Re: [PATCH v7 8/8] s390x/s390-virtio-ccw: add zpcii-disable machine property

[Matthew Rosato <mjrosato@linux.ibm.com>]

On 6/23/22 9:50 AM, Pierre Morel wrote:
> 
> 
> On 6/22/22 17:20, Matthew Rosato wrote:
>> On 6/22/22 4:50 AM, Pierre Morel wrote:
>>>
>>>
>>> On 6/6/22 22:36, Matthew Rosato wrote:
>>>> The zpcii-disable machine property can be used to force-disable the use
>>>> of zPCI interpretation facilities for a VM.  By default, this setting
>>>> will be off for machine 7.1 and newer.
>>>>
>>>> Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
>>>> ---
>>>>   hw/s390x/s390-pci-kvm.c            |  4 +++-
>>>>   hw/s390x/s390-virtio-ccw.c         | 24 ++++++++++++++++++++++++
>>>>   include/hw/s390x/s390-virtio-ccw.h |  1 +
>>>>   qemu-options.hx                    |  8 +++++++-
>>>>   util/qemu-config.c                 |  4 ++++
>>>>   5 files changed, 39 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/hw/s390x/s390-pci-kvm.c b/hw/s390x/s390-pci-kvm.c
>>>> index 9134fe185f..5eb7fd12e2 100644
>>>> --- a/hw/s390x/s390-pci-kvm.c
>>>> +++ b/hw/s390x/s390-pci-kvm.c
>>>> @@ -22,7 +22,9 @@
>>>>   bool s390_pci_kvm_interp_allowed(void)
>>>>   {
>>>> -    return kvm_s390_get_zpci_op() && !s390_is_pv();
>>>> +    return (kvm_s390_get_zpci_op() && !s390_is_pv() &&
>>>> +            !object_property_get_bool(OBJECT(qdev_get_machine()),
>>>> +                                      "zpcii-disable", NULL));
>>>>   }
>>>
>>> Isn't it a duplication of machine_get_zpcii_disable?
>>>
>>
>> No, this will actually trigger machine_get_zpcii_disable -- it was 
>> setup as the 'getter' routine in s390_machine_initfn() -- see below:
> 
> OK, I did not explain myself correctly:
> I was curious why we do not use directly ms->zpci_disabled and use the 
> getter.
> 

To do so, we'd have to either call machine_get_zpcii_disable directly 
from here or duplicate the work machine_get_zpcii_disable does by 
casting the machine to S390CcwMachineState so we could look at 
ms->zpcii_disabled.  We can't call machine_get_zpcii_disable directly 
as-is, it's a static routine in s390-virtio-ccw.c -- making a 'getter' 
routine public seems wrong, so we are left with recreating the cast and 
looking at ms->zpcii_disabled here; but as far as I can figure the point 
is to have a unified interface for querying a machine property value via 
object_property_get_*().  Why wouldn't we use that interface?

FWIW, I modeled this after the way we today handle aes-key-wrap in 
target/s390x/kvm/kvm.c and loadparm in hw/s390x/ipl.c (albeit we use 
object_property_get_str for the latter since it's a different property 
type).


> Does not mean it is false. Far from.
>  >>
>>> Wouldn't it better go to hw/s390x/kvm/ ?
>>>
>>> There get the MachineState *ms = MACHINE(qdev_get_machine()) and 
>>> return the ms->zpcii_disable
>>>
>>> ?
>>>
>>>>   int s390_pci_kvm_aif_enable(S390PCIBusDevice *pbdev, ZpciFib *fib, 
>>>> bool assist)
>>>> diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
>>>> index cc3097bfee..70229b102b 100644
>>>> --- a/hw/s390x/s390-virtio-ccw.c
>>>> +++ b/hw/s390x/s390-virtio-ccw.c
>>>> @@ -645,6 +645,21 @@ static inline void 
>>>> machine_set_dea_key_wrap(Object *obj, bool value,
>>>>       ms->dea_key_wrap = value;
>>>>   }
>>>> +static inline bool machine_get_zpcii_disable(Object *obj, Error 
>>>> **errp)
>>>> +{
>>>> +    S390CcwMachineState *ms = S390_CCW_MACHINE(obj);
>>>> +
>>>> +    return ms->zpcii_disable;
>>>> +}
>>>> +
>>>> +static inline void machine_set_zpcii_disable(Object *obj, bool value,
>>>> +                                             Error **errp)
>>>> +{
>>>> +    S390CcwMachineState *ms = S390_CCW_MACHINE(obj);
>>>> +
>>>> +    ms->zpcii_disable = value;
>>>> +}
>>>> +
>>>>   static S390CcwMachineClass *current_mc;
>>>>   /*
>>>> @@ -740,6 +755,13 @@ static inline void s390_machine_initfn(Object 
>>>> *obj)
>>>>               "Up to 8 chars in set of [A-Za-z0-9. ] (lower case 
>>>> chars converted"
>>>>               " to upper case) to pass to machine loader, boot 
>>>> manager,"
>>>>               " and guest kernel");
>>>> +
>>>> +    object_property_add_bool(obj, "zpcii-disable",
>>>> +                             machine_get_zpcii_disable,
>>
>> ^^ Here.
>