pgtable.h 42.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *  S390 version
3
 *    Copyright IBM Corp. 1999, 2000
L
Linus Torvalds 已提交
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
 *    Author(s): Hartmut Penner (hp@de.ibm.com)
 *               Ulrich Weigand (weigand@de.ibm.com)
 *               Martin Schwidefsky (schwidefsky@de.ibm.com)
 *
 *  Derived from "include/asm-i386/pgtable.h"
 */

#ifndef _ASM_S390_PGTABLE_H
#define _ASM_S390_PGTABLE_H

/*
 * The Linux memory management assumes a three-level page table setup. For
 * s390 31 bit we "fold" the mid level into the top-level page table, so
 * that we physically have the same two-level page table as the s390 mmu
 * expects in 31 bit mode. For s390 64 bit we use three of the five levels
 * the hardware provides (region first and region second tables are not
 * used).
 *
 * The "pgd_xxx()" functions are trivial for a folded two-level
 * setup: the pgd is never bad, and a pmd always exists (as it's folded
 * into the pgd entry)
 *
 * This file contains the functions and defines necessary to modify and use
 * the S390 page table tree.
 */
#ifndef __ASSEMBLY__
30
#include <linux/sched.h>
31
#include <linux/mm_types.h>
L
Linus Torvalds 已提交
32
#include <asm/bug.h>
33
#include <asm/page.h>
L
Linus Torvalds 已提交
34 35 36

extern pgd_t swapper_pg_dir[] __attribute__ ((aligned (4096)));
extern void paging_init(void);
37
extern void vmem_map_init(void);
L
Linus Torvalds 已提交
38 39 40 41 42

/*
 * The S390 doesn't have any external MMU info: the kernel page
 * tables contain all the necessary information.
 */
43
#define update_mmu_cache(vma, address, ptep)     do { } while (0)
44
#define update_mmu_cache_pmd(vma, address, ptep) do { } while (0)
L
Linus Torvalds 已提交
45 46

/*
47
 * ZERO_PAGE is a global shared page that is always zero; used
L
Linus Torvalds 已提交
48 49
 * for zero-mapped memory areas etc..
 */
50 51 52 53 54 55 56

extern unsigned long empty_zero_page;
extern unsigned long zero_page_mask;

#define ZERO_PAGE(vaddr) \
	(virt_to_page((void *)(empty_zero_page + \
	 (((unsigned long)(vaddr)) &zero_page_mask))))
57
#define __HAVE_COLOR_ZERO_PAGE
58

L
Linus Torvalds 已提交
59 60 61 62 63 64 65
#endif /* !__ASSEMBLY__ */

/*
 * PMD_SHIFT determines the size of the area a second-level page
 * table can map
 * PGDIR_SHIFT determines what a third-level page table entry can map
 */
66
#ifndef CONFIG_64BIT
67 68 69
# define PMD_SHIFT	20
# define PUD_SHIFT	20
# define PGDIR_SHIFT	20
70
#else /* CONFIG_64BIT */
71
# define PMD_SHIFT	20
M
Martin Schwidefsky 已提交
72
# define PUD_SHIFT	31
73
# define PGDIR_SHIFT	42
74
#endif /* CONFIG_64BIT */
L
Linus Torvalds 已提交
75 76 77

#define PMD_SIZE        (1UL << PMD_SHIFT)
#define PMD_MASK        (~(PMD_SIZE-1))
M
Martin Schwidefsky 已提交
78 79
#define PUD_SIZE	(1UL << PUD_SHIFT)
#define PUD_MASK	(~(PUD_SIZE-1))
80 81
#define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
#define PGDIR_MASK	(~(PGDIR_SIZE-1))
L
Linus Torvalds 已提交
82 83 84 85 86 87 88

/*
 * entries per page directory level: the S390 is two-level, so
 * we don't really have any PMD directory physically.
 * for S390 segment-table entries are combined to one PGD
 * that leads to 1024 pte per pgd
 */
89
#define PTRS_PER_PTE	256
90
#ifndef CONFIG_64BIT
91
#define PTRS_PER_PMD	1
92
#define PTRS_PER_PUD	1
93
#else /* CONFIG_64BIT */
94
#define PTRS_PER_PMD	2048
95
#define PTRS_PER_PUD	2048
96
#endif /* CONFIG_64BIT */
97
#define PTRS_PER_PGD	2048
L
Linus Torvalds 已提交
98

