embed manager into private context
  ...

9 years agoamdkfd: delete some dead code
Dan Carpenter [Tue, 25 Nov 2014 16:43:29 +0000 (19:43 +0300)] 
amdkfd: delete some dead code

This is dead code.  We don't need to unbind here, we can just return
directly.

Reviewed-by: Oded Gabbay <oded.gabbay@amd.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
9 years agoamdkfd: Fix memory leak of mqds on dqm fini
Oded Gabbay [Tue, 25 Nov 2014 13:16:38 +0000 (15:16 +0200)] 
amdkfd: Fix memory leak of mqds on dqm fini

The mqds array members are not freed when dqm is uninitialized.

Reviewed-by: Ben Goz <Ben.Goz@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
9 years agodrm: Free atomic state during cleanup
Thierry Reding [Tue, 25 Nov 2014 11:09:49 +0000 (12:09 +0100)] 
drm: Free atomic state during cleanup

The current state of CRTCs, planes and connectors currently leaks during
DRM driver ->unload() unless drivers explicitly clean it up. Since there
is nothing driver-specific about it, that cleanup can be done within the
DRM core.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm: Make drm_atomic.h standalone includible
Thierry Reding [Tue, 25 Nov 2014 11:09:48 +0000 (12:09 +0100)] 
drm: Make drm_atomic.h standalone includible

This header file makes use of a bunch of structures declared in the
drm_crtc.h header file. Include that to make sure the drm_atomic.h
header can be included standalone.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm: Make drm_atomic_helper.h standalone includible
Thierry Reding [Tue, 25 Nov 2014 11:09:47 +0000 (12:09 +0100)] 
drm: Make drm_atomic_helper.h standalone includible

This header uses a bunch of declarations from the drm/drm_crtc.h header,
so make sure to include that as well so that drm_atomic_helper.h can be
included standalone.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/plane: Add missing kerneldoc
Thierry Reding [Tue, 25 Nov 2014 11:09:45 +0000 (12:09 +0100)] 
drm/plane: Add missing kerneldoc

The plane helpers aren't pulled into the DocBook yet, so these weren't
noticed.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/plane: Pass old state to ->atomic_update()
Thierry Reding [Tue, 25 Nov 2014 11:09:44 +0000 (12:09 +0100)] 
drm/plane: Pass old state to ->atomic_update()

In most situations it will be useful to have the old state passed to the
->atomic_update() callback. For example if a plane is being disabled the
new state's .crtc field will be NULL, but some drivers may rely on this
field to program the CRTCs registers.

v2: rename variable to old_plane_state and remove redundant comment as
suggested by Daniel Vetter, remove an Exynos hunk that doesn't apply to
drm-next and add a hunk for pending MSM mdp5 changes

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/atomic_helper: Cope with plane->crtc == NULL in disable helper
Jasper St. Pierre [Fri, 21 Nov 2014 03:59:15 +0000 (19:59 -0800)] 
drm/atomic_helper: Cope with plane->crtc == NULL in disable helper

The drm core can call the plane disable hook multiple times, which
means it can get called when plane->crtc is already NULL. That in turn
means we can't get at the implicit acquire ctx we use in the atomic
helpers for legacy entries points.

We could try to pass drm_modeset_legacy_acquire_ctx a drm_device
pointer so that it can cope with a NULL crtc. But that still doesn't
work since the cursor ioctls (remapped with the universal cursor plane
support code) only grabs the crtc locks. So the global acquire context
isn't set eitehr.

The real solution here would be to bite the bullet and wire up
explicit acquire context parameters to all relevant functions. We need
to do that anyway (to be able to get rid of some small allocations
which we can't cope with failing). But that's a lot of work and better
done once atomic has settled a bit.

So meanwhile just catch this case in the helper and bail out.

Signed-off-by: Jasper St. Pierre <jstpierre@mecheye.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
[danvet: Completely rewrite commit message and comment but keep
Jasper's logic and author credits since his patch is the only
short-term solution that works.]
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/atomic: Drop per-plane locking TODO
Daniel Vetter [Mon, 24 Nov 2014 19:51:21 +0000 (20:51 +0100)] 
drm/atomic: Drop per-plane locking TODO

I've forgotten to remove that in my per-plane locking patch.

Reported-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
9 years agodrm/atomic-helper: Skip vblank waits for unchanged fbs
Daniel Vetter [Mon, 24 Nov 2014 19:42:42 +0000 (20:42 +0100)] 
drm/atomic-helper: Skip vblank waits for unchanged fbs

Especially with legacy cursor ioctls existing userspace assumes that
you can pile up lots of updates in one go. The super-proper way to
support this would be a special commit mode which overwrites the last
update. But getting there will be quite a bit of work.

Meanwhile do what pretty much all the drivers have done for the plane
update functions: Simply skip the vblank wait for the buffer cleanup
if the buffer is the same. Since the universal cursor plane code will
not recreate framebuffers needlessly this allows us to not slow down
legacy pageflip events while someone moves the cursor around.

v2: Drop the async plane update hunk from a previous attempt at this
issue.

v3: Fix up kerneldoc.

v4: Don't oops so badly. Reported by Jasper.

