[mirror_ubuntu-artful-kernel.git] / arch / sparc / mm / hugetlbpage.c

/*
 * SPARC64 Huge TLB page support.
 *
 * Copyright (C) 2002, 2003, 2006 David S. Miller (davem@davemloft.net)
 */

#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/sched/mm.h>
#include <linux/hugetlb.h>
#include <linux/pagemap.h>
#include <linux/sysctl.h>

#include <asm/mman.h>
#include <asm/pgalloc.h>
#include <asm/pgtable.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
#include <asm/cacheflush.h>
#include <asm/mmu_context.h>

/* Slightly simplified from the non-hugepage variant because by
 * definition we don't have to worry about any page coloring stuff
 */

static unsigned long hugetlb_get_unmapped_area_bottomup(struct file *filp,
							unsigned long addr,
							unsigned long len,
							unsigned long pgoff,
							unsigned long flags)
{
	struct hstate *h = hstate_file(filp);
	unsigned long task_size = TASK_SIZE;
	struct vm_unmapped_area_info info;

	if (test_thread_flag(TIF_32BIT))
		task_size = STACK_TOP32;

	info.flags = 0;
	info.length = len;
	info.low_limit = TASK_UNMAPPED_BASE;
	info.high_limit = min(task_size, VA_EXCLUDE_START);
	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
	info.align_offset = 0;
	addr = vm_unmapped_area(&info);

	if ((addr & ~PAGE_MASK) && task_size > VA_EXCLUDE_END) {
		VM_BUG_ON(addr != -ENOMEM);
		info.low_limit = VA_EXCLUDE_END;
		info.high_limit = task_size;
		addr = vm_unmapped_area(&info);
	}

	return addr;
}

static unsigned long
hugetlb_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
				  const unsigned long len,
				  const unsigned long pgoff,
				  const unsigned long flags)
{
	struct hstate *h = hstate_file(filp);
	struct mm_struct *mm = current->mm;
	unsigned long addr = addr0;
	struct vm_unmapped_area_info info;

	/* This should only ever run for 32-bit processes.  */
	BUG_ON(!test_thread_flag(TIF_32BIT));

	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	info.length = len;
	info.low_limit = PAGE_SIZE;
	info.high_limit = mm->mmap_base;
	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
	info.align_offset = 0;
	addr = vm_unmapped_area(&info);

	/*
	 * A failed mmap() very likely causes application failure,
	 * so fall back to the bottom-up function here. This scenario
	 * can happen with large stack limits and large mmap()
	 * allocations.
	 */
	if (addr & ~PAGE_MASK) {
		VM_BUG_ON(addr != -ENOMEM);
		info.flags = 0;
		info.low_limit = TASK_UNMAPPED_BASE;
		info.high_limit = STACK_TOP32;
		addr = vm_unmapped_area(&info);
	}

	return addr;
}

unsigned long
hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
		unsigned long len, unsigned long pgoff, unsigned long flags)
{
	struct hstate *h = hstate_file(file);
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	unsigned long task_size = TASK_SIZE;

	if (test_thread_flag(TIF_32BIT))
		task_size = STACK_TOP32;

	if (len & ~huge_page_mask(h))
		return -EINVAL;
	if (len > task_size)
		return -ENOMEM;

	if (flags & MAP_FIXED) {
		if (prepare_hugepage_range(file, addr, len))
			return -EINVAL;
		return addr;
	}

	if (addr) {
		addr = ALIGN(addr, huge_page_size(h));
		vma = find_vma(mm, addr);
		if (task_size - len >= addr &&
		    (!vma || addr + len <= vma->vm_start))
			return addr;
	}
	if (mm->get_unmapped_area == arch_get_unmapped_area)
		return hugetlb_get_unmapped_area_bottomup(file, addr, len,
				pgoff, flags);
	else
		return hugetlb_get_unmapped_area_topdown(file, addr, len,
				pgoff, flags);
}

static pte_t sun4u_hugepage_shift_to_tte(pte_t entry, unsigned int shift)
{
	return entry;
}