99 100
#define FIRST_USER_ADDRESS  0

L
Linus Torvalds 已提交
101 102 103 104
#define pte_ERROR(e) \
	printk("%s:%d: bad pte %p.\n", __FILE__, __LINE__, (void *) pte_val(e))
#define pmd_ERROR(e) \
	printk("%s:%d: bad pmd %p.\n", __FILE__, __LINE__, (void *) pmd_val(e))
M
Martin Schwidefsky 已提交
105 106
#define pud_ERROR(e) \
	printk("%s:%d: bad pud %p.\n", __FILE__, __LINE__, (void *) pud_val(e))
L
Linus Torvalds 已提交
107 108 109 110 111
#define pgd_ERROR(e) \
	printk("%s:%d: bad pgd %p.\n", __FILE__, __LINE__, (void *) pgd_val(e))

#ifndef __ASSEMBLY__
/*
112 113 114 115 116 117
 * The vmalloc and module area will always be on the topmost area of the kernel
 * mapping. We reserve 96MB (31bit) / 128GB (64bit) for vmalloc and modules.
 * On 64 bit kernels we have a 2GB area at the top of the vmalloc area where
 * modules will reside. That makes sure that inter module branches always
 * happen without trampolines and in addition the placement within a 2GB frame
 * is branch prediction unit friendly.
H
Heiko Carstens 已提交
118
 */
119
extern unsigned long VMALLOC_START;
120 121
extern unsigned long VMALLOC_END;
extern struct page *vmemmap;
122

123
#define VMEM_MAX_PHYS ((unsigned long) vmemmap)
124

125 126 127 128 129 130 131 132
#ifdef CONFIG_64BIT
extern unsigned long MODULES_VADDR;
extern unsigned long MODULES_END;
#define MODULES_VADDR	MODULES_VADDR
#define MODULES_END	MODULES_END
#define MODULES_LEN	(1UL << 31)
#endif

L
Linus Torvalds 已提交
133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166
/*
 * A 31 bit pagetable entry of S390 has following format:
 *  |   PFRA          |    |  OS  |
 * 0                   0IP0
 * 00000000001111111111222222222233
 * 01234567890123456789012345678901
 *
 * I Page-Invalid Bit:    Page is not available for address-translation
 * P Page-Protection Bit: Store access not possible for page
 *
 * A 31 bit segmenttable entry of S390 has following format:
 *  |   P-table origin      |  |PTL
 * 0                         IC
 * 00000000001111111111222222222233
 * 01234567890123456789012345678901
 *
 * I Segment-Invalid Bit:    Segment is not available for address-translation
 * C Common-Segment Bit:     Segment is not private (PoP 3-30)
 * PTL Page-Table-Length:    Page-table length (PTL+1*16 entries -> up to 256)
 *
 * The 31 bit segmenttable origin of S390 has following format:
 *
 *  |S-table origin   |     | STL |
 * X                   **GPS
 * 00000000001111111111222222222233
 * 01234567890123456789012345678901
 *
 * X Space-Switch event:
 * G Segment-Invalid Bit:     *
 * P Private-Space Bit:       Segment is not private (PoP 3-30)
 * S Storage-Alteration:
 * STL Segment-Table-Length:  Segment-table length (STL+1*16 entries -> up to 2048)
 *
 * A 64 bit pagetable entry of S390 has following format:
167
 * |			 PFRA			      |0IPC|  OS  |
L
Linus Torvalds 已提交
168 169 170 171 172
 * 0000000000111111111122222222223333333333444444444455555555556666
 * 0123456789012345678901234567890123456789012345678901234567890123
 *
 * I Page-Invalid Bit:    Page is not available for address-translation
 * P Page-Protection Bit: Store access not possible for page
173
 * C Change-bit override: HW is not required to set change bit
L
Linus Torvalds 已提交
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
 *
 * A 64 bit segmenttable entry of S390 has following format:
 * |        P-table origin                              |      TT
 * 0000000000111111111122222222223333333333444444444455555555556666
 * 0123456789012345678901234567890123456789012345678901234567890123
 *
 * I Segment-Invalid Bit:    Segment is not available for address-translation
 * C Common-Segment Bit:     Segment is not private (PoP 3-30)
 * P Page-Protection Bit: Store access not possible for page
 * TT Type 00
 *
 * A 64 bit region table entry of S390 has following format:
 * |        S-table origin                             |   TF  TTTL
 * 0000000000111111111122222222223333333333444444444455555555556666
 * 0123456789012345678901234567890123456789012345678901234567890123
 *
 * I Segment-Invalid Bit:    Segment is not available for address-translation
 * TT Type 01
 * TF
M
Martin Schwidefsky 已提交
193
 * TL Table length
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
 *
 * The 64 bit regiontable origin of S390 has following format:
 * |      region table origon                          |       DTTL
 * 0000000000111111111122222222223333333333444444444455555555556666
 * 0123456789012345678901234567890123456789012345678901234567890123
 *
 * X Space-Switch event:
 * G Segment-Invalid Bit:  
 * P Private-Space Bit:    
 * S Storage-Alteration:
 * R Real space
 * TL Table-Length:
 *
 * A storage key has the following format:
 * | ACC |F|R|C|0|
 *  0   3 4 5 6 7
 * ACC: access key
 * F  : fetch protection bit
 * R  : referenced bit
 * C  : changed bit
 */

