[deliverable/linux.git] / arch / x86 / mm / init.c

#include <linux/gfp.h>
#include <linux/initrd.h>
#include <linux/ioport.h>
#include <linux/swap.h>
#include <linux/memblock.h>
#include <linux/bootmem.h>	/* for max_low_pfn */

#include <asm/cacheflush.h>
#include <asm/e820.h>
#include <asm/init.h>
#include <asm/page.h>
#include <asm/page_types.h>
#include <asm/sections.h>
#include <asm/setup.h>
#include <asm/tlbflush.h>
#include <asm/tlb.h>
#include <asm/proto.h>
#include <asm/dma.h>		/* for MAX_DMA_PFN */

#include "mm_internal.h"

static unsigned long __initdata pgt_buf_start;
static unsigned long __initdata pgt_buf_end;
static unsigned long __initdata pgt_buf_top;

static unsigned long min_pfn_mapped;

/*
 * Pages returned are already directly mapped.
 *
 * Changing that is likely to break Xen, see commit:
 *
 *    279b706 x86,xen: introduce x86_init.mapping.pagetable_reserve
 *
 * for detailed information.
 */
__ref void *alloc_low_pages(unsigned int num)
{
	unsigned long pfn;
	int i;

#ifdef CONFIG_X86_64
	if (after_bootmem) {
		unsigned int order;

		order = get_order((unsigned long)num << PAGE_SHIFT);
		return (void *)__get_free_pages(GFP_ATOMIC | __GFP_NOTRACK |
						__GFP_ZERO, order);
	}
#endif

	if ((pgt_buf_end + num) >= pgt_buf_top) {
		unsigned long ret;
		if (min_pfn_mapped >= max_pfn_mapped)
			panic("alloc_low_page: ran out of memory");
		ret = memblock_find_in_range(min_pfn_mapped << PAGE_SHIFT,
					max_pfn_mapped << PAGE_SHIFT,
					PAGE_SIZE * num , PAGE_SIZE);
		if (!ret)
			panic("alloc_low_page: can not alloc memory");
		memblock_reserve(ret, PAGE_SIZE * num);
		pfn = ret >> PAGE_SHIFT;
	} else {
		pfn = pgt_buf_end;
		pgt_buf_end += num;
	}

	for (i = 0; i < num; i++) {
		void *adr;

		adr = __va((pfn + i) << PAGE_SHIFT);
		clear_page(adr);
	}

	return __va(pfn << PAGE_SHIFT);
}

/* need 4 4k for initial PMD_SIZE, 4k for 0-ISA_END_ADDRESS */
#define INIT_PGT_BUF_SIZE	(5 * PAGE_SIZE)
RESERVE_BRK(early_pgt_alloc, INIT_PGT_BUF_SIZE);
void  __init early_alloc_pgt_buf(void)
{
	unsigned long tables = INIT_PGT_BUF_SIZE;
	phys_addr_t base;

	base = __pa(extend_brk(tables, PAGE_SIZE));

	pgt_buf_start = base >> PAGE_SHIFT;
	pgt_buf_end = pgt_buf_start;
	pgt_buf_top = pgt_buf_start + (tables >> PAGE_SHIFT);
}

int after_bootmem;

int direct_gbpages
#ifdef CONFIG_DIRECT_GBPAGES
				= 1
#endif
;

static void __init init_gbpages(void)
{
#ifdef CONFIG_X86_64
	if (direct_gbpages && cpu_has_gbpages)
		printk(KERN_INFO "Using GB pages for direct mapping\n");
	else
		direct_gbpages = 0;
#endif
}

struct map_range {
	unsigned long start;
	unsigned long end;
	unsigned page_size_mask;
};

static int page_size_mask;

static void __init probe_page_size_mask(void)
{
	init_gbpages();

#if !defined(CONFIG_DEBUG_PAGEALLOC) && !defined(CONFIG_KMEMCHECK)
	/*
	 * For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
	 * This will simplify cpa(), which otherwise needs to support splitting
	 * large pages into small in interrupt context, etc.
	 */
	if (direct_gbpages)
		page_size_mask |= 1 << PG_LEVEL_1G;
	if (cpu_has_pse)
		page_size_mask |= 1 << PG_LEVEL_2M;
#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);
		__supported_pte_mask |= _PAGE_GLOBAL;
	}
}

#ifdef CONFIG_X86_32
#define NR_RANGE_MR 3
#else /* CONFIG_X86_64 */
#define NR_RANGE_MR 5
#endif

static int __meminit save_mr(struct map_range *mr, int nr_range,
			     unsigned long start_pfn, unsigned long end_pfn,
			     unsigned long page_size_mask)
{
	if (start_pfn < end_pfn) {
		if (nr_range >= NR_RANGE_MR)
			panic("run out of range for init_memory_mapping\n");
		mr[nr_range].start = start_pfn<<PAGE_SHIFT;
		mr[nr_range].end   = end_pfn<<PAGE_SHIFT;
		mr[nr_range].page_size_mask = page_size_mask;
		nr_range++;
	}

	return nr_range;
}

/*
 * adjust the page_size_mask for small range to go with
 *	big page size instead small one if nearby are ram too.
 */
static void __init_refok adjust_range_page_size_mask(struct map_range *mr,
							 int nr_range)
{
	int i;

	for (i = 0; i < nr_range; i++) {
		if ((page_size_mask & (1<<PG_LEVEL_2M)) &&
		    !(mr[i].page_size_mask & (1<<PG_LEVEL_2M))) {
			unsigned long start = round_down(mr[i].start, PMD_SIZE);
			unsigned long end = round_up(mr[i].end, PMD_SIZE);

