[mirror_ubuntu-jammy-kernel.git] / arch / x86 / mm / dump_pagetables.c

/*
 * 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.
 */

#include <linux/debugfs.h>
#include <linux/mm.h>
#include <linux/module.h>
#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;
	const struct addr_marker *marker;
	unsigned long lines;
	bool to_dmesg;
};

struct addr_marker {
	unsigned long start_address;
	const char *name;
	unsigned long max_lines;
};

/* indices for address_markers; keep sync'd w/ address_markers below */
enum address_markers_idx {
	USER_SPACE_NR = 0,
#ifdef CONFIG_X86_64
	KERNEL_SPACE_NR,
	LOW_KERNEL_NR,
	VMALLOC_START_NR,
	VMEMMAP_START_NR,
# ifdef CONFIG_X86_ESPFIX64
	ESPFIX_START_NR,
# endif
	HIGH_KERNEL_NR,
	MODULES_VADDR_NR,
	MODULES_END_NR,
#else
	KERNEL_SPACE_NR,
	VMALLOC_START_NR,
	VMALLOC_END_NR,
# ifdef CONFIG_HIGHMEM
	PKMAP_BASE_NR,
# endif
	FIXADDR_START_NR,
#endif
};

/* Address space markers hints */
static struct addr_marker address_markers[] = {
	{ 0, "User Space" },
#ifdef CONFIG_X86_64
	{ 0x8000000000000000UL, "Kernel Space" },
	{ PAGE_OFFSET,		"Low Kernel Mapping" },
	{ VMALLOC_START,        "vmalloc() Area" },
	{ VMEMMAP_START,        "Vmemmap" },
# ifdef CONFIG_X86_ESPFIX64
	{ ESPFIX_BASE_ADDR,	"ESPfix Area", 16 },
# endif
# ifdef CONFIG_EFI
	{ EFI_VA_END,		"EFI Runtime Services" },
# endif
	{ __START_KERNEL_map,   "High Kernel Mapping" },
	{ MODULES_VADDR,        "Modules" },
	{ MODULES_END,          "End Modules" },
#else
	{ PAGE_OFFSET,          "Kernel Mapping" },
	{ 0/* VMALLOC_START */, "vmalloc() Area" },
	{ 0/*VMALLOC_END*/,     "vmalloc() End" },
# ifdef CONFIG_HIGHMEM
	{ 0/*PKMAP_BASE*/,      "Persisent kmap() Area" },
# endif
	{ 0/*FIXADDR_START*/,   "Fixmap Area" },
#endif
	{ -1, NULL }		/* End of list */
};

/* 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)
#define PGD_LEVEL_MULT (PTRS_PER_PUD * PUD_LEVEL_MULT)

#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);		\
})

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

	if (!pgprot_val(prot)) {
		/* Not present */
		pt_dump_cont_printf(m, dmsg, "                              ");
	} else {
		if (pr & _PAGE_USER)
			pt_dump_cont_printf(m, dmsg, "USR ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
		if (pr & _PAGE_RW)
			pt_dump_cont_printf(m, dmsg, "RW ");
		else
			pt_dump_cont_printf(m, dmsg, "ro ");
		if (pr & _PAGE_PWT)
			pt_dump_cont_printf(m, dmsg, "PWT ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
		if (pr & _PAGE_PCD)
			pt_dump_cont_printf(m, dmsg, "PCD ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");

		/* Bit 7 has a different meaning on level 3 vs 4 */
		if (level <= 3 && pr & _PAGE_PSE)
			pt_dump_cont_printf(m, dmsg, "PSE ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
		if ((level == 4 && pr & _PAGE_PAT) ||
		    ((level == 3 || level == 2) && pr & _PAGE_PAT_LARGE))
			pt_dump_cont_printf(m, dmsg, "pat ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
		if (pr & _PAGE_GLOBAL)
			pt_dump_cont_printf(m, dmsg, "GLB ");
		else
			pt_dump_cont_printf(m, dmsg, "    ");
		if (pr & _PAGE_NX)
			pt_dump_cont_printf(m, dmsg, "NX ");
		else
			pt_dump_cont_printf(m, dmsg, "x  ");
	}
	pt_dump_cont_printf(m, dmsg, "%s\n", level_name[level]);
}

