1 files changed, 145 insertions, 162 deletions

diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c
index 3d6926ba899..5a02bf4c91e 100644
--- a/arch/x86/mm/numa_64.c
+++ b/arch/x86/mm/numa_64.c
@@ -1,7 +1,7 @@
-/* 
+/*
  * Generic VM initialization for x86-64 NUMA setups.
  * Copyright 2002,2003 Andi Kleen, SuSE Labs.
- */ 
+ */
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/string.h>
@@ -11,35 +11,45 @@
 #include <linux/ctype.h>
 #include <linux/module.h>
 #include <linux/nodemask.h>
+#include <linux/sched.h>
 
 #include <asm/e820.h>
 #include <asm/proto.h>
 #include <asm/dma.h>
 #include <asm/numa.h>
 #include <asm/acpi.h>
+#include <asm/k8.h>
 
 #ifndef Dprintk
 #define Dprintk(x...)
 #endif
 
 struct pglist_data *node_data[MAX_NUMNODES] __read_mostly;
+EXPORT_SYMBOL(node_data);
+
 bootmem_data_t plat_node_bdata[MAX_NUMNODES];
 
 struct memnode memnode;
 
-unsigned char cpu_to_node[NR_CPUS] __read_mostly = {
+int x86_cpu_to_node_map_init[NR_CPUS] = {
         [0 ... NR_CPUS-1] = NUMA_NO_NODE
 };
-unsigned char apicid_to_node[MAX_LOCAL_APIC] __cpuinitdata = {
-        [0 ... MAX_LOCAL_APIC-1] = NUMA_NO_NODE
+void *x86_cpu_to_node_map_early_ptr;
+DEFINE_PER_CPU(int, x86_cpu_to_node_map) = NUMA_NO_NODE;
+EXPORT_PER_CPU_SYMBOL(x86_cpu_to_node_map);
+EXPORT_SYMBOL(x86_cpu_to_node_map_early_ptr);
+
+s16 apicid_to_node[MAX_LOCAL_APIC] __cpuinitdata = {
+        [0 ... MAX_LOCAL_APIC-1] = NUMA_NO_NODE
 };
-cpumask_t node_to_cpumask[MAX_NUMNODES] __read_mostly;
+
+cpumask_t node_to_cpumask_map[MAX_NUMNODES] __read_mostly;
+EXPORT_SYMBOL(node_to_cpumask_map);
 
 int numa_off __initdata;
 unsigned long __initdata nodemap_addr;
 unsigned long __initdata nodemap_size;
 
-
 /*
  * Given a shift value, try to populate memnodemap[]
  * Returns :
@@ -47,14 +57,13 @@ unsigned long __initdata nodemap_size;
  * 0 if memnodmap[] too small (of shift too small)
  * -1 if node overlap or lost ram (shift too big)
  */
-static int __init
-populate_memnodemap(const struct bootnode *nodes, int numnodes, int shift)
+static int __init populate_memnodemap(const struct bootnode *nodes,
+                                      int numnodes, int shift)
 {
-        int i; 
-        int res = -1;
         unsigned long addr, end;
+        int i, res = -1;
 
-        memset(memnodemap, 0xff, memnodemapsize);
+        memset(memnodemap, 0xff, sizeof(s16)*memnodemapsize);
         for (i = 0; i < numnodes; i++) {
                 addr = nodes[i].start;
                 end = nodes[i].end;
@@ -63,37 +72,36 @@ populate_memnodemap(const struct bootnode *nodes, int numnodes, int shift)
                 if ((end >> shift) >= memnodemapsize)
                         return 0;
                 do {
-                        if (memnodemap[addr >> shift] != 0xff)
+                        if (memnodemap[addr >> shift] != NUMA_NO_NODE)
                                 return -1;
                         memnodemap[addr >> shift] = i;
                         addr += (1UL << shift);
                 } while (addr < end);
                 res = 1;
-        } 
+        }
         return res;
 }
 
 static int __init allocate_cachealigned_memnodemap(void)
 {
-        unsigned long pad, pad_addr;
+        unsigned long addr;
 
         memnodemap = memnode.embedded_map;
-        if (memnodemapsize <= 48)
+        if (memnodemapsize <= ARRAY_SIZE(memnode.embedded_map))
                 return 0;
 
