[net-next-2.6.git] / arch / i386 / mm / init.c

/*
 *  linux/arch/i386/mm/init.c
 *
 *  Copyright (C) 1995  Linus Torvalds
 *
 *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
 */

#include <linux/config.h>
#include <linux/module.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/ptrace.h>
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/swap.h>
#include <linux/smp.h>
#include <linux/init.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/bootmem.h>
#include <linux/slab.h>
#include <linux/proc_fs.h>
#include <linux/efi.h>
#include <linux/memory_hotplug.h>
#include <linux/initrd.h>
#include <linux/cpumask.h>

#include <asm/processor.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/dma.h>
#include <asm/fixmap.h>
#include <asm/e820.h>
#include <asm/apic.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
#include <asm/sections.h>

unsigned int __VMALLOC_RESERVE = 128 << 20;

DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
unsigned long highstart_pfn, highend_pfn;

static int noinline do_test_wp_bit(void);

/*
 * Creates a middle page table and puts a pointer to it in the
 * given global directory entry. This only returns the gd entry
 * in non-PAE compilation mode, since the middle layer is folded.
 */
static pmd_t * __init one_md_table_init(pgd_t *pgd)
{
	pud_t *pud;
	pmd_t *pmd_table;
		
#ifdef CONFIG_X86_PAE
	pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
	set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
	pud = pud_offset(pgd, 0);
	if (pmd_table != pmd_offset(pud, 0)) 
		BUG();
#else
	pud = pud_offset(pgd, 0);
	pmd_table = pmd_offset(pud, 0);
#endif

	return pmd_table;
}

/*
 * Create a page table and place a pointer to it in a middle page
 * directory entry.
 */
static pte_t * __init one_page_table_init(pmd_t *pmd)
{
	if (pmd_none(*pmd)) {
		pte_t *page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
		set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE));
		if (page_table != pte_offset_kernel(pmd, 0))
			BUG();	

		return page_table;
	}
	
	return pte_offset_kernel(pmd, 0);
}

/*
 * This function initializes a certain range of kernel virtual memory 
 * with new bootmem page tables, everywhere page tables are missing in
 * the given range.
 */

/*
 * NOTE: The pagetables are allocated contiguous on the physical space 
 * so we can cache the place of the first one and move around without 
 * checking the pgd every time.
 */
static void __init page_table_range_init (unsigned long start, unsigned long end, pgd_t *pgd_base)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	int pgd_idx, pmd_idx;
	unsigned long vaddr;

	vaddr = start;
	pgd_idx = pgd_index(vaddr);
	pmd_idx = pmd_index(vaddr);
	pgd = pgd_base + pgd_idx;

	for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) {
		if (pgd_none(*pgd)) 
			one_md_table_init(pgd);
		pud = pud_offset(pgd, vaddr);
		pmd = pmd_offset(pud, vaddr);
		for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) {
			if (pmd_none(*pmd)) 
				one_page_table_init(pmd);

			vaddr += PMD_SIZE;
		}
		pmd_idx = 0;
	}
}

static inline int is_kernel_text(unsigned long addr)
{
	if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end)
		return 1;
	return 0;
}

/*
 * This maps the physical memory to kernel virtual address space, a total 
 * of max_low_pfn pages, by creating page tables starting from address 
 * PAGE_OFFSET.
 */
static void __init kernel_physical_mapping_init(pgd_t *pgd_base)
{
	unsigned long pfn;
	pgd_t *pgd;
	pmd_t *pmd;
	pte_t *pte;
	int pgd_idx, pmd_idx, pte_ofs;

	pgd_idx = pgd_index(PAGE_OFFSET);
	pgd = pgd_base + pgd_idx;
	pfn = 0;

	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
		pmd = one_md_table_init(pgd);
		if (pfn >= max_low_pfn)
			continue;
		for (pmd_idx = 0; pmd_idx < PTRS_PER_PMD && pfn < max_low_pfn; pmd++, pmd_idx++) {
			unsigned int address = pfn * PAGE_SIZE + PAGE_OFFSET;

			/* Map with big pages if possible, otherwise create normal page tables. */
			if (cpu_has_pse) {
				unsigned int address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE + PAGE_OFFSET + PAGE_SIZE-1;

				if (is_kernel_text(address) || is_kernel_text(address2))
					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE_EXEC));
				else
					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE));
				pfn += PTRS_PER_PTE;
			} else {
				pte = one_page_table_init(pmd);

				for (pte_ofs = 0; pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn; pte++, pfn++, pte_ofs++) {
						if (is_kernel_text(address))
							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC));
						else
							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL));
				}
			}
		}
	}
}

static inline int page_kills_ppro(unsigned long pagenr)
{
	if (pagenr >= 0x70000 && pagenr <= 0x7003F)
		return 1;
	return 0;
}

extern int is_available_memory(efi_memory_desc_t *);

int page_is_ram(unsigned long pagenr)
{
	int i;
	unsigned long addr, end;

	if (efi_enabled) {
		efi_memory_desc_t *md;
		void *p;

		for (p = memmap.map; p < memmap.map_end; p += memmap.desc_size) {
			md = p;
			if (!is_available_memory(md))
				continue;
			addr = (md->phys_addr+PAGE_SIZE-1) >> PAGE_SHIFT;
			end = (md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) >> PAGE_SHIFT;

			if ((pagenr >= addr) && (pagenr < end))
				return 1;
		}
		return 0;
	}

	for (i = 0; i < e820.nr_map; i++) {

		if (e820.map[i].type != E820_RAM)	/* not usable memory */
			continue;
		/*
		 *	!!!FIXME!!! Some BIOSen report areas as RAM that
		 *	are not. Notably the 640->1Mb area. We need a sanity
		 *	check here.
		 */
		addr = (e820.map[i].addr+PAGE_SIZE-1) >> PAGE_SHIFT;
		end = (e820.map[i].addr+e820.map[i].size) >> PAGE_SHIFT;
		if  ((pagenr >= addr) && (pagenr < end))
			return 1;
	}
	return 0;
}

#ifdef CONFIG_HIGHMEM
pte_t *kmap_pte;
pgprot_t kmap_prot;

#define kmap_get_fixmap_pte(vaddr)					\
	pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr), vaddr), (vaddr)), (vaddr))

static void __init kmap_init(void)
{
	unsigned long kmap_vstart;