#ifdef CONFIG_X86_32
			if ((end >> PAGE_SHIFT) > max_low_pfn)
				continue;
#endif

			if (memblock_is_region_memory(start, end - start))
				mr[i].page_size_mask |= 1<<PG_LEVEL_2M;
		}
		if ((page_size_mask & (1<<PG_LEVEL_1G)) &&
		    !(mr[i].page_size_mask & (1<<PG_LEVEL_1G))) {
			unsigned long start = round_down(mr[i].start, PUD_SIZE);
			unsigned long end = round_up(mr[i].end, PUD_SIZE);

			if (memblock_is_region_memory(start, end - start))
				mr[i].page_size_mask |= 1<<PG_LEVEL_1G;
		}
	}
}

static int __meminit split_mem_range(struct map_range *mr, int nr_range,
				     unsigned long start,
				     unsigned long end)
{
	unsigned long start_pfn, end_pfn;
	unsigned long pfn;
	int i;

	/* head if not big page alignment ? */
	pfn = start_pfn = PFN_DOWN(start);
#ifdef CONFIG_X86_32
	/*
	 * Don't use a large page for the first 2/4MB of memory
	 * because there are often fixed size MTRRs in there
	 * and overlapping MTRRs into large pages can cause
	 * slowdowns.
	 */
	if (pfn == 0)
		end_pfn = PFN_DOWN(PMD_SIZE);
	else
		end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
#else /* CONFIG_X86_64 */
	end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
#endif
	if (end_pfn > PFN_DOWN(end))
		end_pfn = PFN_DOWN(end);
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
		pfn = end_pfn;
	}

	/* big page (2M) range */
	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
#ifdef CONFIG_X86_32
	end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
#else /* CONFIG_X86_64 */
	end_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
	if (end_pfn > PFN_DOWN(round_down(end, PMD_SIZE)))
		end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
#endif

	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask & (1<<PG_LEVEL_2M));
		pfn = end_pfn;
	}

#ifdef CONFIG_X86_64
	/* big page (1G) range */
	start_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
	end_pfn = PFN_DOWN(round_down(end, PUD_SIZE));
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask &
				 ((1<<PG_LEVEL_2M)|(1<<PG_LEVEL_1G)));
		pfn = end_pfn;
	}

	/* tail is not big page (1G) alignment */
	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
	end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask & (1<<PG_LEVEL_2M));
		pfn = end_pfn;
	}
#endif

	/* tail is not big page (2M) alignment */
	start_pfn = pfn;
	end_pfn = PFN_DOWN(end);
	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);

	/* try to merge same page size and continuous */
	for (i = 0; nr_range > 1 && i < nr_range - 1; i++) {
		unsigned long old_start;
		if (mr[i].end != mr[i+1].start ||
		    mr[i].page_size_mask != mr[i+1].page_size_mask)
			continue;
		/* move it */
		old_start = mr[i].start;
		memmove(&mr[i], &mr[i+1],
			(nr_range - 1 - i) * sizeof(struct map_range));
		mr[i--].start = old_start;
		nr_range--;
	}

	if (!after_bootmem)
		adjust_range_page_size_mask(mr, nr_range);

	for (i = 0; i < nr_range; i++)
		printk(KERN_DEBUG " [mem %#010lx-%#010lx] page %s\n",
				mr[i].start, mr[i].end - 1,
			(mr[i].page_size_mask & (1<<PG_LEVEL_1G))?"1G":(
			 (mr[i].page_size_mask & (1<<PG_LEVEL_2M))?"2M":"4k"));

	return nr_range;
}

static struct range pfn_mapped[E820_X_MAX];
static int nr_pfn_mapped;

static void add_pfn_range_mapped(unsigned long start_pfn, unsigned long end_pfn)
{
	nr_pfn_mapped = add_range_with_merge(pfn_mapped, E820_X_MAX,
					     nr_pfn_mapped, start_pfn, end_pfn);
	nr_pfn_mapped = clean_sort_range(pfn_mapped, E820_X_MAX);

	max_pfn_mapped = max(max_pfn_mapped, end_pfn);

	if (start_pfn < (1UL<<(32-PAGE_SHIFT)))
		max_low_pfn_mapped = max(max_low_pfn_mapped,
					 min(end_pfn, 1UL<<(32-PAGE_SHIFT)));
}

bool pfn_range_is_mapped(unsigned long start_pfn, unsigned long end_pfn)
{
	int i;

	for (i = 0; i < nr_pfn_mapped; i++)
		if ((start_pfn >= pfn_mapped[i].start) &&
		    (end_pfn <= pfn_mapped[i].end))
			return true;

	return false;
}

/*
 * Setup the direct mapping of the physical memory at PAGE_OFFSET.
 * This runs before bootmem is initialized and gets pages directly from
 * the physical memory. To access them they are temporarily mapped.
 */
unsigned long __init_refok init_memory_mapping(unsigned long start,
					       unsigned long end)
{
	struct map_range mr[NR_RANGE_MR];
	unsigned long ret = 0;
	int nr_range, i;

