[linux.git] / mm / mremap.c

/*
 *	mm/mremap.c
 *
 *	(C) Copyright 1996 Linus Torvalds
 *
 *	Address space accounting code	<[email protected]>
 *	(C) Copyright 2002 Red Hat Inc, All Rights Reserved
 */

#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/slab.h>
#include <linux/shm.h>
#include <linux/mman.h>
#include <linux/swap.h>
#include <linux/capability.h>
#include <linux/fs.h>
#include <linux/highmem.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/mmu_notifier.h>

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

static pmd_t *get_old_pmd(struct mm_struct *mm, unsigned long addr)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;

	pgd = pgd_offset(mm, addr);
	if (pgd_none_or_clear_bad(pgd))
		return NULL;

	pud = pud_offset(pgd, addr);
	if (pud_none_or_clear_bad(pud))
		return NULL;

	pmd = pmd_offset(pud, addr);
	if (pmd_none_or_clear_bad(pmd))
		return NULL;

	return pmd;
}

static pmd_t *alloc_new_pmd(struct mm_struct *mm, unsigned long addr)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;

	pgd = pgd_offset(mm, addr);
	pud = pud_alloc(mm, pgd, addr);
	if (!pud)
		return NULL;

	pmd = pmd_alloc(mm, pud, addr);
	if (!pmd)
		return NULL;

	if (!pmd_present(*pmd) && __pte_alloc(mm, pmd, addr))
		return NULL;

	return pmd;
}

static void move_ptes(struct vm_area_struct *vma, pmd_t *old_pmd,
		unsigned long old_addr, unsigned long old_end,
		struct vm_area_struct *new_vma, pmd_t *new_pmd,
		unsigned long new_addr)
{
	struct address_space *mapping = NULL;
	struct mm_struct *mm = vma->vm_mm;
	pte_t *old_pte, *new_pte, pte;
	spinlock_t *old_ptl, *new_ptl;
	unsigned long old_start;

	old_start = old_addr;
	mmu_notifier_invalidate_range_start(vma->vm_mm,
					    old_start, old_end);
	if (vma->vm_file) {
		/*
		 * Subtle point from Rajesh Venkatasubramanian: before
		 * moving file-based ptes, we must lock vmtruncate out,
		 * since it might clean the dst vma before the src vma,
		 * and we propagate stale pages into the dst afterward.
		 */
		mapping = vma->vm_file->f_mapping;
		spin_lock(&mapping->i_mmap_lock);
		if (new_vma->vm_truncate_count &&
		    new_vma->vm_truncate_count != vma->vm_truncate_count)
			new_vma->vm_truncate_count = 0;
	}

	/*
	 * We don't have to worry about the ordering of src and dst
	 * pte locks because exclusive mmap_sem prevents deadlock.
	 */
	old_pte = pte_offset_map_lock(mm, old_pmd, old_addr, &old_ptl);
 	new_pte = pte_offset_map_nested(new_pmd, new_addr);
	new_ptl = pte_lockptr(mm, new_pmd);
	if (new_ptl != old_ptl)
		spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
	arch_enter_lazy_mmu_mode();

	for (; old_addr < old_end; old_pte++, old_addr += PAGE_SIZE,
				   new_pte++, new_addr += PAGE_SIZE) {
		if (pte_none(*old_pte))
			continue;
		pte = ptep_clear_flush(vma, old_addr, old_pte);
		pte = move_pte(pte, new_vma->vm_page_prot, old_addr, new_addr);
		set_pte_at(mm, new_addr, new_pte, pte);
	}

	arch_leave_lazy_mmu_mode();
	if (new_ptl != old_ptl)
		spin_unlock(new_ptl);
	pte_unmap_nested(new_pte - 1);
	pte_unmap_unlock(old_pte - 1, old_ptl);
	if (mapping)
		spin_unlock(&mapping->i_mmap_lock);
	mmu_notifier_invalidate_range_end(vma->vm_mm, old_start, old_end);
}

