[PATCH 0/2] Add device tree for Asus IPMI Card

[PATCH 0/2] Add device tree for Asus IPMI Card

[Anirudh Srinivasan]

Adds support for Asus IPMI Card [1], which is a PCIe card with an
AST2600 on it that provides BMC functionality to any host without an
onboard BMC as long as it has a PCIe slot.

Currently supported functionality includes UART, NIC, booting from SPI,
KVM functionality(usb gadget for host, reading VGA framebuffer from
host) and LED/Power Control via GPIOs. This functionality has been
tested on an OpenBMC build available here [2]. This requires a modified
u-boot device tree [3], which I can also submit if needed.

The card supports some additional functionality in it's stock firmware
like fan headers with fan speed control, host bios flashing via SPI, PSU
monitoring via an SMBUS connector, a custom ASUS SMBUS connector to talk
to select ASUS motherboards that support it and exposing a IPMI device
via PCIe to the host. These are unsupported at the moment.


[1] https://www.asus.com/supportonly/ipmi%20expansion%20card/helpdesk_manual/
[2] https://github.com/openbmc/openbmc/compare/master...Genius1237:openbmc:asus-ipmi-card
[3] https://github.com/Genius1237/u-boot/commit/1b1b7daa85f6c998e5f404296b3da43077a2758e


Anirudh Srinivasan (2):
  dt-bindings: arm: aspeed: Add Asus IPMI card
  ARM: dts: aspeed: Add Asus IPMI card BMC

 .../bindings/arm/aspeed/aspeed.yaml           |   1 +
 arch/arm/boot/dts/aspeed/Makefile             |   1 +
 .../dts/aspeed/aspeed-bmc-asus-ipmi-card.dts  | 137 ++++++++++++++++++
 3 files changed, 139 insertions(+)
 create mode 100644 arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts

-- 
2.52.0

[PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Document the new comptaibles for Asus IPMI card

Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
---
 Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
index 9298c1a75dd1..b2d20341f8eb 100644
--- a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
+++ b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
@@ -79,6 +79,7 @@ properties:
               - ampere,mtmitchell-bmc
               - aspeed,ast2600-evb
               - aspeed,ast2600-evb-a1
+              - asus,ipmi-card-bmc
               - asus,x4tf-bmc
               - facebook,bletchley-bmc
               - facebook,catalina-bmc
-- 
2.52.0

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Andrew Jeffery]

On Sun, 2026-01-11 at 14:10 -0600, Anirudh Srinivasan wrote:
> Document the new comptaibles for Asus IPMI card
> 
> Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
> ---
>  Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> index 9298c1a75dd1..b2d20341f8eb 100644
> --- a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> +++ b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> @@ -79,6 +79,7 @@ properties:
>                - ampere,mtmitchell-bmc
>                - aspeed,ast2600-evb
>                - aspeed,ast2600-evb-a1
> +              - asus,ipmi-card-bmc

It's a bit bike-sheddy, however: the pattern tends to be
${vendor},${platform}-bmc, but as the platform can't be specified and
the card's function is the BMC itself, I'd go with "asus,ipmi-
expansion-card" (I couldn't immediately find a useful identifier other
than the product name).

Andrew

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Krzysztof Kozlowski]

On Mon, Jan 12, 2026 at 12:01:27PM +1030, Andrew Jeffery wrote:
> On Sun, 2026-01-11 at 14:10 -0600, Anirudh Srinivasan wrote:
> > Document the new comptaibles for Asus IPMI card
> > 
> > Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
> > ---
> >  Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> > index 9298c1a75dd1..b2d20341f8eb 100644
> > --- a/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> > +++ b/Documentation/devicetree/bindings/arm/aspeed/aspeed.yaml
> > @@ -79,6 +79,7 @@ properties:
> >                - ampere,mtmitchell-bmc
> >                - aspeed,ast2600-evb
> >                - aspeed,ast2600-evb-a1
> > +              - asus,ipmi-card-bmc
> 
> It's a bit bike-sheddy, however: the pattern tends to be
> ${vendor},${platform}-bmc, but as the platform can't be specified and
> the card's function is the BMC itself, I'd go with "asus,ipmi-
> expansion-card" (I couldn't immediately find a useful identifier other
> than the product name).

This should include some model name or at least soc. What if you have
IPMI card for ast2600 and later completely different for ast2700?