/* Hardware bits in the page table entry */
217
#define _PAGE_CO	0x100		/* HW Change-bit override */
M
Martin Schwidefsky 已提交
218 219
#define _PAGE_RO	0x200		/* HW read-only bit  */
#define _PAGE_INVALID	0x400		/* HW invalid bit    */
220 221

/* Software bits in the page table entry */
M
Martin Schwidefsky 已提交
222 223
#define _PAGE_SWT	0x001		/* SW pte type bit t */
#define _PAGE_SWX	0x002		/* SW pte type bit x */
224 225 226
#define _PAGE_SWC	0x004		/* SW pte changed bit (for KVM) */
#define _PAGE_SWR	0x008		/* SW pte referenced bit (for KVM) */
#define _PAGE_SPECIAL	0x010		/* SW associated with special page */
N
Nick Piggin 已提交
227
#define __HAVE_ARCH_PTE_SPECIAL
L
Linus Torvalds 已提交
228

229
/* Set of bits not changed in pte_modify */
230
#define _PAGE_CHG_MASK	(PAGE_MASK | _PAGE_SPECIAL | _PAGE_SWC | _PAGE_SWR)
231

M
Martin Schwidefsky 已提交
232
/* Six different types of pages. */
233 234
#define _PAGE_TYPE_EMPTY	0x400
#define _PAGE_TYPE_NONE		0x401
M
Martin Schwidefsky 已提交
235 236
#define _PAGE_TYPE_SWAP		0x403
#define _PAGE_TYPE_FILE		0x601	/* bit 0x002 is used for offset !! */
237 238
#define _PAGE_TYPE_RO		0x200
#define _PAGE_TYPE_RW		0x000
L
Linus Torvalds 已提交
239

240 241 242 243 244 245 246 247 248
/*
 * Only four types for huge pages, using the invalid bit and protection bit
 * of a segment table entry.
 */
#define _HPAGE_TYPE_EMPTY	0x020	/* _SEGMENT_ENTRY_INV */
#define _HPAGE_TYPE_NONE	0x220
#define _HPAGE_TYPE_RO		0x200	/* _SEGMENT_ENTRY_RO  */
#define _HPAGE_TYPE_RW		0x000

