Apparently some DP sinks are a little nuts and cause HPD to drop
intermittently during modesets. This happens eg. on an ASUS PB287Q.
In oder to recover from this we can't really use the previous
connector status to determine if the link needs retraining, so let's
just ignore that piece of information and do the retrain
unconditionally. We do of course still check whether the link is
supposed to be running or not.
To actually get read out the EDID and update things properly we
also need to nuke the goto out added by commit 7d23e3c37b
("drm/i915: Cleaning up intel_dp_hpd_pulse"). I'm actually not sure
why that was there. Perhaps to avoid an EDID read if the connector
status didn't appear to change, but that sort of thing is quite racy
and would have failed anyway if we failed to keep up with the
hotplugs (if we missed the HPD down in between two HPD ups). And
now that we take this codepath unconditionally we definitely need
to drop the goto as otherwise we would never do the EDID read.
v2: Drop the goto that made us skip EDID reads entirely. Doh!
v3: Rebase due to locking changes
s/apparely/apparently/ in the comment (Chris)
Cc: stable@vger.kernel.org
Cc: Manasi Navare <manasi.d.navare@intel.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Reported-by: Palmer Dabbelt <palmer@dabbelt.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99766
References: https://lists.freedesktop.org/archives/intel-gfx/2017-February/119779.html
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170412193017.21029-1-ville.syrjala@linux.intel.com
(cherry picked from commit 1a36147bb9)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
If the engine is continually completing nops, we can saturate the
signaler and keep it working indefinitely. This angers the NMI watchdog!
A good example is to disable semaphores on snb and run igt/gem_exec_nop -
the parallel, multi-engine workloads are more than sufficient to hog the
CPU, preventing the system from even processing ICMP echo replies.
v2: Tvrtko dug into cond_resched() on x86 and found that it only
depended upon preempt_count and not tif_need_resched() - which means
that we would always call schedule() at that point.
Fixes: c81d46138d ("drm/i915: Convert trace-irq to the breadcrumb waiter")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170404120531.10737-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
(cherry picked from commit a7980a640c)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
gvt-next-fixes-2017-04-20
- some code optimization from Changbin
- debug message cleanup after QoS merge
- misc fixes for display mmio init, reset vgpu warning, etc.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
When downscaling, mclk needs to be sufficiently higher than pxlclk in
order to be able to fetch the higher-resolution data and produce output
pixels. When not scaling, or when upscaling, mclk can be equal to
pxlclk. Since the driver doesn't control mclk, just ensure that the
requirement is satisfied with the current clock rate.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Apply image enhacement when we are upscaling by a factor of 2
or more in either direction.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Enable the scaling engine for upscaling a single plane using the polyphase
scaler. No image enhancement support or downscaling yet*, and composition
result scaling is not implemented.
* Downscaling a plane requires mclk > pxlclk.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Add a core_id file in the driver's sysfs directory,
exposing the hardware CORE ID. This is useful to allow
userspace to discover the hardware version used.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
All DPs have a COLORADJ matrix which is applied prior to output gamma.
Attach that to the CTM property. Also, ensure the input CTM's coefficients
can fit in the DP registers' Q3.12 format.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Add gamma via the DRM GAMMA_LUT/GAMMA_LUT_SIZE CRTC
properties. The expected LUT size is 4096 in order
to produce as accurate a set of segments as possible.
This version uses only the green channel's gamma curve
to set the hardware curve on DP550/650. For the sake of
simplicity, it uses the same table of coefficients for
all 3 curves on DP500.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Add a custom CRTC state struct to enable storing driver's private
state. This patch only adds the base drm_crtc_state struct and
the atomic functions that handle it.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Now that we are using a custom plane state in mali-dp, we need a
bespoke reset that takes into account the larger structure.
Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
[Updated commit message]
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
The newly introduced function causes a harmless build warning:
drivers/gpu/drm/arm/malidp_planes.c: In function 'malidp_plane_atomic_print_state':
drivers/gpu/drm/arm/malidp_planes.c:98:23: error: unused variable 'mp' [-Werror=unused-variable]
The variable serves no purpose here and can be removed.
Fixes: 242f43b69c61 ("drm: mali-dp: add atomic_print_state for planes")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Enable runtime and system Power Management. Clocks are now managed
from malidp_crtc_{enable,disable} functions. Suspend-to-RAM tested
as working on Juno.
