> 48-bit lba has a non-significant overhead (twice the outb's, 12 instead
s/non-// ?
> of 6 per command), so it makes sense to use 28-bit lba commands whenever
> we can.
>
> + if (drive->addressing == 1 && block > 0xfffffff)
> + lba48 = 1;
Hmm. I wonder whether we should be more cautious, and ask for lba48
as soon as some part of the interval is past this limit.
(say, block+nsectors > 0xfffffff)
I don't know whether the standard spells out what happens
at the boundary, but for example the LBA low/mid/high, DEV is required
to contain the sector number at the place the error occurred,
and that is possible only if one stays below the 28-byte sector limit.
Andries
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/