Re: [smartmontools-support] SMART to USB/sata dock in 2.6.27-rc5 doesn't work

[Mark Lord <liml@rtr.ca>]

Mark Lord wrote:
> matthieu castet wrote:
>> Mark Lord wrote:
>>> matthieu castet wrote:
>>>> Hi,
>>>>
>>>> Tejun Heo wrote:
>>>>> David Greaves wrote:
>>>>>> Error Counter logging not supported
>>>>>> Device does not support Self Test logging
>>>>>
>>>>> The USB bridge should support SAT (SCSI-ATA translation) and many
>>>>> don't.  Nothing much can be done about it.  The hardware just don't 
>>>>> know
>>>>> how to do it.
>>>>>
>>>> That true that the hardware should support SAT, but even if it 
>>>> support it, linux need some patches to support SAT over USB. See 
>>>> http://thread.gmane.org/gmane.linux.scsi/43688 and the patch 
>>>> mentioned in the thread.
>>> ...
>>>
>>> Ahh.. Great!  This mostly fixes smartctl operation on at least two
>>> of my USB drive enclosures here:
>>>
>>>    1. a recent rev.  2.5" NexStar-3 SATA enclosure.
>>>    2. a 250GB 2.5" WD "Passport" external USB drive/enclosure.
>>>
>>> With the patch, this command line works for me:
>>>
>>>    smartctl -a -dsat -T permissive /dev/sdb
>>> And hdparm just works regardless, with or without the patch.
>>>
>>>
>> It should work without permissive, permissive was need without the patch.
>>
>> You said hdparm works without the patch. I wonder how it work. IRRC it 
>> asks check condition stuff for every command. I am interested to see 
>> the hdparm --verbose output without the patch.
> ..
> 
> Mmm... actually, the "-dsat" flag seems to make smartctl mostly work
> with and without the patch.  This is on 2.6.25.6 on my notebook here.
> Some month I'll update the kernel and fix VMware and wireless again
> for a newer kernel, but for now that's what it runs.
> 
> Here's hdparm --verbose, sans patch:
> 
> hdparm --verbose -I /dev/sdb
> 
> /dev/sdb:
> outgoing cdb:  85 08 2e 00 00 00 00 00 00 00 00 00 00 40 ec 00
> SG_IO: ATA_16 status=0x2, host_status=0x0, driver_status=0x8
> SG_IO: sb[] = {72 00 00 00 00 00 00 0e}
> SG_IO: desc[] = {09 0c .. }
>      ATA_16 tf->status=0x00 tf->error=0x00
> 
> ATA device, with non-removable media
>        Model Number:       Hitachi HTS541612J9SA00
>        Serial Number:      SB2D11E4GESWAB
>        Firmware Revision:  SBDOC74P

..

Here it is again, with the entire returned sense data shown:

hdparm --verbose -I /dev/sdb

/dev/sdb:
outgoing cdb:  85 08 2e 00 00 00 00 00 00 00 00 00 00 40 ec 00
SG_IO: ATA_16 status=0x2, host_status=0x0, driver_status=0x8
SG_IO: sb[] = {72 00 00 00 00 00 00 0e}
SG_IO: desc[] = {09 0c 00 00 50 01 40 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00}
      ATA_16 tf->status=0x00 tf->error=0x00

ATA device, with non-removable media
        Model Number:       Hitachi HTS541612J9SA00
        Serial Number:      SB2D11E4GESWAB
        Firmware Revision:  SBDOC74P
...

I guess it "works" by accident, since the 00 device status
doesn't have the ATA error bit set.  But I suppose this also
means that a failed ATA command would not be noticed either..
At least not without the patch.

Cheers