drm/edid: convert drm_mode_detailed() to drm_edid

We'll need to propagate drm_edid everywhere.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/6e4625d529ee4bfb2f6ebbfef5fb707d65a0554d.1652097712.git.jani.nikula@intel.com

drm/edid: convert struct detailed_mode_closure to drm_edid

We'll need to propagate drm_edid everywhere.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/da4c927cd973766568a79ef044254fff20317e7f.1652097712.git.jani.nikula@intel.com

drm/edid: convert drm_edid_connector_update() to drm_edid fully

We'll need to propagate drm_edid everywhere.

v3: s/edid/drm_edid/ in comment too (Ankit)

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220511092650.4193330-1-jani.nikula@intel.com

drm/edid: propagate drm_edid to drm_edid_to_eld()

We'll need to propagate drm_edid everywhere.

v3: Rebase

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220511092349.4186498-1-jani.nikula@intel.com

drm/edid: keep propagating drm_edid to display info

We'll need to propagate drm_edid everywhere.

v3: Update comment to refer to update_display_info() (Ankit)

v2: Use drm_edid_legacy_init()

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220511092104.4179637-1-jani.nikula@intel.com

drm/edid: start propagating drm_edid to lower levels

We'll need to propagate drm_edid everywhere. This is a clunky start, but
a start nonetheless. We'll eventually convert all of the EDID parsing to
struct drm_edid.

Initially, we'll just create the struct drm_edid in stack. This will be
the compat layer for legacy struct edid code. In the future, we'll have
EDID read return drm_edid objects.

v2: Add legacy init helper.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/000452fddedbaf7f473ac25d4dde2502e60b8e39.1652097712.git.jani.nikula@intel.com

drm/edid: add struct drm_edid container

Introduce new opaque type struct drm_edid to encapsulate the EDID data
and the size allocated for it. The contents will be private to
drm_edid.c.

There are a number of reasons for adding a container around struct edid:

* struct edid is a raw blob pointer to data that usually originates
  outside of the kernel. Its size is contained within the structure.

* There's no way to attach meta information (such as allocated memory
  size) to struct edid.

* Validation of the EDID blob and its size become crucial, and it's
  spread all over the subsystem, with varying levels of accuracy.

* HDMI Forum has introduced an HF-EEODB extension that defines an
  override EDID size within an EDID extension. The size allocated for an
  EDID depends on whether the allocator understands the HF-EEODB
  extension. Given a struct edid *, it's impossible to know how much
  memory was actually allocated for it.

There are also some reasons for making the container type struct
drm_edid opaque and private to drm_edid.c:

* Have only one place for creating and parsing the EDID, to avoid
  duplicating bugs.

* Prepare for reading a pure DisplayID 2.0 from its own DDC address, and
  adding it within the same struct drm_edid container, transparently,
  and for all drivers.

* With the idea that the drm_edid objects are immutable during their
  lifetimes, it will be possible to refcount them and reduce EDID
  copying everywhere (this is left for future work).

Initially, just add the type. In follow-up, we'll start converting the
guts of drm_edid.c to use it, and finally add interfaces around it.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/f3ecabd8a219aea678ad00f7bcdecf77b27b3c78.1652097712.git.jani.nikula@intel.com

drm/edid: convert drm_for_each_detailed_block() to edid iter

We have an iterator for this, use it. It does include the base block,
but its tag is 0 and will be skipped.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/a68718819a3dbfbdaaaddca7f73afa5ac6b33d07.1652097712.git.jani.nikula@intel.com

drm/edid: use else-if in CTA extension parsing

Only one of the conditions can be true.

Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/3a7ba2217134b0ab4c4c72e4333231a4e0616e88.1652097712.git.jani.nikula@intel.com

fbdev: vesafb: Allow to be built if COMPILE_TEST is enabled

The driver has runtime but no build time dependency with X86, so it can
be built for testing purposes if the COMPILE_TEST option is enabled.

This is useful to have more build coverage and make sure that the driver
is not affected by changes that could cause build regressions.

Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505120419.314136-1-javierm@redhat.com

fbdev: Restart conflicting fb removal loop when unregistering devices

Drivers that want to remove registered conflicting framebuffers prior to
register their own framebuffer, call to remove_conflicting_framebuffers().

This function takes the registration_lock mutex, to prevent a race when
drivers register framebuffer devices. But if a conflicting framebuffer
device is found, the underlaying platform device is unregistered and this
will lead to the platform driver .remove callback to be called. Which in
turn will call to unregister_framebuffer() that takes the same lock.

To prevent this, a struct fb_info.forced_out field was used as indication
to unregister_framebuffer() whether the mutex has to be grabbed or not.

But this could be unsafe, since the fbdev core is making assumptions about
what drivers may or may not do in their .remove callbacks. Allowing to run
these callbacks with the registration_lock held can cause deadlocks, since
the fbdev core has no control over what drivers do in their removal path.

A better solution is to drop the lock before platform_device_unregister(),
so unregister_framebuffer() can take it when called from the fbdev driver.
The lock is acquired again after the device has been unregistered and at
this point the removal loop can be restarted.

Since the conflicting framebuffer device has already been removed, the
loop would just finish when no more conflicting framebuffers are found.

Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220511113039.1252432-1-javierm@redhat.com

drm/bridge: tc358767: Add DSI-to-(e)DP mode support

Implement DSI-to-e(DP) mode, which is a mix of currently supported
DSI-to-DPI and DPI-to-(e)DP modes. The input side is configured as
either DSI or DPI, the DP AUX channel is registered for both input
side options, and the DSI host is attached for both DPI and (e)DP
output side options.