	pr_info("init_memory_mapping: [mem %#010lx-%#010lx]\n",
	       start, end - 1);

	memset(mr, 0, sizeof(mr));
	nr_range = split_mem_range(mr, 0, start, end);

	for (i = 0; i < nr_range; i++)
		ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
						   mr[i].page_size_mask);

	add_pfn_range_mapped(start >> PAGE_SHIFT, ret >> PAGE_SHIFT);

	return ret >> PAGE_SHIFT;
}

/*
 * would have hole in the middle or ends, and only ram parts will be mapped.
 */
static unsigned long __init init_range_memory_mapping(
					   unsigned long range_start,
					   unsigned long range_end)
{
	unsigned long start_pfn, end_pfn;
	unsigned long mapped_ram_size = 0;
	int i;

	for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, NULL) {
		u64 start = (u64)start_pfn << PAGE_SHIFT;
		u64 end = (u64)end_pfn << PAGE_SHIFT;

		if (end <= range_start)
			continue;

		if (start < range_start)
			start = range_start;

		if (start >= range_end)
			continue;

		if (end > range_end)
			end = range_end;

		init_memory_mapping(start, end);

		mapped_ram_size += end - start;
	}

	return mapped_ram_size;
}

/* (PUD_SHIFT-PMD_SHIFT)/2 */
#define STEP_SIZE_SHIFT 5
void __init init_mem_mapping(void)
{
	unsigned long end, real_end, start, last_start;
	unsigned long step_size;
	unsigned long addr;
	unsigned long mapped_ram_size = 0;
	unsigned long new_mapped_ram_size;

	probe_page_size_mask();

#ifdef CONFIG_X86_64
	end = max_pfn << PAGE_SHIFT;
#else
	end = max_low_pfn << PAGE_SHIFT;
#endif

	/* the ISA range is always mapped regardless of memory holes */
	init_memory_mapping(0, ISA_END_ADDRESS);

	/* xen has big range in reserved near end of ram, skip it at first */
	addr = memblock_find_in_range(ISA_END_ADDRESS, end, PMD_SIZE,
			 PAGE_SIZE);
	real_end = addr + PMD_SIZE;

	/* step_size need to be small so pgt_buf from BRK could cover it */
	step_size = PMD_SIZE;
	max_pfn_mapped = 0; /* will get exact value next */
	min_pfn_mapped = real_end >> PAGE_SHIFT;
	last_start = start = real_end;
	while (last_start > ISA_END_ADDRESS) {
		if (last_start > step_size) {
			start = round_down(last_start - 1, step_size);
			if (start < ISA_END_ADDRESS)
				start = ISA_END_ADDRESS;
		} else
			start = ISA_END_ADDRESS;
		new_mapped_ram_size = init_range_memory_mapping(start,
							last_start);
		last_start = start;
		min_pfn_mapped = last_start >> PAGE_SHIFT;
		/* only increase step_size after big range get mapped */
		if (new_mapped_ram_size > mapped_ram_size)
			step_size <<= STEP_SIZE_SHIFT;
		mapped_ram_size += new_mapped_ram_size;
	}

	if (real_end < end)
		init_range_memory_mapping(real_end, end);

#ifdef CONFIG_X86_64
	if (max_pfn > max_low_pfn) {
		/* can we preseve max_low_pfn ?*/
		max_low_pfn = max_pfn;
	}
#else
	early_ioremap_page_table_range_init();
	load_cr3(swapper_pg_dir);
	__flush_tlb_all();
#endif

	early_memtest(0, max_pfn_mapped << PAGE_SHIFT);
}

/*
 * devmem_is_allowed() checks to see if /dev/mem access to a certain address
 * is valid. The argument is a physical page number.
 *
 *
 * On x86, access has to be given to the first megabyte of ram because that area
 * contains bios code and data regions used by X and dosemu and similar apps.
 * Access has to be given to non-kernel-ram areas as well, these contain the PCI
 * mmio resources as well as potential bios/acpi data regions.
 */
int devmem_is_allowed(unsigned long pagenr)
{
	if (pagenr < 256)
		return 1;
	if (iomem_is_exclusive(pagenr << PAGE_SHIFT))
		return 0;
	if (!page_is_ram(pagenr))
		return 1;
	return 0;
}

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

	/* Make sure boundaries are page aligned */
	begin_aligned = PAGE_ALIGN(begin);
	end_aligned   = end & PAGE_MASK;

	if (WARN_ON(begin_aligned != begin || end_aligned != end)) {
		begin = begin_aligned;
		end   = end_aligned;
	}

	if (begin >= end)
		return;

	addr = begin;

	/*
	 * If debugging page accesses then do not free this memory but
	 * mark them not present - any buggy init-section access will
	 * create a kernel page fault:
	 */
#ifdef CONFIG_DEBUG_PAGEALLOC
	printk(KERN_INFO "debug: unmapping init [mem %#010lx-%#010lx]\n",
		begin, end - 1);
	set_memory_np(begin, (end - begin) >> PAGE_SHIFT);
#else
	/*
	 * We just marked the kernel text read only above, now that
	 * we are going to free part of that, we need to make that
	 * writeable and non-executable first.
	 */
	set_memory_nx(begin, (end - begin) >> PAGE_SHIFT);
	set_memory_rw(begin, (end - begin) >> PAGE_SHIFT);

	printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);

	for (; addr < end; addr += PAGE_SIZE) {
		ClearPageReserved(virt_to_page(addr));
		init_page_count(virt_to_page(addr));
		memset((void *)addr, POISON_FREE_INITMEM, PAGE_SIZE);
		free_page(addr);
		totalram_pages++;
	}
#endif
}

