[PATCH 0/5] ARM: dts: bcm470x: set MACs where missing

[PATCH 0/5] ARM: dts: bcm470x: set MACs where missing

[Rosen Penev]

I believe these predate the bcm,nvram driver, which makes this
possible. Downstream OpenWrt currently assigns these in userspace.

Rosen Penev (5):
  ARM: dts: bcm4709-linksys-ea9200: set WAN MAC
  ARM: dts: bcm47094-linksys-panamera: set WAN MAC
  ARM: dts: bcm4709-asus-rt-ac87u: specify switch
  ARM: dts: bcm4708-linksys-ea6300-v1: specify switch
  ARM: dts: bcm47094-phicomm-k3: specify switch

 .../broadcom/bcm4708-linksys-ea6300-v1.dts    | 45 +++++++++++++++++++
 .../dts/broadcom/bcm4709-asus-rt-ac87u.dts    | 37 +++++++++++++++
 .../dts/broadcom/bcm4709-linksys-ea9200.dts   |  7 +++
 .../broadcom/bcm47094-linksys-panamera.dts    |  7 +++
 .../boot/dts/broadcom/bcm47094-phicomm-k3.dts | 43 ++++++++++++++++++
 5 files changed, 139 insertions(+)

--
2.53.0

[PATCH 1/5] ARM: dts: bcm4709-linksys-ea9200: set WAN MAC

[Rosen Penev]

The MAC address from the stock firmare is offset by 1. Define it
properly to avoid having to override it in userspace.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 arch/arm/boot/dts/broadcom/bcm4709-linksys-ea9200.dts | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/arm/boot/dts/broadcom/bcm4709-linksys-ea9200.dts b/arch/arm/boot/dts/broadcom/bcm4709-linksys-ea9200.dts
index 2ba5adf2b7e7..7c6ad7787945 100644
--- a/arch/arm/boot/dts/broadcom/bcm4709-linksys-ea9200.dts
+++ b/arch/arm/boot/dts/broadcom/bcm4709-linksys-ea9200.dts
@@ -25,6 +25,10 @@ memory@0 {
 	nvram@1c080000 {
 		compatible = "brcm,nvram";
 		reg = <0x1c080000 0x180000>;
+
+		et2macaddr: et2macaddr {
+			#nvmem-cell-cells = <1>;
+		};
 	};
 
 	gpio-keys {
@@ -70,6 +74,9 @@ port@3 {
 
 		port@4 {
 			label = "wan";
+
+			nvmem-cells = <&et2macaddr 1>;
+			nvmem-cell-names = "mac-address";
 		};
 
 		port@5 {
-- 
2.53.0

[PATCH 2/5] ARM: dts: bcm47094-linksys-panamera: set WAN MAC

[Rosen Penev]

The MAC address from the stock firmare is offset by 1. Define it
properly to avoid having to override it in userspace.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 arch/arm/boot/dts/broadcom/bcm47094-linksys-panamera.dts | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/arm/boot/dts/broadcom/bcm47094-linksys-panamera.dts b/arch/arm/boot/dts/broadcom/bcm47094-linksys-panamera.dts
index 2b5c80d835e9..74161b76008a 100644
--- a/arch/arm/boot/dts/broadcom/bcm47094-linksys-panamera.dts
+++ b/arch/arm/boot/dts/broadcom/bcm47094-linksys-panamera.dts
@@ -25,6 +25,10 @@ memory@0 {
 	nvram@1c080000 {
 		compatible = "brcm,nvram";
 		reg = <0x1c080000 0x100000>;
+
+		et2macaddr: et2macaddr {
+			#nvmem-cell-cells = <1>;
+		};
 	};
 
 	gpio-keys {
@@ -230,6 +234,9 @@ port@3 {
 
 		port@4 {
 			label = "wan";
+
+			nvmem-cells = <&et2macaddr 1>;
+			nvmem-cell-names = "mac-address";
 		};
 
 		port@5 {
-- 
2.53.0

[PATCH 3/5] ARM: dts: bcm4709-asus-rt-ac87u: specify switch

[Rosen Penev]

bcm-ns.dtsi specifies a default layout that is not correct for the
RT-AC87U. Also allows setting the WAN MAC address properly.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 .../dts/broadcom/bcm4709-asus-rt-ac87u.dts    | 37 +++++++++++++++++++
 1 file changed, 37 insertions(+)

diff --git a/arch/arm/boot/dts/broadcom/bcm4709-asus-rt-ac87u.dts b/arch/arm/boot/dts/broadcom/bcm4709-asus-rt-ac87u.dts
index 59400217f8c3..f48c44ff01c0 100644
--- a/arch/arm/boot/dts/broadcom/bcm4709-asus-rt-ac87u.dts
+++ b/arch/arm/boot/dts/broadcom/bcm4709-asus-rt-ac87u.dts
@@ -77,6 +77,43 @@ &usb3_phy {
 	status = "okay";
 };
 
