perf_counter: fix counter freeing logic

Fix counter lifetime bugs which explain the crashes reported by
Marcelo Tosatti and Arnaldo Carvalho de Melo.

The new rule is: flushing + freeing is only done for a task's
own counters, never for other tasks.

[ Impact: fix crashes/lockups with inherited counters ]

Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Reported-by: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

kernel/exit.c

@@ -178,12 +178,6 @@ void release_task(struct task_struct * p)
 
 	proc_flush_task(p);
 
-	/*
-	 * Flush inherited counters to the parent - before the parent
-	 * gets woken up by child-exit notifications.
-	 */
-	perf_counter_exit_task(p);
-
 	write_lock_irq(&tasklist_lock);
 	tracehook_finish_release_task(p);
 	__exit_signal(p);
@@ -985,6 +979,13 @@ NORET_TYPE void do_exit(long code)
 		module_put(tsk->binfmt->module);
 
 	proc_exit_connector(tsk);
+
+	/*
+	 * Flush inherited counters to the parent - before the parent
+	 * gets woken up by child-exit notifications.
+	 */
+	perf_counter_exit_task(tsk);
+
 	exit_notify(tsk, group_dead);
 #ifdef CONFIG_NUMA
 	mpol_put(tsk->mempolicy);
@@ -1257,12 +1258,6 @@ static int wait_task_zombie(struct task_struct *p, int options,
 	 */
 	read_unlock(&tasklist_lock);
 
-	/*
-	 * Flush inherited counters to the parent - before the parent
-	 * gets woken up by child-exit notifications.
-	 */
-	perf_counter_exit_task(p);
-
 	retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0;
 	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
 		? p->signal->group_exit_code : p->exit_code;

kernel/perf_counter.c

@@ -3299,6 +3299,8 @@ void perf_counter_exit_task(struct task_struct *child)
 	struct perf_counter *child_counter, *tmp;
 	struct perf_counter_context *child_ctx;
 
+	WARN_ON_ONCE(child != current);
+
 	child_ctx = &child->perf_counter_ctx;
 
 	if (likely(!child_ctx->nr_counters))