Skip to contentSkip to conversation

ARM: dts: imx: Fix the AR803X phy-mode

RFC In review Fabio Estevam 7 participants 11 messages Mar 30, 2019 – Apr 1, 2019

Subsystems (2)
lore
Share & tools
lore

Fabio Estevam

author reviewer
#0 lore raw

RFC 1/1 ARM: dts: imx: Fix the AR803X phy-mode

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

Quoting Andrew Lunn:

"The problem here is, all the DTs were broken since day 0. However, because the PHY driver was also broken, nobody noticed and it worked. Now that the PHY driver has been fixed, all the bugs in the DTs now become an issue"

To fix this problem, the phy-mode property needs to be "rgmii-id", which has the following meaning as per Documentation/devicetree/bindings/net/ethernet.txt:

"RGMII with internal RX and TX delays provided by the PHY, the MAC should not add the RX or TX delays in this case)"

Tested on imx6-sabresd, imx6sx-sdb and imx7d-pico boards with successfully restored networking.

---
Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

It's not possible to know if there are other imx boards in tree that use AR803X, so here is an attempt.

mx6-sabresd, imx6sx-sdb and imx7d-pico have been successfully tested.

Thanks

 arch/arm/boot/dts/imx6q-tbs2910.dts      | 2 +-
 arch/arm/boot/dts/imx6qdl-sabresd.dtsi   | 2 +-
 arch/arm/boot/dts/imx6qdl-sr-som.dtsi    | 2 +-
 arch/arm/boot/dts/imx6qdl-wandboard.dtsi | 2 +-
 arch/arm/boot/dts/imx6sx-sdb.dtsi        | 2 +-
 arch/arm/boot/dts/imx7d-pico.dtsi        | 2 +-
 6 files changed, 6 insertions(+), 6 deletions(-)
arch/arm/boot/dts/imx6q-tbs2910.dts +1 -1
index 2ce8399a10ba..bfff87ce2e1f 100644
@@ -98,7 +98,7 @@
 &fec {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_enet>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-reset-gpios = <&gpio1 25 GPIO_ACTIVE_LOW>;
 	status = "okay";
 };
arch/arm/boot/dts/imx6qdl-sabresd.dtsi +1 -1
index a0705066ccba..185fb17a3500 100644
@@ -202,7 +202,7 @@
 &fec {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_enet>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-reset-gpios = <&gpio1 25 GPIO_ACTIVE_LOW>;
 	status = "okay";
 };
arch/arm/boot/dts/imx6qdl-sr-som.dtsi +1 -1
index 4ccb7afc4b35..6d7f6b9035bc 100644
@@ -53,7 +53,7 @@
 &fec {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_microsom_enet_ar8035>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-reset-duration = <2>;
 	phy-reset-gpios = <&gpio4 15 GPIO_ACTIVE_LOW>;
 	status = "okay";
arch/arm/boot/dts/imx6qdl-wandboard.dtsi +1 -1
index b7d5fb421404..50d9a989e06a 100644
@@ -224,7 +224,7 @@
 &fec {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_enet>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-reset-gpios = <&gpio3 29 GPIO_ACTIVE_LOW>;
 	interrupts-extended = <&gpio1 6 IRQ_TYPE_LEVEL_HIGH>,
 			      <&intc 0 119 IRQ_TYPE_LEVEL_HIGH>;
arch/arm/boot/dts/imx6sx-sdb.dtsi +1 -1
index 08ede56c3f10..f6972deb5e39 100644
@@ -191,7 +191,7 @@
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_enet1>;
 	phy-supply = <&reg_enet_3v3>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-handle = <&ethphy1>;
 	phy-reset-gpios = <&gpio2 7 GPIO_ACTIVE_LOW>;
 	status = "okay";
arch/arm/boot/dts/imx7d-pico.dtsi +1 -1
index 3fd595a71202..6f50ebf31a0a 100644
@@ -92,7 +92,7 @@
 			  <&clks IMX7D_ENET1_TIME_ROOT_CLK>;
 	assigned-clock-parents = <&clks IMX7D_PLL_ENET_MAIN_100M_CLK>;
 	assigned-clock-rates = <0>, <100000000>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-handle = <&ethphy0>;
 	fsl,magic-packet;
 	phy-reset-gpios = <&gpio6 11 GPIO_ACTIVE_LOW>;
Signature · 7 lines
2.17.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#1 lore raw
Quoted 4 lines

Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

I think it should be safe to propose these patches go to stable.

RGMII-ID has been supported since

2e5f9f281ee8 ("net: phy: at803x: Allow specifying the RGMII RX clock delay via phy mode")

Which is somewhere around v4.4. So back porting them should be fine.

Andrew

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

author reviewer
#2 lore raw

Hi Andrew,

Quoted 6 lines · Andrew Lunn
Quoted 5 lines

Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

Quoted 7 lines

I think it should be safe to propose these patches go to stable.

RGMII-ID has been supported since

2e5f9f281ee8 ("net: phy: at803x: Allow specifying the RGMII RX clock delay via phy mode")

Which is somewhere around v4.4. So back porting them should be fine.

This commit landed in v4.5, so I marked it as:
Cc: <[email protected]> # 4.9+

I probably will need to manually generate versions for the some of the stable versions.

Thanks

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#3 lore raw
In reply to Fabio Estevam

Hi Fabio,

Quoted 27 lines · Fabio Estevam

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

Quoting Andrew Lunn:

"The problem here is, all the DTs were broken since day 0. However, because the PHY driver was also broken, nobody noticed and it worked. Now that the PHY driver has been fixed, all the bugs in the DTs now become an issue"

To fix this problem, the phy-mode property needs to be "rgmii-id", which has the following meaning as per Documentation/devicetree/bindings/net/ethernet.txt:

"RGMII with internal RX and TX delays provided by the PHY, the MAC should not add the RX or TX delays in this case)"

Tested on imx6-sabresd, imx6sx-sdb and imx7d-pico boards with successfully restored networking.

---
Hi,

Quoted 7 lines

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

It's not possible to know if there are other imx boards in tree that use AR803X, so here is an attempt.

mx6-sabresd, imx6sx-sdb and imx7d-pico have been successfully tested.

The SolidRun SOM is also affected, I believe (imx6qdl-sr-som.dtsi, AR8035). Would you be able to test it on your side?

Thanks,
baruch

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

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

author reviewer
#4 lore raw

Hi Baruch,

Quoted 2 lines · Baruch Siach

The SolidRun SOM is also affected, I believe (imx6qdl-sr-som.dtsi, AR8035). Would you be able to test it on your side?

This patch does touch imx6qdl-sr-som.dtsi, but I don't have access to my Cubox at the moment.

Could you please test it?

Thanks

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Tested-by
#5 lore raw

Hi Fabio,

Quoted 7 lines · Fabio Estevam
Quoted 2 lines · Baruch Siach

The SolidRun SOM is also affected, I believe (imx6qdl-sr-som.dtsi, AR8035). Would you be able to test it on your side?

This patch does touch imx6qdl-sr-som.dtsi, but I don't have access to my Cubox at the moment.

I missed that.

Quoted 1 line

Could you please test it?

Sure. Tested on Hummingboard2. I verified that networking is broken without this patch (v5.1-rc2). This patch fixes the problem.

Thanks,
baruch

Signature · 8 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - [email protected] - tel: +972.52.368.4656, http://www.tkos.co.il -

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Tested-by on v1 1/1
#6 lore raw
In reply to Fabio Estevam
Quoted 23 lines · Fabio Estevam

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

Quoting Andrew Lunn:

"The problem here is, all the DTs were broken since day 0. However, because the PHY driver was also broken, nobody noticed and it worked. Now that the PHY driver has been fixed, all the bugs in the DTs now become an issue"

To fix this problem, the phy-mode property needs to be "rgmii-id", which has the following meaning as per Documentation/devicetree/bindings/net/ethernet.txt:

"RGMII with internal RX and TX delays provided by the PHY, the MAC should not add the RX or TX delays in this case)"

Tested on imx6-sabresd, imx6sx-sdb and imx7d-pico boards with successfully restored networking.

Quoted 6 lines

---
Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

I can confirm this regression on a tbs2910 board, and also thatthis patch fixes this.

I also think it is the right solution to fix the broken dtbs, like in this patch, instead of adding more complicated backward compatibility stuff.

Thanks,
Soeren

Quoted 8 lines

It's not possible to know if there are other imx boards in tree that use AR803X, so here is an attempt.

mx6-sabresd, imx6sx-sdb and imx7d-pico have been successfully tested.

Thanks

Quoted 8 lines
 arch/arm/boot/dts/imx6q-tbs2910.dts      | 2 +-
 arch/arm/boot/dts/imx6qdl-sabresd.dtsi   | 2 +-
 arch/arm/boot/dts/imx6qdl-sr-som.dtsi    | 2 +-
 arch/arm/boot/dts/imx6qdl-wandboard.dtsi | 2 +-
 arch/arm/boot/dts/imx6sx-sdb.dtsi        | 2 +-
 arch/arm/boot/dts/imx7d-pico.dtsi        | 2 +-
 6 files changed, 6 insertions(+), 6 deletions(-)

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Tested-by Tested-by Adam Thomson on v1 1/1
#7 lore raw
In reply to Fabio Estevam

Hi Fabio,

Quoted 7 lines · On 30 March 2019 15:51 Fabio Estevam, wrote:

Subject: [RFC] ARM: dts: imx: Fix the AR803X phy-mode

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

[...]

