diff options
Diffstat (limited to 'arch/powerpc/mm/numa.c')
| -rw-r--r-- | arch/powerpc/mm/numa.c | 556 |
1 files changed, 422 insertions, 134 deletions
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index bba87ca2b4d..3b181b22cd4 100644 --- a/arch/powerpc/mm/numa.c +++ b/arch/powerpc/mm/numa.c @@ -22,13 +22,22 @@ #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/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; @@ -51,7 +60,7 @@ static int form1_affinity; #define MAX_DISTANCE_REF_POINTS 4 static int distance_ref_points_depth; -static const unsigned int *distance_ref_points; +static const __be32 *distance_ref_points; static int distance_lookup_table[MAX_NUMNODES][MAX_DISTANCE_REF_POINTS]; /* @@ -62,14 +71,11 @@ static int distance_lookup_table[MAX_NUMNODES][MAX_DISTANCE_REF_POINTS]; */ static void __init setup_node_to_cpumask_map(void) { - unsigned int node, num = 0; + unsigned int node; /* setup nr_node_ids if not done yet */ - if (nr_node_ids == MAX_NUMNODES) { - for_each_node_mask(node, node_possible_map) - num = node; - nr_node_ids = num + 1; - } + if (nr_node_ids == MAX_NUMNODES) + setup_nr_node_ids(); /* allocate the map */ for (node = 0; node < nr_node_ids; node++) @@ -79,7 +85,7 @@ static void __init setup_node_to_cpumask_map(void) dbg("Node to cpumask map for %d nodes\n", nr_node_ids); } -static int __cpuinit fake_numa_create_new_node(unsigned long end_pfn, +static int __init fake_numa_create_new_node(unsigned long end_pfn, unsigned int *nid) { unsigned long long mem; @@ -148,9 +154,22 @@ static void __init get_node_active_region(unsigned long pfn, } } -static void map_cpu_to_node(int cpu, int node) +static void reset_numa_cpu_lookup_table(void) +{ + unsigned int cpu; + + for_each_possible_cpu(cpu) + numa_cpu_lookup_table[cpu] = -1; +} + +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); @@ -175,7 +194,7 @@ static void unmap_cpu_from_node(unsigned long 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); } @@ -185,13 +204,13 @@ 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; } @@ -201,7 +220,7 @@ int __node_distance(int a, int b) int distance = LOCAL_DISTANCE; if (!form1_affinity) - return distance; + 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]) @@ -213,9 +232,10 @@ int __node_distance(int a, int b) return distance; } +EXPORT_SYMBOL(__node_distance); static void initialize_distance_lookup_table(int nid, - const unsigned int *associativity) + const __be32 *associativity) { int i; @@ -223,29 +243,32 @@ static void initialize_distance_lookup_table(int nid, return; for (i = 0; i < distance_ref_points_depth; i++) { - distance_lookup_table[nid][i] = - associativity[distance_ref_points[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 associativity_to_nid(const unsigned int *associativity) +static int associativity_to_nid(const __be32 *associativity) { int nid = -1; if (min_common_depth == -1) goto out; - if (associativity[0] >= min_common_depth) - nid = associativity[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 && associativity[0] >= distance_ref_points_depth) + if (nid > 0 && + of_read_number(associativity, 1) >= distance_ref_points_depth) initialize_distance_lookup_table(nid, associativity); out: @@ -258,7 +281,7 @@ out: static int of_node_to_nid_single(struct device_node *device) { int nid = -1; - const unsigned int *tmp; + const __be32 *tmp; tmp = of_get_associativity(device); if (tmp) @@ -291,9 +314,7 @@ EXPORT_SYMBOL_GPL(of_node_to_nid); static int __init find_min_common_depth(void) { int depth; - struct device_node *chosen; struct device_node *root; - const char *vec5; if (firmware_has_feature(FW_FEATURE_OPAL)) root = of_find_node_by_path("/ibm,opal"); @@ -325,28 +346,14 @@ static int __init find_min_common_depth(void) distance_ref_points_depth /= sizeof(int); -#define VEC5_AFFINITY_BYTE 5 -#define VEC5_AFFINITY 0x80 - - if (firmware_has_feature(FW_FEATURE_OPAL)) + if (firmware_has_feature(FW_FEATURE_OPAL) || + firmware_has_feature(FW_FEATURE_TYPE1_AFFINITY)) { + dbg("Using form 1 affinity\n"); form1_affinity = 1; - else { - chosen = of_find_node_by_path("/chosen"); - if (chosen) { - vec5 = of_get_property(chosen, - "ibm,architecture-vec-5", NULL); - if (vec5 && (vec5[VEC5_AFFINITY_BYTE] & - VEC5_AFFINITY)) { - dbg("Using form 1 affinity\n"); - form1_affinity = 1; - } - - of_node_put(chosen); - } } if (form1_affinity) { - depth = distance_ref_points[0]; + depth = of_read_number(distance_ref_points, 1); } else { if (distance_ref_points_depth < 2) { printk(KERN_WARNING "NUMA: " @@ -354,7 +361,7 @@ static int __init find_min_common_depth(void) goto err; } - depth = distance_ref_points[1]; + depth = of_read_number(&distance_ref_points[1], 1); } /* @@ -388,12 +395,12 @@ static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells) of_node_put(memory); } -static unsigned long 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; @@ -403,17 +410,17 @@ static unsigned long read_n_cells(int n, const unsigned int **buf) * 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; } @@ -425,16 +432,16 @@ static void read_drconf_cell(struct of_drconf_cell *drmem, const u32 **cellp) * 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 @@ -452,7 +459,7 @@ static int of_get_drconf_memory(struct device_node *memory, const u32 **dm) */ 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); @@ -465,7 +472,7 @@ 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; }; /* @@ -481,15 +488,15 @@ 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 arrays and size of each array, * revalidate the size of the property read in. @@ -516,7 +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; @@ -529,13 +536,26 @@ 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; } @@ -551,17 +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 @@ -608,7 +649,7 @@ static unsigned long __init numa_enforce_memory_limit(unsigned long 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 @@ -625,7 +666,7 @@ static inline int __init read_usm_ranges(const u32 **usm) */ static void __init parse_drconf_memory(struct device_node *memory) { - const u32 *uninitialized_var(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; @@ -680,7 +721,8 @@ static void __init parse_drconf_memory(struct device_node *memory) node_set_online(nid); sz = numa_enforce_memory_limit(base, size); if (sz) - memblock_set_node(base, sz, nid); + memblock_set_node(base, sz, + &memblock.memory, nid); } while (--ranges); } } @@ -734,7 +776,7 @@ static int __init parse_numa_properties(void) 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, @@ -770,7 +812,7 @@ new_range: continue; } - memblock_set_node(start, size, nid); + memblock_set_node(start, size, &memblock.memory, nid); if (--ranges) goto new_range; @@ -807,7 +849,8 @@ static void __init setup_nonnuma(void) fake_numa_create_new_node(end_pfn, &nid); memblock_set_node(PFN_PHYS(start_pfn), - PFN_PHYS(end_pfn - start_pfn), nid); + PFN_PHYS(end_pfn - start_pfn), + &memblock.memory, nid); node_set_online(nid); } } @@ -932,7 +975,7 @@ 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. */ }; @@ -948,8 +991,7 @@ static void __init mark_reserved_regions_for_nid(int nid) 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 memblock.reserved area is @@ -1078,6 +1120,7 @@ void __init do_init_bootmem(void) */ 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); @@ -1119,7 +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; @@ -1164,7 +1207,7 @@ static int hot_add_drconf_scn_to_nid(struct device_node *memory, * represented in the device tree as a node (i.e. memory@XXXX) for * 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; int nid = -1; @@ -1172,7 +1215,7 @@ int hot_add_node_scn_to_nid(unsigned long scn_addr) 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); @@ -1245,7 +1288,7 @@ static u64 hot_add_drconf_memory_max(void) struct device_node *memory = NULL; unsigned int drconf_cell_cnt = 0; u64 lmb_size = 0; - const u32 *dm = 0; + const __be32 *dm = NULL; memory = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory"); if (memory) { @@ -1270,10 +1313,18 @@ u64 memory_hotplug_max(void) /* 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 void set_topology_timer(void); +static int prrn_enabled; +static void reset_topology_timer(void); /* * Store the current values of the associativity change counters in the @@ -1309,11 +1360,9 @@ static void setup_cpu_associativity_change_counters(void) */ static int update_cpu_associativity_changes_mask(void) { - int cpu, nr_cpus = 0; + int cpu; cpumask_t *changes = &cpu_associativity_changes_mask; - cpumask_clear(changes); - for_each_possible_cpu(cpu) { int i, changed = 0; u8 *counts = vphn_cpu_change_counts[cpu]; @@ -1326,12 +1375,12 @@ static int update_cpu_associativity_changes_mask(void) } } if (changed) { - cpumask_set_cpu(cpu, changes); - nr_cpus++; + cpumask_or(changes, changes, cpu_sibling_mask(cpu)); + cpu = cpu_last_thread_sibling(cpu); } } - return nr_cpus; + return cpumask_weight(changes); } /* @@ -1344,40 +1393,41 @@ static int update_cpu_associativity_changes_mask(void) * 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, unsigned int *unpacked) +static int vphn_unpack_associativity(const long *packed, __be32 *unpacked) { int i, nr_assoc_doms = 0; - const u16 *field = (const u16*) packed; + 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 (*field == VPHN_FIELD_UNUSED) { + 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] = *((u32*)field); + unpacked[i] = *((__be32 *)field); field += 2; - } else if (*field & VPHN_FIELD_MSB) { + } else if (be16_to_cpup(field) & VPHN_FIELD_MSB) { /* Data is in the lower 15 bits of this field */ - unpacked[i] = *field & VPHN_FIELD_MASK; + 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] = *((u32*)field); + unpacked[i] = *((__be32 *)field); field += 2; nr_assoc_doms++; } } /* The first cell contains the length of the property */ - unpacked[0] = nr_assoc_doms; + unpacked[0] = cpu_to_be32(nr_assoc_doms); return nr_assoc_doms; } @@ -1386,7 +1436,7 @@ static int vphn_unpack_associativity(const long *packed, unsigned int *unpacked) * Retrieve the new associativity information for a virtual processor's * home node. */ -static long hcall_vphn(unsigned long cpu, unsigned int *associativity) +static long hcall_vphn(unsigned long cpu, __be32 *associativity) { long rc; long retbuf[PLPAR_HCALL9_BUFSIZE] = {0}; @@ -1400,7 +1450,7 @@ static long hcall_vphn(unsigned long cpu, unsigned int *associativity) } static long vphn_get_associativity(unsigned long cpu, - unsigned int *associativity) + __be32 *associativity) { long rc; @@ -1423,40 +1473,162 @@ static long vphn_get_associativity(unsigned long cpu, } /* + * 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) { - int cpu, nid, old_nid, changed = 0; - unsigned int associativity[VPHN_ASSOC_BUFSIZE] = {0}; + 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; - for_each_cpu(cpu,&cpu_associativity_changes_mask) { - vphn_get_associativity(cpu, associativity); - nid = associativity_to_nid(associativity); + weight = cpumask_weight(&cpu_associativity_changes_mask); + if (!weight) + return 0; - if (nid < 0 || !node_online(nid)) - nid = first_online_node; + updates = kzalloc(weight * (sizeof(*updates)), GFP_KERNEL); + if (!updates) + return 0; - old_nid = numa_cpu_lookup_table[cpu]; + cpumask_clear(&updated_cpus); - /* Disable hotplug while we update the cpu - * masks and sysfs. + 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. */ - get_online_cpus(); - unregister_cpu_under_node(cpu, old_nid); - unmap_cpu_from_node(cpu); - map_cpu_to_node(cpu, nid); - register_cpu_under_node(cpu, nid); - put_online_cpus(); - - dev = get_cpu_device(cpu); + 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; } @@ -1466,56 +1638,172 @@ static void topology_work_fn(struct work_struct *work) } static DECLARE_WORK(topology_work, topology_work_fn); -void topology_schedule_update(void) +static void topology_schedule_update(void) { schedule_work(&topology_work); } static void topology_timer_fn(unsigned long ignored) { - if (!vphn_enabled) - return; - if (update_cpu_associativity_changes_mask() > 0) + if (prrn_enabled && cpumask_weight(&cpu_associativity_changes_mask)) topology_schedule_update(); - set_topology_timer(); + 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 set_topology_timer(void) +static void reset_topology_timer(void) { topology_timer.data = 0; topology_timer.expires = jiffies + 60 * HZ; - add_timer(&topology_timer); + 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 VPHN associativity changes. + * Start polling for associativity changes. */ int start_topology_update(void) { int rc = 0; - /* Disabled until races with load balancing are fixed */ - if (0 && firmware_has_feature(FW_FEATURE_VPHN) && - get_lppaca()->shared_proc) { - vphn_enabled = 1; - setup_cpu_associativity_change_counters(); - init_timer_deferrable(&topology_timer); - set_topology_timer(); - rc = 1; + 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; } -__initcall(start_topology_update); /* * Disable polling for VPHN associativity changes. */ int stop_topology_update(void) { - vphn_enabled = 0; - return del_timer_sync(&topology_timer); + 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 */ |