Messages in this thread | | | Subject | Re: 2.6.19-rc5: known regressions | From | Tim Chen <> | Date | Mon, 13 Nov 2006 14:46:44 -0800 |
| |
On Wed, 2006-11-08 at 22:10 -0700, Eric W. Biederman wrote:
> > Cool. I'm glad to know it was simply a buggy lmbench. > > What is sysconf(_SN_NPROCESSORS_ONLN) doing that it slows down as the > number of irqs increase? It is a slow path certainly but possibly > something we should fix. My hunch is cat /proc/cpuinfo... >
The increase in time of sysconf(_SN_NPROCESSORS_ONLN) call is within "show_stat" function after looking at profiling data. There are a couple of loops that iterate over kstat_irqs interrupt statistics and depend on NR_IRQS. Doesn't look like something we need to fix.
Tim - 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/
| |