diff options
Diffstat (limited to 'arch/powerpc/mm/numa.c')
| -rw-r--r-- | arch/powerpc/mm/numa.c | 1057 |
1 files changed, 861 insertions, 196 deletions
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index b037d95eead..3b181b22cd4 100644 --- a/arch/powerpc/mm/numa.c +++ b/arch/powerpc/mm/numa.c @@ -13,17 +13,31 @@ #include <linux/init.h> #include <linux/mm.h> #include <linux/mmzone.h> -#include <linux/module.h> +#include <linux/export.h> #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 <linux/cpuset.h> +#include <linux/node.h> +#include <linux/stop_machine.h> +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/slab.h> +#include <asm/cputhreads.h> #include <asm/sparsemem.h> #include <asm/prom.h> -#include <asm/system.h> #include <asm/smp.h> +#include <asm/cputhreads.h> +#include <asm/topology.h> +#include <asm/firmware.h> +#include <asm/paca.h> +#include <asm/hvcall.h> +#include <asm/setup.h> +#include <asm/vdso.h> static int numa_enabled = 1; @@ -33,17 +47,45 @@ static int numa_debug; #define dbg(args...) if (numa_debug) { printk(KERN_INFO args); } int numa_cpu_lookup_table[NR_CPUS]; -cpumask_t numa_cpumask_lookup_table[MAX_NUMNODES]; +cpumask_var_t node_to_cpumask_map[MAX_NUMNODES]; struct pglist_data *node_data[MAX_NUMNODES]; EXPORT_SYMBOL(numa_cpu_lookup_table); -EXPORT_SYMBOL(numa_cpumask_lookup_table); +EXPORT_SYMBOL(node_to_cpumask_map); EXPORT_SYMBOL(node_data); static int min_common_depth; static int n_mem_addr_cells, n_mem_size_cells; +static int form1_affinity; -static int __cpuinit fake_numa_create_new_node(unsigned long end_pfn, +#define MAX_DISTANCE_REF_POINTS 4 +static int distance_ref_points_depth; +static const __be32 *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 + * Requires node_possible_map to be valid. + * + * Note: cpumask_of_node() is not valid until after this is done. + */ +static void __init setup_node_to_cpumask_map(void) +{ + unsigned int node; + + /* setup nr_node_ids if not done yet */ + if (nr_node_ids == MAX_NUMNODES) + setup_nr_node_ids(); + + /* allocate the map */ + for (node = 0; node < nr_node_ids; node++) + alloc_bootmem_cpumask_var(&node_to_cpumask_map[node]); + + /* cpumask_of_node() will now work */ + dbg("Node to cpumask map for %d nodes\n", nr_node_ids); +} + +static int __init fake_numa_create_new_node(unsigned long end_pfn, unsigned int *nid) { unsigned long long mem; @@ -91,75 +133,68 @@ static int __cpuinit fake_numa_create_new_node(unsigned long end_pfn, } /* - * get_active_region_work_fn - A helper function for get_node_active_region - * Returns datax set to the start_pfn and end_pfn if they contain - * the initial value of datax->start_pfn between them - * @start_pfn: start page(inclusive) of region to check - * @end_pfn: end page(exclusive) of region to check - * @datax: comes in with ->start_pfn set to value to search for and - * goes out with active range if it contains it - * Returns 1 if search value is in range else 0 + * get_node_active_region - Return active region containing pfn + * Active range returned is empty if none found. + * @pfn: The page to return the region for + * @node_ar: Returned set to the active region containing @pfn */ -static int __init get_active_region_work_fn(unsigned long start_pfn, - unsigned long end_pfn, void *datax) +static void __init get_node_active_region(unsigned long pfn, + struct node_active_region *node_ar) { - struct node_active_region *data; - data = (struct node_active_region *)datax; + unsigned long start_pfn, end_pfn; + int i, nid; - if (start_pfn <= data->start_pfn && end_pfn > data->start_pfn) { - data->start_pfn = start_pfn; - data->end_pfn = end_pfn; - return 1; + for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, &nid) { + if (pfn >= start_pfn && pfn < end_pfn) { + node_ar->nid = nid; + node_ar->start_pfn = start_pfn; + node_ar->end_pfn = end_pfn; + break; + } } - return 0; - } -/* - * get_node_active_region - Return active region containing start_pfn - * Active range returned is empty if none found. - * @start_pfn: The page to return the region for. - * @node_ar: Returned set to the active region containing start_pfn - */ -static void __init get_node_active_region(unsigned long start_pfn, - struct node_active_region *node_ar) +static void reset_numa_cpu_lookup_table(void) { - int nid = early_pfn_to_nid(start_pfn); + unsigned int cpu; - node_ar->nid = nid; - node_ar->start_pfn = start_pfn; - node_ar->end_pfn = start_pfn; - work_with_active_regions(nid, get_active_region_work_fn, node_ar); + for_each_possible_cpu(cpu) + numa_cpu_lookup_table[cpu] = -1; } -static void __cpuinit map_cpu_to_node(int cpu, int node) +static void update_numa_cpu_lookup_table(unsigned int cpu, int node) { numa_cpu_lookup_table[cpu] = node; +} + +static void map_cpu_to_node(int cpu, int node) +{ + update_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]); + if (!(cpumask_test_cpu(cpu, node_to_cpumask_map[node]))) + cpumask_set_cpu(cpu, node_to_cpumask_map[node]); } -#ifdef CONFIG_HOTPLUG_CPU +#if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_PPC_SPLPAR) static void unmap_cpu_from_node(unsigned long cpu) { int node = numa_cpu_lookup_table[cpu]; dbg("removing cpu %lu from node %d\n", cpu, node); - if (cpu_isset(cpu, numa_cpumask_lookup_table[node])) { - cpu_clear(cpu, numa_cpumask_lookup_table[node]); + if (cpumask_test_cpu(cpu, node_to_cpumask_map[node])) { + cpumask_clear_cpu(cpu, node_to_cpumask_map[node]); } else { printk(KERN_ERR "WARNING: cpu %lu not found in node %d\n", cpu, node); } } -#endif /* CONFIG_HOTPLUG_CPU */ +#endif /* CONFIG_HOTPLUG_CPU || CONFIG_PPC_SPLPAR */ /* must hold reference to node during call */ -static const int *of_get_associativity(struct device_node *dev) +static const __be32 *of_get_associativity(struct device_node *dev) { return of_get_property(dev, "ibm,associativity", NULL); } @@ -169,41 +204,91 @@ static const int *of_get_associativity(struct device_node *dev) * it exists (the property exists only in kexec/kdump kernels, * added by kexec-tools) */ -static const u32 *of_get_usable_memory(struct device_node *memory) +static const __be32 *of_get_usable_memory(struct device_node *memory) { - const u32 *prop; + const __be32 *prop; u32 len; prop = of_get_property(memory, "linux,drconf-usable-memory", &len); if (!prop || len < sizeof(unsigned int)) - return 0; + return NULL; return prop; } +int __node_distance(int a, int b) +{ + int i; + int distance = LOCAL_DISTANCE; + + if (!form1_affinity) + return ((a == b) ? LOCAL_DISTANCE : REMOTE_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; +} +EXPORT_SYMBOL(__node_distance); + +static void initialize_distance_lookup_table(int nid, + const __be32 *associativity) +{ + int i; + + if (!form1_affinity) + return; + + for (i = 0; i < distance_ref_points_depth; i++) { + const __be32 *entry; + + entry = &associativity[be32_to_cpu(distance_ref_points[i])]; + distance_lookup_table[nid][i] = of_read_number(entry, 1); + } +} + /* Returns nid in the range [0..MAX_NUMNODES-1], or -1 if no useful numa * info is found. */ -static int of_node_to_nid_single(struct device_node *device) +static int associativity_to_nid(const __be32 *associativity) { int nid = -1; - const unsigned int *tmp; if (min_common_depth == -1) goto out; - tmp = of_get_associativity(device); - if (!tmp) - goto out; - - if (tmp[0] >= min_common_depth) - nid = tmp[min_common_depth]; + if (of_read_number(associativity, 1) >= min_common_depth) + nid = of_read_number(&associativity[min_common_depth], 1); /* POWER4 LPAR uses 0xffff as invalid node */ if (nid == 0xffff || nid >= MAX_NUMNODES) nid = -1; + + if (nid > 0 && + of_read_number(associativity, 1) >= distance_ref_points_depth) + initialize_distance_lookup_table(nid, associativity); + out: return nid; } +/* Returns the nid associated with the given device tree node, + * or -1 if not found. + */ +static int of_node_to_nid_single(struct device_node *device) +{ + int nid = -1; + const __be32 *tmp; + + tmp = of_get_associativity(device); + if (tmp) + nid = associativity_to_nid(tmp); + return nid; +} + /* Walk the device tree upwards, looking for an associativity id */ int of_node_to_nid(struct device_node *device) { @@ -226,50 +311,75 @@ 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; - const unsigned int *ref_points; - struct device_node *rtas_root; - unsigned int len; - - rtas_root = of_find_node_by_path("/rtas"); + struct device_node *root; - if (!rtas_root) - return -1; + if (firmware_has_feature(FW_FEATURE_OPAL)) + root = of_find_node_by_path("/ibm,opal"); + else + root = of_find_node_by_path("/rtas"); + if (!root) + root = of_find_node_by_path("/"); /* - * 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. */ - ref_points = of_get_property(rtas_root, - "ibm,associativity-reference-points", &len); + distance_ref_points = of_get_property(root, + "ibm,associativity-reference-points", + &distance_ref_points_depth); - if ((len >= 2 * sizeof(unsigned int)) && ref_points) { - depth = ref_points[1]; - } else { + if (!distance_ref_points) { dbg("NUMA: ibm,associativity-reference-points not found.