[deliverable/linux.git] / arch / powerpc / include / asm / pgtable-ppc64.h

#ifndef _ASM_POWERPC_PGTABLE_PPC64_H_
#define _ASM_POWERPC_PGTABLE_PPC64_H_
/*
 * This file contains the functions and defines necessary to modify and use
 * the ppc64 hashed page table.
 */

#ifdef CONFIG_PPC_64K_PAGES
#include <asm/pgtable-ppc64-64k.h>
#else
#include <asm/pgtable-ppc64-4k.h>
#endif

#define FIRST_USER_ADDRESS	0

/*
 * Size of EA range mapped by our pagetables.
 */
#define PGTABLE_EADDR_SIZE (PTE_INDEX_SIZE + PMD_INDEX_SIZE + \
                	    PUD_INDEX_SIZE + PGD_INDEX_SIZE + PAGE_SHIFT)
#define PGTABLE_RANGE (ASM_CONST(1) << PGTABLE_EADDR_SIZE)


/*
 * Define the address range of the kernel non-linear virtual area
 */

#ifdef CONFIG_PPC_BOOK3E
#define KERN_VIRT_START ASM_CONST(0x8000000000000000)
#else
#define KERN_VIRT_START ASM_CONST(0xD000000000000000)
#endif
#define KERN_VIRT_SIZE	ASM_CONST(0x0000100000000000)

/*
 * The vmalloc space starts at the beginning of that region, and
 * occupies half of it on hash CPUs and a quarter of it on Book3E
 * (we keep a quarter for the virtual memmap)
 */
#define VMALLOC_START	KERN_VIRT_START
#ifdef CONFIG_PPC_BOOK3E
#define VMALLOC_SIZE	(KERN_VIRT_SIZE >> 2)
#else
#define VMALLOC_SIZE	(KERN_VIRT_SIZE >> 1)
#endif
#define VMALLOC_END	(VMALLOC_START + VMALLOC_SIZE)

/*
 * The second half of the kernel virtual space is used for IO mappings,
 * it's itself carved into the PIO region (ISA and PHB IO space) and
 * the ioremap space
 *
 *  ISA_IO_BASE = KERN_IO_START, 64K reserved area
 *  PHB_IO_BASE = ISA_IO_BASE + 64K to ISA_IO_BASE + 2G, PHB IO spaces
 * IOREMAP_BASE = ISA_IO_BASE + 2G to VMALLOC_START + PGTABLE_RANGE
 */
#define KERN_IO_START	(KERN_VIRT_START + (KERN_VIRT_SIZE >> 1))
#define FULL_IO_SIZE	0x80000000ul
#define  ISA_IO_BASE	(KERN_IO_START)
#define  ISA_IO_END	(KERN_IO_START + 0x10000ul)
#define  PHB_IO_BASE	(ISA_IO_END)
#define  PHB_IO_END	(KERN_IO_START + FULL_IO_SIZE)
#define IOREMAP_BASE	(PHB_IO_END)
#define IOREMAP_END	(KERN_VIRT_START + KERN_VIRT_SIZE)


/*
 * Region IDs
 */
#define REGION_SHIFT		60UL
#define REGION_MASK		(0xfUL << REGION_SHIFT)
#define REGION_ID(ea)		(((unsigned long)(ea)) >> REGION_SHIFT)

#define VMALLOC_REGION_ID	(REGION_ID(VMALLOC_START))
#define KERNEL_REGION_ID	(REGION_ID(PAGE_OFFSET))
#define VMEMMAP_REGION_ID	(0xfUL)	/* Server only */
#define USER_REGION_ID		(0UL)

/*
 * Defines the address of the vmemap area, in its own region on
 * hash table CPUs and after the vmalloc space on Book3E
 */
#ifdef CONFIG_PPC_BOOK3E
#define VMEMMAP_BASE		VMALLOC_END
#define VMEMMAP_END		KERN_IO_START
#else
#define VMEMMAP_BASE		(VMEMMAP_REGION_ID << REGION_SHIFT)
#endif
#define vmemmap			((struct page *)VMEMMAP_BASE)


/*
 * Include the PTE bits definitions
 */
#ifdef CONFIG_PPC_BOOK3S
#include <asm/pte-hash64.h>
#else
#include <asm/pte-book3e.h>
#endif
#include <asm/pte-common.h>

#ifdef CONFIG_PPC_MM_SLICES
#define HAVE_ARCH_UNMAPPED_AREA
#define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
#endif /* CONFIG_PPC_MM_SLICES */

