Seagate Ultra160 Product Manual page 223

SCSI Interface Product Manual, Rev. B
from the medium). The valid data is placed in the reallocated block. This function doesn't apply to the
FORMAT UNIT command. When set to zero, the drive shall not perform automatic reallocation but shall
create CHECK CONDITION status with sense key of Medium Error instead.
[4] The Automatic Read Reallocation of defective data blocks Enabled (ARRE) bit, when set to one, allows
the drive to automatically relocate bad blocks detected during read operations. Automatic reallocation is
performed only if the drive successfully recovers the data and is able to place it in the reallocated block.
When set to zero, the drive shall not perform automatic reallocation but shall create CHECK CONDITION
status with sense key of Medium Error instead.
[5] The Transfer Block (TB) bit, when set to one, indicates the data block that is not recovered shall be trans-
ferred to the initiator. When set to zero, the failing data block shall not be transferred.
[6] The Read Continuous (RC) bit, when set to one, requests the drive to transfer the requested data length
without adding delays (for retries or ECC correction) that may be required to ensure data integrity. The
drive may send erroneous data in order to maintain the continuous flow of data. This bit shall override the
DTE bit if it is set. RC bit has priority also over ARRE, AWRE, EER, DCR, and PER bits. If the RC bit is set
to one, no Auto Reallocation will be attempted. When set to zero, recovery actions during data transfer are
allowed. This bit is set to zero and is not changeable in most if not all of the drive models covered by this
manual. See individual drive's Product Manual, Volume 1, section showing changeable values.
[7] The Enable Early Recovery (EER) bit, when set to one, allows the drive to apply maximum T level ECC
correction on the fly, before attempting other retry mechanisms. SEEK error retries and message system
errors are not affected by this bit. When this bit is set, the DCR bit must be zero. When the EER bit is set
to zero, the drive shall apply ECC correction before other retry mechanisms, but shall perform only normal
T level ECC corrections on the fly (Normal T level is product-specific, but is typically max T–1). In either
case, any successful correction on the fly will not be reported to the host as an error.
[8] The Post Error (PER) bit, when set to one, indicates the drive reports CHECK CONDITION status and
appropriate sense key for any recovered errors encountered. Reporting of unrecoverable errors has prior-
ity over reporting of recoverable errors. When set to zero, any errors recovered within the limits estab-
lished by the other Error Recovery Flags are not reported. Any unrecoverable errors are reported.
[9] The Disable Transfer on Error (DTE) bit is valid only when the PER bit is set to one. When the DTE bit is
set to one, it indicates the drive terminates data transfer even for recoverable errors (the drive will transfer
the data for the recovered error before terminating the transfer). When DTE is set to zero, data transfer
continues if recoverable errors are encountered. If the PER bit is one and the DTE bit is zero, recoverable
errors are reported after all data has been transferred.
[10] The Disable Correction (DCR) bit, when set to one, indicates that only on-the-fly ECC correction shall be
applied. The exact behavior is product specific. When set to zero, maximum ECC correction shall be
applied if correction is possible.
[11] The READ RETRY COUNT sets up the maximum amount of error recovery effort to be applied for each
LBA that could not be recovered during a read operation. The hex value in this field specifies the maxi-
mum error recovery level that the drive applies during a read operation to the recovery of an LBA needing
recovery effort. Each level may consist of multiple error recovery steps. See individual drive's Product
Manual, Volume 1, for more details on the levels of error recovery available.
[12] The CORRECTION SPAN is the size of the largest read data error, in bits, on which ECC correction is to
be attempted. Errors longer than this span are reported as unrecoverable. This value is drive dependent.
If this field is zero, the drive uses its default value.
[13] The drive HEAD OFFSET COUNT is a default of zero and not changeable to signify that this feature is not
programmable by the initiator. Head offsets are performed as part of the drive's retry algorithms.
[14] The drive DATA STROBE OFFSET COUNT is a default of zero and not changeable to signify that this fea-
ture is not programmable by the initiator. Data strobe offsets are performed as part of the drive's retry
algorithms.
[15] The WRITE RETRY COUNT sets up the maximum amount of error recovery to be applied for each LBA
that could not be recovered during a write operation. The hex value in this field specifies the maximum
error recovery level that the drive applies during a write operation to the recovery of an LBA needing
recovery effort. Each level may consist of multiple error recovery steps. See individual drive's Product
Manual, Volume 1, for more details on the levels of error recovery available.
205