1ST(4) Linux Programmer's Manual ST(4)
2
6 st - SCSI tape device
7
9 #include <sys/mtio.h>
10 int ioctl(int fd, int request [, (void *)arg3]);
12 int ioctl(int fd, MTIOCTOP, (struct mtop *)mt_cmd);
13 int ioctl(int fd, MTIOCGET, (struct mtget *)mt_status);
14 int ioctl(int fd, MTIOCPOS, (struct mtpos *)mt_pos);
15
17 The st driver provides the interface to a variety of SCSI tape devices.
18 Currently, the driver takes control of all detected devices of type
19 “sequential-access.” The st driver uses major device number 9.
20 Each device uses eight minor device numbers. The lowermost five bits in
22 the minor numbers are assigned sequentially in the order of detection.
23 In the 2.6 kernel, the bits above the eight lowermost bits are concate‐
24 nated to the five lowermost bits to form the tape number. The minor
25 numbers can be grouped into two sets of four numbers: the principal
26 (auto-rewind) minor device numbers, n, and the “no-rewind” device num‐
27 bers, (n+ 128). Devices opened using the principal device number will
28 be sent a REWIND command when they are closed. Devices opened using
29 the “no-rewind” device number will not. (Note that using an auto-
30 rewind device for positioning the tape with, for instance, mt does not
31 lead to the desired result: the tape is rewound after the mt command
32 and the next command starts from the beginning of the tape).
33 Within each group, four minor numbers are available to define devices
35 with different characteristics (block size, compression, density, etc.)
36 When the system starts up, only the first device is available. The
37 other three are activated when the default characteristics are defined
38 (see below). (By changing compile-time constants, it is possible to
39 change the balance between the maximum number of tape drives and the
40 number of minor numbers for each drive. The default allocation allows
41 control of 32 tape drives. For instance, it is possible to control up
42 to 64 tape drives with two minor numbers for different options.)
43 Devices are typically created by:
45 mknod -m 666 /dev/st0 c 9 0
46 mknod -m 666 /dev/st0l c 9 32
47 mknod -m 666 /dev/st0m c 9 64
48 mknod -m 666 /dev/st0a c 9 96
49 mknod -m 666 /dev/nst0 c 9 128
50 mknod -m 666 /dev/nst0l c 9 160
51 mknod -m 666 /dev/nst0m c 9 192
52 mknod -m 666 /dev/nst0a c 9 224
53 There is no corresponding block device.
55 The driver uses an internal buffer that has to be large enough to hold
57 at least one tape block. In kernels before 2.1.121, the buffer is allo‐
58 cated as one contiguous block. This limits the block size to the
59 largest contiguous block of memory the kernel allocator can provide.
60 The limit is currently 128 kB for 32-bit architectures and 256 kB for
61 64-bit architectures. In newer kernels the driver allocates the buffer
62 in several parts if necessary. By default, the maximum number of parts
63 is 16. This means that the maximum block size is very large (2 MB if
64 allocation of 16 blocks of 128 kB succeeds).
65 The driver's internal buffer size is determined by a compile-time con‐
67 stant which can be overridden with a kernel startup option. In addi‐
68 tion to this, the driver tries to allocate a larger temporary buffer at
69 run-time if necessary. However, run-time allocation of large contiguous
70 blocks of memory may fail and it is advisable not to rely too much on
71 dynamic buffer allocation with kernels older than 2.1.121 (this applies
72 also to demand-loading the driver with kerneld or kmod).
73 The driver does not specifically support any tape drive brand or model.
75 After system start-up the tape device options are defined by the drive
76 firmware. For example, if the drive firmware selects fixed-block mode,
77 the tape device uses fixed-block mode. The options can be changed with
78 explicit ioctl() calls and remain in effect when the device is closed
79 and reopened. Setting the options affects both the auto-rewind and the
80 non-rewind device.
