Re: Bug in "genirq: record trigger type"

[David Brownell <david-b@pacbell.net>]

On Monday 20 October 2008, Benjamin Herrenschmidt wrote:
> This one is obviously broken and breaks booting on a whole bunch of
> machines (including powermac's and thus my G5, it's never good when my
> own machine breaks !).
> 
> Nice to see 3 SOB's and one Ack and nobody caught the obvious bug :-)

As you saw, that one's fixed.  Chris' patch unfortunately didn't
get integrated right away.


I'm a bit more curious about another potential issue though ... as
described in the patch comment:

     - Make set_irq_type() usage match request_irq() usage:
        * IRQ_TYPE_NONE should be a NOP; succeed, so irq_chip methods
          won't have to handle that case any more (many do it wrong).

It might be a bit more accurate to say irq_chip.set_type() methods
are *inconsistent* in handling IRQ_TYPE_NONE.  Previously the
set_irq_type() method would pass that down to irq_chip code. 

I had observed two behaviors, but I thought I observed a third one
in some of the PowerPC code:

 (1) ignore it ... matching request_irq() usage
 (2) return an error ... nasty
 (3) assign some irq_chip-specific trigger mode

That third behavior might cause a bit of trouble, but I think
it was only used during platform init.  Someone more attuned
to PowerPC might want to check ...

- Dave