/*
 * On 64 bits, sign-extend the 48 bit address to 64 bit
 */
static unsigned long normalize_addr(unsigned long u)
{
#ifdef CONFIG_X86_64
	return (signed long)(u << 16) >> 16;
#else
	return u;
#endif
}

/*
 * 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,
		      pgprot_t new_prot, int level)
{
	pgprotval_t prot, cur;
	static const char units[] = "BKMGTPE";

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

	if (!st->level) {
		/* First entry */
		st->current_prot = new_prot;
		st->level = level;
		st->marker = address_markers;
		st->lines = 0;
		pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
				   st->marker->name);
	} else if (prot != cur || level != st->level ||
		   st->current_address >= st->marker[1].start_address) {
		const char *unit = units;
		unsigned long delta;
		int width = sizeof(unsigned long) * 2;

		/*
		 * Now print the actual finished series
		 */
		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);

			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);
		}
		st->lines++;

		/*
		 * 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) {
			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");
			}
			st->marker++;
			st->lines = 0;
			pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
					   st->marker->name);
		}

		st->start_address = st->current_address;
		st->current_prot = new_prot;
		st->level = level;
	}
}

static void walk_pte_level(struct seq_file *m, struct pg_state *st, pmd_t addr,
							unsigned long P)
{
	int i;
	pte_t *start;

	start = (pte_t *) pmd_page_vaddr(addr);
	for (i = 0; i < PTRS_PER_PTE; i++) {
		pgprot_t prot = pte_pgprot(*start);

		st->current_address = normalize_addr(P + i * PTE_LEVEL_MULT);
		note_page(m, st, prot, 4);
		start++;
	}
}

#if PTRS_PER_PMD > 1

static void walk_pmd_level(struct seq_file *m, struct pg_state *st, pud_t addr,
							unsigned long P)
{
	int i;
	pmd_t *start;

	start = (pmd_t *) pud_page_vaddr(addr);
	for (i = 0; i < PTRS_PER_PMD; i++) {
		st->current_address = normalize_addr(P + i * PMD_LEVEL_MULT);
		if (!pmd_none(*start)) {
			pgprotval_t prot = pmd_val(*start) & PTE_FLAGS_MASK;

			if (pmd_large(*start) || !pmd_present(*start))
				note_page(m, st, __pgprot(prot), 3);
			else
				walk_pte_level(m, st, *start,
					       P + i * PMD_LEVEL_MULT);
		} else
			note_page(m, st, __pgprot(0), 3);
		start++;
	}
}

#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

#if PTRS_PER_PUD > 1

static void walk_pud_level(struct seq_file *m, struct pg_state *st, pgd_t addr,
							unsigned long P)
{
	int i;
	pud_t *start;

	start = (pud_t *) pgd_page_vaddr(addr);

	for (i = 0; i < PTRS_PER_PUD; i++) {
		st->current_address = normalize_addr(P + i * PUD_LEVEL_MULT);
		if (!pud_none(*start)) {
			pgprotval_t prot = pud_val(*start) & PTE_FLAGS_MASK;

			if (pud_large(*start) || !pud_present(*start))
				note_page(m, st, __pgprot(prot), 2);
			else
				walk_pmd_level(m, st, *start,
					       P + i * PUD_LEVEL_MULT);
		} else
			note_page(m, st, __pgprot(0), 2);

		start++;
	}
}

#else
#define walk_pud_level(m,s,a,p) walk_pmd_level(m,s,__pud(pgd_val(a)),p)
#define pgd_large(a) pud_large(__pud(pgd_val(a)))
#define pgd_none(a)  pud_none(__pud(pgd_val(a)))
#endif

void ptdump_walk_pgd_level(struct seq_file *m, pgd_t *pgd)
{
#ifdef CONFIG_X86_64
	pgd_t *start = (pgd_t *) &init_level4_pgt;
#else
	pgd_t *start = swapper_pg_dir;
#endif
	int i;
	struct pg_state st = {};

	if (pgd) {
		start = pgd;
		st.to_dmesg = true;
	}

	for (i = 0; i < PTRS_PER_PGD; i++) {
		st.current_address = normalize_addr(i * PGD_LEVEL_MULT);
		if (!pgd_none(*start)) {
			pgprotval_t prot = pgd_val(*start) & PTE_FLAGS_MASK;

