2.1.116/7 mis-defaults to LBA on non-LBA partitions

Anil B. Somayaji (soma@cs.unm.edu)
20 Aug 1998 22:31:19 -0600


--pgp-sign-Multipart_Thu_Aug_20_22:31:12_1998-1
Content-Type: text/plain; charset=US-ASCII

I've managed to trash my disk today, and now I think I have a vague
understanding why. (Luckily, I made a backup before I really did some
damage.) I was installing 2.1.117 on a machine running 2.1.116, and
lilo complained about the partitioning. I tried to have lilo fix
it...anyway, here's what happened.

My machine is a HP Pavilion 8260, which is a PII/266 with a LX chipset
and a IDE Quantum Bigfoot CY4320A. I had my BIOS set to do "DOS"
geometry translation, giving the drive a 560/240/63 geometry.
Everyone seemed happy with this, until 2.1.116.

Even though it detected the geometry as 560/240/63, and there was an
existing, fully populated partition with this geometry, it decided to
do LBA addressing, reporting a line:

hda: [LBA] [527/255/63]

which is wrong!

I had been running 2.1.115 with no problems on this machine, and I
noticed a comment in the 2.1.116 patch for linux/drivers/block/ide.c,
which says:

+ * Version 6.18 default unpartitioned-disk translation now "BIOS LBA"

It seems like this happened to me on a regularly partitioned disk.

Now, if I create partitions under 2.1.117, 2.0.34 (which is on my
rescue disk) recognizes the LBA geometry (saying [PTBL] instead of
[LBA]), and Linux fdisk causes no problems. Partition Magic, however,
considers this partitioning to be broken.

If I set the bios translation to "Other", on boot I get a line saying
CHS=8960/15/63, and on a clean disk I consistently get 527/255/63
geometry from Linux fdisk, no matter what kernel. However, if
partitions exist from the 560/240/63 geometry, 2.0.34 correctly
defaults to it, but 2.1.117 reports the LBA geometry.

I can provide more details if needed. Also, I'm willing to do some
destructive experiments before I restore the drive. Any suggestions?

--Anil

-- 
Anil Somayaji (soma@cs.unm.edu)
http://www.cs.unm.edu/~soma
+1 505 872 3150

--pgp-sign-Multipart_Thu_Aug_20_22:31:12_1998-1 Content-Type: application/pgp-signature Content-Transfer-Encoding: 7bit

-----BEGIN PGP MESSAGE----- Version: 2.6.2

iQCVAwUBNdz4FkLkmkLHxWM5AQFPmwQAx9LR/YWV5fmLS07siKV8DqPrytM51RyV IE6RToN+5T5TNVrRhW7jYSnCDwp/Jnn2fsv888sAFQ8VxMV+WAe/qA5C4UKygAva KPgEIQO4fXY5qB0TCn3KJ+RwfcJnODQdW/qfsx2WwjjLz7Tn46FEASOF1OajgQj4 IUhrLhHzhZs= =FB80 -----END PGP MESSAGE-----

--pgp-sign-Multipart_Thu_Aug_20_22:31:12_1998-1--

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html