	/* cache the first kmap pte */
	kmap_vstart = __fix_to_virt(FIX_KMAP_BEGIN);
	kmap_pte = kmap_get_fixmap_pte(kmap_vstart);

	kmap_prot = PAGE_KERNEL;
}

static void __init permanent_kmaps_init(pgd_t *pgd_base)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
	unsigned long vaddr;

	vaddr = PKMAP_BASE;
	page_table_range_init(vaddr, vaddr + PAGE_SIZE*LAST_PKMAP, pgd_base);

	pgd = swapper_pg_dir + pgd_index(vaddr);
	pud = pud_offset(pgd, vaddr);
	pmd = pmd_offset(pud, vaddr);
	pte = pte_offset_kernel(pmd, vaddr);
	pkmap_page_table = pte;	
}

static void __meminit free_new_highpage(struct page *page)
{
	init_page_count(page);
	__free_page(page);
	totalhigh_pages++;
}

void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro)
{
	if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
		ClearPageReserved(page);
		free_new_highpage(page);
	} else
		SetPageReserved(page);
}

static int add_one_highpage_hotplug(struct page *page, unsigned long pfn)
{
	free_new_highpage(page);
	totalram_pages++;
#ifdef CONFIG_FLATMEM
	max_mapnr = max(pfn, max_mapnr);
#endif
	num_physpages++;
	return 0;
}

/*
 * Not currently handling the NUMA case.
 * Assuming single node and all memory that
 * has been added dynamically that would be
 * onlined here is in HIGHMEM
 */
void online_page(struct page *page)
{
	ClearPageReserved(page);
	add_one_highpage_hotplug(page, page_to_pfn(page));
}


#ifdef CONFIG_NUMA
extern void set_highmem_pages_init(int);
#else
static void __init set_highmem_pages_init(int bad_ppro)
{
	int pfn;
	for (pfn = highstart_pfn; pfn < highend_pfn; pfn++)
		add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
	totalram_pages += totalhigh_pages;
}
#endif /* CONFIG_FLATMEM */

#else
#define kmap_init() do { } while (0)
#define permanent_kmaps_init(pgd_base) do { } while (0)
#define set_highmem_pages_init(bad_ppro) do { } while (0)
#endif /* CONFIG_HIGHMEM */

unsigned long long __PAGE_KERNEL = _PAGE_KERNEL;
EXPORT_SYMBOL(__PAGE_KERNEL);
unsigned long long __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;

#ifdef CONFIG_NUMA
extern void __init remap_numa_kva(void);
#else
#define remap_numa_kva() do {} while (0)
#endif

static void __init pagetable_init (void)
{
	unsigned long vaddr;
	pgd_t *pgd_base = swapper_pg_dir;

#ifdef CONFIG_X86_PAE
	int i;
	/* Init entries of the first-level page table to the zero page */
	for (i = 0; i < PTRS_PER_PGD; i++)
		set_pgd(pgd_base + i, __pgd(__pa(empty_zero_page) | _PAGE_PRESENT));
#endif

	/* Enable PSE if available */
	if (cpu_has_pse) {
		set_in_cr4(X86_CR4_PSE);
	}

	/* Enable PGE if available */
	if (cpu_has_pge) {
		set_in_cr4(X86_CR4_PGE);
		__PAGE_KERNEL |= _PAGE_GLOBAL;
		__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
	}

	kernel_physical_mapping_init(pgd_base);
	remap_numa_kva();

	/*
	 * Fixed mappings, only the page table structure has to be
	 * created - mappings will be set by set_fixmap():
	 */
	vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
	page_table_range_init(vaddr, 0, pgd_base);

	permanent_kmaps_init(pgd_base);

#ifdef CONFIG_X86_PAE
	/*
	 * Add low memory identity-mappings - SMP needs it when
	 * starting up on an AP from real-mode. In the non-PAE
	 * case we already have these mappings through head.S.
	 * All user-space mappings are explicitly cleared after
	 * SMP startup.
	 */
	set_pgd(&pgd_base[0], pgd_base[USER_PTRS_PER_PGD]);
#endif
}

#if defined(CONFIG_SOFTWARE_SUSPEND) || defined(CONFIG_ACPI_SLEEP)
/*
 * Swap suspend & friends need this for resume because things like the intel-agp
 * driver might have split up a kernel 4MB mapping.
 */
char __nosavedata swsusp_pg_dir[PAGE_SIZE]
	__attribute__ ((aligned (PAGE_SIZE)));

static inline void save_pg_dir(void)
{
	memcpy(swsusp_pg_dir, swapper_pg_dir, PAGE_SIZE);
}
#else
static inline void save_pg_dir(void)
{
}
#endif

void zap_low_mappings (void)
{
	int i;

	save_pg_dir();

	/*
	 * Zap initial low-memory mappings.
	 *
	 * Note that "pgd_clear()" doesn't do it for
	 * us, because pgd_clear() is a no-op on i386.
	 */
	for (i = 0; i < USER_PTRS_PER_PGD; i++)
#ifdef CONFIG_X86_PAE
		set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
#else
		set_pgd(swapper_pg_dir+i, __pgd(0));
#endif
	flush_tlb_all();
}

static int disable_nx __initdata = 0;
u64 __supported_pte_mask __read_mostly = ~_PAGE_NX;

/*
 * noexec = on|off
 *
 * Control non executable mappings.
 *
 * on      Enable
 * off     Disable
 */
void __init noexec_setup(const char *str)
{
	if (!strncmp(str, "on",2) && cpu_has_nx) {
		__supported_pte_mask |= _PAGE_NX;
		disable_nx = 0;
	} else if (!strncmp(str,"off",3)) {
		disable_nx = 1;
		__supported_pte_mask &= ~_PAGE_NX;
	}
}

int nx_enabled = 0;
#ifdef CONFIG_X86_PAE

static void __init set_nx(void)
{
	unsigned int v[4], l, h;

	if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
		cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
		if ((v[3] & (1 << 20)) && !disable_nx) {
			rdmsr(MSR_EFER, l, h);
			l |= EFER_NX;
			wrmsr(MSR_EFER, l, h);
			nx_enabled = 1;
			__supported_pte_mask |= _PAGE_NX;
		}
	}
}

