Club Amiga de Montreal - CAM/CAM_CD_1.iso/files/571.lha/SCSI_II_Documentation/SCSI2.LZH/SCSIDOC.009

Text File | 1991-08-11 | 154KB | 3,600 lines

Sequential-Access Devices                                               3/9/90

9. Sequential-Access Devices


9.1. Sequential-Access Device Model


9.1.1. Physical Elements

Sequential-access devices (called devices below) optimize their use in storing
or retrieving user data in a sequential manner.  Since access is sequential,
position changes typically take a long time, when compared to direct-access
devices.

Sequential-access devices are usually tape devices.  The remainder of this
description is from the point of view of a tape device; however, other
implementations are not precluded.

The recording medium for tape devices consists of various widths and lengths
of a flexible substrate coated with a semi-permanent magnetic material.  The
recording medium may be wound onto single reels or encapsulated into
cartridges containing both a supply reel and a take-up reel.  Several American
National Standards exist covering the construction of reels and cartridges for
interchange as well as recording techniques for many of the format or density
combinations.

A complete unit composed of the recording medium and its physical carrier
(e.g., reel, cartridge, cassette) is called a volume.  Volumes have an
attribute of being mounted or demounted on a suitable transport mechanism.

Mounted is the state of a volume when the device is physically capable of
executing commands that cause the the medium to be moved.  A volume is
demounted when it is being loaded, threaded, unloaded, unthreaded, or when not
attached to the device.

Ready is the state of the device when medium access and non-medium access
commands can be executed.  The device is not ready when no volume is mounted
or, from the initiator's perspective, whenever all medium access commands
report CHECK CONDITION status and a NOT READY sense key.  Some devices may
have a separate switch function which places the device in a not ready state
even when a volume is mounted.

The write enabled or write protected state determines when an initiator may
write information on a volume.  This attribute is usually controlled by the
user of the volume through manual intervention (e.g., thumbwheel switch).

The recording medium has two physical attributes called beginning-of-medium





SCSI-2 draft proposed American National Standard        9-1      Revision 10c1







Sequential-Access Devices                                               3/9/90

(BOM) and end-of-medium (EOM).  Beginning-of-medium is at the end of the
medium that is attached to the take-up reel.  End-of-medium is at the end of
the medium that is attached to the supply reel.  In some cases, the medium is
permanently affixed to one or both of the reel hubs.

As shown in Figure 9-1, the entire physical length of medium is not usable for
recording data. For most volumes, a length of the medium is reserved before
the beginning-of-medium and after the end-of-medium position. This is done to
provide sufficient tape wraps onto the reel hub(s) and to ensure that
recording starts in an undamaged section of the medium.

       BOM                                                          EOM
==============================================================================
          ----------------------------------------------------------
 Reserved ----------Usable------------------------------------------ Reserved
  Medium  --------------------------Recording-----------------------  Medium
   Area   --------------------------------------------Zone----------   Area
          ----------------------------------------------------------
==============================================================================
<--- Take-up Hub                                               Supply Hub --->

                      Figure 9-1: Typical Volume Layout


9.1.2. Data Storage Characteristics

The position on the medium where a pattern of recorded signal may be written
by one write component is called a track (Figure 9-2).  A device may write or
read from one or more tracks at a time, depending on the format.

       BOM                                                          EOM
==============================================================================
Track 1   ----------------------------------------------------------
Track 2   ----------------------------------------------------------
  :::     ----------------------------------------------------------
Track n-1 ----------------------------------------------------------
Track n   ----------------------------------------------------------
==============================================================================

                   Figure 9-2: Typical Medium Track Layout


On a new volume, recording of one or more tracks begins after mounting the
volume and moving from beginning-of-medium toward end-of-medium. The number of
tracks written at one time is called a track group (TrkGrp).  For recorded
volumes, reading in the forward direction follows the same course of tracks as
when writing.

If not all tracks are recorded at the same time, and the device reverses


SCSI-2 draft proposed American National Standard        9-2      Revision 10c2







Sequential-Access Devices                                               3/9/90

direction when approaching end-of-medium and begins writing on remaining
tracks, the recording method is called serpentine. For serpentine devices that
record only one track at a time, each physical track represents one track
group (Figure 9-3).

        BOM                                                         EOM
==============================================================================
 Track 1   --------------------------------------------------------> TrkGrp 1
 Track 2   <-------------------------------------------------------- TrkGrp 2
   :::     -------------------------------------------------------->   :::
   :::     <--------------------------------------------------------   :::
 Track n   --------------------------------------------------------- TrkGrp n
==============================================================================

                   Figure 9-3: Serpentine Recording Example


Some multi-track devices have only one track group, using a parallel storage
format that supports the simultaneous recording of all available tracks
(Figure 9-4).

        BOM                                                         EOM
==============================================================================
 Track 1   -------------------------------------------------------->\
 Track 2   --------------------------------------------------------> \
   :::     -------------------------------------------------------->  >TrkGrp1
   :::     --------------------------------------------------------> /
 Track n   -------------------------------------------------------->/
==============================================================================

                    Figure 9-4: Parallel Recording Example


The serpentine and parallel recording formats shown in the previous examples
define tracks as longitudinal patterns of recorded information.  One other
storage format used by some devices records tracks diagonally across the
medium.  This recording technique is known as helical scan (Figure 9-5).


==============================================================================
        BOM    / / / / / / / / / / / / / / / / / / / / / / / / / /  EOM
              / / / / / / / / / / / / / / / / / / / / / / / / / /
             / / / / / / / / / / / / / / / / / / / / / / / / / /
            / / / / / / / / / / / / / / / / / / / / / / / / / /
Track 1 -->/ / / / / / / / / / / / / / / / / / / / / / / / / /<-- Track n
==============================================================================

                  Figure 9-5: Helical Scan Recording Example



SCSI-2 draft proposed American National Standard        9-3      Revision 10c3







Sequential-Access Devices                                               3/9/90

For most recording formats, an area at beginning-of-medium contains a format
identification in the form of a tone burst or some other recognizable pattern.
User data is not recorded in this area.  The format identification is an
attribute of a volume used for interchange purposes and is defined in
applicable standards.

When writing, the initiator needs an indication that it is approaching the end
of the permissible recording area.  This position, called early-warning (EW),
is typically reported to the initiator at a position early enough for the
device to write any buffered data to the medium while still leaving enough
room for additional recorded labels or filemarks.  Some American National
Standards include physical requirements for a marker placed on the medium to
be detected by the device as early-warning (Figure 9-6).

                  BOM                           EW       EOM
                  -------------------------------->-------->


                      Figure 9-6: Early-Warning Example


For devices which implement large data buffers, the early-warning position
defined by a physical marker may be too close to the end of the recording
region to permit emptying the data buffer(s).  For these devices, a logical
concept of early-warning is often used to signal the initiator at an
appropriate location prior to the physical marker.


9.1.3. Partitions within a Volume

Another attribute of a volume is called a partition.  Partitions consist of
one or more non-overlapped mini-volumes, each with its own beginning and
ending points, occupying a single physical volume.  Each partition (x) within
a volume has a defined beginning-of-partition (BOP x), an early-warning
position (EW x), and an end-of-partition (EOP x).

All volumes have a minimum of one partition called partition zero, the default
data partition.  For devices which support only one partition, the
beginning-of-partition zero (BOP 0) may be equivalent to the beginning-of-
medium and the end-of-partition zero (EOP 0) may be equivalent to the end-of-
medium.

When a volume is mounted, it is logically positioned to beginning of the
default data partition (BOP 0).  When a REWIND command is received in any
partition (x), the device positions to the beginning-of-partition (BOP x).

 Partitions on a volume do not need to be recorded in any defined order, nor




SCSI-2 draft proposed American National Standard        9-4      Revision 10c4







Sequential-Access Devices                                               3/9/90

do all partition numbers in a sequence need to be present on a volume.  It is
sufficient for a device to be able to locate a partition, given its code
value, or determine that it does not exist on the volume.  For interchange,
information about which partitions are present on a volume may be stored on
the volume in a device-defined area (possibly unavailable to the initiator) or
the information may be an intrinsic attribute of the device implementation.

Figure 9-7 shows a possible partition implementation for a four-track
serpentine recording device, assuming that each track group defines a
partition.

         BOM                                                            EOM
==============================================================================
TrkGrp1  BOP0 ----------------------------------------------->EW0-----> EOP0
TrkGrp2  EOP1 <-----EW1<----------------------------------------------- BOP1
TrkGrp3  BOP2 ----------------------------------------------->EW2-----> EOP2
TrkGrp4  EOP3 <-----EW3<----------------------------------------------- BOP3
==============================================================================

       Figure 9-7: Partitioning Example - One Partition per Track Group


Another possible partition implementation for this four-track serpentine
recording device is shown in Figure 9-8, using two track groups to define each
partition.

         BOM                                                            EOM
==============================================================================
TrkGrp1  BOP0 -------------------------------------------------------->
TrkGrp2  EOP0 <-----EW0<-----------------------------------------------
TrkGrp3  BOP1 -------------------------------------------------------->
TrkGrp4  EOP1 <-----EW1<-----------------------------------------------
==============================================================================

    Figure 9-8: Partitioning Example - One Partition per Two Track Groups


The previous examples show the beginning and ending points for a partition
aligned with physical bounds of the medium.  This is not a mandatory
requirement for partitioning; it is sufficient for a device to be able to
locate to and stay in any partition bounded by a BOP x and EOP x.  In this
case, a recorded mark or some other device-recognizable attribute could be
used to delineate the partitions.  Figure 9-9 shows a possible two-partition
implementation for a device with only one track group.

         BOM                                                            EOM





SCSI-2 draft proposed American National Standard        9-5      Revision 10c5







Sequential-Access Devices                                               3/9/90

==============================================================================
              --------->------->----------------------------->-------->
              --------->------->----------------------------->-------->
TrkGrp1  BOP0 --------->EW0---->EOP0/BOP1-------------------->EW1-----> EOP1
              --------->------->----------------------------->-------->
              --------->------->----------------------------->-------->
==============================================================================

      Figure 9-9: Partitioning Example - Two Partitions per Track Group


Three methods are defined in the MODE SENSE and MODE SELECT commands for
managing partitions; each method is progressively more difficult to implement
and manage:

(1) device-defined fixed locations;
(2) device-defined based on an initiator supplied number of partitions and a
device specific allocation algorithm; and,
(3) definition by name and capacity by an initiator.


9.1.4. Logical Elements within a Partition

The area between BOP x and EOP x on a typical recorded volume contains at
least two types of initiator accessible elements, data blocks and tape marks.
These elements are controlled and transferred between the initiator and the
medium using READ, READ REVERSE, WRITE, and WRITE FILEMARKS commands.

A unit of data supplied or requested by an initiator is called a logical
block.  Logical blocks are stored according to the specifications of the
format for the volume and may be recorded as one or more physical blocks on
the medium.  When the physical block and the logical block are not recorded in
a one-to-one relationship, it is the responsibility of the device to perform
all blocking, de-blocking, padding, stripping, splitting or rebuilding of the
logical data block(s) sent by an initiator.

Filemarks are special recorded elements containing no user data.  The filemark
format is defined in some American National Standards.  Initiators
traditionally use filemarks to separate user data from labels and logical
groupings of data from each other.  Since some format standards do not define
an explicit end-of-data (EOD), host software has often used conventions with
filemarks to represent an EOD indication.  At least one American National
Standard specifically defines filemark use for this purpose.  In some
implementations, the device's EOD definition may be specified the initiator
using the MODE SELECT command.

A setmark is another type of special recorded element containing no user data,




SCSI-2 draft proposed American National Standard        9-6      Revision 10c6







Sequential-Access Devices                                               3/9/90

providing a segmentation scheme hierarchically superior to filemarks. This
level of segmentation is useful for some high capacity storage devices to
provide concise addressing and fast positioning to specific sets of data
within a partition.  In some implementations, the detection and reporting of
setmarks may be controlled by the initiator using the MODE SELECT command.

Inter-block gaps, the gaps between blocks, filemarks, and setmarks, are
introduced on the medium at the time a block or mark is written without
explicit action by the initiator.  Minimum and maximum lengths for inter-block
gaps are defined in some American National Standards.  In some devices, the
length of inter-block gaps may be selected by the initiator using the MODE
SELECT command while in other devices the gaps are fixed and non-changeable.

In addition to blocks, filemarks, and setmarks, erase gaps may be recorded on
the medium through use of the ERASE command or device-initiated error recovery
actions.  Although explicitly recorded on the medium, there is normally no
distinction between two contiguous erase gaps.  An erase gap may be a length
of erased medium or a recorded pattern not distinguishable as a block or mark.
Minimum and maximum lengths for erase gaps are defined in some American
National Standards while some devices may have no implementation of an erase
gap.

