1SGM_DD(8) SG3_UTILS SGM_DD(8)
2
6 sgm_dd - copy data to and from files and devices, especially SCSI
7 devices
8
10 sgm_dd [bs=BS] [count=COUNT] [ibs=BS] [if=IFILE] [iflag=FLAGS] [obs=BS]
11 [of=OFILE] [oflag=FLAGS] [seek=SEEK] [skip=SKIP] [--help] [--version]
12 [bpt=BPT] [cdbsz=6|10|12|16] [dio=0|1] [sync=0|1] [time=0|1] [ver‐
14 bose=VERB]
15
17 Copy data to and from any files. Specialized for "files" that are Linux
18 SCSI generic (sg) devices and raw devices. Uses memory mapped transfers
19 on sg devices. Similar syntax and semantics to dd(1) but does not per‐
20 form any conversions.
21 Will only perform memory mapped transfers when IFILE or OFILE are SCSI
23 generic (sg) devices.
24 If both IFILE and OFILE are sg devices then memory mapped transfers are
26 performed on IFILE. If no other flags are specified then indirect IO is
27 performed on OFILE. If 'oflag=dio' is given then direct IO is attempted
28 on OFILE. If 'oflag=smmap' is given then shared mmap-ed IO (sharing the
29 mmap-ed reserve buffer associated with IFILE) is attempted. In both
30 latter cases if the faster IO option is not available, they fall back
31 to indirect IO and report this at the end of the copy.
32 The first group in the synopsis above are "standard" Unix dd(1) oper‐
34 ands. The second group are extra options added by this utility. Both
35 groups are defined below.
36
38 bpt=BPT
39 each IO transaction will be made using BPT blocks (or less if
40 near the end of the copy). Default is 128 for block sizes less
41 that 2048 bytes, otherwise the default is 32. So for bs=512 the
42 reads and writes will each convey 64 KiB of data by default
43 (less if near the end of the transfer or memory restrictions).
44 When cd/dvd drives are accessed, the block size is typically
45 2048 bytes and bpt defaults to 32 which again implies 64 KiB
46 transfers.
47 bs=BS where BS must be the block size of the physical device. Note
49 that this differs from dd(1) which permits BS to be an integral
50 multiple. Default is 512 which is usually correct for disks but
51 incorrect for cdroms (which normally have 2048 byte blocks). For
52 this utility the maximum size of each individual IO operation is
53 BS * BPT bytes.
54 cdbsz=6 | 10 | 12 | 16
56 size of SCSI READ and/or WRITE commands issued on sg device
57 names. Default is 10 byte SCSI command blocks (unless calcula‐
58 tions indicate that a 4 byte block number may be exceeded, in
59 which case it defaults to 16 byte SCSI commands).
60 count=COUNT
62 copy COUNT blocks from IFILE to OFILE. Default is the minimum
63 (of IFILE and OFILE) number of blocks that sg devices report
64 from SCSI READ CAPACITY commands or that block devices (or their
65 partitions) report. Normal files are not probed for their size.
66 If skip=SKIP or skip=SEEK are given and the count is derived
67 (i.e. not explicitly given) then the derived count is scaled
68 back so that the copy will not overrun the device. If the file
69 name is a block device partition and COUNT is not given then the
70 size of the partition rather than the size of the whole device
71 is used. If COUNT is not given and cannot be derived then an
72 error message is issued and no copy takes place.
73 dio=0 | 1
75 permits direct IO to be selected on the write-side (i.e. on
76 OFILE). Only allowed when the read-side (i.e. IFILE) is a sg
77 device. When 1 there may be a "zero copy" copy (i.e. mmap-ed
78 transfer on the read into the user space and direct IO from
79 there on the write, potentially two DMAs and no data copying
80 from the CPU). Default is 0. The same action as 'dio=1' is also
81 available with 'oflag=dio'.
82 ibs=BS if given must be the same as BS given to 'bs=' option.
84 if=IFILE
86 read from IFILE instead of stdin. If IFILE is '-' then stdin is
87 read. Starts reading at the beginning of IFILE unless SKIP is
88 given.
89 iflag=FLAGS
91 where FLAGS is a comma separated list of one or more flags out‐
92 lined below. These flags are associated with IFILE and are
93 ignored when IFILE is stdin.
94 obs=BS if given must be the same as BS given to 'bs=' option.
96 of=OFILE
98 write to OFILE instead of stdout. If OFILE is '-' then writes to
99 stdout. If OFILE is /dev/null then no actual writes are per‐
100 formed. If OFILE is '.' (period) then it is treated the same
101 way as /dev/null (this is a shorthand notation). If OFILE exists
102 then it is _not_ truncated; it is overwritten from the start of
103 OFILE unless 'oflag=append' or SEEK is given.
104 oflag=FLAGS
106 where FLAGS is a comma separated list of one or more flags out‐
107 lined below. These flags are associated with OFILE and are
108 ignored when OFILE is /dev/null, '.' (period), or stdout.
109 seek=SEEK
111 start writing SEEK bs-sized blocks from the start of OFILE.
112 Default is block 0 (i.e. start of file).
113 skip=SKIP
115 start reading SKIP bs-sized blocks from the start of IFILE.
116 Default is block 0 (i.e. start of file).
117 sync=0 | 1
119 when 1, does SYNCHRONIZE CACHE command on OFILE at the end of
120 the transfer. Only active when OFILE is a sg device file name.
121 time=0 | 1
123 when 1, times transfer and does throughput calculation, out‐
124 putting the results (to stderr) at completion. When 0 (default)
125 doesn't perform timing.
