[Arm-netbook] eoma68-a20 ethernet testing

[luke.leighton]

On Sat, Aug 31, 2013 at 5:51 AM, Ryan Mullen <rmmullen at gmail.com> wrote:
> Hi Luke and all,
>
> I was unable to get my ethernet working, but I do not have a
> micro-engineering board. (I did order one - did everyone receive them
> but me?)

 ... talk to chris.

> Here's a photo of my setup: http://i.imgur.com/FdogfjM.jpg

 neat!

> The ethernet jack has integrated magnetics, and all six relevant
> signals are connected.

 ok.

> The clip leads are for the UART. Obviously the wirewrap wire could
> have been made shorter and more orderly, so maybe that's the issue.
> The wirewrap is soldered to solid wire that fits the connector well,
> and then the solder point is heatshrunk.

 much more professionally done than anything i would have set up :)

> I used the sunxi-tools usb-boot script to load up a uboot on the
> eoma68-a20.

 awesome.

> Once at the uboot prompt, I did this:
>
> setenv ethaddr 00:00:00:00:00:01 (note: earlier instructions in this
> thread said macaddr instead of ethaddr)
> setenv ipaddr 10.0.0.6
> setenv netmask 255.255.255.0
> ping 10.0.0.1

 ok i must try this and confirm it works on a known-working board.

> The ping fails. Additionally, I do not get an LED on my network
> switch's front panel indicating link presence.

 right: if you don't get a link light then either a) there's a cabling
issue b) the eth jack's faulty c) uboot isn't bringing up the phy
properly.  still quite a lot to eliminate, let me do some testing here
ok?

> At the moment there is plenty of reason to doubt the integrity of my
> ethernet jack, but I just thought I would throw in my datapoint. At
> least it was nice to get the practice connecting to the CPU card and
> it's nice to see it show some signs of life :)

yaaay :)

ok, first thing: check the cables are the right way up.  it's easy to
put the cables in the top row rather than the bottom row and
vice-versa.  i can never remember which way round it is and have to
work it out every time sigh.  so.  looking _again_ at the PCB
layout.... :)

... pin 68 is on the end, and that's SATA.  so you can tell from the 4
capacitors which is pin 68.  by "pin 68 is on the end" i mean "the
pins when they come out on TOP are staggered,  so the ordering is
1,35,2,36.... 33,67,34,68.

to identify which one is 68 on the *connector*, look veeeeery very
carefully from the side, holding the board face UP with the pcmcia
connector on the RIGHT, and you'll see that the 68th pin's metal lead
dips *DOWN* whereas pin 34 does a dog-leg thing but basically stays
flat and goes into the *TOP* row of pins.

so now you know that when you can read the A20 CPU the right way up,
the top row of pins from furthest-away right-most corner to
closest-to-you is 1-34, and the row underneath it is 35-68.

so, with that in mind, could you please check ryan that you've got the
+/- in the right holes, and then also with a multimeter verify that
there are no short-outs between the pins.  put the multimeter on the
top of the pcmcia header (very tricky to get them to stay there!) and
onto each of the 6 wires of the ethernet jack's pins, go through all
6x6 combinations just to be absolutely sure.

i'll check the known-working card with u-boot in the meantime ok?

l.