[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/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 int zap_pte(struct mm_struct *mm, struct vm_area_struct *vma,
			unsigned long addr, pte_t *ptep)
{
	pte_t pte = *ptep;
	struct page *page = NULL;

	if (pte_present(pte)) {
		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);
		}
	} else {
		if (!pte_file(pte))
			free_swap_and_cache(pte_to_swp_entry(pte));
		pte_clear_not_present_full(mm, addr, ptep, 0);
	}
	return !!page;
}

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

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

	/*
	 * This page may have been truncated. Tell the
	 * caller about it.
	 */
	err = -EINVAL;
	inode = vma->vm_file->f_mapping->host;
	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
	if (!page->mapping || page->index >= size)
		goto unlock;
	err = -ENOMEM;
	if (page_mapcount(page) > INT_MAX/2)
		goto unlock;

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

	flush_icache_page(vma, page);
	pte_val = mk_pte(page, prot);
	set_pte_at(mm, addr, pte, pte_val);
	page_add_file_rmap(page);
	update_mmu_cache(vma, addr, pte_val);
	lazy_mmu_prot_update(pte_val);
	err = 0;
unlock:
	pte_unmap_unlock(pte, ptl);
out:
	return err;
}
EXPORT_SYMBOL(install_page);

/*
 * Install a file pte to a given virtual memory address, release any
 * previously existing mapping.
 */
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)) {
		update_hiwater_rss(mm);
		dec_mm_counter(mm, file_rss);
	}

	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;
}

/***
 * sys_remap_file_pages - remap arbitrary pages of a shared backing store
 *                        file within an existing vma.
 * @start: start of the remapped virtual memory range
 * @size: size of the remapped virtual memory range
 * @prot: new protection bits of the range
 * @pgoff: to be mapped page of the backing store file
 * @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
 *
 * 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, 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) &&
		(!vma->vm_private_data || (vma->vm_flags & VM_NONLINEAR)) &&
		vma->vm_ops && vma->vm_ops->populate &&
			end > start && start >= vma->vm_start &&
				end <= vma->vm_end) {

		/* Must set VM_NONLINEAR before any pages are populated. */
		if (pgoff != linear_page_index(vma, start) &&
		    !(vma->vm_flags & VM_NONLINEAR)) {
			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;
			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 = vma->vm_ops->populate(vma, start, size,
					    vma->vm_page_prot,
					    pgoff, flags & MAP_NONBLOCK);

