Hi,

Ethernet is working, albeit with some FIFO under/overflow problems that stem 
from bugs in the FPGA design (not the PCB) that were already present in the 
ML401 design and appear more frequently here because of the lower system 
frequency used at the moment (83MHz instead of 100MHz). I will fix them later. 
Nonetheless, I have just been able to netboot a small Linux system on the M1.

I have also spotted a minor PCB error. At power-up, the KSZ8001 expects its 
(negative logic) reset pin to be low, but it is currently driven high by the 
pre-configuration weak pull-up resistors of the FPGA. This means that 
immediately after power is applied, there is a brief time during which the 
KSZ8001 is in an undefined state until the FPGA gets configured and resets the 
KSZ8001 properly. This does not prevent at all the prototype from working, 
however, I would prefer to stick closely to the chip specifications to avoid 
bad surprises during production. A simple fix can be to add a 10K pull-down 
resistor on ETH_RESET_N that will hold the KSZ8001 in reset until the 
configured FPGA drives the signal high.

Sébastien
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkym...@freenode
Webchat: www.milkymist.org/irc.html
Wiki: www.milkymist.org/wiki

Reply via email to