* [PATCH 1/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add LPO clock for AP6212
2026-11-20 11:40 [PATCH 0/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix Bluetooth and USB gadget Ivan Pfarher via B4 Relay
@ 2026-11-20 11:40 ` Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 2/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add AP6212 Bluetooth Ivan Pfarher via B4 Relay
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Ivan Pfarher via B4 Relay @ 2026-11-20 11:40 UTC (permalink / raw)
To: Rob Herring, Krzysztof Kozlowski, Conor Dooley, Chen-Yu Tsai,
Jernej Skrabec, Samuel Holland
Cc: devicetree, linux-arm-kernel, linux-sunxi, Ivan Pfarher
From: Ivan Pfarher <ipfarher@gmail.com>
The AP6212 WiFi/BT module needs a 32.768 kHz low power oscillator (LPO)
clock, which on this board comes from the RTC clock output (osc32k-out).
Nothing in the device tree references that clock, so it is disabled as
unused during late boot.
WiFi keeps working without it, but the Bluetooth core does not: with the
LPO gated, the BCM43430A1 ROM answers HCI_Reset and nothing else (no
Read Local Version, no firmware download), so the controller can never
be brought up. Turning the clock output on by hand (RTC LOSC_OUT_GATING)
makes the controller answer normally.
Reference the clock from the WiFi power sequence, as other boards using
the same module do (e.g. sun8i-h2-plus-bananapi-m2-zero.dts).
Tested on an Orange Pi Zero Plus2 (H5) running Linux 6.6.160 with this
change applied: osc32k-out stays enabled and the Bluetooth controller
answers HCI commands. Build-tested only against current mainline.
This problem was investigated, and the patch drafted, in an interactive
session with an AI coding assistant (Claude by Anthropic). The submitter
reviewed the change and ran the tests described above.
Fixes: cf8e19e1ec15 ("arm64: dts: orange-pi-zero-plus2: enable AP6212a WiFi/BT combo")
Signed-off-by: Ivan Pfarher <ipfarher@gmail.com>
---
arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
index 3eb986c35..34d6c1468 100644
--- a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
+++ b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
@@ -5,6 +5,7 @@
#include "sun50i-h5.dtsi"
+#include <dt-bindings/clock/sun6i-rtc.h>
#include <dt-bindings/gpio/gpio.h>
/ {
@@ -54,6 +55,8 @@
wifi_pwrseq: pwrseq {
compatible = "mmc-pwrseq-simple";
+ clocks = <&rtc CLK_OSC32K_FANOUT>;
+ clock-names = "ext_clock";
reset-gpios = <&pio 0 9 GPIO_ACTIVE_LOW>; /* PA9 */
post-power-on-delay-ms = <200>;
};
--
2.43.0
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH 2/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add AP6212 Bluetooth
2026-11-20 11:40 [PATCH 0/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix Bluetooth and USB gadget Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 1/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add LPO clock for AP6212 Ivan Pfarher via B4 Relay
@ 2026-11-20 11:40 ` Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 3/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix USB peripheral mode Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 4/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add MMC aliases Ivan Pfarher via B4 Relay
3 siblings, 0 replies; 5+ messages in thread
From: Ivan Pfarher via B4 Relay @ 2026-11-20 11:40 UTC (permalink / raw)
To: Rob Herring, Krzysztof Kozlowski, Conor Dooley, Chen-Yu Tsai,
Jernej Skrabec, Samuel Holland
Cc: devicetree, linux-arm-kernel, linux-sunxi, Ivan Pfarher
From: Ivan Pfarher <ipfarher@gmail.com>
UART1 is connected to the Bluetooth part of the AP6212 module, but the
controller itself is not described, so nothing powers it up and the
kernel never registers it.
Add it as a serdev child of UART1. BT_REG_ON is connected to PA10:
on the board, driving PA10 low makes the controller drop CTS and stop
answering, driving it high brings it back. It uses the same RTC LPO
clock as the WiFi part. The wake-up lines are not described because
they could not be identified.
Tested on an Orange Pi Zero Plus2 (H5) running Linux 6.6.160 with this
change applied: hci_bcm powers the controller, loads
brcm/BCM43430A1.hcd, switches the UART to 1.5 Mbaud and the adapter
works with BlueZ (scan finds nearby devices):
Bluetooth: hci0: BCM43430A1 'brcm/BCM43430A1.hcd' Patch
Bluetooth: hci0: BCM43438A1 26MHz AP6212A1_CL1 BT4.0 OTP-BD-0034
Build-tested only against current mainline.
This problem was investigated, and the patch drafted, in an interactive
session with an AI coding assistant (Claude by Anthropic). The submitter
reviewed the change and ran the tests described above.
