[PATCH] drm/msm/dp: do not initialize phy until plugin interrupt received

[Stephen Boyd]

Quoting Bjorn Andersson (2021-11-05 10:48:37)
> On Fri 05 Nov 10:28 PDT 2021, Kuogee Hsieh wrote:
>
> > From: Kuogee Hsieh <khsieh at codeaurora.org>
> >
> > Combo phy supports both USB and DP simultaneously. There may has a
> > possible conflict during phy initialization phase between USB and
> > DP driver which may cause USB phy timeout when USB tries to power
> > up its phy. This patch has the DP driver not initialize its phy
> > during DP driver initialization phase. Instead DP driver only enable
> > required regulators and clocks so that it is able to receive HPD
> > interrupts after completion of initialization phase. DP driver will
> > initialize its phy when HPD plug-in interrupt received.
>
> Is this a hardware requirement, or is this a issue with the current
> implementation of the QMP combo phy driver? We should not hack up the DP
> driver to circumvent the latter.
>
> Also, I don't suppose there's anything here that prevents the HPD to
> come before the USB PHY is powered up? Even though that seems less
> likely in practice...
>
> > This patch also provides a positive side effects which balance regulator
> > enable count since regulator only enabled at initialize phase and resume
> > and disabled at followed suspend.
> >
>
> Is this something that needs to be fixed separately, so that it can be
> backported to stable kernels?

Agreed. Please split the regulator balance problem from the DP
initialization delay.