Re: [PATCH] PCI/pwrctl: pwrseq: abandon probe on pre-pwrseq device-trees

[Bjorn Andersson]

On Fri, Oct 04, 2024 at 07:59:41PM GMT, Bartosz Golaszewski wrote:
> On Fri, Oct 4, 2024 at 7:31 PM Bjorn Andersson <andersson@xxxxxxxxxx> wrote:
> >
> > >
> > > +     /*
> > > +      * Old device trees for some platforms already define wifi nodes for
> > > +      * the WCN family of chips since before power sequencing was added
> > > +      * upstream.
> > > +      *
> > > +      * These nodes don't consume the regulator outputs from the PMU and
> > > +      * if we allow this driver to bind to one of such "incomplete" nodes,
> > > +      * we'll see a kernel log error about the indefinite probe deferral.
> > > +      *
> > > +      * Let's check the existence of the regulator supply that exists on all
> > > +      * WCN models before moving forward.
> > > +      *
> > > +      * NOTE: If this driver is ever used to support a device other than
> > > +      * a WCN chip, the following lines should become conditional and depend
> > > +      * on the compatible string.
> >
> > What do you mean "is ever used ... other than WCN chip"?
> >
> 
> This driver was released as part of v6.11 and so far (until v6.12) is
> only used to support the WCN chips. That's not to say that it cannot
> be extended to support more hardware. I don't know how to put it in
> simpler words.
> 
> > This driver and the power sequence framework was presented as a
> > completely generic solution to solve all kinds of PCI power sequence
> > problems - upon which the WCN case was built.
> >
> 
> I never presented anything as "completely generic". You demanded that
> I make it into a miraculous catch-all solution.

That is correct. I strongly requested that you would come up with a
solution that worked for BOTH (all two!) use cases we had on the table
for PCI power sequencing.

> I argued that there's no such thing and this kind of attitude is
> precisely why it's so hard to get anything done in the kernel.

I'm sorry that you feel it's my attitude that's the problem here. I
don't think that is what make this hard, but rather the technical
challenges of the problem itself.

Regards,
Bjorn