diff options
Diffstat (limited to 'arch/powerpc/mm/numa.c')
| -rw-r--r-- | arch/powerpc/mm/numa.c | 182 |
1 files changed, 119 insertions, 63 deletions
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index 80d110635d2..002878ccf90 100644 --- a/arch/powerpc/mm/numa.c +++ b/arch/powerpc/mm/numa.c @@ -17,7 +17,7 @@ #include <linux/nodemask.h> #include <linux/cpu.h> #include <linux/notifier.h> -#include <linux/lmb.h> +#include <linux/memblock.h> #include <linux/of.h> #include <linux/pfn.h> #include <asm/sparsemem.h> @@ -42,6 +42,12 @@ EXPORT_SYMBOL(node_data); static int min_common_depth; static int n_mem_addr_cells, n_mem_size_cells; +static int form1_affinity; + +#define MAX_DISTANCE_REF_POINTS 4 +static int distance_ref_points_depth; +static const unsigned int *distance_ref_points; +static int distance_lookup_table[MAX_NUMNODES][MAX_DISTANCE_REF_POINTS]; /* * Allocate node_to_cpumask_map based on number of available nodes @@ -204,6 +210,39 @@ static const u32 *of_get_usable_memory(struct device_node *memory) return prop; } +int __node_distance(int a, int b) +{ + int i; + int distance = LOCAL_DISTANCE; + + if (!form1_affinity) + return distance; + + for (i = 0; i < distance_ref_points_depth; i++) { + if (distance_lookup_table[a][i] == distance_lookup_table[b][i]) + break; + + /* Double the distance for each NUMA level */ + distance *= 2; + } + + return distance; +} + +static void initialize_distance_lookup_table(int nid, + const unsigned int *associativity) +{ + int i; + + if (!form1_affinity) + return; + + for (i = 0; i < distance_ref_points_depth; i++) { + distance_lookup_table[nid][i] = + associativity[distance_ref_points[i]]; + } +} + /* Returns nid in the range [0..MAX_NUMNODES-1], or -1 if no useful numa * info is found. */ @@ -225,6 +264,10 @@ static int of_node_to_nid_single(struct device_node *device) /* POWER4 LPAR uses 0xffff as invalid node */ if (nid == 0xffff || nid >= MAX_NUMNODES) nid = -1; + + if (nid > 0 && tmp[0] >= distance_ref_points_depth) + initialize_distance_lookup_table(nid, tmp); + out: return nid; } @@ -251,26 +294,10 @@ int of_node_to_nid(struct device_node *device) } EXPORT_SYMBOL_GPL(of_node_to_nid); -/* - * In theory, the "ibm,associativity" property may contain multiple - * associativity lists because a resource may be multiply connected - * into the machine. This resource then has different associativity - * characteristics relative to its multiple connections. We ignore - * this for now. We also assume that all cpu and memory sets have - * their distances represented at a common level. This won't be - * true for hierarchical NUMA. - * - * In any case the ibm,associativity-reference-points should give - * the correct depth for a normal NUMA system. - * - * - Dave Hansen <haveblue@us.ibm.com> - */ static int __init find_min_common_depth(void) { - int depth, index; - const unsigned int *ref_points; + int depth; struct device_node *rtas_root; - unsigned int len; struct device_node *chosen; const char *vec5; @@ -280,18 +307,28 @@ static int __init find_min_common_depth(void) return -1; /* - * this property is 2 32-bit integers, each representing a level of - * depth in the associativity nodes. The first is for an SMP - * configuration (should be all 0's) and the second is for a normal - * NUMA configuration. + * This property is a set of 32-bit integers, each representing + * an index into the ibm,associativity nodes. + * + * With form 0 affinity the first integer is for an SMP configuration + * (should be all 0's) and the second is for a normal NUMA + * configuration. We have only one level of NUMA. + * + * With form 1 affinity the first integer is the most significant + * NUMA boundary and the following are progressively less significant + * boundaries. There can be more than one level of NUMA. */ - index = 1; - ref_points = of_get_property(rtas_root, - "ibm,associativity-reference-points", &len); + distance_ref_points = of_get_property(rtas_root, + "ibm,associativity-reference-points", + &distance_ref_points_depth); + + if (!distance_ref_points) { + dbg("NUMA: ibm,associativity-reference-points not found.\n"); + goto err; + } + + distance_ref_points_depth /= sizeof(int); - /* - * For form 1 affinity information we want the first field - */ #define VEC5_AFFINITY_BYTE 5 #define VEC5_AFFINITY 0x80 chosen = of_find_node_by_path("/chosen"); @@ -299,19 +336,38 @@ static int __init find_min_common_depth(void) vec5 = of_get_property(chosen, "ibm,architecture-vec-5", NULL); if (vec5 && (vec5[VEC5_AFFINITY_BYTE] & VEC5_AFFINITY)) { dbg("Using form 1 affinity\n"); - index = 0; + form1_affinity = 1; } } - if ((len >= 2 * sizeof(unsigned int)) && ref_points) { - depth = ref_points[index]; + if (form1_affinity) { + depth = distance_ref_points[0]; } else { - dbg("NUMA: ibm,associativity-reference-points not found.\n"); - depth = -1; + if (distance_ref_points_depth < 2) { + printk(KERN_WARNING "NUMA: " + "short ibm,associativity-reference-points\n"); + goto err; + } + + depth = distance_ref_points[1]; } - of_node_put(rtas_root); + /* + * Warn and cap if the hardware supports more than + * MAX_DISTANCE_REF_POINTS domains. + */ + if (distance_ref_points_depth > MAX_DISTANCE_REF_POINTS) { + printk(KERN_WARNING "NUMA: distance array capped at " + "%d entries\n", MAX_DISTANCE_REF_POINTS); + distance_ref_points_depth = MAX_DISTANCE_REF_POINTS; + } + + of_node_put(rtas_root); return depth; + +err: + of_node_put(rtas_root); + return -1; } static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells) @@ -351,7 +407,7 @@ struct of_drconf_cell { #define DRCONF_MEM_RESERVED 0x00000080 /* - * Read the next lmb list entry from the ibm,dynamic-memory property + * Read the next memblock list entry from the ibm,dynamic-memory property * and return the information in the provided of_drconf_cell structure. */ static void read_drconf_cell(struct of_drconf_cell *drmem, const u32 **cellp) @@ -372,8 +428,8 @@ static void read_drconf_cell(struct of_drconf_cell *drmem, const u32 **cellp) /* * Retreive and validate the ibm,dynamic-memory property of the device tree. * - * The layout of the ibm,dynamic-memory property is a number N of lmb - * list entries followed by N lmb list entries. Each lmb list entry + * The layout of the ibm,dynamic-memory property is a number N of memblock + * list entries followed by N memblock list entries. Each memblock list entry * contains information as layed out in the of_drconf_cell struct above. */ static int of_get_drconf_memory(struct device_node *memory, const u32 **dm) @@ -540,19 +596,19 @@ static unsigned long __init numa_enforce_memory_limit(unsigned long start, unsigned long size) { /* - * We use lmb_end_of_DRAM() in here instead of memory_limit because + * We use memblock_end_of_DRAM() in here instead of memory_limit because * we've already adjusted it for the limit and it takes care of * having memory holes below the limit. Also, in the case of * iommu_is_off, memory_limit is not set but is implicitly enforced. */ - if (start + size <= lmb_end_of_DRAM()) + if (start + size <= memblock_end_of_DRAM()) return size; - if (start >= lmb_end_of_DRAM()) + if (start >= memblock_end_of_DRAM()) return 0; - return lmb_end_of_DRAM() - start; + return memblock_end_of_DRAM() - start; } /* @@ -731,7 +787,7 @@ new_range: } /* - * Now do the same thing for each LMB listed in the ibm,dynamic-memory + * Now do the same thing for each MEMBLOCK listed in the ibm,dynamic-memory * property in the ibm,dynamic-reconfiguration-memory node. */ memory = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory"); @@ -743,8 +799,8 @@ new_range: static void __init setup_nonnuma(void) { - unsigned long top_of_ram = lmb_end_of_DRAM(); - unsigned long total_ram = lmb_phys_mem_size(); + unsigned long top_of_ram = memblock_end_of_DRAM(); + unsigned long total_ram = memblock_phys_mem_size(); unsigned long start_pfn, end_pfn; unsigned int i, nid = 0; @@ -753,9 +809,9 @@ static void __init setup_nonnuma(void) printk(KERN_DEBUG "Memory hole size: %ldMB\n", (top_of_ram - total_ram) >> 20); - for (i = 0; i < lmb.memory.cnt; ++i) { - start_pfn = lmb.memory.region[i].base >> PAGE_SHIFT; - end_pfn = start_pfn + lmb_size_pages(&lmb.memory, i); + for (i = 0; i < memblock.memory.cnt; ++i) { + start_pfn = memblock.memory.region[i].base >> PAGE_SHIFT; + end_pfn = start_pfn + memblock_size_pages(&memblock.memory, i); fake_numa_create_new_node(end_pfn, &nid); add_active_range(nid, start_pfn, end_pfn); @@ -813,7 +869,7 @@ static void __init dump_numa_memory_topology(void) count = 0; - for (i = 0; i < lmb_end_of_DRAM(); + for (i = 0; i < memblock_end_of_DRAM(); i += (1 << SECTION_SIZE_BITS)) { if (early_pfn_to_nid(i >> PAGE_SHIFT) == node) { if (count == 0) @@ -833,7 +889,7 @@ static void __init dump_numa_memory_topology(void) } /* - * Allocate some memory, satisfying the lmb or bootmem allocator where + * Allocate some memory, satisfying the memblock or bootmem allocator where * required. nid is the preferred node and end is the physical address of * the highest address in the node. * @@ -847,11 +903,11 @@ static void __init *careful_zallocation(int nid, unsigned long size, int new_nid; unsigned long ret_paddr; - ret_paddr = __lmb_alloc_base(size, align, end_pfn << PAGE_SHIFT); + ret_paddr = __memblock_alloc_base(size, align, end_pfn << PAGE_SHIFT); /* retry over all memory */ if (!ret_paddr) - ret_paddr = __lmb_alloc_base(size, align, lmb_end_of_DRAM()); + ret_paddr = __memblock_alloc_base(size, align, memblock_end_of_DRAM()); if (!ret_paddr) panic("numa.c: cannot allocate %lu bytes for node %d", @@ -861,14 +917,14 @@ static void __init *careful_zallocation(int nid, unsigned long size, /* * We initialize the nodes in numeric order: 0, 1, 2... - * and hand over control from the LMB allocator to the + * and hand over control from the MEMBLOCK allocator to the * bootmem allocator. If this function is called for * node 5, then we know that all nodes <5 are using the - * bootmem allocator instead of the LMB allocator. + * bootmem allocator instead of the MEMBLOCK allocator. * * So, check the nid from which this allocation came * and double check to see if we need to use bootmem - * instead of the LMB. We don't free the LMB memory + * instead of the MEMBLOCK. We don't free the MEMBLOCK memory * since it would be useless. */ new_nid = early_pfn_to_nid(ret_paddr >> PAGE_SHIFT); @@ -893,9 +949,9 @@ static void mark_reserved_regions_for_nid(int nid) struct pglist_data *node = NODE_DATA(nid); int i; - for (i = 0; i < lmb.reserved.cnt; i++) { - unsigned long physbase = lmb.reserved.region[i].base; - unsigned long size = lmb.reserved.region[i].size; + for (i = 0; i < memblock.reserved.cnt; i++) { + unsigned long physbase = memblock.reserved.region[i].base; + unsigned long size = memblock.reserved.region[i].size; unsigned long start_pfn = physbase >> PAGE_SHIFT; unsigned long end_pfn = PFN_UP(physbase + size); struct node_active_region node_ar; @@ -903,7 +959,7 @@ static void mark_reserved_regions_for_nid(int nid) node->node_spanned_pages; /* - * Check to make sure that this lmb.reserved area is + * Check to make sure that this memblock.reserved area is * within the bounds of the node that we care about. * Checking the nid of the start and end points is not * sufficient because the reserved area could span the @@ -961,7 +1017,7 @@ void __init do_init_bootmem(void) int nid; min_low_pfn = 0; - max_low_pfn = lmb_end_of_DRAM() >> PAGE_SHIFT; + max_low_pfn = memblock_end_of_DRAM() >> PAGE_SHIFT; max_pfn = max_low_pfn; if (parse_numa_properties()) @@ -1038,7 +1094,7 @@ void __init paging_init(void) { unsigned long max_zone_pfns[MAX_NR_ZONES]; memset(max_zone_pfns, 0, sizeof(max_zone_pfns)); - max_zone_pfns[ZONE_DMA] = lmb_end_of_DRAM() >> PAGE_SHIFT; + max_zone_pfns[ZONE_DMA] = memblock_end_of_DRAM() >> PAGE_SHIFT; free_area_init_nodes(max_zone_pfns); } @@ -1113,7 +1169,7 @@ static int hot_add_drconf_scn_to_nid(struct device_node *memory, /* * Find the node associated with a hot added memory section for memory * represented in the device tree as a node (i.e. memory@XXXX) for - * each lmb. + * each memblock. */ int hot_add_node_scn_to_nid(unsigned long scn_addr) { @@ -1154,8 +1210,8 @@ int hot_add_node_scn_to_nid(unsigned long scn_addr) /* * 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. + * corresponds to a SPARSEMEM section, not an MEMBLOCK. It is assumed that + * sections are fully contained within a single MEMBLOCK. */ int hot_add_scn_to_nid(unsigned long scn_addr) { |