[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 <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;

			flags &= MAP_NONBLOCK;
			addr = mmap_region(vma->vm_file, start, size,
					flags, vma->vm_flags, pgoff, 1);
			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);
	}

	err = populate_range(mm, vma, start, size, pgoff);
	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>
	18
	19	#include <asm/mmu_context.h>
	20	#include <asm/cacheflush.h>
	21	#include <asm/tlbflush.h>
	22
d0217ac0	23	static void zap_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	24	unsigned long addr, pte_t *ptep)
	25	{
	26	pte_t pte = *ptep;
	27
1da177e4	28	if (pte_present(pte)) {
d0217ac0 NP	29	struct page *page;
d0217ac0 NP	30
6aab341e	31	flush_cache_page(vma, addr, pte_pfn(pte));
1da177e4	32	pte = ptep_clear_flush(vma, addr, ptep);
6aab341e LT	33	page = vm_normal_page(vma, addr, pte);
	34	if (page) {
	35	if (pte_dirty(pte))
	36	set_page_dirty(page);
7de6b805	37	page_remove_rmap(page, vma);
6aab341e	38	page_cache_release(page);
d0217ac0 NP	39	update_hiwater_rss(mm);
d0217ac0 NP	40	dec_mm_counter(mm, file_rss);
1da177e4 LT	41	}
	42	} else {
	43	if (!pte_file(pte))
	44	free_swap_and_cache(pte_to_swp_entry(pte));
9888a1ca	45	pte_clear_not_present_full(mm, addr, ptep, 0);
1da177e4 LT	46	}
	47	}
	48
1da177e4 LT	49	/*
	50	* Install a file pte to a given virtual memory address, release any
	51	* previously existing mapping.
	52	*/
d0217ac0	53	static int install_file_pte(struct mm_struct mm, struct vm_area_struct vma,
1da177e4 LT	54	unsigned long addr, unsigned long pgoff, pgprot_t prot)
	55	{
	56	int err = -ENOMEM;
	57	pte_t *pte;
c74df32c	58	spinlock_t *ptl;
1da177e4	59
c9cfcddf	60	pte = get_locked_pte(mm, addr, &ptl);
1da177e4	61	if (!pte)
c74df32c	62	goto out;
1da177e4	63
d0217ac0 NP	64	if (!pte_none(*pte))
d0217ac0 NP	65	zap_pte(mm, vma, addr, pte);
1da177e4 LT	66
1da177e4 LT	67	set_pte_at(mm, addr, pte, pgoff_to_pte(pgoff));
668e0d8f HD	68	/*
	69	* We don't need to run update_mmu_cache() here because the "file pte"
	70	* being installed by install_file_pte() is not a real pte - it's a
	71	* non-present entry (like a swap entry), noting what file offset should
	72	* be mapped there when there's a fault (in a non-linear vma where
	73	* that's not obvious).
	74	*/
c74df32c HD	75	pte_unmap_unlock(pte, ptl);
	76	err = 0;
	77	out:
1da177e4 LT	78	return err;
	79	}
	80
54cb8821 NP	81	static int populate_range(struct mm_struct mm, struct vm_area_struct vma,
	82	unsigned long addr, unsigned long size, pgoff_t pgoff)
	83	{
	84	int err;
	85
	86	do {
	87	err = install_file_pte(mm, vma, addr, pgoff, vma->vm_page_prot);
	88	if (err)
	89	return err;
	90
	91	size -= PAGE_SIZE;
	92	addr += PAGE_SIZE;
	93	pgoff++;
	94	} while (size);
	95
	96	return 0;
	97
	98	}
	99
8d63494f RD	100	/**
8d63494f RD	101	* sys_remap_file_pages - remap arbitrary pages of an existing VM_SHARED vma
1da177e4 LT	102	* @start: start of the remapped virtual memory range
1da177e4 LT	103	* @size: size of the remapped virtual memory range
8d63494f RD	104	* @prot: new protection bits of the range (see NOTE)
8d63494f RD	105	* @pgoff: to-be-mapped page of the backing store file
1da177e4 LT	106	* @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
1da177e4 LT	107	*
8d63494f RD	108	* sys_remap_file_pages remaps arbitrary pages of an existing VM_SHARED vma
	109	* (shared backing store file).
	110	*
	111	* This syscall works purely via pagetables, so it's the most efficient
1da177e4 LT	112	* way to map the same (large) file into a given virtual window. Unlike
	113	* mmap()/mremap() it does not create any new vmas. The new mappings are
	114	* also safe across swapout.
	115	*
8d63494f RD	116	* NOTE: the 'prot' parameter right now is ignored (but must be zero),
	117	* and the vma's default protection is used. Arbitrary protections
	118	* might be implemented in the future.