-        pad = L1_CACHE_BYTES - 1;
-        pad_addr = 0x8000;
-        nodemap_size = pad + memnodemapsize;
-        nodemap_addr = find_e820_area(pad_addr, end_pfn<<PAGE_SHIFT,
-                                      nodemap_size);
+        addr = 0x8000;
+        nodemap_size = round_up(sizeof(s16) * memnodemapsize, L1_CACHE_BYTES);
+        nodemap_addr = find_e820_area(addr, end_pfn<<PAGE_SHIFT,
+                                      nodemap_size, L1_CACHE_BYTES);
         if (nodemap_addr == -1UL) {
                 printk(KERN_ERR
                        "NUMA: Unable to allocate Memory to Node hash map\n");
                 nodemap_addr = nodemap_size = 0;
                 return -1;
         }
-        pad_addr = (nodemap_addr + pad) & ~pad;
-        memnodemap = phys_to_virt(pad_addr);
+        memnodemap = phys_to_virt(nodemap_addr);
+        reserve_early(nodemap_addr, nodemap_addr + nodemap_size, "MEMNODEMAP");
 
         printk(KERN_DEBUG "NUMA: Allocated memnodemap from %lx - %lx\n",
                nodemap_addr, nodemap_addr + nodemap_size);
@@ -104,8 +112,8 @@ static int __init allocate_cachealigned_memnodemap(void)
  * The LSB of all start and end addresses in the node map is the value of the
  * maximum possible shift.
  */
-static int __init
-extract_lsb_from_nodes (const struct bootnode *nodes, int numnodes)
+static int __init extract_lsb_from_nodes(const struct bootnode *nodes,
+                                         int numnodes)
 {
         int i, nodes_used = 0;
         unsigned long start, end;
@@ -140,59 +148,62 @@ int __init compute_hash_shift(struct bootnode *nodes, int numnodes)
                 shift);
 
         if (populate_memnodemap(nodes, numnodes, shift) != 1) {
-                printk(KERN_INFO
-        "Your memory is not aligned you need to rebuild your kernel "
-        "with a bigger NODEMAPSIZE shift=%d\n",
-                        shift);
+                printk(KERN_INFO "Your memory is not aligned you need to "
+                       "rebuild your kernel with a bigger NODEMAPSIZE "
+                       "shift=%d\n", shift);
                 return -1;
         }
         return shift;
 }
 
-#ifdef CONFIG_SPARSEMEM
 int early_pfn_to_nid(unsigned long pfn)
 {
         return phys_to_nid(pfn << PAGE_SHIFT);
 }
-#endif
 
-static void * __init
-early_node_mem(int nodeid, unsigned long start, unsigned long end,
-              unsigned long size)
+static void * __init early_node_mem(int nodeid, unsigned long start,
+                                    unsigned long end, unsigned long size,
+                                    unsigned long align)
 {
-        unsigned long mem = find_e820_area(start, end, size);
+        unsigned long mem = find_e820_area(start, end, size, align);
         void *ptr;
+
         if (mem != -1L)
                 return __va(mem);
-        ptr = __alloc_bootmem_nopanic(size,
-                                SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS));
+
+        ptr = __alloc_bootmem_nopanic(size, align, __pa(MAX_DMA_ADDRESS));
         if (ptr == NULL) {
                 printk(KERN_ERR "Cannot find %lu bytes in node %d\n",
-                        size, nodeid);
+                       size, nodeid);
                 return NULL;
         }
         return ptr;
 }
 
 /* Initialize bootmem allocator for a node */
-void __init setup_node_bootmem(int nodeid, unsigned long start, unsigned long end)
-{ 
-        unsigned long start_pfn, end_pfn, bootmap_pages, bootmap_size, bootmap_start; 
-        unsigned long nodedata_phys;
+void __init setup_node_bootmem(int nodeid, unsigned long start,
+                               unsigned long end)
+{
+        unsigned long start_pfn, end_pfn, bootmap_pages, bootmap_size;
+        unsigned long bootmap_start, nodedata_phys;
         void *bootmap;
         const int pgdat_size = round_up(sizeof(pg_data_t), PAGE_SIZE);
 