			if (pgd_large(*start) || !pgd_present(*start))
				note_page(m, &st, __pgprot(prot), 1);
			else
				walk_pud_level(m, &st, *start,
					       i * PGD_LEVEL_MULT);
		} else
			note_page(m, &st, __pgprot(0), 1);

		start++;
	}

	/* Flush out the last page */
	st.current_address = normalize_addr(PTRS_PER_PGD*PGD_LEVEL_MULT);
	note_page(m, &st, __pgprot(0), 0);
}

static int ptdump_show(struct seq_file *m, void *v)
{
	ptdump_walk_pgd_level(m, NULL);
	return 0;
}

static int ptdump_open(struct inode *inode, struct file *filp)
{
	return single_open(filp, ptdump_show, NULL);
}

static const struct file_operations ptdump_fops = {
	.open		= ptdump_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
};

static int pt_dump_init(void)
{
	struct dentry *pe;

#ifdef CONFIG_X86_32
	/* Not a compile-time constant on x86-32 */
	address_markers[VMALLOC_START_NR].start_address = VMALLOC_START;
	address_markers[VMALLOC_END_NR].start_address = VMALLOC_END;
# ifdef CONFIG_HIGHMEM
	address_markers[PKMAP_BASE_NR].start_address = PKMAP_BASE;
# endif
	address_markers[FIXADDR_START_NR].start_address = FIXADDR_START;
#endif

	pe = debugfs_create_file("kernel_page_tables", 0600, NULL, NULL,
				 &ptdump_fops);
	if (!pe)
		return -ENOMEM;

