[deliverable/linux.git] / arch / um / kernel / tlb.c

/*
 * Copyright (C) 2000 - 2007 Jeff Dike (jdike@{addtoit,linux.intel}.com)
 * Licensed under the GPL
 */

#include <linux/mm.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
#include <as-layout.h>
#include <mem_user.h>
#include <os.h>
#include <skas.h>

struct host_vm_change {
	struct host_vm_op {
		enum { NONE, MMAP, MUNMAP, MPROTECT } type;
		union {
			struct {
				unsigned long addr;
				unsigned long len;
				unsigned int prot;
				int fd;
				__u64 offset;
			} mmap;
			struct {
				unsigned long addr;
				unsigned long len;
			} munmap;
			struct {
				unsigned long addr;
				unsigned long len;
				unsigned int prot;
			} mprotect;
		} u;
	} ops[1];
	int index;
	struct mm_id *id;
	void *data;
	int force;
};

#define INIT_HVC(mm, force) \
	((struct host_vm_change) \
	 { .ops		= { { .type = NONE } },	\
	   .id		= &mm->context.id, \
       	   .data	= NULL, \
	   .index	= 0, \
	   .force	= force })

static int do_ops(struct host_vm_change *hvc, int end,
		  int finished)
{
	struct host_vm_op *op;
	int i, ret = 0;

	for (i = 0; i < end && !ret; i++) {
		op = &hvc->ops[i];
		switch (op->type) {
		case MMAP:
			ret = map(hvc->id, op->u.mmap.addr, op->u.mmap.len,
				  op->u.mmap.prot, op->u.mmap.fd,
				  op->u.mmap.offset, finished, &hvc->data);
			break;
		case MUNMAP:
			ret = unmap(hvc->id, op->u.munmap.addr,
				    op->u.munmap.len, finished, &hvc->data);
			break;
		case MPROTECT:
			ret = protect(hvc->id, op->u.mprotect.addr,
				      op->u.mprotect.len, op->u.mprotect.prot,
				      finished, &hvc->data);
			break;
		default:
			printk(KERN_ERR "Unknown op type %d in do_ops\n",
			       op->type);
			BUG();
			break;
		}
	}

	return ret;
}

static int add_mmap(unsigned long virt, unsigned long phys, unsigned long len,
		    unsigned int prot, struct host_vm_change *hvc)
{
	__u64 offset;
	struct host_vm_op *last;
	int fd, ret = 0;

	fd = phys_mapping(phys, &offset);
	if (hvc->index != 0) {
		last = &hvc->ops[hvc->index - 1];
		if ((last->type == MMAP) &&
		   (last->u.mmap.addr + last->u.mmap.len == virt) &&
		   (last->u.mmap.prot == prot) && (last->u.mmap.fd == fd) &&
		   (last->u.mmap.offset + last->u.mmap.len == offset)) {
			last->u.mmap.len += len;
			return 0;
		}
	}

	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
		ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
		hvc->index = 0;
	}

	hvc->ops[hvc->index++] = ((struct host_vm_op)
				  { .type	= MMAP,
				    .u = { .mmap = { .addr	= virt,
						     .len	= len,
						     .prot	= prot,
						     .fd	= fd,
						     .offset	= offset }
			   } });
	return ret;
}

static int add_munmap(unsigned long addr, unsigned long len,
		      struct host_vm_change *hvc)
{
	struct host_vm_op *last;
	int ret = 0;

	if (hvc->index != 0) {
		last = &hvc->ops[hvc->index - 1];
		if ((last->type == MUNMAP) &&
		   (last->u.munmap.addr + last->u.mmap.len == addr)) {
			last->u.munmap.len += len;
			return 0;
		}
	}

	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
		ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
		hvc->index = 0;
	}

	hvc->ops[hvc->index++] = ((struct host_vm_op)
				  { .type	= MUNMAP,
			     	    .u = { .munmap = { .addr	= addr,
						       .len	= len } } });
	return ret;
}

static int add_mprotect(unsigned long addr, unsigned long len,
			unsigned int prot, struct host_vm_change *hvc)
{
	struct host_vm_op *last;
	int ret = 0;

	if (hvc->index != 0) {
		last = &hvc->ops[hvc->index - 1];
		if ((last->type == MPROTECT) &&
		   (last->u.mprotect.addr + last->u.mprotect.len == addr) &&
		   (last->u.mprotect.prot == prot)) {
			last->u.mprotect.len += len;
			return 0;
		}
	}

	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
		ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
		hvc->index = 0;
	}

	hvc->ops[hvc->index++] = ((struct host_vm_op)
				  { .type	= MPROTECT,
			     	    .u = { .mprotect = { .addr	= addr,
							 .len	= len,
							 .prot	= prot } } });
	return ret;
}

