2018-05-26 02:14:14

by Andrey Smirnov

[permalink] [raw]
Subject: [PATCH v2] ARM: dts: imx51-zii-rdu1: Make sure SD1_WP is low

Make sure that MX51_PAD_GPIO1_1 does not remain configure as
ALT0/SD1_WP (it is out of reset). This is needed because of external
pull-up resistor attached to that pad that, when left unchanged, will
drive SD1_WP high preventing eSDHC1/eMMC from working correctly.

To fix that add a pinmux configuration line configureing the pad to
function as a GPIO. While we are at it, add a corresponding
output-high GPIO hog in an effort to minimize current consumption.

Cc: Nikita Yushchenko <[email protected]>
Cc: Shawn Guo <[email protected]>
Cc: Fabio Estevam <[email protected]>
Cc: Lucas Stach <[email protected]>
Cc: Chris Healy <[email protected]>
Cc: Rob Herring <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Andrey Smirnov <[email protected]>
---

Changes since [v1]:

- Switched GPIO hog to be output-high

- Removed whitespace damage

[v1] lkml.kernel.org/r/[email protected]

arch/arm/boot/dts/imx51-zii-rdu1.dts | 28 ++++++++++++++++++++++++++++
1 file changed, 28 insertions(+)

diff --git a/arch/arm/boot/dts/imx51-zii-rdu1.dts b/arch/arm/boot/dts/imx51-zii-rdu1.dts
index df9eca94d812..1e343f35a9d9 100644
--- a/arch/arm/boot/dts/imx51-zii-rdu1.dts
+++ b/arch/arm/boot/dts/imx51-zii-rdu1.dts
@@ -476,6 +476,17 @@
status = "okay";
};

+&gpio1 {
+ unused-sd3-wp-gpio {
+ /*
+ * See pinctrl_esdhc1 below for more details on this
+ */
+ gpio-hog;
+ gpios = <1 GPIO_ACTIVE_HIGH>;
+ output-high;
+ };
+};
+
&i2c2 {
pinctrl-names = "default";
pinctrl-0 = <&pinctrl_i2c2>;
@@ -660,6 +671,23 @@
MX51_PAD_SD1_DATA1__SD1_DATA1 0x20d5
MX51_PAD_SD1_DATA2__SD1_DATA2 0x20d5
MX51_PAD_SD1_DATA3__SD1_DATA3 0x20d5
+ /*
+ * GPIO1_1 is not directly used by eSDHC1 in
+ * any capacity, but earlier versions of RDU1
+ * used that pin as WP GPIO for eSDHC3 and
+ * because of that that pad has an external
+ * pull-up resistor. This is problematic
+ * because out of reset the pad is configured
+ * as ALT0 which serves as SD1_WP, which, when
+ * pulled high by and external pull-up, will
+ * inhibit execution of any write request to
+ * attached eMMC device.
+ *
+ * To avoid this problem we configure the pad
+ * to ALT1/GPIO and avoid driving SD1_WP
+ * signal high.
+ */
+ MX51_PAD_GPIO1_1__GPIO1_1 0x0000
>;
};

--
2.17.0



2018-05-31 09:13:45

by Nikita Yushchenko

[permalink] [raw]
Subject: Re: [PATCH v2] ARM: dts: imx51-zii-rdu1: Make sure SD1_WP is low

> Make sure that MX51_PAD_GPIO1_1 does not remain configure as
> ALT0/SD1_WP (it is out of reset). This is needed because of external
> pull-up resistor attached to that pad that, when left unchanged, will
> drive SD1_WP high preventing eSDHC1/eMMC from working correctly.
>
> To fix that add a pinmux configuration line configureing the pad to
> function as a GPIO. While we are at it, add a corresponding
> output-high GPIO hog in an effort to minimize current consumption.
>
> Cc: Nikita Yushchenko <[email protected]>
> Cc: Shawn Guo <[email protected]>
> Cc: Fabio Estevam <[email protected]>
> Cc: Lucas Stach <[email protected]>
> Cc: Chris Healy <[email protected]>
> Cc: Rob Herring <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Cc: [email protected]
> Signed-off-by: Andrey Smirnov <[email protected]>

Tested-By: Nikita Yushchenko <[email protected]

Tested on 8.9'' RDU1. Without this patch, eMMC does not work, get

[ 40.801367] mmc0: Timeout waiting for hardware interrupt.
[ 40.806789] mmc0: sdhci: ============ SDHCI REGISTER DUMP ===========
... <dump follows>

With this patch, eMMC works correctly.

Nikita

2018-06-17 07:17:24

by Shawn Guo

[permalink] [raw]
Subject: Re: [PATCH v2] ARM: dts: imx51-zii-rdu1: Make sure SD1_WP is low

On Fri, May 25, 2018 at 07:12:38PM -0700, Andrey Smirnov wrote:
> Make sure that MX51_PAD_GPIO1_1 does not remain configure as
> ALT0/SD1_WP (it is out of reset). This is needed because of external
> pull-up resistor attached to that pad that, when left unchanged, will
> drive SD1_WP high preventing eSDHC1/eMMC from working correctly.
>
> To fix that add a pinmux configuration line configureing the pad to
> function as a GPIO. While we are at it, add a corresponding
> output-high GPIO hog in an effort to minimize current consumption.
>
> Cc: Nikita Yushchenko <[email protected]>
> Cc: Shawn Guo <[email protected]>
> Cc: Fabio Estevam <[email protected]>
> Cc: Lucas Stach <[email protected]>
> Cc: Chris Healy <[email protected]>
> Cc: Rob Herring <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Cc: [email protected]
> Signed-off-by: Andrey Smirnov <[email protected]>

Applied, thanks.