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

#ifdef CONFIG_BLK_DEV_INITRD
void __init free_initrd_mem(unsigned long start, unsigned long end)
{
	/*
	 * end could be not aligned, and We can not align that,
	 * decompresser could be confused by aligned initrd_end
	 * We already reserve the end partial page before in
	 *   - i386_start_kernel()
	 *   - x86_64_start_kernel()
	 *   - relocate_initrd()
	 * So here We can do PAGE_ALIGN() safely to get partial page to be freed
	 */
	free_init_pages("initrd memory", start, PAGE_ALIGN(end));
}
#endif

void __init zone_sizes_init(void)
{
	unsigned long max_zone_pfns[MAX_NR_ZONES];

	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));

#ifdef CONFIG_ZONE_DMA
	max_zone_pfns[ZONE_DMA]		= MAX_DMA_PFN;
#endif
#ifdef CONFIG_ZONE_DMA32
	max_zone_pfns[ZONE_DMA32]	= MAX_DMA32_PFN;
#endif
	max_zone_pfns[ZONE_NORMAL]	= max_low_pfn;
#ifdef CONFIG_HIGHMEM
	max_zone_pfns[ZONE_HIGHMEM]	= max_pfn;
#endif

	free_area_init_nodes(max_zone_pfns);
}
Commit	Line	Data
5a0e3ad6	1	#include <linux/gfp.h>
2c1b284e	2	#include <linux/initrd.h>
540aca06	3	#include <linux/ioport.h>
e5b2bb55	4	#include <linux/swap.h>
a9ce6bc1	5	#include <linux/memblock.h>
17623915	6	#include <linux/bootmem.h> /* for max_low_pfn */
540aca06	7
e5b2bb55	8	#include <asm/cacheflush.h>
f765090a	9	#include <asm/e820.h>
4fcb2083	10	#include <asm/init.h>
e5b2bb55	11	#include <asm/page.h>
540aca06	12	#include <asm/page_types.h>
e5b2bb55	13	#include <asm/sections.h>
49834396	14	#include <asm/setup.h>
f765090a	15	#include <asm/tlbflush.h>
9518e0e4	16	#include <asm/tlb.h>
76c06927	17	#include <asm/proto.h>
17623915	18	#include <asm/dma.h> /* for MAX_DMA_PFN */
9518e0e4	19
5c51bdbe YL	20	#include "mm_internal.h"
5c51bdbe YL	21
cf470659 YL	22	static unsigned long __initdata pgt_buf_start;
	23	static unsigned long __initdata pgt_buf_end;
	24	static unsigned long __initdata pgt_buf_top;
f765090a	25
9985b4c6 YL	26	static unsigned long min_pfn_mapped;
9985b4c6 YL	27
ddd3509d SS	28	/*
	29	* Pages returned are already directly mapped.
	30	*
	31	* Changing that is likely to break Xen, see commit:
	32	*
	33	* 279b706 x86,xen: introduce x86_init.mapping.pagetable_reserve
	34	*
	35	* for detailed information.
	36	*/
22c8ca2a	37	__ref void *alloc_low_pages(unsigned int num)
5c51bdbe YL	38	{
5c51bdbe YL	39	unsigned long pfn;
22c8ca2a	40	int i;
5c51bdbe YL	41
	42	#ifdef CONFIG_X86_64
	43	if (after_bootmem) {
22c8ca2a	44	unsigned int order;
5c51bdbe	45
22c8ca2a YL	46	order = get_order((unsigned long)num << PAGE_SHIFT);
	47	return (void *)__get_free_pages(GFP_ATOMIC \| __GFP_NOTRACK \|
	48	__GFP_ZERO, order);
5c51bdbe YL	49	}
	50	#endif
	51
22c8ca2a	52	if ((pgt_buf_end + num) >= pgt_buf_top) {
5c51bdbe YL	53	unsigned long ret;
	54	if (min_pfn_mapped >= max_pfn_mapped)
	55	panic("alloc_low_page: ran out of memory");
	56	ret = memblock_find_in_range(min_pfn_mapped << PAGE_SHIFT,
	57	max_pfn_mapped << PAGE_SHIFT,
22c8ca2a	58	PAGE_SIZE * num , PAGE_SIZE);
5c51bdbe YL	59	if (!ret)
5c51bdbe YL	60	panic("alloc_low_page: can not alloc memory");
22c8ca2a	61	memblock_reserve(ret, PAGE_SIZE * num);
5c51bdbe	62	pfn = ret >> PAGE_SHIFT;
22c8ca2a YL	63	} else {
	64	pfn = pgt_buf_end;
	65	pgt_buf_end += num;
	66	}
	67
	68	for (i = 0; i < num; i++) {
	69	void *adr;
	70
	71	adr = __va((pfn + i) << PAGE_SHIFT);
	72	clear_page(adr);
	73	}
5c51bdbe	74
22c8ca2a	75	return __va(pfn << PAGE_SHIFT);
5c51bdbe YL	76	}
5c51bdbe YL	77
8d57470d YL	78	/* need 4 4k for initial PMD_SIZE, 4k for 0-ISA_END_ADDRESS */
	79	#define INIT_PGT_BUF_SIZE (5 * PAGE_SIZE)
	80	RESERVE_BRK(early_pgt_alloc, INIT_PGT_BUF_SIZE);
	81	void __init early_alloc_pgt_buf(void)
	82	{
	83	unsigned long tables = INIT_PGT_BUF_SIZE;
	84	phys_addr_t base;
	85
	86	base = __pa(extend_brk(tables, PAGE_SIZE));
	87
	88	pgt_buf_start = base >> PAGE_SHIFT;
	89	pgt_buf_end = pgt_buf_start;
	90	pgt_buf_top = pgt_buf_start + (tables >> PAGE_SHIFT);
	91	}
