Messages in this thread | | | Subject | Re: [PATCH] rtmutex: multiple candidate owners without unrelated boosting | From | Steven Rostedt <> | Date | Wed, 15 Dec 2010 09:32:37 -0500 |
| |
On Wed, 2010-12-15 at 15:16 +0100, Thomas Gleixner wrote: > On Wed, 15 Dec 2010, Steven Rostedt wrote: >
> That does not matter. The interesting part is whether the lock on > which orig_waiter is blocked on was unlocked. Lai's follow up patch does: > > + if (orig_waiter && !rt_mutex_owner(orig_lock)) > goto out_unlock_pi; >
I haven't looked at his new patch yet. But if this is the case, then we can nuke all cand_* fields. This was the only reason I kept the cand_owner in the first place.
Thanks,
-- Steve
| |