Dave Airlie [Wed, 30 Nov 2016 23:25:58 +0000 (09:25 +1000)]
Merge branch 'msm-next' of git://people.freedesktop.org/~robclark/linux into drm-next
On the userspace side, all the basics are working, and most of glmark2
is working. I've been working through deqp, and I've got a couple more
things to fix (but we've gone from 70% to 80+% pass in last day, and
current deqp run that is going should pick up another 5-10%). I expect
to push the mesa patches today or tomorrow.
There are a couple more a5xx related patches to take the gpu out of
secure mode (for the devices that come up in secure mode, like the hw
I have), but those depend on an scm patch that would come in through
another tree. If that can land in the next day or two, there might
be a second late pull request for drm/msm.
In addition to the new-shiny, there have also been a lot of overlay/
plane related fixes for issues found using drm-hwc2 (in the process of
testing/debugging the atomic/kms fence patches), resulting in rework
to assign hwpipes to kms planes dynamically (as part of global atomic
state) and also handling SMP (fifo) block allocation atomically as
part of the ->atomic_check() step. All those patches should also help
out atomic weston (when those patches eventually land).
* 'msm-next' of git://people.freedesktop.org/~robclark/linux: (36 commits)
drm/msm: gpu: Add support for the GPMU
drm/msm: gpu: Add A5XX target support
drm/msm: Disable interrupts during init
drm/msm: Remove 'src_clk' from adreno configuration
drm/msm: gpu: Add OUT_TYPE4 and OUT_TYPE7
drm/msm: Add adreno_gpu_write64()
drm/msm: gpu Add new gpu register read/write functions
drm/msm: gpu: Return error on hw_init failure
drm/msm: gpu: Cut down the list of "generic" registers to the ones we use
drm/msm: update generated headers
drm/msm/adreno: move scratch register dumping to per-gen code
drm/msm/rd: support for 64b iova
drm/msm: convert iova to 64b
drm/msm: set dma_mask properly
drm/msm: Remove bad calls to of_node_put()
drm/msm/mdp5: move LM bounds check into plane->atomic_check()
drm/msm/mdp5: dump smp state on errors too
drm/msm/mdp5: add debugfs to show smp block status
drm/msm/mdp5: handle SMP block allocations "atomically"
drm/msm/mdp5: dynamically assign hw pipes to planes
...
Dave Airlie [Wed, 30 Nov 2016 04:28:20 +0000 (14:28 +1000)]
Merge tag 'drm-misc-next-2016-11-29' of git://anongit.freedesktop.org/git/drm-misc into drm-next
Big thing is that drm-misc is now officially a group maintainer/committer
model thing, with MAINTAINERS suitably updated. Otherwise just the usual
pile of misc things all over, nothing that stands out this time around.
* tag 'drm-misc-next-2016-11-29' of git://anongit.freedesktop.org/git/drm-misc: (33 commits)
drm: Introduce drm_framebuffer_assign()
drm/bridge: adv7511: Enable the audio data and clock pads on adv7533
drm/bridge: adv7511: Add Audio support
drm/edid: Consider alternate cea timings to be the same VIC
drm/atomic: Constify drm_atomic_crtc_needs_modeset()
drm: bridge: dw-hdmi: add ASoC dependency
drm: Fix shift operations for drm_fb_helper::drm_target_preferred()
drm: Avoid NULL dereference for DRM_LEGACY debug message
drm: Use u64_to_user_ptr() helper for blob ioctls
drm: Fix conflicting macro parameter in drm_mm_for_each_node_in_range()
drm: Fixup kernel doc for driver->gem_create_object
drm/hisilicon/hibmc: mark PM functions __maybe_unused
drm/hisilicon/hibmc: Checking for NULL instead of IS_ERR()
drm: bridge: add DesignWare HDMI I2S audio support
drm: Check against color expansion in drm_mm_reserve_node()
drm: Define drm_mm_for_each_node_in_range()
drm/doc: Fix links in drm_property.c
MAINTAINERS: Add link to drm-misc documentation
vgaarb: use valid dev pointer in vgaarb_info()
drm/atomic: Unconfuse the old_state mess in commmit_tail
...
Dave Airlie [Wed, 30 Nov 2016 04:21:35 +0000 (14:21 +1000)]
Merge tag 'drm-intel-next-2016-11-21' of git://anongit.freedesktop.org/git/drm-intel into drm-next
Final 4.10 updates:
- fine-tune fb flushing and tracking (Chris Wilson)
- refactor state check dumper code for more conciseness (Tvrtko)
- roll out dev_priv all over the place (Tvrkto)
- finally remove __i915__ magic macro (Tvrtko)
- more gvt bugfixes (Zhenyu&team)
- better opregion CADL handling (Jani)
- refactor/clean up wm programming (Maarten)
- gpu scheduler + priority boosting for flips as first user (Chris
Wilson)
- make fbc use more atomic (Paulo)
- initial kvm-gvt framework, but not yet complete (Zhenyu&team)
* tag 'drm-intel-next-2016-11-21' of git://anongit.freedesktop.org/git/drm-intel: (127 commits)
drm/i915: Update DRIVER_DATE to 20161121
drm/i915: Skip final clflush if LLC is coherent
drm/i915: Always flush the dirty CPU cache when pinning the scanout
drm/i915: Don't touch NULL sg on i915_gem_object_get_pages_gtt() error
drm/i915: Check that each request phase is completed before retiring
drm/i915: i915_pages_create_for_stolen should return err ptr
drm/i915: Enable support for nonblocking modeset
drm/i915: Be more careful to drop the GT wakeref
drm/i915: Move frontbuffer CS write tracking from ggtt vma to object
drm/i915: Only dump dp_m2_n2 configuration when drrs is used
drm/i915: don't leak global_timeline
drm/i915: add i915_address_space_fini
drm/i915: Add a few more sanity checks for stolen handling
drm/i915: Waterproof verification of gen9 forcewake table ranges
drm/i915: Introduce enableddisabled helper
drm/i915: Only dump possible panel fitter config for the platform
drm/i915: Only dump scaler config where supported
drm/i915: Compact a few pipe config debug lines
drm/i915: Don't log pipe config kernel pointer and duplicated pipe name
drm/i915: Dump FDI config only where applicable
...
John Stultz [Tue, 29 Nov 2016 01:22:31 +0000 (17:22 -0800)]
drm/bridge: adv7511: Add Audio support
This patch adds support to Audio for both adv7511 and adv7533
bridge chips.
This patch was originally from [1] by Lars-Peter Clausen <lars@metafoo.de>
and was adapted by Archit Taneja <architt@codeaurora.org> and
Srinivas Kandagatla <srinivas.kandagatla@linaro.org>.
Then I heavily reworked it to use the hdmi-codec driver. And also
folded in some audio packet initialization done by Andy Green
<andy.green@linaro.org>. So credit to them, but blame to me.
Jordan Crouse [Mon, 28 Nov 2016 19:28:34 +0000 (12:28 -0700)]
drm/msm: gpu: Add support for the GPMU
Most 5XX targets have GPMU (Graphics Power Management Unit) that
handles a lot of the heavy lifting for power management including
thermal and limits management and dynamic power collapse. While
the GPMU itself is optional, it is usually nessesary to hit
aggressive power targets.
The GPMU firmware needs to be loaded into the GPMU at init time via a
shared hardware block of registers. Using the GPU to write the microcode
is more efficient than using the CPU so at first load create an indirect
buffer that can be executed during subsequent initalization sequences.
After loading the GPMU gets initalized through a shared register
interface and then we mostly get out of its way and let it do
its thing.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:32 +0000 (12:28 -0700)]
drm/msm: Disable interrupts during init
Disable the interrupt during the init sequence to avoid having
interrupts fired for errors and other things that we are not
ready to handle while initializing.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:31 +0000 (12:28 -0700)]
drm/msm: Remove 'src_clk' from adreno configuration
The adreno code inherited a silly workaround from downstream
from the bad old days before decent clock control. grp_clk[0]
(named 'src_clk') doesn't actually exist - it was used as a proxy
for whatever the core clock actually was (usually 'core_clk').
All targets should be able to correctly request 'core_clk' and
get the right thing back so zap the anachronism and directly
use grp_clk[0] to control the clock rate.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:29 +0000 (12:28 -0700)]
drm/msm: Add adreno_gpu_write64()
Add a new generic function to write a "64" bit value. This isn't
actually a 64 bit operation, it just writes the upper and lower
32 bit of a 64 bit value to a specified LO and HI register. If
a particular target doesn't support one of the registers it can
mark that register as SKIP and writes/reads from that register
will be quietly dropped.
This can be immediately put in place for the ringbuffer base and
the RPTR address. Both writes are converted to use
adreno_gpu_write64() with their respective high and low registers
and the high register appropriately marked as SKIP for both 32 bit
targets (a3xx and a4xx). When a5xx comes it will define valid target
registers for the 'hi' option and everything else will just work.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:28 +0000 (12:28 -0700)]
drm/msm: gpu Add new gpu register read/write functions
Add some new functions to manipulate GPU registers. gpu_read64 and
gpu_write64 can read/write a 64 bit value to two 32 bit registers.
