RE: [EXT] Re: [PATCH net-next 0/1] net: fec: add C45 MDIO read/write support

[Andy Duan]

From: Andrew Lunn <andrew@xxxxxxx> Sent: Tuesday, August 20, 2019 9:04 PM
> On Tue, Aug 20, 2019 at 02:32:26AM +0000, Andy Duan wrote:
> > From: Andrew Lunn <andrew@xxxxxxx>
> > > On Mon, Aug 19, 2019 at 05:11:14PM +0000, Marco Hartmann wrote:
> > > > As of yet, the Fast Ethernet Controller (FEC) driver only supports
> > > > Clause 22 conform MDIO transactions. IEEE 802.3ae Clause 45
> > > > defines a modified MDIO protocol that uses a two staged access
> > > > model in order to increase the address space.
> > > >
> > > > This patch adds support for Clause 45 conform MDIO read and write
> > > > operations to the FEC driver.
> > >
> > > Hi Marco
> > >
> > > Do all versions of the FEC hardware support C45? Or do we need to
> > > make use of the quirk support in this driver to just enable it for some
> revisions of FEC?
> > >
> > > Thanks
> > >         Andrew
> >
> > i.MX legacy platforms like i.MX6/7 series, they doesn't support Write & Read
> Increment.
> > But for i.MX8MQ/MM series, it support C45 full features like Write & Read
> Increment.
> >
> > For the patch itself, it doesn't support Write & Read Increment, so I
> > think the patch doesn't need to add quirk support.
> 
> Hi Andy
> 
> So what happens with something older than a i.MX8MQ/MM when a C45
> transfer is attempted? This patch adds a new write. Does that write
> immediately trigger a completion interrupt? Does it never trigger an interrupt,
> and we have to wait FEC_MII_TIMEOUT?
> 
> Ideally, if the hardware does not support C45, we want it to return
> EOPNOTSUPP.
> 
> Thanks
>         Andrew

It still trigger an interrupt to wakeup the completion, we have to wait FEC_MII_TIMEOUT.
Older chips just support part of C45 feature just like the patch implementation.