[net-next-2.6.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/module.h>
#include <linux/syscalls.h>
#include <linux/mmu_notifier.h>

#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.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, vma);
			page_cache_release(page);
			update_hiwater_rss(mm);
			dec_mm_counter(mm, file_rss);
		}
	} 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.
 */
asmlinkage long sys_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;
	unsigned long end = start + size;
	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;

	/* 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 (end <= start || start < vma->vm_start || end > 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, 1);
			fput(file);
			if (IS_ERR_VALUE(addr)) {
				err = addr;
			} else {
				BUG_ON(addr != start);
				err = 0;
			}
			goto out;
		}
		spin_lock(&mapping->i_mmap_lock);
		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);
		spin_unlock(&mapping->i_mmap_lock);
	}

	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 (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>
	16	#include <linux/module.h>
	17	#include <linux/syscalls.h>
cddb8a5c	18	#include <linux/mmu_notifier.h>
1da177e4 LT	19
	20	#include <asm/mmu_context.h>
	21	#include <asm/cacheflush.h>
	22	#include <asm/tlbflush.h>
	23
d0217ac0	24	static void zap_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	25	unsigned long addr, pte_t *ptep)
	26	{
	27	pte_t pte = *ptep;
	28
1da177e4	29	if (pte_present(pte)) {
d0217ac0 NP	30	struct page *page;
d0217ac0 NP	31
6aab341e	32	flush_cache_page(vma, addr, pte_pfn(pte));
1da177e4	33	pte = ptep_clear_flush(vma, addr, ptep);
6aab341e LT	34	page = vm_normal_page(vma, addr, pte);
	35	if (page) {
	36	if (pte_dirty(pte))
	37	set_page_dirty(page);
7de6b805	38	page_remove_rmap(page, vma);
6aab341e	39	page_cache_release(page);
d0217ac0 NP	40	update_hiwater_rss(mm);
d0217ac0 NP	41	dec_mm_counter(mm, file_rss);
1da177e4 LT	42	}
	43	} else {
	44	if (!pte_file(pte))
	45	free_swap_and_cache(pte_to_swp_entry(pte));
9888a1ca	46	pte_clear_not_present_full(mm, addr, ptep, 0);
1da177e4 LT	47	}
	48	}
	49
1da177e4 LT	50	/*
	51	* Install a file pte to a given virtual memory address, release any
	52	* previously existing mapping.
	53	*/
d0217ac0	54	static int install_file_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	55	unsigned long addr, unsigned long pgoff, pgprot_t prot)
	56	{
	57	int err = -ENOMEM;
	58	pte_t *pte;
c74df32c	59	spinlock_t *ptl;
1da177e4	60
c9cfcddf	61	pte = get_locked_pte(mm, addr, &ptl);
1da177e4	62	if (!pte)
c74df32c	63	goto out;
1da177e4	64
d0217ac0 NP	65	if (!pte_none(*pte))
d0217ac0 NP	66	zap_pte(mm, vma, addr, pte);
1da177e4 LT	67
1da177e4 LT	68	set_pte_at(mm, addr, pte, pgoff_to_pte(pgoff));
668e0d8f HD	69	/*
	70	* We don't need to run update_mmu_cache() here because the "file pte"
	71	* being installed by install_file_pte() is not a real pte - it's a
	72	* non-present entry (like a swap entry), noting what file offset should
	73	* be mapped there when there's a fault (in a non-linear vma where
	74	* that's not obvious).
	75	*/
c74df32c HD	76	pte_unmap_unlock(pte, ptl);
	77	err = 0;
	78	out:
1da177e4 LT	79	return err;
	80	}
	81
54cb8821 NP	82	static int populate_range(struct mm_struct mm, struct vm_area_struct vma,
	83	unsigned long addr, unsigned long size, pgoff_t pgoff)
	84	{
	85	int err;
	86
	87	do {
	88	err = install_file_pte(mm, vma, addr, pgoff, vma->vm_page_prot);
	89	if (err)
	90	return err;
	91
	92	size -= PAGE_SIZE;
	93	addr += PAGE_SIZE;
	94	pgoff++;
	95	} while (size);
	96
	97	return 0;
	98
	99	}
	100
8d63494f RD	101	/**
8d63494f RD	102	* sys_remap_file_pages - remap arbitrary pages of an existing VM_SHARED vma
1da177e4 LT	103	* @start: start of the remapped virtual memory range
1da177e4 LT	104	* @size: size of the remapped virtual memory range
8d63494f RD	105	* @prot: new protection bits of the range (see NOTE)
8d63494f RD	106	* @pgoff: to-be-mapped page of the backing store file
1da177e4 LT	107	* @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
1da177e4 LT	108	*
8d63494f RD	109	* sys_remap_file_pages remaps arbitrary pages of an existing VM_SHARED vma
