[linux-2.6-block.git] / arch / x86 / mm / hugetlbpage.c

/*
 * IA-32 Huge TLB Page Support for Kernel.
 *
 * Copyright (C) 2002, Rohit Seth <rohit.seth@intel.com>
 */

#include <linux/init.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/pagemap.h>
#include <linux/err.h>
#include <linux/sysctl.h>
#include <asm/mman.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
#include <asm/pgalloc.h>

#if 0	/* This is just for testing */
struct page *
follow_huge_addr(struct mm_struct *mm, unsigned long address, int write)
{
	unsigned long start = address;
	int length = 1;
	int nr;
	struct page *page;
	struct vm_area_struct *vma;

	vma = find_vma(mm, addr);
	if (!vma || !is_vm_hugetlb_page(vma))
		return ERR_PTR(-EINVAL);

	pte = huge_pte_offset(mm, address);

	/* hugetlb should be locked, and hence, prefaulted */
	WARN_ON(!pte || pte_none(*pte));

	page = &pte_page(*pte)[vpfn % (HPAGE_SIZE/PAGE_SIZE)];

	WARN_ON(!PageHead(page));

	return page;
}

int pmd_huge(pmd_t pmd)
{
	return 0;
}

int pud_huge(pud_t pud)
{
	return 0;
}

#else

/*
 * pmd_huge() returns 1 if @pmd is hugetlb related entry, that is normal
 * hugetlb entry or non-present (migration or hwpoisoned) hugetlb entry.
 * Otherwise, returns 0.
 */
int pmd_huge(pmd_t pmd)
{
	return !pmd_none(pmd) &&
		(pmd_val(pmd) & (_PAGE_PRESENT|_PAGE_PSE)) != _PAGE_PRESENT;
}

int pud_huge(pud_t pud)
{
	return !!(pud_val(pud) & _PAGE_PSE);
}
#endif

#ifdef CONFIG_HUGETLB_PAGE
static unsigned long hugetlb_get_unmapped_area_bottomup(struct file *file,
		unsigned long addr, unsigned long len,
		unsigned long pgoff, unsigned long flags)
{
	struct hstate *h = hstate_file(file);
	struct vm_unmapped_area_info info;

	info.flags = 0;
	info.length = len;
	info.low_limit = current->mm->mmap_legacy_base;
	info.high_limit = TASK_SIZE;
	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
	info.align_offset = 0;
	return vm_unmapped_area(&info);
}

static unsigned long hugetlb_get_unmapped_area_topdown(struct file *file,
		unsigned long addr0, unsigned long len,
		unsigned long pgoff, unsigned long flags)
{
	struct hstate *h = hstate_file(file);
	struct vm_unmapped_area_info info;
	unsigned long addr;

	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	info.length = len;
	info.low_limit = PAGE_SIZE;
	info.high_limit = current->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 = TASK_SIZE;
		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;

	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);
}
#endif /* CONFIG_HUGETLB_PAGE */

#ifdef CONFIG_X86_64
static __init int setup_hugepagesz(char *opt)
{
	unsigned long ps = memparse(opt, &opt);
	if (ps == PMD_SIZE) {
		hugetlb_add_hstate(PMD_SHIFT - PAGE_SHIFT);
	} else if (ps == PUD_SIZE && boot_cpu_has(X86_FEATURE_GBPAGES)) {
		hugetlb_add_hstate(PUD_SHIFT - PAGE_SHIFT);
	} else {
		printk(KERN_ERR "hugepagesz: Unsupported page size %lu M\n",
			ps >> 20);
		return 0;
	}
	return 1;
}
__setup("hugepagesz=", setup_hugepagesz);