For 4XX and older these are normally perfcounter registers, but
future targets will use 64 bit addressing so there will be many
more spots where a 64 bit read and write are needed.
gpu_rmw() does a read/modify/write on a 32 bit register given a mask
and bits to OR in.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:27 +0000 (12:28 -0700)]
drm/msm: gpu: Return error on hw_init failure
When the GPU hardware init function fails (like say, ME_INIT timed
out) return error instead of blindly continuing on. This gives us
a small chance of saving the system before it goes boom.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Jordan Crouse [Mon, 28 Nov 2016 19:28:26 +0000 (12:28 -0700)]
drm/msm: gpu: Cut down the list of "generic" registers to the ones we use
There are very few register accesses in the common code. Cut down
the list of common registers to just those that are used. This
saves const space and saves us the effort of maintaining registers
for A3XX and A4XX that don't exist or are unused.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Fri, 11 Nov 2016 17:06:46 +0000 (12:06 -0500)]
drm/msm: convert iova to 64b
For a5xx the gpu is 64b so we need to change iova to 64b everywhere. On
the display side, iova is still 32b so it can ignore the upper bits.
(Although all the armv8 devices have an iommu that can map 64b pa to 32b
iova.)
Ville Syrjälä [Thu, 3 Nov 2016 12:53:29 +0000 (14:53 +0200)]
drm/edid: Consider alternate cea timings to be the same VIC
CEA-861 specifies that the vertical front porch may vary by one or two
lines for specific VICs. Up to now we've only considered a mode to match
the VIC if it matched the shortest possible vertical front porch length
(as that is the variant we store in cea_modes[]). Let's allow our VIC
matching to work with the other timings variants as well so that that
we'll send out the correct VIC if the variant actually used isn't the
one with the shortest vertical front porch.
Chris Wilson [Sun, 27 Nov 2016 17:09:10 +0000 (17:09 +0000)]
drm: Fix shift operations for drm_fb_helper::drm_target_preferred()
smatch correctly warns:
drivers/gpu/drm/drm_fb_helper.c:1960 drm_target_preferred() warn: should '1 << i' be a 64 bit type?
drivers/gpu/drm/drm_fb_helper.c:2001 drm_target_preferred() warn: should '1 << i' be a 64 bit type?
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Sun, 27 Nov 2016 11:16:23 +0000 (11:16 +0000)]
drm: Fix conflicting macro parameter in drm_mm_for_each_node_in_range()
start is being used as both a macro parameter and as a member of struct
drm_mm_node (node->start). This causes a conflict as cpp then tries to
replace node->start with the passed in string for "start". Work just
fine so long as you also happened to using local variables called start!
Fixes: 522e85dd8677 ("drm: Define drm_mm_for_each_node_in_range()") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Christian König <christian.koenig@amd.com>.
[danvet: Fixup kerneldoc.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161127111623.11124-1-chris@chris-wilson.co.uk
Archit Taneja [Thu, 17 Nov 2016 06:42:03 +0000 (12:12 +0530)]
drm/msm: Remove bad calls to of_node_put()
In add_components_mdp, we parse the endpoints in MDP output ports
using the helper for_each_endpoint_of_node(). Our function calls
of_node_put() on the endpoint node before we iterate over the
next one. This is already done by the helper, and results in
trying to decrement the refcount twice.
Remove the extra of_node_put calls. This fixes warnings seen when
we try to insert the driver as a module on IFC6410.
Reported-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Archit Taneja <architt@codeaurora.org> Signed-off-by: Rob Clark <robdclark@gmail.com>
Rob Clark [Sat, 5 Nov 2016 14:43:55 +0000 (10:43 -0400)]
drm/msm/mdp5: move LM bounds check into plane->atomic_check()
The mode_config->max_{width,height} is for the maximum size of a fb, not
the max scanout limits (of the layer-mixer). It is legal, and in fact
common, to create a larger fb, only only scan-out a smaller part of it.
For example multi-monitor configurations for x11, or android wallpaper
layer (which is created larger than the screen resolution for fast
scrolling by just changing the src x/y coordinates).
Previously, SMP block allocation was not checked in the plane's
atomic_check() fxn, so we could fail allocation SMP block allocation at
atomic_update() time. Re-work the block allocation to request blocks
during atomic_check(), but not update the hw until committing the atomic
update.