	110	* (shared backing store file).
	111	*
	112	* This syscall works purely via pagetables, so it's the most efficient
1da177e4 LT	113	* way to map the same (large) file into a given virtual window. Unlike
	114	* mmap()/mremap() it does not create any new vmas. The new mappings are
	115	* also safe across swapout.
	116	*
7682486b	117	* NOTE: the @prot parameter right now is ignored (but must be zero),
8d63494f RD	118	* and the vma's default protection is used. Arbitrary protections
8d63494f RD	119	* might be implemented in the future.
1da177e4 LT	120	*/
1da177e4 LT	121	asmlinkage long sys_remap_file_pages(unsigned long start, unsigned long size,
8d63494f	122	unsigned long prot, unsigned long pgoff, unsigned long flags)
1da177e4 LT	123	{
	124	struct mm_struct *mm = current->mm;
	125	struct address_space *mapping;
	126	unsigned long end = start + size;
	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
	143	/* Can we represent this offset inside this architecture's pte's? */
	144	#if PTE_FILE_MAX_BITS < BITS_PER_LONG
	145	if (pgoff + (size >> PAGE_SHIFT) >= (1UL << PTE_FILE_MAX_BITS))
	146	return err;
	147	#endif
	148
	149	/* We need down_write() to change vma->vm_flags. */
	150	down_read(&mm->mmap_sem);
	151	retry:
	152	vma = find_vma(mm, start);
	153
	154	/*
	155	* Make sure the vma is shared, that it supports prefaulting,
	156	* and that the remapped range is valid and fully within
	157	* the single existing vma. vm_private_data is used as a
101d2be7	158	* swapout cursor in a VM_NONLINEAR vma.
1da177e4	159	*/
54cb8821 NP	160	if (!vma \|\| !(vma->vm_flags & VM_SHARED))
	161	goto out;
	162
	163	if (vma->vm_private_data && !(vma->vm_flags & VM_NONLINEAR))
	164	goto out;
	165
dd204d63	166	if (!(vma->vm_flags & VM_CAN_NONLINEAR))
54cb8821 NP	167	goto out;
	168
	169	if (end <= start \|\| start < vma->vm_start \|\| end > vma->vm_end)
	170	goto out;
	171
	172	/* Must set VM_NONLINEAR before any pages are populated. */
	173	if (!(vma->vm_flags & VM_NONLINEAR)) {
	174	/* Don't need a nonlinear mapping, exit success */
	175	if (pgoff == linear_page_index(vma, start)) {
	176	err = 0;
	177	goto out;
	178	}
	179
	180	if (!has_write_lock) {
	181	up_read(&mm->mmap_sem);
	182	down_write(&mm->mmap_sem);
	183	has_write_lock = 1;
	184	goto retry;
	185	}
	186	mapping = vma->vm_file->f_mapping;
3ee6dafc MS	187	/*
	188	* page_mkclean doesn't work on nonlinear vmas, so if
	189	* dirty pages need to be accounted, emulate with linear
	190	* vmas.
	191	*/
	192	if (mapping_cap_account_dirty(mapping)) {
	193	unsigned long addr;
8a459e44	194	struct file *file = vma->vm_file;
3ee6dafc MS	195
3ee6dafc MS	196	flags &= MAP_NONBLOCK;
8a459e44 ON	197	get_file(file);
8a459e44 ON	198	addr = mmap_region(file, start, size,
3ee6dafc	199	flags, vma->vm_flags, pgoff, 1);
8a459e44	200	fput(file);
3ee6dafc MS	201	if (IS_ERR_VALUE(addr)) {
	202	err = addr;
	203	} else {
	204	BUG_ON(addr != start);
	205	err = 0;
	206	}
	207	goto out;
	208	}
54cb8821 NP	209	spin_lock(&mapping->i_mmap_lock);
	210	flush_dcache_mmap_lock(mapping);
	211	vma->vm_flags \|= VM_NONLINEAR;
	212	vma_prio_tree_remove(vma, &mapping->i_mmap);
	213	vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
	214	flush_dcache_mmap_unlock(mapping);
	215	spin_unlock(&mapping->i_mmap_lock);
	216	}
	217
cddb8a5c	218	mmu_notifier_invalidate_range_start(mm, start, start + size);
d0217ac0	219	err = populate_range(mm, vma, start, size, pgoff);
cddb8a5c	220	mmu_notifier_invalidate_range_end(mm, start, start + size);
d0217ac0 NP	221	if (!err && !(flags & MAP_NONBLOCK)) {
	222	if (unlikely(has_write_lock)) {
	223	downgrade_write(&mm->mmap_sem);
	224	has_write_lock = 0;
1da177e4	225	}
d0217ac0 NP	226	make_pages_present(start, start+size);
d0217ac0 NP	227	}
1da177e4	228
54cb8821 NP	229	/*
	230	* We can't clear VM_NONLINEAR because we'd have to do
	231	* it after ->populate completes, and that would prevent
	232	* downgrading the lock. (Locks can't be upgraded).
	233	*/
1da177e4	234
54cb8821	235	out:
1da177e4 LT	236	if (likely(!has_write_lock))
	237	up_read(&mm->mmap_sem);
	238	else
	239	up_write(&mm->mmap_sem);
	240
	241	return err;
	242	}
	243