1SG_READ(8) SG3_UTILS SG_READ(8)
2
3
4
6 sg_read - read multiple blocks of data, optionally with SCSI READ com‐
7 mands
8
10 sg_read [blk_sgio=0|1] [bpt=BPT] [bs=BS] [cdbsz=6|10|12|16] count=COUNT
11 [dio=0|1] [dpo=0|1] [fua=0|1] if=IFILE [mmap=0|1] [no_dxfer=0|1]
12 [odir=0|1] [skip=SKIP] [time=TI] [verbose=VERB] [--help] [--version]
13
15 Read data from a Linux SCSI generic (sg) device, a block device or a
16 normal file with each read command issued to the same offset or logical
17 block address (lba). This can be used to test (or time) disk caching,
18 SCSI (or some other) transport throughput, and/or SCSI command over‐
19 head.
20
21 When the COUNT value is positive, then up to BPT blocks are read at a
22 time, until the COUNT is exhausted. Each read operation starts at the
23 same lba which, if SKIP is not given, is the beginning of the file or
24 device.
25
26 The COUNT value may be negative when IFILE is a sg device or is a block
27 device with 'blk_sgio=1' set. Alternatively 'bpt=0' may be given. In
28 these cases |COUNT| "zero block" SCSI READ commands are issued. "Zero
29 block" means "do nothing" for SCSI READ 10, 12 and 16 byte commands
30 (but not for the 6 byte variant). In practice "zero block" SCSI READ
31 commands have low latency and so are one way to measure SCSI command
32 overhead.
33
35 blk_sgio=0 | 1
36 The default action of this utility is to use the Unix read()
37 command when the IFILE is a block device. In lk 2.6 many block
38 devices can handle SCSI commands issued via the SG_IO ioctl. So
39 when this option is set the SG_IO ioctl sends SCSI READ commands
40 to IFILE if it is a block device.
41
42 bpt=BPT
43 where BPT is the maximum number of blocks each read operation
44 fetches. Fewer blocks will be fetched when the remaining COUNT
45 is less than BPT. The default value for BPT is 128. Note that
46 each read operation starts at the same lba (as given by
47 skip=SKIP or 0). If 'bpt=0' then the COUNT is interpreted as
48 the number of zero block SCSI READ commands to issue.
49
50 bs=BS where BS is the size (in bytes) of each block read. This must be
51 the block size of the physical device (defaults to 512) if SCSI
52 commands are being issued to IFILE.
53
54 cdbsz=6 | 10 | 12 | 16
55 size of SCSI READ commands issued on sg device names, or block
56 devices if 'blk_sgio=1' is given. Default is 10 byte SCSI READ
57 cdbs.
58
59 count=COUNT
60 when COUNT is a positive number, read that number of blocks,
61 typically with multiple read operations. When COUNT is negative
62 then |COUNT| SCSI READ commands are performed requesting zero
63 blocks to be transferred. This option is mandatory.
64
65 dio=0 | 1
66 default is 0 which selects indirect IO. Value of 1 attempts
67 direct IO which, if not available, falls back to indirect IO and
68 notes this at completion. This option is only active if IFILE is
69 an sg device. If direct IO is selected and
70 /proc/scsi/sg/allow_dio has the value of 0 then a warning is
71 issued (and indirect IO is performed)
72
73 dpo=0 | 1
74 when set the disable page out (DPO) bit in SCSI READ commands is
75 set. Otherwise the DPO bit is cleared (default).
76
77 fua=0 | 1
78 when set the force unit access (FUA) bit in SCSI READ commands
79 is set. Otherwise the FUA bit is cleared (default).
80
81 if=IFILE
82 read from this IFILE. This argument must be given. If the IFILE
83 is a normal file then it must be seekable (if (COUNT > BPT) or
84 skip=SKIP is given). Hence stdin is not acceptable (and giving
85 "-" as the IFILE argument is reported as an error).
