kernel.git/drivers/pinctrl/pinctrl-amd.c, branch linux-5.11.y

Merge tag 'pinctrl-v5.11-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl

2020-12-16T23:02:49Z

Pull pin control updates from Linus Walleij: "This is the bulk of pin control changes for the v5.11 kernel. Drivers, drivers and drivers. Not a single core change. Some new stuff, especially a bunch of new Intel, Qualcomm and Ocelot SoCs. As part of the modularization attempt, I applied one patch affecting the firmware subsystem as a functional (not syntactic/semantic) dependency and then it blew up in our face, so I had to revert it, bummer. It will come in later, through that subsystem, I guess. New drivers: - New driver for the Microchip Serial GPIO "SGPIO". - Qualcomm SM8250 LPASS (Low Power Audio Subsystem) GPIO driver. New subdrivers: - Intel Lakefield subdriver. - Intel Elkhart Lake subdriver. - Intel Alder Lake-S subdriver. - Qualcomm MSM8953 subdriver. - Qualcomm SDX55 subdriver. - Qualcomm SDX55 PMIC subdriver. - Ocelot Luton SoC subdriver. - Ocelot Serval SoC subdriver. Modularization: - The Meson driver can now be built as modules. - The Qualcomm driver(s) can now be built as modules. Incremental improvements: - The Intel driver now supports pin configuration for GPIO-related configurations. - A bunch of Renesas PFC drivers have been augmented with support for QSPI pins, groups and functions. - Non-critical fixes to the irq handling in the Allwinner Sunxi driver" * tag 'pinctrl-v5.11-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (80 commits) pinctrl/spear: simplify the return expression of spear300_pinctrl_probe() pinctrl: mediatek: simplify the return expression of mtk_pinconf_bias_disable_set_rev1() dt-bindings: pinctrl: pinctrl-microchip-sgpio: Add irq support pinctrl: pinctrl-microchip-sgpio: Add irq support (for sparx5) pinctrl: qcom: Add sm8250 lpass lpi pinctrl driver dt-bindings: pinctrl: qcom: Add sm8250 lpass lpi pinctrl bindings pinctrl: qcom-pmic-gpio: Add support for pmx55 dt-bindings: pinctrl: qcom-pmic-gpio: Add pmx55 support pinctrl: pinctrl-microchip-sgpio: Mark some symbols with static keyword pinctrl: at91-pio4: Make PINCTRL_AT91PIO4 depend on HAS_IOMEM to fix build error pinctrl: mtk: Fix low level output voltage issue pinctrl: falcon: add missing put_device() call in pinctrl_falcon_probe() pinctrl: actions: pinctrl-s500: Constify s500_padinfo[] pinctrl: pinctrl-microchip-sgpio: Add OF config dependency pinctrl: pinctrl-microchip-sgpio: Add pinctrl driver for Microsemi Serial GPIO dt-bindings: pinctrl: Add bindings for pinctrl-microchip-sgpio driver pinctrl: at91-pio4: add support for fewer lines on last PIO bank pinctrl: sunxi: Always call chained_irq_{enter, exit} in sunxi_pinctrl_irq_handler pinctrl: sunxi: Mark the irq bank not found in sunxi_pinctrl_irq_handler() with WARN_ON pinctrl: sunxi: fix irq bank map for the Allwinner A100 pin controller ...

pinctrl: amd: remove debounce filter setting in IRQ type setting

2020-12-04T09:02:18Z

Debounce filter setting should be independent from IRQ type setting because according to the ACPI specs, there are separate arguments for specifying debounce timeout and IRQ type in GpioIo() and GpioInt(). Together with commit 06abe8291bc31839950f7d0362d9979edc88a666 ("pinctrl: amd: fix incorrect way to disable debounce filter") and Andy's patch "gpiolib: acpi: Take into account debounce settings" [1], this will fix broken touchpads for laptops whose BIOS set the debounce timeout to a relatively large value. For example, the BIOS of Lenovo AMD gaming laptops including Legion-5 15ARH05 (R7000), Legion-5P (R7000P) and IdeaPad Gaming 3 15ARH05, set the debounce timeout to 124.8ms. This led to the kernel receiving only ~7 HID reports per second from the Synaptics touchpad (MSFT0001:00 06CB:7F28). Existing touchpads like [2][3] are not troubled by this bug because the debounce timeout has been set to 0 by the BIOS before enabling the debounce filter in setting IRQ type. [1] https://lore.kernel.org/linux-gpio/20201111222008.39993-11-andriy.shevchenko@linux.intel.com/ 8dcb7a15a585 ("gpiolib: acpi: Take into account debounce settings") [2] https://github.com/Syniurge/i2c-amd-mp2/issues/11#issuecomment-721331582 [3] https://forum.manjaro.org/t/random-short-touchpad-freezes/30832/28 Signed-off-by: Coiby Xu Reviewed-by: Andy Shevchenko Cc: Hans de Goede Cc: Andy Shevchenko Cc: Benjamin Tissoires Cc: stable@vger.kernel.org Link: https://lore.kernel.org/linux-gpio/CAHp75VcwiGREBUJ0A06EEw-SyabqYsp%2Bdqs2DpSrhaY-2GVdAA%40mail.gmail.com/ BugLink: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1887190 Link: https://lore.kernel.org/r/20201125130320.311059-1-coiby.xu@gmail.com Signed-off-by: Linus Walleij

pinctrl: amd: print debounce filter info in debugfs

2020-11-10T13:58:14Z

pinctrl: amd: use higher precision for 512 RtcClk

2020-11-10T13:19:10Z

RTC is 32.768kHz thus 512 RtcClk equals 15625 usec. The documentation likely has dropped precision and that's why the driver mistakenly took the slightly deviated value. Cc: stable@vger.kernel.org Reported-by: Andy Shevchenko Suggested-by: Andy Shevchenko Suggested-by: Hans de Goede Signed-off-by: Coiby Xu Reviewed-by: Andy Shevchenko Reviewed-by: Hans de Goede Link: https://lore.kernel.org/linux-gpio/2f4706a1-502f-75f0-9596-cc25b4933b6c@redhat.com/ Link: https://lore.kernel.org/r/20201105231912.69527-3-coiby.xu@gmail.com Signed-off-by: Linus Walleij

pinctrl: amd: fix incorrect way to disable debounce filter

2020-11-10T13:18:37Z

The correct way to disable debounce filter is to clear bit 5 and 6 of the register. Cc: stable@vger.kerne.org Signed-off-by: Coiby Xu Reviewed-by: Hans de Goede Cc: Hans de Goede Link: https://lore.kernel.org/linux-gpio/df2c008b-e7b5-4fdd-42ea-4d1c62b52139@redhat.com/ Link: https://lore.kernel.org/r/20201105231912.69527-2-coiby.xu@gmail.com Signed-off-by: Linus Walleij

pinctrl: amd: Use irqchip template

2020-08-03T23:29:10Z

This makes the driver use the irqchip template to assign properties to the gpio_irq_chip instead of using the explicit call to gpiochip_irqchip_add(). The irqchip is instead added while adding the gpiochip. Signed-off-by: Linus Walleij Cc: Shyam Sundar S K Cc: Sandeep Singh Link: https://lore.kernel.org/r/20200722101545.144373-1-linus.walleij@linaro.org

pinctrl: pinctrl-amd: Do not define 'struct acpi_device_id' when !CONFIG_ACPI

2020-07-16T13:58:12Z

Since ACPI_PTR() is used to NULLify the value when !CONFIG_ACPI, 'struct amd_gpio_acpi_match' becomes defined but unused. Fixes the following W=1 kernel build warning(s): drivers/pinctrl/pinctrl-amd.c:959:36: warning: ‘amd_gpio_acpi_match’ defined but not used [-Wunused-const-variable=] 959 | static const struct acpi_device_id amd_gpio_acpi_match[] = { Signed-off-by: Lee Jones Cc: Ken Xue Cc: "Wu, Jeff" Cc: Nehal Shah Cc: Sundar S K Link: https://lore.kernel.org/r/20200713144930.1034632-26-lee.jones@linaro.org Signed-off-by: Linus Walleij

pinctrl: amd: Honor IRQ trigger type requested by the caller

2020-07-07T12:49:31Z

This change drops the override in `amd_gpio_irq_set_type()` that ignores the IRQ trigger type settings from the caller. The device driver (caller) is in a better position to identify the right trigger type for the device based on the usage as well as the information exposed by the BIOS. There are instances where the device driver might want to configure the trigger type differently in different modes. An example of this is gpio-keys driver which configures IRQ type as trigger on both edges (to identify assert and deassert events) when in S0 and reconfigures the trigger type using the information provided by the BIOS when going into suspend to ensure that the wake happens on the required edge. This override in `amd_gpio_irq_set_type()` prevents the caller from being able to reconfigure trigger type once it is set either based on ACPI information or the type used by the first caller for IRQ on a given GPIO line. Without this change, pen-insert gpio key (used by garaged stylus on a Chromebook) works fine in S0 (i.e. insert and eject events are correctly identified), however, BIOS configuration for wake on only pen eject i.e. only-rising edge or only-falling edge is not honored. With this change, it was verified that pen-insert gpio key behavior is correct in both S0 and for wakeup from S3. Signed-off-by: Furquan Shaikh Signed-off-by: Shyam Sundar S K Link: https://lore.kernel.org/r/20200626211026.513520-1-furquan@google.com Signed-off-by: Linus Walleij

pinctrl: Use new GPIO_LINE_DIRECTION

2020-02-21T14:19:42Z

Use newly added GPIO defines GPIO_LINE_DIRECTION_IN and GPIO_LINE_DIRECTION_OUT instead of using hard-coded 1 and 0. Main benefit is to make it easier to see which values mean IN and which OUT. As a side effect this helps GPIO framework to change the direction defines to something else if ever needed. Please note that return value from get_direction call on pinctrl-axp209 driver was changed. Previously pinctrl-axp209 might have returned value 2 for direction INPUT. Signed-off-by: Matti Vaittinen Reported-by: kbuild test robot Reviewed-by: Geert Uytterhoeven Acked-by: Geert Uytterhoeven Reviewed-by: Bjorn Andersson Reviewed-by: Jacopo Mondi Link: https://lore.kernel.org/r/20200214135712.GA14557@localhost.localdomain Signed-off-by: Linus Walleij

remove ioremap_nocache and devm_ioremap_nocache

2020-01-06T08:45:59Z

ioremap has provided non-cached semantics by default since the Linux 2.6 days, so remove the additional ioremap_nocache interface. Signed-off-by: Christoph Hellwig Acked-by: Arnd Bergmann