Re: [PATCH] drivers/irqchip: Remove function callback casts

[Marc Zyngier]

On 2020-05-24 17:06, Oscar Carter wrote:
> Hi Marc,
>
> On Sun, May 24, 2020 at 12:46:34PM +0100, Marc Zyngier wrote:
> > On Sun, 24 May 2020 10:09:10 +0200
> > Oscar Carter <oscar.carter@xxxxxxx> wrote:
> >
> > Hi Oscar,
> >
> > Thanks for this. Comments below.
> >
> > > In an effort to enable -Wcast-function-type in the top-level Makefile to
> > > support Control Flow Integrity builds, remove all the function callback
> > > casts.
> > >
> > > To do this, modify the IRQCHIP_ACPI_DECLARE macro initializing the
> > > acpi_probe_entry struct directly instead of use the existent macro
> > > ACPI_DECLARE_PROBE_ENTRY.
> > >
> > > In this new initialization use the probe_subtbl field instead of the
> > > probe_table field use in the ACPI_DECLARE_PROBE_ENTRY macro.
> >
> > Please add *why* this is a valid transformation (probe_table and
> > probe_subtbl are part of a union).
>
> Ok, I will add a more detailed explanation.
>
> > > Signed-off-by: Oscar Carter <oscar.carter@xxxxxxx>
> > > ---
> > >  include/linux/irqchip.h | 10 ++++++++--
> > >  1 file changed, 8 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/include/linux/irqchip.h b/include/linux/irqchip.h
> > > index 950e4b2458f0..1f464fd10df0 100644
> > > --- a/include/linux/irqchip.h
> > > +++ b/include/linux/irqchip.h
> > > @@ -39,8 +39,14 @@
> > >   * @fn: initialization function
> > >   */
> > >  #define IRQCHIP_ACPI_DECLARE(name, subtable, validate, data, fn)	\
> > > -	ACPI_DECLARE_PROBE_ENTRY(irqchip, name, ACPI_SIG_MADT, 		\
> > > -				 subtable, validate, data, fn)
> > > +	static const struct acpi_probe_entry __acpi_probe_##name	\
> > > +		__used __section(__irqchip_acpi_probe_table) = {	\
> > > +			.id = ACPI_SIG_MADT,				\
> > > +			.type = subtable,				\
> > > +			.subtable_valid = validate,			\
> > > +			.probe_subtbl = (acpi_tbl_entry_handler)fn,	\
> > > +			.driver_data = data,				\
> > > +		}
> > >
> >
> > I'd rather you add an ACPI_DECLARE_SUBTABLE_PROBE_ENTRY to acpi.h, and
> > use that here so that we can keep the ACPI gunk in a single place.
>
> Ok, I will do the changes you suggested and I will resend a new 
> version.
> Later, I will also send a series to clean up the checkpatch warnings 
> and
> errors for the acpi.h header.

Not necessarily a good idea. Churn for the sake of keeping checkpatch
at bay is pretty pointless. Do fix bugs if you spot any, but please
exercise judgement.

Thanks,

        M.
--
Jazz is not dead. It just smells funny...