#define LATENCY_LIMIT	(64 * PAGE_SIZE)

unsigned long move_page_tables(struct vm_area_struct *vma,
		unsigned long old_addr, struct vm_area_struct *new_vma,
		unsigned long new_addr, unsigned long len)
{
	unsigned long extent, next, old_end;
	pmd_t *old_pmd, *new_pmd;

	old_end = old_addr + len;
	flush_cache_range(vma, old_addr, old_end);

	for (; old_addr < old_end; old_addr += extent, new_addr += extent) {
		cond_resched();
		next = (old_addr + PMD_SIZE) & PMD_MASK;
		if (next - 1 > old_end)
			next = old_end;
		extent = next - old_addr;
		old_pmd = get_old_pmd(vma->vm_mm, old_addr);
		if (!old_pmd)
			continue;
		new_pmd = alloc_new_pmd(vma->vm_mm, new_addr);
		if (!new_pmd)
			break;
		next = (new_addr + PMD_SIZE) & PMD_MASK;
		if (extent > next - new_addr)
			extent = next - new_addr;
		if (extent > LATENCY_LIMIT)
			extent = LATENCY_LIMIT;
		move_ptes(vma, old_pmd, old_addr, old_addr + extent,
				new_vma, new_pmd, new_addr);
	}

	return len + old_addr - old_end;	/* how much done */
}

static unsigned long move_vma(struct vm_area_struct *vma,
		unsigned long old_addr, unsigned long old_len,
		unsigned long new_len, unsigned long new_addr)
{
	struct mm_struct *mm = vma->vm_mm;
	struct vm_area_struct *new_vma;
	unsigned long vm_flags = vma->vm_flags;
	unsigned long new_pgoff;
	unsigned long moved_len;
	unsigned long excess = 0;
	unsigned long hiwater_vm;
	int split = 0;

	/*
	 * We'd prefer to avoid failure later on in do_munmap:
	 * which may split one vma into three before unmapping.
	 */
	if (mm->map_count >= sysctl_max_map_count - 3)
		return -ENOMEM;

	new_pgoff = vma->vm_pgoff + ((old_addr - vma->vm_start) >> PAGE_SHIFT);
	new_vma = copy_vma(&vma, new_addr, new_len, new_pgoff);
	if (!new_vma)
		return -ENOMEM;

	moved_len = move_page_tables(vma, old_addr, new_vma, new_addr, old_len);
	if (moved_len < old_len) {
		/*
		 * On error, move entries back from new area to old,
		 * which will succeed since page tables still there,
		 * and then proceed to unmap new area instead of old.
		 */
		move_page_tables(new_vma, new_addr, vma, old_addr, moved_len);
		vma = new_vma;
		old_len = new_len;
		old_addr = new_addr;
		new_addr = -ENOMEM;
	}

	/* Conceal VM_ACCOUNT so old reservation is not undone */
	if (vm_flags & VM_ACCOUNT) {
		vma->vm_flags &= ~VM_ACCOUNT;
		excess = vma->vm_end - vma->vm_start - old_len;
		if (old_addr > vma->vm_start &&
		    old_addr + old_len < vma->vm_end)
			split = 1;
	}

	/*
	 * If we failed to move page tables we still do total_vm increment
	 * since do_munmap() will decrement it by old_len == new_len.
	 *
	 * Since total_vm is about to be raised artificially high for a
	 * moment, we need to restore high watermark afterwards: if stats
	 * are taken meanwhile, total_vm and hiwater_vm appear too high.
	 * If this were a serious issue, we'd add a flag to do_munmap().
	 */
	hiwater_vm = mm->hiwater_vm;
	mm->total_vm += new_len >> PAGE_SHIFT;
	vm_stat_account(mm, vma->vm_flags, vma->vm_file, new_len>>PAGE_SHIFT);

	if (do_munmap(mm, old_addr, old_len) < 0) {
		/* OOM: unable to split vma, just get accounts right */
		vm_unacct_memory(excess >> PAGE_SHIFT);
		excess = 0;
	}
	mm->hiwater_vm = hiwater_vm;

