The drm_atomic_get_plane_state() function calls the deprecated
drm_atomic_get_existing_plane_state() helper to get find if a plane
state had already been allocated and was part of the given
drm_atomic_state.
At the point in time where drm_atomic_get_plane_state() can be called
(ie, during atomic_check), the existing state is the new state and
drm_atomic_get_existing_plane_state() can thus be replaced by
drm_atomic_get_new_plane_state().
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-5-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
The __drm_atomic_get_current_plane_state() function tries to get and
return the existing plane state, and if it doesn't exist returns the one
stored in the drm_plane->state field.
Using the current nomenclature, it tries to get the existing plane state
with an ad-hoc implementation of drm_atomic_get_existing_plane_state(),
and falls back to either the old or new plane state, depending on
whether it is called before or after drm_atomic_helper_swap_state().
The existing plane state itself is deprecated, because it also changes
when swapping states from the new state to the old state.
Fortunately for us, we can simplify things. Indeed,
__drm_atomic_get_current_plane_state() is only used in two macros:
intel_atomic_crtc_state_for_each_plane_state and
drm_atomic_crtc_state_for_each_plane_state().
The intel variant is only used through the intel_wm_compute() function
that is only ever called in intel_crtc_atomic_check().
The generic variant is more widely used, and can be found in the malidp,
msm, tegra and vc4 drivers. All of these call sites though are during
atomic_check(), so we end up in the same situation than Intel's.
Thus, we only ever use the existing state as the new state, and
plane->state is always going to be the old state. Any plane isn't
guaranteed to be part of the state though, so we can't rely on
drm_atomic_get_old_plane_state() and we still need to use plane->state.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-4-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
The drm_atomic_get_connector_state() function calls a hand-rolled
implementation of the deprecated
drm_atomic_get_existing_connector_state() helper to get find if a
connector state had already been allocated and was part of the given
drm_atomic_state.
At the point in time where drm_atomic_get_connector_state() can be
called (ie, during atomic_check), the existing state is the new state
and drm_atomic_get_existing_connector_state() can thus be replaced by
drm_atomic_get_new_connector_state().
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com> # on imx8mp
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-1-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Add panthor_soc_data to control custom ASN_HASH. Add compatible string
for "mediatek,mt8196-mali" and enable custom ASN_HASH for the soc.
Without custom ASN_HASH, FW fails to boot
panthor 48000000.gpu: [drm] *ERROR* Unhandled Page fault in AS0 at VA 0x0000000000000000
panthor 48000000.gpu: [drm] *ERROR* Failed to boot MCU (status=fatal)
panthor 48000000.gpu: probe with driver panthor failed with error -110
With custom ASN_HASH, panthor probes fine and userspace boots to ui just
fine as well
panthor 48000000.gpu: [drm] clock rate = 0
panthor 48000000.gpu: EM: created perf domain
panthor 48000000.gpu: [drm] Mali-G925-Immortalis id 0xd830 major 0x0 minor 0x1 status 0x5
panthor 48000000.gpu: [drm] Features: L2:0x8130306 Tiler:0x809 Mem:0x301 MMU:0x2830 AS:0xff
panthor 48000000.gpu: [drm] shader_present=0xee0077 l2_present=0x1 tiler_present=0x1
panthor 48000000.gpu: [drm] Firmware protected mode entry not be supported, ignoring
panthor 48000000.gpu: [drm] Firmware git sha: 27713280172c742d467a4b7d11180930094092ec
panthor 48000000.gpu: [drm] CSF FW using interface v3.13.0, Features 0x10 Instrumentation features 0x71
[drm] Initialized panthor 1.5.0 for 48000000.gpu on minor 1
Note that the clock and the regulator drivers are not upstreamed yet.
They might as well take a different form when upstreamed.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://lore.kernel.org/r/20250913002155.1163908-3-olvaffe@gmail.com
drm_bridge_connector_init() takes eight pointers to various bridges, some
of which can be identical, and stores them in pointers inside struct
drm_bridge_connector. Get a reference to each of the taken bridges and put
it on cleanup.
This is tricky because the pointers are currently stored directly in the
drm_bridge_connector in the loop, but there is no nice and clean way to put
those pointers on error return paths. To overcome this, store all pointers
in temporary local variables with a cleanup action, and only on success
copy them into struct drm_bridge_connector (getting another ref while
copying).
Additionally four of these pointers (edid, hpd, detect and modes) can be
written in multiple loop iterations, in order to eventually store the last
matching bridge. However, when one of those pointers is overwritten, we
need to put the reference that we got during the previous assignment. Add a
drm_bridge_put() before writing them to handle this.
Finally, there is also a function-local panel_bridge pointer taken inside
the loop and used after the loop. Use a cleanup action as well to ensure it
is put on return.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250926-drm-bridge-alloc-getput-bridge-connector-v2-1-138b4bb70576@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Ensure that imported buffers are properly mapped and unmapped in
the same way as regular buffers to properly handle buffers during
device's bind and unbind operations to prevent resource leaks and
inconsistent buffer states.
Imported buffers are now dma_mapped before submission and
dma_unmapped in ivpu_bo_unbind(), guaranteeing they are unmapped
when the device is unbound.
Add also imported buffers to vdev->bo_list for consistent unmapping
on device unbind. The bo->ctx_id is set in open() so imported
buffers have a valid context ID.
Debug logs have been updated to match the new code structure.
The function ivpu_bo_pin() has been renamed to ivpu_bo_bind()
to better reflect its purpose, and unbind tests have been refactored
for improved coverage and clarity.
Signed-off-by: Jacek Lawrynowicz <jacek.lawrynowicz@linux.intel.com>
Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>
Reviewed-by: Karol Wachowski <karol.wachowski@linux.intel.com>
Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com>
Link: https://lore.kernel.org/r/20250925145059.1446243-1-maciej.falkowski@linux.intel.com
According to the eDP specification (VESA Embedded DisplayPort Standard
v1.4b, Section 3.3.10.2), if the value of DP_EDP_PWMGEN_BIT_COUNT is
less than DP_EDP_PWMGEN_BIT_COUNT_CAP_MIN, the sink is required to use
the MIN value as the effective PWM bit count.
This commit updates the logic to clamp the reported
DP_EDP_PWMGEN_BIT_COUNT to the range defined by _CAP_MIN and _CAP_MAX.
As part of this change, the behavior is modified such that reading both
_CAP_MIN and _CAP_MAX registers is now required to succeed, otherwise
bl->max value could end up being not set although
drm_edp_backlight_probe_max() returned success.
This ensures correct handling of eDP panels that report a zero PWM
bit count but still provide valid non-zero MIN and MAX capability
values. Without this clamping, brightness values may be interpreted
incorrectly, leading to a dim or non-functional backlight.
For example, the Samsung ATNA40YK20 OLED panel used in the Lenovo
ThinkPad T14s Gen6 (Snapdragon) reports a PWM bit count of 0, but
supports AUX backlight control and declares a valid 11-bit range.
Clamping ensures brightness scaling works as intended on such panels.
Co-developed-by: Rui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: Rui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: Christopher Obbard <christopher.obbard@linaro.org>
Tested-by: Christopher Obbard <christopher.obbard@linaro.org>
Reviewed-by: Christopher Obbard <christopher.obbard@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20250814-topic-x1e80100-t14s-oled-dp-brightness-v7-1-b3d7b4dfe8c5@linaro.org
Replace vkms' vblank timer with the DRM implementation. The DRM
code is identical in concept, but differs in implementation.
Vblank timers are covered in vblank helpers and initializer macros,
so remove the corresponding hrtimer in struct vkms_output. The
vblank timer calls vkms' custom timeout code via handle_vblank_timeout
in struct drm_crtc_helper_funcs.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://lore.kernel.org/r/20250916083816.30275-4-tzimmermann@suse.de
Implement atomic_flush, atomic_enable and atomic_disable of struct
drm_crtc_helper_funcs for vblank handling. Driver with no further
requirements can use these functions instead of adding their own.
Also simplifies the use of vblank timers.
The code has been adopted from vkms, which added the funtionality
in commit 3a0709928b ("drm/vkms: Add vblank events simulated by
hrtimers").
v3:
- mention vkms (Javier)
v2:
- fix docs
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>
Link: https://lore.kernel.org/r/20250916083816.30275-3-tzimmermann@suse.de
The vblank timer simulates a vblank interrupt for hardware without
support. Rate-limits the display update frequency.
DRM drivers for hardware without vblank support apply display updates
ASAP. A vblank event informs DRM clients of the completed update.
Userspace compositors immediately schedule the next update, which
creates significant load on virtualization outputs. Display updates
are usually fast on virtualization outputs, as their framebuffers are
in regular system memory and there's no hardware vblank interrupt to
throttle the update rate.
The vblank timer is a HR timer that signals the vblank in software.
It limits the update frequency of a DRM driver similar to a hardware
vblank interrupt. The timer is not synchronized to the actual vblank
interval of the display.
The code has been adopted from vkms, which added the funtionality
in commit 3a0709928b ("drm/vkms: Add vblank events simulated by
hrtimers").
The new implementation is part of the existing vblank support,
which sets up the timer automatically. Drivers only have to start
and cancel the vblank timer as part of enabling and disabling the
CRTC. The new vblank helper library provides callbacks for struct
drm_crtc_funcs.
The standard way for handling vblank is to call drm_crtc_handle_vblank().
Drivers that require additional processing, such as vkms, can init
handle_vblank_timeout in struct drm_crtc_helper_funcs to refer to
their timeout handler.
There's a possible deadlock between drm_crtc_handle_vblank() and
hrtimer_cancel(). [1] The implementation avoids to call hrtimer_cancel()
directly and instead signals to the timer function to not restart
itself.
v4:
- fix possible race condition between timeout and atomic commit (Michael)
v3:
- avoid deadlock when cancelling timer (Ville, Lyude)
v2:
- implement vblank timer entirely in vblank helpers
- downgrade overrun warning to debug
- fix docs
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>
Link: https://lore.kernel.org/all/20250510094757.4174662-1-zengheng4@huawei.com/ # [1]
Link: https://lore.kernel.org/r/20250916083816.30275-2-tzimmermann@suse.de
The tidss_crtc_reset() function will (rightfully) destroy any
pre-existing state.
However, the tidss CRTC driver has its own CRTC state structure that
subclasses drm_crtc_state, and yet will destroy the previous state
by calling __drm_atomic_helper_crtc_destroy_state() and kfree() on its
drm_crtc_state pointer.
It works only because the drm_crtc_state is the first field in the
structure, and thus its offset is 0. It's incredibly fragile however, so
let's call our destroy implementation in such a case to deal with it
properly.
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20250902-drm-state-readout-v1-22-14ad5315da3f@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250902-drm-state-readout-v1-22-14ad5315da3f@kernel.org
