Re: [BUG] libahci returns stale result tf much of the time.

[Jeff Garzik <jgarzik@pobox.com>]

On 10/04/2010 03:35 PM, Mark Lord wrote:
> On 10-10-04 03:27 PM, Jeff Garzik wrote:
>> On 10/04/2010 02:50 PM, Mark Lord wrote:
>>> I've dug a bit deeper, and worked around the issue.
>>> It seems that libata is not always happy about DATA-xfer commands
>>> that also have the check-sense bit set (0x20 in cdb[2]).
>>> Perhaps that's not even valid in SCSI ??
>>
>>
>> Very interesting... poking at this now on pre- and post-libahci
>> platforms. As you said, don't see much difference in behavior WRT
>> libahci -- which I expected/hoped, since that split shouldn't have
>> changed behavior at all[1].
>
>
> Yeah. Non-data commands still get a (correct) updated result_tf from AHCI,
> but data commands don't get one, unless they fail.
>
> Weird, but I've worked around it now.

Might be that we just get a single-bit "OK" notification from hardware 
for successfully completed commands, a la NCQ's SDB FIS.

Which opcodes are you using?  I see set-acoustic-mgmt in one email...

	Jeff