#define ADD_ROUND(n, inc) (((n) + (inc)) & ~((inc) - 1))

static inline int update_pte_range(pmd_t *pmd, unsigned long addr,
				   unsigned long end,
				   struct host_vm_change *hvc)
{
	pte_t *pte;
	int r, w, x, prot, ret = 0;

	pte = pte_offset_kernel(pmd, addr);
	do {
		if ((addr >= STUB_START) && (addr < STUB_END))
			continue;

		r = pte_read(*pte);
		w = pte_write(*pte);
		x = pte_exec(*pte);
		if (!pte_young(*pte)) {
			r = 0;
			w = 0;
		} else if (!pte_dirty(*pte))
			w = 0;

		prot = ((r ? UM_PROT_READ : 0) | (w ? UM_PROT_WRITE : 0) |
			(x ? UM_PROT_EXEC : 0));
		if (hvc->force || pte_newpage(*pte)) {
			if (pte_present(*pte))
				ret = add_mmap(addr, pte_val(*pte) & PAGE_MASK,
					       PAGE_SIZE, prot, hvc);
			else
				ret = add_munmap(addr, PAGE_SIZE, hvc);
		} else if (pte_newprot(*pte))
			ret = add_mprotect(addr, PAGE_SIZE, prot, hvc);
		*pte = pte_mkuptodate(*pte);
	} while (pte++, addr += PAGE_SIZE, ((addr < end) && !ret));
	return ret;
}

static inline int update_pmd_range(pud_t *pud, unsigned long addr,
				   unsigned long end,
				   struct host_vm_change *hvc)
{
	pmd_t *pmd;
	unsigned long next;
	int ret = 0;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (!pmd_present(*pmd)) {
			if (hvc->force || pmd_newpage(*pmd)) {
				ret = add_munmap(addr, next - addr, hvc);
				pmd_mkuptodate(*pmd);
			}
		}
		else ret = update_pte_range(pmd, addr, next, hvc);
	} while (pmd++, addr = next, ((addr < end) && !ret));
	return ret;
}

static inline int update_pud_range(pgd_t *pgd, unsigned long addr,
				   unsigned long end,
				   struct host_vm_change *hvc)
{
	pud_t *pud;
	unsigned long next;
	int ret = 0;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (!pud_present(*pud)) {
			if (hvc->force || pud_newpage(*pud)) {
				ret = add_munmap(addr, next - addr, hvc);
				pud_mkuptodate(*pud);
			}
		}
		else ret = update_pmd_range(pud, addr, next, hvc);
	} while (pud++, addr = next, ((addr < end) && !ret));
	return ret;
}

void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
		      unsigned long end_addr, int force)
{
	pgd_t *pgd;
	struct host_vm_change hvc;
	unsigned long addr = start_addr, next;
	int ret = 0;

	hvc = INIT_HVC(mm, force);
	pgd = pgd_offset(mm, addr);
	do {
		next = pgd_addr_end(addr, end_addr);
		if (!pgd_present(*pgd)) {
			if (force || pgd_newpage(*pgd)) {
				ret = add_munmap(addr, next - addr, &hvc);
				pgd_mkuptodate(*pgd);
			}
		}
		else ret = update_pud_range(pgd, addr, next, &hvc);
	} while (pgd++, addr = next, ((addr < end_addr) && !ret));

	if (!ret)
		ret = do_ops(&hvc, hvc.index, 1);

	/* This is not an else because ret is modified above */
	if (ret) {
		printk(KERN_ERR "fix_range_common: failed, killing current "
		       "process\n");
		force_sig(SIGKILL, current);
	}
}

static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
{
	struct mm_struct *mm;
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
	unsigned long addr, last;
	int updated = 0, err;

	mm = &init_mm;
	for (addr = start; addr < end;) {
		pgd = pgd_offset(mm, addr);
		if (!pgd_present(*pgd)) {
			last = ADD_ROUND(addr, PGDIR_SIZE);
			if (last > end)
				last = end;
			if (pgd_newpage(*pgd)) {
				updated = 1;
				err = os_unmap_memory((void *) addr,
						      last - addr);
				if (err < 0)
					panic("munmap failed, errno = %d\n",
					      -err);
			}
			addr = last;
			continue;
		}

		pud = pud_offset(pgd, addr);
		if (!pud_present(*pud)) {
			last = ADD_ROUND(addr, PUD_SIZE);
			if (last > end)
				last = end;
			if (pud_newpage(*pud)) {
				updated = 1;
				err = os_unmap_memory((void *) addr,
						      last - addr);
				if (err < 0)
					panic("munmap failed, errno = %d\n",
					      -err);
			}
			addr = last;
			continue;
		}

		pmd = pmd_offset(pud, addr);
		if (!pmd_present(*pmd)) {
			last = ADD_ROUND(addr, PMD_SIZE);
			if (last > end)
				last = end;
			if (pmd_newpage(*pmd)) {
				updated = 1;
				err = os_unmap_memory((void *) addr,
						      last - addr);
				if (err < 0)
					panic("munmap failed, errno = %d\n",
					      -err);
			}
			addr = last;
			continue;
		}

		pte = pte_offset_kernel(pmd, addr);
		if (!pte_present(*pte) || pte_newpage(*pte)) {
			updated = 1;
			err = os_unmap_memory((void *) addr,
					      PAGE_SIZE);
			if (err < 0)
				panic("munmap failed, errno = %d\n",
				      -err);
			if (pte_present(*pte))
				map_memory(addr,
					   pte_val(*pte) & PAGE_MASK,
					   PAGE_SIZE, 1, 1, 1);
		}
		else if (pte_newprot(*pte)) {
			updated = 1;
			os_protect_memory((void *) addr, PAGE_SIZE, 1, 1, 1);
		}
		addr += PAGE_SIZE;
	}
	return updated;
}

void flush_tlb_page(struct vm_area_struct *vma, unsigned long address)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
	struct mm_struct *mm = vma->vm_mm;
	void *flush = NULL;
	int r, w, x, prot, err = 0;
	struct mm_id *mm_id;

	address &= PAGE_MASK;
	pgd = pgd_offset(mm, address);
	if (!pgd_present(*pgd))
		goto kill;

	pud = pud_offset(pgd, address);
	if (!pud_present(*pud))
		goto kill;

	pmd = pmd_offset(pud, address);
	if (!pmd_present(*pmd))
		goto kill;

	pte = pte_offset_kernel(pmd, address);

	r = pte_read(*pte);
	w = pte_write(*pte);
	x = pte_exec(*pte);
	if (!pte_young(*pte)) {
		r = 0;
		w = 0;
	} else if (!pte_dirty(*pte)) {
		w = 0;
	}

	mm_id = &mm->context.id;
	prot = ((r ? UM_PROT_READ : 0) | (w ? UM_PROT_WRITE : 0) |
		(x ? UM_PROT_EXEC : 0));
	if (pte_newpage(*pte)) {
		if (pte_present(*pte)) {
			unsigned long long offset;
			int fd;

			fd = phys_mapping(pte_val(*pte) & PAGE_MASK, &offset);
			err = map(mm_id, address, PAGE_SIZE, prot, fd, offset,
				  1, &flush);
		}
		else err = unmap(mm_id, address, PAGE_SIZE, 1, &flush);
	}
	else if (pte_newprot(*pte))
		err = protect(mm_id, address, PAGE_SIZE, prot, 1, &flush);

	if (err)
		goto kill;

	*pte = pte_mkuptodate(*pte);

	return;

kill:
	printk(KERN_ERR "Failed to flush page for address 0x%lx\n", address);
	force_sig(SIGKILL, current);
}

pgd_t *pgd_offset_proc(struct mm_struct *mm, unsigned long address)
{
	return pgd_offset(mm, address);
}

pud_t *pud_offset_proc(pgd_t *pgd, unsigned long address)
{
	return pud_offset(pgd, address);
}

pmd_t *pmd_offset_proc(pud_t *pud, unsigned long address)
{
	return pmd_offset(pud, address);
}

pte_t *pte_offset_proc(pmd_t *pmd, unsigned long address)
{
	return pte_offset_kernel(pmd, address);
}

pte_t *addr_pte(struct task_struct *task, unsigned long addr)
{
	pgd_t *pgd = pgd_offset(task->mm, addr);
	pud_t *pud = pud_offset(pgd, addr);
	pmd_t *pmd = pmd_offset(pud, addr);

	return pte_offset_map(pmd, addr);
}

void flush_tlb_all(void)
{
	flush_tlb_mm(current->mm);
}

void flush_tlb_kernel_range(unsigned long start, unsigned long end)
{
	flush_tlb_kernel_range_common(start, end);
}

void flush_tlb_kernel_vm(void)
{
	flush_tlb_kernel_range_common(start_vm, end_vm);
}

void __flush_tlb_one(unsigned long addr)
{
	flush_tlb_kernel_range_common(addr, addr + PAGE_SIZE);
}

static void fix_range(struct mm_struct *mm, unsigned long start_addr,
		      unsigned long end_addr, int force)
{
	fix_range_common(mm, start_addr, end_addr, force);
}

void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
		     unsigned long end)
{
	if (vma->vm_mm == NULL)
		flush_tlb_kernel_range_common(start, end);
	else fix_range(vma->vm_mm, start, end, 0);
}
EXPORT_SYMBOL(flush_tlb_range);

