4 Apr
2013
4 Apr
'13
4:49 p.m.
On Thu, 2013-04-04 at 15:31 +0200, Daniel Vetter wrote:
We've discussed this approach of using (rt-prio, age) instead of just age to determine the the "oldness" of a task for deadlock-breaking with -EAGAIN. The problem is that through PI boosting or normal rt-prio changes while tasks are trying to acquire ww_mutexes you can create acyclic loops in the blocked-on graph of ww_mutexes after the fact and so cause deadlocks. So we've convinced ourselves that this approche doesn't work.
Could you pretty please draw me a picture, I'm having trouble seeing what you mean.
AFAICT as long as you boost Y while its the lock owner things should work out, no?