On Mon, Feb 5, 2018 at 4:10 PM, Philippe Mathieu-Daudé <f4...@amsat.org> wrote:
> Hi Christoffer,
>
> On 02/01/2018 05:53 PM, Christoffer Dall wrote:
>> KVM doesn't support emulating a GICv3 in userspace, only GICv2. We
>> currently attempt this anyway, and as a result a KVM guest doesn't
>> receive interrupts and the user is left wondering why. Report an error
>> to the user if this particular combination is requested.
>>
>> Signed-off-by: Christoffer Dall <christoffer.d...@linaro.org>
>> ---
>> target/arm/kvm_arm.h | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h
>> index ff53e9fafb..cfb7e5af72 100644
>> --- a/target/arm/kvm_arm.h
>> +++ b/target/arm/kvm_arm.h
>> @@ -234,6 +234,10 @@ static inline const char *gicv3_class_name(void)
>
> Any reason why these functions are inlined in this include?
>
> Also - not related to your patch - while gicv3_class_name() is
> self-explicit, gic_class_name() isn't (to me at least).
>
> There are many check for gic_version 1/2/3 in virt.c, which we could
> clean using more generic functions such gic_class_name(int gic_version)
> and Co.
>
> Alistair/Edgar:
> Why the ZynqMP is not using the GICv3 for the APU? I remember hearing
> something about not fully implemented so not working correctly.
The ZynqMP only has a GICv2.
Maybe you are thinking of the virtualisation features? Mainline
doesn't have those fully implemented so they don't work correctly.
Alistair
>
>> exit(1);
>> #endif
>> } else {
>> + if (kvm_enabled()) {
>> + error_report("Userspace GICv3 is not supported with KVM");
>> + exit(1);
>
> Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org>
>
>> + }
>> return "arm-gicv3";
>> }
>> }
>>
>
> Regards,
>
> Phil.
>
