Skip to contentSkip to conversation

arm64: dts: clearfog-gt-8k: fix ge phy reset pin

PATCH In review [email protected] 3 participants 4 messages Mar 28, 2020 – Mar 29, 2020

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

PATCH 1/1 arm64: dts: clearfog-gt-8k: fix ge phy reset pin

From: Stefan Eichenberger <[email protected]>

According to the ClearFog-GT-8K-rev-1_1-Simplified-Schematic the reset pin for the gigabit phy is MPP62 and not MPP43.

---
 .../dts/marvell/armada-8040-clearfog-gt-8k.dts     | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
arch/arm64/boot/dts/marvell/armada-8040-clearfog-gt-8k.dts +7 -7
index b90d78a5724b..d371d938b41e 100644
@@ -144,7 +144,6 @@
 	 * [35-38] CP0 I2C1 and I2C0
 	 * [39] GPIO reset button
 	 * [40,41] LED0 and LED1
-	 * [43] 1512 phy reset
 	 * [47] USB VBUS EN (active low)
 	 * [48] FAN PWM
 	 * [49] SFP+ present signal
@@ -155,6 +154,7 @@
 	 * [54] NFC reset
 	 * [55] Micro SD card detect
 	 * [56-61] Micro SD
+	 * [62] 1512 phy reset
 	 */
 
 	cp0_pci0_reset_pins: pci0-reset-pins {
@@ -197,11 +197,6 @@
 		marvell,function = "gpio";
 	};
 
-	cp0_copper_eth_phy_reset: copper-eth-phy-reset {
-		marvell,pins = "mpp43";
-		marvell,function = "gpio";
-	};
-
 	cp0_xhci_vbus_pins: xhci0-vbus-pins {
 		marvell,pins = "mpp47";
 		marvell,function = "gpio";
@@ -232,6 +227,11 @@
 			       "mpp60", "mpp61";
 		marvell,function = "sdio";
 	};
+
+	cp0_copper_eth_phy_reset: copper-eth-phy-reset {
+		marvell,pins = "mpp62";
+		marvell,function = "gpio";
+	};
 };
 
 &cp0_pcie0 {
@@ -365,7 +365,7 @@
 		reg = <0>;
 		pinctrl-names = "default";
 		pinctrl-0 = <&cp0_copper_eth_phy_reset>;
-		reset-gpios = <&cp0_gpio2 11 GPIO_ACTIVE_LOW>;
+		reset-gpios = <&cp0_gpio2 30 GPIO_ACTIVE_LOW>;
 		reset-assert-us = <10000>;
 		reset-deassert-us = <10000>;
 	};
Signature · 1 line
2.20.1
#1 lore raw

Hi Stefan,

Quoted 4 lines · On Sun, Mar 29 2020, [email protected] wrote:

From: Stefan Eichenberger <[email protected]>

According to the ClearFog-GT-8K-rev-1_1-Simplified-Schematic the reset pin for the gigabit phy is MPP62 and not MPP43.

Have you tested that on real hardware?

The 1Gb PHY reset on my Clearfog GT-8K is connected to MPP43. Russell's commit 46f94c7818e7 ("arm64: dts: clearfog-gt-8k: set gigabit PHY reset deassert delay") indicates that this is the case on his board as well.

In case there was a hardware change between board revisions, we need another dtb for that revision.

baruch

Quoted 50 lines
---
 .../dts/marvell/armada-8040-clearfog-gt-8k.dts     | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
arch/arm64/boot/dts/marvell/armada-8040-clearfog-gt-8k.dts +6 -6
index b90d78a5724b..d371d938b41e 100644
@@ -144,7 +144,6 @@
 	 * [35-38] CP0 I2C1 and I2C0
 	 * [39] GPIO reset button
 	 * [40,41] LED0 and LED1
-	 * [43] 1512 phy reset
 	 * [47] USB VBUS EN (active low)
 	 * [48] FAN PWM
 	 * [49] SFP+ present signal
@@ -155,6 +154,7 @@
 	 * [54] NFC reset
 	 * [55] Micro SD card detect
 	 * [56-61] Micro SD
+	 * [62] 1512 phy reset
 	 */
 
 	cp0_pci0_reset_pins: pci0-reset-pins {
@@ -197,11 +197,6 @@
 		marvell,function = "gpio";
 	};
 
-	cp0_copper_eth_phy_reset: copper-eth-phy-reset {
-		marvell,pins = "mpp43";
-		marvell,function = "gpio";
-	};
-
 	cp0_xhci_vbus_pins: xhci0-vbus-pins {
 		marvell,pins = "mpp47";
 		marvell,function = "gpio";
@@ -232,6 +227,11 @@
 			       "mpp60", "mpp61";
 		marvell,function = "sdio";
 	};
+
+	cp0_copper_eth_phy_reset: copper-eth-phy-reset {
+		marvell,pins = "mpp62";
+		marvell,function = "gpio";
+	};
 };
 
 &cp0_pcie0 {

arch/arm64/boot/dts/marvell/armada-8040-clearfog-gt-8k.dts line 371

Quoted 8 lines
arch/arm64/boot/dts/marvell/armada-8040-clearfog-gt-8k.dts +1 -1
@@ -365,7 +365,7 @@
 		reg = <0>;
 		pinctrl-names = "default";
 		pinctrl-0 = <&cp0_copper_eth_phy_reset>;
-		reset-gpios = <&cp0_gpio2 11 GPIO_ACTIVE_LOW>;
+		reset-gpios = <&cp0_gpio2 30 GPIO_ACTIVE_LOW>;
 		reset-assert-us = <10000>;
 		reset-deassert-us = <10000>;
 	};

--
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- [email protected] - tel: +972.52.368.4656, http://www.tkos.co.il -

#2 lore raw
Quoted 9 lines · Baruch Siach

Hi Stefan,

Quoted 4 lines · On Sun, Mar 29 2020, [email protected] wrote:

From: Stefan Eichenberger <[email protected]>

According to the ClearFog-GT-8K-rev-1_1-Simplified-Schematic the reset pin for the gigabit phy is MPP62 and not MPP43.

Quoted 8 lines

Have you tested that on real hardware?

The 1Gb PHY reset on my Clearfog GT-8K is connected to MPP43. Russell's commit 46f94c7818e7 ("arm64: dts: clearfog-gt-8k: set gigabit PHY reset deassert delay") indicates that this is the case on his board as well.

In case there was a hardware change between board revisions, we need another dtb for that revision.

It's a bug in the simplified schematics since that schematics is based on rev 1.0 and not rev 1.1 as claimed.

In rev 1.0; the 1Gbps phy reset was connected to MPP62; but that MPP is not functional as a GPIO when selecting MPP[56:61] as SD card. Due to that we manually rewired ALL rev 1.0 PCBs 1Gbps phy to be connected to MPP43 via R8038 pads.

Rev 1.1 fixes this by that by disconnecting 1Gbps phy reset from MPP62 and wiring it to MPP43.
So basically rev 1.0 and rev 1.1 are compatible software wise. We will fix the schematics.

Rabeeh

#3 lore raw

Hi Rabeeh and Baruch

Quoted 27 lines · Rabeeh Khoury
Quoted 9 lines · Baruch Siach

Hi Stefan,

Quoted 4 lines · On Sun, Mar 29 2020, [email protected] wrote:

From: Stefan Eichenberger <[email protected]>

According to the ClearFog-GT-8K-rev-1_1-Simplified-Schematic the reset pin for the gigabit phy is MPP62 and not MPP43.

Quoted 8 lines

Have you tested that on real hardware?

The 1Gb PHY reset on my Clearfog GT-8K is connected to MPP43. Russell's commit 46f94c7818e7 ("arm64: dts: clearfog-gt-8k: set gigabit PHY reset deassert delay") indicates that this is the case on his board as well.

In case there was a hardware change between board revisions, we need another dtb for that revision.

It's a bug in the simplified schematics since that schematics is based on rev 1.0 and not rev 1.1 as claimed.

In rev 1.0; the 1Gbps phy reset was connected to MPP62; but that MPP is not functional as a GPIO when selecting MPP[56:61] as SD card. Due to that we manually rewired ALL rev 1.0 PCBs 1Gbps phy to be connected to MPP43 via R8038 pads.

Quoted 4 lines

Rev 1.1 fixes this by that by disconnecting 1Gbps phy reset from MPP62 and wiring it to MPP43.
So basically rev 1.0 and rev 1.1 are compatible software wise. We will fix the schematics.

Ahh now I see, I didn't enable the phy driver when I did the test with the default devicetree and then when I changed the devicetree I also enabled the driver, that's my fault.

Sorry for the confusion... I can confirm that it works with MPP43. Thanks for the clarification!

Regards,
Stefan