[v5.4-rhel] Fix: Use SIGKILL instead of SIGTERM when ExecStopContainer timeout is 0

Aligns behavior with documentation stating SIGKILL should be sent immediately if the timeout is zero.

Fixes: https://issues.redhat.com/browse/RHEL-96916
Fixes: https://issues.redhat.com/browse/RHEL-96917

Signed-off-by: Jan Rodák <[email protected]>
(cherry picked from commit 077649f)

Author: Honny1

libpod/oci_conmon_exec_common.go

@@ -250,7 +250,7 @@ func (r *ConmonOCIRuntime) ExecStopContainer(ctr *Container, sessionID string, t
 
 	// SIGTERM did not work. On to SIGKILL.
 	logrus.Debugf("Killing exec session %s (PID %d) of container %s with SIGKILL", sessionID, pid, ctr.ID())
-	if err := unix.Kill(pid, unix.SIGTERM); err != nil {
+	if err := unix.Kill(pid, unix.SIGKILL); err != nil {
 		if err == unix.ESRCH {
 			return nil
 		}