Eric Anholt [Thu, 8 Jun 2017 00:13:36 +0000 (17:13 -0700)]
drm/vc4: Add get/set tiling ioctls.
This allows mesa to set the tiling format for a BO and have that
tiling format be respected by mesa on the other side of an
import/export (and by vc4 scanout in the kernel), without defining a
protocol to pass the tiling through userspace.
Eric Anholt [Thu, 8 Jun 2017 00:13:35 +0000 (17:13 -0700)]
drm/vc4: Add T-format scanout support.
The T tiling format is what V3D uses for textures, with no raster
support at all until later revisions of the hardware (and always at a
large 3D performance penalty). If we can't scan out V3D's format,
then we often need to do a relayout at some stage of the pipeline,
either right before texturing from the scanout buffer (common in X11
without a compositor) or between a tiled screen buffer right before
scanout (an option I've considered in trying to resolve this
inconsistency, but which means needing to use the dirty fb ioctl and
having some update policy).
T-format scanout lets us avoid either of those shadow copies, for a
massive, obvious performance improvement to X11 window dragging
without a compositor. Unfortunately, enabling a compositor to work
around the discrepancy has turned out to be too costly in memory
consumption for the Raspbian distribution.
Because the HVS operates a scanline at a time, compositing from T does
increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp
display on a RPi3, we go from about 15% of system memory bandwidth
with linear to about 20% with tiled. However, for X11 this still ends
up being a huge performance win in active usage.
This patch doesn't yet handle src_x/src_y offsetting within the tiled
buffer. However, we fail to do so for untiled buffers already.
Fabio Estevam [Fri, 5 May 2017 18:01:41 +0000 (15:01 -0300)]
drm: mxsfb_crtc: Reset the eLCDIF controller
According to the eLCDIF initialization steps listed in the MX6SX
Reference Manual the eLCDIF block reset is mandatory.
Without performing the eLCDIF reset the display shows garbage content
when the kernel boots.
In earlier tests this issue has not been observed because the bootloader
was previously showing a splash screen and the bootloader display driver
does properly implement the eLCDIF reset.
Add the eLCDIF reset to the driver, so that it can operate correctly
independently of the bootloader.
drm/panel: s6e3ha2: Add support for s6e3hf2 panel on TM2e board
This patch supports TM2e panel and the panel has 1600x2560 resolution
in 5.65" physical.
This identify panel type with compatibility string, also invoke
display mode that matches the type. So add the check code for s6e3ha2
compatibility and s6e3hf2 type and select the drm_display_mode of
default and edge type.
drm/panel: add backlight dependency for sitronix-st7789v
Without the dependency, we run into a link error:
drivers/gpu/drm/panel/panel-sitronix-st7789v.o: In function `st7789v_probe':
panel-sitronix-st7789v.c:(.text.st7789v_probe+0xc0): undefined reference to `of_find_backlight_by_node'
Lucas Stach [Thu, 8 Jun 2017 18:07:56 +0000 (20:07 +0200)]
dt-bindings: add vendor prefix for NLT Technologies, Ltd.
NLT technologies is the former NEC display business, but changed its
name to NLT Technologies when forming a joint venture with
Shenzhen AVIC OPTOELECTRONICS, Ltd.
drm/vc4: Fix resource leak in 'vc4_get_hang_state_ioctl()' in error handling path
If one 'drm_gem_handle_create()' fails, we leak somes handles and some
memory.
In order to fix it:
- move the 'free(bo_state)' at the end of the function so that it is also
called in the eror handling path. This has the side effect to also try
to free it if the first 'kcalloc' fails. This is harmless.
- add a new label, err_delete_handle, in order to delete already
allocated handles in error handling path
- remove the now useless 'err' label
The way the code is now written will also delete the handles if the
'copy_to_user()' call fails.