Cc: Rob Clark <robdclark@gmail.com>
Cc: "Jasper St. Pierre" <jstpierre@mecheye.net>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Tested-by: Jasper St. Pierre <jstpierre@mecheye.net>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
9 years agodrm: Document that drm_dev_alloc doesn't need a parent
Daniel Vetter [Mon, 24 Nov 2014 19:01:58 +0000 (20:01 +0100)] 
drm: Document that drm_dev_alloc doesn't need a parent

Possible for purely virtual debug devices.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agoMerge branch 'msm-next' of git://people.freedesktop.org/~robclark/linux into drm...
Dave Airlie [Tue, 25 Nov 2014 12:10:53 +0000 (22:10 +1000)] 
Merge branch 'msm-next' of git://people.freedesktop.org/~robclark/linux into drm-next

Now that we have the bits needed for mdp5 atomic, here is the followup
pull request I mentioned.  Main highlights are:

1) mdp5 multiple crtc and public plane support (no more hard-coded mixer setup!)
2) mdp5 atomic conversion
3) couple atomic helper fixes for issues found during mdp5 atomic
debug (reviewed by danvet.. but he didn't plane to send an
atomic-fixes pull request so I agreed to tack them on to mine)

* 'msm-next' of git://people.freedesktop.org/~robclark/linux:
  drm/atomic: shutdown *current* encoder
  drm/atomic: check mode_changed *after* atomic_check
  drm/msm/mdp4: fix mixer setup for multi-crtc + planes
  drm/msm/mdp5: dpms(OFF) cleanups
  drm/msm/mdp5: atomic
  drm/msm: atomic fixes
  drm/msm/mdp5: remove global mdp5_ctl_mgr
  drm/msm/mdp5: don't use void * for opaque types
  drm/msm: add multiple CRTC and overlay support
  drm/msm/mdp5: set rate before enabling clk
  drm/msm/mdp5: introduce mdp5_cfg module
  drm/msm/mdp5: make SMP module dynamically configurable
  drm/msm/hdmi: remove useless kref
  drm/msm/mdp5: get the core clock rate from MDP5 config
  drm/msm/mdp5: use irqdomains

9 years agoamdkfd: fix an error handling bug in pqm_create_queue()
Dan Carpenter [Tue, 25 Nov 2014 10:24:51 +0000 (13:24 +0300)] 
amdkfd: fix an error handling bug in pqm_create_queue()

The call to kernel_queue_uninit(NULL) will trigger a BUG(), and also the
error code is incorrect.

Fixes: 45102048f77e ('amdkfd: Add process queue manager module')
Reviewed-by: Oded Gabbay <oded.gabbay@amd.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
9 years agoamdkfd: fix some error handling in ioctl
Dan Carpenter [Tue, 25 Nov 2014 10:21:30 +0000 (13:21 +0300)] 
amdkfd: fix some error handling in ioctl

There is a typo here so the errors from kfd_bind_process_to_device()
are not detected.

Reviewed-by: Oded Gabbay <oded.gabbay@amd.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
9 years agodrm/exynos: avoid leak if exynos_dpi_probe() fails
Gustavo Padovan [Mon, 24 Nov 2014 18:23:30 +0000 (16:23 -0200)] 
drm/exynos: avoid leak if exynos_dpi_probe() fails

The component must be deleted if the probe fails.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: Fix exynos_dpi_remove() parameter
Gustavo Padovan [Mon, 24 Nov 2014 17:19:49 +0000 (15:19 -0200)] 
drm/exynos: Fix exynos_dpi_remove() parameter

exynos_dpi_remove() should receive a exynos_drm_display but when
DRM_EXYNOS_DPI was disabled it was receiving a struct device resulting in
ia compiler warning.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: vidi: add component support
Inki Dae [Mon, 24 Nov 2014 05:55:41 +0000 (14:55 +0900)] 
drm/exynos: vidi: add component support

This patch adds component support for vidi driver.

vidi driver is a kms driver so it doesn't need to be registered
to exynos_drm_subdrv_list. For this, it changes for the component
framework to be used for vidi driver.

This patch fixes below error also,

# echo 1 > /sys/devices/platform/exynos-drm-vidi/connection
[   55.618529] ------------[ cut here ]------------
[   55.621960] WARNING: CPU: 0 PID: 1397 at drivers/gpu/drm/drm_irq.c:1203 exynos_drm_crtc_dpms+0x88/0x17c()
[   55.631268] Modules linked in:
[   55.634278] CPU: 0 PID: 1397 Comm: sh Not tainted 3.18.0-rc2-146253-g31449d7 #1154
[   55.641885] [<c0014400>] (unwind_backtrace) from [<c0011570>] (show_stack+0x10/0x14)
[   55.649597] [<c0011570>] (show_stack) from [<c04764f4>] (dump_stack+0x84/0xc4)
[   55.656802] [<c04764f4>] (dump_stack) from [<c00218b8>] (warn_slowpath_common+0x6c/0x88)
[   55.664866] [<c00218b8>] (warn_slowpath_common) from [<c0021970>] (warn_slowpath_null+0x1c/0x24)
[   55.673632] [<c0021970>] (warn_slowpath_null) from [<c027a780>] (exynos_drm_crtc_dpms+0x88/0x17c)
[   55.682482] [<c027a780>] (exynos_drm_crtc_dpms) from [<c027a910>] (exynos_drm_crtc_commit+0x14/0x44)
[   55.691622] [<c027a910>] (exynos_drm_crtc_commit) from [<c025521c>] (drm_crtc_helper_set_mode+0x3d0/0x51c)
[   55.701233] [<c025521c>] (drm_crtc_helper_set_mode) from [<c0255d68>] (drm_crtc_helper_set_config+0x87c/0x9dc)
[   55.711230] [<c0255d68>] (drm_crtc_helper_set_config) from [<c026afa8>] (drm_mode_set_config_internal+0x58/0xd4)
[   55.721380] [<c026afa8>] (drm_mode_set_config_internal) from [<c025c208>] (restore_fbdev_mode+0xcc/0xec)
[   55.730834] [<c025c208>] (restore_fbdev_mode) from [<c025c244>] (drm_fb_helper_restore_fbdev_mode_unlocked+0x1c/0x30)
[   55.741424] [<c025c244>] (drm_fb_helper_restore_fbdev_mode_unlocked) from [<c025e0a8>] (drm_fb_helper_set_par+0x1c/0x60)
[   55.752271] [<c025e0a8>] (drm_fb_helper_set_par) from [<c025e174>] (drm_fb_helper_hotplug_event+0x88/0xc4)
[   55.761906] [<c025e174>] (drm_fb_helper_hotplug_event) from [<c02571c4>] (drm_helper_hpd_irq_event+0xc8/0x134)
[   55.771898] [<c02571c4>] (drm_helper_hpd_irq_event) from [<c028e27c>] (vidi_store_connection+0x90/0xc8)
[   55.781268] [<c028e27c>] (vidi_store_connection) from [<c0125f80>] (kernfs_fop_write+0xc0/0x180)
[   55.790045] [<c0125f80>] (kernfs_fop_write) from [<c00cdf60>] (vfs_write+0xa0/0x1ac)
[   55.797757] [<c00cdf60>] (vfs_write) from [<c00ce468>] (SyS_write+0x44/0x9c)
[   55.804790] [<c00ce468>] (SyS_write) from [<c000e6a0>] (ret_fast_syscall+0x0/0x30)
[   55.812328] ---[ end trace 3c0fe4386702d4dd ]---

This issue occurs when modeset to vidi is tried in case that drm_vblank_init
is called prior to crtc creation of vidi driver. In this case, crtc number
of vidi is invalid so any requests with the crtc number will fail.
This patch guarantees drm_vblank_init to be called after all kms drivers
are ready by using component framework.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fix exynos_drm_component_del
Inki Dae [Mon, 24 Nov 2014 07:58:48 +0000 (16:58 +0900)] 
drm/exynos: fix exynos_drm_component_del

This patch resolves the issue that component object isn't removed
correctly.

A given component object couldn't be placed to head of drm_component_list
so all component objects added to the drm_component_list should be checked
to remove the given component object.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/ipp: fix error return code
Julia Lawall [Sun, 23 Nov 2014 13:11:15 +0000 (14:11 +0100)] 
drm/exynos/ipp: fix error return code

Propagate the returned error code on failure.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@@
identifier ret; expression e1,e2;
@@
(
if (\(ret < 0\|ret != 0\))
 { ... return ret; }
|
ret = 0
)
... when != ret = e1
    when != &ret
*if(...)
{
  ... when != ret = e2
      when forall
 return ret;
}
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: clean up machine compatible string check
Inki Dae [Mon, 24 Nov 2014 08:08:00 +0000 (17:08 +0900)] 
drm/exynos: clean up machine compatible string check

Use 'for' statemant instead of hard-coded 'if' statement.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: move Exynos platform drivers registration to init
Gustavo Padovan [Mon, 24 Nov 2014 07:37:26 +0000 (16:37 +0900)] 
drm/exynos: move Exynos platform drivers registration to init

Registering the Exynos DRM subdevices platform drivers in the probe
function is causing an infinite loop. Fix this by moving it to the
exynos_drm_init() function to register the drivers on module init.

Registering drivers in the probe functions causes a deadlock in the parent
device lock. See Grant Likely explanation on the topic:

"I think the problem is that exynos_drm_init() is registering a normal
(non-OF) platform device, so the parent will be /sys/devices/platform.
It immediately gets bound against exynos_drm_platform_driver which
calls the exynos drm_platform_probe() hook. The driver core obtains
device_lock() on the device *and on the device parent*.

Inside the probe hook, additional platform_drivers get registered.
Each time one does, it tries to bind against every platform device in
the system, which includes the ones created by OF. When it attempts to
bind, it obtains device_lock() on the device *and on the device
parent*.

Before the change to move of-generated platform devices into
/sys/devices/platform, the devices had different parents. Now both
devices have /sys/devices/platform as the parent, so yes they are
going to deadlock.

The real problem is registering drivers from within a probe hook. That
is completely wrong for the above deadlock reason. __driver_attach()
will deadlock. Those registrations must be pulled out of .probe().

Registering devices in .probe() is okay because __device_attach()
doesn't try to obtain device_lock() on the parent."

 INFO: task swapper/0:1 blocked for more than 120 seconds.
       Not tainted 3.18.0-rc3-next-20141105 #794
 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
 swapper/0       D c052534c     0     1      0 0x00000000
 [<c052534c>] (__schedule) from [<c0525b34>] (schedule_preempt_disabled+0x14/0x20)
 [<c0525b34>] (schedule_preempt_disabled) from [<c0526d44>] (mutex_lock_nested+0x1c4/0x464

 [<c0526d44>] (mutex_lock_nested) from [<c02be908>] (__driver_attach+0x48/0x98)
 [<c02be908>] (__driver_attach) from [<c02bcc00>] (bus_for_each_dev+0x54/0x88)
 [<c02bcc00>] (bus_for_each_dev) from [<c02bdce0>] (bus_add_driver+0xe4/0x200)
 [<c02bdce0>] (bus_add_driver) from [<c02bef94>] (driver_register+0x78/0xf4)
 [<c02bef94>] (driver_register) from [<c029e99c>] (exynos_drm_platform_probe+0x34/0x234)
 [<c029e99c>] (exynos_drm_platform_probe) from [<c02bfcf0>] (platform_drv_probe+0x48/0xa4)
 [<c02bfcf0>] (platform_drv_probe) from [<c02be680>] (driver_probe_device+0x13c/0x37c)
 [<c02be680>] (driver_probe_device) from [<c02be954>] (__driver_attach+0x94/0x98)
 [<c02be954>] (__driver_attach) from [<c02bcc00>] (bus_for_each_dev+0x54/0x88)
 [<c02bcc00>] (bus_for_each_dev) from [<c02bdce0>] (bus_add_driver+0xe4/0x200)
 [<c02bdce0>] (bus_add_driver) from [<c02bef94>] (driver_register+0x78/0xf4)
 [<c02bef94>] (driver_register) from [<c029e938>] (exynos_drm_init+0x70/0xa0)
 [<c029e938>] (exynos_drm_init) from [<c00089b0>] (do_one_initcall+0xac/0x1f0)
 [<c00089b0>] (do_one_initcall) from [<c074bd90>] (kernel_init_freeable+0x10c/0x1d8)
 [<c074bd90>] (kernel_init_freeable) from [<c051eabc>] (kernel_init+0x8/0xec)
 [<c051eabc>] (kernel_init) from [<c000f268>] (ret_from_fork+0x14/0x2c)
 3 locks held by swapper/0/1:
  #0:  (&dev->mutex){......}, at: [<c02be908>] __driver_attach+0x48/0x98
  #1:  (&dev->mutex){......}, at: [<c02be918>] __driver_attach+0x58/0x98
  #2:  (&dev->mutex){......}, at: [<c02be908>] __driver_attach+0x48/0x98

Changelog v2:
- call platform_driver_register after all kms and non kms drivers are
  registered
- rebased it to exynos-drm-next

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agoRevert "drm/exynos: fix null pointer dereference issue"
Gustavo Padovan [Thu, 20 Nov 2014 23:42:55 +0000 (21:42 -0200)] 
Revert "drm/exynos: fix null pointer dereference issue"

This reverts commit cea24824ab432f8acabb254d6805e9aa756de6af.

Moving subdriver probe to exynos_drm_platform_probe() was making
exynos_drm_device_subdrv_probe() fail because the platform data wasn't set
yet. It only gets set in exynos_drm_load.

We need to find a smarter way to fix this issue.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/dpi: stop using display->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:27 +0000 (09:54 +0100)] 
drm/exynos/dpi: stop using display->ctx pointer

The patch replaces accesses to display->ctx pointer by container_of
construct. The field is removed as well as dpi was the last user of it.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/dpi: embed display into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:26 +0000 (09:54 +0100)] 
drm/exynos/dpi: embed display into private context

exynos_drm_display is used by internal Exynos DRM framework for
representing encoder:connector pair. As it should be mapped 1:1 to dpi
private context it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple dpi devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/dp: stop using display->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:25 +0000 (09:54 +0100)] 
drm/exynos/dp: stop using display->ctx pointer

The patch replaces accesses to display->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/dp: embed display into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:24 +0000 (09:54 +0100)] 
drm/exynos/dp: embed display into private context

exynos_drm_display is used by internal Exynos DRM framework for
representing encoder:connector pair. As it should be mapped 1:1 to dp
private context it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple dp devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/vidi: stop using display->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:23 +0000 (09:54 +0100)] 
drm/exynos/vidi: stop using display->ctx pointer

