On Wed, Aug 25, 2021 at 05:18:16PM +0300, Vladimir Oltean wrote:

On Wed, Aug 25, 2021 at 10:00:45AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 03:58:10PM +0200, Michael Walle wrote:

...

Hi,

I noticed that there is a fallback to the u-boot device tree for linux (esp. EFI boot) if no other device tree was found, see [1]. It seems this is working fine for imx devices, for example, where you can just boot a stock installer iso via EFI. It will just work and it is quite a nice feature as a fallback.

Now for the layerscape architecture, the ls1028a in my case, things are more difficult because the bindings differ between u-boot and linux - one which comes to mind is DSA and ethernet.

Which begs the general question, is it encouraged to have both bindings diverge? To me it seems, that most bindings in u-boot are ad-hoc and there is no real review or alignment but just added as needed, which is ok if they are local to u-boot. But since they are nowadays passed to linux (by default!) I'm not so sure anymore.

OTOH The whole structure around a .dts{,i} and -u-boot.dtsi looks like they should (could?) be shared between linux and u-boot.

-michael

[1] https://elixir.bootlin.com/u-boot/v2021.10-rc2/source/common/board_r.c#L471

The U-Boot device tree is supposed to be able to be passed on to Linux and Just Work. The bindings are not supposed to be different between the two (except for when we take the binding while it's being hashed out upstream BUT THEN RESYNCED).

You might need to spell that out a bit clearer.

You are saying that both U-Boot and Linux are allowed to have their own custom properties (like 'u-boot,dm-spl' for U-Boot, and 'managed = "in-band-status"' for Linux), as long as the device tree files themselves are in sync, and the subset of the device tree blob understood by Linux (i.e. the U-Boot blob sans the U-Boot specifics) is compatible with the Linux DT blob?

I don't know what about the Linux example makes it Linux specific. But yes, 'u-boot,dm-spl' is clearly in our namespace and should be ignored by Linux. The whole reason we have the -u-boot.dtsi automatic drop-in logic (as much as it can be used is that device trees are device trees and describe the hardware and developers don't need to write a device tree for Linux and a device tree for U-Boot and a device tree for FreeBSD and ... So yes, you're supposed to use the device tree for a platform and it works here and there and every where.

To expand even further on that, it means we should put 'managed = "in-band-status"' in U-Boot, which is a Linux phylink device tree property, even if U-Boot does not use phylink?

We should be able to drop in the device trees from Linux and use them. Custodians should be re-syncing them periodically. Some are, even.

...

Incompatible device trees / bindings are a bug that needs to be fixed.

Curious that Michael mentions Ethernet and DSA on LS1028A.

I decompiled the two:

dtc -I dtb -O dts < arch/arm/dts/fsl-ls1028a-rdb.dtb > u-boot.dts dtc -I dtb -O dts < arch/arm64/boot/dts/freescale/fsl-ls1028a-rdb.dtb > linux.dts

and analyzed the Ethernet portion.

U-Boot looks like this:

/dts-v1/;

/ { compatible = "fsl,ls1028a-rdb", "fsl,ls1028a"; interrupt-parent = <0x1>; #address-cells = <0x2>; #size-cells = <0x2>; model = "NXP Layerscape 1028a RDB Board";

pcie@1f0000000 { compatible = "pci-host-ecam-generic"; bus-range = <0x0 0x0>; reg = <0x1 0xf0000000 0x0 0x100000>; #address-cells = <0x3>; #size-cells = <0x2>; device_type = "pci"; ranges = <0x82000000 0x0 0x0 0x1 0xf8000000 0x0 0x160000>;

pci@0,0 {
	reg = <0x0 0x0 0x0 0x0 0x0>;
	status = "okay";
	phy-mode = "sgmii";
	phy-handle = <0x4>;
	phandle = <0x11>;
};

pci@0,1 {
	reg = <0x100 0x0 0x0 0x0 0x0>;
	status = "disabled";
	phandle = <0x12>;
};

pci@0,2 {
	reg = <0x200 0x0 0x0 0x0 0x0>;
	status = "okay";
	phy-mode = "internal";
	phandle = <0x9>;

	fixed-link {
		speed = <0x9c4>;
		full-duplex;
	};
};

pci@0,3 {
	#address-cells = <0x0>;
	#size-cells = <0x1>;
	reg = <0x300 0x0 0x0 0x0 0x0>;
	status = "okay";
	phandle = <0x13>;

