[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

RE: LILO build error in pure64



>From Kurt Roeckx (at various times):
> > I've been trying to make a static version of it on i386 and then
> > set to arch field to amd64 but it failed.
> > There is no libdevmapper.a, so linking it staticly fails.
> I've build it without devmapper support now and it's available in
> the archive.  It's not tested though.

I installed it on my triple boot machine on the 64 bit partition.  The lilo
command ran fine with the config file I provided.  It complained and wanted
confirmation because
(1) it isn't an active partition (which is true, of course), and
(2) it didn't like partition type "0C" (which is odd; it's an ext3
partition).

I listed the 64 bit partition as an "other=" line in the lilo configuration
on the 32 bit partition, next to the lines that explicitly call out the 64
bit kernel and root partition.  Both of the configurations should be booting
the same 2.6.6 kernel image with the same options.

The one from 32-bit partition directly to 64-bit kernel still works.  The
one from 32-bit partition via 64-bit partition to 64-bit kernel works to the
point where the kernel has been loaded into memory and the boot starts.  At
that point, I get an immediate warm reboot to the BIOS logo screen.

It may not be the same problem, but the symptom described above can also be
generated by trying to boot Linux without writing "pci=noacpi" appended to
the kernel command line.  This is because the ACPI data is wrong on the
M6805 laptop - a known problem/workaround.  The configuration file has that
append information specified, so either lilo is not passing it along or
there is something else wrong with how lilo was placing the kernel image in
memory.

Hope that helps ...
    Alex.

PS:
If you want this filed as a bug against the package, let me know ... or do
it yourself.
8-)



Reply to: