I have kernel 2.2.12-20 running on two machines - a HP 7280 Pavilion, and a separate one built using a Celeron 450 MHz and a Maxtor 20G drive. On both machines, after compiling a kernel specific for the machine, whenusing linuxconf to set it up as bootable, it informs me that either the partition is bad or is not a valid ext2 partition. It works just fine if I manually edit /etc/lilo.conf then run /sbin/lilo. I have only recently upgraded from the RedHat 6.0 to the 6.1 package. linuxconf worked fine when I was using the 6.0 kernel. (I believe the kernel was 2.2.5-15.)
The same happended to me (Toshiba 4000CDT Laptop, ran under RH 5.2 with self-compiled 2.2.13-Kernel before). After RH install had destroyed my lilo.conf, I couldnt4t enter the exisiting 2.2.13 Kernel into lilo.conf anymore
The same happended to me (Toshiba 4000CDT Laptop, ran under RH 5.2 with self-compiled 2.2.13-Kernel before). After RH install had destroyed my lilo.conf, I couldnt4t enter the exisiting 2.2.13 Kernel into lilo.conf anymore with linuxconf, as /dev/hda5 was, according to linuxconf, an invalid partition. (Nor could I set /dev/hda5 as default root partition for lilo). Hand editing worked fine. (sorry I submitted my last incomplete comment - Bugzilla newbie. please delete! Martin Wilck, Inst. for Tropospheric Research, Leipzig, Germany
Just the same on a networked workstation with SCSI disks - linuxconf doesn4t accept /dev/sda5 as root partition any more.
Just discovered that this seems to be the same bug as 5737.
*** This bug has been marked as a duplicate of 5737 ***