	return 0;
}

__initcall(pt_dump_init);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Arjan van de Ven <arjan@linux.intel.com>");
MODULE_DESCRIPTION("Kernel debugging helper that dumps pagetables");
Commit	Line	Data
926e5392 AV	1	/*
	2	* Debug helper to dump the current kernel pagetables of the system
	3	* so that we can see what the various memory ranges are set to.
	4	*
	5	* (C) Copyright 2008 Intel Corporation
	6	*
	7	* Author: Arjan van de Ven <arjan@linux.intel.com>
	8	*
	9	* This program is free software; you can redistribute it and/or
	10	* modify it under the terms of the GNU General Public License
	11	* as published by the Free Software Foundation; version 2
	12	* of the License.
	13	*/
	14
fe770bf0 PA	15	#include <linux/debugfs.h>
fe770bf0 PA	16	#include <linux/mm.h>
926e5392 AV	17	#include <linux/module.h>
926e5392 AV	18	#include <linux/seq_file.h>
926e5392 AV	19
	20	#include <asm/pgtable.h>
	21
	22	/*
	23	* The dumper groups pagetable entries of the same type into one, and for
	24	* that it needs to keep some state when walking, and flush this state
	25	* when a "break" in the continuity is found.
	26	*/
	27	struct pg_state {
	28	int level;
	29	pgprot_t current_prot;
	30	unsigned long start_address;
	31	unsigned long current_address;
fe770bf0	32	const struct addr_marker *marker;
3891a04a	33	unsigned long lines;
ef6bea6d	34	bool to_dmesg;
926e5392 AV	35	};
926e5392 AV	36
fe770bf0 PA	37	struct addr_marker {
	38	unsigned long start_address;
	39	const char *name;
3891a04a	40	unsigned long max_lines;
fe770bf0 PA	41	};
fe770bf0 PA	42
92851e2f AS	43	/* indices for address_markers; keep sync'd w/ address_markers below */
	44	enum address_markers_idx {
	45	USER_SPACE_NR = 0,
	46	#ifdef CONFIG_X86_64
	47	KERNEL_SPACE_NR,
	48	LOW_KERNEL_NR,
	49	VMALLOC_START_NR,
	50	VMEMMAP_START_NR,
8a5a5d15	51	# ifdef CONFIG_X86_ESPFIX64
3891a04a	52	ESPFIX_START_NR,
8a5a5d15	53	# endif
92851e2f AS	54	HIGH_KERNEL_NR,
	55	MODULES_VADDR_NR,
	56	MODULES_END_NR,
	57	#else
	58	KERNEL_SPACE_NR,
	59	VMALLOC_START_NR,
	60	VMALLOC_END_NR,
	61	# ifdef CONFIG_HIGHMEM
	62	PKMAP_BASE_NR,
	63	# endif
	64	FIXADDR_START_NR,
	65	#endif
	66	};
	67
fe770bf0 PA	68	/* Address space markers hints */
	69	static struct addr_marker address_markers[] = {
	70	{ 0, "User Space" },
	71	#ifdef CONFIG_X86_64
	72	{ 0x8000000000000000UL, "Kernel Space" },
684eb016	73	{ PAGE_OFFSET, "Low Kernel Mapping" },
fe770bf0	74	{ VMALLOC_START, "vmalloc() Area" },
fe770bf0	75	{ VMEMMAP_START, "Vmemmap" },
8a5a5d15	76	# ifdef CONFIG_X86_ESPFIX64
3891a04a	77	{ ESPFIX_BASE_ADDR, "ESPfix Area", 16 },
8266e31e MK	78	# endif
	79	# ifdef CONFIG_EFI
	80	{ EFI_VA_END, "EFI Runtime Services" },
8a5a5d15	81	# endif
fe770bf0	82	{ __START_KERNEL_map, "High Kernel Mapping" },
9a79cf9c YL	83	{ MODULES_VADDR, "Modules" },
9a79cf9c YL	84	{ MODULES_END, "End Modules" },
fe770bf0 PA	85	#else
	86	{ PAGE_OFFSET, "Kernel Mapping" },
	87	{ 0/* VMALLOC_START */, "vmalloc() Area" },
	88	{ 0/VMALLOC_END/, "vmalloc() End" },
	89	# ifdef CONFIG_HIGHMEM
	90	{ 0/PKMAP_BASE/, "Persisent kmap() Area" },
	91	# endif
	92	{ 0/FIXADDR_START/, "Fixmap Area" },
	93	#endif
	94	{ -1, NULL } /* End of list */
	95	};
926e5392	96
fe770bf0 PA	97	/* Multipliers for offsets within the PTEs */
	98	#define PTE_LEVEL_MULT (PAGE_SIZE)
	99	#define PMD_LEVEL_MULT (PTRS_PER_PTE * PTE_LEVEL_MULT)
	100	#define PUD_LEVEL_MULT (PTRS_PER_PMD * PMD_LEVEL_MULT)
	101	#define PGD_LEVEL_MULT (PTRS_PER_PUD * PUD_LEVEL_MULT)
926e5392	102