/*
 * Enables/disables executability of a given kernel page and
 * returns the previous setting.
 */
int __init set_kernel_exec(unsigned long vaddr, int enable)
{
	pte_t *pte;
	int ret = 1;

	if (!nx_enabled)
		goto out;

	pte = lookup_address(vaddr);
	BUG_ON(!pte);

	if (!pte_exec_kernel(*pte))
		ret = 0;

	if (enable)
		pte->pte_high &= ~(1 << (_PAGE_BIT_NX - 32));
	else
		pte->pte_high |= 1 << (_PAGE_BIT_NX - 32);
	__flush_tlb_all();
out:
	return ret;
}

#endif

/*
 * paging_init() sets up the page tables - note that the first 8MB are
 * already mapped by head.S.
 *
 * This routines also unmaps the page at virtual kernel address 0, so
 * that we can trap those pesky NULL-reference errors in the kernel.
 */
void __init paging_init(void)
{
#ifdef CONFIG_X86_PAE
	set_nx();
	if (nx_enabled)
		printk("NX (Execute Disable) protection: active\n");
#endif

	pagetable_init();

	load_cr3(swapper_pg_dir);

#ifdef CONFIG_X86_PAE
	/*
	 * We will bail out later - printk doesn't work right now so
	 * the user would just see a hanging kernel.
	 */
	if (cpu_has_pae)
		set_in_cr4(X86_CR4_PAE);
#endif
	__flush_tlb_all();

	kmap_init();
}

/*
 * Test if the WP bit works in supervisor mode. It isn't supported on 386's
 * and also on some strange 486's (NexGen etc.). All 586+'s are OK. This
 * used to involve black magic jumps to work around some nasty CPU bugs,
 * but fortunately the switch to using exceptions got rid of all that.
 */

static void __init test_wp_bit(void)
{
	printk("Checking if this processor honours the WP bit even in supervisor mode... ");

	/* Any page-aligned address will do, the test is non-destructive */
	__set_fixmap(FIX_WP_TEST, __pa(&swapper_pg_dir), PAGE_READONLY);
	boot_cpu_data.wp_works_ok = do_test_wp_bit();
	clear_fixmap(FIX_WP_TEST);

	if (!boot_cpu_data.wp_works_ok) {
		printk("No.\n");
#ifdef CONFIG_X86_WP_WORKS_OK
		panic("This kernel doesn't support CPU's with broken WP. Recompile it for a 386!");
#endif
	} else {
		printk("Ok.\n");
	}
}

static void __init set_max_mapnr_init(void)
{
#ifdef CONFIG_HIGHMEM
	num_physpages = highend_pfn;
#else
	num_physpages = max_low_pfn;
#endif
#ifdef CONFIG_FLATMEM
	max_mapnr = num_physpages;
#endif
}

static struct kcore_list kcore_mem, kcore_vmalloc; 

void __init mem_init(void)
{
	extern int ppro_with_ram_bug(void);
	int codesize, reservedpages, datasize, initsize;
	int tmp;
	int bad_ppro;

#ifdef CONFIG_FLATMEM
	if (!mem_map)
		BUG();
#endif
	
	bad_ppro = ppro_with_ram_bug();

#ifdef CONFIG_HIGHMEM
	/* check that fixmap and pkmap do not overlap */
	if (PKMAP_BASE+LAST_PKMAP*PAGE_SIZE >= FIXADDR_START) {
		printk(KERN_ERR "fixmap and kmap areas overlap - this will crash\n");
		printk(KERN_ERR "pkstart: %lxh pkend: %lxh fixstart %lxh\n",
				PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE, FIXADDR_START);
		BUG();
	}
#endif
 
	set_max_mapnr_init();

#ifdef CONFIG_HIGHMEM
	high_memory = (void *) __va(highstart_pfn * PAGE_SIZE - 1) + 1;
#else
	high_memory = (void *) __va(max_low_pfn * PAGE_SIZE - 1) + 1;
#endif

	/* this will put all low memory onto the freelists */
	totalram_pages += free_all_bootmem();

	reservedpages = 0;
	for (tmp = 0; tmp < max_low_pfn; tmp++)
		/*
		 * Only count reserved RAM pages
		 */
		if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp)))
			reservedpages++;

	set_highmem_pages_init(bad_ppro);

	codesize =  (unsigned long) &_etext - (unsigned long) &_text;
	datasize =  (unsigned long) &_edata - (unsigned long) &_etext;
	initsize =  (unsigned long) &__init_end - (unsigned long) &__init_begin;

	kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT); 
	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START, 
		   VMALLOC_END-VMALLOC_START);

	printk(KERN_INFO "Memory: %luk/%luk available (%dk kernel code, %dk reserved, %dk data, %dk init, %ldk highmem)\n",
		(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
		num_physpages << (PAGE_SHIFT-10),
		codesize >> 10,
		reservedpages << (PAGE_SHIFT-10),
		datasize >> 10,
		initsize >> 10,
		(unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))
	       );

#ifdef CONFIG_X86_PAE
	if (!cpu_has_pae)
		panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
#endif
	if (boot_cpu_data.wp_works_ok < 0)
		test_wp_bit();

	/*
	 * Subtle. SMP is doing it's boot stuff late (because it has to
	 * fork idle threads) - but it also needs low mappings for the
	 * protected-mode entry to work. We zap these entries only after
	 * the WP-bit has been tested.
	 */
#ifndef CONFIG_SMP
	zap_low_mappings();
#endif
}

/*
 * this is for the non-NUMA, single node SMP system case.
 * Specifically, in the case of x86, we will always add
 * memory to the highmem for now.
 */
#ifdef CONFIG_MEMORY_HOTPLUG
#ifndef CONFIG_NEED_MULTIPLE_NODES
int add_memory(u64 start, u64 size)
{
	struct pglist_data *pgdata = &contig_page_data;
	struct zone *zone = pgdata->node_zones + MAX_NR_ZONES-1;
	unsigned long start_pfn = start >> PAGE_SHIFT;
	unsigned long nr_pages = size >> PAGE_SHIFT;

	return __add_pages(zone, start_pfn, nr_pages);
}

int remove_memory(u64 start, u64 size)
{
	return -EINVAL;
}
#endif
#endif

kmem_cache_t *pgd_cache;
kmem_cache_t *pmd_cache;

void __init pgtable_cache_init(void)
{
	if (PTRS_PER_PMD > 1) {
		pmd_cache = kmem_cache_create("pmd",
					PTRS_PER_PMD*sizeof(pmd_t),
					PTRS_PER_PMD*sizeof(pmd_t),
					0,
					pmd_ctor,
					NULL);
		if (!pmd_cache)
			panic("pgtable_cache_init(): cannot create pmd cache");
	}
	pgd_cache = kmem_cache_create("pgd",
				PTRS_PER_PGD*sizeof(pgd_t),
				PTRS_PER_PGD*sizeof(pgd_t),
				0,
				pgd_ctor,
				PTRS_PER_PMD == 1 ? pgd_dtor : NULL);
	if (!pgd_cache)
		panic("pgtable_cache_init(): Cannot create pgd cache");
}

