On Wed, Dec 19, 2018 at 9:58 PM Christoph Hellwig <[email protected]> wrote: > > On Tue, Dec 18, 2018 at 02:20:10PM +0530, Anup Patel wrote: > > Actually these functions should not be inline because plic_toggle() uses > > raw_spin_lock() and plic_irq_toggle() uses for-loop. > > So? It still inlines the all of two instances into each caller > for slightly different but related work. Not sure it is 100% worth > it, but probably more than the one to move the calculations to init > time..
Not just at init time but these functions will also be used when irq_affinity is changed by IRQ balancer at runtime. Regards, Anup