#if (defined(CONFIG_MEMORY_ISOLATION) && defined(CONFIG_COMPACTION)) || defined(CONFIG_CMA)
static __init int gigantic_pages_init(void)
{
	/* With compaction or CMA we can allocate gigantic pages at runtime */
	if (boot_cpu_has(X86_FEATURE_GBPAGES) && !size_to_hstate(1UL << PUD_SHIFT))
		hugetlb_add_hstate(PUD_SHIFT - PAGE_SHIFT);
	return 0;
}
arch_initcall(gigantic_pages_init);
#endif
#endif
Commit	Line	Data
1da177e4 LT	1	/*
	2	* IA-32 Huge TLB Page Support for Kernel.
	3	*
	4	* Copyright (C) 2002, Rohit Seth <rohit.seth@intel.com>
	5	*/
	6
1da177e4 LT	7	#include <linux/init.h>
	8	#include <linux/fs.h>
	9	#include <linux/mm.h>
	10	#include <linux/hugetlb.h>
	11	#include <linux/pagemap.h>
1da177e4 LT	12	#include <linux/err.h>
	13	#include <linux/sysctl.h>
	14	#include <asm/mman.h>
	15	#include <asm/tlb.h>
	16	#include <asm/tlbflush.h>
a5a19c63	17	#include <asm/pgalloc.h>
1da177e4	18
1da177e4 LT	19	#if 0 /* This is just for testing */
	20	struct page *
	21	follow_huge_addr(struct mm_struct *mm, unsigned long address, int write)
	22	{
	23	unsigned long start = address;
	24	int length = 1;
	25	int nr;
	26	struct page *page;
	27	struct vm_area_struct *vma;
	28
	29	vma = find_vma(mm, addr);
	30	if (!vma \|\| !is_vm_hugetlb_page(vma))
	31	return ERR_PTR(-EINVAL);
	32
	33	pte = huge_pte_offset(mm, address);
	34
	35	/* hugetlb should be locked, and hence, prefaulted */
	36	WARN_ON(!pte \|\| pte_none(*pte));
	37
	38	page = &pte_page(*pte)[vpfn % (HPAGE_SIZE/PAGE_SIZE)];
	39
25e59881	40	WARN_ON(!PageHead(page));
1da177e4 LT	41
	42	return page;
	43	}
	44
	45	int pmd_huge(pmd_t pmd)
	46	{
	47	return 0;
	48	}
	49
ceb86879 AK	50	int pud_huge(pud_t pud)
	51	{
	52	return 0;
	53	}
	54
1da177e4 LT	55	#else
1da177e4 LT	56
cbef8478 NH	57	/*
	58	* pmd_huge() returns 1 if @pmd is hugetlb related entry, that is normal
	59	* hugetlb entry or non-present (migration or hwpoisoned) hugetlb entry.
	60	* Otherwise, returns 0.
	61	*/
1da177e4 LT	62	int pmd_huge(pmd_t pmd)
1da177e4 LT	63	{
cbef8478 NH	64	return !pmd_none(pmd) &&
cbef8478 NH	65	(pmd_val(pmd) & (_PAGE_PRESENT\|_PAGE_PSE)) != _PAGE_PRESENT;
1da177e4 LT	66	}
1da177e4 LT	67
ceb86879 AK	68	int pud_huge(pud_t pud)
ceb86879 AK	69	{
39c11e6c	70	return !!(pud_val(pud) & _PAGE_PSE);
ceb86879	71	}
1da177e4 LT	72	#endif
1da177e4 LT	73
fd8526ad	74	#ifdef CONFIG_HUGETLB_PAGE
1da177e4 LT	75	static unsigned long hugetlb_get_unmapped_area_bottomup(struct file *file,
	76	unsigned long addr, unsigned long len,
	77	unsigned long pgoff, unsigned long flags)
	78	{
39c11e6c	79	struct hstate *h = hstate_file(file);
cdc17344 ML	80	struct vm_unmapped_area_info info;
	81
	82	info.flags = 0;
	83	info.length = len;
fd8526ad	84	info.low_limit = current->mm->mmap_legacy_base;
cdc17344 ML	85	info.high_limit = TASK_SIZE;
	86	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
	87	info.align_offset = 0;
	88	return vm_unmapped_area(&info);
1da177e4 LT	89	}
	90
	91	static unsigned long hugetlb_get_unmapped_area_topdown(struct file *file,
	92	unsigned long addr0, unsigned long len,
	93	unsigned long pgoff, unsigned long flags)
	94	{