	92
f765090a PE	93	int after_bootmem;
	94
	95	int direct_gbpages
	96	#ifdef CONFIG_DIRECT_GBPAGES
	97	= 1
	98	#endif
	99	;
	100
148b2098 YL	101	static void __init init_gbpages(void)
	102	{
	103	#ifdef CONFIG_X86_64
	104	if (direct_gbpages && cpu_has_gbpages)
	105	printk(KERN_INFO "Using GB pages for direct mapping\n");
	106	else
	107	direct_gbpages = 0;
	108	#endif
	109	}
	110
844ab6f9 JS	111	struct map_range {
	112	unsigned long start;
	113	unsigned long end;
	114	unsigned page_size_mask;
	115	};
	116
fa62aafe	117	static int page_size_mask;
f765090a	118
22ddfcaa	119	static void __init probe_page_size_mask(void)
fa62aafe	120	{
148b2098 YL	121	init_gbpages();
148b2098 YL	122
fa62aafe YL	123	#if !defined(CONFIG_DEBUG_PAGEALLOC) && !defined(CONFIG_KMEMCHECK)
	124	/*
	125	* For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
	126	* This will simplify cpa(), which otherwise needs to support splitting
	127	* large pages into small in interrupt context, etc.
	128	*/
	129	if (direct_gbpages)
	130	page_size_mask \|= 1 << PG_LEVEL_1G;
	131	if (cpu_has_pse)
	132	page_size_mask \|= 1 << PG_LEVEL_2M;
	133	#endif
	134
	135	/* Enable PSE if available */
	136	if (cpu_has_pse)
	137	set_in_cr4(X86_CR4_PSE);
	138
	139	/* Enable PGE if available */
	140	if (cpu_has_pge) {
	141	set_in_cr4(X86_CR4_PGE);
	142	__supported_pte_mask \|= _PAGE_GLOBAL;
	143	}
	144	}
279b706b	145
f765090a PE	146	#ifdef CONFIG_X86_32
	147	#define NR_RANGE_MR 3
	148	#else /* CONFIG_X86_64 */
	149	#define NR_RANGE_MR 5
	150	#endif
	151
dc9dd5cc JB	152	static int __meminit save_mr(struct map_range *mr, int nr_range,
	153	unsigned long start_pfn, unsigned long end_pfn,
	154	unsigned long page_size_mask)
f765090a PE	155	{
	156	if (start_pfn < end_pfn) {
	157	if (nr_range >= NR_RANGE_MR)
	158	panic("run out of range for init_memory_mapping\n");
	159	mr[nr_range].start = start_pfn<<PAGE_SHIFT;
	160	mr[nr_range].end = end_pfn<<PAGE_SHIFT;
	161	mr[nr_range].page_size_mask = page_size_mask;
	162	nr_range++;
	163	}
	164
	165	return nr_range;
	166	}
	167
aeebe84c YL	168	/*
	169	* adjust the page_size_mask for small range to go with
	170	* big page size instead small one if nearby are ram too.
	171	*/
	172	static void __init_refok adjust_range_page_size_mask(struct map_range *mr,
	173	int nr_range)
	174	{
	175	int i;
	176
	177	for (i = 0; i < nr_range; i++) {
	178	if ((page_size_mask & (1<<PG_LEVEL_2M)) &&
	179	!(mr[i].page_size_mask & (1<<PG_LEVEL_2M))) {
	180	unsigned long start = round_down(mr[i].start, PMD_SIZE);
	181	unsigned long end = round_up(mr[i].end, PMD_SIZE);
	182
	183	#ifdef CONFIG_X86_32
	184	if ((end >> PAGE_SHIFT) > max_low_pfn)
	185	continue;
	186	#endif
	187
	188	if (memblock_is_region_memory(start, end - start))
	189	mr[i].page_size_mask \|= 1<<PG_LEVEL_2M;
	190	}
	191	if ((page_size_mask & (1<<PG_LEVEL_1G)) &&
	192	!(mr[i].page_size_mask & (1<<PG_LEVEL_1G))) {
	193	unsigned long start = round_down(mr[i].start, PUD_SIZE);
	194	unsigned long end = round_up(mr[i].end, PUD_SIZE);
	195
	196	if (memblock_is_region_memory(start, end - start))
	197	mr[i].page_size_mask \|= 1<<PG_LEVEL_1G;
	198	}
	199	}
	200	}
	201
4e33e065 YL	202	static int __meminit split_mem_range(struct map_range *mr, int nr_range,
	203	unsigned long start,
	204	unsigned long end)
f765090a	205	{
f765090a	206	unsigned long start_pfn, end_pfn;
1829ae9a	207	unsigned long pfn;
4e33e065	208	int i;
f765090a PE	209
f765090a PE	210	/* head if not big page alignment ? */
1829ae9a	211	pfn = start_pfn = PFN_DOWN(start);
f765090a PE	212	#ifdef CONFIG_X86_32
	213	/*
	214	* Don't use a large page for the first 2/4MB of memory
	215	* because there are often fixed size MTRRs in there
	216	* and overlapping MTRRs into large pages can cause
	217	* slowdowns.
	218	*/
1829ae9a	219	if (pfn == 0)
84d77001	220	end_pfn = PFN_DOWN(PMD_SIZE);
f765090a	221	else
1829ae9a	222	end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
f765090a	223	#else /* CONFIG_X86_64 */