Best regards,
Krzysztof

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Hi Krzysztof,

On Mon, Jan 12, 2026 at 5:47 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> >
> > It's a bit bike-sheddy, however: the pattern tends to be
> > ${vendor},${platform}-bmc, but as the platform can't be specified and
> > the card's function is the BMC itself, I'd go with "asus,ipmi-
> > expansion-card" (I couldn't immediately find a useful identifier other
> > than the product name).
>
> This should include some model name or at least soc. What if you have
> IPMI card for ast2600 and later completely different for ast2700?

Personally, I was okay with Andrew's suggestion of naming it
"asus,ipmi-expansion-card" because that's the official product name
and that exact term works well for online searches.

This seems like a pretty niche product Asus announced 4 years ago and
made in very low volumes. I'm not sure if they'll make a new one with
an ast2700, given how the modern trend is to have BMCs on
motherboards. What do you think about the naming Andrew?

-- 
Regards
Anirudh Srinivasan

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Andrew Jeffery]

On Mon, 2026-01-12 at 15:21 -0600, Anirudh Srinivasan wrote:
> Hi Krzysztof,
> 
> On Mon, Jan 12, 2026 at 5:47 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > 
> > > 
> > > It's a bit bike-sheddy, however: the pattern tends to be
> > > ${vendor},${platform}-bmc, but as the platform can't be specified and
> > > the card's function is the BMC itself, I'd go with "asus,ipmi-
> > > expansion-card" (I couldn't immediately find a useful identifier other
> > > than the product name).
> > 
> > This should include some model name or at least soc. What if you have
> > IPMI card for ast2600 and later completely different for ast2700?
> 
> Personally, I was okay with Andrew's suggestion of naming it
> "asus,ipmi-expansion-card" because that's the official product name
> and that exact term works well for online searches.
> 
> This seems like a pretty niche product Asus announced 4 years ago and
> made in very low volumes. I'm not sure if they'll make a new one with
> an ast2700, given how the modern trend is to have BMCs on
> motherboards. What do you think about the naming Andrew?

I agree that the product seems niche. There's practically nothing on
their website on the marketing side - from searching around the details
are pretty much limited to the support pages. I can't really speak to
future stuff like an AST2700-based design though, who knows.

A couple of numbers turned up that might be helpful:

   1. The Quick Start Guide[1] seems to use "E21524", which is also
      used by some random blog[2] to identify it. 
   2. There's "R1.04" on the silkscreen. 

Perhaps we could incorporate either of those?

 * asus,e21524-ipmi-expansion-card
 * asus,ipmi-expansion-card-r1-04

However, they're not without some risk:

   1. It's hard to tell whether E21524 is properly representative
   2. R1.04 may also problematic as an AST2700-based card will likely
      restart the numbering and risk a collision

Otherwise, I guess there's:

 * asus,ipmi-expansion-card-ast2600

Andrew

[1]: https://www.asus.com/us/supportonly/ipmi%20expansion%20card/helpdesk_manual/
[2]: https://blog.liaosirui.com/%E7%B3%BB%E7%BB%9F%E8%BF%90%E7%BB%B4/A.%E6%9C%8D%E5%8A%A1%E5%99%A8/%E5%B9%B3%E5%8F%B0%E7%AE%A1%E7%90%86/BMC%E5%BA%95%E6%9D%BF%E7%AE%A1%E7%90%86%E6%8E%A7%E5%88%B6%E5%99%A8/IPMI%E8%BF%9C%E7%A8%8B%E7%AE%A1%E7%90%86%E5%8D%A1.html

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Hi Andrew

On Mon, Jan 12, 2026 at 6:57 PM Andrew Jeffery
<andrew@codeconstruct.com.au> wrote:
>
> Perhaps we could incorporate either of those?
>
>  * asus,e21524-ipmi-expansion-card
>  * asus,ipmi-expansion-card-r1-04
>
> However, they're not without some risk:
>
>    1. It's hard to tell whether E21524 is properly representative
>    2. R1.04 may also problematic as an AST2700-based card will likely
>       restart the numbering and risk a collision
>
> Otherwise, I guess there's:
>
>  * asus,ipmi-expansion-card-ast2600
>

There seem to be multiple revisions of this card. The photos have the
revision number printed on the board and the manuals have the E21XXX
number on them.
R1.01 - E21362 photo [1] manual[2]
R1.04 - E21524 photo [3] manual[4]

Mine is the latter. There seem to be some differences in the layout of
the different headers on the board, but the same set of headers are on
both boards. There is only one fw image on Asus's site, so I guess
they're the same from a sw point of view.

I've linked a screenshot [5] from the webui on the stock fw and it
says that the firmware model is "KOMMANDO". Asus also seems to name
the fw update file for this board KODO1140.ima (KODO is short for
KOMMANDO?, 1140 is the version number).

I looked at other Asus Motherboard onboard BMC fw updates and they
seem to be named <Mobo Model><version>.ima. Example board "Pro WS
W680M" [6], fw update name W680M1121.ima, another example board "Pro
WS W790E" [7], fw update name W7901167.ima

"asus,kommando-ipmi-expansion-card" is a bit long, so maybe we should
go with "asus,kodo-ipmi-expansion-card". SEO for either of the 2 terms
is not very good, but that's not what we're looking for here are we?

[1] https://uk.store.asus.com/media/catalog/product/9/0/90mc0ah0-mvuby0-7.jpg
[2] https://manuals.plus/m/8bcc92c2f7a875eda34d546e0e297d339ea55863fea0c6abcf0f338cd26f299d_optim.pdf
[3] https://i.ebayimg.com/images/g/Um8AAeSw27JpOYcU/s-l1600.jpg
[4] https://dlcdnets.asus.com/pub/ASUS/mb/Add-on_card/IPMI_EXPANSION_CARD/E21524_IPMI_Card_QSG_V2_WEB.pdf?model=IPMI%20EXPANSION%20CARD
[5] https://ibb.co/kg96CsYY
[6] https://www.asus.com/us/motherboards-components/motherboards/workstation/pro-ws-w680m-ace-se/helpdesk_bios?model2Name=Pro-WS-W680M-ACE-SE
[7] https://www.asus.com/us/motherboards-components/motherboards/workstation/pro-ws-w790e-sage-se/helpdesk_bios?model2Name=Pro-WS-W790E-SAGE-SE

-- 
Regards
Anirudh Srinivasan

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Andrew Jeffery]

On Tue, 2026-01-13 at 16:28 -0600, Anirudh Srinivasan wrote:
> Hi Andrew
> 
> On Mon, Jan 12, 2026 at 6:57 PM Andrew Jeffery
> <andrew@codeconstruct.com.au> wrote:
> > 
> > Perhaps we could incorporate either of those?
> > 
> >  * asus,e21524-ipmi-expansion-card
> >  * asus,ipmi-expansion-card-r1-04
> > 
> > However, they're not without some risk:
> > 
> >    1. It's hard to tell whether E21524 is properly representative
> >    2. R1.04 may also problematic as an AST2700-based card will likely
> >       restart the numbering and risk a collision
> > 
> > Otherwise, I guess there's:
> > 
> >  * asus,ipmi-expansion-card-ast2600
> > 
> 
> There seem to be multiple revisions of this card. The photos have the
> revision number printed on the board and the manuals have the E21XXX
> number on them.
> R1.01 - E21362 photo [1] manual[2]
> R1.04 - E21524 photo [3] manual[4]
> 
> Mine is the latter. There seem to be some differences in the layout of
> the different headers on the board, but the same set of headers are on
> both boards. There is only one fw image on Asus's site, so I guess
> they're the same from a sw point of view.

Ah, interesting. So the R1.01 card also has an 'E' number silkscreened
on it - 'E248779', though this is different to E21362 from the manual.
E248779 seems to alias to a lot of other ASUS things. Given the
variance and aliasing it's probably a good idea to avoid them.

> 
> I've linked a screenshot [5] from the webui on the stock fw and it
> says that the firmware model is "KOMMANDO". Asus also seems to name
> the fw update file for this board KODO1140.ima (KODO is short for
> KOMMANDO?, 1140 is the version number).

> I looked at other Asus Motherboard onboard BMC fw updates and they
> seem to be named <Mobo Model><version>.ima. Example board "Pro WS
> W680M" [6], fw update name W680M1121.ima, another example board "Pro
> WS W790E" [7], fw update name W7901167.ima
> 
> "asus,kommando-ipmi-expansion-card" is a bit long, so maybe we should
> go with "asus,kodo-ipmi-expansion-card". SEO for either of the 2 terms
> is not very good, but that's not what we're looking for here are we?

So downloading the releases and poking around a bit, there's this:

   > for f in */*.ima; echo $f; dd if=$f bs=$(math 0x100) count=1 skip=$(math 0x3ff00) 2>/dev/null | strings | sed 's/^/\t/'; echo; end
   IPMI_EXPANSION_CARD-1.1.26/KODO1126.ima
   	$MODULE$
   	ast2600e
   	UFW_VERSION=1.01.26
   	FW_DATE=Jan 9 2023
   	FW_BUILDTIME=10:23:36 UTC
   	FW_DESC=AST2600EVB SPX-13 TB2
   	FW_PRODUCTID=1
   	FW_RELEASEID=RR9
   	FW_CODEBASEVERSION=5.X
   