Signed-off-by: Ivan Pfarher <ipfarher@gmail.com>
---
.../boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
index 34d6c1468..5697251b9 100644
--- a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
+++ b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
@@ -128,7 +128,18 @@
&uart1 {
pinctrl-names = "default";
pinctrl-0 = <&uart1_pins>, <&uart1_rts_cts_pins>;
+ uart-has-rtscts;
status = "okay";
+
+ bluetooth {
+ compatible = "brcm,bcm43438-bt";
+ max-speed = <1500000>;
+ clocks = <&rtc CLK_OSC32K_FANOUT>;
+ clock-names = "lpo";
+ vbat-supply = <®_vcc3v3>;
+ vddio-supply = <®_vcc3v3>;
+ shutdown-gpios = <&pio 0 10 GPIO_ACTIVE_HIGH>; /* PA10 */
+ };
};
&usb_otg {
--
2.43.0
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH 3/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix USB peripheral mode
2026-11-20 11:40 [PATCH 0/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix Bluetooth and USB gadget Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 1/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add LPO clock for AP6212 Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 2/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add AP6212 Bluetooth Ivan Pfarher via B4 Relay
@ 2026-11-20 11:40 ` Ivan Pfarher via B4 Relay
2026-11-20 11:40 ` [PATCH 4/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add MMC aliases Ivan Pfarher via B4 Relay
3 siblings, 0 replies; 5+ messages in thread
From: Ivan Pfarher via B4 Relay @ 2026-11-20 11:40 UTC (permalink / raw)
To: Rob Herring, Krzysztof Kozlowski, Conor Dooley, Chen-Yu Tsai,
Jernej Skrabec, Samuel Holland
Cc: devicetree, linux-arm-kernel, linux-sunxi, Ivan Pfarher
From: Ivan Pfarher <ipfarher@gmail.com>
The micro USB port is set to dr_mode = "peripheral" because it cannot
supply VBUS, but EHCI0/OHCI0, which share USB PHY 0 with the MUSB
controller, are enabled too.
When they register, usb_add_hcd() asks the shared PHY for host mode
(since commit e671765e521c ("usb: core: Try generic PHY_MODE_USB_HOST
if usb_phy_roothub_set_mode fails")), and phy-sun4i-usb switches PHY 0
to host ("Changing dr_mode to 1"). MUSB then stays in a_wait_vrise, the
UDC stays "not attached", and a gadget never shows up on the host.
The port cannot work as a host on this board anyway, so drop EHCI0 and
OHCI0.
Tested on an Orange Pi Zero Plus2 (H5) running Linux 6.6.160 with this
change applied: g_cdc enumerates on the PC (CDC ECM + ACM), the UDC
state is "configured" and MUSB is in b_peripheral. Build-tested only
against current mainline.
The H3 version of the board (sun8i-h3-orangepi-zero-plus2.dts) has the
same setup from the same commit but is not changed here, as it could
not be tested.
This problem was investigated, and the patch drafted, in an interactive
session with an AI coding assistant (Claude by Anthropic). The submitter
reviewed the change and ran the tests described above.
Fixes: 43aad09c7928 ("ARM: dts: orange-pi-zero-plus2: enable USB OTG port")
Signed-off-by: Ivan Pfarher <ipfarher@gmail.com>
---
arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts | 8 --------
1 file changed, 8 deletions(-)
diff --git a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
index 5697251b9..892723b46 100644
--- a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
+++ b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
@@ -66,10 +66,6 @@
status = "okay";
};
-&ehci0 {
- status = "okay";
-};
-
&hdmi {
status = "okay";
};
@@ -115,10 +111,6 @@
status = "okay";
};
-&ohci0 {
- status = "okay";
-};
-
&uart0 {
pinctrl-names = "default";
pinctrl-0 = <&uart0_pa_pins>;
--
2.43.0
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH 4/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Add MMC aliases
2026-11-20 11:40 [PATCH 0/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix Bluetooth and USB gadget Ivan Pfarher via B4 Relay
` (2 preceding siblings ...)
2026-11-20 11:40 ` [PATCH 3/4] arm64: dts: allwinner: h5: orangepi-zero-plus2: Fix USB peripheral mode Ivan Pfarher via B4 Relay
@ 2026-11-20 11:40 ` Ivan Pfarher via B4 Relay
3 siblings, 0 replies; 5+ messages in thread
From: Ivan Pfarher via B4 Relay @ 2026-11-20 11:40 UTC (permalink / raw)
To: Rob Herring, Krzysztof Kozlowski, Conor Dooley, Chen-Yu Tsai,
Jernej Skrabec, Samuel Holland
Cc: devicetree, linux-arm-kernel, linux-sunxi, Ivan Pfarher
From: Ivan Pfarher <ipfarher@gmail.com>
sunxi-mmc probes asynchronously, so without aliases the microSD card,
the SDIO WiFi and the eMMC get their mmc numbers in probe order. A boot
script using root=/dev/mmcblk0p2 for the microSD card can then end up
on the eMMC instead.
Add aliases so that the microSD card is always mmc0 and the eMMC mmc1,
matching the numbering U-Boot uses on this board, with the SDIO WiFi
as mmc2.
Tested on an Orange Pi Zero Plus2 (H5) running Linux 6.6.160 with this
change applied: the microSD card is mmcblk0 and the eMMC mmcblk1,
booting from either. Build-tested only against current mainline.
This problem was investigated, and the patch drafted, in an interactive
session with an AI coding assistant (Claude by Anthropic). The submitter
reviewed the change and ran the tests described above.
Signed-off-by: Ivan Pfarher <ipfarher@gmail.com>
---
arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
index 892723b46..4a26863e8 100644
--- a/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
+++ b/arch/arm64/boot/dts/allwinner/sun50i-h5-orangepi-zero-plus2.dts
@@ -13,6 +13,9 @@
compatible = "xunlong,orangepi-zero-plus2", "allwinner,sun50i-h5";
aliases {
+ mmc0 = &mmc0;
+ mmc1 = &mmc2;
+ mmc2 = &mmc1;
serial0 = &uart0;
};
--
2.43.0
^ permalink raw reply related [flat|nested] 5+ messages in thread