On 9/20/19 3:25 PM, Tony Asleson wrote:
> On 9/20/19 11:18 AM, John Snow wrote:
>> For sure -- I handle the normative cases, but I don't test what happens
>> if you issue an unsupported NCQ command. (I don't know what real
>> hardware does right now, either. I'm sure I could read the spec and find
>> out, but don't have a testing setup that lets me analyze real hardware
>> anymore.)
>
> Regardless of what actual hardware does, it's always good to see what
> the spec says as hardware folks get it wrong too sometimes :-)
>
> -Tony
>
That depends. If we're emulating an "AHCI device", we should follow the
spec, but the one we instantiate in QEMU is the ICH9 version, so we try
to follow the hardware.
I do wind up looking at both, but because the relevant specs are split
across SATA, AHCI, ATA, ATAPI and into SCSI sometimes, (plus sub-specs)
it can be hard to piece together a holistic picture of what should
happen sometimes.
Looking at hardware is sometimes quicker and more definitive :)
Anyway, none of this helps you out much, but I will at least stay tuned
to review AHCI code and can try to help you make better incisions.
Enjoy your weekend!
--js