One notable detail is that the DSI-to-(e)DP mode requires the Pixel
PLL to be always enabled, which is not needed for DPI-to-(e)DP mode
which gets the matching clock direct from DPI Pixel Clock instead.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Jonas Karlman <jonas@kwiboo.se>
Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Robert Foss <robert.foss@linaro.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429205644.245480-2-marex@denx.de

drm/bridge: tc358767: Factor out DSI and DPI RX enablement

Factor out register programming to configure the chip video RX side for
reception of video data from DSI or DPI. This is particularly useful in
the (e)DP output mode, where the video data can be received from either
DPI or DSI. While only the former is supported in (e)DP output mode so
far, this patch is added in preparation for addition of the later.

There is a change in the order or register programming in case of the
DSI-to-DPI mode. The DSI RX side is now programmed and enabled all in
one place after the output mode has been configured. Before this change,
the DSI RX has been programmed before the output mode has been set and
only enabled afterward. The order makes no difference however, since the
DSI RX is only enabled at the end either way.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Jonas Karlman <jonas@kwiboo.se>
Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Robert Foss <robert.foss@linaro.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429205644.245480-1-marex@denx.de

drm/mgag200: Warn once if trying to set start address on broken hardware.

Some MGA200 hardware is broken and can't use a start address > 0.

v2: Warn if startaddr is not 0, and hw doesn't support it.
  (instead of removing MGAG200_FLAG_HW_BUG_NO_STARTADD)

Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504134026.283417-4-jfalempe@redhat.com

drm/mgag200: Optimize damage clips

When there are multiple damage clips, previous code merged them into one
big rectangle. As the Matrox memory is very slow, it's faster to copy each
damage clip.

Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504134026.283417-2-jfalempe@redhat.com

drm/mgag200: Add FB_DAMAGE_CLIPS support

The driver does support damage clips, but doesn't advertise it.
So when running gnome/wayland on Matrox hardware, the full frame is
copied to the slow Matrox memory, which leads to very poor performances.

Add drm_plane_enable_fb_damage_clips() to advertise this capability to
userspace.

With this patch, gnome/wayland becomes usable on Matrox GPU.

Suggested-by: Jonas Ådahl <jadahl@gmail.com>
Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504134026.283417-2-jfalempe@redhat.com

MAINTAINERS: add Melissa to V3D maintainers

I've been contributing to v3d through improvements, reviews, testing,
debugging, etc. So, I'm adding myself as a co-maintainer of the V3D
driver.

Signed-off-by: Melissa Wen <mwen@igalia.com>
Acked-by: Juan A. Suarez <jasuarez@igalia.com>
Acked-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429193317.ol3u4e5jpt5jucox@mail.igalia.com

drm/bridge: cdns-dsi: Add support for pre_enable and post_enable control functions.

Add support for pre_enable and post_enable drm bridge control functions.
Making sure that host to be prepared before panel is powered up,
for the panels like TC358762.

Signed-off-by: Jayshri Pawar <jpawar@cadence.com>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220411140606.15654-1-jpawar@cadence.com

drm/todo: Add entry for converting kselftests to kunit

Many of the kselftests in DRM can be converted to kunit tests instead,
since that framework is more suitable for unit testing.

Suggested-by: Maxime Ripard <maxime@cerno.tech>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20220509130846.694397-1-javierm@redhat.com

MAINTAINERS: Add simpledrm driver co-maintainer

Thomas asked me to serve as co-maintainer for the simpledrm driver.

Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505172610.338299-1-javierm@redhat.com

drm/amdgpu: move internal vram_mgr function into the C file

No need to have those in the header.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220510113649.879821-2-christian.koenig@amd.com

drm/amdgpu: fix start calculation in amdgpu_vram_mgr_new

We still need to calculate a virtual start address for the resource to
aid checking of it is visible or not. Only matters on small BAR systems,
but better save than sorry.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220510113649.879821-1-christian.koenig@amd.com

drm/amdgpu: switch DM to atomic fence helpers v2

This gives us the standard atomic implicit and explicit fencing rules.

v2: move the wait to amdgpu_dm_atomic_commit_tail

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Cc: Roman Li <Roman.Li@amd.com>
Cc: Qingqing Zhuo <qingqing.zhuo@amd.com>
Cc: Jude Shih <shenshih@amd.com>
Cc: Wayne Lin <Wayne.Lin@amd.com>
Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220509074712.163899-1-christian.koenig@amd.com

drm/bridge: simplify the return expression of ps8640_bridge_host_attach

Simplify the return expression.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429055145.3852271-1-chi.minghao@zte.com.cn

dma-buf: add the name field to the table header

'cat /sys/kernel/debug/dma_buf/bufinfo' will print the Dma-buf
Objects' information when the CONFIG_DEBUG_FS=y.
However, the printed table header information does not contain
the name field. So we need to add the name field to the table
header and use the '<none>' to replace the empty buf_obj->name.

Signed-off-by: Yuanzheng Song <songyuanzheng@huawei.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220428063924.3570409-1-songyuanzheng@huawei.com

drm/ttm: move default BO destructor into VMWGFX v2

It's the only driver using this.

v2: use BUG_ON() in vmw_bo_create() as suggested by Zack

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220329110243.6335-5-christian.koenig@amd.com

drm/rockchip: Fix spelling mistake "aligened" -> "aligned"

There is a spelling mistake in a drm_err message. Fix it.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505111044.374174-1-colin.i.king@gmail.com

drm/rockchip: remove unneeded semicolon from vop2 driver

Eliminate the following coccicheck warning:
./drivers/gpu/drm/rockchip/rockchip_drm_vop2.c:1476:2-3: Unneeded
semicolon

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505232659.4405-1-yang.lee@linux.alibaba.com

drm/rockchip: Fix Kconfig dependencies for display-port encoders

The DP-helper module has been replaced by the display-helper module.
So the driver have to select it.