#ifndef __ASSEMBLY__

/*
 * This is the default implementation of various PTE accessors, it's
 * used in all cases except Book3S with 64K pages where we have a
 * concept of sub-pages
 */
#ifndef __real_pte

#ifdef STRICT_MM_TYPECHECKS
#define __real_pte(e,p)		((real_pte_t){(e)})
#define __rpte_to_pte(r)	((r).pte)
#else
#define __real_pte(e,p)		(e)
#define __rpte_to_pte(r)	(__pte(r))
#endif
#define __rpte_to_hidx(r,index)	(pte_val(__rpte_to_pte(r)) >> 12)

#define pte_iterate_hashed_subpages(rpte, psize, va, index, shift)       \
	do {							         \
		index = 0;					         \
		shift = mmu_psize_defs[psize].shift;		         \

#define pte_iterate_hashed_end() } while(0)

#ifdef CONFIG_PPC_HAS_HASH_64K
#define pte_pagesize_index(mm, addr, pte)	get_slice_psize(mm, addr)
#else
#define pte_pagesize_index(mm, addr, pte)	MMU_PAGE_4K
#endif

#endif /* __real_pte */


/* pte_clear moved to later in this file */

#define PMD_BAD_BITS		(PTE_TABLE_SIZE-1)
#define PUD_BAD_BITS		(PMD_TABLE_SIZE-1)

#define pmd_set(pmdp, pmdval) 	(pmd_val(*(pmdp)) = (pmdval))
#define pmd_none(pmd)		(!pmd_val(pmd))
#define	pmd_bad(pmd)		(!is_kernel_addr(pmd_val(pmd)) \
				 || (pmd_val(pmd) & PMD_BAD_BITS))
#define	pmd_present(pmd)	(pmd_val(pmd) != 0)
#define	pmd_clear(pmdp)		(pmd_val(*(pmdp)) = 0)
#define pmd_page_vaddr(pmd)	(pmd_val(pmd) & ~PMD_MASKED_BITS)
#define pmd_page(pmd)		virt_to_page(pmd_page_vaddr(pmd))

#define pud_set(pudp, pudval)	(pud_val(*(pudp)) = (pudval))
#define pud_none(pud)		(!pud_val(pud))
#define	pud_bad(pud)		(!is_kernel_addr(pud_val(pud)) \
				 || (pud_val(pud) & PUD_BAD_BITS))
#define pud_present(pud)	(pud_val(pud) != 0)
#define pud_clear(pudp)		(pud_val(*(pudp)) = 0)
#define pud_page_vaddr(pud)	(pud_val(pud) & ~PUD_MASKED_BITS)
#define pud_page(pud)		virt_to_page(pud_page_vaddr(pud))

#define pgd_set(pgdp, pudp)	({pgd_val(*(pgdp)) = (unsigned long)(pudp);})

/*
 * Find an entry in a page-table-directory.  We combine the address region
 * (the high order N bits) and the pgd portion of the address.
 */
/* to avoid overflow in free_pgtables we don't use PTRS_PER_PGD here */
#define pgd_index(address) (((address) >> (PGDIR_SHIFT)) & 0x1ff)

#define pgd_offset(mm, address)	 ((mm)->pgd + pgd_index(address))

#define pmd_offset(pudp,addr) \
  (((pmd_t *) pud_page_vaddr(*(pudp))) + (((addr) >> PMD_SHIFT) & (PTRS_PER_PMD - 1)))

#define pte_offset_kernel(dir,addr) \
  (((pte_t *) pmd_page_vaddr(*(dir))) + (((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))

#define pte_offset_map(dir,addr)	pte_offset_kernel((dir), (addr))
#define pte_unmap(pte)			do { } while(0)

/* to find an entry in a kernel page-table-directory */
/* This now only contains the vmalloc pages */
#define pgd_offset_k(address) pgd_offset(&init_mm, address)
extern void hpte_need_flush(struct mm_struct *mm, unsigned long addr,
			    pte_t *ptep, unsigned long pte, int huge);

/* Atomic PTE updates */
static inline unsigned long pte_update(struct mm_struct *mm,
				       unsigned long addr,
				       pte_t *ptep, unsigned long clr,
				       int huge)
{
#ifdef PTE_ATOMIC_UPDATES
	unsigned long old, tmp;