\n"); - depth = -1; + goto err; + } + + distance_ref_points_depth /= sizeof(int); + + if (firmware_has_feature(FW_FEATURE_OPAL) || + firmware_has_feature(FW_FEATURE_TYPE1_AFFINITY)) { + dbg("Using form 1 affinity\n"); + form1_affinity = 1; + } + + if (form1_affinity) { + depth = of_read_number(distance_ref_points, 1); + } else { + if (distance_ref_points_depth < 2) { + printk(KERN_WARNING "NUMA: " + "short ibm,associativity-reference-points\n"); + goto err; + } + + depth = of_read_number(&distance_ref_points[1], 1); + } + + /* + * 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); + of_node_put(root); return depth; + +err: + of_node_put(root); + return -1; } static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells) @@ -285,65 +395,53 @@ static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells) of_node_put(memory); } -static unsigned long __devinit read_n_cells(int n, const unsigned int **buf) +static unsigned long read_n_cells(int n, const __be32 **buf) { unsigned long result = 0; while (n--) { - result = (result << 32) | **buf; + result = (result << 32) | of_read_number(*buf, 1); (*buf)++; } return result; } -struct of_drconf_cell { - u64 base_addr; - u32 drc_index; - u32 reserved; - u32 aa_index; - u32 flags; -}; - -#define DRCONF_MEM_ASSIGNED 0x00000008 -#define DRCONF_MEM_AI_INVALID 0x00000040 -#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) +static void read_drconf_cell(struct of_drconf_cell *drmem, const __be32 **cellp) { - const u32 *cp; + const __be32 *cp; drmem->base_addr = read_n_cells(n_mem_addr_cells, cellp); cp = *cellp; - drmem->drc_index = cp[0]; - drmem->reserved = cp[1]; - drmem->aa_index = cp[2]; - drmem->flags = cp[3]; + drmem->drc_index = of_read_number(cp, 1); + drmem->reserved = of_read_number(&cp[1], 1); + drmem->aa_index = of_read_number(&cp[2], 1); + drmem->flags = of_read_number(&cp[3], 1); *cellp = cp + 4; } /* - * Retreive and validate the ibm,dynamic-memory property of the device tree. + * Retrieve 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 - * contains information as layed out in the of_drconf_cell struct above. + * 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 laid out in the of_drconf_cell struct above. */ -static int of_get_drconf_memory(struct device_node *memory, const u32 **dm) +static int of_get_drconf_memory(struct device_node *memory, const __be32 **dm) { - const u32 *prop; + const __be32 *prop; u32 len, entries; prop = of_get_property(memory, "ibm,dynamic-memory", &len); if (!prop || len < sizeof(unsigned int)) return 0; - entries = *prop++; + entries = of_read_number(prop++, 1); /* Now that we know the number of entries, revalidate the size * of the property read in to ensure we have everything @@ -356,12 +454,12 @@ static int of_get_drconf_memory(struct device_node *memory, const u32 **dm) } /* - * Retreive and validate the ibm,lmb-size property for drconf memory + * Retrieve and validate the ibm,lmb-size property for drconf memory * from the device tree. */ static u64 of_get_lmb_size(struct device_node *memory) { - const u32 *prop; + const __be32 *prop; u32 len; prop = of_get_property(memory, "ibm,lmb-size", &len); @@ -374,11 +472,11 @@ static u64 of_get_lmb_size(struct device_node *memory) struct assoc_arrays { u32 n_arrays; u32 array_sz; - const u32 *arrays; + const __be32 *arrays; }; /* - * Retreive and validate the list of associativity arrays for drconf + * Retrieve and validate the list of associativity arrays for drconf * memory from the ibm,associativity-lookup-arrays property of the * device tree.. * @@ -390,17 +488,17 @@ struct assoc_arrays { static int of_get_assoc_arrays(struct device_node *memory, struct assoc_arrays *aa) { - const u32 *prop; + const __be32 *prop; u32 len; prop = of_get_property(memory, "ibm,associativity-lookup-arrays", &len); if (!prop || len < 2 * sizeof(unsigned int)) return -1; - aa->n_arrays = *prop++; - aa->array_sz = *prop++; + aa->n_arrays = of_read_number(prop++, 1); + aa->array_sz = of_read_number(prop++, 1); - /* Now that we know the number of arrrays and size of each array, + /* Now that we know the number of arrays and size of each array, * revalidate the size of the property read in. */ if (len < (aa->n_arrays * aa->array_sz + 2) * sizeof(unsigned int)) @@ -425,7 +523,7 @@ static int of_drconf_to_nid_single(struct of_drconf_cell *drmem, !(drmem->flags & DRCONF_MEM_AI_INVALID) && drmem->aa_index < aa->n_arrays) { index = drmem->aa_index * aa->array_sz + min_common_depth - 1; - nid = aa->arrays[index]; + nid = of_read_number(&aa->arrays[index], 1); if (nid == 0xffff || nid >= MAX_NUMNODES) nid = default_nid; @@ -438,20 +536,33 @@ static int of_drconf_to_nid_single(struct of_drconf_cell *drmem, * Figure out to which domain a cpu belongs and stick it there. * Return the id of the domain used. */ -static int __cpuinit numa_setup_cpu(unsigned long lcpu) +static int numa_setup_cpu(unsigned long lcpu) { - int nid = 0; - struct device_node *cpu = of_get_cpu_node(lcpu, NULL); + int nid; + struct device_node *cpu; + + /* + * If a valid cpu-to-node mapping is already available, use it + * directly instead of querying the firmware, since it represents + * the most recent mapping notified to us by the platform (eg: VPHN). + */ + if ((nid = numa_cpu_lookup_table[lcpu]) >= 0) { + map_cpu_to_node(lcpu, nid); + return nid; + } + + cpu = of_get_cpu_node(lcpu, NULL); if (!cpu) { WARN_ON(1); + nid = 0; goto out; } nid = of_node_to_nid_single(cpu); if (nid < 0 || !node_online(nid)) - nid = any_online_node(NODE_MASK_ALL); + nid = first_online_node; out: map_cpu_to_node(lcpu, nid); @@ -460,17 +571,38 @@ out: return nid; } -static int __cpuinit cpu_numa_callback(struct notifier_block *nfb, - unsigned long action, +static void verify_cpu_node_mapping(int cpu, int node) +{ + int base, sibling, i; + + /* Verify that all the threads in the core belong to the same node */ + base = cpu_first_thread_sibling(cpu); + + for (i = 0; i < threads_per_core; i++) { + sibling = base + i; + + if (sibling == cpu || cpu_is_offline(sibling)) + continue; + + if (cpu_to_node(sibling) != node) { + WARN(1, "CPU thread siblings %d and %d don't belong" + " to the same node!\n", cpu, sibling); + break; + } + } +} + +static int cpu_numa_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { unsigned long lcpu = (unsigned long)hcpu; - int ret = NOTIFY_DONE; + int ret = NOTIFY_DONE, nid; switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: - numa_setup_cpu(lcpu); + nid = numa_setup_cpu(lcpu); + verify_cpu_node_mapping((int)lcpu, nid); ret = NOTIFY_OK; break; #ifdef CONFIG_HOTPLUG_CPU @@ -492,32 +624,32 @@ static int __cpuinit cpu_numa_callback(struct notifier_block *nfb, * Returns the size the region should have to enforce the memory limit. * This will either be the original value of size, a truncated value, * or zero. If the returned value of size is 0 the region should be - * discarded as it lies wholy above the memory limit. + * discarded as it lies wholly above the memory limit. */ 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; } /* * Reads the counter for a given entry in * linux,drconf-usable-memory property */ -static inline int __init read_usm_ranges(const u32 **usm) +static inline int __init read_usm_ranges(const __be32 **usm) { /* * For each lmb in ibm,dynamic-memory a corresponding @@ -534,11 +666,11 @@ static inline int __init read_usm_ranges(const u32 **usm) */ static void __init parse_drconf_memory(struct device_node *memory) { - const u32 *dm, *usm; + const __be32 *uninitialized_var(dm), *usm; unsigned int n, rc, ranges, is_kexec_kdump = 0; unsigned long lmb_size, base, size, sz; int nid; - struct assoc_arrays aa; + struct assoc_arrays aa = { .arrays = NULL }; n = of_get_drconf_memory(memory, &dm); if (!n) @@ -589,17 +721,15 @@ static void __init parse_drconf_memory(struct device_node *memory) node_set_online(nid); sz = numa_enforce_memory_limit(base, size); if (sz) - add_active_range(nid, base >> PAGE_SHIFT, - (base >> PAGE_SHIFT) - + (sz >> PAGE_SHIFT)); + memblock_set_node(base, sz, + &memblock.memory, nid); } while (--ranges); } } static int __init parse_numa_properties(void) { - struct device_node *cpu = NULL; - struct device_node *memory = NULL; + struct device_node *memory; int default_nid = 0; unsigned long i; @@ -621,6 +751,7 @@ static int __init parse_numa_properties(void) * each node to be onlined must have NODE_DATA etc backing it. */ for_each_present_cpu(i) { + struct device_node *cpu; int nid; cpu = of_get_cpu_node(i, NULL); @@ -639,13 +770,13 @@ static int __init parse_numa_properties(void) } get_n_mem_cells(&n_mem_addr_cells, &n_mem_size_cells); - memory = NULL; - while ((memory = of_find_node_by_type(memory, "memory")) != NULL) { + + for_each_node_by_type(memory, "memory") { unsigned long start; unsigned long size; int nid; int ranges; - const unsigned int *memcell_buf; + const __be32 *memcell_buf; unsigned int len; memcell_buf = of_get_property(memory, @@ -681,16 +812,16 @@ new_range: continue; } - add_active_range(nid, start >> PAGE_SHIFT, - (start >> PAGE_SHIFT) + (size >> PAGE_SHIFT)); + memblock_set_node(start, size, &memblock.memory, nid); if (--ranges) goto new_range; } /* - * Now do the same thing for each LMB listed in the ibm,dynamic-memory - * property in the ibm,dynamic-reconfiguration-memory node. + * 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"); if (memory) @@ -701,22 +832,25 @@ 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; + unsigned int nid = 0; + struct memblock_region *reg; printk(KERN_DEBUG "Top of RAM: 0x%lx, Total RAM: 0x%lx\n", top_of_ram, total_ram); 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_each_memblock(memory, reg) { + start_pfn = memblock_region_memory_base_pfn(reg); + end_pfn = memblock_region_memory_end_pfn(reg); fake_numa_create_new_node(end_pfn, &nid); - add_active_range(nid, start_pfn, end_pfn); + memblock_set_node(PFN_PHYS(start_pfn), + PFN_PHYS(end_pfn - start_pfn), + &memblock.memory, nid); node_set_online(nid); } } @@ -737,8 +871,9 @@ void __init dump_numa_cpu_topology(void) * 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])) { + for (cpu = 0; cpu < nr_cpu_ids; cpu++) { + if (cpumask_test_cpu(cpu, + node_to_cpumask_map[node])) { if (count == 0) printk(" %u", cpu); ++count; @@ -750,7 +885,7 @@ void __init dump_numa_cpu_topology(void) } if (count > 1) - printk("-%u", NR_CPUS - 1); + printk("-%u", nr_cpu_ids - 1); printk("\n"); } } @@ -770,7 +905,7 @@ 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) @@ -790,7 +925,7 @@ 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. * @@ -804,11 +939,11 @@ 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", @@ -818,14 +953,14 @@ 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); @@ -840,27 +975,26 @@ static void __init *careful_zallocation(int nid, unsigned long size, return ret; } -static struct notifier_block __cpuinitdata ppc64_numa_nb = { +static struct notifier_block ppc64_numa_nb = { .notifier_call = cpu_numa_callback, .priority = 1 /* Must run before sched domains notifier. */ }; -static void mark_reserved_regions_for_nid(int nid) +static void __init mark_reserved_regions_for_nid(int nid) { struct pglist_data *node = NODE_DATA(nid); - int i; + struct memblock_region *reg; - 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_each_memblock(reserved, reg) { + unsigned long physbase = reg->base; + unsigned long size = reg->size; unsigned long start_pfn = physbase >> PAGE_SHIFT; unsigned long end_pfn = PFN_UP(physbase + size); struct node_active_region node_ar; - unsigned long node_end_pfn = node->node_start_pfn + - node->node_spanned_pages; + unsigned long node_end_pfn = pgdat_end_pfn(node); /* - * 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 @@ -918,7 +1052,7 @@ 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()) @@ -926,10 +1060,6 @@ void __init do_init_bootmem(void) else 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; void *bootmem_vaddr; @@ -983,13 +1113,24 @@ void __init do_init_bootmem(void) } init_bootmem_done = 1; + + /* + * Now bootmem is initialised we can create the node to cpumask + * lookup tables and setup the cpu callback to populate them. + */ + setup_node_to_cpumask_map(); + + reset_numa_cpu_lookup_table(); + register_cpu_notifier(&ppc64_numa_nb); + cpu_numa_callback(&ppc64_numa_nb, CPU_UP_PREPARE, + (void *)(unsigned long)boot_cpuid); } 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); } @@ -1021,7 +1162,7 @@ early_param("numa", early_numa); static int hot_add_drconf_scn_to_nid(struct device_node *memory, unsigned long scn_addr) { - const u32 *dm; + const __be32 *dm; unsigned int drconf_cell_cnt, rc; unsigned long lmb_size; struct assoc_arrays aa; @@ -1064,17 +1205,17 @@ 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) +static int hot_add_node_scn_to_nid(unsigned long scn_addr) { - struct device_node *memory = NULL; + struct device_node *memory; int nid = -1; - while ((memory = of_find_node_by_type(memory, "memory")) != NULL) { + for_each_node_by_type(memory, "memory") { unsigned long start, size; int ranges; - const unsigned int *memcell_buf; + const __be32 *memcell_buf; unsigned int len; memcell_buf = of_get_property(memory, "reg", &len); @@ -1095,18 +1236,19 @@ int hot_add_node_scn_to_nid(unsigned long scn_addr) break; } - of_node_put(memory); if (nid >= 0) break; } + of_node_put(memory); + return nid; } /* * 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) { @@ -1114,7 +1256,7 @@ int hot_add_scn_to_nid(unsigned long scn_addr) int nid, found = 0; if (!numa_enabled || (min_common_depth < 0)) - return any_online_node(NODE_MASK_ALL); + return first_online_node; memory = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory"); if (memory) { @@ -1125,7 +1267,7 @@ int hot_add_scn_to_nid(unsigned long scn_addr) } if (nid < 0 || !node_online(nid)) - nid = any_online_node(NODE_MASK_ALL); + nid = first_online_node; if (NODE_DATA(nid)->node_spanned_pages) return nid; @@ -1141,4 +1283,527 @@ int hot_add_scn_to_nid(unsigned long scn_addr) return nid; } +static u64 hot_add_drconf_memory_max(void) +{ + struct device_node *memory = NULL; + unsigned int drconf_cell_cnt = 0; + u64 lmb_size = 0; + const __be32 *dm = NULL; + + memory = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory"); + if (memory) { + drconf_cell_cnt = of_get_drconf_memory(memory, &dm); + lmb_size = of_get_lmb_size(memory); + of_node_put(memory); + } + return lmb_size * drconf_cell_cnt; +} + +/* + * memory_hotplug_max - return max address of memory that may be added + * + * This is currently only used on systems that support drconfig memory + * hotplug. + */ +u64 memory_hotplug_max(void) +{ + return max(hot_add_drconf_memory_max(), memblock_end_of_DRAM()); +} #endif /* CONFIG_MEMORY_HOTPLUG */ + +/* Virtual Processor Home Node (VPHN) support */ +#ifdef CONFIG_PPC_SPLPAR +struct topology_update_data { + struct topology_update_data *next; + unsigned int cpu; + int old_nid; + int new_nid; +}; + +static u8 vphn_cpu_change_counts[NR_CPUS][MAX_DISTANCE_REF_POINTS]; +static cpumask_t cpu_associativity_changes_mask; +static int vphn_enabled; +static int prrn_enabled; +static void reset_topology_timer(void); + +/* + * Store the current values of the associativity change counters in the + * hypervisor. + */ +static void setup_cpu_associativity_change_counters(void) +{ + int cpu; + + /* The VPHN feature supports a maximum of 8 reference points */ + BUILD_BUG_ON(MAX_DISTANCE_REF_POINTS > 8); + + for_each_possible_cpu(cpu) { + int i; + u8 *counts = vphn_cpu_change_counts[cpu]; + volatile u8 *hypervisor_counts = lppaca[cpu].vphn_assoc_counts; + + for (i = 0; i < distance_ref_points_depth; i++) + counts[i] = hypervisor_counts[i]; + } +} + +/* + * The hypervisor maintains a set of 8 associativity change counters in + * the VPA of each cpu that correspond to the associativity levels in the + * ibm,associativity-reference-points property. When an associativity + * level changes, the corresponding counter is incremented. + * + * Set a bit in cpu_associativity_changes_mask for each cpu whose home + * node associativity levels have changed. + * + * Returns the number of cpus with unhandled associativity changes. + */ +static int update_cpu_associativity_changes_mask(void) +{ + int cpu; + cpumask_t *changes = &cpu_associativity_changes_mask; + + for_each_possible_cpu(cpu) { + int i, changed = 0; + u8 *counts = vphn_cpu_change_counts[cpu]; + volatile u8 *hypervisor_counts = lppaca[cpu].vphn_assoc_counts; + + for (i = 0; i < distance_ref_points_depth; i++) { + if (hypervisor_counts[i] != counts[i]) { + counts[i] = hypervisor_counts[i]; + changed = 1; + } + } + if (changed) { + cpumask_or(changes, changes, cpu_sibling_mask(cpu)); + cpu = cpu_last_thread_sibling(cpu); + } + } + + return cpumask_weight(changes); +} + +/* + * 6 64-bit registers unpacked into 12 32-bit associativity values. To form + * the complete property we have to add the length in the first cell. + */ +#define VPHN_ASSOC_BUFSIZE (6*sizeof(u64)/sizeof(u32) + 1) + +/* + * Convert the associativity domain numbers returned from the hypervisor + * to the sequence they would appear in the ibm,associativity property. + */ +static int vphn_unpack_associativity(const long *packed, __be32 *unpacked) +{ + int i, nr_assoc_doms = 0; + const __be16 *field = (const __be16 *) packed; + +#define VPHN_FIELD_UNUSED (0xffff) +#define VPHN_FIELD_MSB (0x8000) +#define VPHN_FIELD_MASK (~VPHN_FIELD_MSB) + + for (i = 1; i < VPHN_ASSOC_BUFSIZE; i++) { + if (be16_to_cpup(field) == VPHN_FIELD_UNUSED) { + /* All significant fields processed, and remaining + * fields contain the reserved value of all 1's. + * Just store them. + */ + unpacked[i] = *((__be32 *)field); + field += 2; + } else if (be16_to_cpup(field) & VPHN_FIELD_MSB) { + /* Data is in the lower 15 bits of this field */ + unpacked[i] = cpu_to_be32( + be16_to_cpup(field) & VPHN_FIELD_MASK); + field++; + nr_assoc_doms++; + } else { + /* Data is in the lower 15 bits of this field + * concatenated with the next 16 bit field + */ + unpacked[i] = *((__be32 *)field); + field += 2; + nr_assoc_doms++; + } + } + + /* The first cell contains the length of the property */ + unpacked[0] = cpu_to_be32(nr_assoc_doms); + + return