81 Different options can be specified for the different devices within the
83 subgroup of four. The options take effect when the device is opened.
84 For example, the system administrator can define one device that writes
85 in fixed-block mode with a certain block size, and one which writes in
86 variable-block mode (if the drive supports both modes).
87 The driver supports tape partitions if they are supported by the drive.
89 (Note that the tape partitions have nothing to do with disk partitions.
90 A partitioned tape can be seen as several logical tapes within one
91 medium.) Partition support has to be enabled with an ioctl(). The tape
92 location is preserved within each partition across partition changes.
93 The partition used for subsequent tape operations is selected with an
94 ioctl(). The partition switch is executed together with the next tape
95 operation in order to avoid unnecessary tape movement. The maximum num‐
96 ber of partitions on a tape is defined by a compile-time constant
97 (originally four). The driver contains an ioctl() that can format a
98 tape with either one or two partitions.
99 Device /dev/tape is usually created as a hard or soft link to the
101 default tape device on the system.
102 Starting from kernel 2.6.2, the driver exports in the sysfs directory
104 /sys/class/scsi_tape the attached devices and some parameters assigned
105 to the devices.
106
108 The driver supports operation in both fixed-block mode and variable-
109 block mode (if supported by the drive). In fixed-block mode the drive
110 writes blocks of the specified size and the block size is not dependent
111 on the byte counts of the write system calls. In variable-block mode
112 one tape block is written for each write call and the byte count deter‐
113 mines the size of the corresponding tape block. Note that the blocks on
114 the tape don't contain any information about the writing mode: when
115 reading, the only important thing is to use commands that accept the
116 block sizes on the tape.
117 In variable-block mode the read byte count does not have to match the
119 tape block size exactly. If the byte count is larger than the next
120 block on tape, the driver returns the data and the function returns the
121 actual block size. If the block size is larger than the byte count, the
122 requested amount of data from the start of the block is returned and
123 the rest of the block is discarded.
124 In fixed-block mode the read byte counts can be arbitrary if buffering
126 is enabled, or a multiple of the tape block size if buffering is dis‐
127 abled. Kernels before 2.1.121 allow writes with arbitrary byte count if
128 buffering is enabled. In all other cases (kernel before 2.1.121 with
129 buffering disabled or newer kernel) the write byte count must be a mul‐
130 tiple of the tape block size.
131 In the 2.6 kernel, the driver tries to use direct transfers between the
133 user buffer and the device. If this is not possible, the driver's
134 internal buffer is used. The reasons for not using direct transfers
135 include improper alignment of the user buffer (default is 512 bytes but
136 this can be changed by the HBA driver), one of more pages of the user
137 buffer not reachable by the SCSI adapter, etc.
138 A filemark is automatically written to tape if the last tape operation
140 before close was a write.
141 When a filemark is encountered while reading, the following happens. If
143 there are data remaining in the buffer when the filemark is found, the
144 buffered data is returned. The next read returns zero bytes. The fol‐
145 lowing read returns data from the next file. The end of recorded data
146 is signaled by returning zero bytes for two consecutive read calls. The
147 third read returns an error.
148
150 The driver supports three ioctl() requests. Requests not recognized by
151 the st driver are passed to the SCSI driver. The definitions below are
152 from /usr/include/linux/mtio.h:
153 MTIOCTOP — Perform a tape operation
155 This request takes an argument of type (struct mtop *). Not all drives
156 support all operations. The driver returns an EIO error if the drive
157 rejects an operation.
158 /* Structure for MTIOCTOP - mag tape op command: */
160 struct mtop {
161 short mt_op; /* operations defined below */
162 int mt_count; /* how many of them */
163 };
164 Magnetic Tape operations for normal tape use:
166 MTBSF Backward space over mt_count filemarks.
167 MTBSFM Backward space over mt_count filemarks. Reposition the
168 tape to the EOT side of the last filemark.
