Re: network install problem - UltraSparc 60
On Thu, 24 Dec 2009 04:13 -0800, "Willie" <tumbleweed@fastmail.net>
wrote:
>
>
> On Wed, 23 Dec 2009 18:40 +0100, "BERTRAND Joel"
> <joel.bertrand@systella.fr> wrote:
> > Willie a écrit :
> > >
> > >
> > > On Wed, 23 Dec 2009 12:44 +0100, "BERTRAND Joel"
> > > <joel.bertrand@systella.fr> wrote:
> > >> Willie a écrit :
> > >>> Good day.
> > >>>
> > >>> I've managed to set up working rarp and tftp servers on an x86 debian
> > >>> laptop, and I see the headless U60 booting across the network via a
> > >>> terminal server on ttyA, but eventually the boot process hangs. Here are
> > >>> the last few lines of boot messages:
> > >>>
> > >>> --
> > >>>
> > >>> [ 0.000000] Zone PFN ranges:
> > >>> [ 0.000000] Normal 0 -> 393078
> > >>> [ 0.000000] Movable zone start PFN for each node
> > >>> [ 0.000000] early_node_map[2] active PFN ranges
> > >>> [ 0.000000] 0: 0 -> 131072
> > >>> [ 0.000000] 0: 262144 -> 393078
> > >>> [ 0.000000] Booting Linux...
> > >>> [ 0.000000] Built 1 zonelists in Zone order, mobility grouping on.
> > >>> Total pages: 259318
> > >>> [ 0.000000] Kernel command line:
> > >>> [ 0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
> > >>> [ 0.000000] clocksource: mult[238db] shift[16]
> > >>> [ 0.000000] clockevent: mult[7334e15e] shift[32]
> > >>> [ 266.802741] Console: colour dummy device 80x25
> > >>> [ 266.855795] console handover: boot [earlyprom0] -> real [tty0]
> > >>>
> > >>> --
> > >>>
> > >>>
> > >>> The boot server directory /var/lib/tftpboot contains:
> > >>>
> > >>> boot.img
> > >>> C0A89E64 --> boot.img
> > >>>
> > >>>
> > >>> ...which I obtained from
> > >>> /debian/dists/squeeze/main/installer-sparc/current/images/netboot
> > >>>
> > >>
> > >> Hello,
> > >>
> > >> It's not a tftp trouble. Your U60 is able to find your boot image. You
> > >> say that you don't have any framebuffer and output you have posted shows
> > >> that kernel tries to switch to a framebuffer to attach tty0. Can you try
> > >> with a frambuffer ? If not, you have to be sure that your boot image
> > >> contains all modules required to support console on serial line. If I
> > >> have more time this afternoon, I shall check on my U60.
> > >>
> > >> Regards,
> > >>
> > >> JKB
> > >>
> > >
> > >
> > > Thank you Joel,
> > >
> > > In fact there is an Expert 3D graphics card in the machine, but I don't
> > > have either a monitor or keyboard attached. I understand that in the
> > > absence of a real console a Sparc will drop back to ttyA for
> > > input/output?
> >
> > Of course, if you don't have any keyboard, openprom switch to ttya by
> > default, but you can configure this line. And in your case, openprom
> > switch to serial line. But if you want to access to your OS, it has to
> > support this serial line. Solaris support it by default, but I'm not
> > sure that debian does. Thus, you can see all messages sent by kernel to
> > openprom until kernel switches to framebuffer. I have installed some
> > T1000 without any trouble, but hardware support differs. Maybe there is
> > an kernel option for your case.
> >
> > If I remember, you should have a SunSU serial and builtin support in
> > boot kernel. If boot image does not contains support for console on
> > serial line, you have to write a bug report ;-)
> >
> > Regards,
> >
> > JKB
>
>
>
> Thanks again for your time.
>
> I'm not really interested in serial port access other tha using it to
> drive the installation process, after that I'll be able to get to the
> machine over the network.
>
> I've tried 'boot net console=ttya' and I've tried setting the bootprom
> to input-device=ttya, output-device=ttya, but neither of these made any
> difference to the OS trying to switch to the frame buffer and
> non-existent monitor.
>
> I can't think of anything else to try unless someone can prod me in the
> right direction.
>
Apologies and a correction to my last post...
Either passing "console=ttya" to the kernel, or setting
output-device=ttya in the bootprom gave a different outcome - lots of
infinitely fast streaming errors to my install console like so:
--
[ 518.944227] /pci@1f,4000: PCI AFAR [000001ff880c0800]
[ 518.944237] /pci@1f,4000: PCI Secondary errors [(Master Abort)]
[ 518.944267] /pci@1f,2000: PCI Error, primary error type[Master Abort]
[ 518.944280] /pci@1f,2000: bytemask[000f] UPA_MID[00] was_block(0)
[ 518.944292] /pci@1f,2000: PCI AFAR [000001ff410a0000]
[ 518.944301] /pci@1f,2000: PCI Secondary errors [(Master Abort)]
[ 519.506530] /pci@1f,4000: PCI Error, primary error type[Master Abort]
[ 519.506545] /pci@1f,4000: bytemask[000f] UPA_MID[00] was_block(0)
[ 519.506557] /pci@1f,4000: PCI AFAR [000001ff880c0800]
[ 519.506566] /pci@1f,4000: PCI Secondary errors [(Master Abort)]
[ 519.506597] /pci@1f,2000: PCI Error, primary error type[Master Abort]
[ 519.506610] /pci@1f,2000: bytemask[000f] UPA_MID[00] was_block(0)
[ 519.506621] /pci@1f,2000: PCI AFAR [000001ff410a0000]
[ 519.506631] /pci@1f,2000: PCI Secondary errors [(Master Abort)]
[ 520.068858] /pci@1f,4000: PCI Error, primary error type[Master Abort]
[ 520.068873] /pci@1f,4000: bytemask[000f] UPA_MID[00] was_block(0)
[ 520.068884] /pci@1f,4000: PCI AFAR [000001ff880c0800]
[ 520.068894] /pci@1f,4000: PCI Secondary errors [(Master Abort)]
[ 520.068923] /pci@1f,2000: PCI Error, primary error type[Master Abort]
--
Does this mean anything to anyone?
--
http://www.fastmail.fm - A fast, anti-spam email service.
Reply to: