On Fri, 13 Jan 2012, Andi Kleen wrote: > On Fri, Jan 13, 2012 at 09:28:20AM -0600, Christoph Lameter wrote: > > On Thu, 12 Jan 2012, Andi Kleen wrote: > > > > > The problem is that then there will be nothing left that actually > > > tests interleaving. The numactl has caught kernel regressions in the past. > > > > How about adding a CONFIG_NUMA_DEBUG option and have it only available > > then? I think there is no general use case. > > For a few lines of code? And making it harder to test? For now yes. We can then add more debugging stuff. Right now there is no framework for that. > > > I don't think disabling useful regression tests is a good idea. > > > In contrary the kernel needs far more of them, not less. > > > > True. Some more debugging code for the NUMA features would be appreciated > > but that does not need to be enabled by default. Lately I have become a > > bit concerned about the number of statistics we are adding. The > > per_cpu_pageset structure should not get too large. > > I don't think the single counter is a problem. I never said that .... There are multiple counters that may not be too useful in that structure. Not just the one thats useless. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>