169 MTBSR Backward space over mt_count records (tape blocks).
170 MTBSS Backward space over mt_count setmarks.
171 MTCOMPRESSION Enable compression of tape data within the drive if
172 mt_count is non-zero and disable compression if mt_count
173 is zero. This command uses the MODE page 15 supported by
174 most DATs.
175 MTEOM Go to the end of the recorded media (for appending
176 files).
177 MTERASE Erase tape. With 2.6 kernel, short erase (mark tape
178 empty) is performed if the argument is zero. Otherwise
179 long erase (erase all) is done.
180 MTFSF Forward space over mt_count filemarks.
181 MTFSFM Forward space over mt_count filemarks. Reposition the
182 tape to the BOT side of the last filemark.
183 MTFSR Forward space over mt_count records (tape blocks).
184 MTFSS Forward space over mt_count setmarks.
185 MTLOAD Execute the SCSI load command. A special case is avail‐
186 able for some HP autoloaders. If mt_count is the constant
187 MT_ST_HPLOADER_OFFSET plus a number, the number is sent
188 to the drive to control the autoloader.
189 MTLOCK Lock the tape drive door.
190 MTMKPART Format the tape into one or two partitions. If mt_count
191 is non-zero, it gives the size of the first partition and
192 the second partition contains the rest of the tape. If
193 mt_count is zero, the tape is formatted into one parti‐
194 tion. This command is not allowed for a drive unless the
195 partition support is enabled for the drive (see
196 MT_ST_CAN_PARTITIONS below).
197 MTNOP No op — flushes the driver's buffer as a side effect.
198 Should be used before reading status with MTIOCGET.
199 MTOFFL Rewind and put the drive off line.
200 MTRESET Reset drive.
201 MTRETEN Re-tension tape.
202 MTREW Rewind.
203 MTSEEK Seek to the tape block number specified in mt_count.
204 This operation requires either a SCSI-2 drive that sup‐
205 ports the LOCATE command (device-specific address) or a
206 Tandberg-compatible SCSI-1 drive (Tandberg, Archive
207 Viper, Wangtek, ... ). The block number should be one
208 that was previously returned by MTIOCPOS if device-spe‐
209 cific addresses are used.
210 MTSETBLK Set the drive's block length to the value specified in
211 mt_count. A block length of zero sets the drive to vari‐
212 able block size mode.
213 MTSETDENSITY Set the tape density to the code in mt_count. The den‐
214 sity codes supported by a drive can be found from the
215 drive documentation.
216 MTSETPART The active partition is switched to mt_count. The parti‐
217 tions are numbered from zero. This command is not allowed
218 for a drive unless the partition support is enabled for
219 the drive (see MT_ST_CAN_PARTITIONS below).
220 MTUNLOAD Execute the SCSI unload command (does not eject the
221 tape).
222 MTUNLOCK Unlock the tape drive door.
223 MTWEOF Write mt_count filemarks.
224 MTWSM Write mt_count setmarks.
225 Magnetic Tape operations for setting of device options (by the supe‐
227 ruser):
228 MTSETDRVBUFFER
229 Set various drive and driver options according to bits encoded
230 in mt_count. These consist of the drive's buffering mode, a
231 set of Boolean driver options, the buffer write threshold,
232 defaults for the block size and density, and timeouts (only in
233 kernels >= 2.1). A single operation can affect only one item
234 in the list above (the Booleans counted as one item.)
235 A value having zeros in the high-order 4 bits will be used to
237 set the drive's buffering mode. The buffering modes are:
238 0 The drive will not report GOOD status on write commands
240 until the data blocks are actually written to the
241 medium.
242 1 The drive may report GOOD status on write commands as
243 soon as all the data has been transferred to the
244 drive's internal buffer.
245 2 The drive may report GOOD status on write commands as
246 soon as (a) all the data has been transferred to the
247 drive's internal buffer, and (b) all buffered data from
248 different initiators has been successfully written to
249 the medium.