+&srab {
+	status = "okay";
+
+	ports {
+		port@0 {
+			label = "wan";
+
+			nvmem-cells = <&et1macaddr 1>;
+			nvmem-cell-names = "mac-address";
+		};
+
+		port@1 {
+			label = "lan1";
+		};
+
+		port@2 {
+			label = "lan2";
+		};
+
+		port@3 {
+			label = "lan3";
+		};
+
+		port@5 {
+			status = "disabled";
+		};
+
+		port@7 {
+			label = "cpu";
+		};
+
+		port@8 {
+			status = "disabled";
+		};
+	};
+};
+
 &nandcs {
 	partitions {
 		compatible = "fixed-partitions";
-- 
2.53.0

Re: [PATCH 3/5] ARM: dts: bcm4709-asus-rt-ac87u: specify switch

[Rafał Miłecki]

pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> bcm-ns.dtsi specifies a default layout that is not correct for the
> RT-AC87U. Also allows setting the WAN MAC address properly.

How the basic ports info from bcm-ns.dtsi is not correct for RT-AC87U?
I don't see you doing any overwriting or /delete-node/. It seems like
a standard NS device with BCM53011 and 8 ports (0, 1, 2, 3, 4, 5, 7,
8), three of them connected to no-SoC Ethernet interfaces.

Except for this unclear commit description, your changes look good, thanks!

-- 
Rafał

Re: [PATCH 3/5] ARM: dts: bcm4709-asus-rt-ac87u: specify switch

[Rosen Penev]

On Fri, Feb 20, 2026 at 3:18 AM Rafał Miłecki <zajec5@gmail.com> wrote:
>
> pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> > bcm-ns.dtsi specifies a default layout that is not correct for the
> > RT-AC87U. Also allows setting the WAN MAC address properly.
>
> How the basic ports info from bcm-ns.dtsi is not correct for RT-AC87U?
> I don't see you doing any overwriting or /delete-node/. It seems like
> a standard NS device with BCM53011 and 8 ports (0, 1, 2, 3, 4, 5, 7,
> 8), three of them connected to no-SoC Ethernet interfaces.
This comes from a local OpenWrt patch. I just added the nvmem
definition for WAN.

There's also the addition of labels.
>
> Except for this unclear commit description, your changes look good, thanks!
>
> --
> Rafał

Re: [PATCH 3/5] ARM: dts: bcm4709-asus-rt-ac87u: specify switch

[Rafał Miłecki]

pt., 20 lut 2026 o 22:18 Rosen Penev <rosenp@gmail.com> napisał(a):
> On Fri, Feb 20, 2026 at 3:18 AM Rafał Miłecki <zajec5@gmail.com> wrote:
> > pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> > > bcm-ns.dtsi specifies a default layout that is not correct for the
> > > RT-AC87U. Also allows setting the WAN MAC address properly.
> >
> > How the basic ports info from bcm-ns.dtsi is not correct for RT-AC87U?
> > I don't see you doing any overwriting or /delete-node/. It seems like
> > a standard NS device with BCM53011 and 8 ports (0, 1, 2, 3, 4, 5, 7,
> > 8), three of them connected to no-SoC Ethernet interfaces.
> This comes from a local OpenWrt patch. I just added the nvmem
> definition for WAN.
>
> There's also the addition of labels.

Then I guess someone else didn't try hard enough for proper
description. It happens with downstream code. Sometimes it suffers
from lower quality & hacks.

When submitting for upstream you can improve it though :-) You can
(and rather should) reword commit description so it's accurate.

-- 
Rafał

[PATCH 4/5] ARM: dts: bcm4708-linksys-ea6300-v1: specify switch

[Rosen Penev]

bcm-ns.dtsi specifies a default layout that is not correct for the
EA6300. Also allows setting the WAN MAC address properly.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 .../broadcom/bcm4708-linksys-ea6300-v1.dts    | 45 +++++++++++++++++++
 1 file changed, 45 insertions(+)

diff --git a/arch/arm/boot/dts/broadcom/bcm4708-linksys-ea6300-v1.dts b/arch/arm/boot/dts/broadcom/bcm4708-linksys-ea6300-v1.dts
index 0ed25bf71f0d..44cf867b223d 100644
--- a/arch/arm/boot/dts/broadcom/bcm4708-linksys-ea6300-v1.dts
+++ b/arch/arm/boot/dts/broadcom/bcm4708-linksys-ea6300-v1.dts
@@ -24,6 +24,10 @@ memory@0 {
 	nvram@1c080000 {
 		compatible = "brcm,nvram";
 		reg = <0x1c080000 0x180000>;
+
+		et0macaddr: et0macaddr {
+			#nvmem-cell-cells = <1>;
+		};
 	};
 
 	gpio-keys {
@@ -46,3 +50,44 @@ button-restart {
 &usb3_phy {
 	status = "okay";
 };
+
+&srab {
+	status = "okay";
+
+	ports {
+		port@0 {
+			label = "lan1";
+		};
+
+		port@1 {
+			label = "lan2";
+		};
+
+		port@2 {
+			label = "lan3";
+		};
+
+		port@3 {
+			label = "lan4";
+		};
+
+		port@4 {
+			label = "wan";
+
+			nvmem-cells = <&et0macaddr 1>;
+			nvmem-cell-names = "mac-address";
+		};
+
+		port@5 {
+			label = "cpu";
+		};
+
+		port@7 {
+			status = "disabled";
+		};
+
+		port@8 {
+			status = "disabled";
+		};
+	};
+};
-- 
2.53.0