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Mali DP needs to have all the planes that are becoming inactive in the
new state disabled before re-enabling the active CRTC, otherwise we
start streaming out data from old pointers in memory.
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Change "S_IRUSR | S_IWUSR" to "0600", it's easier to parse mentally.
This change should be part of commit 50f837371d ("drm/vmwgfx: Revert
"drm/vmwgfx: Replace numeric parameter like 0444 with macro""), but the
patch was truncated somewhere in the patch route, so add the missing
change. Extract from the original commit message:
commit 50f837371d
Author: Øyvind A. Holm <sunny@sunbase.org>
Date: Thu Mar 23 14:54:48 2017 -0700
drm/vmwgfx: Revert "drm/vmwgfx: Replace numeric parameter like 0444
with macro"
This reverts commit 2d8e60e8b0 ("drm/vmwgfx: Replace numeric
parameter like 0444 with macro")
The commit belongs to the series of 1285 patches sent to LKML on
2016-08-02, it changes the representation of file permissions from
the octal value "0600" to "S_IRUSR | S_IWUSR".
The general consensus was that the changes does not increase
readability, quite the opposite; 0600 is easier to parse mentally
than S_IRUSR | S_IWUSR.
Signed-off-by: Øyvind A. Holm <sunny@sunbase.org>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Few parts of kernel define their own macro for aligning down so provide
a common define for this, with the same usage and assumptions as existing
ALIGN.
Convert also three existing implementations to this one.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
drm-misc-next-fixes-2017-04-20
Core changes:
- Maintain sti via drm-misc (Vincent)
- Rename dma_buf_ops->kmap_* to avoid naming collision (Logan)
Driver changes:
- Fix UHD displays on stih407 (Vincent)
- Fix uninitialized var return in atmel-hlcdc (Dan)
* tag 'drm-misc-next-fixes-2017-04-20' of git://anongit.freedesktop.org/git/drm-misc:
dma-buf: Rename dma-ops to prevent conflict with kunmap_atomic macro
drm: atmel-hlcdc: Uninitialized return in atmel_hlcdc_create_outputs()
drm/sti: fix GDP size to support up to UHD resolution
MAINTAINERS: add drm/sti driver into drm-misc
The contents of a ring are only valid between HEAD and TAIL, when the
ring is idle (HEAD == TAIL) we can simply let the pages go under memory
pressure if they are not pinned by an active context. Any new content
will be written after HEAD and so the ring will again be valid between
HEAD and TAIL, everything outside can be discarded.
Note that we take care of ensuring that we do not reset the HEAD
backwards following a GPU hang on an idle ring.
The same precautions are what enable us to use stolen memory for rings.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170420101709.27250-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Currently, the audio sampler clock is enabled from dw_hdmi_setup() at
step E. and is kept enabled for later use. This clock should be enabled
and disabled along with the actual audio stream and not always on (that
is bad for PM). Furthermore, as described by the datasheet, the I2S
variant needs to gate/ungate the clock when the stream is
enabled/disabled.
This commit adds a parameter to hdmi_audio_enable_clk() that controls
when the audio sample clock must be enabled or disabled. Then, it adds
the call to this function from dw_hdmi_i2s_audio_enable() and
dw_hdmi_i2s_audio_disable().
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Romain Perier <romain.perier@collabora.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170414083113.4255-3-romain.perier@collabora.com
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Currently, CTS+N is forced to zero as a workaround of the IP block for
i.MX platforms. This is requested in the datasheet of the corresponding
IP for AHB mode only. However, we have seen that it introduces glitches
or delays when playing a sound on HDMI for I2S mode. This proves that we
cannot keep the current functions for handling audio stream as-is if
these contain workaround that are specific to a mode.
This commit introduces two callbacks, one for each variant.
dw_hdmi_setup defines the right function depending on the detected
variant. Then, the exported functions dw_hdmi_audio_enable and
dw_hdmi_audio_disable calls the corresponding callbacks
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Romain Perier <romain.perier@collabora.com>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20170414083113.4255-2-romain.perier@collabora.com
Allwinner DRM changes for 4.12
Not any functional changes, but a lot of preliminary rework in order to
support multiple display pipelines.
