On Thu, Aug 24, 2023 at 01:51:59PM -0700, Doug Anderson wrote:
> On Thu, Aug 24, 2023 at 11:26 AM Laurent Pinchart wrote:
> > On Thu, Aug 24, 2023 at 07:15:46PM +0100, Biju Das wrote:
> > > The driver has an ID table, but it uses the wrong API for retrieving match
> > > data and that will lead to a crash, if it is instantiated by user space or
> > > using ID. From this, there is no user for the ID table and let's drop it
> > > from the driver as it saves some memory.
> > >
> > > Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> >
> > Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@xxxxxxxxxxxxxxxx>
> >
> > I wonder, as the device can only be instantiated from OF, should we add
> >
> > depends on OF
> >
> > to Kconfig, and drop the
> >
> > #if IS_ENABLED(CONFIG_OF)
> >
> > from the driver ?
>
> In my opinion we shouldn't add the "depends on OF" since that will
> decrease the amount of compile testing. It's somewhat the opposite of
> adding "if COMPILE_TEST" to your driver. ;-)
We could add a || COMPILE_TEST :-)
> I think we could get rid of one of the "#if" statements in the driver
> anyway as of commit c9e358dfc4a8 ("driver-core: remove conditionals
> around devicetree pointers") from ~12 years ago. If we did something
> similar in "struct drm_bridge" we could drop both #ifs.
I'd be fine with that too.
--
Regards,
Laurent Pinchart
