signal: give SEND_SIG_FORCED more power to beat SIGNAL_UNKILLABLE

force_sig_info() and friends have the special semantics for synchronous
signals, this interface should not be used if the target is not current.
And it needs the fixes, in particular the clearing of SIGNAL_UNKILLABLE
is not exactly right.

However there are callers which have to use force_ exactly because it
clears SIGNAL_UNKILLABLE and thus it can kill the CLONE_NEWPID tasks,
although this is almost always is wrong by various reasons.

With this patch SEND_SIG_FORCED ignores SIGNAL_UNKILLABLE, like we do if
the signal comes from the ancestor namespace.

This makes the naming in prepare_signal() paths insane, fixed by the
next cleanup.

Note: this only affects SIGKILL/SIGSTOP, but this is enough for
force_sig() abusers.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Anton Vorontsov <anton.vorontsov@linaro.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/kernel/signal.c b/kernel/signal.c
index e76001ccf5cd8b848fdb48306f61e95c11cc0009..2584f5a91fbee3681497bbf086a49eb16e01d645 100644 (file)
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -1059,7 +1059,8 @@ static int __send_signal(int sig, struct siginfo *info, struct task_struct *t,
        assert_spin_locked(&t->sighand->siglock);
 
        result = TRACE_SIGNAL_IGNORED;
-       if (!prepare_signal(sig, t, from_ancestor_ns))
+       if (!prepare_signal(sig, t,
+                       from_ancestor_ns || (info == SEND_SIG_FORCED)))
                goto ret;
 
        pending = group ? &t->signal->shared_pending : &t->pending;