126 verbose=VERB
128 as VERB increases so does the amount of debug output sent to
129 stderr. Default value is zero which yields the minimum amount
130 of debug output. A value of 1 reports extra information that is
131 not repetitive. A value 2 reports cdbs and responses for SCSI
132 commands that are not repetitive (i.e. other that READ and
133 WRITE). Error processing is not considered repetitive. Values of
134 3 and 4 yield output for all SCSI commands (and Unix read() and
135 write() calls) so there can be a lot of output.
136 --help outputs usage message and exits.
138 --version
140 outputs version number information and exits.
141
143 Here is a list of flags and their meanings:
144 append causes the O_APPEND flag to be added to the open of OFILE. For
146 normal files this will lead to data appended to the end of any
147 existing data. Cannot be used together with the seek=SEEK
148 option as they conflict. The default action of this utility is
149 to overwrite any existing data from the beginning of the file
150 or, if SEEK is given, starting at block SEEK. Note that attempt‐
151 ing to 'append' to a device file (e.g. a disk) will usually be
152 ignored or may cause an error to be reported.
153 dio is only active with oflag (i.e. 'oflag=dio'). Its action is
155 described in the 'dio=1' option description above.
156 direct causes the O_DIRECT flag to be added to the open of IFILE and/or
158 OFILE. This flag requires some memory alignment on IO. Hence
159 user memory buffers are aligned to the page size. Has no effect
160 on sg, normal or raw files.
161 dpo set the DPO bit (disable page out) in SCSI READ and WRITE com‐
163 mands. Not supported for 6 byte cdb variants of READ and WRITE.
164 Indicates that data is unlikely to be required to stay in device
165 (e.g. disk) cache. May speed media copy and/or cause a media
166 copy to have less impact on other device users.
167 dsync causes the O_SYNC flag to be added to the open of IFILE and/or
169 OFILE. The "d" is prepended to lower confusion with the
170 'sync=0|1' option which has another action (i.e. a synchronisa‐
171 tion to media at the end of the transfer).
172 excl causes the O_EXCL flag to be added to the open of IFILE and/or
174 OFILE.
175 fua causes the FUA (force unit access) bit to be set in SCSI READ
177 and/or WRITE commands. This only has effect with sg devices. The
178 6 byte variants of the SCSI READ and WRITE commands do not sup‐
179 port the FUA bit. Only active for sg device file names.
180 null has no affect, just a placeholder.
182 smmap is only active for oflag. It sets shared mmap IO usage on OFILE
184 if it is a sg device node. The IFILE also needs to be a sg
185 device node (or there is no mmap-ed reserve buffer to share).
186
188 Here are some retired options that are still present:
189 fua=0 | 1 | 2 | 3
191 force unit access bit. When 3, fua is set on both IFILE and
192 OFILE; when 2, fua is set on IFILE; when 1, fua is set on OFILE;
193 when 0 (default), fua is cleared on both. See the 'fua' flag.
194
196 A raw device must be bound to a block device prior to using sgm_dd.
197 See raw(8) for more information about binding raw devices. To be safe,
198 the sg device mapping to SCSI block devices should be checked with 'cat
199 /proc/scsi/scsi' before use.
200 Raw device partition information can often be found with fdisk(8) [the
202 "-ul" argument is useful in this respect].
203 Various numeric arguments (e.g. SKIP) may include multiplicative suf‐
205 fixes or be given in hexadecimal. See the "NUMERIC ARGUMENTS" section
206 in the sg3_utils(8) man page.
207 The count, skip and seek parameters can take 64 bit values (i.e. very
209 big numbers). Other values are limited to what can fit in a signed 32
210 bit number.
211 Data usually gets to the user space in a 2 stage process: first the
213 SCSI adapter DMAs into kernel buffers and then the sg driver copies
214 this data into user memory (write operations reverse this sequence).
215 With memory mapped transfers a kernel buffer reserved by sg is memory
216 mapped (see the mmap(2) system call) into the user space. When this is
217 done the second (redundant) copy from kernel buffers to user space is
218 not needed. Hence the transfer is faster and requires less "grunt" from
219 the CPU.
220 All informative, warning and error output is sent to stderr so that
222 dd's output file can be stdout and remain unpolluted. If no options are
223 given, then the usage message is output and nothing else happens.
224 For sg devices this utility issues SCSI READ and WRITE (SBC) commands
226 which are appropriate for disks and reading from CD/DVD/BD drives.
227 Those commands are not formatted correctly for tape devices so sgm_dd
228 should not be used on tape devices.
229 This utility stops the copy if any error is encountered. For more
231 advanced "copy on error" logic see the sg_dd utility (and its 'coe'
232 flag).
233
235 See the examples given in the man page for sg_dd(8).
236
238 The signal handling has been borrowed from dd: SIGINT, SIGQUIT and SIG‐
239 PIPE output the number of remaining blocks to be transferred and the
240 records in + out counts; then they have their default action. SIGUSR1
241 causes the same information to be output yet the copy continues. All
242 output caused by signals is sent to stderr.
243
245 The exit status of sgm_dd is 0 when it is successful. Otherwise see the
246 sg3_utils(8) man page. Since this utility works at a higher level than
247 individual commands, and there are 'coe' and 'retries' flags, individ‐
248 ual SCSI command failures do not necessary cause the process to exit.
249
251 Written by Douglas Gilbert and Peter Allworth.
252
254 Report bugs to <dgilbert at interlog dot com>.
255
257 Copyright © 2000-2012 Douglas Gilbert
258 This software is distributed under the GPL version 2. There is NO war‐
259 ranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PUR‐
260 POSE.
261
263 The simplest variant of this utility is called sg_dd. A POSIX threads
264 version of this utility called sgp_dd is in the sg3_utils package. The
265 lmbench package contains lmdd which is also interesting. raw(8), dd(1)
266sg3_utils-1.35 November 2019 SGM_DD(8)