Since SMP blocks allocated at atomic_check() time, we need to manage the
SMP state as part of mdp5_state (global atomic state). This actually
ends up significantly simplifying the SMP management, as the SMP module
does not need to manage the intermediate state between assigning new
blocks before setting flush bits and releasing old blocks after vblank.
(The SMP registers and SMP allocation is not double-buffered, so newly
allocated blocks need to be updated in kms->prepare_commit() released
blocks in kms->complete_commit().)
Rob Clark [Tue, 1 Nov 2016 15:56:54 +0000 (11:56 -0400)]
drm/msm/mdp5: dynamically assign hw pipes to planes
(re)assign the hw pipes to planes based on required caps, and to handle
situations where we could not modify an in-use plane (ie. SMP block
reallocation).
This means all planes advertise the superset of formats and properties.
Userspace must (as always) use atomic TEST_ONLY step for atomic updates,
as not all planes may be available for use on every frame.
The mapping of hwpipe to plane is stored in mdp5_state, so that state
updates are atomically committed in the same way that plane/etc state
updates are managed. This is needed because the mdp5_plane_state keeps
a pointer to the hwpipe, and we don't want global state to become out
of sync with the plane state if an atomic update fails, we hit deadlock/
backoff scenario, etc. The use of state_lock keeps multiple parallel
updates which both re-assign hwpipes properly serialized.
Rob Clark [Fri, 4 Nov 2016 18:50:08 +0000 (14:50 -0400)]
drm/msm/mdp5: add skeletal mdp5_state
Add basic state duplication/apply mechanism. Following commits will
move actual global hw state into this.
The state_lock allows multiple concurrent updates to proceed as long as
they don't both try to alter global state. The ww_mutex mechanism will
trigger backoff in case of deadlock between multiple threads trying to
update state.
Signed-off-by: Rob Clark <robdclark@gmail.com> Reviewed-by: Archit Taneja <architt@codeaurora.org>
Rob Clark [Tue, 1 Nov 2016 13:56:51 +0000 (09:56 -0400)]
drm/msm/mdp5: introduce mdp5_hw_pipe
Split out the hardware pipe specifics from mdp5_plane. To start, the hw
pipes are statically assigned to planes, but next step is to assign the
hw pipes during plane->atomic_check() based on requested caps (scaling,
YUV, etc). And then hw pipe re-assignment if required if required SMP
blocks changes.
Signed-off-by: Rob Clark <robdclark@gmail.com> Reviewed-by: Archit Taneja <architt@codeaurora.org>
Rob Clark [Tue, 1 Nov 2016 13:31:21 +0000 (09:31 -0400)]
drm/msm/mdp5: drop mdp5_plane::name
Just use plane->name now that it is a thing. In a following patch, once
we dynamically assign hw pipes to planes, it won't make sense to name
planes the way we do, so this also partly reduces churn in following
patch.
Rob Clark [Wed, 28 Sep 2016 23:58:32 +0000 (19:58 -0400)]
drm/msm: support multiple address spaces
We can have various combinations of 64b and 32b address space, ie. 64b
CPU but 32b display and gpu, or 64b CPU and GPU but 32b display. So
best to decouple the device iova's from mmap offset.
Rob Clark [Wed, 19 Oct 2016 21:53:53 +0000 (17:53 -0400)]
drm/msm: update uapi header license
The same file in libdrm is, as is the tradition with the rest of libdrm,
etc, using an MIT license. To avoid complications in the future with
sync'ing the uapi header to libdrm, lets fix the license mismatch now
before there are any non-trivial commits from someone other than myself.
Cc: Emil Velikov <emil.l.velikov@gmail.com> Cc: Gabriel Laskar <gabriel@lse.epita.fr> Cc: Mikko Rapeli <mikko.rapeli@iki.fi> Signed-off-by: Rob Clark <robdclark@gmail.com> Acked-by: Emil Velikov <emil.l.velikov@gmail.com>
Rob Clark [Thu, 13 Oct 2016 16:43:17 +0000 (12:43 -0400)]
drm/msm/mdp5: handle non-fullscreen base plane case
If the bottom-most layer is not fullscreen, we need to use the BASE
mixer stage for solid fill (ie. MDP5_CTL_BLEND_OP_FLAG_BORDER_OUT). The
blend_setup() code pretty much handled this already, we just had to
figure this out in _atomic_check() and assign the stages appropriately.
Also fix the case where there are zero enabled planes, where we also
need to enable BORDER_OUT.