ef6bea6d BP	103	#define pt_dump_seq_printf(m, to_dmesg, fmt, args...) \
	104	({ \
	105	if (to_dmesg) \
	106	printk(KERN_INFO fmt, ##args); \
	107	else \
	108	if (m) \
	109	seq_printf(m, fmt, ##args); \
	110	})
	111
	112	#define pt_dump_cont_printf(m, to_dmesg, fmt, args...) \
	113	({ \
	114	if (to_dmesg) \
	115	printk(KERN_CONT fmt, ##args); \
	116	else \
	117	if (m) \
	118	seq_printf(m, fmt, ##args); \
	119	})
	120
926e5392 AV	121	/*
	122	* Print a readable form of a pgprot_t to the seq_file
	123	*/
ef6bea6d	124	static void printk_prot(struct seq_file *m, pgprot_t prot, int level, bool dmsg)
926e5392	125	{
fe770bf0 PA	126	pgprotval_t pr = pgprot_val(prot);
	127	static const char * const level_name[] =
	128	{ "cr3", "pgd", "pud", "pmd", "pte" };
	129
	130	if (!pgprot_val(prot)) {
	131	/* Not present */
f439c429	132	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0 PA	133	} else {
fe770bf0 PA	134	if (pr & _PAGE_USER)
ef6bea6d	135	pt_dump_cont_printf(m, dmsg, "USR ");
926e5392	136	else
ef6bea6d	137	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0	138	if (pr & _PAGE_RW)
ef6bea6d	139	pt_dump_cont_printf(m, dmsg, "RW ");
fe770bf0	140	else
ef6bea6d	141	pt_dump_cont_printf(m, dmsg, "ro ");
fe770bf0	142	if (pr & _PAGE_PWT)
ef6bea6d	143	pt_dump_cont_printf(m, dmsg, "PWT ");
fe770bf0	144	else
ef6bea6d	145	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0	146	if (pr & _PAGE_PCD)
ef6bea6d	147	pt_dump_cont_printf(m, dmsg, "PCD ");
926e5392	148	else
ef6bea6d	149	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0	150
f439c429 JG	151	/* Bit 7 has a different meaning on level 3 vs 4 */
	152	if (level <= 3 && pr & _PAGE_PSE)
	153	pt_dump_cont_printf(m, dmsg, "PSE ");
	154	else
	155	pt_dump_cont_printf(m, dmsg, " ");
	156	if ((level == 4 && pr & _PAGE_PAT) \|\|
	157	((level == 3 \|\| level == 2) && pr & _PAGE_PAT_LARGE))
	158	pt_dump_cont_printf(m, dmsg, "pat ");
	159	else
	160	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0	161	if (pr & _PAGE_GLOBAL)
ef6bea6d	162	pt_dump_cont_printf(m, dmsg, "GLB ");
fe770bf0	163	else
ef6bea6d	164	pt_dump_cont_printf(m, dmsg, " ");
fe770bf0	165	if (pr & _PAGE_NX)
ef6bea6d	166	pt_dump_cont_printf(m, dmsg, "NX ");
fe770bf0	167	else
ef6bea6d	168	pt_dump_cont_printf(m, dmsg, "x ");
926e5392	169	}
ef6bea6d	170	pt_dump_cont_printf(m, dmsg, "%s\n", level_name[level]);
926e5392 AV	171	}
	172
	173	/*
fe770bf0	174	* On 64 bits, sign-extend the 48 bit address to 64 bit
926e5392	175	*/
fe770bf0	176	static unsigned long normalize_addr(unsigned long u)
926e5392	177	{
fe770bf0 PA	178	#ifdef CONFIG_X86_64
	179	return (signed long)(u << 16) >> 16;
	180	#else
926e5392	181	return u;
fe770bf0	182	#endif
926e5392 AV	183	}
	184
	185	/*
	186	* This function gets called on a break in a continuous series
	187	* of PTE entries; the next one is different so we need to
	188	* print what we collected so far.
	189	*/
	190	static void note_page(struct seq_file m, struct pg_state st,
fe770bf0	191	pgprot_t new_prot, int level)
926e5392	192	{
fe770bf0	193	pgprotval_t prot, cur;
3891a04a	194	static const char units[] = "BKMGTPE";
926e5392 AV	195
	196	/*
	197	* If we have a "break" in the series, we need to flush the state that
fe770bf0 PA	198	* we have now. "break" is either changing perms, levels or
fe770bf0 PA	199	* address space marker.
926e5392	200	*/
27990eac JF	201	prot = pgprot_val(new_prot) & PTE_FLAGS_MASK;
27990eac JF	202	cur = pgprot_val(st->current_prot) & PTE_FLAGS_MASK;
926e5392	203
fe770bf0 PA	204	if (!st->level) {
	205	/* First entry */
	206	st->current_prot = new_prot;