M
Martin Schwidefsky 已提交
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
/*
 * PTE type bits are rather complicated. handle_pte_fault uses pte_present,
 * pte_none and pte_file to find out the pte type WITHOUT holding the page
 * table lock. ptep_clear_flush on the other hand uses ptep_clear_flush to
 * invalidate a given pte. ipte sets the hw invalid bit and clears all tlbs
 * for the page. The page table entry is set to _PAGE_TYPE_EMPTY afterwards.
 * This change is done while holding the lock, but the intermediate step
 * of a previously valid pte with the hw invalid bit set can be observed by
 * handle_pte_fault. That makes it necessary that all valid pte types with
 * the hw invalid bit set must be distinguishable from the four pte types
 * empty, none, swap and file.
 *
 *			irxt  ipte  irxt
 * _PAGE_TYPE_EMPTY	1000   ->   1000
 * _PAGE_TYPE_NONE	1001   ->   1001
 * _PAGE_TYPE_SWAP	1011   ->   1011
 * _PAGE_TYPE_FILE	11?1   ->   11?1
 * _PAGE_TYPE_RO	0100   ->   1100
 * _PAGE_TYPE_RW	0000   ->   1000
 *
G
Gerald Schaefer 已提交
269
 * pte_none is true for bits combinations 1000, 1010, 1100, 1110
M
Martin Schwidefsky 已提交
270 271
 * pte_present is true for bits combinations 0000, 0010, 0100, 0110, 1001
 * pte_file is true for bits combinations 1101, 1111
G
Gerald Schaefer 已提交
272
 * swap pte is 1011 and 0001, 0011, 0101, 0111 are invalid.
M
Martin Schwidefsky 已提交
273 274
 */

275
#ifndef CONFIG_64BIT
L
Linus Torvalds 已提交
276

277 278 279 280 281 282
/* Bits in the segment table address-space-control-element */
#define _ASCE_SPACE_SWITCH	0x80000000UL	/* space switch event	    */
#define _ASCE_ORIGIN_MASK	0x7ffff000UL	/* segment table origin	    */
#define _ASCE_PRIVATE_SPACE	0x100	/* private space control	    */
#define _ASCE_ALT_EVENT		0x80	/* storage alteration event control */
#define _ASCE_TABLE_LENGTH	0x7f	/* 128 x 64 entries = 8k	    */
L
Linus Torvalds 已提交
283

284 285
/* Bits in the segment table entry */
#define _SEGMENT_ENTRY_ORIGIN	0x7fffffc0UL	/* page table origin	    */
286
#define _SEGMENT_ENTRY_RO	0x200	/* page protection bit		    */
287 288 289
#define _SEGMENT_ENTRY_INV	0x20	/* invalid segment table entry	    */
#define _SEGMENT_ENTRY_COMMON	0x10	/* common segment bit		    */
#define _SEGMENT_ENTRY_PTL	0x0f	/* page table length		    */
L
Linus Torvalds 已提交
290

291 292
#define _SEGMENT_ENTRY		(_SEGMENT_ENTRY_PTL)
#define _SEGMENT_ENTRY_EMPTY	(_SEGMENT_ENTRY_INV)
L
Linus Torvalds 已提交
293

294 295 296 297 298 299 300 301 302 303 304 305 306
/* Page status table bits for virtualization */
#define RCP_ACC_BITS	0xf0000000UL
#define RCP_FP_BIT	0x08000000UL
#define RCP_PCL_BIT	0x00800000UL
#define RCP_HR_BIT	0x00400000UL
#define RCP_HC_BIT	0x00200000UL
#define RCP_GR_BIT	0x00040000UL
#define RCP_GC_BIT	0x00020000UL

/* User dirty / referenced bit for KVM's migration feature */
#define KVM_UR_BIT	0x00008000UL
#define KVM_UC_BIT	0x00004000UL

307
#else /* CONFIG_64BIT */
L
Linus Torvalds 已提交
308

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
/* Bits in the segment/region table address-space-control-element */
#define _ASCE_ORIGIN		~0xfffUL/* segment table origin		    */
#define _ASCE_PRIVATE_SPACE	0x100	/* private space control	    */
#define _ASCE_ALT_EVENT		0x80	/* storage alteration event control */
#define _ASCE_SPACE_SWITCH	0x40	/* space switch event		    */
#define _ASCE_REAL_SPACE	0x20	/* real space control		    */
#define _ASCE_TYPE_MASK		0x0c	/* asce table type mask		    */
#define _ASCE_TYPE_REGION1	0x0c	/* region first table type	    */
#define _ASCE_TYPE_REGION2	0x08	/* region second table type	    */
#define _ASCE_TYPE_REGION3	0x04	/* region third table type	    */
#define _ASCE_TYPE_SEGMENT	0x00	/* segment table type		    */
#define _ASCE_TABLE_LENGTH	0x03	/* region table length		    */

