[PATCH 1/2] drivers: serial: earlycon: Pass device-tree node

Andy Shevchenko andy.shevchenko at gmail.com
Fri Jul 29 10:47:35 UTC 2022


On Fri, Jul 29, 2022 at 9:57 AM Greg Kroah-Hartman
<gregkh at linuxfoundation.org> wrote:
> On Thu, Jul 28, 2022 at 11:04:24PM +0200, Andy Shevchenko wrote:
> > On Thu, Jul 28, 2022 at 4:41 PM Greg Kroah-Hartman
> > <gregkh at linuxfoundation.org> wrote:
> > > On Thu, Jul 28, 2022 at 05:28:18PM +0300, Markuss Broks wrote:

...

> > > > +     unsigned long node;
> > >
> > > That should not be an unsigned long, but rather an 'int'.  Something got
> > > messed up, of_setup_earlycon() should be changed to reflect this before
> > > propagating the error to other places in the kernel.
> >
> > It's a pointer, but what puzzles me, why it can't be declared as a such:
> >
> >  struct device_node *node;
> >
> > ?
>
> It should not be a pointer, trace things backwards, it comes from a call
> to of_setup_earlycon() from early_init_dt_scan_chosen_stdout() which has
> offset declared as an int, and then does:
>         if (of_setup_earlycon(match, offset, options) == 0)
>
> So why would it be a node?

This is a very good question.

> > > And it's not really a "node" but an "offset", right?
> >
> > Seems no.
>
> Really?  What am I missing here?

It's me who is missing something here, thanks for your elaboration!
After it it becomes clear that your first question should be
addressed.

-- 
With Best Regards,
Andy Shevchenko


More information about the dri-devel mailing list