Quoted 1 line
 arch/arm/boot/dts/imx6qdl-sabresd.dtsi   | 2 +-

[...]

Quoted 6 lines
arch/arm/boot/dts/imx6qdl-sabresd.dtsi +0 -0
index a0705066ccba..185fb17a3500 100644

arch/arm/boot/dts/imx6qdl-sabresd.dtsi line 208

Quoted 8 lines
arch/arm/boot/dts/imx6qdl-sabresd.dtsi +1 -1
@@ -202,7 +202,7 @@
 &fec {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_enet>;
-	phy-mode = "rgmii";
+	phy-mode = "rgmii-id";
 	phy-reset-gpios = <&gpio1 25 GPIO_ACTIVE_LOW>;
 	status = "okay";
 };

[...]

This patch looks surprisingly similar to this one ;)
https://lore.kernel.org/patchwork/patch/1052588/

Regards,
Steve

author reviewer
#8 lore raw

Hi Steve,

Quoted 7 lines · Steve Twiss

This patch looks surprisingly similar to this one ;)
https://lore.kernel.org/patchwork/patch/1052588/

I put your attributions in the formal patch that I have just submitted.

Thanks

#9 lore raw
In reply to Fabio Estevam

Hi Fabio,

Quoted 30 lines · Fabio Estevam

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

Quoting Andrew Lunn:

"The problem here is, all the DTs were broken since day 0. However, because the PHY driver was also broken, nobody noticed and it worked. Now that the PHY driver has been fixed, all the bugs in the DTs now become an issue"

To fix this problem, the phy-mode property needs to be "rgmii-id", which has the following meaning as per Documentation/devicetree/bindings/net/ethernet.txt:

"RGMII with internal RX and TX delays provided by the PHY, the MAC should not add the RX or TX delays in this case)"

Tested on imx6-sabresd, imx6sx-sdb and imx7d-pico boards with successfully restored networking.

---
Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

Quoted 6 lines

It's not possible to know if there are other imx boards in tree that use AR803X, so here is an attempt.

mx6-sabresd, imx6sx-sdb and imx7d-pico have been successfully tested.

Thanks

APF6 SoM has AR8035. I confirm the issue and modifying the phy-mode property to "rgmii-id" fixes it.

Can you please apply the fix to the APF6 SoM (arch/arm/boot/dts/imx6qdl-apf6.dtsi) in your patch ?

Regards,

Signature · 9 lines
Sébastien Szymanski
Software engineer, Armadeus Systems
Tel: +33 (0)9 72 29 41 44
Fax: +33 (0)9 72 28 79 26

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Tested-by on v1 1/1
#10 lore raw
In reply to Fabio Estevam
Quoted 24 lines · Fabio Estevam

Commit 6d4cd041f0af ("net: phy: at803x: disable delay only for RGMII mode") exposed an issue on imx DTS files using AR8031/AR8035 PHYs.

The end result is that the boards can no longer obtain an IP address via UDHCP, for example.

Quoting Andrew Lunn:

"The problem here is, all the DTs were broken since day 0. However, because the PHY driver was also broken, nobody noticed and it worked. Now that the PHY driver has been fixed, all the bugs in the DTs now become an issue"

To fix this problem, the phy-mode property needs to be "rgmii-id", which has the following meaning as per Documentation/devicetree/bindings/net/ethernet.txt:

"RGMII with internal RX and TX delays provided by the PHY, the MAC should not add the RX or TX delays in this case)"

Tested on imx6-sabresd, imx6sx-sdb and imx7d-pico boards with successfully restored networking.

Quoted 8 lines

---
Hi,

Marking it as RFC in the hope of getting some feedback as to how to proceed with fixing this regression on 5.1-rc1.

It's not possible to know if there are other imx boards in tree that use AR803X, so here is an attempt.

The riotboard is affected, too:

arch/arm/boot/dts/imx6dl-riotboard.dts +1 -1
@@ -92,7 +92,7 @@
 &fec {
        pinctrl-names = "default";
        pinctrl-0 = <&pinctrl_enet>;
-       phy-mode = "rgmii";
+       phy-mode = "rgmii-id";
        phy-reset-gpios = <&gpio3 31 GPIO_ACTIVE_LOW>;
        interrupts-extended = <&gpio1 6 IRQ_TYPE_LEVEL_HIGH>,
                              <&intc 0 119 IRQ_TYPE_LEVEL_HIGH>

Marc

Signature · 4 lines
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel