On 6/15/2012 2:08 PM, Lars-Peter Clausen wrote:
hi,
The sysfs trigger uses iio_trigger_poll_chained which calls
handle_nested_irq. The problem now is that for nested IRQs the primary
handler is not called. Which obviously breaks drivers which have a bottom
half trigger function.
This behaviour was introduced in commit 1f785681 ("staging:iio:trigger sysfs
userspace trigger rework."). The commit message says you are "awaiting
comments on using the nested_irq_trick", but not why it is necessary to use
nested IRQs. And honestly I don't get why it would be necessary either.
handle_nested_irq is supposed to be used with chained IRQs if we are already
running in a thread handler of parent. Neither seems to be true here.
It was a while ago so my memory is rather sketchy.
do you meant the bottom half? Unless I'm very confused its the top half
that doesn't
get called.. handle_nested_irq calls thread_fn which is the bottom half.
It's the only way I've come up with for cleanly running through our
trigger distribution
given that is all irq based. Top halves expect to be run in interrupt
mode. I don't know
of any way to ensure this is true if one is 'creating' the interrupt
from software.
I did wonder at the time about putting an explicit call of the interrupt
handler in
to deal with the missing top halves. It's rather uggly though and
suffers from things
not being run in the state they expect to be run in....
Other suggestions welcome.
--
To unsubscribe from this list: send the line "unsubscribe linux-iio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
