Cylinder/Head/Size Hell Is Not Over
pixel@mandrakesoft.com
$Date$
Nowadays, you can boot Linux without ever relying on CHS
(Cylinder/Head/Size):
- CHS information on partition table is obsolete and not used anymore.
It was causing the 1024th cylinder limitation:
- disallowing partitions above 8GiG. Problem overcome using 28bits Linear
Block Addressing (LBA) in partition table.
- disallowing to boot on partition above 8GiG (due to int13 limitation used
in bootloaders). Problem overcome using int13 extensions (EDD)
Logical CHS (Cylinder/Head/Size) and physical CHS are not related anymore.
Especially the size of a hard drive is not a multiple of the number of
Cylinders.
The problem is that many tools still rely somewhat heavily on CHS information.
- most tools adjust partitions on cylinder boundary
- some tools compute the size of the disk based on CHS (eg: fdisk)
Resulting problems in various tools:
- last partition not cylinder boundary adjusted, and overrun CHS computed size
- fdisk
> Total allocated sectors %d greater than the maximum %d
- cfdisk (need confirmation)
> FATAL ERROR: Bad primary partition %d: Partition ends after end-of-disk
or
> FATAL ERROR: Bad logical partition %d: Partition ends after end-of-disk
- Partition Magic (version?)
> Error #109: Partition ends after end of disk
- Windows <= 98se, Windows Me
(doesn't like it, what error message?)
Tools handling it correctly:
- parted and so diskdruid in RedHat >= 7.2 (or is it 7.x?)
- diskdrake >= Mandrake 8.2
- Windows XP (need confirmation)
Tools using the full size of the disk, and so causing pbs with fdisk/cfdisk/...:
- diskdrake == Mandrake 8.2
- disagreement between CHS and linear address
- parted
> "Warning: Unable to align partition properly. This probably means that
> another partitioning tool generated an
> incorrect partition table, because it didn't have the correct BIOS
> geometry. It is safe to ignore,but
> ignoring may cause (fixable) problems with some boot loaders.
> Ignore Cancel ? C
- lilo
> Warning: Kernel & BIOS return differing head/sector geometries for device 0x80
> Kernel: 38792 cylinders, 16 heads, 63 sectors
> BIOS: 1023 cylinders, 255 heads, 63 sectors
> Device 0x2100: Inconsistent partition table, 2nd entry
> CHS address in PT: 48:0:1 --> LBA (771120)
> LBA address in PT: 48384 --> CHS (3:3:1)
> The partition table is *NOT* being adjusted.
tect-lst/log/dkms-modules-info/2.6.33.4-desktop586-1mnb/vboxvideo.alias?h=0.6.40'>logtreecommitdiffstats