blob: 51ea7f9d3af50c08b3bdee3da39cc3dae4c609a0 [file] [log] [blame]
.\" Copyright 1995 Robert K. Nichols (
.\" Copyright 1999-2005 Kai Mäkisara (
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date. The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein. The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.TH ST 4 2019-03-06 "Linux" "Linux Programmer's Manual"
st \- SCSI tape device
.B #include <sys/mtio.h>
.BI "int ioctl(int " fd ", int " request " [, (void *)" arg3 "]);"
.BI "int ioctl(int " fd ", MTIOCTOP, (struct mtop *)" mt_cmd );
.BI "int ioctl(int " fd ", MTIOCGET, (struct mtget *)" mt_status );
.BI "int ioctl(int " fd ", MTIOCPOS, (struct mtpos *)" mt_pos );
.B st
driver provides the interface to a variety of SCSI tape devices.
Currently, the driver takes control of all detected devices of type
.B st
driver uses major device number 9.
Each device uses eight minor device numbers.
The lowermost five bits
in the minor numbers are assigned sequentially in the order of
In the 2.6 kernel, the bits above the eight lowermost bits are
concatenated to the five lowermost bits to form the tape number.
The minor numbers can be grouped into
two sets of four numbers: the principal (auto-rewind) minor device numbers,
.IR n ,
and the \(lqno-rewind\(rq device numbers,
.RI ( n " + 128)."
Devices opened using the principal device number will be sent a
command when they are closed.
Devices opened using the \(lqno-rewind\(rq device number will not.
(Note that using an auto-rewind device for positioning the tape with,
for instance, mt does not lead to the desired result: the tape is
rewound after the mt command and the next command starts from the
beginning of the tape).
Within each group, four minor numbers are available to define
devices with different characteristics (block size, compression,
density, etc.)
When the system starts up, only the first device is available.
The other three are activated when the default
characteristics are defined (see below).
(By changing compile-time
constants, it is possible to change the balance between the maximum
number of tape drives and the number of minor numbers for each
The default allocation allows control of 32 tape drives.
For instance, it is possible to control up to 64 tape drives
with two minor numbers for different options.)
Devices are typically created by:
.in +4n
mknod \-m 666 /dev/st0 c 9 0
mknod \-m 666 /dev/st0l c 9 32
mknod \-m 666 /dev/st0m c 9 64
mknod \-m 666 /dev/st0a c 9 96
mknod \-m 666 /dev/nst0 c 9 128
mknod \-m 666 /dev/nst0l c 9 160
mknod \-m 666 /dev/nst0m c 9 192
mknod \-m 666 /dev/nst0a c 9 224
There is no corresponding block device.
The driver uses an internal buffer that has to be large enough to hold
at least one tape block.
In kernels before 2.1.121, the buffer is
allocated as one contiguous block.
This limits the block size to the
largest contiguous block of memory the kernel allocator can provide.
The limit is currently 128\ kB for 32-bit architectures and
256\ kB for 64-bit architectures.
In newer kernels the driver
allocates the buffer in several parts if necessary.
By default, the
maximum number of parts is 16.
This means that the maximum block size
is very large (2\ MB if allocation of 16 blocks of 128\ kB succeeds).
The driver's internal buffer size is determined by a compile-time
constant which can be overridden with a kernel startup option.
In addition to this, the driver tries to allocate a larger temporary
buffer at run time if necessary.
However, run-time allocation of large
contiguous blocks of memory may fail and it is advisable not to rely
too much on dynamic buffer allocation with kernels older than 2.1.121
(this applies also to demand-loading the driver with kerneld or kmod).
The driver does not specifically support any tape drive brand or
After system start-up the tape device options are defined by
the drive firmware.
For example, if the drive firmware selects fixed-block mode,
the tape device uses fixed-block mode.
The options can
be changed with explicit
.BR ioctl (2)
calls and remain in effect when the device is closed and reopened.
Setting the options affects both the auto-rewind and the nonrewind
Different options can be specified for the different devices within
the subgroup of four.
The options take effect when the device is
For example, the system administrator can define
one device that writes in fixed-block mode with a certain block size,
and one which writes in variable-block mode (if the drive supports
both modes).
The driver supports
.B tape partitions
if they are supported by the drive.
(Note that the tape partitions
have nothing to do with disk partitions.
A partitioned tape can be
seen as several logical tapes within one medium.)
Partition support has to be enabled with an
.BR ioctl (2).
The tape
location is preserved within each partition across partition changes.
The partition used for subsequent tape operations is
selected with an
.BR ioctl (2).
