Messages in this thread | | | Date | Fri, 17 Apr 1998 13:13:13 +0200 | From | Martin Mares <> | Subject | Re: 2.1.93.. |
| |
Hello,
> <groan> You would have to ask that. Unfortunately, there's no way to predict > what the system BIOS will do. It depends on the order that each of the BIOS > extensions on the host adapters are executed by the system BIOS, and there's no > standard for this at all. Some system BIOS identify all the BIOS extensions on > peripheral cards and then sort them by increasing size, others by decreasing > size, and some simply by slot order. Then they must be copied to low memory > for execution. Whichever BIOS extension gets control first will generally > probe for devices and then install the disks the BIOS understands by hooking > the appropriate interrupt(s). Note also that once a BIOS extension executes it > can shrink itself back to a smaller resident size, so fitting everything > sometimes depends on the recognition order. It's not pretty at all. Certin > combinations of cards can only be used in a specific order in certain > motherboards because of this. It's a tech support nightmare.
# Since there is nothing in Linux that ensures that devices are numbered # according to the _actual_ BIOSes order, any tool based on such an # assumption is essentially broken.
IMHO only few users have more than one SCSI controllers and I think we can ask them to specify BIOS number for the disk they want to boot from manually.
Anyway, I've written a patch to PCI subsystem which uses BIOS ordering while still using direct hw access for touching configuration registers, but I think we should remove all this compatibility glue in 2.3.
| Hmmm... not mentioned in the man page. Also, we're straying
Probably not in the man page, but I'm sure it's mentioned in the documentation.
Have a nice fortnight -- Martin `MJ' Mares <mj@ucw.cz> http://atrey.karlin.mff.cuni.cz/~mj/ Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth "If at first you don't succeed, you must be using Windows ..."
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |