On Fri, Mar 17, 2017 at 03:09:09PM +0100, Peter Zijlstra wrote: > On Fri, Mar 17, 2017 at 09:24:11PM +0800, Wanpeng Li wrote: > > I have another version of patch which utilizes raw_spin_lock_irqsave() > > instead of rq_lock_irqsave() in __balance_callback() as before, which > > one do you like, Peterz? > > Hurm.. the raw_spin_lock_irqsave() one I suspect. No point in pinning > and then unpinning.
Took your other version and edited that to make it do this.
