static bootmem_data_t __initdata plat_node_bdata[MAX_NUMNODES];
static int min_common_depth;
+static int n_mem_addr_cells, n_mem_size_cells;
/*
* We need somewhere to store start/end/node for each region until we have
*start_pfn = 0;
}
-static inline void map_cpu_to_node(int cpu, int node)
+static void __cpuinit map_cpu_to_node(int cpu, int node)
{
numa_cpu_lookup_table[cpu] = node;
+ dbg("adding cpu %d to node %d\n", cpu, node);
+
if (!(cpu_isset(cpu, numa_cpumask_lookup_table[node])))
cpu_set(cpu, numa_cpumask_lookup_table[node]);
}
}
#endif /* CONFIG_HOTPLUG_CPU */
-static struct device_node *find_cpu_node(unsigned int cpu)
+static struct device_node * __cpuinit find_cpu_node(unsigned int cpu)
{
unsigned int hw_cpuid = get_hard_smp_processor_id(cpu);
struct device_node *cpu_node = NULL;
return (unsigned int *)get_property(dev, "ibm,associativity", NULL);
}
-static int of_node_numa_domain(struct device_node *device)
+/* Returns nid in the range [0..MAX_NUMNODES-1], or -1 if no useful numa
+ * info is found.
+ */
+static int of_node_to_nid(struct device_node *device)
{
- int numa_domain;
+ int nid = -1;
unsigned int *tmp;
if (min_common_depth == -1)
- return 0;
+ goto out;
tmp = of_get_associativity(device);
- if (tmp && (tmp[0] >= min_common_depth)) {
- numa_domain = tmp[min_common_depth];
- } else {
- dbg("WARNING: no NUMA information for %s\n",
- device->full_name);
- numa_domain = 0;
- }
- return numa_domain;
+ if (!tmp)
+ goto out;
+
+ if (tmp[0] >= min_common_depth)
+ nid = tmp[min_common_depth];
+
+ /* POWER4 LPAR uses 0xffff as invalid node */
+ if (nid == 0xffff || nid >= MAX_NUMNODES)
+ nid = -1;
+out:
+ return nid;
}
/*
if ((len >= 1) && ref_points) {
depth = ref_points[1];
} else {
- dbg("WARNING: could not find NUMA "
- "associativity reference point\n");
+ dbg("NUMA: ibm,associativity-reference-points not found.\n");
depth = -1;
}
of_node_put(rtas_root);
return depth;
}
-static int __init get_mem_addr_cells(void)
+static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells)
{
struct device_node *memory = NULL;
- int rc;
memory = of_find_node_by_type(memory, "memory");
if (!memory)
- return 0; /* it won't matter */
+ panic("numa.c: No memory nodes found!");
- rc = prom_n_addr_cells(memory);
- return rc;
+ *n_addr_cells = prom_n_addr_cells(memory);
+ *n_size_cells = prom_n_size_cells(memory);
+ of_node_put(memory);
}
-static int __init get_mem_size_cells(void)
-{
- struct device_node *memory = NULL;
- int rc;
-
- memory = of_find_node_by_type(memory, "memory");
- if (!memory)
- return 0; /* it won't matter */
- rc = prom_n_size_cells(memory);
- return rc;
-}
-
-static unsigned long __init read_n_cells(int n, unsigned int **buf)
+static unsigned long __devinit read_n_cells(int n, unsigned int **buf)
{
unsigned long result = 0;
* Figure out to which domain a cpu belongs and stick it there.
* Return the id of the domain used.
*/
-static int numa_setup_cpu(unsigned long lcpu)
+static int __cpuinit numa_setup_cpu(unsigned long lcpu)
{
- int numa_domain = 0;
+ int nid = 0;
struct device_node *cpu = find_cpu_node(lcpu);
if (!cpu) {
goto out;
}
- numa_domain = of_node_numa_domain(cpu);
+ nid = of_node_to_nid(cpu);
- if (numa_domain >= num_online_nodes()) {
- /*
- * POWER4 LPAR uses 0xffff as invalid node,
- * dont warn in this case.