static pte_t sun4v_hugepage_shift_to_tte(pte_t entry, unsigned int shift)
{
	unsigned long hugepage_size = _PAGE_SZ4MB_4V;

	pte_val(entry) = pte_val(entry) & ~_PAGE_SZALL_4V;

	switch (shift) {
	case HPAGE_256MB_SHIFT:
		hugepage_size = _PAGE_SZ256MB_4V;
		pte_val(entry) |= _PAGE_PMD_HUGE;
		break;
	case HPAGE_SHIFT:
		pte_val(entry) |= _PAGE_PMD_HUGE;
		break;
	case HPAGE_64K_SHIFT:
		hugepage_size = _PAGE_SZ64K_4V;
		break;
	default:
		WARN_ONCE(1, "unsupported hugepage shift=%u\n", shift);
	}

	pte_val(entry) = pte_val(entry) | hugepage_size;
	return entry;
}

static pte_t hugepage_shift_to_tte(pte_t entry, unsigned int shift)
{
	if (tlb_type == hypervisor)
		return sun4v_hugepage_shift_to_tte(entry, shift);
	else
		return sun4u_hugepage_shift_to_tte(entry, shift);
}

pte_t arch_make_huge_pte(pte_t entry, struct vm_area_struct *vma,
			 struct page *page, int writeable)
{
	unsigned int shift = huge_page_shift(hstate_vma(vma));

	return hugepage_shift_to_tte(entry, shift);
}

static unsigned int sun4v_huge_tte_to_shift(pte_t entry)
{
	unsigned long tte_szbits = pte_val(entry) & _PAGE_SZALL_4V;
	unsigned int shift;

	switch (tte_szbits) {
	case _PAGE_SZ256MB_4V:
		shift = HPAGE_256MB_SHIFT;
		break;
	case _PAGE_SZ4MB_4V:
		shift = REAL_HPAGE_SHIFT;
		break;
	case _PAGE_SZ64K_4V:
		shift = HPAGE_64K_SHIFT;
		break;
	default:
		shift = PAGE_SHIFT;
		break;
	}
	return shift;
}

static unsigned int sun4u_huge_tte_to_shift(pte_t entry)
{
	unsigned long tte_szbits = pte_val(entry) & _PAGE_SZALL_4U;
	unsigned int shift;

	switch (tte_szbits) {
	case _PAGE_SZ256MB_4U:
		shift = HPAGE_256MB_SHIFT;
		break;
	case _PAGE_SZ4MB_4U:
		shift = REAL_HPAGE_SHIFT;
		break;
	case _PAGE_SZ64K_4U:
		shift = HPAGE_64K_SHIFT;
		break;
	default:
		shift = PAGE_SHIFT;
		break;
	}
	return shift;
}

static unsigned int huge_tte_to_shift(pte_t entry)
{
	unsigned long shift;

	if (tlb_type == hypervisor)
		shift = sun4v_huge_tte_to_shift(entry);
	else
		shift = sun4u_huge_tte_to_shift(entry);

	if (shift == PAGE_SHIFT)
		WARN_ONCE(1, "tto_to_shift: invalid hugepage tte=0x%lx\n",
			  pte_val(entry));

	return shift;
}

static unsigned long huge_tte_to_size(pte_t pte)
{
	unsigned long size = 1UL << huge_tte_to_shift(pte);

	if (size == REAL_HPAGE_SIZE)
		size = HPAGE_SIZE;
	return size;
}

pte_t *huge_pte_alloc(struct mm_struct *mm,
			unsigned long addr, unsigned long sz)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte = NULL;

	pgd = pgd_offset(mm, addr);
	pud = pud_alloc(mm, pgd, addr);
	if (pud) {
		pmd = pmd_alloc(mm, pud, addr);
		if (!pmd)
			return NULL;

		if (sz == PMD_SHIFT)
			pte = (pte_t *)pmd;
		else
			pte = pte_alloc_map(mm, pmd, addr);
	}

	return pte;
}

pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte = NULL;

