3-Jun-00 02:35 you wrote:
> * #61513: enabling `lba-support-bitmap-check' marks some systems unbootable
> That was just because he forgot to specify the option
> `--enable-lba-support-bitmap-check' for the configure script. Not a
> GRUB bug. Moreover, I've already changed the check to an
> installation-time option from a build-time option, so we don't have to
> keep it open.
Someone can explain from WHERE this weird check come anyway ?
-- cut --
/* Make sure that LBA read/write functions are supported. */
if (drp.flags & 1)
-- cut --
Ok. From Ralf Brown's interrupt list:
-- cut --
Bitfields for IBM/MS INT 13 Extensions information flags:
Bit(s) Description (Table 00274)
0 DMA boundary errors handled transparently
1 cylinder/head/sectors-per-track information is valid
2 removable drive
3 write with verify supported
4 drive has change-line support (required if drive >= 80h is removable)
5 drive can be locked (required if drive >= 80h is removable)
6 CHS information set to maximum supported values, not current media
15-7 reserved (0)
SeeAlso: #00273
-- cut --
So LBA read/write functions are supported if and only if "DMA boundary
errors handled transparently" ??? What a strange logic! Who's bright mind
generated such a great check ?
P.S. Yes, I know: Ralf Brown's list is not official documentation. From past
expirience I know though that if official documentation and Ralf Brown's
interrupt list say different things then 9 times out of 10 interrupt list is
correct and official documentation wrong.
P.P.S. At least my Asus K7M motherboard return "0" there while LBA works
just fine.