/* Bits in the region table entry */
#define _REGION_ENTRY_ORIGIN	~0xfffUL/* region/segment table origin	    */
#define _REGION_ENTRY_INV	0x20	/* invalid region table entry	    */
#define _REGION_ENTRY_TYPE_MASK	0x0c	/* region/segment table type mask   */
#define _REGION_ENTRY_TYPE_R1	0x0c	/* region first table type	    */
#define _REGION_ENTRY_TYPE_R2	0x08	/* region second table type	    */
#define _REGION_ENTRY_TYPE_R3	0x04	/* region third table type	    */
#define _REGION_ENTRY_LENGTH	0x03	/* region third length		    */

#define _REGION1_ENTRY		(_REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_LENGTH)
#define _REGION1_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_INV)
#define _REGION2_ENTRY		(_REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_LENGTH)
#define _REGION2_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_INV)
#define _REGION3_ENTRY		(_REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_LENGTH)
#define _REGION3_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_INV)

338 339
#define _REGION3_ENTRY_LARGE	0x400	/* RTTE-format control, large page  */

L
Linus Torvalds 已提交
340
/* Bits in the segment table entry */
341 342 343
#define _SEGMENT_ENTRY_ORIGIN	~0x7ffUL/* segment table origin		    */
#define _SEGMENT_ENTRY_RO	0x200	/* page protection bit		    */
#define _SEGMENT_ENTRY_INV	0x20	/* invalid segment table entry	    */
L
Linus Torvalds 已提交
344

345 346 347
#define _SEGMENT_ENTRY		(0)
#define _SEGMENT_ENTRY_EMPTY	(_SEGMENT_ENTRY_INV)

348 349
#define _SEGMENT_ENTRY_LARGE	0x400	/* STE-format control, large page   */
#define _SEGMENT_ENTRY_CO	0x100	/* change-recording override   */
350 351
#define _SEGMENT_ENTRY_SPLIT_BIT 0	/* THP splitting bit number */
#define _SEGMENT_ENTRY_SPLIT	(1UL << _SEGMENT_ENTRY_SPLIT_BIT)
352

353 354 355 356
/* Set of bits not changed in pmd_modify */
#define _SEGMENT_CHG_MASK	(_SEGMENT_ENTRY_ORIGIN | _SEGMENT_ENTRY_LARGE \
				 | _SEGMENT_ENTRY_SPLIT | _SEGMENT_ENTRY_CO)

357 358 359 360 361 362 363 364 365 366 367 368 369
/* Page status table bits for virtualization */
#define RCP_ACC_BITS	0xf000000000000000UL
#define RCP_FP_BIT	0x0800000000000000UL
#define RCP_PCL_BIT	0x0080000000000000UL
#define RCP_HR_BIT	0x0040000000000000UL
#define RCP_HC_BIT	0x0020000000000000UL
#define RCP_GR_BIT	0x0004000000000000UL
#define RCP_GC_BIT	0x0002000000000000UL

/* User dirty / referenced bit for KVM's migration feature */
#define KVM_UR_BIT	0x0000800000000000UL
#define KVM_UC_BIT	0x0000400000000000UL

370
#endif /* CONFIG_64BIT */
L
Linus Torvalds 已提交
371 372

/*
373 374 375
 * A user page table pointer has the space-switch-event bit, the
 * private-space-control bit and the storage-alteration-event-control
 * bit set. A kernel page table pointer doesn't need them.
L
Linus Torvalds 已提交
376
 */
377 378
#define _ASCE_USER_BITS		(_ASCE_SPACE_SWITCH | _ASCE_PRIVATE_SPACE | \
				 _ASCE_ALT_EVENT)
L
Linus Torvalds 已提交
379 380

/*
381
 * Page protection definitions.
L
Linus Torvalds 已提交
382
 */
383 384 385 386 387 388
#define PAGE_NONE	__pgprot(_PAGE_TYPE_NONE)
#define PAGE_RO		__pgprot(_PAGE_TYPE_RO)
#define PAGE_RW		__pgprot(_PAGE_TYPE_RW)