250 To control the write threshold the value in mt_count must
252 include the constant MT_ST_WRITE_THRESHOLD logically ORed with
253 a block count in the low 28 bits. The block count refers to
254 1024-byte blocks, not the physical block size on the tape. The
255 threshold cannot exceed the driver's internal buffer size (see
256 DESCRIPTION, above).
257 To set and clear the Boolean options the value in mt_count must
259 include one of the constants MT_ST_BOOLEANS, MT_ST_SETBOOLEANS,
260 MT_ST_CLEARBOOLEANS, or MT_ST_DEFBOOLEANS logically or'ed with
261 whatever combination of the following options is desired.
262 Using MT_ST_BOOLEANS the options can be set to the values
263 defined in the corresponding bits. With MT_ST_SETBOOLEANS the
264 options can be selectively set and with MT_ST_DEFBOOLEANS
265 selectively cleared.
266 The default options for a tape device are set with MT_ST_DEF‐
268 BOOLEANS. A non-active tape device (e.g., device with minor 32
269 or 160) is activated when the default options for it are
270 defined the first time. An activated device inherits from the
271 device activated at start-up the options not set explicitly.
272 The Boolean options are:
274 MT_ST_BUFFER_WRITES (Default: true)
276 Buffer all write operations in fixed-block mode. If
277 this option is false and the drive uses a fixed block
278 size, then all write operations must be for a multiple
279 of the block size. This option must be set false to
280 write reliable multi-volume archives.
281 MT_ST_ASYNC_WRITES (Default: true)
282 When this option is true, write operations return imme‐
283 diately without waiting for the data to be transferred
284 to the drive if the data fits into the driver's buffer.
285 The write threshold determines how full the buffer must
286 be before a new SCSI write command is issued. Any
287 errors reported by the drive will be held until the next
288 operation. This option must be set false to write reli‐
289 able multi-volume archives.
290 MT_ST_READ_AHEAD (Default: true)
291 This option causes the driver to provide read buffering
292 and read-ahead in fixed-block mode. If this option is
293 false and the drive uses a fixed block size, then all
294 read operations must be for a multiple of the block
295 size.
296 MT_ST_TWO_FM (Default: false)
297 This option modifies the driver behavior when a file is
298 closed. The normal action is to write a single file‐
299 mark. If the option is true the driver will write two
300 filemarks and backspace over the second one.
301 Note: This option should not be set true for QIC tape
303 drives since they are unable to overwrite a filemark.
304 These drives detect the end of recorded data by testing
305 for blank tape rather than two consecutive filemarks.
306 Most other current drives also detect the end of
307 recorded data and using two filemarks is usually neces‐
308 sary only when interchanging tapes with some other sys‐
309 tems.
310 MT_ST_DEBUGGING (Default: false)
312 This option turns on various debugging messages from the
313 driver (effective only if the driver was compiled with
314 DEBUG defined non-zero).
315 MT_ST_FAST_EOM (Default: false)
316 This option causes the MTEOM operation to be sent
317 directly to the drive, potentially speeding up the oper‐
318 ation but causing the driver to lose track of the cur‐
319 rent file number normally returned by the MTIOCGET
320 request. If MT_ST_FAST_EOM is false the driver will
321 respond to an MTEOM request by forward spacing over
322 files.
323 MT_ST_AUTO_LOCK (Default: false)
324 When this option is true, the drive door is locked when
325 the device is opened and unlocked when it is closed.
326 MT_ST_DEF_WRITES (Default: false)
327 The tape options (block size, mode, compression, etc.)
328 may change when changing from one device linked to a
329 drive to another device linked to the same drive depend‐
330 ing on how the devices are defined. This option defines
331 when the changes are enforced by the driver using SCSI-
332 commands and when the drives auto-detection capabilities
333 are relied upon. If this option is false, the driver
334 sends the SCSI-commands immediately when the device is
335 changed. If the option is true, the SCSI-commands are
336 not sent until a write is requested. In this case the
337 drive firmware is allowed to detect the tape structure
338 when reading and the SCSI-commands are used only to make
339 sure that a tape is written according to the correct
340 specification.