* tag 'sunxi-drm-for-4.12' of https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux: (26 commits)
MAINTAINERS: Add sun4i-drm git repo
drm/sun4i: Pass pointer for underlying backend into layer init
drm/sun4i: Pass pointers for associated backend and tcon into crtc init
drm/sun4i: tv: Get tcon and backend pointers from associated crtc
drm/sun4i: Use embedded tcon pointer to get the tcon's output port node
drm/sun4i: Fix tcon channel 0 comment about backporch = backporch + hsync
drm/sun4i: Fix TCON clock and regmap initialization sequence
drm/sun4i: Grab reserved memory region
drm/sun4i: Add backend and tcon pointers to sun4i_crtc
drm/sun4i: Add backend pointer to sun4i_layer
drm/sun4i: rgb: Pass tcon pointer when initializing RGB encoder
drm/sun4i: tv: Switch to drm_of_find_possible_crtcs
drm/sun4i: Drop hardcoded .possible_crtcs values from layers
drm/sun4i: Drop primary layer pointer from sun4i_drv
drm/sun4i: Initialize crtc from tcon bind function
drm/sun4i: Move layers from sun4i_drv to sun4i_crtc
drm/sun4i: Add end of list element for sun4i_layers_init's returned list
drm/sun4i: Set drm_crtc.port to the underlying TCON's output port node
drm/sun4i: Make sunxi_rgb2yuv_coef constant
drm/sun4i: Make sun4i_crtc_init return ERR_PTR style error codes
...
Until now, we've had to limit Raspberry Pi to 256MB of CMA memory to
keep from triggering the hardware addressing bug between the tile
binner and the tile alloc memory (where the top 4 bits come from the
tile state data array's address).
To work around that and allow more memory to be reserved for graphics,
allocate a single BO to store tile state data arrays and tile
alloc/overflow memory while the GPU is active, and make sure that that
one BO doesn't happen to cross a 256MB boundary. With that in place,
we can allocate textures and shaders anywhere in system memory (still
contiguous, of course).
Signed-off-by: Eric Anholt <eric@anholt.net>
Link: http://patchwork.freedesktop.org/patch/msgid/20170327231025.19391-1-eric@anholt.net
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
The HDMI driver is currently enabling all clocks at probe time and
keeps the power-domain connected to the HDMI encoder enabled.
Move all activation code to vc4_hdmi_encoder_enable() and make sure
the clks and power domain are released when the HDMI encoder is not used
by adding deactivation steps in vc4_hdmi_encoder_disable().
Note that the sequencing imposed by the IP requires that we move
vc4_hdmi_encoder_mode_set() code into vc4_hdmi_encoder_enable().
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
A group of Linux kernel hackers reported chasing a bug that resulted
from their assumption that SLAB_DESTROY_BY_RCU provided an existence
guarantee, that is, that no block from such a slab would be reallocated
during an RCU read-side critical section. Of course, that is not the
case. Instead, SLAB_DESTROY_BY_RCU only prevents freeing of an entire
slab of blocks.
However, there is a phrase for this, namely "type safety". This commit
therefore renames SLAB_DESTROY_BY_RCU to SLAB_TYPESAFE_BY_RCU in order
to avoid future instances of this sort of confusion.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: <linux-mm@kvack.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
[ paulmck: Add comments mentioning the old name, as requested by Eric
Dumazet, in order to help people familiar with the old name find
the new one. ]
Acked-by: David Rientjes <rientjes@google.com>
All the error codes we (ab)use are strictly not the right ones (since
they're all for the vfs, and the only thing we're allowed to do from
an ioctl is EINVAL). But ENOENT is the common error code for failed to
look up an object throughout drm, so let's use it in the cma helpers,
too.
Note that this does change uabi for some drivers, but that shouldn't
ba a problem:
- other drivers (e.g. i915) already return ENOENT for invalid gem bo
- userspace tends to only check for errors, not specific ones (except
when ioctl restarting is needed)
- if the gem bo isn't there a previous step (either creating or
importing it) failed already, this situation should always be a
userspace bug.
Changing it does help with debugging userspace bugs though, by making
error reporting more consistent.
Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170418121120.31862-1-daniel.vetter@ffwll.ch
There are two bugs here. The && should be || and the > is off by one so
it should be >= ARRAY_SIZE().
Fixes: 8453d674ae ("drm/i915/gvt: vGPU execlist virtualization")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