	pgd = pgd_offset(mm, addr);
	if (!pgd_none(*pgd)) {
		pud = pud_offset(pgd, addr);
		if (!pud_none(*pud)) {
			pmd = pmd_offset(pud, addr);
			if (!pmd_none(*pmd)) {
				if (is_hugetlb_pmd(*pmd))
					pte = (pte_t *)pmd;
				else
					pte = pte_offset_map(pmd, addr);
			}
		}
	}

	return pte;
}

void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
		     pte_t *ptep, pte_t entry)
{
	unsigned int i, nptes, orig_shift, shift;
	unsigned long size;
	pte_t orig;

	size = huge_tte_to_size(entry);
	shift = size >= HPAGE_SIZE ? PMD_SHIFT : PAGE_SHIFT;
	nptes = size >> shift;

	if (!pte_present(*ptep) && pte_present(entry))
		mm->context.hugetlb_pte_count += nptes;

	addr &= ~(size - 1);
	orig = *ptep;
	orig_shift = pte_none(orig) ? PAGE_SHIFT : huge_tte_to_shift(orig);

	for (i = 0; i < nptes; i++)
		ptep[i] = __pte(pte_val(entry) + (i << shift));

	maybe_tlb_batch_add(mm, addr, ptep, orig, 0, orig_shift);
	/* An HPAGE_SIZE'ed page is composed of two REAL_HPAGE_SIZE'ed pages */
	if (size == HPAGE_SIZE)
		maybe_tlb_batch_add(mm, addr + REAL_HPAGE_SIZE, ptep, orig, 0,
				    orig_shift);
}

pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr,
			      pte_t *ptep)
{
	unsigned int i, nptes, hugepage_shift;
	unsigned long size;
	pte_t entry;

	entry = *ptep;
	size = huge_tte_to_size(entry);
	if (size >= HPAGE_SIZE)
		nptes = size >> PMD_SHIFT;
	else
		nptes = size >> PAGE_SHIFT;

	hugepage_shift = pte_none(entry) ? PAGE_SHIFT :
		huge_tte_to_shift(entry);

	if (pte_present(entry))
		mm->context.hugetlb_pte_count -= nptes;

	addr &= ~(size - 1);
	for (i = 0; i < nptes; i++)
		ptep[i] = __pte(0UL);

	maybe_tlb_batch_add(mm, addr, ptep, entry, 0, hugepage_shift);
	/* An HPAGE_SIZE'ed page is composed of two REAL_HPAGE_SIZE'ed pages */
	if (size == HPAGE_SIZE)
		maybe_tlb_batch_add(mm, addr + REAL_HPAGE_SIZE, ptep, entry, 0,
				    hugepage_shift);

	return entry;
}

int pmd_huge(pmd_t pmd)
{
	return !pmd_none(pmd) &&
		(pmd_val(pmd) & (_PAGE_VALID|_PAGE_PMD_HUGE)) != _PAGE_VALID;
}

int pud_huge(pud_t pud)
{
	return 0;
}

static void hugetlb_free_pte_range(struct mmu_gather *tlb, pmd_t *pmd,
			   unsigned long addr)
{
	pgtable_t token = pmd_pgtable(*pmd);

	pmd_clear(pmd);
	pte_free_tlb(tlb, token, addr);
	atomic_long_dec(&tlb->mm->nr_ptes);
}

static void hugetlb_free_pmd_range(struct mmu_gather *tlb, pud_t *pud,
				   unsigned long addr, unsigned long end,
				   unsigned long floor, unsigned long ceiling)
{
	pmd_t *pmd;
	unsigned long next;
	unsigned long start;

	start = addr;
	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none(*pmd))
			continue;
		if (is_hugetlb_pmd(*pmd))
			pmd_clear(pmd);
		else
			hugetlb_free_pte_range(tlb, pmd, addr);
	} while (pmd++, addr = next, addr != end);

	start &= PUD_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PUD_MASK;
		if (!ceiling)
			return;
	}
	if (end - 1 > ceiling - 1)
		return;

	pmd = pmd_offset(pud, start);
	pud_clear(pud);
	pmd_free_tlb(tlb, pmd, start);
	mm_dec_nr_pmds(tlb->mm);
}

