Skip to contentSkip to conversation

arm64: dts: allwinner: a523: fix SD card detect pull resistor

PATCH In review Andre Przywara 3 participants 5 messages Apr 25, 2025 – Apr 28, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 arm64: dts: allwinner: a523: fix SD card detect pull resistor

Trying to use the SD card on the Radxa board revealed that the card detect wouldn't work as expected (insert not detected). Looking at the schematic shows that the pull-up resistor is actually not populated ("NC"), and the transistor just pulls the GPIO pin to GND, but it's floating otherwise.
So using the pull-down flag is definitely wrong, we need the internal pull up to get a reliable signal. The same is true for the Avaota board (there is no transistor there, but it's floating in the same way). There is no schematic for the X96QPro+ board, but experiments show it's the same behaviour.

So change the GPIO flag for the card detect GPIO property to activate the pull-up resistor for that pin.

Fixes 80e0fb4e491b ("arm64: dts: allwinner: a523: add Radxa A5E support")

---
Hi,

please let me know if I should split this up into 3 patches, with proper
Fixes: tags, or if you can maybe squash this into the original commits still?

Cheers,
Andre

 arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts  | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts +1 -1
index 03c9a9ef5adc2..2d2f3af91d05e 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
 	vmmc-supply = <&reg_cldo3>;
-	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
 	bus-width = <4>;
 	status = "okay";
 };
arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts +1 -1
index c0bce3f4fa925..59db103546f65 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
 	vmmc-supply = <&reg_vcc3v3>;
-	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
 	bus-width = <4>;
 	disable-wp;
 	status = "okay";
arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts +1 -1
index 85a546aecdbe1..dea2acc1849bb 100644
@@ -66,7 +66,7 @@ &ehci1 {
 
 &mmc0 {
 	vmmc-supply = <&reg_cldo3>;
-	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+	cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
 	bus-width = <4>;
 	status = "okay";
 };

base-commit: 1e5a69d67d1b3c55c9b0cd3933af1436b5d52aa1

Signature · 1 line
2.46.3
#1 lore raw
Quoted 15 lines · Andre Przywara

Trying to use the SD card on the Radxa board revealed that the card detect wouldn't work as expected (insert not detected). Looking at the schematic shows that the pull-up resistor is actually not populated ("NC"), and the transistor just pulls the GPIO pin to GND, but it's floating otherwise.
So using the pull-down flag is definitely wrong, we need the internal pull up to get a reliable signal. The same is true for the Avaota board (there is no transistor there, but it's floating in the same way). There is no schematic for the X96QPro+ board, but experiments show it's the same behaviour.

So change the GPIO flag for the card detect GPIO property to activate the pull-up resistor for that pin.

Quoted 8 lines
Fixes 80e0fb4e491b ("arm64: dts: allwinner: a523: add Radxa A5E support")

---
Hi,

please let me know if I should split this up into 3 patches, with proper
Fixes: tags, or if you can maybe squash this into the original commits still?

I can squash them in if you prefer.

ChenYu

Quoted 52 lines

Cheers,
Andre

 arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts  | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts +1 -1
index 03c9a9ef5adc2..2d2f3af91d05e 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_cldo3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        status = "okay";
 };
arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts +1 -1
index c0bce3f4fa925..59db103546f65 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_vcc3v3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        disable-wp;
        status = "okay";
arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts +1 -1
index 85a546aecdbe1..dea2acc1849bb 100644
@@ -66,7 +66,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_cldo3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        status = "okay";
 };

base-commit: 1e5a69d67d1b3c55c9b0cd3933af1436b5d52aa1 --
2.46.3

#2 lore raw
Quoted 22 lines · Chen-Yu Tsai
Quoted 20 lines · Andre Przywara

Trying to use the SD card on the Radxa board revealed that the card detect wouldn't work as expected (insert not detected). Looking at the schematic shows that the pull-up resistor is actually not populated ("NC"), and the transistor just pulls the GPIO pin to GND, but it's floating otherwise.
So using the pull-down flag is definitely wrong, we need the internal pull up to get a reliable signal. The same is true for the Avaota board (there is no transistor there, but it's floating in the same way). There is no schematic for the X96QPro+ board, but experiments show it's the same behaviour.

So change the GPIO flag for the card detect GPIO property to activate the pull-up resistor for that pin.

Fixes 80e0fb4e491b ("arm64: dts: allwinner: a523: add Radxa A5E support")

---
Hi,

Quoted 5 lines
Quoted 3 lines

please let me know if I should split this up into 3 patches, with proper
Fixes: tags, or if you can maybe squash this into the original commits still?

I can squash them in if you prefer.

Fixes separated and squashed into the original patches.

ChenYu

Quoted 54 lines

ChenYu

Quoted 52 lines

Cheers,
Andre

 arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts  | 2 +-
 arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
arch/arm64/boot/dts/allwinner/sun55i-a527-radxa-a5e.dts +1 -1
index 03c9a9ef5adc2..2d2f3af91d05e 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_cldo3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        status = "okay";
 };
arch/arm64/boot/dts/allwinner/sun55i-h728-x96qpro+.dts +1 -1
index c0bce3f4fa925..59db103546f65 100644
@@ -56,7 +56,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_vcc3v3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        disable-wp;
        status = "okay";
arch/arm64/boot/dts/allwinner/sun55i-t527-avaota-a1.dts +1 -1
index 85a546aecdbe1..dea2acc1849bb 100644
@@ -66,7 +66,7 @@ &ehci1 {
 
 &mmc0 {
        vmmc-supply = <&reg_cldo3>;
-       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_DOWN)>; /* PF6 */
+       cd-gpios = <&pio 5 6 (GPIO_ACTIVE_LOW | GPIO_PULL_UP)>; /* PF6 */
        bus-width = <4>;
        status = "okay";
 };

base-commit: 1e5a69d67d1b3c55c9b0cd3933af1436b5d52aa1 --
2.46.3

#4 lore raw
Quoted 5 lines · Chukun Pan

Hi,

Quoted 1 line

Fixes separated and squashed into the original patches.

Quoted 7 lines

https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux.git/commit/?h=sunxi/dt-for-6.16&id=dd97ddc4da0af02139b0e97e98c210219631e161

In the above commit, dts was renamed but Makefile was not changed.

Quoted 1 line

+dtb-$(CONFIG_ARCH_SUNXI) += sun55i-a527-radxa-a5e.dtb

Do you want a patch or will you rebase again?

Fixed now, though today's linux-next might be broken. Sorry for the mess.

ChenYu