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 1e0f66420b13("drm/display: Introduce a DRM display-helper module")
Signed-off-by: Ren Zhijie <renzhijie2@huawei.com>
Reviewed-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220507100910.93705-1-renzhijie2@huawei.com

dt-bindings: display: bridge: ldb: Fill in reg property

Add missing reg and reg-names properties for both 'LDB_CTRL'
and 'LVDS_CTRL' registers.

Fixes: 463db5c2ed4ae ("drm: bridge: ldb: Implement simple Freescale i.MX8MP LDB bridge")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Robby Cai <robby.cai@nxp.com>
Cc: Robert Foss <robert.foss@linaro.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: devicetree@vger.kernel.org
To: dri-devel@lists.freedesktop.org
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504012601.423644-1-marex@denx.de

drm/bridge: nwl-dsi: Set PHY mode in nwl_dsi_mode_set()

The Northwest Logic MIPI DSI host controller embedded in i.MX8qxp
works with a Mixel MIPI DPHY + LVDS PHY combo to support either
a MIPI DSI display or a LVDS display.  So, this patch calls
phy_set_mode() from nwl_dsi_mode_set() to set PHY mode to MIPI DPHY
explicitly.

Cc: Guido Günther <agx@sigxcpu.org>
Cc: Robert Chiras <robert.chiras@nxp.com>
Cc: Martin Kepplinger <martin.kepplinger@puri.sm>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
Cc: Jonas Karlman <jonas@kwiboo.se>
Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: NXP Linux Team <linux-imx@nxp.com>
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Acked-by: Robert Foss <robert.foss@linaro.org>
Reviewed-by: Guido Günther <agx@sigxcpu.org>
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220419010852.452169-2-victor.liu@nxp.com

drm/nouveau/gr/gf100-: Clean up some inconsistent indenting

Eliminate the follow smatch warning:

drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c:1925
gf100_gr_oneinit_tiles() warn: inconsistent indenting.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505081345.89762-1-jiapeng.chong@linux.alibaba.com

drm/edid: sunset drm_find_cea_extension()

Convert drm_find_cea_extension() to a predicate function to check if the
EDID has a CTA extension or a DisplayID CTA data block. This is mainly
to avoid adding new users that only find the first CTA extension.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/a5bf228942e6bd0fc70d5cf7a14c249a14a7afcd.1651569697.git.jani.nikula@intel.com

drm/edid: skip CTA extension scan in drm_edid_to_eld() just for CTA rev

The DisplayID CTA data block version does not necessarily match the CTA
revision. Simplify by postponing drm_edid_to_eld() slightly, and reusing
the CTA revision extracted by drm_parse_cea_ext().

By not bailing out early in drm_edid_to_eld() we may end up filling
meaningless data to the ELD. However, the main decision for audio is not
the ELD, but rather drm_detect_monitor_audio() called by drivers.

(Arguably a future cleanup could do that in drm_add_edid_modes() and
cache the result in the connector.)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/163ebbfd765066acbfc267256fb3b67fc711a78d.1651569697.git.jani.nikula@intel.com

drm/edid: detect color formats and CTA revision in all CTA extensions

Convert drm_find_cea_extension() to EDID block iterator in color format
and CTA revision detection. Detect them in all CTA extensions.

Also parse CTA Data Blocks in DisplayID even if there's no CTA EDID
extension.

v2:
- Don't assume DRM_COLOR_FORMAT_RGB444 support if there's only DisplayID
  CTA Data Blocks (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505105242.1198521-1-jani.nikula@intel.com

drm/edid: detect basic audio in all CEA extensions

Convert drm_find_cea_extension() to EDID block iterator in basic audio
detection. Detect basic audio in all CEA extensions.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/59481d0397de4b91b3e6ea665882334e92538f40.1651569697.git.jani.nikula@intel.com

drm/edid: restore some type safety to cea_db_*() functions

During the transition, we accepted a void pointer for a poor C
programmer's version of polymorphism. Switch the functions to use struct
cea_db * to regain some more type safety.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/c6ad496cdadea1bb598711a65ef536f4a43b74a7.1651569697.git.jani.nikula@intel.com

drm/edid: sunset the old unused cea data block iterators

All CTA data block iteration has now been converted to the new cea db
iterators.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/a5b35ed0e711bd874f76ae87ee0b70150d0a77c4.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_edid_to_eld() to use cea db iter

Iterate through all CTA data blocks across all CTA extensions and
DisplayID data blocks. This may gather more data than before, and if
there's duplicated data, some is overwritten by whichever comes last.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/8bdc67aa731857111eddd08a9c192d41d40b5f5a.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_parse_cea_ext() to use cea db iter

Iterate through all CTA data blocks across all CTA Extensions and
DisplayID data blocks.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/7400792525c13f58652f288b891d6057637ed4c8.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_detect_monitor_audio() to use cea db iter

Iterate through all CEA data blocks.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/a7f0c380da9526f8dd6f758d7a748bca7b4da6ce.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_detect_hdmi_monitor() to use cea db iter

Iterate through all CTA data blocks, not just the first CTA extension.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/b867e7b628189d2f8fa7eac5b9aa701892724711.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_edid_to_sad() to use cea db iter

Use the cea db iterator for short audio descriptors. We'll still stop at
the first audio data block, but not at the first CTA Extension if that
doesn't have the info.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/237e4b1de1567903d37ce1d1bb830020b8fd6690.1651569697.git.jani.nikula@intel.com

drm/edid: convert drm_edid_to_speaker_allocation() to use cea db iter

Use the cea db iterator for speaker allocation. We'll still stop at the
first speaker data block, but not at the first CTA extension if that
doesn't have the info.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/0b9e7f136854055a14b826097160fe0b43b9f3d1.1651569697.git.jani.nikula@intel.com

drm/edid: convert add_cea_modes() to use cea db iter

Iterate through all CTA EDID extension blocks and DisplayID CTA data
blocks to add CEA modes.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/dfa2d79db61d5eea543bef7aca7d80da3e71f99d.1651569697.git.jani.nikula@intel.com

drm/edid: clean up cea_db_is_*() functions

Abstract helpers for matching vendor data blocks and extended tags, and
use them to simplify all the cea_db_is_*() functions.