	__asm__ __volatile__(
	"1:	ldarx	%0,0,%3		# pte_update\n\
	andi.	%1,%0,%6\n\
	bne-	1b \n\
	andc	%1,%0,%4 \n\
	stdcx.	%1,0,%3 \n\
	bne-	1b"
	: "=&r" (old), "=&r" (tmp), "=m" (*ptep)
	: "r" (ptep), "r" (clr), "m" (*ptep), "i" (_PAGE_BUSY)
	: "cc" );
#else
	unsigned long old = pte_val(*ptep);
	*ptep = __pte(old & ~clr);
#endif
	/* huge pages use the old page table lock */
	if (!huge)
		assert_pte_locked(mm, addr);

#ifdef CONFIG_PPC_STD_MMU_64
	if (old & _PAGE_HASHPTE)
		hpte_need_flush(mm, addr, ptep, old, huge);
#endif

	return old;
}

static inline int __ptep_test_and_clear_young(struct mm_struct *mm,
					      unsigned long addr, pte_t *ptep)
{
	unsigned long old;

       	if ((pte_val(*ptep) & (_PAGE_ACCESSED | _PAGE_HASHPTE)) == 0)
		return 0;
	old = pte_update(mm, addr, ptep, _PAGE_ACCESSED, 0);
	return (old & _PAGE_ACCESSED) != 0;
}
#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
#define ptep_test_and_clear_young(__vma, __addr, __ptep)		   \
({									   \
	int __r;							   \
	__r = __ptep_test_and_clear_young((__vma)->vm_mm, __addr, __ptep); \
	__r;								   \
})

#define __HAVE_ARCH_PTEP_SET_WRPROTECT
static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr,
				      pte_t *ptep)
{

	if ((pte_val(*ptep) & _PAGE_RW) == 0)
		return;

	pte_update(mm, addr, ptep, _PAGE_RW, 0);
}

static inline void huge_ptep_set_wrprotect(struct mm_struct *mm,
					   unsigned long addr, pte_t *ptep)
{
	if ((pte_val(*ptep) & _PAGE_RW) == 0)
		return;

	pte_update(mm, addr, ptep, _PAGE_RW, 1);
}

/*
 * We currently remove entries from the hashtable regardless of whether
 * the entry was young or dirty. The generic routines only flush if the
 * entry was young or dirty which is not good enough.
 *
 * We should be more intelligent about this but for the moment we override
 * these functions and force a tlb flush unconditionally
 */
#define __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH
#define ptep_clear_flush_young(__vma, __address, __ptep)		\
({									\
	int __young = __ptep_test_and_clear_young((__vma)->vm_mm, __address, \
						  __ptep);		\
	__young;							\
})

#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
static inline pte_t ptep_get_and_clear(struct mm_struct *mm,
				       unsigned long addr, pte_t *ptep)
{
	unsigned long old = pte_update(mm, addr, ptep, ~0UL, 0);
	return __pte(old);
}

static inline void pte_clear(struct mm_struct *mm, unsigned long addr,
			     pte_t * ptep)
{
	pte_update(mm, addr, ptep, ~0UL, 0);
}


/* Set the dirty and/or accessed bits atomically in a linux PTE, this
 * function doesn't need to flush the hash entry
 */
static inline void __ptep_set_access_flags(pte_t *ptep, pte_t entry)
{
	unsigned long bits = pte_val(entry) &
		(_PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_RW | _PAGE_EXEC);

#ifdef PTE_ATOMIC_UPDATES
	unsigned long old, tmp;

	__asm__ __volatile__(
	"1:	ldarx	%0,0,%4\n\
		andi.	%1,%0,%6\n\
		bne-	1b \n\
		or	%0,%3,%0\n\
		stdcx.	%0,0,%4\n\
		bne-	1b"
	:"=&r" (old), "=&r" (tmp), "=m" (*ptep)
	:"r" (bits), "r" (ptep), "m" (*ptep), "i" (_PAGE_BUSY)
	:"cc");
#else
	unsigned long old = pte_val(*ptep);
	*ptep = __pte(old | bits);
#endif
}

#define __HAVE_ARCH_PTE_SAME
#define pte_same(A,B)	(((pte_val(A) ^ pte_val(B)) & ~_PAGE_HPTEFLAGS) == 0)

#define pte_ERROR(e) \
	printk("%s:%d: bad pte %08lx.\n", __FILE__, __LINE__, pte_val(e))
#define pmd_ERROR(e) \
	printk("%s:%d: bad pmd %08lx.\n", __FILE__, __LINE__, pmd_val(e))
#define pgd_ERROR(e) \
	printk("%s:%d: bad pgd %08lx.\n", __FILE__, __LINE__, pgd_val(e))