static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd,
				   unsigned long addr, unsigned long end,
				   unsigned long floor, unsigned long ceiling)
{
	pud_t *pud;
	unsigned long next;
	unsigned long start;

	start = addr;
	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
		hugetlb_free_pmd_range(tlb, pud, addr, next, floor,
				       ceiling);
	} while (pud++, addr = next, addr != end);

	start &= PGDIR_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PGDIR_MASK;
		if (!ceiling)
			return;
	}
	if (end - 1 > ceiling - 1)
		return;

	pud = pud_offset(pgd, start);
	pgd_clear(pgd);
	pud_free_tlb(tlb, pud, start);
}

void hugetlb_free_pgd_range(struct mmu_gather *tlb,
			    unsigned long addr, unsigned long end,
			    unsigned long floor, unsigned long ceiling)
{
	pgd_t *pgd;
	unsigned long next;

	pgd = pgd_offset(tlb->mm, addr);
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
		hugetlb_free_pud_range(tlb, pgd, addr, next, floor, ceiling);
	} while (pgd++, addr = next, addr != end);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* SPARC64 Huge TLB page support.
	3	*
f6b83f07	4	* Copyright (C) 2002, 2003, 2006 David S. Miller (davem@davemloft.net)
1da177e4 LT	5	*/
1da177e4 LT	6
1da177e4 LT	7	#include <linux/fs.h>
1da177e4 LT	8	#include <linux/mm.h>
01042607	9	#include <linux/sched/mm.h>
1da177e4 LT	10	#include <linux/hugetlb.h>
1da177e4 LT	11	#include <linux/pagemap.h>
1da177e4 LT	12	#include <linux/sysctl.h>
	13
	14	#include <asm/mman.h>
	15	#include <asm/pgalloc.h>
7bc3777c	16	#include <asm/pgtable.h>
1da177e4 LT	17	#include <asm/tlb.h>
	18	#include <asm/tlbflush.h>
	19	#include <asm/cacheflush.h>
	20	#include <asm/mmu_context.h>
	21
f6b83f07 DM	22	/* Slightly simplified from the non-hugepage variant because by
	23	* definition we don't have to worry about any page coloring stuff
	24	*/
f6b83f07 DM	25
	26	static unsigned long hugetlb_get_unmapped_area_bottomup(struct file *filp,
	27	unsigned long addr,
	28	unsigned long len,
	29	unsigned long pgoff,
	30	unsigned long flags)
	31	{
c7d9f77d	32	struct hstate *h = hstate_file(filp);
f6b83f07	33	unsigned long task_size = TASK_SIZE;
2aea28b9	34	struct vm_unmapped_area_info info;
f6b83f07 DM	35
	36	if (test_thread_flag(TIF_32BIT))
	37	task_size = STACK_TOP32;
f6b83f07	38
2aea28b9 ML	39	info.flags = 0;
	40	info.length = len;
	41	info.low_limit = TASK_UNMAPPED_BASE;
	42	info.high_limit = min(task_size, VA_EXCLUDE_START);
c7d9f77d	43	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
2aea28b9 ML	44	info.align_offset = 0;
	45	addr = vm_unmapped_area(&info);
	46
	47	if ((addr & ~PAGE_MASK) && task_size > VA_EXCLUDE_END) {
	48	VM_BUG_ON(addr != -ENOMEM);
	49	info.low_limit = VA_EXCLUDE_END;
	50	info.high_limit = task_size;
	51	addr = vm_unmapped_area(&info);
f6b83f07 DM	52	}
f6b83f07 DM	53
2aea28b9	54	return addr;
f6b83f07 DM	55	}
	56