Jyri Sarha [Tue, 30 May 2017 16:29:55 +0000 (19:29 +0300)]
drm: Add const to name field declaration in struct drm_prop_enum_list
There is no reason why the name field should not be const, but
several why it should. The struct should only be used by
drm_property_create_enum() and there the name-field from the struct
is passed to drm_property_add_enum(), which takes a const char * as
a parameter.
drm/bridge: Build the panel wrapper in drm_kms_helper
This fixes the following depmod error when building drm as a module:
depmod: ERROR: Found 6 modules in dependency cycles!
depmod: ERROR: Cycle detected: drm -> drm_kms_helper -> drm
Fixes: 65f5a41cde8b ("drm/bridge: Refactor out the panel wrapper from the lvds-encoder bridge.") Tested-by: Lofstedt, Marta <marta.lofstedt@intel.com> Reviewed-by: Eric Anholt <eric@anholt.net> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Archit Taneja <architt@codeaurora.org> Link: http://patchwork.freedesktop.org/patch/msgid/3fd262cf-1db6-4335-320c-af92f9014502@linux.intel.com
Eric Anholt [Thu, 11 May 2017 18:31:28 +0000 (11:31 -0700)]
drm/atmel-hlcdc: Replace the panel usage with drm_panel_bridge.
This cuts 135 lines of boilerplate, at the cost of losing the
filtering of get_modes() using atmel_hlcdc_dc_mode_valid(). The
atomic check will still check that we don't set an invalid mode,
though.
Eric Anholt [Thu, 11 May 2017 18:31:23 +0000 (11:31 -0700)]
drm/vc4: Switch DSI to the panel-bridge layer, and support bridges.
The newer version of the RPi panel driver is going to be a combination
of a bridge and a panel, but we should also support panels without a
bridge, so the panel-bridge layer lets us do that cleanly.
Eric Anholt [Fri, 2 Jun 2017 20:25:14 +0000 (13:25 -0700)]
drm/bridge: Refactor out the panel wrapper from the lvds-encoder bridge.
Many DRM drivers have common code to make a stub connector
implementation that wraps a drm_panel. By wrapping the panel in a DRM
bridge, all of the connector code (including calls during encoder
enable/disable) goes away.
v2: Fix build with CONFIG_DRM=m, drop "dev" argument that should just
be the panel's dev, move kerneldoc up a level and document
_remove().
v3: Fix another breakage with CONFIG_DRM=m, fix breakage with
CONFIG_OF=n, move protos under CONFIG_DRM_PANEL_BRIDGE, wrap a
line.
Signed-off-by: Eric Anholt <eric@anholt.net> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v1) Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com> (v2) Acked-by: Archit Taneja <architt@codeaurora.org> (v2) Signed-off-by: Archit Taneja <architt@codeaurora.org> Link: http://patchwork.freedesktop.org/patch/msgid/20170602202514.11900-1-eric@anholt.net
Jose Abreu [Thu, 25 May 2017 14:19:19 +0000 (15:19 +0100)]
drm/bridge/synopsys: dw-hdmi: Use bridge->mode_valid() callback
Now that we have a callback to check if bridge supports a given mode
we can use it in Synopsys Designware HDMI bridge so that we restrict
the number of probbed modes to the ones we can actually display.
Also, there is no need to use mode_fixup() callback as mode_valid()
will handle the mode validation.
NOTE: I also had to change the pdata declaration of mode_valid
custom callback so that the passed modes are const. I also changed
in the platforms I found. Not even compiled it though.
Signed-off-by: Jose Abreu <joabreu@synopsys.com> Acked-by: Neil Armstrong <narmstrong@baylibre.com> Acked-by: Philipp Zabel <p.zabel@pengutronix.de> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Archit Taneja <architt@codeaurora.org> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: David Airlie <airlied@linux.ie> Cc: Philipp Zabel <p.zabel@pengutronix.de> Cc: Carlo Caione <carlo@caione.org> Cc: Kevin Hilman <khilman@baylibre.com> Cc: Mark Yao <mark.yao@rock-chips.com> Cc: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Archit Taneja <architt@codeaurora.org> Link: http://patchwork.freedesktop.org/patch/msgid/3d8d449e4d13d2535fa292c75f5fa931de4a4fa8.1495720737.git.joabreu@synopsys.com
Jose Abreu [Thu, 25 May 2017 14:19:21 +0000 (15:19 +0100)]
drm/atmel-hlcdc: Use crtc->mode_valid() callback
Now that we have a callback to check if crtc supports a given mode
we can use it in atmel-hlcdc so that we restrict the number of probbed
modes to the ones we can actually display.