341 MT_ST_CAN_BSR (Default: false)
342 When read-ahead is used, the tape must sometimes be
343 spaced backward to the correct position when the device
344 is closed and the SCSI command to space backwards over
345 records is used for this purpose. Some older drives
346 can't process this command reliably and this option can
347 be used to instruct the driver not to use the command.
348 The end result is that, with read-ahead and fixed-block
349 mode, the tape may not be correctly positioned within a
350 file when the device is closed. With 2.6 kernel, the
351 default is true for drives supporting SCSI-3.
352 MT_ST_NO_BLKLIMS (Default: false)
353 Some drives don't accept the READ BLOCK LIMITS SCSI com‐
354 mand. If this is used, the driver does not use the com‐
355 mand. The drawback is that the driver can't check before
356 sending commands if the selected block size is accept‐
357 able to the drive.
358 MT_ST_CAN_PARTITIONS (Default: false)
359 This option enables support for several partitions
360 within a tape. The option applies to all devices linked
361 to a drive.
362 MT_ST_SCSI2LOGICAL (Default: false)
363 This option instructs the driver to use the logical
364 block addresses defined in the SCSI-2 standard when per‐
365 forming the seek and tell operations (both with MTSEEK
366 and MTIOCPOS commands and when changing tape partition).
367 Otherwise the device-specific addresses are used. It is
368 highly advisable to set this option if the drive sup‐
369 ports the logical addresses because they count also
370 filemarks. There are some drives that only support the
371 logical block addresses.
372 MT_ST_SYSV (Default: false)
373 When this option is enabled, the tape devices use the
374 SystemV semantics. Otherwise the BSD semantics are used.
375 The most important difference between the semantics is
376 what happens when a device used for reading is closed:
377 in System V semantics the tape is spaced forward past
378 the next filemark if this has not happened while using
379 the device. In BSD semantics the tape position is not
380 changed.
381 MT_NO_WAIT (Default: false)
382 Enables immediate mode (i.e., don't wait for the command
383 to finish) for some commands (e.g., rewind).
384 EXAMPLE
385 struct mtop mt_cmd;
386 mt_cmd.mt_op = MTSETDRVBUFFER;
387 mt_cmd.mt_count = MT_ST_BOOLEANS |
388 MT_ST_BUFFER_WRITES |
389 MT_ST_ASYNC_WRITES;
390 ioctl(fd, MTIOCTOP, &mt_cmd);
391 The default block size for a device can be set with
393 MT_ST_DEF_BLKSIZE and the default density code can be set with
394 MT_ST_DEFDENSITY. The values for the parameters are or'ed with
395 the operation code.
396 With kernels 2.1.x and later, the timeout values can be set
398 with the subcommand MT_ST_SET_TIMEOUT ORed with the timeout in
399 seconds. The long timeout (used for rewinds and other commands
400 that may take a long time) can be set with MT_ST_SET_LONG_TIME‐
401 OUT. The kernel defaults are very long to make sure that a suc‐
402 cessful command is not timed out with any drive. Because of
403 this the driver may seem stuck even if it is only waiting for
404 the timeout. These commands can be used to set more practical
405 values for a specific drive. The timeouts set for one device
406 apply for all devices linked to the same drive.
407 Starting from kernels 2.4.19 and 2.5.43, the driver supports a
409 status bit which indicates whether the drive requests cleaning.