/*
 * This function cannot be __init, since exceptions don't work in that
 * section.  Put this after the callers, so that it cannot be inlined.
 */
static int noinline do_test_wp_bit(void)
{
	char tmp_reg;
	int flag;

	__asm__ __volatile__(
		"	movb %0,%1	\n"
		"1:	movb %1,%0	\n"
		"	xorl %2,%2	\n"
		"2:			\n"
		".section __ex_table,\"a\"\n"
		"	.align 4	\n"
		"	.long 1b,2b	\n"
		".previous		\n"
		:"=m" (*(char *)fix_to_virt(FIX_WP_TEST)),
		 "=q" (tmp_reg),
		 "=r" (flag)
		:"2" (1)
		:"memory");
	
	return flag;
}

#ifdef CONFIG_DEBUG_RODATA

extern char __start_rodata, __end_rodata;
void mark_rodata_ro(void)
{
	unsigned long addr = (unsigned long)&__start_rodata;

	for (; addr < (unsigned long)&__end_rodata; addr += PAGE_SIZE)
		change_page_attr(virt_to_page(addr), 1, PAGE_KERNEL_RO);

	printk ("Write protecting the kernel read-only data: %luk\n",
			(unsigned long)(&__end_rodata - &__start_rodata) >> 10);

	/*
	 * change_page_attr() requires a global_flush_tlb() call after it.
	 * We do this after the printk so that if something went wrong in the
	 * change, the printk gets out at least to give a better debug hint
	 * of who is the culprit.
	 */
	global_flush_tlb();
}
#endif

void free_init_pages(char *what, unsigned long begin, unsigned long end)
{
	unsigned long addr;

	for (addr = begin; addr < end; addr += PAGE_SIZE) {
		ClearPageReserved(virt_to_page(addr));
		init_page_count(virt_to_page(addr));
		memset((void *)addr, 0xcc, PAGE_SIZE);
		free_page(addr);
		totalram_pages++;
	}
	printk(KERN_INFO "Freeing %s: %ldk freed\n", what, (end - begin) >> 10);
}

void free_initmem(void)
{
	free_init_pages("unused kernel memory",
			(unsigned long)(&__init_begin),
			(unsigned long)(&__init_end));
}