	/* Restore VM_ACCOUNT if one or two pieces of vma left */
	if (excess) {
		vma->vm_flags |= VM_ACCOUNT;
		if (split)
			vma->vm_next->vm_flags |= VM_ACCOUNT;
	}

	if (vm_flags & VM_LOCKED) {
		mm->locked_vm += new_len >> PAGE_SHIFT;
		if (new_len > old_len)
			make_pages_present(new_addr + old_len,
					   new_addr + new_len);
	}

	return new_addr;
}

/*
 * Expand (or shrink) an existing mapping, potentially moving it at the
 * same time (controlled by the MREMAP_MAYMOVE flag and available VM space)
 *
 * MREMAP_FIXED option added 5-Dec-1999 by Benjamin LaHaise
 * This option implies MREMAP_MAYMOVE.
 */
unsigned long do_mremap(unsigned long addr,
	unsigned long old_len, unsigned long new_len,
	unsigned long flags, unsigned long new_addr)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	unsigned long ret = -EINVAL;
	unsigned long charged = 0;

	if (flags & ~(MREMAP_FIXED | MREMAP_MAYMOVE))
		goto out;

	if (addr & ~PAGE_MASK)
		goto out;

	old_len = PAGE_ALIGN(old_len);
	new_len = PAGE_ALIGN(new_len);

	/*
	 * We allow a zero old-len as a special case
	 * for DOS-emu "duplicate shm area" thing. But
	 * a zero new-len is nonsensical.
	 */
	if (!new_len)
		goto out;

	/* new_addr is only valid if MREMAP_FIXED is specified */
	if (flags & MREMAP_FIXED) {
		if (new_addr & ~PAGE_MASK)
			goto out;
		if (!(flags & MREMAP_MAYMOVE))
			goto out;

		if (new_len > TASK_SIZE || new_addr > TASK_SIZE - new_len)
			goto out;

		/* Check if the location we're moving into overlaps the
		 * old location at all, and fail if it does.
		 */
		if ((new_addr <= addr) && (new_addr+new_len) > addr)
			goto out;

		if ((addr <= new_addr) && (addr+old_len) > new_addr)
			goto out;

		ret = security_file_mmap(NULL, 0, 0, 0, new_addr, 1);
		if (ret)
			goto out;

		ret = do_munmap(mm, new_addr, new_len);
		if (ret)
			goto out;
	}

	/*
	 * Always allow a shrinking remap: that just unmaps
	 * the unnecessary pages..
	 * do_munmap does all the needed commit accounting
	 */
	if (old_len >= new_len) {
		ret = do_munmap(mm, addr+new_len, old_len - new_len);
		if (ret && old_len != new_len)
			goto out;
		ret = addr;
		if (!(flags & MREMAP_FIXED) || (new_addr == addr))
			goto out;
		old_len = new_len;
	}

	/*
	 * Ok, we need to grow..  or relocate.
	 */
	ret = -EFAULT;
	vma = find_vma(mm, addr);
	if (!vma || vma->vm_start > addr)
		goto out;
	if (is_vm_hugetlb_page(vma)) {
		ret = -EINVAL;
		goto out;
	}
	/* We can't remap across vm area boundaries */
	if (old_len > vma->vm_end - addr)
		goto out;
	if (vma->vm_flags & (VM_DONTEXPAND | VM_PFNMAP)) {
		if (new_len > old_len)
			goto out;
	}
	if (vma->vm_flags & VM_LOCKED) {
		unsigned long locked, lock_limit;
		locked = mm->locked_vm << PAGE_SHIFT;
		lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur;
		locked += new_len - old_len;
		ret = -EAGAIN;
		if (locked > lock_limit && !capable(CAP_IPC_LOCK))
			goto out;
	}
	if (!may_expand_vm(mm, (new_len - old_len) >> PAGE_SHIFT)) {
		ret = -ENOMEM;
		goto out;
	}

	if (vma->vm_flags & VM_ACCOUNT) {
		charged = (new_len - old_len) >> PAGE_SHIFT;
		if (security_vm_enough_memory(charged))
			goto out_nc;
	}

	/* old_len exactly to the end of the area..
	 * And we're not relocating the area.
	 */
	if (old_len == vma->vm_end - addr &&
	    !((flags & MREMAP_FIXED) && (addr != new_addr)) &&
	    (old_len != new_len || !(flags & MREMAP_MAYMOVE))) {
		unsigned long max_addr = TASK_SIZE;
		if (vma->vm_next)
			max_addr = vma->vm_next->vm_start;
		/* can we just expand the current mapping? */
		if (max_addr - addr >= new_len) {
			int pages = (new_len - old_len) >> PAGE_SHIFT;

			vma_adjust(vma, vma->vm_start,
				addr + new_len, vma->vm_pgoff, NULL);

			mm->total_vm += pages;
			vm_stat_account(mm, vma->vm_flags, vma->vm_file, pages);
			if (vma->vm_flags & VM_LOCKED) {
				mm->locked_vm += pages;
				make_pages_present(addr + old_len,
						   addr + new_len);
			}
			ret = addr;
			goto out;
		}
	}

