Bill Studenmund wrote:

>One of the things I'd like to do as part of fixing up installboot to
>support Apple partitioned disks is get the "driver" table entries, located
>in sector 0, correct. I'd like to fix this as it might be partly why we
>can't boot on OF 2.4 disks. :-)
>
>There is a problem though. A driver consists of three fields, a start
>(int32_t), a length (int16_t), and a type (int16_t). The problem is that
>the way the structures are layed out, the type filed of the first driver
>will overlap with the first 2 bytes of the magic number in the disklabel
>we stick there.
>
>MacOS is OS type 1. I'm not sure if the boot loaders care about this
>number or not..

If you're talking about the same field I think you are there are some 
additional values that Apple is also using.  0x701, 0xffff, and 0xf8ff.  
The original value of 0x0001 meant to try to load/boot from a non-chained 
SCSI driver; the complement (0xffff) is reserved for the chained variant 
of the driver.  The 0x701 is the non-chained driver for the IDE, and the 
0xf8ff is the chained variant of the IDE driver.  This is documented, 
although not very well, in some of the new Tech Notes.

-bob