dump_pagetables.c 15.4 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
/*
 * Debug helper to dump the current kernel pagetables of the system
 * so that we can see what the various memory ranges are set to.
 *
 * (C) Copyright 2008 Intel Corporation
 *
 * Author: Arjan van de Ven <arjan@linux.intel.com>
 *
 * 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; version 2
 * of the License.
 */

15
#include <linux/debugfs.h>
16
#include <linux/kasan.h>
17
#include <linux/mm.h>
18
#include <linux/init.h>
19
#include <linux/sched.h>
20
21
22
23
24
25
26
27
28
29
30
31
32
33
#include <linux/seq_file.h>

#include <asm/pgtable.h>

/*
 * The dumper groups pagetable entries of the same type into one, and for
 * that it needs to keep some state when walking, and flush this state
 * when a "break" in the continuity is found.
 */
struct pg_state {
	int level;
	pgprot_t current_prot;
	unsigned long start_address;
	unsigned long current_address;
34
	const struct addr_marker *marker;
35
	unsigned long lines;
36
	bool to_dmesg;
Stephen Smalley's avatar
Stephen Smalley committed
37
38
	bool check_wx;
	unsigned long wx_pages;
39
40
};

41
42
43
struct addr_marker {
	unsigned long start_address;
	const char *name;
44
	unsigned long max_lines;
45
46
};

47
48
49
50
/* Address space markers hints */

#ifdef CONFIG_X86_64

51
52
53
54
enum address_markers_idx {
	USER_SPACE_NR = 0,
	KERNEL_SPACE_NR,
	LOW_KERNEL_NR,
55
56
57
#if defined(CONFIG_MODIFY_LDT_SYSCALL) && defined(CONFIG_X86_5LEVEL)
	LDT_NR,
#endif
58
59
	VMALLOC_START_NR,
	VMEMMAP_START_NR,
60
61
62
#ifdef CONFIG_KASAN
	KASAN_SHADOW_START_NR,
	KASAN_SHADOW_END_NR,
63
#endif
64
	CPU_ENTRY_AREA_NR,
65
66
#if defined(CONFIG_MODIFY_LDT_SYSCALL) && !defined(CONFIG_X86_5LEVEL)
	LDT_NR,
67
#endif
68
#ifdef CONFIG_X86_ESPFIX64
69
	ESPFIX_START_NR,
70
71
72
73
#endif
#ifdef CONFIG_EFI
	EFI_END_NR,
#endif
74
75
76
	HIGH_KERNEL_NR,
	MODULES_VADDR_NR,
	MODULES_END_NR,
77
78
79
80
81
82
83
84
85
86
87
88
89
	FIXADDR_START_NR,
	END_OF_SPACE_NR,
};

static struct addr_marker address_markers[] = {
	[USER_SPACE_NR]		= { 0,			"User Space" },
	[KERNEL_SPACE_NR]	= { (1UL << 63),	"Kernel Space" },
	[LOW_KERNEL_NR]		= { 0UL,		"Low Kernel Mapping" },
	[VMALLOC_START_NR]	= { 0UL,		"vmalloc() Area" },
	[VMEMMAP_START_NR]	= { 0UL,		"Vmemmap" },
#ifdef CONFIG_KASAN
	[KASAN_SHADOW_START_NR]	= { KASAN_SHADOW_START,	"KASAN shadow" },
	[KASAN_SHADOW_END_NR]	= { KASAN_SHADOW_END,	"KASAN shadow end" },
90
91
#endif
#ifdef CONFIG_MODIFY_LDT_SYSCALL
92
	[LDT_NR]		= { 0UL,		"LDT remap" },
93
#endif
94
	[CPU_ENTRY_AREA_NR]	= { CPU_ENTRY_AREA_BASE,"CPU entry Area" },
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
#ifdef CONFIG_X86_ESPFIX64
	[ESPFIX_START_NR]	= { ESPFIX_BASE_ADDR,	"ESPfix Area", 16 },
#endif
#ifdef CONFIG_EFI
	[EFI_END_NR]		= { EFI_VA_END,		"EFI Runtime Services" },
#endif
	[HIGH_KERNEL_NR]	= { __START_KERNEL_map,	"High Kernel Mapping" },
	[MODULES_VADDR_NR]	= { MODULES_VADDR,	"Modules" },
	[MODULES_END_NR]	= { MODULES_END,	"End Modules" },
	[FIXADDR_START_NR]	= { FIXADDR_START,	"Fixmap Area" },
	[END_OF_SPACE_NR]	= { -1,			NULL }
};

#else /* CONFIG_X86_64 */

enum address_markers_idx {
	USER_SPACE_NR = 0,
112
113
114
	KERNEL_SPACE_NR,
	VMALLOC_START_NR,
	VMALLOC_END_NR,
115
#ifdef CONFIG_HIGHMEM
116
117
	PKMAP_BASE_NR,
#endif
118
	CPU_ENTRY_AREA_NR,
119
120
	FIXADDR_START_NR,
	END_OF_SPACE_NR,
121
122
};

123
static struct addr_marker address_markers[] = {
124
125
126
127
128
129
	[USER_SPACE_NR]		= { 0,			"User Space" },
	[KERNEL_SPACE_NR]	= { PAGE_OFFSET,	"Kernel Mapping" },
	[VMALLOC_START_NR]	= { 0UL,		"vmalloc() Area" },
	[VMALLOC_END_NR]	= { 0UL,		"vmalloc() End" },
#ifdef CONFIG_HIGHMEM
	[PKMAP_BASE_NR]		= { 0UL,		"Persistent kmap() Area" },
130
#endif
131
	[CPU_ENTRY_AREA_NR]	= { 0UL,		"CPU entry area" },
132
133
	[FIXADDR_START_NR]	= { 0UL,		"Fixmap area" },
	[END_OF_SPACE_NR]	= { -1,			NULL }
134
};
135

136
137
#endif /* !CONFIG_X86_64 */

138
139
140
141
/* Multipliers for offsets within the PTEs */
#define PTE_LEVEL_MULT (PAGE_SIZE)
#define PMD_LEVEL_MULT (PTRS_PER_PTE * PTE_LEVEL_MULT)
#define PUD_LEVEL_MULT (PTRS_PER_PMD * PMD_LEVEL_MULT)
142
#define P4D_LEVEL_MULT (PTRS_PER_PUD * PUD_LEVEL_MULT)
143
#define PGD_LEVEL_MULT (PTRS_PER_P4D * P4D_LEVEL_MULT)
144

145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
#define pt_dump_seq_printf(m, to_dmesg, fmt, args...)		\
({								\
	if (to_dmesg)					\
		printk(KERN_INFO fmt, ##args);			\
	else							\
		if (m)						\
			seq_printf(m, fmt, ##args);		\
})

#define pt_dump_cont_printf(m, to_dmesg, fmt, args...)		\
({								\
	if (to_dmesg)					\
		printk(KERN_CONT fmt, ##args);			\
	else							\
		if (m)						\
			seq_printf(m, fmt, ##args);		\
})

163
164
165
/*
 * Print a readable form of a pgprot_t to the seq_file
 */
166
static void printk_prot(struct seq_file *m, pgprot_t prot, int level, bool dmsg)
167
{
168
169
	pgprotval_t pr = pgprot_val(prot);
	static const char * const level_name[] =
170
		{ "cr3", "pgd", "p4d", "pud", "pmd", "pte" };
171

172
	if (!(pr & _PAGE_PRESENT)) {
173
		/* Not present */
174
		pt_dump_cont_printf(m, dmsg, "                              ");
175
176
	} else {
		if (pr & _PAGE_USER)
177
			pt_dump_cont_printf(m, dmsg, "USR ");
178
		else
179
			pt_dump_cont_printf(m, dmsg, "    ");
180
		if (pr & _PAGE_RW)
181
			pt_dump_cont_printf(m, dmsg, "RW ");
182
		else
183
			pt_dump_cont_printf(m, dmsg, "ro ");
184
		if (pr & _PAGE_PWT)
185
			pt_dump_cont_printf(m, dmsg, "PWT ");
186
		else
187
			pt_dump_cont_printf(m, dmsg, "    ");
188
		if (pr & _PAGE_PCD)
189
			pt_dump_cont_printf(m, dmsg, "PCD ");
190
		else
191
			pt_dump_cont_printf(m, dmsg, "    ");
192