Also, remove the mode_fixup() callback as this is no longer needed
because mode_valid() will be called before.
Daniel Vetter [Wed, 24 May 2017 14:51:51 +0000 (16:51 +0200)]
drm/hdlcd|mali: Drop drm_vblank_cleanup
IRQs are properly shut down, so it almost works as race-free shutdown.
Except the irq is stopped after the vblank stuff, so boom anyway.
Proper way would be to call drm_atomic_helper_shutdown before any of
the kms things gets stopped. So no harm in removing the
drm_vblank_cleanup here really.
Same story for both hdlcd and mali.
v2: Move misplaced malidp hunk to this patch (Liviu).
Daniel Vetter [Wed, 31 May 2017 09:21:46 +0000 (11:21 +0200)]
drm: Extract drm_vblank.[hc]
drm_irq.c contains both the irq helper library (optional) and the
vblank support (optional, but part of the modeset uapi, and doesn't
require the use of the irq helpers at all.
Split this up for more clarity of the scope of the individual bits.
Arnd Bergmann [Wed, 24 May 2017 15:49:58 +0000 (17:49 +0200)]
drm/pl111: fix warnings without CONFIG_ARM_AMBA
The driver is written in a way to enable compile-testing without CONFIG_ARM_AMBA,
but it just causes needless warnings:
drivers/gpu/drm/pl111/pl111_drv.c:149:26: error: 'pl111_drm_driver' defined but not used [-Werror=unused-variable]
drivers/gpu/drm/pl111/pl111_drv.c:81:12: error: 'pl111_modeset_init' defined but not used [-Werror=unused-function]
This unhides the probe/remove functions again and just leaves the driver
object as unused when CONFIG_ARM_AMBA is disabled, with a __maybe_unused
annotation to shut up the warning.
Daniel Vetter [Wed, 24 May 2017 14:51:50 +0000 (16:51 +0200)]
drm/arcgpu: Drop drm_vblank_cleanup
CRTC don't seem to get shut down in a controlled fashion, but no one
bothers to stop interrupts either so this races no matter what. Might
as well remove it. A call to drm_atomic_helper_shutdown would be
pretty sweet somewhere (and maybe getting rid of the load/unload
callbacks while at it).
Daniel Vetter [Wed, 24 May 2017 14:51:52 +0000 (16:51 +0200)]
drm/atmel: Drop drm_vblank_cleanup
Again almost correct, but since interrupts are shut down after vblank
still a race. Proper cleanup would call drm_atomic_helper_shutdown to
make sure this really is safe.
Daniel Vetter [Wed, 24 May 2017 14:51:58 +0000 (16:51 +0200)]
drm/imx: Drop drm_vblank_cleanup
It's only done in the driver load error path, where vblanks don't need
to be quiescent anyway. And that's all drm_vblank_cleanup does, since
the core will release the vblank allocations on its own already. So
drop it.
Daniel Vetter [Wed, 24 May 2017 14:51:36 +0000 (16:51 +0200)]
drm/doc: move printf helpers out of drmP.h
And document them lightly. Unfortunately kernel-doc isn't the most
awesome for documenting #defines that don't look like functions, it
makes functions out of them :-/
Arnd Bergmann [Tue, 30 May 2017 09:22:08 +0000 (11:22 +0200)]
drm/pl111: select DRM_PANEL
When DRM_PANEL is disabled, we get a link error for pl111:
drivers/gpu/built-in.o: In function `pl111_connector_destroy':
pl111_connector.c:(.text+0x3487e6): undefined reference to `drm_panel_detach'
For some reason this only appears in the latest linux-next
although the driver appears to have used the symbol for a few
weeks already. The solution however is simple enough, we just
need to add a 'select' statement.
Jose Abreu [Thu, 25 May 2017 14:19:18 +0000 (15:19 +0100)]
drm/bridge: analogix-anx78xx: Use bridge->mode_valid() callback
Now that we have a callback to check if bridge supports a given mode
we can use it in Analogix bridge so that we restrict the number of
probbed modes to the ones we can actually display.