/* Encode and de-code a swap entry */
#define __swp_type(entry)	(((entry).val >> 1) & 0x3f)
#define __swp_offset(entry)	((entry).val >> 8)
#define __swp_entry(type, offset) ((swp_entry_t){((type)<< 1)|((offset)<<8)})
#define __pte_to_swp_entry(pte)	((swp_entry_t){pte_val(pte) >> PTE_RPN_SHIFT})
#define __swp_entry_to_pte(x)	((pte_t) { (x).val << PTE_RPN_SHIFT })
#define pte_to_pgoff(pte)	(pte_val(pte) >> PTE_RPN_SHIFT)
#define pgoff_to_pte(off)	((pte_t) {((off) << PTE_RPN_SHIFT)|_PAGE_FILE})
#define PTE_FILE_MAX_BITS	(BITS_PER_LONG - PTE_RPN_SHIFT)

void pgtable_cache_add(unsigned shift, void (*ctor)(void *));
void pgtable_cache_init(void);

/*
 * find_linux_pte returns the address of a linux pte for a given
 * effective address and directory.  If not found, it returns zero.
 */
static inline pte_t *find_linux_pte(pgd_t *pgdir, unsigned long ea)
{
	pgd_t *pg;
	pud_t *pu;
	pmd_t *pm;
	pte_t *pt = NULL;

	pg = pgdir + pgd_index(ea);
	if (!pgd_none(*pg)) {
		pu = pud_offset(pg, ea);
		if (!pud_none(*pu)) {
			pm = pmd_offset(pu, ea);
			if (pmd_present(*pm))
				pt = pte_offset_kernel(pm, ea);
		}
	}
	return pt;
}

#ifdef CONFIG_HUGETLB_PAGE
pte_t *find_linux_pte_or_hugepte(pgd_t *pgdir, unsigned long ea,
				 unsigned *shift);
#else
static inline pte_t *find_linux_pte_or_hugepte(pgd_t *pgdir, unsigned long ea,
					       unsigned *shift)
{
	if (shift)
		*shift = 0;
	return find_linux_pte(pgdir, ea);
}
#endif /* !CONFIG_HUGETLB_PAGE */

#endif /* __ASSEMBLY__ */

#endif /* _ASM_POWERPC_PGTABLE_PPC64_H_ */
Commit	Line	Data
f88df14b DG	1	#ifndef _ASM_POWERPC_PGTABLE_PPC64_H_
	2	#define _ASM_POWERPC_PGTABLE_PPC64_H_
	3	/*
	4	* This file contains the functions and defines necessary to modify and use
	5	* the ppc64 hashed page table.
	6	*/
	7
f88df14b	8	#ifdef CONFIG_PPC_64K_PAGES
c605782b	9	#include <asm/pgtable-ppc64-64k.h>
f88df14b	10	#else
c605782b	11	#include <asm/pgtable-ppc64-4k.h>
f88df14b DG	12	#endif
	13
	14	#define FIRST_USER_ADDRESS 0
	15
	16	/*
	17	* Size of EA range mapped by our pagetables.
	18	*/
	19	#define PGTABLE_EADDR_SIZE (PTE_INDEX_SIZE + PMD_INDEX_SIZE + \
	20	PUD_INDEX_SIZE + PGD_INDEX_SIZE + PAGE_SHIFT)
3d5134ee	21	#define PGTABLE_RANGE (ASM_CONST(1) << PGTABLE_EADDR_SIZE)
f88df14b	22
c605782b	23
f88df14b	24	/*
57e2a99f	25	* Define the address range of the kernel non-linear virtual area
f88df14b	26	*/
57e2a99f BH	27
	28	#ifdef CONFIG_PPC_BOOK3E
	29	#define KERN_VIRT_START ASM_CONST(0x8000000000000000)
	30	#else
	31	#define KERN_VIRT_START ASM_CONST(0xD000000000000000)
	32	#endif
67550080	33	#define KERN_VIRT_SIZE ASM_CONST(0x0000100000000000)
f88df14b DG	34
f88df14b DG	35	/*
57e2a99f BH	36	* The vmalloc space starts at the beginning of that region, and
57e2a99f BH	37	* occupies half of it on hash CPUs and a quarter of it on Book3E
32a74949	38	* (we keep a quarter for the virtual memmap)
57e2a99f BH	39	*/
	40	#define VMALLOC_START KERN_VIRT_START
	41	#ifdef CONFIG_PPC_BOOK3E
	42	#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 2)
	43	#else
	44	#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 1)
	45	#endif
	46	#define VMALLOC_END (VMALLOC_START + VMALLOC_SIZE)
	47
	48	/*
	49	* The second half of the kernel virtual space is used for IO mappings,
	50	* it's itself carved into the PIO region (ISA and PHB IO space) and
	51	* the ioremap space
3d5134ee	52	*
57e2a99f	53	* ISA_IO_BASE = KERN_IO_START, 64K reserved area
3d5134ee BH	54	* PHB_IO_BASE = ISA_IO_BASE + 64K to ISA_IO_BASE + 2G, PHB IO spaces
3d5134ee BH	55	* IOREMAP_BASE = ISA_IO_BASE + 2G to VMALLOC_START + PGTABLE_RANGE
f88df14b	56	*/
57e2a99f	57	#define KERN_IO_START (KERN_VIRT_START + (KERN_VIRT_SIZE >> 1))
3d5134ee	58	#define FULL_IO_SIZE 0x80000000ul
57e2a99f BH	59	#define ISA_IO_BASE (KERN_IO_START)
57e2a99f BH	60	#define ISA_IO_END (KERN_IO_START + 0x10000ul)
3d5134ee	61	#define PHB_IO_BASE (ISA_IO_END)
57e2a99f	62	#define PHB_IO_END (KERN_IO_START + FULL_IO_SIZE)
3d5134ee	63	#define IOREMAP_BASE (PHB_IO_END)
57e2a99f BH	64	#define IOREMAP_END (KERN_VIRT_START + KERN_VIRT_SIZE)
57e2a99f BH	65
f88df14b DG	66
	67	/*
	68	* Region IDs
	69	*/
	70	#define REGION_SHIFT 60UL
	71	#define REGION_MASK (0xfUL << REGION_SHIFT)
	72	#define REGION_ID(ea) (((unsigned long)(ea)) >> REGION_SHIFT)
	73
	74	#define VMALLOC_REGION_ID (REGION_ID(VMALLOC_START))
	75	#define KERNEL_REGION_ID (REGION_ID(PAGE_OFFSET))
32a74949	76	#define VMEMMAP_REGION_ID (0xfUL) /* Server only */
f88df14b DG	77	#define USER_REGION_ID (0UL)
f88df14b DG	78
d29eff7b	79	/*
57e2a99f BH	80	* Defines the address of the vmemap area, in its own region on
57e2a99f BH	81	* hash table CPUs and after the vmalloc space on Book3E
d29eff7b	82	*/
57e2a99f BH	83	#ifdef CONFIG_PPC_BOOK3E
	84	#define VMEMMAP_BASE VMALLOC_END
	85	#define VMEMMAP_END KERN_IO_START
	86	#else
cec08e7a	87	#define VMEMMAP_BASE (VMEMMAP_REGION_ID << REGION_SHIFT)
57e2a99f	88	#endif
cec08e7a BH	89	#define vmemmap ((struct page *)VMEMMAP_BASE)
cec08e7a BH	90
d29eff7b	91
f88df14b	92	/*
c605782b	93	* Include the PTE bits definitions
f88df14b	94	*/
57e2a99f	95	#ifdef CONFIG_PPC_BOOK3S
c605782b	96	#include <asm/pte-hash64.h>
57e2a99f BH	97	#else
	98	#include <asm/pte-book3e.h>
	99	#endif
71087002	100	#include <asm/pte-common.h>
c605782b	101
94ee815c	102	#ifdef CONFIG_PPC_MM_SLICES
f88df14b DG	103	#define HAVE_ARCH_UNMAPPED_AREA
f88df14b DG	104	#define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
94ee815c	105	#endif /* CONFIG_PPC_MM_SLICES */
f88df14b DG	106
	107	#ifndef __ASSEMBLY__
	108
c605782b BH	109	/*
	110	* This is the default implementation of various PTE accessors, it's
	111	* used in all cases except Book3S with 64K pages where we have a
	112	* concept of sub-pages
	113	*/
	114	#ifndef __real_pte
	115
	116	#ifdef STRICT_MM_TYPECHECKS
	117	#define __real_pte(e,p) ((real_pte_t){(e)})
	118	#define __rpte_to_pte(r) ((r).pte)
	119	#else
	120	#define __real_pte(e,p) (e)
	121	#define __rpte_to_pte(r) (__pte(r))
	122	#endif
	123	#define __rpte_to_hidx(r,index) (pte_val(__rpte_to_pte(r)) >> 12)
	124
	125	#define pte_iterate_hashed_subpages(rpte, psize, va, index, shift) \
	126	do { \
	127	index = 0; \
	128	shift = mmu_psize_defs[psize].shift; \
	129
	130	#define pte_iterate_hashed_end() } while(0)
	131
	132	#ifdef CONFIG_PPC_HAS_HASH_64K
	133	#define pte_pagesize_index(mm, addr, pte) get_slice_psize(mm, addr)
	134	#else
	135	#define pte_pagesize_index(mm, addr, pte) MMU_PAGE_4K
	136	#endif
	137
	138	#endif /* __real_pte */
	139
	140
f88df14b DG	141	/* pte_clear moved to later in this file */
f88df14b DG	142
f88df14b DG	143	#define PMD_BAD_BITS (PTE_TABLE_SIZE-1)
	144	#define PUD_BAD_BITS (PMD_TABLE_SIZE-1)
	145
	146	#define pmd_set(pmdp, pmdval) (pmd_val(*(pmdp)) = (pmdval))
	147	#define pmd_none(pmd) (!pmd_val(pmd))
	148	#define pmd_bad(pmd) (!is_kernel_addr(pmd_val(pmd)) \
	149	\|\| (pmd_val(pmd) & PMD_BAD_BITS))
	150	#define pmd_present(pmd) (pmd_val(pmd) != 0)
	151	#define pmd_clear(pmdp) (pmd_val(*(pmdp)) = 0)
	152	#define pmd_page_vaddr(pmd) (pmd_val(pmd) & ~PMD_MASKED_BITS)
	153	#define pmd_page(pmd) virt_to_page(pmd_page_vaddr(pmd))
	154
	155	#define pud_set(pudp, pudval) (pud_val(*(pudp)) = (pudval))
	156	#define pud_none(pud) (!pud_val(pud))
	157	#define pud_bad(pud) (!is_kernel_addr(pud_val(pud)) \
	158	\|\| (pud_val(pud) & PUD_BAD_BITS))
	159	#define pud_present(pud) (pud_val(pud) != 0)
	160	#define pud_clear(pudp) (pud_val(*(pudp)) = 0)
	161	#define pud_page_vaddr(pud) (pud_val(pud) & ~PUD_MASKED_BITS)
	162	#define pud_page(pud) virt_to_page(pud_page_vaddr(pud))
	163
	164	#define pgd_set(pgdp, pudp) ({pgd_val(*(pgdp)) = (unsigned long)(pudp);})
	165
	166	/*
	167	* Find an entry in a page-table-directory. We combine the address region
	168	* (the high order N bits) and the pgd portion of the address.
	169	*/
	170	/* to avoid overflow in free_pgtables we don't use PTRS_PER_PGD here */
	171	#define pgd_index(address) (((address) >> (PGDIR_SHIFT)) & 0x1ff)
	172
	173	#define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address))
	174
	175	#define pmd_offset(pudp,addr) \
	176	(((pmd_t ) pud_page_vaddr((pudp))) + (((addr) >> PMD_SHIFT) & (PTRS_PER_PMD - 1)))
	177
	178	#define pte_offset_kernel(dir,addr) \
	179	(((pte_t ) pmd_page_vaddr((dir))) + (((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
	180
	181	#define pte_offset_map(dir,addr) pte_offset_kernel((dir), (addr))
f88df14b	182	#define pte_unmap(pte) do { } while(0)
f88df14b DG	183
	184	/* to find an entry in a kernel page-table-directory */
	185	/* This now only contains the vmalloc pages */
	186	#define pgd_offset_k(address) pgd_offset(&init_mm, address)
78f1dbde AK	187	extern void hpte_need_flush(struct mm_struct *mm, unsigned long addr,
78f1dbde AK	188	pte_t *ptep, unsigned long pte, int huge);
f88df14b DG	189
	190	/* Atomic PTE updates */
	191	static inline unsigned long pte_update(struct mm_struct *mm,
	192	unsigned long addr,
	193	pte_t *ptep, unsigned long clr,
	194	int huge)
	195	{
a033a487	196	#ifdef PTE_ATOMIC_UPDATES
f88df14b DG	197	unsigned long old, tmp;
	198
	199	__asm__ __volatile__(
	200	"1: ldarx %0,0,%3 # pte_update\n\
	201	andi. %1,%0,%6\n\
	202	bne- 1b \n\
	203	andc %1,%0,%4 \n\
	204	stdcx. %1,0,%3 \n\
	205	bne- 1b"
	206	: "=&r" (old), "=&r" (tmp), "=m" (*ptep)
	207	: "r" (ptep), "r" (clr), "m" (*ptep), "i" (_PAGE_BUSY)
	208	: "cc" );
a033a487 BH	209	#else
	210	unsigned long old = pte_val(*ptep);
	211	*ptep = __pte(old & ~clr);
	212	#endif
8d30c14c BH	213	/* huge pages use the old page table lock */
	214	if (!huge)
	215	assert_pte_locked(mm, addr);
	216
94491685	217	#ifdef CONFIG_PPC_STD_MMU_64
f88df14b DG	218	if (old & _PAGE_HASHPTE)
f88df14b DG	219	hpte_need_flush(mm, addr, ptep, old, huge);
94491685 BH	220	#endif
94491685 BH	221
f88df14b DG	222	return old;
	223	}
	224
	225	static inline int __ptep_test_and_clear_young(struct mm_struct *mm,
	226	unsigned long addr, pte_t *ptep)
	227	{
	228	unsigned long old;
	229
	230	if ((pte_val(*ptep) & (_PAGE_ACCESSED \| _PAGE_HASHPTE)) == 0)
	231	return 0;
	232	old = pte_update(mm, addr, ptep, _PAGE_ACCESSED, 0);
	233	return (old & _PAGE_ACCESSED) != 0;
	234	}
	235	#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
	236	#define ptep_test_and_clear_young(__vma, __addr, __ptep) \
	237	({ \
	238	int __r; \
	239	__r = __ptep_test_and_clear_young((__vma)->vm_mm, __addr, __ptep); \
	240	__r; \
	241	})
	242
f88df14b DG	243	#define __HAVE_ARCH_PTEP_SET_WRPROTECT
	244	static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr,
	245	pte_t *ptep)
	246	{
f88df14b	247
2a2c29c1 SP	248	if ((pte_val(*ptep) & _PAGE_RW) == 0)
	249	return;
	250
	251	pte_update(mm, addr, ptep, _PAGE_RW, 0);
f88df14b DG	252	}
f88df14b DG	253
016b33c4 AW	254	static inline void huge_ptep_set_wrprotect(struct mm_struct *mm,
	255	unsigned long addr, pte_t *ptep)
	256	{
86df8642 DG	257	if ((pte_val(*ptep) & _PAGE_RW) == 0)
86df8642 DG	258	return;
2a2c29c1 SP	259
2a2c29c1 SP	260	pte_update(mm, addr, ptep, _PAGE_RW, 1);
016b33c4	261	}
f88df14b DG	262
	263	/*
	264	* We currently remove entries from the hashtable regardless of whether
	265	* the entry was young or dirty. The generic routines only flush if the
	266	* entry was young or dirty which is not good enough.
	267	*
	268	* We should be more intelligent about this but for the moment we override
	269	* these functions and force a tlb flush unconditionally
	270	*/
	271	#define __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH
	272	#define ptep_clear_flush_young(__vma, __address, __ptep) \
	273	({ \
	274	int __young = __ptep_test_and_clear_young((__vma)->vm_mm, __address, \
	275	__ptep); \
	276	__young; \
	277	})
	278
f88df14b DG	279	#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
	280	static inline pte_t ptep_get_and_clear(struct mm_struct *mm,
	281	unsigned long addr, pte_t *ptep)
	282	{
	283	unsigned long old = pte_update(mm, addr, ptep, ~0UL, 0);
	284	return __pte(old);
	285	}
	286
	287	static inline void pte_clear(struct mm_struct *mm, unsigned long addr,
	288	pte_t * ptep)
	289	{
	290	pte_update(mm, addr, ptep, ~0UL, 0);
	291	}
	292
f88df14b DG	293
	294	/* Set the dirty and/or accessed bits atomically in a linux PTE, this
	295	* function doesn't need to flush the hash entry
	296	*/
8d30c14c	297	static inline void __ptep_set_access_flags(pte_t *ptep, pte_t entry)
f88df14b DG	298	{
f88df14b DG	299	unsigned long bits = pte_val(entry) &
ea3cc330	300	(_PAGE_DIRTY \| _PAGE_ACCESSED \| _PAGE_RW \| _PAGE_EXEC);
a033a487 BH	301
a033a487 BH	302	#ifdef PTE_ATOMIC_UPDATES
f88df14b DG	303	unsigned long old, tmp;
	304
	305	__asm__ __volatile__(
	306	"1: ldarx %0,0,%4\n\
	307	andi. %1,%0,%6\n\
	308	bne- 1b \n\
	309	or %0,%3,%0\n\
	310	stdcx. %0,0,%4\n\
	311	bne- 1b"
	312	:"=&r" (old), "=&r" (tmp), "=m" (*ptep)
	313	:"r" (bits), "r" (ptep), "m" (*ptep), "i" (_PAGE_BUSY)
	314	:"cc");
a033a487 BH	315	#else
	316	unsigned long old = pte_val(*ptep);
	317	*ptep = __pte(old \| bits);
	318	#endif
f88df14b	319	}
f88df14b	320
f88df14b DG	321	#define __HAVE_ARCH_PTE_SAME
	322	#define pte_same(A,B) (((pte_val(A) ^ pte_val(B)) & ~_PAGE_HPTEFLAGS) == 0)
	323
	324	#define pte_ERROR(e) \
	325	printk("%s:%d: bad pte %08lx.\n", __FILE__, __LINE__, pte_val(e))
	326	#define pmd_ERROR(e) \
	327	printk("%s:%d: bad pmd %08lx.\n", __FILE__, __LINE__, pmd_val(e))
	328	#define pgd_ERROR(e) \
	329	printk("%s:%d: bad pgd %08lx.\n", __FILE__, __LINE__, pgd_val(e))
	330
f88df14b DG	331	/* Encode and de-code a swap entry */
	332	#define __swp_type(entry) (((entry).val >> 1) & 0x3f)
	333	#define __swp_offset(entry) ((entry).val >> 8)
	334	#define __swp_entry(type, offset) ((swp_entry_t){((type)<< 1)\|((offset)<<8)})
	335	#define __pte_to_swp_entry(pte) ((swp_entry_t){pte_val(pte) >> PTE_RPN_SHIFT})
	336	#define __swp_entry_to_pte(x) ((pte_t) { (x).val << PTE_RPN_SHIFT })
	337	#define pte_to_pgoff(pte) (pte_val(pte) >> PTE_RPN_SHIFT)
	338	#define pgoff_to_pte(off) ((pte_t) {((off) << PTE_RPN_SHIFT)\|_PAGE_FILE})
	339	#define PTE_FILE_MAX_BITS (BITS_PER_LONG - PTE_RPN_SHIFT)
	340
a0668cdc	341	void pgtable_cache_add(unsigned shift, void (ctor)(void ));
f88df14b DG	342	void pgtable_cache_init(void);
	343
	344	/*
	345	* find_linux_pte returns the address of a linux pte for a given
	346	* effective address and directory. If not found, it returns zero.
72632ce5 BB	347	*/
72632ce5 BB	348	static inline pte_t find_linux_pte(pgd_t pgdir, unsigned long ea)
f88df14b DG	349	{
	350	pgd_t *pg;
	351	pud_t *pu;
	352	pmd_t *pm;
	353	pte_t *pt = NULL;
	354
	355	pg = pgdir + pgd_index(ea);
	356	if (!pgd_none(*pg)) {
	357	pu = pud_offset(pg, ea);
	358	if (!pud_none(*pu)) {
	359	pm = pmd_offset(pu, ea);
	360	if (pmd_present(*pm))
	361	pt = pte_offset_kernel(pm, ea);
	362	}
	363	}
	364	return pt;
	365	}
	366
a4fe3ce7 DG	367	#ifdef CONFIG_HUGETLB_PAGE
	368	pte_t find_linux_pte_or_hugepte(pgd_t pgdir, unsigned long ea,
	369	unsigned *shift);
	370	#else
	371	static inline pte_t find_linux_pte_or_hugepte(pgd_t pgdir, unsigned long ea,
	372	unsigned *shift)
	373	{
	374	if (shift)
	375	*shift = 0;
	376	return find_linux_pte(pgdir, ea);
	377	}
	378	#endif /* !CONFIG_HUGETLB_PAGE */
ce0ad7f0	379
f88df14b DG	380	#endif /* __ASSEMBLY__ */
	381
	382	#endif /* _ASM_POWERPC_PGTABLE_PPC64_H_ */