Re: [PATCH] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Joonhoe Kim <26rote@gmail.com>]

On 10/22/26 9:14 PM, Val Packett wrote:
> Good catch. I don't think I've seen system hangs from this on Hamoa, but
> I've had USB stop working on the ports where 4xDP+USB2 was left
> connected during suspend..

That may be the same path: on resume the USB PHY init programs the USB
tables over the DisplayPort configuration and times out. On the Y700
the xHCI resume that followed hung; a USB port that stays dead
afterwards would fit as well. A "phy initialization timed-out" from the
QMP PHY on resume would confirm it.

> The USB4 chunk above (juuuust before the start of this diff's context)
> does a similar thing in a different style: it increments both init_count
> and usb_init_count and early-returns, skipping over qmp_combo_com_init
> as well.
>
> I think both USB4 and DP should be written similarly but should
> qmp_combo_com_init be skipped here? No in both cases or does USB4 skip
> it for a good reason?
>
> (I'm especially wondering about the impact of that on the USB HS PHY
> clock, since it's actually sourced from here without a properly modeled
> dependency right now…)

DP-only needs it, and USB4 does not really skip it: the bare
init_count++ is exactly what qmp_combo_com_init(qmp, false) does when
the count is already non-zero, and qmp_combo_usb4_init() holds that
first reference. In DP-only mode nothing guarantees another holder: on
resume the USB controller can init its PHY before DisplayPort does, and
then qmp_combo_com_init() is what turns the clocks, resets and
regulators on.

So I kept the two branches different: the USB4 one is a separate early
return because it must not reprogram the PHY at all, while DP-only only
leaves the USB PCS alone. The mux path already keeps the USB PCS
powered down in DP-only mode; v2 makes usb_init/usb_exit match it, and
the common block stays on while USB holds a reference, which looks like
the safer side for the HS PHY clock.

v2 also skips the USB PCS in qmp_combo_typec_switch_set() and
qmp_combo_reconfigure_phy() in DP-only mode, as Sashiko pointed out on
v1:

https://lore.kernel.org/all/20261023123512.1912-1-26rote@gmail.com/

Thanks,
Joonhoe Kim