Arnd Bergmann [Thu, 24 Nov 2016 16:30:26 +0000 (17:30 +0100)]
drm/hisilicon/hibmc: mark PM functions __maybe_unused
When CONFIG_PM_SLEEP is disabled, we get a harmless warning
drm/hisilicon/hibmc/hibmc_drm_drv.c:115:12: error: ‘hibmc_pm_resume’ defined but not used [-Werror=unused-function]
drm/hisilicon/hibmc/hibmc_drm_drv.c:97:12: error: ‘hibmc_pm_suspend’ defined but not used [-Werror=unused-function]
Marking the functions as __maybe_unused avoids the warning without
having to add an #ifdef.
Dave Airlie [Thu, 24 Nov 2016 23:55:33 +0000 (09:55 +1000)]
Merge branch 'drm-tda998x-devel' of git://git.armlinux.org.uk/~rmk/linux-arm into drm-next
These updates:
* improve the robustness of the driver wrt races
* improve the compliance for sending infoframes and audio
* re-organise the function order in the driver to group like functions
together. (This unfortunately causes a conflict with the change in
drm-misc, but it should be trivial to solve, although it looks more
scarey than it really is - sfr has already sent two reports about
this, one earlier today.)
* simplify tda998x_audio_get_eld and DPMS handling
* power down sections of the chip that we never use
* add some initial preparation for supporting the CEC driver
* 'drm-tda998x-devel' of git://git.armlinux.org.uk/~rmk/linux-arm:
drm/i2c: tda998x: fix spelling mistake
drm/i2c: tda998x: allow sharing of the CEC device accesses
drm/i2c: tda998x: allow interrupt to be shared
drm/i2c: tda998x: power down pre-filter and color conversion
drm/i2c: tda998x: switch to boolean is_on
drm/i2c: tda998x: remove complexity from tda998x_audio_get_eld()
drm/i2c: tda998x: group audio functions together
drm/i2c: tda998x: separate connector initialisation
drm/i2c: tda998x: group connector functions and funcs together
drm/i2c: tda998x: move and rename tda998x_encoder_set_config()
drm/i2c: tda998x: correct function name in comments
drm/i2c: tda998x: only enable audio if supported by sink
drm/i2c: tda998x: only configure infoframes and audio if supported
drm/i2c: tda998x: avoid race when programming audio
drm/i2c: tda998x: avoid racy access to mode clock
drm/i2c: tda998x: avoid race in tda998x_encoder_mode_set()
drm/i2c: tda998x: move audio mutex initialisation
Dave Airlie [Thu, 24 Nov 2016 23:03:27 +0000 (09:03 +1000)]
Merge branch 'drm-armada-devel' of git://git.armlinux.org.uk/~rmk/linux-arm into drm-next
Building on top of the MALI change previously merged, these changes:
* add tracing support for overlay updates
* refactor some of the plane support code
* de-midlayer the driver
* cleanups from other folk reviewing the code
* 'drm-armada-devel' of git://git.armlinux.org.uk/~rmk/linux-arm:
drm/armada: fix NULL pointer comparison warning
drm/armada: use DRM_FB_HELPER_DEFAULT_OPS for fb_ops
drm/armada: remove some dead code
drm/armada: mark symbols static where possible
drm/armada: de-midlayer armada
drm/armada: use common helper for plane base address
drm/armada: move setting primary plane position to armada_drm_primary_set()
drm/armada: split out primary plane update
drm/armada: move plane state to struct armada_plane
drm/armada: clean up armada_drm_plane_work_run()
drm/armada: add tracing support
drm: bridge: add DesignWare HDMI I2S audio support
Current dw-hdmi is supporting sound via AHB bus, but it has
I2S audio feature too. This patch adds I2S audio support to dw-hdmi.
This HDMI I2S is supported by using ALSA SoC common HDMI encoder
driver.
Chris Wilson [Wed, 23 Nov 2016 14:11:15 +0000 (14:11 +0000)]
drm: Check against color expansion in drm_mm_reserve_node()
Use the color_adjust callback when reserving a node to check if
inserting a node into this hole requires any additional space, and so if
that space then conflicts with an existing allocation.
Chris Wilson [Wed, 23 Nov 2016 14:11:14 +0000 (14:11 +0000)]
drm: Define drm_mm_for_each_node_in_range()
Some clients would like to iterate over every node within a certain
range. Make a nice little macro for them to hide the mixing of the
rbtree search and linear walk.