-        start = round_up(start, ZONE_ALIGN); 
+        start = round_up(start, ZONE_ALIGN);
 
-        printk(KERN_INFO "Bootmem setup node %d %016lx-%016lx\n", nodeid, start, end);
+        printk(KERN_INFO "Bootmem setup node %d %016lx-%016lx\n", nodeid,
+               start, end);
 
         start_pfn = start >> PAGE_SHIFT;
         end_pfn = end >> PAGE_SHIFT;
 
-        node_data[nodeid] = early_node_mem(nodeid, start, end, pgdat_size);
+        node_data[nodeid] = early_node_mem(nodeid, start, end, pgdat_size,
+                                           SMP_CACHE_BYTES);
         if (node_data[nodeid] == NULL)
                 return;
         nodedata_phys = __pa(node_data[nodeid]);
+        printk(KERN_INFO "  NODE_DATA [%016lx - %016lx]\n", nodedata_phys,
+                nodedata_phys + pgdat_size - 1);
 
         memset(NODE_DATA(nodeid), 0, sizeof(pg_data_t));
         NODE_DATA(nodeid)->bdata = &plat_node_bdata[nodeid];
@@ -200,75 +211,62 @@ void __init setup_node_bootmem(int nodeid, unsigned long start, unsigned long en
         NODE_DATA(nodeid)->node_spanned_pages = end_pfn - start_pfn;
 
         /* Find a place for the bootmem map */
-        bootmap_pages = bootmem_bootmap_pages(end_pfn - start_pfn); 
+        bootmap_pages = bootmem_bootmap_pages(end_pfn - start_pfn);
         bootmap_start = round_up(nodedata_phys + pgdat_size, PAGE_SIZE);
+        /*
+         * SMP_CAHCE_BYTES could be enough, but init_bootmem_node like
+         * to use that to align to PAGE_SIZE
+         */
         bootmap = early_node_mem(nodeid, bootmap_start, end,
-                                        bootmap_pages<<PAGE_SHIFT);
+                                 bootmap_pages<<PAGE_SHIFT, PAGE_SIZE);
         if (bootmap == NULL)  {
                 if (nodedata_phys < start || nodedata_phys >= end)
-                        free_bootmem((unsigned long)node_data[nodeid],pgdat_size);
+                        free_bootmem((unsigned long)node_data[nodeid],
+                                     pgdat_size);
                 node_data[nodeid] = NULL;
                 return;
         }
         bootmap_start = __pa(bootmap);
-        Dprintk("bootmap start %lu pages %lu\n", bootmap_start, bootmap_pages); 
-        
+
         bootmap_size = init_bootmem_node(NODE_DATA(nodeid),
-                                         bootmap_start >> PAGE_SHIFT, 
-                                         start_pfn, end_pfn); 
+                                         bootmap_start >> PAGE_SHIFT,
+                                         start_pfn, end_pfn);
+
+        printk(KERN_INFO "  bootmap [%016lx -  %016lx] pages %lx\n",
+                 bootmap_start, bootmap_start + bootmap_size - 1,
+                 bootmap_pages);
 
         free_bootmem_with_active_regions(nodeid, end);
 