[PATCH 5/5] ARM: dts: bcm47094-phicomm-k3: specify switch

[Rosen Penev]

bcm-ns.dtsi specifies a default layout that is not correct for the
K3. Also allows setting the WAN MAC address properly.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 .../boot/dts/broadcom/bcm47094-phicomm-k3.dts | 43 +++++++++++++++++++
 1 file changed, 43 insertions(+)

diff --git a/arch/arm/boot/dts/broadcom/bcm47094-phicomm-k3.dts b/arch/arm/boot/dts/broadcom/bcm47094-phicomm-k3.dts
index bb1bc4e61bc2..5ed284b8f156 100644
--- a/arch/arm/boot/dts/broadcom/bcm47094-phicomm-k3.dts
+++ b/arch/arm/boot/dts/broadcom/bcm47094-phicomm-k3.dts
@@ -19,6 +19,12 @@ memory@0 {
 		      <0x88000000 0x18000000>;
 	};
 
+	nvram@1c080000 {
+		et0macaddr: et0macaddr {
+			#nvmem-cell-cells = <1>;
+		};
+	};
+
 	gpio-keys {
 		compatible = "gpio-keys";
 
@@ -38,6 +44,43 @@ &usb3_phy {
 	status = "okay";
 };
 
+&srab {
+	status = "okay";
+
+	ports {
+		port@0 {
+			label = "lan1";
+		};
+
+		port@1 {
+			label = "lan2";
+		};
+
+		port@2 {
+			label = "lan3";
+		};
+
+		port@3 {
+			label = "wan";
+
+			nvmem-cells = <&et0macaddr 1>;
+			nvmem-cell-names = "mac-address";
+		};
+
+		port@5 {
+			label = "cpu";
+		};
+
+		port@7 {
+			status = "disabled";
+		};
+
+		port@8 {
+			status = "disabled";
+		};
+	};
+};
+
 &nandcs {
 	partitions {
 		compatible = "fixed-partitions";
-- 
2.53.0

Re: [PATCH 0/5] ARM: dts: bcm470x: set MACs where missing

[Rafał Miłecki]

pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> I believe these predate the bcm,nvram driver, which makes this
> possible. Downstream OpenWrt currently assigns these in userspace.

Changes look good, I just posted one comment about commit description.
Thanks for submitting this!

If you check
git log --oneline arch/arm/boot/dts/broadcom/bcm470*dts
git log --oneline -- arch/arm/boot/dts/bcm470*dts
the prefix always used for NS patches was:
"ARM: dts: BCM5301X: "

I guess it may be preferred / required to follow that.

-- 
Rafał

Re: [PATCH 0/5] ARM: dts: bcm470x: set MACs where missing

[Rosen Penev]

On Fri, Feb 20, 2026 at 3:21 AM Rafał Miłecki <zajec5@gmail.com> wrote:
>
> pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> > I believe these predate the bcm,nvram driver, which makes this
> > possible. Downstream OpenWrt currently assigns these in userspace.
>
> Changes look good, I just posted one comment about commit description.
> Thanks for submitting this!
>
> If you check
> git log --oneline arch/arm/boot/dts/broadcom/bcm470*dts
> git log --oneline -- arch/arm/boot/dts/bcm470*dts
> the prefix always used for NS patches was:
> "ARM: dts: BCM5301X: "
>
> I guess it may be preferred / required to follow that.
I didn't want to make the subject line huge.

Actually in such a case, would it make sense to fold into one commit?
>
> --
> Rafał

Re: [PATCH 0/5] ARM: dts: bcm470x: set MACs where missing

[Rafał Miłecki]

pt., 20 lut 2026 o 22:20 Rosen Penev <rosenp@gmail.com> napisał(a):
> On Fri, Feb 20, 2026 at 3:21 AM Rafał Miłecki <zajec5@gmail.com> wrote:
> >
> > pt., 20 lut 2026 o 03:14 Rosen Penev <rosenp@gmail.com> napisał(a):
> > > I believe these predate the bcm,nvram driver, which makes this
> > > possible. Downstream OpenWrt currently assigns these in userspace.
> >
> > Changes look good, I just posted one comment about commit description.
> > Thanks for submitting this!
> >
> > If you check
> > git log --oneline arch/arm/boot/dts/broadcom/bcm470*dts
> > git log --oneline -- arch/arm/boot/dts/bcm470*dts
> > the prefix always used for NS patches was:
> > "ARM: dts: BCM5301X: "
> >
> > I guess it may be preferred / required to follow that.
> I didn't want to make the subject line huge.
>
> Actually in such a case, would it make sense to fold into one commit?

I'll leave it to Florian to comment on. He's more in place to decide than me.

--
Rafał