signals: protect init from unwanted signals more
authorOleg Nesterov <oleg@redhat.com>
Thu, 2 Apr 2009 23:58:02 +0000 (16:58 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 3 Apr 2009 02:04:58 +0000 (19:04 -0700)
(This is a modified version of the patch submitted by Oleg Nesterov
http://lkml.org/lkml/2008/11/18/249 and tries to address comments that
came up in that discussion)

init ignores the SIG_DFL signals but we queue them anyway, including
SIGKILL.  This is mostly OK, the signal will be dropped silently when
dequeued, but the pending SIGKILL has 2 bad implications:

        - it implies fatal_signal_pending(), so we confuse things
          like wait_for_completion_killable/lock_page_killable.

        - for the sub-namespace inits, the pending SIGKILL can
          mask (legacy_queue) the subsequent SIGKILL from the
          parent namespace which must kill cinit reliably.
          (preparation, cinits don't have SIGNAL_UNKILLABLE yet)

The patch can't help when init is ptraced, but ptracing of init is not
"safe" anyway.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Acked-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/signal.c

index 92a1ab00449818a59f85ccee75870c2e67f9a772..8bf7a40e5c71486080151c372b464a67a2cb7878 100644 (file)
@@ -55,10 +55,21 @@ static int sig_handler_ignored(void __user *handler, int sig)
                (handler == SIG_DFL && sig_kernel_ignore(sig));
 }
 
-static int sig_ignored(struct task_struct *t, int sig)
+static int sig_task_ignored(struct task_struct *t, int sig)
 {
        void __user *handler;
 
+       handler = sig_handler(t, sig);
+
+       if (unlikely(t->signal->flags & SIGNAL_UNKILLABLE) &&
+                       handler == SIG_DFL)
+               return 1;
+
+       return sig_handler_ignored(handler, sig);
+}
+
+static int sig_ignored(struct task_struct *t, int sig)
+{
        /*
         * Blocked signals are never ignored, since the
         * signal handler may change by the time it is
@@ -67,8 +78,7 @@ static int sig_ignored(struct task_struct *t, int sig)
        if (sigismember(&t->blocked, sig) || sigismember(&t->real_blocked, sig))
                return 0;
 
-       handler = sig_handler(t, sig);
-       if (!sig_handler_ignored(handler, sig))
+       if (!sig_task_ignored(t, sig))
                return 0;
 
        /*