Messages in this thread | | | Date | Wed, 14 Oct 2015 11:28:14 +0200 | From | Peter Zijlstra <> | Subject | Re: [PATCH v7 4/5] locking/pvqspinlock: Allow 1 lock stealing attempt |
| |
On Tue, Oct 13, 2015 at 04:50:25PM -0400, Waiman Long wrote: > On 10/13/2015 03:56 PM, Peter Zijlstra wrote:
> >So the below is exactly duplicated from the normal slowpath, so why > >don't you keep that there? > > > >It would get you something like: > > > > if (pv_wait_head_or_steal(..)) > > goto stolen; > > > > > >stolen: > >>+ /* > >>+ * contended path; wait for next, release. > >>+ */ > >>+ while (!(next = READ_ONCE(node->next))) > >>+ cpu_relax(); > >>+ > >>+ arch_mcs_spin_unlock_contended(&next->locked); > >>+ pv_kick_node(lock, next); > >release: > > ... > > Yes, it is largely the same. I thought that you don't like too much change > in the logic flow of the generic qspinlock code. I will make the change in > the next revision.
Well, you already put the branch in there, the only difference here is an 'extra' label. OTOH that extra label avoids duplicating some hairy code. So over all I would say its a definite win.
And its easy to see it will compile away on the native case where:
#define pv_wait_head_or_steal(l, n, t) (false)
| |