[mirror_ubuntu-zesty-kernel.git] / mm / pagewalk.c

#include <linux/mm.h>
#include <linux/highmem.h>
#include <linux/sched.h>
#include <linux/hugetlb.h>

static int walk_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	pte_t *pte;
	int err = 0;

	pte = pte_offset_map(pmd, addr);
	for (;;) {
		err = walk->pte_entry(pte, addr, addr + PAGE_SIZE, walk);
		if (err)
		       break;
		addr += PAGE_SIZE;
		if (addr == end)
			break;
		pte++;
	}

	pte_unmap(pte);
	return err;
}

static int walk_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	pmd_t *pmd;
	unsigned long next;
	int err = 0;

	pmd = pmd_offset(pud, addr);
	do {
again:
		next = pmd_addr_end(addr, end);
		if (pmd_none(*pmd)) {
			if (walk->pte_hole)
				err = walk->pte_hole(addr, next, walk);
			if (err)
				break;
			continue;
		}
		/*
		 * This implies that each ->pmd_entry() handler
		 * needs to know about pmd_trans_huge() pmds
		 */
		if (walk->pmd_entry)
			err = walk->pmd_entry(pmd, addr, next, walk);
		if (err)
			break;

		/*
		 * Check this here so we only break down trans_huge
		 * pages when we _need_ to
		 */
		if (!walk->pte_entry)
			continue;

		split_huge_page_pmd_mm(walk->mm, addr, pmd);
		if (pmd_trans_unstable(pmd))
			goto again;
		err = walk_pte_range(pmd, addr, next, walk);
		if (err)
			break;
	} while (pmd++, addr = next, addr != end);

	return err;
}

static int walk_pud_range(pgd_t *pgd, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	pud_t *pud;
	unsigned long next;
	int err = 0;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud)) {
			if (walk->pte_hole)
				err = walk->pte_hole(addr, next, walk);
			if (err)
				break;
			continue;
		}
		if (walk->pmd_entry || walk->pte_entry)
			err = walk_pmd_range(pud, addr, next, walk);
		if (err)
			break;
	} while (pud++, addr = next, addr != end);

	return err;
}

static int walk_pgd_range(unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	pgd_t *pgd;
	unsigned long next;
	int err = 0;

	pgd = pgd_offset(walk->mm, addr);
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd)) {
			if (walk->pte_hole)
				err = walk->pte_hole(addr, next, walk);
			if (err)
				break;
			continue;
		}
		if (walk->pmd_entry || walk->pte_entry)
			err = walk_pud_range(pgd, addr, next, walk);
		if (err)
			break;
	} while (pgd++, addr = next, addr != end);

	return err;
}

#ifdef CONFIG_HUGETLB_PAGE
static unsigned long hugetlb_entry_end(struct hstate *h, unsigned long addr,
				       unsigned long end)
{
	unsigned long boundary = (addr & huge_page_mask(h)) + huge_page_size(h);
	return boundary < end ? boundary : end;
}

static int walk_hugetlb_range(unsigned long addr, unsigned long end,
			      struct mm_walk *walk)
{
	struct vm_area_struct *vma = walk->vma;
	struct hstate *h = hstate_vma(vma);
	unsigned long next;
	unsigned long hmask = huge_page_mask(h);
	pte_t *pte;
	int err = 0;

	do {
		next = hugetlb_entry_end(h, addr, end);
		pte = huge_pte_offset(walk->mm, addr & hmask);
		if (pte && walk->hugetlb_entry)
			err = walk->hugetlb_entry(pte, hmask, addr, next, walk);
		if (err)
			break;
	} while (addr = next, addr != end);

	return err;
}

#else /* CONFIG_HUGETLB_PAGE */
static int walk_hugetlb_range(unsigned long addr, unsigned long end,
			      struct mm_walk *walk)
{
	return 0;
}

#endif /* CONFIG_HUGETLB_PAGE */

/*
 * Decide whether we really walk over the current vma on [@start, @end)
 * or skip it via the returned value. Return 0 if we do walk over the
 * current vma, and return 1 if we skip the vma. Negative values means
 * error, where we abort the current walk.
 *
 * Default check (only VM_PFNMAP check for now) is used when the caller
 * doesn't define test_walk() callback.
 */
static int walk_page_test(unsigned long start, unsigned long end,
			struct mm_walk *walk)
{
	struct vm_area_struct *vma = walk->vma;

	if (walk->test_walk)
		return walk->test_walk(start, end, walk);