-        reserve_bootmem_node(NODE_DATA(nodeid), nodedata_phys, pgdat_size); 
-        reserve_bootmem_node(NODE_DATA(nodeid), bootmap_start, bootmap_pages<<PAGE_SHIFT);
+        reserve_bootmem_node(NODE_DATA(nodeid), nodedata_phys, pgdat_size);
+        reserve_bootmem_node(NODE_DATA(nodeid), bootmap_start,
+                             bootmap_pages<<PAGE_SHIFT);
 #ifdef CONFIG_ACPI_NUMA
         srat_reserve_add_area(nodeid);
 #endif
         node_set_online(nodeid);
-} 
-
-/* Initialize final allocator for a zone */
-void __init setup_node_zones(int nodeid)
-{ 
-        unsigned long start_pfn, end_pfn, memmapsize, limit;
-
-        start_pfn = node_start_pfn(nodeid);
-        end_pfn = node_end_pfn(nodeid);
-
-        Dprintk(KERN_INFO "Setting up memmap for node %d %lx-%lx\n",
-                nodeid, start_pfn, end_pfn);
-
-        /* Try to allocate mem_map at end to not fill up precious <4GB
-           memory. */
-        memmapsize = sizeof(struct page) * (end_pfn-start_pfn);
-        limit = end_pfn << PAGE_SHIFT;
-#ifdef CONFIG_FLAT_NODE_MEM_MAP
-        NODE_DATA(nodeid)->node_mem_map = 
-                __alloc_bootmem_core(NODE_DATA(nodeid)->bdata, 
-                                memmapsize, SMP_CACHE_BYTES, 
-                                round_down(limit - memmapsize, PAGE_SIZE), 
-                                limit);
-#endif
-} 
+}
 
+/*
+ * There are unfortunately some poorly designed mainboards around that
+ * only connect memory to a single CPU. This breaks the 1:1 cpu->node
+ * mapping. To avoid this fill in the mapping for all possible CPUs,
+ * as the number of CPUs is not known yet. We round robin the existing
+ * nodes.
+ */
 void __init numa_init_array(void)
 {
         int rr, i;
-        /* There are unfortunately some poorly designed mainboards around
-           that only connect memory to a single CPU. This breaks the 1:1 cpu->node
-           mapping. To avoid this fill in the mapping for all possible
-           CPUs, as the number of CPUs is not known yet. 
-           We round robin the existing nodes. */
+
         rr = first_node(node_online_map);
         for (i = 0; i < NR_CPUS; i++) {
-                if (cpu_to_node(i) != NUMA_NO_NODE)
+                if (early_cpu_to_node(i) != NUMA_NO_NODE)
                         continue;
-                numa_set_node(i, rr);
+                numa_set_node(i, rr);
                 rr = next_node(rr, node_online_map);
                 if (rr == MAX_NUMNODES)
                         rr = first_node(node_online_map);
         }
-
 }
 
 #ifdef CONFIG_NUMA_EMU
@@ -276,15 +274,17 @@ void __init numa_init_array(void)
 char *cmdline __initdata;
 
 /*
- * Setups up nid to range from addr to addr + size.  If the end boundary is
- * greater than max_addr, then max_addr is used instead.  The return value is 0
- * if there is additional memory left for allocation past addr and -1 otherwise.
- * addr is adjusted to be at the end of the node.
+ * Setups up nid to range from addr to addr + size.  If the end
+ * boundary is greater than max_addr, then max_addr is used instead.
+ * The return value is 0 if there is additional memory left for
+ * allocation past addr and -1 otherwise.  addr is adjusted to be at
+ * the end of the node.
  */
 static int __init setup_node_range(int nid, struct bootnode *nodes, u64 *addr,
                                    u64 size, u64 max_addr)
 {
         int ret = 0;
+
         nodes[nid].start = *addr;
         *addr += size;
         if (*addr >= max_addr) {
@@ -335,6 +335,7 @@ static int __init split_nodes_equally(struct bootnode *nodes, u64 *addr,
 
         for (i = node_start; i < num_nodes + node_start; i++) {
                 u64 end = *addr + size;
+
                 if (i < big)
                         end += FAKE_NODE_MIN_SIZE;
                 /*
@@ -380,14 +381,9 @@ static int __init split_nodes_by_size(struct bootnode *nodes, u64 *addr,
 static int __init numa_emulation(unsigned long start_pfn, unsigned long end_pfn)
 {
         struct bootnode nodes[MAX_NUMNODES];
-        u64 addr = start_pfn << PAGE_SHIFT;
+        u64 size, addr = start_pfn << PAGE_SHIFT;
         u64 max_addr = end_pfn << PAGE_SHIFT;
-        int num_nodes = 0;
-        int coeff_flag;
-        int coeff = -1;
-        int num = 0;
-        u64 size;
-        int i;
+        int num_nodes = 0, num = 0, coeff_flag, coeff = -1, i;
 
         memset(&nodes, 0, sizeof(nodes));
         /*
@@ -395,8 +391,9 @@ static int __init numa_emulation(unsigned long start_pfn, unsigned long end_pfn)
          * system RAM into N fake nodes.
          */
         if (!strchr(cmdline, '*') && !strchr(cmdline, ',')) {
-                num_nodes = split_nodes_equally(nodes, &addr, max_addr, 0,
-                                                simple_strtol(cmdline, NULL, 0));
+                long n = simple_strtol(cmdline, NULL, 0);
+
+                num_nodes = split_nodes_equally(nodes, &addr, max_addr, 0, n);
                 if (num_nodes < 0)
                         return num_nodes;
                 goto out;
@@ -483,46 +480,47 @@ out:
         for_each_node_mask(i, node_possible_map) {
                 e820_register_active_regions(i, nodes[i].start >> PAGE_SHIFT,
                                                 nodes[i].end >> PAGE_SHIFT);
-                setup_node_bootmem(i, nodes[i].start, nodes[i].end);
+                setup_node_bootmem(i, nodes[i].start, nodes[i].end);
         }
         acpi_fake_nodes(nodes, num_nodes);
-        numa_init_array();
-        return 0;
+        numa_init_array();
+        return 0;
 }
 #endif /* CONFIG_NUMA_EMU */
 
 void __init numa_initmem_init(unsigned long start_pfn, unsigned long end_pfn)
-{ 
+{
         int i;
 
         nodes_clear(node_possible_map);
 
 #ifdef CONFIG_NUMA_EMU
         if (cmdline && !numa_emulation(start_pfn, end_pfn))
-                return;
+                return;
         nodes_clear(node_possible_map);
 #endif
 
 #ifdef CONFIG_ACPI_NUMA
         if (!numa_off && !acpi_scan_nodes(start_pfn << PAGE_SHIFT,
                                           end_pfn << PAGE_SHIFT))
-                return;
+                return;
         nodes_clear(node_possible_map);
 #endif
 
 #ifdef CONFIG_K8_NUMA
-        if (!numa_off && !k8_scan_nodes(start_pfn<<PAGE_SHIFT, end_pfn<<PAGE_SHIFT))
+        if (!numa_off && !k8_scan_nodes(start_pfn<<PAGE_SHIFT,
+                                        end_pfn<<PAGE_SHIFT))
                 return;
         nodes_clear(node_possible_map);
 #endif
         printk(KERN_INFO "%s\n",
                numa_off ? "NUMA turned off" : "No NUMA configuration found");
 
-        printk(KERN_INFO "Faking a node at %016lx-%016lx\n", 
+        printk(KERN_INFO "Faking a node at %016lx-%016lx\n",
                start_pfn << PAGE_SHIFT,
-               end_pfn << PAGE_SHIFT); 
-                /* setup dummy node covering all memory */ 
-        memnode_shift = 63; 
+               end_pfn << PAGE_SHIFT);
+        /* setup dummy node covering all memory */
+        memnode_shift = 63;
         memnodemap = memnode.embedded_map;
         memnodemap[0] = 0;
         nodes_clear(node_online_map);
@@ -530,36 +528,48 @@ void __init numa_initmem_init(unsigned long start_pfn, unsigned long end_pfn)
         node_set(0, node_possible_map);
         for (i = 0; i < NR_CPUS; i++)
                 numa_set_node(i, 0);
-        node_to_cpumask[0] = cpumask_of_cpu(0);
+        /* cpumask_of_cpu() may not be available during early startup */
+        memset(&node_to_cpumask_map[0], 0, sizeof(node_to_cpumask_map[0]));
+        cpu_set(0, node_to_cpumask_map[0]);
         e820_register_active_regions(0, start_pfn, end_pfn);
         setup_node_bootmem(0, start_pfn << PAGE_SHIFT, end_pfn << PAGE_SHIFT);
 }
 