	/*
	 * We weren't able to just expand or shrink the area,
	 * we need to create a new one and move it..
	 */
	ret = -ENOMEM;
	if (flags & MREMAP_MAYMOVE) {
		if (!(flags & MREMAP_FIXED)) {
			unsigned long map_flags = 0;
			if (vma->vm_flags & VM_MAYSHARE)
				map_flags |= MAP_SHARED;

			new_addr = get_unmapped_area(vma->vm_file, 0, new_len,
						vma->vm_pgoff, map_flags);
			if (new_addr & ~PAGE_MASK) {
				ret = new_addr;
				goto out;
			}

			ret = security_file_mmap(NULL, 0, 0, 0, new_addr, 1);
			if (ret)
				goto out;
		}
		ret = move_vma(vma, addr, old_len, new_len, new_addr);
	}
out:
	if (ret & ~PAGE_MASK)
		vm_unacct_memory(charged);
out_nc:
	return ret;
}

asmlinkage unsigned long sys_mremap(unsigned long addr,
	unsigned long old_len, unsigned long new_len,
	unsigned long flags, unsigned long new_addr)
{
	unsigned long ret;

	down_write(&current->mm->mmap_sem);
	ret = do_mremap(addr, old_len, new_len, flags, new_addr);
	up_write(&current->mm->mmap_sem);
	return ret;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* mm/mremap.c
	3	*
	4	* (C) Copyright 1996 Linus Torvalds
	5	*
	6	* Address space accounting code <[email protected]>
	7	* (C) Copyright 2002 Red Hat Inc, All Rights Reserved
	8	*/
	9
	10	#include <linux/mm.h>
	11	#include <linux/hugetlb.h>
	12	#include <linux/slab.h>
	13	#include <linux/shm.h>
	14	#include <linux/mman.h>
	15	#include <linux/swap.h>
c59ede7b	16	#include <linux/capability.h>
1da177e4 LT	17	#include <linux/fs.h>
	18	#include <linux/highmem.h>
	19	#include <linux/security.h>
	20	#include <linux/syscalls.h>
cddb8a5c	21	#include <linux/mmu_notifier.h>
1da177e4 LT	22
	23	#include <asm/uaccess.h>
	24	#include <asm/cacheflush.h>
	25	#include <asm/tlbflush.h>
	26
7be7a546	27	static pmd_t get_old_pmd(struct mm_struct mm, unsigned long addr)
1da177e4 LT	28	{
	29	pgd_t *pgd;
	30	pud_t *pud;
	31	pmd_t *pmd;
	32
	33	pgd = pgd_offset(mm, addr);
	34	if (pgd_none_or_clear_bad(pgd))
	35	return NULL;
	36
	37	pud = pud_offset(pgd, addr);
	38	if (pud_none_or_clear_bad(pud))
	39	return NULL;
	40
	41	pmd = pmd_offset(pud, addr);
	42	if (pmd_none_or_clear_bad(pmd))
	43	return NULL;
	44
7be7a546	45	return pmd;
1da177e4 LT	46	}
1da177e4 LT	47
7be7a546	48	static pmd_t alloc_new_pmd(struct mm_struct mm, unsigned long addr)
1da177e4 LT	49	{
	50	pgd_t *pgd;
	51	pud_t *pud;
c74df32c	52	pmd_t *pmd;
1da177e4 LT	53
1da177e4 LT	54	pgd = pgd_offset(mm, addr);
1da177e4 LT	55	pud = pud_alloc(mm, pgd, addr);
1da177e4 LT	56	if (!pud)
c74df32c	57	return NULL;
7be7a546	58
1da177e4	59	pmd = pmd_alloc(mm, pud, addr);
7be7a546	60	if (!pmd)
c74df32c	61	return NULL;
7be7a546	62
1bb3630e	63	if (!pmd_present(*pmd) && __pte_alloc(mm, pmd, addr))
c74df32c HD	64	return NULL;
c74df32c HD	65
7be7a546	66	return pmd;
1da177e4 LT	67	}
1da177e4 LT	68
7be7a546 HD	69	static void move_ptes(struct vm_area_struct vma, pmd_t old_pmd,
	70	unsigned long old_addr, unsigned long old_end,
	71	struct vm_area_struct new_vma, pmd_t new_pmd,
