Re: [PATCH v1] mtd: parsers: ofpart: Fix parsing when size-cells is 0

[Miquel Raynal <miquel.raynal@bootlin.com>]

Hi Marek,

marex@denx.de wrote on Fri, 2 Dec 2022 17:52:05 +0100:

> On 12/2/22 17:42, Miquel Raynal wrote:
> > Hi Marek,  
> 
> Hi,
> 
> [...]
> 
> >>> However, it should not be empty, at the very least a reg property
> >>> should indicate on which CS it is wired, as expected there:
> >>> https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git/tree/Documentation/devicetree/bindings/mtd/nand-chip.yaml?h=mtd/next  
> >>
> >> OK, I see your point. So basically this?
> >>
> >> &gpmi {
> >>     #size-cells = <1>;
> >>     ...
> >>     nand-chip@0 {
> >>       reg = <0>;
> >>     };
> >> };
> >>
> >> btw. the GPMI NAND controller supports only one chipselect, so the reg in nand-chip node makes little sense.  
> > 
> > I randomly opened a reference manual (IMX6DQL.pdf), they say:
> > 
> > 	"Up to four NAND devices, supported by four chip-selects and one
> > 	 ganged ready/ busy."  
> 
> Doh, and MX7D has the same controller, so size-cells = <1>; makes sense with nand-chip@N {} .

Actually #address-cells is here for that. You need to point at one CS,
so in most cases this is:

controller {
	#address-cells = <1>;
	#size-cells = <0>;
	chip@N {
		reg = <N>;
	};
};

> 
> > Anyway, the NAND controller generic bindings which require this reg
> > property, what the controller or the driver actually supports, or even
> > how it is used on current designs is not relevant here.
> >   
> >>> But, as nand-chip.yaml references mtd.yaml, you can as well use
> >>> whatever is described here:
> >>> https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git/tree/Documentation/devicetree/bindings/mtd/mtd.yaml?h=mtd/next  
> >>>    >>>> What would be the gpmi controller size cells (X) in that case, still 0, right ? So how does that help solve this problem, wouldn't U-Boot still populate the partitions directly under the gpmi node or into partitions sub-node ?  
> >>>
> >>> The commit that was pointed in the original fix clearly stated that the
> >>> NAND chip node was targeted  
> >>
> >> I think this is another miscommunication here. The commit
> >>
> >> 753395ea1e45 ("ARM: dts: imx7: Fix NAND controller size-cells")
> >>
> >> modifies the size-cells of the NAND controller. The nand-chip is not involved in this at all . In the examples above, it's the "&gpmi" node size-cells that is modified.  
> > 
> > Yes I know. I was referring to this commit, sorry:
> > 36fee2f7621e ("common: fdt_support: add support for "partitions" subnode to fdt_fixup_mtdparts()")
> > 
> > The log says:
> > 
> > 	Listing MTD partitions directly in the flash mode has been
> > 	deprecated for a while for kernel Device Trees. Look for a node "partitions" in the
> > 	found flash nodes and use it instead of the flash node itself for the
> > 	partition list when it exists, so Device Trees following the current
> > 	best practices can be fixed up.
> > 
> > Which (I hope) means U-boot will equivalently try to play with the
> > partitions container, either in the controller node or in the chip node.
> >   
> >>> , not the NAND controller node. I hope this
> >>> is correctly supported in U-Boot though. So if there is a NAND chip
> >>> subnode, I suppose U-Boot would try to create the partitions that are
> >>> inside, or even in the sub "partitions" container.  
> >>
> >> My understanding is that U-Boot checks the nand-controller node size-cells, not the nand-chip{} or partitions{} subnode size-cells .  
> > 
> > I don't think U-Boot cares.
> >   
> >> Francesco, can you please share the DT, including the U-Boot generated partitions, which is passed to Linux on Colibri MX7 ? I think that should make all confusion go away.  
> > 
> > Please also do it with the NAND chip described. If, when the NAND chip
> > is described U-Boot tries to create partitions in the controller node,
> > then the situation is even worse than I thought. But I believe
> > describing the node like a suggest in the DT should prevent the boot
> > failure while still allowing a rather good description of the hardware.
> > 
> > BTW I still think the relevant action right now is to revert the DT
> > patch.  
> 
> I am starting to bank toward that variant as well (thanks for clarifying the rationale in the discussion, that helped a lot).
> 
> But then, the follow up fix would be what exactly, update the binding document to require #size-cells = <1>; ?


Thanks,
Miquèl