x86-64: Move nodenumber from PDA to per-cpu.

tj: * s/nodenumber/node_number/
    * removed now unused pda variable from pda_init()

Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>

arch/x86/include/asm/pda.h

@@ -22,7 +22,6 @@ struct x8664_pda {
 					/* gcc-ABI: this canary MUST be at
 					   offset 40!!! */
 #endif
-	short nodenumber;		/* number of current node (32k max) */
 	short in_bootmem;		/* pda lives in bootmem */
 	short isidle;
 } ____cacheline_aligned_in_smp;

arch/x86/include/asm/topology.h

@@ -83,7 +83,8 @@ extern cpumask_t *node_to_cpumask_map;
 DECLARE_EARLY_PER_CPU(int, x86_cpu_to_node_map);
 
 /* Returns the number of the current Node. */
-#define numa_node_id()		read_pda(nodenumber)
+DECLARE_PER_CPU(int, node_number);
+#define numa_node_id()		percpu_read(node_number)
 
 #ifdef CONFIG_DEBUG_PER_CPU_MAPS
 extern int cpu_to_node(int cpu);

arch/x86/kernel/cpu/common.c

@@ -897,18 +897,11 @@ DEFINE_PER_CPU(unsigned int, irq_count) = -1;
 
 void __cpuinit pda_init(int cpu)
 {
-	struct x8664_pda *pda = cpu_pda(cpu);
-
 	/* Setup up data that may be needed in __get_free_pages early */
 	loadsegment(fs, 0);
 	loadsegment(gs, 0);
 
 	load_pda_offset(cpu);
-
-	if (cpu != 0) {
-		if (pda->nodenumber == 0 && cpu_to_node(cpu) != NUMA_NO_NODE)
-			pda->nodenumber = cpu_to_node(cpu);
-	}
 }
 
 static DEFINE_PER_CPU_PAGE_ALIGNED(char, exception_stacks
@@ -978,6 +971,12 @@ void __cpuinit cpu_init(void)
 	if (cpu != 0)
 		pda_init(cpu);
 
+#ifdef CONFIG_NUMA
+	if (cpu != 0 && percpu_read(node_number) == 0 &&
+	    cpu_to_node(cpu) != NUMA_NO_NODE)
+		percpu_write(node_number, cpu_to_node(cpu));
+#endif
+
 	me = current;
 
 	if (cpumask_test_and_set_cpu(cpu, cpu_initialized_mask))

arch/x86/kernel/setup_percpu.c

@@ -53,6 +53,8 @@ EXPORT_EARLY_PER_CPU_SYMBOL(x86_bios_cpu_apicid);
 
 #if defined(CONFIG_NUMA) && defined(CONFIG_X86_64)
 #define	X86_64_NUMA	1	/* (used later) */
+DEFINE_PER_CPU(int, node_number) = 0;
+EXPORT_PER_CPU_SYMBOL(node_number);
 
 /*
  * Map cpu index to node index
@@ -283,7 +285,7 @@ void __cpuinit numa_set_node(int cpu, int node)
 	per_cpu(x86_cpu_to_node_map, cpu) = node;
 
 	if (node != NUMA_NO_NODE)
-		cpu_pda(cpu)->nodenumber = node;
+		per_cpu(node_number, cpu) = node;
 }
 
 void __cpuinit numa_clear_node(int cpu)