HP DL140 install requires pci=noacpi kernel boot option

Could you please send the output of lspci and dmidecode?

(Or, really, any way we can identify this crazy laptop [apparently people make
laptops that aren't X40s or ShinyBooks -- counter-revolutionaries] to
special-case it.)

(In reply to comment #2)
> (Or, really, any way we can identify this crazy laptop [apparently people make
> laptops that aren't X40s or ShinyBooks -- counter-revolutionaries] to
> special-case it.)

It's a server. I can't give you the requested info because none of our 3 are
plugged in at the moment. Next time I'm in the data centre, I'll get it.

Oh right, a DL -- I should've read more carefully. Good times.

There's no way that we can do any kind of detection to solve this; it breaks
things for the installer. Ideally, the kernel needs to be fixed so that it
doesn't break. Maybe try the current ACPI patch.

Failing that, I suppose we can document it in the isolinux help...

OK, so unless someone wants to ship me a DL140, I can't do anything about this
bug; I'm certainly not rebooting fd.o to try out crazy experimental kernels that
might break the networking. ;)

Failing that, I suppose you can document it in the isolinux help...

Just uploaded d-i with the additional help text.

That's a workaround at least; downgrading. CCing Herbert in hopes that he knows
where to forward this, especially whether it is likely a hardware/BIOS bug or a
kernel bug

(In reply to comment #9)
> That's a workaround at least; downgrading. CCing Herbert in hopes that he knows
> where to forward this, especially whether it is likely a hardware/BIOS bug or a
> kernel bug

As most kernel bugs do, this should be refiled at bugzilla.kernel.org.

But please provide more information when it comes to kernel bugs. dmesg and
dmidecode would be a good start.

And Matt's comment about downgrading is interesting. Which version can work
with these devices using ACPI? If we have a working kernel, then it makes the
job of identifying the problem much easier.

(In reply to comment #10)
> (In reply to comment #9)
> > That's a workaround at least; downgrading. CCing Herbert in hopes that he knows
> > where to forward this, especially whether it is likely a hardware/BIOS bug or a
> > kernel bug
>
> As most kernel bugs do, this should be refiled at bugzilla.kernel.org.
>
> But please provide more information when it comes to kernel bugs. dmesg and
> dmidecode would be a good start.

I'll get that next week when I'm in the data centre.

> And Matt's comment about downgrading is interesting.

I think Matt meant 'downgrading in bugzilla'...

(In reply to comment #11)
> > > That's a workaround at least; downgrading. CCing Herbert in hopes that
> >
> > And Matt's comment about downgrading is interesting.
>
> I think Matt meant 'downgrading in bugzilla'...

I could swear that it was a colon when I looked at it :)

This may be related to:

http://bugzilla.kernel.org/show_bug.cgi?id=2122

That bug describes a failure to boot on another HP DLxxx system unless
pci=noacpi was specified. In that case, a certain driver seemed to get hung up,
but given the similar make, if there is a BIOS compatibility issue, it might be
the same

James, your only symptom was that it didn't detect some PCI devices?

(In reply to comment #13)
> This may be related to:
>
> http://bugzilla.kernel.org/show_bug.cgi?id=2122

Thanks for the link - I'll get the info they ask for in there on Monday.

> That bug describes a failure to boot on another HP DLxxx system unless
> pci=noacpi was specified. In that case, a certain driver seemed to get hung up,
> but given the similar make, if there is a BIOS compatibility issue, it might be
> the same
>
> James, your only symptom was that it didn't detect some PCI devices?

Yes, AFAICR, it couldn't see the Tigon3 card. The machines we have are hideous
cheap-as-dirt 1U boxes that won't have a CCISS RAID card in them.

James, were you able to get the requested information from the machine?

Sorry for the inordinate delay - these boxes install fine with Warty final. Yay.