	72	unsigned long new_addr)
1da177e4 LT	73	{
	74	struct address_space *mapping = NULL;
	75	struct mm_struct *mm = vma->vm_mm;
7be7a546	76	pte_t old_pte, new_pte, pte;
4c21e2f2	77	spinlock_t old_ptl, new_ptl;
cddb8a5c	78	unsigned long old_start;
1da177e4	79
cddb8a5c AA	80	old_start = old_addr;
	81	mmu_notifier_invalidate_range_start(vma->vm_mm,
	82	old_start, old_end);
1da177e4 LT	83	if (vma->vm_file) {
	84	/*
	85	* Subtle point from Rajesh Venkatasubramanian: before
	86	* moving file-based ptes, we must lock vmtruncate out,
	87	* since it might clean the dst vma before the src vma,
	88	* and we propagate stale pages into the dst afterward.
	89	*/
	90	mapping = vma->vm_file->f_mapping;
	91	spin_lock(&mapping->i_mmap_lock);
	92	if (new_vma->vm_truncate_count &&
	93	new_vma->vm_truncate_count != vma->vm_truncate_count)
	94	new_vma->vm_truncate_count = 0;
	95	}
1da177e4	96
4c21e2f2 HD	97	/*
	98	* We don't have to worry about the ordering of src and dst
	99	* pte locks because exclusive mmap_sem prevents deadlock.
	100	*/
c74df32c HD	101	old_pte = pte_offset_map_lock(mm, old_pmd, old_addr, &old_ptl);
c74df32c HD	102	new_pte = pte_offset_map_nested(new_pmd, new_addr);
4c21e2f2 HD	103	new_ptl = pte_lockptr(mm, new_pmd);
4c21e2f2 HD	104	if (new_ptl != old_ptl)
f20dc5f7	105	spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
6606c3e0	106	arch_enter_lazy_mmu_mode();
7be7a546 HD	107
	108	for (; old_addr < old_end; old_pte++, old_addr += PAGE_SIZE,
	109	new_pte++, new_addr += PAGE_SIZE) {
	110	if (pte_none(*old_pte))
	111	continue;
	112	pte = ptep_clear_flush(vma, old_addr, old_pte);
7be7a546 HD	113	pte = move_pte(pte, new_vma->vm_page_prot, old_addr, new_addr);
7be7a546 HD	114	set_pte_at(mm, new_addr, new_pte, pte);
1da177e4	115	}
7be7a546	116
6606c3e0	117	arch_leave_lazy_mmu_mode();
4c21e2f2 HD	118	if (new_ptl != old_ptl)
4c21e2f2 HD	119	spin_unlock(new_ptl);
7be7a546	120	pte_unmap_nested(new_pte - 1);
c74df32c	121	pte_unmap_unlock(old_pte - 1, old_ptl);
1da177e4 LT	122	if (mapping)
1da177e4 LT	123	spin_unlock(&mapping->i_mmap_lock);
cddb8a5c	124	mmu_notifier_invalidate_range_end(vma->vm_mm, old_start, old_end);
1da177e4 LT	125	}
1da177e4 LT	126
7be7a546 HD	127	#define LATENCY_LIMIT (64 * PAGE_SIZE)
7be7a546 HD	128
b6a2fea3	129	unsigned long move_page_tables(struct vm_area_struct *vma,
1da177e4 LT	130	unsigned long old_addr, struct vm_area_struct *new_vma,
	131	unsigned long new_addr, unsigned long len)
	132	{
7be7a546 HD	133	unsigned long extent, next, old_end;
7be7a546 HD	134	pmd_t old_pmd, new_pmd;
1da177e4	135
7be7a546 HD	136	old_end = old_addr + len;
7be7a546 HD	137	flush_cache_range(vma, old_addr, old_end);
1da177e4	138
7be7a546	139	for (; old_addr < old_end; old_addr += extent, new_addr += extent) {
1da177e4	140	cond_resched();
7be7a546 HD	141	next = (old_addr + PMD_SIZE) & PMD_MASK;
	142	if (next - 1 > old_end)
	143	next = old_end;
	144	extent = next - old_addr;
	145	old_pmd = get_old_pmd(vma->vm_mm, old_addr);
	146	if (!old_pmd)
	147	continue;
	148	new_pmd = alloc_new_pmd(vma->vm_mm, new_addr);
	149	if (!new_pmd)
	150	break;
	151	next = (new_addr + PMD_SIZE) & PMD_MASK;
	152	if (extent > next - new_addr)
	153	extent = next - new_addr;
	154	if (extent > LATENCY_LIMIT)
	155	extent = LATENCY_LIMIT;
	156	move_ptes(vma, old_pmd, old_addr, old_addr + extent,
	157	new_vma, new_pmd, new_addr);
1da177e4	158	}
7be7a546 HD	159
7be7a546 HD	160	return len + old_addr - old_end; /* how much done */
1da177e4 LT	161	}
	162
	163	static unsigned long move_vma(struct vm_area_struct *vma,
	164	unsigned long old_addr, unsigned long old_len,
	165	unsigned long new_len, unsigned long new_addr)
	166	{
	167	struct mm_struct *mm = vma->vm_mm;
	168	struct vm_area_struct *new_vma;
	169	unsigned long vm_flags = vma->vm_flags;
	170	unsigned long new_pgoff;
	171	unsigned long moved_len;
	172	unsigned long excess = 0;
365e9c87	173	unsigned long hiwater_vm;
1da177e4 LT	174	int split = 0;
	175
	176	/*
	177	* We'd prefer to avoid failure later on in do_munmap:
	178	* which may split one vma into three before unmapping.
	179	*/
	180	if (mm->map_count >= sysctl_max_map_count - 3)
	181	return -ENOMEM;
	182
	183	new_pgoff = vma->vm_pgoff + ((old_addr - vma->vm_start) >> PAGE_SHIFT);
	184	new_vma = copy_vma(&vma, new_addr, new_len, new_pgoff);
	185	if (!new_vma)
	186	return -ENOMEM;
	187
	188	moved_len = move_page_tables(vma, old_addr, new_vma, new_addr, old_len);
	189	if (moved_len < old_len) {
	190	/*
	191	* On error, move entries back from new area to old,
	192	* which will succeed since page tables still there,
	193	* and then proceed to unmap new area instead of old.
	194	*/
	195	move_page_tables(new_vma, new_addr, vma, old_addr, moved_len);
	196	vma = new_vma;
	197	old_len = new_len;
	198	old_addr = new_addr;
	199	new_addr = -ENOMEM;
	200	}
	201
	202	/* Conceal VM_ACCOUNT so old reservation is not undone */
	203	if (vm_flags & VM_ACCOUNT) {
	204	vma->vm_flags &= ~VM_ACCOUNT;
	205	excess = vma->vm_end - vma->vm_start - old_len;
	206	if (old_addr > vma->vm_start &&
	207	old_addr + old_len < vma->vm_end)
	208	split = 1;
	209	}
	210
71799062	211	/*
365e9c87 HD	212	* If we failed to move page tables we still do total_vm increment
	213	* since do_munmap() will decrement it by old_len == new_len.
	214	*
	215	* Since total_vm is about to be raised artificially high for a
	216	* moment, we need to restore high watermark afterwards: if stats
	217	* are taken meanwhile, total_vm and hiwater_vm appear too high.
	218	* If this were a serious issue, we'd add a flag to do_munmap().
71799062	219	*/
365e9c87	220	hiwater_vm = mm->hiwater_vm;
71799062	221	mm->total_vm += new_len >> PAGE_SHIFT;
ab50b8ed	222	vm_stat_account(mm, vma->vm_flags, vma->vm_file, new_len>>PAGE_SHIFT);
71799062	223
1da177e4 LT	224	if (do_munmap(mm, old_addr, old_len) < 0) {
	225	/* OOM: unable to split vma, just get accounts right */
	226	vm_unacct_memory(excess >> PAGE_SHIFT);
	227	excess = 0;
	228	}
