[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 <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;

	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);
}

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