After writing data from BOP x, the medium is considered to be a contiguous
grouping of blocks, filemarks, setmarks, and gaps.  Certain American National
Standards define gap lengths which, if exceeded, are to be considered as
having reached blank medium.  Depending on the format, this blank medium may
be treated as an end-of-data indication, an error recovery area, or an
unrecoverable medium error causing an interchange error.  Unrecorded volumes
(new or erased) may exhibit blank medium characteristics if an attempt is made
to read or space the volume before data has been written.

A sequential-access device may be capable of supporting fixed or variable
length blocks.  The concept of fixed or variable mode for writing and reading
blocks only indicates the method by which the initiator specifies the size of
a logical block for transfer and not the method of recording physical blocks
on the medium.  However, a device that supports only fixed-length physical
blocks may only be capable of supporting logical blocks of the same length.
The length of a logical block is always described in bytes.  The length of a
physical block may or may not be recorded as an exact byte count, depending on
the format.


9.1.5. Data Buffering








SCSI-2 draft proposed American National Standard        9-7      Revision 10c7







Sequential-Access Devices                                               3/9/90


A device may contain a temporary storage area capable of holding one or more
logical blocks - a data buffer.  A device data buffer may include any
combination of blocks, filemarks, and setmarks in the process of being written
to the medium, or it may contain read-ahead data blocks transferred from the
medium.

A device with a data buffer may be capable of operating in either a buffered
mode or an unbuffered mode.  A device with no data buffer operates only in
unbuffered mode.  Either term is only applicable to the manner in which the
device manages information to be written to the medium. Buffered mode is not
applicable during read commands, regardless of whether read data p asses
through a data buffer.

A device operating in buffered mode may return GOOD status for write
operations when all write data has been successfully transferred from the
initiator into the device data buffer.  For devices operating in unbuffered
mode, GOOD status is not returned until all requested data, filemarks, or
setmarks are successfully recorded on the medium.

When issuing a buffered WRITE FILEMARKS command with the immediate bit set to
one, GOOD status is returned as soon as the command is validated.  A WRITE
FILEMARKS command with the immediate bit set to zero causes any buffered
blocks, filemarks, and setmarks to be written to the medium.  Upon successful
completion of this process, which is called a synchronize operation, no
blocks, filemarks, or setmarks remain in the data buffer which have not been
written to the medium.  A synchronize operation has no effect on a data buffer
which contains only read-ahead data or write data which has already been
successfully written to the medium.

Should an unrecoverable write error occur while in buffered mode, the device
generates an error condition to the current active command.  If no command is
active, the error may be reported on the next applicable operation as a
deferred error (see 7.2.14.2).  For some implementations, asynchronous event
notification (AEN) or extended contingent allegiance (ECA) may be required.
Refer to 6.5.5 and 6.7 for descriptions of AEN and ECA protocol.

The READ POSITION command may be used to determine the number and storage
space of buffered blocks not written before the unrecoverable error was
encountered.

 A device with read-ahead data blocks in the data buffer does not report an
unrecovered read error until the data block in error is requested by an
initiator.


9.1.6. Recorded Object Descriptors (Block Identifiers)




SCSI-2 draft proposed American National Standard        9-8      Revision 10c8







Sequential-Access Devices                                               3/9/90


Some recording formats specify that recorded objects (blocks, filemarks, and
setmarks) have identifiers included in the recorded information to help
determine write sequence and also to help detect device positioning errors.
The identifier values are unique within a partition and may be unique within a
volume.

The use of the term block identifier may imply some arithmetic sequence
applied to the assignment of recorded objects.  The block identifier
assignment algorithm may be defined in an applicable format standard.

For some pre-formatted volumes, the identifiers are associated with physical
blocks.  In variable-length implementations, the identifier can be associated
with a physical block when the logical block and the physical block have a
one-to-one relationship on the medium.

Some formats may carry both physical and logical block identifiers recorded on
the medium.  When a logical block is split over more than one physical block,
or multiple logical blocks are concatenated to form a physical block, the
logical block identifier and the physical block identifier are not the same.
Filemarks and setmarks may or may not have recorded identifiers, but if
identifiers are used in the format, then each mark is assigned a value even if
it is not explicitly recorded.

The READ POSITION and LOCATE commands use four-byte fields to hold these
format dependent identifiers.  For some implementations, this value may
correspond to a real physical location; however, it is sufficient for the
device to map the identifier to a value representing the unique recorded
object.  With this capability, the READ POSITION command may be used to report
a device-defined block identifier and the initiator may use this value with a
LOCATE command to position to the same location at some future time (provided
the volume has not been rewritten in the interim).


9.1.7. Direction and Position Definitions

For sequential-access devices, positioning has the connotation of logically
being in, at, before, or after some defined place within a volume.  This
definition means the position is capable of being repeated under the same
circumstances.  The orientation of usage for the four words (in, at, before,
or after) is in one direction, from BOP x toward EOP x.  All positioning
defined below is worded from this perspective.  Devices without buffers have
some physical position which relates to these logical positions.  However,
these definitions do not require the medium to have a physical position
equivalent to the logical position unless explicitly stated.

The forward direction is defined as logically progressing from BOP x toward




SCSI-2 draft proposed American National Standard        9-9      Revision 10c9







Sequential-Access Devices                                               3/9/90

EOP x. The reverse direction is defined as logically progressing from EOP x
toward BOP x.  In serpentine devices, the logical forward or reverse direction
has an alternating relationship to the physical motion of the medium.

The concept of being "in" some position means not being outside a defined
region.  The definition allows the position to be on the boundary of a defined
region.  When a volume is first mounted, the logical position is always at the
beginning of the default data partition (BOP0).  Whenever a volume is mounted
and the medium motion is stopped, the position is in some partition.  While
moving between partitions, there is no stable position.

The concept of being "at" some position indicates being positioned to a
logical or physical extremity of a partition.  A sequential-access device may
be positioned at beginning-of-medium, at BOP x, at end-of-data (EOD), at EOP
x, or at end-of-medium (EOM), since these are stable positions at extremities
of a partition.

The concept of being "before" some position indicates that there is some
element (data block, filemark, setmark, or other defined point) which may be
encountered when moving toward EOP x, if the proper commands are issued.
Being positioned before a particular data block means that if the device
receives a valid READ command, the data block is transferred to the initiator.
This position may also be before EW x and EOP x, since these are defined
points within any partition.  However, if data has not been written to the
end-of-partition, these points may not be accessible by the initiator.

The concept of being "after" some position indicates that there is some
element (data block, filemark, setmark, or other defined point) on the BOP x
side of the current position which may be encountered if the proper commands
are issued.  When a READ command for a single data block has been successfully
executed, the logical position is after the transferred data block.

9.1.8. Error Reporting

If any of the following conditions occur during the execution of a command the
target shall return CHECK CONDITION status.  The appropriate sense key and
additional sense code should be set.  The following table illustrates some
error conditions and the applicable sense keys.  The table does not provide an
exhaustive enumeration of all conditions that may cause the CHECK CONDITION
status.

Condition                                              Sense Key









SCSI-2 draft proposed American National Standard        9-10      Revision 10c10







Sequential-Access Devices                                               3/9/90

----------------------------------------------------   ---------------
Invalid logical block address                          ILLEGAL REQUEST

Unsupported option requested                           ILLEGAL REQUEST

Target reset or medium change since
last command from this initiator                       UNIT ATTENTION

Self diagnostic failed                                 HARDWARE ERROR

Unrecovered read error                                 MEDIUM ERROR
                                                       HARDWARE ERROR

Recovered read or write error                          RECOVERED ERROR

Overrun or other error that might
be resolved by repeating the command                   ABORTED COMMAND

Attempt a WRITE, READ, READ REVERSE, VERIFY, or        ILLEGAL REQUEST
RECOVER BUFFERED DATA command with the fixed bit
set to zero and variable block mode is not supported

Attempt a WRITE, READ, READ REVERSE, VERIFY, or        ILLEGAL REQUEST
RECOVER BUFFERED DATA command with a fixed bit of
zero and requested block length is not supported

Attempt a WRITE, READ, READ REVERSE, VERIFY, or        ILLEGAL REQUEST
RECOVER BUFFERED DATA command with a fixed bit of
one and MODE SENSE block length set to zero

Attempt to execute an erase or write-type operation    DATA PROTECT
on write protected medium

Deferred write error                                   MEDIUM ERROR
                                                       VOLUME OVERFLOW

In the case of an unrecovered read or write error, if the read-write error
recovery page (9.3.3.4) is implemented, the current values specify the target
error recovery criteria.  If this page is not implemented, the error recovery
is vendor specific.

In the case of an unrecovered read error, if the fixed bit is one, the valid
bit shall be set to one and the information field shall be set to the
requested transfer length minus the actual number of blocks read (not
including the unrecovered block).  If the fixed bit is zero, the valid bit
shall be set to one and the information field shall be set to the requested
transfer length.  Upon termination, the logical position shall be after the
unrecovered block.

In the case of an unrecovered write error, if unbuffered mode is selected and

SCSI-2 draft proposed American National Standard        9-11      Revision 10c11







Sequential-Access Devices                                               3/9/90

the fixed bit is set to one, the valid bit shall be set to one and the
information field shall be set to the requested transfer length minus the
actual number of blocks written.  If unbuffered mode is selected and the fixed
bit is set to zero, the information field shall be set to the requested
transfer length.

In the case of an unrecovered write error or a deferred write error, if
buffered mode is selected and the fixed bit is one, the valid bit shall be set
to one and the information field shall be set to the total number of blocks,
filemarks, and setmarks not written (the number of blocks not transferred from
the initiator plus the number of blocks, filemarks, and setmarks remaining in
the target's buffer).  If buffered mode is selected and the fixed bit is zero,
the information field shall be set to the total number of bytes, filemarks,
and setmarks not written (the number of bytes not transferred from the
initiator plus the number of bytes, filemarks, and setmarks remaining in the
target's buffer).  In both cases, the value in the information field may
exceed the transfer length.

In the case of an unrecovered write error or a deferred write error, if
buffered mode 1h is selected, the error shall be reported to the first
initiator issuing a command (other than INQUIRY or REQUEST SENSE) or the first
initiator responding to asynchronous event notification.  If buffered mode 2h
is selected, the error shall be reported to the initiator with unwritten data
in the buffer.



9.2. Command Descriptions for Sequential-Access Devices























SCSI-2 draft proposed American National Standard        9-12      Revision 10c12







Sequential-Access Devices                                               3/9/90


The commands for sequential-access devices shall be as shown in Table 9-1.

              Table 9-1: Commands for Sequential-Access Devices

==============================================================================
                               Operation
Command Name                     Code         Type    Section
------------------------------------------------------------------------------
 CHANGE DEFINITION                40h          O       7.2.1
 COMPARE                          39h          O       7.2.2
 COPY                             18h          O       7.2.3
 COPY AND VERIFY                  3Ah          O       7.2.4
 ERASE                            19h          M       9.2.1
 INQUIRY                          12h          M       7.2.5
 LOAD UNLOAD                      1Bh          O       9.2.2
 LOCATE                           2Bh          O       9.2.3
 LOG SELECT                       4Ch          O       7.2.6
 LOG SENSE                        4Dh          O       7.2.7
 MODE SELECT(6)                   15h          M       7.2.8
 MODE SELECT(10)                  55h          O       7.2.9
 MODE SENSE(6)                    1Ah          M       7.2.10
 MODE SENSE(10)                   5Ah          O       7.2.11
 PREVENT ALLOW MEDIUM REMOVAL     1Eh          O       8.2.4
 READ                             08h          M       9.2.4
 READ BLOCK LIMITS                05h          M       9.2.5
 READ BUFFER                      3Ch          O       7.2.12
 READ POSITION                    34h          O       9.2.6
 READ REVERSE                     0Fh          O       9.2.7
 RECEIVE DIAGNOSTIC RESULTS       1Ch          O       7.2.13
 RECOVER BUFFERED DATA            14h          O       9.2.8
 RELEASE UNIT                     17h          M       9.2.9
 REQUEST SENSE                    03h          M       7.2.14
 RESERVE UNIT                     16h          M       9.2.10
 REWIND                           01h          M       9.2.11
 SEND DIAGNOSTIC                  1Dh          M       7.2.15
 SPACE                            11h          M       9.2.12
 TEST UNIT READY                  00h          M       7.2.16
 VERIFY                           13h          O       9.2.13
 WRITE                            0Ah          M       9.2.14
 WRITE BUFFER                     3Bh          O       7.2.17
 WRITE FILEMARKS                  10h          M       9.2.15
==============================================================================

Key: M  =  Command implementation is mandatory.
     O  =  Command implementation is optional.

The following command codes are vendor specific: 02h, 06h, 07h, 09h, 0Ch, 0Dh,
0Eh.  All other command codes for sequential-access devices are reserved for
future standardization.

SCSI-2 draft proposed American National Standard        9-13      Revision 10c13







Sequential-Access Devices                                               3/9/90


9.2.1. ERASE Command

                          Table 9-2: ERASE Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (19h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |         Reserved          | Immed |  Long  |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The ERASE command (Table 9-2) causes part or all of the medium to be erased
beginning at the current position on the logical unit.  As used here, "erased"
means either the medium shall be erased or a pattern shall be written on the
medium that appears as a gap to the target.

An immediate (Immed) bit of zero indicates that the target shall not return
status until the erase operation has completed.  An Immed bit of one indicates
that the target shall return status as soon as the command has been validated.
If CHECK CONDITION status is returned for an ERASE command with an Immed bit
of one, the erase operation shall not be performed.

A long bit of one indicates that all remaining medium in the current partition
shall be erased beginning at the current logical position.  If the Immed bit
is one, the target shall return status as soon as all buffered commands have
completed execution and the command descriptor block of the ERASE command has
been validated.  The logical position following an ERASE command with a long
bit of one is not specified by this standard.

IMPLEMENTORS NOTE:  Some targets may reject an ERASE command with the long bit
set to one if the logical unit is not at beginning-of-partition.

A long bit of zero specifies an erase gap defined by the gap size field in the






SCSI-2 draft proposed American National Standard        9-14      Revision 10c14







Sequential-Access Devices                                               3/9/90

device configuration page (9.3.3.1).  If the gap size is zero or the field is
not supported, a device defined erase gap operation shall be performed.  If
the Immed bit is one, the target shall return status as soon as the command
descriptor block has been validated.  Erase gaps may be used in host
controlled error recovery or "update in place" applications.

If the logical unit encounters early-warning during an ERASE command, and any
buffered data, filemarks, or setmarks remain to be written, the target action
shall be as defined for the early-warning condition of the WRITE command
(9.2.14).  If the long bit is zero, the erase operation shall terminate with
CHECK CONDITION status and return sense data as defined for the WRITE command.
Any buffered erases are not reported as part of the information field.

9.2.2. LOAD UNLOAD Command

                       Table 9-3: LOAD UNLOAD Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (1Bh)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |              Reserved             | Immed  |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved         |   EOT  | Re-Ten |  Load  |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The LOAD UNLOAD command (Table 9-3) requests that the target enable or disable
the logical unit for further operations.  This command may also be used to
request a re-tension function.  Prior to performing the load or unload
operation, the target shall ensure that all buffered data, filemarks, and
setmarks have been transferred to the medium.

An immediate (Immed) bit of zero indicates that the target shall not return
status until the load or unload operation has completed.  An Immed bit of one
indicates that the target shall return status as soon as all buffered commands
have completed execution and the command descriptor block of the LOAD UNLOAD
command has been validated.  If CHECK CONDITION status is returned for a LOAD
UNLOAD command with an Immed bit of one, the load or unload operation shall
not be performed.

IMPLEMENTORS NOTE:  For compatibility with devices implemented prior to this

SCSI-2 draft proposed American National Standard        9-15      Revision 10c15







Sequential-Access Devices                                               3/9/90

version of the standard, a WRITE FILEMARKS command with an Immed bit of zero
should be used to ensure that all buffered data, filemarks, or setmarks have
been transferred to the medium prior to issuing a LOAD UNLOAD command with an
Immed bit of one.

An end-of-tape (EOT) bit of one indicates that an unload operation (load bit
set to zero) shall position the medium at end-of-medium for removal from the
device.  An EOT bit of zero indicates that an unload operation shall position
the medium at beginning-of-medium for removal from the device.

An EOT bit of one and a load bit of one shall cause the target to return CHECK
CONDITION status and the sense key shall be set to ILLEGAL REQUEST in the
sense data.

A re-tension (Re-Ten) bit of one indicates that the medium on the logical unit
shall be correctly tensioned.  Implementation of the re-tension function is
device specific.

If the load bit is set to one, the medium in the logical unit shall be loaded
and positioned to the beginning-of-partition zero.  If the load bit is zero,
the medium in the logical unit shall be positioned for removal at the extreme
position along the medium specified by the EOT bit.  Following successful
completion of an unload operation, the target shall return CHECK CONDITION
status with the sense key set to NOT READY for all subsequent medium-access
commands until a new volume is mounted or a load operation is successfully
completed.

When operating in buffered mode 1h or 2h (see 9.3.3), the target shall discard
any unwritten buffered data after the LOAD UNLOAD command is validated if the
previous command was terminated with CHECK CONDITION status and the device is
unable to continue successfully writing.

9.2.3. LOCATE Command


















SCSI-2 draft proposed American National Standard        9-16      Revision 10c16







Sequential-Access Devices                                               3/9/90


                          Table 9-4: LOCATE Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (2Bh)                        |
-----|-----------------------------------------------------------------------|
 1   |    Logical Unit Number   |      Reserved   |   BT   |   CP   | Immed  |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   | (MSB)                                                                 |
-----|---                                                                 ---|
 4   |                                                                       |
-----|---                        Block Address                            ---|
 5   |                                                                       |
-----|---                                                                 ---|
 6   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 7   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 8   |                           Partition                                   |
-----|-----------------------------------------------------------------------|
 9   |                           Control                                     |
==============================================================================


LOCATE command (Table 9-4) causes the target to position the logical unit to
the specified block address in a specified partition.  Upon completion, the
logical position shall be before the specified location.  Prior to performing
the locate operation, the target shall ensure that all buffered data,
filemarks, and setmarks have been transferred to the medium.

A block address type (BT) bit of one indicates the value in the block address
field shall be interpreted as a device-specific value.  A BT bit of zero
indicates the value in the block address field shall be interpreted as an SCSI
logical block address.

A change partition (CP) bit of one indicates that a change to the partition
specified in the partition field is to occur prior to positioning to the block
specified in the block address field.  A CP bit of zero indicates no partition
change is to be made and the partition field is to be ignored.

An immediate (Immed) bit of zero indicates that the target shall not return





SCSI-2 draft proposed American National Standard        9-17      Revision 10c17







Sequential-Access Devices                                               3/9/90

status until the locate operation has completed.  An Immed bit of one
indicates that the target shall return status as soon as all buffered commands
have completed execution and the command descriptor block of the LOCATE
command has been validated.  If CHECK CONDITION status is returned for a
LOCATE command with an Immed bit of one, the locate operation shall not be
performed.

The block address field specifies the block address to which the target shall
position the medium based on the current setting of the BT bit.

The partition field specifies which partition to select if the CP bit is one.
Refer to the sequential-access device model (9.1.3) and the medium partition
pages (9.3.3.2 and 9.3.3.3) for additional information about partitioning.

9.2.4. READ Command

                           Table 9-5: READ Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (08h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number     |        Reserved         |  SILI  | Fixed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Transfer Length                             |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The READ command (Table 9-6) requests that the target transfer one or more
block(s) of data to the initiator beginning with the next block on the logical
unit.

The fixed bit specifies whether fixed-length or variable-length blocks are to
be transferred.  Refer to the READ BLOCK LIMITS command (9.2.5) for additional
information about fixed and variable block mode.

If the fixed bit is one, the transfer length specifies the number of fixed-






SCSI-2 draft proposed American National Standard        9-18      Revision 10c18







Sequential-Access Devices                                               3/9/90

length blocks to be transferred, using the current block length reported in
the mode parameters block descriptor (7.3.3).  If the fixed bit is zero, a
variable-length block is requested with the transfer length specifying the
maximum number of bytes allocated for the returned data.

A successful READ command with a fixed bit of one shall transfer the requested
transfer length times the current block length in bytes to the initiator.  A
successful READ command with a fixed bit of zero shall transfer the requested
transfer length in bytes to the initiator.  Upon completion, the logical
position shall be after the last block transferred (end-of-partition side).

If the suppress incorrect length indicator (SILI) bit is one and the fixed bit
is zero, the target shall:
(1) report CHECK CONDITION status for an incorrect length condition only if
the overlength condition exists (see 9.4) and the block length field in the
mode parameter block descriptor is nonzero (see 7.3.3)
(2) not report CHECK CONDITION status if the only error is the underlength
condition (see 9.4), or if the only error is the overlength condition and the
block length field of the mode parameters block descriptor is zero.

IMPLEMENTORS NOTE:  Since the residue information normally provided in the
information field of the sense data may not be available when the SILI bit is
set, other methods for determining the actual block length should be used
(e.g., including length information in the data block).

If the SILI bit is one and the fixed bit is one, the target shall terminate
the command with CHECK CONDITION status and the sense key shall be set to
ILLEGAL REQUEST with an additional sense code of INVALID FIELD IN CDB.

If the SILI bit is zero and an incorrect length block is read, CHECK CONDITION
status shall be returned and the ILI and valid bits shall be set to one in the
sense data.  Upon termination, the logical position shall be after the
incorrect length block (end-of-partition side).  If the fixed bit is one, the
information field shall be set to the requested transfer length minus the
actual number of blocks read (not including the incorrect length block).  If
the fixed bit is zero, the information field shall be set to the requested
transfer length minus the actual block length.  Targets that do not support
negative values shall set the information field to zero if the overlength
condition exists (see 9.4).

IMPLEMENTORS NOTE:  In the above case with the fixed bit of one, only the
position of the incorrect-length logical block can be determined from the
sense data.  The actual length of the incorrect block is not reported.  Other
means may be used to determine its actual length (e.g., read it again with the
fixed bit set to zero).

A transfer length of zero indicates that no data shall be transferred.  This




SCSI-2 draft proposed American National Standard        9-19      Revision 10c19







Sequential-Access Devices                                               3/9/90

condition shall not be considered an error and the logical position shall not
be changed.

If the logical unit encounters a filemark during a READ command, CHECK
CONDITION status shall be returned and the filemark and valid bits shall be
set to one in the sense data.  The sense key shall be set to NO SENSE or
RECOVERED ERROR, as appropriate.  Upon termination, the logical position shall
be after the filemark (end-of-partition side).  If the fixed bit is one, the
information field shall be set to the requested transfer length minus the
actual number of blocks read (not including the filemark).  If the fixed bit
is zero, the information field shall be set to the requested transfer length.

If the logical unit encounters a setmark during a READ command and the RSmk
bit is set to one in the device configuration page (9.3.3.1), CHECK CONDITION
status shall be returned and the filemark and valid bits shall be set to one
in the sense data.  The sense key shall be set to NO SENSE or RECOVERED ERROR,
as appropriate, and the additional sense code shall be set to SETMARK
DETECTED.  Upon termination, the logical position shall be after the setmark
(end-of-partition side).  If the fixed bit is one, the information field shall
be set to the requested transfer length minus the actual number of blocks read
(not including the setmark).  If the fixed bit is zero, the information field
shall be set to the requested transfer length.  The target shall not return
CHECK CONDITION when a setmark is encountered if the RSmk bit is set to zero
or if this option is not supported.

If the logical unit encounters early-warning during a READ command and the REW
bit is set to one in the device configuration page (see 9.3.3.1), CHECK
CONDITION status shall be returned upon completion of the current block.  The
sense key shall be set to NO SENSE or RECOVERED ERROR, as appropriate.  The
EOM and valid bits shall be set to one in the sense data.  Upon termination,
the logical position shall be after the last block transferred (end-of-
partition side).  If the fixed bit is one, the information field shall be set
to the requested transfer length minus the actual number of blocks read.  If
the fixed bit is zero, the information field shall be set to the requested
transfer length minus the actual block length.  The target shall not return
CHECK CONDITION status when early-warning is encountered if the REW bit is
zero or if the REW option is not supported.

IMPLEMENTORS NOTE:  A REW bit of one is not recommended for most system
applications since read data may be present after early-warning.

If the logical unit encounters end-of-data during a READ command, CHECK









SCSI-2 draft proposed American National Standard        9-20      Revision 10c20







Sequential-Access Devices                                               3/9/90

CONDITION status shall be returned, the sense key shall be set to BLANK CHECK,
and the valid bit shall be set to one in the sense data.  If end-of-data is
encountered at or after early-warning, the EOM bit shall also be set to one.
Upon termination, the logical position shall be after the last recorded
logical block (end-of-partition side).  If the fixed bit is one, the
information field shall be set to the requested transfer length minus the
actual number of blocks read.  If the fixed bit is zero, the information field
shall be set to the requested transfer length.

If the logical unit encounters end-of-partition during a READ command, CHECK
CONDITION status shall be returned, the sense key shall be set to MEDIUM
ERROR, and the EOM and valid bits shall be set to one in the sense data. The
medium position following this condition is not defined.  If the fixed bit is
one, the information field shall be set to the requested transfer length minus
the actual number of blocks read. If the fixed bit is zero, the information
field shall be set to the requested transfer length.


9.2.5. READ BLOCK LIMITS Command

                    Table 9-6: READ BLOCK LIMITS Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (05h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |                  Reserved                  |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The READ BLOCK LIMITS command (Table 9-6) requests that the logical unit's
block length limits capability be returned.  The READ BLOCK LIMITS data shown
in Table 9-7 shall be returned during the DATA IN phase of the command.

                      Table 9-7: READ BLOCK LIMITS Data






SCSI-2 draft proposed American National Standard        9-21      Revision 10c21







Sequential-Access Devices                                               3/9/90


==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 1   | (MSB)                                                                 |
-----|---                                                                 ---|
 2   |                           Maximum Block Length Limit                  |
-----|---                                                                 ---|
 3   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 4   | (MSB)                                                                 |
-----|---                        Minimum Block Length Limit               ---|
 5   |                                                                 (LSB) |
==============================================================================


If the maximum block length limit equals the minimum block length limit, the
logical unit supports the transfer of data in the fixed-block mode only, with
the block length equal to the given limit.  In this case the fixed bit shall
be set to one in the WRITE and READ commands

If the maximum block length limit is not equal to the minimum block length
limit, the logical unit supports the transfer of data in either fixed-block or
variable-block modes, with the block length constrained between the given
limits in either mode.  The transfer mode is controlled by the fixed bit in
the WRITE or READ commands. If the maximum block limit is zero a maximum block
length is not specified.

9.2.6. READ POSITION Command



















SCSI-2 draft proposed American National Standard        9-22      Revision 10c22







Sequential-Access Devices                                               3/9/90


                      Table 9-8: READ POSITION Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (34h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |             Reserved              |   BT   |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 6   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 7   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 8   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 9   |                           Control                                     |
==============================================================================


The READ POSITION command (Table 9-8) reports the current position of the
logical unit and any data blocks in the buffer.  No medium movement shall
occur as a result of the command.

A block address type (BT) bit of one requests the target to return its current
first block location and last block location as a device-specific value.  A BT
bit of zero requests the target to return the first block location and the
last block location as an SCSI logical block address (relative to a
partition).

 If the logical unit is not capable of reporting block locations, the BPU bit
shall be set to one in the returned data.  Support for this option of the READ
POSITION command is indicated by a BIS bit set to one in the device
configuration page (9.3.3.1).

The READ POSITION data shown in Table 9-9 shall be returned during the DATA IN
phase of the command.

                    Table 9-9: READ POSITION Data Format



SCSI-2 draft proposed American National Standard        9-23      Revision 10c23







Sequential-Access Devices                                               3/9/90


==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |  BOP   |   EOP  |        Reserved          |  BPU   |     Reserved    |
-----|-----------------------------------------------------------------------|
 1   |                          Partition Number                             |
-----|-----------------------------------------------------------------------|
 2   |                          Reserved                                     |
-----|-----------------------------------------------------------------------|
 3   |                          Reserved                                     |
-----|-----------------------------------------------------------------------|
 4   | (MSB)                                                                 |
- - -|- -                       First Block Location                      - -|
 7   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 8   | (MSB)                                                                 |
- - -|- -                       Last Block Location                       - -|
 11  |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 12  |                          Reserved                                     |
-----|-----------------------------------------------------------------------|
 13  | (MSB)                                                                 |
- - -|- -                       Number of Blocks in Buffer                - -|
 15  |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 16  | (MSB)                                                                 |
- - -|- -                       Number of Bytes in Buffer                 - -|
 19  |                                                                 (LSB) |
==============================================================================


A beginning of partition (BOP) bit of one indicates that the logical unit is
at the beginning-of-partition in the current partition.  A BOP bit of zero
indicates that the current logical position is not at the beginning-of-
partition.

An end of partition (EOP) bit of one indicates that the logical unit is
positioned between early-warning and end-of-partition in the current
partition.  An EOP bit of zero indicates that the current logical position is
not between early-warning and end-of-partition.

IMPLEMENTORS NOTE:  The BOP and EOP indications are not necessarily a result
of a physical tape marker (e.g., reflective marker).

A block position unknown (BPU) bit of one indicates that the first and last




SCSI-2 draft proposed American National Standard        9-24      Revision 10c24







Sequential-Access Devices                                               3/9/90

block locations are not known or cannot be obtained.  A BPU bit of zero
indicates that the first and last block location fields contain valid position
information.

The partition number field reports the partition number for the current
logical position.  If the logical unit only supports one partition for the
medium, this field shall be set to zero.

The first block location field indicates the block address associated with the
current logical position.  The value shall indicate the block address of the
next data block to be transferred between the initiator and the target if a
READ or WRITE command is issued.

The last block location field indicates the block address associated with the
next block to be transferred from the buffer to the medium.  The value shall
indicate the block address of the next data block to be transferred between
the buffer and the medium.  If the buffer does not contain a whole block of
data or is empty, the value reported for the last block location shall be
equal to the value reported for the first block location.

IMPLEMENTORS NOTE:  The information provided by the first and last block
location fields can be used in conjunction with the LOCATE command to position
the medium at the appropriate logical block on another device in the case of
unrecoverable errors on the first device.

The number of blocks in buffer field indicates the number of data blocks in
the target's buffer that have not been written to the medium.

The number of bytes in buffer field indicates the total number of data bytes
in the target's buffer that have not been written to the medium.

9.2.7. READ REVERSE Command



















SCSI-2 draft proposed American National Standard        9-25      Revision 10c25







Sequential-Access Devices                                               3/9/90


                       Table 9-10: READ REVERSE Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (0Fh)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |         Reserved         |  SILI  | Fixed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Transfer Length                             |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The READ REVERSE command (Table 9-10) requests that the target transfer one or
more block(s) of data to the initiator beginning at the current position on
the logical unit.

The execution of this command is similar to the READ command except that
medium motion is in the reverse direction.  All block(s), and the byte(s)
within the block(s), are transferred in the reverse order.  The order of bits
within each byte shall not be changed.  Upon completion of a READ REVERSE
command, the logical position shall be before the last block transferred
(beginning-of-partition side).

Refer to the READ command (9.2.4) for a description of the fixed bit, the SILI
bit, the transfer length field, and any associated error conditions.

Filemarks, setmarks, incorrect length blocks, and unrecovered read errors are
handled the same as in the READ command, except that upon termination the
logical position shall be before the filemark, setmark, incorrect length
block, or unrecovered block (beginning-of-partition side).

If the logical unit encounters beginning-of-partition during a READ REVERSE
command, CHECK CONDITION status shall be returned and the EOM and valid bits
shall be set to one in the sense data.  The sense key shall be set to NO SENSE
or RECOVERED ERROR, as appropriate.  If the fixed bit is one, the information
field shall be set to the requested transfer length minus the actual number of
blocks transferred.  If the fixed bit is zero, the information field shall be
set to the requested transfer length.


9.2.8. RECOVER BUFFERED DATA Command

SCSI-2 draft proposed American National Standard        9-26      Revision 10c26







Sequential-Access Devices                                               3/9/90


                  Table 9-11: RECOVER BUFFERED DATA Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (14h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |         Reserved         |  SILI  | Fixed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Transfer Length                             |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The RECOVER BUFFERED DATA command (Table 9-11) is used to recover data that
has been transferred to the target's buffer but has not been successfully
written to the medium.  It is normally used to recover from error or exception
conditions that make it impossible to write the buffered data to the medium.
One or more RECOVER BUFFERED DATA commands may be required to recover all
unwritten buffered data.

The execution of this command is similar to the READ command except that the
data is transferred from the target's buffer instead of the medium.  The order
in which block(s) are transferred is defined by the RBO bit in the device
configuration page (9.3.3.1).  If the RBO bit is not implemented, block(s) are
transferred in the same order they would have been transferred to the medium.

Refer to the READ command (9.2.4) for a description of the fixed bit, the SILI
bit, the transfer length field, and any associated error conditions.

If the fixed bit is zero, no more than the requested transfer length shall be
transferred to the initiator.  If the requested transfer length is smaller
than the actual length of the logical block to be recovered, only the
requested transfer length shall be transferred to the initiator and the
remaining data for the current logical block shall be discarded.

IMPLEMENTORS NOTE:  During recovery operations involving unknown block sizes,
the initiator should select the maximum block length supported by the target
to ensure that all buffered data will be transferred.

If a buffered filemark is encountered during a RECOVER BUFFERED DATA command,



SCSI-2 draft proposed American National Standard        9-27      Revision 10c27







Sequential-Access Devices                                               3/9/90

CHECK CONDITION status shall be returned, the sense key shall be set to NO
SENSE, and the filemark and valid bits shall be set to one in the sense data.
Upon termination, the logical position shall be after the filemark.  If the
fixed bit is one, the information field shall be set to the requested transfer
length minus the actual number of blocks transferred (not including the
filemark).  If the fixed bit is zero, the information field shall be set to
the requested transfer length.

If a buffered setmark is encountered during a RECOVER BUFFERED DATA command
and the RSmk bit is set to one in the device configuration page (9.3.3.1),
CHECK CONDITION status shall be returned and the filemark and valid bits shall
be set to one in the sense data.  The sense key shall be set to NO SENSE and
the additional sense code shall be set to SETMARK DETECTED.  Upon termination,
the logical position shall be after the setmark.  If the fixed bit is one, the
information field shall be set to the requested transfer length minus the
actual number of blocks transferred (not including the setmark).  If the fixed
bit is zero, the information field shall be set to the requested transfer
length. The target shall not return CHECK CONDITION when a setmark is
encountered if the RSmk bit is zero or if this option is not supported.

If an attempt is made to recover more logical blocks of data than are
contained in the target's buffer, CHECK CONDITION status shall be returned,
the sense key shall be set to NO SENSE, and the EOM and valid bits shall be
set to one in the sense data.  If the fixed bit is one, the information field
shall be set to the requested transfer length minus the actual number of
blocks transferred.  If the fixed bit is zero, the information field shall be
set to the requested transfer length.

9.2.9. RELEASE UNIT Command






















SCSI-2 draft proposed American National Standard        9-28      Revision 10c28







Sequential-Access Devices                                               3/9/90


                       Table 9-12: RELEASE UNIT Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (17h)                        |
-----|-----------------------------------------------------------------------|
 1   | Logical Unit Number      | 3rdPty | Third Party Device ID    |Reserved|
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The RESERVE UNIT and RELEASE UNIT commands provide the basic mechanism for
contention resolution in multiple-initiator systems.  The RELEASE UNIT command
(Table 9-12) is used to release previously reserved logical units for the
requesting initiator, or if it is a third-party release, to another specified
SCSI device.

If a valid reservation exists for the I_T_L nexus, the target shall release
the reservation and return GOOD status.

It is not an error to attempt to release a reservation that is not currently
valid.  In this case, the target shall return GOOD status without altering any
other reservation.

9.2.9.1. Third-Party Release
Third-party release allows an initiator to release a logical unit that was
previously reserved using a third-party reservation (see 9.2.10.1).

If the third party (3rdPty) bit is zero, then a third-party release is not
requested.  If the 3rdPty bit is one, and if the reservation was made using a
third-party reservation by the initiator that is requesting the release for
the same SCSI device as specified in the third-party device ID, then the
target shall release the reservation.

If the 3rdPty bit is one, the target shall not modify the mode parameters for
commands received from the third-party device even if the target implements
the transfer of mode parameters with a third-party RESERVE UNIT command.

IMPLEMENTORS NOTE:  When a target implements independent storage of mode


SCSI-2 draft proposed American National Standard        9-29      Revision 10c29







Sequential-Access Devices                                               3/9/90

parameters for each initiator, a third-party RESERVE UNIT command effects a
transfer of the current mode parameters.  Those set up by the initiator of the
RESERVE UNIT are to be set as the mode parameters used for commands from the
third-party device (usually a copy master device).  A unit attention condition
notifies the third-party device of the changed mode parameters.  A successful
third-party RELEASE UNIT command leaves the transferred parameters intact.
The third-party device can issue MODE SENSE and MODE SELECT commands to query
and modify the mode parameters.



9.2.10. RESERVE UNIT Command

                       Table 9-13: RESERVE UNIT Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (16h)                        |
-----|-----------------------------------------------------------------------|
 1   | Logical Unit Number      | 3rdPty | Third Party Device ID    |Reserved|
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The RESERVE UNIT and RELEASE UNIT commands provide the basic mechanism for
contention resolution in multiple-initiator systems.  The RESERVE UNIT command
(Table 9-13) is used to reserve logical units for the exclusive use of the
requesting initiator, or if it is a third-party reservation, to another
specified SCSI device.

This command requests that the entire logical unit be reserved for the
exclusive use of the initiator until the reservation is superseded by another
valid RESERVE UNIT command from the initiator that made the reservation or
until released by a RELEASE UNIT command from the same initiator that made the
reservation, by a BUS DEVICE RESET message from any initiator, by a hard reset
condition, or by a power on cycle.  The reservation shall not be granted if
the logical unit is reserved by another initiator.  It shall be permissible
for an initiator to reserve a logical unit that is currently reserved by that
initiator.

If the logical unit is reserved for another initiator, the target shall return

SCSI-2 draft proposed American National Standard        9-30      Revision 10c30







Sequential-Access Devices                                               3/9/90

RESERVATION CONFLICT status.

If, after honoring the reservation, any other initiator attempts to perform
any command on the reserved logical unit other than an INQUIRY, REQUEST SENSE,
ALLOW MEDIUM REMOVAL, RESERVE, or a RELEASE UNIT command, the command shall be
rejected with RESERVATION CONFLICT status.


9.2.10.1. Third-Party Reservation
Third-party reservation allows an initiator to reserve a logical unit for
another SCSI device.  This is intended for use in multiple-initiator systems
that use the COPY command.

If the third-party (3rdPty) bit is zero, a third-party reservation is not
requested.  If the 3rdPty bit is one the target shall reserve the logical unit
for the SCSI device specified in the third-party device ID field. The target
shall preserve the reservation until it is superseded by another valid RESERVE
UNIT command from the initiator that made the reservation or until it is
released by the same initiator, by a BUS DEVICE RESET message from any
initiator, or a hard reset condition.  The target shall ignore any attempt to
release the reservation made by any other initiator.

If independent sets of parameters are implemented, a third party reservation
shall cause the target to transfer the set of parameters in effect for the
initiator of the RESERVE command to the parameters used for commands from the
third party device.  Any subsequent command issued by the third-party device
is executed according to the mode parameters in effect for the initiator that
sent the RESERVE command.

If independent sets of parameters are implemented, a third party reservation
shall cause the target to transfer the set of parameters in effect for the
initiator of the RESERVE command to the parameters used for commands from the
third party device.  Any subsequent command issued by the third-party device
is executed according to the mode parameters in effect for the initiator that
sent the RESERVE command.

IMPLEMENTORS NOTE:  This transfer of the mode parameters is applicable to
target devices which store mode information independently for different
initiators.  This mechanism allows an initiator to set the mode parameters of
a target for the use of a copy master (i.e., the third-party device).  The
third-party copy master may subsequently issue a MODE SELECT command to modify
the mode parameters.

9.2.10.2. Superseding Reservations







SCSI-2 draft proposed American National Standard        9-31      Revision 10c31







Sequential-Access Devices                                               3/9/90

An initiator that currently has a logical unit reserved may modify the current
reservation by issuing another RESERVE UNIT command to the same logical unit.
The superseding reservation shall release the current reservation if the
superseding reservation request is granted.  The current reservation shall not
be modified if the superseding reservation request cannot be granted.  If the
superseding reservation cannot be granted because of conflicts with a previous
reservation (other than the current reservation), then the target shall return
RESERVATION CONFLICT status.

IMPLEMENTORS NOTE:  Superseding reservations allow the SCSI device ID to be
changed on a reservation using the third-party reservation option.  This
capability is necessary for certain situations when using COMPARE, COPY, and
COPY AND VERIFY commands.

9.2.11. REWIND Command

                          Table 9-14: REWIND Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (01h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |              Reserved             | Immed  |
-----|-----------------------------------------------------------------------|
 2   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 3   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The REWIND command (Table 9-14) causes the target to position the logical unit
to the beginning-of-partition in the current partition.  Prior to performing
the rewind operation, the target shall ensure that all buffered data,
filemarks, and setmarks have been transferred to the medium.

An immediate (Immed) bit of zero indicates that the target shall not return
status until the rewind operation has completed.  An Immed bit of one
indicates that the target shall return status as soon as all buffered commands
have completed execution and the command descriptor block of the REWIND
command has been validated.  If CHECK CONDITION status is returned for a
REWIND command with an Immed bit of one, the rewind operation shall not be
performed.

IMPLEMENTORS NOTE:  For compatibility with devices implemented prior to this

SCSI-2 draft proposed American National Standard        9-32      Revision 10c32







Sequential-Access Devices                                               3/9/90

standard, it is suggested that a WRITE FILEMARKS command with an Immed bit of
zero be used to ensure that all buffered data, filemarks, or setmarks have
been transferred to the medium before issuing a REWIND command with an Immed
bit of one.

When operating in buffered mode 1h or 2h (see 9.3.3), the target shall discard
any unwritten buffered data after the REWIND command is validated if the
previous command was terminated with CHECK CONDITION status and the device is
unable to continue successfully writing.

9.2.12. SPACE Command

                          Table 9-15: SPACE Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation  (11h)                            |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |     Reserved    |           Code           |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Count                                       |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The SPACE command (Table 9-15) provides a variety of positioning functions
that are determined by the code and count.  Both forward and reverse
positioning are provided, although some targets may only support a subset of
this command.  If an initiator requests an unsupported function, the command
shall be terminated with CHECK CONDITION status and the sense key shall be set
to ILLEGAL REQUEST.

The code is defined as follows:











SCSI-2 draft proposed American National Standard        9-33      Revision 10c33







Sequential-Access Devices                                               3/9/90


                   Code         Description          Support
                 ---------  --------------------    ---------
                   000b     Blocks                  Mandatory
                   001b     Filemarks               Mandatory
                   010b     Sequential Filemarks    Optional
                   011b     End-of-Data             Optional
                   100b     Setmarks                Optional
                   101b     Sequential Setmarks     Optional
                 110b-111b  Reserved

When spacing over blocks, filemarks, or setmarks, the count field specifies
the number of blocks, filemarks, or setmarks to be spaced over in the current
partition.  A positive value N in the count field shall cause forward
positioning (toward end-of-partition) over N blocks, filemarks, or setmarks
ending on the end-of-partition side of the last block, filemark, or setmark.
A zero value in the count field shall cause no change of logical position.  A
negative value -N (two's complement notation) in the count field shall cause
reverse positioning (toward beginning-of-partition) over N blocks, filemarks,
or setmarks ending on the beginning-of-partition side of the last block,
filemark, or setmark.  Support of spacing in the reverse direction is
optional.

If a filemark is encountered while spacing over blocks, the command shall be
terminated.  The logical position shall be on the end-of-partition side of the
filemark if movement was in the forward direction and on the beginning-of-
partition side of the filemark if movement was in the reverse direction.
CHECK CONDITION status shall be returned to the initiator, the sense key shall
be set to NO SENSE, and the filemark and valid bits shall be set to one in the
sense data.  The information field shall be set to the requested count minus
the actual number of blocks spaced over (not including the filemark).

If a setmark is encountered while spacing over blocks or filemarks and the
RSmk bit is set to one in the device configuration page (9.3.3.1), the command
shall be terminated, CHECK CONDITION status shall be returned to the
initiator, and the filemark and valid bits shall be set to one in the sense
data.  The sense key shall be set to NO SENSE and the additional sense code
shall be set to SETMARK DETECTED.  The information field shall be set to the
requested count minus the actual number of blocks or filemarks spaced over
(not including the setmark).  The logical position shall be on the end-of-
partition side of the setmark if movement was in the forward direction and on
the beginning-of-partition side of the setmark if movement was in the reverse
direction.  The target shall not return CHECK CONDITION status when a setmark
is encountered if the RSmk bit is set to zero or if this option is not
supported.

If early-warning is encountered while spacing over blocks, filemarks, or




SCSI-2 draft proposed American National Standard        9-34      Revision 10c34







Sequential-Access Devices                                               3/9/90

setmarks and the REW bit is set to one in the device configuration page
(9.3.3.1), CHECK CONDITION status shall be returned, the sense key shall be
set to NO SENSE, and the EOM and valid bits shall be set to one in the sense
data.  The information field shall be set to the requested count minus the
actual number of blocks, filemarks, or setmarks spaced over as defined by the
code value.  If the REW bit is zero or the option is not supported by the
target, the target shall not report CHECK CONDITION status at the early-
warning point.

IMPLEMENTORS NOTE:  Setting the REW bit is not recommended for most system
applications since data may be present after early-warning.

If end-of-data is encountered while spacing over blocks, filemarks, or
setmarks, CHECK CONDITION status shall be returned, the sense key shall be set
to BLANK CHECK, and the valid bit shall be set to one in the sense data.
Additionally, the EOM bit shall be set to one if end-of-data is encountered at
or after early-warning.  The information field shall be set to the requested
count minus the actual number of blocks, filemarks, or setmarks spaced over as
defined by the code value.

If the end-of-partition is encountered while spacing forward over blocks,
filemarks, or setmarks, CHECK CONDITION status shall be returned, the sense
key shall be set to MEDIUM ERROR, the EOM bit shall be set to one, and the
valid bit shall be set to one.  The information field shall be set to the
requested count minus the actual number of blocks, filemarks, or setmarks
spaced over as defined by the code value.

If beginning-of-partition is encountered while spacing over blocks, filemarks,
or setmarks in the reverse direction, the target shall return CHECK CONDITION
status and shall set the sense key to NO SENSE.  The EOM and valid bits shall
be set to one, and the information field set to the total number of blocks,
filemarks, or setmarks not spaced over (the requested number of blocks,
filemarks, or setmarks minus the actual number of blocks, filemarks, or
setmarks spaced over).

When spacing over sequential filemarks (or setmarks), the count field is
interpreted as follows:
(1) A positive value N shall cause forward movement to the first occurrence of
N or more consecutive filemarks (or setmarks) being logically positioned after
the Nth filemark (or setmark).
(2) A zero value shall cause no change in the logical position.
(3) A negative value -N (2's complement notation) shall cause reverse movement
to the first occurrence of N or more consecutive filemarks (or setmarks) being
logically positioned on the beginning-of-partition side of the Nth filemark
(or setmark).

If a setmark is encountered while spacing to sequential filemarks and the RSmk




SCSI-2 draft proposed American National Standard        9-35      Revision 10c35







Sequential-Access Devices                                               3/9/90

bit is set to one in the device configuration page (9.3.3.1), CHECK CONDITION
status shall be returned, the filemark bit shall be set to one, and the valid
bit shall be set to zero.  The sense key shall be set to NO SENSE and the
additional sense code shall be set to SETMARK DETECTED.  The target shall not
return CHECK CONDITION status when a setmark is encountered if the RSmk bit is
set to zero or if this option is not supported.

If end-of-partition is encountered while spacing to sequential filemarks or
setmarks, CHECK CONDITION status shall be returned, the sense key shall be set
to MEDIUM ERROR, the EOM bit shall be set to one, and the valid bit shall be
set to zero.

If end-of-data is encountered while spacing to sequential filemarks or
setmarks, CHECK CONDITION status shall be returned, the sense key shall be set
to BLANK CHECK, and the valid bit shall be set to zero.  Additionally, the EOM
bit shall be set to one if end-of-data is encountered at or after early-
warning.

When spacing to end-of-data, the count field is ignored.  Upon successful
completion, the medium shall be positioned such that a subsequent write
operation would append to the last logically recorded information.

If end-of-partition is encountered while spacing to end-of-data, CHECK
CONDITION status shall be returned, the sense key shall be set to MEDIUM
ERROR, the EOM bit shall be set to one, and the valid bit shall be set to
zero.

9.2.13. VERIFY Command























SCSI-2 draft proposed American National Standard        9-36      Revision 10c36







Sequential-Access Devices                                               3/9/90


                          Table 9-16: VERIFY Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (13h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |    Reserved     | Immed  | BytCmp | Fixed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Verification Length                         |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The VERIFY command (Table 9-16) requests that the target verify one or more
block(s) beginning with the next block on the logical unit.

An immediate (Immed) bit of zero indicates that the command shall not return
status until the verify operation has completed.  An Immed bit of one
indicates that status shall be returned as soon as the command descriptor
block has been validated (but after the data transfer from the initiator, if
the BytCmp bit is one).

IMPLEMENTORS NOTE: In order to ensure that no errors are lost, the initiator
should set the Immed bit to zero on the last VERIFY command when issuing a
series of VERIFY commands.

A byte compare (BytCmp) bit of zero indicates that the verification shall be
simply a medium verification (e.g., CRC, ECC).  No data shall be transferred
between the initiator and target.

A BytCmp bit of one indicates that the target shall perform a byte-by-byte
compare of the data on the medium and the data transferred from the initiator
during the DATA OUT phase of the VERIFY command.  Data shall be transferred
from the initiator to the target as in a WRITE command.  If the BytCmp bit is
one and the byte compare option is not supported, the target shall terminate
the command with CHECK CONDITION status and the sense key shall be set to
ILLEGAL REQUEST.

