Arch Linux

beretta

Member

Registered: 2008-04-21

Posts: 133

hexdump vs. od

I've been tinkering around with cryptography, and looking for a way to generate a page of random numbers from /dev/random.  Some quick google work turns up a couple of methods, one using hexdump and one using either od or dd.

So I took a look at the man pages and other info about these utilities to see how they were used, and they seem to me to be redundant; they seem to do the exact same thing, just with different syntax.  Does anyone know more in depth how these functions work, and why I would choose to use one over the other?

Procyon

Member

Registered: 2008-05-07

Posts: 1,819

Re: hexdump vs. od

Something like this?

$ od -t u1 /dev/urandom | head -n 10 | cut -d' ' -f2-
250 174  33  71 131 163 182 134 247 205 138 252 103 103  29 165
 80  56 111  89  89  10 247 166 139 207  81 106 201 181 160 109
 47  95 115 243  54 118  92 108  98 223 174 151 203  91  63 100
100 120  45 115  84 202 224 104 225  23 142 123  83  50 215  67
 77  27 133 254  62  38 129 115 144 240  36 170   5  84  12  74
 12  71 253 132 182 152 100 217  10 247   2 188   9 162  98 253
 62 216   2 155  18   1  11 182 215  59 170  25  62  12 175  85
172 170  26 121   8  84 247  21  73  87 203 214 144  49  99  50
154 237 143 160  15 250  95  64 189 217 139  56 167  56  19  37
187 217 208  33 247 190   3 119 191  41 169 182  65 144 186  48

I tried it with hexdump, but it looks complicated.

beretta

Member

Registered: 2008-04-21

Posts: 133

Re: hexdump vs. od

 $ hexdump -d -n 48 /dev/random
0000000   38030   00633   39946   58926   53927   29523   56775   06475
0000010   34335   32605   01201   49477   04577   03462   38569   27701
0000020   32932   54542   05110   38278   60108   20146   04924   00718
0000030

I found in some ways, hexdump is easier to get a direct result, so I don't know that it's more complicated.  Its output is different: the first column catalogs the number of bits used, essentially.  Each 5 digit number is 2 bits, so 16 bits per line (hence 0x10, 0x20, etc., can be supressed with -s)

The equivalent od command, I think, would be

 $ od -t u2 -N 48 /dev/random
0000000 45388 41311 39918 56995  1742  8573 40478 31762
0000020 27172 18537 40670 23827 62650 48597 39033 43433
0000040 26550 62165  1196  9266 56362 64998 46322 62522
0000060

Only difference is the leading zeros.  (The index is octal for od, I think)

I'm not so much interested in how to get the random numbers as I am why I would choose to use one utility over the other.  Both are required by sysvinit and udev, so it'd be near impossible to have a (Arch) linux system without both, so what's the reason for having two different utilities that seem to do pretty much the same thing?  Hopefully someone knows what the difference is.

For example, is there something you can do with od that you can't with hexdump (or vice versa)?

Last edited by beretta (2009-04-06 17:43:33)

Procyon

Member

Registered: 2008-05-07

Posts: 1,819

Re: hexdump vs. od

hexdump has some nice defaults like -C

This is the first time I had heard of od actually. It seems simpler. That -t u1 looks nice. Can you do that in hexdump? With some complex -e '? ? "%_d"' maybe but I couldn't get it to work at all.