	/*
	 * Do not walk over vma(VM_PFNMAP), because we have no valid struct
	 * page backing a VM_PFNMAP range. See also commit a9ff785e4437.
	 */
	if (vma->vm_flags & VM_PFNMAP)
		return 1;
	return 0;
}

static int __walk_page_range(unsigned long start, unsigned long end,
			struct mm_walk *walk)
{
	int err = 0;
	struct vm_area_struct *vma = walk->vma;

	if (vma && is_vm_hugetlb_page(vma)) {
		if (walk->hugetlb_entry)
			err = walk_hugetlb_range(start, end, walk);
	} else
		err = walk_pgd_range(start, end, walk);

	return err;
}

/**
 * walk_page_range - walk page table with caller specific callbacks
 *
 * Recursively walk the page table tree of the process represented by @walk->mm
 * within the virtual address range [@start, @end). During walking, we can do
 * some caller-specific works for each entry, by setting up pmd_entry(),
 * pte_entry(), and/or hugetlb_entry(). If you don't set up for some of these
 * callbacks, the associated entries/pages are just ignored.
 * The return values of these callbacks are commonly defined like below:
 *  - 0  : succeeded to handle the current entry, and if you don't reach the
 *         end address yet, continue to walk.
 *  - >0 : succeeded to handle the current entry, and return to the caller
 *         with caller specific value.
 *  - <0 : failed to handle the current entry, and return to the caller
 *         with error code.
 *
 * Before starting to walk page table, some callers want to check whether
 * they really want to walk over the current vma, typically by checking
 * its vm_flags. walk_page_test() and @walk->test_walk() are used for this
 * purpose.
 *
 * struct mm_walk keeps current values of some common data like vma and pmd,
 * which are useful for the access from callbacks. If you want to pass some
 * caller-specific data to callbacks, @walk->private should be helpful.
 *
 * Locking:
 *   Callers of walk_page_range() and walk_page_vma() should hold
 *   @walk->mm->mmap_sem, because these function traverse vma list and/or
 *   access to vma's data.
 */
int walk_page_range(unsigned long start, unsigned long end,
		    struct mm_walk *walk)
{
	int err = 0;
	unsigned long next;
	struct vm_area_struct *vma;

	if (start >= end)
		return -EINVAL;

	if (!walk->mm)
		return -EINVAL;

	VM_BUG_ON_MM(!rwsem_is_locked(&walk->mm->mmap_sem), walk->mm);

	vma = find_vma(walk->mm, start);
	do {
		if (!vma) { /* after the last vma */
			walk->vma = NULL;
			next = end;
		} else if (start < vma->vm_start) { /* outside vma */
			walk->vma = NULL;
			next = min(end, vma->vm_start);
		} else { /* inside vma */
			walk->vma = vma;
			next = min(end, vma->vm_end);
			vma = vma->vm_next;