Take void pointer as parameter to allow transitional use for both u8 *
and struct cea_db *.

v2: Remove superfluous parens (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/b6bb813afc35c763e744c6cdb4a787da3adf910b.1651569697.git.jani.nikula@intel.com

drm/edid: add iterator for CTA data blocks

Add an iterator for CTA Data Blocks across EDID CTA Extensions and
DisplayID CTA Data Blocks.

v2: Update references, note why we can trust displayid ranges (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/37fdd2d9eabc73aaa9f95c56246dc47aea0e8e4e.1651569697.git.jani.nikula@intel.com

drm/edid: add iterator for EDID base and extension blocks

Add an iterator abstraction for going through all the EDID blocks.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ce02778b353b906c928268de9d7569d493a9be5d.1651569697.git.jani.nikula@intel.com

drm/edid: clean up CTA data block tag definitions

Add prefixed names, group, sort, add references.

v2:
- Updated references to CTA-861-H
- s/CEA/CTA/ in data block macros

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/c66cc3adeb375f5e60d1a8f91b41580d8ab67442.1651569697.git.jani.nikula@intel.com

drm/edid: rename HDMI Forum VSDB to SCDS

The HDMI spec talks about SCDS, Sink Capability Data Structure, exposed
via HF-VSDB or HF-SCDB. Rename VSDB to SCDS.

Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/f8230e1893400e9a9c5829041a8ab36349182a54.1651569697.git.jani.nikula@intel.com

drm/edid: check for HF-SCDB block

Find HF-SCDB information in CEA extensions block. And retrieve
Max_TMDS_Character_Rate that support by sink device.

v2: HF-SCDB and HF-VSDBS carry the same SCDS data. Reuse
    drm_parse_hdmi_forum_vsdb() to parse this packet.

Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Cc: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Cc: intel-gfx <intel-gfx@lists.freedesktop.org>
Signed-off-by: Lee Shawn C <shawn.c.lee@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/2bf7110cdb56d5dda311480787faf472c9802de4.1651569697.git.jani.nikula@intel.com

drm/edid: reset display info in drm_add_edid_modes() for NULL edid

If a NULL edid gets passed to drm_add_edid_modes(), we should probably
also reset the display info.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/2ac1c55f94a08d5e72c0b518d956a11002ec85c1.1651569697.git.jani.nikula@intel.com

drm/qxl: add drm_gem_plane_helper_prepare_fb

We could need to wait for the pin to complete here.

Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: virtualization@lists.linux-foundation.org
Cc: spice-devel@lists.freedesktop.org
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429134230.24334-4-christian.koenig@amd.com

drm/todo: Add entry for using kunit in the subsystem

The Kernel Unit Testing (KUnit) framework provides a common framework for
unit tests within the Linux kernel. Having a test suite would allow to
identify regressions earlier.

Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504080212.713275-1-javierm@redhat.com

drm: simplify the return expression of ast_drm_resume()

Simplify the return expression.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220505022208.57157-1-chi.minghao@zte.com.cn

drm/mgag200: Protect concurrent access to I/O registers with lock

Add a mutex lock to protect concurrent access to I/O registers
against each other. This happens between invocation of commit-
tail functions and get-mode operations. Both with use the CRTC
index registers MGA1064_GEN_IO_DATA and MGA1064_GEN_IO_CTL.
Concurrent access can lead to failed mode-setting operations.

v2:
* fix typo in commit description (Jocelyn)
* add comment to explain rmmio_lock

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220502142514.2174-4-tzimmermann@suse.de

drm/ast: Protect concurrent access to I/O registers with lock

Add a mutex lock to protect concurrent access to I/O registers
against each other. This happens between invocation of commit-
tail functions and get-mode operations. Both with use the CRTC
index register AST_IO_CRTC_PORT. Concurrent access can lead to
failed mode-setting operations.

v2:
* fix typo in commit description (Jocelyn)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reported-by: KuoHsiang Chou <kuohsiang_chou@aspeedtech.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220502142514.2174-3-tzimmermann@suse.de

drm: Add DRM-managed mutex_init()

Add drmm_mutex_init(), a helper that provides managed mutex cleanup. The
mutex will be destroyed with the final reference of the DRM device.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220502142514.2174-2-tzimmermann@suse.de

drm/format-helper: Share implementation among conversion helpers

Provide format-independent conversion helpers for system and I/O
memory. Implement most existing helpers on top of it. The source and
destination formats of each conversion is handled by a per-line
helper that is given to the generic implementation.

v2:
* remove a blank line

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220427141409.22842-5-tzimmermann@suse.de

drm/format-helper: Unify the parameters of all per-line conversion helpers

Give each per-line conversion helper pointers of type void and the
number of pixels in the line. Remove the unused swab parameters.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220427141409.22842-4-tzimmermann@suse.de

drm/format-helper: Remove optional byte-swap from line convertion

Implement per-pixel byte swapping in a separate conversion helper
for the single function that requires it. Select the correct helper
for each conversion.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220427141409.22842-3-tzimmermann@suse.de

drm/format-helper: Implement drm_fb_swab() with per-line helpers

Replace the inner loop of drm_fb_swab() with helper functions that
swap the bytes in each pixel. This will allow to share the outer
loop with other conversion helpers.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220427141409.22842-2-tzimmermann@suse.de

drm/bridge: tc358767: Fix DP bridge mode detection from DT endpoints

Per toshiba,tc358767.yaml DT binding document, port@2 the output (e)DP
port is optional. In case this port is not described in DT, the bridge
driver operates in DPI-to-DP mode. Make sure the driver treats this as
a valid mode of operation instead of reporting invalid mode.