The verification length specifies the amount of data to verify, in blocks or




SCSI-2 draft proposed American National Standard        9-37      Revision 10c37







Sequential-Access Devices                                               3/9/90

bytes, as indicated by the fixed bit. Refer to the READ command (9.2.4) for a
description of the fixed bit and any associated error conditions. When the
verification length is zero, no data shall be verified and the current logical
position shall not be changed.  This condition shall not be considered as an
error.

The VERIFY command shall terminate when the verification length has been
satisfied, when an incorrect length block is encountered, when a filemark is
encountered, when a setmark is encountered (if the RSmk bit is one in the
device configuration page, 9.3.3.1), when end-of-data is encountered, when the
end-of-partition is encountered, when early-warning is encountered (if the REW
bit is one in the device configuration page, 9.3.3.1), or when an
unrecoverable read error is encountered.  The status and sense data for each
of these conditions are handled in the same manner as in the READ command
(9.2.4).  Upon successful completion of a VERIFY command, the logical position
shall be after the last block verified.

If the data does not compare (BytCmp bit of one), the command shall terminate
with CHECK CONDITION status, the valid bit shall be set to one, and the sense
key shall be set to MISCOMPARE.  If the fixed bit is one, the information
field shall be set to the requested verification length minus the actual
number of blocks successfully verified.  If the fixed bit is zero, the
information field shall be set to the requested verification length minus the
actual number of bytes successfully verified.  This number may be larger than
the requested verification length if the error occurred on a previous VERIFY
command with an Immed bit of one.  Upon termination, the medium shall be
positioned after the block containing the miscompare (end-of-partition side).


9.2.14. WRITE Command





















SCSI-2 draft proposed American National Standard        9-38      Revision 10c38







Sequential-Access Devices                                               3/9/90


                          Table 9-17: WRITE Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (0Ah)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |              Reserved             | Fixed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Transfer Length                             |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The WRITE command (Table 9-17) requests that the target write the data that is
transferred from the initiator to the current position on the logical unit.

The fixed bit specifies whether fixed-length or variable-length blocks are to
be transferred.  See the READ BLOCK LIMITS command (9.2.5) for additional
information about fixed and variable block mode.

If the fixed bit is one, the transfer length specifies the number of fixed-
length blocks to be transferred, using the current block length reported in
the mode parameter block descriptor (7.3.3).  If the fixed bit is zero, a
single block is transferred with the transfer length specifying the block
length in bytes.

If the transfer length is zero, no data shall be transferred and the current
position on the logical unit shall not be changed.  This condition shall not
be considered an error.

A WRITE command may be buffered or unbuffered, as indicated by the buffered
mode field of the mode parameter header (9.3.3).  For unbuffered operation
(buffered mode 0h), the target shall not return GOOD status until all data
block(s) are successfully written to the medium.  For buffered operation
(buffered mode 1h or 2h), the target may return GOOD status as soon as all
data block(s) are successfully transferred to the target's buffer.

IMPLEMENTORS NOTE:  For compatibility with devices implemented prior to this





SCSI-2 draft proposed American National Standard        9-39      Revision 10c39







Sequential-Access Devices                                               3/9/90

version of the standard, a WRITE FILEMARKS command with the Immed bit set to
zero should be issued when completing a buffered write operation to ensure
that all buffered data, filemarks, and setmarks are written to the medium.

If the logical unit encounters early-warning during a WRITE command, an
attempt to finish writing any data may be made, as determined by the current
settings of the REW and SEW bits in the device configuration page (9.3.3.1).
The command shall terminate with CHECK CONDITION status and the EOM and valid
bits shall be set to one in the sense data.  If all data that is to be written
is successfully transferred to the medium, the sense key shall be set to NO
SENSE or RECOVERED ERROR, as appropriate.  If any data that is to be written
cannot be transferred to the medium when early-warning is encountered, the
sense key shall be set to VOLUME OVERFLOW.

The information field shall be defined as follows:
(1)  If unbuffered mode is reported in the mode parameter header and the fixed
bit is set to one, the information field shall be set to the requested
transfer length minus the actual number of blocks written.
(2)  If unbuffered mode is reported and the fixed bit is set to zero, the
information field shall be set to the requested transfer length.
(3)  If buffered mode is reported in the mode parameter header and the fixed
bit is set to one, the information field shall be set to the total number of
blocks, filemarks, and setmarks not written (the number of blocks not
transferred from the initiator plus the number of blocks, filemarks, and
setmarks remaining in the target's buffer).  Note that the value in the
information field may exceed the transfer length.
(4)  If buffered mode is reported and the fixed bit is set to zero, the
information field shall be set to the total number of bytes, filemarks, and
setmarks not written (the number of bytes not transferred from the initiator
plus the number of bytes, filemarks, and setmarks remaining in the target's
buffer).  Note that the value in the information field may exceed the transfer
length.

IMPLEMENTORS NOTE:  The target should ensure that some additional data can be
written to the medium (e.g., labels, filemarks, or setmarks) after the first
early-warning indication has been returned to the initiator.

If a WRITE command is received while the logical unit is positioned between
early-warning and end-of-partition, the target shall return CHECK CONDITION
status after attempting to perform the command.  The EOM and valid bits shall
be set to one in the sense data.  If all data that is to be written is
successfully transferred to the medium, the information field shall be set to
zero.  If any data that is to be written is not transferred to the medium
prior to encountering end-of-partition, the sense key shall be set to VOLUME
OVERFLOW and the information field shall be defined as follows:
(1)  If the fixed bit is one, the information field shall be set to the





SCSI-2 draft proposed American National Standard        9-40      Revision 10c40







Sequential-Access Devices                                               3/9/90

requested transfer length minus the actual number of blocks written to the
medium.
(2)  If the fixed bit is zero, the information field shall be set to the
requested transfer length.

IMPLEMENTORS NOTE:  In some systems it is important to recognize an error if
end-of-partition is encountered during execution of a WRITE command, without
regard for whether all data that is to be written is successfully transferred
to the medium.  By its definition, the VOLUME OVERFLOW sense key may always
validly be returned if end-of-partition is encountered while writing, and such
usage is recommended.  Reporting the MEDIUM ERROR sense key, as was done in
some SCSI-1 implementations, may cause confusion as to whether there was
really defective medium encountered during execution of the last write
command.

9.2.15. WRITE FILEMARKS Command

                     Table 9-18: WRITE FILEMARKS Command

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |                           Operation Code (10h)                        |
-----|-----------------------------------------------------------------------|
 1   |   Logical Unit Number    |         Reserved         |  WSmk  | Immed  |
-----|-----------------------------------------------------------------------|
 2   | (MSB)                                                                 |
-----|---                                                                 ---|
 3   |                           Transfer Length                             |
-----|---                                                                 ---|
 4   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 5   |                           Control                                     |
==============================================================================


The WRITE FILEMARKS command (Table 9-18) requests that the target write the
specified number of filemarks or setmarks to the current position on the
logical unit.

If the Write Setmark (WSmk) bit is one, the transfer length specifies the
number of setmarks to be written.  If the WSmk bit is zero, the transfer
length specifies the number of filemarks to be written.

An immediate (Immed) bit of one indicates that the target shall return status





SCSI-2 draft proposed American National Standard        9-41      Revision 10c41







Sequential-Access Devices                                               3/9/90

as soon as the command descriptor block has been validated.  An Immed bit of
one is only valid if buffered mode is reported in the mode parameter header
(9.3.3).

An Immed bit of zero indicates that the target shall not return status until
the write operation has completed.  Any buffered data, filemarks, and setmarks
shall be written to the medium prior to completing the command.

IMPLEMENTORS NOTE:  Upon completion of any buffered write operation, the
initiator can issue a WRITE FILEMARKS command with the Immed bit set to zero
and the transfer length field set to zero to ensure that all buffered data,
filemarks, and setmarks are successfully written to the medium.

If the logical unit encounters early-warning during a WRITE FILEMARKS command,
an attempt to finish writing any buffered data, filemarks, or setmarks may be
made, as determined by the current settings of the REW and SEW bits in the
device configuration page (9.3.3.1).  The command shall terminate with CHECK
CONDITION status and the EOM and valid bits shall be set to one in the sense
data.  If all buffered data, filemarks, and setmarks are successfully
transferred to the medium, the sense key shall be set to NO SENSE or RECOVERED
ERROR, as appropriate. If any buffered data, filemarks, or setmarks to be
written are not transferred to the medium when early-warning is encountered,
the sense key shall be set to VOLUME OVERFLOW.

The information field shall be defined as follows:
(1) If unbuffered mode is reported in the mode parameter header, the
information field shall be set to the requested transfer length minus the
actual number of filemarks or setmarks written.
(2) If buffered mode is reported in the mode parameter header and the buffered
data was written in variable block mode (see 9.2.14), the information field
shall be set to the total number of bytes, filemarks, and setmarks not written
(the number of filemarks or setmarks not transferred from the initiator plus
the number of bytes, filemarks and setmarks remaining in the target's buffer).
Note that it is possible for the value in the information field to exceed the
transfer length.
(3) If buffered mode is reported and the buffered data was written in fixed
block mode (see 9.2.14), the information field shall be set to the total
number of blocks, filemarks, and setmarks not written (the number filemarks or
setmarks not transferred from the initiator plus the number of blocks,
filemarks, and setmarks remaining in the target's buffer).  Note that it is
possible for the value in the information field to exceed the transfer length.

IMPLEMENTORS NOTE:  The target should ensure that some additional data can be
written to the medium (e.g., labels, filemarks, or setmarks) after the first
early-warning indication has been returned to the initiator.

If a WRITE FILEMARKS command is received while the logical unit is positioned




SCSI-2 draft proposed American National Standard        9-42      Revision 10c42







Sequential-Access Devices                                               3/9/90

between early-warning and end-of-partition, the target shall return CHECK
CONDITION status after attempting to perform the command.  The EOM and valid
bits shall be set to one in the sense data.  If all filemarks or setmarks to
be written are successfully transferred to the medium, the information field
shall be set to zero.  If any filemarks or setmarks to be written are not
transferred to the medium prior to encountering end-of- partition, the sense
key shall be set to VOLUME OVERFLOW and the information field shall be set to
the requested transfer length minus the actual number of filemarks or setmarks
written to the medium.


9.3. Parameters for Sequential-Access Devices

9.3.1. Diagnostic Parameters

This section defines the descriptors and pages for diagnostic parameters used
with sequential-access devices.

The diagnostic page codes for sequential-access devices are defined in Table
9-19.

                      Table 9-19: Diagnostic Page Codes

==============================================================================
Page Code     Description                                           Section
------------------------------------------------------------------------------
00h        Supported Diagnostic Pages                            7.3.1.1
01h - 3Fh     Reserved (for all device type pages)
40h - 7Fh     Reserved
80h - FFh     Vendor-specific pages
==============================================================================


9.3.2. Log Parameters

















SCSI-2 draft proposed American National Standard        9-43      Revision 10c43







Sequential-Access Devices                                               3/9/90


This section defines the descriptors and pages for log parameters used with
sequential-access devices.

The log page codes for sequential-access devices are defined in Table 9-20.

                          Table 9-20: Log Page Codes

==============================================================================
  Page Code    Description                                          Section
------------------------------------------------------------------------------
     01h       Buffer Over-Run/Under-Run Page                       7.3.2.1
     02h       Error Counter Page (Write) Page                      7.3.2.2
     03h       Error Counter Page (Read) Page                       7.3.2.2
     04h       Error Counter Page (Read Reverse) Page               7.3.2.2
     05h       Error Counter Page (Verify) Page                     7.3.2.2
     07h       Last n Error Events Page                             7.3.2.3
     06h       Non-Media Error Page                                 7.3.2.4
     00h       Supported Log Pages                                  7.3.2.5
  08h - 2Fh    Reserved
     3Fh       Reserved
  30h - 3Eh    Vendor Specific
==============================================================================


9.3.3. Mode Parameters

This section defines the descriptors and pages for mode parameters used with
sequential-access devices.

The mode parameter list, including the mode parameter header and mode block
descriptor, are described in 7.3.3.

The medium-type code field in the mode parameter header is reserved for
sequential-access devices.

The device-specific parameter byte of the mode parameter header (see Tables
7-61 and 7-62) is defined in Table 9-21 for sequential-access devices.

                      Table 9-21: Device-Specific Parameter

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
==============================================================================
     |   WP   |      Buffered Mode       |               Speed               |
==============================================================================


For the MODE SENSE command, a write protect (WP) bit of zero indicates that
the medium is write enabled.  A WP bit of one indicates that the medium is

SCSI-2 draft proposed American National Standard        9-44      Revision 10c44







Sequential-Access Devices                                               3/9/90

write protected.  For the MODE SELECT command, this field is ignored.

