linux-kernel - Re: [PATCH v4 11/13] ARM: dts: aspeed: facebook-fuji: Include facebook-fuji-data64.dts

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

Message-ID: <aO7RJs8ceO3mwjn4@localhost.localdomain>
Date: Tue, 14 Oct 2025 15:39:34 -0700
From: Tao Ren <rentao.bupt@...il.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Guenter Roeck <linux@...ck-us.net>, Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>, Joel Stanley <joel@....id.au>,
	Andrew Jeffery <andrew@...econstruct.com.au>,
	devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-aspeed@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
	Tao Ren <taoren@...a.com>
Subject: Re: [PATCH v4 11/13] ARM: dts: aspeed: facebook-fuji: Include
 facebook-fuji-data64.dts

Hi Andrew,

On Tue, Oct 14, 2025 at 09:14:21PM +0200, Andrew Lunn wrote:
> On Tue, Oct 14, 2025 at 09:39:02AM -0700, Guenter Roeck wrote:
> > On 10/14/25 08:11, Andrew Lunn wrote:
> > > > > If it is already in mainline, i don't care too much if it is wrong. We
> > > > > don't want to cause regressions.
> > > > > 
> > > > > I only object when adding new nodes which are wrong. If we keep adding
> > > > > broken nodes, there is no incentive to fix the broken driver to do the
> > > > 
> > > > This wasn't adding an allegedly (sorry, it worked for me) broken node,
> > > > it was removing one that worked for me all along. Obviously I do not know
> > > > if it worked (or if it is even used) on real hardware, but it worked for
> > > > the fuji-bmc qemu emulation.
> > > 
> > > It probably does work on real hardware, because it is one of those
> > > "two wrongs makes a right" cases. So i see this as a regression. The
> > > node should not be removed. It should hopefully get corrected sometime
> > > in the future when somebody actually fixes the aspeed driver, and
> > > fixes both wrongs.
> > 
> > So you are trying to force the issue by disabling the Ethernet interface
> > on fuji-bmc until the problem in the driver (whatever it is) has been fixed ?
> > That just seems odd.
> 
> No, i see this as a regression, it probably used to work, so it should
> still work.
> 
> I'm just pushing back on adding new nodes which are broken. If it is a
> new node, it should not cause a regression.
> 
> 	Andrew

The fuji-data64.dts was copied from fuji.dts mainly for larger
persistent data partition (64MB), and the latest fuji.dts includes
fuji-data64.dts (instead of vice versa) because we wish to deprecate
fuji.dts in the long term.

So it is a regression for fuji, but fuji-data64.dts is a new node..

My original plan is to add mac3 back to fuji-data64.dts (which will fix
fuji.dts) when the rgmii fix is ready in aspeed mac driver, but I can
add it back now if we need to fix the fuji regression asap. Please
suggest.

BTW, the current mac3 settings work in real fuji hardware because rgmii
delay is configured in u-boot.


Thanks,

Tao