mem.c 14.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/*
 *  PowerPC version
 *    Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
 *
 *  Modifications by Paul Mackerras (PowerMac) (paulus@cs.anu.edu.au)
 *  and Cort Dougan (PReP) (cort@cs.nmt.edu)
 *    Copyright (C) 1996 Paul Mackerras
 *  PPC44x/36-bit changes by Matt Porter (mporter@mvista.com)
 *
 *  Derived from "arch/i386/mm/init.c"
 *    Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *
 *  This program is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU General Public License
 *  as published by the Free Software Foundation; either version
 *  2 of the License, or (at your option) any later version.
 *
 */

#include <linux/module.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
25
#include <linux/gfp.h>
26 27 28 29 30 31 32 33
#include <linux/types.h>
#include <linux/mm.h>
#include <linux/stddef.h>
#include <linux/init.h>
#include <linux/bootmem.h>
#include <linux/highmem.h>
#include <linux/initrd.h>
#include <linux/pagemap.h>
34
#include <linux/suspend.h>
Y
Yinghai Lu 已提交
35
#include <linux/memblock.h>
36
#include <linux/hugetlb.h>
37 38 39 40 41 42 43 44 45 46 47

#include <asm/pgalloc.h>
#include <asm/prom.h>
#include <asm/io.h>
#include <asm/mmu_context.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
#include <asm/smp.h>
#include <asm/machdep.h>
#include <asm/btext.h>
#include <asm/tlb.h>
48
#include <asm/sections.h>
49
#include <asm/sparsemem.h>
50
#include <asm/vdso.h>
51
#include <asm/fixmap.h>
52
#include <asm/swiotlb.h>
53 54 55 56 57 58 59 60

#include "mmu_decl.h"

#ifndef CPU_FTR_COHERENT_ICACHE
#define CPU_FTR_COHERENT_ICACHE	0	/* XXX for now */
#define CPU_FTR_NOEXECUTE	0
#endif

61 62
int init_bootmem_done;
int mem_init_done;
63
phys_addr_t memory_limit;
64

65 66 67 68 69 70 71 72 73 74 75 76 77 78
#ifdef CONFIG_HIGHMEM
pte_t *kmap_pte;
pgprot_t kmap_prot;

EXPORT_SYMBOL(kmap_prot);
EXPORT_SYMBOL(kmap_pte);

static inline pte_t *virt_to_kpte(unsigned long vaddr)
{
	return pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr),
			vaddr), vaddr), vaddr);
}
#endif

79 80 81
int page_is_ram(unsigned long pfn)
{
#ifndef CONFIG_PPC64	/* XXX for now */
82
	return pfn < max_pfn;
83
#else
84
	unsigned long paddr = (pfn << PAGE_SHIFT);
85
	int i;
Y
Yinghai Lu 已提交
86
	for (i=0; i < memblock.memory.cnt; i++) {
87 88
		unsigned long base;

89
		base = memblock.memory.regions[i].base;
90 91

		if ((paddr >= base) &&
92
			(paddr < (base + memblock.memory.regions[i].size))) {
93 94 95 96 97 98 99 100
			return 1;
		}
	}

	return 0;
#endif
}

101
pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn,
102 103 104
			      unsigned long size, pgprot_t vma_prot)
{
	if (ppc_md.phys_mem_access_prot)
105
		return ppc_md.phys_mem_access_prot(file, pfn, size, vma_prot);
106

107
	if (!page_is_ram(pfn))
108 109
		vma_prot = pgprot_noncached(vma_prot);

110 111 112 113
	return vma_prot;
}
EXPORT_SYMBOL(phys_mem_access_prot);

P
Paul Mackerras 已提交
114 115
#ifdef CONFIG_MEMORY_HOTPLUG

116 117 118 119 120 121 122
#ifdef CONFIG_NUMA
int memory_add_physaddr_to_nid(u64 start)
{
	return hot_add_scn_to_nid(start);
}
#endif

123
int arch_add_memory(int nid, u64 start, u64 size)
P
Paul Mackerras 已提交
124
{
125
	struct pglist_data *pgdata;
P
Paul Mackerras 已提交
126 127 128 129
	struct zone *zone;
	unsigned long start_pfn = start >> PAGE_SHIFT;
	unsigned long nr_pages = size >> PAGE_SHIFT;

130 131
	pgdata = NODE_DATA(nid);

132
	start = (unsigned long)__va(start);
133 134
	create_section_mapping(start, start + size);

