[linux.git] / mm / fremap.c

/*
 *   linux/mm/fremap.c
 * 
 * Explicit pagetable population and nonlinear (random) mappings support.
 *
 * started by Ingo Molnar, Copyright (C) 2002, 2003
 */
#include <linux/backing-dev.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/file.h>
#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/swapops.h>
#include <linux/rmap.h>
#include <linux/syscalls.h>
#include <linux/mmu_notifier.h>

#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

#include "internal.h"

static void zap_pte(struct mm_struct *mm, struct vm_area_struct *vma,
			unsigned long addr, pte_t *ptep)
{
	pte_t pte = *ptep;

	if (pte_present(pte)) {
		struct page *page;

		flush_cache_page(vma, addr, pte_pfn(pte));
		pte = ptep_clear_flush(vma, addr, ptep);
		page = vm_normal_page(vma, addr, pte);
		if (page) {
			if (pte_dirty(pte))
				set_page_dirty(page);
			page_remove_rmap(page);
			page_cache_release(page);
			update_hiwater_rss(mm);
			dec_mm_counter(mm, MM_FILEPAGES);
		}
	} else {
		if (!pte_file(pte))
			free_swap_and_cache(pte_to_swp_entry(pte));
		pte_clear_not_present_full(mm, addr, ptep, 0);
	}
}

/*
 * Install a file pte to a given virtual memory address, release any
 * previously existing mapping.
 */
static int install_file_pte(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long addr, unsigned long pgoff, pgprot_t prot)
{
	int err = -ENOMEM;
	pte_t *pte;
	spinlock_t *ptl;

	pte = get_locked_pte(mm, addr, &ptl);
	if (!pte)
		goto out;

	if (!pte_none(*pte))
		zap_pte(mm, vma, addr, pte);

	set_pte_at(mm, addr, pte, pgoff_to_pte(pgoff));
	/*
	 * We don't need to run update_mmu_cache() here because the "file pte"
	 * being installed by install_file_pte() is not a real pte - it's a
	 * non-present entry (like a swap entry), noting what file offset should
	 * be mapped there when there's a fault (in a non-linear vma where
	 * that's not obvious).
	 */
	pte_unmap_unlock(pte, ptl);
	err = 0;
out:
	return err;
}

static int populate_range(struct mm_struct *mm, struct vm_area_struct *vma,
			unsigned long addr, unsigned long size, pgoff_t pgoff)
{
	int err;

	do {
		err = install_file_pte(mm, vma, addr, pgoff, vma->vm_page_prot);
		if (err)
			return err;

		size -= PAGE_SIZE;
		addr += PAGE_SIZE;
		pgoff++;
	} while (size);

        return 0;

}

/**
 * sys_remap_file_pages - remap arbitrary pages of an existing VM_SHARED vma
 * @start: start of the remapped virtual memory range
 * @size: size of the remapped virtual memory range
 * @prot: new protection bits of the range (see NOTE)
 * @pgoff: to-be-mapped page of the backing store file
 * @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
 *
 * sys_remap_file_pages remaps arbitrary pages of an existing VM_SHARED vma
 * (shared backing store file).
 *
 * This syscall works purely via pagetables, so it's the most efficient
 * way to map the same (large) file into a given virtual window. Unlike
 * mmap()/mremap() it does not create any new vmas. The new mappings are
 * also safe across swapout.
 *
 * NOTE: the @prot parameter right now is ignored (but must be zero),
 * and the vma's default protection is used. Arbitrary protections
 * might be implemented in the future.
 */
SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
		unsigned long, prot, unsigned long, pgoff, unsigned long, flags)
{
	struct mm_struct *mm = current->mm;
	struct address_space *mapping;
	struct vm_area_struct *vma;
	int err = -EINVAL;
	int has_write_lock = 0;

	if (prot)
		return err;
	/*
	 * Sanitize the syscall parameters:
	 */
	start = start & PAGE_MASK;
	size = size & PAGE_MASK;

	/* Does the address range wrap, or is the span zero-sized? */
	if (start + size <= start)
		return err;

	/* Does pgoff wrap? */
	if (pgoff + (size >> PAGE_SHIFT) < pgoff)
		return err;

	/* Can we represent this offset inside this architecture's pte's? */
#if PTE_FILE_MAX_BITS < BITS_PER_LONG
	if (pgoff + (size >> PAGE_SHIFT) >= (1UL << PTE_FILE_MAX_BITS))
		return err;
#endif

	/* We need down_write() to change vma->vm_flags. */
	down_read(&mm->mmap_sem);
 retry:
	vma = find_vma(mm, start);

	/*
	 * Make sure the vma is shared, that it supports prefaulting,
	 * and that the remapped range is valid and fully within
	 * the single existing vma.  vm_private_data is used as a
	 * swapout cursor in a VM_NONLINEAR vma.
	 */
	if (!vma || !(vma->vm_flags & VM_SHARED))
		goto out;

	if (vma->vm_private_data && !(vma->vm_flags & VM_NONLINEAR))
		goto out;

	if (!(vma->vm_flags & VM_CAN_NONLINEAR))
		goto out;

	if (start < vma->vm_start || start + size > vma->vm_end)
		goto out;

	/* Must set VM_NONLINEAR before any pages are populated. */
	if (!(vma->vm_flags & VM_NONLINEAR)) {
		/* Don't need a nonlinear mapping, exit success */
		if (pgoff == linear_page_index(vma, start)) {
			err = 0;
			goto out;
		}

		if (!has_write_lock) {
			up_read(&mm->mmap_sem);
			down_write(&mm->mmap_sem);
			has_write_lock = 1;
			goto retry;
		}
		mapping = vma->vm_file->f_mapping;
		/*
		 * page_mkclean doesn't work on nonlinear vmas, so if
		 * dirty pages need to be accounted, emulate with linear
		 * vmas.
		 */
		if (mapping_cap_account_dirty(mapping)) {
			unsigned long addr;
			struct file *file = vma->vm_file;

			flags &= MAP_NONBLOCK;
			get_file(file);
			addr = mmap_region(file, start, size,
					flags, vma->vm_flags, pgoff);
			fput(file);
			if (IS_ERR_VALUE(addr)) {
				err = addr;
			} else {
				BUG_ON(addr != start);
				err = 0;
			}
			goto out;
		}
		mutex_lock(&mapping->i_mmap_mutex);
		flush_dcache_mmap_lock(mapping);
		vma->vm_flags |= VM_NONLINEAR;
		vma_prio_tree_remove(vma, &mapping->i_mmap);
		vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
		flush_dcache_mmap_unlock(mapping);
		mutex_unlock(&mapping->i_mmap_mutex);
	}

	if (vma->vm_flags & VM_LOCKED) {
		/*
		 * drop PG_Mlocked flag for over-mapped range
		 */
		vm_flags_t saved_flags = vma->vm_flags;
		munlock_vma_pages_range(vma, start, start + size);
		vma->vm_flags = saved_flags;
	}

	mmu_notifier_invalidate_range_start(mm, start, start + size);
	err = populate_range(mm, vma, start, size, pgoff);
	mmu_notifier_invalidate_range_end(mm, start, start + size);
	if (!err && !(flags & MAP_NONBLOCK)) {
		if (vma->vm_flags & VM_LOCKED) {
			/*
			 * might be mapping previously unmapped range of file
			 */
			mlock_vma_pages_range(vma, start, start + size);
		} else {
			if (unlikely(has_write_lock)) {
				downgrade_write(&mm->mmap_sem);
				has_write_lock = 0;
			}
			make_pages_present(start, start+size);
		}
	}

	/*
	 * We can't clear VM_NONLINEAR because we'd have to do
	 * it after ->populate completes, and that would prevent
	 * downgrading the lock.  (Locks can't be upgraded).
	 */

out:
	if (likely(!has_write_lock))
		up_read(&mm->mmap_sem);
	else
		up_write(&mm->mmap_sem);

