[deliverable/linux.git] / arch / powerpc / include / asm / mmu-hash64.h

#ifndef _ASM_POWERPC_MMU_HASH64_H_
#define _ASM_POWERPC_MMU_HASH64_H_
/*
 * PowerPC64 memory management structures
 *
 * Dave Engebretsen & Mike Corrigan <{engebret|mikejc}@us.ibm.com>
 *   PPC64 rework.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#include <asm/asm-compat.h>
#include <asm/page.h>

/*
 * Segment table
 */

#define STE_ESID_V	0x80
#define STE_ESID_KS	0x20
#define STE_ESID_KP	0x10
#define STE_ESID_N	0x08

#define STE_VSID_SHIFT	12

/* Location of cpu0's segment table */
#define STAB0_PAGE	0x6
#define STAB0_OFFSET	(STAB0_PAGE << 12)
#define STAB0_PHYS_ADDR	(STAB0_OFFSET + PHYSICAL_START)

#ifndef __ASSEMBLY__
extern char initial_stab[];
#endif /* ! __ASSEMBLY */

/*
 * SLB
 */

#define SLB_NUM_BOLTED		3
#define SLB_CACHE_ENTRIES	8
#define SLB_MIN_SIZE		32

/* Bits in the SLB ESID word */
#define SLB_ESID_V		ASM_CONST(0x0000000008000000) /* valid */

/* Bits in the SLB VSID word */
#define SLB_VSID_SHIFT		12
#define SLB_VSID_SHIFT_1T	24
#define SLB_VSID_SSIZE_SHIFT	62
#define SLB_VSID_B		ASM_CONST(0xc000000000000000)
#define SLB_VSID_B_256M		ASM_CONST(0x0000000000000000)
#define SLB_VSID_B_1T		ASM_CONST(0x4000000000000000)
#define SLB_VSID_KS		ASM_CONST(0x0000000000000800)
#define SLB_VSID_KP		ASM_CONST(0x0000000000000400)
#define SLB_VSID_N		ASM_CONST(0x0000000000000200) /* no-execute */
#define SLB_VSID_L		ASM_CONST(0x0000000000000100)
#define SLB_VSID_C		ASM_CONST(0x0000000000000080) /* class */
#define SLB_VSID_LP		ASM_CONST(0x0000000000000030)
#define SLB_VSID_LP_00		ASM_CONST(0x0000000000000000)
#define SLB_VSID_LP_01		ASM_CONST(0x0000000000000010)
#define SLB_VSID_LP_10		ASM_CONST(0x0000000000000020)
#define SLB_VSID_LP_11		ASM_CONST(0x0000000000000030)
#define SLB_VSID_LLP		(SLB_VSID_L|SLB_VSID_LP)

#define SLB_VSID_KERNEL		(SLB_VSID_KP)
#define SLB_VSID_USER		(SLB_VSID_KP|SLB_VSID_KS|SLB_VSID_C)

#define SLBIE_C			(0x08000000)
#define SLBIE_SSIZE_SHIFT	25

/*
 * Hash table
 */

#define HPTES_PER_GROUP 8

#define HPTE_V_SSIZE_SHIFT	62
#define HPTE_V_AVPN_SHIFT	7
#define HPTE_V_AVPN		ASM_CONST(0x3fffffffffffff80)
#define HPTE_V_AVPN_VAL(x)	(((x) & HPTE_V_AVPN) >> HPTE_V_AVPN_SHIFT)
#define HPTE_V_COMPARE(x,y)	(!(((x) ^ (y)) & 0xffffffffffffff80UL))
#define HPTE_V_BOLTED		ASM_CONST(0x0000000000000010)
#define HPTE_V_LOCK		ASM_CONST(0x0000000000000008)
#define HPTE_V_LARGE		ASM_CONST(0x0000000000000004)
#define HPTE_V_SECONDARY	ASM_CONST(0x0000000000000002)
#define HPTE_V_VALID		ASM_CONST(0x0000000000000001)

#define HPTE_R_PP0		ASM_CONST(0x8000000000000000)
#define HPTE_R_TS		ASM_CONST(0x4000000000000000)
#define HPTE_R_RPN_SHIFT	12
#define HPTE_R_RPN		ASM_CONST(0x3ffffffffffff000)
#define HPTE_R_FLAGS		ASM_CONST(0x00000000000003ff)
#define HPTE_R_PP		ASM_CONST(0x0000000000000003)
#define HPTE_R_N		ASM_CONST(0x0000000000000004)
#define HPTE_R_C		ASM_CONST(0x0000000000000080)
#define HPTE_R_R		ASM_CONST(0x0000000000000100)

#define HPTE_V_1TB_SEG		ASM_CONST(0x4000000000000000)
#define HPTE_V_VRMA_MASK	ASM_CONST(0x4001ffffff000000)

/* Values for PP (assumes Ks=0, Kp=1) */
/* pp0 will always be 0 for linux     */
#define PP_RWXX	0	/* Supervisor read/write, User none */
#define PP_RWRX 1	/* Supervisor read/write, User read */
#define PP_RWRW 2	/* Supervisor read/write, User read/write */
#define PP_RXRX 3	/* Supervisor read,       User read */

