Miaoqian Lin [Wed, 11 May 2022 05:40:51 +0000 (09:40 +0400)]
drm/meson: Fix refcount leak in meson_encoder_hdmi_init
of_find_device_by_node() takes reference, we should use put_device()
to release it when not need anymore.
Add missing put_device() in error path to avoid refcount
leak.
Fixes: bcf5446630b8 ("drm/meson: encoder_hdmi: switch to bridge DRM_BRIDGE_ATTACH_NO_CONNECTOR") Signed-off-by: Miaoqian Lin <linmq006@gmail.com> Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Reviewed-by: Neil Armstrong <narmstrong@baylibre.com> Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220511054052.51981-1-linmq006@gmail.com
drm/mgag200: Split up connector's mode_valid helper
Split up the connector's mode_valid helper into a simple-pipe and a
mode-config helper. The simple-pipe helper tests for display-size
limits while the mode-config helper tests for memory-bandwidth limits.
Also add the mgag200_ prefix to mga_vga_calculate_mode_bandwidth() and
comment on the function's purpose.
The memory-bandwidth tests assume that the display uses 4 bytes per
pixel. The first models of G200SE-A only had 1.75 MiB of VRAM, which
limits these devices to 640x480-32.
v2:
* note the memory constraints on early G200SE-A
drm/mgag200: Test memory requirements in drm_mode_config_funcs.mode_valid
Test for a mode's memory requirements in the device-wide mode_valid
helper. For simplicify, always assume a 32-bit color format. While
some rejected modes would work with less colors, implementing this
is probably not worth the effort.
Also remove the memory-related test from the connector's mode_valid
helper. The test uses the bpp value that users can specify on the
kernel's command line. This value is unrelated and the test would
belong into atomic_check.
struct mga_connector has outlived its purpose. Inline the rsp init
helper into the mode-config code and remove the data structure. No
functional changes.
Store the I2C state within struct mga_device and switch I2C to
managed release. Simplifies the related code and lets us remove
mga_connector_destroy().
drm/mgag200: Implement connector's get_modes with helper
Provide drm_connector_helper_get_modes_from_ddc() to implement the
connector's get_modes callback. The new helper updates the connector
from DDC-provided EDID data.
Initialization of the I2C adapter was allowed to fail. The mgag200
driver would have continued without DDC support. Had this happened in
practice, it would have led to segmentation faults in the connector
code. Resolve this problem by failing driver initialization on I2C-
related errors.
v2:
* initialize 'ret' before drm_err() (kernel test robot)
DDC operation conflicts with concurrent mode setting. Acquire the
driver's I/O lock in get_modes to prevent this. This change should
have been part of commit 946965594140 ("drm/mgag200: Protect
concurrent access to I/O registers with lock"), but apparently got
lost somewhere.
v3:
* fix commit message to say 'drm/mgag200' (Jocelyn)
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Fixes: 946965594140 ("drm/mgag200: Protect concurrent access to I/O registers with lock") Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com> Tested-by: Jocelyn Falempe <jfalempe@redhat.com> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Dave Airlie <airlied@redhat.com> Cc: dri-devel@lists.freedesktop.org Link: https://patchwork.freedesktop.org/patch/msgid/20220516134343.6085-2-tzimmermann@suse.de
drm/scheduler: Don't kill jobs in interrupt context
Interrupt context can't sleep. Drivers like Panfrost and MSM are taking
mutex when job is released, and thus, that code can sleep. This results
into "BUG: scheduling while atomic" if locks are contented while job is
freed. There is no good reason for releasing scheduler's jobs in IRQ
context, hence use normal context to fix the trouble.
drm: bridge: DRM_FSL_LDB should depend on ARCH_MXC
The Freescale i.MX8MP LDB bridge is only present on Freescale i.MX8MP
SoCs. Hence add a dependency on ARCH_MXC, to prevent asking the user
about this driver when configuring a kernel without i.MX SoC support.