			err = walk_page_test(start, next, walk);
			if (err > 0)
				continue;
			if (err < 0)
				break;
		}
		if (walk->vma || walk->pte_hole)
			err = __walk_page_range(start, next, walk);
		if (err)
			break;
	} while (start = next, start < end);
	return err;
}
Commit	Line	Data
e6473092 MM	1	#include <linux/mm.h>
	2	#include <linux/highmem.h>
	3	#include <linux/sched.h>
d33b9f45	4	#include <linux/hugetlb.h>
e6473092 MM	5
e6473092 MM	6	static int walk_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
2165009b	7	struct mm_walk *walk)
e6473092 MM	8	{
	9	pte_t *pte;
	10	int err = 0;
	11
	12	pte = pte_offset_map(pmd, addr);
556637cd	13	for (;;) {
2165009b	14	err = walk->pte_entry(pte, addr, addr + PAGE_SIZE, walk);
e6473092 MM	15	if (err)
e6473092 MM	16	break;
556637cd JW	17	addr += PAGE_SIZE;
	18	if (addr == end)
	19	break;
	20	pte++;
	21	}
e6473092 MM	22
	23	pte_unmap(pte);
	24	return err;
	25	}
	26
	27	static int walk_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
2165009b	28	struct mm_walk *walk)
e6473092 MM	29	{
	30	pmd_t *pmd;
	31	unsigned long next;
	32	int err = 0;
	33
	34	pmd = pmd_offset(pud, addr);
	35	do {
03319327	36	again:
e6473092	37	next = pmd_addr_end(addr, end);
03319327	38	if (pmd_none(*pmd)) {
e6473092	39	if (walk->pte_hole)
2165009b	40	err = walk->pte_hole(addr, next, walk);
e6473092 MM	41	if (err)
	42	break;
	43	continue;
	44	}
03319327 DH	45	/*
	46	* This implies that each ->pmd_entry() handler
	47	* needs to know about pmd_trans_huge() pmds
	48	*/
e6473092	49	if (walk->pmd_entry)
2165009b	50	err = walk->pmd_entry(pmd, addr, next, walk);
03319327 DH	51	if (err)
	52	break;
	53
	54	/*
	55	* Check this here so we only break down trans_huge
	56	* pages when we _need_ to
	57	*/
	58	if (!walk->pte_entry)
	59	continue;
	60
e180377f	61	split_huge_page_pmd_mm(walk->mm, addr, pmd);
fafaa426	62	if (pmd_trans_unstable(pmd))
03319327 DH	63	goto again;
03319327 DH	64	err = walk_pte_range(pmd, addr, next, walk);
e6473092 MM	65	if (err)
	66	break;
	67	} while (pmd++, addr = next, addr != end);
	68
	69	return err;
	70	}
	71
	72	static int walk_pud_range(pgd_t *pgd, unsigned long addr, unsigned long end,
2165009b	73	struct mm_walk *walk)
e6473092 MM	74	{
	75	pud_t *pud;
	76	unsigned long next;
	77	int err = 0;
	78
	79	pud = pud_offset(pgd, addr);
	80	do {
	81	next = pud_addr_end(addr, end);
	82	if (pud_none_or_clear_bad(pud)) {
	83	if (walk->pte_hole)
2165009b	84	err = walk->pte_hole(addr, next, walk);
e6473092 MM	85	if (err)
	86	break;
	87	continue;
	88	}
0b1fbfe5	89	if (walk->pmd_entry \|\| walk->pte_entry)
2165009b	90	err = walk_pmd_range(pud, addr, next, walk);
e6473092 MM	91	if (err)
	92	break;
	93	} while (pud++, addr = next, addr != end);
	94
	95	return err;
	96	}
	97
fafaa426 NH	98	static int walk_pgd_range(unsigned long addr, unsigned long end,
	99	struct mm_walk *walk)
	100	{
	101	pgd_t *pgd;
	102	unsigned long next;
	103	int err = 0;
	104
	105	pgd = pgd_offset(walk->mm, addr);
	106	do {
	107	next = pgd_addr_end(addr, end);
	108	if (pgd_none_or_clear_bad(pgd)) {
	109	if (walk->pte_hole)
	110	err = walk->pte_hole(addr, next, walk);
	111	if (err)
	112	break;
	113	continue;
	114	}
	115	if (walk->pmd_entry \|\| walk->pte_entry)
	116	err = walk_pud_range(pgd, addr, next, walk);
	117	if (err)
	118	break;
	119	} while (pgd++, addr = next, addr != end);
	120
	121	return err;
	122	}
	123
116354d1 NH	124	#ifdef CONFIG_HUGETLB_PAGE
	125	static unsigned long hugetlb_entry_end(struct hstate *h, unsigned long addr,
	126	unsigned long end)
	127	{
	128	unsigned long boundary = (addr & huge_page_mask(h)) + huge_page_size(h);
	129	return boundary < end ? boundary : end;
	130	}
	131
fafaa426	132	static int walk_hugetlb_range(unsigned long addr, unsigned long end,
116354d1 NH	133	struct mm_walk *walk)
116354d1 NH	134	{
fafaa426	135	struct vm_area_struct *vma = walk->vma;
116354d1 NH	136	struct hstate *h = hstate_vma(vma);
	137	unsigned long next;
	138	unsigned long hmask = huge_page_mask(h);
	139	pte_t *pte;
	140	int err = 0;
	141
	142	do {
	143	next = hugetlb_entry_end(h, addr, end);
	144	pte = huge_pte_offset(walk->mm, addr & hmask);
	145	if (pte && walk->hugetlb_entry)
	146	err = walk->hugetlb_entry(pte, hmask, addr, next, walk);
	147	if (err)
fafaa426	148	break;
116354d1 NH	149	} while (addr = next, addr != end);
116354d1 NH	150
fafaa426	151	return err;
116354d1	152	}
6c6d5280	153
6c6d5280	154	#else /* CONFIG_HUGETLB_PAGE */
