Re: [PATCH v2 1/2] mfd: ucb1x00: Register software node for GPIO controller

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Thu, Jul 16, 2026 at 03:00:27PM +0100, Lee Jones wrote:
> On Thu, 16 Jul 2026, Arnd Bergmann wrote:
> 
> > On Thu, Jul 16, 2026, at 14:40, Lee Jones wrote:
> > > On Tue, 07 Jul 2026, Arnd Bergmann wrote:
> > >
> > >> On Tue, Jul 7, 2026, at 07:01, Dmitry Torokhov wrote:
> > >> > Define a static software node for the UCB1x00 GPIO controller and attach
> > >> > it to the core MFD device in ucb1x00_probe(). This node will also be
> > >> > used by the created GPIO chip.
> > >> >
> > >> > This allows machine subdrivers (such as Assabet evaluation board
> > >> > support) to reference the UCB1x00 GPIO controller in property entries
> > >> > when converting legacy platform data to software nodes, resolving pin
> > >> > bindings directly via the attached firmware node without relying on
> > >> > name matching.
> > >> >
> > >> > Assisted-by: Antigravity:gemini-3.5-flash
> > >> > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> > >> 
> > >> Acked-by: Arnd Bergmann <arnd@arndb.de>
> > >> 
> > >> Assabet should be going away soon, so it really doesn't
> > >> matter much.
> > 
> > >> > +EXPORT_SYMBOL_GPL(ucb1x00_gpiochip_node);
> > >> 
> > >> It does feel a bit counterproductive if you have to add more
> > >> exported symbols.

Just FTR: I completely agree. However Bartosz is dead set on not
supporting weaker references based on name matching and wants only
identity matching. So we export bunch of stuff and pay for this with 
tighter coupling and potential changes in load order.

> >  
> > >> No need for the seperate declaration of the struct tag,
> > >> you only need this if the first user is in the argument
> > >> list of a function, but not for a global variable.
> > >
> > > Are you planning on fixing these review comments, Dmitry?
> > 
> > Just to clarify, I was not asking for anything to be changed
> > in this patch, just observing. I'm fine with this version
> > getting applied as-is. As I said, the file is probably
> > going away soon, so let's just make sure it's not getting
> > in the way until then.
> 
> The 2 comments above sure look like review comments.
> 

Well, if this all truly goes away in 7.4 there is no much point in
polishing this.

Thanks.

-- 
Dmitry