Fixes: 71f7d9c03118 ("drm/bridge: tc358767: Detect bridge mode from connected endpoints in DT")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Jonas Karlman <jonas@kwiboo.se>
Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Robert Foss <robert.foss@linaro.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429204625.241591-1-marex@denx.de

drm: mxsfb: Implement LCDIF scanout CRC32 support

The LCDIF controller as present in i.MX28/i.MX6SX/i.MX8M Mini/Nano has
CRC_STAT register, which contains CRC32 of the frame as it was clocked
out of the DPI interface of the LCDIF. This is most likely meant as a
functional safety feature.

Unfortunately, there is zero documentation on how the CRC32 is calculated,
there is no documentation of the polynomial, the init value, nor on which
data is the checksum applied.

By applying brute-force on 8 pixel / 2 line frame, which is the minimum
size LCDIF would work with, it turns out the polynomial is CRC32_POLY_LE
0xedb88320 , init value is 0xffffffff , the input data are bitrev32()
of the entire frame and the resulting CRC has to be also bitrev32()ed.

Doing this calculation in kernel for each frame is unrealistic due to the
CPU demand, so attach the CRC collected from hardware to a frame instead.
The DRM subsystem already has an interface for this purpose and the CRC
can be accessed e.g. via debugfs:
"
$ echo auto > /sys/kernel/debug/dri/1/crtc-0/crc/control
$ cat /sys/kernel/debug/dri/1/crtc-0/crc/data
0x0000408c 0xa4e5cdd8
0x0000408d 0x72f537b4
"
The per-frame CRC can be used by userspace e.g. during automated testing,
to verify that whatever buffer was sent to be scanned out was actually
scanned out of the LCDIF correctly.

Acked-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Alexander Stein <alexander.stein@ew.tq-group.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Robby Cai <robby.cai@nxp.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stefan Agner <stefan@agner.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429212313.305556-1-marex@denx.de

drm/nouveau: remove trace_dma_fence_emit

Hardware drivers which don't have much difference between emit and init
shouldn't use this trace point.

Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Karol Herbst <kherbst@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: nouveau@lists.freedesktop.org
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220503085935.11023-1-christian.koenig@amd.com

drm/nouveau/subdev/bus: Ratelimit logging for fault errors

There's plenty of ways to fudge the GPU when developing on nouveau by
mistake, some of which can result in nouveau seriously spamming dmesg with
fault errors. This can be somewhat annoying, as it can quickly overrun the
message buffer (or your terminal emulator's buffer) and get rid of actually
useful feedback from the driver. While working on my new atomic only MST
branch, I ran into this issue a couple of times.

So, let's fix this by adding nvkm_error_ratelimited(), and using it to
ratelimit errors from faults. This should be fine for developers, since
it's nearly always only the first few faults that we care about seeing.
Plus, you can turn off rate limiting in the kernel if you really need to.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Cc: stable@vger.kernel.org
Link: https://patchwork.freedesktop.org/patch/msgid/20220429195350.85620-1-lyude@redhat.com

drm/nouveau/devinit: fix returnvar.cocci warnings

Fix the following coccicheck warnings:

drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gf100.c:71:5-12:
Unneeded variable: "disable". Return "0ULL" on line 90.
drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm107.c:35:5-12:
Unneeded variable: "disable". Return "0ULL" on line 44.
drivers/gpu/drm/nouveau/nvkm/subdev/devinit/g98.c:35:5-12:
Unneeded variable: "disable". Return "0ULL" on line 50.

Signed-off-by: Guo Zhengkui <guozhengkui@vivo.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504161003.9245-1-guozhengkui@vivo.com

drm: exynos: dsi: Use child panel or bridge find helpers

commit <711c7adc4687> ("drm: exynos: dsi: Use drm panel_bridge API")
added devm_drm_of_get_bridge for looking up if child node has panel
or bridge.

However commit <b089c0a9b14c> ("Revert "drm: of: Lookup if child node
has panel or bridge") has reverted panel or bridge child node lookup
from devm_drm_of_get_bridge which eventually failed to find the DSI
devices in exynos drm dsi driver.

So, use the conventional child panel bridge lookup helpers like it
does before.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20220428094808.782938-1-jagan@amarulasolutions.com

Revert "drm: bridge: mcde_dsi: Switch to devm_drm_of_get_bridge"

commit <3d7039e1e649> ("drm: bridge: mcde_dsi: Switch to devm_drm_of_get_bridge")
switched to devm_drm_of_get_bridge for looking up if child node has panel
or bridge.

However commit <b089c0a9b14c> ("Revert "drm: of: Lookup if child node
has panel or bridge") has reverted panel or bridge child node lookup
from devm_drm_of_get_bridge as it breaks the non-trivial cases the
first child node might not be a panel or bridge.

So, revert this commit to switch the previous behavior of looking up
child panel or bridge.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429085947.1699963-2-jagan@amarulasolutions.com

Revert "drm: bridge: mcde_dsi: Drop explicit bridge remove"

commit <3730bc6147b0> ("drm: bridge: mcde_dsi: Drop explicit bridge
remove") has removed downstream bridge as it's prior commit <3d7039e1e649>
("drm: bridge: mcde_dsi: Switch to devm_drm_of_get_bridge") added
devm_drm_of_get_bridge for looking up if child node has panel or bridge.