	return err;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/mm/fremap.c
	3	*
	4	* Explicit pagetable population and nonlinear (random) mappings support.
	5	*
	6	* started by Ingo Molnar, Copyright (C) 2002, 2003
	7	*/
4af3c9cc	8	#include <linux/backing-dev.h>
1da177e4 LT	9	#include <linux/mm.h>
	10	#include <linux/swap.h>
	11	#include <linux/file.h>
	12	#include <linux/mman.h>
	13	#include <linux/pagemap.h>
	14	#include <linux/swapops.h>
	15	#include <linux/rmap.h>
1da177e4	16	#include <linux/syscalls.h>
cddb8a5c	17	#include <linux/mmu_notifier.h>
1da177e4 LT	18
	19	#include <asm/mmu_context.h>
	20	#include <asm/cacheflush.h>
	21	#include <asm/tlbflush.h>
	22
ba470de4 RR	23	#include "internal.h"
ba470de4 RR	24
d0217ac0	25	static void zap_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	26	unsigned long addr, pte_t *ptep)
	27	{
	28	pte_t pte = *ptep;
	29
1da177e4	30	if (pte_present(pte)) {
d0217ac0 NP	31	struct page *page;
d0217ac0 NP	32
6aab341e	33	flush_cache_page(vma, addr, pte_pfn(pte));
1da177e4	34	pte = ptep_clear_flush(vma, addr, ptep);
6aab341e LT	35	page = vm_normal_page(vma, addr, pte);
	36	if (page) {
	37	if (pte_dirty(pte))
	38	set_page_dirty(page);
edc315fd	39	page_remove_rmap(page);
6aab341e	40	page_cache_release(page);
d0217ac0	41	update_hiwater_rss(mm);
d559db08	42	dec_mm_counter(mm, MM_FILEPAGES);
1da177e4 LT	43	}
	44	} else {
	45	if (!pte_file(pte))
	46	free_swap_and_cache(pte_to_swp_entry(pte));
9888a1ca	47	pte_clear_not_present_full(mm, addr, ptep, 0);
1da177e4 LT	48	}
	49	}
	50
1da177e4 LT	51	/*
	52	* Install a file pte to a given virtual memory address, release any
	53	* previously existing mapping.
	54	*/
d0217ac0	55	static int install_file_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	56	unsigned long addr, unsigned long pgoff, pgprot_t prot)
	57	{
	58	int err = -ENOMEM;
	59	pte_t *pte;
c74df32c	60	spinlock_t *ptl;
1da177e4	61
c9cfcddf	62	pte = get_locked_pte(mm, addr, &ptl);
1da177e4	63	if (!pte)
c74df32c	64	goto out;
1da177e4	65
d0217ac0 NP	66	if (!pte_none(*pte))
d0217ac0 NP	67	zap_pte(mm, vma, addr, pte);
1da177e4 LT	68
1da177e4 LT	69	set_pte_at(mm, addr, pte, pgoff_to_pte(pgoff));
668e0d8f HD	70	/*
	71	* We don't need to run update_mmu_cache() here because the "file pte"
	72	* being installed by install_file_pte() is not a real pte - it's a
	73	* non-present entry (like a swap entry), noting what file offset should
	74	* be mapped there when there's a fault (in a non-linear vma where
	75	* that's not obvious).
	76	*/
c74df32c HD	77	pte_unmap_unlock(pte, ptl);
	78	err = 0;
	79	out:
1da177e4 LT	80	return err;
	81	}
	82
54cb8821 NP	83	static int populate_range(struct mm_struct mm, struct vm_area_struct vma,
	84	unsigned long addr, unsigned long size, pgoff_t pgoff)
	85	{
	86	int err;
	87
	88	do {
	89	err = install_file_pte(mm, vma, addr, pgoff, vma->vm_page_prot);
	90	if (err)
	91	return err;
	92
	93	size -= PAGE_SIZE;
	94	addr += PAGE_SIZE;
	95	pgoff++;
	96	} while (size);
	97
	98	return 0;
	99
	100	}
	101
8d63494f RD	102	/**
8d63494f RD	103	* sys_remap_file_pages - remap arbitrary pages of an existing VM_SHARED vma
1da177e4 LT	104	* @start: start of the remapped virtual memory range
1da177e4 LT	105	* @size: size of the remapped virtual memory range
8d63494f RD	106	* @prot: new protection bits of the range (see NOTE)
8d63494f RD	107	* @pgoff: to-be-mapped page of the backing store file
1da177e4 LT	108	* @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
1da177e4 LT	109	*
8d63494f RD	110	* sys_remap_file_pages remaps arbitrary pages of an existing VM_SHARED vma
	111	* (shared backing store file).
	112	*
	113	* This syscall works purely via pagetables, so it's the most efficient
1da177e4 LT	114	* way to map the same (large) file into a given virtual window. Unlike
	115	* mmap()/mremap() it does not create any new vmas. The new mappings are
	116	* also safe across swapout.
	117	*
7682486b	118	* NOTE: the @prot parameter right now is ignored (but must be zero),
8d63494f RD	119	* and the vma's default protection is used. Arbitrary protections
8d63494f RD	120	* might be implemented in the future.
1da177e4	121	*/
6a6160a7 HC	122	SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
6a6160a7 HC	123	unsigned long, prot, unsigned long, pgoff, unsigned long, flags)
1da177e4 LT	124	{
