Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

DRM_VMW_EVENT_FENCE_SIGNALED (struct drm_vmw_event_fence) and
DRM_EVENT_VBLANK (struct drm_event_vblank) pass timestamps in 32-bit
seconds/microseconds format.

As of commit c61eef72 ("drm: add support for monotonic vblank
timestamps"), other DRM drivers use monotonic times for drm_event_vblank,
but vmwgfx still uses CLOCK_REALTIME for both events, which suffers from
the y2038/y2106 overflow as well as time jumps.

For consistency, this changes vmwgfx to use ktime_get_ts64 as well,
which solves those problems and avoids the deprecated do_gettimeofday()
function.

This should be transparent to to user space, as long as it doesn't
compare the time against the result of gettimeofday().
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

We were relying on the pinned screen object backup buffer to be destroyed
when not used. But if we hold a copy of the atomic state, like when
hibernating, the backup buffer might not be destroyed since it's
refcounted by the atomic state. This causes us to hibernate with a
buffer pinned in VRAM.

Fix this by only having the buffer pinned when it is actually used by a
screen object.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>

For legacy surfaces, they were previously registered as device resources
when the driver resources were created. Since they are evictable we instead
register them as device resources once they are created on the device,
just like for guest-backed surfaces. This has implications during
hibernation where we can't hibernate with device resources active.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Reviewed-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>

Use kasprintf instead of combination of kmalloc and sprintf. Also,
remove the local variables used for storing the string length as they
are not required now.
Signed-off-by: NHimanshu Jha <himanshujha199640@gmail.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

It was used to early block fbdev dirty processing. Replace it with an
unprotected check of the par->dirty.active field. While this might
race with the vmw_fb_off() function, we do a protected check later so
the race will at worst lead to grabbing and releasing a couple of locks.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>

Make it possible to hibernate also with masters that don't switch VT at
hibernation time. We save and restore modesetting state unless fbdev is
active and enabled at hibernation time.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>

fbdev framebuffers were previously pinned to be able to keep them mapped
across updates.

This commit introduces a mechanism that instead revalidates the map on
each update, keeping the map cached across updates. The cached map is torn
down if the underlying pages change. Typically on buffer object moves and
swapouts.

This should be nicer to the system when we have resource contention.

Testing done: Basic fbdev functionality under Fedora 27.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NSinclair Yeh <syeh@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Reviewed-by: NDeepak Rawat <drawat@vmware.com>

The start / stop and preempt commands don't honor the context argument
but rather acts on all available contexts.

Also add detection for context 1 availability.

Note that currently there's no driver interface for submitting buffers
using the high-priority command queue (context 1).

Testing done:
Change the default context for command submission to 1 instead of 0,
verify basic desktop functionality including faulty command injection and
recovery.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Reviewed-by: NDeepak Rawat <drawat@vmware.com>

This blit was previously performed using two large vmaps, one of which
was teared down and remapped on each blit. Use the more resource-
conserving TTM cpu blit instead.

The blit is used in boundary-box computing mode which makes it possible
to minimize the bounding box used in host operations.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>

The utility uses kmap_atomic() instead of vmapping the whole buffer
object. As a result there will be more book-keeping but on some
architectures this will help avoid exhausting vmalloc space and also
avoid expensive TLB flushes.

The blit utility also adds a provision to compute a bounding box of
changed content, which is very useful to optimize presentation speed
of ill-behaved applications that don't supply proper damage regions, and
for page-flips. The cost of computing the bounding box is not that
expensive when done in a cpu-blit utility like this.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>

It will be used by vmwgfx cpu blit.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

Use helpers to perform the kmap_atomic_prot() functionality to
a) Avoid in-function ifdefs that violate the kernel coding policy,
b) Facilitate exporting the functionality.

This commit should not change any functionality.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

The current limit of 2 leads to some GPU idle times, as the usual
IRQ latency leads to up to 3 jobs getting signaled at once with some
standard workloads.

