Hi All, Here is version 6 of my series to adjust the INT3472 code's handling of the privacy LED on x86 laptops with MIPI camera(s) so that it will also work on devices which have a privacy-LED GPIO but not a clk-enable GPIO (so that we cannot just tie the LED state to the clk-enable state). Changes in v6: - The LED lookup series has been merged, immutable branch pull-req here: https://lore.kernel.org/platform-driver-x86/Y9QGcA+9nlmOOy2d@xxxxxxxxxx/ - Rework the media-core changes to correctly free the LED reference, this rework also allows dropping the patch to merge the async + fwnode code into videodev.ko - Drop the patch merging the async + fwnode code into videodev.ko Changes in v5: - Rename lookup-table names to match those from the gpio and reset lookups: s/led_name/provider/ s/consumer_dev_name/dev_id/ s/consumer_function/con_id/ - Add static inline wrappers for the v4l2_async debugfs init/exit funcs, to fix build errors when CONFIG_V4L2_ASYNC is not enabled Changes in v4: - Rename new __led_get() helper to led_module_get() - Drop of/devicetree support from "led-class: Add generic [devm_]led_get()" - Add RFC patch to re-add of/devicetree support to show that the new led_get() can easily be extended with dt support when the need for this arises (proof-of-concept dt code, not intended for merging) - New patch to built async and fwnode code into videodev.ko, to avoid issues with some of the new LED code getting builtin vs other parts possibly being in a module - Move the led_get() call to v4l2_async_register_subdev_sensor() - Move the led_disable_sysfs() call to be done at led_get() time - Address some other minor review comments Changes in v3: - Due to popular request by multiple people this new version now models the privacy LED as a LED class device. This requires being able to "tie" the LED class device to a specific camera sensor (some devices have multiple sensors + privacy-LEDs). Patch 1 adds generic privacy-LED support to the v4l2-core/v4l2-subdev.c code automatically enabling the privacy-LED when s_stream(subdev, 1) is called. So that we don't need to add privacy-LED code to all the camera sensor drivers separately (as requested by Sakari). Patches 2-5 are patches to the platform specific INT3472 code to register privacy-LED class devices + lookup table entries for privacy-LEDs described in the special INT3472 ACPI nodes found on x86 devices with MIPI cameras. This depends on the just merged LED lookup code from: https://lore.kernel.org/platform-driver-x86/Y9QGcA+9nlmOOy2d@xxxxxxxxxx/ This series has been tested on: - Lenovo ThinkPad X1 Yoga gen 7, IPU6, front: ov2740 with privacy LED - Dell Latitude 9420, IPU 6, front: ov01a1s with privacy LED - Mirosoft Surface Go, IPU3, front: ov5693 with privacy LED back: ov8865 with privacy LED (pled not yet supported) Regards, Hans Hans de Goede (5): media: v4l2-core: Make the v4l2-core code enable/disable the privacy LED if present platform/x86: int3472/discrete: Refactor GPIO to sensor mapping platform/x86: int3472/discrete: Create a LED class device for the privacy LED platform/x86: int3472/discrete: Move GPIO request to skl_int3472_register_clock() platform/x86: int3472/discrete: Get the polarity from the _DSM entry drivers/media/v4l2-core/v4l2-async.c | 4 + drivers/media/v4l2-core/v4l2-fwnode.c | 7 ++ drivers/media/v4l2-core/v4l2-subdev-priv.h | 14 +++ drivers/media/v4l2-core/v4l2-subdev.c | 44 ++++++++ drivers/platform/x86/intel/int3472/Makefile | 2 +- .../x86/intel/int3472/clk_and_regulator.c | 34 ++++-- drivers/platform/x86/intel/int3472/common.h | 18 +++- drivers/platform/x86/intel/int3472/discrete.c | 100 ++++++++---------- drivers/platform/x86/intel/int3472/led.c | 74 +++++++++++++ include/media/v4l2-subdev.h | 3 + 10 files changed, 234 insertions(+), 66 deletions(-) create mode 100644 drivers/media/v4l2-core/v4l2-subdev-priv.h create mode 100644 drivers/platform/x86/intel/int3472/led.c -- 2.39.1
