init.c 21.5 KB
Newer Older
1
#include <linux/gfp.h>
2
#include <linux/initrd.h>
3
#include <linux/ioport.h>
4
#include <linux/swap.h>
5
#include <linux/memblock.h>
P
Pekka Enberg 已提交
6
#include <linux/bootmem.h>	/* for max_low_pfn */
7

8
#include <asm/cacheflush.h>
9
#include <asm/e820.h>
10
#include <asm/init.h>
11
#include <asm/page.h>
12
#include <asm/page_types.h>
13
#include <asm/sections.h>
14
#include <asm/setup.h>
15
#include <asm/tlbflush.h>
16
#include <asm/tlb.h>
17
#include <asm/proto.h>
P
Pekka Enberg 已提交
18
#include <asm/dma.h>		/* for MAX_DMA_PFN */
19
#include <asm/microcode.h>
20

21 22 23 24 25 26 27
/*
 * We need to define the tracepoints somewhere, and tlb.c
 * is only compied when SMP=y.
 */
#define CREATE_TRACE_POINTS
#include <trace/events/tlb.h>

28 29
#include "mm_internal.h"

30 31
/*
 * Tables translating between page_cache_type_t and pte encoding.
32
 *
33 34 35 36 37
 * The default values are defined statically as minimal supported mode;
 * WC and WT fall back to UC-.  pat_init() updates these values to support
 * more cache modes, WC and WT, when it is safe to do so.  See pat_init()
 * for the details.  Note, __early_ioremap() used during early boot-time
 * takes pgprot_t (pte encoding) and does not use these tables.
38 39 40 41 42
 *
 *   Index into __cachemode2pte_tbl[] is the cachemode.
 *
 *   Index into __pte2cachemode_tbl[] are the caching attribute bits of the pte
 *   (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2.
43 44
 */
uint16_t __cachemode2pte_tbl[_PAGE_CACHE_MODE_NUM] = {
45
	[_PAGE_CACHE_MODE_WB      ]	= 0         | 0        ,
46
	[_PAGE_CACHE_MODE_WC      ]	= 0         | _PAGE_PCD,
47 48 49 50
	[_PAGE_CACHE_MODE_UC_MINUS]	= 0         | _PAGE_PCD,
	[_PAGE_CACHE_MODE_UC      ]	= _PAGE_PWT | _PAGE_PCD,
	[_PAGE_CACHE_MODE_WT      ]	= 0         | _PAGE_PCD,
	[_PAGE_CACHE_MODE_WP      ]	= 0         | _PAGE_PCD,
51
};
52
EXPORT_SYMBOL(__cachemode2pte_tbl);
53

54
uint8_t __pte2cachemode_tbl[8] = {
55
	[__pte2cm_idx( 0        | 0         | 0        )] = _PAGE_CACHE_MODE_WB,
56
	[__pte2cm_idx(_PAGE_PWT | 0         | 0        )] = _PAGE_CACHE_MODE_UC_MINUS,
57 58 59
	[__pte2cm_idx( 0        | _PAGE_PCD | 0        )] = _PAGE_CACHE_MODE_UC_MINUS,
	[__pte2cm_idx(_PAGE_PWT | _PAGE_PCD | 0        )] = _PAGE_CACHE_MODE_UC,
	[__pte2cm_idx( 0        | 0         | _PAGE_PAT)] = _PAGE_CACHE_MODE_WB,
60
	[__pte2cm_idx(_PAGE_PWT | 0         | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS,
61
	[__pte2cm_idx(0         | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS,
62 63
	[__pte2cm_idx(_PAGE_PWT | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC,
};
64
EXPORT_SYMBOL(__pte2cachemode_tbl);
65

66 67 68
static unsigned long __initdata pgt_buf_start;
static unsigned long __initdata pgt_buf_end;
static unsigned long __initdata pgt_buf_top;
69

70 71
static unsigned long min_pfn_mapped;

72 73
static bool __initdata can_use_brk_pgt = true;

74 75 76 77 78 79 80 81 82
/*
 * 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.
 */
Y
Yinghai Lu 已提交
83
__ref void *alloc_low_pages(unsigned int num)
84 85
{
	unsigned long pfn;
Y
Yinghai Lu 已提交
86
	int i;
87 88

	if (after_bootmem) {
Y
Yinghai Lu 已提交
89
		unsigned int order;
90

Y
Yinghai Lu 已提交
91 92 93
		order = get_order((unsigned long)num << PAGE_SHIFT);
		return (void *)__get_free_pages(GFP_ATOMIC | __GFP_NOTRACK |
						__GFP_ZERO, order);
94 95
	}

96
	if ((pgt_buf_end + num) > pgt_buf_top || !can_use_brk_pgt) {
97 98
		unsigned long ret;
		if (min_pfn_mapped >= max_pfn_mapped)
99
			panic("alloc_low_pages: ran out of memory");
100 101
		ret = memblock_find_in_range(min_pfn_mapped << PAGE_SHIFT,
					max_pfn_mapped << PAGE_SHIFT,
Y
Yinghai Lu 已提交
102
					PAGE_SIZE * num , PAGE_SIZE);
103
		if (!ret)
104
			panic("alloc_low_pages: can not alloc memory");
Y
Yinghai Lu 已提交
105
		memblock_reserve(ret, PAGE_SIZE * num);
106
		pfn = ret >> PAGE_SHIFT;
Y
Yinghai Lu 已提交
107 108 109
	} else {
		pfn = pgt_buf_end;
		pgt_buf_end += num;
110 111
		printk(KERN_DEBUG "BRK [%#010lx, %#010lx] PGTABLE\n",
			pfn << PAGE_SHIFT, (pgt_buf_end << PAGE_SHIFT) - 1);
Y
Yinghai Lu 已提交
112 113 114 115 116 117 118 119
	}

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

		adr = __va((pfn + i) << PAGE_SHIFT);
		clear_page(adr);
	}
120

Y
Yinghai Lu 已提交
121
	return __va(pfn << PAGE_SHIFT);
122 123
}

124 125
/* need 3 4k for initial PMD_SIZE,  3 4k for 0-ISA_END_ADDRESS */
#define INIT_PGT_BUF_SIZE	(6 * PAGE_SIZE)
Y
Yinghai Lu 已提交
126 127 128 129 130 131 132 133 134 135 136 137 138
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);
}

139 140
int after_bootmem;

141
early_param_on_off("gbpages", "nogbpages", direct_gbpages, CONFIG_X86_DIRECT_GBPAGES);
142

143 144 145 146 147 148
struct map_range {
	unsigned long start;
	unsigned long end;
	unsigned page_size_mask;
};

149
static int page_size_mask;
150

151
static void __init probe_page_size_mask(void)
152
{
153
#if !defined(CONFIG_KMEMCHECK)
154
	/*
155 156
	 * For CONFIG_KMEMCHECK or pagealloc debugging, identity mapping will
	 * use small pages.
157 158 159
	 * This will simplify cpa(), which otherwise needs to support splitting
	 * large pages into small in interrupt context, etc.
	 */
160
	if (cpu_has_pse && !debug_pagealloc_enabled())
161 162 163 164 165
		page_size_mask |= 1 << PG_LEVEL_2M;
#endif

	/* Enable PSE if available */
	if (cpu_has_pse)
A
Andy Lutomirski 已提交
166
		cr4_set_bits_and_update_boot(X86_CR4_PSE);
167 168

	/* Enable PGE if available */
169
	if (boot_cpu_has(X86_FEATURE_PGE)) {
A
Andy Lutomirski 已提交
170
		cr4_set_bits_and_update_boot(X86_CR4_PGE);
171
		__supported_pte_mask |= _PAGE_GLOBAL;
172 173
	} else
		__supported_pte_mask &= ~_PAGE_GLOBAL;
174 175

	/* Enable 1 GB linear kernel mappings if available: */
176
	if (direct_gbpages && boot_cpu_has(X86_FEATURE_GBPAGES)) {
177 178 179 180 181
		printk(KERN_INFO "Using GB pages for direct mapping\n");
		page_size_mask |= 1 << PG_LEVEL_1G;
	} else {
		direct_gbpages = 0;
	}
182
}
183

