[mirror_ubuntu-artful-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) || !walk->vma) {
			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_pmd(walk->vma, pmd, addr);
		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(p4d_t *p4d, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	pud_t *pud;
	unsigned long next;
	int err = 0;

	pud = pud_offset(p4d, addr);
	do {
 again:
		next = pud_addr_end(addr, end);
		if (pud_none(*pud) || !walk->vma) {
			if (walk->pte_hole)
				err = walk->pte_hole(addr, next, walk);
			if (err)
				break;
			continue;
		}

		if (walk->pud_entry) {
			spinlock_t *ptl = pud_trans_huge_lock(pud, walk->vma);

			if (ptl) {
				err = walk->pud_entry(pud, addr, next, walk);
				spin_unlock(ptl);
				if (err)
					break;
				continue;
			}
		}

		split_huge_pud(walk->vma, pud, addr);
		if (pud_none(*pud))
			goto again;

		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_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	p4d_t *p4d;
	unsigned long next;
	int err = 0;

	p4d = p4d_offset(pgd, addr);
	do {
		next = p4d_addr_end(addr, end);
		if (p4d_none_or_clear_bad(p4d)) {
			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(p4d, addr, next, walk);
		if (err)
			break;
	} while (p4d++, 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_p4d_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);
	unsigned long sz = huge_page_size(h);
	pte_t *pte;
	int err = 0;

	do {
		next = hugetlb_entry_end(h, addr, end);
		pte = huge_pte_offset(walk->mm, addr & hmask, sz);
		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.
 */
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);

	/*
	 * vma(VM_PFNMAP) doesn't have any valid struct pages behind VM_PFNMAP
	 * range, so we don't walk over it as we do for normal vmas. However,
	 * Some callers are interested in handling hole range and they don't
	 * want to just ignore any single address range. Such users certainly
	 * define their ->pte_hole() callbacks, so let's delegate them to handle
	 * vma(VM_PFNMAP).
	 */
	if (vma->vm_flags & VM_PFNMAP) {
		int err = 1;
		if (walk->pte_hole)
			err = walk->pte_hole(start, end, walk);
		return err ? err : 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) {
				/*
				 * positive return values are purely for
				 * controlling the pagewalk, so should never
				 * be passed to the callers.
				 */
				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;
}

int walk_page_vma(struct vm_area_struct *vma, struct mm_walk *walk)
{
	int err;

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

	VM_BUG_ON(!rwsem_is_locked(&walk->mm->mmap_sem));
	VM_BUG_ON(!vma);
	walk->vma = vma;
	err = walk_page_test(vma->vm_start, vma->vm_end, walk);
	if (err > 0)
		return 0;
	if (err < 0)
		return err;
	return __walk_page_range(vma->vm_start, vma->vm_end, walk);
}
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);
48684a65	38	if (pmd_none(*pmd) \|\| !walk->vma) {
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
78ddc534	61	split_huge_pmd(walk->vma, pmd, addr);
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
c2febafc	72	static int walk_pud_range(p4d_t *p4d, 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
c2febafc	79	pud = pud_offset(p4d, addr);
e6473092	80	do {
a00cc7d9	81	again:
e6473092	82	next = pud_addr_end(addr, end);
a00cc7d9	83	if (pud_none(*pud) \|\| !walk->vma) {
e6473092	84	if (walk->pte_hole)
2165009b	85	err = walk->pte_hole(addr, next, walk);
e6473092 MM	86	if (err)
	87	break;
	88	continue;
	89	}
a00cc7d9 MW	90
	91	if (walk->pud_entry) {
	92	spinlock_t *ptl = pud_trans_huge_lock(pud, walk->vma);
	93
	94	if (ptl) {
	95	err = walk->pud_entry(pud, addr, next, walk);
	96	spin_unlock(ptl);
	97	if (err)
	98	break;
	99	continue;
	100	}
	101	}
	102
	103	split_huge_pud(walk->vma, pud, addr);
	104	if (pud_none(*pud))
	105	goto again;
	106
0b1fbfe5	107	if (walk->pmd_entry \|\| walk->pte_entry)
2165009b	108	err = walk_pmd_range(pud, addr, next, walk);
e6473092 MM	109	if (err)
	110	break;
	111	} while (pud++, addr = next, addr != end);
	112
	113	return err;
	114	}
	115
c2febafc KS	116	static int walk_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end,
	117	struct mm_walk *walk)
	118	{
	119	p4d_t *p4d;
	120	unsigned long next;
	121	int err = 0;
	122
	123	p4d = p4d_offset(pgd, addr);
	124	do {
	125	next = p4d_addr_end(addr, end);
	126	if (p4d_none_or_clear_bad(p4d)) {
	127	if (walk->pte_hole)
	128	err = walk->pte_hole(addr, next, walk);
	129	if (err)
	130	break;
	131	continue;
	132	}
	133	if (walk->pmd_entry \|\| walk->pte_entry)
	134	err = walk_pud_range(p4d, addr, next, walk);
	135	if (err)
	136	break;
	137	} while (p4d++, addr = next, addr != end);
	138
	139	return err;
	140	}
	141
fafaa426 NH	142	static int walk_pgd_range(unsigned long addr, unsigned long end,
	143	struct mm_walk *walk)
	144	{
	145	pgd_t *pgd;
	146	unsigned long next;
	147	int err = 0;
	148
	149	pgd = pgd_offset(walk->mm, addr);
	150	do {
	151	next = pgd_addr_end(addr, end);
	152	if (pgd_none_or_clear_bad(pgd)) {
	153	if (walk->pte_hole)
	154	err = walk->pte_hole(addr, next, walk);
	155	if (err)
	156	break;
	157	continue;
	158	}
	159	if (walk->pmd_entry \|\| walk->pte_entry)
c2febafc	160	err = walk_p4d_range(pgd, addr, next, walk);
fafaa426 NH	161	if (err)
	162	break;
	163	} while (pgd++, addr = next, addr != end);
	164
	165	return err;
	166	}
	167
116354d1 NH	168	#ifdef CONFIG_HUGETLB_PAGE
	169	static unsigned long hugetlb_entry_end(struct hstate *h, unsigned long addr,
	170	unsigned long end)
	171	{
	172	unsigned long boundary = (addr & huge_page_mask(h)) + huge_page_size(h);
	173	return boundary < end ? boundary : end;
	174	}
	175
fafaa426	176	static int walk_hugetlb_range(unsigned long addr, unsigned long end,
116354d1 NH	177	struct mm_walk *walk)
116354d1 NH	178	{
fafaa426	179	struct vm_area_struct *vma = walk->vma;
116354d1 NH	180	struct hstate *h = hstate_vma(vma);
	181	unsigned long next;
	182	unsigned long hmask = huge_page_mask(h);
7868a208	183	unsigned long sz = huge_page_size(h);
116354d1 NH	184	pte_t *pte;
	185	int err = 0;
	186
	187	do {
	188	next = hugetlb_entry_end(h, addr, end);
7868a208	189	pte = huge_pte_offset(walk->mm, addr & hmask, sz);
116354d1 NH	190	if (pte && walk->hugetlb_entry)
	191	err = walk->hugetlb_entry(pte, hmask, addr, next, walk);
	192	if (err)
fafaa426	193	break;
116354d1 NH	194	} while (addr = next, addr != end);
116354d1 NH	195
fafaa426	196	return err;
116354d1	197	}
6c6d5280	198
6c6d5280	199	#else /* CONFIG_HUGETLB_PAGE */
fafaa426	200	static int walk_hugetlb_range(unsigned long addr, unsigned long end,
6c6d5280 KM	201	struct mm_walk *walk)
	202	{
	203	return 0;
	204	}
	205
	206	#endif /* CONFIG_HUGETLB_PAGE */
	207
fafaa426 NH	208	/*
	209	* Decide whether we really walk over the current vma on [@start, @end)
	210	* or skip it via the returned value. Return 0 if we do walk over the
	211	* current vma, and return 1 if we skip the vma. Negative values means
	212	* error, where we abort the current walk.
fafaa426 NH	213	*/
	214	static int walk_page_test(unsigned long start, unsigned long end,
	215	struct mm_walk *walk)
	216	{
	217	struct vm_area_struct *vma = walk->vma;
6c6d5280	218
fafaa426 NH	219	if (walk->test_walk)
	220	return walk->test_walk(start, end, walk);
	221
	222	/*
48684a65 NH	223	* vma(VM_PFNMAP) doesn't have any valid struct pages behind VM_PFNMAP
	224	* range, so we don't walk over it as we do for normal vmas. However,
	225	* Some callers are interested in handling hole range and they don't
	226	* want to just ignore any single address range. Such users certainly
	227	* define their ->pte_hole() callbacks, so let's delegate them to handle
	228	* vma(VM_PFNMAP).
fafaa426	229	*/
48684a65 NH	230	if (vma->vm_flags & VM_PFNMAP) {
	231	int err = 1;
	232	if (walk->pte_hole)
	233	err = walk->pte_hole(start, end, walk);
	234	return err ? err : 1;
	235	}
fafaa426 NH	236	return 0;
	237	}
	238
	239	static int __walk_page_range(unsigned long start, unsigned long end,
	240	struct mm_walk *walk)
	241	{
	242	int err = 0;
	243	struct vm_area_struct *vma = walk->vma;
	244
	245	if (vma && is_vm_hugetlb_page(vma)) {
	246	if (walk->hugetlb_entry)
	247	err = walk_hugetlb_range(start, end, walk);
	248	} else
	249	err = walk_pgd_range(start, end, walk);
	250
	251	return err;
	252	}
116354d1	253
e6473092	254	/**
fafaa426	255	* walk_page_range - walk page table with caller specific callbacks
e6473092	256	*
fafaa426 NH	257	* Recursively walk the page table tree of the process represented by @walk->mm
	258	* within the virtual address range [@start, @end). During walking, we can do
	259	* some caller-specific works for each entry, by setting up pmd_entry(),
	260	* pte_entry(), and/or hugetlb_entry(). If you don't set up for some of these
	261	* callbacks, the associated entries/pages are just ignored.
	262	* The return values of these callbacks are commonly defined like below:
	263	* - 0 : succeeded to handle the current entry, and if you don't reach the
	264	* end address yet, continue to walk.
	265	* - >0 : succeeded to handle the current entry, and return to the caller
	266	* with caller specific value.
	267	* - <0 : failed to handle the current entry, and return to the caller
	268	* with error code.
e6473092	269	*
fafaa426 NH	270	* Before starting to walk page table, some callers want to check whether
	271	* they really want to walk over the current vma, typically by checking
	272	* its vm_flags. walk_page_test() and @walk->test_walk() are used for this
	273	* purpose.
e6473092	274	*
fafaa426 NH	275	* struct mm_walk keeps current values of some common data like vma and pmd,
	276	* which are useful for the access from callbacks. If you want to pass some
	277	* caller-specific data to callbacks, @walk->private should be helpful.
c27fe4c8	278	*
fafaa426 NH	279	* Locking:
	280	* Callers of walk_page_range() and walk_page_vma() should hold
	281	* @walk->mm->mmap_sem, because these function traverse vma list and/or
	282	* access to vma's data.
e6473092	283	*/
fafaa426	284	int walk_page_range(unsigned long start, unsigned long end,
2165009b	285	struct mm_walk *walk)
e6473092	286	{
e6473092	287	int err = 0;
fafaa426 NH	288	unsigned long next;
fafaa426 NH	289	struct vm_area_struct *vma;
e6473092	290
fafaa426 NH	291	if (start >= end)
fafaa426 NH	292	return -EINVAL;
e6473092	293
2165009b DH	294	if (!walk->mm)
	295	return -EINVAL;
	296
96dad67f	297	VM_BUG_ON_MM(!rwsem_is_locked(&walk->mm->mmap_sem), walk->mm);
a9ff785e	298
fafaa426	299	vma = find_vma(walk->mm, start);
e6473092	300	do {
fafaa426 NH	301	if (!vma) { /* after the last vma */
	302	walk->vma = NULL;
	303	next = end;
	304	} else if (start < vma->vm_start) { /* outside vma */
	305	walk->vma = NULL;
	306	next = min(end, vma->vm_start);
	307	} else { /* inside vma */
	308	walk->vma = vma;
	309	next = min(end, vma->vm_end);
	310	vma = vma->vm_next;
5f0af70a	311
fafaa426	312	err = walk_page_test(start, next, walk);
f6837395 NH	313	if (err > 0) {
	314	/*
	315	* positive return values are purely for
	316	* controlling the pagewalk, so should never
	317	* be passed to the callers.
	318	*/
	319	err = 0;
a9ff785e	320	continue;
f6837395	321	}
fafaa426	322	if (err < 0)
e6473092	323	break;
e6473092	324	}
fafaa426 NH	325	if (walk->vma \|\| walk->pte_hole)
fafaa426 NH	326	err = __walk_page_range(start, next, walk);
e6473092 MM	327	if (err)
e6473092 MM	328	break;
fafaa426	329	} while (start = next, start < end);
e6473092 MM	330	return err;
e6473092 MM	331	}
900fc5f1 NH	332
	333	int walk_page_vma(struct vm_area_struct vma, struct mm_walk walk)
	334	{
	335	int err;
	336
	337	if (!walk->mm)
	338	return -EINVAL;
	339
	340	VM_BUG_ON(!rwsem_is_locked(&walk->mm->mmap_sem));
	341	VM_BUG_ON(!vma);
	342	walk->vma = vma;
	343	err = walk_page_test(vma->vm_start, vma->vm_end, walk);
	344	if (err > 0)
	345	return 0;
	346	if (err < 0)
	347	return err;
	348	return __walk_page_range(vma->vm_start, vma->vm_end, walk);
	349	}