This function may need to return a value

Signed-off-by: Li kunyu <kunyu@nfschina.com>
---
 arch/x86/kvm/mmu/mmu.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 64a2a7e2be90..68f33b932f94 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -6500,6 +6500,8 @@ static int kvm_nx_lpage_recovery_worker(struct kvm *kvm, uintptr_t data)
 
 		kvm_recover_nx_lpages(kvm);
 	}
+
+	return 0;
 }
 
 int kvm_mmu_post_init_vm(struct kvm *kvm)
-- 
2.18.2

On 05.05.22 13:32, Li kunyu wrote:
> This function may need to return a value
> 
> Signed-off-by: Li kunyu <kunyu@nfschina.com>
> ---
>   arch/x86/kvm/mmu/mmu.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 64a2a7e2be90..68f33b932f94 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -6500,6 +6500,8 @@ static int kvm_nx_lpage_recovery_worker(struct kvm *kvm, uintptr_t data)
>   
>   		kvm_recover_nx_lpages(kvm);
>   	}
> +
> +	return 0;

This statement is not reachable, so the patch is adding unneeded dead
code only.


Juergen

On 5.05.2022 13:36, Juergen Gross wrote:
> On 05.05.22 13:32, Li kunyu wrote:
>> This function may need to return a value
>>
>> Signed-off-by: Li kunyu <kunyu@nfschina.com>
>> ---
>>   arch/x86/kvm/mmu/mmu.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
>> index 64a2a7e2be90..68f33b932f94 100644
>> --- a/arch/x86/kvm/mmu/mmu.c
>> +++ b/arch/x86/kvm/mmu/mmu.c
>> @@ -6500,6 +6500,8 @@ static int kvm_nx_lpage_recovery_worker(struct kvm *kvm, uintptr_t data)
>>           kvm_recover_nx_lpages(kvm);
>>       }
>> +
>> +    return 0;
> 
> This statement is not reachable, so the patch is adding unneeded dead
> code only.

Maybe some static checker isn't smart enough to figure this out.

In this case it would probably be better to also change:
> if (kthread_should_stop())
>     return 0;

into:
> if (kthread_should_stop())
>     break;

so the newly introduced code isn't dead.

> Juergen

Thanks,
Maciej

Hello, senior, I've considered break before, but I'm not sure if I want to execute more instructions.   
Query break executed one more NOP instruction.

On 5/5/22 15:37, Maciej S. Szmigiero wrote:
>> This statement is not reachable, so the patch is adding unneeded dead
>> code only.
> 
> Maybe some static checker isn't smart enough to figure this out.

The static checker really should be improved.  This is a while(true), 
not the halting problem. :)

Paolo