193
		/* Bit 7 has a different meaning on level 3 vs 4 */
194
		if (level <= 4 && pr & _PAGE_PSE)
195
196
197
			pt_dump_cont_printf(m, dmsg, "PSE ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
198
199
		if ((level == 5 && pr & _PAGE_PAT) ||
		    ((level == 4 || level == 3) && pr & _PAGE_PAT_LARGE))
200
			pt_dump_cont_printf(m, dmsg, "PAT ");
201
202
		else
			pt_dump_cont_printf(m, dmsg, "    ");
203
		if (pr & _PAGE_GLOBAL)
204
			pt_dump_cont_printf(m, dmsg, "GLB ");
205
		else
206
			pt_dump_cont_printf(m, dmsg, "    ");
207
		if (pr & _PAGE_NX)
208
			pt_dump_cont_printf(m, dmsg, "NX ");
209
		else
210
			pt_dump_cont_printf(m, dmsg, "x  ");
211
	}
212
	pt_dump_cont_printf(m, dmsg, "%s\n", level_name[level]);
213
214
215
}

/*
216
 * On 64 bits, sign-extend the 48 bit address to 64 bit
217
 */
218
static unsigned long normalize_addr(unsigned long u)
219
{
220
221
222
223
224
225
	int shift;
	if (!IS_ENABLED(CONFIG_X86_64))
		return u;

	shift = 64 - (__VIRTUAL_MASK_SHIFT + 1);
	return (signed long)(u << shift) >> shift;
226
227
228
229
230
231
232
233
}

/*
 * This function gets called on a break in a continuous series
 * of PTE entries; the next one is different so we need to
 * print what we collected so far.
 */
static void note_page(struct seq_file *m, struct pg_state *st,
234
		      pgprot_t new_prot, int level)
235
{
236
	pgprotval_t prot, cur;
237
	static const char units[] = "BKMGTPE";
238
239
240

	/*
	 * If we have a "break" in the series, we need to flush the state that
241
242
	 * we have now. "break" is either changing perms, levels or
	 * address space marker.
243
	 */
244
245
	prot = pgprot_val(new_prot);
	cur = pgprot_val(st->current_prot);
246

247
248
249
250
251
	if (!st->level) {
		/* First entry */
		st->current_prot = new_prot;
		st->level = level;
		st->marker = address_markers;
252
		st->lines = 0;
253
254
		pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
				   st->marker->name);
255
256
257
	} else if (prot != cur || level != st->level ||
		   st->current_address >= st->marker[1].start_address) {
		const char *unit = units;
258
		unsigned long delta;
259
		int width = sizeof(unsigned long) * 2;
Stephen Smalley's avatar
Stephen Smalley committed
260
261
262
263
264
265
266
267
268
269
		pgprotval_t pr = pgprot_val(st->current_prot);

		if (st->check_wx && (pr & _PAGE_RW) && !(pr & _PAGE_NX)) {
			WARN_ONCE(1,
				  "x86/mm: Found insecure W+X mapping at address %p/%pS\n",
				  (void *)st->start_address,
				  (void *)st->start_address);
			st->wx_pages += (st->current_address -
					 st->start_address) / PAGE_SIZE;
		}
270
271
272
273

		/*
		 * Now print the actual finished series
		 */
274
275
276
277
278
279
		if (!st->marker->max_lines ||
		    st->lines < st->marker->max_lines) {
			pt_dump_seq_printf(m, st->to_dmesg,
					   "0x%0*lx-0x%0*lx   ",
					   width, st->start_address,
					   width, st->current_address);
280

281
282
283
284
285
286
287
288
289
			delta = st->current_address - st->start_address;
			while (!(delta & 1023) && unit[1]) {
				delta >>= 10;
				unit++;
			}
			pt_dump_cont_printf(m, st->to_dmesg, "%9lu%c ",
					    delta, *unit);
			printk_prot(m, st->current_prot, st->level,
				    st->to_dmesg);
290
		}
291
		st->lines++;
292
293
294
295
296
297
298

		/*
		 * We print markers for special areas of address space,
		 * such as the start of vmalloc space etc.
		 * This helps in the interpretation.
		 */
		if (st->current_address >= st->marker[1].start_address) {
299
300
301
302
303
304
305
306
307
			if (st->marker->max_lines &&
			    st->lines > st->marker->max_lines) {
				unsigned long nskip =
					st->lines - st->marker->max_lines;
				pt_dump_seq_printf(m, st->to_dmesg,
						   "... %lu entr%s skipped ... \n",
						   nskip,
						   nskip == 1 ? "y" : "ies");
			}
308
			st->marker++;
309
			st->lines = 0;
310
311
			pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
					   st->marker->name);
312
		}
313

314
315
316
		st->start_address = st->current_address;
		st->current_prot = new_prot;
		st->level = level;
317
	}
318
319
}

320
static void walk_pte_level(struct seq_file *m, struct pg_state *st, pmd_t addr, unsigned long P)
321
322
323
{
	int i;
	pte_t *start;
324
	pgprotval_t prot;
325

326
	start = (pte_t *)pmd_page_vaddr(addr);
327
	for (i = 0; i < PTRS_PER_PTE; i++) {
328
		prot = pte_flags(*start);
329
		st->current_address = normalize_addr(P + i * PTE_LEVEL_MULT);
330
		note_page(m, st, __pgprot(prot), 5);
331
332
333
		start++;
	}
}
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
#ifdef CONFIG_KASAN

/*
 * This is an optimization for KASAN=y case. Since all kasan page tables
 * eventually point to the kasan_zero_page we could call note_page()
 * right away without walking through lower level page tables. This saves
 * us dozens of seconds (minutes for 5-level config) while checking for
 * W+X mapping or reading kernel_page_tables debugfs file.
 */
static inline bool kasan_page_table(struct seq_file *m, struct pg_state *st,
				void *pt)
{
	if (__pa(pt) == __pa(kasan_zero_pmd) ||
#ifdef CONFIG_X86_5LEVEL
	    __pa(pt) == __pa(kasan_zero_p4d) ||
#endif
	    __pa(pt) == __pa(kasan_zero_pud)) {
		pgprotval_t prot = pte_flags(kasan_zero_pte[0]);
		note_page(m, st, __pgprot(prot), 5);
		return true;
	}
	return false;
}
#else
static inline bool kasan_page_table(struct seq_file *m, struct pg_state *st,
				void *pt)
{
	return false;
}
#endif
364

365
#if PTRS_PER_PMD > 1
366

367
static void walk_pmd_level(struct seq_file *m, struct pg_state *st, pud_t addr, unsigned long P)
368
369
{
	int i;
370
	pmd_t *start, *pmd_start;
371
	pgprotval_t prot;
372

373
	pmd_start = start = (pmd_t *)pud_page_vaddr(addr);
374
	for (i = 0; i < PTRS_PER_PMD; i++) {
375
		st->current_address = normalize_addr(P + i * PMD_LEVEL_MULT);
376
		if (!pmd_none(*start)) {
377
378
			if (pmd_large(*start) || !pmd_present(*start)) {
				prot = pmd_flags(*start);
379
				note_page(m, st, __pgprot(prot), 4);
380
			} else if (!kasan_page_table(m, st, pmd_start)) {
381
382
				walk_pte_level(m, st, *start,
					       P + i * PMD_LEVEL_MULT);
383
			}
384
		} else
385
			note_page(m, st, __pgprot(0), 4);
386
387
388
389
		start++;
	}
}

390
391
392
393
394
#else
#define walk_pmd_level(m,s,a,p) walk_pte_level(m,s,__pmd(pud_val(a)),p)
#define pud_large(a) pmd_large(__pmd(pud_val(a)))
#define pud_none(a)  pmd_none(__pmd(pud_val(a)))
#endif
395

396
397
#if PTRS_PER_PUD > 1

398
static void walk_pud_level(struct seq_file *m, struct pg_state *st, p4d_t addr, unsigned long P)
399
400
{
	int i;
401
	pud_t *start, *pud_start;
402
	pgprotval_t prot;
403
	pud_t *prev_pud = NULL;
404

405
	pud_start = start = (pud_t *)p4d_page_vaddr(addr);
406
407

	for (i = 0; i < PTRS_PER_PUD; i++) {
408
		st->current_address = normalize_addr(P + i * PUD_LEVEL_MULT);
409
		if (!pud_none(*start)) {
410
411
			if (pud_large(*start) || !pud_present(*start)) {
				prot = pud_flags(*start);
412
				note_page(m, st, __pgprot(prot), 3);
413
			} else if (!kasan_page_table(m, st, pud_start)) {
414
415
				walk_pmd_level(m, st, *start,
					       P + i * PUD_LEVEL_MULT);
416
			}
417
		} else
418
			note_page(m, st, __pgprot(0), 3);
419

420
		prev_pud = start;
421
422
423
424
		start++;
	}
}

425
#else
426
427
428
429
430
431
432
433
#define walk_pud_level(m,s,a,p) walk_pmd_level(m,s,__pud(p4d_val(a)),p)
#define p4d_large(a) pud_large(__pud(p4d_val(a)))
#define p4d_none(a)  pud_none(__pud(p4d_val(a)))
#endif

static void walk_p4d_level(struct seq_file *m, struct pg_state *st, pgd_t addr, unsigned long P)
{
	int i;
434
	p4d_t *start, *p4d_start;
435
436
	pgprotval_t prot;

437
438
439
	if (PTRS_PER_P4D == 1)
		return walk_pud_level(m, st, __p4d(pgd_val(addr)), P);

440
	p4d_start = start = (p4d_t *)pgd_page_vaddr(addr);
441
442
443
444
445
446
447

	for (i = 0; i < PTRS_PER_P4D; i++) {
		st->current_address = normalize_addr(P + i * P4D_LEVEL_MULT);
		if (!p4d_none(*start)) {
			if (p4d_large(*start) || !p4d_present(*start)) {
				prot = p4d_flags(*start);
				note_page(m, st, __pgprot(prot), 2);
448
			} else if (!kasan_page_table(m, st, p4d_start)) {
449
450
451
452
453
454
455
456
457
458
				walk_pud_level(m, st, *start,
					       P + i * P4D_LEVEL_MULT);
			}
		} else
			note_page(m, st, __pgprot(0), 2);

		start++;
	}
}

459
460
#define pgd_large(a) (pgtable_l5_enabled ? pgd_large(a) : p4d_large(__p4d(pgd_val(a))))
#define pgd_none(a)  (pgtable_l5_enabled ? pgd_none(a) : p4d_none(__p4d(pgd_val(a))))
461

462
463
static inline bool is_hypervisor_range(int idx)
{
464
#ifdef CONFIG_X86_64
465
466
467
468
	/*
	 * ffff800000000000 - ffff87ffffffffff is reserved for
	 * the hypervisor.
	 */
469
470
	return	(idx >= pgd_index(__PAGE_OFFSET) - 16) &&
		(idx <  pgd_index(__PAGE_OFFSET));
471
#else
472
	return false;
473
#endif
474
}
475

Stephen Smalley's avatar
Stephen Smalley committed
476
static void ptdump_walk_pgd_level_core(struct seq_file *m, pgd_t *pgd,
477
				       bool checkwx, bool dmesg)
478
{
479
#ifdef CONFIG_X86_64
480
	pgd_t *start = (pgd_t *) &init_top_pgt;
481
482
483
#else
	pgd_t *start = swapper_pg_dir;
#endif
484
	pgprotval_t prot;
485
	int i;
486
	struct pg_state st = {};
487

488
489
	if (pgd) {
		start = pgd;
490
		st.to_dmesg = dmesg;
491
	}
492

Stephen Smalley's avatar
Stephen Smalley committed
493
494
495
496
	st.check_wx = checkwx;
	if (checkwx)
		st.wx_pages = 0;