1829ae9a	224	end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
f765090a	225	#endif
84d77001 YL	226	if (end_pfn > PFN_DOWN(end))
84d77001 YL	227	end_pfn = PFN_DOWN(end);
f765090a PE	228	if (start_pfn < end_pfn) {
f765090a PE	229	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
1829ae9a	230	pfn = end_pfn;
f765090a PE	231	}
	232
	233	/* big page (2M) range */
1829ae9a	234	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
f765090a	235	#ifdef CONFIG_X86_32
84d77001	236	end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
f765090a	237	#else /* CONFIG_X86_64 */
1829ae9a	238	end_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
84d77001 YL	239	if (end_pfn > PFN_DOWN(round_down(end, PMD_SIZE)))
84d77001 YL	240	end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
f765090a PE	241	#endif
	242
	243	if (start_pfn < end_pfn) {
	244	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	245	page_size_mask & (1<<PG_LEVEL_2M));
1829ae9a	246	pfn = end_pfn;
f765090a PE	247	}
	248
	249	#ifdef CONFIG_X86_64
	250	/* big page (1G) range */
1829ae9a	251	start_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
84d77001	252	end_pfn = PFN_DOWN(round_down(end, PUD_SIZE));
f765090a PE	253	if (start_pfn < end_pfn) {
	254	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	255	page_size_mask &
	256	((1<<PG_LEVEL_2M)\|(1<<PG_LEVEL_1G)));
1829ae9a	257	pfn = end_pfn;
f765090a PE	258	}
	259
	260	/* tail is not big page (1G) alignment */
1829ae9a	261	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
84d77001	262	end_pfn = PFN_DOWN(round_down(end, PMD_SIZE));
f765090a PE	263	if (start_pfn < end_pfn) {
	264	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	265	page_size_mask & (1<<PG_LEVEL_2M));
1829ae9a	266	pfn = end_pfn;
f765090a PE	267	}
	268	#endif
	269
	270	/* tail is not big page (2M) alignment */
1829ae9a	271	start_pfn = pfn;
84d77001	272	end_pfn = PFN_DOWN(end);
f765090a PE	273	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
	274
	275	/* try to merge same page size and continuous */
	276	for (i = 0; nr_range > 1 && i < nr_range - 1; i++) {
	277	unsigned long old_start;
	278	if (mr[i].end != mr[i+1].start \|\|
	279	mr[i].page_size_mask != mr[i+1].page_size_mask)
	280	continue;
	281	/* move it */
	282	old_start = mr[i].start;
	283	memmove(&mr[i], &mr[i+1],
	284	(nr_range - 1 - i) * sizeof(struct map_range));
	285	mr[i--].start = old_start;
	286	nr_range--;
	287	}
	288
aeebe84c YL	289	if (!after_bootmem)
	290	adjust_range_page_size_mask(mr, nr_range);
	291
f765090a	292	for (i = 0; i < nr_range; i++)
365811d6 BH	293	printk(KERN_DEBUG " [mem %#010lx-%#010lx] page %s\n",
365811d6 BH	294	mr[i].start, mr[i].end - 1,
f765090a PE	295	(mr[i].page_size_mask & (1<<PG_LEVEL_1G))?"1G":(
	296	(mr[i].page_size_mask & (1<<PG_LEVEL_2M))?"2M":"4k"));
	297
4e33e065 YL	298	return nr_range;
	299	}
	300
66520ebc JS	301	static struct range pfn_mapped[E820_X_MAX];
	302	static int nr_pfn_mapped;
	303
	304	static void add_pfn_range_mapped(unsigned long start_pfn, unsigned long end_pfn)
	305	{
	306	nr_pfn_mapped = add_range_with_merge(pfn_mapped, E820_X_MAX,
	307	nr_pfn_mapped, start_pfn, end_pfn);
	308	nr_pfn_mapped = clean_sort_range(pfn_mapped, E820_X_MAX);
	309
	310	max_pfn_mapped = max(max_pfn_mapped, end_pfn);
	311
	312	if (start_pfn < (1UL<<(32-PAGE_SHIFT)))
	313	max_low_pfn_mapped = max(max_low_pfn_mapped,
	314	min(end_pfn, 1UL<<(32-PAGE_SHIFT)));
	315	}
	316
	317	bool pfn_range_is_mapped(unsigned long start_pfn, unsigned long end_pfn)
	318	{
	319	int i;
	320
	321	for (i = 0; i < nr_pfn_mapped; i++)
	322	if ((start_pfn >= pfn_mapped[i].start) &&
	323	(end_pfn <= pfn_mapped[i].end))
	324	return true;
	325
	326	return false;
	327	}
	328
4e33e065 YL	329	/*
	330	* Setup the direct mapping of the physical memory at PAGE_OFFSET.
	331	* This runs before bootmem is initialized and gets pages directly from
	332	* the physical memory. To access them they are temporarily mapped.
	333	*/
	334	unsigned long __init_refok init_memory_mapping(unsigned long start,
	335	unsigned long end)
	336	{
	337	struct map_range mr[NR_RANGE_MR];
	338	unsigned long ret = 0;
	339	int nr_range, i;
	340
	341	pr_info("init_memory_mapping: [mem %#010lx-%#010lx]\n",
	342	start, end - 1);
	343
	344	memset(mr, 0, sizeof(mr));
	345	nr_range = split_mem_range(mr, 0, start, end);
	346
f765090a PE	347	for (i = 0; i < nr_range; i++)
	348	ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
	349	mr[i].page_size_mask);
f765090a	350
66520ebc JS	351	add_pfn_range_mapped(start >> PAGE_SHIFT, ret >> PAGE_SHIFT);
66520ebc JS	352
c14fa0b6 YL	353	return ret >> PAGE_SHIFT;
	354	}
	355
66520ebc	356	/*
8d57470d	357	* would have hole in the middle or ends, and only ram parts will be mapped.
66520ebc	358	*/
8d57470d YL	359	static unsigned long __init init_range_memory_mapping(
8d57470d YL	360	unsigned long range_start,
f763ad1d	361	unsigned long range_end)
66520ebc JS	362	{
66520ebc JS	363	unsigned long start_pfn, end_pfn;
8d57470d	364	unsigned long mapped_ram_size = 0;
66520ebc JS	365	int i;
66520ebc JS	366
66520ebc JS	367	for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, NULL) {
	368	u64 start = (u64)start_pfn << PAGE_SHIFT;
	369	u64 end = (u64)end_pfn << PAGE_SHIFT;
	370
f763ad1d	371	if (end <= range_start)
66520ebc JS	372	continue;
66520ebc JS	373
f763ad1d YL	374	if (start < range_start)
	375	start = range_start;
	376
	377	if (start >= range_end)
66520ebc JS	378	continue;
66520ebc JS	379
f763ad1d YL	380	if (end > range_end)
f763ad1d YL	381	end = range_end;
66520ebc	382
f763ad1d	383	init_memory_mapping(start, end);
8d57470d YL	384
8d57470d YL	385	mapped_ram_size += end - start;
66520ebc	386	}
8d57470d YL	387
8d57470d YL	388	return mapped_ram_size;
66520ebc JS	389	}
66520ebc JS	390
8d57470d YL	391	/* (PUD_SHIFT-PMD_SHIFT)/2 */
8d57470d YL	392	#define STEP_SIZE_SHIFT 5
c14fa0b6 YL	393	void __init init_mem_mapping(void)
c14fa0b6 YL	394	{
8d57470d YL	395	unsigned long end, real_end, start, last_start;
	396	unsigned long step_size;
	397	unsigned long addr;
	398	unsigned long mapped_ram_size = 0;
	399	unsigned long new_mapped_ram_size;
ab951937	400
c14fa0b6 YL	401	probe_page_size_mask();
c14fa0b6 YL	402
c14fa0b6	403	#ifdef CONFIG_X86_64
ab951937	404	end = max_pfn << PAGE_SHIFT;
c14fa0b6	405	#else
ab951937	406	end = max_low_pfn << PAGE_SHIFT;
c14fa0b6	407	#endif
ab951937	408
f763ad1d YL	409	/* the ISA range is always mapped regardless of memory holes */
f763ad1d YL	410	init_memory_mapping(0, ISA_END_ADDRESS);
8d57470d YL	411
	412	/* xen has big range in reserved near end of ram, skip it at first */
	413	addr = memblock_find_in_range(ISA_END_ADDRESS, end, PMD_SIZE,
	414	PAGE_SIZE);
	415	real_end = addr + PMD_SIZE;
	416
	417	/* step_size need to be small so pgt_buf from BRK could cover it */
	418	step_size = PMD_SIZE;
	419	max_pfn_mapped = 0; /* will get exact value next */
	420	min_pfn_mapped = real_end >> PAGE_SHIFT;
	421	last_start = start = real_end;
	422	while (last_start > ISA_END_ADDRESS) {
	423	if (last_start > step_size) {
	424	start = round_down(last_start - 1, step_size);
	425	if (start < ISA_END_ADDRESS)
	426	start = ISA_END_ADDRESS;
	427	} else
	428	start = ISA_END_ADDRESS;
	429	new_mapped_ram_size = init_range_memory_mapping(start,
	430	last_start);
	431	last_start = start;
	432	min_pfn_mapped = last_start >> PAGE_SHIFT;
	433	/* only increase step_size after big range get mapped */
	434	if (new_mapped_ram_size > mapped_ram_size)
	435	step_size <<= STEP_SIZE_SHIFT;
	436	mapped_ram_size += new_mapped_ram_size;
	437	}
	438
	439	if (real_end < end)
	440	init_range_memory_mapping(real_end, end);
	441
f763ad1d YL	442	#ifdef CONFIG_X86_64
	443	if (max_pfn > max_low_pfn) {
	444	/* can we preseve max_low_pfn ?*/
	445	max_low_pfn = max_pfn;
	446	}
719272c4 YL	447	#else
	448	early_ioremap_page_table_range_init();
	449	load_cr3(swapper_pg_dir);
	450	__flush_tlb_all();
f763ad1d	451	#endif
719272c4	452
c14fa0b6	453	early_memtest(0, max_pfn_mapped << PAGE_SHIFT);
22ddfcaa	454	}
e5b2bb55	455
540aca06 PE	456	/*
	457	* devmem_is_allowed() checks to see if /dev/mem access to a certain address
	458	* is valid. The argument is a physical page number.
	459	*
	460	*
	461	* On x86, access has to be given to the first megabyte of ram because that area
	462	* contains bios code and data regions used by X and dosemu and similar apps.
	463	* Access has to be given to non-kernel-ram areas as well, these contain the PCI
	464	* mmio resources as well as potential bios/acpi data regions.
	465	*/
	466	int devmem_is_allowed(unsigned long pagenr)
	467	{
73e8f3d7	468	if (pagenr < 256)
540aca06 PE	469	return 1;
	470	if (iomem_is_exclusive(pagenr << PAGE_SHIFT))
	471	return 0;
	472	if (!page_is_ram(pagenr))
	473	return 1;
	474	return 0;
	475	}
	476
e5b2bb55 PE	477	void free_init_pages(char *what, unsigned long begin, unsigned long end)
e5b2bb55 PE	478	{
c967da6a YL	479	unsigned long addr;
c967da6a YL	480	unsigned long begin_aligned, end_aligned;
e5b2bb55	481
c967da6a YL	482	/* Make sure boundaries are page aligned */
	483	begin_aligned = PAGE_ALIGN(begin);
	484	end_aligned = end & PAGE_MASK;
	485
	486	if (WARN_ON(begin_aligned != begin \|\| end_aligned != end)) {
	487	begin = begin_aligned;
	488	end = end_aligned;
	489	}
	490
	491	if (begin >= end)
e5b2bb55 PE	492	return;
e5b2bb55 PE	493
c967da6a YL	494	addr = begin;
c967da6a YL	495
e5b2bb55 PE	496	/*
	497	* If debugging page accesses then do not free this memory but
	498	* mark them not present - any buggy init-section access will
	499	* create a kernel page fault:
	500	*/
	501	#ifdef CONFIG_DEBUG_PAGEALLOC
365811d6 BH	502	printk(KERN_INFO "debug: unmapping init [mem %#010lx-%#010lx]\n",
365811d6 BH	503	begin, end - 1);
e5b2bb55 PE	504	set_memory_np(begin, (end - begin) >> PAGE_SHIFT);
	505	#else
	506	/*
	507	* We just marked the kernel text read only above, now that
	508	* we are going to free part of that, we need to make that
5bd5a452	509	* writeable and non-executable first.
e5b2bb55	510	*/
5bd5a452	511	set_memory_nx(begin, (end - begin) >> PAGE_SHIFT);
e5b2bb55 PE	512	set_memory_rw(begin, (end - begin) >> PAGE_SHIFT);
	513
	514	printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
	515
	516	for (; addr < end; addr += PAGE_SIZE) {
	517	ClearPageReserved(virt_to_page(addr));
	518	init_page_count(virt_to_page(addr));
c967da6a	519	memset((void *)addr, POISON_FREE_INITMEM, PAGE_SIZE);
e5b2bb55 PE	520	free_page(addr);
	521	totalram_pages++;
	522	}
	523	#endif
	524	}
	525
	526	void free_initmem(void)
	527	{
	528	free_init_pages("unused kernel memory",
	529	(unsigned long)(&__init_begin),
	530	(unsigned long)(&__init_end));
	531	}
731ddea6 PE	532
731ddea6 PE	533	#ifdef CONFIG_BLK_DEV_INITRD
0d26d1d8	534	void __init free_initrd_mem(unsigned long start, unsigned long end)
731ddea6	535	{
c967da6a YL	536	/*
	537	* end could be not aligned, and We can not align that,
	538	* decompresser could be confused by aligned initrd_end
	539	* We already reserve the end partial page before in
	540	* - i386_start_kernel()
	541	* - x86_64_start_kernel()
	542	* - relocate_initrd()
	543	* So here We can do PAGE_ALIGN() safely to get partial page to be freed
	544	*/
	545	free_init_pages("initrd memory", start, PAGE_ALIGN(end));
731ddea6 PE	546	}
731ddea6 PE	547	#endif
17623915 PE	548
	549	void __init zone_sizes_init(void)
	550	{
	551	unsigned long max_zone_pfns[MAX_NR_ZONES];
	552
	553	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
	554
	555	#ifdef CONFIG_ZONE_DMA
	556	max_zone_pfns[ZONE_DMA] = MAX_DMA_PFN;
	557	#endif
	558	#ifdef CONFIG_ZONE_DMA32
	559	max_zone_pfns[ZONE_DMA32] = MAX_DMA32_PFN;
	560	#endif
	561	max_zone_pfns[ZONE_NORMAL] = max_low_pfn;
	562	#ifdef CONFIG_HIGHMEM
	563	max_zone_pfns[ZONE_HIGHMEM] = max_pfn;
	564	#endif
	565
	566	free_area_init_nodes(max_zone_pfns);
	567	}
	568