[mirror_ubuntu-bionic-kernel.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 */

unsigned long __initdata pgt_buf_start;
unsigned long __meminitdata pgt_buf_end;
unsigned long __meminitdata pgt_buf_top;

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;
	}
}
void __init native_pagetable_reserve(u64 start, u64 end)
{
	memblock_reserve(start, end - start);
}

#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;
}

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--;
	}

	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;
}

/*
 * First calculate space needed for kernel direct mapping page tables to cover
 * mr[0].start to mr[nr_range - 1].end, while accounting for possible 2M and 1GB
 * pages. Then find enough contiguous space for those page tables.
 */
static void __init find_early_table_space(unsigned long start, unsigned long end)
{
	int i;
	unsigned long puds = 0, pmds = 0, ptes = 0, tables;
	unsigned long good_end;
	phys_addr_t base;
	struct map_range mr[NR_RANGE_MR];
	int nr_range;

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

	for (i = 0; i < nr_range; i++) {
		unsigned long range, extra;

		range = mr[i].end - mr[i].start;
		puds += (range + PUD_SIZE - 1) >> PUD_SHIFT;

		if (mr[i].page_size_mask & (1 << PG_LEVEL_1G)) {
			extra = range - ((range >> PUD_SHIFT) << PUD_SHIFT);
			pmds += (extra + PMD_SIZE - 1) >> PMD_SHIFT;
		} else {
			pmds += (range + PMD_SIZE - 1) >> PMD_SHIFT;
		}

		if (mr[i].page_size_mask & (1 << PG_LEVEL_2M)) {
			extra = range - ((range >> PMD_SHIFT) << PMD_SHIFT);
#ifdef CONFIG_X86_32
			extra += PMD_SIZE;
#endif
			ptes += (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
		} else {
			ptes += (range + PAGE_SIZE - 1) >> PAGE_SHIFT;
		}
	}

	tables = roundup(puds * sizeof(pud_t), PAGE_SIZE);
	tables += roundup(pmds * sizeof(pmd_t), PAGE_SIZE);
	tables += roundup(ptes * sizeof(pte_t), PAGE_SIZE);

#ifdef CONFIG_X86_32
	/* for fixmap */
	tables += roundup(__end_of_fixed_addresses * sizeof(pte_t), PAGE_SIZE);
	good_end = max_pfn_mapped << PAGE_SHIFT;
#endif

	base = memblock_find_in_range(start, good_end, tables, PAGE_SIZE);
	if (!base)
		panic("Cannot find space for the kernel page tables");

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

	printk(KERN_DEBUG "kernel direct mapping tables up to %#lx @ [mem %#010lx-%#010lx]\n",
		mr[nr_range - 1].end - 1, pgt_buf_start << PAGE_SHIFT,
		(pgt_buf_top << PAGE_SHIFT) - 1);
}

/*
 * 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);

	/*
	 * Find space for the kernel direct mapping tables.
	 *
	 * Later we should allocate these tables in the local node of the
	 * memory mapped. Unfortunately this is done currently before the
	 * nodes are discovered.
	 */
	if (!after_bootmem)
		find_early_table_space(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);

#ifdef CONFIG_X86_32
	early_ioremap_page_table_range_init();

	load_cr3(swapper_pg_dir);
#endif

	__flush_tlb_all();

	/*
	 * Reserve the kernel pagetable pages we used (pgt_buf_start -
	 * pgt_buf_end) and free the other ones (pgt_buf_end - pgt_buf_top)
	 * so that they can be reused for other purposes.
	 *
	 * On native it just means calling memblock_reserve, on Xen it also
	 * means marking RW the pagetable pages that we allocated before
	 * but that haven't been used.
	 *
	 * In fact on xen we mark RO the whole range pgt_buf_start -
	 * pgt_buf_top, because we have to make sure that when
	 * init_memory_mapping reaches the pagetable pages area, it maps
	 * RO all the pagetable pages, including the ones that are beyond
	 * pgt_buf_end at that time.
	 */
	if (!after_bootmem && pgt_buf_end > pgt_buf_start)
		x86_init.mapping.pagetable_reserve(PFN_PHYS(pgt_buf_start),
				PFN_PHYS(pgt_buf_end));

	if (!after_bootmem)
		early_memtest(start, end);

	return ret >> PAGE_SHIFT;
}

void __init init_mem_mapping(void)
{
	probe_page_size_mask();

