Re: [PATCH 1/1] libata: Fix ATA request sense

[Tejun Heo <tj@kernel.org>]

On Mon, Dec 19, 2016 at 10:17:40AM +0900, Damien Le Moal wrote:
> For an ATA device supporting the sense data reporting feature set,
> a failed command will trigger the execution of ata_eh_request_sense if
> the result task file of the failed command has the ATA_SENSE bit set
> (sense data available bit). ata_eh_request_sense executes the
> REQUEST SENSE DATA EXT command to retrieve the sense data of the failed
> command. On success of REQUEST SENSE DATA EXT, the ATA_SENSE bit will
> NOT be set (the command succeeded) but ata_eh_request_sense
> nevertheless tests the availability of sense data by testing that bit
> presence in the result tf of the REQUEST SENSE DATA EXT command.
> This leads to a falsy assume that request sense data failed and to the
> warning message:
> 
> atax.xx: request sense failed stat 50 emask 0
> 
> Upon success of REQUEST SENSE DATA EXT, set the ATA_SENSE bit in the
> result task file command so that sense data can be returned by
> ata_eh_request_sense.
> 
> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>

Applied to libata/for-4.10-fixes with patch description updated as
suggested by Sergei.

Thanks.

-- 
tejun