The patch replaces accesses to display->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/vidi: embed display into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:22 +0000 (09:54 +0100)] 
drm/exynos/vidi: embed display into private context

exynos_drm_display is used by internal Exynos DRM framework for
representing encoder:connector pair. As it should be mapped 1:1 to vidi
private context it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple vidi devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/hdmi: stop using display->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:21 +0000 (09:54 +0100)] 
drm/exynos/hdmi: stop using display->ctx pointer

The patch replaces accesses to display->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/hdmi: embed display into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:20 +0000 (09:54 +0100)] 
drm/exynos/hdmi: embed display into private context

exynos_drm_display is used by internal Exynos DRM framework for
representing encoder:connector pair. As it should be mapped 1:1 to hdmi
private context it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple hdmi devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/fimd: stop using manager->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:19 +0000 (09:54 +0100)] 
drm/exynos/fimd: stop using manager->ctx pointer

The patch replaces accesses to manager->ctx pointer by container_of
construct. As fimd was the last user of ctx the patch removes
this field as well.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/fimd: embed manager into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:18 +0000 (09:54 +0100)] 
drm/exynos/fimd: embed manager into private context

exynos_drm_manager is used by internal Exynos DRM framework for
representing crtc. As it should be mapped 1:1 to fimd private context
it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple FIMD devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/vidi: stop using manager->ctx pointer
Andrzej Hajda [Mon, 17 Nov 2014 08:54:17 +0000 (09:54 +0100)] 
drm/exynos/vidi: stop using manager->ctx pointer

The patch replaces accesses to manager->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/vidi: embed manager into private context
Andrzej Hajda [Mon, 17 Nov 2014 08:54:16 +0000 (09:54 +0100)] 
drm/exynos/vidi: embed manager into private context

exynos_drm_manager is used by internal Exynos DRM framework for
representing crtc. As it should be mapped 1:1 to vidi private context
it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple mixer devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/mixer: stop using manager->ctx pointer
Andrzej Hajda [Mon, 24 Nov 2014 05:14:49 +0000 (14:14 +0900)] 
drm/exynos/mixer: stop using manager->ctx pointer

The patch replaces accesses to manager->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos/mixer: embed manager into private context
Andrzej Hajda [Mon, 24 Nov 2014 05:12:46 +0000 (14:12 +0900)] 
drm/exynos/mixer: embed manager into private context

exynos_drm_manager is used by internal Exynos DRM framework for
representing crtc. As it should be mapped 1:1 to fimd private context
it seems more reasonable to embed it directly in that context.
As a result further code simplification will be possible.
Moreover it will be possible to handle multiple mixer devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: set TE GPIO IRQ status as IRQ_NOAUTOEN
YoungJun Cho [Mon, 17 Nov 2014 13:00:16 +0000 (22:00 +0900)] 
drm/exynos: dsi: set TE GPIO IRQ status as IRQ_NOAUTOEN

The exynos_dsi_te_irq_handler() works only dsi(DPMS) is on.
So it is enough to enable and disable TE GPIO IRQ in
exynos_dsi_enable(disable)_irq() like DSI IRQ.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>a
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: move TE irq handler registration position
YoungJun Cho [Mon, 17 Nov 2014 13:00:15 +0000 (22:00 +0900)] 
drm/exynos: dsi: move TE irq handler registration position

The drm_helper_hpd_irq_event() does dpms control and
the panel is initialized and displayed on by it.
So the exynos_dsi_te_irq_handler() should be registered
beforehand.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: use irq_flags instead of triggering
Joonyoung Shim [Mon, 17 Nov 2014 13:00:14 +0000 (22:00 +0900)] 
drm/exynos: use irq_flags instead of triggering

The drm_handle_vblank should be called whenever be vsync, te interrupt
means vsync on i80 interface.

Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: add triggering unset routine in fimd_trigger()
YoungJun Cho [Mon, 17 Nov 2014 13:00:13 +0000 (22:00 +0900)] 
drm/exynos: fimd: add triggering unset routine in fimd_trigger()

There is a case like set config which requires triggering
but vblank is not enabled yet.
So triggering unset routine is required to exit from
triggering mode.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: modify I80 i/f irq relevant routine
YoungJun Cho [Mon, 17 Nov 2014 13:00:12 +0000 (22:00 +0900)] 
drm/exynos: fimd: modify I80 i/f irq relevant routine

For the I80 interface, the video interrupt pending register(VIDINTCON1)
should be handled in fimd_irq_handler() and the video interrupt control
register(VIDINTCON0) should be handled in fimd_enable_vblank() and
fimd_disable_vblank() like RGB interface.
So this patch moves each set / unset routines into proper positions.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: add fimd_enable_shadow_channel_path() to cleanup
YoungJun Cho [Mon, 17 Nov 2014 13:00:11 +0000 (22:00 +0900)] 
drm/exynos: fimd: add fimd_enable_shadow_channel_path() to cleanup

This function is valid only the SoC has SHADOWCON register
and it should be used together with fimd_enable_video_output()
to match the ENWIN_F bit in WINCON# and C#_EN_F bit in SHADOWCON.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: add fimd_enable_video_output() to cleanup
YoungJun Cho [Mon, 17 Nov 2014 13:00:10 +0000 (22:00 +0900)] 
drm/exynos: fimd: add fimd_enable_video_output() to cleanup

This bit is used for video output and logic signal control.
So it is better for readability.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: move shadow unprotection position
YoungJun Cho [Mon, 17 Nov 2014 13:00:09 +0000 (22:00 +0900)] 
drm/exynos: fimd: move shadow unprotection position

The C#_EN_F in SHADOWCON register is updated per frame.
So it should be protected by fimd_shadow_protect_win().

