On 01/02/2013 02:18 PM, Michael Brown wrote:
On Wednesday 02 Jan 2013 09:41:54 Shao Miller wrote:
Anyway, since it's working without iPXE, perhaps the iPXE INT 0x13 hook
has a bug where the AH register is being incremented. AH=0x4D means
"Return Boot Catalog," but if it was an accidental increment, it
might've been AH=0x4C, which is "Initiate Disk Emulation & Boot," which
makes a whole lot of sense for the scenario you've described.
I'm pretty sure it will be triggered by iPXE's int13_boot(), which does:
/* Look for a usable boot sector */
if ( ( ( rc = int13_load_mbr ( drive, &address ) ) != 0 ) &&
( ( rc = int13_load_eltorito ( drive, &address ) ) != 0 ) )
return rc;
When int13_load_mbr() fails (since drive 0xa0 does not have a valid MBR
present), we try int13_load_eltorito(). This issues INT 13,4d00 to read the
El Torito boot catalog, relying upon the BIOS to return with an error if the
drive isn't actually a CD-ROM.
I have no idea why VirtualBox would treat INT 13,4d00 on a non-CDROM drive as
being worthy of a fatal error.
You should use function 4B01h first to verify that you are talking to a
CD-ROM.
-hpa
--
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel. I don't speak on their behalf.
_______________________________________________
ipxe-devel mailing list
[email protected]
https://lists.ipxe.org/mailman/listinfo.cgi/ipxe-devel
