task_mmu.c 21.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4
#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/mount.h>
#include <linux/seq_file.h>
M
Mauricio Lin 已提交
5
#include <linux/highmem.h>
K
Kees Cook 已提交
6
#include <linux/ptrace.h>
7
#include <linux/slab.h>
8 9
#include <linux/pagemap.h>
#include <linux/mempolicy.h>
10 11
#include <linux/swap.h>
#include <linux/swapops.h>
M
Mauricio Lin 已提交
12

L
Linus Torvalds 已提交
13 14
#include <asm/elf.h>
#include <asm/uaccess.h>
M
Mauricio Lin 已提交
15
#include <asm/tlbflush.h>
L
Linus Torvalds 已提交
16 17
#include "internal.h"

18
void task_mem(struct seq_file *m, struct mm_struct *mm)
L
Linus Torvalds 已提交
19
{
K
KAMEZAWA Hiroyuki 已提交
20
	unsigned long data, text, lib, swap;
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
	unsigned long hiwater_vm, total_vm, hiwater_rss, total_rss;

	/*
	 * Note: to minimize their overhead, mm maintains hiwater_vm and
	 * hiwater_rss only when about to *lower* total_vm or rss.  Any
	 * collector of these hiwater stats must therefore get total_vm
	 * and rss too, which will usually be the higher.  Barriers? not
	 * worth the effort, such snapshots can always be inconsistent.
	 */
	hiwater_vm = total_vm = mm->total_vm;
	if (hiwater_vm < mm->hiwater_vm)
		hiwater_vm = mm->hiwater_vm;
	hiwater_rss = total_rss = get_mm_rss(mm);
	if (hiwater_rss < mm->hiwater_rss)
		hiwater_rss = mm->hiwater_rss;
L
Linus Torvalds 已提交
36 37 38 39

	data = mm->total_vm - mm->shared_vm - mm->stack_vm;
	text = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK)) >> 10;
	lib = (mm->exec_vm << (PAGE_SHIFT-10)) - text;
K
KAMEZAWA Hiroyuki 已提交
40
	swap = get_mm_counter(mm, MM_SWAPENTS);
41
	seq_printf(m,
42
		"VmPeak:\t%8lu kB\n"
L
Linus Torvalds 已提交
43 44
		"VmSize:\t%8lu kB\n"
		"VmLck:\t%8lu kB\n"
45
		"VmHWM:\t%8lu kB\n"
L
Linus Torvalds 已提交
46 47 48 49 50
		"VmRSS:\t%8lu kB\n"
		"VmData:\t%8lu kB\n"
		"VmStk:\t%8lu kB\n"
		"VmExe:\t%8lu kB\n"
		"VmLib:\t%8lu kB\n"
K
KAMEZAWA Hiroyuki 已提交
51 52
		"VmPTE:\t%8lu kB\n"
		"VmSwap:\t%8lu kB\n",
53 54
		hiwater_vm << (PAGE_SHIFT-10),
		(total_vm - mm->reserved_vm) << (PAGE_SHIFT-10),
L
Linus Torvalds 已提交
55
		mm->locked_vm << (PAGE_SHIFT-10),
56 57
		hiwater_rss << (PAGE_SHIFT-10),
		total_rss << (PAGE_SHIFT-10),
L
Linus Torvalds 已提交
58 59
		data << (PAGE_SHIFT-10),
		mm->stack_vm << (PAGE_SHIFT-10), text, lib,
K
KAMEZAWA Hiroyuki 已提交
60 61
		(PTRS_PER_PTE*sizeof(pte_t)*mm->nr_ptes) >> 10,
		swap << (PAGE_SHIFT-10));
L
Linus Torvalds 已提交
62 63 64 65 66 67 68
}

unsigned long task_vsize(struct mm_struct *mm)
{
	return PAGE_SIZE * mm->total_vm;
}