	fixed-link {
		speed = <0x3e8>;
		full-duplex;
	};

	phy@2 {
		reg = <0x2>;
		phandle = <0x4>;
	};

	phy@10 {
		reg = <0x10>;
		phandle = <0x5>;
	};

	phy@11 {
		reg = <0x11>;
		phandle = <0x6>;
	};

	phy@12 {
		reg = <0x12>;
		phandle = <0x7>;
	};

	phy@13 {
		reg = <0x13>;
		phandle = <0x8>;
	};
};

pci@0,5 {
	reg = <0x500 0x0 0x0 0x0 0x0>;
	status = "okay";
	phandle = <0x14>;

	ports {
		#address-cells = <0x1>;
		#size-cells = <0x0>;

		port@0 {
			reg = <0x0>;
			status = "okay";
			label = "swp0";
			phy-handle = <0x5>;
			phy-mode = "qsgmii";
			phandle = <0x15>;
		};

		port@1 {
			reg = <0x1>;
			status = "okay";
			label = "swp1";
			phy-handle = <0x6>;
			phy-mode = "qsgmii";
			phandle = <0x16>;
		};

		port@2 {
			reg = <0x2>;
			status = "okay";
			label = "swp2";
			phy-handle = <0x7>;
			phy-mode = "qsgmii";
			phandle = <0x17>;
		};

		port@3 {
			reg = <0x3>;
			status = "okay";
			label = "swp3";
			phy-handle = <0x8>;
			phy-mode = "qsgmii";
			phandle = <0x18>;
		};

		port@4 {
			reg = <0x4>;
			phy-mode = "internal";
			status = "okay";
			ethernet = <0x9>;
			phandle = <0x19>;

			fixed-link {
				speed = <0x9c4>;
				full-duplex;
			};
		};

		port@5 {
			reg = <0x5>;
			phy-mode = "internal";
			status = "disabled";
			phandle = <0x1a>;

			fixed-link {
				speed = <0x3e8>;
				full-duplex;
			};
		};
	};
};

pci@0,6 {
	reg = <0x600 0x0 0x0 0x0 0x0>;
	status = "disabled";
	phy-mode = "internal";
	phandle = <0x1b>;
};

}; };

While Linux looks like this:

/dts-v1/;

/ { soc { compatible = "simple-bus"; #address-cells = <0x2>; #size-cells = <0x2>; ranges;

pcie@1f0000000 {
	compatible = "pci-host-ecam-generic";
	reg = <0x1 0xf0000000 0x0 0x100000>;
	#address-cells = <0x3>;
	#size-cells = <0x2>;
	msi-parent = <0x11>;
	device_type = "pci";
	bus-range = <0x0 0x0>;
	dma-coherent;
	msi-map = <0x0 0x11 0x17 0xe>;
	iommu-map = <0x0 0x12 0x17 0xe>;
	ranges = <0x82000000 0x1 0xf8000000 0x1 0xf8000000 0x0 0x160000 0xc2000000 0x1 0xf8160000 0x1 0xf8160000 0x0 0x70000 0x82000000 0x1 0xf81d0000 0x1 0xf81d0000 0x0 0x20000 0xc2000000 0x1 0xf81f0000 0x1 0xf81f0000 0x0 0x20000 0x82000000 0x1 0xf8210000 0x1 0xf8210000 0x0 0x20000 0xc2000000 0x1 0xf8230000 0x1 0xf8230000 0x0 0x20000 0x82000000 0x1 0xfc000000 0x1 0xfc000000 0x0 0x400000>;

	ethernet@0,0 {
		compatible = "fsl,enetc";
		reg = <0x0 0x0 0x0 0x0 0x0>;
		status = "okay";
		phy-handle = <0x13>;
		phy-connection-type = "sgmii";
		managed = "in-band-status";

		mdio {
			#address-cells = <0x1>;
			#size-cells = <0x0>;

			ethernet-phy@2 {
				reg = <0x2>;
				phandle = <0x13>;
			};
		};
	};

	ethernet@0,1 {
		compatible = "fsl,enetc";
		reg = <0x100 0x0 0x0 0x0 0x0>;
		status = "disabled";
	};

	ethernet@0,2 {
		compatible = "fsl,enetc";
		reg = <0x200 0x0 0x0 0x0 0x0>;
		phy-mode = "internal";
		status = "okay";
		phandle = <0x18>;

		fixed-link {
			speed = <0x9c4>;
			full-duplex;
		};
	};