A larger HW job limit might lead to slightly worse QoS, but we accept
that to not sacrifice GPU throughput in the common case.
Signed-off-by: NLucas Stach <l.stach@pengutronix.de>

Use drm_plane_helper_check_update also for the cursor plane.
Some applications, like gdm on gnome shell still uses cursor front-buffer
like rendering without notifying the kernel. We do need some kind of
noficiation, but work around this for now by updating the cursor image on
every cursor move.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NBrian Paul <brianp@vmware.com>

Page flip can be slow for vmwgfx in some cases, like need to do surface
copy to different surface or waiting for IN_FENCE_FD. Enabling
nonblocking commits for vmwgfx in case userspace request it.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

Atomic ioctl can also send the same page flip flags as legacy ioctl.
In those cases also need to send the vblank event to userspace.

vmwgfx does not support flag DRM_MODE_PAGE_FLIP_ASYNC, so this flag is
never expected.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

The dmabuf_dirty/surface_dirty in case of screen object is moved to
plane atomic update, so that page flip in atomic ioctl also works.

vmwgfx does not support DRM_MODE_PAGE_FLIP_ASYNC, so this flag is never
expected.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

The function drm_crtc_arm_vblank_event should be used for the driver
which have vblank interrupt support. In case of vmwgfx we do not have
vblank interrupt.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

When display surface is different than the framebuffer surface, atomic
path do not copy the surface data. This commit moved the code to copy
surface from legacy to atomic path.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

In case of page flip there is no need to iterate over all display unit
in the function "vmw_kms_helper_dirty". If crtc is available then
dirty commands is performed on that crtc only.
Signed-off-by: NDeepak Rawat <drawat@vmware.com>
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>

Replace drm_gem_object_reference/unreference function with *_get/put()
suffixes, because it is shorter and consistent with the kernel
kref_get/put() functions. The following Coccinelle script was used:

@@
expression e;
@@

(
-drm_gem_object_reference(e);
+drm_gem_object_get(e);
|
-drm_gem_object_unreference(e);
+drm_gem_object_put(e);
|
-drm_gem_object_unreference_unlocked(e);
+drm_gem_object_put_unlocked(e);
)
Signed-off-by: NSantha Meena Ramamoorthy <santhameena13@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/1521570567-22519-1-git-send-email-santhameena13@gmail.com

platform_get_resource_byname() may fail and return NULL, so we should
better check it's return value to avoid a NULL pointer dereference
a bit later in the code.

This is detected by Coccinelle semantic patch.

@@
expression pdev, res, n, t, e, e1, e2;
@@

res = platform_get_resource_byname(pdev, t, n);
+ if (!res)
+   return -EINVAL;
... when != res == NULL
e = devm_ioremap(e1, res->start, e2);
Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: NNeil Armstrong <narmstrong@baylibre.com>
Acked-by: NNeil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1521555630-29284-1-git-send-email-weiyongjun1@huawei.com

Now that we have the guarantee that we will have only a single YUV plane,
actually support them. The way it works is not really straightforward,
since we first need to enable the YUV mode in the plane that we want to
setup, and then we have a few registers to setup the YUV buffer and
parameters.

We also need to setup the color correction to actually have something
displayed.
Reviewed-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/66088c1398bd3189123f28a89a7ccc669fe9f296.1519931807.git-series.maxime.ripard@bootlin.com

Just like for the frontend, a single plane can use a YUV format. Make sure
we have that constraint covered in our atomic_check.

This is preliminary to the actual YUV support to make sure we don't end up
in an impossible to support situation.
Reviewed-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/2f8586493d9139b12efe7e94f65e9a149f818e0e.1519931807.git-series.maxime.ripard@bootlin.com

This patch adds support for the compatible strings of the A80 display
pipeline.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180315114136.24747-6-wens@csie.org

The Allwinner A80 SoC has 2 documented TCONs. The display pipeline
diagram from the user manual shows a third TCON, but it's missing
an interrupt line, and its registers are not explained either.
It's also not used in Allwinner's vendor BSP.

The first TCON only has channel 0, for LCD panel output. The TCON
hardware setup is peculiar in that the eDP reset must also be
deasserted to allow access to the TCON. How the eDP module is wired
in the SoC itself is never explained.

The second TCON only has channel 1, and its output is connected to
the HDMI encoder block.

This patch adds a "needs_edp_reset" field to the tcon quirks structure,
and adds quirks and compatible strings for the 2 documented TCONs.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180315114136.24747-4-wens@csie.org

These macros are similar to the DRM_<LEVEL> with the addition
of a struct device * to the arguments.

Convert the single drm_dev_printk function into 2 separate functions.
drm_dev_printk with a KERN_<LEVEL> * for generic use and drm_dev_dbg
for conditional masked use.

Remove the __func__ argument and use __builtin_return_address(0) to be
similar to the DRM_<LEVEL> macros uses.

Convert the DRM_DEV_<LEVEL> macros to remove now unnecessary arguments
and use a consistent style.

These macros are rarely used in the generic gpu/drm code so the code
size does not change much for a defconfig, but when more drivers are
enabled, there is ~4k savings.

Many of these macros have no existing use at all.