	/* max_pfn_mapped is updated here */
	max_low_pfn_mapped = init_memory_mapping(0, max_low_pfn<<PAGE_SHIFT);
	max_pfn_mapped = max_low_pfn_mapped;

#ifdef CONFIG_X86_64
	if (max_pfn > max_low_pfn) {
		max_pfn_mapped = init_memory_mapping(1UL<<32,
						     max_pfn<<PAGE_SHIFT);
		/* can we preseve max_low_pfn ?*/
		max_low_pfn = max_pfn;
	}
#endif
}

/*
 * 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
d1b19426 YL	20	unsigned long __initdata pgt_buf_start;
	21	unsigned long __meminitdata pgt_buf_end;
	22	unsigned long __meminitdata pgt_buf_top;
f765090a PE	23
	24	int after_bootmem;
	25
	26	int direct_gbpages
	27	#ifdef CONFIG_DIRECT_GBPAGES
	28	= 1
	29	#endif
	30	;
	31
844ab6f9 JS	32	struct map_range {
	33	unsigned long start;
	34	unsigned long end;
	35	unsigned page_size_mask;
	36	};
	37
fa62aafe	38	static int page_size_mask;
f765090a	39
22ddfcaa	40	static void __init probe_page_size_mask(void)
fa62aafe YL	41	{
	42	#if !defined(CONFIG_DEBUG_PAGEALLOC) && !defined(CONFIG_KMEMCHECK)
	43	/*
	44	* For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
	45	* This will simplify cpa(), which otherwise needs to support splitting
	46	* large pages into small in interrupt context, etc.
	47	*/
	48	if (direct_gbpages)
	49	page_size_mask \|= 1 << PG_LEVEL_1G;
	50	if (cpu_has_pse)
	51	page_size_mask \|= 1 << PG_LEVEL_2M;
	52	#endif
	53
	54	/* Enable PSE if available */
	55	if (cpu_has_pse)
	56	set_in_cr4(X86_CR4_PSE);
	57
	58	/* Enable PGE if available */
	59	if (cpu_has_pge) {
	60	set_in_cr4(X86_CR4_PGE);
	61	__supported_pte_mask \|= _PAGE_GLOBAL;
	62	}
	63	}
53f8023f	64	void __init native_pagetable_reserve(u64 start, u64 end)
279b706b	65	{
24aa0788	66	memblock_reserve(start, end - start);
279b706b SS	67	}
279b706b SS	68
f765090a PE	69	#ifdef CONFIG_X86_32
	70	#define NR_RANGE_MR 3
	71	#else /* CONFIG_X86_64 */
	72	#define NR_RANGE_MR 5
	73	#endif
	74
dc9dd5cc JB	75	static int __meminit save_mr(struct map_range *mr, int nr_range,
	76	unsigned long start_pfn, unsigned long end_pfn,
	77	unsigned long page_size_mask)
f765090a PE	78	{
	79	if (start_pfn < end_pfn) {
	80	if (nr_range >= NR_RANGE_MR)
	81	panic("run out of range for init_memory_mapping\n");
	82	mr[nr_range].start = start_pfn<<PAGE_SHIFT;
	83	mr[nr_range].end = end_pfn<<PAGE_SHIFT;
	84	mr[nr_range].page_size_mask = page_size_mask;
	85	nr_range++;
	86	}
	87
	88	return nr_range;
	89	}
	90
4e33e065 YL	91	static int __meminit split_mem_range(struct map_range *mr, int nr_range,
	92	unsigned long start,
	93	unsigned long end)
f765090a	94	{
f765090a PE	95	unsigned long start_pfn, end_pfn;
f765090a PE	96	unsigned long pos;
4e33e065	97	int i;
f765090a PE	98
	99	/* head if not big page alignment ? */
	100	start_pfn = start >> PAGE_SHIFT;
	101	pos = start_pfn << PAGE_SHIFT;
	102	#ifdef CONFIG_X86_32
	103	/*
	104	* Don't use a large page for the first 2/4MB of memory
	105	* because there are often fixed size MTRRs in there
	106	* and overlapping MTRRs into large pages can cause
	107	* slowdowns.
	108	*/
	109	if (pos == 0)
	110	end_pfn = 1<<(PMD_SHIFT - PAGE_SHIFT);
	111	else
	112	end_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
	113	<< (PMD_SHIFT - PAGE_SHIFT);
	114	#else /* CONFIG_X86_64 */
	115	end_pfn = ((pos + (PMD_SIZE - 1)) >> PMD_SHIFT)
	116	<< (PMD_SHIFT - PAGE_SHIFT);
	117	#endif
	118	if (end_pfn > (end >> PAGE_SHIFT))
	119	end_pfn = end >> PAGE_SHIFT;
	120	if (start_pfn < end_pfn) {
	121	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
	122	pos = end_pfn << PAGE_SHIFT;
	123	}
	124
	125	/* big page (2M) range */
	126	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
	127	<< (PMD_SHIFT - PAGE_SHIFT);
	128	#ifdef CONFIG_X86_32
	129	end_pfn = (end>>PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
	130	#else /* CONFIG_X86_64 */
	131	end_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
	132	<< (PUD_SHIFT - PAGE_SHIFT);
	133	if (end_pfn > ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT)))
	134	end_pfn = ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT));
	135	#endif
	136
	137	if (start_pfn < end_pfn) {
	138	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	139	page_size_mask & (1<<PG_LEVEL_2M));
	140	pos = end_pfn << PAGE_SHIFT;
	141	}
	142
	143	#ifdef CONFIG_X86_64
	144	/* big page (1G) range */
	145	start_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
	146	<< (PUD_SHIFT - PAGE_SHIFT);
	147	end_pfn = (end >> PUD_SHIFT) << (PUD_SHIFT - PAGE_SHIFT);
	148	if (start_pfn < end_pfn) {
	149	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	150	page_size_mask &
	151	((1<<PG_LEVEL_2M)\|(1<<PG_LEVEL_1G)));
	152	pos = end_pfn << PAGE_SHIFT;
	153	}
	154
	155	/* tail is not big page (1G) alignment */
	156	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
	157	<< (PMD_SHIFT - PAGE_SHIFT);
	158	end_pfn = (end >> PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
	159	if (start_pfn < end_pfn) {
	160	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
	161	page_size_mask & (1<<PG_LEVEL_2M));
162	pos = end_pfn << PAGE_SHIFT;
163	}
164	#endif
165
166	/* tail is not big page (2M) alignment */
167	start_pfn = pos>>PAGE_SHIFT;
168	end_pfn = end>>PAGE_SHIFT;
169	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
170
171	/* try to merge same page size and continuous */
172	for (i = 0; nr_range > 1 && i < nr_range - 1; i++) {
173	unsigned long old_start;
174	if (mr[i].end != mr[i+1].start \|\|
175	mr[i].page_size_mask != mr[i+1].page_size_mask)
176	continue;
177	/* move it */
178	old_start = mr[i].start;
179	memmove(&mr[i], &mr[i+1],
180	(nr_range - 1 - i) * sizeof(struct map_range));
181	mr[i--].start = old_start;
182	nr_range--;
183	}
184
185	for (i = 0; i < nr_range; i++)
365811d6 BH	186	printk(KERN_DEBUG " [mem %#010lx-%#010lx] page %s\n",
365811d6 BH	187	mr[i].start, mr[i].end - 1,
f765090a PE	188	(mr[i].page_size_mask & (1<<PG_LEVEL_1G))?"1G":(
	189	(mr[i].page_size_mask & (1<<PG_LEVEL_2M))?"2M":"4k"));
	190
4e33e065 YL	191	return nr_range;
	192	}
	193
2086fe11 YL	194	/*
	195	* First calculate space needed for kernel direct mapping page tables to cover
	196	* mr[0].start to mr[nr_range - 1].end, while accounting for possible 2M and 1GB
	197	* pages. Then find enough contiguous space for those page tables.
	198	*/
84f1ae30	199	static void __init find_early_table_space(unsigned long start, unsigned long end)
2086fe11 YL	200	{
	201	int i;
	202	unsigned long puds = 0, pmds = 0, ptes = 0, tables;
84f1ae30	203	unsigned long good_end;
2086fe11	204	phys_addr_t base;
84f1ae30 YL	205	struct map_range mr[NR_RANGE_MR];
	206	int nr_range;
	207
	208	memset(mr, 0, sizeof(mr));
	209	nr_range = 0;
	210	nr_range = split_mem_range(mr, nr_range, start, end);
2086fe11 YL	211
	212	for (i = 0; i < nr_range; i++) {
	213	unsigned long range, extra;
	214
	215	range = mr[i].end - mr[i].start;
	216	puds += (range + PUD_SIZE - 1) >> PUD_SHIFT;
	217
	218	if (mr[i].page_size_mask & (1 << PG_LEVEL_1G)) {
	219	extra = range - ((range >> PUD_SHIFT) << PUD_SHIFT);
	220	pmds += (extra + PMD_SIZE - 1) >> PMD_SHIFT;
	221	} else {
	222	pmds += (range + PMD_SIZE - 1) >> PMD_SHIFT;
	223	}
	224
	225	if (mr[i].page_size_mask & (1 << PG_LEVEL_2M)) {
	226	extra = range - ((range >> PMD_SHIFT) << PMD_SHIFT);
	227	#ifdef CONFIG_X86_32
	228	extra += PMD_SIZE;
	229	#endif
	230	ptes += (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
	231	} else {
	232	ptes += (range + PAGE_SIZE - 1) >> PAGE_SHIFT;
	233	}
	234	}
	235
	236	tables = roundup(puds * sizeof(pud_t), PAGE_SIZE);
	237	tables += roundup(pmds * sizeof(pmd_t), PAGE_SIZE);
	238	tables += roundup(ptes * sizeof(pte_t), PAGE_SIZE);
	239
	240	#ifdef CONFIG_X86_32
	241	/* for fixmap */
	242	tables += roundup(__end_of_fixed_addresses * sizeof(pte_t), PAGE_SIZE);
2086fe11	243	good_end = max_pfn_mapped << PAGE_SHIFT;
28b6ff66	244	#endif
2086fe11 YL	245
	246	base = memblock_find_in_range(start, good_end, tables, PAGE_SIZE);
	247	if (!base)
	248	panic("Cannot find space for the kernel page tables");
	249
	250	pgt_buf_start = base >> PAGE_SHIFT;
	251	pgt_buf_end = pgt_buf_start;
	252	pgt_buf_top = pgt_buf_start + (tables >> PAGE_SHIFT);
	253
	254	printk(KERN_DEBUG "kernel direct mapping tables up to %#lx @ [mem %#010lx-%#010lx]\n",
	255	mr[nr_range - 1].end - 1, pgt_buf_start << PAGE_SHIFT,
	256	(pgt_buf_top << PAGE_SHIFT) - 1);
	257	}
	258
4e33e065 YL	259	/*
	260	* Setup the direct mapping of the physical memory at PAGE_OFFSET.
	261	* This runs before bootmem is initialized and gets pages directly from
	262	* the physical memory. To access them they are temporarily mapped.
	263	*/
	264	unsigned long __init_refok init_memory_mapping(unsigned long start,
	265	unsigned long end)
	266	{
	267	struct map_range mr[NR_RANGE_MR];
	268	unsigned long ret = 0;
	269	int nr_range, i;
	270
	271	pr_info("init_memory_mapping: [mem %#010lx-%#010lx]\n",
	272	start, end - 1);
	273
	274	memset(mr, 0, sizeof(mr));
	275	nr_range = split_mem_range(mr, 0, start, end);
	276
f765090a PE	277	/*
	278	* Find space for the kernel direct mapping tables.
	279	*
	280	* Later we should allocate these tables in the local node of the
	281	* memory mapped. Unfortunately this is done currently before the
	282	* nodes are discovered.
	283	*/
	284	if (!after_bootmem)
84f1ae30	285	find_early_table_space(start, end);
f765090a	286
f765090a PE	287	for (i = 0; i < nr_range; i++)
	288	ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
	289	mr[i].page_size_mask);
f765090a PE	290
	291	#ifdef CONFIG_X86_32
	292	early_ioremap_page_table_range_init();
	293
	294	load_cr3(swapper_pg_dir);
	295	#endif
	296
f765090a PE	297	__flush_tlb_all();
f765090a PE	298
279b706b SS	299	/*
	300	* Reserve the kernel pagetable pages we used (pgt_buf_start -
	301	* pgt_buf_end) and free the other ones (pgt_buf_end - pgt_buf_top)
	302	* so that they can be reused for other purposes.
	303	*
24aa0788 TH	304	* On native it just means calling memblock_reserve, on Xen it also
24aa0788 TH	305	* means marking RW the pagetable pages that we allocated before
279b706b SS	306	* but that haven't been used.
	307	*
	308	* In fact on xen we mark RO the whole range pgt_buf_start -
	309	* pgt_buf_top, because we have to make sure that when
	310	* init_memory_mapping reaches the pagetable pages area, it maps
	311	* RO all the pagetable pages, including the ones that are beyond
	312	* pgt_buf_end at that time.
	313	*/
d1b19426	314	if (!after_bootmem && pgt_buf_end > pgt_buf_start)
279b706b SS	315	x86_init.mapping.pagetable_reserve(PFN_PHYS(pgt_buf_start),
279b706b SS	316	PFN_PHYS(pgt_buf_end));
f765090a PE	317
	318	if (!after_bootmem)
	319	early_memtest(start, end);
	320
	321	return ret >> PAGE_SHIFT;
	322	}
	323
22ddfcaa YL	324	void __init init_mem_mapping(void)
	325	{
	326	probe_page_size_mask();
	327
	328	/* max_pfn_mapped is updated here */
	329	max_low_pfn_mapped = init_memory_mapping(0, max_low_pfn<<PAGE_SHIFT);
	330	max_pfn_mapped = max_low_pfn_mapped;
	331
	332	#ifdef CONFIG_X86_64
	333	if (max_pfn > max_low_pfn) {
	334	max_pfn_mapped = init_memory_mapping(1UL<<32,
	335	max_pfn<<PAGE_SHIFT);
	336	/* can we preseve max_low_pfn ?*/
	337	max_low_pfn = max_pfn;
	338	}
	339	#endif
	340	}
e5b2bb55	341
540aca06 PE	342	/*
	343	* devmem_is_allowed() checks to see if /dev/mem access to a certain address
	344	* is valid. The argument is a physical page number.
	345	*
	346	*
	347	* On x86, access has to be given to the first megabyte of ram because that area
	348	* contains bios code and data regions used by X and dosemu and similar apps.
	349	* Access has to be given to non-kernel-ram areas as well, these contain the PCI
	350	* mmio resources as well as potential bios/acpi data regions.
	351	*/
	352	int devmem_is_allowed(unsigned long pagenr)
	353	{
73e8f3d7	354	if (pagenr < 256)
540aca06 PE	355	return 1;
	356	if (iomem_is_exclusive(pagenr << PAGE_SHIFT))
	357	return 0;
	358	if (!page_is_ram(pagenr))
	359	return 1;
	360	return 0;
	361	}
	362
e5b2bb55 PE	363	void free_init_pages(char *what, unsigned long begin, unsigned long end)