69 70 71
unsigned long task_statm(struct mm_struct *mm,
			 unsigned long *shared, unsigned long *text,
			 unsigned long *data, unsigned long *resident)
L
Linus Torvalds 已提交
72
{
K
KAMEZAWA Hiroyuki 已提交
73
	*shared = get_mm_counter(mm, MM_FILEPAGES);
L
Linus Torvalds 已提交
74 75 76
	*text = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK))
								>> PAGE_SHIFT;
	*data = mm->total_vm - mm->shared_vm;
K
KAMEZAWA Hiroyuki 已提交
77
	*resident = *shared + get_mm_counter(mm, MM_ANONPAGES);
L
Linus Torvalds 已提交
78 79 80 81 82 83 84 85 86 87 88
	return mm->total_vm;
}

static void pad_len_spaces(struct seq_file *m, int len)
{
	len = 25 + sizeof(void*) * 6 - len;
	if (len < 1)
		len = 1;
	seq_printf(m, "%*c", len, ' ');
}

89 90 91 92 93 94 95 96
static void vma_stop(struct proc_maps_private *priv, struct vm_area_struct *vma)
{
	if (vma && vma != priv->tail_vma) {
		struct mm_struct *mm = vma->vm_mm;
		up_read(&mm->mmap_sem);
		mmput(mm);
	}
}
97

98
static void *m_start(struct seq_file *m, loff_t *pos)
M
Mauricio Lin 已提交
99
{
100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126
	struct proc_maps_private *priv = m->private;
	unsigned long last_addr = m->version;
	struct mm_struct *mm;
	struct vm_area_struct *vma, *tail_vma = NULL;
	loff_t l = *pos;

	/* Clear the per syscall fields in priv */
	priv->task = NULL;
	priv->tail_vma = NULL;

	/*
	 * We remember last_addr rather than next_addr to hit with
	 * mmap_cache most of the time. We have zero last_addr at
	 * the beginning and also after lseek. We will have -1 last_addr
	 * after the end of the vmas.
	 */

	if (last_addr == -1UL)
		return NULL;

	priv->task = get_pid_task(priv->pid, PIDTYPE_PID);
	if (!priv->task)
		return NULL;

	mm = mm_for_maps(priv->task);
	if (!mm)
		return NULL;
127
	down_read(&mm->mmap_sem);
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188

	tail_vma = get_gate_vma(priv->task);
	priv->tail_vma = tail_vma;

	/* Start with last addr hint */
	vma = find_vma(mm, last_addr);
	if (last_addr && vma) {
		vma = vma->vm_next;
		goto out;
	}

	/*
	 * Check the vma index is within the range and do
	 * sequential scan until m_index.
	 */
	vma = NULL;
	if ((unsigned long)l < mm->map_count) {
		vma = mm->mmap;
		while (l-- && vma)
			vma = vma->vm_next;
		goto out;
	}

	if (l != mm->map_count)
		tail_vma = NULL; /* After gate vma */

out:
	if (vma)
		return vma;

	/* End of vmas has been reached */
	m->version = (tail_vma != NULL)? 0: -1UL;
	up_read(&mm->mmap_sem);
	mmput(mm);
	return tail_vma;
}

static void *m_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct proc_maps_private *priv = m->private;
	struct vm_area_struct *vma = v;
	struct vm_area_struct *tail_vma = priv->tail_vma;

	(*pos)++;
	if (vma && (vma != tail_vma) && vma->vm_next)
		return vma->vm_next;
	vma_stop(priv, vma);
	return (vma != tail_vma)? tail_vma: NULL;
}

static void m_stop(struct seq_file *m, void *v)
{
	struct proc_maps_private *priv = m->private;
	struct vm_area_struct *vma = v;

	vma_stop(priv, vma);
	if (priv->task)
		put_task_struct(priv->task);
}

static int do_maps_open(struct inode *inode, struct file *file,
189
			const struct seq_operations *ops)
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205
{
	struct proc_maps_private *priv;
	int ret = -ENOMEM;
	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
	if (priv) {
		priv->pid = proc_pid(inode);
		ret = seq_open(file, ops);
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = priv;
		} else {
			kfree(priv);
		}
	}
	return ret;
}
M
Mauricio Lin 已提交
206

207
static void show_map_vma(struct seq_file *m, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
208
{
M
Mauricio Lin 已提交
209 210 211
	struct mm_struct *mm = vma->vm_mm;
	struct file *file = vma->vm_file;
	int flags = vma->vm_flags;
L
Linus Torvalds 已提交
212
	unsigned long ino = 0;
213
	unsigned long long pgoff = 0;
214
	unsigned long start;
L
Linus Torvalds 已提交
215 216 217 218
	dev_t dev = 0;
	int len;

	if (file) {
219
		struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
220 221
		dev = inode->i_sb->s_dev;
		ino = inode->i_ino;
222
		pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT;
L
Linus Torvalds 已提交
223 224
	}

225 226 227
	/* We don't show the stack guard page in /proc/maps */
	start = vma->vm_start;
	if (vma->vm_flags & VM_GROWSDOWN)
228 229
		if (!vma_stack_continue(vma->vm_prev, vma->vm_start))
			start += PAGE_SIZE;
230

231
	seq_printf(m, "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu %n",
232
			start,
M
Mauricio Lin 已提交
233
			vma->vm_end,
L
Linus Torvalds 已提交
234 235 236 237
			flags & VM_READ ? 'r' : '-',
			flags & VM_WRITE ? 'w' : '-',
			flags & VM_EXEC ? 'x' : '-',
			flags & VM_MAYSHARE ? 's' : 'p',
238
			pgoff,
L
Linus Torvalds 已提交
239 240 241 242 243 244
			MAJOR(dev), MINOR(dev), ino, &len);

	/*
	 * Print the dentry name for named mappings, and a
	 * special [heap] marker for the heap:
	 */
M
Mauricio Lin 已提交
245
	if (file) {
L
Linus Torvalds 已提交
246
		pad_len_spaces(m, len);
247
		seq_path(m, &file->f_path, "\n");
L
Linus Torvalds 已提交
248
	} else {
249 250 251 252
		const char *name = arch_vma_name(vma);
		if (!name) {
			if (mm) {
				if (vma->vm_start <= mm->start_brk &&
M
Mauricio Lin 已提交
253
						vma->vm_end >= mm->brk) {
254 255 256 257
					name = "[heap]";
				} else if (vma->vm_start <= mm->start_stack &&
					   vma->vm_end >= mm->start_stack) {
					name = "[stack]";
L
Linus Torvalds 已提交
258
				}
259 260
			} else {
				name = "[vdso]";
L
Linus Torvalds 已提交
261
			}
262 263
		}
		if (name) {
L
Linus Torvalds 已提交
264
			pad_len_spaces(m, len);
265
			seq_puts(m, name);
L
Linus Torvalds 已提交
266 267 268
		}
	}
	seq_putc(m, '\n');
269 270 271 272 273 274 275 276 277
}

static int show_map(struct seq_file *m, void *v)
{
	struct vm_area_struct *vma = v;
	struct proc_maps_private *priv = m->private;
	struct task_struct *task = priv->task;

	show_map_vma(m, vma);
M
Mauricio Lin 已提交
278 279 280

	if (m->count < m->size)  /* vma is copied successfully */
		m->version = (vma != get_gate_vma(task))? vma->vm_start: 0;
L
Linus Torvalds 已提交
281 282 283
	return 0;
}

284
static const struct seq_operations proc_pid_maps_op = {
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= show_map
};

static int maps_open(struct inode *inode, struct file *file)
{
	return do_maps_open(inode, file, &proc_pid_maps_op);
}

const struct file_operations proc_maps_operations = {
	.open		= maps_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= seq_release_private,
};

/*
 * Proportional Set Size(PSS): my share of RSS.
 *
 * PSS of a process is the count of pages it has in memory, where each
 * page is divided by the number of processes sharing it.  So if a
 * process has 1000 pages all to itself, and 1000 shared with one other
 * process, its PSS will be 1500.
 *
 * To keep (accumulated) division errors low, we adopt a 64bit
 * fixed-point pss counter to minimize division errors. So (pss >>
 * PSS_SHIFT) would be the real byte count.
 *
 * A shift of 12 before division means (assuming 4K page size):
 * 	- 1M 3-user-pages add up to 8KB errors;
 * 	- supports mapcount up to 2^24, or 16M;
 * 	- supports PSS up to 2^52 bytes, or 4PB.
 */
#define PSS_SHIFT 12