410 The method used by the drive to return cleaning information is
411 set using the MT_ST_SEL_CLN subcommand. If the value is zero,
412 the cleaning bit is always zero. If the value is one, the
413 TapeAlert data defined in the SCSI-3 standard is used (not yet
414 implemented). Values 2-17 are reserved. If the lowest eight
415 bits are >= 18, bits from the extended sense data are used. The
416 bits 9-16 specify a mask to select the bits to look at and the
417 bits 17-23 specify the bit pattern to look for. If the bit
418 pattern is zero, one or more bits under the mask indicate the
419 cleaning request. If the pattern is non-zero, the pattern must
420 match the masked sense data byte.
421 MTIOCGET — Get status
423 This request takes an argument of type (struct mtget *).
424 /* structure for MTIOCGET - mag tape get status command */
426 struct mtget {
427 long mt_type;
428 long mt_resid;
429 /* the following registers are device dependent */
430 long mt_dsreg;
431 long mt_gstat;
432 long mt_erreg;
433 /* The next two fields are not always used */
434 daddr_t mt_fileno;
435 daddr_t mt_blkno;
436 };
437 mt_type The header file defines many values for mt_type, but the
439 current driver reports only the generic types MT_ISSCSI1
440 (Generic SCSI-1 tape) and MT_ISSCSI2 (Generic SCSI-2 tape).
441 mt_resid contains the current tape partition number.
442 mt_dsreg reports the drive's current settings for block size (in the
443 low 24 bits) and density (in the high 8 bits). These fields
444 are defined by MT_ST_BLKSIZE_SHIFT, MT_ST_BLKSIZE_MASK,
445 MT_ST_DENSITY_SHIFT, and MT_ST_DENSITY_MASK.
446 mt_gstat reports generic (device independent) status information.
447 The header file defines macros for testing these status
448 bits:
449 GMT_EOF(x): The tape is positioned just after a filemark
450 (always false after an MTSEEK operation).
451 GMT_BOT(x): The tape is positioned at the beginning of the
452 first file (always false after an MTSEEK operation).
453 GMT_EOT(x): A tape operation has reached the physical End Of
454 Tape.
455 GMT_SM(x): The tape is currently positioned at a setmark
456 (always false after an MTSEEK operation).
457 GMT_EOD(x): The tape is positioned at the end of recorded
458 data.
459 GMT_WR_PROT(x): The drive is write-protected. For some
460 drives this can also mean that the drive does not sup‐
461 port writing on the current medium type.
462 GMT_ONLINE(x): The last open() found the drive with a tape
463 in place and ready for operation.
464 GMT_D_6250(x), GMT_D_1600(x), GMT_D_800(x): This “generic”
465 status information reports the current density setting
466 for 9-track ½" tape drives only.
467 GMT_DR_OPEN(x): The drive does not have a tape in place.
468 GMT_IM_REP_EN(x): Immediate report mode. This bit is set if
469 there are no guarantees that the data has been physi‐
470 cally written to the tape when the write call returns.
471 It is set zero only when the driver does not buffer data
472 and the drive is set not to buffer data.
473 GMT_CLN(x): The drive has requested cleaning. Implemented in
474 kernels >= 2.4.19 and 2.5.43.
475 mt_erreg The only field defined in mt_erreg is the recovered error
476 count in the low 16 bits (as defined by MT_ST_SOFTERR_SHIFT
477 and MT_ST_SOFTERR_MASK). Due to inconsistencies in the way
478 drives report recovered errors, this count is often not
479 maintained (most drives do not by default report soft errors
480 but this can be changed with a SCSI MODE SELECT command).
481 mt_fileno reports the current file number (zero-based). This value is
482 set to -1 when the file number is unknown (e.g., after MTBSS
483 or MTSEEK).
484 mt_blkno reports the block number (zero-based) within the current
485 file. This value is set to -1 when the block number is
486 unknown (e.g., after MTBSF, MTBSS, or MTSEEK).
