LinuxCNC-RIO - RealtimeIO for LinuxCNC based on FPGA (ICE40 / ECP5)

@Happppppppy

you need to configure your network !

Argh, @meister thankyou, it works now. Took a bit of googling to figure out how to do that.

For other beginners, what I did in the terminal:
sudo ip addr add 192.168.10.1/24 dev eno1
sudo ip link set eno1 up

I also added the following to the "interfaces" file in /etc/network/
iface eno1 inet static
address 192.168.10.1
netmask 255.255.255.0

Change eno1 to your adaptor name

Setting up a machine, have most things going, can see encoders in the test gui, inputs etc.  As soon as I add a bit out to turn on the spindle the compile fails with :

Reading constraint file: "/home/me/riocore/Output/TangNano9K/Gateware/pins.cst"
ERROR  (CT1135) : "/home/me/riocore/Output/TangNano9K/Gateware/pins.cst":54 | Can't find object named 'PINOUT_BITOUT0_BIT'
ERROR  (CT1135) : "/home/me/riocore/Output/TangNano9K/Gateware/pins.cst":55 | Can't find object named 'PINOUT_BITOUT0_BIT'
Physical Constraint parsed completed with errors

As far as I know I am doing as I always have, I did have a few silly mistakes but I have double checked the pin is valid and I tried not naming the plugin to spindle to see if that changed anything but no luck.

Cheers.
Russell.

Deleted...

Hi Russell,
please add your config

Cheers Olli

Hmm I worked it out, I don't really want to talk about it but I was trying to add outputs on pins that were already assigned to inputs, and strangely enough that didn't work...been a long weekend 

was my mistake, normally I give a warning/error if a pin is used twice, but I had once installed that you can use input pins more than once and forgot to check if both pins are marked as input.

is fixed in dev branch

I still have a weird issue that I am thinking is my hardware somehow, I can connect in the test GUI, see inputs, quadrature encoders and toggle outputs, but the terminal window keeps saying received data have wrong header. I have triple checked pins for conflicts. Apart from the terminal spamming that message it seems to test OK. 
Unfortunately LinuxCNC won't get out of estop, and spams wrong header size.

I am still treating it as my hardware, going to setup a different PC, I took out the Tang Nano that is in my milling machine and flashed it but the fault persists. If not the PC, I'll start swapping out PCB's from my two working machines.

Worst part is it was all working nicely last night.

if the testgui works, the hardware and gateware should be ok.

the c/riocomp.c must always match the gateware.
so if you have made changes to the config.json and start your LinuxCNC GUI on a different computer or from a different directory, it cannot establish a connection because the protocol no longer matches.

another reason may be that linuxcnc is trying to connect to the wrong IP.

you can look in the riocomp.c to see which ip is there:

#define UDP_IP “192.168.11.194”

and check in config.json whether UDP is really being used:

"protocol": "UDP",

I did a complete re-install and made sure I selected the right board starting config, and now everything works as normal. I must have accidentally started with a tangBOB starting config instead of the TangNano9K, or something along those lines, so as I was re-installing linux/checking hardware I brought the problem along with the original JSON file every time I tried fault finding.

If nothing else I can serve as a bad example !
Cheers.
Russell.