Arnd Bergmann [Tue, 22 Nov 2016 14:34:19 +0000 (15:34 +0100)]
vgaarb: use valid dev pointer in vgaarb_info()
We now pass the device to the debug messages, but on non-x86,
this is an invalid pointer in vga_arb_device_init:
drivers/gpu/vga/vgaarb.c: In function 'vga_arb_device_init':
drivers/gpu/vga/vgaarb.c:1467:4: error: 'dev' may be used uninitialized in this function [-Werror=maybe-uninitialized]
This moves the initialization of the dev pointer outside of the
architecture #ifdef.
Daniel Vetter [Mon, 21 Nov 2016 17:18:02 +0000 (18:18 +0100)]
drm/atomic: Unconfuse the old_state mess in commmit_tail
I totally butcherd the job on typing the kernel-doc for these, and no
one realized. Noticed by Russell. Maarten has a more complete approach
to this confusion, by making it more explicit what the new/old state
is, instead of this magic switching behaviour.
v2:
- Liviu pointed out that wait_for_fences is even more magic. Leave
that as @state, and document @pre_swap better.
- While at it, patch in header for the reference section.
- Fix spelling issues Russell noticed.
v3: Fix up the @pre_swap note (Liviu): Also s/synchronous/blocking/,
since async flip is something else than non-blocking.
Cc: Liviu Dudau <liviu.dudau@arm.com> Reported-by: Russell King - ARM Linux <linux@armlinux.org.uk> Cc: Russell King - ARM Linux <linux@armlinux.org.uk> Fixes: 9f2a7950e77a ("drm/atomic-helper: nonblocking commit support") Cc: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Tomeu Vizoso <tomeu.vizoso@gmail.com> Cc: Daniel Stone <daniels@collabora.com> Reviewed-by: Liviu Dudau <liviu.dudau@arm.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161121171802.24147-1-daniel.vetter@ffwll.ch
Liviu Dudau [Thu, 17 Nov 2016 11:41:29 +0000 (11:41 +0000)]
drm/atomic: cleanup debugfs entries on un-registering the driver.
Cleanup the debugfs entries created by
commit 6559c901cb48: drm/atomic: add debugfs file to dump out atomic state
when the driver's minor gets un-registered. Without it, DRM drivers
compiled as modules cannot be rmmod-ed and modprobed again.
Dave Airlie [Mon, 21 Nov 2016 01:42:17 +0000 (11:42 +1000)]
Merge tag 'drm-hisilicon-next-2016-11-17' of http://github.com/zourongrong/linux into drm-next
hibmc drm driver for hisilicon.
* tag 'drm-hisilicon-next-2016-11-17' of http://github.com/zourongrong/linux:
MAINTAINERS: Update HISILICON DRM entries
drm/hisilicon/hibmc: Add support for vblank interrupt
drm/hisilicon/hibmc: Add support for VDAC
drm/hisilicon/hibmc: Add support for display engine
drm/hisilicon/hibmc: Add support for frame buffer
drm/hisilicon/hibmc: Add video memory management
drm/hisilicon/hibmc: Add hisilicon hibmc drm master driver
Chris Wilson [Fri, 18 Nov 2016 21:17:47 +0000 (21:17 +0000)]
drm/i915: Skip final clflush if LLC is coherent
If the LLC is coherent with the object, we do not need to worry about
whether main memory and cache mismatch when we hand the object back to
the system.
Chris Wilson [Fri, 18 Nov 2016 21:17:46 +0000 (21:17 +0000)]
drm/i915: Always flush the dirty CPU cache when pinning the scanout
Currently we only clflush the scanout if it is in the CPU domain. Also
flush if we have a pending CPU clflush. We also want to treat the
dirtyfb path similar, and flush any pending writes there as well.
v2: Only send the fb flush message if flushing the dirt on flip
v3: Make flush-for-flip and dirtyfb look more alike since they serve
similar roles as end-of-frame marker.
Let's add a couple more BUG_ONs before this to ascertain that the request
did make it to hardware. The impossible part of this stacktrace is that
request must have been considered completed by the i915_request_wait()
before we tried to retire it.
Matthew Auld [Fri, 18 Nov 2016 17:02:16 +0000 (17:02 +0000)]
drm/i915: i915_pages_create_for_stolen should return err ptr
When gathering the pages from our backing storage we expect get_pages()
to either give us our sg_table or an err ptr. However when gathering our
fake pages for stolen memory we may return NULL in the event of a
failure. To prevent any funny business we should therefore return the
proper err ptr value.
Chris Wilson [Tue, 15 Nov 2016 16:46:20 +0000 (16:46 +0000)]
drm/i915: Be more careful to drop the GT wakeref
Since we can retire requests from multiple paths, we cannot assume that
i915_gem_retire_requests() is the sole path on which we can transition
to gt.active_requests == 0. A consequence of this is that we would skip
the function if we had already retired all the requests and not
scheduled the idle worker.