nr_assoc_doms; +} + +/* + * Retrieve the new associativity information for a virtual processor's + * home node. + */ +static long hcall_vphn(unsigned long cpu, __be32 *associativity) +{ + long rc; + long retbuf[PLPAR_HCALL9_BUFSIZE] = {0}; + u64 flags = 1; + int hwcpu = get_hard_smp_processor_id(cpu); + + rc = plpar_hcall9(H_HOME_NODE_ASSOCIATIVITY, retbuf, flags, hwcpu); + vphn_unpack_associativity(retbuf, associativity); + + return rc; +} + +static long vphn_get_associativity(unsigned long cpu, + __be32 *associativity) +{ + long rc; + + rc = hcall_vphn(cpu, associativity); + + switch (rc) { + case H_FUNCTION: + printk(KERN_INFO + "VPHN is not supported. Disabling polling...\n"); + stop_topology_update(); + break; + case H_HARDWARE: + printk(KERN_ERR + "hcall_vphn() experienced a hardware fault " + "preventing VPHN. Disabling polling...\n"); + stop_topology_update(); + } + + return rc; +} + +/* + * Update the CPU maps and sysfs entries for a single CPU when its NUMA + * characteristics change. This function doesn't perform any locking and is + * only safe to call from stop_machine(). + */ +static int update_cpu_topology(void *data) +{ + struct topology_update_data *update; + unsigned long cpu; + + if (!data) + return -EINVAL; + + cpu = smp_processor_id(); + + for (update = data; update; update = update->next) { + if (cpu != update->cpu) + continue; + + unmap_cpu_from_node(update->cpu); + map_cpu_to_node(update->cpu, update->new_nid); + vdso_getcpu_init(); + } + + return 0; +} + +static int update_lookup_table(void *data) +{ + struct topology_update_data *update; + + if (!data) + return -EINVAL; + + /* + * Upon topology update, the numa-cpu lookup table needs to be updated + * for all threads in the core, including offline CPUs, to ensure that + * future hotplug operations respect the cpu-to-node associativity + * properly. + */ + for (update = data; update; update = update->next) { + int nid, base, j; + + nid = update->new_nid; + base = cpu_first_thread_sibling(update->cpu); + + for (j = 0; j < threads_per_core; j++) { + update_numa_cpu_lookup_table(base + j, nid); + } + } + + return 0; +} + +/* + * Update the node maps and sysfs entries for each cpu whose home node + * has changed. Returns 1 when the topology has changed, and 0 otherwise. + */ +int arch_update_cpu_topology(void) +{ + unsigned int cpu, sibling, changed = 0; + struct topology_update_data *updates, *ud; + __be32 associativity[VPHN_ASSOC_BUFSIZE] = {0}; + cpumask_t updated_cpus; + struct device *dev; + int weight, new_nid, i = 0; + + weight = cpumask_weight(&cpu_associativity_changes_mask); + if (!weight) + return 0; + + updates = kzalloc(weight * (sizeof(*updates)), GFP_KERNEL); + if (!updates) + return 0; + + cpumask_clear(&updated_cpus); + + for_each_cpu(cpu, &cpu_associativity_changes_mask) { + /* + * If siblings aren't flagged for changes, updates list + * will be too short. Skip on this update and set for next + * update. + */ + if (!cpumask_subset(cpu_sibling_mask(cpu), + &cpu_associativity_changes_mask)) { + pr_info("Sibling bits not set for associativity " + "change, cpu%d\n", cpu); + cpumask_or(&cpu_associativity_changes_mask, + &cpu_associativity_changes_mask, + cpu_sibling_mask(cpu)); + cpu = cpu_last_thread_sibling(cpu); + continue; + } + + /* Use associativity from first thread for all siblings */ + vphn_get_associativity(cpu, associativity); + new_nid = associativity_to_nid(associativity); + if (new_nid < 0 || !node_online(new_nid)) + new_nid = first_online_node; + + if (new_nid == numa_cpu_lookup_table[cpu]) { + cpumask_andnot(&cpu_associativity_changes_mask, + &cpu_associativity_changes_mask, + cpu_sibling_mask(cpu)); + cpu = cpu_last_thread_sibling(cpu); + continue; + } + + for_each_cpu(sibling, cpu_sibling_mask(cpu)) { + ud = &updates[i++]; + ud->cpu = sibling; + ud->new_nid = new_nid; + ud->old_nid = numa_cpu_lookup_table[sibling]; + cpumask_set_cpu(sibling, &updated_cpus); + if (i < weight) + ud->next = &updates[i]; + } + cpu = cpu_last_thread_sibling(cpu); + } + + /* + * In cases where we have nothing to update (because the updates list + * is too short or because the new topology is same as the old one), + * skip invoking update_cpu_topology() via stop-machine(). This is + * necessary (and not just a fast-path optimization) since stop-machine + * can end up electing a random CPU to run update_cpu_topology(), and + * thus trick us into setting up incorrect cpu-node mappings (since + * 'updates' is kzalloc()'ed). + * + * And for the similar reason, we will skip