Gentoo LiveCD wont boot after BootX

chapium · n00b Joined: 03 Jun 2004 Posts: 11

Hello,

I am using the 2005.0 universal cd on an oldworld beige 266mhz PowerPC. After following the instructions from the manual and suggestions from many pages i've searched in google and this forum I am still stuck with the same symptoms.

There will be around 10-15 lines of text, and then Tux will appear in the top left corner. My CD-ROM light is on for around 15-30 seconds and then turns off. Tux is still grinning at me in his island of blackness.

Any ideas? :roll:

In BootX we have tried the following

lazloman · Posted: Sun May 22, 2005 3:54 pm Post subject:

I'm not looking at it right now, but I believe the boot kernel you want is called something like G3G4. Try that one, I'll take a look at my Old World box and get the boot loader arguments and send that along later. I know it boots, I rebuilt my Old World box using that CD just last month.

Here you go:
init=/linuxrc looptype=squashfs loop=/livecd.squashfs udev nodevfs cdroot root=/dev/ram0 video=atyfb:1280x1024

The kernel you want to use is G3G4 and the image is G3G4.igz. Make sure both of these are in your Linux Kernel folder.
I use the built-in video, so I have atyfb in my args. You'll need to adjust it and the resolution for your setup. There is radeon support in the kernle if you want to use that try radeonfb with your resolution settings. Good luck!
_________________
Gentoo x86
2x Dell Optiplex GX200

"My Break-Dancing days are over, but there's always the Funky Chicken."
--The Full Monty

chapium · n00b Joined: 03 Jun 2004 Posts: 11

Thanks, I'll give that a try. I've put similar settings (minus the video part) and had trouble. Interestingly I get other problems running yellowdog linux on it. Hopefully this works.

JoseJX · Retired Dev Joined: 28 Apr 2002 Posts: 2774

First, you should use this kernel instead on OW machines if you're having trouble booting:
http://dev.gentoo.org/~josejx/OW-Kernel.bz2 It is a drop in replacement for G3G4 and fixes a booting issue on some OW machines.

A few more things to try:

Try another video driver. You may need to use the exact video card for the framebuffer instead of the generic OpenFirmware one. I can't tell you your exact video card but it shouldn't be too hard to figure out. It's probably a Mach64 chipset, in which case you should try the atyfb driver instead.

Try passing 'console=tty0' as well as your other parameters, sometimes it outputs kernel messages to the serial port instead.

Let us know if anything worked for you and good luck!
_________________
Gentoo PPC FAQ: http://www.gentoo.org/doc/en/gentoo-ppc-faq.xml

dobbs · Posted: Tue Sep 20, 2005 12:42 pm Post subject:

I just now ran into this problem on a PowerMac 9600. In addition to chapium's symptoms, my system automatically reboots back to MacOS after three or four minutes. Maybe chapium didn't wait long enough to discover this. I'm attempting to boot the 2005.1 LiveCD on a 9600 (OF v1.0.5) with 256MB (interleaved) RAM and a 400MHz G4 upgrade (ZIF slocket, replaces the original CPU) with BootX v1.2.2. The kernel and .igz are copied to the HD and I'm using a 32768 "Ramdisk size". Kernel arguments follow:

disposableHero · n00b Joined: 07 Jul 2003 Posts: 13 Location: Austria

Hello!

I've got the same troubles with my PowerMacintosh G3 rev2, 266MHz, 128MB RAM, ATI mach64_3DU (as System Profiler in OS 9.1 says)

I've tried multiple kernels (ppc32, G3G5, OW-Kernel) and also tried different kernel args for booting the liveCD (2005.0, 2005.1) but after tux appears on a black background nothing more happens.

