[Bug 216648] man 5 core incorrectly states circumstances in which a core is not produced

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=216648

--- Comment #2 from Luca Versari (veluca93@gmail.com) ---
> Do you have any experiments demonstrating the actual behavior? Could you show
> them?
On a system with systemd (or where /proc/sys/kernel/core_pattern is configured
to pipe to a program), run `ulimit -c 0`, then run a program that segfaults
(writing `int main() { *(volatile int *)0; }` in a `segfault.c` file and
running `gcc segfault.c; ./a.out` should be sufficient). This creates a core
dump, despite RLIMIT_CORE being 0.

> And do you suggest any specific patch for that page?
Something like adding ", unless the system is configured to pipe core dumps to
a program" in that bullet point should be sufficient.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.