I don't know how well UFS keeps up with SCSI architecture, but SCSI
has expanded from the old 18-byte fixed format sense data to a
variable-length descriptor format sense data, in which an 8 byte
header precedes a variable number of descriptors. This was devised
to support 8-byte LBAs in the INFORMATION field and
COMMAND-SPECIFIC INFORMATION field for >2 TiB drives.
A few other new features:
1. Extended INQUIRY Data VPD page
MAXIMUM SUPPORTED SENSE DATA LENGTH field:
Report how much sense data the device server is capable of returning
2. Control Extensions mode page
MAXIMUM SENSE DATA LENGTH field:
Specify how much sense data the device server is allowed to return
3. Descriptor format sense data header
SDAT_OVFL bit:
Reports that the device server had to drop some descriptors
to fit in the specified length
4. "Direct-access block device sense data descriptor"
Devised for the use by Base feature set compliant designs.
Combined with the 8 byte header, results in 32 bytes of sense data.
Always carries an INFORMATION field, COMMAND-SPECIFIC
INFORMATION field, sense-key specific information, field
replaceable unit code, and ILI bit.
> -----Original Message-----
> From: linux-scsi-ow...@vger.kernel.org [mailto:linux-scsi-
> ow...@vger.kernel.org] On Behalf Of Seungwon Jeon
> Sent: Tuesday, 30 July, 2013 8:03 AM
> To: 'Santosh Y'
> Cc: linux-scsi@vger.kernel.org; 'Vinayak Holikatti'; 'James E.J. Bottomley'
> Subject: RE: [PATCH 2/7] scsi: ufs: find out sense data over scsi status
> values
>
> On Tue, July 30, 2013, Santosh Y wrote:
> > On Fri, Jul 26, 2013 at 7:16 PM, Seungwon Jeon <tgih....@samsung.com>
> wrote:
> > > Except for 'GOOD' and 'CHECK CONDITION', other status value
> > > in Response UPIU may or may contain sense data. If a non-zero
> > > value is in the Data Segment Length field, it means that UPIU
> > > has Sense Data in the Data Segment area.
> > >
> > > Signed-off-by: Seungwon Jeon <tgih....@samsung.com>
> > > ---
> > > drivers/scsi/ufs/ufs.h | 1 +
> > > drivers/scsi/ufs/ufshcd.c | 27 +++++++++++++++++++--------
> > > 2 files changed, 20 insertions(+), 8 deletions(-)
> > >
> > > diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
> > > index 139bc06..737c31b 100644
> > > --- a/drivers/scsi/ufs/ufs.h
> > > +++ b/drivers/scsi/ufs/ufs.h
> > > @@ -114,6 +114,7 @@ enum {
> > > MASK_SCSI_STATUS = 0xFF,
> > > MASK_TASK_RESPONSE = 0xFF00,
> > > MASK_RSP_UPIU_RESULT = 0xFFFF,
> > > + MASK_RSP_UPIU_DATA_SEG_LEN = 0xFFFF,
> > > };
> > >
> > > /* Task management service response */
> > > diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> > > index 4cf3a2d..688ae0e 100644
> > > --- a/drivers/scsi/ufs/ufshcd.c
> > > +++ b/drivers/scsi/ufs/ufshcd.c
> > > @@ -218,6 +218,20 @@ ufshcd_get_rsp_upiu_result(struct utp_upiu_rsp
> *ucd_rsp_ptr)
> > > return be32_to_cpu(ucd_rsp_ptr->header.dword_1) &
> MASK_RSP_UPIU_RESULT;
> > > }
> > >
> > > +/*
> > > + * ufshcd_get_rsp_upiu_data_seg_len - Get the data segment length
> > > + * from response UPIU
> > > + * @ucd_rsp_ptr: pointer to response UPIU
> > > + *
> > > + * Return the data segment length.
> > > + */
> > > +static inline int
> > > +ufshcd_get_rsp_upiu_data_seg_len(struct utp_upiu_rsp *ucd_rsp_ptr)
> > > +{
> > > + return be32_to_cpu(ucd_rsp_ptr->header.dword_2) &
> > > + MASK_RSP_UPIU_DATA_SEG_LEN;
> > > +}
> > > +
> > > /**
> > > * ufshcd_config_int_aggr - Configure interrupt aggregation values.
> > > * Currently there is no use case where we want to configure
> > > @@ -298,7 +312,8 @@ void ufshcd_send_command(struct ufs_hba
> *hba, unsigned int task_tag)
> > > static inline void ufshcd_copy_sense_data(struct ufshcd_lrb *lrbp)
> > > {
> > > int len;
> > > - if (lrbp->sense_buffer) {
> > > + if (lrbp->sense_buffer &&
> > > + !ufshcd_get_rsp_upiu_data_seg_len(lrbp->ucd_rsp_ptr)) {
> > > len = be16_to_cpu(lrbp->ucd_rsp_ptr->sense_data_len);
> > > memcpy(lrbp->sense_buffer,
> > > lrbp->ucd_rsp_ptr->sense_data,
> >
> > Also please change the SCSI_SENSE_BUFFERSIZE to 18-bytes as per the
> > spec, *in case* if wrong 'data segment length' is returned, there
> > won't be a need to memcpy extra bytes ;-).
> >
> > #define UFS_SENSE_BUFFERSIZE 18
> > memcpy(lrbp->sense_buffer,
> > lrbp->ucd_rsp_ptr->sense_data,
> > min_t(int, len, UFS_SENSE_BUFFERSIZE));
> Ok. I'll check.
> But considering the extension with spec. revision, it's not bad to be kept.
> Current: [#define SCSI_SENSE_BUFFERSIZE 96]
>
> Thanks,
> Seungwon Jeon
>
> >
> >
> > > @@ -1156,21 +1171,17 @@ ufshcd_scsi_cmd_status(struct ufshcd_lrb
> *lrbp, int scsi_status)
> > > SAM_STAT_CHECK_CONDITION;
> > > ufshcd_copy_sense_data(lrbp);
> > > break;
> > > - case SAM_STAT_BUSY:
> > > - result |= SAM_STAT_BUSY;
> > > - break;
> > > case SAM_STAT_TASK_SET_FULL:
> > > -
> > > /*
> > > * If a LUN reports SAM_STAT_TASK_SET_FULL, then the LUN
> queue
> > > * depth needs to be adjusted to the exact number of
> > > * outstanding commands the LUN can handle at any given
> > > time.
> > > */
> > > ufshcd_adjust_lun_qdepth(lrbp->cmd);
> > > - result |= SAM_STAT_TASK_SET_FULL;
> > > - break;
> > > + case SAM_STAT_BUSY:
> > > case SAM_STAT_TASK_ABORTED:
> > > - result |= SAM_STAT_TASK_ABORTED;
> > > + result |= scsi_status;
> > > + ufshcd_copy_sense_data(lrbp);
> > > break;
> > > default:
> > > result |= DID_ERROR << 16;
> > > --
> > > 1.7.0.4
> > >
> > >
> >
> >
> >
> > --
> > ~Santosh
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> > the body of a message to majord...@vger.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majord...@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