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: move triggering checking
Joonyoung Shim [Mon, 17 Nov 2014 13:00:08 +0000 (22:00 +0900)] 
drm/exynos: move triggering checking

It's better to be checking whether triggerring in fimd_trigger function.
Also it will return if in triggerring on fimd_te_handler, then it can't
execute remain codes.

Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: add has_vtsel flag
Joonyoung Shim [Fri, 14 Nov 2014 02:36:02 +0000 (11:36 +0900)] 
drm/exynos: add has_vtsel flag

The exynos fimd provides video type selection bits from system register
but exynos3 series don't has it, so needs has_vtsel flag and we can
distinguish whether set video type selection bits.

Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: Fix DSI resuming fail because power domain being off
Krzysztof Kozlowski [Fri, 7 Nov 2014 13:53:58 +0000 (14:53 +0100)] 
drm/exynos: Fix DSI resuming fail because power domain being off

During system resume from suspend to RAM the Exynos DRM driver forced
CRTC mode thus turning display on (DPMS_ON). This lead to runtime resuming
of DSI which failed because whole LCD power domain was off and it was
not allowed to turn on because of system resume in progress.

Forcing mode should not be needed and removing it solves this particular
problem.

This necessary fix for following scenario reproduced on Exynos DRM:
1. Power domain is off before suspending the system.
2. System is suspended to RAM.
3. Resuming starts. The Exynos DRM driver resume callback is called.
4. The Exynos DRM driver calls drm_helper_resume_force_mode() which turns
   on the screen by calling exynos_dsi_dpms with DRM_MODE_DPMS_ON.
5. The Exynos DSI driver calls pm_runtime_get. The driver runtime
   resumes and this should turn LCD power domain on.
6. Unfortunately the domain cannot be turned on because system resume is
   in progress and genpd->prepared_count is positive.

Steps to reproduce:
1. Add runtime PM to Exynos DSI driver.
2. Build Exynos DRM/FB without FRAMEBUFFER_CONSOLE.
3. Enable the connector and screen (e.g. with modeset-vsync).
4. echo 3 > /sys/devices/platform/exynos-drm/graphics/fb0/blank
5. echo mem > /sys/power/state
6. Resume.
[   77.712469] PM: early resume of devices complete after 3.854 msecs
[   77.712739] exynos-dsi 11c80000.dsi: pm_genpd_resume()
[   77.712758] exynos4-fimc 11800000.fimc: pm_genpd_resume()
[   77.712774] exynos4-fimc 11810000.fimc: pm_genpd_resume()
[   77.712787] exynos-drm-fimc 11820000.fimc: pm_genpd_resume()
[   77.712802] exynos-drm-fimc 11830000.fimc: pm_genpd_resume()
[   77.712815] s5p-mipi-csis 11880000.csis: pm_genpd_resume()
[   77.712829] s5p-mipi-csis 11890000.csis: pm_genpd_resume()
[   77.712843] exynos-fimc-lite 12390000.fimc-lite: pm_genpd_resume()
[   77.712856] exynos-fimc-lite 123a0000.fimc-lite: pm_genpd_resume()
[   77.713788] exynos4-fb 11c00000.fimd: pm_genpd_resume()
[   77.713912] wake disabled for irq 184
[   77.713923] wake disabled for irq 185
[   77.714082] wake disabled for irq 173
[   77.715676] wake disabled for irq 176
[   77.718540] exynos4-fb 11c00000.fimd: pm_genpd_runtime_resume()
[   77.718567] exynos4-fb 11c00000.fimd: state restore latency exceeded, new value 1708 ns
[   77.718636] exynos-dsi 11c80000.dsi: pm_genpd_runtime_resume()
[   77.892366] exynos-dsi 11c80000.dsi: PLL failed to stabilize
[   77.892377] exynos-dsi 11c80000.dsi: failed to configure DSI PLL
[   78.192168] exynos-dsi 11c80000.dsi: timeout waiting for reset
[   78.211578] exynos-dsi 11c80000.dsi: waiting for bus lanes timed out
[   78.307173] exynos-dsi 11c80000.dsi: xfer timed out: d1 00 (null)
[   78.307190] panel_s6e8aa0 11c80000.dsi.0: error -110 reading dcs seq(0xd1)
[   78.307199] panel_s6e8aa0 11c80000.dsi.0: read id failed

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: move DSIM_STATE_ENABLED set position
YoungJun Cho [Wed, 1 Oct 2014 06:19:13 +0000 (15:19 +0900)] 
drm/exynos: dsi: move DSIM_STATE_ENABLED set position

The command mode panel should draw image earlier than the display
on command execution to prevent showing garbage GRAM screen data.
So should set dsi->state as DSIM_STATE_ENABLED between calling
exynos_dsi_set_display_enable() and drm_panel_enable() to transmit
image data before executing display on command.
And moves the display on command execution routine from prepare()
to enable() in drm_panel_funcs also.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: move handle vblank position in TE handler
YoungJun Cho [Wed, 1 Oct 2014 06:19:10 +0000 (15:19 +0900)] 
drm/exynos: fimd: move handle vblank position in TE handler

For providing VBLANK information, drm_handle_vblank() should
be called properly, but it is blocked by wait_vsync_event
condition which is set by manager_ops->wait_for_vblank().
So moves it out from wait_vsync_event routine.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: remove unnecessary waiting vblank routine
YoungJun Cho [Wed, 1 Oct 2014 06:19:07 +0000 (15:19 +0900)] 
drm/exynos: fimd: remove unnecessary waiting vblank routine

The exynos_drm_crtc_dpms() waits until pended page flip
queue is empty, calls the drm_vblank_off() then calls
manager->ops->dpms() when mode is DRM_MODE_DPMS_OFF.
The fimd_dpms() is one of manager->ops->dpms()s and
finally calls fimd_window_suspend().
But there is no active window and vblank is already off
when it is called.
So addtional waiting vblank is not necessary any more.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Inki Dae <inki.dae@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fimd: support Exynos4415 SoC
YoungJun Cho [Fri, 7 Nov 2014 06:12:25 +0000 (15:12 +0900)] 
drm/exynos: fimd: support Exynos4415 SoC

This patch supports Exynos4415 SoC.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: support Exynos4415 SoC
YoungJun Cho [Fri, 7 Nov 2014 06:12:24 +0000 (15:12 +0900)] 
drm/exynos: dsi: support Exynos4415 SoC

This patch supports Exynos4415 SoC.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dp: Remove support for unused dptx-phy
Vivek Gautam [Wed, 12 Nov 2014 09:42:10 +0000 (15:12 +0530)] 
drm/exynos: dp: Remove support for unused dptx-phy

Now that we have moved to generic phy based bindings,
we don't need to have any code related to older dptx-phy.
Nobody is using this dptx-phy anymore, so removing the
same.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove leftover hdmi function declarations
Gustavo Padovan [Fri, 31 Oct 2014 14:17:43 +0000 (14:17 +0000)] 
drm/exynos: remove leftover hdmi function declarations

They are not implemented anywhere, so wipe them out.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: update documentation to reflect code changes
Gustavo Padovan [Fri, 31 Oct 2014 14:17:42 +0000 (14:17 +0000)] 
drm/exynos: update documentation to reflect code changes

Description of the @create_connector callback was missing,
and the @manager was no longer needed.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: Save up space using bool var as bitfields
Gustavo Padovan [Thu, 13 Nov 2014 08:25:53 +0000 (17:25 +0900)] 
drm/exynos: Save up space using bool var as bitfields

Save a few bytes by compiling them all in the same byte.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove unused wait_for macro
Gustavo Padovan [Fri, 31 Oct 2014 14:17:40 +0000 (14:17 +0000)] 
drm/exynos: remove unused wait_for macro

This is a leftover, all code using this macro have been removed/
changed already.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: Replace repeated declarations by #include "exynos_drm_drv.h"
Gustavo Padovan [Thu, 13 Nov 2014 08:24:06 +0000 (17:24 +0900)] 
drm/exynos: Replace repeated declarations by #include "exynos_drm_drv.h"

Re-declare struct is not a good practice, let's use the original
drm and exynos declarations.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: Replace repeated declaration by include <drm/drmP.h>
Gustavo Padovan [Fri, 31 Oct 2014 14:17:38 +0000 (14:17 +0000)] 
drm/exynos: Replace repeated declaration by include <drm/drmP.h>

Re-declare struct is not a good practice, let's use the original
drm declarations.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove extra declaration of struct exynos_overlay
Gustavo Padovan [Fri, 31 Oct 2014 14:17:37 +0000 (14:17 +0000)] 
drm/exynos: remove extra declaration of struct exynos_overlay

The struct is defined in the same file, declare it here is just
unnecessary

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove extra declaration of struct exynos_drm_manager
Gustavo Padovan [Fri, 31 Oct 2014 14:17:36 +0000 (14:17 +0000)] 
drm/exynos: remove extra declaration of struct exynos_drm_manager

The struct is defined in the same file, declare it here is just
unnecessary.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove uneeded declaration of struct dma_iommu_mapping
Gustavo Padovan [Fri, 31 Oct 2014 14:17:35 +0000 (14:17 +0000)] 
drm/exynos: remove uneeded declaration of struct dma_iommu_mapping

