Messages in this thread | | | Date | Tue, 3 Jan 2006 19:41:01 +0530 | From | Dipankar Sarma <> | Subject | Re: [patch] latency tracer, 2.6.15-rc7 |
| |
On Tue, Jan 03, 2006 at 05:28:15AM -0800, David Lang wrote: > On Tue, 3 Jan 2006, Dipankar Sarma wrote: > > >I do agree that the two layers of batching really makes things > >subtle. I think the best we can do is to figure out some way of > >automatic throttling in RCU and forced quiescent state under extreme > >conditions. That way we will have less dependency on softirq > >throttling. > > would it make sense to have the RCU subsystems with a threshold so that > when more then X items are outstanding they trigger a premption of all > other CPU's ASAP (forceing the scheduling break needed to make progress on > clearing RCU)? This wouldn't work in all cases, but it could significantly > reduce the problem situations.
Yes, I think it would make sense to try something like that. See Paul's earlier mail in this thread for an example code snippet.
Thanks Dipankar - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |