On Fri, 2006-08-25 at 12:33 -0400, William L. Maltby wrote:
On Fri, 2006-08-25 at 17:15 +0100, Will McDonald wrote:
On 25/08/06, William L. Maltby BillsCentOS@triad.rr.com wrote:
And I still don't, but maybe a clue? Kernel parameter *if* allowed. I'm looking for the list of those on my machine now. Haven't found them yet, but ...
Was having so much fun, forgot to make my suggestion, based upon the man page. Maybe TERM=linux-m at end of grub line vmlinuz... ?
Tried this, NG. But with what Will discovered below, maybe the OP is approaching nirvana?
I think I need to read up on grub. I also tried the "append=" stuff, but no help (I didn't read up on it, I wonder if I used the right syntax? Maybe that's only a LILO construct?).
Nope. My original "TERM=" and "append=..." looks OK.
There's further documentation in /usr/share/doc/initscripts-7.93.24.EL/sysconfig.txt about the options
<snip>
There could be more in there but my "info" navigation's useless :)
Well, I navigated. But it didn't help. Good navigation doesn't help the case of no valid destination! ;-)
IIUC, all this, including the "man init" statement about receiving CONSOLE from kernel, relates only to devices, but doesn't help on the "terminal type" (TERM=).
This reinforced by both the grub stuff and the sysconfig text that Will referenced.
I'll take one last stab by "info grub", viewing the /usr/share/doc/initscripts-7.93.24.EL/sysconfig.txt Will discovered and then move on to things I need to do.
Kernel param list didn't help. The closest I can see is what Dag originally found, the BOOTUP in sysconfig for init. And that comes too late to address output before init is invoked. Kernel seems to have a *few* "pre-defined" types of displays it supports. VGA, some serial, a couple specials (for HP) and MDA (IIRC, a really old type?).
With the assumption that everything is VT100 (or ANSI) compatible and no way I discovered to pass a kernel param that changes the behavior (or TERM= equivalent value), I think the BOOTUP is as close as you get *unless* you make your own kernel. In there you should be able to suppress the "printk" escape sequences that produce "bold", etc.
The is a parameter, "earlyprintk=" that *might* offer a slim hope. But you'll have to chase that, I'm on to other things I need to do.
Will
<snip sig stuff>