From: "Paul E. McKenney" <paulmck@us.ibm.com>

The following patch fixes a locking issue in -rt signals, where the
job-control signals drop sighand_lock momentarily during signal handling. 
This interacts badly with the RCU-ification of unicast signal delivery. 
The fix is simply to acquire tasklist_lock for job-control signals, since
these typically do not have stringent scalability or latency requirements.

Signed-off-by: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland McGrath <roland@redhat.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---

 kernel/signal.c |    7 +++++++
 1 files changed, 7 insertions(+)

diff -puN kernel/signal.c~fix-sigstop-locking-issue kernel/signal.c
--- devel/kernel/signal.c~fix-sigstop-locking-issue	2005-11-22 22:30:39.000000000 -0800
+++ devel-akpm/kernel/signal.c	2005-11-22 22:30:39.000000000 -0800
@@ -1158,13 +1158,20 @@ int
 kill_proc_info(int sig, struct siginfo *info, pid_t pid)
 {
 	int error;
+	int acquired_tasklist_lock = 0;
 	struct task_struct *p;
 
 	rcu_read_lock();
+	if (unlikely(sig_kernel_stop(sig) || sig == SIGCONT)) {
+		read_lock(&tasklist_lock);
+		acquired_tasklist_lock = 1;
+	}
 	p = find_task_by_pid(pid);
 	error = -ESRCH;
 	if (p)
 		error = group_send_sig_info(sig, info, p);
+	if (unlikely(acquired_tasklist_lock))
+		read_unlock(&tasklist_lock);
 	rcu_read_unlock();
 	return error;
 }
_