> > > as it apparently makes CONFIG_IDEDMA_IVB a complete no-op?
> >
> > Exactly what it is designed to do, Ignore Validity Bits, because the whole
> > damn messedup the rules between ATA-4 and ATA-6
>
> I think the question is more - so why not lose the ifdef
> -
Because there are the exceptions that get it correct based on the level of
ATA support reported in the IDENTIFY page.
When I state that it is all screwed up, I mean in a contigious nature of
the Standard. The rules for ATA-4 with ATA-4 limited support is correct
as is ATA-5 with ATA-5, and ATA-6 with ATA-6, but ATA-4 rules do not mix
with ATA-5 nor ATA-6. This is the mess in front of me to sort.
So we default a full test of both bits 13 and 14, but it you have a
hardware combination that fails the rules.
ATA-4 is HOST side and Device side based on Bit 13 only
ATA-5 is HOST side and Device side based on Bit 14 only
ATA-6 is HOST side and Device side based on Bit 14 and Bit 13
ATA-6 is the correct method...
Cheers,
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/