   IPMI_EXPANSION_CARD-1.1.33/KODO1133.ima
   	$MODULE$
   	ast2600e
   	UFW_VERSION=1.01.33
   	FW_DATE=Jul 17 2023
   	FW_BUILDTIME=09:25:13 UTC
   	FW_DESC=AST2600EVB SPX-13 TB2
   	FW_PRODUCTID=1
   	FW_RELEASEID=RR9
   	FW_CODEBASEVERSION=5.X
   
   IPMI_EXPANSION_CARD-1.1.34/KODO1134.ima
   	$MODULE$
   	ast2600e
   	UFW_VERSION=1.01.34
   	FW_DATE=Feb 27 2024
   	FW_BUILDTIME=07:45:26 UTC
   	FW_DESC=AST2600EVB SPX-13 TB2
   	FW_PRODUCTID=1
   	FW_RELEASEID=RR9
   	FW_CODEBASEVERSION=5.X
   
   IPMI_EXPANSION_CARD-1.1.35/KODO1135.ima
   	$MODULE$
   	ast2600e
   	UFW_VERSION=1.01.35
   	FW_DATE=Aug 20 2024
   	FW_BUILDTIME=08:32:35 UTC
   	FW_DESC=AST2600EVB SPX-13 TB2
   	FW_PRODUCTID=1
   	FW_RELEASEID=RR9
   	FW_CODEBASEVERSION=5.X
   
   IPMI_EXPANSION_CARD-1.1.40/KODO1140.ima
   	$MODULE$
   	Oast2600e
   	UFW_VERSION=1.01.40
   	FW_DATE=Apr 15 2025
   	FW_BUILDTIME=09:52:26 UTC
   	FW_DESC=AST2600EVB KOMMANDO SPX-13 TB2
   	FW_PRODUCTID=1
   	FW_RELEASEID=RR9
   	FW_CODEBASEVERSION=5.X

KODO is pretty consistent in the firmware update file name. KOMMANDO as
the product identifier also seems reasonable, even if it only appears
in FW_DESC in the latest release. KODO -> KOMMANDO seems reasonable.

A brief bit of binwalking suggests it's an AMI MegaRAC implementation,
and so I expect 'SPX-13' in FW_DESC refers to MegaRAC SP-X[1]. The
significance of 'TB2' isn't yet unclear to me.

I'm not super concerned about the length of "asus,kommando-ipmi-
expansion-card" - I'd prefer it over incorporating the 'kodo'
contraction.

I guess it's still unclear whether 'kommando' is a (future) line of
products or refers to the specific AST2600-based design. Perhaps we
could solve a few problems with 'asus,kommando-ast2600' and including
the phrase 'IPMI expansion card' in the binding description for
something searchable.

Andrew

[1]: https://9443417.fs1.hubspotusercontent-na1.net/hubfs/9443417/Data_Sheets/Firmware_Solutions/MegaRAC_SP-X_Data_Sheet_PUB.pdf

Re: [PATCH 1/2] dt-bindings: arm: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

On Wed, Jan 14, 2026 at 7:03 PM Andrew Jeffery
<andrew@codeconstruct.com.au> wrote:
>
> KODO is pretty consistent in the firmware update file name. KOMMANDO as
> the product identifier also seems reasonable, even if it only appears
> in FW_DESC in the latest release. KODO -> KOMMANDO seems reasonable.
>
> A brief bit of binwalking suggests it's an AMI MegaRAC implementation,
> and so I expect 'SPX-13' in FW_DESC refers to MegaRAC SP-X[1]. The
> significance of 'TB2' isn't yet unclear to me.

Yup. The screenshot in the PDF you linked looks a lot like the webui
on stock firmware. AMI has sources for a bunch of stuff ( "unmodified"
source tarballs and patches to apply on top of them manually) over
here [1].

>
> I'm not super concerned about the length of "asus,kommando-ipmi-
> expansion-card" - I'd prefer it over incorporating the 'kodo'
> contraction.

Understood.

>
> I guess it's still unclear whether 'kommando' is a (future) line of
> products or refers to the specific AST2600-based design. Perhaps we
> could solve a few problems with 'asus,kommando-ast2600' and including
> the phrase 'IPMI expansion card' in the binding description for
> something searchable.

I will make the compatible "asus,ast2600-kommando-ipmi-card" and have
the model name in the DT contain the phrase "IPMI expansion card" for
searchability. You mention about adding something to the binding
description (that'd end up being more than a one line change to the
bindings) but we could just have the DT model name have that
information.

I'll send out a second version of this series shortly. Thank you for
bearing with me on this. Naming things is hard.

