Is there a known recent Linux locking bug that affects the OCaml runtime?

Ecosystem
21

Finally some good news!

The glibc fixes have been committed upstream, see commit hashes listed here:
https://sourceware.org/bugzilla/show_bug.cgi?id=25847#c72

They have also been backported to the glibc package that will be in Fedora 42:
https://bugzilla.redhat.com/show_activity.cgi?id=1889892
https://src.fedoraproject.org/rpms/glibc/c/f7aa0c142598fe50cbaa3430cecda7e96ce7027f?branch=rawhide
Fixed In Version glibc-2.40.9000-30.fc42

(note that upstream’s 2.40.9000 doesn’t include these yet, just Fedora’s package)

22

Please don’t ban me for linking to a talk at a C++ convention but I thought this would be of general interest to us elite language enjoyers who have been haunted by this bug.

Malte Skarupke gave a presentation on how he diagnosed and (provably!) fixed this bug. He insists mere mortals can learn some basic TLA+ to solve their own excruciatingly rare pthread condition variable corruption issues in standard glibc infrastructure.

I appreciate the talk and the closure!