487 MTIOCPOS — Get tape position
489 This request takes an argument of type (struct mtpos *) and reports the
490 drive's notion of the current tape block number, which is not the same
491 as mt_blkno returned by MTIOCGET. This drive must be a SCSI-2 drive
492 that supports the READ POSITION command (device-specific address) or a
493 Tandberg-compatible SCSI-1 drive (Tandberg, Archive Viper, Wangtek, ...
494 ).
495 /* structure for MTIOCPOS - mag tape get position command */
497 struct mtpos {
498 long mt_blkno; /* current block number */
499 };
500
503 EIO The requested operation could not be completed.
504 ENOSPC A write operation could not be completed because the tape
506 reached end-of-medium.
507 ENOMEM The byte count in read() is smaller than the next physi‐
509 cal block on the tape. (Before 2.2.18 and 2.4.0-test6 the
510 extra bytes have been silently ignored.)
511 EACCES An attempt was made to write or erase a write-protected
513 tape. (This error is not detected during open().)
514 EFAULT The command parameters point to memory not belonging to
516 the calling process.
517 ENXIO During opening, the tape device does not exist.
519 EBUSY The device is already in use or the driver was unable to
521 allocate a buffer.
522 EOVERFLOW An attempt was made to read or write a variable-length
524 block that is larger than the driver's internal buffer.
525 EINVAL An ioctl() had an illegal argument, or a requested block
527 size was illegal.
528 ENOSYS Unknown ioctl().
530 EROFS Open is attempted with O_WRONLY or O_RDWR when the tape
532 in the drive is write-protected.
533
535 /dev/st* : the auto-rewind SCSI tape devices
536 /dev/nst* : the non-rewind SCSI tape devices
537
539 The driver has been written by Kai Mäkisara (Kai.Makisara@metla.fi)
540 starting from a driver written by Dwayne Forsyth. Several other people
541 have also contributed to the driver.
542
544 mt(1)
545 The file README.st or st.txt (kernel >= 2.6) in the kernel sources con‐
547 tains the most recent information about the driver and its configura‐
548 tion possibilities.
549
551 1. When exchanging data between systems, both systems have to agree on
552 the physical tape block size. The parameters of a drive after startup
553 are often not the ones most operating systems use with these devices.
554 Most systems use drives in variable-block mode if the drive supports
555 that mode. This applies to most modern drives, including DATs, 8mm
556 helical scan drives, DLTs, etc. It may be advisable to use these drives
557 in variable-block mode also in Linux (i.e., use MTSETBLK or MTSETDEFBLK
558 at system startup to set the mode), at least when exchanging data with
559 a foreign system. The drawback of this is that a fairly large tape
560 block size has to be used to get acceptable data transfer rates on the
561 SCSI bus.
562 2. Many programs (e.g., tar) allow the user to specify the blocking
564 factor on the command line. Note that this determines the physical
565 block size on tape only in variable-block mode.
566 3. In order to use SCSI tape drives, the basic SCSI driver, a SCSI-
568 adapter driver and the SCSI tape driver must be either configured into
569 the kernel or loaded as modules. If the SCSI-tape driver is not
570 present, the drive is recognized but the tape support described in this
571 page is not available.
572 4. The driver writes error messages to the console/log. The SENSE codes
574 written into some messages are automatically translated to text if ver‐
575 bose SCSI messages are enabled in kernel configuration.
576 5. The driver's internal buffering allows good throughput in fixed-
578 block mode also with small read() and write() byte counts. With direct
579 transfers this is not possible and may cause a surprise when moving to
580 the 2.6 kernel. The solution is to tell the software to use larger
581 transfers (often telling it to use larger blocks). If this is not pos‐
582 sible, direct transfers can be disabled.
583
585 Copyright © 1995 Robert K. Nichols.
586 Copyright © 1999-2005 Kai Mäkisara.
587 Permission is granted to make and distribute verbatim copies of this
589 manual provided the copyright notice and this permission notice are
590 preserved on all copies. Additional permissions are contained in the
591 header of the source file.
592Linux 2.0 - 2.6 2005-03-13 ST(4)