39c11e6c	95	struct hstate *h = hstate_file(file);
cdc17344 ML	96	struct vm_unmapped_area_info info;
cdc17344 ML	97	unsigned long addr;
1da177e4	98
cdc17344 ML	99	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	100	info.length = len;
	101	info.low_limit = PAGE_SIZE;
	102	info.high_limit = current->mm->mmap_base;
	103	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
	104	info.align_offset = 0;
	105	addr = vm_unmapped_area(&info);
1da177e4	106
1da177e4 LT	107	/*
	108	* A failed mmap() very likely causes application failure,
	109	* so fall back to the bottom-up function here. This scenario
	110	* can happen with large stack limits and large mmap()
	111	* allocations.
	112	*/
cdc17344 ML	113	if (addr & ~PAGE_MASK) {
	114	VM_BUG_ON(addr != -ENOMEM);
	115	info.flags = 0;
	116	info.low_limit = TASK_UNMAPPED_BASE;
	117	info.high_limit = TASK_SIZE;
	118	addr = vm_unmapped_area(&info);
	119	}
1da177e4 LT	120
	121	return addr;
	122	}
	123
	124	unsigned long
	125	hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
	126	unsigned long len, unsigned long pgoff, unsigned long flags)
	127	{
39c11e6c	128	struct hstate *h = hstate_file(file);
1da177e4 LT	129	struct mm_struct *mm = current->mm;
	130	struct vm_area_struct *vma;
	131
39c11e6c	132	if (len & ~huge_page_mask(h))
1da177e4 LT	133	return -EINVAL;
	134	if (len > TASK_SIZE)
	135	return -ENOMEM;
	136
5a8130f2	137	if (flags & MAP_FIXED) {
a5516438	138	if (prepare_hugepage_range(file, addr, len))
5a8130f2 BH	139	return -EINVAL;
	140	return addr;
	141	}
	142
1da177e4	143	if (addr) {
39c11e6c	144	addr = ALIGN(addr, huge_page_size(h));
1da177e4 LT	145	vma = find_vma(mm, addr);
	146	if (TASK_SIZE - len >= addr &&
	147	(!vma \|\| addr + len <= vma->vm_start))
	148	return addr;
	149	}
	150	if (mm->get_unmapped_area == arch_get_unmapped_area)
	151	return hugetlb_get_unmapped_area_bottomup(file, addr, len,
	152	pgoff, flags);
	153	else
	154	return hugetlb_get_unmapped_area_topdown(file, addr, len,
	155	pgoff, flags);
	156	}
fd8526ad	157	#endif /* CONFIG_HUGETLB_PAGE */
1da177e4	158
b4718e62 AK	159	#ifdef CONFIG_X86_64
	160	static __init int setup_hugepagesz(char *opt)
	161	{
	162	unsigned long ps = memparse(opt, &opt);
	163	if (ps == PMD_SIZE) {
	164	hugetlb_add_hstate(PMD_SHIFT - PAGE_SHIFT);
b8291adc	165	} else if (ps == PUD_SIZE && boot_cpu_has(X86_FEATURE_GBPAGES)) {
b4718e62 AK	166	hugetlb_add_hstate(PUD_SHIFT - PAGE_SHIFT);
	167	} else {
	168	printk(KERN_ERR "hugepagesz: Unsupported page size %lu M\n",
	169	ps >> 20);
	170	return 0;
	171	}
	172	return 1;
	173	}
	174	__setup("hugepagesz=", setup_hugepagesz);
ece84b39	175
080fe206	176	#if (defined(CONFIG_MEMORY_ISOLATION) && defined(CONFIG_COMPACTION)) \|\| defined(CONFIG_CMA)
ece84b39 KS	177	static __init int gigantic_pages_init(void)
ece84b39 KS	178	{
080fe206	179	/* With compaction or CMA we can allocate gigantic pages at runtime */
b8291adc	180	if (boot_cpu_has(X86_FEATURE_GBPAGES) && !size_to_hstate(1UL << PUD_SHIFT))
ece84b39 KS	181	hugetlb_add_hstate(PUD_SHIFT - PAGE_SHIFT);
	182	return 0;
	183	}
	184	arch_initcall(gigantic_pages_init);
	185	#endif
b4718e62	186	#endif