Hi,
On 06-11-2019 17:17, Andy Shevchenko wrote:
On Wed, Nov 06, 2019 at 06:16:22PM +0200, Andy Shevchenko wrote:
On Wed, Nov 06, 2019 at 04:47:15PM +0100, Hans de Goede wrote:
We need to convert all old gpio irqchips to pass the irqchip
setup along when adding the gpio_chip. For more info see
drivers/gpio/TODO.
For chained irqchips this is a pretty straight-forward conversion.
+ chip->irq.chip = &pctrl->irqchip;
+ if (pctrl->need_valid_mask)
+ chip->irq.init_valid_mask = chv_init_irq_valid_mask;
I just realize we probably may assign here unconditionally
Continuing...
+ chip->irq.init_hw = chv_gpio_irq_init_hw;
+ chip->irq.parent_handler = chv_gpio_irq_handler;
+ chip->irq.num_parents = 1;
+ chip->irq.parents = &pctrl->irq;
+ chip->irq.default_type = IRQ_TYPE_NONE;
+ chip->irq.handler = handle_bad_irq;
if (!pctrl->need_valid_mask) {
And here turn it back to NULL and check the pointer against NULL instead of
additional variable.
What do you think?
I think that first setting it and then clearing it again is not
very pretty. But ...
I do think you are on to something, we can use pctrl->chip.irq.init_valid_mask
instead of storing the dmi quirk in the chv_pinctrl struct.
Then we can leave the dmi handling as is, and replace later checks for
the dmi quirk (in callbacks) with a check for pctrl->chip.irq.init_valid_mask
I do believe that that is better then adding a need_validmask member to
the chv_pinctrl struct, I will prepare a v3 of the series with this change.
Regards,
Hans
