> > When I run Linux with a customized hal and config file it will crash > > with the window with the errors. But after that I can't even run the > > stock single parallel port version without it crashing. It's like the > > 0x378 port has now vanished. > > > It has. You will need to blacklist ppdev as its grabbing the port before > LCNC can. The manpage for modprobe should explain how to blacklist it. >
Gene, I think this is a problem within LinuxCNC itself and MachineKit because I ran into this with the BBB too. If the hal/ini file has a certain type of error it won't run LinuxCNC and instead throws up a dialog box with all the error information. Very extensive error information too. So imagine this scenario. 1. Run LinuxCNC for the parallel port with the standard Stepper Motor example. 2. LinuxCNC runs. G-Code can be loaded. 3. Exit LinuxCNC and run it again but this time with a hal/ini file that has an error. 4. I get the window with the debugging information. 5. Run LinuxCNC for the parallel port with the standard Stepper Motor example as step 1. 6. Get the same debug window as step 4. I believe what happens is that with the bad hal file, Linux CNC manages to capture the parallel port at 0x378 but then chokes on the second one at 0x2000. Or if there is only one parallel port it captures it but then fails on some improperly formed hal command. Then it throws up the diagnostic window. At that point it should have released ownership of the 0x378 port but it doesn’t (I think). And so the next time you try and run with a good hal/ini file the failure window is put up because now the 0x378 port isn't available anymore still held by something that didn't release it. But it was there once. So far only a reboot has fixed this. I'm not sure why. John _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
