linux-kernel - Re: Disabling graph endpoints in device trees

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]

Date:	Mon, 22 Feb 2016 17:26:00 +0000
From:	John Keeping <john@...anate.com>
To:	Philipp Zabel <p.zabel@...gutronix.de>
Cc:	devicetree@...r.kernel.org, Rob Herring <robh+dt@...nel.org>,
	Frank Rowand <frowand.list@...il.com>,
	Grant Likely <grant.likely@...aro.org>,
	linux-kernel@...r.kernel.org, dri-devel@...ts.freedesktop.org
Subject: Re: Disabling graph endpoints in device trees

Hi Phillipp,

On Mon, 22 Feb 2016 17:12:27 +0100, Philipp Zabel wrote:
> Am Montag, den 22.02.2016, 14:14 +0000 schrieb John Keeping:
> > Is there a reason why endpoints in a device tree graph can't be
> > disabled?  
> 
> You can always remove them using /delete-node/, which also has the
> advantage of reminding you not to leave a single dangling endpoint.

Thanks, I wasn't aware of /delete-node/.  It does indeed do what I want.

> > I would like to be able to force the use of a particular CRTC for
> > certain outputs even though the hardware is capable of connecting any
> > CRTC to any output.  In this case I need to be able to support a wide
> > range of frequencies for external HDMI monitors so I will configure one
> > of the CRTCs to be able to generate these while the other will be tied
> > into a limited set of clock rates as a result of the overall system
> > clock setup.
> > 
> > Currently this can only be achieved by removing the endpoints from the
> > base SoC .dtsi file but it feels like it should be possible to add
> > 'status = "disabled"' to the nodes in the board-specific .dts in order
> > to disable undesirable configurations.
> >
> > I tested the change below and it behaves exactly as I want, but I don't
> > claim to understand all of the users of these functions to know if it
> > will break something else (hence this isn't a formal patch).  
> 
> I don't know that any driver depends on being able to parse disabled
> endpoints, but given the above I'm not sure that keeping disabled
> endpoints in the device tree is a useful feature.
> Disabling ports makes more sense to me. It should be documented in
> Documentation/devicetree/bindings/graph.txt though.

That isn't relevent for my scenario because this case has a port with
multiple endpoints and we only want one of the endpoints to be
available:

	vopb_out: port {
		#address-cells = <1>;
		#size-cells = <0>;

		vopb_out_hdmi: endpoint@0 {
			reg = <0>;
			remote-endpoint = <&hdmi_in_vopb>;
		};
		vopb_out_mipi: endpoint@1 {
			reg = <1>;
			remote-endpoint = <&mipi_in_vopb>;
		};
	};

But I'm happy that /delete-node/ works and seems to be the right thing
to do.


Thanks,
John