linux-kernel - [PATCH 1/1] signals: introduce send

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 10 Jun 2010 03:00:23 +0200
From:	Oleg Nesterov <oleg@...hat.com>
To:	Andrew Morton <akpm@...ux-foundation.org>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	Roland McGrath <roland@...hat.com>
Cc:	"Luis Claudio R. Goncalves" <lclaudio@...g.org>,
	LKML <linux-kernel@...r.kernel.org>,
	linux-mm <linux-mm@...ck.org>,
	David Rientjes <rientjes@...gle.com>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>,
	Nick Piggin <npiggin@...e.de>,
	Minchan Kim <minchan.kim@...il.com>
Subject: [PATCH 1/1] signals: introduce send_sigkill() helper

Cleanup, no functional changes.

There are a lot of buggy SIGKILL users in kernel. For example, almost
every force_sig(SIGKILL) is wrong. force_sig() is not safe, it assumes
that the task has the valid ->sighand, and in general it should be used
only for synchronous signals. send_sig(SIGKILL, p, 1) or
send_xxx(SEND_SIG_FORCED/SEND_SIG_PRIV) is not right too but this is not
immediately obvious.

The only way to correctly send SIGKILL is send_sig_info(SEND_SIG_NOINFO)
but we do not want to use this directly, because we can optimize this
case later. For example, zap_pid_ns_processes() allocates sigqueue for
each process in namespace, this is unneeded.

Introduce the trivial send_sigkill() helper on top of send_sig_info()
and change zap_pid_ns_processes() as an example.

Note: we need more cleanups here, this is only the first change.

Signed-off-by: Oleg Nesterov <oleg@...hat.com>
---

 include/linux/sched.h  |    9 +++++++++
 kernel/pid_namespace.c |    8 +-------
 2 files changed, 10 insertions(+), 7 deletions(-)

--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2053,6 +2053,15 @@ static inline int kill_cad_pid(int sig, 
 #define SEND_SIG_PRIV	((struct siginfo *) 1)
 #define SEND_SIG_FORCED	((struct siginfo *) 2)
 
+static inline int send_sigkill(struct task_struct * p)
+{
+	/*
+	 * Kills any user-space task, even SIGNAL_UNKILLABLE.
+	 * We use SEND_SIG_NOINFO to make si_fromuser() true.
+	 */
+	return send_sig_info(SIGKILL, SEND_SIG_NOINFO, p);
+}
+
 /*
  * True if we are on the alternate signal stack.
  */
--- a/kernel/pid_namespace.c
+++ b/kernel/pid_namespace.c
@@ -160,15 +160,9 @@ void zap_pid_ns_processes(struct pid_nam
 	nr = next_pidmap(pid_ns, 1);
 	while (nr > 0) {
 		rcu_read_lock();
-
-		/*
-		 * Any nested-container's init processes won't ignore the
-		 * SEND_SIG_NOINFO signal, see send_signal()->si_fromuser().
-		 */
 		task = pid_task(find_vpid(nr), PIDTYPE_PID);
 		if (task)
-			send_sig_info(SIGKILL, SEND_SIG_NOINFO, task);
-
+			send_sigkill(task);
 		rcu_read_unlock();
 
 		nr = next_pidmap(pid_ns, nr);

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/