P
Paul Mackerras 已提交
135 136 137
	/* this should work for most non-highmem platforms */
	zone = pgdata->node_zones;

138
	return __add_pages(nid, zone, start_pfn, nr_pages);
P
Paul Mackerras 已提交
139
}
140
#endif /* CONFIG_MEMORY_HOTPLUG */
141 142 143

/*
 * walk_memory_resource() needs to make sure there is no holes in a given
144
 * memory range.  PPC64 does not maintain the memory layout in /proc/iomem.
Y
Yinghai Lu 已提交
145
 * Instead it maintains it in memblock.memory structures.  Walk through the
146
 * memory regions, find holes and callback for contiguous regions.
147 148
 */
int
K
KAMEZAWA Hiroyuki 已提交
149 150
walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages,
		void *arg, int (*func)(unsigned long, unsigned long, void *))
151
{
152
	struct memblock_region res;
153 154 155 156 157 158 159 160
	unsigned long pfn, len;
	u64 end;
	int ret = -1;

	res.base = (u64) start_pfn << PAGE_SHIFT;
	res.size = (u64) nr_pages << PAGE_SHIFT;

	end = res.base + res.size - 1;
Y
Yinghai Lu 已提交
161
	while ((res.base < end) && (memblock_find(&res) >= 0)) {
162 163 164 165 166 167 168 169 170
		pfn = (unsigned long)(res.base >> PAGE_SHIFT);
		len = (unsigned long)(res.size >> PAGE_SHIFT);
		ret = (*func)(pfn, len, arg);
		if (ret)
			break;
		res.base += (res.size + 1);
		res.size = (end - res.base + 1);
	}
	return ret;
171
}
K
KAMEZAWA Hiroyuki 已提交
172
EXPORT_SYMBOL_GPL(walk_system_ram_range);
173

174 175 176 177 178 179 180 181 182 183 184 185 186
/*
 * Initialize the bootmem system and give it all the memory we
 * have available.  If we are using highmem, we only put the
 * lowmem into the bootmem system.
 */
#ifndef CONFIG_NEED_MULTIPLE_NODES
void __init do_init_bootmem(void)
{
	unsigned long i;
	unsigned long start, bootmap_pages;
	unsigned long total_pages;
	int boot_mapsize;

Y
Yinghai Lu 已提交
187 188
	max_low_pfn = max_pfn = memblock_end_of_DRAM() >> PAGE_SHIFT;
	total_pages = (memblock_end_of_DRAM() - memstart_addr) >> PAGE_SHIFT;
189 190
#ifdef CONFIG_HIGHMEM
	total_pages = total_lowmem >> PAGE_SHIFT;
191
	max_low_pfn = lowmem_end_addr >> PAGE_SHIFT;
192 193 194 195 196 197 198 199 200
#endif

	/*
	 * Find an area to use for the bootmem bitmap.  Calculate the size of
	 * bitmap required as (Total Memory) / PAGE_SIZE / BITS_PER_BYTE.
	 * Add 1 additional page in case the address isn't page-aligned.
	 */
	bootmap_pages = bootmem_bootmap_pages(total_pages);

Y
Yinghai Lu 已提交
201
	start = memblock_alloc(bootmap_pages << PAGE_SHIFT, PAGE_SIZE);
202

203 204
	min_low_pfn = MEMORY_START >> PAGE_SHIFT;
	boot_mapsize = init_bootmem_node(NODE_DATA(0), start >> PAGE_SHIFT, min_low_pfn, max_low_pfn);
205

206
	/* Add active regions with valid PFNs */
Y
Yinghai Lu 已提交
207
	for (i = 0; i < memblock.memory.cnt; i++) {
208
		unsigned long start_pfn, end_pfn;
209
		start_pfn = memblock.memory.regions[i].base >> PAGE_SHIFT;
Y
Yinghai Lu 已提交
210
		end_pfn = start_pfn + memblock_size_pages(&memblock.memory, i);
211 212 213
		add_active_range(0, start_pfn, end_pfn);
	}

214 215 216 217
	/* Add all physical memory to the bootmem map, mark each area
	 * present.
	 */
#ifdef CONFIG_HIGHMEM
218
	free_bootmem_with_active_regions(0, lowmem_end_addr >> PAGE_SHIFT);
219 220

