On Thu, Aug 27, 2015 at 01:44:45PM +0200, Christian Hartmann wrote:
Hey,
2015-08-26 12:01 GMT+02:00 Mika Westerberg mika.westerberg@linux.intel.com:
The master should already have IRQ. This is about assigning IRQ for the SPI childe device.
ok, *saved*
You probably want to do same to SPI core that was done for I2C:
845c877009cf014b9 ("i2c / ACPI: Assign IRQ for devices that have GpioInt automatically")
I looked at this log and have added or better set spi->irq now in arizon-spi.c arizona_spi_probe() and its that what I have already via local patches added: (ACPI) using irq_gpio GPIO = 146 The question for me is, this looks like an GPIO number, not like an IRQ number. Also it seems better to look again into pxa2xx.c or spi.c (not done yet) and get the irq there.
If you are calling acpi_dev_gpio_irq_get() for the SPI device it should return you valid Linux IRQ number (or fail with an error). But it requires that your device has GpioInt resource somewhere in _CRS.
Then it will automatically assign first found GpioInt to the SPI device in question.
I again saw also the following files, which are not used under this device or they need some attention
drivers/gpio/gpio-arizona.c and drivers/extcon/extcon-arizona.c
I have already CONFIG_GPIO_ARIZONA=y, but it does not be used at all. Is it save to ignore ?? Or should I use them as well ??
If the SPI device interrupt is connected to arizona GPIO pin, then you probably need that.