Using substring and iterating over the string seems simple enough, and it would work alright.

If you want to use a regex, you should use look ahead assertions if you want to find occurrences that can overlap (and for pos to give you the start of the match). :

perl -E 'for("bababa") { say pos while /(?=baba)/g }'
0
2
[download]

If you don't want to find overlapping occurrences, you can use @- to find the start of the last match.

perl -E 'for("bababa") { say $-[0] while /baba/g }'
0
[download]