e5b2bb55 PE	364	{
c967da6a YL	365	unsigned long addr;
c967da6a YL	366	unsigned long begin_aligned, end_aligned;
e5b2bb55	367
c967da6a YL	368	/* Make sure boundaries are page aligned */
	369	begin_aligned = PAGE_ALIGN(begin);
	370	end_aligned = end & PAGE_MASK;
	371
	372	if (WARN_ON(begin_aligned != begin \|\| end_aligned != end)) {
	373	begin = begin_aligned;
	374	end = end_aligned;
	375	}
	376
	377	if (begin >= end)
e5b2bb55 PE	378	return;
e5b2bb55 PE	379
c967da6a YL	380	addr = begin;
c967da6a YL	381
e5b2bb55 PE	382	/*
	383	* If debugging page accesses then do not free this memory but
	384	* mark them not present - any buggy init-section access will
	385	* create a kernel page fault:
	386	*/
	387	#ifdef CONFIG_DEBUG_PAGEALLOC
365811d6 BH	388	printk(KERN_INFO "debug: unmapping init [mem %#010lx-%#010lx]\n",
365811d6 BH	389	begin, end - 1);
e5b2bb55 PE	390	set_memory_np(begin, (end - begin) >> PAGE_SHIFT);
	391	#else
	392	/*
	393	* We just marked the kernel text read only above, now that
	394	* we are going to free part of that, we need to make that
5bd5a452	395	* writeable and non-executable first.
e5b2bb55	396	*/
5bd5a452	397	set_memory_nx(begin, (end - begin) >> PAGE_SHIFT);
e5b2bb55 PE	398	set_memory_rw(begin, (end - begin) >> PAGE_SHIFT);
	399
	400	printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
	401
	402	for (; addr < end; addr += PAGE_SIZE) {
	403	ClearPageReserved(virt_to_page(addr));
	404	init_page_count(virt_to_page(addr));
c967da6a	405	memset((void *)addr, POISON_FREE_INITMEM, PAGE_SIZE);
e5b2bb55 PE	406	free_page(addr);
	407	totalram_pages++;
	408	}
	409	#endif
	410	}
	411
	412	void free_initmem(void)
	413	{
	414	free_init_pages("unused kernel memory",
	415	(unsigned long)(&__init_begin),
	416	(unsigned long)(&__init_end));
	417	}
731ddea6 PE	418
731ddea6 PE	419	#ifdef CONFIG_BLK_DEV_INITRD
0d26d1d8	420	void __init free_initrd_mem(unsigned long start, unsigned long end)
731ddea6	421	{
c967da6a YL	422	/*
	423	* end could be not aligned, and We can not align that,
	424	* decompresser could be confused by aligned initrd_end
	425	* We already reserve the end partial page before in
	426	* - i386_start_kernel()
	427	* - x86_64_start_kernel()
	428	* - relocate_initrd()
	429	* So here We can do PAGE_ALIGN() safely to get partial page to be freed
	430	*/
	431	free_init_pages("initrd memory", start, PAGE_ALIGN(end));
731ddea6 PE	432	}
731ddea6 PE	433	#endif
17623915 PE	434
	435	void __init zone_sizes_init(void)
	436	{
	437	unsigned long max_zone_pfns[MAX_NR_ZONES];
	438
	439	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
	440
	441	#ifdef CONFIG_ZONE_DMA
	442	max_zone_pfns[ZONE_DMA] = MAX_DMA_PFN;
	443	#endif
	444	#ifdef CONFIG_ZONE_DMA32
	445	max_zone_pfns[ZONE_DMA32] = MAX_DMA32_PFN;
	446	#endif
	447	max_zone_pfns[ZONE_NORMAL] = max_low_pfn;
	448	#ifdef CONFIG_HIGHMEM
	449	max_zone_pfns[ZONE_HIGHMEM] = max_pfn;
	450	#endif
	451
	452	free_area_init_nodes(max_zone_pfns);
	453	}
	454