On Tue, 2014-01-14 at 09:52 +0800, Dave Young wrote: > On 01/13/14 at 06:09pm, Toshi Kani wrote: > > On Mon, 2014-01-13 at 16:53 -0800, H. Peter Anvin wrote: > > > On 01/13/2014 04:33 PM, Toshi Kani wrote: > > > > > > > > I do not think it makes sense. You needed memmap=exactmap as a > > > > workaround because the kernel did not boot with the firmware's memory > > > > info. So, it's broken, and you requested the kernel to ignore the > > > > firmware info. > > > > > > > > Why do you think memory hotplug needs to be supported under such > > > > condition, which has to use the broken firmware info? > > > > > > > > > > Even more than memory hotplug: what do we do with NUMA? Since we have > > > already told the kernel "the firmware is bogus" it would seem that any > > > NUMA optimizations would be a bit ... cantankerous at best, no? > > > > Agreed that NUMA info can be bogus in this case, but is probably not > > critical. > > > > In majority of the cases, memmap=exactmap is used for kdump and the > > firmware info is sane. So, I think we should keep NUMA enabled since it > > could be useful when multiple CPUs are enabled for kdump. > > In Fedora kdump, we by default add numa=off to 2nd kernel cmdline because > enabling numa will use a lot more memory, at the same time we have only 128M > reserved by default.. That quite makes sense as we only enable a single CPU today. Thanks, -Toshi -- 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/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>