Re: [net-next RFC PATCH 0/6] Add support for qca8k mdio rw in Ethernet packet

[Ansuel Smith]

On Tue, Dec 07, 2021 at 11:22:41PM +0100, Andrew Lunn wrote:
> > The main problem here is that we really need a way to have shared data
> > between tagger and dsa driver. I also think that it would be limiting
> > using this only for mdio. For example qca8k can autocast mib with
> > Ethernet port and that would be another feature that the tagger would
> > handle.
> 
> The Marvell switches also have an efficient way to get the whole MIB
> table. So this is something i would also want.
>

Again same think... they just put the type in the qca hdr (placed in the
EthType position) and everything else is a mib. The switch send 7 packet
and each correspond to the MIB for the port (the port number is
comunicated in the qca hdr)

> > I like the idea of tagger-owend per-switch-tree private data.
> > Do we really need to hook logic?
> 
> We have two different things here.
> 
> 1) The tagger needs somewhere to store its own private data.
> 2) The tagger needs to share state with the switch driver.
> 
> We can probably have the DSA core provide 1). Add the size to
> dsa_device_ops structure, and provide helpers to go from either a
> master or a slave netdev to the private data.

I'm just implementing this. It doesn't look that hard.

> 
> 2) is harder. But as far as i know, we have an 1:N setup.  One switch
> driver can use N tag drivers. So we need the switch driver to be sure
> the tag driver is what it expects. We keep the shared state in the tag
> driver, so it always has valid data, but when the switch driver wants
> to get a pointer to it, it needs to pass a enum dsa_tag_protocol and
> if it does not match, the core should return -EINVAL or similar.
> 

Mhh this looks a bit complex. I'm probably missing something but why the
tagger needs to share a state? To check if it does support some feature?
If it's ready to be used for mdio Ethernet? Or just to be future-proof?

>    Andrew

-- 
	Ansuel