Re: Cannot boot 2.6.26 kernel on HPPA N4000
Moritz Muehlenhoff wrote:
> On Sun, Jul 26, 2009 at 08:39:22AM -0600, Grant Grundler wrote:
> > Moritz,
> > thanks for forwarding...
> > On Sat, Jul 25, 2009 at 04:10:04PM +0200, Moritz Muehlenhoff wrote:
> > > > Attempting to boot a 2.6.26 kernel on an HP N4000 machine (64 bit PA-RISC)
> > > > yields the following results (beginning at system startup):
> > > > <<Beginning of error>>
> > ...
> > > > Elroy version TR3.0 (0x4) found at 0xfffffffffecf4000
> > > > PCI: Address space collision on region 0 of device 0000:d0:00.0
> > > > [70e0800:70e08ff]
> > ...
> > > > sym53c8xx 0000:d0:00.0: enabling device (0000 -> 0003)
> > > > sym53c8xx 0000:d0:00.0: enabling SERR and PARITY (0003 -> 0143)
> > > >
> > > > ************* SYSTEM ALERT **************
> > ...
> > > > 0x0000187000FF6292 00000000 00000000 - type 0 = Data Field Unused
> > > > 0x5800187000FF6292 00006D02 100F120F - type 11 = Timestamp 03/16/2009
> > The address space collision is likely the cause of this HPMC.
> > (0xff6292 == HPMC)
> > With HPMC's, "PIM" info is worth collecting. See
> > http://www.parisc-linux.org/faq/kernelbug-howto.html
> > for details on how to collect PIM info.
> > It would also be helpful to collect "in io" output from the same PDC
> > prompt that allows one to run "ser pim" and "clearpim".
> > Lastly, if an older kernel does boot, "lspci -v" would be helpful.
> > > Is this a known issue, has it been fixed in current kernels from unstable?
> > The "parenting" of resources in the generic PCI support has been changed.
> > I can't say if it fixes this problem. Is it possible to test a 2.6.30
> > or 2.6.31-rc kernel?
> That will be difficult for the bug submitter, since the current versions
> of the debian installer are not yet based on 2.6.30 AFAICT.
> However, maybe the PCI card, which causes the collision can be temporarily
> removed, so that the installation proceeds. After that, the 2.6.30 kernel
> from unstable could be installed and the PCI card re-plugged in.
> Marc, would you be able to test this?
Marc, would you be able to test this?