	/* reserve the sections we're already using */
Y
Yinghai Lu 已提交
221
	for (i = 0; i < memblock.reserved.cnt; i++) {
222
		unsigned long addr = memblock.reserved.regions[i].base +
Y
Yinghai Lu 已提交
223
				     memblock_size_bytes(&memblock.reserved, i) - 1;
224
		if (addr < lowmem_end_addr)
225
			reserve_bootmem(memblock.reserved.regions[i].base,
Y
Yinghai Lu 已提交
226
					memblock_size_bytes(&memblock.reserved, i),
227
					BOOTMEM_DEFAULT);
228
		else if (memblock.reserved.regions[i].base < lowmem_end_addr) {
229
			unsigned long adjusted_size = lowmem_end_addr -
230 231
				      memblock.reserved.regions[i].base;
			reserve_bootmem(memblock.reserved.regions[i].base,
232
					adjusted_size, BOOTMEM_DEFAULT);
233 234
		}
	}
235 236
#else
	free_bootmem_with_active_regions(0, max_pfn);
237 238

	/* reserve the sections we're already using */
Y
Yinghai Lu 已提交
239
	for (i = 0; i < memblock.reserved.cnt; i++)
240
		reserve_bootmem(memblock.reserved.regions[i].base,
Y
Yinghai Lu 已提交
241
				memblock_size_bytes(&memblock.reserved, i),
242
				BOOTMEM_DEFAULT);
243

244
#endif
245
	/* XXX need to clip this if using highmem? */
246 247
	sparse_memory_present_with_active_regions(0);

248 249 250
	init_bootmem_done = 1;
}

251 252 253
/* mark pages that don't exist as nosave */
static int __init mark_nonram_nosave(void)
{
Y
Yinghai Lu 已提交
254 255
	unsigned long memblock_next_region_start_pfn,
		      memblock_region_max_pfn;
256 257
	int i;

Y
Yinghai Lu 已提交
258 259
	for (i = 0; i < memblock.memory.cnt - 1; i++) {
		memblock_region_max_pfn =
260 261
			(memblock.memory.regions[i].base >> PAGE_SHIFT) +
			(memblock.memory.regions[i].size >> PAGE_SHIFT);
Y
Yinghai Lu 已提交
262
		memblock_next_region_start_pfn =
263
			memblock.memory.regions[i+1].base >> PAGE_SHIFT;
264

Y
Yinghai Lu 已提交
265 266 267
		if (memblock_region_max_pfn < memblock_next_region_start_pfn)
			register_nosave_region(memblock_region_max_pfn,
					       memblock_next_region_start_pfn);
268 269 270 271 272
	}

	return 0;
}

273 274 275 276 277
/*
 * paging_init() sets up the page tables - in fact we've already done this.
 */
void __init paging_init(void)
{
Y
Yinghai Lu 已提交
278 279
	unsigned long total_ram = memblock_phys_mem_size();
	phys_addr_t top_of_ram = memblock_end_of_DRAM();
280
	unsigned long max_zone_pfns[MAX_NR_ZONES];
281

282 283 284 285 286 287 288 289
#ifdef CONFIG_PPC32
	unsigned long v = __fix_to_virt(__end_of_fixed_addresses - 1);
	unsigned long end = __fix_to_virt(FIX_HOLE);

	for (; v < end; v += PAGE_SIZE)
		map_page(v, 0, 0); /* XXX gross */
#endif

290 291
#ifdef CONFIG_HIGHMEM
	map_page(PKMAP_BASE, 0, 0);	/* XXX gross */
292 293 294
	pkmap_page_table = virt_to_kpte(PKMAP_BASE);

	kmap_pte = virt_to_kpte(__fix_to_virt(FIX_KMAP_BEGIN));
295 296 297
	kmap_prot = PAGE_KERNEL;
#endif /* CONFIG_HIGHMEM */

298
	printk(KERN_DEBUG "Top of RAM: 0x%llx, Total RAM: 0x%lx\n",
299
	       (unsigned long long)top_of_ram, total_ram);
300
	printk(KERN_DEBUG "Memory hole size: %ldMB\n",
301
	       (long int)((top_of_ram - total_ram) >> 20));
302
	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
303
#ifdef CONFIG_HIGHMEM
304
	max_zone_pfns[ZONE_DMA] = lowmem_end_addr >> PAGE_SHIFT;
305
	max_zone_pfns[ZONE_HIGHMEM] = top_of_ram >> PAGE_SHIFT;
306
#else
307
	max_zone_pfns[ZONE_DMA] = top_of_ram >> PAGE_SHIFT;
308 309
#endif
	free_area_init_nodes(max_zone_pfns);
310 311

	mark_nonram_nosave();
312 313 314 315 316 317 318 319 320 321 322 323 324
}
#endif /* ! CONFIG_NEED_MULTIPLE_NODES */

void __init mem_init(void)
{
#ifdef CONFIG_NEED_MULTIPLE_NODES
	int nid;
#endif
	pg_data_t *pgdat;
	unsigned long i;
	struct page *page;
	unsigned long reservedpages = 0, codesize, initsize, datasize, bsssize;