[1] https://github.com/ami-megarac/OSSW-v13/tree/master/Core/

--
Regards
Anirudh Srinivasan

[PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Add device tree for Asus IPMI card, an AST2600 BMC PCIe card

Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
---
 arch/arm/boot/dts/aspeed/Makefile             |   1 +
 .../dts/aspeed/aspeed-bmc-asus-ipmi-card.dts  | 136 ++++++++++++++++++
 2 files changed, 137 insertions(+)
 create mode 100644 arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts

diff --git a/arch/arm/boot/dts/aspeed/Makefile b/arch/arm/boot/dts/aspeed/Makefile
index 9adf9278dc94..ff40d9ab88b7 100644
--- a/arch/arm/boot/dts/aspeed/Makefile
+++ b/arch/arm/boot/dts/aspeed/Makefile
@@ -15,6 +15,7 @@ dtb-$(CONFIG_ARCH_ASPEED) += \
 	aspeed-bmc-asrock-spc621d8hm3.dtb \
 	aspeed-bmc-asrock-x570d4u.dtb \
 	aspeed-bmc-asus-x4tf.dtb \
+	aspeed-bmc-asus-ipmi-card.dtb \
 	aspeed-bmc-bytedance-g220a.dtb \
 	aspeed-bmc-delta-ahe50dc.dtb \
 	aspeed-bmc-facebook-bletchley.dtb \
diff --git a/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
new file mode 100644
index 000000000000..fea94360245c
--- /dev/null
+++ b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
@@ -0,0 +1,136 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+// Copyright 2025 Anirudh Srinivasan
+
+/dts-v1/;
+
+#include "aspeed-g6.dtsi"
+#include "aspeed-g6-pinctrl.dtsi"
+#include <dt-bindings/i2c/i2c.h>
+#include <dt-bindings/gpio/aspeed-gpio.h>
+
+/ {
+	model = "ASUS IPMI Card BMC";
+	compatible = "asus,ipmi-card-bmc", "aspeed,ast2600";
+
+	aliases {
+		serial4 = &uart5;
+	};
+
+	chosen {
+		stdout-path = "serial4:115200n8";
+	};
+
+	memory@80000000 {
+		device_type = "memory";
+		reg = <0x80000000 0x40000000>;
+	};
+
+	reserved-memory {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+
+		video_engine_memory: video {
+			size = <0x04000000>;
+			alignment = <0x01000000>;
+			compatible = "shared-dma-pool";
+			reusable;
+		};
+	};
+
+	leds {
+		compatible = "gpio-leds";
+
+		led-fault {
+			gpios = <&gpio1 ASPEED_GPIO(C, 5) GPIO_ACTIVE_HIGH>;
+			panic-indicator;
+			default-state = "off";
+		};
+
+		led-heartbeat {
+			gpios = <&gpio0 ASPEED_GPIO(A, 7) GPIO_ACTIVE_LOW>;
+			linux,default-trigger = "timer";
+		};
+
+	};
+};
+
+&mdio2 {
+	status = "okay";
+
+	ethphy2: ethernet-phy@0 {
+		compatible = "ethernet-phy-ieee802.3-c22";
+		reg = <0>;
+	};
+};
+
+&mac2 {
+	status = "okay";
+
+	/* Bootloader sets up the MAC to insert delay */
+	phy-mode = "rgmii";
+	phy-handle = <&ethphy2>;
+	pinctrl-names = "default";
+	pinctrl-0 = <&pinctrl_rgmii3_default>;
+};
+
+
+&fmc {
+	status = "okay";
+
+	flash@0 {
+		status = "okay";
+		m25p,fast-read;
+		label = "bmc";
+		spi-max-frequency = <50000000>;
+#include "openbmc-flash-layout-64.dtsi"
+	};
+};
+
+&video {
+	status = "okay";
+	memory-region = <&video_engine_memory>;
+};
+
+&vhub {
+	status = "okay";
+};
+
+&gpio0 {
+	gpio-line-names =
+	/*A0 0*/	"", "", "", "", "", "", "", "BMC_HBLED",
+	/*B0 8*/	"", "", "", "", "", "", "", "",
+	/*C0 16*/	"", "", "", "", "", "", "", "",
+	/*D0 24*/	"", "", "", "", "", "", "", "",
+	/*E0 32*/	"", "", "", "", "", "", "", "",
+	/*F0 40*/	"", "", "", "", "", "", "", "",
+	/*G0 48*/	"", "", "", "", "", "", "", "",
+	/*H0 56*/	"", "", "", "", "", "", "", "",
+	/*I0 64*/	"", "", "", "BMC_RSTBTN", "", "", "", "",
+	/*J0 72*/	"", "", "", "", "", "", "", "",
+	/*K0 80*/	"", "", "", "", "", "", "", "",
+	/*L0 88*/	"", "", "", "", "", "", "", "",
+	/*M0 96*/	"", "", "", "", "", "", "", "",
+	/*N0 104*/	"", "", "", "", "", "", "", "",
+	/*O0 112*/	"", "", "", "", "", "", "", "",
+	/*P0 120*/	"", "", "", "", "", "", "", "",
+	/*Q0 128*/	"", "", "", "", "", "", "", "",
+	/*R0 136*/	"", "", "", "", "", "", "", "",
+	/*S0 144*/	"", "", "", "", "", "", "", "",
+	/*T0 152*/	"", "", "", "", "", "", "", "",
+	/*U0 160*/	"", "", "", "", "", "", "", "",
+	/*V0 168*/	"", "", "", "", "BMC_PWRBTN", "", "MB_S0_DETECT", "",
+	/*W0 176*/	"", "", "", "", "", "", "", "",
+	/*X0 184*/	"", "", "", "", "", "", "", "",
+	/*Y0 192*/	"", "", "", "", "", "", "", "",
+	/*Z0 200*/	"", "", "", "", "", "", "", "";
+};
+
+&gpio1 {
+	gpio-line-names =
+	/*18A0 0*/  "","","","","","","","",
+	/*18B0 8*/  "","","","","","","","",
+	/*18C0 16*/ "","","","","","BMC_MLED","","",
+	/*18D0 24*/ "","","","","","","","",
+	/*18E0 32*/ "","","","","","","","";
+};
-- 
2.52.0

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Andrew Lunn]

> +&mac2 {
> +	status = "okay";
> +
> +	/* Bootloader sets up the MAC to insert delay */
> +	phy-mode = "rgmii";

There has been many discussions about this. This is broken, 'rgmii'
means the PCB adds the delays, not the MAC/PHY pair.

Please drop Ethernet support until aspeed fix the broken MAC driver.

       Andrew

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

On Sun, Jan 11, 2026 at 3:09 PM Andrew Lunn <andrew@lunn.ch> wrote:
>
> > +&mac2 {
> > +     status = "okay";
> > +
> > +     /* Bootloader sets up the MAC to insert delay */
> > +     phy-mode = "rgmii";
>
> There has been many discussions about this. This is broken, 'rgmii'
> means the PCB adds the delays, not the MAC/PHY pair.
>
> Please drop Ethernet support until aspeed fix the broken MAC driver.
>
>        Andrew

Understood, I will remove it for now.

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Andrew Jeffery]

Hi Anirudh,

On Sun, 2026-01-11 at 14:10 -0600, Anirudh Srinivasan wrote:
> Add device tree for Asus IPMI card, an AST2600 BMC PCIe card
> 
> Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
> ---
>  arch/arm/boot/dts/aspeed/Makefile             |   1 +
>  .../dts/aspeed/aspeed-bmc-asus-ipmi-card.dts  | 136 ++++++++++++++++++
>  2 files changed, 137 insertions(+)
>  create mode 100644 arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> 
> diff --git a/arch/arm/boot/dts/aspeed/Makefile b/arch/arm/boot/dts/aspeed/Makefile
> index 9adf9278dc94..ff40d9ab88b7 100644
> --- a/arch/arm/boot/dts/aspeed/Makefile
> +++ b/arch/arm/boot/dts/aspeed/Makefile
> @@ -15,6 +15,7 @@ dtb-$(CONFIG_ARCH_ASPEED) += \
>  	aspeed-bmc-asrock-spc621d8hm3.dtb \
>  	aspeed-bmc-asrock-x570d4u.dtb \
>  	aspeed-bmc-asus-x4tf.dtb \
> +	aspeed-bmc-asus-ipmi-card.dtb \
>  	aspeed-bmc-bytedance-g220a.dtb \
>  	aspeed-bmc-delta-ahe50dc.dtb \
>  	aspeed-bmc-facebook-bletchley.dtb \
> diff --git a/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> new file mode 100644
> index 000000000000..fea94360245c
> --- /dev/null
> +++ b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> @@ -0,0 +1,136 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +// Copyright 2025 Anirudh Srinivasan
> +
> +/dts-v1/;
> +
> +#include "aspeed-g6.dtsi"
> +#include "aspeed-g6-pinctrl.dtsi"
> +#include <dt-bindings/i2c/i2c.h>
> +#include <dt-bindings/gpio/aspeed-gpio.h>
> +
> +/ {
> +	model = "ASUS IPMI Card BMC";

I'd drop the "BMC" from the model description as well (wrt the
discussion on patch 1/2).

> +	compatible = "asus,ipmi-card-bmc", "aspeed,ast2600";
> +
> +	aliases {
> +		serial4 = &uart5;
> +	};
> +
> +	chosen {
> +		stdout-path = "serial4:115200n8";
> +	};
> +
> +	memory@80000000 {
> +		device_type = "memory";
> +		reg = <0x80000000 0x40000000>;
> +	};
> +
> +	reserved-memory {
> +		#address-cells = <1>;
> +		#size-cells = <1>;
> +		ranges;
> +
> +		video_engine_memory: video {
> +			size = <0x04000000>;
> +			alignment = <0x01000000>;
> +			compatible = "shared-dma-pool";
> +			reusable;
> +		};
> +	};
> +
> +	leds {
> +		compatible = "gpio-leds";
> +
> +		led-fault {
> +			gpios = <&gpio1 ASPEED_GPIO(C, 5) GPIO_ACTIVE_HIGH>;
> +			panic-indicator;
> +			default-state = "off";
> +		};
> +
> +		led-heartbeat {
> +			gpios = <&gpio0 ASPEED_GPIO(A, 7) GPIO_ACTIVE_LOW>;
> +			linux,default-trigger = "timer";
> +		};
> +
> +	};
> +};
> +
> +&mdio2 {
> +	status = "okay";
> +
> +	ethphy2: ethernet-phy@0 {
> +		compatible = "ethernet-phy-ieee802.3-c22";
> +		reg = <0>;
> +	};
> +};
> +
> +&mac2 {

This looks good in general.

For the purpose of review I've started requesting that phandle
references be ordered alphabetically, because it's tedious to confirm
they're ordered by unit address, and it's nice to have consistency
across the files. Can you please fix that along with Andrew Lunn's
concern about the phy-mode?

Andrew

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Hi Andrew,

On Sun, Jan 11, 2026 at 7:47 PM Andrew Jeffery
<andrew@codeconstruct.com.au> wrote:
>
> > +/ {
> > +     model = "ASUS IPMI Card BMC";
>
> I'd drop the "BMC" from the model description as well (wrt the
> discussion on patch 1/2).

Yes, I realized sticking to its official name everywhere might just
make better sense.

I'll make this name "ASUS IPMI Expansion Card", change the compatible
to "asus,ipmi-expansion-card" and the dtb name to
"aspeed-bmc-asus-ipmi-expansion-card.dtb".

>
> This looks good in general.
>
> For the purpose of review I've started requesting that phandle
> references be ordered alphabetically, because it's tedious to confirm
> they're ordered by unit address, and it's nice to have consistency
> across the files. Can you please fix that along with Andrew Lunn's
> concern about the phy-mode?

Will fix this up and resend another series. Thanks for your quick response.

--
Regards
Anirudh Srinivasan

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Krzysztof Kozlowski]

On Sun, Jan 11, 2026 at 02:10:32PM -0600, Anirudh Srinivasan wrote:
> Add device tree for Asus IPMI card, an AST2600 BMC PCIe card
> 
> Signed-off-by: Anirudh Srinivasan <anirudhsriniv@gmail.com>
> ---
>  arch/arm/boot/dts/aspeed/Makefile             |   1 +
>  .../dts/aspeed/aspeed-bmc-asus-ipmi-card.dts  | 136 ++++++++++++++++++
>  2 files changed, 137 insertions(+)
>  create mode 100644 arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> 
> diff --git a/arch/arm/boot/dts/aspeed/Makefile b/arch/arm/boot/dts/aspeed/Makefile
> index 9adf9278dc94..ff40d9ab88b7 100644
> --- a/arch/arm/boot/dts/aspeed/Makefile
> +++ b/arch/arm/boot/dts/aspeed/Makefile
> @@ -15,6 +15,7 @@ dtb-$(CONFIG_ARCH_ASPEED) += \
>  	aspeed-bmc-asrock-spc621d8hm3.dtb \
>  	aspeed-bmc-asrock-x570d4u.dtb \
>  	aspeed-bmc-asus-x4tf.dtb \
> +	aspeed-bmc-asus-ipmi-card.dtb \
>  	aspeed-bmc-bytedance-g220a.dtb \
>  	aspeed-bmc-delta-ahe50dc.dtb \
>  	aspeed-bmc-facebook-bletchley.dtb \
> diff --git a/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> new file mode 100644
> index 000000000000..fea94360245c
> --- /dev/null
> +++ b/arch/arm/boot/dts/aspeed/aspeed-bmc-asus-ipmi-card.dts
> @@ -0,0 +1,136 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +// Copyright 2025 Anirudh Srinivasan
> +
> +/dts-v1/;
> +
> +#include "aspeed-g6.dtsi"
> +#include "aspeed-g6-pinctrl.dtsi"
> +#include <dt-bindings/i2c/i2c.h>
> +#include <dt-bindings/gpio/aspeed-gpio.h>
> +
> +/ {
> +	model = "ASUS IPMI Card BMC";
> +	compatible = "asus,ipmi-card-bmc", "aspeed,ast2600";
> +
> +	aliases {
> +		serial4 = &uart5;
> +	};
> +
> +	chosen {
> +		stdout-path = "serial4:115200n8";
> +	};
> +
> +	memory@80000000 {
> +		device_type = "memory";
> +		reg = <0x80000000 0x40000000>;
> +	};
> +
> +	reserved-memory {
> +		#address-cells = <1>;
> +		#size-cells = <1>;
> +		ranges;
> +
> +		video_engine_memory: video {
> +			size = <0x04000000>;
> +			alignment = <0x01000000>;
> +			compatible = "shared-dma-pool";
> +			reusable;
> +		};
> +	};
> +
> +	leds {
> +		compatible = "gpio-leds";
> +
> +		led-fault {
> +			gpios = <&gpio1 ASPEED_GPIO(C, 5) GPIO_ACTIVE_HIGH>;
> +			panic-indicator;
> +			default-state = "off";