365e9c87	229	mm->hiwater_vm = hiwater_vm;
1da177e4 LT	230
	231	/* Restore VM_ACCOUNT if one or two pieces of vma left */
	232	if (excess) {
	233	vma->vm_flags \|= VM_ACCOUNT;
	234	if (split)
	235	vma->vm_next->vm_flags \|= VM_ACCOUNT;
	236	}
	237
1da177e4 LT	238	if (vm_flags & VM_LOCKED) {
	239	mm->locked_vm += new_len >> PAGE_SHIFT;
	240	if (new_len > old_len)
	241	make_pages_present(new_addr + old_len,
	242	new_addr + new_len);
	243	}
	244
	245	return new_addr;
	246	}
	247
	248	/*
	249	* Expand (or shrink) an existing mapping, potentially moving it at the
	250	* same time (controlled by the MREMAP_MAYMOVE flag and available VM space)
	251	*
	252	* MREMAP_FIXED option added 5-Dec-1999 by Benjamin LaHaise
	253	* This option implies MREMAP_MAYMOVE.
	254	*/
	255	unsigned long do_mremap(unsigned long addr,
	256	unsigned long old_len, unsigned long new_len,
	257	unsigned long flags, unsigned long new_addr)
	258	{
d0de32d9	259	struct mm_struct *mm = current->mm;
1da177e4 LT	260	struct vm_area_struct *vma;
	261	unsigned long ret = -EINVAL;
	262	unsigned long charged = 0;
	263
	264	if (flags & ~(MREMAP_FIXED \| MREMAP_MAYMOVE))
	265	goto out;
	266
	267	if (addr & ~PAGE_MASK)
	268	goto out;
	269
	270	old_len = PAGE_ALIGN(old_len);
	271	new_len = PAGE_ALIGN(new_len);
	272
	273	/*
	274	* We allow a zero old-len as a special case
	275	* for DOS-emu "duplicate shm area" thing. But
	276	* a zero new-len is nonsensical.
	277	*/
	278	if (!new_len)
	279	goto out;
	280
	281	/* new_addr is only valid if MREMAP_FIXED is specified */
	282	if (flags & MREMAP_FIXED) {
	283	if (new_addr & ~PAGE_MASK)
	284	goto out;
	285	if (!(flags & MREMAP_MAYMOVE))
	286	goto out;
	287
	288	if (new_len > TASK_SIZE \|\| new_addr > TASK_SIZE - new_len)
	289	goto out;
	290
	291	/* Check if the location we're moving into overlaps the
	292	* old location at all, and fail if it does.
	293	*/
	294	if ((new_addr <= addr) && (new_addr+new_len) > addr)
	295	goto out;
	296
	297	if ((addr <= new_addr) && (addr+old_len) > new_addr)
	298	goto out;
	299
c80544dc	300	ret = security_file_mmap(NULL, 0, 0, 0, new_addr, 1);
ed032189 EP	301	if (ret)
	302	goto out;
	303
d0de32d9	304	ret = do_munmap(mm, new_addr, new_len);
1da177e4 LT	305	if (ret)
	306	goto out;
	307	}
	308
	309	/*
	310	* Always allow a shrinking remap: that just unmaps
	311	* the unnecessary pages..
	312	* do_munmap does all the needed commit accounting
	313	*/
	314	if (old_len >= new_len) {
d0de32d9	315	ret = do_munmap(mm, addr+new_len, old_len - new_len);
1da177e4 LT	316	if (ret && old_len != new_len)
	317	goto out;
	318	ret = addr;
	319	if (!(flags & MREMAP_FIXED) \|\| (new_addr == addr))
	320	goto out;
	321	old_len = new_len;
	322	}
	323
	324	/*
	325	* Ok, we need to grow.. or relocate.
	326	*/
	327	ret = -EFAULT;
d0de32d9	328	vma = find_vma(mm, addr);
1da177e4 LT	329	if (!vma \|\| vma->vm_start > addr)
	330	goto out;
	331	if (is_vm_hugetlb_page(vma)) {
	332	ret = -EINVAL;
	333	goto out;
	334	}
