[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
;

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)
{
#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 pos;
	int i;

	/* head if not big page alignment ? */
	start_pfn = start >> PAGE_SHIFT;
	pos = start_pfn << PAGE_SHIFT;
#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 (pos == 0)
		end_pfn = 1<<(PMD_SHIFT - PAGE_SHIFT);
	else
		end_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
				 << (PMD_SHIFT - PAGE_SHIFT);
#else /* CONFIG_X86_64 */
	end_pfn = ((pos + (PMD_SIZE - 1)) >> PMD_SHIFT)
			<< (PMD_SHIFT - PAGE_SHIFT);
#endif
	if (end_pfn > (end >> PAGE_SHIFT))
		end_pfn = end >> PAGE_SHIFT;
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
		pos = end_pfn << PAGE_SHIFT;
	}

	/* big page (2M) range */
	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
			 << (PMD_SHIFT - PAGE_SHIFT);
#ifdef CONFIG_X86_32
	end_pfn = (end>>PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
#else /* CONFIG_X86_64 */
	end_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
			 << (PUD_SHIFT - PAGE_SHIFT);
	if (end_pfn > ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT)))
		end_pfn = ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT));
#endif

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

#ifdef CONFIG_X86_64
	/* big page (1G) range */
	start_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
			 << (PUD_SHIFT - PAGE_SHIFT);
	end_pfn = (end >> PUD_SHIFT) << (PUD_SHIFT - PAGE_SHIFT);
	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)));
		pos = end_pfn << PAGE_SHIFT;
	}

	/* tail is not big page (1G) alignment */
	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
			 << (PMD_SHIFT - PAGE_SHIFT);
	end_pfn = (end >> PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask & (1<<PG_LEVEL_2M));
		pos = end_pfn << PAGE_SHIFT;
	}
#endif

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