void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
			unsigned long end)
{
	/*
	 * Don't bother flushing if this address space is about to be
	 * destroyed.
	 */
	if (atomic_read(&mm->mm_users) == 0)
		return;

	fix_range(mm, start, end, 0);
}

void flush_tlb_mm(struct mm_struct *mm)
{
	struct vm_area_struct *vma = mm->mmap;

	while (vma != NULL) {
		fix_range(mm, vma->vm_start, vma->vm_end, 0);
		vma = vma->vm_next;
	}
}

void force_flush_all(void)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma = mm->mmap;

	while (vma != NULL) {
		fix_range(mm, vma->vm_start, vma->vm_end, 1);
		vma = vma->vm_next;
	}
}
Commit	Line	Data
5e1f65a6	1	/*
ba180fd4	2	* Copyright (C) 2000 - 2007 Jeff Dike (jdike@{addtoit,linux.intel}.com)
1da177e4 LT	3	* Licensed under the GPL
	4	*/
	5
8192ab42	6	#include <linux/mm.h>
73395a00	7	#include <linux/module.h>
8192ab42 JD	8	#include <linux/sched.h>
	9	#include <asm/pgtable.h>
	10	#include <asm/tlbflush.h>
37185b33 AV	11	#include <as-layout.h>
	12	#include <mem_user.h>
	13	#include <os.h>
	14	#include <skas.h>
1da177e4	15
1466abf2 JD	16	struct host_vm_change {
	17	struct host_vm_op {
	18	enum { NONE, MMAP, MUNMAP, MPROTECT } type;
	19	union {
	20	struct {
	21	unsigned long addr;
	22	unsigned long len;
	23	unsigned int prot;
	24	int fd;
	25	__u64 offset;
	26	} mmap;
	27	struct {
	28	unsigned long addr;
	29	unsigned long len;
	30	} munmap;
	31	struct {
	32	unsigned long addr;
	33	unsigned long len;
	34	unsigned int prot;
	35	} mprotect;
	36	} u;
	37	} ops[1];
	38	int index;
	39	struct mm_id *id;
	40	void *data;
	41	int force;
	42	};
	43
	44	#define INIT_HVC(mm, force) \
	45	((struct host_vm_change) \
	46	{ .ops = { { .type = NONE } }, \
	47	.id = &mm->context.id, \
	48	.data = NULL, \
	49	.index = 0, \
	50	.force = force })
	51
	52	static int do_ops(struct host_vm_change *hvc, int end,
	53	int finished)
	54	{
	55	struct host_vm_op *op;
	56	int i, ret = 0;
	57
	58	for (i = 0; i < end && !ret; i++) {
	59	op = &hvc->ops[i];
c5d4bb17	60	switch (op->type) {
1466abf2 JD	61	case MMAP:
	62	ret = map(hvc->id, op->u.mmap.addr, op->u.mmap.len,
	63	op->u.mmap.prot, op->u.mmap.fd,
	64	op->u.mmap.offset, finished, &hvc->data);
	65	break;
	66	case MUNMAP:
	67	ret = unmap(hvc->id, op->u.munmap.addr,
	68	op->u.munmap.len, finished, &hvc->data);
	69	break;
	70	case MPROTECT:
	71	ret = protect(hvc->id, op->u.mprotect.addr,
	72	op->u.mprotect.len, op->u.mprotect.prot,
	73	finished, &hvc->data);
	74	break;
	75	default:
	76	printk(KERN_ERR "Unknown op type %d in do_ops\n",
	77	op->type);
62179d47	78	BUG();
1466abf2 JD	79	break;
	80	}
	81	}
	82
	83	return ret;
	84	}
	85
c5600490	86	static int add_mmap(unsigned long virt, unsigned long phys, unsigned long len,
1466abf2	87	unsigned int prot, struct host_vm_change *hvc)
c5600490	88	{
5e1f65a6	89	__u64 offset;
c5600490	90	struct host_vm_op *last;
07bf731e	91	int fd, ret = 0;
c5600490 JD	92
c5600490 JD	93	fd = phys_mapping(phys, &offset);
1466abf2 JD	94	if (hvc->index != 0) {
1466abf2 JD	95	last = &hvc->ops[hvc->index - 1];
ba180fd4	96	if ((last->type == MMAP) &&
c5600490	97	(last->u.mmap.addr + last->u.mmap.len == virt) &&
16dd07bc	98	(last->u.mmap.prot == prot) && (last->u.mmap.fd == fd) &&
ba180fd4	99	(last->u.mmap.offset + last->u.mmap.len == offset)) {
c5600490	100	last->u.mmap.len += len;
07bf731e	101	return 0;
c5600490 JD	102	}
	103	}
	104
1466abf2 JD	105	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
	106	ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
	107	hvc->index = 0;
c5600490 JD	108	}
c5600490 JD	109
1466abf2 JD	110	hvc->ops[hvc->index++] = ((struct host_vm_op)
	111	{ .type = MMAP,
	112	.u = { .mmap = { .addr = virt,
	113	.len = len,
	114	.prot = prot,
	115	.fd = fd,
	116	.offset = offset }
07bf731e BS	117	} });
07bf731e BS	118	return ret;
c5600490 JD	119	}
	120
	121	static int add_munmap(unsigned long addr, unsigned long len,
1466abf2	122	struct host_vm_change *hvc)
c5600490 JD	123	{
c5600490 JD	124	struct host_vm_op *last;
07bf731e	125	int ret = 0;
c5600490	126
1466abf2 JD	127	if (hvc->index != 0) {
1466abf2 JD	128	last = &hvc->ops[hvc->index - 1];
ba180fd4 JD	129	if ((last->type == MUNMAP) &&
ba180fd4 JD	130	(last->u.munmap.addr + last->u.mmap.len == addr)) {
c5600490	131	last->u.munmap.len += len;
07bf731e	132	return 0;
c5600490 JD	133	}
	134	}
	135
1466abf2 JD	136	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
	137	ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
	138	hvc->index = 0;
c5600490 JD	139	}
c5600490 JD	140
1466abf2 JD	141	hvc->ops[hvc->index++] = ((struct host_vm_op)
	142	{ .type = MUNMAP,
	143	.u = { .munmap = { .addr = addr,
	144	.len = len } } });
07bf731e	145	return ret;
c5600490 JD	146	}
c5600490 JD	147
16dd07bc	148	static int add_mprotect(unsigned long addr, unsigned long len,
1466abf2	149	unsigned int prot, struct host_vm_change *hvc)
c5600490 JD	150	{
c5600490 JD	151	struct host_vm_op *last;
07bf731e	152	int ret = 0;
c5600490	153
1466abf2 JD	154	if (hvc->index != 0) {
1466abf2 JD	155	last = &hvc->ops[hvc->index - 1];
ba180fd4	156	if ((last->type == MPROTECT) &&
c5600490	157	(last->u.mprotect.addr + last->u.mprotect.len == addr) &&
ba180fd4	158	(last->u.mprotect.prot == prot)) {
c5600490	159	last->u.mprotect.len += len;
07bf731e	160	return 0;
c5600490 JD	161	}
	162	}
	163
1466abf2 JD	164	if (hvc->index == ARRAY_SIZE(hvc->ops)) {
	165	ret = do_ops(hvc, ARRAY_SIZE(hvc->ops), 0);
	166	hvc->index = 0;
c5600490 JD	167	}
c5600490 JD	168
1466abf2 JD	169	hvc->ops[hvc->index++] = ((struct host_vm_op)
	170	{ .type = MPROTECT,
	171	.u = { .mprotect = { .addr = addr,
	172	.len = len,
	173	.prot = prot } } });
07bf731e	174	return ret;
c5600490 JD	175	}
c5600490 JD	176
1da177e4 LT	177	#define ADD_ROUND(n, inc) (((n) + (inc)) & ~((inc) - 1))
1da177e4 LT	178
7f0536f8	179	static inline int update_pte_range(pmd_t *pmd, unsigned long addr,
1466abf2 JD	180	unsigned long end,
1466abf2 JD	181	struct host_vm_change *hvc)
7f0536f8 JD	182	{
7f0536f8 JD	183	pte_t *pte;
16dd07bc	184	int r, w, x, prot, ret = 0;
7f0536f8 JD	185
	186	pte = pte_offset_kernel(pmd, addr);
	187	do {
3963333f JD	188	if ((addr >= STUB_START) && (addr < STUB_END))
	189	continue;
	190
7f0536f8 JD	191	r = pte_read(*pte);
	192	w = pte_write(*pte);
	193	x = pte_exec(*pte);
	194	if (!pte_young(*pte)) {
	195	r = 0;
	196	w = 0;
0b4e273f	197	} else if (!pte_dirty(*pte))
7f0536f8	198	w = 0;
0b4e273f	199
16dd07bc JD	200	prot = ((r ? UM_PROT_READ : 0) \| (w ? UM_PROT_WRITE : 0) \|
16dd07bc JD	201	(x ? UM_PROT_EXEC : 0));
1466abf2	202	if (hvc->force \|\| pte_newpage(*pte)) {
ba180fd4	203	if (pte_present(*pte))
