With OCaml 5 (not 4) my program deadlocks in caml_acquire_runtime_system (aka caml_leave_blocking_section). The example below demonstrates what I mean: test.ml contains: Printf.eprintf "hello from OCaml\n%!" test-ocaml-5.c contains: #include <stdio.h> #include <stdlib.h> #include <caml/misc.h> …

Indeed. caml_c_thread_register always defaults to registering the given thread in domain 0. This was a pragmatic backwards compatible choice. Until your example, we haven’t had the need for registering C threads with other domains. This needs a little bit of work as we really don’t have a user-progr…

This patch series (see bottom of the page) is the solution I came up with, and it appears to work: https://listman.redhat.com/archives/libguestfs/2023-June/thread.html#31879 The final code is upstream here: nbdkit/nbdkit/-/blob/master/plugins/ocaml/plugin.c …

There is an additional problem, but it only happens sometimes and only with OCaml < 5. That is that after fork (called from C, not OCaml), something in the OCaml runtime gets very confused and hangs when entering the thread. I guess that caml_atfork_handler is not being run to reinitialize the thr…

[image] rwmjones: The Libguestfs June 2023 Archive by thread It might be useful to make an issue for this on the OCaml issue tracker so that we can investigate and fix any issues if we find them.

Do you mean this issue? [Libguestfs] nbdkit | Failed pipeline for master | 9d4b87e0 To be honest although I am able to reproduce the problem easily, I’m not at all sure that it is related to fork, that might merely be a coincidence. Edit: Let me see if I can make a standalone reproducer first.

The only thing you can do formally in the child process after forking in a multi-threaded program is to apply async-signal-safe functions and then exec. Could your issue be related to that?

I can reproduce this in a program without threads, but using fork. NB. This only reproduces with OCaml 4, not OCaml 5. #include <stdio.h> #include <stdlib.h> #include <string.h> #include <unistd.h> #include <sys/wait.h> #include <caml/callback.h> #include <caml/memory.h> #include <caml/misc.h> #in…

Actually here’s an even simpler version: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <sys/wait.h> #include <caml/callback.h> #include <caml/misc.h> #include <caml/threads.h> int main (int argc, char *argv[]) { pid_t pid; fprintf (stderr, "startup & release runtime syst…

I have to use threads.cmxa because the real program needs to call caml_c_thread_register.

Fatal error: Fatal error during lock: Resource deadlock avoided

Learning

cvine June 22, 2023, 2:08pm 15

In your last example, if you don’t link in threads.cmxa it seems to work OK. I’ve not tested your other one.

Topic		Replies	Views
Test caml_state and conditionally caml_acquire_runtime_system - good or bad? Learning multicore , threading	9	602	June 27, 2023
Signal handling checkpoints and Mutex.lock Learning	20	777	June 24, 2024
Trying to understand `caml_acquire_runtime_system` when called from C threads Learning	7	344	May 26, 2025
A tutorial on parallel programming in OCaml 5 Learning multicore	34	8674	December 21, 2022
OCaml 5.0 and C interface Community	6	990	November 25, 2022

Fatal error: Fatal error during lock: Resource deadlock avoided

Related topics