#ifdef CONFIG_BLK_DEV_INITRD
void free_initrd_mem(unsigned long start, unsigned long end)
{
	free_init_pages("initrd memory", start, end);
}
#endif
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/arch/i386/mm/init.c
	3	*
	4	* Copyright (C) 1995 Linus Torvalds
	5	*
	6	* Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
	7	*/
	8
	9	#include <linux/config.h>
	10	#include <linux/module.h>
	11	#include <linux/signal.h>
	12	#include <linux/sched.h>
	13	#include <linux/kernel.h>
	14	#include <linux/errno.h>
	15	#include <linux/string.h>
	16	#include <linux/types.h>
	17	#include <linux/ptrace.h>
	18	#include <linux/mman.h>
	19	#include <linux/mm.h>
	20	#include <linux/hugetlb.h>
	21	#include <linux/swap.h>
	22	#include <linux/smp.h>
	23	#include <linux/init.h>
	24	#include <linux/highmem.h>
	25	#include <linux/pagemap.h>
	26	#include <linux/bootmem.h>
	27	#include <linux/slab.h>
	28	#include <linux/proc_fs.h>
	29	#include <linux/efi.h>
05039b92	30	#include <linux/memory_hotplug.h>
27d99f7e	31	#include <linux/initrd.h>
55b2355e	32	#include <linux/cpumask.h>
1da177e4 LT	33
	34	#include <asm/processor.h>
	35	#include <asm/system.h>
	36	#include <asm/uaccess.h>
	37	#include <asm/pgtable.h>
	38	#include <asm/dma.h>
	39	#include <asm/fixmap.h>
	40	#include <asm/e820.h>
	41	#include <asm/apic.h>
	42	#include <asm/tlb.h>
	43	#include <asm/tlbflush.h>
	44	#include <asm/sections.h>
	45
	46	unsigned int __VMALLOC_RESERVE = 128 << 20;
	47
	48	DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
	49	unsigned long highstart_pfn, highend_pfn;
	50
	51	static int noinline do_test_wp_bit(void);
	52
	53	/*
	54	* Creates a middle page table and puts a pointer to it in the
	55	* given global directory entry. This only returns the gd entry
	56	* in non-PAE compilation mode, since the middle layer is folded.
	57	*/
	58	static pmd_t * __init one_md_table_init(pgd_t *pgd)
	59	{
	60	pud_t *pud;
	61	pmd_t *pmd_table;
	62
	63	#ifdef CONFIG_X86_PAE
	64	pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
	65	set_pgd(pgd, __pgd(__pa(pmd_table) \| _PAGE_PRESENT));
	66	pud = pud_offset(pgd, 0);
	67	if (pmd_table != pmd_offset(pud, 0))
	68	BUG();
	69	#else
	70	pud = pud_offset(pgd, 0);
	71	pmd_table = pmd_offset(pud, 0);
	72	#endif
	73
	74	return pmd_table;
	75	}
	76
	77	/*
	78	* Create a page table and place a pointer to it in a middle page
	79	* directory entry.
	80	*/
	81	static pte_t * __init one_page_table_init(pmd_t *pmd)
	82	{
	83	if (pmd_none(*pmd)) {
	84	pte_t page_table = (pte_t ) alloc_bootmem_low_pages(PAGE_SIZE);
	85	set_pmd(pmd, __pmd(__pa(page_table) \| _PAGE_TABLE));
	86	if (page_table != pte_offset_kernel(pmd, 0))
	87	BUG();
	88
	89	return page_table;
	90	}
	91
	92	return pte_offset_kernel(pmd, 0);
	93	}
	94
	95	/*
	96	* This function initializes a certain range of kernel virtual memory
97	* with new bootmem page tables, everywhere page tables are missing in
98	* the given range.
99	*/
100
101	/*
102	* NOTE: The pagetables are allocated contiguous on the physical space
103	* so we can cache the place of the first one and move around without
104	* checking the pgd every time.
105	*/
106	static void __init page_table_range_init (unsigned long start, unsigned long end, pgd_t *pgd_base)
107	{
108	pgd_t *pgd;
109	pud_t *pud;
110	pmd_t *pmd;
111	int pgd_idx, pmd_idx;
112	unsigned long vaddr;
113
114	vaddr = start;
115	pgd_idx = pgd_index(vaddr);
116	pmd_idx = pmd_index(vaddr);
117	pgd = pgd_base + pgd_idx;
118
119	for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) {
120	if (pgd_none(*pgd))
121	one_md_table_init(pgd);
122	pud = pud_offset(pgd, vaddr);
123	pmd = pmd_offset(pud, vaddr);
124	for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) {
125	if (pmd_none(*pmd))
126	one_page_table_init(pmd);
127
128	vaddr += PMD_SIZE;
129	}
130	pmd_idx = 0;
131	}
132	}
133
134	static inline int is_kernel_text(unsigned long addr)
135	{
136	if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end)
137	return 1;
138	return 0;
139	}
140
141	/*
142	* This maps the physical memory to kernel virtual address space, a total
143	* of max_low_pfn pages, by creating page tables starting from address
144	* PAGE_OFFSET.
145	*/
146	static void __init kernel_physical_mapping_init(pgd_t *pgd_base)
147	{
148	unsigned long pfn;
149	pgd_t *pgd;
150	pmd_t *pmd;
151	pte_t *pte;
152	int pgd_idx, pmd_idx, pte_ofs;
153
154	pgd_idx = pgd_index(PAGE_OFFSET);
155	pgd = pgd_base + pgd_idx;
156	pfn = 0;
157
158	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
159	pmd = one_md_table_init(pgd);
160	if (pfn >= max_low_pfn)
161	continue;
162	for (pmd_idx = 0; pmd_idx < PTRS_PER_PMD && pfn < max_low_pfn; pmd++, pmd_idx++) {
163	unsigned int address = pfn * PAGE_SIZE + PAGE_OFFSET;
164
165	/* Map with big pages if possible, otherwise create normal page tables. */
166	if (cpu_has_pse) {
167	unsigned int address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE + PAGE_OFFSET + PAGE_SIZE-1;
168
169	if (is_kernel_text(address) \|\| is_kernel_text(address2))
170	set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE_EXEC));
171	else
172	set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE));
173	pfn += PTRS_PER_PTE;
174	} else {
175	pte = one_page_table_init(pmd);
176
177	for (pte_ofs = 0; pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn; pte++, pfn++, pte_ofs++) {
178	if (is_kernel_text(address))
179	set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC));
180	else
181	set_pte(pte, pfn_pte(pfn, PAGE_KERNEL));
182	}
183	}
184	}
185	}
186	}
187
