+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/crash_dump.h>
#include <linux/smp.h>
#include <linux/topology.h>
+#include <linux/pfn.h>
#include <asm/sections.h>
#include <asm/processor.h>
#include <asm/setup.h>
#include <asm/highmem.h>
#include <asm/proto.h>
#include <asm/cpumask.h>
-
-#ifdef CONFIG_DEBUG_PER_CPU_MAPS
-# define DBG(x...) printk(KERN_DEBUG x)
-#else
-# define DBG(x...)
-#endif
+#include <asm/cpu.h>
+#include <asm/stackprotector.h>
DEFINE_PER_CPU(int, cpu_number);
EXPORT_PER_CPU_SYMBOL(cpu_number);
-#ifdef CONFIG_HAVE_SETUP_PER_CPU_AREA
-
#ifdef CONFIG_X86_64
+#define BOOT_PERCPU_OFFSET ((unsigned long)__per_cpu_load)
+#else
+#define BOOT_PERCPU_OFFSET 0
+#endif
+
+DEFINE_PER_CPU(unsigned long, this_cpu_off) = BOOT_PERCPU_OFFSET;
+EXPORT_PER_CPU_SYMBOL(this_cpu_off);
+
unsigned long __per_cpu_offset[NR_CPUS] __read_mostly = {
- [0] = (unsigned long)__per_cpu_load,
+ [0 ... NR_CPUS-1] = BOOT_PERCPU_OFFSET,
};
+EXPORT_SYMBOL(__per_cpu_offset);
+
+/*
+ * On x86_64 symbols referenced from code should be reachable using
+ * 32bit relocations. Reserve space for static percpu variables in
+ * modules so that they are always served from the first chunk which
+ * is located at the percpu segment base. On x86_32, anything can
+ * address anywhere. No need to reserve space in the first chunk.
+ */
+#ifdef CONFIG_X86_64
+#define PERCPU_FIRST_CHUNK_RESERVE PERCPU_MODULE_RESERVE
#else
-unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
+#define PERCPU_FIRST_CHUNK_RESERVE 0
#endif
-EXPORT_SYMBOL(__per_cpu_offset);
+
+#ifdef CONFIG_X86_32
+/**
+ * pcpu_need_numa - determine percpu allocation needs to consider NUMA
+ *
+ * If NUMA is not configured or there is only one NUMA node available,
+ * there is no reason to consider NUMA. This function determines
+ * whether percpu allocation should consider NUMA or not.
+ *
+ * RETURNS:
+ * true if NUMA should be considered; otherwise, false.
+ */
+static bool __init pcpu_need_numa(void)
+{
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+ pg_data_t *last = NULL;
+ unsigned int cpu;
+
+ for_each_possible_cpu(cpu) {
+ int node = early_cpu_to_node(cpu);
+
+ if (node_online(node) && NODE_DATA(node) &&
+ last && last != NODE_DATA(node))
+ return true;
+
+ last = NODE_DATA(node);
+ }
+#endif
+ return false;
+}
+#endif
+
+/**
+ * pcpu_alloc_bootmem - NUMA friendly alloc_bootmem wrapper for percpu
+ * @cpu: cpu to allocate for
+ * @size: size allocation in bytes
+ * @align: alignment
+ *
+ * Allocate @size bytes aligned at @align for cpu @cpu. This wrapper
+ * does the right thing for NUMA regardless of the current
+ * configuration.
+ *
+ * RETURNS:
+ * Pointer to the allocated area on success, NULL on failure.
+ */
+static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size,
+ unsigned long align)
+{
+ const unsigned long goal = __pa(MAX_DMA_ADDRESS);
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+ int node = early_cpu_to_node(cpu);
+ void *ptr;
+
+ if (!node_online(node) || !NODE_DATA(node)) {
+ ptr = __alloc_bootmem_nopanic(size, align, goal);
+ pr_info("cpu %d has no node %d or node-local memory\n",
+ cpu, node);
+ pr_debug("per cpu data for cpu%d %lu bytes at %016lx\n",
+ cpu, size, __pa(ptr));
+ } else {
+ ptr = __alloc_bootmem_node_nopanic(NODE_DATA(node),
+ size, align, goal);
+ pr_debug("per cpu data for cpu%d %lu bytes on node%d at %016lx\n",
+ cpu, size, node, __pa(ptr));
+ }
+ return ptr;
+#else
+ return __alloc_bootmem_nopanic(size, align, goal);
+#endif
+}
/*
- * Great future plan:
- * Declare PDA itself and support (irqstack,tss,pgd) as per cpu data.
- * Always point %gs to its beginning
+ * Helpers for first chunk memory allocation
*/
-void __init setup_per_cpu_areas(void)
+static void * __init pcpu_fc_alloc(unsigned int cpu, size_t size, size_t align)
+{
+ return pcpu_alloc_bootmem(cpu, size, align);
+}
+
+static void __init pcpu_fc_free(void *ptr, size_t size)
+{
+ free_bootmem(__pa(ptr), size);
+}
+
+static int __init pcpu_cpu_distance(unsigned int from, unsigned int to)
+{
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+ if (early_cpu_to_node(from) == early_cpu_to_node(to))
+ return LOCAL_DISTANCE;
+ else
+ return REMOTE_DISTANCE;
+#else
+ return LOCAL_DISTANCE;
+#endif
+}
+
+static void __init pcpup_populate_pte(unsigned long addr)
+{
+ populate_extra_pte(addr);
+}
+
+static inline void setup_percpu_segment(int cpu)
{
- ssize_t size;
- char *ptr;
- int cpu;
+#ifdef CONFIG_X86_32
+ struct desc_struct gdt;
- /* Copy section for each CPU (we discard the original) */
- size = roundup(PERCPU_ENOUGH_ROOM, PAGE_SIZE);
+ pack_descriptor(&gdt, per_cpu_offset(cpu), 0xFFFFF,
+ 0x2 | DESCTYPE_S, 0x8);
+ gdt.s = 1;
+ write_gdt_entry(get_cpu_gdt_table(cpu),
+ GDT_ENTRY_PERCPU, &gdt, DESCTYPE_S);
+#endif
+}
+
+void __init setup_per_cpu_areas(void)
+{
+ unsigned int cpu;
+ unsigned long delta;
+ int rc;
pr_info("NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n",
NR_CPUS, nr_cpumask_bits, nr_cpu_ids, nr_node_ids);
- pr_info("PERCPU: Allocating %zd bytes of per cpu data\n", size);
+ /*
+ * Allocate percpu area. Embedding allocator is our favorite;
+ * however, on NUMA configurations, it can result in very
+ * sparse unit mapping and vmalloc area isn't spacious enough
+ * on 32bit. Use page in that case.
+ */
+#ifdef CONFIG_X86_32
+ if (pcpu_chosen_fc == PCPU_FC_AUTO && pcpu_need_numa())
+ pcpu_chosen_fc = PCPU_FC_PAGE;
+#endif
+ rc = -EINVAL;
+ if (pcpu_chosen_fc != PCPU_FC_PAGE) {
+ const size_t dyn_size = PERCPU_MODULE_RESERVE +
+ PERCPU_DYNAMIC_RESERVE - PERCPU_FIRST_CHUNK_RESERVE;
+ size_t atom_size;
- for_each_possible_cpu(cpu) {
-#ifndef CONFIG_NEED_MULTIPLE_NODES
- ptr = alloc_bootmem_pages(size);
+ /*
+ * On 64bit, use PMD_SIZE for atom_size so that embedded
+ * percpu areas are aligned to PMD. This, in the future,
+ * can also allow using PMD mappings in vmalloc area. Use
+ * PAGE_SIZE on 32bit as vmalloc space is highly contended
+ * and large vmalloc area allocs can easily fail.