Missing color and function. Or at least label.

> +		};
> +
> +		led-heartbeat {
> +			gpios = <&gpio0 ASPEED_GPIO(A, 7) GPIO_ACTIVE_LOW>;
> +			linux,default-trigger = "timer";
> +		};
> +

Drop redundant blank lines.

> +	};
> +};
> +
> +&mdio2 {
> +	status = "okay";
> +
> +	ethphy2: ethernet-phy@0 {
> +		compatible = "ethernet-phy-ieee802.3-c22";
> +		reg = <0>;
> +	};
> +};
> +
> +&mac2 {
> +	status = "okay";
> +
> +	/* Bootloader sets up the MAC to insert delay */
> +	phy-mode = "rgmii";
> +	phy-handle = <&ethphy2>;
> +	pinctrl-names = "default";
> +	pinctrl-0 = <&pinctrl_rgmii3_default>;
> +};
> +
> +

Like this as well. It's a sign of sloppy coding.

> +&fmc {
> +	status = "okay";
> +
> +	flash@0 {
> +		status = "okay";
> +		m25p,fast-read;
> +		label = "bmc";
> +		spi-max-frequency = <50000000>;
> +#include "openbmc-flash-layout-64.dtsi"
> +	};
> +};

Best regards,
Krzysztof

Re: [PATCH 2/2] ARM: dts: aspeed: Add Asus IPMI card

[Anirudh Srinivasan]

Hi Krzysztof,

On Mon, Jan 12, 2026 at 5:48 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> > +
> > +     leds {
> > +             compatible = "gpio-leds";
> > +
> > +             led-fault {
> > +                     gpios = <&gpio1 ASPEED_GPIO(C, 5) GPIO_ACTIVE_HIGH>;
> > +                     panic-indicator;
> > +                     default-state = "off";
>
> Missing color and function. Or at least label.

I will add this.

>
> > +             };
> > +
> > +             led-heartbeat {
> > +                     gpios = <&gpio0 ASPEED_GPIO(A, 7) GPIO_ACTIVE_LOW>;
> > +                     linux,default-trigger = "timer";
> > +             };
> > +
>
> Drop redundant blank lines.
>
> > +&mac2 {
> > +     status = "okay";
> > +
> > +     /* Bootloader sets up the MAC to insert delay */
> > +     phy-mode = "rgmii";
> > +     phy-handle = <&ethphy2>;
> > +     pinctrl-names = "default";
> > +     pinctrl-0 = <&pinctrl_rgmii3_default>;
> > +};
> > +
> > +
>
> Like this as well. It's a sign of sloppy coding.

Will remove

>
> > +&fmc {
> > +     status = "okay";
> > +
> > +     flash@0 {
> > +             status = "okay";
> > +             m25p,fast-read;
> > +             label = "bmc";
> > +             spi-max-frequency = <50000000>;
> > +#include "openbmc-flash-layout-64.dtsi"
> > +     };
> > +};
>
> Best regards,
> Krzysztof
>

-- 
Regards
Anirudh Srinivasan