linux/drivers/gpio, branch v5.4

Merge tag 'gpio-v5.4-rc8-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux into fixes

2019-11-13T21:58:01Z

gpio fixes for v5.4-rc8 - fix debounce times in max776520 and bd70528 - fix parallel build of gpio tools

gpiolib: acpi: Add Terra Pad 1061 to the run_edge_events_on_boot_blacklist

2019-11-13T21:57:26Z

The Terra Pad 1061 has the usual micro-USB-B id-pin handler, but instead of controlling the actual micro-USB-B it turns the 5V boost for the tablet's USB-A connector and its keyboard-cover connector off. The actual micro-USB-B connector on the tablet is wired for charging only, and its id pin is *not* connected to the GPIO which is used for the (broken) id-pin event handler in the DSDT. While at it not only add a comment why the Terra Pad 1061 is on the blacklist, but also fix the missing comment for the Minix Neo Z83-4 entry. Fixes: 61f7f7c8f978 ("gpiolib: acpi: Add gpiolib_acpi_run_edge_events_on_boot option and blacklist") Signed-off-by: Hans de Goede Reviewed-by: Andy Shevchenko Acked-by: Mika Westerberg Signed-off-by: Linus Walleij

gpio: bd70528: Use correct unit for debounce times

2019-11-12T10:18:52Z

The debounce time passed to gpiod_set_debounce() is specified in microseconds, so make sure to use the correct unit when computing the register values, which denote delays in milliseconds. Signed-off-by: Thierry Reding Cc: Fixes: 18bc64b3aebf ("gpio: Initial support for ROHM bd70528 GPIO block") [Bartosz: fixed a typo in commit message] Signed-off-by: Bartosz Golaszewski

gpio: max77620: Fixup debounce delays

2019-11-12T10:13:54Z

When converting milliseconds to microseconds in commit fffa6af94894 ("gpio: max77620: Use correct unit for debounce times") some ~1 ms gaps were introduced between the various ranges supported by the controller. Fix this by changing the start of each range to the value immediately following the end of the previous range. This way a debounce time of, say 8250 us will translate into 16 ms instead of returning an -EINVAL error. Typically the debounce delay is only ever set through device tree and specified in milliseconds, so we can never really hit this issue because debounce times are always a multiple of 1000 us. The only notable exception for this is drivers/mmc/host/mmc-spi.c where the CD GPIO is requested, which passes a 1 us debounce time. According to a comment preceeding that code this should actually be 1 ms (i.e. 1000 us). Reported-by: Pavel Machek Signed-off-by: Thierry Reding Acked-by: Pavel Machek Cc: Signed-off-by: Bartosz Golaszewski

Revert "gpio: merrifield: Pass irqchip when adding gpiochip"

2019-11-03T22:41:11Z

This reverts commit 8f86a5b4ad679e4836733b47414226074eee4e4d. It has been established that this causes a boot regression on both Baytrail and Cherrytrail SoCs, and we can't have that in the final kernel release, so we need to revert it. Reported-by: Hans de Goede Acked-by: Andy Shevchenko Signed-off-by: Linus Walleij

Revert "gpio: merrifield: Restore use of irq_base"

2019-11-03T22:40:48Z

This reverts commit 6658f87f219427ee776c498e07c878eb5cad1be2. This revert is a prerequisite for the later revert of commit 8f86a5b4ad679e4836733b47414226074eee4e4d. Reported-by: Hans de Goede Acked-by: Andy Shevchenko Signed-off-by: Linus Walleij

Revert "gpio: merrifield: Move hardware initialization to callback"

2019-11-03T22:38:39Z

This reverts commit 4c87540940cbc7ddbe9674087919c605fd5c2ef1. This revert is a prerequisite for the later revert of commit 8f86a5b4ad679e4836733b47414226074eee4e4d. Reported-by: Hans de Goede Acked-by: Andy Shevchenko Signed-off-by: Linus Walleij

gpio: lynxpoint: set default handler to be handle_bad_irq()

2019-10-14T23:19:05Z

We switch the default handler to be handle_bad_irq() instead of handle_simple_irq() (which was not correct anyway). Signed-off-by: Andy Shevchenko Signed-off-by: Linus Walleij

gpio: merrifield: Move hardware initialization to callback

2019-10-14T23:19:01Z

The driver wants to initialize related registers before IRQ chip will be added. That's why move it to a corresponding callback. It also fixes the NULL pointer dereference. Fixes: 8f86a5b4ad67 ("gpio: merrifield: Pass irqchip when adding gpiochip") Signed-off-by: Andy Shevchenko Signed-off-by: Linus Walleij

gpio: lynxpoint: Move hardware initialization to callback

2019-10-14T23:18:57Z

The driver wants to initialize related registers before IRQ chip will be added. That's why move it to a corresponding callback. It also fixes the NULL pointer dereference. Fixes: 7b1e889436a1 ("gpio: lynxpoint: Pass irqchip when adding gpiochip") Signed-off-by: Andy Shevchenko Signed-off-by: Linus Walleij