7f0536f8	204	ret = add_mmap(addr, pte_val(*pte) & PAGE_MASK,
1466abf2	205	PAGE_SIZE, prot, hvc);
0b4e273f JD	206	else
	207	ret = add_munmap(addr, PAGE_SIZE, hvc);
	208	} else if (pte_newprot(*pte))
1466abf2	209	ret = add_mprotect(addr, PAGE_SIZE, prot, hvc);
7f0536f8	210	pte = pte_mkuptodate(pte);
909e90d3	211	} while (pte++, addr += PAGE_SIZE, ((addr < end) && !ret));
7f0536f8 JD	212	return ret;
	213	}
	214
	215	static inline int update_pmd_range(pud_t *pud, unsigned long addr,
1466abf2 JD	216	unsigned long end,
1466abf2 JD	217	struct host_vm_change *hvc)
7f0536f8 JD	218	{
	219	pmd_t *pmd;
	220	unsigned long next;
	221	int ret = 0;
	222
	223	pmd = pmd_offset(pud, addr);
	224	do {
	225	next = pmd_addr_end(addr, end);
ba180fd4	226	if (!pmd_present(*pmd)) {
1466abf2 JD	227	if (hvc->force \|\| pmd_newpage(*pmd)) {
1466abf2 JD	228	ret = add_munmap(addr, next - addr, hvc);
7f0536f8 JD	229	pmd_mkuptodate(*pmd);
	230	}
	231	}
1466abf2	232	else ret = update_pte_range(pmd, addr, next, hvc);
909e90d3	233	} while (pmd++, addr = next, ((addr < end) && !ret));
7f0536f8 JD	234	return ret;
	235	}
	236
	237	static inline int update_pud_range(pgd_t *pgd, unsigned long addr,
1466abf2 JD	238	unsigned long end,
1466abf2 JD	239	struct host_vm_change *hvc)
7f0536f8 JD	240	{
	241	pud_t *pud;
	242	unsigned long next;
	243	int ret = 0;
	244
	245	pud = pud_offset(pgd, addr);
	246	do {
	247	next = pud_addr_end(addr, end);
ba180fd4	248	if (!pud_present(*pud)) {
1466abf2 JD	249	if (hvc->force \|\| pud_newpage(*pud)) {
1466abf2 JD	250	ret = add_munmap(addr, next - addr, hvc);
7f0536f8 JD	251	pud_mkuptodate(*pud);
	252	}
	253	}
1466abf2	254	else ret = update_pmd_range(pud, addr, next, hvc);
909e90d3	255	} while (pud++, addr = next, ((addr < end) && !ret));
7f0536f8 JD	256	return ret;
	257	}
	258
1da177e4	259	void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
1466abf2	260	unsigned long end_addr, int force)
1da177e4	261	{
7f0536f8	262	pgd_t *pgd;
1466abf2	263	struct host_vm_change hvc;
7f0536f8	264	unsigned long addr = start_addr, next;
1466abf2	265	int ret = 0;
1da177e4	266
1466abf2	267	hvc = INIT_HVC(mm, force);
7f0536f8 JD	268	pgd = pgd_offset(mm, addr);
	269	do {
	270	next = pgd_addr_end(addr, end_addr);
ba180fd4 JD	271	if (!pgd_present(*pgd)) {
ba180fd4 JD	272	if (force \|\| pgd_newpage(*pgd)) {
1466abf2	273	ret = add_munmap(addr, next - addr, &hvc);
7f0536f8	274	pgd_mkuptodate(*pgd);
5e1f65a6	275	}
5e1f65a6	276	}
1466abf2	277	else ret = update_pud_range(pgd, addr, next, &hvc);
909e90d3	278	} while (pgd++, addr = next, ((addr < end_addr) && !ret));
5e1f65a6	279
ba180fd4	280	if (!ret)
1466abf2	281	ret = do_ops(&hvc, hvc.index, 1);
07bf731e	282
7f0536f8	283	/* This is not an else because ret is modified above */
ba180fd4 JD	284	if (ret) {
	285	printk(KERN_ERR "fix_range_common: failed, killing current "
	286	"process\n");
07bf731e BS	287	force_sig(SIGKILL, current);
07bf731e BS	288	}
1da177e4 LT	289	}
1da177e4 LT	290
c75d053b	291	static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
1da177e4	292	{
5e1f65a6 JD	293	struct mm_struct *mm;
	294	pgd_t *pgd;
	295	pud_t *pud;
	296	pmd_t *pmd;
	297	pte_t *pte;
	298	unsigned long addr, last;
	299	int updated = 0, err;
	300
	301	mm = &init_mm;
ba180fd4	302	for (addr = start; addr < end;) {
5e1f65a6	303	pgd = pgd_offset(mm, addr);