	335	/* We can't remap across vm area boundaries */
	336	if (old_len > vma->vm_end - addr)
	337	goto out;
4d7672b4	338	if (vma->vm_flags & (VM_DONTEXPAND \| VM_PFNMAP)) {
1da177e4 LT	339	if (new_len > old_len)
	340	goto out;
	341	}
	342	if (vma->vm_flags & VM_LOCKED) {
	343	unsigned long locked, lock_limit;
d0de32d9	344	locked = mm->locked_vm << PAGE_SHIFT;
1da177e4 LT	345	lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur;
	346	locked += new_len - old_len;
	347	ret = -EAGAIN;
	348	if (locked > lock_limit && !capable(CAP_IPC_LOCK))
	349	goto out;
	350	}
d0de32d9	351	if (!may_expand_vm(mm, (new_len - old_len) >> PAGE_SHIFT)) {
119f657c	352	ret = -ENOMEM;
1da177e4	353	goto out;
119f657c	354	}
1da177e4 LT	355
	356	if (vma->vm_flags & VM_ACCOUNT) {
	357	charged = (new_len - old_len) >> PAGE_SHIFT;
	358	if (security_vm_enough_memory(charged))
	359	goto out_nc;
	360	}
	361
	362	/* old_len exactly to the end of the area..
	363	* And we're not relocating the area.
	364	*/
	365	if (old_len == vma->vm_end - addr &&
	366	!((flags & MREMAP_FIXED) && (addr != new_addr)) &&
	367	(old_len != new_len \|\| !(flags & MREMAP_MAYMOVE))) {
	368	unsigned long max_addr = TASK_SIZE;
	369	if (vma->vm_next)
	370	max_addr = vma->vm_next->vm_start;
	371	/* can we just expand the current mapping? */
	372	if (max_addr - addr >= new_len) {
	373	int pages = (new_len - old_len) >> PAGE_SHIFT;
	374
	375	vma_adjust(vma, vma->vm_start,
	376	addr + new_len, vma->vm_pgoff, NULL);
	377
d0de32d9 HD	378	mm->total_vm += pages;
d0de32d9 HD	379	vm_stat_account(mm, vma->vm_flags, vma->vm_file, pages);
1da177e4	380	if (vma->vm_flags & VM_LOCKED) {
d0de32d9	381	mm->locked_vm += pages;
1da177e4 LT	382	make_pages_present(addr + old_len,
	383	addr + new_len);
	384	}
	385	ret = addr;
	386	goto out;
	387	}
	388	}
	389
	390	/*
	391	* We weren't able to just expand or shrink the area,
	392	* we need to create a new one and move it..
	393	*/
	394	ret = -ENOMEM;
	395	if (flags & MREMAP_MAYMOVE) {
	396	if (!(flags & MREMAP_FIXED)) {
	397	unsigned long map_flags = 0;
	398	if (vma->vm_flags & VM_MAYSHARE)
	399	map_flags \|= MAP_SHARED;
	400
	401	new_addr = get_unmapped_area(vma->vm_file, 0, new_len,
	402	vma->vm_pgoff, map_flags);
ed032189 EP	403	if (new_addr & ~PAGE_MASK) {
	404	ret = new_addr;
	405	goto out;
	406	}
	407
c80544dc	408	ret = security_file_mmap(NULL, 0, 0, 0, new_addr, 1);
ed032189	409	if (ret)
1da177e4 LT	410	goto out;
	411	}
	412	ret = move_vma(vma, addr, old_len, new_len, new_addr);
	413	}
	414	out:
	415	if (ret & ~PAGE_MASK)
	416	vm_unacct_memory(charged);
	417	out_nc:
	418	return ret;
	419	}
	420
	421	asmlinkage unsigned long sys_mremap(unsigned long addr,
	422	unsigned long old_len, unsigned long new_len,
	423	unsigned long flags, unsigned long new_addr)
	424	{
	425	unsigned long ret;
	426
	427	down_write(&current->mm->mmap_sem);
	428	ret = do_mremap(addr, old_len, new_len, flags, new_addr);
	429	up_write(&current->mm->mmap_sem);
	430	return ret;
	431	}