Re: [Qemu-devel] [PATCH v3 30/30] target/s390x: update maximum TCG model to z800

[David Hildenbrand <david@redhat.com>]

On 01.06.2017 21:17, Aurelien Jarno wrote:
> On 2017-06-01 10:38, David Hildenbrand wrote:
>> On 01.06.2017 00:01, Aurelien Jarno wrote:
>>> At the same time fix the TCG version of get_max_cpu_model to return the
>>> maximum model like on KVM. Remove the ETF2 and long-displacement
>>
>> I don't understand the part
>> "fix the TCG version of get_max_cpu_model to return the maximum model
>> like on KVM".
>>
>> Can you elaborate?
> 
> Currently get_max_cpu_model returns the features of the base model, so
> for example the one of a z900 even on a z800. This makes impossible to
> enable the features that are provided by a z800 like etf2 or ldisp.
> 

Right, you can always change the max_cpu_model, e.g. bumping up the
version or adding new features, that is just fine.

> For what I understand from the KVM code (but I haven't tested), the
> function return all the features that are supported by the current CPU,
> not all the features that are supported by the base model of the current
> CPU.

Correct, for KVM it is the detected model, that means: Base features +
optional features.


> 
> 
>>> facilities from the additional features as it is included in the z800.
>>>
>>> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
>>> ---
>>>  target/s390x/cpu_models.c | 13 ++++++-------
>>>  1 file changed, 6 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c
>>> index fc3cb25cc3..c13bbd852c 100644
>>> --- a/target/s390x/cpu_models.c
>>> +++ b/target/s390x/cpu_models.c
>>> @@ -668,8 +668,6 @@ static void add_qemu_cpu_model_features(S390FeatBitmap fbm)
>>>      static const int feats[] = {
>>>          S390_FEAT_STFLE,
>>>          S390_FEAT_EXTENDED_IMMEDIATE,
>>> -        S390_FEAT_EXTENDED_TRANSLATION_2,
>>> -        S390_FEAT_LONG_DISPLACEMENT,
>>>          S390_FEAT_LONG_DISPLACEMENT_FAST,
>>>          S390_FEAT_ETF2_ENH,
>>>          S390_FEAT_STORE_CLOCK_FAST,
>>> @@ -696,9 +694,9 @@ static S390CPUModel *get_max_cpu_model(Error **errp)
>>>      if (kvm_enabled()) {
>>>          kvm_s390_get_host_cpu_model(&max_model, errp);
>>>      } else {
>>> -        /* TCG emulates a z900 (with some optional additional features) */
>>> -        max_model.def = &s390_cpu_defs[0];
>>> -        bitmap_copy(max_model.features, max_model.def->default_feat,
>>> +        /* TCG emulates a z800 (with some optional additional features) */
>>> +        max_model.def = s390_find_cpu_def(0x2066, 7, 3, NULL);
>>> +        bitmap_copy(max_model.features, max_model.def->full_feat,
>>>                      S390_FEAT_MAX);
>>>          add_qemu_cpu_model_features(max_model.features);
>>>      }
>>> @@ -956,8 +954,9 @@ static void s390_qemu_cpu_model_initfn(Object *obj)
>>>      S390CPU *cpu = S390_CPU(obj);
>>>  
>>>      cpu->model = g_malloc0(sizeof(*cpu->model));
>>> -    /* TCG emulates a z900 (with some optional additional features) */
>>> -    memcpy(&s390_qemu_cpu_defs, &s390_cpu_defs[0], sizeof(s390_qemu_cpu_defs));
>>> +    /* TCG emulates a z800 (with some optional additional features) */
>>> +    memcpy(&s390_qemu_cpu_defs, s390_find_cpu_def(0x2066, 7, 3, NULL),
>>> +           sizeof(s390_qemu_cpu_defs));
>>
>> No changing the qemu model without compatibility handling.
> 
> This patch series is based on the patch from Thomas Huth. It means the
> QEMU model is still based on a z900, but that it is possible to enable
> some more features like etf2.

Thomas' code did neither change features nor the "model definition". It
just allows for some more feature to be set. It is a hack.

I am pretty sure that expanding the "qemu" CPU model now (QMP
query-cpu-model-expansion) will indicate a z800, not a z900.

See cpu_info_from_model(). And that is a problem, because the QEMU CPU
model is a "migration-safe" CPU model, meaning it must remain equal for
every compatibility machine.

Thanks.

> 
> Aurelien
> 


-- 

Thanks,

David