On 07/08/2019 14:39, Steven Price wrote:
> On 03/08/2019 18:34, Marc Zyngier wrote:
>> On Sat, 3 Aug 2019 13:51:13 +0100
>> Marc Zyngier <maz@xxxxxxxxxx> wrote:
>>
>> [forgot that one]
>>
>>> On Fri, 2 Aug 2019 15:50:14 +0100
>>> Steven Price <steven.price@xxxxxxx> wrote:
>>
>> [...]
>>
>>>> +static int __init kvm_pvtime_init(void)
>>>> +{
>>>> + kvm_register_device_ops(&pvtime_ops, KVM_DEV_TYPE_ARM_PV_TIME);
>>>> +
>>>> + return 0;
>>>> +}
>>>> +
>>>> +late_initcall(kvm_pvtime_init);
>>
>> Why is it an initcall? So far, the only initcall we've used is the one
>> that initializes KVM itself. Can't we just the device_ops just like we
>> do for the vgic?
>
> So would you prefer a direct call from init_subsystems() in
> virt/kvm/arm/arm.c?
Yes. Consistency is important.
> The benefit of initcall is just that it keeps the code self-contained.
> In init_subsystems() I'd either need a #ifdef CONFIG_ARM64 or a dummy
> function for arm.
Having a dummy function for 32bit ARM is fine. Most of the code we add
to the 32bit port is made of empty stubs anyway.
Thanks,
M.
--
Jazz is not dead, it just smells funny...