$ size -t drivers/gpu/drm/built-in.a | tail -1
1877530	  44651	    995	1923176	 1d5868	(TOTALS)

$ size -t drivers/gpu/drm/built-in.a | tail -1
1877527	  44651	    995	1923173	 1d5865	(TOTALS)

$ size -t drivers/gpu/drm/built-in.a | tail -1
17166750	2689238	 108352	19964340	130a1b4	(TOTALS)

$ size -t drivers/gpu/drm/built-in.a | tail -1
17168888	2691734	 108352	19968974	130b3ce	(TOTALS)
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/e5c164946e15375ac71b69b75f296efdf0b76e6d.1521233717.git.joe@perches.com

Fix a couple of checkpatch issues
Signed-off-by: NPaul McQuade <paulmcquad@gmail.com>
[seanpaul squashed series of 4 into one patch, and changed commit msg]
Signed-off-by: NSean Paul <seanpaul@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20180319005225.1545-1-paulmcquad@gmail.com

The adreno driver stopped building when CONFIG_DEBUGFS is disabled:

drivers/gpu/drm/msm/adreno/adreno_device.c: In function 'adreno_load_gpu':
drivers/gpu/drm/msm/adreno/adreno_device.c:153:16: error: 'const struct msm_gpu_funcs' has no member named 'debugfs_init'
  if (gpu->funcs->debugfs_init) {
                ^~
drivers/gpu/drm/msm/adreno/adreno_device.c:154:13: error: 'const struct msm_gpu_funcs' has no member named 'debugfs_init'
   gpu->funcs->debugfs_init(gpu, dev->primary);
             ^~

This adds an #ifdef around the code that references the hidden
pointer.

Fixes: 331dc0bc ("drm/msm: add a5xx specific debugfs")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NRob Clark <robdclark@gmail.com>

If there is only a single DSI interface, don't reserve the first two
layer-mixers for the dual-DSI use-case.

This was causing problems for WB, not being able to assign a LM, on
8x16, which has only two LM's and a single DSI.
Signed-off-by: NRob Clark <robdclark@gmail.com>

For some reason, layer-mixer 3 and 4 were missing.  LM3 is used for
writeback on 8x16.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Signed-off-by: NRob Clark <robdclark@gmail.com>

For DSI cmd-mode and writeback, we need to write the CTL's START
register to kick things off, but we only want to do that once both
the encoder and the crtc have a chance to write their corresponding
flush bits.  The difficulty is that when there is a full modeset
(ie. encoder state has changed) we want to defer the start until
encoder->enable().  But if only plane's have changed, we want to do
this from crtc->commit().

The start_mask was a previous attempt to handle this, but it didn't
really do the right thing since atomic conversion.

Instead track in the crtc state that the start should be deferred,
set to try from encoder's (or in future writeback's) atomic_check().
This way the state is part of the atomic state, and rollback can
work properly if an atomic test fails.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Interrupt commands causes the CP to trigger an interrupt as the command
is processed, regardless of the GPU being done processing previous
commands. This is seen by the interrupt being delivered before the
fence is written on 8974 and is likely the cause of the additional
CP_WAIT_FOR_IDLE workaround found for a306, which would cause the CP to
wait for the GPU to go idle before triggering the interrupt.

Instead we can set the (undocumented) BIT(31) of the CACHE_FLUSH_TS
which will cause a special CACHE_FLUSH_TS interrupt to be triggered from
the GPU as the write event is processed.

Add CACHE_FLUSH_TS to the IRQ masks of A3xx and A4xx and remove the
workaround for A306.
Suggested-by: NJordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: NRob Clark <robdclark@gmail.com>

This should be using drm_gem_object_put().  Also since this is done only
in driver unload path, we don't need to synchronize setting tx_gem_obj
to NULL, so juse use the _unlocked() variant.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Remnants of pre-dma_fence fencing which got left behind by mistake.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Since new display controller is called "dpu" instead of "mdp".  Lets
make the name of the toplevel directory for the display controllers a
bit more generic.
Signed-off-by: NRob Clark <robdclark@gmail.com>
Reviewed-by: NSean Paul <seanpaul@chromium.org>

_dev_ is being dereferenced before it is null checked, hence there
is a potential null pointer dereference.

Fix this by moving the pointer dereference after _dev_ has been
null checked.

Fixes: d4e7f38d70ef ("drm/msm/dsi: check msm_dsi and dsi pointers before use")
Signed-off-by: NGustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>

_minor_ is being dereferenced before it is null checked, hence there
is a potential null pointer dereference. Fix this by moving the pointer
dereference after _minor_ has been null checked.

Fixes: 024ad8df763f ("drm/msm: add a5xx specific debugfs")
Signed-off-by: NGustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>