However commit <b089c0a9b14c> ("Revert "drm: of: Lookup if child node
has panel or bridge") has reverted panel or bridge child node lookup
from devm_drm_of_get_bridge as it breaks the non-trivial cases the
first child node might not be a panel or bridge.

So, revert this commit to switch the previous behavior of looking up
child panel or bridge.

Link: https://lore.kernel.org/all/20220429085947.1699963-1-jagan@amarulasolutions.com/
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>

drm/etnaviv: avoid cleaning up sched_job when submit succeeded

While the labels may mislead the casual reader, the tail of the function
etnaviv_ioctl_gem_submit is always executed, as a lot of the structures
set up in this function need to be cleaned up regardless of whether the
submit succeeded or failed.

An exception is the newly added drm_sched_job_cleanup, which must only
be called when the submit failed before handing the job to the
scheduler.

Fixes: b827c84f5e84 ("drm/etnaviv: Use scheduler dependency handling")
Reported-by: Michael Walle <michael@walle.cc>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220504090229.2506560-1-l.stach@pengutronix.de

drm: fix typo in comment

Fix typo in a doxygen comment (replaced green by blue) in drm_color_mgmt.c

Signed-off-by: Raphael Gallais-Pou <raphael.gallais-pou@foss.st.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220426131912.3850354-1-raphael.gallais-pou@foss.st.com

drm/rockchip: Add VOP2 driver

The VOP2 unit is found on Rockchip SoCs beginning with rk3566/rk3568.
It replaces the VOP unit found in the older Rockchip SoCs.

This driver has been derived from the downstream Rockchip Kernel and
heavily modified:

- All nonstandard DRM properties have been removed
- dropped struct vop2_plane_state and pass around less data between
  functions
- Dropped all DRM_FORMAT_* not known on upstream
- rework register access to get rid of excessively used macros
- Drop all waiting for framesyncs

The driver is tested with HDMI and MIPI-DSI display on a RK3568-EVB
board. Overlay support is tested with the modetest utility. AFBC support
on the cluster windows is tested with weston-simple-dmabuf-egl on
weston using the (yet to be upstreamed) panfrost driver support.

Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Co-Developed-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
[dt-binding-header:]
Acked-by: Rob Herring <robh@kernel.org>
[moved dt-binding header from dt-nodes patch to here
 and made checkpatch --strict happier]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-23-s.hauer@pengutronix.de

drm/rockchip: Make VOP driver optional

With upcoming VOP2 support VOP won't be the only choice anymore, so make
the VOP driver optional.

This also adds a dependency from ROCKCHIP_ANALOGIX_DP to ROCKCHIP_VOP,
because that driver currently only links and works with the VOP driver.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-22-s.hauer@pengutronix.de

dt-bindings: display: rockchip: dw-hdmi: fix ports description

Current port description doesn't cover all possible cases. It currently
expects one single port with two endpoints.

When the HDMI connector is described in the device tree there can be two
ports, first one going to the VOP and the second one going to the connector.

Also on SoCs which only have a single VOP there will be only one
endpoint instead of two.

This patch addresses both issues. With this there can either be a single
port ("port") , or two of them ("port@0", "port@1") when the connector
is also in the device tree. Also the first or only port can either have
one endpoint ("endpoint") for single VOP SoCs or two ("endpoint@0",
"endpoint@1") for dual VOP SoCs.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-25-s.hauer@pengutronix.de

dt-bindings: display: rockchip: Add binding for VOP2

The VOP2 is found on newer Rockchip SoCs like the rk3568 or the rk3566.
The binding differs slightly from the existing VOP binding, so add a new
binding file for it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-24-s.hauer@pengutronix.de

dma-buf/sync_file: cleanup fence merging a bit

krealloc_array() ignores attempts to reduce the array size, so the attempt
to save memory is completely pointless here.

Also move testing for the no fence case into sync_file_set_fence(), this
way we don't even touch the fence array when we don't have any fences.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220426124637.329764-1-christian.koenig@amd.com

drm/simpledrm: Use fbdev defaults for shadow buffering

Don't select shadow buffering for the fbdev console explicitly. The
fbdev emulation's heuristic will enable it for any framebuffer with
.dirty callback.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220303205839.28484-2-tzimmermann@suse.de

fbdev: Use pageref offset for deferred-I/O writeback

Use pageref->offset instead of page->index for deferred-I/O writeback
where appropriate. Distinguishes between file-mapping offset and video-
memory offset. While at it, also remove unnecessary references to
struct page.

Fbdev's deferred-I/O code uses the two related page->index and
pageref->offset. The former is the page offset in the mapped file,
the latter is the byte offset in the video memory (or fbdev screen
buffer). It's the same value for fbdev drivers, but for DRM the values
can be different. Because GEM buffer objects are mapped at an offset
in the DRM device file, page->index has this offset added to it as well.
We currently don't hit this case in DRM, because all affected mappings
of GEM memory are performed with an internal, intermediate shadow buffer.

The value of page->index is required by page_mkclean(), which we
call to reset the mappings during the writeback phase of the deferred
I/O. The value of pageref->offset is for conveniently getting an offset
into video memory in fb helpers.

v4:
* fix commit message (Javier)

Suggested-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429100834.18898-6-tzimmermann@suse.de

fbdev: Rename pagelist to pagereflist for deferred I/O

Rename various instances of pagelist to pagereflist. The list now
stores pageref structures, so the new name is more appropriate.

