Kernel Bug Tracker – Bug 41522
Sun Blade 100 kernel panics during boot
Last modified: 2011-09-02 01:12:55 UTC
Created attachment 69562 [details]
blade 100 kernel panic image 1
My Sun Blade 100 hangs when I boot any recent kernel. I've tried many Debian and Gentoo installation CDs, but every CD made since 2008 or so prints "console [tty0] enabled, bootconsole disabled" and hangs. I have to use Gentoo's 2007.0 install CD to in order to complete booting and not disable the bootconsole.
I've built Linux 3.0.3 and tried it with the same results. When I boot with the "keep_bootcon" kernel parameter I see that Linux is kernel panicking in pci_scan_one_pbm. I'm attaching three photos of the full kernel panic.
Created attachment 69572 [details]
blade 100 kernel panic image 2
Created attachment 69582 [details]
blade 100 kernel panic image 3
Created attachment 69592 [details]
I'm not sure that any sparc maintainers follow bugzilla. You'll probably have better luck using the firstname.lastname@example.org mailing list.
Created attachment 70002 [details]
kernel panic w/ boot_delay image 1
Created attachment 70012 [details]
kernel panic w/ boot_delay image 2
Created attachment 70042 [details]
kernel panic w/ boot_delay image 3
Created attachment 70062 [details]
kernel panic w/ boot_delay image 4
Created attachment 70082 [details]
kernel panic w/ boot_delay image 5
Created attachment 70092 [details]
kernel panic w/ boot_delay image 6
Created attachment 70102 [details]
kernel panic w/ boot_delay image 7
Created attachment 70112 [details]
kernel panic w/ boot_delay image 8
Created attachment 70212 [details]
On Wed, Aug 31, 2011 at 03:59:56PM -0400, David Miller wrote:
> There is a non-trivial problem with the device tree reported by your
> firmware, several PCI device nodes are reported multiple times.
> If you'll look 'isa' appears twice, and so does 'pmu'.
> This is a very serious issue.
> And that's what is causing all of these boot failures.
> I have two theories, either the Tulip card makes the firmware corrupt
> the device tree like this. Or, alternatively, there is some bug in
> the version of OBP installed on this machine.
> This really isn't a kernel bug. And if it worked in the past, it worked
> entirely by accident.
Thanks so much for the info! I finally have 3.0.3 booting normally.
First thing I did was update to the latest firmware, 4.17.1, which can
be found at http://ftp.sunet.se (the readme is here:
This was a bit of a chore because neither the SILO nor the Solaris
loader wanted to boot the firmware update. I had to setup a tftp
netboot server in order to patch the firmware.
But, that didn't fix the problem.
Next, I yanked the tulip card. That didn't fix the problem either.
Finally, I reset nvram by typing this at openprom's "ok" prompt:
A reset-all and... it works! I can boot 3.0.3!