Hi,
I am using busybox in my embedded project. I was interested in knowing how busybox is linking multiple utilities to a single executable.

http://en.wikipedia.org/wiki/BusyBox#Examples

According to the above link. If I call "/bin/ls" then in turn it is being translated to "/bin/busybox ls" (I am not sure whether my understanding is correct.).

Can someone please tell me how busybox is converting "/bin/ls" to "/bin/busybox"

Thanks

There are two components to the answer.
1. For each utility that will be implemented by busybox, a filesystem link from some standard directory such as /bin, /usr/bin, /sbin, etc, will point to the busybox binary.
2. Any application in Linux can look at its argv[0], and from that see what the parent process used to invoke the application. By doing this, busybox is able to know what the parent process was expecting, and is able to provide the accordant behavior.
There are other applications, such as bash which use the same procedure to provide an alter ego, of sorts.

--- rod.

2 members found this post helpful.

Thanks theNbomr...I had forgotten about argv[0]... :P

I'm curious, is the busybox code re-entrant in that many command calls only loads one instance of busybox or if say 2 processes call busybox ls then its loaded twice?

The reason I'm curious is that if loaded twice then isn't that actually quite inefficient as it means that two copies of a "large" program are loaded as apposed to two smaller specialised programs as busy box has to include code for every command as apposed to say ls as a standalone that could be quite small in comparison.

Also wouldn't loading one "large" program take a lot more resources and time than two smaller programs.