	57	static unsigned long
	58	hugetlb_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
	59	const unsigned long len,
	60	const unsigned long pgoff,
	61	const unsigned long flags)
	62	{
c7d9f77d	63	struct hstate *h = hstate_file(filp);
f6b83f07 DM	64	struct mm_struct *mm = current->mm;
f6b83f07 DM	65	unsigned long addr = addr0;
2aea28b9	66	struct vm_unmapped_area_info info;
f6b83f07 DM	67
	68	/* This should only ever run for 32-bit processes. */
	69	BUG_ON(!test_thread_flag(TIF_32BIT));
	70
2aea28b9 ML	71	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	72	info.length = len;
	73	info.low_limit = PAGE_SIZE;
	74	info.high_limit = mm->mmap_base;
c7d9f77d	75	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
2aea28b9 ML	76	info.align_offset = 0;
2aea28b9 ML	77	addr = vm_unmapped_area(&info);
f6b83f07	78
f6b83f07 DM	79	/*
	80	* A failed mmap() very likely causes application failure,
	81	* so fall back to the bottom-up function here. This scenario
	82	* can happen with large stack limits and large mmap()
	83	* allocations.
	84	*/
2aea28b9 ML	85	if (addr & ~PAGE_MASK) {
	86	VM_BUG_ON(addr != -ENOMEM);
	87	info.flags = 0;
	88	info.low_limit = TASK_UNMAPPED_BASE;
	89	info.high_limit = STACK_TOP32;
	90	addr = vm_unmapped_area(&info);
	91	}
f6b83f07 DM	92
	93	return addr;
	94	}
	95
	96	unsigned long
	97	hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
	98	unsigned long len, unsigned long pgoff, unsigned long flags)
	99	{
c7d9f77d	100	struct hstate *h = hstate_file(file);
f6b83f07 DM	101	struct mm_struct *mm = current->mm;
	102	struct vm_area_struct *vma;
	103	unsigned long task_size = TASK_SIZE;
	104
	105	if (test_thread_flag(TIF_32BIT))
	106	task_size = STACK_TOP32;
	107
c7d9f77d	108	if (len & ~huge_page_mask(h))
f6b83f07 DM	109	return -EINVAL;
	110	if (len > task_size)
	111	return -ENOMEM;
	112
ac35ee48	113	if (flags & MAP_FIXED) {
a5516438	114	if (prepare_hugepage_range(file, addr, len))
ac35ee48 BH	115	return -EINVAL;
	116	return addr;
	117	}
	118
f6b83f07	119	if (addr) {
c7d9f77d	120	addr = ALIGN(addr, huge_page_size(h));
f6b83f07 DM	121	vma = find_vma(mm, addr);
	122	if (task_size - len >= addr &&
	123	(!vma \|\| addr + len <= vma->vm_start))
	124	return addr;
	125	}
	126	if (mm->get_unmapped_area == arch_get_unmapped_area)
	127	return hugetlb_get_unmapped_area_bottomup(file, addr, len,
	128	pgoff, flags);
	129	else
	130	return hugetlb_get_unmapped_area_topdown(file, addr, len,
	131	pgoff, flags);
	132	}
	133
c7d9f77d NG	134	static pte_t sun4u_hugepage_shift_to_tte(pte_t entry, unsigned int shift)
	135	{
	136	return entry;
	137	}
	138
	139	static pte_t sun4v_hugepage_shift_to_tte(pte_t entry, unsigned int shift)
	140	{
	141	unsigned long hugepage_size = _PAGE_SZ4MB_4V;
	142
	143	pte_val(entry) = pte_val(entry) & ~_PAGE_SZALL_4V;
	144
	145	switch (shift) {
	146	case HPAGE_256MB_SHIFT:
	147	hugepage_size = _PAGE_SZ256MB_4V;