 __cpuinit void numa_add_cpu(int cpu)
 {
-        set_bit(cpu, &node_to_cpumask[cpu_to_node(cpu)]);
-} 
+        set_bit(cpu,
+                (unsigned long *)&node_to_cpumask_map[early_cpu_to_node(cpu)]);
+}
 
 void __cpuinit numa_set_node(int cpu, int node)
 {
+        int *cpu_to_node_map = x86_cpu_to_node_map_early_ptr;
+
         cpu_pda(cpu)->nodenumber = node;
-        cpu_to_node(cpu) = node;
+
+        if(cpu_to_node_map)
+                cpu_to_node_map[cpu] = node;
+        else if(per_cpu_offset(cpu))
+                per_cpu(x86_cpu_to_node_map, cpu) = node;
+        else
+                Dprintk(KERN_INFO "Setting node for non-present cpu %d\n", cpu);
 }
 
-unsigned long __init numa_free_all_bootmem(void) 
-{ 
-        int i;
+unsigned long __init numa_free_all_bootmem(void)
+{
         unsigned long pages = 0;
-        for_each_online_node(i) {
+        int i;
+
+        for_each_online_node(i)
                 pages += free_all_bootmem_node(NODE_DATA(i));
-        }
+
         return pages;
-} 
+}
 
 void __init paging_init(void)
-{ 
-        int i;
+{
         unsigned long max_zone_pfns[MAX_NR_ZONES];
+
         memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
         max_zone_pfns[ZONE_DMA] = MAX_DMA_PFN;
         max_zone_pfns[ZONE_DMA32] = MAX_DMA32_PFN;
@@ -568,32 +578,27 @@ void __init paging_init(void)
         sparse_memory_present_with_active_regions(MAX_NUMNODES);
         sparse_init();
 
-        for_each_online_node(i) {
-                setup_node_zones(i); 
-        }
-
         free_area_init_nodes(max_zone_pfns);
-} 
+}
 
 static __init int numa_setup(char *opt)
-{ 
+{
         if (!opt)
                 return -EINVAL;
-        if (!strncmp(opt,"off",3))
+        if (!strncmp(opt, "off", 3))
                 numa_off = 1;
 #ifdef CONFIG_NUMA_EMU
         if (!strncmp(opt, "fake=", 5))
                 cmdline = opt + 5;
 #endif
 #ifdef CONFIG_ACPI_NUMA
-        if (!strncmp(opt,"noacpi",6))
-                acpi_numa = -1;
-        if (!strncmp(opt,"hotadd=", 7))
+        if (!strncmp(opt, "noacpi", 6))
+                acpi_numa = -1;
+        if (!strncmp(opt, "hotadd=", 7))
                 hotadd_percent = simple_strtoul(opt+7, NULL, 10);
 #endif
         return 0;
-} 
-
+}
 early_param("numa", numa_setup);
 
 /*
@@ -611,38 +616,16 @@ early_param("numa", numa_setup);
 void __init init_cpu_to_node(void)
 {
         int i;
-        for (i = 0; i < NR_CPUS; i++) {
-                u8 apicid = x86_cpu_to_apicid_init[i];
+
+        for (i = 0; i < NR_CPUS; i++) {
+                u16 apicid = x86_cpu_to_apicid_init[i];
+
                 if (apicid == BAD_APICID)
                         continue;
                 if (apicid_to_node[apicid] == NUMA_NO_NODE)
                         continue;
-                numa_set_node(i,apicid_to_node[apicid]);
+                numa_set_node(i, apicid_to_node[apicid]);
         }
 }
 
-EXPORT_SYMBOL(cpu_to_node);
-EXPORT_SYMBOL(node_to_cpumask);
-EXPORT_SYMBOL(memnode);
-EXPORT_SYMBOL(node_data);
 
-#ifdef CONFIG_DISCONTIGMEM
-/*
- * Functions to convert PFNs from/to per node page addresses.
- * These are out of line because they are quite big.
- * They could be all tuned by pre caching more state.
- * Should do that.
- */
-
-int pfn_valid(unsigned long pfn)
-{
-        unsigned nid;
-        if (pfn >= num_physpages)
-                return 0;
-        nid = pfn_to_nid(pfn);
-        if (nid == 0xff)
-                return 0;
-        return pfn >= node_start_pfn(nid) && (pfn) < node_end_pfn(nid);
-}
-EXPORT_SYMBOL(pfn_valid);
-#endif