184 185 186 187 188 189
#ifdef CONFIG_X86_32
#define NR_RANGE_MR 3
#else /* CONFIG_X86_64 */
#define NR_RANGE_MR 5
#endif

190 191 192
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)
193 194 195 196 197 198 199 200 201 202 203 204 205
{
	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;
}

206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
/*
 * 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;
		}
	}
}

240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
static const char *page_size_string(struct map_range *mr)
{
	static const char str_1g[] = "1G";
	static const char str_2m[] = "2M";
	static const char str_4m[] = "4M";
	static const char str_4k[] = "4k";

	if (mr->page_size_mask & (1<<PG_LEVEL_1G))
		return str_1g;
	/*
	 * 32-bit without PAE has a 4M large page size.
	 * PG_LEVEL_2M is misnamed, but we can at least
	 * print out the right size in the string.
	 */
	if (IS_ENABLED(CONFIG_X86_32) &&
	    !IS_ENABLED(CONFIG_X86_PAE) &&
	    mr->page_size_mask & (1<<PG_LEVEL_2M))
		return str_4m;

	if (mr->page_size_mask & (1<<PG_LEVEL_2M))
		return str_2m;

	return str_4k;
}

265 266 267
static int __meminit split_mem_range(struct map_range *mr, int nr_range,
				     unsigned long start,
				     unsigned long end)
268
{
Y
Yinghai Lu 已提交
269
	unsigned long start_pfn, end_pfn, limit_pfn;
270
	unsigned long pfn;
271
	int i;
272

Y
Yinghai Lu 已提交
273 274
	limit_pfn = PFN_DOWN(end);

275
	/* head if not big page alignment ? */
276
	pfn = start_pfn = PFN_DOWN(start);
277 278 279 280 281 282 283
#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.
	 */
284
	if (pfn == 0)
285
		end_pfn = PFN_DOWN(PMD_SIZE);
286
	else
287
		end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
288
#else /* CONFIG_X86_64 */
289
	end_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
290
#endif
Y
Yinghai Lu 已提交
291 292
	if (end_pfn > limit_pfn)
		end_pfn = limit_pfn;
293 294
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
295
		pfn = end_pfn;
296 297 298
	}

	/* big page (2M) range */
299
	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
300
#ifdef CONFIG_X86_32
Y
Yinghai Lu 已提交
301
	end_pfn = round_down(limit_pfn, PFN_DOWN(PMD_SIZE));
302
#else /* CONFIG_X86_64 */
303
	end_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
Y
Yinghai Lu 已提交
304 305
	if (end_pfn > round_down(limit_pfn, PFN_DOWN(PMD_SIZE)))
		end_pfn = round_down(limit_pfn, PFN_DOWN(PMD_SIZE));
306 307 308 309 310
#endif

	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask & (1<<PG_LEVEL_2M));
311
		pfn = end_pfn;
312 313 314 315
	}

#ifdef CONFIG_X86_64
	/* big page (1G) range */
316
	start_pfn = round_up(pfn, PFN_DOWN(PUD_SIZE));
Y
Yinghai Lu 已提交
317
	end_pfn = round_down(limit_pfn, PFN_DOWN(PUD_SIZE));
318 319 320 321
	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)));
322
		pfn = end_pfn;
323 324 325
	}

	/* tail is not big page (1G) alignment */
326
	start_pfn = round_up(pfn, PFN_DOWN(PMD_SIZE));
Y
Yinghai Lu 已提交
327
	end_pfn = round_down(limit_pfn, PFN_DOWN(PMD_SIZE));
328 329 330
	if (start_pfn < end_pfn) {
		nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
				page_size_mask & (1<<PG_LEVEL_2M));
331
		pfn = end_pfn;
332 333 334 335
	}
#endif

	/* tail is not big page (2M) alignment */
336
	start_pfn = pfn;
Y
Yinghai Lu 已提交
337
	end_pfn = limit_pfn;
338 339
	nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);

340 341 342
	if (!after_bootmem)
		adjust_range_page_size_mask(mr, nr_range);

343 344 345 346 347 348 349 350 351 352 353 354 355 356 357
	/* 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++)
D
Dan Williams 已提交
358
		pr_debug(" [mem %#010lx-%#010lx] page %s\n",
359
				mr[i].start, mr[i].end - 1,
360
				page_size_string(&mr[i]));
361

362 363 364
	return nr_range;
}

365 366
struct range pfn_mapped[E820_X_MAX];
int nr_pfn_mapped;
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392

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;
}

393 394 395 396 397 398 399 400 401 402 403 404
/*
 * 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;

D
Dan Williams 已提交
405
	pr_debug("init_memory_mapping: [mem %#010lx-%#010lx]\n",
406 407 408 409 410
	       start, end - 1);

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

411 412 413 414
	for (i = 0; i < nr_range; i++)
		ret = kernel_physical_mapping_init(mr[i].start, mr[i].end,
						   mr[i].page_size_mask);

415 416
	add_pfn_range_mapped(start >> PAGE_SHIFT, ret >> PAGE_SHIFT);

417 418 419
	return ret >> PAGE_SHIFT;
}

420
/*
421 422 423 424 425 426 427 428 429 430 431
 * We need to iterate through the E820 memory map and create direct mappings
 * for only E820_RAM and E820_KERN_RESERVED regions. We cannot simply
 * create direct mappings for all pfns from [0 to max_low_pfn) and
 * [4GB to max_pfn) because of possible memory holes in high addresses
 * that cannot be marked as UC by fixed/variable range MTRRs.
 * Depending on the alignment of E820 ranges, this may possibly result
 * in using smaller size (i.e. 4K instead of 2M or 1G) page tables.
 *
 * init_mem_mapping() calls init_range_memory_mapping() with big range.
 * That range would have hole in the middle or ends, and only ram parts
 * will be mapped in init_range_memory_mapping().
432
 */
Y
Yinghai Lu 已提交
433
static unsigned long __init init_range_memory_mapping(
434 435
					   unsigned long r_start,
					   unsigned long r_end)
436 437
{
	unsigned long start_pfn, end_pfn;
Y
Yinghai Lu 已提交
438
	unsigned long mapped_ram_size = 0;
439 440 441
	int i;

	for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, NULL) {
442 443 444
		u64 start = clamp_val(PFN_PHYS(start_pfn), r_start, r_end);
		u64 end = clamp_val(PFN_PHYS(end_pfn), r_start, r_end);
		if (start >= end)
445 446
			continue;

447 448 449 450 451 452
		/*
		 * if it is overlapping with brk pgt, we need to
		 * alloc pgt buf from memblock instead.
		 */
		can_use_brk_pgt = max(start, (u64)pgt_buf_end<<PAGE_SHIFT) >=
				    min(end, (u64)pgt_buf_top<<PAGE_SHIFT);
453
		init_memory_mapping(start, end);
Y
Yinghai Lu 已提交
454
		mapped_ram_size += end - start;
455
		can_use_brk_pgt = true;
456
	}
Y
Yinghai Lu 已提交
457 458

	return mapped_ram_size;
459 460
}

461 462 463
static unsigned long __init get_new_step_size(unsigned long step_size)
{
	/*
464
	 * Initial mapped size is PMD_SIZE (2M).
465 466 467
	 * We can not set step_size to be PUD_SIZE (1G) yet.
	 * In worse case, when we cross the 1G boundary, and
	 * PG_LEVEL_2M is not set, we will need 1+1+512 pages (2M + 8k)
468 469
	 * to map 1G range with PTE. Hence we use one less than the
	 * difference of page table level shifts.
470
	 *
471 472 473 474 475
	 * Don't need to worry about overflow in the top-down case, on 32bit,
	 * when step_size is 0, round_down() returns 0 for start, and that
	 * turns it into 0x100000000ULL.
	 * In the bottom-up case, round_up(x, 0) returns 0 though too, which
	 * needs to be taken into consideration by the code below.
476
	 */
477
	return step_size << (PMD_SHIFT - PAGE_SHIFT - 1);
478 479
}