Values for the buffered mode field are defined as follows:
0h       The target shall not report GOOD status on write commands until the
data blocks are actually written on the medium.

1h       The target may report GOOD status on write commands as soon as all
the data specified in the write command has been transferred to the target's
buffer.  One or more blocks may be buffered prior to writing the block(s) to
the medium.

2h       The target may report GOOD status on write commands as soon as:
         (1) All the data specified in the write command has been
successfully transferred to the target's buffer, and
         (2) All buffered data from different initiators has been successfully
written to the medium.

3h - 7h  Reserved

Values for the speed field shall be assigned as follows:
0h       Default (Use the peripheral device's default speed).
1h       Use the peripheral device's lowest speed.
2h - Fh  Use increasing peripheral device speeds.

For the MODE SELECT command, the density code field of the sequential-access
device block descriptor (7.3.3) indicates the density selected by the
initiator for use in subsequent read and write operations.  For devices
capable of automatic density recognition, the density code selected by the
initiator may be overridden by the target for a subsequent read operation if
the selected value does not match the current recorded density of the medium.
If the MODE SELECT command specifies the default density code the device
selects the actual density code to be used in a vendor specific manner.  The
value is expected to be the principal density code (or an optimal density
code).

For the MODE SENSE command, the density code field reflects the current
operating density of the device.  If a current operating density has not been
selected, either because no medium is installed or because the density of the
installed medium has not been determined, the density code field should be set
to the principal density code value.  The principal density code is the
highest density code supported, the optimal density code supported, or the
most common density code supported.  In some devices, the principal density
code value returned in response to a MODE SENSE command may change dynamically
to match the most recently selected density.  The density code value returned
in response to a MODE SENSE command shall be as described below:
(1) Following a UNIT ATTENTION condition for a power on or hard reset
condition, while not ready, the target shall report the principal density.
(2)  Following a UNIT ATTENTION condition for a not-ready-to-ready transition,



SCSI-2 draft proposed American National Standard        9-45      Revision 10c45







Sequential-Access Devices                                               3/9/90

the target shall:
(a) report the principal density if no attempt has been made by the target to
determine the density.
(b) report the principal density if the target cannot automatically determine
the density from the medium.
(c) report the current medium density if the target can determine the density
from the medium.
(3) Following a successful read operation at or after beginning-of-medium, the
target shall report a density code value reflecting the recorded density of
the medium.  For some implementations, the target may automatically determine
this value from the medium.  For devices not capable of automatic density
determination, the principal density is reported if the density code value is
not provided by the preceding MODE SELECT command.
(4) Following an unsuccessful read operation or a successful write operation,
while at beginning-of-partition, the target shall:
(a) report a density code value as described for item (2) if a previous MODE
SELECT command has not established a density code for the currently mounted
volume.
(b) report a density code value as provided by the last successful MODE SELECT
command for the currently mounted volume.
(5) Following a successful unload operation the target shall report the most
recent density code value as determined by items (2) through (4) above.

Table 9-22 lists the sequential-access device density codes



























SCSI-2 draft proposed American National Standard        9-46      Revision 10c46







Sequential-Access Devices                                               3/9/90


                 Table 9-22: Sequential-Access Density Codes

==============================================================================
Code Value                             Density
----------  ------------------------------------------------------------------
   00h      Default density (MODE SELECT command only)

                                    Magnetic Tapes
            Width                  Density
            mm  (Inch)   Tracks  bpmm    bpi   Code  Type  Reference     Note
            ----------   ------  ------------  ----  ----  ------------  ----
   01h      12.7 (0.5)      9     32    (800)  NRZI   R    X3.22-1983      2
   02h      12.7 (0.5)      9     63   (1600)  PE     R    X3.39-1986      2
   03h      12.7 (0.5)      9    246   (6250)  GCR    R    X3.54-1986      2
   04h       6.3 (0.25)    4/9   315   (8000)  GCR    C    X3.136-1986    1,3
   05h       6.3 (0.25)    4/9   315   (8000)  GCR    C    X3.136-1986     1
   06h      12.7 (0.5)      9    126   (3200)  PE     R    X3.157-1987    2,4
   07h       6.3 (0.25)     4    252   (6400)  IMFM   C    X3.116-1986     1
   08h       3.81 (0.15)    4    315   (8000)  GCR    CS   X3.158-1987    1,4
   09h      12.7 (0.5)     18   1491  (37871)  GCR    C    X3B5/87-099    2,4
   0Ah      12.7 (0.5)     22    262   (6667)  MFM    C    X3B5/86-199    1,4
   0Bh       6.3 (0.25)     4     63   (1600)  PE     C    X3.56-1986      1
   0Ch      12.7 (0.5)     24    500  (12690)  GCR    C    HI-TC1         1,6
   0Dh      12.7 (0.5)     24    999  (25380)  GCR    C    HI-TC2         1,6
   0Eh      Reserved for ECMA
   0Fh       6.3 (0.25)    15    394  (10000)  GCR    C    QIC-120        1,6
   10h       6.3 (0.25)    18    394  (10000)  GCR    C    QIC-150        1,6
   11h       6.3 (0.25)    26    630  (16000)  GCR    C    QIC-320        1,6
   12h       6.3 (0.25)    30   2034  (51667)  RLL    C    QIC-1350       1,6
   13h       3.81 (0.15)    1   2400  (61000)  DDS    CS   X3B5/88-185A    5
   14h       8.00 (0.315)   1   2126  (54000)         CS   X3B5/88-036     5
15h - 7Eh   Reserved
   7Fh      No change from previous density (NO-OP)                        7
80h - FFh   Vendor unique
==============================================================================

Key:
               Code                                       Type
--------------------------------------------       -----------------------
NRZI  Non Return to Zero, change on ones           R   Reel-to-Reel
GCR   Group Code Recording                         C   Cartridge
PE    Phase Encoded                                CS  Cassette
IMFM  Inverted Modified Frequency Modulation
MFM   Modified Frequency Modulation
DDS   DAT Data Storage
RLL   Run Length Limited
NOTES:



SCSI-2 draft proposed American National Standard        9-47      Revision 10c47







Sequential-Access Devices                                               3/9/90

(1)  Serial Recorded.
(2)  Parallel Recorded.
(3)  Old format known as QIC-11.
(4)  See Appendix D for additional standards information.
(5)  Helical Scan
(6)  This is not an American National Standard.  The reference is based on an
industry standard definition of the media format.
(7)  This density code value is defined for the MODE SELECT command and
shall not be returned by the MODE SENSE command.


The mode page codes for sequential-access devices are defined in Table 9-23.

                         Table 9-23: Mode Page Codes

==============================================================================
Page Code     Description                                           Section
------------------------------------------------------------------------------
0Ah        Control Mode Page                                     7.3.3.1
10h        Device Configuration Page                             9.3.3.1
02h        Disconnect-Reconnect Page                             7.3.3.2
11h        Medium Partition Page(1)                              9.3.3.2
12h        Medium Partition Page(2)                              9.3.3.3
13h        Medium Partition Page(3)                              9.3.3.3
14h        Medium Partition Page(4)                              9.3.3.3
09h        Peripheral Device Page                                7.3.3.3
01h        Read-Write Error Recovery Page                        9.3.3.4
00h        Vendor specific (does not require page format)        7.3.3
03h - 08h     Reserved
0Bh - 10h     Reserved
15h - 3Eh     Vendor Specific (page format required)
3Fh        Returns all pages (valid only for the MODE SENSE command)
==============================================================================


9.3.3.1. Device Configuration Page















SCSI-2 draft proposed American National Standard        9-48      Revision 10c48







Sequential-Access Devices                                               3/9/90


                    Table 9-24: Device Configuration Page

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |   PS   |Reserved|         Page Code (10h)                             |
-----|-----------------------------------------------------------------------|
 1   |                           Page Length (0Eh)                           |
-----|-----------------------------------------------------------------------|
 2   |Reserved|  CAP   |  CAF   |                Active Format               |
-----|-----------------------------------------------------------------------|
 3   |                           Active Partition                            |
-----|-----------------------------------------------------------------------|
 4   |                           Write Buffer Full Ratio                     |
-----|-----------------------------------------------------------------------|
 5   |                           Read Buffer Empty Ratio                     |
-----|-----------------------------------------------------------------------|
 6   | (MSB)                                                                 |
-----|---                        Write Delay Time                         ---|
 7   |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 8   |  DBR   |  BIS   |  RSmk  |  AVC   |      SOCF       |   RBO  |  REW   |
-----|-----------------------------------------------------------------------|
 9   |                           Gap Size                                    |
-----|-----------------------------------------------------------------------|
 10  |      EOD Defined         |  EEG   |  SEW   |        Reserved          |
-----|-----------------------------------------------------------------------|
 11  | (MSB)                                                                 |
-----|---                                                                 ---|
 12  |                           Buffer Size at Early Warning                |
-----|---                                                                 ---|
 13  |                                                                 (LSB) |
-----|-----------------------------------------------------------------------|
 14  |                           Select Data Compression Algorithm           |
-----|-----------------------------------------------------------------------|
 15  |                           Reserved                                    |
==============================================================================


This page (Table 9-24) is used to specify the appropriate sequential-access
device configuration.

A change active partition (CAP) bit of one indicates that the logical
partition is to be changed to the one specified by the active partition field.
A CAP bit of zero indicates no partition change is specified.

A change active format (CAF) bit of one indicates that the active format is to


SCSI-2 draft proposed American National Standard        9-49      Revision 10c49







Sequential-Access Devices                                               3/9/90

be changed to the value specified in the active format field.  A CAF bit of
zero indicates no active format change is specified.  For some devices, the
format may only be changed when the logical unit is at beginning-of-partition.

 The active format field indicates which recording format is in use for the
selected density code when reading or writing data on a logical unit.  The
value of the active format field is vendor specific.

The active partition field indicates the current logical partition number in
use on the medium.

The write buffer full ratio, on WRITE commands, indicates to the target how
full the buffer shall be before writing data to the medium.  A value of zero
indicates that the value is not specified.

The read buffer empty ratio, on READ commands, indicates to the target how
empty the buffer shall be before retrieving additional data from the medium.
A value of zero indicates that the value is not specified.

The write delay time indicates the maximum time, in 100 millisecond
increments, that the target should wait before any buffered data that is to be
written, is forced to the medium after the last buffered WRITE command that
did not cause the buffer to exceed the buffer full ratio.  A value of zero
indicates that the target shall never force buffered data to the medium under
these conditions.

A data buffer recovery (DBR) bit of one indicates that the target supports
data buffer recovery using the RECOVER BUFFERED DATA command.  A DBR bit of
zero indicates that the target does not support data buffer recovery.  This
bit is target-defined.

A block identifiers supported (BIS) bit of zero indicates that block IDs are
not supported in the format written on the medium.  A BIS bit of one indicates
that the format on the medium has recorded information about the block IDs
relative to a partition.  This bit is target-defined.

A report setmarks (RSmk) bit of one indicates that the target shall recognize
and report setmarks during appropriate read or space operations.  A RSmk bit
of zero indicates that the target shall not report setmarks.

The automatic velocity control (AVC) bit of one, indicates that the device
shall select the speed (if the device supports more than one speed) based on
the data transfer rate that should optimize streaming activity and minimize
medium repositioning.  An AVC bit of zero indicates the speed chosen should be
the device's default speed.

A stop on consecutive filemarks (SOCF) field of 00b indicates that the device




SCSI-2 draft proposed American National Standard        9-50      Revision 10c50







Sequential-Access Devices                                               3/9/90

shall pre-read data from the medium in buffered mode to the limits of the
buffer capacity without regard for filemarks.  This implies that the device
can differentiate between data blocks and filemarks in the buffer.  Values
01b, 10b, and 11b specify that the device shall terminate the pre-read
operation if 1, 2, or 3 consecutive filemarks are detected, respectively.  If
the RSmk bit is one, the target shall interpret this field as stop on
consecutive setmarks.

A recover buffer order (RBO) bit of one indicates that data blocks shall be
returned from the target's buffer on a RECOVERED BUFFERED DATA command in LIFO
order (last-in-first-out) from which they were written to the buffer.  A RBO
bit of zero indicates data blocks shall be returned in FIFO (first-in-first-
out) order.

A report early-warning (REW) bit of zero indicates that the target shall not
report the early-warning condition for read operations and it shall report
early-warning at or before any medium-defined early-warning position during
write operations.

A REW bit of one indicates that the target shall return CHECK CONDITION status
with the EOM bit set in the sense data when the early-warning position is
encountered during read and write operations.  If the REW bit is one and the
SEW bit is zero, the target shall return CHECK CONDITION status with the sense
key set to VOLUME OVERFLOW when early-warning is encountered during write
operations.

IMPLEMENTORS NOTE: A REW bit of one is intended for compatibility with those
systems using old tape formats that require an early-warning indication during
read operations.  Other systems should set this bit to zero to avoid potential
data loss when interchanging tapes between peripheral devices.