	mdio@0,3 {
		compatible = "fsl,enetc-mdio";
		reg = <0x300 0x0 0x0 0x0 0x0>;
		#address-cells = <0x1>;
		#size-cells = <0x0>;

		ethernet-phy@10 {
			reg = <0x10>;
			phandle = <0x14>;
		};

		ethernet-phy@11 {
			reg = <0x11>;
			phandle = <0x15>;
		};

		ethernet-phy@12 {
			reg = <0x12>;
			phandle = <0x16>;
		};

		ethernet-phy@13 {
			reg = <0x13>;
			phandle = <0x17>;
		};
	};

	ethernet@0,4 {
		compatible = "fsl,enetc-ptp";
		reg = <0x400 0x0 0x0 0x0 0x0>;
		clocks = <0x2 0x2 0x3>;
		little-endian;
		fsl,extts-fifo;
	};

	ethernet-switch@0,5 {
		reg = <0x500 0x0 0x0 0x0 0x0>;
		interrupts = <0x0 0x5f 0x4>;
		status = "okay";

		ports {
			#address-cells = <0x1>;
			#size-cells = <0x0>;

			port@0 {
				reg = <0x0>;
				status = "okay";
				label = "swp0";
				managed = "in-band-status";
				phy-handle = <0x14>;
				phy-mode = "qsgmii";
			};

			port@1 {
				reg = <0x1>;
				status = "okay";
				label = "swp1";
				managed = "in-band-status";
				phy-handle = <0x15>;
				phy-mode = "qsgmii";
			};

			port@2 {
				reg = <0x2>;
				status = "okay";
				label = "swp2";
				managed = "in-band-status";
				phy-handle = <0x16>;
				phy-mode = "qsgmii";
			};

			port@3 {
				reg = <0x3>;
				status = "okay";
				label = "swp3";
				managed = "in-band-status";
				phy-handle = <0x17>;
				phy-mode = "qsgmii";
			};

			port@4 {
				reg = <0x4>;
				phy-mode = "internal";
				status = "okay";
				ethernet = <0x18>;

				fixed-link {
					speed = <0x9c4>;
					full-duplex;
				};
			};

			port@5 {
				reg = <0x5>;
				phy-mode = "internal";
				status = "disabled";

				fixed-link {
					speed = <0x3e8>;
					full-duplex;
				};
			};
		};
	};

	ethernet@0,6 {
		compatible = "fsl,enetc";
		reg = <0x600 0x0 0x0 0x0 0x0>;
		phy-mode = "internal";
		status = "disabled";

		fixed-link {
			speed = <0x3e8>;
			full-duplex;
		};
	};

	rcec@1f,0 {
		reg = <0xf800 0x0 0x0 0x0 0x0>;
		interrupts = <0x0 0x5e 0x4>;
	};
};

}; };

I mean, they look pretty similar to me? The biggest difference is that the ENETC ECAM is under the /soc node in Linux, but under / in U-Boot, as well as some BAR memory areas that are not marked as prefetchable or non-prefetchable in the U-Boot device tree. But excuse me, that isn't an Ethernet/DSA problem, is it?

I'll leave this part to Michael.

On Wed, Aug 25, 2021 at 06:12:20PM +0300, Vladimir Oltean wrote:

On Wed, Aug 25, 2021 at 10:26:10AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 05:18:16PM +0300, Vladimir Oltean wrote:

...

On Wed, Aug 25, 2021 at 10:00:45AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 03:58:10PM +0200, Michael Walle wrote:

...

Hi,

I noticed that there is a fallback to the u-boot device tree for linux (esp. EFI boot) if no other device tree was found, see [1]. It seems this is working fine for imx devices, for example, where you can just boot a stock installer iso via EFI. It will just work and it is quite a nice feature as a fallback.

Now for the layerscape architecture, the ls1028a in my case, things are more difficult because the bindings differ between u-boot and linux - one which comes to mind is DSA and ethernet.

Which begs the general question, is it encouraged to have both bindings diverge? To me it seems, that most bindings in u-boot are ad-hoc and there is no real review or alignment but just added as needed, which is ok if they are local to u-boot. But since they are nowadays passed to linux (by default!) I'm not so sure anymore.

OTOH The whole structure around a .dts{,i} and -u-boot.dtsi looks like they should (could?) be shared between linux and u-boot.

-michael

[1] https://elixir.bootlin.com/u-boot/v2021.10-rc2/source/common/board_r.c#L471

The U-Boot device tree is supposed to be able to be passed on to Linux and Just Work. The bindings are not supposed to be different between the two (except for when we take the binding while it's being hashed out upstream BUT THEN RESYNCED).

You might need to spell that out a bit clearer.

You are saying that both U-Boot and Linux are allowed to have their own custom properties (like 'u-boot,dm-spl' for U-Boot, and 'managed = "in-band-status"' for Linux), as long as the device tree files themselves are in sync, and the subset of the device tree blob understood by Linux (i.e. the U-Boot blob sans the U-Boot specifics) is compatible with the Linux DT blob?

I don't know what about the Linux example makes it Linux specific. But yes, 'u-boot,dm-spl' is clearly in our namespace and should be ignored by Linux. The whole reason we have the -u-boot.dtsi automatic drop-in logic (as much as it can be used is that device trees are device trees

    ^
    I don't think this parenthesis ever closes...

Ah, whoops. Should have been "(as much as it can be used)" because it does get #included instead in some cases, for reasons.

...

and describe the hardware and developers don't need to write a device tree for Linux and a device tree for U-Boot and a device tree for FreeBSD and ... So yes, you're supposed to use the device tree for a

            ^
            so I never get the answer to "the whole reason is...".

...

platform and it works here and there and every where.

The fact that only Linux uses it makes it Linux specific.

...

...

To expand even further on that, it means we should put 'managed = "in-band-status"' in U-Boot, which is a Linux phylink device tree property, even if U-Boot does not use phylink?

We should be able to drop in the device trees from Linux and use them. Custodians should be re-syncing them periodically. Some are, even.

Are you ready to take up device tree bindings for PTP timers, PCIe root complex event collectors, cascaded interrupt controllers, things which U-Boot will never ever need to support?

At least in Linux there is a policy to not add device tree nodes that do not have drivers. Is the same policy not true for U-Boot? At least your ./scripts/checkpatch.pl does have the same "check for DT compatible documentation" section as Linux. You might consider removing it if you want people to not strip the DTs they submit to U-Boot.

And why do we even maintain the device tree bindings in Linux at all? It seems rather counter-productive for both ends to do that, if it is expected that the kernel works with DT blobs provided by third parties too, and if all third parties need to resync with it (there are other boot loaders too beyond U-Boot, and other kernels beyond Linux). Somehow it doesn't feel right for the reference to be the Linux kernel. Maybe this is something that needs to be brought up with higher-level Linux maintainers.

I have no problem at all with structuring the device tree in the same way in U-Boot as in Linux, as long as that proves to not be a foolish endeavor.

DT is ABI is hardware description and OS-agnostic has been the rule for 10+ years. If that's no longer the case, can someone please tell me?

On Wed, Aug 25, 2021 at 06:43:23PM +0300, Vladimir Oltean wrote:

On Wed, Aug 25, 2021 at 11:24:08AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 06:12:20PM +0300, Vladimir Oltean wrote:

...

On Wed, Aug 25, 2021 at 10:26:10AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 05:18:16PM +0300, Vladimir Oltean wrote:

...

On Wed, Aug 25, 2021 at 10:00:45AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 03:58:10PM +0200, Michael Walle wrote:

> Hi, > > I noticed that there is a fallback to the u-boot device tree for linux > (esp. EFI boot) if no other device tree was found, see [1]. It seems this > is working fine for imx devices, for example, where you can just boot a > stock installer iso via EFI. It will just work and it is quite a nice > feature as a fallback. > > Now for the layerscape architecture, the ls1028a in my case, things are > more difficult because the bindings differ between u-boot and linux - one > which comes to mind is DSA and ethernet. > > Which begs the general question, is it encouraged to have both bindings > diverge? To me it seems, that most bindings in u-boot are ad-hoc and there > is no real review or alignment but just added as needed, which is ok if > they are local to u-boot. But since they are nowadays passed to linux > (by default!) I'm not so sure anymore. > > OTOH The whole structure around a .dts{,i} and -u-boot.dtsi looks like > they should (could?) be shared between linux and u-boot. > > -michael > > [1] > https://elixir.bootlin.com/u-boot/v2021.10-rc2/source/common/board_r.c#L471

The U-Boot device tree is supposed to be able to be passed on to Linux and Just Work. The bindings are not supposed to be different between the two (except for when we take the binding while it's being hashed out upstream BUT THEN RESYNCED).

You might need to spell that out a bit clearer.

You are saying that both U-Boot and Linux are allowed to have their own custom properties (like 'u-boot,dm-spl' for U-Boot, and 'managed = "in-band-status"' for Linux), as long as the device tree files themselves are in sync, and the subset of the device tree blob understood by Linux (i.e. the U-Boot blob sans the U-Boot specifics) is compatible with the Linux DT blob?

I don't know what about the Linux example makes it Linux specific. But yes, 'u-boot,dm-spl' is clearly in our namespace and should be ignored by Linux. The whole reason we have the -u-boot.dtsi automatic drop-in logic (as much as it can be used is that device trees are device trees

    ^
    I don't think this parenthesis ever closes...

Ah, whoops. Should have been "(as much as it can be used)" because it does get #included instead in some cases, for reasons.

...

...

and describe the hardware and developers don't need to write a device tree for Linux and a device tree for U-Boot and a device tree for FreeBSD and ... So yes, you're supposed to use the device tree for a

            ^
            so I never get the answer to "the whole reason is...".

...

platform and it works here and there and every where.

The fact that only Linux uses it makes it Linux specific.

...

...

To expand even further on that, it means we should put 'managed = "in-band-status"' in U-Boot, which is a Linux phylink device tree property, even if U-Boot does not use phylink?

We should be able to drop in the device trees from Linux and use them. Custodians should be re-syncing them periodically. Some are, even.

Are you ready to take up device tree bindings for PTP timers, PCIe root complex event collectors, cascaded interrupt controllers, things which U-Boot will never ever need to support?

At least in Linux there is a policy to not add device tree nodes that do not have drivers. Is the same policy not true for U-Boot? At least your ./scripts/checkpatch.pl does have the same "check for DT compatible documentation" section as Linux. You might consider removing it if you want people to not strip the DTs they submit to U-Boot.

And why do we even maintain the device tree bindings in Linux at all? It seems rather counter-productive for both ends to do that, if it is expected that the kernel works with DT blobs provided by third parties too, and if all third parties need to resync with it (there are other boot loaders too beyond U-Boot, and other kernels beyond Linux). Somehow it doesn't feel right for the reference to be the Linux kernel. Maybe this is something that needs to be brought up with higher-level Linux maintainers.

I have no problem at all with structuring the device tree in the same way in U-Boot as in Linux, as long as that proves to not be a foolish endeavor.

DT is ABI is hardware description and OS-agnostic has been the rule for 10+ years. If that's no longer the case, can someone please tell me?

So if Michael's board with DT provided by U-Boot doesn't work for some stupid reason like "Linux expects the pcie node to be under /soc", or "Linux wants all PCIe BARs of a RCIEP ECAM to be spelled out in the 'ranges' property, because it's too dumb to detect them itself", or something like that, I've got no argument against that, let's go ahead and resync U-Boot with Linux.

But "DT is ABI is hardware description" is a pretty vague truism that does not actually help here.

I'm saying that because it's what's been said for what feels like 10+ years. I don't want to think how many countless hours have been spent on that point at conferences over the years. It's not even a Linux thing. I would swear you can (or could, unless it got broken) take the same DTB for some platforms and boot Linux or FreeBSD or some other BSD or maybe even VxWorks and it works.

Will you accept device trees with devices for which a driver will never probe in U-Boot,

Yes, I will absolutely take device trees that have devices we don't need in U-Boot since the point is, and many SoC vendors are doing (and the ones that aren't are, I am not happy about / with) right now.

and will you remove the checkpatch warnings about those, to not discourage people from submitting them prior to the actual public review?

With respect to checkpatch.pl, maybe I'm just missing the line in question? Or maybe it's a kernel-related warning we need to disable in our .checkpatch.conf. But I don't want to side-track over this part.

If a U-Boot driver will be written for a device that does not have a driver yet in Linux, then the Linux driver will be written but with DT bindings incompatible with what was established in U-Boot, will you wake up the U-Boot developer from the grave and ask them to resync the driver to follow Linux? Or will you accept drivers at all for hardware that is not supported by Linux?

What I've said for years (but yes, I've missed changes, maybe the yaml dt binding stuff would help so I could make CI fail or at least require manual override?) is that U-Boot will take immature bindings but it's on the developer to re-sync once the bindings are fully reviewed. This is to help with the chicken-and-egg problem. But old bindings are not intended to be supported, once it's finalized. That is part of the bargain.

I also think there are various degrees of what it means "to work" with a device tree provided directly by U-Boot. Distros like Arch Linux ARM have a package for device tree blobs, and it is expected that these are updated by the distro, and U-Boot just loads them. As Michael points out, the DT provided by U-Boot is just a fallback. The OS should boot to prompt and have a network connection to set itself up properly again. But we need to draw a harder line on what we _actually_ desire to work beyond that.

Every distribution has a package for device tree binaries, because that's how you get a device tree on the still vast majority of platforms that don't ship one in-flash (RPis being the modern exception) and to my knowledge none of them are happy about having to build and pass and make sure the right one is used on a given board at boot. So yes, U-Boot being able to pass a device tree on to the next stage is one of those things to help everything Just Work and be boring.

Am 2021-08-26 01:03, schrieb Vladimir Oltean:

On Wed, Aug 25, 2021 at 04:09:50PM -0400, Tom Rini wrote: In any case, it doesn't sound absurd at all, with a bit of passion it could be done on all Layerscapes. I would be absolutely glad to help on the Ethernet / DSA side of things (which I believe is the reason why Michael summoned me into this thread),

;) and because I thought you might be interested in the answer to the initial question. After all, you also worked on the device trees in linux and u-boot.

but I don't believe that's where the problem is right now. When I added the DM_DSA uclass to U-Boot I did my best to pick a reasonable subset of Linux DSA, and with compatible device tree bindings. Also maintaining the Linux side of things, I did provide feedback to Tim Harvey for the Microchip KSZ switches as to what is the subset supported by U-Boot that would also be DT-compatible with Linux. If it turns out that I failed at that, I am willing to rework what we have.

I started to convert the u-boot device tree yesterday - and it doesn't look too bad for now. I was already able to copy the kernel soc dtsi and u-boot is still booting and working.

Theres still one catch at the moment, AFAIK in linux you can put the PHYs either in the mdio controller node or in a "mdio" subnode within the ethernet controller node. I'm not sure wether the latter works in u-boot, but [1] looks promising. At least, linux dtbs are using the mdio subnodes and u-boot put the phys into the mdio controller node.

Maybe sharing the device tree between linux and u-boot isn't that hard for the ls1028a after all and its just that nobody did it for now. Renaming the reference here and there and introducing the linux compatible strings may do it. I'll come back to you if there are problems with ethernet (or DSA).

-michael

I have been known on a few occasions to say "U-Boot does not parse this part of the device tree, you can just strip it away", but I will keep my mouth shut from now on.

[1] https://lists.denx.de/pipermail/u-boot/2020-May/410169.html

Am 2021-08-26 18:32, schrieb Vladimir Oltean:

On Thu, Aug 26, 2021 at 09:35:12AM +0200, Michael Walle wrote:

...

Theres still one catch at the moment, AFAIK in linux you can put the PHYs either in the mdio controller node or in a "mdio" subnode within the ethernet controller node. I'm not sure wether the latter works in u-boot, but [1] looks promising. At least, linux dtbs are using the mdio subnodes and u-boot put the phys into the mdio controller node.

Please change Linux for that, move the PHYs from the per-port MDIO node to the PF 3 central MDIO controller node. Due to hardware reasons, the per-port MDIO controller registers are in fact de-featured and should be hidden from new LS1028A reference manuals.

Care to share some more details? There should be some more information besides "for hardware reasons" which should go into the commit message. Is there an erratum?

In fact, I guess in Rev0 of the RM it has already been removed, at least from the port memory map (in the RevB RM it was still present).

-michael

Date: Wed, 25 Aug 2021 18:12:20 +0300 From: Vladimir Oltean olteanv@gmail.com

On Wed, Aug 25, 2021 at 10:26:10AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 05:18:16PM +0300, Vladimir Oltean wrote:

...

On Wed, Aug 25, 2021 at 10:00:45AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 03:58:10PM +0200, Michael Walle wrote:

...

Hi,

I noticed that there is a fallback to the u-boot device tree for linux (esp. EFI boot) if no other device tree was found, see [1]. It seems this is working fine for imx devices, for example, where you can just boot a stock installer iso via EFI. It will just work and it is quite a nice feature as a fallback.

Now for the layerscape architecture, the ls1028a in my case, things are more difficult because the bindings differ between u-boot and linux - one which comes to mind is DSA and ethernet.

Which begs the general question, is it encouraged to have both bindings diverge? To me it seems, that most bindings in u-boot are ad-hoc and there is no real review or alignment but just added as needed, which is ok if they are local to u-boot. But since they are nowadays passed to linux (by default!) I'm not so sure anymore.

OTOH The whole structure around a .dts{,i} and -u-boot.dtsi looks like they should (could?) be shared between linux and u-boot.

-michael

[1] https://elixir.bootlin.com/u-boot/v2021.10-rc2/source/common/board_r.c#L471

The U-Boot device tree is supposed to be able to be passed on to Linux and Just Work. The bindings are not supposed to be different between the two (except for when we take the binding while it's being hashed out upstream BUT THEN RESYNCED).

You might need to spell that out a bit clearer.

You are saying that both U-Boot and Linux are allowed to have their own custom properties (like 'u-boot,dm-spl' for U-Boot, and 'managed = "in-band-status"' for Linux), as long as the device tree files themselves are in sync, and the subset of the device tree blob understood by Linux (i.e. the U-Boot blob sans the U-Boot specifics) is compatible with the Linux DT blob?

I don't know what about the Linux example makes it Linux specific. But yes, 'u-boot,dm-spl' is clearly in our namespace and should be ignored by Linux. The whole reason we have the -u-boot.dtsi automatic drop-in logic (as much as it can be used is that device trees are device trees

    ^
    I don't think this parenthesis ever closes...

...

and describe the hardware and developers don't need to write a device tree for Linux and a device tree for U-Boot and a device tree for FreeBSD and ... So yes, you're supposed to use the device tree for a

            ^
            so I never get the answer to "the whole reason is...".

...

platform and it works here and there and every where.

The fact that only Linux uses it makes it Linux specific.

Huh? Other OSes use device trees as well on various platforms. There are (somewhat) linux-specific properties in those device trees in the "linux," namespace.

For OpenBSD we don't maintain our own set of device trees. Instead we rely on the device tree provided by U-Boot and distribute device trees built from the Linux tree as an alternative that users can install if the U-Boot device tree is a bit behind.

...

...

To expand even further on that, it means we should put 'managed = "in-band-status"' in U-Boot, which is a Linux phylink device tree property, even if U-Boot does not use phylink?

We should be able to drop in the device trees from Linux and use them. Custodians should be re-syncing them periodically. Some are, even.

Are you ready to take up device tree bindings for PTP timers, PCIe root complex event collectors, cascaded interrupt controllers, things which U-Boot will never ever need to support?

The policy is that U-Boot uses a verbatim copy of the device tree source files from Linux (.dts and .dtsi files) and augments that device tree if necessary by using a *-u-boot.dts files that ideally only add properties in the "u-boot," namespace.

That policy isn't always followed though. In some cases support for certain hardware is added to U-Boot early, before "official" DT bindings have been established. These should be replaced with the "official" device tree from Linux as soon as possible, but that doesn't always happen in a timely fashion.

Unfortunately what happens quite often is that hardware vendors provide patches with preliminary bindings from their own Linux fork before their devices are fully supported in the mainline Linux tree.

What also happens is that the DT bindings in mainline Linux evolve over time. Sometimes there is a good reason for that, but in other cases maintainers seem to be a little bit too eager to "clean things" up in a way that breaks backwards compatibility.

At least in Linux there is a policy to not add device tree nodes that do not have drivers. Is the same policy not true for U-Boot? At least your ./scripts/checkpatch.pl does have the same "check for DT compatible documentation" section as Linux. You might consider removing it if you want people to not strip the DTs they submit to U-Boot.

There is no such policy in U-Boot. And even in Linux the devicetree maintainers seem to be willing to accept DT bindings and nodes for things that have a driver elsewhere.

And why do we even maintain the device tree bindings in Linux at all? It seems rather counter-productive for both ends to do that, if it is expected that the kernel works with DT blobs provided by third parties too, and if all third parties need to resync with it (there are other boot loaders too beyond U-Boot, and other kernels beyond Linux). Somehow it doesn't feel right for the reference to be the Linux kernel. Maybe this is something that needs to be brought up with higher-level Linux maintainers.

IMHO it would be a good idea if the device trees would be maintained in a separate source tree from the Linux kernel.

I have no problem at all with structuring the device tree in the same way in U-Boot as in Linux, as long as that proves to not be a foolish endeavor.

As I said above, U-Boot should simply use a verbatim copy of the Linux device tree.

On 8/31/21 9:35 AM, Rob Herring wrote:

On Wed, Aug 25, 2021 at 10:12 AM Vladimir Oltean olteanv@gmail.com wrote:

...

On Wed, Aug 25, 2021 at 10:26:10AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 05:18:16PM +0300, Vladimir Oltean wrote:

...

On Wed, Aug 25, 2021 at 10:00:45AM -0400, Tom Rini wrote:

...

On Wed, Aug 25, 2021 at 03:58:10PM +0200, Michael Walle wrote:

...

Hi,

I noticed that there is a fallback to the u-boot device tree for linux (esp. EFI boot) if no other device tree was found, see [1]. It seems this is working fine for imx devices, for example, where you can just boot a stock installer iso via EFI. It will just work and it is quite a nice feature as a fallback.

Now for the layerscape architecture, the ls1028a in my case, things are more difficult because the bindings differ between u-boot and linux - one which comes to mind is DSA and ethernet.

Which begs the general question, is it encouraged to have both bindings diverge? To me it seems, that most bindings in u-boot are ad-hoc and there is no real review or alignment but just added as needed, which is ok if they are local to u-boot. But since they are nowadays passed to linux (by default!) I'm not so sure anymore.

OTOH The whole structure around a .dts{,i} and -u-boot.dtsi looks like they should (could?) be shared between linux and u-boot.

-michael

[1] https://elixir.bootlin.com/u-boot/v2021.10-rc2/source/common/board_r.c#L471

The U-Boot device tree is supposed to be able to be passed on to Linux and Just Work. The bindings are not supposed to be different between the two (except for when we take the binding while it's being hashed out upstream BUT THEN RESYNCED).

You might need to spell that out a bit clearer.

You are saying that both U-Boot and Linux are allowed to have their own custom properties (like 'u-boot,dm-spl' for U-Boot, and 'managed = "in-band-status"' for Linux), as long as the device tree files themselves are in sync, and the subset of the device tree blob understood by Linux (i.e. the U-Boot blob sans the U-Boot specifics) is compatible with the Linux DT blob?

I don't know what about the Linux example makes it Linux specific. But yes, 'u-boot,dm-spl' is clearly in our namespace and should be ignored by Linux. The whole reason we have the -u-boot.dtsi automatic drop-in logic (as much as it can be used is that device trees are device trees

     ^
     I don't think this parenthesis ever closes...

...

and describe the hardware and developers don't need to write a device tree for Linux and a device tree for U-Boot and a device tree for FreeBSD and ... So yes, you're supposed to use the device tree for a

             ^
             so I never get the answer to "the whole reason is...".

...

platform and it works here and there and every where.

The fact that only Linux uses it makes it Linux specific.

...

...

To expand even further on that, it means we should put 'managed = "in-band-status"' in U-Boot, which is a Linux phylink device tree property, even if U-Boot does not use phylink?

We should be able to drop in the device trees from Linux and use them. Custodians should be re-syncing them periodically. Some are, even.

Are you ready to take up device tree bindings for PTP timers, PCIe root complex event collectors, cascaded interrupt controllers, things which U-Boot will never ever need to support?

At least in Linux there is a policy to not add device tree nodes that do not have drivers.

That is not the policy. The policy is DT nodes must have binding (schema) documentation and the binding should be complete as possible (not what some driver currently uses). However, for complex bindings, it might be difficult to write the binding in absence of a driver.

It is effective policy for some arches...

When the K210 patches were submitted, any bindings for devices without enabled drivers were requested to be (and subsequently were) removed, even though many of those bindings were based off of existing documentation. This is the primary cause of divergence between the Linux and U-Boot devicetrees for this platform. It is also the main reason that I have not bothered putting together a sync patch.

--Sean

Am 2021-08-25 16:18, schrieb Vladimir Oltean:

Curious that Michael mentions Ethernet and DSA on LS1028A.

[..]

I mean, they look pretty similar to me? The biggest difference is that the ENETC ECAM is under the /soc node in Linux, but under / in U-Boot, as well as some BAR memory areas that are not marked as prefetchable or non-prefetchable in the U-Boot device tree. But excuse me, that isn't an Ethernet/DSA problem, is it?

Mh, you're right. I had something in mind you've mentioned during the initial DSA patches for u-boot. I must have remembered incorrectly.

My intention was to have a clear message if the u-boot device tree should be following the linux one before I'm digging deeper into what is different between these two and try to sync both.

-michael