Are there any new informations about that problem?

thx

lazloman · Posted: Mon Nov 07, 2005 1:59 am Post subject:

Is this a beige G3? What are you using for your video args?
_________________
Gentoo x86
2x Dell Optiplex GX200

"My Break-Dancing days are over, but there's always the Funky Chicken."
--The Full Monty

disposableHero · n00b Joined: 07 Jul 2003 Posts: 13 Location: Austria

yes it is a beige G3 (rev2)
i've tried video=atyfb, video=ofonly and even none, they did all the same

dobbs · Posted: Fri Apr 21, 2006 7:42 am Post subject:

Hmm, the 2006.0 LiveCD boots the kernel fine (on my previously described 9600). Except it doesn't seem to detect SCSI or mount a ramdisk, thusly I'm SoL for a root partition. By the way, swapping out the upgrade card for the original 266MHz 604 didn't help the previous issue. Also of note, the kernel will sometimes load on one video card (the one specified in OpenFirmware) and then swap over to the second when it starts detecting hardware.

As a side question, is it normal for BootX to not remember the specified ramdisk size? It keeps resetting to 8192 no matter what I do. I guess I could ResEdit the damned thing...

Any information on configuring loopback devices with the BusyBox shell would be helpful (in terms of booting off a USB or FireWire storage device).

Edit: Nevermind, I'm just being stupid. Forgot to copy the new kernel into BootX's folder. Although it's still not loading the ramdisk: "Invalid loop location: livecd.squashfs"

Edit: Gah! More stupidity! It's supposed to be "image.squashfs"! I'm not having a very good night...

lazloman · Posted: Fri Apr 21, 2006 10:25 pm Post subject:

you can slo pass the ramdisk as a kernel argument and not worry about the pref holding. If I remember the parameter correctly, it should like this:
cdroot root=/dev/ram0 init=linuxrc ramdisk_size=32000 loop=livecd.squashfs looptype=squashfs console=tty0 nodevfs udev video=ofonly:1024x768@75
_________________
Gentoo x86
2x Dell Optiplex GX200

"My Break-Dancing days are over, but there's always the Funky Chicken."
--The Full Monty

dobbs · Posted: Wed May 03, 2006 9:54 am Post subject:

Ok, this one's getting on my nerves. It was working fine, save for not giving the SCSI HD a /dev entry, but that's not the point... Worked fine across several reboots. And now I'm back to the "invalid loop location" error. And yes, I checked the filename. Many times.

The trouble started when I tried adding "ramdisk_size=32768" to the BootX kernel args. Unless this somehow collided with the non-volitile memory space, that's just a coincidence; it looked like it was ignoring that parameter anyway.

The LiveCD is detected (/dev/sr0) and mounted at /newroot/mnt/cdrom. Current kernel parameters are "cdroot root=dev/ram0 init=linuxrc loop=image.squashfs looptype=squashfs console=tty0 nodevfs udev", though I've tried it with variations including doscsi, dofirewire, dousb, noide, dopcmcia (don't ask), scandelay=5, docache and with ramdisk sizes ranging from 16384 (yes, I know that's low) to 65536. ppc32 and ppc32.igz have been copied from the 2006.0 LiveCD to the HD (repeatedly). I even tried moving the drives to the MESH SCSI bus rather than the 53C94 SCSI bus (BeOS didn't much like the MESH bus...) without effect. And now I've removed all extra PCI cards, replaced the original 266MHz CPU, zapped the PRAM (apologies for the Mac term) AND booted MacOS without extensions and still no dice.

I'm out of ideas for a while.

Is /dev/ram0 supposed to show up in the list of mounted filesystems (calling mount without any parameters)? It doesn't, but I'm tired enough to wonder.

Do you guys have something against ELF images? Sorry, my mind wanders.

JoseJX · Retired Dev Joined: 28 Apr 2002 Posts: 2774

I know this is an old thread, but it came up on one of my OW machines so I thought I'd document a workaround. If you get the "invalid loop location" error, use the console it gives you to mount the image.squashfs image at /newroot/mnt/livecd (mount -o loop /newroot/cdrom/image.squashfs /newroot/mnt/livecd). Then, just exit the console to continue booting. It should work. :)
_________________
Gentoo PPC FAQ: http://www.gentoo.org/doc/en/gentoo-ppc-faq.xml