497
	for (i = 0; i < PTRS_PER_PGD; i++) {
498
		st.current_address = normalize_addr(i * PGD_LEVEL_MULT);
499
		if (!pgd_none(*start) && !is_hypervisor_range(i)) {
500
501
			if (pgd_large(*start) || !pgd_present(*start)) {
				prot = pgd_flags(*start);
502
				note_page(m, &st, __pgprot(prot), 1);
503
			} else {
504
				walk_p4d_level(m, &st, *start,
505
					       i * PGD_LEVEL_MULT);
506
			}
507
		} else
508
			note_page(m, &st, __pgprot(0), 1);
509

510
		cond_resched();
511
512
		start++;
	}
513
514
515
516

	/* Flush out the last page */
	st.current_address = normalize_addr(PTRS_PER_PGD*PGD_LEVEL_MULT);
	note_page(m, &st, __pgprot(0), 0);
Stephen Smalley's avatar
Stephen Smalley committed
517
518
519
520
521
522
523
524
525
526
527
	if (!checkwx)
		return;
	if (st.wx_pages)
		pr_info("x86/mm: Checked W+X mappings: FAILED, %lu W+X pages found.\n",
			st.wx_pages);
	else
		pr_info("x86/mm: Checked W+X mappings: passed, no W+X pages found.\n");
}

void ptdump_walk_pgd_level(struct seq_file *m, pgd_t *pgd)
{
528
529
530
	ptdump_walk_pgd_level_core(m, pgd, false, true);
}

531
void ptdump_walk_pgd_level_debugfs(struct seq_file *m, pgd_t *pgd, bool user)
532
{
533
534
535
536
#ifdef CONFIG_PAGE_TABLE_ISOLATION
	if (user && static_cpu_has(X86_FEATURE_PTI))
		pgd = kernel_to_user_pgdp(pgd);
#endif
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
	ptdump_walk_pgd_level_core(m, pgd, false, false);
}
EXPORT_SYMBOL_GPL(ptdump_walk_pgd_level_debugfs);

static void ptdump_walk_user_pgd_level_checkwx(void)
{
#ifdef CONFIG_PAGE_TABLE_ISOLATION
	pgd_t *pgd = (pgd_t *) &init_top_pgt;

	if (!static_cpu_has(X86_FEATURE_PTI))
		return;

	pr_info("x86/mm: Checking user space page tables\n");
	pgd = kernel_to_user_pgdp(pgd);
	ptdump_walk_pgd_level_core(NULL, pgd, true, false);
#endif
553
554
}

Stephen Smalley's avatar
Stephen Smalley committed
555
556
void ptdump_walk_pgd_level_checkwx(void)
{
557
558
	ptdump_walk_pgd_level_core(NULL, NULL, true, false);
	ptdump_walk_user_pgd_level_checkwx();
Stephen Smalley's avatar
Stephen Smalley committed
559
560
}

561
static int __init pt_dump_init(void)
562
{
563
564
565
566
567
568
569
570
	/*
	 * Various markers are not compile-time constants, so assign them
	 * here.
	 */
#ifdef CONFIG_X86_64
	address_markers[LOW_KERNEL_NR].start_address = PAGE_OFFSET;
	address_markers[VMALLOC_START_NR].start_address = VMALLOC_START;
	address_markers[VMEMMAP_START_NR].start_address = VMEMMAP_START;
571
572
573
#ifdef CONFIG_MODIFY_LDT_SYSCALL
	address_markers[LDT_NR].start_address = LDT_BASE_ADDR;
#endif
574
#endif
575
#ifdef CONFIG_X86_32
576
577
	address_markers[VMALLOC_START_NR].start_address = VMALLOC_START;
	address_markers[VMALLOC_END_NR].start_address = VMALLOC_END;
578
# ifdef CONFIG_HIGHMEM
579
	address_markers[PKMAP_BASE_NR].start_address = PKMAP_BASE;
580
# endif
581
	address_markers[FIXADDR_START_NR].start_address = FIXADDR_START;
582
	address_markers[CPU_ENTRY_AREA_NR].start_address = CPU_ENTRY_AREA_BASE;
583
#endif
584
585
586
	return 0;
}
__initcall(pt_dump_init);