Also, there is no need to use mode_fixup() callback as mode_valid()
will handle the mode validation.
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com> Signed-off-by: Jose Abreu <joabreu@synopsys.com> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Archit Taneja <architt@codeaurora.org> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: David Airlie <airlied@linux.ie> Signed-off-by: Archit Taneja <architt@codeaurora.org> Link: http://patchwork.freedesktop.org/patch/msgid/1d0ed1858ae56c827bd09cc1fa6ff4a05d1530eb.1495720737.git.joabreu@synopsys.com
Jose Abreu [Thu, 25 May 2017 14:19:16 +0000 (15:19 +0100)]
drm: Use mode_valid() in atomic modeset
This patches makes use of the new mode_valid() callbacks introduced
previously to validate the full video pipeline when modesetting.
This calls the connector->mode_valid(), encoder->mode_valid(),
bridge->mode_valid() and crtc->mode_valid() so that we can
make sure that the mode will be accepted in every components.
Signed-off-by: Jose Abreu <joabreu@synopsys.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Dave Airlie <airlied@linux.ie> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Archit Taneja <architt@codeaurora.org> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Changes v1->v2:
- Removed call to connector->mode_valid (Ville, Daniel)
- Changed function name (Ville)
- Use for_each_new_connector_in_state (Ville)
- Do not validate if connector and mode didn't change (Ville)
- Use new helpers to call mode_valid
Jose Abreu [Thu, 25 May 2017 14:19:15 +0000 (15:19 +0100)]
drm: Use new mode_valid() helpers in connector probe helper
This changes the connector probe helper function to use the new
encoder->mode_valid(), bridge->mode_valid() and crtc->mode_valid()
helper callbacks to validate the modes.
The new callbacks are optional so the behaviour remains the same
if they are not implemented. If they are, then the code loops
through all the connector's encodersXbridgesXcrtcs and calls the
callback.
If at least a valid encoderXbridgeXcrtc combination is found which
accepts the mode then the function returns MODE_OK.
Signed-off-by: Jose Abreu <joabreu@synopsys.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Dave Airlie <airlied@linux.ie> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Changes v3->v4:
- Change function name (Laurent)
Changes v2->v3:
- Call also bridge->mode_valid (Daniel)
Changes v1->v2:
- Use new helpers suggested by Ville
- Change documentation (Daniel) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/d4e3ba87d822fa92f1b8773e441b9a02af3bde71.1495720737.git.joabreu@synopsys.com
Add a new helper to call crtc->mode_valid, connector->mode_valid
and encoder->mode_valid callbacks.
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Jose Abreu <joabreu@synopsys.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Dave Airlie <airlied@linux.ie>
Changes v2->v3:
- Move helpers to drm_probe_helper.c (Daniel)
- Squeeze patches that introduce helpers into a single
one (Daniel)
Daniel Vetter [Wed, 24 May 2017 14:51:37 +0000 (16:51 +0200)]
drm: Remove drm_device->virtdev
This is a leftover from the drm_bus days, where we've had a
bus-specific device type for every bus type in drm_device. Except for
pci (which we can't remove because dri1 drivers) this is all gone. And
the virt driver also doesn't really need it, dev_to_virtio works
perfectly fine.
Chris Wilson [Tue, 16 May 2017 11:10:42 +0000 (12:10 +0100)]
dma-buf/sync-file: Defer creation of sync_file->name
Constructing the name takes the majority of the time for allocating a
sync_file to wrap a fence, and the name is very rarely used (only via
the sync_file status user interface). To reduce the impact on the common
path (that of creating sync_file to pass around), defer the construction
of the name until it is first used.
v2: Update kerneldoc (kbuild test robot)
v3: sync_debug.c was peeking at the name
v4: Comment upon the potential race between two users of
sync_file_get_name() and claim that such a race is below the level of
notice. However, to prevent any future nuisance, use a global spinlock
to serialize the assignment of the name.
v5: Completely avoid the read/write race by only storing the name passed
in from the user inside sync_file->user_name and passing in a buffer to
dynamically construct the name otherwise.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Sumit Semwal <sumit.semwal@linaro.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Herrmann <dh.herrmann@gmail.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170516111042.24719-1-chris@chris-wilson.co.uk
Each text file under Documentation follows a different
format. Some doesn't even have titles!
