On 04/24/2018 11:09 AM, Steffen Maier wrote:
On 11/04/2016 05:35 PM, Martin K. Petersen wrote:
"Hannes" == Hannes Reinecke <hare@xxxxxxx> writes:
Hannes> Checking with SAT-3 (section 6.2.4: Commands the SATL queues
Hannes> internally) the implemented behaviour is standards conformant,
Hannes> although the standard also allows for returning 'TASK SET FULL'
Hannes> or 'BUSY' in these cases. Doing so would nicely solve this
Hannes> issue.
I agree with Hannes that it would be appropriate for the SATL to report
busy when it makes an non-queued command queueable.
Wouldn't this potentially still cause problems if the secure erase takes
longer than max_retries * scmd_tmo. I.e. the command timing out by
default after 180 seconds as in
https://www.spinics.net/lists/linux-block/msg24837.html ?
The fix approach here seems to also handle this gracefully.
Well, yes, of course the command will be terminated after it timed out.
But typically secure erase is invoked from userspace via sg ioctls, and
it's in the responsibility of the application to set the correct timeout.
Cheers,
Hannes