The GapSize field value determines the size of the inter-block gap when
writing data.  A value of 00h specifies the device's defined gap size.  A
value of 01h specifies a device defined gap size sufficiently long to support
update-in-place.  Values of 02h through 0Fh are multipliers on the device's
defined gap size.  Values 10h through 7Fh are reserved.  Values 80h through
FFh are vendor specific.

The end-of-data (EOD) defined field indicates which format type the logical












SCSI-2 draft proposed American National Standard        9-51      Revision 10c51







Sequential-Access Devices                                               3/9/90

unit shall use to detect and generate the EOD area.  The types of EOD formats
are specified below:

           Value                   Description
         ---------   -----------------------------------------------
            000b     Logical unit's default EOD definition
            001b     Format-defined erased area of medium
            010b     As specified in the SOCF field
            011b     EOD recognition and generation is not supported
         100b-111b   Reserved

An enable EOD generation (EEG) bit set to one indicates that the logical unit
shall generate the appropriate EOD area, as determined by the EOD defined
field. A value of zero indicates that EOD generation is disabled.

IMPLEMENTORS NOTE:  Some logical units may not generate EOD at the completion
of any write-type operation.

The synchronize at early-warning (SEW) bit set to one indicates that the
target shall cause any buffered write data, filemarks, or setmarks to be
transferred to the medium when early-warning is encountered. A value of zero
indicates that the target shall retain any unwritten buffered data, filemarks,
or setmarks in the buffer when early-warning is encountered (see the WRITE
command, 9.2.14, and the WRITE FILEMARKS command, 9.2.15).

The buffer size at early-warning field indicates the value, in bytes, to which
the target shall reduce its logical buffer size when writing.  The target
should reduce the buffer size only when the logical unit is positioned between
its early-warning and end-of-partition.  A value of zero indicates that the
implementation of this function is device specific.

NOTE: The intent is to prevent the loss of data by limiting the size of the
buffer when near the end-of-partition.

The select data compression algorithm field set to 00h indicates that the
target shall not use a compression algorithm on any data sent to it prior to
writing the data to the medium.  A value of 01h indicates that the data to be
written shall be compressed using the target's default compression algorithm.
Values 02h through 7Fh are reserved.  Values 80h through FFh are vendor
specific.

9.3.3.2. Medium Partition Page(1)









SCSI-2 draft proposed American National Standard        9-52      Revision 10c52







Sequential-Access Devices                                               3/9/90


                     Table 9-25: Medium Partition Page(1)

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |   PS   |Reserved|         Page Code (11h)                             |
-----|-----------------------------------------------------------------------|
 1   |                           Page Length (n-1)                           |
-----|-----------------------------------------------------------------------|
 2   |                           Maximum Additional Partitions               |
-----|-----------------------------------------------------------------------|
 3   |                           Additional Partitions Defined               |
-----|-----------------------------------------------------------------------|
 4   |  FDP   |  SDP   |  IDP   |       PSUM      |       Reserved           |
-----|-----------------------------------------------------------------------|
 5   |                           Medium Format Recognition                   |
-----|-----------------------------------------------------------------------|
 6   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 7   |                           Reserved                                    |
==============================================================================
8 - n|                   Partition Size Descriptor(s)                        |
==============================================================================
 0   | (MSB)                                                                 |
-----|---                        Partition Size                           ---|
 1   |                                                                 (LSB) |
==============================================================================


This page (Table 9-25) is used to specify the first group of medium
partitions.  Additional groups are specified in medium partition pages(2-4).

The maximum additional partitions field is a target-defined value indicating
the maximum number of additional partitions supported by the logical unit.  A
value of zero indicates that no value is specified.

The additional partitions defined field specifies the number of additional
partitions to be defined for a volume when the SDP or IDP bit is set to one.
The maximum value allowed is the value returned in the maximum additional
partitions field.

A fixed data partitions (FDP) bit of one indicates that the device assigns
partitions based on its fixed definition of partitions.  Setting this bit to
one may only be valid at beginning-of-partition and is mutually exclusive with
the SDP and IDP bits.

A select data partitions (SDP) bit of one indicates that the device is to


SCSI-2 draft proposed American National Standard        9-53      Revision 10c53







Sequential-Access Devices                                               3/9/90

partition the medium into the the number of partitions as specified by the
additional partitions defined field using partition sizes defined by the
device.  Setting this bit to one may only be valid at beginning-of-partition
and it is mutually exclusive with the FDP and IDP fields.

An initiator-defined partitions (IDP) bit of one indicates that the initiator
is defining the number and size of the data partitions using the additional
partitions defined field and the partition size descriptors. Setting this bit
to one may only be valid at beginning-of-partition and is mutually exclusive
with the FDP and SDP fields.

IMPLEMENTORS NOTE: Since defining partitions may require re-formatting the
medium for some implementations, an implicit write to the medium may occur as
a result of a MODE SELECT command that supplies any of these parameters.

Partition size unit of measure (PSUM) field defines the units in which the
partition size descriptors select the partition size.  The values 00b, 01b,
and 10b define the units as bytes, kilobytes, and megabytes, respectively.
The value 11b is reserved.

 The medium format recognition field is a target-defined value indicating the
device's capability to automatically identify the medium format and partition
information when reading an unknown volume.

Values for the medium format recognition field shall be assigned as follows:

  00h    Logical unit is incapable of format or partition recognition.
  01h    Logical unit is capable of format recognition only.
  02h    Logical unit is capable of partition recognition only.
  03h    Logical unit is capable of format and partition recognition.
04h-FFh  Reserved.

IMPLEMENTORS NOTE: If a target indicates that it is not capable of medium
format recognition, the initiator must supply all necessary parameters for the
device to identify the specific format.  The value in this field may be
different following a medium change.

Partition size descriptors define the size of the respective partitions in the
units specified in the PSUM field.  Up to 64 partitions may be defined using
this page.  If additional partitions need to be defined, up to 64 partitions
can be defined in each of the medium partition pages(2-4) defined in 9.3.3.3.

9.3.3.3. Medium Partition Page(2-4)








SCSI-2 draft proposed American National Standard        9-54      Revision 10c54







Sequential-Access Devices                                               3/9/90


                    Table 9-26: Medium Partition Page(2-4)

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |   PS   |Reserved|         Page Code (p)                               |
-----|-----------------------------------------------------------------------|
 1   |                           Page Length (n-1)                           |
==============================================================================
2 - n|                       Partition Size Descriptor(s)                    |
==============================================================================
 0   | (MSB)                                                                 |
-----|---                        Partition Size                           ---|
 1   |                                                                 (LSB) |
==============================================================================


The medium partition pages (Table 9-26) are used to specify additional groups
of medium partitions.  The first group is specified in the medium partition
page(1) (see 9.3.3.2).

The additional page codes (p) defined for partition size definition are 12h,
13h, and 14h.  Up to a maximum of 64 partitions can be defined in each of
these pages.  The partition size is defined by the value of the partition size
field.  The units of size used by the partition size field is specified in the
PSUM field of the medium partition page(1) (9.3.3.2).

IMPLEMENTORS NOTE: Since defining partitions may require re-formatting the
medium for some implementations, an implicit write to the medium may occur as
a result of a MODE SELECT command that supplies these parameters.

9.3.3.4. Read-Write Error Recovery Page

















SCSI-2 draft proposed American National Standard        9-55      Revision 10c55







Sequential-Access Devices                                               3/9/90


                  Table 9-27: Read-Write Error Recovery Page

==============================================================================
  Bit|   7    |   6    |   5    |   4    |   3    |   2    |   1    |   0    |
Byte |        |        |        |        |        |        |        |        |
==============================================================================
 0   |   PS   |Reserved|         Page Code (01h)                             |
-----|-----------------------------------------------------------------------|
 1   |                           Page Length (0Ah)                           |
-----|-----------------------------------------------------------------------|
 2   |Reserved|Reserved|   TB   |Reserved|  EER   |  PER   |  DTE   |  DCR   |
-----|-----------------------------------------------------------------------|
 3   |                           Read Retry Count                            |
-----|-----------------------------------------------------------------------|
 4   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 5   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 6   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 7   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 8   |                           Write Retry Count                           |
-----|-----------------------------------------------------------------------|
 9   |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 10  |                           Reserved                                    |
-----|-----------------------------------------------------------------------|
 11  |                           Reserved                                    |
==============================================================================


The read-write error recovery page (Table 9-27) specifies the error recovery
and reporting parameters that the target shall use when transferring data
between the initiator and the medium.  These parameters only apply to read-
write errors and do not affect message system retries or positioning error
recovery procedures.

NOTE: The parameters in this page also apply to verify operations.

A transfer block (TB) bit of one indicates that a data block that is not
recoverable within the specified recovery limits shall be transferred to the
initiator before CHECK CONDITION status is returned.  A TB bit of zero
indicates that the unrecoverable data block shall not be transferred to the
initiator.  Data blocks that can be recovered within the recovery limits are
always transferred, regardless of the value of the TB bit.

An enable early recovery (EER) bit of one indicates that the target shall use


SCSI-2 draft proposed American National Standard        9-56      Revision 10c56







Sequential-Access Devices                                               3/9/90

the most expedient error recovery algorithm (e.g., attempt error correction
prior to retries).  An EER bit of zero indicates that the target shall use the
most deliberate error recovery algorithm, within the limits established by the
other error recovery parameters (e.g., attempt to recover the block error-free
prior to using error correction).

A post error (PER) bit of one indicates that the target shall return CHECK
CONDITION status to report recovered errors.  A PER bit of zero indicates that
the target shall not report errors recovered within the limits established by
the error recovery parameters.  If this bit is zero, the DTE bit must also be
set to zero.

A disable transfer on error (DTE) bit of one indicates that the target shall
terminate the data transfer after a recovered read-write error occurs.  All
data from the recovered block shall be transferred prior to terminating the
read-write operation.  A DTE bit of zero indicates that the target shall not
terminate the transfer for errors recovered within the limits established by
the read-write error recovery parameters.

A disable correction (DCR) bit of one indicates that the target shall not use
error correction codes during error recovery.  A DCR bit of zero allows the
use of error correction codes for error recovery.

The read retry count specifies the number of times that the target should
attempt its recovery algorithm during a read operation before an unrecoverable
error is reported.  A read retry count of zero indicates that the target shall
not use its recovery algorithm during read operations.

The write retry count specifies the number of times that the target should
attempt its recovery algorithm during a write operation before an
unrecoverable error is reported.  A write retry count of zero indicates that
the target shall not use its recovery algorithm during write operations.

9.4. Glossary for Sequential Access Devices

















SCSI-2 draft proposed American National Standard        9-57      Revision 10c57







Sequential-Access Devices                                               3/9/90


beginning-of-partition. The position at the beginning of the permissible
recording region of a partition.  If only one partition is defined, this
position is typically equivalent to the beginning-of-medium.

beginning-of-medium. The extreme position along the medium in the direction
away from the supply reel which can be accessed by the device.

buffered mode. A mode of data transfer in write operations which facilitates
tape streaming (9.1.5), as reported in the mode parameter header device-
specific parameter (9.3.3).

early-warning. A physical mark or device computed position near but logically
before the end-of-partition (independent of physical direction).  See the REW
bit in the device configuration page in 9.3.3.1.

end-of-data. End of data in a partition is denoted in format-specific manner.
See the EOD defined field in the device configuration page in 9.3.3.1.

end-of-medium. The extreme position along the medium in the direction away
from the take-up reel which can be accessed by the device.  This position may
be accessed by devices that support the LOAD UNLOAD command with the EOT bit
set to one.

end-of-partition. The position at the end of the permissible recording region
of a partition.

filemark. A special recorded element within a partition, containing no user
data, which provides a segmentation scheme.

overlength. The incorrect length condition that exists after executing a read
command when the the length of the actual block read exceeds the requested
transfer length in the command descriptor block.

partition. The entire usable region of recording and reading paths in a volume
or in a portion of a volume, defined in a device-specific manner.  If there is
more than one partition, they shall be numbered starting with zero (i.e.,
beginning-of-partition-zero).

setmark.  A special recorded element within a partition, containing no user
data, which provides a segmentation scheme hierarchically superior to
filemarks for use in addressing or fast positioning on high capacity storage
devices.

spacing.  The act of positioning the medium on a sequential access device.

underlength. The incorrect length condition that exists after executing a read
command when the requested transfer length in the command descriptor block
exceeds the length of the actual block read.


SCSI-2 draft proposed American National Standard        9-58      Revision 10c58







Sequential-Access Devices                                               3/9/90

volume. A recording medium together with its physical carrier.


















































SCSI-2 draft proposed American National Standard        9-59      Revision 10c59