Skip to content
Tutorials Center
My learning

Topics

Linux Command Reference ยท Files and Directories

Linux mkfifo command

Reference for the Linux mkfifo command: create named pipes (FIFOs), connect two processes through a file name, understand blocking, and avoid deadlocks.

mkfifo creates a named pipe, also called a FIFO (first in, first out). It looks like a file in a directory, but it holds no data on disk: bytes written into it by one process come out of it, in order, for another process. It is the same mechanism as the | in a pipeline, except that the two processes need not be started by the same command line, and they find each other by file name.

The important rule: opening a FIFO blocks until the other end is open too. A writer waits for a reader, and a reader waits for a writer. Most surprises come from forgetting this.

Terminal
mkfifo [OPTION]... NAME...

Try it

Bash Two processes talking through a FIFO

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.

Bash Mode, a producer and filter, and cleanup

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.

Syntax

Terminal
mkfifo [-m MODE] NAME...
Option Description
-m MODE, --mode=MODE Set permissions (octal or symbolic) instead of 0666 minus the umask.
-Z Set the SELinux context to the default type (where supported).

Returns: 0 if every FIFO was created; non-zero otherwise ("File exists", "Permission denied", or "No such file or directory" for a missing directory).

Usage notes

  • Blocking open: echo hi > fifo waits until some process opens the FIFO for reading. Run the writer in the background (&) or in a second terminal, or both ends will wait forever.
  • Data lives only in memory, in a kernel buffer (64 KiB by default on Linux). A writer that outruns the reader is paused until space frees up. Nothing is stored in the file system, so the size shows as 0.
  • End of file: a reader sees EOF when the last writer closes the pipe. A shell loop while read ... done < fifo therefore ends after the first writer finishes. To keep reading across writers, reopen the FIFO in a loop.
  • A reader and writer in one shell can deadlock. The classic trick to avoid blocking is to open it read-write with exec 3<>fifo, which succeeds without a partner.
  • Permissions work like for files: use mode 600 for private channels. Anyone who can write to the FIFO can inject data into your reader.
  • Remove it when done. A FIFO stays in the directory until you rm it. A script that creates one should clean up with a trap, see Traps and signals.
  • Compared with a regular pipe: use | when one command line starts both sides; use a FIFO when processes start independently, or to build non-linear flows (one producer, several consumers each with its own FIFO).
  • Related tools: process substitution <(cmd) creates a temporary FIFO or /dev/fd path for you; tee can feed several FIFOs from one stream.
  • No seeking: a FIFO is a stream, so tools that need to jump around in their input (or read it twice) cannot work on it.
  • Portability: mkfifo is POSIX. Named pipes work on Linux, BSD and macOS but not on every network file system.