Hennerich, Michael wrote on 2011-04-19: > Jonathan Cameron wrote on 2011-04-19: >> On 04/19/11 16:22, Hennerich, Michael wrote: >>> Hi Jonathan, >>> >>> The AD7606 ring buffer doesn't use the thread, and installs only >>> the hard handler. >>> >>> indio_dev->pollfunc->h = &ad7606_trigger_handler_th; >>> indio_dev->pollfunc->thread = NULL; This crashes the system >>> in handle_nested_irq (null pointer >>> action->thread_fn) called from iio_trigger_poll_chained(). >> I knew that wouldn't work, but didn't realize it wouldn't just fail >> with an error... >> >> The only thing I can think to do is to actually set both h and >> thread to ad7606_trigger_handler_th. >> >> As it returns IRQ_HANDLED, if it is called via irq_trigger_poll, it >> will happen in interrupt context and thread will never run. >> >> If it is called via irq_trigger_poll_handler (e.g. for non interrupt >> context) it'll happen outside interrupt context. Given timing is never >> going to be that tight for userspace triggers, this probably isn't a >> problem. >> >> Can you try that out and see if it works? > > I know that setting the thread function will effectively avoid the > crash. However I actually haven't traced if it's actually being called > once the Hard handler returned IRQ_HANDLED. > > I'll have try. I did some tests, using the sysfs trigger. The top-half handler is never being called. That's due to the fact that iio_trigger_poll_chained() calls handle_nested_irq(), which will only call the bottom-half thread function. I guess there is some fundamental issue here. Greetings, Michael -- Analog Devices GmbH Wilhelm-Wagenfeld-Str. 6 80807 Muenchen Sitz der Gesellschaft: Muenchen; Registergericht: Muenchen HRB 40368; Geschaeftsfuehrer:Dr.Carsten Suckrow, Thomas Wessel, William A. Martin, Margaret Seif -- 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
