Re: fwnode_for_each_child_node() and OF backend discrepancy

[Michael Walle]

Am 2022-06-30 22:16, schrieb Horatiu Vultur:
> The 06/28/2022 23:07, Michael Walle wrote:
> > EXTERNAL EMAIL: Do not click links or open attachments unless you know 
> > the content is safe
> >
> > Am 2022-06-28 22:52, schrieb Horatiu Vultur:
> > > The 06/28/2022 22:28, Andy Shevchenko wrote:
> > > > EXTERNAL EMAIL: Do not click links or open attachments unless you know
> > > > the content is safe
> > > >
> > > > On Tue, Jun 28, 2022 at 5:17 PM Krzysztof Kozlowski
> > > > <krzysztof.kozlowski@xxxxxxxxxx> wrote:
> > > > > On 28/06/2022 17:09, Michael Walle wrote:
> > >
> > > Hi,
> > >
> > > Sorry for joint this late.
> > >
> > > >
> > > > ...
> > > >
> > > > > > Mh. Assume a SoC with an integrated ethernet switch. Some ports
> > > > > > are externally connected, some don't. I'd think they should be disabled,
> > > > > > no? Until now, all bindings I know, treat them as disabled. But OTOH
> > > > > > you still need to do some configurations on them, like disable port
> > > > > > forwarding, disable them or whatever. So the hardware is present, but
> > > > > > it is not connected to anything.
> > > > >
> > > > > I see your point and the meaning is okay... except that drivers don't
> > > > > touch disabled nodes. If a device (with some address space) is disabled,
> > > > > you do not write there "please be power off". Here the case is a bit
> > > > > different, because I think ports do not have their own address space.
> > > > > Yet it contradicts the logic - something is disabled in DT and you
> > > > > expect to perform actual operations on it.
> > > >
> > > > You beat me up to this comment, I also see a contradiction of what
> > > > "disabled" means in your, Michael, case and what it should be.
> > > >
> > > > If you need to perform an operation on some piece of HW, it has not to
> > > > be disabled.
> > > >
> > > > Or, you may deduce them by knowing how many ports in hardware (this is
> > > > usually done not by counting the nodes, but by a property) and do
> > > > whatever you want on ones, you have  not listed (by port_num) in the
> > > > array of parsed children.
> > >
> > > It is not possible to have a defined for the MAX number of ports that
> > > supported by lan966x. Which is 8. And assigned that define to
> > > num_phys_ports instead of counting the entries in DT?
> >
> > You mean also for the lan9662? I'm pretty sure that doesn't
> > work. Have a look where num_phys_ports is used. One random
> > example:
> > https://elixir.bootlin.com/linux/latest/source/drivers/net/ethernet/microchip/lan966x/lan966x_main.c#L874
> >
> > So if your switch only has 4 ports, then I'd guess you'll
> > access a non-existing register.
>
> Underneath lan662 and lan668 is the same chip. The HW people disable
> some ports/features on each platform but from what I know you will 
> still
> be able to access the registers.

I noticed that there are still 8 ports in the register description and
assumed that it was wrong [1]. But ok, that makes sense in some way.
OTOH that means, we cannot do the guesswork Vladimir proposed.

-michael

[1] https://microchip-ung.github.io/lan9662_reginfo/reginfo_LAN9662.html