[PATCH 1/3] dt-bindings: reset: mediatek: add MT7629 TOPRGU reset indices

[Chuanhong Guo]

Hi!

On Tue, Sep 29, 2026 at 4:44 PM Krzysztof Kozlowski <krzk at kernel.org> wrote:
>
> On Sat, Sep 26, 2026 at 07:13:59PM +0800, Chuanhong Guo wrote:
> > Define all seven reset IDs documented in the MT7629 TOPRGU SWSYSRST
> > register: INFRA, ETHDMA, DDRPHY, INFRA_AO, CONN, APMIXED and CONN_MCU.
> > Use their register bit positions as reset IDs.
>
> If so, I say you don't need bindings then... but fine except:
>
> >
> > Signed-off-by: Chuanhong Guo <gch981213 at gmail.com>
> > ---
> >  include/dt-bindings/reset/mt7629-resets.h | 9 +++++++++
> >  1 file changed, 9 insertions(+)
> >
> > diff --git a/include/dt-bindings/reset/mt7629-resets.h b/include/dt-bindings/reset/mt7629-resets.h
> > index 6bb85734f68d..828919eee298 100644
> > --- a/include/dt-bindings/reset/mt7629-resets.h
> > +++ b/include/dt-bindings/reset/mt7629-resets.h
> > @@ -53,6 +53,15 @@
> >  #define MT7629_PCIE0_USER_RST                        30
> >  #define MT7629_PCIE0_PIPE_RST                        31
> >
> > +/* TOPRGU resets */
>
> Didn't new bindings switch to continues indexes?
>

This is written for the existing reset driver in mtk_wdt.c. I don't
want to add a reset id remap when all the other SoCs in that
driver use binding ids matching their hardware bits.

Is this OK for existing drivers?
I can also live with plain numbers in the device tree if that's preferred.

Another option would be to include mt7622-resets.h for mt7629 dt
as their TOPRGU definitions match exactly. But that would bring other
unused mt7622 definitions into mt7629 dt, probably not ideal.

> > +#define MT7629_TOPRGU_INFRA_SW_RST           0
> > +#define MT7629_TOPRGU_ETHDMA_SW_RST          1
> > +#define MT7629_TOPRGU_DDRPHY_SW_RST          6
> > +#define MT7629_TOPRGU_INFRA_AO_SW_RST                8
>
> Best regards,
> Krzysztof
>

--
Regards,
Chuanhong Guo