> On 10/22/2013 01:30 PM, David Cohen wrote:
> > On 10/22/2013 12:46 PM, Alexander Shiyan wrote:
> >>> The change in the max3110 driver makes the IRQ handling threaded, now
> >>> the handler is called only once per received character. Without that
> >>> change, we had many (more than 100) interrupts per one received
> >>> character.
> >>>
> >>> Unfortunately, SFI interface does not support IRQ polarity and
> >>> triggering modes, so we have to keep the hacks as hard-coded device
> >>> names and IRQ numbers until we switch to ACPI.
> >>>
> >>> Edge-triggered IRQ still supported to keep old platforms working.
> >>> Use platform data to pass the irq mode argument.
> >>>
> >>> Signed-off-by: Ivan Gorinov <ivan.gorinov@xxxxxxxxx>
> >>> Signed-off-by: Li Ning <ning.li@xxxxxxxxx>
> >>> Signed-off-by: David Cohen <david.a.cohen@xxxxxxxxxxxxxxx>
> >> ...
> >>> +++ b/include/linux/serial_max3110.h
> >>> @@ -0,0 +1,16 @@
> >>> +#ifndef _LINUX_SERIAL_MAX3110_H
> >>> +#define _LINUX_SERIAL_MAX3110_H
> >>> +
> >>> +/**
> >>> + * struct plat_max3110 - MAX3110 SPI UART platform data
> >>> + * @irq_edge_trigger: if IRQ is edge triggered
> >>> + *
> >>> + * You should use this structure in your machine description to specify
> >>> + * how the MAX3110 is connected.
> >>> + *
> >>> + */
> >>> +struct plat_max3110 {
> >>> + int irq_edge_triggered;
> >>> +};
> >>> +
> >>> +#endif
> >>> --
> >>
> >> Is just resource->flags for IRQ can be reused for handle such case?
> >
> > I believe your suggestion makes perfect sense. I'll rework it.
>
> Looks like isp_device has no place for 'resource'. In this case pdata
> seems to be the way to go here.
> Or maybe there's a better way to recommend?
What do you mean by the isp_device?
My idea is always use threaded irq and passing flags into request.
Like as:
unsigned long flags = res->flags & IORESOURCE_BITS;
...
request_threaded_irq(max->irq, serial_m3110_irq, IRQF_ONESHOT | flags, "max3110", max);
---
��.n��������+%������w��{.n�����{��ǫ����{ay� ʇڙ���f���h������_�(�階�ݢj"��������G����?���&��
