linux-hardening - Re: [PATCH] ARM: dts: aspeed: Add device tree for Nvidia's GB200NVL BMC

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <7b47725f-fb57-454c-82f6-859202060243@lunn.ch>
Date: Fri, 24 Jan 2025 15:56:18 +0100
From: Andrew Lunn <andrew@...n.ch>
To: Willie Thai <wthai@...dia.com>
Cc: robh@...nel.org, krzk+dt@...nel.org, conor+dt@...nel.org,
	joel@....id.au, andrew@...econstruct.com.au, kees@...nel.org,
	tony.luck@...el.com, gpiccoli@...lia.com,
	devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-aspeed@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
	linux-hardening@...r.kernel.org, openbmc@...ts.ozlabs.org,
	leohu@...dia.com, tingkaic@...dia.com, dkodihalli@...dia.com,
	wthai <wthai@...lie-obmc-builder.nvidia.com>
Subject: Re: [PATCH] ARM: dts: aspeed: Add device tree for Nvidia's GB200NVL
 BMC

On Fri, Jan 24, 2025 at 05:18:19AM +0000, Willie Thai wrote:
> From: wthai <wthai@...lie-obmc-builder.nvidia.com>
> 
> The GB200NVL BMC is an Aspeed Ast2600 based BMC
> for Nvidia Blackwell GB200NVL platform.
> 
> Signed-off-by: wthai <wthai@...dia.com>

I'm not sure that is a valid name for a Signed-off-by. I would expect
something like:

Signed-off-by: Willie Thai <wthai@...dia.com>

> +&mac0 {
> +	status = "okay";
> +	pinctrl-names = "default";
> +	phy-mode = "rgmii-rxid";
> +	max-speed = <1000>;
> +	phy-handle = <&ethphy3>;
> +	pinctrl-0 = <&pinctrl_rgmii1_default>;
> +};

This phy-mode is wrong. There is a long discussion about this, started
from a BMC submitted by IBM. Please wait until Aspeed fix the MAC
driver before adding support for RGMII based ethernet. You are
probably innocent here, apart from failing to sanity check vendor
advise, which happens to be wrong.

https://lore.kernel.org/netdev/20250107162350.1281165-10-ninad@linux.ibm.com/T/
https://lore.kernel.org/linux-arm-kernel/bebbba7b-f86e-4dc4-8253-65d34cb84804@linux.ibm.com/T/

The max-speed parameter should also be unneeded. Unless you have a
real need for it. Does this MAC/PHY combination support higher speeds,
but they are broken for some reason, so you need to avoid them?

	Andrew