This is fallout from changing the routine from considering active_engines
(for which it was the only consumer) to active_requests.
v2: Move kicking the idle working to i915_gem_request_retire() otherwise
we could postpone the idle callback everytime we called retire_requests
even though we did no work.
v3: We only need to move the idle work kicking!
v4: Drop the BUG_ON(!awake) as we may be called from the shrinker in the
middle of constructing a request before we have marked the device awake.
v5: Add a BUG_ON() for active_requests underflow upon retirement (Joonas)
Chris Wilson [Wed, 16 Nov 2016 19:07:04 +0000 (19:07 +0000)]
drm/i915: Move frontbuffer CS write tracking from ggtt vma to object
I tried to avoid having to track the write for every VMA by only
tracking writes to the ggtt. However, for the purposes of frontbuffer
tracking this is insufficient as we need to invalidate around writes not
just to the the ggtt but all aliased ppgtt views of the framebuffer. By
moving the critical section to the object and only doing so for
framebuffer writes we can reduce the tracking even further by only
watching framebuffers and not vma.
Bjorn Helgaas [Thu, 17 Nov 2016 17:47:58 +0000 (11:47 -0600)]
vgaarb: Use dev_printk() when possible
Use dev_printk() when possible. This makes messages more consistent with
other device-related messages and, in some cases, adds useful information.
This changes messages like this:
vgaarb: failed to allocate pci device
vgaarb: setting as boot device: PCI:0000:01:00.0
vgaarb: device added: PCI:0000:01:00.0,decodes=io+mem,owns=io+mem,locks=none
vgaarb: bridge control possible 0000:01:00.0
to this:
pci 0000:01:00.0: vgaarb: failed to allocate VGA arbiter data
pci 0000:01:00.0: vgaarb: setting as boot VGA device
pci 0000:01:00.0: vgaarb: VGA device added: decodes=io+mem,owns=io+mem,locks=none
pci 0000:01:00.0: vgaarb: bridge control possible
Daniel Vetter [Thu, 17 Nov 2016 08:56:48 +0000 (09:56 +0100)]
drm: document standard connector properties
There's a really big pile of additional connector properties, a lot of
them standardized. But they're all for specific outputs (panels, TV,
scaling, ...) so I left them out for now since this is enough for a
start.
I typed this to give Manasi a place to add her new link status
property documentation.
v2: forgot to git add all the bits (Manasi).
v3: Be more epxlicit about integrated tiled panels (Archit)
Russell King [Thu, 3 Nov 2016 10:16:17 +0000 (10:16 +0000)]
drm/i2c: tda998x: allow sharing of the CEC device accesses
cec_read() is non-atomic in the presence of other I2C bus transactions
to the same device. This presents a problem when we add support for
the TDA9950 CEC engine part - both drivers can be trying to access the
device.
Avoid the inherent problems by switching to i2c_transfer() instead,
which allows us to perform more than one bus transaction atomically.
As this means we will be using I2C transactions rather than SMBUS, we
have to check that the host supports I2C functionality.
Tested-by: Brian Starkey <brian.starkey@arm.com> Reviewed-by: Brian Starkey <brian.starkey@arm.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Thu, 3 Nov 2016 08:58:04 +0000 (08:58 +0000)]
drm/i2c: tda998x: allow interrupt to be shared
Some TDA998x contain several different I2C devices - there is the HDMI
encoder, and there is a TDA9950 CEC engine. These two share the same
interrupt signal.
In order to allow a driver for the CEC engine to work, we need to be
able to share the interrupt with the CEC driver, so convert the handler
and registration to allow this to happen.
Tested-by: Brian Starkey <brian.starkey@arm.com> Reviewed-by: Brian Starkey <brian.starkey@arm.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Thu, 3 Nov 2016 15:19:06 +0000 (15:19 +0000)]
drm/i2c: tda998x: power down pre-filter and color conversion
Disabling the pre-filter block of the TDA998x saves 40mW and the colour
conversion block saves 15mW. As we always disable these two blocks, we
can power these sections of the chip down to save 55mW of unnecessary
power consumption.
Tested-by: Brian Starkey <brian.starkey@arm.com> Reviewed-by: Brian Starkey <brian.starkey@arm.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:39:04 +0000 (11:39 +0100)]
drm/i2c: tda998x: switch to boolean is_on
Rather than storing the DPMS mode (which will always be on or off) use a
boolean to store this instead.
