I've noticed that the system's idea of the IDE disk geometry seems to
have changed from 2.4.xx.
In particular, I've got my 40G HD on a Promise PDC20268 PIC, and find
that the (logical) CHS is different, while the LBA sectors remains constant
(and correct according to the HD label).
2.5.xx does have the right values in that CxHxS = LBA blocks, but fdisk and lilo
get a bit stroppy about the partition boundaries that were created from the CHS
that were generated under 2.4.xx, so I will have to patch them up.
Question is - what is determining that initial value that becomes the "logical"
CHS, and does it matter?
The fdisk and friends man pages are a bit vague on this, cfdisk says "picking
255 heads and 63 sectors/track is always a good idea", but I find 2.4.xx uses
nn/255/63, while 2.4.xx uses mm/16/63, and as I dual boot with win98, consistent
partition table behaviour is important to me.
Advice anyone?
Aside - RedHat has dropped cfdisk from util-linux in their distro versions 7.2 ff.
Given the bad words said about fdisk, what did cfdisk do to be ostracised?
On Fri, Oct 04, 2002 at 11:47:16PM +1000, Allan Duncan wrote:
> Question is - what is determining that initial value that becomes the "logical"
> CHS, and does it matter?
No, it does not matter at all.
CHS are meaningless numbers not used anywhere anymore in Linux.
If you want to influence what geometry *fdisk will use, give it
the appropriate options or commands. No need to go via the kernel.
But only in rare cases is it necessary to worry about geometry.
Andries
> Aside - RedHat has dropped cfdisk from util-linux in their distro versions 7.2 ff.
> Given the bad words said about fdisk, what did cfdisk do to be ostracised?
RedHat thought cfdisk is buggy.
They were mistaken.
Andries,
If CHS is truly meaningless (less drives smaller than 8.4GB) why can we
not specify forced LBA geometry reporting? Also any drive supporting
48-bit feature sets are forbidden to use CHS.
Just a comment, not bait for a lesson or lecture :-)
On Fri, 4 Oct 2002, Andries Brouwer wrote:
> On Fri, Oct 04, 2002 at 11:47:16PM +1000, Allan Duncan wrote:
>
> > Question is - what is determining that initial value that becomes the "logical"
> > CHS, and does it matter?
>
> No, it does not matter at all.
> CHS are meaningless numbers not used anywhere anymore in Linux.
>
> If you want to influence what geometry *fdisk will use, give it
> the appropriate options or commands. No need to go via the kernel.
> But only in rare cases is it necessary to worry about geometry.
>
> Andries
>
> > Aside - RedHat has dropped cfdisk from util-linux in their distro versions 7.2 ff.
> > Given the bad words said about fdisk, what did cfdisk do to be ostracised?
>
> RedHat thought cfdisk is buggy.
> They were mistaken.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
Andre Hedrick
LAD Storage Consulting Group
Andries Brouwer wrote:
>
> On Fri, Oct 04, 2002 at 11:47:16PM +1000, Allan Duncan wrote:
>
> > Question is - what is determining that initial value that becomes the "logical"
> > CHS, and does it matter?
>
> No, it does not matter at all.
> CHS are meaningless numbers not used anywhere anymore in Linux.
>
> If you want to influence what geometry *fdisk will use, give it
> the appropriate options or commands. No need to go via the kernel.
> But only in rare cases is it necessary to worry about geometry.
Like LILO. It complains that the partition tables don't match the
geometry, or somesuch, despite an explicit "lba32".
Maybe I should start looking at grub, assuming it doesn't do the same.
On Mon, Oct 07, 2002 at 07:48:46AM +1000, Allan Duncan wrote:
> Like LILO. It complains
Maybe sufficiently recent LILO is OK.
> Maybe I should start looking at grub, assuming it doesn't do the same.
"lilo -R" is very useful. I don't think grub has that capability.
Andries
Allan Duncan wrote:
> Like LILO. It complains that the partition tables don't match the
> geometry, or somesuch, despite an explicit "lba32".
If everything else is fine, and if this is one of the errors from
the partition table checks, you could turn it simply into a
non-fatal warning with the config option "IGNORE-TABLE".
- Werner
--
_________________________________________________________________________
/ Werner Almesberger, Buenos Aires, Argentina [email protected] /
/_http://www.almesberger.net/____________________________________________/