[Top] [All Lists]

RE: dependency on rs_init

Subject: RE: dependency on rs_init
From: Harald Koerfgen <>
Date: Wed, 26 Aug 1998 20:00:09 +0200 (MEST)
In-reply-to: <>
Organization: none
Reply-to: "Harald Koerfgen" <>
Hi all,

On 26-Aug-98 Thomas Riemer wrote:
> The dependency on rs_init is slightly more complicated than 
> first glance.

Yes, Tom, the situation is actually somewhat confusing, due to lack of device
drivers. There are two possibilities to get output from a DECstation at the

One is the real(tm) SERIAL_CONSOLE in drivers/tc/decserial.c. This works only
on DS5K/1xx and DS5k/2x0 and would, with a little hack, on DS5k/xx (all have
Z85C30 based serial ports). (If you're interested: replace 0xbc18000? with
0xbc10000? in serial_putc(...) for your maxine, but I don't know if that will

The other one is using the PROM routines (works on all DECstations). This is
what promcon.c in arch/mips/dec is for and is intended as a temporary
solution until we have charakter device drivers to use as console devices for
all DECstations. From the kernels point of view, this is a SERIAL_CONSOLE too,
even if the PROM directs the output to your framebuffer.

To make a long story short, at the beginning of arch/mips/promcon.c you'll

#if 0
__initfunc(int rs_init(void))
        return 0;

Replace the #if 0 with #if 1 and disable CONFIG_TC. Your DS 2100 doesn't have a
TURBOchannel anyway. I didn't want to make all this configurable, because
promcon.c will go away as soon as we have at least serial console code for DZ
based DECstations.

Olivier, what do you think about this? 
> drivers/Makefile has an ifdef CONFIG_TC to include the /drivers/tc
> subdirectory.  
> drivers/tc/decserial.c contains the function rs_init required
> if CONFIG_SERIAL is set in drivers/char/tty_io.c.
> -Tom

I apologize for the confusion and hope this helps.

Keep hacking.

<Prev in Thread] Current Thread [Next in Thread>