- */
- if (numa_domain != 0xffff)
- printk(KERN_ERR "WARNING: cpu %ld "
- "maps to invalid NUMA node %d\n",
- lcpu, numa_domain);
- numa_domain = 0;
- }
+ if (nid < 0 || !node_online(nid))
+ nid = any_online_node(NODE_MASK_ALL);
out:
- node_set_online(numa_domain);
-
- map_cpu_to_node(lcpu, numa_domain);
+ map_cpu_to_node(lcpu, nid);
of_node_put(cpu);
- return numa_domain;
+ return nid;
}
static int cpu_numa_callback(struct notifier_block *nfb,
switch (action) {
case CPU_UP_PREPARE:
- if (min_common_depth == -1 || !numa_enabled)
- map_cpu_to_node(lcpu, 0);
- else
- numa_setup_cpu(lcpu);
+ numa_setup_cpu(lcpu);
ret = NOTIFY_OK;
break;
#ifdef CONFIG_HOTPLUG_CPU
{
struct device_node *cpu = NULL;
struct device_node *memory = NULL;
- int addr_cells, size_cells;
- int max_domain;
+ int default_nid = 0;
unsigned long i;
if (numa_enabled == 0) {
min_common_depth = find_min_common_depth();
- dbg("NUMA associativity depth for CPU/Memory: %d\n", min_common_depth);
if (min_common_depth < 0)
return min_common_depth;
- max_domain = numa_setup_cpu(boot_cpuid);
+ dbg("NUMA associativity depth for CPU/Memory: %d\n", min_common_depth);
/*
- * Even though we connect cpus to numa domains later in SMP init,
- * we need to know the maximum node id now. This is because each
- * node id must have NODE_DATA etc backing it.
- * As a result of hotplug we could still have cpus appear later on
- * with larger node ids. In that case we force the cpu into node 0.
+ * Even though we connect cpus to numa domains later in SMP
+ * init, we need to know the node ids now. This is because
+ * each node to be onlined must have NODE_DATA etc backing it.
*/
- for_each_cpu(i) {
- int numa_domain;
+ for_each_present_cpu(i) {
+ int nid;
cpu = find_cpu_node(i);
+ BUG_ON(!cpu);
+ nid = of_node_to_nid(cpu);
+ of_node_put(cpu);
- if (cpu) {
- numa_domain = of_node_numa_domain(cpu);
- of_node_put(cpu);
-
- if (numa_domain < MAX_NUMNODES &&
- max_domain < numa_domain)
- max_domain = numa_domain;
- }
+ /*
+ * Don't fall back to default_nid yet -- we will plug
+ * cpus into nodes once the memory scan has discovered
+ * the topology.
+ */
+ if (nid < 0)
+ continue;
+ node_set_online(nid);
}
- addr_cells = get_mem_addr_cells();
- size_cells = get_mem_size_cells();
+ get_n_mem_cells(&n_mem_addr_cells, &n_mem_size_cells);
memory = NULL;
while ((memory = of_find_node_by_type(memory, "memory")) != NULL) {
unsigned long start;
unsigned long size;
- int numa_domain;
+ int nid;
int ranges;
unsigned int *memcell_buf;
unsigned int len;
- memcell_buf = (unsigned int *)get_property(memory, "reg", &len);
+ memcell_buf = (unsigned int *)get_property(memory,
+ "linux,usable-memory", &len);
+ if (!memcell_buf || len <= 0)
+ memcell_buf =
+ (unsigned int *)get_property(memory, "reg",
+ &len);
if (!memcell_buf || len <= 0)
continue;
- ranges = memory->n_addrs;
+ /* ranges in cell */
+ ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
new_range:
/* these are order-sensitive, and modify the buffer pointer */
- start = read_n_cells(addr_cells, &memcell_buf);
- size = read_n_cells(size_cells, &memcell_buf);
+ start = read_n_cells(n_mem_addr_cells, &memcell_buf);
+ size = read_n_cells(n_mem_size_cells, &memcell_buf);
- numa_domain = of_node_numa_domain(memory);
-
- if (numa_domain >= MAX_NUMNODES) {
- if (numa_domain != 0xffff)
- printk(KERN_ERR "WARNING: memory at %lx maps "
- "to invalid NUMA node %d\n", start,
- numa_domain);
- numa_domain = 0;
- }
-
- if (max_domain < numa_domain)
- max_domain = numa_domain;
+ /*
+ * Assumption: either all memory nodes or none will
+ * have associativity properties. If none, then
+ * everything goes to default_nid.
+ */
+ nid = of_node_to_nid(memory);
+ if (nid < 0)
+ nid = default_nid;
+ node_set_online(nid);
if (!(size = numa_enforce_memory_limit(start, size))) {
if (--ranges)
continue;
}
- add_region(numa_domain, start >> PAGE_SHIFT,
+ add_region(nid, start >> PAGE_SHIFT,
size >> PAGE_SHIFT);
if (--ranges)
goto new_range;
}
- for (i = 0; i <= max_domain; i++)
- node_set_online(i);
-
return 0;
}
unsigned long total_ram = lmb_phys_mem_size();
unsigned int i;
- printk(KERN_INFO "Top of RAM: 0x%lx, Total RAM: 0x%lx\n",
+ printk(KERN_DEBUG "Top of RAM: 0x%lx, Total RAM: 0x%lx\n",
top_of_ram, total_ram);
- printk(KERN_INFO "Memory hole size: %ldMB\n",
+ printk(KERN_DEBUG "Memory hole size: %ldMB\n",
(top_of_ram - total_ram) >> 20);
- map_cpu_to_node(boot_cpuid, 0);
for (i = 0; i < lmb.memory.cnt; ++i)
add_region(0, lmb.memory.region[i].base >> PAGE_SHIFT,
lmb_size_pages(&lmb.memory, i));
node_set_online(0);
}
-static void __init dump_numa_topology(void)
+void __init dump_numa_cpu_topology(void)
+{
+ unsigned int node;
+ unsigned int cpu, count;
+
+ if (min_common_depth == -1 || !numa_enabled)
+ return;
+
+ for_each_online_node(node) {
+ printk(KERN_DEBUG "Node %d CPUs:", node);
+
+ count = 0;
+ /*
+ * If we used a CPU iterator here we would miss printing
+ * the holes in the cpumap.
+ */
+ for (cpu = 0; cpu < NR_CPUS; cpu++) {
+ if (cpu_isset(cpu, numa_cpumask_lookup_table[node])) {
+ if (count == 0)
+ printk(" %u", cpu);
+ ++count;
+ } else {
+ if (count > 1)
+ printk("-%u", cpu - 1);
+ count = 0;
+ }
+ }
+
+ if (count > 1)
+ printk("-%u", NR_CPUS - 1);
+ printk("\n");
+ }
+}
+
+static void __init dump_numa_memory_topology(void)
{
unsigned int node;
unsigned int count;
for_each_online_node(node) {
unsigned long i;
- printk(KERN_INFO "Node %d Memory:", node);
+ printk(KERN_DEBUG "Node %d Memory:", node);
count = 0;
printk("-0x%lx", i);
printk("\n");
}
- return;
}
/*
unsigned long end_pfn)
{
int new_nid;
- unsigned long ret = lmb_alloc_base(size, align, end_pfn << PAGE_SHIFT);
+ unsigned long ret = __lmb_alloc_base(size, align, end_pfn << PAGE_SHIFT);
/* retry over all memory */
if (!ret)
- ret = lmb_alloc_base(size, align, lmb_end_of_DRAM());
+ ret = __lmb_alloc_base(size, align, lmb_end_of_DRAM());
if (!ret)
panic("numa.c: cannot allocate %lu bytes on node %d",
if (parse_numa_properties())
setup_nonnuma();
else
- dump_numa_topology();
+ dump_numa_memory_topology();
register_cpu_notifier(&ppc64_numa_nb);
+ cpu_numa_callback(&ppc64_numa_nb, CPU_UP_PREPARE,
+ (void *)(unsigned long)boot_cpuid);
for_each_online_node(nid) {
unsigned long start_pfn, end_pfn, pages_present;
return 0;
}
early_param("numa", early_numa);
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+/*
+ * Find the node associated with a hot added memory section. Section
+ * corresponds to a SPARSEMEM section, not an LMB. It is assumed that
+ * sections are fully contained within a single LMB.
+ */
+int hot_add_scn_to_nid(unsigned long scn_addr)
+{
+ struct device_node *memory = NULL;
+ nodemask_t nodes;
+ int default_nid = any_online_node(NODE_MASK_ALL);
+ int nid;
+
+ if (!numa_enabled || (min_common_depth < 0))
+ return default_nid;
+
+ while ((memory = of_find_node_by_type(memory, "memory")) != NULL) {
+ unsigned long start, size;
+ int ranges;
+ unsigned int *memcell_buf;
+ unsigned int len;
+
+ memcell_buf = (unsigned int *)get_property(memory, "reg", &len);
+ if (!memcell_buf || len <= 0)
+ continue;
+
+ /* ranges in cell */
+ ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
+ha_new_range:
+ start = read_n_cells(n_mem_addr_cells, &memcell_buf);
+ size = read_n_cells(n_mem_size_cells, &memcell_buf);
+ nid = of_node_to_nid(memory);
+
+ /* Domains not present at boot default to 0 */
+ if (nid < 0 || !node_online(nid))
+ nid = default_nid;
+
+ if ((scn_addr >= start) && (scn_addr < (start + size))) {
+ of_node_put(memory);
+ goto got_nid;
+ }
+
+ if (--ranges) /* process all ranges in cell */
+ goto ha_new_range;
+ }
+ BUG(); /* section address should be found above */
+ return 0;
+
+ /* Temporary code to ensure that returned node is not empty */
+got_nid:
+ nodes_setall(nodes);
+ while (NODE_DATA(nid)->node_spanned_pages == 0) {
+ node_clear(nid, nodes);
+ nid = any_online_node(nodes);
+ }
+ return nid;
+}
+#endif /* CONFIG_MEMORY_HOTPLUG */