	207	st->level = level;
	208	st->marker = address_markers;
3891a04a	209	st->lines = 0;
ef6bea6d BP	210	pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
ef6bea6d BP	211	st->marker->name);
fe770bf0 PA	212	} else if (prot != cur \|\| level != st->level \|\|
	213	st->current_address >= st->marker[1].start_address) {
	214	const char *unit = units;
926e5392	215	unsigned long delta;
6424fb38	216	int width = sizeof(unsigned long) * 2;
926e5392	217
926e5392 AV	218	/*
	219	* Now print the actual finished series
	220	*/
3891a04a PA	221	if (!st->marker->max_lines \|\|
	222	st->lines < st->marker->max_lines) {
	223	pt_dump_seq_printf(m, st->to_dmesg,
	224	"0x%0lx-0x%0lx ",
	225	width, st->start_address,
	226	width, st->current_address);
926e5392	227
3891a04a PA	228	delta = st->current_address - st->start_address;
	229	while (!(delta & 1023) && unit[1]) {
	230	delta >>= 10;
	231	unit++;
	232	}
	233	pt_dump_cont_printf(m, st->to_dmesg, "%9lu%c ",
	234	delta, *unit);
	235	printk_prot(m, st->current_prot, st->level,
	236	st->to_dmesg);
926e5392	237	}
3891a04a	238	st->lines++;
fe770bf0 PA	239
	240	/*
	241	* We print markers for special areas of address space,
	242	* such as the start of vmalloc space etc.
	243	* This helps in the interpretation.
	244	*/
	245	if (st->current_address >= st->marker[1].start_address) {
3891a04a PA	246	if (st->marker->max_lines &&
	247	st->lines > st->marker->max_lines) {
	248	unsigned long nskip =
	249	st->lines - st->marker->max_lines;
	250	pt_dump_seq_printf(m, st->to_dmesg,
	251	"... %lu entr%s skipped ... \n",
	252	nskip,
	253	nskip == 1 ? "y" : "ies");
	254	}
fe770bf0	255	st->marker++;
3891a04a	256	st->lines = 0;
ef6bea6d BP	257	pt_dump_seq_printf(m, st->to_dmesg, "---[ %s ]---\n",
ef6bea6d BP	258	st->marker->name);
926e5392	259	}
fe770bf0	260
926e5392 AV	261	st->start_address = st->current_address;
	262	st->current_prot = new_prot;
	263	st->level = level;
fe770bf0	264	}
926e5392 AV	265	}
926e5392 AV	266
fe770bf0	267	static void walk_pte_level(struct seq_file m, struct pg_state st, pmd_t addr,
926e5392 AV	268	unsigned long P)
	269	{
	270	int i;
	271	pte_t *start;
	272
	273	start = (pte_t *) pmd_page_vaddr(addr);
	274	for (i = 0; i < PTRS_PER_PTE; i++) {
	275	pgprot_t prot = pte_pgprot(*start);
	276
fe770bf0	277	st->current_address = normalize_addr(P + i * PTE_LEVEL_MULT);
926e5392 AV	278	note_page(m, st, prot, 4);
	279	start++;
	280	}
	281	}
	282
fe770bf0	283	#if PTRS_PER_PMD > 1
926e5392	284
fe770bf0	285	static void walk_pmd_level(struct seq_file m, struct pg_state st, pud_t addr,
926e5392 AV	286	unsigned long P)
	287	{
	288	int i;
	289	pmd_t *start;
	290
	291	start = (pmd_t *) pud_page_vaddr(addr);
	292	for (i = 0; i < PTRS_PER_PMD; i++) {
fe770bf0	293	st->current_address = normalize_addr(P + i * PMD_LEVEL_MULT);
926e5392	294	if (!pmd_none(*start)) {
77be1fab	295	pgprotval_t prot = pmd_val(*start) & PTE_FLAGS_MASK;
926e5392	296
fe770bf0	297	if (pmd_large(start) \|\| !pmd_present(start))
926e5392 AV	298	note_page(m, st, __pgprot(prot), 3);
926e5392 AV	299	else
fe770bf0 PA	300	walk_pte_level(m, st, *start,
fe770bf0 PA	301	P + i * PMD_LEVEL_MULT);
926e5392 AV	302	} else
	303	note_page(m, st, __pgprot(0), 3);
	304	start++;
	305	}
	306	}
	307
fe770bf0 PA	308	#else
	309	#define walk_pmd_level(m,s,a,p) walk_pte_level(m,s,__pmd(pud_val(a)),p)
	310	#define pud_large(a) pmd_large(__pmd(pud_val(a)))
	311	#define pud_none(a) pmd_none(__pmd(pud_val(a)))
	312	#endif
926e5392	313
fe770bf0 PA	314	#if PTRS_PER_PUD > 1
	315
