Messages in this thread | | | Date | Fri, 20 Aug 2004 12:11:06 +1000 | From | Nick Piggin <> | Subject | Re: [PATCH] add scheduler domains for ia64 |
| |
Jesse Barnes wrote: > On Saturday, August 14, 2004 8:54 pm, Nick Piggin wrote: > >>Yeah, all the SD_*_INIT values are overridable. We could even say, put >>in an SD_NODE2_INIT for a 2nd level NUMA domain in the generic code, >>for example. > > > Yeah, we'll need different values for each level in the hierarchy. > > >>I'd say your closest-node setup would probably get close to what you want. >>The main thing you want is to not do huge amounts of balancing work in >>interrupt context, and also not to move a task from one side of the >>system to the other when one node is a little bit out of balance. >> >>I guess if you want to do anything fancier then we can take a look at >>re-exporting the domain setup. > > > Ok, sounds good. How does this look? It sits on top of 2.6.8.1-mm1, ripping > out the ia64 specific bits and moving things to sched.c. I've also added an > ia64 specific SD_NODE_INIT and an #if !defined to sched.c >
Sorry I haven't replied earlier. I think this looks good, provided it does the right thing for you (I can't test it myself). Send it to Andrew to get merged if you'd like. - 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/
| |