322
#ifdef CONFIG_PROC_PAGE_MONITOR
P
Peter Zijlstra 已提交
323
struct mem_size_stats {
324 325 326 327 328 329 330
	struct vm_area_struct *vma;
	unsigned long resident;
	unsigned long shared_clean;
	unsigned long shared_dirty;
	unsigned long private_clean;
	unsigned long private_dirty;
	unsigned long referenced;
331
	unsigned long anonymous;
P
Peter Zijlstra 已提交
332
	unsigned long swap;
333 334 335
	u64 pss;
};

336
static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
D
Dave Hansen 已提交
337
			   struct mm_walk *walk)
M
Mauricio Lin 已提交
338
{
D
Dave Hansen 已提交
339
	struct mem_size_stats *mss = walk->private;
340
	struct vm_area_struct *vma = mss->vma;
M
Mauricio Lin 已提交
341
	pte_t *pte, ptent;
342
	spinlock_t *ptl;
M
Mauricio Lin 已提交
343
	struct page *page;
344
	int mapcount;
M
Mauricio Lin 已提交
345

346 347
	split_huge_page_pmd(walk->mm, pmd);

348
	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
349
	for (; addr != end; pte++, addr += PAGE_SIZE) {
M
Mauricio Lin 已提交
350
		ptent = *pte;
P
Peter Zijlstra 已提交
351 352 353 354 355 356

		if (is_swap_pte(ptent)) {
			mss->swap += PAGE_SIZE;
			continue;
		}

357
		if (!pte_present(ptent))
M
Mauricio Lin 已提交
358 359
			continue;

N
Nick Piggin 已提交
360 361
		page = vm_normal_page(vma, addr, ptent);
		if (!page)
M
Mauricio Lin 已提交
362 363
			continue;

364 365 366
		if (PageAnon(page))
			mss->anonymous += PAGE_SIZE;

M
Minchan Kim 已提交
367
		mss->resident += PAGE_SIZE;
368 369 370
		/* Accumulate the size in pages that have been accessed. */
		if (pte_young(ptent) || PageReferenced(page))
			mss->referenced += PAGE_SIZE;
371 372
		mapcount = page_mapcount(page);
		if (mapcount >= 2) {
373
			if (pte_dirty(ptent) || PageDirty(page))
M
Mauricio Lin 已提交
374 375 376
				mss->shared_dirty += PAGE_SIZE;
			else
				mss->shared_clean += PAGE_SIZE;
377
			mss->pss += (PAGE_SIZE << PSS_SHIFT) / mapcount;
M
Mauricio Lin 已提交
378
		} else {
379
			if (pte_dirty(ptent) || PageDirty(page))
M
Mauricio Lin 已提交
380 381 382
				mss->private_dirty += PAGE_SIZE;
			else
				mss->private_clean += PAGE_SIZE;
383
			mss->pss += (PAGE_SIZE << PSS_SHIFT);
M
Mauricio Lin 已提交
384
		}
385
	}
386 387
	pte_unmap_unlock(pte - 1, ptl);
	cond_resched();
388
	return 0;
M
Mauricio Lin 已提交
389 390 391 392
}

static int show_smap(struct seq_file *m, void *v)
{
393 394
	struct proc_maps_private *priv = m->private;
	struct task_struct *task = priv->task;
M
Mauricio Lin 已提交
395 396
	struct vm_area_struct *vma = v;
	struct mem_size_stats mss;
D
Dave Hansen 已提交
397 398 399 400 401
	struct mm_walk smaps_walk = {
		.pmd_entry = smaps_pte_range,
		.mm = vma->vm_mm,
		.private = &mss,
	};
M
Mauricio Lin 已提交
402 403

	memset(&mss, 0, sizeof mss);
404
	mss.vma = vma;
405
	/* mmap_sem is held in m_start */
N
Nick Piggin 已提交
406
	if (vma->vm_mm && !is_vm_hugetlb_page(vma))
D
Dave Hansen 已提交
407
		walk_page_range(vma->vm_start, vma->vm_end, &smaps_walk);
408

