Re: [Qemu-devel] [PATCH] vl.c: make sure maxcpus matches topology to prevent migration failure

[Igor Mammedov]

On Thu, 23 Aug 2018 15:03:07 -0300
Eduardo Habkost <address@hidden> wrote:

> On Thu, Aug 23, 2018 at 06:32:41PM +0200, Paolo Bonzini wrote:
> > On 23/08/2018 16:51, Igor Mammedov wrote:  
> > > Topology (threads*cores*sockets) must match maxcpus to be valid,
> > > otherwise we could start QEMU with invalid topology that throws
> > > a error on migration destination side, that should not be reachable:
> > > Source:
> > >   -smp 8,maxcpus=64,cores=1,threads=8,sockets=1
> > > // hotplug cpus upto maxcpus
> > > Destination:
> > >   -smp 64,maxcpus=64,cores=1,threads=8,sockets=1
> > >   qemu: cpu topology: sockets (1) * cores (1) * threads (8) < smp_cpus 
> > > (64)  
> > 
> > The destination should have sockets=8, shouldn't it?
> > 
> > It seems to me that, at startup, you should have cpus = s*t*c and cpus
> > <= maxcpus.  Currently we check cpus <= s*t*c <= maxcpus, which doesn't
> > make much sense.  
> 
> Most of the incompleteness of input validation at smp_parse() can
> be explained by our fear of breaking existing configurations and
> making existing running VMs not runnable.
> 
> But now we have a deprecation policy.  If we're still afraid of
> breaking peoples' existing configurations, we should at least
> deprecate those configurations as soon as possible (and make QEMU
> at least emit a warning).
Sure, I can send a deprecation patch first.