all the following updating. + */ + if (!cpumask_weight(&updated_cpus)) + goto out; + + stop_machine(update_cpu_topology, &updates[0], &updated_cpus); + + /* + * Update the numa-cpu lookup table with the new mappings, even for + * offline CPUs. It is best to perform this update from the stop- + * machine context. + */ + stop_machine(update_lookup_table, &updates[0], + cpumask_of(raw_smp_processor_id())); + + for (ud = &updates[0]; ud; ud = ud->next) { + unregister_cpu_under_node(ud->cpu, ud->old_nid); + register_cpu_under_node(ud->cpu, ud->new_nid); + + dev = get_cpu_device(ud->cpu); + if (dev) + kobject_uevent(&dev->kobj, KOBJ_CHANGE); + cpumask_clear_cpu(ud->cpu, &cpu_associativity_changes_mask); + changed = 1; + } + +out: + kfree(updates); + return changed; +} + +static void topology_work_fn(struct work_struct *work) +{ + rebuild_sched_domains(); +} +static DECLARE_WORK(topology_work, topology_work_fn); + +static void topology_schedule_update(void) +{ + schedule_work(&topology_work); +} + +static void topology_timer_fn(unsigned long ignored) +{ + if (prrn_enabled && cpumask_weight(&cpu_associativity_changes_mask)) + topology_schedule_update(); + else if (vphn_enabled) { + if (update_cpu_associativity_changes_mask() > 0) + topology_schedule_update(); + reset_topology_timer(); + } +} +static struct timer_list topology_timer = + TIMER_INITIALIZER(topology_timer_fn, 0, 0); + +static void reset_topology_timer(void) +{ + topology_timer.data = 0; + topology_timer.expires = jiffies + 60 * HZ; + mod_timer(&topology_timer, topology_timer.expires); +} + +#ifdef CONFIG_SMP + +static void stage_topology_update(int core_id) +{ + cpumask_or(&cpu_associativity_changes_mask, + &cpu_associativity_changes_mask, cpu_sibling_mask(core_id)); + reset_topology_timer(); +} + +static int dt_update_callback(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct of_prop_reconfig *update; + int rc = NOTIFY_DONE; + + switch (action) { + case OF_RECONFIG_UPDATE_PROPERTY: + update = (struct of_prop_reconfig *)data; + if (!of_prop_cmp(update->dn->type, "cpu") && + !of_prop_cmp(update->prop->name, "ibm,associativity")) { + u32 core_id; + of_property_read_u32(update->dn, "reg", &core_id); + stage_topology_update(core_id); + rc = NOTIFY_OK; + } + break; + } + + return rc; +} + +static struct notifier_block dt_update_nb = { + .notifier_call = dt_update_callback, +}; + +#endif + +/* + * Start polling for associativity changes. + */ +int start_topology_update(void) +{ + int rc = 0; + + if (firmware_has_feature(FW_FEATURE_PRRN)) { + if (!prrn_enabled) { + prrn_enabled = 1; + vphn_enabled = 0; +#ifdef CONFIG_SMP + rc = of_reconfig_notifier_register(&dt_update_nb); +#endif + } + } else if (firmware_has_feature(FW_FEATURE_VPHN) && + lppaca_shared_proc(get_lppaca())) { + if (!vphn_enabled) { + prrn_enabled = 0; + vphn_enabled = 1; + setup_cpu_associativity_change_counters(); + init_timer_deferrable(&topology_timer); + reset_topology_timer(); + } + } + + return rc; +} + +/* + * Disable polling for VPHN associativity changes. + */ +int stop_topology_update(void) +{ + int rc = 0; + + if (prrn_enabled) { + prrn_enabled = 0; +#ifdef CONFIG_SMP + rc = of_reconfig_notifier_unregister(&dt_update_nb); +#endif + } else if (vphn_enabled) { + vphn_enabled = 0; + rc = del_timer_sync(&topology_timer); + } + + return rc; +} + +int prrn_is_enabled(void) +{ + return prrn_enabled; +} + +static int topology_read(struct seq_file *file, void *v) +{ + if (vphn_enabled || prrn_enabled) + seq_puts(file, "on\n"); + else + seq_puts(file, "off\n"); + + return 0; +} + +static int topology_open(struct inode *inode, struct file *file) +{ + return single_open(file, topology_read, NULL); +} + +static ssize_t topology_write(struct file *file, const char __user *buf, + size_t count, loff_t *off) +{ + char kbuf[4]; /* "on" or "off" plus null. */ + int read_len; + + read_len = count < 3 ? count : 3; + if (copy_from_user(kbuf, buf, read_len)) + return -EINVAL; + + kbuf[read_len] = '\0'; + + if (!strncmp(kbuf, "on", 2)) + start_topology_update(); + else if (!strncmp(kbuf, "off", 3)) + stop_topology_update(); + else + return -EINVAL; + + return count; +} + +static const struct file_operations topology_ops = { + .read = seq_read, + .write = topology_write, + .open = topology_open, + .release = single_release +}; + +static int topology_update_init(void) +{ + start_topology_update(); + proc_create("powerpc/topology_updates", 0644, NULL, &topology_ops); + + return 0; +} +device_initcall(topology_update_init); +#endif /* CONFIG_PPC_SPLPAR */ |