On Thu, Aug 30, 2012 at 7:23 AM, Geanta Neag Horia Ioan-B05471
[off-list ref] wrote:
Besides these:
1. As Dan Williams mentioned, you should explain why you are using
both spin_lock_bh and spin_lock_irqsave on the same lock.
It looks like talitos_process_pending() can be called from hardirq
context via talitos_error(). So spin_lock_bh is not sufficient.