Re: [PATCH 2/2] target/mips/kvm: Assert unreachable code is not used

[Philippe Mathieu-Daudé]

On 11/24/20 12:02 PM, Paolo Bonzini wrote:
> On 24/11/20 11:41, Philippe Mathieu-Daudé wrote:
>> Huacai, ping?
>>
>> On 5/12/20 9:09 AM, Philippe Mathieu-Daudé wrote:
>>> +Paolo
>>>
>>> On 4/29/20 10:29 AM, Philippe Mathieu-Daudé wrote:
>>>> This code must not be used outside of KVM. Abort if it is.
>>>>
>>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>>> ---
>>>>    target/mips/kvm.c | 8 ++------
>>>>    1 file changed, 2 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/target/mips/kvm.c b/target/mips/kvm.c
>>>> index de3e26ef1f..050bfbd7fa 100644
>>>> --- a/target/mips/kvm.c
>>>> +++ b/target/mips/kvm.c
>>>> @@ -196,9 +196,7 @@ int kvm_mips_set_interrupt(MIPSCPU *cpu, int irq,
>>>> int level)
>>>>        CPUState *cs = CPU(cpu);
>>>>        struct kvm_mips_interrupt intr;
>>>>    -    if (!kvm_enabled()) {
>>>> -        return 0;
>>>> -    }
>>>> +    assert(kvm_enabled());
>>>>          intr.cpu = -1;
>>>>    @@ -219,9 +217,7 @@ int kvm_mips_set_ipi_interrupt(MIPSCPU *cpu, int
>>>> irq, int level)
>>>>        CPUState *dest_cs = CPU(cpu);
>>>>        struct kvm_mips_interrupt intr;
>>>>    -    if (!kvm_enabled()) {
>>>> -        return 0;
>>>> -    }
>>>> +    assert(kvm_enabled());
>>>>          intr.cpu = dest_cs->cpu_index;
>>>>   
>>>
>>
> 
> Acked-by: Paolo Bonzini <pbonzini@redhat.com>
> 
> For kvm_mips_set_ipi_interrupt, however, it would be nicer if
> hw/intc/mips_gic.c always used gic->vps[vp].env->irq[], and the qemu_irq
> handler took care of calling kvm_mips_set_ipi_interrupt.
> 
> Likewise, there is some duplication between kvm_mips_interrupt's caller
> and kvm_arch_pre_run.  I'm not sure if kvm_arch_pre_run is needed at all.

OK, I'll have a look and ask Huacai to test.

Meanwhile, patch applied to mips-next.

Thanks,

Phil.