I'm learning how Linux processes and pipes work. I understand that a fork bomb consumes process limits by repeatedly creating more processes, and I'm wondering whether repeatedly connecting a process's output back to its own input could create a similar resource-exhaustion effect. Would a self-referential pipe merely keep one process running, or could it cause additional processes, threads, memory, or file descriptors to grow? I'm interested in the underlying mechanism rather than trying this on a real system, and I'd especially like to understand why a short shell pipeline would or would not behave explosively.
3 Answers
If your goal is to learn safely, inspect concepts such as pipe buffers, blocking reads and writes, process creation, thread limits, and per-user resource limits. You can experiment with bounded input and operating-system limits inside a disposable container or virtual machine, then clean it up, rather than testing an unbounded feedback loop on your main Linux installation.
A pipe by itself doesn’t create new processes or duplicate work. It’s just a kernel-managed data channel between processes. Connecting a command’s output back to its input would generally create a feedback loop involving the same small set of processes, so it would more likely block, consume memory or I/O, or run indefinitely rather than multiply the process count. The explosive behavior of a fork bomb comes specifically from repeatedly creating additional processes.
A shell pipeline normally starts a separate process for each pipeline stage, but the shell does not keep creating new stages just because data flows through the pipe. A recursive function that invokes another shell could increase the process count, yet that is essentially a less direct form of process spawning. Without explicit process creation, self-piping is not equivalent to a fork bomb.
The practical failure mode depends on buffering and scheduling: the loop might stall when a pipe buffer fills, keep producing data, or exhaust another limit such as memory or open file descriptors. Those are different resource limits from the process-table exhaustion caused by a fork bomb.
Parallel recursion or repeatedly launching subshells could consume more resources, but that changes the example into a process- or thread-spawning program rather than a pipe-only effect. It’s safer to study this with diagrams, a disposable virtual machine, and strict resource limits instead of running an unbounded command.