If msm_gem_address_space_create() fails, then return right away.
Otherwise it leads to a Oops when we dereference "aspace" on the next
line.
Fixes: eabba31a83 ("drm/msm: register a fault handler for display mmu faults")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/639357/
Link: https://lore.kernel.org/r/3221e88c-3351-42e6-aeb1-69f4f014b509@stanley.mountain
[DB: fixed commit id]
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
In order to simplify the driver even further and to remove the
boilerplate code, rewrite the audio interface to use the DRM HDMI Audio
framework.
Audio InfoFames are controlled centrally via the DRM HDMI framework.
Correct InfoFrame data is programmed at the atomic_pre_enable() time (if
it was set before, drm_atomic_helper_connector_hdmi_update_infoframes()
takes care of writing all InfoFrames, including the Audio one.) or
during msm_hdmi_bridge_audio_prepare() when the new stream is started.
All audio data frame management is deferred to
msm_hdmi_bridge_audio_prepare() and msm_hdmi_bridge_audio_shutdown().
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/639663/
Link: https://lore.kernel.org/r/20250226-bridge-hdmi-connector-v8-7-340af24b35cc@linaro.org
Extend the driver to send SPD and HDMI Vendor Specific InfoFrames.
While the HDMI block has special block to send HVS InfoFrame, use
GENERIC0 block instead. VENSPEC_INFO registers pack frame data in a way
that requires manual repacking in the driver, while GENERIC0 doesn't
have such format requirements. The msm-4.4 kernel uses GENERIC0 to send
HDR InfoFrame which we do not at this point anyway.
Acked-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/639661/
Link: https://lore.kernel.org/r/20250226-bridge-hdmi-connector-v8-6-340af24b35cc@linaro.org
The mode_set callback is deprecated, it doesn't get the
drm_bridge_state, just mode-related argumetns. Also Abhinav pointed out
that HDMI timings should be programmed before setting up HDMI PHY and
PLL. Rework the code to program HDMI timings at the start of
atomic_pre_enable().
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/639652/
Link: https://lore.kernel.org/r/20250226-bridge-hdmi-connector-v8-2-340af24b35cc@linaro.org
Add a helper that will handle the correct order of the encoder kickoffs
for concurrent writeback.
For concurrent writeback, the realtime encoder must always kickoff last
as it will call the trigger flush and start.
This avoids the following scenario where the writeback encoder
increments the pending kickoff count after the WB_DONE interrupt is
fired:
If the realtime encoder is kicked off first, the encoder kickoff will
flush/start the encoder and increment the pending kickoff count. The
WB_DONE interrupt then fires (before the writeback encoder is kicked
off). When the writeback encoder enters its kickoff, it will skip the
flush/start (due to CWB being enabled) and hit a frame done timeout
as the frame was kicked off (and the WB_DONE interrupt fired) without
the pending kickoff count being incremented.
In addition, the writeback timer should only start after the realtime
encoder is kicked off to ensure that we don't get timeouts when the
system has a heavy load (ex. when debug logs are enabled)
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637491/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-13-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
For concurrent writeback, the real time encoder is responsible for
trigger flush and trigger start. Return early for trigger start and
trigger flush for the concurrent writeback encoders.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637505/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-12-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Starting the frame done timer before the encoder is finished kicking off
can lead to unnecessary frame done timeouts when the device is
experiencing heavy load (ex. when debug logs are enabled).
Thus, create a separate API for starting the encoder frame done timer and
call it after the encoder kickoff is finished
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637502/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-11-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Cache the CWB block mask in the DPU virtual encoder and configure CWB
according to the CWB block mask within the writeback phys encoder
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637501/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-8-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Add support for RM to reserve dedicated CWB PINGPONGs and CWB muxes
For concurrent writeback, even-indexed CWB muxes must be assigned to
even-indexed LMs and odd-indexed CWB muxes for odd-indexed LMs. The same
even/odd rule applies for dedicated CWB PINGPONGs.
Track the CWB muxes in the global state and add a CWB-specific helper to
reserve the correct CWB muxes and dedicated PINGPONGs following the
even/odd rule.
