[PATCH] of: support passing console options with stdout-path

Grant Likely grant.likely at linaro.org
Tue Nov 25 07:39:49 PST 2014


On Tue, Nov 25, 2014 at 3:24 PM, Ian Campbell <ijc at debian.org> wrote:
> On Tue, 2014-11-25 at 15:20 +0000, Grant Likely wrote:
>> On Tue, Nov 25, 2014 at 3:15 PM, Leif Lindholm <leif.lindholm at linaro.org> wrote:
>> > On Tue, Nov 25, 2014 at 02:58:54PM +0000, Grant Likely wrote:
>> >> > +           len = strchrnul(path, ':') - path;
>> >> > +
>> >> >             for_each_property_of_node(of_aliases, pp) {
>> >> >                     if (strlen(pp->name) == len && !strncmp(pp->name, path, len)) {
>> >> >                             np = of_find_node_by_path(pp->value);
>> >>
>> >> Can you add a testcase to drivers/of/unittests.c for this new path
>> >> parsing? Then we know it's correct!
>> >
>> > Will do.
>> >
>> >> > @@ -1830,8 +1835,12 @@ void of_alias_scan(void * (*dt_alloc)(u64 size, u64 align))
>> >> >                     name = of_get_property(of_chosen, "linux,stdout-path", NULL);
>> >> >             if (IS_ENABLED(CONFIG_PPC) && !name)
>> >> >                     name = of_get_property(of_aliases, "stdout", NULL);
>> >> > -           if (name)
>> >> > +           if (name) {
>> >> >                     of_stdout = of_find_node_by_path(name);
>> >> > +                   of_stdout_options = strchr(name, ':');
>> >> > +                   if (of_stdout_options != NULL)
>> >> > +                           of_stdout_options++;
>> >> > +           }
>> >>
>> >> Not so good. The ':' should actually be a generic part of
>> >> of_find_node_by_path(), it doesn't have to be limited to stdout parsing.
>> >> There are other places that would use it. I would add a second,
>> >> optional, argument to of_find_node_by_path() that will update a pointer
>> >> to the arguments after the ':'.
>> >
>> > So, I agree this would be nicer...
>> >
>> > However, I'm counting 157 callers of this function outside of
>> > drivers/of and 43 inside. Any chance you'd let me get away with a
>> > separate of_find_extra_options_by_path()?
>>
>> I'd rather not. It is a trivial change and can be pretty much done
>> mechanically. Something like (I think; I always have to look up the
>> sed syntax):
>>
>> sed s/of_find_node_by_path\((.*)\)/of_find_node_by_path\(\1, NULL\)
>
> Coccinelle rules for this sort of transformation...

/me still hasn't gotten his head around how to use Coccinelle properly. :-(

g.



More information about the linux-arm-kernel mailing list