linux-origin
[Top] [All Lists]

Re: CVS Update@oss.sgi.com: linux

To: ralf@xxxxxxxxxxx (Ralf Baechle)
Subject: Re: CVS Update@oss.sgi.com: linux
From: Kanoj Sarcar <kanoj@xxxxxxxxxxxxxxxxxxx>
Date: Sun, 27 Aug 2000 09:54:57 -0700 (PDT)
Cc: linux-origin@xxxxxxxxxxx
In-reply-to: <20000827000332Z42234-31375+281@oss.sgi.com> from "Ralf Baechle" at Aug 26, 2000 05:03:31 PM
Sender: owner-linux-origin@xxxxxxxxxxx
> 
> CVSROOT:      /home/pub/cvs
> Module name:  linux
> Changes by:   ralf@xxxxxxxxxxx        00/08/26 17:03:21
> 
> Modified files:
>       arch/mips64/arc: Makefile console.c 
>       arch/mips64/kernel: setup.c 
>       arch/mips64/sgi-ip27: Makefile ip27-setup.c 
> Added files:
>       arch/mips64/sgi-ip27: ip27-console.c 
> 
> Log message:
>       Desinfect console code from #ifdefmania.
> 
> 

Ralf,

This does not work, I tried it before I checked the console
init code in. It _seems_ to be working, it is not _really_ 
working. This is part of the work I am doing to determine
serial port addresses at run time. 

In any case, I am trying to work around other ioc3 oddities 
currently. Once I resolve that, I will most probably revert
the code back to the original form I checked it in (after I 
convince myself, again, that the console init setup in ip27_setup
will not work). More discussions of this deferred till that time ...

If you want to argue what I am saying is wrong, here's a simple
experiment: change ip27prom_console_write to do

        prom_printf("IP27 CON: %s", s);

Then all kernel printks should print the "IP27 CON" before the
printk strings. 

The register_console at least must be done after the paging_init()
has completed in setup_arch(). 

Kanoj


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