1da177e4 LT	119	*/
1da177e4 LT	120	asmlinkage long sys_remap_file_pages(unsigned long start, unsigned long size,
8d63494f	121	unsigned long prot, unsigned long pgoff, unsigned long flags)
1da177e4 LT	122	{
	123	struct mm_struct *mm = current->mm;
	124	struct address_space *mapping;
	125	unsigned long end = start + size;
	126	struct vm_area_struct *vma;
	127	int err = -EINVAL;
	128	int has_write_lock = 0;
	129
8d63494f	130	if (prot)
1da177e4 LT	131	return err;
	132	/*
	133	* Sanitize the syscall parameters:
	134	*/
	135	start = start & PAGE_MASK;
	136	size = size & PAGE_MASK;
	137
	138	/* Does the address range wrap, or is the span zero-sized? */
	139	if (start + size <= start)
	140	return err;
	141
	142	/* Can we represent this offset inside this architecture's pte's? */
	143	#if PTE_FILE_MAX_BITS < BITS_PER_LONG
	144	if (pgoff + (size >> PAGE_SHIFT) >= (1UL << PTE_FILE_MAX_BITS))
	145	return err;
	146	#endif
	147
	148	/* We need down_write() to change vma->vm_flags. */
	149	down_read(&mm->mmap_sem);
	150	retry:
	151	vma = find_vma(mm, start);
	152
	153	/*
	154	* Make sure the vma is shared, that it supports prefaulting,
	155	* and that the remapped range is valid and fully within
	156	* the single existing vma. vm_private_data is used as a
101d2be7	157	* swapout cursor in a VM_NONLINEAR vma.
1da177e4	158	*/
54cb8821 NP	159	if (!vma \|\| !(vma->vm_flags & VM_SHARED))
	160	goto out;
	161
	162	if (vma->vm_private_data && !(vma->vm_flags & VM_NONLINEAR))
	163	goto out;
	164
dd204d63	165	if (!(vma->vm_flags & VM_CAN_NONLINEAR))
54cb8821 NP	166	goto out;
	167
	168	if (end <= start \|\| start < vma->vm_start \|\| end > vma->vm_end)
	169	goto out;
	170
	171	/* Must set VM_NONLINEAR before any pages are populated. */
	172	if (!(vma->vm_flags & VM_NONLINEAR)) {
	173	/* Don't need a nonlinear mapping, exit success */
	174	if (pgoff == linear_page_index(vma, start)) {
	175	err = 0;
	176	goto out;
	177	}
	178
	179	if (!has_write_lock) {
	180	up_read(&mm->mmap_sem);
	181	down_write(&mm->mmap_sem);
	182	has_write_lock = 1;
	183	goto retry;
	184	}
	185	mapping = vma->vm_file->f_mapping;
3ee6dafc MS	186	/*
	187	* page_mkclean doesn't work on nonlinear vmas, so if
	188	* dirty pages need to be accounted, emulate with linear
	189	* vmas.
	190	*/
	191	if (mapping_cap_account_dirty(mapping)) {
	192	unsigned long addr;
	193
	194	flags &= MAP_NONBLOCK;
	195	addr = mmap_region(vma->vm_file, start, size,
	196	flags, vma->vm_flags, pgoff, 1);
	197	if (IS_ERR_VALUE(addr)) {
	198	err = addr;
	199	} else {
	200	BUG_ON(addr != start);
	201	err = 0;
	202	}
	203	goto out;
	204	}
54cb8821 NP	205	spin_lock(&mapping->i_mmap_lock);
	206	flush_dcache_mmap_lock(mapping);
	207	vma->vm_flags \|= VM_NONLINEAR;
	208	vma_prio_tree_remove(vma, &mapping->i_mmap);
	209	vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
	210	flush_dcache_mmap_unlock(mapping);
	211	spin_unlock(&mapping->i_mmap_lock);
	212	}
	213
d0217ac0 NP	214	err = populate_range(mm, vma, start, size, pgoff);
	215	if (!err && !(flags & MAP_NONBLOCK)) {
	216	if (unlikely(has_write_lock)) {
	217	downgrade_write(&mm->mmap_sem);
	218	has_write_lock = 0;
1da177e4	219	}
d0217ac0 NP	220	make_pages_present(start, start+size);
d0217ac0 NP	221	}
1da177e4	222
54cb8821 NP	223	/*
	224	* We can't clear VM_NONLINEAR because we'd have to do
	225	* it after ->populate completes, and that would prevent
	226	* downgrading the lock. (Locks can't be upgraded).
	227	*/
1da177e4	228
54cb8821	229	out:
1da177e4 LT	230	if (likely(!has_write_lock))
	231	up_read(&mm->mmap_sem);
	232	else
	233	up_write(&mm->mmap_sem);
	234
	235	return err;
	236	}
	237