In their write-back helpers, several fbdev drivers refer to the
pageref list in struct fb_deferred_io instead of using the one
supplied as argument to the function. Convert them over to the
supplied one. It's the same instance, so no change of behavior
occurs.

v4:
* fix commit message (Javier)

Suggested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429100834.18898-5-tzimmermann@suse.de

fbdev: Refactor implementation of page_mkwrite

Refactor the page-write handler for deferred I/O. Drivers use the
function to let fbdev track written pages of mmap'ed framebuffer
memory.

v3:
* keep locking within track-pages function for readability (Sam)
v2:
* don't export the helper until we have an external caller

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429100834.18898-4-tzimmermann@suse.de

fbdev: Track deferred-I/O pages in pageref struct

Store the per-page state for fbdev's deferred I/O in struct
fb_deferred_io_pageref. Maintain a list of pagerefs for the pages
that have to be written back to video memory. Update all affected
drivers.

As with pages before, fbdev acquires a pageref when an mmaped page
of the framebuffer is being written to. It holds the pageref in a
list of all currently written pagerefs until it flushes the written
pages to video memory. Writeback occurs periodically. After writeback
fbdev releases all pagerefs and builds up a new dirty list until the
next writeback occurs.

Using pagerefs has a number of benefits.

For pages of the framebuffer, the deferred I/O code used struct
page.lru as an entry into the list of dirty pages. The lru field is
owned by the page cache, which makes deferred I/O incompatible with
some memory pages (e.g., most notably DRM's GEM SHMEM allocator).
struct fb_deferred_io_pageref now provides an entry into a list of
dirty framebuffer pages, freeing lru for use with the page cache.

Drivers also assumed that struct page.index is the page offset into
the framebuffer. This is not true for DRM buffers, which are located
at various offset within a mapped area. struct fb_deferred_io_pageref
explicitly stores an offset into the framebuffer. struct page.index
is now only the page offset into the mapped area.

These changes will allow DRM to use fbdev deferred I/O without an
intermediate shadow buffer.

v3:
* use pageref->offset for sorting
* fix grammar in comment
v2:
* minor fixes in commit message

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429100834.18898-3-tzimmermann@suse.de

fbdev: Put mmap for deferred I/O into drivers

The fbdev mmap function fb_mmap() unconditionally overrides the
driver's implementation if deferred I/O has been activated. This
makes it hard to implement mmap with anything but a vmalloc()'ed
software buffer. That is specifically a problem for DRM, where
video memory is maintained by a memory manager.

Leave the mmap handling to drivers and expect them to call the
helper for deferred I/O by thmeselves.

v4:
* unlock mm_lock in fb_mmap() error path (Dan)
v3:
* fix warning if fb_mmap is missing (kernel test robot)
v2:
* print a helpful error message if the defio setup is
  incorrect (Javier)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220429100834.18898-2-tzimmermann@suse.de

drm/ast: Create the driver for ASPEED proprietory Display-Port

V1:
1. The MCU FW controling ASPEED DP is loaded by BMC boot loader.
2. Driver starts after CR[3:1] == 111b that indicates Tx is ASTDP,
   and CRD1[5] has been asserted by BMVC boot loader.
3. EDID is prioritized by DP monitor.
4. DP's EDID has high priority to decide resolution supporting.

V2:
Modules description:
1. ASTDP (ASPEED DisplayPort) is controlled by dedicated
   AST-MCU (ASPEED propriatary MCU).
2. MCU is looping in charged of HPD, Read EDID, Link Training with
   DP sink.
3. ASTDP and AST-MUC reside in BMC (Baseboard Management controller)
   addressing-space.
4. ASPEED DRM driver requests MCU to get HPD and EDID by CR-scratched
   register.

Booting sequence:
1. Check if TX is ASTDP // ast_dp_launch()
2. Check if DP-MCU FW has loaded // ast_dp_launch()
3. Read EDID // ast_dp_read_edid()
4. Resolution switch // ast_dp_SetOutput()

V3:
1. Remove unneeded semicolon.
2. Apply to git://anongit.freedesktop.org/drm/drm, instead of
   git://anongit.freedesktop.org/drm/drm-misc
3. Resolve auto build test WARNINGs on V1 patch.

V4:
1. Sync code-base with kernel 5.17_rc6
2. Remove the define of DPControlPower, because DP chips need to be
   powered on to be used.
3. Remove the switches of PHY and Display from EDID procedure.
4. Revise increaing delay to fixed delay, because this version kernel
   doesn't detect minitor consistenntly.
5. Create clean-up code used for reset of power state on errors with
   -EIO manner.
6. Revise the DP detection by TX type and its DP-FW status during
   booting and resume.
7. Correct the CamelCase Style.
8. Use register reading while needing, and remove to hold full
   register.
9. Instead of 'u8', revise to 'bool' on swwitch of PHY and video.
10.Correct typo
11.Remove the duplicated copy of TX definition.
12.Use EDID_LENGTH as the constant of 128.

