On Wed, Feb 23, 2022, Nathan Chancellor wrote:
> On Wed, Feb 23, 2022 at 05:59:05PM +0000, Sean Christopherson wrote:
> > On Wed, Feb 23, 2022, Nathan Chancellor wrote:
> > > Hi Sean,
> > >
> > > On Wed, Feb 23, 2022 at 04:23:55PM +0000, Sean Christopherson wrote:
> > > > Cast kvm_x86_ops.func to 'void *' when updating KVM static calls that are
> > > > conditionally patched to __static_call_return0(). clang complains about
> > > > using mismatching pointers in the ternary operator, which breaks the
> > > > build when compiling with CONFIG_KVM_WERROR=y.
> > > >
> > > > >> arch/x86/include/asm/kvm-x86-ops.h:82:1: warning: pointer type mismatch
> > > > ('bool (*)(struct kvm_vcpu *)' and 'void *') [-Wpointer-type-mismatch]
> > > >
> > > > Fixes: 5be2226f417d ("KVM: x86: allow defining return-0 static calls")
> > > > Reported-by: Like Xu <like.xu.linux@xxxxxxxxx>
> > > > Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > > > Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx>
> > >
> > > Thank you for the patch! Is this a bug in clang?
> >
> > IMO, no. I think it's completely reasonable for the compiler to complain that KVM
> > is generating two different pointer types out of a ternary operator.
> >
> > clang is somewhat inconsistent, though it may be deliberate. clang doesn't complain
> > about implicitly casting a 'void *' to another data type, e.g. this complies clean,
> > where "data" is a 'void *'
> >
> > struct kvm_vcpu *x = vcpu ? : data;
>
> Right, I would assume this is deliberate. I think warning in this case
> might be quite noisy, as the kernel implicitly converts 'void *' to
> typed pointers for certain function pointer callbacks (although this
> particular case is probably pretty rare).
Aha! Looks like clang's behavior is correct, assuming a function is not considered
an "object". From C99 "6.5.15 Conditional operator":
One of the following shall hold for the second and third operands:
— both operands have arithmetic type;
— both operands have the same structure or union type;
— both operands have void type;
— both operands are pointers to qualified or unqualified versions of compatible types;
— one operand is a pointer and the other is a null pointer constant; or
— one operand is a pointer to an object or incomplete type and the other is a pointer to a
qualified or unqualified version of void.
That last case would explain why clang warns about a function pointer but not a
object pointer when the third operand is a 'void *'.