409
	show_map_vma(m, vma);
410 411 412 413 414 415 416 417 418

	seq_printf(m,
		   "Size:           %8lu kB\n"
		   "Rss:            %8lu kB\n"
		   "Pss:            %8lu kB\n"
		   "Shared_Clean:   %8lu kB\n"
		   "Shared_Dirty:   %8lu kB\n"
		   "Private_Clean:  %8lu kB\n"
		   "Private_Dirty:  %8lu kB\n"
P
Peter Zijlstra 已提交
419
		   "Referenced:     %8lu kB\n"
420
		   "Anonymous:      %8lu kB\n"
421
		   "Swap:           %8lu kB\n"
422
		   "KernelPageSize: %8lu kB\n"
423 424
		   "MMUPageSize:    %8lu kB\n"
		   "Locked:         %8lu kB\n",
425 426 427 428 429 430 431
		   (vma->vm_end - vma->vm_start) >> 10,
		   mss.resident >> 10,
		   (unsigned long)(mss.pss >> (10 + PSS_SHIFT)),
		   mss.shared_clean  >> 10,
		   mss.shared_dirty  >> 10,
		   mss.private_clean >> 10,
		   mss.private_dirty >> 10,
P
Peter Zijlstra 已提交
432
		   mss.referenced >> 10,
433
		   mss.anonymous >> 10,
434
		   mss.swap >> 10,
435
		   vma_kernel_pagesize(vma) >> 10,
436 437 438
		   vma_mmu_pagesize(vma) >> 10,
		   (vma->vm_flags & VM_LOCKED) ?
			(unsigned long)(mss.pss >> (10 + PSS_SHIFT)) : 0);
439

440 441 442
	if (m->count < m->size)  /* vma is copied successfully */
		m->version = (vma != get_gate_vma(task)) ? vma->vm_start : 0;
	return 0;
M
Mauricio Lin 已提交
443 444
}

445
static const struct seq_operations proc_pid_smaps_op = {
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= show_smap
};

static int smaps_open(struct inode *inode, struct file *file)
{
	return do_maps_open(inode, file, &proc_pid_smaps_op);
}

const struct file_operations proc_smaps_operations = {
	.open		= smaps_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= seq_release_private,
};

static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
D
Dave Hansen 已提交
465
				unsigned long end, struct mm_walk *walk)
466
{
D
Dave Hansen 已提交
467
	struct vm_area_struct *vma = walk->private;
468 469 470 471
	pte_t *pte, ptent;
	spinlock_t *ptl;
	struct page *page;

472 473
	split_huge_page_pmd(walk->mm, pmd);

474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
	for (; addr != end; pte++, addr += PAGE_SIZE) {
		ptent = *pte;
		if (!pte_present(ptent))
			continue;

		page = vm_normal_page(vma, addr, ptent);
		if (!page)
			continue;

		/* Clear accessed and referenced bits. */
		ptep_test_and_clear_young(vma, addr, pte);
		ClearPageReferenced(page);
	}
	pte_unmap_unlock(pte - 1, ptl);
	cond_resched();
	return 0;
}

493 494 495 496
#define CLEAR_REFS_ALL 1
#define CLEAR_REFS_ANON 2
#define CLEAR_REFS_MAPPED 3

497 498
static ssize_t clear_refs_write(struct file *file, const char __user *buf,
				size_t count, loff_t *ppos)
499
{
500
	struct task_struct *task;
501
	char buffer[PROC_NUMBUF];
502
	struct mm_struct *mm;
503
	struct vm_area_struct *vma;
504
	long type;
505

506 507 508 509 510
	memset(buffer, 0, sizeof(buffer));
	if (count > sizeof(buffer) - 1)
		count = sizeof(buffer) - 1;
	if (copy_from_user(buffer, buf, count))
		return -EFAULT;
511 512
	if (strict_strtol(strstrip(buffer), 10, &type))
		return -EINVAL;
513
	if (type < CLEAR_REFS_ALL || type > CLEAR_REFS_MAPPED)
514 515 516 517 518 519
		return -EINVAL;
	task = get_proc_task(file->f_path.dentry->d_inode);
	if (!task)
		return -ESRCH;
	mm = get_task_mm(task);
	if (mm) {
520 521 522 523
		struct mm_walk clear_refs_walk = {
			.pmd_entry = clear_refs_pte_range,
			.mm = mm,
		};
524
		down_read(&mm->mmap_sem);
D
Dave Hansen 已提交
525 526
		for (vma = mm->mmap; vma; vma = vma->vm_next) {
			clear_refs_walk.private = vma;
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
			if (is_vm_hugetlb_page(vma))
				continue;
			/*
			 * Writing 1 to /proc/pid/clear_refs affects all pages.
			 *
			 * Writing 2 to /proc/pid/clear_refs only affects
			 * Anonymous pages.
			 *
			 * Writing 3 to /proc/pid/clear_refs only affects file
			 * mapped pages.
			 */
			if (type == CLEAR_REFS_ANON && vma->vm_file)
				continue;
			if (type == CLEAR_REFS_MAPPED && !vma->vm_file)
				continue;
			walk_page_range(vma->vm_start, vma->vm_end,
					&clear_refs_walk);
D
Dave Hansen 已提交
544
		}
545 546 547 548 549
		flush_tlb_mm(mm);
		up_read(&mm->mmap_sem);
		mmput(mm);
	}
	put_task_struct(task);
550 551

	return count;
552 553
}

554 555
const struct file_operations proc_clear_refs_operations = {
	.write		= clear_refs_write,
556
	.llseek		= noop_llseek,
557 558
};

559
struct pagemapread {
560 561
	int pos, len;
	u64 *buffer;
562 563
};

564 565 566 567 568 569 570 571 572 573 574 575 576 577 578
#define PM_ENTRY_BYTES      sizeof(u64)
#define PM_STATUS_BITS      3
#define PM_STATUS_OFFSET    (64 - PM_STATUS_BITS)
#define PM_STATUS_MASK      (((1LL << PM_STATUS_BITS) - 1) << PM_STATUS_OFFSET)
#define PM_STATUS(nr)       (((nr) << PM_STATUS_OFFSET) & PM_STATUS_MASK)
#define PM_PSHIFT_BITS      6
#define PM_PSHIFT_OFFSET    (PM_STATUS_OFFSET - PM_PSHIFT_BITS)
#define PM_PSHIFT_MASK      (((1LL << PM_PSHIFT_BITS) - 1) << PM_PSHIFT_OFFSET)
#define PM_PSHIFT(x)        (((u64) (x) << PM_PSHIFT_OFFSET) & PM_PSHIFT_MASK)
#define PM_PFRAME_MASK      ((1LL << PM_PSHIFT_OFFSET) - 1)
#define PM_PFRAME(x)        ((x) & PM_PFRAME_MASK)

#define PM_PRESENT          PM_STATUS(4LL)
#define PM_SWAP             PM_STATUS(2LL)
#define PM_NOT_PRESENT      PM_PSHIFT(PAGE_SHIFT)
579 580 581 582 583
#define PM_END_OF_BUFFER    1

static int add_to_pagemap(unsigned long addr, u64 pfn,
			  struct pagemapread *pm)
{
584 585
	pm->buffer[pm->pos++] = pfn;
	if (pm->pos >= pm->len)
586
		return PM_END_OF_BUFFER;
587 588 589 590
	return 0;
}

static int pagemap_pte_hole(unsigned long start, unsigned long end,
D
Dave Hansen 已提交
591
				struct mm_walk *walk)
592
{
D
Dave Hansen 已提交
593
	struct pagemapread *pm = walk->private;
594 595 596 597 598 599 600 601 602 603
	unsigned long addr;
	int err = 0;
	for (addr = start; addr < end; addr += PAGE_SIZE) {
		err = add_to_pagemap(addr, PM_NOT_PRESENT, pm);
		if (err)
			break;
	}
	return err;
}

604
static u64 swap_pte_to_pagemap_entry(pte_t pte)
605 606
{
	swp_entry_t e = pte_to_swp_entry(pte);
607
	return swp_type(e) | (swp_offset(e) << MAX_SWAPFILES_SHIFT);
608 609
}

610
static u64 pte_to_pagemap_entry(pte_t pte)
611
{
612
	u64 pme = 0;
613 614 615 616 617 618 619 620 621
	if (is_swap_pte(pte))
		pme = PM_PFRAME(swap_pte_to_pagemap_entry(pte))
			| PM_PSHIFT(PAGE_SHIFT) | PM_SWAP;
	else if (pte_present(pte))
		pme = PM_PFRAME(pte_pfn(pte))
			| PM_PSHIFT(PAGE_SHIFT) | PM_PRESENT;
	return pme;
}

622
static int pagemap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
D
Dave Hansen 已提交
623
			     struct mm_walk *walk)
624
{
625
	struct vm_area_struct *vma;
D
Dave Hansen 已提交
626
	struct pagemapread *pm = walk->private;
627 628 629
	pte_t *pte;
	int err = 0;

630 631
	split_huge_page_pmd(walk->mm, pmd);

632 633
	/* find the first VMA at or above 'addr' */
	vma = find_vma(walk->mm, addr);
634 635
	for (; addr != end; addr += PAGE_SIZE) {
		u64 pfn = PM_NOT_PRESENT;
636 637 638 639 640 641 642 643 644 645 646 647 648 649 650

		/* check to see if we've left 'vma' behind
		 * and need a new, higher one */
		if (vma && (addr >= vma->vm_end))
			vma = find_vma(walk->mm, addr);

		/* check that 'vma' actually covers this address,
		 * and that it isn't a huge page vma */
		if (vma && (vma->vm_start <= addr) &&
		    !is_vm_hugetlb_page(vma)) {
			pte = pte_offset_map(pmd, addr);
			pfn = pte_to_pagemap_entry(*pte);
			/* unmap before userspace copy */
			pte_unmap(pte);
		}
651 652 653 654 655 656 657 658 659 660
		err = add_to_pagemap(addr, pfn, pm);
		if (err)
			return err;
	}

	cond_resched();

	return err;
}

661
#ifdef CONFIG_HUGETLB_PAGE
662 663 664 665 666 667 668 669 670
static u64 huge_pte_to_pagemap_entry(pte_t pte, int offset)
{
	u64 pme = 0;
	if (pte_present(pte))
		pme = PM_PFRAME(pte_pfn(pte) + offset)
			| PM_PSHIFT(PAGE_SHIFT) | PM_PRESENT;
	return pme;
}

671 672 673 674
/* This function walks within one hugetlb entry in the single call */
static int pagemap_hugetlb_range(pte_t *pte, unsigned long hmask,
				 unsigned long addr, unsigned long end,
				 struct mm_walk *walk)
675 676 677
{
	struct pagemapread *pm = walk->private;
	int err = 0;
678
	u64 pfn;
679 680

	for (; addr != end; addr += PAGE_SIZE) {
681 682
		int offset = (addr & ~hmask) >> PAGE_SHIFT;
		pfn = huge_pte_to_pagemap_entry(*pte, offset);
683 684 685 686 687 688 689 690 691
		err = add_to_pagemap(addr, pfn, pm);
		if (err)
			return err;
	}

	cond_resched();

	return err;
}
692
#endif /* HUGETLB_PAGE */
693

694 695 696
/*
 * /proc/pid/pagemap - an array mapping virtual pages to pfns
 *
697 698 699 700 701 702 703 704 705 706 707 708 709 710
 * For each page in the address space, this file contains one 64-bit entry
 * consisting of the following:
 *
 * Bits 0-55  page frame number (PFN) if present
 * Bits 0-4   swap type if swapped
 * Bits 5-55  swap offset if swapped
 * Bits 55-60 page shift (page size = 1<<page shift)
 * Bit  61    reserved for future use
 * Bit  62    page swapped
 * Bit  63    page present
 *
 * If the page is not present but in swap, then the PFN contains an
 * encoding of the swap file number and the page's offset into the
 * swap. Unmapped pages return a null PFN. This allows determining
711 712 713 714 715 716 717
 * precisely which pages are mapped (or in swap) and comparing mapped
 * pages between processes.
 *
 * Efficient users of this interface will use /proc/pid/maps to
 * determine which areas of memory are actually mapped and llseek to
 * skip over unmapped regions.
 */
718
#define PAGEMAP_WALK_SIZE	(PMD_SIZE)
719
#define PAGEMAP_WALK_MASK	(PMD_MASK)
720 721 722 723 724 725 726
static ssize_t pagemap_read(struct file *file, char __user *buf,
			    size_t count, loff_t *ppos)
{
	struct task_struct *task = get_proc_task(file->f_path.dentry->d_inode);
	struct mm_struct *mm;
	struct pagemapread pm;
	int ret = -ESRCH;
727
	struct mm_walk pagemap_walk = {};
728 729 730 731
	unsigned long src;
	unsigned long svpfn;
	unsigned long start_vaddr;
	unsigned long end_vaddr;
732
	int copied = 0;
733 734 735 736 737

	if (!task)
		goto out;

	ret = -EACCES;
738
	if (!ptrace_may_access(task, PTRACE_MODE_READ))
739
		goto out_task;
740 741 742

	ret = -EINVAL;
	/* file position must be aligned */
743
	if ((*ppos % PM_ENTRY_BYTES) || (count % PM_ENTRY_BYTES))
744
		goto out_task;
745 746

	ret = 0;
747 748 749 750

	if (!count)
		goto out_task;

751 752
	mm = get_task_mm(task);
	if (!mm)
753
		goto out_task;
754

755 756
	pm.len = PM_ENTRY_BYTES * (PAGEMAP_WALK_SIZE >> PAGE_SHIFT);
	pm.buffer = kmalloc(pm.len, GFP_TEMPORARY);
757
	ret = -ENOMEM;
758
	if (!pm.buffer)
759
		goto out_mm;
760

761 762
	pagemap_walk.pmd_entry = pagemap_pte_range;
	pagemap_walk.pte_hole = pagemap_pte_hole;
763
#ifdef CONFIG_HUGETLB_PAGE
764
	pagemap_walk.hugetlb_entry = pagemap_hugetlb_range;
765
#endif
766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783
	pagemap_walk.mm = mm;
	pagemap_walk.private = &pm;

	src = *ppos;
	svpfn = src / PM_ENTRY_BYTES;
	start_vaddr = svpfn << PAGE_SHIFT;
	end_vaddr = TASK_SIZE_OF(task);

	/* watch out for wraparound */
	if (svpfn > TASK_SIZE_OF(task) >> PAGE_SHIFT)
		start_vaddr = end_vaddr;

	/*
	 * The odds are that this will stop walking way
	 * before end_vaddr, because the length of the
	 * user buffer is tracked in "pm", and the walk
	 * will stop when we hit the end of the buffer.
	 */
784 785 786 787 788 789
	ret = 0;
	while (count && (start_vaddr < end_vaddr)) {
		int len;
		unsigned long end;

		pm.pos = 0;
790
		end = (start_vaddr + PAGEMAP_WALK_SIZE) & PAGEMAP_WALK_MASK;
791 792 793 794 795 796 797 798 799
		/* overflow ? */
		if (end < start_vaddr || end > end_vaddr)
			end = end_vaddr;
		down_read(&mm->mmap_sem);
		ret = walk_page_range(start_vaddr, end, &pagemap_walk);
		up_read(&mm->mmap_sem);
		start_vaddr = end;

		len = min(count, PM_ENTRY_BYTES * pm.pos);
800
		if (copy_to_user(buf, pm.buffer, len)) {
801 802 803 804 805 806
			ret = -EFAULT;
			goto out_free;
		}
		copied += len;
		buf += len;
		count -= len;
807
	}
808 809 810 811
	*ppos += copied;
	if (!ret || ret == PM_END_OF_BUFFER)
		ret = copied;

812
out_free:
813
	kfree(pm.buffer);
814 815
out_mm:
	mmput(mm);
816 817 818 819 820 821 822 823 824 825
out_task:
	put_task_struct(task);
out:
	return ret;
}

const struct file_operations proc_pagemap_operations = {
	.llseek		= mem_lseek, /* borrow this */
	.read		= pagemap_read,
};
826
#endif /* CONFIG_PROC_PAGE_MONITOR */
827

828
#ifdef CONFIG_NUMA
829
extern int show_numa_map(struct seq_file *m, void *v);
830

831
static const struct seq_operations proc_pid_numa_maps_op = {
832 833 834
        .start  = m_start,
        .next   = m_next,
        .stop   = m_stop,
A
Alexey Dobriyan 已提交
835
        .show   = show_numa_map,
836
};
837 838 839 840 841 842

static int numa_maps_open(struct inode *inode, struct file *file)
{
	return do_maps_open(inode, file, &proc_pid_numa_maps_op);
}

843
const struct file_operations proc_numa_maps_operations = {
844 845 846
	.open		= numa_maps_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
847
	.release	= seq_release_private,
848
};
849
#endif