480 481 482 483 484 485 486 487 488 489 490 491
/**
 * memory_map_top_down - Map [map_start, map_end) top down
 * @map_start: start address of the target memory range
 * @map_end: end address of the target memory range
 *
 * This function will setup direct mapping for memory range
 * [map_start, map_end) in top-down. That said, the page tables
 * will be allocated at the end of the memory, and we map the
 * memory in top-down.
 */
static void __init memory_map_top_down(unsigned long map_start,
				       unsigned long map_end)
492
{
493
	unsigned long real_end, start, last_start;
Y
Yinghai Lu 已提交
494 495 496
	unsigned long step_size;
	unsigned long addr;
	unsigned long mapped_ram_size = 0;
497

498
	/* xen has big range in reserved near end of ram, skip it at first.*/
499
	addr = memblock_find_in_range(map_start, map_end, PMD_SIZE, PMD_SIZE);
Y
Yinghai Lu 已提交
500 501 502 503 504 505 506
	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;
507 508 509 510 511 512 513

	/*
	 * We start from the top (end of memory) and go to the bottom.
	 * The memblock_find_in_range() gets us a block of RAM from the
	 * end of RAM in [min_pfn_mapped, max_pfn_mapped) used as new pages
	 * for page table.
	 */
514
	while (last_start > map_start) {
Y
Yinghai Lu 已提交
515 516
		if (last_start > step_size) {
			start = round_down(last_start - 1, step_size);
517 518
			if (start < map_start)
				start = map_start;
Y
Yinghai Lu 已提交
519
		} else
520
			start = map_start;
521
		mapped_ram_size += init_range_memory_mapping(start,
Y
Yinghai Lu 已提交
522 523 524
							last_start);
		last_start = start;
		min_pfn_mapped = last_start >> PAGE_SHIFT;
525
		if (mapped_ram_size >= step_size)
526
			step_size = get_new_step_size(step_size);
Y
Yinghai Lu 已提交
527 528
	}

529 530 531 532
	if (real_end < map_end)
		init_range_memory_mapping(real_end, map_end);
}

533 534 535 536 537 538 539 540 541 542 543 544 545 546
/**
 * memory_map_bottom_up - Map [map_start, map_end) bottom up
 * @map_start: start address of the target memory range
 * @map_end: end address of the target memory range
 *
 * This function will setup direct mapping for memory range
 * [map_start, map_end) in bottom-up. Since we have limited the
 * bottom-up allocation above the kernel, the page tables will
 * be allocated just above the kernel and we map the memory
 * in [map_start, map_end) in bottom-up.
 */
static void __init memory_map_bottom_up(unsigned long map_start,
					unsigned long map_end)
{
547
	unsigned long next, start;
548 549 550 551 552 553 554 555 556 557 558 559 560 561
	unsigned long mapped_ram_size = 0;
	/* step_size need to be small so pgt_buf from BRK could cover it */
	unsigned long step_size = PMD_SIZE;

	start = map_start;
	min_pfn_mapped = start >> PAGE_SHIFT;

	/*
	 * We start from the bottom (@map_start) and go to the top (@map_end).
	 * The memblock_find_in_range() gets us a block of RAM from the
	 * end of RAM in [min_pfn_mapped, max_pfn_mapped) used as new pages
	 * for page table.
	 */
	while (start < map_end) {
562
		if (step_size && map_end - start > step_size) {
563 564 565
			next = round_up(start + 1, step_size);
			if (next > map_end)
				next = map_end;
566
		} else {
567
			next = map_end;
568
		}
569

570
		mapped_ram_size += init_range_memory_mapping(start, next);
571 572
		start = next;

573
		if (mapped_ram_size >= step_size)
574 575 576 577
			step_size = get_new_step_size(step_size);
	}
}

578 579 580 581 582 583 584 585 586 587 588 589 590 591 592
void __init init_mem_mapping(void)
{
	unsigned long end;

	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);

593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
	/*
	 * If the allocation is in bottom-up direction, we setup direct mapping
	 * in bottom-up, otherwise we setup direct mapping in top-down.
	 */
	if (memblock_bottom_up()) {
		unsigned long kernel_end = __pa_symbol(_end);

