We first INIT all IDE-code.
Then we do the SCSI-code.
Next we register SCSI-disks.
Finally we register IDE-disks.
In the meantime, we later poke the partition table.
I may not have the exact order, but it goes something like above.
On Thu, 21 Jan 1999 Andries.Brouwer@cwi.nl wrote:
> After installing a new large IDE disk in some machine, I found
> to my surprise that Linux detects the geometry as 1020/200/62
> while the BIOS had 16383/16/63, as was to be expected for
> a large disk. The CMOS has the right values.
> What goes wrong?
> Well, the IDE code gets the BIOS information from setup.S,
> and the code there gets hd0 and hd1 data via the INT 41 and INT 46
> addresses. But who says that hd0 and hd1 are IDE disks?
> This machine (Pentium with Award BIOS) has both IDE and SCSI
> disks, and boots from SCSI. And indeed, /dev/sda has geometry
> Thus, the IDE driver mistook "first disk" for "first IDE disk".
> Clearly, we would like to know which disks the BIOS called
> "first" and "second" disk. I guess that the first disk is
> the one on which the BIOS starts the boot loader.
> But the kernel no longer has this information.
> Moreover, maybe it was booted by loadlin or so.
> Is there something meaningful the kernel can do?
Nope.......a question of who gets first crack at booting is a major
rehack and a tought one too.
> Or should the user just tell the kernel on the command line:
> "first disk is sda"?
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to email@example.com
> Please read the FAQ at http://www.tux.org/lkml/
The (NEW) Linux IDE guy
The APC UPS Specialist for Linux
You just need a bigger hammer, or learn how to swing the one you have better.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to firstname.lastname@example.org
Please read the FAQ at http://www.tux.org/lkml/