Re: [RESOLVED] Re: libata errors with smartctl on 2.6.24, WD3200AAKS and nVidia AHCI

[Tejun Heo <htejun@gmail.com>]

Jim Paris wrote:
> Tejun Heo wrote:
>> Tvrtko A. Ursulin wrote:
>>> On Wednesday 14 May 2008 08:58:56 Erik Inge Bolsø wrote:
>>>> The drive does say
>>>>
>>>>         Enabled Supported:
>>>>                 SMART feature set
>>>>                 Security Mode feature set
>>>>            *    Power Management feature set
>>>>            *    Write cache
>>>>            *    Look-ahead
>>>>
>>>> ... so doesn't look like SMART is enabled on the drive.
>>>>
>>>> Does the problem go away if you do a
>>>>
>>>> smartctl -s on /dev/sda
>>>>
>>>> first? :)
>>> Indeed it does, shame on me for overseeing that small detail. :) 
>> Oh... shame on me too.  :-)
> 
> Why didn't smartctl notice that?  In my experience it usually does
> report when SMART is disabled.  Is the error not being properly
> reported back to userspace maybe?  Tvrtko's original problem report
> has lots of uninitialized (or leaked?) data, and a lot of it looks
> like ASCII (for example his bogus RAW_VALUEs contain pieces of the
> model number).

Let's ask Bruce.  Bruce, SMART was disabled on WD3200AAKS but smartctl 
-a didn't detect it and issued SMART commands anyway triggering 
timeouts.  Is this expected behavior?  Or is libata feeding wrong status 
to smartctl?

Thanks.

-- 
tejun