	125	struct mm_struct *mm = current->mm;
	126	struct address_space *mapping;
1da177e4 LT	127	struct vm_area_struct *vma;
	128	int err = -EINVAL;
	129	int has_write_lock = 0;
	130
8d63494f	131	if (prot)
1da177e4 LT	132	return err;
	133	/*
	134	* Sanitize the syscall parameters:
	135	*/
	136	start = start & PAGE_MASK;
	137	size = size & PAGE_MASK;
	138
	139	/* Does the address range wrap, or is the span zero-sized? */
	140	if (start + size <= start)
	141	return err;
	142
5ec1055a LW	143	/* Does pgoff wrap? */
	144	if (pgoff + (size >> PAGE_SHIFT) < pgoff)
	145	return err;
	146
1da177e4 LT	147	/* Can we represent this offset inside this architecture's pte's? */
	148	#if PTE_FILE_MAX_BITS < BITS_PER_LONG
	149	if (pgoff + (size >> PAGE_SHIFT) >= (1UL << PTE_FILE_MAX_BITS))
	150	return err;
	151	#endif
	152
	153	/* We need down_write() to change vma->vm_flags. */
	154	down_read(&mm->mmap_sem);
	155	retry:
	156	vma = find_vma(mm, start);
	157
	158	/*
	159	* Make sure the vma is shared, that it supports prefaulting,
	160	* and that the remapped range is valid and fully within
	161	* the single existing vma. vm_private_data is used as a
101d2be7	162	* swapout cursor in a VM_NONLINEAR vma.
1da177e4	163	*/
54cb8821 NP	164	if (!vma \|\| !(vma->vm_flags & VM_SHARED))
	165	goto out;
	166
	167	if (vma->vm_private_data && !(vma->vm_flags & VM_NONLINEAR))
	168	goto out;
	169
dd204d63	170	if (!(vma->vm_flags & VM_CAN_NONLINEAR))
54cb8821 NP	171	goto out;
54cb8821 NP	172
e92b05de	173	if (start < vma->vm_start \|\| start + size > vma->vm_end)
54cb8821 NP	174	goto out;
	175
	176	/* Must set VM_NONLINEAR before any pages are populated. */
	177	if (!(vma->vm_flags & VM_NONLINEAR)) {
	178	/* Don't need a nonlinear mapping, exit success */
	179	if (pgoff == linear_page_index(vma, start)) {
	180	err = 0;
	181	goto out;
	182	}
	183
	184	if (!has_write_lock) {
	185	up_read(&mm->mmap_sem);
	186	down_write(&mm->mmap_sem);
	187	has_write_lock = 1;
	188	goto retry;
	189	}
	190	mapping = vma->vm_file->f_mapping;
3ee6dafc MS	191	/*
	192	* page_mkclean doesn't work on nonlinear vmas, so if
	193	* dirty pages need to be accounted, emulate with linear
	194	* vmas.
	195	*/
	196	if (mapping_cap_account_dirty(mapping)) {
	197	unsigned long addr;
8a459e44	198	struct file *file = vma->vm_file;
3ee6dafc MS	199
3ee6dafc MS	200	flags &= MAP_NONBLOCK;
8a459e44 ON	201	get_file(file);
8a459e44 ON	202	addr = mmap_region(file, start, size,
5a6fe125	203	flags, vma->vm_flags, pgoff);
8a459e44	204	fput(file);
3ee6dafc MS	205	if (IS_ERR_VALUE(addr)) {
	206	err = addr;
	207	} else {
	208	BUG_ON(addr != start);
	209	err = 0;
	210	}
	211	goto out;
	212	}
3d48ae45	213	mutex_lock(&mapping->i_mmap_mutex);
54cb8821 NP	214	flush_dcache_mmap_lock(mapping);
	215	vma->vm_flags \|= VM_NONLINEAR;
	216	vma_prio_tree_remove(vma, &mapping->i_mmap);
	217	vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
	218	flush_dcache_mmap_unlock(mapping);
3d48ae45	219	mutex_unlock(&mapping->i_mmap_mutex);
54cb8821 NP	220	}
54cb8821 NP	221
ba470de4 RR	222	if (vma->vm_flags & VM_LOCKED) {
	223	/*
	224	* drop PG_Mlocked flag for over-mapped range
	225	*/
ca16d140	226	vm_flags_t saved_flags = vma->vm_flags;
ba470de4 RR	227	munlock_vma_pages_range(vma, start, start + size);
	228	vma->vm_flags = saved_flags;
	229	}
	230
cddb8a5c	231	mmu_notifier_invalidate_range_start(mm, start, start + size);
d0217ac0	232	err = populate_range(mm, vma, start, size, pgoff);
cddb8a5c	233	mmu_notifier_invalidate_range_end(mm, start, start + size);
d0217ac0	234	if (!err && !(flags & MAP_NONBLOCK)) {
ba470de4 RR	235	if (vma->vm_flags & VM_LOCKED) {
	236	/*
	237	* might be mapping previously unmapped range of file
	238	*/
	239	mlock_vma_pages_range(vma, start, start + size);
	240	} else {
	241	if (unlikely(has_write_lock)) {
	242	downgrade_write(&mm->mmap_sem);
	243	has_write_lock = 0;
	244	}
	245	make_pages_present(start, start+size);
1da177e4	246	}
d0217ac0	247	}
1da177e4	248
54cb8821 NP	249	/*
	250	* We can't clear VM_NONLINEAR because we'd have to do
	251	* it after ->populate completes, and that would prevent
	252	* downgrading the lock. (Locks can't be upgraded).
	253	*/
1da177e4	254
54cb8821	255	out:
1da177e4 LT	256	if (likely(!has_write_lock))
	257	up_read(&mm->mmap_sem);
	258	else
	259	up_write(&mm->mmap_sem);
	260
	261	return err;
	262	}