Marek Vasut [Sat, 30 Apr 2022 02:50:20 +0000 (04:50 +0200)]
drm: bridge: icn6211: Register macro clean up
Drop two unused register macros, ICN6211_MAX_REGISTER and MIPI_ATE_STATUS_1,
neither of which is used and where the later should be specified using macro
MIPI_ATE_STATUS(1) instead. Drop the _(n) underscore and keep only the (n)
part of register macros. No functional change.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Jagan Teki <jagan@amarulasolutions.com> Cc: Mark Brown <broonie@kernel.org> Cc: Maxime Ripard <maxime@cerno.tech> Cc: Robert Foss <robert.foss@linaro.org> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: Thomas Zimmermann <tzimmermann@suse.de>
To: dri-devel@lists.freedesktop.org Signed-off-by: Robert Foss <robert.foss@linaro.org> Link: https://patchwork.freedesktop.org/patch/msgid/20220430025020.640277-1-marex@denx.de
Jocelyn Falempe [Fri, 13 May 2022 08:49:00 +0000 (10:49 +0200)]
drm/mgag200: Enable atomic gamma lut update
Add support for atomic update of gamma lut.
With this patch the "Night light" feature of gnome3
is working properly on mgag200.
v2:
- Add a default linear gamma function
- renamed functions with mgag200 prefix
- use format's 4cc code instead of bit depth
- use better interpolation for 16bits gamma
- remove legacy function mga_crtc_load_lut()
- can't remove the call to drm_mode_crtc_set_gamma_size()
because it doesn't work with userspace.
- other small refactors
v3:
- change mgag200_crtc_set_gamma*() argument
to struct drm_format_info *format
- fix printk format to %p4cc for 4cc and %zu for size_t
- rebased to drm-misc-next.
Don't add a mode for the kernel's command-line parameters from
within the DRM client code. Doing so can result in an unusable
display. If there's no compatible command-line mode, the client
will use one of the connector's preferred modes.
All mode creation and validation has to be performed by the
connector. When clients run, the connector's fill_modes callback
has already processed the kernel parameters and validated each
mode before adding it. The connector's mode list does not contain
invalid modes.
When picking a mode, first look for modes that have been specified
by the user on the kernel's command line. Only if that fails, use
the existing heuristic of picking a nearby mode from it's various
parameters.
drm: Always warn if user-defined modes are not supported
Print a warning if a user-specifed display mode is not supported by
the display pipeline. Users specified the display mode on the kernel
command line with the use of the video= parameter. Setting an
unsupported mode will leave the console blank, so we should at least
let the user know why.
Jani Nikula [Mon, 9 May 2022 12:03:03 +0000 (15:03 +0300)]
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.
Jani Nikula [Mon, 9 May 2022 12:03:02 +0000 (15:03 +0300)]
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.
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.
Marek Vasut [Fri, 29 Apr 2022 20:56:44 +0000 (22:56 +0200)]
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
Marek Vasut [Fri, 29 Apr 2022 20:56:43 +0000 (22:56 +0200)]
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
Jocelyn Falempe [Wed, 4 May 2022 13:40:25 +0000 (15:40 +0200)]
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.
Jocelyn Falempe [Wed, 4 May 2022 13:40:24 +0000 (15:40 +0200)]
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.
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.
Christian König [Tue, 10 May 2022 10:52:05 +0000 (12:52 +0200)]
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.
Christian König [Mon, 9 May 2022 07:47:12 +0000 (09:47 +0200)]
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
Yuanzheng Song [Thu, 28 Apr 2022 06:39:24 +0000 (06:39 +0000)]
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.
Liu Ying [Tue, 19 Apr 2022 01:08:48 +0000 (09:08 +0800)]
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
Jani Nikula [Tue, 3 May 2022 09:24:05 +0000 (12:24 +0300)]
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.
Jani Nikula [Tue, 3 May 2022 09:24:04 +0000 (12:24 +0300)]
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.)
Jani Nikula [Tue, 3 May 2022 09:24:01 +0000 (12:24 +0300)]
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.
Jani Nikula [Tue, 3 May 2022 09:23:59 +0000 (12:23 +0300)]
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.
Jani Nikula [Tue, 3 May 2022 09:23:55 +0000 (12:23 +0300)]
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.
Jani Nikula [Tue, 3 May 2022 09:23:54 +0000 (12:23 +0300)]
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.
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.
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
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.
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.
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.
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.
Marek Vasut [Fri, 29 Apr 2022 20:46:25 +0000 (22:46 +0200)]
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: d6341964b973 ("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
projects / kernel.git / log