On Thu, Nov 30, 2017 at 07:26:21PM +0100, Andrew Lunn wrote:

On Thu, Nov 30, 2017 at 06:15:20PM +0000, Russell King - ARM Linux wrote:

quoted

On Thu, Nov 30, 2017 at 06:41:27PM +0100, Andrew Lunn wrote:

quoted

SGMII supports passing auto-negotiation results from the PHY to the
MAC. 1000BASE-X does not.

SGMII supports the PHY running at 10, 100, and 1000 Mbps. But to
support this, the MAC needs to replicate the bits 100, or 10 times
when the PHY is running in 10 or 100Mbps mode.

So with your 2.5G SGMII, you need to replicate the bits 250, 25, or
2.5 times if they PHY is running at lower speeds. This last one is
interesting.

That's not what I've read so far - but I don't know about the PHY
in this exact case because the docs are only available under NDA
(which makes it incredibly difficult to have this discussion.)

However, from what I can ascertain from a Xilinx document, 2.5G is
1G SGMII clocked 2.5x faster.  When in 2.5G mode, the other modes
are unavailable.

Hi Russell

Thanks for looking into the details.

So you need the PHY driver to see what it has negotiated, and when it
calls the adjust_link callback, the MAC needs look at the
phydev->interface and set the MAC to 2.5G SGMII or SGMII.

Yes, it /looks/ that way from what I've read so far.  I'd like to do
a more comprehensive look around at what various manufacturers are
doing before saying that definitively.  Even better would be to find
a specification for 2.5G SGMII!

Same as the Marvell 10G PHY driver flips between
PHY_INTERFACE_MODE_10GKR and PHY_INTERFACE_MODE_SGMII depending on
what it has negotiated.

Having played extensively with the Marvell 88x3310 and thoroughly
inspected its registers and analysed its behaviour, I'd be hessitant
to compare other PHYs with it.  The 3310 PHY seems to be multiple PHY
blocks (from two vendors) in a single package, and it dynamically
switches in the appropriate hardware blocks according to the negotiated
link modes.

However, given that 1G SGMII is clocked at a different rate from 2.5G
SGMII, both ends would need to be configured the same.

For example, both mvneta and mvpp2.2 support 2.5G modes merely by
reconfiguring the comphy block, and all that does is increase the
serdes clock.  That's used for 2500BASE-X (which I have had working
locally through a devmem2 pokes on Clearfog and the older Marvell
pp2x driver on Macchiatobin) but it's more difficult to find 2.5G
copper SFP modules to test with - which would presumably talk 2.5G
SGMII.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 8.8Mbps down 630kbps up
According to speedtest.net: 8.21Mbps down 510kbps up