86
87 mmap=0 | 1
88 default is 0 which selects indirect IO. Value of 1 causes memory
89 mapped IO to be performed. Selecting both dio and mmap is an
90 error. This option is only active if IFILE is an sg device.
91
92 no_dxfer=0 | 1
93 when set then DMA transfers from the device are made into kernel
94 buffers but no further (i.e. there is no second copy into the
95 user space). The default value is 0 in which case transfers are
96 made into the user space. When neither mmap nor dio is set then
97 data transfer are copied via kernel buffers (i.e. a double
98 copy). Mainly for testing.
99
100 odir=0 | 1
101 when set opens an IFILE which is a block device with an addi‐
102 tional O_DIRECT flag. The default value is 0 (i.e. don't open
103 block devices O_DIRECT).
104
105 skip=SKIP
106 all read operations will start offset by SKIP bs-sized blocks
107 from the start of the input file (or device).
108
109 time=TI
110 When TI is 0 (default) doesn't perform timing. When 1, times
111 transfer and does throughput calculation, starting at the first
112 issued command until completion. When 2, times transfer and does
113 throughput calculation, starting at the second issued command
114 until completion. When 3 times from third command, etc. An aver‐
115 age number of commands (SCSI READs or Unix read()s) executed per
116 second is also output.
117
118 verbose=VERB
119 as VERB increases so does the amount of debug output sent to
120 stderr. Default value is zero which yields the minimum amount
121 of debug output. A value of 1 reports extra information that is
122 not repetitive.
123
124 --help Output the usage message then exit.
125
126 --version
127 Output the version string then exit.
128
130 Various numeric arguments (e.g. SKIP) may include multiplicative suf‐
131 fixes or be given in hexadecimal. See the "NUMERIC ARGUMENTS" section
132 in the sg3_utils(8) man page.
133
134 Data usually gets to the user space in a 2 stage process: first the
135 SCSI adapter DMAs into kernel buffers and then the sg driver copies
136 this data into user memory. This is called "indirect IO" and there is
137 a "dio" option to select "direct IO" which will DMA directly into user
138 memory. Due to some issues "direct IO" is disabled in the sg driver and
139 needs a configuration change to activate it. This is typically done
140 with "echo 1 > /proc/scsi/sg/allow_dio". An alternate way to avoid the
141 2 stage copy is to select memory mapped IO with 'mmap=1'.
142
144 The signal handling has been borrowed from dd: SIGINT, SIGQUIT and SIG‐
145 PIPE output the number of remaining blocks to be transferred; then they
146 have their default action. SIGUSR1 causes the same information to be
147 output yet the copy continues. All output caused by signals is sent to
148 stderr.
149
151 Let us assume that /dev/sg0 is a disk and we wish to time the disk's
152 cache performance.
153
154 sg_read if=/dev/sg0 bs=512 count=1MB mmap=1 time=2
155
156 This command will continually read 128 512 byte blocks from block 0.
157 The "128" is the default value for 'bpt' while "block 0" is chosen
158 because the 'skip' argument was not given. This will continue until
159 1,000,000 blocks are read. The idea behind using 'time=2' is that the
160 first 64 KiB read operation will involve reading the magnetic media
161 while the remaining read operations will "hit" the disk's cache. The
162 output of third command will look like this:
163
164 time from second command to end was 4.50 secs, 113.70 MB/sec
165 Average number of READ commands per second was 1735.27
166 1000000+0 records in, SCSI commands issued: 7813
167
169 The exit status of sg_read is 0 when it is successful. Otherwise see
170 the sg3_utils(8) man page.
171
173 Written by Douglas Gilbert.
174
176 Report bugs to <dgilbert at interlog dot com>.
177
179 Copyright © 2000-2012 Douglas Gilbert
180 This software is distributed under the GPL version 2. There is NO war‐
181 ranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PUR‐
182 POSE.
183
185 To time streaming media read or write time see sg_dd is in the
186 sg3_utils package. The lmbench package contains lmdd which is also
187 interesting. raw(8), dd(1)
188
189
190
191sg3_utils-1.35 November 2012 SG_READ(8)