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

#include <linux/initrd.h>
#include <linux/ioport.h>
#include <linux/swap.h>

#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/system.h>
#include <asm/tlbflush.h>

unsigned long __initdata e820_table_start;
unsigned long __meminitdata e820_table_end;
unsigned long __meminitdata e820_table_top;

int after_bootmem;

int direct_gbpages
#ifdef CONFIG_DIRECT_GBPAGES
				= 1
#endif
;

static void __init find_early_table_space(unsigned long end, int use_pse,
					  int use_gbpages)
{
	unsigned long puds, pmds, ptes, tables, start;

	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
	tables = roundup(puds * sizeof(pud_t), PAGE_SIZE);

	if (use_gbpages) {
		unsigned long extra;

		extra = end - ((end>>PUD_SHIFT) << PUD_SHIFT);
		pmds = (extra + PMD_SIZE - 1) >> PMD_SHIFT;
	} else
		pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;

	tables += roundup(pmds * sizeof(pmd_t), PAGE_SIZE);

	if (use_pse) {
		unsigned long extra;

		extra = end - ((end>>PMD_SHIFT) << PMD_SHIFT);
#ifdef CONFIG_X86_32
		extra += PMD_SIZE;
#endif
		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
	} else
		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;

	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);
#endif

	/*
	 * RED-PEN putting page tables only on node 0 could
	 * cause a hotspot and fill up ZONE_DMA. The page tables
	 * need roughly 0.5KB per GB.
	 */
#ifdef CONFIG_X86_32
	start = 0x7000;
	e820_table_start = find_e820_area(start, max_pfn_mapped<<PAGE_SHIFT,
					tables, PAGE_SIZE);
#else /* CONFIG_X86_64 */
	start = 0x8000;
	e820_table_start = find_e820_area(start, end, tables, PAGE_SIZE);
#endif
	if (e820_table_start == -1UL)
		panic("Cannot find space for the kernel page tables");

	e820_table_start >>= PAGE_SHIFT;
	e820_table_end = e820_table_start;
	e820_table_top = e820_table_start + (tables >> PAGE_SHIFT);

	printk(KERN_DEBUG "kernel direct mapping tables up to %lx @ %lx-%lx\n",
		end, e820_table_start << PAGE_SHIFT, e820_table_top << PAGE_SHIFT);
}

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

#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;
}

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

/*
 * 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)
{
	unsigned long page_size_mask = 0;
	unsigned long start_pfn, end_pfn;
	unsigned long ret = 0;
	unsigned long pos;

	struct map_range mr[NR_RANGE_MR];
	int nr_range, i;
	int use_pse, use_gbpages;

	printk(KERN_INFO "init_memory_mapping: %016lx-%016lx\n", start, end);

	if (!after_bootmem)
		init_gbpages();

#ifdef CONFIG_DEBUG_PAGEALLOC
	/*
	 * 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.
	 */
	use_pse = use_gbpages = 0;
#else
	use_pse = cpu_has_pse;
	use_gbpages = direct_gbpages;
#endif

#ifdef CONFIG_X86_32
#ifdef CONFIG_X86_PAE
	set_nx();
	if (nx_enabled)
		printk(KERN_INFO "NX (Execute Disable) protection: active\n");
#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;
	}
#endif

	if (use_gbpages)
		page_size_mask |= 1 << PG_LEVEL_1G;
	if (use_pse)
		page_size_mask |= 1 << PG_LEVEL_2M;

	memset(mr, 0, sizeof(mr));
	nr_range = 0;

	/* 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 " %010lx - %010lx page %s\n",
				mr[i].start, mr[i].end,
			(mr[i].page_size_mask & (1<<PG_LEVEL_1G))?"1G":(
			 (mr[i].page_size_mask & (1<<PG_LEVEL_2M))?"2M":"4k"));

	/*
	 * 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(end, use_pse, use_gbpages);

#ifdef CONFIG_X86_32
	for (i = 0; i < nr_range; i++)
		kernel_physical_mapping_init(mr[i].start, mr[i].end,
					     mr[i].page_size_mask);
	ret = end;
#else /* CONFIG_X86_64 */
	for (i = 0; i < nr_range; i++)
		ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
						   mr[i].page_size_mask);
#endif

#ifdef CONFIG_X86_32
	early_ioremap_page_table_range_init();

	load_cr3(swapper_pg_dir);
#endif

#ifdef CONFIG_X86_64
	if (!after_bootmem)
		mmu_cr4_features = read_cr4();
#endif
	__flush_tlb_all();

	if (!after_bootmem && e820_table_end > e820_table_start)
		reserve_early(e820_table_start << PAGE_SHIFT,
				 e820_table_end << PAGE_SHIFT, "PGTABLE");

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

	return ret >> 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 = begin;

	if (addr >= end)
		return;

	/*
	 * 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 memory %08lx..%08lx\n",
		begin, PAGE_ALIGN(end));
	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 first.
	 */
	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 & ~(PAGE_SIZE-1)),
			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 free_initrd_mem(unsigned long start, unsigned long end)
{
	free_init_pages("initrd memory", start, end);
}
#endif
Commit	Line	Data
2c1b284e	1	#include <linux/initrd.h>
540aca06	2	#include <linux/ioport.h>
e5b2bb55	3	#include <linux/swap.h>
540aca06	4
e5b2bb55	5	#include <asm/cacheflush.h>
f765090a	6	#include <asm/e820.h>
4fcb2083	7	#include <asm/init.h>
e5b2bb55	8	#include <asm/page.h>
540aca06	9	#include <asm/page_types.h>
e5b2bb55 PE	10	#include <asm/sections.h>
e5b2bb55 PE	11	#include <asm/system.h>
f765090a PE	12	#include <asm/tlbflush.h>
f765090a PE	13
298af9d8 PE	14	unsigned long __initdata e820_table_start;
	15	unsigned long __meminitdata e820_table_end;
	16	unsigned long __meminitdata e820_table_top;
f765090a PE	17
	18	int after_bootmem;
	19
	20	int direct_gbpages
	21	#ifdef CONFIG_DIRECT_GBPAGES
	22	= 1
	23	#endif
	24	;
	25
	26	static void __init find_early_table_space(unsigned long end, int use_pse,
	27	int use_gbpages)
	28	{
	29	unsigned long puds, pmds, ptes, tables, start;
	30
	31	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
	32	tables = roundup(puds * sizeof(pud_t), PAGE_SIZE);
	33
	34	if (use_gbpages) {
	35	unsigned long extra;
	36
	37	extra = end - ((end>>PUD_SHIFT) << PUD_SHIFT);
	38	pmds = (extra + PMD_SIZE - 1) >> PMD_SHIFT;
	39	} else
	40	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
	41
	42	tables += roundup(pmds * sizeof(pmd_t), PAGE_SIZE);
	43
	44	if (use_pse) {
	45	unsigned long extra;
	46
	47	extra = end - ((end>>PMD_SHIFT) << PMD_SHIFT);
	48	#ifdef CONFIG_X86_32
	49	extra += PMD_SIZE;
	50	#endif
	51	ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
	52	} else
	53	ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
	54
	55	tables += roundup(ptes * sizeof(pte_t), PAGE_SIZE);
	56
	57	#ifdef CONFIG_X86_32
	58	/* for fixmap */
	59	tables += roundup(__end_of_fixed_addresses * sizeof(pte_t), PAGE_SIZE);
	60	#endif
	61
	62	/*
	63	* RED-PEN putting page tables only on node 0 could
	64	* cause a hotspot and fill up ZONE_DMA. The page tables
	65	* need roughly 0.5KB per GB.
	66	*/
	67	#ifdef CONFIG_X86_32
	68	start = 0x7000;
298af9d8	69	e820_table_start = find_e820_area(start, max_pfn_mapped<<PAGE_SHIFT,
f765090a PE	70	tables, PAGE_SIZE);
	71	#else /* CONFIG_X86_64 */
	72	start = 0x8000;
298af9d8	73	e820_table_start = find_e820_area(start, end, tables, PAGE_SIZE);
f765090a	74	#endif
298af9d8	75	if (e820_table_start == -1UL)
f765090a PE	76	panic("Cannot find space for the kernel page tables");
f765090a PE	77
298af9d8 PE	78	e820_table_start >>= PAGE_SHIFT;
	79	e820_table_end = e820_table_start;
	80	e820_table_top = e820_table_start + (tables >> PAGE_SHIFT);
f765090a PE	81
f765090a PE	82	printk(KERN_DEBUG "kernel direct mapping tables up to %lx @ %lx-%lx\n",
298af9d8	83	end, e820_table_start << PAGE_SHIFT, e820_table_top << PAGE_SHIFT);
f765090a PE	84	}
	85
	86	struct map_range {
	87	unsigned long start;
	88	unsigned long end;
	89	unsigned page_size_mask;
	90	};
	91
	92	#ifdef CONFIG_X86_32
	93	#define NR_RANGE_MR 3
	94	#else /* CONFIG_X86_64 */
	95	#define NR_RANGE_MR 5
	96	#endif
	97
dc9dd5cc JB	98	static int __meminit save_mr(struct map_range *mr, int nr_range,
	99	unsigned long start_pfn, unsigned long end_pfn,
	100	unsigned long page_size_mask)
f765090a PE	101	{
	102	if (start_pfn < end_pfn) {
	103	if (nr_range >= NR_RANGE_MR)
	104	panic("run out of range for init_memory_mapping\n");
	105	mr[nr_range].start = start_pfn<<PAGE_SHIFT;
	106	mr[nr_range].end = end_pfn<<PAGE_SHIFT;
	107	mr[nr_range].page_size_mask = page_size_mask;
	108	nr_range++;
	109	}
	110
	111	return nr_range;
	112	}
	113
	114	#ifdef CONFIG_X86_64
	115	static void __init init_gbpages(void)
	116	{
	117	if (direct_gbpages && cpu_has_gbpages)
	118	printk(KERN_INFO "Using GB pages for direct mapping\n");
	119	else
	120	direct_gbpages = 0;
	121	}
	122	#else
	123	static inline void init_gbpages(void)
	124	{
	125	}
	126	#endif
	127
	128	/*
	129	* Setup the direct mapping of the physical memory at PAGE_OFFSET.
	130	* This runs before bootmem is initialized and gets pages directly from
	131	* the physical memory. To access them they are temporarily mapped.
	132	*/
	133	unsigned long __init_refok init_memory_mapping(unsigned long start,
	134	unsigned long end)
	135	{
	136	unsigned long page_size_mask = 0;
	137	unsigned long start_pfn, end_pfn;
c77a3b59	138	unsigned long ret = 0;
f765090a	139	unsigned long pos;
f765090a PE	140
	141	struct map_range mr[NR_RANGE_MR];
	142	int nr_range, i;
	143	int use_pse, use_gbpages;
	144
	145	printk(KERN_INFO "init_memory_mapping: %016lx-%016lx\n", start, end);
	146
	147	if (!after_bootmem)
	148	init_gbpages();
	149
	150	#ifdef CONFIG_DEBUG_PAGEALLOC
	151	/*
	152	* For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
	153	* This will simplify cpa(), which otherwise needs to support splitting
	154	* large pages into small in interrupt context, etc.
	155	*/
	156	use_pse = use_gbpages = 0;
	157	#else
	158	use_pse = cpu_has_pse;
	159	use_gbpages = direct_gbpages;
	160	#endif
	161
	162	#ifdef CONFIG_X86_32
	163	#ifdef CONFIG_X86_PAE
	164	set_nx();
	165	if (nx_enabled)
	166	printk(KERN_INFO "NX (Execute Disable) protection: active\n");
	167	#endif
	168
	169	/* Enable PSE if available */
	170	if (cpu_has_pse)
	171	set_in_cr4(X86_CR4_PSE);
	172
	173	/* Enable PGE if available */
	174	if (cpu_has_pge) {
	175	set_in_cr4(X86_CR4_PGE);
	176	__supported_pte_mask \|= _PAGE_GLOBAL;
	177	}
	178	#endif
	179
	180	if (use_gbpages)
	181	page_size_mask \|= 1 << PG_LEVEL_1G;
	182	if (use_pse)
	183	page_size_mask \|= 1 << PG_LEVEL_2M;
	184
	185	memset(mr, 0, sizeof(mr));
	186	nr_range = 0;
	187
	188	/* head if not big page alignment ? */
	189	start_pfn = start >> PAGE_SHIFT;
	190	pos = start_pfn << PAGE_SHIFT;
	191	#ifdef CONFIG_X86_32
	192	/*
	193	* Don't use a large page for the first 2/4MB of memory
	194	* because there are often fixed size MTRRs in there
	195	* and overlapping MTRRs into large pages can cause
	196	* slowdowns.
	197	*/
	198	if (pos == 0)
	199	end_pfn = 1<<(PMD_SHIFT - PAGE_SHIFT);
	200	else
	201	end_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
	202	<< (PMD_SHIFT - PAGE_SHIFT);
	203	#else /* CONFIG_X86_64 */
204	end_pfn = ((pos + (PMD_SIZE - 1)) >> PMD_SHIFT)
205	<< (PMD_SHIFT - PAGE_SHIFT);
206	#endif
207	if (end_pfn > (end >> PAGE_SHIFT))
208	end_pfn = end >> PAGE_SHIFT;
209	if (start_pfn < end_pfn) {
210	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
211	pos = end_pfn << PAGE_SHIFT;
212	}
213
214	/* big page (2M) range */
215	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
216	<< (PMD_SHIFT - PAGE_SHIFT);
217	#ifdef CONFIG_X86_32
218	end_pfn = (end>>PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
219	#else /* CONFIG_X86_64 */
220	end_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
221	<< (PUD_SHIFT - PAGE_SHIFT);
222	if (end_pfn > ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT)))
223	end_pfn = ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT));
224	#endif
225
226	if (start_pfn < end_pfn) {
227	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
228	page_size_mask & (1<<PG_LEVEL_2M));
229	pos = end_pfn << PAGE_SHIFT;
230	}
231
232	#ifdef CONFIG_X86_64
233	/* big page (1G) range */
234	start_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
235	<< (PUD_SHIFT - PAGE_SHIFT);
236	end_pfn = (end >> PUD_SHIFT) << (PUD_SHIFT - PAGE_SHIFT);
237	if (start_pfn < end_pfn) {
238	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
239	page_size_mask &
240	((1<<PG_LEVEL_2M)\|(1<<PG_LEVEL_1G)));
241	pos = end_pfn << PAGE_SHIFT;
242	}
243
244	/* tail is not big page (1G) alignment */
245	start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
246	<< (PMD_SHIFT - PAGE_SHIFT);
247	end_pfn = (end >> PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
248	if (start_pfn < end_pfn) {
249	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
250	page_size_mask & (1<<PG_LEVEL_2M));
251	pos = end_pfn << PAGE_SHIFT;
252	}
253	#endif
254
255	/* tail is not big page (2M) alignment */
256	start_pfn = pos>>PAGE_SHIFT;
257	end_pfn = end>>PAGE_SHIFT;
258	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
259
260	/* try to merge same page size and continuous */
261	for (i = 0; nr_range > 1 && i < nr_range - 1; i++) {
262	unsigned long old_start;
263	if (mr[i].end != mr[i+1].start \|\|
264	mr[i].page_size_mask != mr[i+1].page_size_mask)
265	continue;
266	/* move it */
267	old_start = mr[i].start;
268	memmove(&mr[i], &mr[i+1],
269	(nr_range - 1 - i) * sizeof(struct map_range));
270	mr[i--].start = old_start;
271	nr_range--;
272	}
273
274	for (i = 0; i < nr_range; i++)
275	printk(KERN_DEBUG " %010lx - %010lx page %s\n",
276	mr[i].start, mr[i].end,
277	(mr[i].page_size_mask & (1<<PG_LEVEL_1G))?"1G":(
278	(mr[i].page_size_mask & (1<<PG_LEVEL_2M))?"2M":"4k"));
279
280	/*
281	* Find space for the kernel direct mapping tables.
282	*
283	* Later we should allocate these tables in the local node of the
284	* memory mapped. Unfortunately this is done currently before the
