preload: Reinitialize mutexes after forking

[vanvugt]

Or else there's a risk the child process has inherited mutexes that are already locked, resulting in a deadlock as soon as it tries to use them.

https://bugs.launchpad.net/bugs/2144723

[martinpitt]

Thanks @vanvugt ! I'm not familiar with that glibc on_load() constructor API magic. I'll do some RTFM, but also ask the robodude -- please don't take it as bad faith.

[martinpitt]

Ah nevermind, of course on_load() isn't magic at all. It's just a random name for the constructor. Makes sense!

[Copilot]

Pull request overview

This PR addresses a fork-safety issue in the libumockdev preload library by ensuring mutex state in the child process cannot remain inherited/locked, which could otherwise deadlock the child when it next uses the preload interposition logic (per Launchpad #2144723).

Changes:

• Add a pthread_atfork() child handler to reset internal mutexes after fork().
• Register the at-fork handler via a constructor (on_load) when the preload library is loaded.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[martinpitt]

See https://manpages.debian.org/testing/glibc-doc/pthread_atfork.3.en.html

[martinpitt]

Thank you! I'll do a release now.