[PATCH v7] wdt: nuvoton: Add support for Nuvoton

[劉家君]

Hi  Michael

thanks for your reply.
the wdt design of npcm750 and npcm845 is the same.
so the driver can work on npcm750 and npcm845.
about the npcm845 wdt dtsi , i will followed kernel upstream name and
use nuvoton,npcm750-wdt.

i think the v6 version is better for now.
should i use v6 patch to make new v8 version?

Your comments are welcome.


Michael Walle <michael at walle.cc> 於 2022年3月22日 週二 下午8:03寫道：

> Am 2022-03-22 11:54, schrieb Andre Przywara:
> > On Tue, 22 Mar 2022 17:19:53 +0800
> > Jim Liu <jim.t90615 at gmail.com> wrote:
> >
> > Hi,
> >
> >> Add watchdog controller driver for NPCM7xx/npcm8xx
> >
> > So this is now the same as v4, with both compatible strings listed.
> > Which *might* be right, but my question still stands:
> >
> > Is the 845 watchdog compatible to the 740 version? Does the current
> > Linux driver work?
> >
> > If yes (and your driver suggests so, because you treat the two
> > compatible strings the same), then you would not need to list both *in
> > the driver*, but just do so in the .dts file.
> >
> > So the driver just matches "nuvoton,npcm750-wdt" (as does the Linux
> > driver), the nuvoton-common-npcm7xx.dtsi stays the same, but in the
> > npcm845.dtsi you write:
> >       watchdog: watchdog at xxx {
> >               compatible = "nuvoton,npcm845-wdt",
> >                            "nuvoton,npcm750-wdt";
> >               interrupts = ...
> >               ...
> >       };
> >
> > This way you don't need to change the (Linux) driver, and get away with
> > one compatible string, for both devices.
>
> Ahh didn't noticed that there are different compatible strings
> in the linux kernel driver.
>
> And even
> >               compatible = "nuvoton,npcm845-wdt",
> >                            "nuvoton,npcm750-wdt";
> would be undocumented if it is not upstreamed to linux.
>
> > So can you please confirm that both are compatible. I couldn't find any
> > information on that.
> >
> > Please just don't send another patch without answering the question ;-)
>
> I'd suggest to convert the binding to yaml format so it is unambiguous
> what is supported and what isn't.
>
> -michael
>