		/*
		 * 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).
		 */
	}
	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	*/
	8
	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
861f2fb8	23	static int 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;
861f2fb8	27	struct page *page = NULL;
1da177e4	28
1da177e4	29	if (pte_present(pte)) {
6aab341e	30	flush_cache_page(vma, addr, pte_pfn(pte));
1da177e4	31	pte = ptep_clear_flush(vma, addr, ptep);
6aab341e LT	32	page = vm_normal_page(vma, addr, pte);
	33	if (page) {
	34	if (pte_dirty(pte))
	35	set_page_dirty(page);
	36	page_remove_rmap(page);
	37	page_cache_release(page);
1da177e4 LT	38	}
	39	} else {
	40	if (!pte_file(pte))
	41	free_swap_and_cache(pte_to_swp_entry(pte));
9888a1ca	42	pte_clear_not_present_full(mm, addr, ptep, 0);
1da177e4	43	}
861f2fb8	44	return !!page;
1da177e4 LT	45	}
	46
	47	/*
	48	* Install a file page to a given virtual memory address, release any
	49	* previously existing mapping.
	50	*/
	51	int install_page(struct mm_struct mm, struct vm_area_struct vma,
	52	unsigned long addr, struct page *page, pgprot_t prot)
	53	{
	54	struct inode *inode;
	55	pgoff_t size;
	56	int err = -ENOMEM;
	57	pte_t *pte;
1da177e4	58	pte_t pte_val;
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 LT	64
	65	/*
	66	* This page may have been truncated. Tell the
	67	* caller about it.
	68	*/
	69	err = -EINVAL;
	70	inode = vma->vm_file->f_mapping->host;
	71	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
	72	if (!page->mapping \|\| page->index >= size)
c74df32c	73	goto unlock;
f5154a98 HD	74	err = -ENOMEM;
f5154a98 HD	75	if (page_mapcount(page) > INT_MAX/2)
c74df32c	76	goto unlock;
1da177e4	77
861f2fb8 HD	78	if (pte_none(*pte) \|\| !zap_pte(mm, vma, addr, pte))
861f2fb8 HD	79	inc_mm_counter(mm, file_rss);
1da177e4	80
1da177e4	81	flush_icache_page(vma, page);
e88dd6c1 PZ	82	pte_val = mk_pte(page, prot);
e88dd6c1 PZ	83	set_pte_at(mm, addr, pte, pte_val);
1da177e4	84	page_add_file_rmap(page);
1da177e4	85	update_mmu_cache(vma, addr, pte_val);
668e0d8f	86	lazy_mmu_prot_update(pte_val);
1da177e4	87	err = 0;
c74df32c HD	88	unlock:
	89	pte_unmap_unlock(pte, ptl);
	90	out:
1da177e4 LT	91	return err;
	92	}
	93	EXPORT_SYMBOL(install_page);
	94
1da177e4 LT	95	/*
	96	* Install a file pte to a given virtual memory address, release any
	97	* previously existing mapping.
	98	*/
	99	int install_file_pte(struct mm_struct mm, struct vm_area_struct vma,
	100	unsigned long addr, unsigned long pgoff, pgprot_t prot)
	101	{
	102	int err = -ENOMEM;
	103	pte_t *pte;
c74df32c	104	spinlock_t *ptl;
1da177e4	105
c9cfcddf	106	pte = get_locked_pte(mm, addr, &ptl);
1da177e4	107	if (!pte)
c74df32c	108	goto out;
1da177e4	109
365e9c87 HD	110	if (!pte_none(*pte) && zap_pte(mm, vma, addr, pte)) {
365e9c87 HD	111	update_hiwater_rss(mm);
861f2fb8	112	dec_mm_counter(mm, file_rss);
365e9c87	113	}
1da177e4 LT	114
1da177e4 LT	115	set_pte_at(mm, addr, pte, pgoff_to_pte(pgoff));
668e0d8f HD	116	/*
	117	* We don't need to run update_mmu_cache() here because the "file pte"
	118	* being installed by install_file_pte() is not a real pte - it's a
	119	* non-present entry (like a swap entry), noting what file offset should
	120	* be mapped there when there's a fault (in a non-linear vma where
	121	* that's not obvious).
	122	*/
c74df32c HD	123	pte_unmap_unlock(pte, ptl);
	124	err = 0;
	125	out:
1da177e4 LT	126	return err;
	127	}
	128
1da177e4 LT	129	/***
	130	* sys_remap_file_pages - remap arbitrary pages of a shared backing store
	131	* file within an existing vma.
	132	* @start: start of the remapped virtual memory range
	133	* @size: size of the remapped virtual memory range
	134	* @prot: new protection bits of the range
	135	* @pgoff: to be mapped page of the backing store file
	136	* @flags: 0 or MAP_NONBLOCKED - the later will cause no IO.
	137	*
	138	* this syscall works purely via pagetables, so it's the most efficient
	139	* way to map the same (large) file into a given virtual window. Unlike
	140	* mmap()/mremap() it does not create any new vmas. The new mappings are
	141	* also safe across swapout.
	142	*
	143	* NOTE: the 'prot' parameter right now is ignored, and the vma's default
	144	* protection is used. Arbitrary protections might be implemented in the
	145	* future.
	146	*/
	147	asmlinkage long sys_remap_file_pages(unsigned long start, unsigned long size,
	148	unsigned long __prot, unsigned long pgoff, unsigned long flags)
	149	{
	150	struct mm_struct *mm = current->mm;
	151	struct address_space *mapping;
	152	unsigned long end = start + size;
	153	struct vm_area_struct *vma;
	154	int err = -EINVAL;
	155	int has_write_lock = 0;
	156
	157	if (__prot)
	158	return err;
	159	/*
	160	* Sanitize the syscall parameters:
	161	*/
	162	start = start & PAGE_MASK;
	163	size = size & PAGE_MASK;
	164
	165	/* Does the address range wrap, or is the span zero-sized? */
	166	if (start + size <= start)
	167	return err;
	168
	169	/* Can we represent this offset inside this architecture's pte's? */
	170	#if PTE_FILE_MAX_BITS < BITS_PER_LONG
	171	if (pgoff + (size >> PAGE_SHIFT) >= (1UL << PTE_FILE_MAX_BITS))
	172	return err;
	173	#endif
	174
	175	/* We need down_write() to change vma->vm_flags. */
	176	down_read(&mm->mmap_sem);
	177	retry:
	178	vma = find_vma(mm, start);
	179
	180	/*
	181	* Make sure the vma is shared, that it supports prefaulting,
	182	* and that the remapped range is valid and fully within
	183	* the single existing vma. vm_private_data is used as a
101d2be7	184	* swapout cursor in a VM_NONLINEAR vma.
1da177e4 LT	185	*/
1da177e4 LT	186	if (vma && (vma->vm_flags & VM_SHARED) &&
101d2be7	187	(!vma->vm_private_data \|\| (vma->vm_flags & VM_NONLINEAR)) &&
1da177e4 LT	188	vma->vm_ops && vma->vm_ops->populate &&
	189	end > start && start >= vma->vm_start &&
	190	end <= vma->vm_end) {
	191
	192	/* Must set VM_NONLINEAR before any pages are populated. */
	193	if (pgoff != linear_page_index(vma, start) &&
	194	!(vma->vm_flags & VM_NONLINEAR)) {
	195	if (!has_write_lock) {
	196	up_read(&mm->mmap_sem);
	197	down_write(&mm->mmap_sem);
	198	has_write_lock = 1;
	199	goto retry;
	200	}
	201	mapping = vma->vm_file->f_mapping;
	202	spin_lock(&mapping->i_mmap_lock);
	203	flush_dcache_mmap_lock(mapping);
	204	vma->vm_flags \|= VM_NONLINEAR;
	205	vma_prio_tree_remove(vma, &mapping->i_mmap);
	206	vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
	207	flush_dcache_mmap_unlock(mapping);
	208	spin_unlock(&mapping->i_mmap_lock);
	209	}
	210
	211	err = vma->vm_ops->populate(vma, start, size,
	212	vma->vm_page_prot,
	213	pgoff, flags & MAP_NONBLOCK);
	214
	215	/*
	216	* We can't clear VM_NONLINEAR because we'd have to do
	217	* it after ->populate completes, and that would prevent
	218	* downgrading the lock. (Locks can't be upgraded).
	219	*/
	220	}
	221	if (likely(!has_write_lock))
	222	up_read(&mm->mmap_sem);
	223	else
	224	up_write(&mm->mmap_sem);
	225
	226	return err;
	227	}
	228