	148	pte_val(entry) \|= _PAGE_PMD_HUGE;
	149	break;
	150	case HPAGE_SHIFT:
	151	pte_val(entry) \|= _PAGE_PMD_HUGE;
	152	break;
dcd1912d NG	153	case HPAGE_64K_SHIFT:
	154	hugepage_size = _PAGE_SZ64K_4V;
	155	break;
c7d9f77d NG	156	default:
	157	WARN_ONCE(1, "unsupported hugepage shift=%u\n", shift);
	158	}
	159
	160	pte_val(entry) = pte_val(entry) \| hugepage_size;
	161	return entry;
	162	}
	163
	164	static pte_t hugepage_shift_to_tte(pte_t entry, unsigned int shift)
	165	{
	166	if (tlb_type == hypervisor)
	167	return sun4v_hugepage_shift_to_tte(entry, shift);
	168	else
	169	return sun4u_hugepage_shift_to_tte(entry, shift);
	170	}
	171
	172	pte_t arch_make_huge_pte(pte_t entry, struct vm_area_struct *vma,
	173	struct page *page, int writeable)
	174	{
	175	unsigned int shift = huge_page_shift(hstate_vma(vma));
	176
	177	return hugepage_shift_to_tte(entry, shift);
	178	}
	179
	180	static unsigned int sun4v_huge_tte_to_shift(pte_t entry)
	181	{
	182	unsigned long tte_szbits = pte_val(entry) & _PAGE_SZALL_4V;
	183	unsigned int shift;
	184
	185	switch (tte_szbits) {
	186	case _PAGE_SZ256MB_4V:
	187	shift = HPAGE_256MB_SHIFT;
	188	break;
	189	case _PAGE_SZ4MB_4V:
	190	shift = REAL_HPAGE_SHIFT;
	191	break;
dcd1912d NG	192	case _PAGE_SZ64K_4V:
	193	shift = HPAGE_64K_SHIFT;
	194	break;
c7d9f77d NG	195	default:
	196	shift = PAGE_SHIFT;
	197	break;
	198	}
	199	return shift;
	200	}
	201
	202	static unsigned int sun4u_huge_tte_to_shift(pte_t entry)
	203	{
	204	unsigned long tte_szbits = pte_val(entry) & _PAGE_SZALL_4U;
	205	unsigned int shift;
	206
	207	switch (tte_szbits) {
	208	case _PAGE_SZ256MB_4U:
	209	shift = HPAGE_256MB_SHIFT;
	210	break;
	211	case _PAGE_SZ4MB_4U:
	212	shift = REAL_HPAGE_SHIFT;
	213	break;
dcd1912d NG	214	case _PAGE_SZ64K_4U:
	215	shift = HPAGE_64K_SHIFT;
	216	break;
c7d9f77d NG	217	default:
	218	shift = PAGE_SHIFT;
	219	break;
	220	}
	221	return shift;
	222	}
	223
	224	static unsigned int huge_tte_to_shift(pte_t entry)
	225	{
	226	unsigned long shift;
	227
	228	if (tlb_type == hypervisor)
	229	shift = sun4v_huge_tte_to_shift(entry);
	230	else
	231	shift = sun4u_huge_tte_to_shift(entry);
	232
	233	if (shift == PAGE_SHIFT)
	234	WARN_ONCE(1, "tto_to_shift: invalid hugepage tte=0x%lx\n",
	235	pte_val(entry));
	236
	237	return shift;
	238	}
	239
	240	static unsigned long huge_tte_to_size(pte_t pte)
	241	{
	242	unsigned long size = 1UL << huge_tte_to_shift(pte);
	243
	244	if (size == REAL_HPAGE_SIZE)
	245	size = HPAGE_SIZE;
	246	return size;
	247	}
	248
a5516438 AK	249	pte_t huge_pte_alloc(struct mm_struct mm,
a5516438 AK	250	unsigned long addr, unsigned long sz)
1da177e4 LT	251	{
	252	pgd_t *pgd;
	253	pud_t *pud;
dcd1912d	254	pmd_t *pmd;
1da177e4 LT	255	pte_t *pte = NULL;
	256
	257	pgd = pgd_offset(mm, addr);
dcc1e8dd	258	pud = pud_alloc(mm, pgd, addr);
dcd1912d NG	259	if (pud) {
	260	pmd = pmd_alloc(mm, pud, addr);