It is not even used in this header anymore.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: stop using display->ctx pointer
Andrzej Hajda [Tue, 7 Oct 2014 12:01:11 +0000 (14:01 +0200)] 
drm/exynos: dsi: stop using display->ctx pointer

The patch replaces accesses to display->ctx pointer by container_of
construct. It will allow to remove ctx field in the future.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: remove redundant encoder field
Andrzej Hajda [Tue, 7 Oct 2014 12:01:10 +0000 (14:01 +0200)] 
drm/exynos: dsi: remove redundant encoder field

The patch removes redundant encoder field from private DSI context.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: simplify device pointer evaluation
Andrzej Hajda [Tue, 7 Oct 2014 12:01:09 +0000 (14:01 +0200)] 
drm/exynos: dsi: simplify device pointer evaluation

The patch replaces multiple evaluation of device address
with local variable.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: dsi: remove global variable exynos_dsi_display
Andrzej Hajda [Tue, 7 Oct 2014 12:01:08 +0000 (14:01 +0200)] 
drm/exynos: dsi: remove global variable exynos_dsi_display

exynos_dsi_display is used by internal Exynos DRM framework for
representing pair encoder->connecter. As it should be mapped 1:1 to dsi
private context it seems more reasonable to embed it directly
in that context. As a result further code simplification will be possible.
Moreover it will be possible to handle multiple DSI devices in the system.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: remove ifdeferry from initialization code
Andrzej Hajda [Thu, 13 Nov 2014 07:37:57 +0000 (16:37 +0900)] 
drm/exynos: remove ifdeferry from initialization code

The patch replaces separate calls to driver (de)registration by
loops over the array of drivers. As a result it significantly
decreases number of ifdefs. Additionally it moves device registration
related ifdefs to header file.

Changelog v2:
- Rebased.
- Consider non kms driver in respect to infinite loop issue.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fix null pointer dereference issue
Inki Dae [Thu, 13 Nov 2014 07:05:12 +0000 (16:05 +0900)] 
drm/exynos: fix null pointer dereference issue

This patch fixes null pointer dereference issue incurred
when ipp driver is enabled and Exynos drm driver is closed.

Non kms driver should register its own sub driver to setup necessary
resources, which is done by load(). So null pointer dereference
occurs when ipp driver is enabled and Exynos drm driver is closed
because ipp core device is registered after component_master_add_with_match
call.

This patch makes exynos_drm_device_subdrv_probe() to be called after all non
kms drivers are registered.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: fix possible infinite loop issue
Inki Dae [Fri, 7 Nov 2014 12:32:34 +0000 (21:32 +0900)] 
drm/exynos: fix possible infinite loop issue

This patch fixes possible infinite loop issue by postponing
registration to non kms drivers after component_master_add_with_match
call, which can be incurred in all cases that non kms driver is probed
and then component bind is failed

This patch should be applied on top of below patches,
http://comments.gmane.org/gmane.comp.video.dri.devel/117740
http://www.spinics.net/lists/linux-samsung-soc/msg38624.html

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: g2d: fix null pointer dereference
Inki Dae [Fri, 7 Nov 2014 11:31:08 +0000 (20:31 +0900)] 
drm/exynos: g2d: fix null pointer dereference

This patch fixes a null pointer dereference issue incurred by
calling g2d_remove when exynos_drm_platform_probe is failed.

cmdlist_pool of g2d is allocated when g2d sub driver is probed.
So if exynos_drm_platform_probe is failed, the g2d sub driver is
not probed and the cmdlist_pool is still NULL.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: resolve infinite loop issue on non multi-platform
Inki Dae [Thu, 6 Nov 2014 14:00:37 +0000 (23:00 +0900)] 
drm/exynos: resolve infinite loop issue on non multi-platform

This patch resovles the infinite loop issue incurred
when Exyno drm driver is enabled but all kms drivers
are disabled on Exynos board by returning -EPROBE_DEFER
only in case that there is kms device registered.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/exynos: resolve infinite loop issue on multi-platform
Inki Dae [Thu, 6 Nov 2014 10:23:35 +0000 (19:23 +0900)] 
drm/exynos: resolve infinite loop issue on multi-platform

This patch resolves temporarily infinite loop issue incurred
when Exynos drm driver is enabled and multi-platform kernel
is used by registering Exynos drm device object only in case
of Exynos SoC. So this patch will be replaced with more generic
way later.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
9 years agodrm/atomic: shutdown *current* encoder
Rob Clark [Thu, 20 Nov 2014 20:40:36 +0000 (15:40 -0500)] 
drm/atomic: shutdown *current* encoder

In disable_outputs() we need to shut down the outgoing encoder, not the
incoming one (we have already swapped-state at this point).  Without
this, we end up telling the driver to crtc->dpms(OFF) without first
encoder->dpms(OFF), and that makes some hw quite unhappy.

v2: missing WARN_ON() hunk and comment

Reviewed-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Rob Clark <robdclark@gmail.com>
This page took 0.048528 seconds and 5 git commands to generate.