The partition switch is executed together with
the next tape operation in order to avoid unnecessary tape
The maximum number of partitions on a tape is defined by a
compile-time constant (originally four).
The driver contains an
.BR ioctl (2)
that can format a tape with either one or two partitions.
.I /dev/tape
is usually created as a hard or soft link to the default tape device
on the system.
Starting from kernel 2.6.2, the driver exports in the sysfs directory
.I /sys/class/scsi_tape
the attached devices and some parameters assigned to the devices.
.SS Data transfer
The driver supports operation in both fixed-block mode and
variable-block mode (if supported by the drive).
In fixed-block mode the drive
writes blocks of the specified size and the block size is not
dependent on the byte counts of the write system calls.
In variable-block mode one tape block is written for each write call
and the byte
count determines the size of the corresponding tape block.
Note that
the blocks on the tape don't contain any information about the
writing mode: when reading, the only important thing is to use
commands that accept the block sizes on the tape.
In variable-block mode the read byte count does not have to match
the tape block size exactly.
If the byte count is larger than the
next block on tape, the driver returns the data and the function
returns the actual block size.
If the block size is larger than the
byte count, an error is returned.
In fixed-block mode the read byte counts can be arbitrary if
buffering is enabled, or a multiple of the tape block size if
buffering is disabled.
Kernels before 2.1.121 allow writes with
arbitrary byte count if buffering is enabled.
In all other cases
(kernel before 2.1.121 with buffering disabled or newer kernel) the
write byte count must be a multiple of the tape block size.
In the 2.6 kernel, the driver tries to use direct transfers between the user
buffer and the device.
If this is not possible, the driver's internal buffer
is used.
The reasons for not using direct transfers include improper alignment
of the user buffer (default is 512 bytes but this can be changed by the HBA
driver), one or more pages of the user buffer not reachable by the
SCSI adapter, and so on.
A filemark is automatically written to tape if the last tape operation
before close was a write.
When a filemark is encountered while reading, the following
If there are data remaining in the buffer when the filemark
is found, the buffered data is returned.
The next read returns zero
The following read returns data from the next file.
The end of
recorded data is signaled by returning zero bytes for two consecutive
read calls.
The third read returns an error.
.SS Ioctls
The driver supports three
.BR ioctl (2)
Requests not recognized by the
.B st
driver are passed to the
The definitions below are from
.IR /usr/include/linux/mtio.h :
.SS MTIOCTOP \(em perform a tape operation
This request takes an argument of type
.IR "(struct mtop\ *)" .
Not all drives support all operations.
The driver returns an
error if the drive rejects an operation.
.in +4n
/* Structure for MTIOCTOP \- mag tape op command: */
struct mtop {
short mt_op; /* operations defined below */
int mt_count; /* how many of them */
Magnetic Tape operations for normal tape use:
.TP 14
Backward space over
.I mt_count
Backward space over
.I mt_count
Reposition the tape to the EOT side of the last filemark.
Backward space over
.I mt_count
records (tape blocks).
Backward space over
.I mt_count
Enable compression of tape data within the drive if
.I mt_count
is nonzero and disable compression if
.I mt_count
is zero.
This command uses the MODE page 15 supported by most DATs.
Go to the end of the recorded media (for appending files).
Erase tape.
With 2.6 kernel, short erase (mark tape empty) is performed if the
argument is zero.
Otherwise, long erase (erase all) is done.
Forward space over
.I mt_count
Forward space over
.I mt_count
Reposition the tape to the BOT side of the last filemark.
Forward space over
.I mt_count
records (tape blocks).
Forward space over
.I mt_count
Execute the SCSI load command.
A special case is available for some HP
.I mt_count
is the constant
plus a number, the number is
sent to the drive to control the autoloader.
Lock the tape drive door.
Format the tape into one or two partitions.
.I mt_count
is positive, it gives the size of partition 1 and partition
0 contains the rest of the tape.
.I mt_count
is zero, the tape is formatted into one partition.
From kernel version 4.6,
.\" commit 8038e6456a3e6f5c4759e0d73c4f9165b90c93e7
a negative
.I mt_count
specifies the size of partition 0 and
the rest of the tape contains partition 1.
The physical ordering of partitions depends on the drive.
This command is not allowed for a drive unless the partition support
is enabled for the drive (see
No op\(emflushes the driver's buffer as a side effect.
Should be used before reading status with
Rewind and put the drive off line.
Reset drive.
Re-tension tape.
Seek to the tape block number specified in
.IR mt_count .
This operation requires either a SCSI-2 drive that supports the
command (device-specific address)
or a Tandberg-compatible SCSI-1 drive (Tandberg, Archive
Viper, Wangtek, ...).