188	static inline int page_kills_ppro(unsigned long pagenr)
189	{
190	if (pagenr >= 0x70000 && pagenr <= 0x7003F)
191	return 1;
192	return 0;
193	}
194
195	extern int is_available_memory(efi_memory_desc_t *);
196
5b505b90	197	int page_is_ram(unsigned long pagenr)
1da177e4 LT	198	{
	199	int i;
	200	unsigned long addr, end;
	201
	202	if (efi_enabled) {
	203	efi_memory_desc_t *md;
7ae65fd3	204	void *p;
1da177e4	205
7ae65fd3 MT	206	for (p = memmap.map; p < memmap.map_end; p += memmap.desc_size) {
7ae65fd3 MT	207	md = p;
1da177e4 LT	208	if (!is_available_memory(md))
	209	continue;
	210	addr = (md->phys_addr+PAGE_SIZE-1) >> PAGE_SHIFT;
	211	end = (md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) >> PAGE_SHIFT;
	212
	213	if ((pagenr >= addr) && (pagenr < end))
	214	return 1;
	215	}
	216	return 0;
	217	}
	218
	219	for (i = 0; i < e820.nr_map; i++) {
	220
	221	if (e820.map[i].type != E820_RAM) /* not usable memory */
	222	continue;
	223	/*
	224	* !!!FIXME!!! Some BIOSen report areas as RAM that
	225	* are not. Notably the 640->1Mb area. We need a sanity
	226	* check here.
	227	*/
	228	addr = (e820.map[i].addr+PAGE_SIZE-1) >> PAGE_SHIFT;
	229	end = (e820.map[i].addr+e820.map[i].size) >> PAGE_SHIFT;
	230	if ((pagenr >= addr) && (pagenr < end))
	231	return 1;
	232	}
	233	return 0;
	234	}
	235
	236	#ifdef CONFIG_HIGHMEM
	237	pte_t *kmap_pte;
	238	pgprot_t kmap_prot;
	239
	240	#define kmap_get_fixmap_pte(vaddr) \
	241	pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr), vaddr), (vaddr)), (vaddr))
	242
	243	static void __init kmap_init(void)
	244	{
	245	unsigned long kmap_vstart;
	246
	247	/* cache the first kmap pte */
	248	kmap_vstart = __fix_to_virt(FIX_KMAP_BEGIN);
	249	kmap_pte = kmap_get_fixmap_pte(kmap_vstart);
	250
	251	kmap_prot = PAGE_KERNEL;
	252	}
	253
	254	static void __init permanent_kmaps_init(pgd_t *pgd_base)
	255	{
	256	pgd_t *pgd;
	257	pud_t *pud;
	258	pmd_t *pmd;
	259	pte_t *pte;
	260	unsigned long vaddr;
	261
	262	vaddr = PKMAP_BASE;
	263	page_table_range_init(vaddr, vaddr + PAGE_SIZE*LAST_PKMAP, pgd_base);
	264
	265	pgd = swapper_pg_dir + pgd_index(vaddr);
	266	pud = pud_offset(pgd, vaddr);
	267	pmd = pmd_offset(pud, vaddr);
	268	pte = pte_offset_kernel(pmd, vaddr);
	269	pkmap_page_table = pte;
	270	}
	271
c09b4240	272	static void __meminit free_new_highpage(struct page *page)
05039b92	273	{
7835e98b	274	init_page_count(page);
05039b92 DH	275	__free_page(page);
	276	totalhigh_pages++;
	277	}
	278
	279	void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro)
1da177e4 LT	280	{
	281	if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
	282	ClearPageReserved(page);
05039b92	283	free_new_highpage(page);
1da177e4 LT	284	} else
	285	SetPageReserved(page);
	286	}
	287
05039b92 DH	288	static int add_one_highpage_hotplug(struct page *page, unsigned long pfn)
	289	{
	290	free_new_highpage(page);
	291	totalram_pages++;
	292	#ifdef CONFIG_FLATMEM
	293	max_mapnr = max(pfn, max_mapnr);
	294	#endif
	295	num_physpages++;
	296	return 0;
	297	}
	298
	299	/*
	300	* Not currently handling the NUMA case.
	301	* Assuming single node and all memory that
	302	* has been added dynamically that would be
	303	* onlined here is in HIGHMEM
	304	*/
	305	void online_page(struct page *page)
	306	{
	307	ClearPageReserved(page);
	308	add_one_highpage_hotplug(page, page_to_pfn(page));
	309	}
	310
	311
05b79bdc AW	312	#ifdef CONFIG_NUMA
	313	extern void set_highmem_pages_init(int);
	314	#else
1da177e4 LT	315	static void __init set_highmem_pages_init(int bad_ppro)
	316	{
	317	int pfn;
	318	for (pfn = highstart_pfn; pfn < highend_pfn; pfn++)
05039b92	319	add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
1da177e4 LT	320	totalram_pages += totalhigh_pages;
1da177e4 LT	321	}
05b79bdc	322	#endif /* CONFIG_FLATMEM */
1da177e4 LT	323
	324	#else
	325	#define kmap_init() do { } while (0)
	326	#define permanent_kmaps_init(pgd_base) do { } while (0)
	327	#define set_highmem_pages_init(bad_ppro) do { } while (0)
	328	#endif /* CONFIG_HIGHMEM */
	329
	330	unsigned long long __PAGE_KERNEL = _PAGE_KERNEL;
129f6946	331	EXPORT_SYMBOL(__PAGE_KERNEL);
1da177e4 LT	332	unsigned long long __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
1da177e4 LT	333
05b79bdc	334	#ifdef CONFIG_NUMA
1da177e4	335	extern void __init remap_numa_kva(void);
05b79bdc AW	336	#else
05b79bdc AW	337	#define remap_numa_kva() do {} while (0)
1da177e4 LT	338	#endif
	339
	340	static void __init pagetable_init (void)
	341	{
	342	unsigned long vaddr;
	343	pgd_t *pgd_base = swapper_pg_dir;
	344
	345	#ifdef CONFIG_X86_PAE
	346	int i;
	347	/* Init entries of the first-level page table to the zero page */
	348	for (i = 0; i < PTRS_PER_PGD; i++)
	349	set_pgd(pgd_base + i, __pgd(__pa(empty_zero_page) \| _PAGE_PRESENT));
	350	#endif
	351
	352	/* Enable PSE if available */
	353	if (cpu_has_pse) {
	354	set_in_cr4(X86_CR4_PSE);
	355	}
	356
	357	/* Enable PGE if available */
	358	if (cpu_has_pge) {
	359	set_in_cr4(X86_CR4_PGE);
	360	__PAGE_KERNEL \|= _PAGE_GLOBAL;
	361	__PAGE_KERNEL_EXEC \|= _PAGE_GLOBAL;
	362	}
	363
	364	kernel_physical_mapping_init(pgd_base);
	365	remap_numa_kva();
	366
	367	/*
	368	* Fixed mappings, only the page table structure has to be
	369	* created - mappings will be set by set_fixmap():
	370	*/
	371	vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
	372	page_table_range_init(vaddr, 0, pgd_base);
	373
	374	permanent_kmaps_init(pgd_base);
	375
	376	#ifdef CONFIG_X86_PAE
	377	/*
	378	* Add low memory identity-mappings - SMP needs it when
	379	* starting up on an AP from real-mode. In the non-PAE
