[systemd-devel] [PATCH] core: send sigabrt on watchdog timeout

[Lennart Poettering]

On Sat, 25.10.14 23:58, Umut Tezduyar Lindskog (umut.tezduyar at axis.com) wrote:

> if sigabrt doesn't do the job, follow regular shutdown
> routine, sigterm > sigkill.

I like it, with one exception.

> 
> diff --git a/src/core/unit.c b/src/core/unit.c
> index e40e6f2..66804c9 100644
> --- a/src/core/unit.c
> +++ b/src/core/unit.c
> @@ -3305,6 +3305,7 @@ int unit_kill_context(
>                  Unit *u,
>                  KillContext *c,
>                  bool sigkill,
> +                bool sigabrt,

Hmm, I think it would be better to replace these two bool's with a
single enum. Maybe like this:

typedef enum KillOperation = {
        KILL_TERMINATE,
        KILL_KILL,
        KILL_ABORT,
} KillOperation;

Where KILL_TERMINATE would use c->kill_signal, while KILL_KILL would
use SIGKILL, and KILL_ABORT would use SIGABRT?

Of course, KILL_KILL is an awful name, but I guess that's Unix, we
just inherit that from the weirdly named "kill(pid, SIGKILL)"...

Makes sense?

Lennart

-- 
Lennart Poettering, Red Hat