Re: [vox-tech] RHL 7 and kernel 2.4.0 woes

[Peter Jay Salzman <p@dirac.oMAPSrg>]

dang,  hate to do follow up on my own post, but...

On Thu 15 Feb 01, 11:37 AM, Peter Jay Salzman said: 
> hi jan,
> 
> jeff took the words right out of my mouth.   :)
> 
> but let me add to it.  "root=" is indeed a _VALID_ parameter to pass to the
> kernel, but it doesn't do what you think it should.
> 
> when you pass "root=" to the kernel, it tells the kernel HOW to mount the
> root filesystem.  the valid entries are "ro", "rw" or "debug".   as you can
> imagine, ro means to mount / as read only.   rw means to mount / as read
> write, and debug effects the "level of importance" for which messages are
> printed to the console during kernel initialization.
> 
> what you want to do is tell the kernel where to FIND the filesystem, not how
> to mount it.   to do that, you use the "root=" parameter as a lilo
> directive, not as a kernel argument (which is what append= does).
> 
> some lilo commands get written directly into the kernel itself.  that's why
> you can willy-nilly

of course, i meant "can't".   looks like i'm the willy.  :(

also, i should've said "with all other things being equal".  can't expect a
kernel compiled without ext2 support to work on all machines either.   :)

p

> compile a kernel and have it work on all machines.  the
> location of / is hardwired into the kernel image.  looking at the issue this
> way, it should make more sense that you want to issue a lilo directive, not
> an argument to be passed to the kernel at boot time.
> 
> pete
> 
> On Thu 15 Feb 01, 11:11 AM, jdnewmil@dcn.davis.ca.us said: 
> > On Thu, 15 Feb 2001, Jan Wynholds wrote:
> > 
> > [...]
> > 
> > > Here's my scoop:
> > > 
> > > Last night I was trying to compile with old modutils
> > > (not 2.4.0 as recommended), old e2fsprogs (not 1.19 as
> > > recommended), and old ppp (not 2.4.0 as recommended). 
> > > So I upgraded all those, and tried again.  
> > > 
> > > And got a bit farther.  Makes and such were done to
> > > completion and without terrible errors.  So, I
> > > modified lilo.conf, ran lilo (in the make install
> > > script), and crossed my fingers.  Anyhow, when I tried
> > > to boot, the messages went somewhat like:
> > > 
> > > Uncompressing linux.......
> > > Ok booting the kernel
> > > TCP Stuff....
> > > IPv4 blah blah
> > > [more lines of stuff]
> > > kernel panic: VFS: unable to mount root file system:
> > > /dev/hda2 bad parameter
> > > 
> > > Or some such.  It's roughly 3:30 am, and I don't
> > > remember the exact messages, but it hangs there.  I
> > > have to reboot and use the old kernel :(.  It's weird,
> > > cause it begins to load some stuff, but when it tries
> > > to find the root file system, it hangs.  In the docs
> > > it says to use a line in lilo.conf to pass the boot
> > > partition to the kernel so I added the line:
> > > 
> > > image=vmlinux-2.4.1
> > >   label=linux241
> > >   append = "root=/dev/hda2"
> > 
> > mmm... I think the "root=" is lilo information, not kernel information.
> > Put it in just like the label line... no append.
> > 
> > ---------------------------------------------------------------------------
> > Jeff Newmiller                        The     .....       .....  Go Live...
> > DCN:<jdnewmil@dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
> > Work:<JeffN@endecon.com>              Live:   OO#.. Dead: OO#..  Playing
> > Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
> > /Software/Embedded Controllers)               .OO#.       .OO#.  rocks...2k
> > ---------------------------------------------------------------------------
> > 
> 
> -- 
> "It's better to be safe than assimilated."                   p@dirac.org
>                       -- Chakotay                            www.dirac.org/p



-- 
"It's better to be safe than assimilated."                   p@dirac.org
                      -- Chakotay                            www.dirac.org/p