On Wed, August 29, 2007 10:29 am, Jiri Kosina wrote: > On Tue, 28 Aug 2007, Adrian McMenamin wrote: > >> works fine for a minute or two with events/0 reporting, via top, CPU >> usage of about 0.3%. Then, without warning, events/0 whacks up to 99% >> and the system gets locked > > The first and easiest way of detecting what's going on might be sticking a > few printk()s into your code which is being scheduled in workqueue, to > make sure that it's not your code that triggers inifinite loop somewhere. >
I tried that, the printks were the fastest way to lock the system up - as they are essentially output at the same rate as vblanks. The code is obviously a bit racey in some way and as always that is very difficult to debug :( -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