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:31:44 +0000 (11:31 +0100)]
drm/i2c: tda998x: remove complexity from tda998x_audio_get_eld()
tda998x_audio_get_eld() is needlessly complex - the connector associated
with the encoder is always our own priv->connector. Remove this
complexity, but ensure that there are no races when copying out the ELD.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:30:56 +0000 (11:30 +0100)]
drm/i2c: tda998x: group audio functions together
Group the TDA998x audio functions together rather than split between
two different locations in the file, keeping like code together.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:30:56 +0000 (11:30 +0100)]
drm/i2c: tda998x: separate connector initialisation
Separate out the connector initialisation from the rest of the drivers
initialisation.
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:29:59 +0000 (11:29 +0100)]
drm/i2c: tda998x: group connector functions and funcs together
Group the TDA998x connector functions and funcs structures together
before the encoder support, rather than scattered amongst the rest of
the file. This keeps like code together.
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:25:02 +0000 (11:25 +0100)]
drm/i2c: tda998x: move and rename tda998x_encoder_set_config()
The naming of tda998x_encoder_set_config() is a left-over from when
TDA998x was a slave encoder. Since this is part of the initialisation,
drop the _encoder from the name, and move it near tda998x_bind().
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:24:22 +0000 (11:24 +0100)]
drm/i2c: tda998x: correct function name in comments
Correct two references to tda998x_connector_get_modes() which were
incorrectly referring to tda998x_encoder_get_modes().
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Wed, 2 Nov 2016 21:15:04 +0000 (21:15 +0000)]
drm/i2c: tda998x: only enable audio if supported by sink
Check for audio support by the attached sink by consulting the EDID
prior to enabling audio over the TMDS link. We must consult the EDID
after calling drm_helper_probe_single_connector_modes(), as this can
use an override EDID, or load a replacement EDID.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:32:42 +0000 (11:32 +0100)]
drm/i2c: tda998x: only configure infoframes and audio if supported
The CEA 861B specification indicates the situations when we are able to
send each infoframe based on the version of the EDID's CEA extension.
Update the tda998x driver to follow the CEA specification wrt sending
of infoframes.
Since we only support the generation of AVI version 2, this limits us
to CEA extension version 3, so we treat CEA extension version 2 as
CEA 861 (no infoframes, no audio.)
Tested-by: Robin Murphy <robin.murphy@arm.com> Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Wed, 2 Nov 2016 21:38:34 +0000 (21:38 +0000)]
drm/i2c: tda998x: avoid race when programming audio
Avoid a race between programming audio and an in-progress mode set.
A mode set is complex, and disables the ability to send infoframes
to the sink, and is disruptive to audio - we have to mute the audio
FIFO while doing a mode set.
If an attempt is made to start up the audio side, we will undo the
audio FIFO mute before the mode set has completed.
Move the lock so that we prevent audio interfering with an in-progress
mode set.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:32:43 +0000 (11:32 +0100)]
drm/i2c: tda998x: avoid racy access to mode clock
Avoid a racy access to the mode clock by storing the current mode clock
during a mode set under the audio mutex. This allows us to access it
from the audio path in a safe way.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Russell King [Sun, 23 Oct 2016 10:31:44 +0000 (11:31 +0100)]
drm/i2c: tda998x: avoid race in tda998x_encoder_mode_set()
As priv->audio_params can now be changed at run time, we need to be more
careful about how we deal with a mode set. We must take the audio lock
while checking if there's a valid audio configuration.
However, it's slightly worse than that - during mode set, we mute the
audio, and it must not be unmuted until we have finished the mode set.
It is possible that the audio side may start while a mode set is in
progress, so take the audio_mutex lock around the whole mode setting
procedure.
Tested-by: Jon Medhurst <tixy@linaro.org> Acked-by: Jon Medhurst <tixy@linaro.org> Tested-by: Jyri Sarha <jsarha@ti.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Matthew Auld [Thu, 17 Nov 2016 21:04:10 +0000 (21:04 +0000)]
drm/i915: add i915_address_space_fini
We already have an i915_address_space_init, so for symmetry we should
also have a _fini, plus we already open code it twice. This then also
fixes a bug where we leak the timeline for the ggtt vm.
v2: don't forget about the struct_mutex for the ggtt path.
Chris Wilson [Thu, 17 Nov 2016 15:58:46 +0000 (15:58 +0000)]
drm/i915: Add a few more sanity checks for stolen handling
We should never be called via obj->ops->release() on anything other than
a fully formed stolen object, so raise that to an assert. In the process
tidy up a comment and variable no longer used outside of a conditional
BUG.