This blog post is going to cover a little trick I developed recently that might be useful to others. Let’s say you have a hexdump of a binary. For example, you create a hexdump using xxd that looks like this:
00000000: 6000 1c00 008e 0800 2222 1221 e000 0000 `......."".!....
00000010: 008e 0063 004e 57e0 0052 5201 0000 0000 ...c.NW..RR.....
00000020: c000 00e4 3112 fff7 02e7 3102 1600 12ff ....1.....1.....
00000030: f615 fff6 16ff f7c0 30d7 2005 6601 6000 ........0. .f.`.
00000040: 2100 0000 0000 0000 0000 0000 0000 0000 !...............
How does one convert data in this format back to binary? xxd -r -p doesn’t handle the address column well in many cases, instead interpretting it as hex data into the final binary output. For example, running xxd -r -p over the above content results in a binary in this format (hexdumping the output from xxd -r -p):
00000000: 0000 0000 6000 1c00 008e 0800 2222 1221 ....`......."".!
00000010: e000 0000 0000 0010 008e 0063 004e 57e0 ...........c.NW.
00000020: 0052 5201 0000 0000 0000 0020 c000 00e4 .RR........ ....
00000030: 3112 fff7 02e7 3102 1600 12ff 0000 0030 1.....1........0
00000040: f615 fff6 16ff f7c0 30d7 2005 6601 6000 ........0. .f.`.
00000050: 0000 0040 2100 0000 0000 0000 0000 0000 ...@!...........
00000060: 0000 0000 ....
I looked extensively at tools that can convert a hexdump back to binary, and I could not find one that worked well. Then I realized I could easily do this with awk and xxd. Here is a one liner to convert hexdump back to binary:
$ awk -F' ' '{print $2$3$4$5$6$7$8$9}' input.txt | xxd -r -p > output.bin
The awk command splits each line on the space character as a delimiter. Then it outputs the bytes columnns from input.txt. Lastly, we pipe the output to xxd -r -p which can now easily convert the hex-only output back to binary! The output is redirected then to output.bin.
Jared Brees (https://twitter.com/mejaredbrees) alerted me to a shorter one-liner that also gets the job done:
$ cut -d' ' -f2-9 input.txt | xxd -r -p > output.bin
Thanks Jared!