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