#ifndef __ASSEMBLY__

struct hash_pte {
	unsigned long v;
	unsigned long r;
};

extern struct hash_pte *htab_address;
extern unsigned long htab_size_bytes;
extern unsigned long htab_hash_mask;

/*
 * Page size definition
 *
 *    shift : is the "PAGE_SHIFT" value for that page size
 *    sllp  : is a bit mask with the value of SLB L || LP to be or'ed
 *            directly to a slbmte "vsid" value
 *    penc  : is the HPTE encoding mask for the "LP" field:
 *
 */
struct mmu_psize_def
{
	unsigned int	shift;	/* number of bits */
	unsigned int	penc;	/* HPTE encoding */
	unsigned int	tlbiel;	/* tlbiel supported for that page size */
	unsigned long	avpnm;	/* bits to mask out in AVPN in the HPTE */
	unsigned long	sllp;	/* SLB L||LP (exact mask to use in slbmte) */
};

#endif /* __ASSEMBLY__ */

/*
 * Segment sizes.
 * These are the values used by hardware in the B field of
 * SLB entries and the first dword of MMU hashtable entries.
 * The B field is 2 bits; the values 2 and 3 are unused and reserved.
 */
#define MMU_SEGSIZE_256M	0
#define MMU_SEGSIZE_1T		1


#ifndef __ASSEMBLY__

/*
 * The current system page and segment sizes
 */
extern struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT];
extern int mmu_linear_psize;
extern int mmu_virtual_psize;
extern int mmu_vmalloc_psize;
extern int mmu_vmemmap_psize;
extern int mmu_io_psize;
extern int mmu_kernel_ssize;
extern int mmu_highuser_ssize;
extern u16 mmu_slb_size;
extern unsigned long tce_alloc_start, tce_alloc_end;

/*
 * If the processor supports 64k normal pages but not 64k cache
 * inhibited pages, we have to be prepared to switch processes
 * to use 4k pages when they create cache-inhibited mappings.
 * If this is the case, mmu_ci_restrictions will be set to 1.
 */
extern int mmu_ci_restrictions;

/*
 * This function sets the AVPN and L fields of the HPTE  appropriately
 * for the page size
 */
static inline unsigned long hpte_encode_v(unsigned long va, int psize,
					  int ssize)
{
	unsigned long v;
	v = (va >> 23) & ~(mmu_psize_defs[psize].avpnm);
	v <<= HPTE_V_AVPN_SHIFT;
	if (psize != MMU_PAGE_4K)
		v |= HPTE_V_LARGE;
	v |= ((unsigned long) ssize) << HPTE_V_SSIZE_SHIFT;
	return v;
}

/*
 * This function sets the ARPN, and LP fields of the HPTE appropriately
 * for the page size. We assume the pa is already "clean" that is properly
 * aligned for the requested page size
 */
static inline unsigned long hpte_encode_r(unsigned long pa, int psize)
{
	unsigned long r;

	/* A 4K page needs no special encoding */
	if (psize == MMU_PAGE_4K)
		return pa & HPTE_R_RPN;
	else {
		unsigned int penc = mmu_psize_defs[psize].penc;
		unsigned int shift = mmu_psize_defs[psize].shift;
		return (pa & ~((1ul << shift) - 1)) | (penc << 12);
	}
	return r;
}

/*
 * Build a VA given VSID, EA and segment size
 */
static inline unsigned long hpt_va(unsigned long ea, unsigned long vsid,
				   int ssize)
{
	if (ssize == MMU_SEGSIZE_256M)
		return (vsid << 28) | (ea & 0xfffffffUL);
	return (vsid << 40) | (ea & 0xffffffffffUL);
}

/*
 * This hashes a virtual address
 */

static inline unsigned long hpt_hash(unsigned long va, unsigned int shift,
				     int ssize)
{
	unsigned long hash, vsid;

	if (ssize == MMU_SEGSIZE_256M) {
		hash = (va >> 28) ^ ((va & 0x0fffffffUL) >> shift);
	} else {
		vsid = va >> 40;
		hash = vsid ^ (vsid << 25) ^ ((va & 0xffffffffffUL) >> shift);
	}
	return hash & 0x7fffffffffUL;
}

extern int __hash_page_4K(unsigned long ea, unsigned long access,
			  unsigned long vsid, pte_t *ptep, unsigned long trap,
			  unsigned int local, int ssize, int subpage_prot);
extern int __hash_page_64K(unsigned long ea, unsigned long access,
			   unsigned long vsid, pte_t *ptep, unsigned long trap,
			   unsigned int local, int ssize);
struct mm_struct;
extern int hash_page(unsigned long ea, unsigned long access, unsigned long trap);
int __hash_page_huge(unsigned long ea, unsigned long access, unsigned long vsid,
		     pte_t *ptep, unsigned long trap, int local, int ssize,
		     unsigned int shift, unsigned int mmu_psize);

extern int htab_bolt_mapping(unsigned long vstart, unsigned long vend,
			     unsigned long pstart, unsigned long prot,
			     int psize, int ssize);
extern void add_gpage(unsigned long addr, unsigned long page_size,
			  unsigned long number_of_pages);
extern void demote_segment_4k(struct mm_struct *mm, unsigned long addr);

extern void hpte_init_native(void);
extern void hpte_init_lpar(void);
extern void hpte_init_iSeries(void);
extern void hpte_init_beat(void);
extern void hpte_init_beat_v3(void);

extern void stabs_alloc(void);
extern void slb_initialize(void);
extern void slb_flush_and_rebolt(void);
extern void stab_initialize(unsigned long stab);

extern void slb_vmalloc_update(void);
extern void slb_set_size(u16 size);
#endif /* __ASSEMBLY__ */

/*
 * VSID allocation
 *
 * We first generate a 36-bit "proto-VSID".  For kernel addresses this
 * is equal to the ESID, for user addresses it is:
 *	(context << 15) | (esid & 0x7fff)
 *
 * The two forms are distinguishable because the top bit is 0 for user
 * addresses, whereas the top two bits are 1 for kernel addresses.
 * Proto-VSIDs with the top two bits equal to 0b10 are reserved for
 * now.
 *
 * The proto-VSIDs are then scrambled into real VSIDs with the
 * multiplicative hash:
 *
 *	VSID = (proto-VSID * VSID_MULTIPLIER) % VSID_MODULUS
 *	where	VSID_MULTIPLIER = 268435399 = 0xFFFFFC7
 *		VSID_MODULUS = 2^36-1 = 0xFFFFFFFFF
 *
 * This scramble is only well defined for proto-VSIDs below
 * 0xFFFFFFFFF, so both proto-VSID and actual VSID 0xFFFFFFFFF are
 * reserved.  VSID_MULTIPLIER is prime, so in particular it is
 * co-prime to VSID_MODULUS, making this a 1:1 scrambling function.
 * Because the modulus is 2^n-1 we can compute it efficiently without
 * a divide or extra multiply (see below).
 *
 * This scheme has several advantages over older methods:
 *
 * 	- We have VSIDs allocated for every kernel address
 * (i.e. everything above 0xC000000000000000), except the very top
 * segment, which simplifies several things.
 *
 * 	- We allow for 15 significant bits of ESID and 20 bits of
 * context for user addresses.  i.e. 8T (43 bits) of address space for
 * up to 1M contexts (although the page table structure and context
 * allocation will need changes to take advantage of this).
 *
 * 	- The scramble function gives robust scattering in the hash
 * table (at least based on some initial results).  The previous
 * method was more susceptible to pathological cases giving excessive
 * hash collisions.
 */
/*
 * WARNING - If you change these you must make sure the asm
 * implementations in slb_allocate (slb_low.S), do_stab_bolted
 * (head.S) and ASM_VSID_SCRAMBLE (below) are changed accordingly.
 *
 * You'll also need to change the precomputed VSID values in head.S
 * which are used by the iSeries firmware.
 */

#define VSID_MULTIPLIER_256M	ASM_CONST(200730139)	/* 28-bit prime */
#define VSID_BITS_256M		36
#define VSID_MODULUS_256M	((1UL<<VSID_BITS_256M)-1)

#define VSID_MULTIPLIER_1T	ASM_CONST(12538073)	/* 24-bit prime */
#define VSID_BITS_1T		24
#define VSID_MODULUS_1T		((1UL<<VSID_BITS_1T)-1)

#define CONTEXT_BITS		19
#define USER_ESID_BITS		16
#define USER_ESID_BITS_1T	4

#define USER_VSID_RANGE	(1UL << (USER_ESID_BITS + SID_SHIFT))

/*
 * This macro generates asm code to compute the VSID scramble
 * function.  Used in slb_allocate() and do_stab_bolted.  The function
 * computed is: (protovsid*VSID_MULTIPLIER) % VSID_MODULUS
 *
 *	rt = register continaing the proto-VSID and into which the
 *		VSID will be stored
 *	rx = scratch register (clobbered)
 *
 * 	- rt and rx must be different registers
 * 	- The answer will end up in the low VSID_BITS bits of rt.  The higher
 * 	  bits may contain other garbage, so you may need to mask the
 * 	  result.
 */
#define ASM_VSID_SCRAMBLE(rt, rx, size)					\
	lis	rx,VSID_MULTIPLIER_##size@h;				\
	ori	rx,rx,VSID_MULTIPLIER_##size@l;				\
	mulld	rt,rt,rx;		/* rt = rt * MULTIPLIER */	\
									\
	srdi	rx,rt,VSID_BITS_##size;					\
	clrldi	rt,rt,(64-VSID_BITS_##size);				\
	add	rt,rt,rx;		/* add high and low bits */	\
	/* Now, r3 == VSID (mod 2^36-1), and lies between 0 and		\
	 * 2^36-1+2^28-1.  That in particular means that if r3 >=	\
	 * 2^36-1, then r3+1 has the 2^36 bit set.  So, if r3+1 has	\
	 * the bit clear, r3 already has the answer we want, if it	\
	 * doesn't, the answer is the low 36 bits of r3+1.  So in all	\
	 * cases the answer is the low 36 bits of (r3 + ((r3+1) >> 36))*/\
	addi	rx,rt,1;						\
	srdi	rx,rx,VSID_BITS_##size;	/* extract 2^VSID_BITS bit */	\
	add	rt,rt,rx


#ifndef __ASSEMBLY__

typedef unsigned long mm_context_id_t;

typedef struct {
	mm_context_id_t id;
	u16 user_psize;		/* page size index */

#ifdef CONFIG_PPC_MM_SLICES
	u64 low_slices_psize;	/* SLB page size encodings */
	u64 high_slices_psize;  /* 4 bits per slice for now */
#else
	u16 sllp;		/* SLB page size encoding */
#endif
	unsigned long vdso_base;
} mm_context_t;


#if 0
/*
 * The code below is equivalent to this function for arguments
 * < 2^VSID_BITS, which is all this should ever be called
 * with.  However gcc is not clever enough to compute the
 * modulus (2^n-1) without a second multiply.
 */
#define vsid_scrample(protovsid, size) \
	((((protovsid) * VSID_MULTIPLIER_##size) % VSID_MODULUS_##size))

#else /* 1 */
#define vsid_scramble(protovsid, size) \
	({								 \
		unsigned long x;					 \
		x = (protovsid) * VSID_MULTIPLIER_##size;		 \
		x = (x >> VSID_BITS_##size) + (x & VSID_MODULUS_##size); \
		(x + ((x+1) >> VSID_BITS_##size)) & VSID_MODULUS_##size; \
	})
#endif /* 1 */

/* This is only valid for addresses >= PAGE_OFFSET */
static inline unsigned long get_kernel_vsid(unsigned long ea, int ssize)
{
	if (ssize == MMU_SEGSIZE_256M)
		return vsid_scramble(ea >> SID_SHIFT, 256M);
	return vsid_scramble(ea >> SID_SHIFT_1T, 1T);
}

/* Returns the segment size indicator for a user address */
static inline int user_segment_size(unsigned long addr)
{
	/* Use 1T segments if possible for addresses >= 1T */
	if (addr >= (1UL << SID_SHIFT_1T))
		return mmu_highuser_ssize;
	return MMU_SEGSIZE_256M;
}

/* This is only valid for user addresses (which are below 2^44) */
static inline unsigned long get_vsid(unsigned long context, unsigned long ea,
				     int ssize)
{
	if (ssize == MMU_SEGSIZE_256M)
		return vsid_scramble((context << USER_ESID_BITS)
				     | (ea >> SID_SHIFT), 256M);
	return vsid_scramble((context << USER_ESID_BITS_1T)
			     | (ea >> SID_SHIFT_1T), 1T);
}

/*
 * This is only used on legacy iSeries in lparmap.c,
 * hence the 256MB segment assumption.
 */
#define VSID_SCRAMBLE(pvsid)	(((pvsid) * VSID_MULTIPLIER_256M) %	\
				 VSID_MODULUS_256M)
#define KERNEL_VSID(ea)		VSID_SCRAMBLE(GET_ESID(ea))

#endif /* __ASSEMBLY__ */

#endif /* _ASM_POWERPC_MMU_HASH64_H_ */
Commit	Line	Data
8d2169e8 DG	1	#ifndef _ASM_POWERPC_MMU_HASH64_H_
	2	#define _ASM_POWERPC_MMU_HASH64_H_
	3	/*
	4	* PowerPC64 memory management structures
	5	*
	6	* Dave Engebretsen & Mike Corrigan <{engebret\|mikejc}@us.ibm.com>
	7	* PPC64 rework.
	8	*
	9	* This program is free software; you can redistribute it and/or
	10	* modify it under the terms of the GNU General Public License
	11	* as published by the Free Software Foundation; either version
	12	* 2 of the License, or (at your option) any later version.
	13	*/
	14
	15	#include <asm/asm-compat.h>
	16	#include <asm/page.h>
	17
	18	/*
	19	* Segment table
	20	*/
	21
	22	#define STE_ESID_V 0x80
	23	#define STE_ESID_KS 0x20
	24	#define STE_ESID_KP 0x10
	25	#define STE_ESID_N 0x08
	26
	27	#define STE_VSID_SHIFT 12
	28
	29	/* Location of cpu0's segment table */
	30	#define STAB0_PAGE 0x6
	31	#define STAB0_OFFSET (STAB0_PAGE << 12)
	32	#define STAB0_PHYS_ADDR (STAB0_OFFSET + PHYSICAL_START)
	33
	34	#ifndef __ASSEMBLY__
	35	extern char initial_stab[];
	36	#endif /* ! __ASSEMBLY */
	37
	38	/*
	39	* SLB
	40	*/
	41
	42	#define SLB_NUM_BOLTED 3
	43	#define SLB_CACHE_ENTRIES 8
46db2f86	44	#define SLB_MIN_SIZE 32
8d2169e8 DG	45
	46	/* Bits in the SLB ESID word */
	47	#define SLB_ESID_V ASM_CONST(0x0000000008000000) /* valid */
	48
	49	/* Bits in the SLB VSID word */
	50	#define SLB_VSID_SHIFT 12
1189be65 PM	51	#define SLB_VSID_SHIFT_1T 24
1189be65 PM	52	#define SLB_VSID_SSIZE_SHIFT 62
8d2169e8 DG	53	#define SLB_VSID_B ASM_CONST(0xc000000000000000)
	54	#define SLB_VSID_B_256M ASM_CONST(0x0000000000000000)
	55	#define SLB_VSID_B_1T ASM_CONST(0x4000000000000000)
	56	#define SLB_VSID_KS ASM_CONST(0x0000000000000800)
	57	#define SLB_VSID_KP ASM_CONST(0x0000000000000400)
	58	#define SLB_VSID_N ASM_CONST(0x0000000000000200) /* no-execute */
	59	#define SLB_VSID_L ASM_CONST(0x0000000000000100)
	60	#define SLB_VSID_C ASM_CONST(0x0000000000000080) /* class */
	61	#define SLB_VSID_LP ASM_CONST(0x0000000000000030)
	62	#define SLB_VSID_LP_00 ASM_CONST(0x0000000000000000)
	63	#define SLB_VSID_LP_01 ASM_CONST(0x0000000000000010)
	64	#define SLB_VSID_LP_10 ASM_CONST(0x0000000000000020)
	65	#define SLB_VSID_LP_11 ASM_CONST(0x0000000000000030)
	66	#define SLB_VSID_LLP (SLB_VSID_L\|SLB_VSID_LP)
	67
	68	#define SLB_VSID_KERNEL (SLB_VSID_KP)
	69	#define SLB_VSID_USER (SLB_VSID_KP\|SLB_VSID_KS\|SLB_VSID_C)
	70
	71	#define SLBIE_C (0x08000000)
1189be65	72	#define SLBIE_SSIZE_SHIFT 25
8d2169e8 DG	73
	74	/*
	75	* Hash table
	76	*/
	77
	78	#define HPTES_PER_GROUP 8
	79
2454c7e9	80	#define HPTE_V_SSIZE_SHIFT 62
8d2169e8	81	#define HPTE_V_AVPN_SHIFT 7
2454c7e9	82	#define HPTE_V_AVPN ASM_CONST(0x3fffffffffffff80)
8d2169e8	83	#define HPTE_V_AVPN_VAL(x) (((x) & HPTE_V_AVPN) >> HPTE_V_AVPN_SHIFT)
91bbbe22	84	#define HPTE_V_COMPARE(x,y) (!(((x) ^ (y)) & 0xffffffffffffff80UL))
8d2169e8 DG	85	#define HPTE_V_BOLTED ASM_CONST(0x0000000000000010)
	86	#define HPTE_V_LOCK ASM_CONST(0x0000000000000008)
	87	#define HPTE_V_LARGE ASM_CONST(0x0000000000000004)
	88	#define HPTE_V_SECONDARY ASM_CONST(0x0000000000000002)
	89	#define HPTE_V_VALID ASM_CONST(0x0000000000000001)
	90
	91	#define HPTE_R_PP0 ASM_CONST(0x8000000000000000)
	92	#define HPTE_R_TS ASM_CONST(0x4000000000000000)
	93	#define HPTE_R_RPN_SHIFT 12
	94	#define HPTE_R_RPN ASM_CONST(0x3ffffffffffff000)
	95	#define HPTE_R_FLAGS ASM_CONST(0x00000000000003ff)
	96	#define HPTE_R_PP ASM_CONST(0x0000000000000003)
	97	#define HPTE_R_N ASM_CONST(0x0000000000000004)
	98	#define HPTE_R_C ASM_CONST(0x0000000000000080)
	99	#define HPTE_R_R ASM_CONST(0x0000000000000100)
	100
b7abc5c5 SS	101	#define HPTE_V_1TB_SEG ASM_CONST(0x4000000000000000)
	102	#define HPTE_V_VRMA_MASK ASM_CONST(0x4001ffffff000000)
	103
8d2169e8 DG	104	/* Values for PP (assumes Ks=0, Kp=1) */
	105	/* pp0 will always be 0 for linux */
	106	#define PP_RWXX 0 /* Supervisor read/write, User none */
	107	#define PP_RWRX 1 /* Supervisor read/write, User read */
	108	#define PP_RWRW 2 /* Supervisor read/write, User read/write */
	109	#define PP_RXRX 3 /* Supervisor read, User read */
	110
	111	#ifndef __ASSEMBLY__
	112
8e561e7e	113	struct hash_pte {
8d2169e8 DG	114	unsigned long v;
8d2169e8 DG	115	unsigned long r;
8e561e7e	116	};
8d2169e8	117
8e561e7e	118	extern struct hash_pte *htab_address;
8d2169e8 DG	119	extern unsigned long htab_size_bytes;
	120	extern unsigned long htab_hash_mask;
	121
	122	/*
	123	* Page size definition
	124	*
	125	* shift : is the "PAGE_SHIFT" value for that page size
	126	* sllp : is a bit mask with the value of SLB L \|\| LP to be or'ed
	127	* directly to a slbmte "vsid" value
	128	* penc : is the HPTE encoding mask for the "LP" field:
	129	*
	130	*/
	131	struct mmu_psize_def
	132	{
	133	unsigned int shift; /* number of bits */
	134	unsigned int penc; /* HPTE encoding */
	135	unsigned int tlbiel; /* tlbiel supported for that page size */
	136	unsigned long avpnm; /* bits to mask out in AVPN in the HPTE */
	137	unsigned long sllp; /* SLB L\|\|LP (exact mask to use in slbmte) */
	138	};
	139
	140	#endif /* __ASSEMBLY__ */
	141
2454c7e9 PM	142	/*
	143	* Segment sizes.
	144	* These are the values used by hardware in the B field of
	145	* SLB entries and the first dword of MMU hashtable entries.
	146	* The B field is 2 bits; the values 2 and 3 are unused and reserved.
	147	*/
	148	#define MMU_SEGSIZE_256M 0
	149	#define MMU_SEGSIZE_1T 1
	150
1189be65	151
8d2169e8 DG	152	#ifndef __ASSEMBLY__
	153
	154	/*
1189be65	155	* The current system page and segment sizes
8d2169e8 DG	156	*/
	157	extern struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT];
	158	extern int mmu_linear_psize;
	159	extern int mmu_virtual_psize;
	160	extern int mmu_vmalloc_psize;
cec08e7a	161	extern int mmu_vmemmap_psize;
8d2169e8	162	extern int mmu_io_psize;
1189be65 PM	163	extern int mmu_kernel_ssize;
1189be65 PM	164	extern int mmu_highuser_ssize;
584f8b71	165	extern u16 mmu_slb_size;
572fb578	166	extern unsigned long tce_alloc_start, tce_alloc_end;
8d2169e8 DG	167
	168	/*
	169	* If the processor supports 64k normal pages but not 64k cache
	170	* inhibited pages, we have to be prepared to switch processes
	171	* to use 4k pages when they create cache-inhibited mappings.
	172	* If this is the case, mmu_ci_restrictions will be set to 1.
	173	*/
	174	extern int mmu_ci_restrictions;
	175
8d2169e8 DG	176	/*
	177	* This function sets the AVPN and L fields of the HPTE appropriately
	178	* for the page size
	179	*/
1189be65 PM	180	static inline unsigned long hpte_encode_v(unsigned long va, int psize,
1189be65 PM	181	int ssize)
8d2169e8	182	{
1189be65	183	unsigned long v;
8d2169e8 DG	184	v = (va >> 23) & ~(mmu_psize_defs[psize].avpnm);
	185	v <<= HPTE_V_AVPN_SHIFT;
	186	if (psize != MMU_PAGE_4K)
	187	v \|= HPTE_V_LARGE;
1189be65	188	v \|= ((unsigned long) ssize) << HPTE_V_SSIZE_SHIFT;
8d2169e8 DG	189	return v;
	190	}
	191
	192	/*
	193	* This function sets the ARPN, and LP fields of the HPTE appropriately
	194	* for the page size. We assume the pa is already "clean" that is properly
	195	* aligned for the requested page size
	196	*/
	197	static inline unsigned long hpte_encode_r(unsigned long pa, int psize)
	198	{
	199	unsigned long r;
	200
	201	/* A 4K page needs no special encoding */
	202	if (psize == MMU_PAGE_4K)
	203	return pa & HPTE_R_RPN;
	204	else {
	205	unsigned int penc = mmu_psize_defs[psize].penc;
	206	unsigned int shift = mmu_psize_defs[psize].shift;
	207	return (pa & ~((1ul << shift) - 1)) \| (penc << 12);
	208	}
	209	return r;
	210	}
	211
	212	/*
1189be65	213	* Build a VA given VSID, EA and segment size
8d2169e8	214	*/
1189be65 PM	215	static inline unsigned long hpt_va(unsigned long ea, unsigned long vsid,
	216	int ssize)
	217	{
	218	if (ssize == MMU_SEGSIZE_256M)
	219	return (vsid << 28) \| (ea & 0xfffffffUL);
	220	return (vsid << 40) \| (ea & 0xffffffffffUL);
	221	}
8d2169e8	222
1189be65 PM	223	/*
	224	* This hashes a virtual address
	225	*/
	226
	227	static inline unsigned long hpt_hash(unsigned long va, unsigned int shift,
	228	int ssize)
8d2169e8	229	{
1189be65 PM	230	unsigned long hash, vsid;
	231
	232	if (ssize == MMU_SEGSIZE_256M) {
	233	hash = (va >> 28) ^ ((va & 0x0fffffffUL) >> shift);
	234	} else {
	235	vsid = va >> 40;
	236	hash = vsid ^ (vsid << 25) ^ ((va & 0xffffffffffUL) >> shift);
	237	}
	238	return hash & 0x7fffffffffUL;
8d2169e8 DG	239	}
	240
	241	extern int __hash_page_4K(unsigned long ea, unsigned long access,
	242	unsigned long vsid, pte_t *ptep, unsigned long trap,
fa28237c	243	unsigned int local, int ssize, int subpage_prot);
8d2169e8 DG	244	extern int __hash_page_64K(unsigned long ea, unsigned long access,
8d2169e8 DG	245	unsigned long vsid, pte_t *ptep, unsigned long trap,
1189be65	246	unsigned int local, int ssize);
8d2169e8 DG	247	struct mm_struct;
8d2169e8 DG	248	extern int hash_page(unsigned long ea, unsigned long access, unsigned long trap);
a4fe3ce7 DG	249	int __hash_page_huge(unsigned long ea, unsigned long access, unsigned long vsid,
	250	pte_t *ptep, unsigned long trap, int local, int ssize,
	251	unsigned int shift, unsigned int mmu_psize);
8d2169e8 DG	252
8d2169e8 DG	253	extern int htab_bolt_mapping(unsigned long vstart, unsigned long vend,
bc033b63	254	unsigned long pstart, unsigned long prot,
1189be65	255	int psize, int ssize);
658013e9 JT	256	extern void add_gpage(unsigned long addr, unsigned long page_size,
658013e9 JT	257	unsigned long number_of_pages);
fa28237c	258	extern void demote_segment_4k(struct mm_struct *mm, unsigned long addr);
8d2169e8	259
8d2169e8 DG	260	extern void hpte_init_native(void);
	261	extern void hpte_init_lpar(void);
	262	extern void hpte_init_iSeries(void);
	263	extern void hpte_init_beat(void);
7f2c8577	264	extern void hpte_init_beat_v3(void);
8d2169e8 DG	265
	266	extern void stabs_alloc(void);
	267	extern void slb_initialize(void);
	268	extern void slb_flush_and_rebolt(void);
	269	extern void stab_initialize(unsigned long stab);
	270
67439b76	271	extern void slb_vmalloc_update(void);
46db2f86	272	extern void slb_set_size(u16 size);
8d2169e8 DG	273	#endif /* __ASSEMBLY__ */
	274
	275	/*
	276	* VSID allocation
	277	*
	278	* We first generate a 36-bit "proto-VSID". For kernel addresses this
	279	* is equal to the ESID, for user addresses it is:
	280	* (context << 15) \| (esid & 0x7fff)
	281	*
	282	* The two forms are distinguishable because the top bit is 0 for user
	283	* addresses, whereas the top two bits are 1 for kernel addresses.
	284	* Proto-VSIDs with the top two bits equal to 0b10 are reserved for
	285	* now.
	286	*
	287	* The proto-VSIDs are then scrambled into real VSIDs with the
	288	* multiplicative hash:
	289	*
	290	* VSID = (proto-VSID * VSID_MULTIPLIER) % VSID_MODULUS
	291	* where VSID_MULTIPLIER = 268435399 = 0xFFFFFC7
	292	* VSID_MODULUS = 2^36-1 = 0xFFFFFFFFF
	293	*
	294	* This scramble is only well defined for proto-VSIDs below
	295	* 0xFFFFFFFFF, so both proto-VSID and actual VSID 0xFFFFFFFFF are
	296	* reserved. VSID_MULTIPLIER is prime, so in particular it is
	297	* co-prime to VSID_MODULUS, making this a 1:1 scrambling function.
	298	* Because the modulus is 2^n-1 we can compute it efficiently without
	299	* a divide or extra multiply (see below).
	300	*
	301	* This scheme has several advantages over older methods:
	302	*
	303	* - We have VSIDs allocated for every kernel address
	304	* (i.e. everything above 0xC000000000000000), except the very top
	305	* segment, which simplifies several things.
	306	*
	307	* - We allow for 15 significant bits of ESID and 20 bits of
	308	* context for user addresses. i.e. 8T (43 bits) of address space for
	309	* up to 1M contexts (although the page table structure and context
	310	* allocation will need changes to take advantage of this).
	311	*
	312	* - The scramble function gives robust scattering in the hash
	313	* table (at least based on some initial results). The previous
	314	* method was more susceptible to pathological cases giving excessive
	315	* hash collisions.
	316	*/
	317	/*
	318	* WARNING - If you change these you must make sure the asm
	319	* implementations in slb_allocate (slb_low.S), do_stab_bolted
	320	* (head.S) and ASM_VSID_SCRAMBLE (below) are changed accordingly.
	321	*
	322	* You'll also need to change the precomputed VSID values in head.S
	323	* which are used by the iSeries firmware.
	324	*/
	325
1189be65 PM	326	#define VSID_MULTIPLIER_256M ASM_CONST(200730139) /* 28-bit prime */
	327	#define VSID_BITS_256M 36
	328	#define VSID_MODULUS_256M ((1UL<<VSID_BITS_256M)-1)
8d2169e8	329
1189be65 PM	330	#define VSID_MULTIPLIER_1T ASM_CONST(12538073) /* 24-bit prime */
	331	#define VSID_BITS_1T 24
	332	#define VSID_MODULUS_1T ((1UL<<VSID_BITS_1T)-1)
	333
	334	#define CONTEXT_BITS 19
	335	#define USER_ESID_BITS 16
	336	#define USER_ESID_BITS_1T 4
8d2169e8 DG	337
	338	#define USER_VSID_RANGE (1UL << (USER_ESID_BITS + SID_SHIFT))
	339
	340	/*
	341	* This macro generates asm code to compute the VSID scramble
	342	* function. Used in slb_allocate() and do_stab_bolted. The function
	343	* computed is: (protovsid*VSID_MULTIPLIER) % VSID_MODULUS
	344	*
	345	* rt = register continaing the proto-VSID and into which the
	346	* VSID will be stored
	347	* rx = scratch register (clobbered)
	348	*
	349	* - rt and rx must be different registers
1189be65	350	* - The answer will end up in the low VSID_BITS bits of rt. The higher
8d2169e8 DG	351	* bits may contain other garbage, so you may need to mask the
	352	* result.
	353	*/
1189be65 PM	354	#define ASM_VSID_SCRAMBLE(rt, rx, size) \
	355	lis rx,VSID_MULTIPLIER_##size@h; \
	356	ori rx,rx,VSID_MULTIPLIER_##size@l; \
8d2169e8 DG	357	mulld rt,rt,rx; /* rt = rt * MULTIPLIER */ \
8d2169e8 DG	358	\
1189be65 PM	359	srdi rx,rt,VSID_BITS_##size; \
1189be65 PM	360	clrldi rt,rt,(64-VSID_BITS_##size); \
8d2169e8 DG	361	add rt,rt,rx; /* add high and low bits */ \
	362	/* Now, r3 == VSID (mod 2^36-1), and lies between 0 and \
	363	* 2^36-1+2^28-1. That in particular means that if r3 >= \
	364	* 2^36-1, then r3+1 has the 2^36 bit set. So, if r3+1 has \
	365	* the bit clear, r3 already has the answer we want, if it \
	366	* doesn't, the answer is the low 36 bits of r3+1. So in all \
	367	* cases the answer is the low 36 bits of (r3 + ((r3+1) >> 36))*/\
	368	addi rx,rt,1; \
1189be65	369	srdi rx,rx,VSID_BITS_##size; /* extract 2^VSID_BITS bit */ \
8d2169e8 DG	370	add rt,rt,rx
	371
	372
	373	#ifndef __ASSEMBLY__
	374
	375	typedef unsigned long mm_context_id_t;
	376
	377	typedef struct {
	378	mm_context_id_t id;
d0f13e3c BH	379	u16 user_psize; /* page size index */
	380
	381	#ifdef CONFIG_PPC_MM_SLICES
	382	u64 low_slices_psize; /* SLB page size encodings */
	383	u64 high_slices_psize; /* 4 bits per slice for now */
	384	#else
	385	u16 sllp; /* SLB page size encoding */
8d2169e8 DG	386	#endif
	387	unsigned long vdso_base;
	388	} mm_context_t;
	389
	390
8d2169e8	391	#if 0
1189be65 PM	392	/*
	393	* The code below is equivalent to this function for arguments
	394	* < 2^VSID_BITS, which is all this should ever be called
	395	* with. However gcc is not clever enough to compute the
	396	* modulus (2^n-1) without a second multiply.
	397	*/
	398	#define vsid_scrample(protovsid, size) \
	399	((((protovsid) * VSID_MULTIPLIER_##size) % VSID_MODULUS_##size))
8d2169e8	400
1189be65 PM	401	#else /* 1 */
	402	#define vsid_scramble(protovsid, size) \
	403	({ \
	404	unsigned long x; \
	405	x = (protovsid) * VSID_MULTIPLIER_##size; \
	406	x = (x >> VSID_BITS_##size) + (x & VSID_MODULUS_##size); \
	407	(x + ((x+1) >> VSID_BITS_##size)) & VSID_MODULUS_##size; \
	408	})
8d2169e8	409	#endif /* 1 */
8d2169e8	410
549e8152	411	/* This is only valid for addresses >= PAGE_OFFSET */
1189be65	412	static inline unsigned long get_kernel_vsid(unsigned long ea, int ssize)
8d2169e8	413	{
1189be65 PM	414	if (ssize == MMU_SEGSIZE_256M)
	415	return vsid_scramble(ea >> SID_SHIFT, 256M);
	416	return vsid_scramble(ea >> SID_SHIFT_1T, 1T);
8d2169e8 DG	417	}
8d2169e8 DG	418
1189be65 PM	419	/* Returns the segment size indicator for a user address */
1189be65 PM	420	static inline int user_segment_size(unsigned long addr)
8d2169e8	421	{
1189be65 PM	422	/* Use 1T segments if possible for addresses >= 1T */
	423	if (addr >= (1UL << SID_SHIFT_1T))
	424	return mmu_highuser_ssize;
	425	return MMU_SEGSIZE_256M;
8d2169e8 DG	426	}
8d2169e8 DG	427
1189be65 PM	428	/* This is only valid for user addresses (which are below 2^44) */
	429	static inline unsigned long get_vsid(unsigned long context, unsigned long ea,
	430	int ssize)
	431	{
	432	if (ssize == MMU_SEGSIZE_256M)
	433	return vsid_scramble((context << USER_ESID_BITS)
	434	\| (ea >> SID_SHIFT), 256M);
	435	return vsid_scramble((context << USER_ESID_BITS_1T)
	436	\| (ea >> SID_SHIFT_1T), 1T);
	437	}
	438
	439	/*
	440	* This is only used on legacy iSeries in lparmap.c,
	441	* hence the 256MB segment assumption.
	442	*/
	443	#define VSID_SCRAMBLE(pvsid) (((pvsid) * VSID_MULTIPLIER_256M) % \
	444	VSID_MODULUS_256M)
8d2169e8 DG	445	#define KERNEL_VSID(ea) VSID_SCRAMBLE(GET_ESID(ea))
8d2169e8 DG	446
8d2169e8 DG	447	#endif /* __ASSEMBLY__ */
	448
	449	#endif /* _ASM_POWERPC_MMU_HASH64_H_ */