Hi Grant,
On Monday 17 February 2014 18:18:29 Grant Likely wrote:
> On Tue, 11 Feb 2014 13:36:51 +0100, Laurent Pinchart wrote:
> > when CONFIG_OF is disabled of_match_node is defined as a macro that
> > evaluates to NULL. This breaks compilation of drivers that dereference
> > the function's return value directly. Fix it by turning the macro into a
> > static inline function that returns NULL.
>
> Is it not more problematic that code is dereferencing the return value
> directly *without* checking to see if it is NULL first? I suppose it
> will cause a runtime oops on all platforms except no-mmu, but still.
> Compile time catches are better.
There should be no risk of oops if the of_match_ptr() call is guarded by a
compiler (as opposed to preprocessor) IS_ENABLED(CONFIG_OF) check. Now, if the
check is missing, we'll oops at runtime, but the problem already exists before
this patch.
Anyway I've decided to drop this patch and use an intermediate variable.
> > Signed-off-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> > ---
> >
> > include/linux/of.h | 8 +++++++-
> > 1 file changed, 7 insertions(+), 1 deletion(-)
> >
> > diff --git a/include/linux/of.h b/include/linux/of.h
> > index 70c64ba..719e69f 100644
> > --- a/include/linux/of.h
> > +++ b/include/linux/of.h
> > @@ -560,7 +560,13 @@ static inline const char *of_prop_next_string(struct
> > property *prop,>
> > }
> >
> > #define of_match_ptr(_ptr) NULL
> >
> > -#define of_match_node(_matches, _node) NULL
> > +
> > +static inline const struct of_device_id *of_match_node(
> > + const struct of_device_id *matches, const struct device_node *node)
> > +{
> > + return NULL;
> > +}
> > +
> >
> > #endif /* CONFIG_OF */
> >
> > #if defined(CONFIG_OF) && defined(CONFIG_NUMA)
--
Regards,
Laurent Pinchart
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