		/*
		 * we need two separate calls here. This is because we want to
		 * allocate page tables above the kernel. So we first map
		 * [kernel_end, end) to make memory above the kernel be mapped
		 * as soon as possible. And then use page tables allocated above
		 * the kernel to map [ISA_END_ADDRESS, kernel_end).
		 */
		memory_map_bottom_up(kernel_end, end);
		memory_map_bottom_up(ISA_END_ADDRESS, kernel_end);
	} else {
		memory_map_top_down(ISA_END_ADDRESS, end);
	}
Y
Yinghai Lu 已提交
612

613 614 615 616 617
#ifdef CONFIG_X86_64
	if (max_pfn > max_low_pfn) {
		/* can we preseve max_low_pfn ?*/
		max_low_pfn = max_pfn;
	}
618 619
#else
	early_ioremap_page_table_range_init();
620 621
#endif

622 623 624
	load_cr3(swapper_pg_dir);
	__flush_tlb_all();

625
	early_memtest(0, max_pfn_mapped << PAGE_SHIFT);
626
}
627

628 629 630 631 632 633
/*
 * 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
634
 * contains BIOS code and data regions used by X and dosemu and similar apps.
635 636 637 638 639
 * 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)
{
640
	if (pagenr < 256)
641 642 643 644 645 646 647 648
		return 1;
	if (iomem_is_exclusive(pagenr << PAGE_SHIFT))
		return 0;
	if (!page_is_ram(pagenr))
		return 1;
	return 0;
}

649 650
void free_init_pages(char *what, unsigned long begin, unsigned long end)
{
651
	unsigned long begin_aligned, end_aligned;
652

653 654 655 656 657 658 659 660 661 662
	/* 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)
663 664 665 666 667 668 669
		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:
	 */
670 671 672 673 674 675 676 677 678 679 680 681
	if (debug_pagealloc_enabled()) {
		pr_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);
682

683 684 685
		free_reserved_area((void *)begin, (void *)end,
				   POISON_FREE_INITMEM, what);
	}
686 687 688 689
}

void free_initmem(void)
{
690
	free_init_pages("unused kernel",
691 692 693
			(unsigned long)(&__init_begin),
			(unsigned long)(&__init_end));
}
694 695

#ifdef CONFIG_BLK_DEV_INITRD
696
void __init free_initrd_mem(unsigned long start, unsigned long end)
697
{
698 699 700 701 702 703 704
	/*
	 * Remember, initrd memory may contain microcode or other useful things.
	 * Before we lose initrd mem, we need to find a place to hold them
	 * now that normal virtual memory is enabled.
	 */
	save_microcode_in_initrd();

705 706 707 708 709 710 711 712 713
	/*
	 * 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
	 */
714
	free_init_pages("initrd", start, PAGE_ALIGN(end));
715 716
}
#endif
P
Pekka Enberg 已提交
717 718 719 720 721 722 723 724

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
X
Xishi Qiu 已提交
725
	max_zone_pfns[ZONE_DMA]		= min(MAX_DMA_PFN, max_low_pfn);
P
Pekka Enberg 已提交
726 727
#endif
#ifdef CONFIG_ZONE_DMA32
X
Xishi Qiu 已提交
728
	max_zone_pfns[ZONE_DMA32]	= min(MAX_DMA32_PFN, max_low_pfn);
P
Pekka Enberg 已提交
729 730 731 732 733 734 735 736 737
#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);
}

738 739 740 741 742 743 744 745 746
DEFINE_PER_CPU_SHARED_ALIGNED(struct tlb_state, cpu_tlbstate) = {
#ifdef CONFIG_SMP
	.active_mm = &init_mm,
	.state = 0,
#endif
	.cr4 = ~0UL,	/* fail hard if we screw up cr4 shadow initialization */
};
EXPORT_SYMBOL_GPL(cpu_tlbstate);

747 748 749 750 751 752 753 754
void update_cache_mode_entry(unsigned entry, enum page_cache_mode cache)
{
	/* entry 0 MUST be WB (hardwired to speed up translations) */
	BUG_ON(!entry && cache != _PAGE_CACHE_MODE_WB);

	__cachemode2pte_tbl[cache] = __cm_idx2pte(entry);
	__pte2cachemode_tbl[entry] = cache;
}