How to Write Block of Random Data to File using Bash-CodePudding

I need to write a 5MB block of data obtained from /dev/urandom to a partition, at a specific location in the partition. I then need to check that the write executed correctly. I was previously successfully doing this in C , but now I would like to implement it in a bash script.

My C code consisted of:

create 10MB array of data populated from /dev/urandom (RANDOM_ARRAY)
open partition with open()
use lseek() to navigate to desired position in partition
use write() to write the array into the partition
close and reopen partition, use lseek to navigate back to desired position
use read() to read 5MB at this position and populate another array with this data (WRITTEN_ARRAY)
compare each element in (RANDOM_ARRAY) with (WRITTEN_ARRAY)

Im not experienced with writing bash scripts but this is what I've got so far, although it doesn't seem to work:

random_data="$(cat /dev/urandom | head -c<5MB>)"

printf $random_data | dd of=<partition_path> bs=1 seek=<position_in_partition> count=<5MB>

file_data=$(dd if=<partition_path> bs=1 skip=<position_in_partition> count=<5MB>)

if [ "$random_data" == "$file_data" ]
then
 echo "data write successful"
fi

Thanks to the helpful commenters my script now looks something like this:

# get 10MB random data 
head -c<10MB> /dev/urandom > random.bin
# write 10MB random data to partition
dd if=random.bin of=<partition_location>
# copy the written data
dd if=<partition_location> count=<10MB/512 bytes> of=newdata.bin
# compare 
cmp random.bin newdata.bin

At this point cmp returns that the first char is different. Looking at a verbose output of cmp and turns out all values in newdata.bin are 0.

CodePudding user response：

Bash strings cannot hold arbitrary binary data because the ASCII NUL character is used as a string terminator.

One way to do what you want to do is to put the data in files instead of variables and use cmp to compare the files.

Another option is to store cryptographic hashes of the data in Bash variables. This Shellcheck-clean code demonstrates the idea:

#! /bin/bash -p

partition_path=testpart
position_in_partition=10

sha256_1=$( exec 3>&1
            head -c 5MB /dev/urandom                \
                | tee >(sha256sum >&3)              \
                | dd of="$partition_path" bs=1  \
                        seek="$position_in_partition" count=5MB)

sha256_2=$(dd if="$partition_path" bs=1 \
                skip="$position_in_partition" count=5MB \
            | sha256sum)

[[ $sha256_1 == "$sha256_2" ]] && echo 'data write successful'

You'll need to set the partition_path and position_in_partion variables to values that are appropriate for you.
exec 3>&1 connects file descriptor 3 to the stream that is used to read the value of sha256_1.
tee >(sha256sum >&3) uses the standard tee utility and Bash process substitution to copy the pipeline data as input to a sha256sum process whose output is redirected to file descriptor 3. The effect of this is that the sha256sum output (with the trailing newline removed) becomes the value of the sha256_1 variable.
You can use a stronger cryptographic hash function by replacing sha256sum with (for example) sha512sum.

CodePudding user response：

Here's a simpler approach which just saves the data in a temporary file.

#!/bin/sh

set -e

random_data=$(mktemp -t ideone.XXXXXXXX) || exit
trap 'rm -rf "$t"' EXIT

dd if=/dev/urandom bs=10240 count=512 of="$random_data"

dd if="$random_data" of=<partition_path> bs=1 seek=<position_in_partition> count=<5MB>

if dd if=<partition_path> bs=1 skip=<position_in_partition> count=<5MB> |
    cmp "$random_data"
then
 echo "data write successful"
fi