Re: 5.2.1 Unbootable on my Hardware

On Sun, 7 Mar 2004, Brian F. Feldman wrote:

> Andre Guibert de Bruet <andy_at_siliconlandmark.com> wrote:
> >
> > On Fri, 5 Mar 2004, Doug White wrote:
> >
> > > On Fri, 5 Mar 2004, Nick Nelson wrote:
> > >
> > > > Hi folks,
> > > >
> > > > I've recently tried to install 5.2.1 on a new system of mine, basically
> > > > it's an AMD 3000XP, Radeon9800 Pro, Corsair XMS RAM, Lanparty Ultra II
> > > > (http://www.dfi.com.tw/Product/xx_product_spec_details_r_us.jsp?PRODUCT_ID=1524)
> > > > with a SerialATA (Marvell 88i8030 chip), I have one SerialATA drive and
> > > > one IDE drive.
> > >
> > > The Marvell chip is the SATA bridge, not the ATA controller.  Its probably
> > > the standard built-in nForce 2 ATA controller.
> > >
> > > > I've tried booting off both the SATA drive and the IDE drive, each time
> > > > it gets to:
> > > >
> > > > ---
> > > >
> > > > Timecounters tick every 10.000 msec
> > > > lo0: bpf attached
> > > >
> > > > ---
> > > >
> > > > and stops..I ran it with logging to get that much information. I've
> > > > unplugged the SerialATA, which I thought would resolve it, but didn't.
> > > > Any ideas?
> > >
> > > Hm ... try booting without ACPI, and if that doesn't work, try safe mode.
> > > Looks like the controller is locking up or something...
> >
> > This hang could be ULE-related.
> >
> > I'm saying this because it's the exact same thing that I see when I try a
> > ULE kernel on my Dual Athlon 2000+ MP running on a Asus A7M266-D using
> > regular ATA133 disks.  Luckily, it's been tracking current since before
> > the switchover, so I run a custom generic with 4BSD in it and everything
> > works fine.
> >
> > I wish I could offer more debugging information such a verbose boot or a
> > ddb ps when it hangs but I'm swamped with work from my job and cannot
> > seem to find the time to debug such problems.
>
> Are you seeing it lock up when /sbin/init executes /etc/rc?  When that would
> normally start doing the fsck -p output, I see nothing, and I can
> Ctrl-Alt-Esc to DDB and see the processes that should be there, but they
> just don't do anything.  If I could someone encode the dumpdev in the kernel
> (hint, hint) I could panic it there and get something more useful to go by.
>
> This only happens when I try to use SCHED_ULE.  I've always used SCHED_4BSD
> (and indeed, am still doing so now).  This is my normal boot output:

Last time I checked, I didn't get that far. This was a kernel from the the
25 of February. I guess I could give it another whirl sometime.

Regards,
Andy

> Andre Guibert de Bruet | Enterprise Software Consultant >
> Silicon Landmark, LLC. | http://siliconlandmark.com/    >