[PATCH v2 0/2] phy: sun4i-usb: add the Allwinner A733 USB PHY
From: Vinicius Pedrosa
Date: Mon Nov 16 2026 - 05:56:07 EST
This adds the A733 to the sun4i-usb PHY driver. It works with the D1
settings, except that the A733 gets SIDDQ cleared before the PHY reset
is released (patch 2). That made an intermittent USB1 enumeration
failure go away on my board.
I only have one Cubie A7Z, running a local DT, since there's no A733
dtsi upstream yet. Only USB1 is enabled in it, so PHY0 (the OTG port)
never went through the new order. The boots ran the same reorder on
top of 7.3-rc6, before I turned it into a per-SoC flag.
The numbers are in patch 2. The comparison inside the last run alone
(1/31 vs 0/60) isn't statistically significant, so it leans on the
earlier 3/40 too.
One question. The vendor BSP uses this order in common code, for every
SoC. I kept it per SoC because I can't test the others and the manual
says nothing about the order. Would you prefer it for all the SoCs that
clear PHY_CTL_SIDDQ (D1, H616, and A100/A523 through the D1 fallback),
or is a per-SoC flag fine?
Changes in v2:
- Reword the binding commit message into a single paragraph (Krzysztof)
- Rebase on phy/next and drop the dependency on Xu Rao's "phy:
allwinner: sun4i-usb: disable the PHY2 PMU clock after SIDDQ setup".
Patch 2 only overlaps with it in context.
v1: https://lore.kernel.org/all/20261029172325.3425046-1-vinicius.eduardo.pedrosa@xxxxxxxxx/
Vinicius
Vinicius Pedrosa (2):
dt-bindings: phy: document Allwinner A733 USB-2.0 PHY
phy: sun4i-usb: clear SIDDQ before the PHY reset on A733
.../phy/allwinner,sun50i-a64-usb-phy.yaml | 1 +
drivers/phy/allwinner/phy-sun4i-usb.c | 35 ++++++++++++++++---
2 files changed, 31 insertions(+), 5 deletions(-)
base-commit: 6bdfbe223ed4735073c299a6645ebd3faac3f51f
--
2.43.0