	380	* case we already have these mappings through head.S.
	381	* All user-space mappings are explicitly cleared after
	382	* SMP startup.
	383	*/
c9b02a24	384	set_pgd(&pgd_base[0], pgd_base[USER_PTRS_PER_PGD]);
1da177e4 LT	385	#endif
	386	}
	387
55b2355e	388	#if defined(CONFIG_SOFTWARE_SUSPEND) \|\| defined(CONFIG_ACPI_SLEEP)
1da177e4 LT	389	/*
	390	* Swap suspend & friends need this for resume because things like the intel-agp
	391	* driver might have split up a kernel 4MB mapping.
	392	*/
	393	char __nosavedata swsusp_pg_dir[PAGE_SIZE]
	394	__attribute__ ((aligned (PAGE_SIZE)));
	395
	396	static inline void save_pg_dir(void)
	397	{
	398	memcpy(swsusp_pg_dir, swapper_pg_dir, PAGE_SIZE);
	399	}
	400	#else
	401	static inline void save_pg_dir(void)
	402	{
	403	}
	404	#endif
	405
	406	void zap_low_mappings (void)
	407	{
	408	int i;
	409
	410	save_pg_dir();
	411
	412	/*
	413	* Zap initial low-memory mappings.
	414	*
	415	* Note that "pgd_clear()" doesn't do it for
	416	* us, because pgd_clear() is a no-op on i386.
	417	*/
	418	for (i = 0; i < USER_PTRS_PER_PGD; i++)
	419	#ifdef CONFIG_X86_PAE
	420	set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
	421	#else
	422	set_pgd(swapper_pg_dir+i, __pgd(0));
	423	#endif
	424	flush_tlb_all();
	425	}
	426
	427	static int disable_nx __initdata = 0;
6c231b7b	428	u64 __supported_pte_mask __read_mostly = ~_PAGE_NX;
1da177e4 LT	429
	430	/*
	431	* noexec = on\|off
	432	*
	433	* Control non executable mappings.
	434	*
	435	* on Enable
	436	* off Disable
	437	*/
	438	void __init noexec_setup(const char *str)
	439	{
	440	if (!strncmp(str, "on",2) && cpu_has_nx) {
	441	__supported_pte_mask \|= _PAGE_NX;
	442	disable_nx = 0;
	443	} else if (!strncmp(str,"off",3)) {
	444	disable_nx = 1;
	445	__supported_pte_mask &= ~_PAGE_NX;
	446	}
	447	}
	448
	449	int nx_enabled = 0;
	450	#ifdef CONFIG_X86_PAE
	451
	452	static void __init set_nx(void)
	453	{
	454	unsigned int v[4], l, h;
	455
	456	if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
	457	cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
	458	if ((v[3] & (1 << 20)) && !disable_nx) {
	459	rdmsr(MSR_EFER, l, h);
	460	l \|= EFER_NX;
	461	wrmsr(MSR_EFER, l, h);
	462	nx_enabled = 1;
	463	__supported_pte_mask \|= _PAGE_NX;
	464	}
	465	}
	466	}
	467
	468	/*
	469	* Enables/disables executability of a given kernel page and
	470	* returns the previous setting.
	471	*/
	472	int __init set_kernel_exec(unsigned long vaddr, int enable)
	473	{
	474	pte_t *pte;
	475	int ret = 1;
	476
	477	if (!nx_enabled)
	478	goto out;
	479
	480	pte = lookup_address(vaddr);
	481	BUG_ON(!pte);
	482
	483	if (!pte_exec_kernel(*pte))
	484	ret = 0;
	485
	486	if (enable)
	487	pte->pte_high &= ~(1 << (_PAGE_BIT_NX - 32));
	488	else
	489	pte->pte_high \|= 1 << (_PAGE_BIT_NX - 32);
	490	__flush_tlb_all();
	491	out:
	492	return ret;
493	}
494
495	#endif
496
497	/*
498	* paging_init() sets up the page tables - note that the first 8MB are
499	* already mapped by head.S.
500	*
501	* This routines also unmaps the page at virtual kernel address 0, so
502	* that we can trap those pesky NULL-reference errors in the kernel.
503	*/
504	void __init paging_init(void)
505	{
506	#ifdef CONFIG_X86_PAE
507	set_nx();
508	if (nx_enabled)
509	printk("NX (Execute Disable) protection: active\n");
510	#endif
511
512	pagetable_init();
513
514	load_cr3(swapper_pg_dir);
515
516	#ifdef CONFIG_X86_PAE
517	/*
518	* We will bail out later - printk doesn't work right now so
519	* the user would just see a hanging kernel.
520	*/
521	if (cpu_has_pae)
522	set_in_cr4(X86_CR4_PAE);
523	#endif
524	__flush_tlb_all();
525
526	kmap_init();
527	}
528
529	/*
530	* Test if the WP bit works in supervisor mode. It isn't supported on 386's
531	* and also on some strange 486's (NexGen etc.). All 586+'s are OK. This
532	* used to involve black magic jumps to work around some nasty CPU bugs,
533	* but fortunately the switch to using exceptions got rid of all that.
534	*/
535
536	static void __init test_wp_bit(void)
537	{
538	printk("Checking if this processor honours the WP bit even in supervisor mode... ");
539
540	/* Any page-aligned address will do, the test is non-destructive */
541	__set_fixmap(FIX_WP_TEST, __pa(&swapper_pg_dir), PAGE_READONLY);
542	boot_cpu_data.wp_works_ok = do_test_wp_bit();
543	clear_fixmap(FIX_WP_TEST);
544
545	if (!boot_cpu_data.wp_works_ok) {
546	printk("No.\n");
547	#ifdef CONFIG_X86_WP_WORKS_OK
548	panic("This kernel doesn't support CPU's with broken WP. Recompile it for a 386!");
549	#endif
550	} else {
551	printk("Ok.\n");
552	}
553	}
554
555	static void __init set_max_mapnr_init(void)
556	{
557	#ifdef CONFIG_HIGHMEM
558	num_physpages = highend_pfn;
559	#else
560	num_physpages = max_low_pfn;
561	#endif
05b79bdc	562	#ifdef CONFIG_FLATMEM
1da177e4 LT	563	max_mapnr = num_physpages;
	564	#endif
	565	}
	566
	567	static struct kcore_list kcore_mem, kcore_vmalloc;
	568
	569	void __init mem_init(void)
	570	{
	571	extern int ppro_with_ram_bug(void);
	572	int codesize, reservedpages, datasize, initsize;
	573	int tmp;
	574	int bad_ppro;
	575
05b79bdc	576	#ifdef CONFIG_FLATMEM
1da177e4 LT	577	if (!mem_map)
	578	BUG();
	579	#endif
	580
	581	bad_ppro = ppro_with_ram_bug();
	582
	583	#ifdef CONFIG_HIGHMEM
	584	/* check that fixmap and pkmap do not overlap */
	585	if (PKMAP_BASE+LAST_PKMAP*PAGE_SIZE >= FIXADDR_START) {
	586	printk(KERN_ERR "fixmap and kmap areas overlap - this will crash\n");
	587	printk(KERN_ERR "pkstart: %lxh pkend: %lxh fixstart %lxh\n",
	588	PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE, FIXADDR_START);