fafaa426	155	static int walk_hugetlb_range(unsigned long addr, unsigned long end,
6c6d5280 KM	156	struct mm_walk *walk)
	157	{
	158	return 0;
	159	}
	160
	161	#endif /* CONFIG_HUGETLB_PAGE */
	162
fafaa426 NH	163	/*
	164	* Decide whether we really walk over the current vma on [@start, @end)
	165	* or skip it via the returned value. Return 0 if we do walk over the
	166	* current vma, and return 1 if we skip the vma. Negative values means
	167	* error, where we abort the current walk.
	168	*
	169	* Default check (only VM_PFNMAP check for now) is used when the caller
	170	* doesn't define test_walk() callback.
	171	*/
	172	static int walk_page_test(unsigned long start, unsigned long end,
	173	struct mm_walk *walk)
	174	{
	175	struct vm_area_struct *vma = walk->vma;
6c6d5280	176
fafaa426 NH	177	if (walk->test_walk)
	178	return walk->test_walk(start, end, walk);
	179
	180	/*
	181	* Do not walk over vma(VM_PFNMAP), because we have no valid struct
	182	* page backing a VM_PFNMAP range. See also commit a9ff785e4437.
	183	*/
	184	if (vma->vm_flags & VM_PFNMAP)
	185	return 1;
	186	return 0;
	187	}
	188
	189	static int __walk_page_range(unsigned long start, unsigned long end,
	190	struct mm_walk *walk)
	191	{
	192	int err = 0;
	193	struct vm_area_struct *vma = walk->vma;
	194
	195	if (vma && is_vm_hugetlb_page(vma)) {
	196	if (walk->hugetlb_entry)
	197	err = walk_hugetlb_range(start, end, walk);
	198	} else
	199	err = walk_pgd_range(start, end, walk);
	200
	201	return err;
	202	}
116354d1	203
e6473092	204	/**
fafaa426	205	* walk_page_range - walk page table with caller specific callbacks
e6473092	206	*
fafaa426 NH	207	* Recursively walk the page table tree of the process represented by @walk->mm
	208	* within the virtual address range [@start, @end). During walking, we can do
	209	* some caller-specific works for each entry, by setting up pmd_entry(),
	210	* pte_entry(), and/or hugetlb_entry(). If you don't set up for some of these
	211	* callbacks, the associated entries/pages are just ignored.
	212	* The return values of these callbacks are commonly defined like below:
	213	* - 0 : succeeded to handle the current entry, and if you don't reach the
	214	* end address yet, continue to walk.
	215	* - >0 : succeeded to handle the current entry, and return to the caller
	216	* with caller specific value.
	217	* - <0 : failed to handle the current entry, and return to the caller
	218	* with error code.
e6473092	219	*
fafaa426 NH	220	* Before starting to walk page table, some callers want to check whether
	221	* they really want to walk over the current vma, typically by checking
	222	* its vm_flags. walk_page_test() and @walk->test_walk() are used for this
	223	* purpose.
e6473092	224	*
fafaa426 NH	225	* struct mm_walk keeps current values of some common data like vma and pmd,
	226	* which are useful for the access from callbacks. If you want to pass some
	227	* caller-specific data to callbacks, @walk->private should be helpful.
c27fe4c8	228	*
fafaa426 NH	229	* Locking:
	230	* Callers of walk_page_range() and walk_page_vma() should hold
	231	* @walk->mm->mmap_sem, because these function traverse vma list and/or
	232	* access to vma's data.
e6473092	233	*/
fafaa426	234	int walk_page_range(unsigned long start, unsigned long end,
2165009b	235	struct mm_walk *walk)
e6473092	236	{
e6473092	237	int err = 0;
fafaa426 NH	238	unsigned long next;
fafaa426 NH	239	struct vm_area_struct *vma;
e6473092	240
fafaa426 NH	241	if (start >= end)
fafaa426 NH	242	return -EINVAL;
e6473092	243
2165009b DH	244	if (!walk->mm)
	245	return -EINVAL;
	246
96dad67f	247	VM_BUG_ON_MM(!rwsem_is_locked(&walk->mm->mmap_sem), walk->mm);
a9ff785e	248
fafaa426	249	vma = find_vma(walk->mm, start);
e6473092	250	do {
fafaa426 NH	251	if (!vma) { /* after the last vma */
	252	walk->vma = NULL;
	253	next = end;
	254	} else if (start < vma->vm_start) { /* outside vma */
	255	walk->vma = NULL;
	256	next = min(end, vma->vm_start);
	257	} else { /* inside vma */
	258	walk->vma = vma;
	259	next = min(end, vma->vm_end);
	260	vma = vma->vm_next;
5f0af70a	261
fafaa426 NH	262	err = walk_page_test(start, next, walk);
fafaa426 NH	263	if (err > 0)
a9ff785e	264	continue;
fafaa426	265	if (err < 0)
e6473092	266	break;
e6473092	267	}
fafaa426 NH	268	if (walk->vma \|\| walk->pte_hole)
fafaa426 NH	269	err = __walk_page_range(start, next, walk);
e6473092 MM	270	if (err)
e6473092 MM	271	break;
fafaa426	272	} while (start = next, start < end);
e6473092 MM	273	return err;
e6473092 MM	274	}