	261	if (!pmd)
	262	return NULL;
	263
	264	if (sz == PMD_SHIFT)
	265	pte = (pte_t *)pmd;
	266	else
	267	pte = pte_alloc_map(mm, pmd, addr);
	268	}
7bc3777c	269
1da177e4 LT	270	return pte;
	271	}
	272
63551ae0	273	pte_t huge_pte_offset(struct mm_struct mm, unsigned long addr)
1da177e4 LT	274	{
	275	pgd_t *pgd;
	276	pud_t *pud;
dcd1912d	277	pmd_t *pmd;
1da177e4 LT	278	pte_t *pte = NULL;
	279
	280	pgd = pgd_offset(mm, addr);
f6b83f07	281	if (!pgd_none(*pgd)) {
1da177e4	282	pud = pud_offset(pgd, addr);
dcd1912d NG	283	if (!pud_none(*pud)) {
	284	pmd = pmd_offset(pud, addr);
	285	if (!pmd_none(*pmd)) {
	286	if (is_hugetlb_pmd(*pmd))
	287	pte = (pte_t *)pmd;
	288	else
	289	pte = pte_offset_map(pmd, addr);
	290	}
	291	}
1da177e4	292	}
dcd1912d	293
1da177e4 LT	294	return pte;
	295	}
	296
63551ae0 DG	297	void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
63551ae0 DG	298	pte_t *ptep, pte_t entry)
1da177e4	299	{
dcd1912d	300	unsigned int i, nptes, orig_shift, shift;
c7d9f77d	301	unsigned long size;
7bc3777c	302	pte_t orig;
63551ae0	303
c7d9f77d	304	size = huge_tte_to_size(entry);
dcd1912d NG	305	shift = size >= HPAGE_SIZE ? PMD_SHIFT : PAGE_SHIFT;
dcd1912d NG	306	nptes = size >> shift;
c7d9f77d	307
dcc1e8dd	308	if (!pte_present(*ptep) && pte_present(entry))
c7d9f77d	309	mm->context.hugetlb_pte_count += nptes;
dcc1e8dd	310
c7d9f77d	311	addr &= ~(size - 1);
7bc3777c	312	orig = *ptep;
ac65e282	313	orig_shift = pte_none(orig) ? PAGE_SHIFT : huge_tte_to_shift(orig);
24e49ee3	314
c7d9f77d	315	for (i = 0; i < nptes; i++)
dcd1912d	316	ptep[i] = __pte(pte_val(entry) + (i << shift));
c7d9f77d	317
dcd1912d	318	maybe_tlb_batch_add(mm, addr, ptep, orig, 0, orig_shift);
c7d9f77d NG	319	/* An HPAGE_SIZE'ed page is composed of two REAL_HPAGE_SIZE'ed pages */
	320	if (size == HPAGE_SIZE)
	321	maybe_tlb_batch_add(mm, addr + REAL_HPAGE_SIZE, ptep, orig, 0,
dcd1912d	322	orig_shift);
63551ae0	323	}
1da177e4	324
63551ae0 DG	325	pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr,
	326	pte_t *ptep)
	327	{
c7d9f77d NG	328	unsigned int i, nptes, hugepage_shift;
c7d9f77d NG	329	unsigned long size;
63551ae0	330	pte_t entry;
1da177e4	331
63551ae0	332	entry = *ptep;
c7d9f77d	333	size = huge_tte_to_size(entry);
dcd1912d NG	334	if (size >= HPAGE_SIZE)
	335	nptes = size >> PMD_SHIFT;
	336	else
	337	nptes = size >> PAGE_SHIFT;
	338
ac65e282 NG	339	hugepage_shift = pte_none(entry) ? PAGE_SHIFT :
ac65e282 NG	340	huge_tte_to_shift(entry);
c7d9f77d	341
dcc1e8dd	342	if (pte_present(entry))
c7d9f77d	343	mm->context.hugetlb_pte_count -= nptes;
1da177e4	344
c7d9f77d NG	345	addr &= ~(size - 1);
	346	for (i = 0; i < nptes; i++)
	347	ptep[i] = __pte(0UL);
63551ae0	348
c7d9f77d NG	349	maybe_tlb_batch_add(mm, addr, ptep, entry, 0, hugepage_shift);
	350	/* An HPAGE_SIZE'ed page is composed of two REAL_HPAGE_SIZE'ed pages */
	351	if (size == HPAGE_SIZE)