Change its representation to follow the adopted standard,
using ReST markups for it to be parseable by Sphinx:
- Use markup for document title and authorship;
- Mark literal blocks;
- Use a numbered list for references.
Arnd Bergmann [Mon, 22 May 2017 20:30:20 +0000 (22:30 +0200)]
gpu: drm: gma500: remove two more dead variable
The dead code removal left two unused variables:
drivers/gpu/drm/gma500/mdfld_tpo_vid.c: In function 'tpo_vid_get_config_mode':
drivers/gpu/drm/gma500/mdfld_tpo_vid.c:34:31: error: unused variable 'ti' [-Werror=unused-variable]
Daniel Vetter [Mon, 15 May 2017 09:11:36 +0000 (11:11 +0200)]
drm/doc: Clarify mode_fixup vs. atomic_check a bit more
Brought up by both Laurent and Andrzej when reviewing the new
->mode_valid hooks. Since mode_fixup is just a simpler version of the
much more generic atomic_check we can't really unify it with
mode_valid. Most drivers should probably switch their current
mode_fixup code to either the new mode_valid or the atomic_check
hooks, but e.g. that doesn't exist yet for bridges, and for CRTCs the
situation is a bit more complicated. Hence there's no clear
equivalence between mode_fixup and mode_valid, even if it looks like
that at first glance.
v2: Fix accidental double-dot (Adnrzej).
Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Jose Abreu <Jose.Abreu@synopsys.com> Reviewed-by: Jose Abreu <joabreu@synopsys.com> Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170515091136.26307-2-daniel.vetter@ffwll.ch
This adds a new callback to crtc, encoder and bridge helper functions
called mode_valid(). This callback shall be implemented if the
corresponding component has some sort of restriction in the modes
that can be displayed. A NULL callback implicates that the component
can display all the modes.
We also change the documentation so that the new and old callbacks
are correctly documented.
Only the callbacks were implemented to simplify review process,
following patches will make use of them.
Changes in v2 from Daniel:
- Update the warning about how modes aren't filtered in atomic_check -
the heleprs help out a lot more now.
- Consistenly roll out that warning, crtc/encoder's atomic_check
missed it.
- Sprinkle more links all over the place, so it's easier to see where
this stuff is used and how the differen hooks are related.
- Note that ->mode_valid is optional everywhere.
- Explain why the connector's mode_valid is special and does _not_ get
called in atomic_check.
v3: Document what can and cannot be checked in mode_valid a bit better
(Andrjez). Answer: Only allowed to look at the mode, nothing else.
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Jose Abreu <joabreu@synopsys.com> Cc: Jose Abreu <joabreu@synopsys.com> Cc: Carlos Palminha <palminha@synopsys.com> Cc: Alexey Brodkin <abrodkin@synopsys.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Dave Airlie <airlied@linux.ie> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Archit Taneja <architt@codeaurora.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v2) Reviewed-by: Jose Abreu <joabreu@synopsys.com> (v2) Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170515093347.31098-1-daniel.vetter@ffwll.ch
Wei Yongjun [Sun, 21 May 2017 01:01:52 +0000 (01:01 +0000)]
drm/pl111: Fix return value check in pl111_amba_probe()
In case of error, the function devm_ioremap_resource() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().
Robert Foss [Fri, 19 May 2017 20:50:17 +0000 (16:50 -0400)]
drm: Add DRM_MODE_ROTATE_ and DRM_MODE_REFLECT_ to UAPI
Add DRM_MODE_ROTATE_ and DRM_MODE_REFLECT_ defines to the UAPI
as a convenience.
Ideally the DRM_ROTATE_ and DRM_REFLECT_ property ids are looked up
through the atomic API, but realizing that userspace is likely to take
shortcuts and assume that the enum values are what is sent over the
wire.
As a result these defines are provided purely as a convenience to
userspace applications.
