+Peter, Marc

On Mon, Nov 22, 2021, kernel test robot wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core
> head:   a9f4a6e92b3b319296fb078da2615f618f6cd80c
> commit: e1bfc24577cc65c95dc519d7621a9c985b97e567 [12/17] KVM: Move x86's perf guest info callbacks to generic KVM
> All warnings (new ones prefixed by >>):
> 
> >> arch/arm64/kvm/arm.c:499:15: warning: no previous prototype for function 'kvm_arch_vcpu_get_ip' [-Wmissing-prototypes]
>    unsigned long kvm_arch_vcpu_get_ip(struct kvm_vcpu *vcpu)
>                  ^
>    arch/arm64/kvm/arm.c:499:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
>    unsigned long kvm_arch_vcpu_get_ip(struct kvm_vcpu *vcpu)
>    ^
>    static 
>    1 warning generated.

Argh, the declaration is behind CONFIG_GUEST_PERF_EVENTS=y so that architectures
that don't support GUEST_PERF_EVENTS at all didn't need to provide implementations.

There are a few ways to solve this, none of which I particularly love.  I think my
first choice would be to wrap arm's implementation with an #ifdef.  Stubbing out a
getter like this feels all kinds of wrong, and moving the declaration outside of
CONFIG_GUEST_PERF_EVENTS=y will set someone up to try and use the helper when it's
not available.

---
 arch/arm64/kvm/arm.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index 8129ee1ed3a4..e3b46951a92d 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
@@ -496,10 +496,12 @@ bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu)
 	return vcpu_mode_priv(vcpu);
 }

+#ifdef CONFIG_GUEST_PERF_EVENTS
 unsigned long kvm_arch_vcpu_get_ip(struct kvm_vcpu *vcpu)
 {
 	return *vcpu_pc(vcpu);
 }
+#endif

 /* Just ensure a guest exit from a particular CPU */
 static void exit_vm_noop(void *info)
--