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/637495/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-7-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Currently, our hardware only supports a single output using CDM block at
most. Because of this, we cannot support cases where both writeback and DP
output request CDM simultaneously
To avoid this happening when CWB is enabled, change
msm_display_topoloy.needs_cdm into a num_cdm counter to track how many
outputs are requesting CDM block. Return EINVAL if multiple outputs are
trying to reserve CDM.
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/637499/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-6-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Currently, the topology is calculated based on the assumption that the
user cannot request real-time and writeback simultaneously. For example,
the number of LMs and CTLs are currently based off the number of phys
encoders under the assumption there will be at least 1 LM/CTL per phys
encoder.
This will not hold true for concurrent writeback as both phys encoders
(1 real-time and 1 writeback) must be driven by 1 LM/CTL when concurrent
writeback is enabled.
To account for this, add a cwb_enabled flag and only adjust the number of
CTL/LMs needed by a given topology based on the number of phys encoders
only if CWB is not enabled.
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637486/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-4-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Up to now the driver has been using encoder to allocate hardware
resources. Switch it to use CRTC id in preparation for the next step.
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/637503/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-3-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
All resource allocation is centered around the LMs. Then other blocks
(except DSCs) are allocated basing on the LMs that was selected, and LM
powers up the CRTC rather than the encoder.
Moreover if at some point the driver supports encoder cloning,
allocating resources from the encoder will be incorrect, as all clones
will have different encoder IDs, while LMs are to be shared by these
encoders.
In addition, move mode_changed() to dpu_crtc as encoder no longer has
access to topology information
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
[quic_abhinavk@quicinc.com: Refactored resource allocation for CDM]
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
[quic_jesszhan@quicinc.com: Changed to grabbing exising global state]
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
[DB: rebased on top of msm-next]
[DB: fixed resource allcoation to ignore the active_changed flag]
Patchwork: https://patchwork.freedesktop.org/patch/637487/
Link: https://lore.kernel.org/r/20250214-concurrent-wb-v6-2-a44c293cf422@quicinc.com
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Any rules using engine matching are currently broken due RTP processing
happening too in early init, before the list of hardware engines has been
initialised.
Fix this by moving workaround processing to later in the driver probe
sequence, to just before the processed list is used for the first time.
Looking at the debugfs gt0/workarounds on ADL-P we notice 14011060649
should be present while we see, before:
GT Workarounds
14011059788
14015795083
And with the patch:
GT Workarounds
14011060649
14011059788
14015795083
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: stable@vger.kernel.org # v6.11+
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250227101304.46660-2-tvrtko.ursulin@igalia.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
(cherry picked from commit 25d434cef7)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Currently we just leave it uninitialised, which at first looks harmless,
however we also don't zero out the pfn array, and with pfn_flags_mask
the idea is to be able set individual flags for a given range of pfn or
completely ignore them, outside of default_flags. So here we end up with
pfn[i] & pfn_flags_mask, and if both are uninitialised we might get back
an unexpected flags value, like asking for read only with default_flags,
but getting back write on top, leading to potentially bogus behaviour.
To fix this ensure we zero the pfn_flags_mask, such that hmm only
considers the default_flags and not also the initial pfn[i] value.
v2 (Thomas):
- Prefer proper initializer.
Fixes: 81e058a3e7 ("drm/xe: Introduce helper to populate userptr")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@intel.com>
Cc: <stable@vger.kernel.org> # v6.10+
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Tejas Upadhyay <tejas.upadhyay@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250226174748.294285-2-matthew.auld@intel.com
(cherry picked from commit dd8c01e42f)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
We create the stream encoders and attach connectors for each pipe we
have. As the number of pipes has increased, we've failed to update the
topology manager maximum number of payloads to match that. Bump up the
max stream count to match number of pipes, enabling the fourth stream on
platforms that support four pipes.
Cc: stable@vger.kernel.org
Cc: Imre Deak <imre.deak@intel.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250226135626.1956012-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
(cherry picked from commit 15bccbfb78)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This is already handled below in the code by fixup_initial_plane_config.
Fixes: a815362752 ("drm/i915: Try to relocate the BIOS fb to the start of ggtt")
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241210083111.230484-3-dev@lankhorst.se
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
(cherry picked from commit 2218704997)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
While missing lvds pinctrl is unexpected and is reported, we nevertheless
don't fail setting up the device and instead continue without explicit
pinctrl handling. So lower the log-level from error to warning to reflect
that.
Suggested-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Heiko Stuebner <heiko.stuebner@cherry.de>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20250304124418.111061-4-heiko@sntech.de
Commit 52d11c863a ("drm/rockchip: lvds: do not print scary message when
probing defer") already started hiding scary messages that are not relevant
if the requested supply just returned EPROBE_DEFER, but there are more
possible sources - like the phy.
So modernize the whole logging in the probe path by replacing the
remaining deprecated DRM_DEV_ERROR with appropriate dev_err(_probe)
and drm_err calls.
The distinction here is that all messages talking about mishaps of the
lvds element use dev_err(_probe) while messages caused by interaction
with the main Rockchip drm-device use drm_err.
Reviewed-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko.stuebner@cherry.de>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20250304124418.111061-3-heiko@sntech.de
The LVDS block needs a separate pclk only on some socs, so currently
requests and prepares it in the soc-specific probe function, but common
code is required to unprepare it in the error path or on driver remove.
While this works because clk_unprepare just does nothing if clk is NULL,
this mismatch of who is responsible still is not very nice.
The clock-framework already has a helper for clk-get-and-prepare even
with devres support in devm_clk_get_prepared().
This will get and prepare the clock and also unprepare it on driver
removal, saving the driver from having to handle it "manually".
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Reviewed-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko.stuebner@cherry.de>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20250304124418.111061-2-heiko@sntech.de
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(sched, ...) to dev_err(sched->dev, ...) and
similar. This matches current usage, as struct drm_device is not
available, but drops "[drm]" or "[drm] *ERROR*" prefix from logging.
Unfortunately, there's no dev_WARN_ON(), so the conversion is not
exactly the same.
Reviewed-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/fe441dd1469d2b03e6b2ff247078bdde2011c6e3.1737644530.git.jani.nikula@intel.com
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(hdmi, ...) to dev_err(hdmi->dev, ...). This matches
current usage, but drops "[drm] *ERROR*" prefix from logging.
Reviewed-by: Simona Vetter <simona.vetter@ffwll.ch>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/f42da4c9943a2f2a9de4272b7849e72236d4c3f9.1737644530.git.jani.nikula@intel.com
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(host, ...) to dev_err(host->dev, ...). This matches
current usage, as struct drm_device is not available, but drops "[drm]
*ERROR*" from logs.
Reviewed-by: Simona Vetter <simona.vetter@ffwll.ch>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/842f97ade87d6f0c4b1de12e8ed5610a1b07fd8c.1737644530.git.jani.nikula@intel.com
When building for a 32-bit platform, there are some warnings (or errors
with CONFIG_WERROR=y) due to an incorrect specifier for 'size_t'
variables, which is typedef'd as 'unsigned int' for these architectures:
drivers/gpu/drm/tiny/appletbdrm.c:171:17: error: format specifies type 'unsigned long' but the argument has type 'size_t' (aka 'unsigned int') [-Werror,-Wformat]
170 | drm_err(drm, "Actual size (%d) doesn't match expected size (%lu)\n",
| ~~~
| %zu
171 | actual_size, size);
| ^~~~
...
drivers/gpu/drm/tiny/appletbdrm.c:212:17: error: format specifies type 'unsigned long' but the argument has type 'size_t' (aka 'unsigned int') [-Werror,-Wformat]
211 | drm_err(drm, "Actual size (%d) doesn't match expected size (%lu)\n",
| ~~~
| %zu
212 | actual_size, size);
| ^~~~
Use '%zu' as suggested, clearing up the warnings.
Fixes: 0670c2f56e ("drm/tiny: add driver for Apple Touch Bars in x86 Macs")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Aditya Garg <gargaditya08@live.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250304-appletbdrm-fix-size_t-specifier-v1-1-94fe1d2c91f8@kernel.org
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
We customarily define the bits of a register in big endian
order. Reorder the gen9+ timestamp freq register bits to match.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250211231941.22769-13-ville.syrjala@linux.intel.com