	352	maybe_tlb_batch_add(mm, addr + REAL_HPAGE_SIZE, ptep, entry, 0,
	353	hugepage_shift);
24e49ee3	354
63551ae0	355	return entry;
1da177e4 LT	356	}
1da177e4 LT	357
1da177e4 LT	358	int pmd_huge(pmd_t pmd)
1da177e4 LT	359	{
7bc3777c NG	360	return !pmd_none(pmd) &&
7bc3777c NG	361	(pmd_val(pmd) & (_PAGE_VALID\|_PAGE_PMD_HUGE)) != _PAGE_VALID;
1da177e4 LT	362	}
1da177e4 LT	363
ceb86879 AK	364	int pud_huge(pud_t pud)
	365	{
	366	return 0;
	367	}
7bc3777c NG	368
	369	static void hugetlb_free_pte_range(struct mmu_gather tlb, pmd_t pmd,
	370	unsigned long addr)
	371	{
	372	pgtable_t token = pmd_pgtable(*pmd);
	373
	374	pmd_clear(pmd);
	375	pte_free_tlb(tlb, token, addr);
	376	atomic_long_dec(&tlb->mm->nr_ptes);
	377	}
	378
	379	static void hugetlb_free_pmd_range(struct mmu_gather tlb, pud_t pud,
	380	unsigned long addr, unsigned long end,
	381	unsigned long floor, unsigned long ceiling)
	382	{
	383	pmd_t *pmd;
	384	unsigned long next;
	385	unsigned long start;
	386
	387	start = addr;
	388	pmd = pmd_offset(pud, addr);
	389	do {
	390	next = pmd_addr_end(addr, end);
	391	if (pmd_none(*pmd))
	392	continue;
	393	if (is_hugetlb_pmd(*pmd))
	394	pmd_clear(pmd);
	395	else
	396	hugetlb_free_pte_range(tlb, pmd, addr);
	397	} while (pmd++, addr = next, addr != end);
	398
	399	start &= PUD_MASK;
	400	if (start < floor)
	401	return;
	402	if (ceiling) {
	403	ceiling &= PUD_MASK;
	404	if (!ceiling)
	405	return;
	406	}
	407	if (end - 1 > ceiling - 1)
	408	return;
	409
	410	pmd = pmd_offset(pud, start);
	411	pud_clear(pud);
	412	pmd_free_tlb(tlb, pmd, start);
	413	mm_dec_nr_pmds(tlb->mm);
	414	}
	415
	416	static void hugetlb_free_pud_range(struct mmu_gather tlb, pgd_t pgd,
	417	unsigned long addr, unsigned long end,
	418	unsigned long floor, unsigned long ceiling)
	419	{
	420	pud_t *pud;
	421	unsigned long next;
	422	unsigned long start;
	423
	424	start = addr;
	425	pud = pud_offset(pgd, addr);
	426	do {
	427	next = pud_addr_end(addr, end);
	428	if (pud_none_or_clear_bad(pud))
	429	continue;
	430	hugetlb_free_pmd_range(tlb, pud, addr, next, floor,
	431	ceiling);
432	} while (pud++, addr = next, addr != end);
433
434	start &= PGDIR_MASK;
435	if (start < floor)
436	return;
437	if (ceiling) {
438	ceiling &= PGDIR_MASK;
439	if (!ceiling)
440	return;
441	}
442	if (end - 1 > ceiling - 1)
443	return;
444
445	pud = pud_offset(pgd, start);
446	pgd_clear(pgd);
447	pud_free_tlb(tlb, pud, start);
448	}
449
450	void hugetlb_free_pgd_range(struct mmu_gather *tlb,
451	unsigned long addr, unsigned long end,
452	unsigned long floor, unsigned long ceiling)
453	{
454	pgd_t *pgd;
455	unsigned long next;
456
457	pgd = pgd_offset(tlb->mm, addr);
458	do {
459	next = pgd_addr_end(addr, end);
460	if (pgd_none_or_clear_bad(pgd))
461	continue;
462	hugetlb_free_pud_range(tlb, pgd, addr, next, floor, ceiling);
463	} while (pgd++, addr = next, addr != end);
464	}