In the last couple of days we've been having a discussion in the
FreeBSD mobile list about a bug in newer IBM laptops.  See
http://www.pc.ibm.com/qtechinfo/MIGR-4QHLS4.html for more details.
Basically, it seems that the IBM BIOS uses partition type 165 for its
own purposes, and goes crazy if it doesn't find what it wants in it:

----- Forwarded message from Ken Key <key@network-alchemy.com> -----
> I spent most of last week repeatedly converting an A21P and a T21 into
> wortless bricks and back, trying to find a way to get FreeBSD running on
> them.  I distilled the death point to:
>
> - Do the install (I've used both boot0 and  LILO, doesn't matter), it's
>   a brick - won't even get into the BIOS menu.
> - Pull drive out and boot on my 600X.  Fire up fdisk and change it from 165
>   to something else.  Do not touch MBR or anything else.
> - Put drive back in (A21P and T21) and I can now get into the BIOS screens
>   or actually boot via boot0 or LILO into Win2K.
> - Put drive back in 600x, change back to 165, back in *21
> - It is a brick.
>
> So, while I don't know what IBM is doing with partition type 165,
> I am convinced it is a key to the boot lock-up problem for the
> T21 and A21P.  Now boot0 has grown from 1 to 2 sectors in length,
> maybe that's what "the story inside IBM" is trying to talk about -
> I don't know.  However, the above pattern was reproduced with LILO,
> independant of boot0, so I really don't think boot0 is related to
> the problem.
>
> We have a couple of T20s' running FreeBSD, so this particular behavior
> seems to be related to changes for the *21 models.  The T20 guys are
> never going to "upgrade" their BIOS after this mess.

If this analysis is correct, there should be no problems running
NetBSD or OpenBSD on these machines, since they use different
partition numbers.  Can anybody confirm or deny?

Greg
--
Finger grog@lemis.com for PGP public key
See complete headers for address and phone numbers