Re: [PATCH RFC] usb: typec: ucsi: Schedule connector worker on freezable workqueue

[Vladyslav Tymoshchuk]

Hi Joonhoe, Abel,

On Tue, 20 Oct 2026 22:24:30 +0900, Joonhoe Kim wrote:
> Abel, the RFC has not moved since February 2025 and two more boards
> now need it. Do you still plan to respin it? On Dmitry's question,
> 7d0767c5cd87 took the same approach for the altmode worker as a
> stop-gap until the PMIC GLINK rework, which may help here too.

I have sent an RFC for that rework:

  https://lore.kernel.org/all/20261022164151.592-1-tvlad744@xxxxxxxxx/

pmic_glink holds the notifications back from ->prepare() on, and
passes them on from system_freezable_wq after resume, for all its
clients. For a notification that arrives during system sleep, the
UCSI connector work then runs only after the devices it calls into
have resumed, so this patch would not be needed. A notification can
still wake up the system, if the client's device may wake up the
system.

On SM8750 the series avoids the reset on an unplug during
suspend-to-idle, also with wakeup disabled for both clients, on a
kernel without the dwc3-qcom fix.

Joonhoe, if you can test it on your board, that would show whether it
also covers the FSA4480 case. It is on top of next-20261021; if it
does not apply to your tree, I can send it rebased on v7.3-rc5.

Thanks,
Vladyslav

(I used an AI assistant, Claude, for the analysis and to draft this
mail; the tests ran on my device.)