sparc64: use embedding percpu first chunk allocator

sparc64 currently allocates a large page for each cpu and partially
remap them into vmalloc area much like what lpage first chunk
allocator did.  As a 4M page is used for each cpu, this results in
very large unit size and also adds TLB pressure due to the double
mapping of pages in the first chunk.

This patch converts sparc64 to use the embedding percpu first chunk
allocator which now knows how to handle NUMA configurations.  This
simplifies the code a lot, doesn't incur any extra TLB pressure and
results in better utilization of address space.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: David S. Miller <davem@davemloft.net>

diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index 4f6ed0f113f09eb3d0167d163714022b4b06bb06..fbd1233b392dda3a97b452df0b3853a3d1797dab 100644 (file)
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -95,6 +95,9 @@ config AUDIT_ARCH
 config HAVE_SETUP_PER_CPU_AREA
        def_bool y if SPARC64
 
+config NEED_PER_CPU_EMBED_FIRST_CHUNK
+       def_bool y if SPARC64
+
 config GENERIC_HARDIRQS_NO__DO_IRQ
        bool
        def_bool y if SPARC64

diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c
index b03fd362c629cd1b54dcbda47452e2ef4057d541..ff68373ce6d67615f4ad0bf88061bd04730a1673 100644 (file)
--- a/arch/sparc/kernel/smp_64.c
+++ b/arch/sparc/kernel/smp_64.c
@@ -1389,8 +1389,8 @@ void smp_send_stop(void)
  * 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)
+static void * __init pcpu_alloc_bootmem(unsigned int cpu, size_t size,
+                                       size_t align)
 {
        const unsigned long goal = __pa(MAX_DMA_ADDRESS);
 #ifdef CONFIG_NEED_MULTIPLE_NODES
@@ -1415,123 +1415,31 @@ static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size,
 #endif
 }
 
-#define PCPU_CHUNK_SIZE (4UL * 1024UL * 1024UL)
-
-static void __init pcpu_map_range(unsigned long start, unsigned long end,
-                                 struct page *page)
+static void __init pcpu_free_bootmem(void *ptr, size_t size)
 {
-       unsigned long pfn = page_to_pfn(page);
-       unsigned long pte_base;
-
-       BUG_ON((pfn<<PAGE_SHIFT)&(PCPU_CHUNK_SIZE - 1UL));
-
-       pte_base = (_PAGE_VALID | _PAGE_SZ4MB_4U |
-                   _PAGE_CP_4U | _PAGE_CV_4U |
-                   _PAGE_P_4U | _PAGE_W_4U);
-       if (tlb_type == hypervisor)
-               pte_base = (_PAGE_VALID | _PAGE_SZ4MB_4V |
-                           _PAGE_CP_4V | _PAGE_CV_4V |
-                           _PAGE_P_4V | _PAGE_W_4V);
-
-       while (start < end) {
-               pgd_t *pgd = pgd_offset_k(start);
-               unsigned long this_end;
-               pud_t *pud;
-               pmd_t *pmd;
-               pte_t *pte;
-
-               pud = pud_offset(pgd, start);
-               if (pud_none(*pud)) {
-                       pmd_t *new;
-
-                       new = __alloc_bootmem(PAGE_SIZE, PAGE_SIZE, PAGE_SIZE);
-                       pud_populate(&init_mm, pud, new);
-               }
-
-               pmd = pmd_offset(pud, start);
-               if (!pmd_present(*pmd)) {
-                       pte_t *new;
-
-                       new = __alloc_bootmem(PAGE_SIZE, PAGE_SIZE, PAGE_SIZE);
-                       pmd_populate_kernel(&init_mm, pmd, new);
-               }
-
-               pte = pte_offset_kernel(pmd, start);
-               this_end = (start + PMD_SIZE) & PMD_MASK;
-               if (this_end > end)
-                       this_end = end;
-
-               while (start < this_end) {
-                       unsigned long paddr = pfn << PAGE_SHIFT;
-
-                       pte_val(*pte) = (paddr | pte_base);
+       free_bootmem(__pa(ptr), size);
+}
 
-                       start += PAGE_SIZE;
-                       pte++;
-                       pfn++;
-               }
-       }
+static int pcpu_cpu_distance(unsigned int from, unsigned int to)
+{
+       if (cpu_to_node(from) == cpu_to_node(to))
+               return LOCAL_DISTANCE;
+       else
+               return REMOTE_DISTANCE;
 }
 
 void __init setup_per_cpu_areas(void)
 {
-       static struct vm_struct vm;
-       struct pcpu_alloc_info *ai;
-       unsigned long delta, cpu;
-       size_t size_sum;
-       size_t ptrs_size;
-       void **ptrs;
+       unsigned long delta;
+       unsigned int cpu;
        int rc;
 
-       ai = pcpu_alloc_alloc_info(1, nr_cpu_ids);
-
-       ai->static_size = __per_cpu_end - __per_cpu_start;
-       ai->reserved_size = PERCPU_MODULE_RESERVE;
-
-       size_sum = PFN_ALIGN(ai->static_size + ai->reserved_size +
-                            PERCPU_DYNAMIC_RESERVE);
-
-       ai->dyn_size = size_sum - ai->static_size - ai->reserved_size;
-       ai->unit_size = PCPU_CHUNK_SIZE;
-       ai->atom_size = PCPU_CHUNK_SIZE;
-       ai->alloc_size = PCPU_CHUNK_SIZE;
-       ai->groups[0].nr_units = nr_cpu_ids;
-
-       for_each_possible_cpu(cpu)
-               ai->groups[0].cpu_map[cpu] = cpu;
-
-       ptrs_size = PFN_ALIGN(nr_cpu_ids * sizeof(ptrs[0]));
-       ptrs = alloc_bootmem(ptrs_size);
-
-       for_each_possible_cpu(cpu) {
-               ptrs[cpu] = pcpu_alloc_bootmem(cpu, PCPU_CHUNK_SIZE,
-                                              PCPU_CHUNK_SIZE);
-
-               free_bootmem(__pa(ptrs[cpu] + size_sum),
-                            PCPU_CHUNK_SIZE - size_sum);
-
-               memcpy(ptrs[cpu], __per_cpu_load, ai->static_size);
-       }
-
-       /* allocate address and map */
-       vm.flags = VM_ALLOC;
-       vm.size = nr_cpu_ids * PCPU_CHUNK_SIZE;
-       vm_area_register_early(&vm, PCPU_CHUNK_SIZE);
-
-       for_each_possible_cpu(cpu) {
-               unsigned long start = (unsigned long) vm.addr;
-               unsigned long end;
-
-               start += cpu * PCPU_CHUNK_SIZE;
-               end = start + PCPU_CHUNK_SIZE;
-               pcpu_map_range(start, end, virt_to_page(ptrs[cpu]));
-       }
-
-       rc = pcpu_setup_first_chunk(ai, vm.addr);
+       rc = pcpu_embed_first_chunk(PERCPU_MODULE_RESERVE,
+                                   PERCPU_DYNAMIC_RESERVE, 4 << 20,
+                                   pcpu_cpu_distance, pcpu_alloc_bootmem,
+                                   pcpu_free_bootmem);
        if (rc)
-               panic("failed to setup percpu first chunk (%d)", rc);
-
-       free_bootmem(__pa(ptrs), ptrs_size);
+               panic("failed to initialize first chunk (%d)", rc);
 
        delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
        for_each_possible_cpu(cpu)