Re: [Qemu-devel] [PATCH 0/7] spapr: rework memory nodes

[Nishanth Aravamudan]

On 16.06.2014 [18:16:29 +1000], Alexey Kardashevskiy wrote:
> On 06/16/2014 05:53 PM, Alexey Kardashevskiy wrote:
> > c4177479 "spapr: make sure RMA is in first mode of first memory node"
> > introduced regression which prevents from running guests with memoryless
> > NUMA node#0 which may happen on real POWER8 boxes and which would make
> > sense to debug in QEMU.
> > 
> > This patchset aim is to fix that and also fix various code problems in
> > memory nodes generation.
> > 
> > These 2 patches could be merged (the resulting patch looks rather ugly):
> > spapr: Use DT memory node rendering helper for other nodes
> > spapr: Move DT memory node rendering to a helper
> > 
> > Please comment. Thanks!
> > 
> 
> Sure I forgot to add an example of what I am trying to run without errors
> and warnings:

<snip>

> -numa node,nodeid=0,cpus=0-7,memory=0 \
> -numa node,nodeid=2,cpus=0-3,mem=1040 \
> -numa node,nodeid=4,cpus=4-7,mem=1040

Semantically, what does this mean? CPUs 0-3 are on both node 0 and node
2? I didn't think the NUMA spec allowed that? Or does qemu's
command-line take the "last" specified assignment of a CPU to a nodeid?

Perhaps unrelated to your changes, but I think it would be most sensible
here to error out if a CPU is assigned to multiple NUMA nodes.

<snip>

> address@hidden ~]# numactl --hardware
> 
> available: 3 nodes (0,2,4)
> node 0 cpus:
> node 0 size: 0 MB
> node 0 free: 0 MB
> node 2 cpus: 0 1 2 3
> node 2 size: 1021 MB
> node 2 free: 610 MB
> node 4 cpus: 4 5 6 7
> node 4 size: 1038 MB
> node 4 free: 881 MB
> node distances:
> node   0   2   4
>   0:  10  40  40
>   2:  40  10  40
>   4:  40  40  10
> 
> 
> Seems correct except that weird node#0 which comes I do not where from.

Well, Linux has a statically online Node 0, which if no CPUs or memory
are assigned to it, will show up as above as a cpuless and memoryless
node. That's not a bug in qemu, and is something I'm looking into
upstream in the kernel.

> And the patchset is made agains agraf/ppc-next tree.