The block number should be one that was previously returned by
if device-specific addresses are used.
Set the drive's block length to the value specified in
.IR mt_count .
A block length of zero sets the drive to variable block size mode.
Set the tape density to the code in
.IR mt_count .
The density codes supported by a drive can be found from the drive
The active partition is switched to
.IR mt_count .
The partitions are numbered from zero.
This command is not allowed for
a drive unless the partition support is enabled for the drive (see
Execute the SCSI unload command (does not eject the tape).
Unlock the tape drive door.
.I mt_count
.I mt_count
Magnetic Tape operations for setting of device options (by the superuser):
.TP 8
Set various drive and driver options according to bits encoded in
.IR mt_count .
These consist of the drive's buffering mode, a set of Boolean driver
options, the buffer write threshold, defaults for the block size and
density, and timeouts (only in kernels 2.1 and later).
A single operation can affect only one item in the list above (the
Booleans counted as one item.)
A value having zeros in the high-order 4 bits will be used to set the
drive's buffering mode.
The buffering modes are:
.RS 12
.IP 0 4
The drive will not report
status on write commands until the data
blocks are actually written to the medium.
.IP 1
The drive may report
status on write commands as soon as all the
data has been transferred to the drive's internal buffer.
.IP 2
The drive may report
status on write commands as soon as (a) all
the data has been transferred to the drive's internal buffer, and
(b) all buffered data from different initiators has been successfully
written to the medium.
To control the write threshold the value in
.I mt_count
must include the constant
bitwise ORed with a block count in the low 28 bits.
The block count refers to 1024-byte blocks, not the physical block
size on the tape.
The threshold cannot exceed the driver's internal buffer size (see
To set and clear the Boolean options
the value in
.I mt_count
must include one of the constants
bitwise ORed with
whatever combination of the following options is desired.
the options can be set to the values
defined in the corresponding bits.
the options can be selectively set and with
selectively cleared.
.IP ""
The default options for a tape device are set with
A nonactive tape device (e.g., device with
minor 32 or 160) is activated when the default options for it are
defined the first time.
An activated device inherits from the device
activated at start-up the options not set explicitly.
.IP ""
The Boolean options are:
.BR MT_ST_BUFFER_WRITES " (Default: true)"
Buffer all write operations in fixed-block mode.
If this option is false and the drive uses a fixed block size, then
all write operations must be for a multiple of the block size.
This option must be set false to write reliable multivolume archives.
.BR MT_ST_ASYNC_WRITES " (Default: true)"
When this option is true, write operations return immediately without
waiting for the data to be transferred to the drive if the data fits
into the driver's buffer.
The write threshold determines how full the buffer must be before a
new SCSI write command is issued.
Any errors reported by the drive will be held until the next
This option must be set false to write reliable multivolume archives.
.BR MT_ST_READ_AHEAD " (Default: true)"
This option causes the driver to provide read buffering and
read-ahead in fixed-block mode.
If this option is false and the drive uses a fixed block size, then
all read operations must be for a multiple of the block size.
.BR MT_ST_TWO_FM " (Default: false)"
This option modifies the driver behavior when a file is closed.
The normal action is to write a single filemark.
If the option is true, the driver will write two filemarks and
backspace over the second one.
This option should not be set true for QIC tape drives since they are
unable to overwrite a filemark.
These drives detect the end of recorded data by testing for blank tape
rather than two consecutive filemarks.
Most other current drives also
detect the end of recorded data and using two filemarks is usually
necessary only when interchanging tapes with some other systems.
.BR MT_ST_DEBUGGING " (Default: false)"
This option turns on various debugging messages from the driver
(effective only if the driver was compiled with
defined nonzero).
.BR MT_ST_FAST_EOM " (Default: false)"
This option causes the
operation to be sent directly to the
drive, potentially speeding up the operation but causing the driver to
lose track of the current file number normally returned by the
is false, the driver will respond to an
request by forward spacing over files.
.BR MT_ST_AUTO_LOCK " (Default: false)"
When this option is true, the drive door is locked when the device is
opened and unlocked when it is closed.
.BR MT_ST_DEF_WRITES " (Default: false)"
The tape options (block size, mode, compression, etc.) may change
when changing from one device linked to a drive to another device
linked to the same drive depending on how the devices are
This option defines when the changes are enforced by the
driver using SCSI-commands and when the drives auto-detection
capabilities are relied upon.
If this option is false, the driver
sends the SCSI-commands immediately when the device is changed.
If the
option is true, the SCSI-commands are not sent until a write is
In this case, the drive firmware is allowed to detect the
tape structure when reading and the SCSI-commands are used only to
make sure that a tape is written according to the correct specification.
