Hello,

> Hm, it should work the same way as the smg stuff does.  I don't have
> a diff handy but it should be like:

I added a similar chunk of code SMG uses in conf.c, substituting smg
with lcg in declarations. I also added needs-flag to conf/files.vax for
lcg so that lcg.h is generated containing #define NLCG 1. I have
verified that i have lcgcnprobe() function in place and it's there,
similar to that of lcspx or smg. In theory it should all work now i
suppose but in reality it doesn't.

How can i verify whether lcgcnprobe has been called at all (without
using global variable to be set in lcgcnprobe and its contents printed
out afterwards, when DZ console is already working ... this woudl be
sooo kludgy) ?

> Note that two of them are required:
> 
>         lcgcnprobe() which is called early in the boot process
>         (to find out where the console is).
> 
>         lcgprobe() which is called during autoconfig.

Um, i don't seem to have the latter. Neither does smg.c for that matter.
Are you referring to *_match() function by any chance here ?

Anyway, seeing Michaels comment the other day (about DZ ports used for
LK keyboard, depending on position of S3 switch) i was thinking changing
where lkkbd is attached might solve the problem. Can anybody confirm
which serial port (starting with 0) it is that keyboard is connected to
when in serial console mode (so i can avoid data getting dumped into
keyboard when it's set as line 0) ? If it is port 3 should this get it
working in kernel config file:

lkkbd0	at dz0 line 3

? It's set to line 0 in GENERIC config file. I haven't tried it out yet,
just asking ahead of time when i find it doesn't work :-)

Blaz Antonic
-- 
Hi! I'm a signature virus!
Copy me into your signature to help me spread!