Messages in this thread | | | Date | Mon, 28 Jan 2013 10:07:07 +0800 | From | Tang Chen <> | Subject | Re: [PATCH 3/3] acpi, memory-hotplug: Support getting hotplug info from SRAT. |
| |
On 01/26/2013 09:29 AM, H. Peter Anvin wrote: > On 01/25/2013 05:12 PM, Andrew Morton wrote: >> On Fri, 25 Jan 2013 17:42:09 +0800 >> Tang Chen<tangchen@cn.fujitsu.com> wrote: >> >>> NOTE: Using this way will cause NUMA performance down because the whole node >>> will be set as ZONE_MOVABLE, and kernel cannot use memory on it. >>> If users don't want to lose NUMA performance, just don't use it. >> >> I agree with this, but it means that nobody will test any of your new code. >> >> To get improved testing coverage, can you think of any temporary >> testing-only patch which will cause testers to exercise the >> memory-hotplug changes? >> > > There is another problem: if ALL the nodes in the system support > hotpluggable memory, what happens? >
Hi HPA,
I think I missed this case. If all the memory is hotpluggable, and user specified movablemem_map=acpi, all the memory could be set as movable, and the kernel will fail to start.
I will post a patch to fix it. How about always keep node0 unhotpluggable ?
Thanks. :)
| |