Re: [PATCH] ata: libata-scsi: Set INFORMATION sense data field consistently

[Igor Pylypiv <ipylypiv@google.com>]

On Tue, Apr 01, 2025 at 04:33:55PM +0900, Damien Le Moal wrote:
> On 3/29/25 03:30, Igor Pylypiv wrote:
> > Populate the INFORMATION field with ATA LBA when sense data is obtained
> > by the ata_eh_request_sense(). Kernel already populates the INFORMATION
> 
> s/by the/with
> s/Kernel/The kernel or s/Kernel/libata-eh
> 
> > field when sense data is reported via autosense or when it is generated
> > by the ata_gen_ata_sense().
> > 
> > ATA PASS-THROUGH commands, unlike regular ATA commands, populate
> > the INFORMATION field with ATA ERROR, STATUS, DEVICE, and COUNT(7:0)
> > fields thus setting ATA LBA into the INFORMATION field is incorrect.
> 
> Could you rephrase all of this is a manner that is clearer ? That is, describe
> the problem first and then how you fix it. As-is, this is hard to understand.

Thanks, Damien. I'll rephrase it in v2.

> 
> > diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
> > index 2796c0da8257..7e93581439b2 100644
> > --- a/drivers/ata/libata-scsi.c
> > +++ b/drivers/ata/libata-scsi.c
> > @@ -216,17 +216,30 @@ void ata_scsi_set_sense(struct ata_device *dev, struct scsi_cmnd *cmd,
> >  	scsi_build_sense(cmd, d_sense, sk, asc, ascq);
> >  }
> >  
> > -void ata_scsi_set_sense_information(struct ata_device *dev,
> > -				    struct scsi_cmnd *cmd,
> > -				    const struct ata_taskfile *tf)
> > +/**
> > + *	ata_scsi_set_sense_information - Populate INFORMATION field
> > + *	@qc: ATA command
> > + *
> > + *	Populates the INFORMATION field with ATA LBA.
> > + *
> > + *	LOCKING:
> > + *	None.
> > + */
> 
> The function name is clear enough. We do not need this kdoc block to clarify it
> and this is a static function anyway.
> 
> > +static void ata_scsi_set_sense_information(struct ata_queued_cmd *qc)
> >  {
> >  	u64 information;
> >  
> > -	information = ata_tf_read_block(tf, dev);
> > +	if (!(qc->flags & ATA_QCFLAG_RTF_FILLED)) {
> > +		ata_dev_dbg(qc->dev,
> > +			    "missing result TF: can't set INFORMATION sense field\n");
> 
> Why ata_dev_dbg() ? This should be ata_dev_err(), no ? With your change, this is
> not supposed to be called without the rtf filled...

I used ata_dev_dbg() to match ata_scsi_set_passthru_sense_fields(),
ata_gen_passthru_sense(), and ata_gen_ata_sense().

Last year we settled on using the ata_dev_dbg() to avoid a potential log spam:
https://lore.kernel.org/lkml/ab9f6564-3df1-4061-93e7-32a59aacb205@kernel.org/

Thanks,
Igor
 
> > +		return;
> > +	}
> > +
> > +	information = ata_tf_read_block(&qc->result_tf, qc->dev);
> >  	if (information == U64_MAX)
> >  		return;
> >  
> > -	scsi_set_sense_information(cmd->sense_buffer,
> > +	scsi_set_sense_information(qc->scsicmd->sense_buffer,
> >  				   SCSI_SENSE_BUFFERSIZE, information);
> >  }
> >  
> > @@ -971,8 +984,7 @@ static void ata_gen_passthru_sense(struct ata_queued_cmd *qc)
> >   *	ata_gen_ata_sense - generate a SCSI fixed sense block
> >   *	@qc: Command that we are erroring out
> >   *
> > - *	Generate sense block for a failed ATA command @qc.  Descriptor
> > - *	format is used to accommodate LBA48 block address.
> > + *	Generate sense block for a failed ATA command @qc.
> >   *
> >   *	LOCKING:
> >   *	None.
> > @@ -982,8 +994,6 @@ static void ata_gen_ata_sense(struct ata_queued_cmd *qc)
> >  	struct ata_device *dev = qc->dev;
> >  	struct scsi_cmnd *cmd = qc->scsicmd;
> >  	struct ata_taskfile *tf = &qc->result_tf;
> > -	unsigned char *sb = cmd->sense_buffer;
> > -	u64 block;
> >  	u8 sense_key, asc, ascq;
> >  
> >  	if (ata_dev_disabled(dev)) {
> > @@ -1014,12 +1024,6 @@ static void ata_gen_ata_sense(struct ata_queued_cmd *qc)
> >  		ata_scsi_set_sense(dev, cmd, ABORTED_COMMAND, 0, 0);
> >  		return;
> >  	}
> > -
> > -	block = ata_tf_read_block(&qc->result_tf, dev);
> > -	if (block == U64_MAX)
> > -		return;
> > -
> > -	scsi_set_sense_information(sb, SCSI_SENSE_BUFFERSIZE, block);
> >  }
> >  
> >  void ata_scsi_sdev_config(struct scsi_device *sdev)
> > @@ -1679,8 +1683,10 @@ static void ata_scsi_qc_complete(struct ata_queued_cmd *qc)
> >  		ata_scsi_set_passthru_sense_fields(qc);
> >  		if (is_ck_cond_request)
> >  			set_status_byte(qc->scsicmd, SAM_STAT_CHECK_CONDITION);
> > -	} else if (is_error && !have_sense) {
> > -		ata_gen_ata_sense(qc);
> > +	} else if (is_error) {
> > +		if (!have_sense)
> > +			ata_gen_ata_sense(qc);
> > +		ata_scsi_set_sense_information(qc);
> >  	}
> >  
> >  	ata_qc_done(qc);
> > diff --git a/drivers/ata/libata.h b/drivers/ata/libata.h
> > index 0337be4faec7..ce5c628fa6fd 100644
> > --- a/drivers/ata/libata.h
> > +++ b/drivers/ata/libata.h
> > @@ -141,9 +141,6 @@ extern int ata_scsi_offline_dev(struct ata_device *dev);
> >  extern bool ata_scsi_sense_is_valid(u8 sk, u8 asc, u8 ascq);
> >  extern void ata_scsi_set_sense(struct ata_device *dev,
> >  			       struct scsi_cmnd *cmd, u8 sk, u8 asc, u8 ascq);
> > -extern void ata_scsi_set_sense_information(struct ata_device *dev,
> > -					   struct scsi_cmnd *cmd,
> > -					   const struct ata_taskfile *tf);
> >  extern void ata_scsi_media_change_notify(struct ata_device *dev);
> >  extern void ata_scsi_hotplug(struct work_struct *work);
> >  extern void ata_scsi_dev_rescan(struct work_struct *work);
> 
> 
> -- 
> Damien Le Moal
> Western Digital Research