285	* nodes are discovered.
286	*/
287	if (!after_bootmem)
288	find_early_table_space(end, use_pse, use_gbpages);
289
290	#ifdef CONFIG_X86_32
291	for (i = 0; i < nr_range; i++)
e53fb04f PE	292	kernel_physical_mapping_init(mr[i].start, mr[i].end,
e53fb04f PE	293	mr[i].page_size_mask);
f765090a PE	294	ret = end;
	295	#else /* CONFIG_X86_64 */
	296	for (i = 0; i < nr_range; i++)
	297	ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
	298	mr[i].page_size_mask);
	299	#endif
	300
	301	#ifdef CONFIG_X86_32
	302	early_ioremap_page_table_range_init();
	303
	304	load_cr3(swapper_pg_dir);
	305	#endif
	306
	307	#ifdef CONFIG_X86_64
	308	if (!after_bootmem)
	309	mmu_cr4_features = read_cr4();
	310	#endif
	311	__flush_tlb_all();
	312
298af9d8 PE	313	if (!after_bootmem && e820_table_end > e820_table_start)
	314	reserve_early(e820_table_start << PAGE_SHIFT,
	315	e820_table_end << PAGE_SHIFT, "PGTABLE");
f765090a PE	316
	317	if (!after_bootmem)
	318	early_memtest(start, end);
	319
	320	return ret >> PAGE_SHIFT;
	321	}
	322
e5b2bb55	323
540aca06 PE	324	/*
	325	* devmem_is_allowed() checks to see if /dev/mem access to a certain address
	326	* is valid. The argument is a physical page number.
	327	*
	328	*
	329	* On x86, access has to be given to the first megabyte of ram because that area
	330	* contains bios code and data regions used by X and dosemu and similar apps.
	331	* Access has to be given to non-kernel-ram areas as well, these contain the PCI
	332	* mmio resources as well as potential bios/acpi data regions.
	333	*/
	334	int devmem_is_allowed(unsigned long pagenr)
	335	{
	336	if (pagenr <= 256)
	337	return 1;
	338	if (iomem_is_exclusive(pagenr << PAGE_SHIFT))
	339	return 0;
	340	if (!page_is_ram(pagenr))
	341	return 1;
	342	return 0;
	343	}
	344
e5b2bb55 PE	345	void free_init_pages(char *what, unsigned long begin, unsigned long end)
	346	{
	347	unsigned long addr = begin;
	348
	349	if (addr >= end)
	350	return;
	351
	352	/*
	353	* If debugging page accesses then do not free this memory but
	354	* mark them not present - any buggy init-section access will
	355	* create a kernel page fault:
	356	*/
	357	#ifdef CONFIG_DEBUG_PAGEALLOC
	358	printk(KERN_INFO "debug: unmapping init memory %08lx..%08lx\n",
	359	begin, PAGE_ALIGN(end));
	360	set_memory_np(begin, (end - begin) >> PAGE_SHIFT);
	361	#else
	362	/*
	363	* We just marked the kernel text read only above, now that
	364	* we are going to free part of that, we need to make that
	365	* writeable first.
	366	*/
	367	set_memory_rw(begin, (end - begin) >> PAGE_SHIFT);
	368
	369	printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
	370
	371	for (; addr < end; addr += PAGE_SIZE) {
	372	ClearPageReserved(virt_to_page(addr));
	373	init_page_count(virt_to_page(addr));
	374	memset((void *)(addr & ~(PAGE_SIZE-1)),
	375	POISON_FREE_INITMEM, PAGE_SIZE);
	376	free_page(addr);
	377	totalram_pages++;
	378	}
	379	#endif
	380	}
	381
	382	void free_initmem(void)
	383	{
	384	free_init_pages("unused kernel memory",
	385	(unsigned long)(&__init_begin),
	386	(unsigned long)(&__init_end));
	387	}
731ddea6 PE	388
	389	#ifdef CONFIG_BLK_DEV_INITRD
	390	void free_initrd_mem(unsigned long start, unsigned long end)
	391	{
	392	free_init_pages("initrd memory", start, end);
	393	}
	394	#endif