On Fri, Jul 24, 2015 at 05:40:13PM +0200, Daniel Vetter wrote: > Afaict intel_irq_fini never existed. No idea how that one came > about. It's notable by its absence. We should write it! There are a few steps in intel_irq_init() that would be best undone intel_irq_fini(). The work handlers could be dealt with by a comment that the core does a flush_wq() - that at leasts says we have done due diligence and checked that the work will be complete before module unload. But we should move the pm_qos teardown here for instance. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