	316	static void walk_pud_level(struct seq_file m, struct pg_state st, pgd_t addr,
926e5392 AV	317	unsigned long P)
	318	{
	319	int i;
	320	pud_t *start;
	321
	322	start = (pud_t *) pgd_page_vaddr(addr);
	323
	324	for (i = 0; i < PTRS_PER_PUD; i++) {
fe770bf0	325	st->current_address = normalize_addr(P + i * PUD_LEVEL_MULT);
926e5392	326	if (!pud_none(*start)) {
77be1fab	327	pgprotval_t prot = pud_val(*start) & PTE_FLAGS_MASK;
926e5392	328
fe770bf0	329	if (pud_large(start) \|\| !pud_present(start))
926e5392 AV	330	note_page(m, st, __pgprot(prot), 2);
926e5392 AV	331	else
fe770bf0 PA	332	walk_pmd_level(m, st, *start,
fe770bf0 PA	333	P + i * PUD_LEVEL_MULT);
926e5392 AV	334	} else
	335	note_page(m, st, __pgprot(0), 2);
	336
	337	start++;
	338	}
	339	}
	340
fe770bf0 PA	341	#else
	342	#define walk_pud_level(m,s,a,p) walk_pmd_level(m,s,__pud(pgd_val(a)),p)
	343	#define pgd_large(a) pud_large(__pud(pgd_val(a)))
	344	#define pgd_none(a) pud_none(__pud(pgd_val(a)))
	345	#endif
	346
ef6bea6d	347	void ptdump_walk_pgd_level(struct seq_file m, pgd_t pgd)
926e5392	348	{
fe770bf0	349	#ifdef CONFIG_X86_64
926e5392	350	pgd_t start = (pgd_t ) &init_level4_pgt;
fe770bf0 PA	351	#else
	352	pgd_t *start = swapper_pg_dir;
	353	#endif
926e5392	354	int i;
ef6bea6d	355	struct pg_state st = {};
926e5392	356
ef6bea6d BP	357	if (pgd) {
	358	start = pgd;
	359	st.to_dmesg = true;
	360	}
926e5392 AV	361
926e5392 AV	362	for (i = 0; i < PTRS_PER_PGD; i++) {
fe770bf0 PA	363	st.current_address = normalize_addr(i * PGD_LEVEL_MULT);
fe770bf0 PA	364	if (!pgd_none(*start)) {
77be1fab	365	pgprotval_t prot = pgd_val(*start) & PTE_FLAGS_MASK;
fe770bf0 PA	366
	367	if (pgd_large(start) \|\| !pgd_present(start))
	368	note_page(m, &st, __pgprot(prot), 1);
	369	else
	370	walk_pud_level(m, &st, *start,
	371	i * PGD_LEVEL_MULT);
	372	} else
926e5392	373	note_page(m, &st, __pgprot(0), 1);
fe770bf0	374
926e5392 AV	375	start++;
926e5392 AV	376	}
fe770bf0 PA	377
	378	/* Flush out the last page */
	379	st.current_address = normalize_addr(PTRS_PER_PGD*PGD_LEVEL_MULT);
	380	note_page(m, &st, __pgprot(0), 0);
926e5392 AV	381	}
	382
	383	static int ptdump_show(struct seq_file m, void v)
	384	{
ef6bea6d	385	ptdump_walk_pgd_level(m, NULL);
926e5392 AV	386	return 0;
	387	}
	388
	389	static int ptdump_open(struct inode inode, struct file filp)
	390	{
	391	return single_open(filp, ptdump_show, NULL);
	392	}
	393
	394	static const struct file_operations ptdump_fops = {
	395	.open = ptdump_open,
	396	.read = seq_read,
	397	.llseek = seq_lseek,
	398	.release = single_release,
	399	};
	400
a4928cff	401	static int pt_dump_init(void)
926e5392 AV	402	{
	403	struct dentry *pe;
	404
fe770bf0 PA	405	#ifdef CONFIG_X86_32
fe770bf0 PA	406	/* Not a compile-time constant on x86-32 */
92851e2f AS	407	address_markers[VMALLOC_START_NR].start_address = VMALLOC_START;
92851e2f AS	408	address_markers[VMALLOC_END_NR].start_address = VMALLOC_END;
fe770bf0	409	# ifdef CONFIG_HIGHMEM
92851e2f	410	address_markers[PKMAP_BASE_NR].start_address = PKMAP_BASE;
fe770bf0	411	# endif
92851e2f	412	address_markers[FIXADDR_START_NR].start_address = FIXADDR_START;
fe770bf0 PA	413	#endif
fe770bf0 PA	414
926e5392 AV	415	pe = debugfs_create_file("kernel_page_tables", 0600, NULL, NULL,
	416	&ptdump_fops);
	417	if (!pe)
	418	return -ENOMEM;
	419
	420	return 0;
	421	}
	422
	423	__initcall(pt_dump_init);
	424	MODULE_LICENSE("GPL");
	425	MODULE_AUTHOR("Arjan van de Ven <arjan@linux.intel.com>");
	426	MODULE_DESCRIPTION("Kernel debugging helper that dumps pagetables");