Signed-off-by: KuoHsiang Chou <kuohsiang_chou@aspeedtech.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220428075603.20904-1-kuohsiang_chou@aspeedtech.com

dt-bindings: display: rockchip: dw-hdmi: Make unwedge pinctrl optional

None of the upstream device tree files has a "unwedge" pinctrl
specified. Make it optional.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Rob Herring <robh@kernel.org>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-16-s.hauer@pengutronix.de

drm/rockchip: dw_hdmi: add regulator support

The RK3568 has HDMI_TX_AVDD0V9 and HDMI_TX_AVDD_1V8 supply inputs needed
for the HDMI port. add support for these to the driver for boards which
have them supplied by switchable regulators.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-10-s.hauer@pengutronix.de

dt-bindings: display: rockchip: dw-hdmi: Add regulator support

The RK3568 has HDMI_TX_AVDD0V9 and HDMI_TX_AVDD_1V8 supply inputs
needed for the HDMI port. Add the binding for these supplies.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Rob Herring <robh@kernel.org>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-11-s.hauer@pengutronix.de

drm/rockchip: dw_hdmi: add rk3568 support

Add a new dw_hdmi_plat_data struct and new compatible for rk3568.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-8-s.hauer@pengutronix.de

dt-bindings: display: rockchip: dw-hdmi: Add compatible for rk3568 HDMI

Define a new compatible for rk3568 HDMI.
This version of HDMI hardware block needs two new clocks hclk_vio and hclk
to provide phy reference clocks.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-9-s.hauer@pengutronix.de

Merge drm/drm-next into drm-misc-next

Christian needs a backmerge to avoid a merge conflict for amdgpu.

Signed-off-by: Maxime Ripard <maxime@cerno.tech>

dt-bindings: display: rockchip: dw-hdmi: use "ref" as clock name

"vpll" is a misnomer. A clock input to a device should be named after
the usage in the device, not after the clock that drives it. On the
rk3568 the same clock is driven by the HPLL.
This patch adds "ref" as a new alternative clock name for "vpll"

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Rob Herring <robh@kernel.org>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-6-s.hauer@pengutronix.de

drm/rockchip: dw_hdmi: rename vpll clock to reference clock

"vpll" is a misnomer. A clock input to a device should be named after
the usage in the device, not after the clock that drives it. On the
rk3568 the same clock is driven by the HPLL.
To fix that, this patch renames the vpll clock to ref clock. The clock
name "vpll" is left for compatibility to old device trees.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-5-s.hauer@pengutronix.de

drm/rockchip: Add crtc_endpoint_id to rockchip_encoder

The VOP2 has an interface mux which decides to which encoder(s) a CRTC
is routed to. The encoders and CRTCs are connected via of_graphs in the
device tree. When given an encoder the VOP2 driver needs to know to
which internal register setting this encoder matches. For this the VOP2
binding offers different endpoints, one for each possible encoder. The
endpoint ids of these endpoints are used as a key from an encoders
device tree description to the internal register setting.

This patch adds the key aka endpoint id to struct rockchip_encoder plus
a function to read the endpoint id starting from the encoders device
node.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-4-s.hauer@pengutronix.de

drm/rockchip: Embed drm_encoder into rockchip_decoder

The VOP2 driver needs rockchip specific information for a drm_encoder.

This patch creates a struct rockchip_encoder with a struct drm_encoder
embedded in it. This is used throughout the rockchip driver instead of
struct drm_encoder directly.

The information the VOP2 drivers needs is the of_graph endpoint node
of the encoder. To ease bisectability this is added here.

While at it convert the different encoder-to-driverdata macros to
static inline functions in order to gain type safety and readability.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Michael Riesch <michael.riesch@wolfvision.net>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220422072841.2206452-3-s.hauer@pengutronix.de

Backmerge tag 'v5.18-rc5' of git://git./linux/kernel/git/torvalds/linux into drm-next

Linux 5.18-rc5

There was a build fix for arm I wanted in drm-next, so backmerge rather then cherry-pick.

Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/rockchip: Support YUV formats with U/V swapped

The VOP in Rockchip SoCs that support YUV planes also support swapping
of the U and V elements. Supporting the swapped variants, especially
NV21, would be beneficial for multimedia applications, as the hardware
video decoders only output NV21, and supporting this pixel format in
the display pipeline would allow the decoded video frames to be output
directly.

Add support for this to support the various formats that have U/V
swapped.

Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220114074038.2633848-1-wenst@chromium.org

drm/rockchip: rk3066_hdmi: Replace drm_detect_hdmi_monitor() with is_hdmi

Once EDID is parsed, the monitor HDMI support information is available
through drm_display_info.is_hdmi.

This driver calls drm_detect_hdmi_monitor() to receive the same
information and stores its own cached value, which is less efficient.

Avoid calling drm_detect_hdmi_monitor() and use drm_display_info.is_hdmi
instead and also remove hdmi_data_info.sink_is_hdmi as it is no longer
necessary.

Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220421170725.903361-4-jose.exposito89@gmail.com

drm/rockchip: inno_hdmi: Replace drm_detect_hdmi_monitor() with is_hdmi

Once EDID is parsed, the monitor HDMI support information is available
through drm_display_info.is_hdmi.

This driver calls drm_detect_hdmi_monitor() to receive the same
information and stores its own cached value, which is less efficient.

Avoid calling drm_detect_hdmi_monitor() and use drm_display_info.is_hdmi
instead and also remove hdmi_data_info.sink_is_hdmi as it is no longer
necessary.

Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220421170725.903361-3-jose.exposito89@gmail.com

drm/rockchip: cdn-dp: change rk3399_cdn_dp from global to static

Smatch reports this issue
cdn-dp-core.c:51:20: warning: symbol 'rk3399_cdn_dp' was not declared. Should it be static?

rk3399_cdn_dp is only used in cdn-dp-core.c so change
its storge-class specifier to static.

Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220421144304.586396-1-trix@redhat.com

drm/rockchip: Refactor IOMMU initialisation

Defer the IOMMU domain setup until after successfully binding
components, so we can figure out IOMMU support directly from the VOP
devices themselves, rather than manually inferring it from the DT (which
also fails to account for whether the IOMMU driver is actually loaded).
Although this is somewhat of a logical cleanup, the main motivation is
to prepare for a change in the iommu_domain_alloc() interface.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/94eee7ab434fe11eb0787f691e9f1ab03a2e91be.1649168685.git.robin.murphy@arm.com