On Thursday 06 January 2005 3:05 pm, Ian Campbell wrote: > On Thu, 2005-01-06 at 13:10 -0800, David Brownell wrote: > > > But surely it'd be better to have one arch-neutral call in > > the chip's driver; nobody else needs to know how that > > hardware signal works. > > I don't know that you have a choice on all architectures. e.g. an ISA > IRQ is always rising edge. (or is it falling, in any case it is fixed). > PXA255 can only do rising and/or falling but not proper level triggered.
So the set_irq_type() call could have specialized implementations, with failure codes for "we can't do that on this hardware"; and the driver would have logic to cope with quirks like "can't set level triggered". That can probably be done so at least the PXA style quirks can be dealt with at compile time. (The IRQ handling is already platform-specific; PXA works differently from OMAP works differently from LH7A40x etc.) There's no rush on resolving this, but I did see some comments on LKML recently to the effect that even x86 needs such a call. I figure it'll happen someday, and until then I'llnot worry about the details. - Dave ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
