Re: [PATCH v2 3/3] mm/page_alloc: Keep memoryless cpuless node 0 offline

[Srikar Dronamraju <srikar@linux.vnet.ibm.com>]

* David Hildenbrand <david@redhat.com> [2020-05-12 09:49:05]:

> On 11.05.20 19:47, Srikar Dronamraju wrote:
> > * David Hildenbrand <david@redhat.com> [2020-05-08 15:42:12]:
> > 
> > 
> > [root@localhost ~]# cat /sys/devices/system/node/online
> > 0
> > [root@localhost ~]# cat /sys/devices/system/node/possible
> > 0-1
> > 
> > Even without my patch, both the combinations, I am still unable to see a
> > cpuless, memoryless node being online. And the interesting part being even
> 
> Yeah, I think on x86, all memory-less and cpu-less nodes are offline as
> default. Especially when hotunplugging cpus/memory, we set them offline
> as well.

I also came to the same conclusion that we may not have a cpuless,memoryless
node on x86.

> 
> But as Michal mentioned, the node handling code is complicated and
> differs between various architectures.
> 

I do agree that node handling code differs across various architectures and
quite complicated.

> > if I mark node 0 as cpuless,memoryless and node 1 as actual node, the system
> > somewhere marks node 0 as the actual node.
> 
> Is the kernel maybe mapping PXM 1 to node 0 in that case, because it
> always requires node 0 to be online/contain memory? Would be interesting
> what happens if you hotplug a DIMM to (QEMU )node 0 - if PXM 0 will be
> mapped to node 1 then as well.
> 

Satheesh Rajendra had tried with cpu hotplug on a similar setup and we found
that it crashes the x86 system.
reference: https://bugzilla.kernel.org/show_bug.cgi?id=202187

Even if we were able to hotplug 1 DIMM memory into node 1, that would no
more be a memoryless node.

-- 
Thanks and Regards
Srikar Dronamraju