ba180fd4	304	if (!pgd_present(*pgd)) {
5e1f65a6	305	last = ADD_ROUND(addr, PGDIR_SIZE);
ba180fd4	306	if (last > end)
5e1f65a6	307	last = end;
ba180fd4	308	if (pgd_newpage(*pgd)) {
5e1f65a6 JD	309	updated = 1;
	310	err = os_unmap_memory((void *) addr,
	311	last - addr);
ba180fd4	312	if (err < 0)
5e1f65a6 JD	313	panic("munmap failed, errno = %d\n",
	314	-err);
	315	}
	316	addr = last;
	317	continue;
	318	}
	319
	320	pud = pud_offset(pgd, addr);
ba180fd4	321	if (!pud_present(*pud)) {
5e1f65a6	322	last = ADD_ROUND(addr, PUD_SIZE);
ba180fd4	323	if (last > end)
5e1f65a6	324	last = end;
ba180fd4	325	if (pud_newpage(*pud)) {
5e1f65a6 JD	326	updated = 1;
	327	err = os_unmap_memory((void *) addr,
	328	last - addr);
ba180fd4	329	if (err < 0)
5e1f65a6 JD	330	panic("munmap failed, errno = %d\n",
	331	-err);
	332	}
	333	addr = last;
	334	continue;
	335	}
	336
	337	pmd = pmd_offset(pud, addr);
ba180fd4	338	if (!pmd_present(*pmd)) {
5e1f65a6	339	last = ADD_ROUND(addr, PMD_SIZE);
ba180fd4	340	if (last > end)
5e1f65a6	341	last = end;
ba180fd4	342	if (pmd_newpage(*pmd)) {
5e1f65a6 JD	343	updated = 1;
	344	err = os_unmap_memory((void *) addr,
	345	last - addr);
ba180fd4	346	if (err < 0)
5e1f65a6 JD	347	panic("munmap failed, errno = %d\n",
	348	-err);
	349	}
	350	addr = last;
	351	continue;
	352	}
	353
	354	pte = pte_offset_kernel(pmd, addr);
ba180fd4	355	if (!pte_present(pte) \|\| pte_newpage(pte)) {
5e1f65a6 JD	356	updated = 1;
	357	err = os_unmap_memory((void *) addr,
	358	PAGE_SIZE);
ba180fd4	359	if (err < 0)
5e1f65a6 JD	360	panic("munmap failed, errno = %d\n",
5e1f65a6 JD	361	-err);
ba180fd4	362	if (pte_present(*pte))
5e1f65a6 JD	363	map_memory(addr,
	364	pte_val(*pte) & PAGE_MASK,
	365	PAGE_SIZE, 1, 1, 1);
	366	}
ba180fd4	367	else if (pte_newprot(*pte)) {
5e1f65a6 JD	368	updated = 1;
	369	os_protect_memory((void *) addr, PAGE_SIZE, 1, 1, 1);
	370	}
	371	addr += PAGE_SIZE;
	372	}
ba180fd4	373	return updated;
1da177e4 LT	374	}
1da177e4 LT	375
77bf4400 JD	376	void flush_tlb_page(struct vm_area_struct *vma, unsigned long address)
	377	{
	378	pgd_t *pgd;
	379	pud_t *pud;
	380	pmd_t *pmd;
	381	pte_t *pte;
	382	struct mm_struct *mm = vma->vm_mm;
	383	void *flush = NULL;
	384	int r, w, x, prot, err = 0;
	385	struct mm_id *mm_id;
	386
	387	address &= PAGE_MASK;
	388	pgd = pgd_offset(mm, address);
ba180fd4	389	if (!pgd_present(*pgd))
77bf4400 JD	390	goto kill;
	391
	392	pud = pud_offset(pgd, address);
ba180fd4	393	if (!pud_present(*pud))
77bf4400 JD	394	goto kill;
	395
	396	pmd = pmd_offset(pud, address);
ba180fd4	397	if (!pmd_present(*pmd))
77bf4400 JD	398	goto kill;
	399
	400	pte = pte_offset_kernel(pmd, address);
	401
	402	r = pte_read(*pte);
	403	w = pte_write(*pte);
	404	x = pte_exec(*pte);
	405	if (!pte_young(*pte)) {
	406	r = 0;
	407	w = 0;
	408	} else if (!pte_dirty(*pte)) {
	409	w = 0;
	410	}
	411
6c738ffa	412	mm_id = &mm->context.id;
77bf4400 JD	413	prot = ((r ? UM_PROT_READ : 0) \| (w ? UM_PROT_WRITE : 0) \|
77bf4400 JD	414	(x ? UM_PROT_EXEC : 0));
ba180fd4 JD	415	if (pte_newpage(*pte)) {
ba180fd4 JD	416	if (pte_present(*pte)) {
77bf4400 JD	417	unsigned long long offset;
	418	int fd;
	419
	420	fd = phys_mapping(pte_val(*pte) & PAGE_MASK, &offset);
	421	err = map(mm_id, address, PAGE_SIZE, prot, fd, offset,
	422	1, &flush);
	423	}
	424	else err = unmap(mm_id, address, PAGE_SIZE, 1, &flush);
	425	}
ba180fd4	426	else if (pte_newprot(*pte))
77bf4400 JD	427	err = protect(mm_id, address, PAGE_SIZE, prot, 1, &flush);
77bf4400 JD	428
ba180fd4	429	if (err)
77bf4400 JD	430	goto kill;
	431
	432	pte = pte_mkuptodate(pte);
	433
	434	return;
	435
	436	kill:
ba180fd4	437	printk(KERN_ERR "Failed to flush page for address 0x%lx\n", address);
77bf4400 JD	438	force_sig(SIGKILL, current);
	439	}
	440
1da177e4 LT	441	pgd_t pgd_offset_proc(struct mm_struct mm, unsigned long address)
1da177e4 LT	442	{
ba180fd4	443	return pgd_offset(mm, address);
1da177e4 LT	444	}
	445
	446	pud_t pud_offset_proc(pgd_t pgd, unsigned long address)
	447	{
ba180fd4	448	return pud_offset(pgd, address);
1da177e4 LT	449	}
	450
	451	pmd_t pmd_offset_proc(pud_t pud, unsigned long address)
	452	{
ba180fd4	453	return pmd_offset(pud, address);
1da177e4 LT	454	}
	455
	456	pte_t pte_offset_proc(pmd_t pmd, unsigned long address)
	457	{
ba180fd4	458	return pte_offset_kernel(pmd, address);
1da177e4 LT	459	}
	460
	461	pte_t addr_pte(struct task_struct task, unsigned long addr)
	462	{
5e1f65a6 JD	463	pgd_t *pgd = pgd_offset(task->mm, addr);
	464	pud_t *pud = pud_offset(pgd, addr);
	465	pmd_t *pmd = pmd_offset(pud, addr);
1da177e4	466
ba180fd4	467	return pte_offset_map(pmd, addr);
1da177e4 LT	468	}
1da177e4 LT	469
d67b569f JD	470	void flush_tlb_all(void)
d67b569f JD	471	{
5e1f65a6	472	flush_tlb_mm(current->mm);
d67b569f JD	473	}
	474
	475	void flush_tlb_kernel_range(unsigned long start, unsigned long end)
	476	{
6aa802ce	477	flush_tlb_kernel_range_common(start, end);
d67b569f JD	478	}
	479
	480	void flush_tlb_kernel_vm(void)
	481	{
6aa802ce	482	flush_tlb_kernel_range_common(start_vm, end_vm);
d67b569f JD	483	}
	484
	485	void __flush_tlb_one(unsigned long addr)
	486	{
ba180fd4	487	flush_tlb_kernel_range_common(addr, addr + PAGE_SIZE);
77bf4400 JD	488	}
77bf4400 JD	489
77bf4400 JD	490	static void fix_range(struct mm_struct *mm, unsigned long start_addr,
	491	unsigned long end_addr, int force)
	492	{
1466abf2	493	fix_range_common(mm, start_addr, end_addr, force);
d67b569f JD	494	}
	495
	496	void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
	497	unsigned long end)
	498	{
ba180fd4 JD	499	if (vma->vm_mm == NULL)
	500	flush_tlb_kernel_range_common(start, end);
	501	else fix_range(vma->vm_mm, start, end, 0);
d67b569f	502	}
73395a00	503	EXPORT_SYMBOL(flush_tlb_range);
d67b569f	504
0b4e273f JD	505	void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
0b4e273f JD	506	unsigned long end)
d67b569f	507	{
ba180fd4 JD	508	/*
	509	* Don't bother flushing if this address space is about to be
	510	* destroyed.
	511	*/
	512	if (atomic_read(&mm->mm_users) == 0)
	513	return;
77bf4400	514
0b4e273f JD	515	fix_range(mm, start, end, 0);
	516	}
	517
	518	void flush_tlb_mm(struct mm_struct *mm)
	519	{
	520	struct vm_area_struct *vma = mm->mmap;
	521
	522	while (vma != NULL) {
	523	fix_range(mm, vma->vm_start, vma->vm_end, 0);
	524	vma = vma->vm_next;
	525	}
d67b569f JD	526	}
	527
	528	void force_flush_all(void)
	529	{
77bf4400 JD	530	struct mm_struct *mm = current->mm;
	531	struct vm_area_struct *vma = mm->mmap;
	532
ba180fd4	533	while (vma != NULL) {
77bf4400 JD	534	fix_range(mm, vma->vm_start, vma->vm_end, 1);
	535	vma = vma->vm_next;
	536	}
d67b569f	537	}