Re: [9fans] Grub Boot Problem
- From: leimy2k@xxxxxxxxx (David Leimbach)
- Date: Tue, 28 Mar 2006 19:56:47 GMT
On 3/28/06, Russ Cox <rsc@xxxxxxxxx> wrote:
It's really not that magical. The BIOS itself has to map interruptI think I get what you mean, but why would it swap the drive letters ?
I mean if grub already found the right partition and started it, why
would it then mess up while calling 9load ?
PCs are only meant to boot from the first hard drive.
Grub does some kind of magic that I don't fully understand
to convince the BIOS to pretend that the second hard drive
is the first and vice versa. If that doesn't get through to
pbs/pbslba or gets through the wrong way, then pbs/pbslba
could easily end up looking at the wrong disk for the 9fat
partition.
Russ
0x13 (which is originally the floppy controller from way back before
hard disks were available in PCs) to 0x40. Then 0x13 has to deal with
things that are really floppy requests and subsequently call 0x40.
(bit 7 being high in the DL register meant it was a hard disk...
anything else was a floppy drive)
Replacing interrupt handlers in DOS was very commonplace. It should
be possible to remap which disk is being booted from (CMOS parameters
do this, and so can GRUB, just like you could in DOS.)
0x80 is typically C:
0x81 is typically D:
Machines (and BIOSes) after 1995 typically could support 4 drives.
Hmmm, that semester of 16bit PC assembly in college just came in handy
I guess.... (I admit, I had to look at my notes :-)
It used to be fun to rewrite the keyboard interrupt (0x9 IIRC) such
that it would dork around with the vga panning registers everytime
someone hit spacebar. DOS was so hackable.
Dave
.
- References:
- Re: [9fans] Grub Boot Problem
- From: Russ Cox
- Re: [9fans] Grub Boot Problem
- Prev by Date: Re: [9fans] [OT] hypothetical questions
- Next by Date: Re: [9fans] [OT] linus, the early years
- Previous by thread: Re: [9fans] Grub Boot Problem
- Next by thread: [9fans] something like cscope/cbrowser?
- Index(es):
Relevant Pages
|