325 326 327 328 329
#ifdef CONFIG_SWIOTLB
	if (ppc_swiotlb_enable)
		swiotlb_init(1);
#endif

Y
Yinghai Lu 已提交
330
	num_physpages = memblock.memory.size >> PAGE_SHIFT;
331 332 333 334 335
	high_memory = (void *) __va(max_low_pfn * PAGE_SIZE);

#ifdef CONFIG_NEED_MULTIPLE_NODES
        for_each_online_node(nid) {
		if (NODE_DATA(nid)->node_spanned_pages != 0) {
336
			printk("freeing bootmem node %d\n", nid);
337 338 339 340 341
			totalram_pages +=
				free_all_bootmem_node(NODE_DATA(nid));
		}
	}
#else
342
	max_mapnr = max_pfn;
343 344
	totalram_pages += free_all_bootmem();
#endif
345
	for_each_online_pgdat(pgdat) {
346
		for (i = 0; i < pgdat->node_spanned_pages; i++) {
347 348
			if (!pfn_valid(pgdat->node_start_pfn + i))
				continue;
349 350 351 352 353 354 355
			page = pgdat_page_nr(pgdat, i);
			if (PageReserved(page))
				reservedpages++;
		}
	}

	codesize = (unsigned long)&_sdata - (unsigned long)&_stext;
356
	datasize = (unsigned long)&_edata - (unsigned long)&_sdata;
357 358 359 360 361 362 363
	initsize = (unsigned long)&__init_end - (unsigned long)&__init_begin;
	bsssize = (unsigned long)&__bss_stop - (unsigned long)&__bss_start;

#ifdef CONFIG_HIGHMEM
	{
		unsigned long pfn, highmem_mapnr;

364
		highmem_mapnr = lowmem_end_addr >> PAGE_SHIFT;
365 366
		for (pfn = highmem_mapnr; pfn < max_mapnr; ++pfn) {
			struct page *page = pfn_to_page(pfn);
Y
Yinghai Lu 已提交
367
			if (memblock_is_reserved(pfn << PAGE_SHIFT))
368
				continue;
369
			ClearPageReserved(page);
370
			init_page_count(page);
371 372
			__free_page(page);
			totalhigh_pages++;
373
			reservedpages--;
374 375
		}
		totalram_pages += totalhigh_pages;
376
		printk(KERN_DEBUG "High memory: %luk\n",
377 378 379 380 381 382
		       totalhigh_pages << (PAGE_SHIFT-10));
	}
#endif /* CONFIG_HIGHMEM */

	printk(KERN_INFO "Memory: %luk/%luk available (%luk kernel code, "
	       "%luk reserved, %luk data, %luk bss, %luk init)\n",
383
		nr_free_pages() << (PAGE_SHIFT-10),
384 385 386 387 388 389 390
		num_physpages << (PAGE_SHIFT-10),
		codesize >> 10,
		reservedpages << (PAGE_SHIFT-10),
		datasize >> 10,
		bsssize >> 10,
		initsize >> 10);

391 392 393 394 395 396 397
#ifdef CONFIG_PPC32
	pr_info("Kernel virtual memory layout:\n");
	pr_info("  * 0x%08lx..0x%08lx  : fixmap\n", FIXADDR_START, FIXADDR_TOP);
#ifdef CONFIG_HIGHMEM
	pr_info("  * 0x%08lx..0x%08lx  : highmem PTEs\n",
		PKMAP_BASE, PKMAP_ADDR(LAST_PKMAP));
#endif /* CONFIG_HIGHMEM */
398 399 400 401
#ifdef CONFIG_NOT_COHERENT_CACHE
	pr_info("  * 0x%08lx..0x%08lx  : consistent mem\n",
		IOREMAP_TOP, IOREMAP_TOP + CONFIG_CONSISTENT_SIZE);
#endif /* CONFIG_NOT_COHERENT_CACHE */
402 403 404 405 406 407
	pr_info("  * 0x%08lx..0x%08lx  : early ioremap\n",
		ioremap_bot, IOREMAP_TOP);
	pr_info("  * 0x%08lx..0x%08lx  : vmalloc & ioremap\n",
		VMALLOC_START, VMALLOC_END);
#endif /* CONFIG_PPC32 */

408 409 410
	mem_init_done = 1;
}

411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
/*
 * This is called when a page has been modified by the kernel.
 * It just marks the page as not i-cache clean.  We do the i-cache
 * flush later when the page is given to a user process, if necessary.
 */
void flush_dcache_page(struct page *page)
{
	if (cpu_has_feature(CPU_FTR_COHERENT_ICACHE))
		return;
	/* avoid an atomic op if possible */
	if (test_bit(PG_arch_1, &page->flags))
		clear_bit(PG_arch_1, &page->flags);
}
EXPORT_SYMBOL(flush_dcache_page);

void flush_dcache_icache_page(struct page *page)
{
428 429 430 431 432 433
#ifdef CONFIG_HUGETLB_PAGE
	if (PageCompound(page)) {
		flush_dcache_icache_hugepage(page);
		return;
	}
#endif
434
#ifdef CONFIG_BOOKE
435 436 437 438 439
	{
		void *start = kmap_atomic(page, KM_PPC_SYNC_ICACHE);
		__flush_dcache_icache(start);
		kunmap_atomic(start, KM_PPC_SYNC_ICACHE);
	}
440
#elif defined(CONFIG_8xx) || defined(CONFIG_PPC64)
441 442 443 444 445 446
	/* On 8xx there is no need to kmap since highmem is not supported */
	__flush_dcache_icache(page_address(page)); 
#else
	__flush_dcache_icache_phys(page_to_pfn(page) << PAGE_SHIFT);
#endif
}
447

448 449 450 451 452 453 454 455 456
void clear_user_page(void *page, unsigned long vaddr, struct page *pg)
{
	clear_page(page);

	/*
	 * We shouldnt have to do this, but some versions of glibc
	 * require it (ld.so assumes zero filled pages are icache clean)
	 * - Anton
	 */
457
	flush_dcache_page(pg);
458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
}
EXPORT_SYMBOL(clear_user_page);

void copy_user_page(void *vto, void *vfrom, unsigned long vaddr,
		    struct page *pg)
{
	copy_page(vto, vfrom);

	/*
	 * We should be able to use the following optimisation, however
	 * there are two problems.
	 * Firstly a bug in some versions of binutils meant PLT sections
	 * were not marked executable.
	 * Secondly the first word in the GOT section is blrl, used
	 * to establish the GOT address. Until recently the GOT was
	 * not marked executable.
	 * - Anton
	 */
#if 0
	if (!vma->vm_file && ((vma->vm_flags & VM_EXEC) == 0))
		return;
#endif

481
	flush_dcache_page(pg);
482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
}

void flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
			     unsigned long addr, int len)
{
	unsigned long maddr;

	maddr = (unsigned long) kmap(page) + (addr & ~PAGE_MASK);
	flush_icache_range(maddr, maddr + len);
	kunmap(page);
}
EXPORT_SYMBOL(flush_icache_user_range);

/*
 * This is called at the end of handling a user page fault, when the
 * fault has been handled by updating a PTE in the linux page tables.
 * We use it to preload an HPTE into the hash table corresponding to
 * the updated linux PTE.
 * 
H
Hugh Dickins 已提交
501
 * This must always be called with the pte lock held.
502 503
 */
void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
504
		      pte_t *ptep)
505
{
506 507
#ifdef CONFIG_PPC_STD_MMU
	unsigned long access = 0, trap;
508 509

	/* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */
510
	if (!pte_young(*ptep) || address >= TASK_SIZE)
511 512
		return;

513 514 515 516 517 518 519 520
	/* We try to figure out if we are coming from an instruction
	 * access fault and pass that down to __hash_page so we avoid
	 * double-faulting on execution of fresh text. We have to test
	 * for regs NULL since init will get here first thing at boot
	 *
	 * We also avoid filling the hash if not coming from a fault
	 */
	if (current->thread.regs == NULL)
521
		return;
522 523 524 525 526 527 528
	trap = TRAP(current->thread.regs);
	if (trap == 0x400)
		access |= _PAGE_EXEC;
	else if (trap != 0x300)
		return;
	hash_preload(vma->vm_mm, address, access, trap);
#endif /* CONFIG_PPC_STD_MMU */
529
}