Re: [Qemu-devel] [qemu-s390x] [PATCH v1] s390x/cpu_models: Add "-cpu max" support

From: David Hildenbrand <david@redhat.com>
To: Eduardo Habkost <ehabkost@redhat.com>
Cc: "Thomas Huth" <thuth@redhat.com>,
	"Chris Venteicher" <cventeic@redhat.com>,
	"Daniel P. Berrangé" <berrange@redhat.com>,
	"Cornelia Huck" <cohuck@redhat.com>,
	"Alexander Graf" <agraf@suse.de>,
	qemu-devel@nongnu.org,
	"Christian Borntraeger" <borntraeger@de.ibm.com>,
	qemu-s390x@nongnu.org, "Collin Walling" <walling@linux.ibm.com>,
	"Richard Henderson" <rth@twiddle.net>
Subject: Re: [Qemu-devel] [qemu-s390x] [PATCH v1] s390x/cpu_models: Add "-cpu max" support
Date: Fri, 27 Jul 2018 16:57:32 +0200	[thread overview]
Message-ID: <5a5f4272-907c-d2d3-a7b2-30366eeed4b5@redhat.com> (raw)
In-Reply-To: <20180726150754.GL12380@localhost.localdomain>

> If we had a S390KVMAccelerator object on machine->accelerator,
> S390KVMAccelerator::host_model would be a good candidate?

Depends if at that point the machine would already be initialized (we
disable CPU model support for KVM on some legacy machine due to
interactions). It's complicated :)
[...]

>> Right now the semantics are clear: if we have "!cpu->model" after the
>> object has been created, details about the host CPU model are not
>> available (models unavailable/unsupported). Modifying properties,
>> baselining, expanding is not possible with that model then. But it can
>> be used for execution.
> 
> This is interesting.  If most users of cpu->model don't care
> about kvm_s390_get_host_cpu_model() errors at all, the current
> solution sounds more reasonable.
> 
> Except for the error_report_err() call inside instance_init.
> This still bothers me, but it's not a big deal.

Yes, we should refactor that. I'll add this to my TODO list!

-- 

Thanks,

David / dhildenb