Changes since v3:
- Switched away from past tense in comments
- Add define name change to previously mis-spelled DRM_REFLECT_X comment
- Improved the comment for the DRM_MODE_REFLECT_<axis> comment
Changes since v2:
- Changed define prefix from DRM_MODE_PROP_ to DRM_MODE_
- Fix compilation errors
- Changed comment formatting
- Deduplicated comment lines
- Clarified DRM_MODE_PROP_REFLECT_ comment
Changes since v1:
- Moved defines from drm.h to drm_mode.h
- Changed define prefix from DRM_ to DRM_MODE_PROP_
- Updated uses of the defines to the new prefix
- Removed include from drm_rect.c
- Stopped using the BIT() macro
Wei Yongjun [Sun, 21 May 2017 01:19:39 +0000 (01:19 +0000)]
drm/vgem: Fix return value check in vgem_init()
In case of error, the function platform_device_register_simple() returns
ERR_PTR() and never returns NULL. The NULL test in the return value
check should be replaced with IS_ERR().
Fixes: ed15c6395ba7 ("drm/vgem: Enable dmabuf import interfaces") Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Fix fixes: tag per Chris' review.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20170521011939.8111-1-weiyj.lk@gmail.com
Eric Anholt [Thu, 18 May 2017 00:56:40 +0000 (17:56 -0700)]
drm/pl111: Add a debugfs node to dump our registers.
While debugging an X11 display failure, I wanted to see where we were
actually scanning out from. This is probably generally useful to
others that might be working on this device.
Eric Anholt [Mon, 8 May 2017 19:33:48 +0000 (12:33 -0700)]
drm/pl111: Register the clock divider and use it.
This is required for the panel to work on bcm911360, where CLCDCLK is
the fixed 200Mhz AXI41 clock. The rate set is still passed up to the
CLCDCLK, for platforms that have a settable rate on that one.
v2: Set SET_RATE_PARENT (caught by Linus Walleij), depend on
COMMON_CLK.
v3: Mark the clk_ops static (caught by Stephen).
Michal Hocko [Wed, 17 May 2017 12:23:12 +0000 (14:23 +0200)]
drm: drop drm_[cm]alloc* helpers
Now that drm_[cm]alloc* helpers are simple one line wrappers around
kvmalloc_array and drm_free_large is just kvfree alias we can drop
them and replace by their native forms.
This shouldn't introduce any functional change.
Changes since v1
- fix typo in drivers/gpu//drm/etnaviv/etnaviv_gem.c - noticed by 0day
build robot
Suggested-by: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Michal Hocko <mhocko@suse.com>drm: drop drm_[cm]alloc* helpers
[danvet: Fixup vgem which grew another user very recently.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Acked-by: Christian König <christian.koenig@amd.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170517122312.GK18247@dhcp22.suse.cz
Michal Hocko [Wed, 17 May 2017 06:55:08 +0000 (08:55 +0200)]
drm: replace drm_[cm]alloc* by kvmalloc alternatives
drm_[cm]alloc* has grown their own kvmalloc with vmalloc fallback
implementations. MM has grown kvmalloc* helpers in the meantime. Let's
use those because it a) reduces the code and b) MM has a better idea
how to implement fallbacks (e.g. do not vmalloc before kmalloc is tried
with __GFP_NORETRY).
drm_calloc_large needs to get __GFP_ZERO explicitly but it is the same
thing as kvmalloc_array in principle.
When IOMMU_IOVA is not built-in but host1x is, we get a link error:
drivers/gpu/host1x/dev.o: In function `host1x_remove':
dev.c:(.text.host1x_remove+0x50): undefined reference to `put_iova_domain'
drivers/gpu/host1x/dev.o: In function `host1x_probe':
dev.c:(.text.host1x_probe+0x31c): undefined reference to `init_iova_domain'
dev.c:(.text.host1x_probe+0x38c): undefined reference to `put_iova_domain'
drivers/gpu/host1x/cdma.o: In function `host1x_cdma_init':
cdma.c:(.text.host1x_cdma_init+0x238): undefined reference to `alloc_iova'
cdma.c:(.text.host1x_cdma_init+0x2c0): undefined reference to `__free_iova'
drivers/gpu/host1x/cdma.o: In function `host1x_cdma_deinit':
cdma.c:(.text.host1x_cdma_deinit+0xb0): undefined reference to `free_iova'
This adds the same select statement that we have for drm_tegra.