sparc64: Call xcall_deliver() directly in some cases.

For these cases the callers make sure:

1) The cpus indicated are online.

2) The current cpu is not in the list of indicated cpus.

Therefore we can pass a pointer to the mask directly.

One of the motivations in this transformation is to make use of
"&cpumask_of_cpu(cpu)" which evaluates to a pointer to constant
data in the kernel and thus takes up no stack space.

Hopefully someone in the future will change the interface of
arch_send_call_function_ipi() such that it passes a const cpumask_t
pointer so that this will optimize ever further.

Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c
index 868625e..47b0738 100644
--- a/arch/sparc64/kernel/smp.c
+++ b/arch/sparc64/kernel/smp.c
@@ -792,16 +792,15 @@
 
 void arch_send_call_function_ipi(cpumask_t mask)
 {
-	smp_cross_call_masked(&xcall_call_function, 0, 0, 0, mask);
+	xcall_deliver((u64) &xcall_call_function, 0, 0, &mask);
 }
 
 extern unsigned long xcall_call_function_single;
 
 void arch_send_call_function_single_ipi(int cpu)
 {
-	cpumask_t mask = cpumask_of_cpu(cpu);
-
-	smp_cross_call_masked(&xcall_call_function_single, 0, 0, 0, mask);
+	xcall_deliver((u64) &xcall_call_function_single, 0, 0,
+		      &cpumask_of_cpu(cpu));
 }
 
 /* Send cross call to all processors except self. */
@@ -959,24 +958,6 @@
 	put_cpu();
 }
 
-static void __smp_receive_signal_mask(cpumask_t mask)
-{
-	smp_cross_call_masked(&xcall_receive_signal, 0, 0, 0, mask);
-}
-
-void smp_receive_signal(int cpu)
-{
-	cpumask_t mask = cpumask_of_cpu(cpu);
-
-	if (cpu_online(cpu))
-		__smp_receive_signal_mask(mask);
-}
-
-void smp_receive_signal_client(int irq, struct pt_regs *regs)
-{
-	clear_softint(1 << irq);
-}
-
 void smp_new_mmu_context_version_client(int irq, struct pt_regs *regs)
 {
 	struct mm_struct *mm;
@@ -1374,7 +1355,13 @@
 
 void smp_send_reschedule(int cpu)
 {
-	smp_receive_signal(cpu);
+	xcall_deliver((u64) &xcall_receive_signal, 0, 0,
+		      &cpumask_of_cpu(cpu));
+}
+
+void smp_receive_signal_client(int irq, struct pt_regs *regs)
+{
+	clear_softint(1 << irq);
 }
 
 /* This is a nop because we capture all other cpus