.BR MT_ST_CAN_BSR " (Default: false)"
When read-ahead is used, the tape must sometimes be spaced backward to the
correct position when the device is closed and the SCSI command to
space backward over records is used for this purpose.
Some older
drives can't process this command reliably and this option can be used
to instruct the driver not to use the command.
The end result is that,
with read-ahead and fixed-block mode, the tape may not be correctly
positioned within a file when the device is closed.
With 2.6 kernel, the
default is true for drives supporting SCSI-3.
.BR MT_ST_NO_BLKLIMS " (Default: false)"
Some drives don't accept the
SCSI command.
If this is used, the driver does not use the command.
The drawback is
that the driver can't check before sending commands if the selected
block size is acceptable to the drive.
.BR MT_ST_CAN_PARTITIONS " (Default: false)"
This option enables support for several partitions within a
The option applies to all devices linked to a drive.
.BR MT_ST_SCSI2LOGICAL " (Default: false)"
This option instructs the driver to use the logical block addresses
defined in the SCSI-2 standard when performing the seek and tell
operations (both with
commands and when changing tape
Otherwise, the device-specific addresses are used.
It is highly advisable to set this option if the drive supports the
logical addresses because they count also filemarks.
There are some
drives that support only the logical block addresses.
.BR MT_ST_SYSV " (Default: false)"
When this option is enabled, the tape devices use the SystemV
Otherwise, the BSD semantics are used.
The most important
difference between the semantics is what happens when a device used
for reading is closed: in System V semantics the tape is spaced forward
past the next filemark if this has not happened while using the
In BSD semantics the tape position is not changed.
.BR MT_NO_WAIT " (Default: false)"
Enables immediate mode (i.e., don't wait for the command to finish) for some
commands (e.g., rewind).
An example:
.in +4n
struct mtop mt_cmd;
mt_cmd.mt_op = MTSETDRVBUFFER;
mt_cmd.mt_count = MT_ST_BOOLEANS |
ioctl(fd, MTIOCTOP, mt_cmd);
.IP ""
The default block size for a device can be set with
and the default density code can be set with
The values for the parameters are or'ed
with the operation code.
.IP ""
With kernels 2.1.x and later, the timeout values can be set with the
ORed with the timeout in seconds.
The long timeout (used for rewinds and other commands
that may take a long time) can be set with
The kernel defaults are very long to
make sure that a successful command is not timed out with any
Because of this, the driver may seem stuck even if it is only
waiting for the timeout.
These commands can be used to set more
practical values for a specific drive.
The timeouts set for one device
apply for all devices linked to the same drive.
.IP ""
Starting from kernels 2.4.19 and 2.5.43, the driver supports a status
bit which indicates whether the drive requests cleaning.
The method used by the
drive to return cleaning information is set using the
If the value is zero, the cleaning
bit is always zero.
If the value is one, the TapeAlert data defined
in the SCSI-3 standard is used (not yet implemented).
Values 2\(en17 are
If the lowest eight bits are >= 18, bits from the extended
sense data are used.
The bits 9\(en16 specify a mask to select the bits
to look at and the bits 17\(en23 specify the bit pattern to look for.
If the bit pattern is zero, one or more bits under the mask indicate
the cleaning request.
If the pattern is nonzero, the pattern must match
the masked sense data byte.
.SS MTIOCGET \(em get status
This request takes an argument of type
.IR "(struct mtget\ *)" .
.in +4n
/* structure for MTIOCGET \- mag tape get status command */
struct mtget {
long mt_type;
long mt_resid;
/* the following registers are device dependent */
long mt_dsreg;
long mt_gstat;
long mt_erreg;
/* The next two fields are not always used */
daddr_t mt_fileno;
daddr_t mt_blkno;
.IP \fImt_type\fP 11
The header file defines many values for
.IR mt_type ,
but the current driver reports only the generic types
(Generic SCSI-1 tape)
(Generic SCSI-2 tape).
.IP \fImt_resid\fP
contains the current tape partition number.
.IP \fImt_dsreg\fP
reports the drive's current settings for block size (in the low 24
bits) and density (in the high 8 bits).
These fields are defined by
.IP \fImt_gstat\fP
reports generic (device independent) status information.
The header file defines macros for testing these status bits:
.HP 4
The tape is positioned just after a filemark
(always false after an
The tape is positioned at the beginning of the first file (always false
after an
A tape operation has reached the physical End Of Tape.
The tape is currently positioned at a setmark
(always false after an
The tape is positioned at the end of recorded data.
The drive is write-protected.
