smp: Allocate smp_call_on_cpu() workqueue on stack too

The SMP IPI struct descriptor is allocated on the stack except for the
workqueue and lockdep complains:

  INFO: trying to register non-static key.
  the code is fine but needs lockdep annotation.
  turning off the locking correctness validator.
  CPU: 0 PID: 110 Comm: kworker/0:1 Not tainted 4.8.0-rc5+ #14
  Hardware name: Dell Inc. Precision T3600/0PTTT9, BIOS A13 05/11/2014
  Workqueue: events smp_call_on_cpu_callback
  ...
  Call Trace:
    dump_stack
    register_lock_class
    ? __lock_acquire
    __lock_acquire
    ? __lock_acquire
    lock_acquire
    ? process_one_work
    process_one_work
    ? process_one_work
    worker_thread
    ? process_one_work
    ? process_one_work
    kthread
    ? kthread_create_on_node
    ret_from_fork

So allocate it on the stack too.

Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
[ Test and write commit message. ]
Signed-off-by: Borislav Petkov <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>

Author: Peter Zijlstra

kernel/smp.c

@@ -759,13 +759,14 @@ static void smp_call_on_cpu_callback(struct work_struct *work)
 int smp_call_on_cpu(unsigned int cpu, int (*func)(void *), void *par, bool phys)
 {
 	struct smp_call_on_cpu_struct sscs = {
-		.work = __WORK_INITIALIZER(sscs.work, smp_call_on_cpu_callback),
 		.done = COMPLETION_INITIALIZER_ONSTACK(sscs.done),
 		.func = func,
 		.data = par,
 		.cpu  = phys ? cpu : -1,
 	};
 
+	INIT_WORK_ONSTACK(&sscs.work, smp_call_on_cpu_callback);
+
 	if (cpu >= nr_cpu_ids || !cpu_online(cpu))
 		return -ENXIO;