Messages in this thread | | | Date | Fri, 14 Nov 2003 11:28:58 -0800 | From | Mike Fedyk <> | Subject | Re: hot cache line due to note_interrupt() |
| |
On Sat, Nov 15, 2003 at 06:10:46AM +1100, Anton Blanchard wrote: > > > Probably too late for 2.6.0, but here is a patch that disables noirqdebug: > > Why dont you just disable it during boot somewhere in sgi ia64 specific > code? It doesnt seem right to disable this after all the driver effort > it took to make it work. > > And yes Im a paid up member of the "we build stupidly big machines" > club. I'll disable where applicable in ppc64 specific code.
Since the problem only starts when the cache line bounces between the CPUs, then it might be better to just disable if more than 150 processors are detected?
If it's a layering voilation to disable this during SMP init, then maybe there can be a smp-quirks.c file where exceptions can be put. - 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/
| |