	589	BUG();
	590	}
	591	#endif
	592
	593	set_max_mapnr_init();
	594
	595	#ifdef CONFIG_HIGHMEM
	596	high_memory = (void ) __va(highstart_pfn PAGE_SIZE - 1) + 1;
	597	#else
	598	high_memory = (void ) __va(max_low_pfn PAGE_SIZE - 1) + 1;
	599	#endif
	600
	601	/* this will put all low memory onto the freelists */
	602	totalram_pages += free_all_bootmem();
	603
	604	reservedpages = 0;
	605	for (tmp = 0; tmp < max_low_pfn; tmp++)
	606	/*
	607	* Only count reserved RAM pages
	608	*/
	609	if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp)))
	610	reservedpages++;
	611
	612	set_highmem_pages_init(bad_ppro);
	613
	614	codesize = (unsigned long) &_etext - (unsigned long) &_text;
	615	datasize = (unsigned long) &_edata - (unsigned long) &_etext;
	616	initsize = (unsigned long) &__init_end - (unsigned long) &__init_begin;
	617
	618	kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT);
	619	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
	620	VMALLOC_END-VMALLOC_START);
	621
	622	printk(KERN_INFO "Memory: %luk/%luk available (%dk kernel code, %dk reserved, %dk data, %dk init, %ldk highmem)\n",
	623	(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
	624	num_physpages << (PAGE_SHIFT-10),
	625	codesize >> 10,
	626	reservedpages << (PAGE_SHIFT-10),
	627	datasize >> 10,
	628	initsize >> 10,
	629	(unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))
	630	);
	631
	632	#ifdef CONFIG_X86_PAE
	633	if (!cpu_has_pae)
	634	panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
	635	#endif
	636	if (boot_cpu_data.wp_works_ok < 0)
	637	test_wp_bit();
	638
	639	/*
	640	* Subtle. SMP is doing it's boot stuff late (because it has to
641	* fork idle threads) - but it also needs low mappings for the
642	* protected-mode entry to work. We zap these entries only after
643	* the WP-bit has been tested.
644	*/
645	#ifndef CONFIG_SMP
646	zap_low_mappings();
647	#endif
648	}
649
05039b92 DH	650	/*
	651	* this is for the non-NUMA, single node SMP system case.
	652	* Specifically, in the case of x86, we will always add
	653	* memory to the highmem for now.
	654	*/
ad8f5797	655	#ifdef CONFIG_MEMORY_HOTPLUG
05039b92 DH	656	#ifndef CONFIG_NEED_MULTIPLE_NODES
	657	int add_memory(u64 start, u64 size)
	658	{
	659	struct pglist_data *pgdata = &contig_page_data;
	660	struct zone *zone = pgdata->node_zones + MAX_NR_ZONES-1;
	661	unsigned long start_pfn = start >> PAGE_SHIFT;
	662	unsigned long nr_pages = size >> PAGE_SHIFT;
	663
	664	return __add_pages(zone, start_pfn, nr_pages);
	665	}
	666
	667	int remove_memory(u64 start, u64 size)
	668	{
	669	return -EINVAL;
	670	}
	671	#endif
9d99aaa3	672	#endif
05039b92	673
1da177e4 LT	674	kmem_cache_t *pgd_cache;
	675	kmem_cache_t *pmd_cache;
	676
	677	void __init pgtable_cache_init(void)
	678	{
	679	if (PTRS_PER_PMD > 1) {
	680	pmd_cache = kmem_cache_create("pmd",
	681	PTRS_PER_PMD*sizeof(pmd_t),
	682	PTRS_PER_PMD*sizeof(pmd_t),
	683	0,
	684	pmd_ctor,
	685	NULL);
	686	if (!pmd_cache)
	687	panic("pgtable_cache_init(): cannot create pmd cache");
	688	}
	689	pgd_cache = kmem_cache_create("pgd",
	690	PTRS_PER_PGD*sizeof(pgd_t),
	691	PTRS_PER_PGD*sizeof(pgd_t),
	692	0,
	693	pgd_ctor,
	694	PTRS_PER_PMD == 1 ? pgd_dtor : NULL);
	695	if (!pgd_cache)
	696	panic("pgtable_cache_init(): Cannot create pgd cache");
	697	}
	698
	699	/*
	700	* This function cannot be __init, since exceptions don't work in that
	701	* section. Put this after the callers, so that it cannot be inlined.
	702	*/
	703	static int noinline do_test_wp_bit(void)
	704	{
	705	char tmp_reg;
	706	int flag;
	707
	708	__asm__ __volatile__(
	709	" movb %0,%1 \n"
	710	"1: movb %1,%0 \n"
	711	" xorl %2,%2 \n"
	712	"2: \n"
	713	".section __ex_table,\"a\"\n"
	714	" .align 4 \n"
	715	" .long 1b,2b \n"
	716	".previous \n"
	717	:"=m" ((char )fix_to_virt(FIX_WP_TEST)),
	718	"=q" (tmp_reg),
	719	"=r" (flag)
	720	:"2" (1)
	721	:"memory");
	722
	723	return flag;
	724	}
	725
63aaf308 AV	726	#ifdef CONFIG_DEBUG_RODATA
	727
	728	extern char __start_rodata, __end_rodata;
	729	void mark_rodata_ro(void)
	730	{
	731	unsigned long addr = (unsigned long)&__start_rodata;
	732
	733	for (; addr < (unsigned long)&__end_rodata; addr += PAGE_SIZE)
	734	change_page_attr(virt_to_page(addr), 1, PAGE_KERNEL_RO);
	735
	736	printk ("Write protecting the kernel read-only data: %luk\n",
	737	(unsigned long)(&__end_rodata - &__start_rodata) >> 10);
	738
	739	/*
	740	* change_page_attr() requires a global_flush_tlb() call after it.
	741	* We do this after the printk so that if something went wrong in the
	742	* change, the printk gets out at least to give a better debug hint
	743	* of who is the culprit.
	744	*/
	745	global_flush_tlb();
	746	}
	747	#endif
	748
9a0b5817 GH	749	void free_init_pages(char *what, unsigned long begin, unsigned long end)
	750	{
	751	unsigned long addr;
	752
	753	for (addr = begin; addr < end; addr += PAGE_SIZE) {
	754	ClearPageReserved(virt_to_page(addr));
	755	init_page_count(virt_to_page(addr));
	756	memset((void *)addr, 0xcc, PAGE_SIZE);
	757	free_page(addr);
	758	totalram_pages++;
	759	}
	760	printk(KERN_INFO "Freeing %s: %ldk freed\n", what, (end - begin) >> 10);
	761	}
	762
	763	void free_initmem(void)
	764	{
	765	free_init_pages("unused kernel memory",
	766	(unsigned long)(&__init_begin),
	767	(unsigned long)(&__init_end));
	768	}
63aaf308	769
1da177e4 LT	770	#ifdef CONFIG_BLK_DEV_INITRD
	771	void free_initrd_mem(unsigned long start, unsigned long end)
	772	{
9a0b5817	773	free_init_pages("initrd memory", start, end);
1da177e4 LT	774	}
1da177e4 LT	775	#endif
9a0b5817	776