+ */
+#ifdef CONFIG_X86_64
+ atom_size = PMD_SIZE;
#else
- int node = early_cpu_to_node(cpu);
- if (!node_online(node) || !NODE_DATA(node)) {
- ptr = alloc_bootmem_pages(size);
- pr_info("cpu %d has no node %d or node-local memory\n",
- cpu, node);
- pr_debug("per cpu data for cpu%d at %016lx\n",
- cpu, __pa(ptr));
- } else {
- ptr = alloc_bootmem_pages_node(NODE_DATA(node), size);
- pr_debug("per cpu data for cpu%d on node%d at %016lx\n",
- cpu, node, __pa(ptr));
- }
-#endif
-
- memcpy(ptr, __per_cpu_load, __per_cpu_end - __per_cpu_start);
- per_cpu_offset(cpu) = ptr - __per_cpu_start;
+ atom_size = PAGE_SIZE;
+#endif
+ rc = pcpu_embed_first_chunk(PERCPU_FIRST_CHUNK_RESERVE,
+ dyn_size, atom_size,
+ pcpu_cpu_distance,
+ pcpu_fc_alloc, pcpu_fc_free);
+ if (rc < 0)
+ pr_warning("%s allocator failed (%d), falling back to page size\n",
+ pcpu_fc_names[pcpu_chosen_fc], rc);
+ }
+ if (rc < 0)
+ rc = pcpu_page_first_chunk(PERCPU_FIRST_CHUNK_RESERVE,
+ pcpu_fc_alloc, pcpu_fc_free,
+ pcpup_populate_pte);
+ if (rc < 0)
+ panic("cannot initialize percpu area (err=%d)", rc);
+
+ /* alrighty, percpu areas up and running */
+ delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
+ for_each_possible_cpu(cpu) {
+ per_cpu_offset(cpu) = delta + pcpu_unit_offsets[cpu];
per_cpu(this_cpu_off, cpu) = per_cpu_offset(cpu);
per_cpu(cpu_number, cpu) = cpu;
+ setup_percpu_segment(cpu);
+ setup_stack_canary_segment(cpu);
/*
- * Copy data used in early init routines from the initial arrays to the
- * per cpu data areas. These arrays then become expendable and the
- * *_early_ptr's are zeroed indicating that the static arrays are gone.
+ * Copy data used in early init routines from the
+ * initial arrays to the per cpu data areas. These
+ * arrays then become expendable and the *_early_ptr's
+ * are zeroed indicating that the static arrays are
+ * gone.
*/
+#ifndef CONFIG_XEN
#ifdef CONFIG_X86_LOCAL_APIC
per_cpu(x86_cpu_to_apicid, cpu) =
- early_per_cpu_map(x86_cpu_to_apicid, cpu);
+ early_per_cpu_map(x86_cpu_to_apicid, cpu);
per_cpu(x86_bios_cpu_apicid, cpu) =
- early_per_cpu_map(x86_bios_cpu_apicid, cpu);
+ early_per_cpu_map(x86_bios_cpu_apicid, cpu);
+#endif
+#ifdef CONFIG_X86_32
+ per_cpu(x86_cpu_to_logical_apicid, cpu) =
+ early_per_cpu_map(x86_cpu_to_logical_apicid, cpu);
+#endif
#endif
#ifdef CONFIG_X86_64
per_cpu(irq_stack_ptr, cpu) =
- per_cpu(irq_stack_union.irq_stack, cpu) + IRQ_STACK_SIZE - 64;
+ per_cpu(irq_stack_union.irq_stack, cpu) +
+ IRQ_STACK_SIZE - 64;
+#endif
#ifdef CONFIG_NUMA
per_cpu(x86_cpu_to_node_map, cpu) =
- early_per_cpu_map(x86_cpu_to_node_map, cpu);
-#endif
+ early_per_cpu_map(x86_cpu_to_node_map, cpu);
/*
- * Up to this point, CPU0 has been using .data.init
- * area. Reload %gs offset for CPU0.
+ * Ensure that the boot cpu numa_node is correct when the boot
+ * cpu is on a node that doesn't have memory installed.
+ * Also cpu_up() will call cpu_to_node() for APs when
+ * MEMORY_HOTPLUG is defined, before per_cpu(numa_node) is set
+ * up later with c_init aka intel_init/amd_init.
+ * So set them all (boot cpu and all APs).
*/
- if (cpu == 0)
- load_gs_base(cpu);
+ set_cpu_numa_node(cpu, early_cpu_to_node(cpu));
#endif
-
- DBG("PERCPU: cpu %4d %p\n", cpu, ptr);
+ /*
+ * Up to this point, the boot CPU has been using .init.data
+ * area. Reload any changed state for the boot CPU.
+ */
+ if (!cpu)
+ switch_to_new_gdt(cpu);
}
/* indicate the early static arrays will soon be gone */
+#ifndef CONFIG_XEN
+#ifdef CONFIG_X86_LOCAL_APIC
early_per_cpu_ptr(x86_cpu_to_apicid) = NULL;
early_per_cpu_ptr(x86_bios_cpu_apicid) = NULL;
-#if defined(CONFIG_X86_64) && defined(CONFIG_NUMA)
+#endif
+#ifdef CONFIG_X86_32
+ early_per_cpu_ptr(x86_cpu_to_logical_apicid) = NULL;
+#endif
+#endif
+#ifdef CONFIG_NUMA
early_per_cpu_ptr(x86_cpu_to_node_map) = NULL;
#endif
/* Setup cpu initialized, callin, callout masks */
setup_cpu_local_masks();
}
-
-#endif
-