Messages in this thread | | | Date | Sat, 24 Jul 1999 14:15:39 +0200 (CEST) | From | Rik van Riel <> | Subject | Re: low priority soft RT? |
| |
On Fri, 23 Jul 1999, Kurt Garloff wrote: > On Fri, Jul 23, 1999 at 05:21:03PM +0100, Stephen C. Tweedie wrote: > > > I have made a fairly substantial SCHED_IDLE patch, but due to > > > the (never resolved) deadlock issues I haven't submitted it. > > > > Yes, it really is a loaded gun. A large SCHED_IDLE process can consume > > arbitrary system resources, and if there is any single cpu-bound process > > preventing it from being scheduled, you won't even be able to kill -9 > > the SCHED_IDLE task (the kill signal isn't delivered until the task is > > next scheduled). > > The kill -9 can be handled by promoting this process to another > scheduling class, when there are signals pending. Especially with > SIGKILL. Note that some of the OOM patches discussed for inclusion > into the kernel are promoted to SCHED_RR when being killed.
Promotion to a higher scheduling class shouldn't be too difficult. We can use a construction not unlike the goto theme park in schedule() to keep additional costs in goodness() down to an absolute minimum.
Please wait while I grab 2.3.11 and code up a patch...
Rik -- Open Source: you deserve to be in control of your data. +-------------------------------------------------------------------+ | Le Reseau netwerksystemen BV: http://www.reseau.nl/ | | Linux Memory Management site: http://www.linux.eu.org/Linux-MM/ | | Nederlandse Linux documentatie: http://www.nl.linux.org/ | +-------------------------------------------------------------------+
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |