nvmem-cells regression after adding 'call of_platform_populate() for MTD partitions'

[Miquel Raynal]

Hi Maxim,

fido_max at inbox.ru wrote on Mon, 12 Dec 2022 20:57:49 +0300:

> Hi, Miquel!
> 
> On 12.12.2022 19:37, Miquel Raynal wrote:
> 
> > Let me try to recap the situation for all the people I just involved:
> > 
> > * An Ethernet driver gets its mac address from an nvmem cell. The
> >    Ethernet controller DT node then has an "nvmem-cells" property
> >    pointing towards an nvmem cell.
> > * The nvmem cell comes from an mtd partition.
> > * The mtd partition is flagged with a particular compatible
> >    (which is also named "nvmem-cells") to tell the kernel that the node
> >    produces nvmem cells.
> > * The mtd partition itself has no driver, but is the child node of a
> >    "partitions" container which has one (in this case,
> >    "fixed-partitions", see the snippet below).
> > 
> > Because the "nvmem-cells" property of the Ethernet node points at the
> > nvmem-cell node, the core create a device link between the Ethernet
> > controller (consumer) and the mtd partition (producer).
> > 
> > The device link in this case will never be satisfied because no driver
> > matches the "nvmem-cells" compatible of the partition node.
> > 
> > Reverting commit bcdf0315a61a ("mtd: call of_platform_populate() for MTD
> > partitions") would IMHO not make much sense, the problem comes from the
> > device link side and even there, there is nothing really "wrong",
> > because I really expect the mtd device to be ready before the
> > Ethernet controller probe, the device link is legitimate.
> > 
> > So I would like to explore other alternatives. Here are a bunch of
> > ideas, but I'm open:  
> 
> How about to create simple driver with compatible="nvmem-cell" and to move all the suff from main mtd driver which serves nvmem-cell to the probe function?

This is probably worth the try but I doubt you can make it work without
regressions because IIRC the nvmem registration happens no matter the
compatible (not mentioning the user-otp and factory-otp cases). You can
definitely try this out if you think you can come up with something
though.

But I would like to hear from the device-link gurus :) because even if
we fix mtd with a "trick" like above, I guess we'll very likely find
other corner cases like that and I am interested in understanding the
rationale of what could be a proper fix.

Thanks,
Miquèl