[1/3] arm64: efi: Simplify arch_efi_call_virt macro by not using efi_##f##_t type

Currently, the arch_efi_call_virt() assumes all users of it will have
defined a type 'efi_##f##_t' to make use of it. It is unnecessarily
forcing the users to create a new typedef when __efi_rt_asm_wrapper()
actually expects void pointer.

Simplify the arch_efi_call_virt() macro by typecasting p->f to (void *)
as required by __efi_rt_asm_wrapper() and eliminate the explicit need
for efi_##f##_t type for every user of this macro.

This is needed now in preparation to enable PRMT support on ARM64.

Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 arch/arm64/include/asm/efi.h | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

Hi Ard,

I am not sure if you prefer to add type for each users of this or this
is acceptable. I see only compile time advantage but I am not sure if
it make sense to add typedefs in ACPI PRMT driver just for this reason.

Let me know.

Regards,
Sudeep

On Fri, 24 Jun 2022 at 17:23, Sudeep Holla <sudeep.holla@arm.com> wrote:
>
> Currently, the arch_efi_call_virt() assumes all users of it will have
> defined a type 'efi_##f##_t' to make use of it. It is unnecessarily
> forcing the users to create a new typedef when __efi_rt_asm_wrapper()
> actually expects void pointer.
>
> Simplify the arch_efi_call_virt() macro by typecasting p->f to (void *)
> as required by __efi_rt_asm_wrapper() and eliminate the explicit need
> for efi_##f##_t type for every user of this macro.
>

Can't we just use typeof() here?

__efi_rt_asm_wrapper() was intended as a temporary thing, so I'd
prefer to avoid starting to rely on the void* type of its first
argument.


> This is needed now in preparation to enable PRMT support on ARM64.
>
> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
> ---
>  arch/arm64/include/asm/efi.h | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> Hi Ard,
>
> I am not sure if you prefer to add type for each users of this or this
> is acceptable. I see only compile time advantage but I am not sure if
> it make sense to add typedefs in ACPI PRMT driver just for this reason.
>
> Let me know.
>
> Regards,
> Sudeep
>
> diff --git a/arch/arm64/include/asm/efi.h b/arch/arm64/include/asm/efi.h
> index ad55079abe47..263d7fd67207 100644
> --- a/arch/arm64/include/asm/efi.h
> +++ b/arch/arm64/include/asm/efi.h
> @@ -29,9 +29,7 @@ int efi_set_mapping_permissions(struct mm_struct *mm, efi_memory_desc_t *md);
>
>  #define arch_efi_call_virt(p, f, args...)                              \
>  ({                                                                     \
> -       efi_##f##_t *__f;                                               \
> -       __f = p->f;                                                     \
> -       __efi_rt_asm_wrapper(__f, #f, args);                            \
> +       __efi_rt_asm_wrapper((void *)p->f, #f, args);                   \
>  })
>
>  #define arch_efi_call_virt_teardown()                                  \
> --
> 2.36.1
>

On Fri, Jun 24, 2022 at 07:45:14PM +0200, Ard Biesheuvel wrote:
> On Fri, 24 Jun 2022 at 17:23, Sudeep Holla <sudeep.holla@arm.com> wrote:
> >
> > Currently, the arch_efi_call_virt() assumes all users of it will have
> > defined a type 'efi_##f##_t' to make use of it. It is unnecessarily
> > forcing the users to create a new typedef when __efi_rt_asm_wrapper()
> > actually expects void pointer.
> >
> > Simplify the arch_efi_call_virt() macro by typecasting p->f to (void *)
> > as required by __efi_rt_asm_wrapper() and eliminate the explicit need
> > for efi_##f##_t type for every user of this macro.
> >
>
> Can't we just use typeof() here?

I had tried that, but unless p->f is pointer of some type, we will get
the warning as it is passed without a cast to __efi_rt_asm_wrapper().

> __efi_rt_asm_wrapper() was intended as a temporary thing, so I'd
> prefer to avoid starting to rely on the void* type of its first
> argument.
>

Fair enough. Can we expect p->f to be some pointer then ? If yes, then
PRMT driver needs to change the handler_addr from u64 to some pointer
which sounds OK to me.

--
Regards,
Sudeep

On Sat, 25 Jun 2022 at 12:11, Sudeep Holla <sudeep.holla@arm.com> wrote:
>
> On Fri, Jun 24, 2022 at 07:45:14PM +0200, Ard Biesheuvel wrote:
> > On Fri, 24 Jun 2022 at 17:23, Sudeep Holla <sudeep.holla@arm.com> wrote:
> > >
> > > Currently, the arch_efi_call_virt() assumes all users of it will have
> > > defined a type 'efi_##f##_t' to make use of it. It is unnecessarily
> > > forcing the users to create a new typedef when __efi_rt_asm_wrapper()
> > > actually expects void pointer.
> > >
> > > Simplify the arch_efi_call_virt() macro by typecasting p->f to (void *)
> > > as required by __efi_rt_asm_wrapper() and eliminate the explicit need
> > > for efi_##f##_t type for every user of this macro.
> > >
> >
> > Can't we just use typeof() here?
>
> I had tried that, but unless p->f is pointer of some type, we will get
> the warning as it is passed without a cast to __efi_rt_asm_wrapper().
>
> > __efi_rt_asm_wrapper() was intended as a temporary thing, so I'd
> > prefer to avoid starting to rely on the void* type of its first
> > argument.
> >
>
> Fair enough. Can we expect p->f to be some pointer then ? If yes, then
> PRMT driver needs to change the handler_addr from u64 to some pointer
> which sounds OK to me.
>

We are dealing with function pointers here, so passing those as u64 is
just sloppy.