For some drives this can also mean that the drive does not support
writing on the current medium type.
The last
.BR open (2)
found the drive with a tape in place and ready for operation.
\fBGMT_D_6250\fP(\fIx\fP), \fBGMT_D_1600\fP(\fIx\fP), \fBGMT_D_800\fP(\fIx\fP):
This \(lqgeneric\(rq status information reports the current
density setting for 9-track \(12" tape drives only.
The drive does not have a tape in place.
Immediate report mode.
This bit is set if there are no guarantees that
the data has been physically written to the tape when the write call
It is set zero only when the driver does not buffer data and
the drive is set not to buffer data.
The drive has requested cleaning.
Implemented in kernels since 2.4.19 and 2.5.43.
.IP \fImt_erreg\fP
The only field defined in
.I mt_erreg
is the recovered error count in the low 16 bits (as defined by
Due to inconsistencies in the way drives report recovered errors, this
count is often not maintained (most drives do not by default report
soft errors but this can be changed with a SCSI MODE SELECT command).
.IP \fImt_fileno\fP
reports the current file number (zero-based).
This value is set to \-1 when the file number is unknown (e.g., after
.IP \fImt_blkno\fP
reports the block number (zero-based) within the current file.
This value is set to \-1 when the block number is unknown (e.g., after
.SS MTIOCPOS \(em get tape position
This request takes an argument of type
.I "(struct mtpos\ *)"
and reports the drive's notion of the current tape block number,
which is not the same as
.I mt_blkno
returned by
This drive must be a SCSI-2 drive that supports the
command (device-specific address)
or a Tandberg-compatible SCSI-1 drive (Tandberg, Archive
Viper, Wangtek, ... ).
.in +4n
/* structure for MTIOCPOS \- mag tape get position command */
struct mtpos {
long mt_blkno; /* current block number */
.TP 14
An attempt was made to write or erase a write-protected tape.
(This error is not detected during
.BR open (2).)
The device is already in use or the driver was unable to allocate a
The command parameters point to memory not belonging to the calling
.BR ioctl (2)
had an invalid argument, or a requested block size was invalid.
The requested operation could not be completed.
The byte count in
.BR read (2)
is smaller than the next physical block on the tape.
(Before 2.2.18 and 2.4.0 the extra bytes have been
.\" Precisely: Linux 2.6.0-test6
silently ignored.)
A write operation could not be completed because the tape reached
.BR ioctl (2).
During opening, the tape device does not exist.
An attempt was made to read or write a variable-length block that is
larger than the driver's internal buffer.
Open is attempted with
when the tape in the drive is write-protected.
.I /dev/st*
the auto-rewind SCSI tape devices
.I /dev/nst*
the nonrewind SCSI tape devices
.\" The driver has been written by Kai M\(:akisara (
.\" starting from a driver written by Dwayne Forsyth.
.\" Several other
.\" people have also contributed to the driver.
.IP 1. 4
When exchanging data between systems, both systems have to agree on
the physical tape block size.
The parameters of a drive after startup
are often not the ones most operating systems use with these
Most systems use drives in variable-block mode if the drive
supports that mode.
This applies to most modern drives, including
DATs, 8mm helical scan drives, DLTs, etc.
It may be advisable to use
these drives in variable-block mode also in Linux (i.e., use
at system startup to set the mode), at least when
exchanging data with a foreign system.
The drawback of
this is that a fairly large tape block size has to be used to get
acceptable data transfer rates on the SCSI bus.
.IP 2.
Many programs (e.g.,
.BR tar (1))
allow the user to specify the blocking
factor on the command line.
Note that this determines the physical block
size on tape only in variable-block mode.
.IP 3.
In order to use SCSI tape drives, the basic SCSI driver,
a SCSI-adapter driver and the SCSI tape driver must be either
configured into the kernel or loaded as modules.
If the SCSI-tape
driver is not present, the drive is recognized but the tape support
described in this page is not available.
.IP 4.
The driver writes error messages to the console/log.
codes written into some messages are automatically translated to text
if verbose SCSI messages are enabled in kernel configuration.
.IP 5.
The driver's internal buffering allows good throughput in fixed-block
mode also with small
.BR read (2)
.BR write (2)
byte counts.
With direct transfers
this is not possible and may cause a surprise when moving to the 2.6
The solution is to tell the software to use larger transfers (often
telling it to use larger blocks).
If this is not possible, direct transfers can be disabled.
.BR mt (1)
The file
.I drivers/scsi/
.I Documentation/scsi/st.txt
(kernel >= 2.6) in the Linux kernel source tree contains
the most recent information about the driver and its configuration