#define PAGE_KERNEL	PAGE_RW
#define PAGE_COPY	PAGE_RO
L
Linus Torvalds 已提交
389 390

/*
391 392 393
 * On s390 the page table entry has an invalid bit and a read-only bit.
 * Read permission implies execute permission and write permission
 * implies read permission.
L
Linus Torvalds 已提交
394 395
 */
         /*xwr*/
396 397 398 399
#define __P000	PAGE_NONE
#define __P001	PAGE_RO
#define __P010	PAGE_RO
#define __P011	PAGE_RO
400 401 402 403
#define __P100	PAGE_RO
#define __P101	PAGE_RO
#define __P110	PAGE_RO
#define __P111	PAGE_RO
404 405 406 407 408

#define __S000	PAGE_NONE
#define __S001	PAGE_RO
#define __S010	PAGE_RW
#define __S011	PAGE_RW
409 410 411 412
#define __S100	PAGE_RO
#define __S101	PAGE_RO
#define __S110	PAGE_RW
#define __S111	PAGE_RW
L
Linus Torvalds 已提交
413

414
static inline int mm_exclusive(struct mm_struct *mm)
L
Linus Torvalds 已提交
415
{
416 417
	return likely(mm == current->active_mm &&
		      atomic_read(&mm->context.attach_count) <= 1);
L
Linus Torvalds 已提交
418 419
}

420 421 422 423 424 425 426 427
static inline int mm_has_pgste(struct mm_struct *mm)
{
#ifdef CONFIG_PGSTE
	if (unlikely(mm->context.has_pgste))
		return 1;
#endif
	return 0;
}
L
Linus Torvalds 已提交
428 429 430
/*
 * pgd/pmd/pte query functions
 */
431
#ifndef CONFIG_64BIT
L
Linus Torvalds 已提交
432

433 434 435
static inline int pgd_present(pgd_t pgd) { return 1; }
static inline int pgd_none(pgd_t pgd)    { return 0; }
static inline int pgd_bad(pgd_t pgd)     { return 0; }
L
Linus Torvalds 已提交
436

M
Martin Schwidefsky 已提交
437 438
static inline int pud_present(pud_t pud) { return 1; }
static inline int pud_none(pud_t pud)	 { return 0; }
439
static inline int pud_large(pud_t pud)	 { return 0; }
M
Martin Schwidefsky 已提交
440 441
static inline int pud_bad(pud_t pud)	 { return 0; }

442
#else /* CONFIG_64BIT */
L
Linus Torvalds 已提交
443

444 445
static inline int pgd_present(pgd_t pgd)
{
M
Martin Schwidefsky 已提交
446 447
	if ((pgd_val(pgd) & _REGION_ENTRY_TYPE_MASK) < _REGION_ENTRY_TYPE_R2)
		return 1;
448 449 450 451 452
	return (pgd_val(pgd) & _REGION_ENTRY_ORIGIN) != 0UL;
}

static inline int pgd_none(pgd_t pgd)
{
M
Martin Schwidefsky 已提交
453 454
	if ((pgd_val(pgd) & _REGION_ENTRY_TYPE_MASK) < _REGION_ENTRY_TYPE_R2)
		return 0;
455 456 457 458 459
	return (pgd_val(pgd) & _REGION_ENTRY_INV) != 0UL;
}

static inline int pgd_bad(pgd_t pgd)
{
M
Martin Schwidefsky 已提交
460 461 462 463 464
	/*
	 * With dynamic page table levels the pgd can be a region table
	 * entry or a segment table entry. Check for the bit that are
	 * invalid for either table entry.
	 */
465
	unsigned long mask =
M
Martin Schwidefsky 已提交
466
		~_SEGMENT_ENTRY_ORIGIN & ~_REGION_ENTRY_INV &
467 468 469
		~_REGION_ENTRY_TYPE_MASK & ~_REGION_ENTRY_LENGTH;
	return (pgd_val(pgd) & mask) != 0;
}
M
Martin Schwidefsky 已提交
470 471

static inline int pud_present(pud_t pud)
L
Linus Torvalds 已提交
472
{
M
Martin Schwidefsky 已提交
473 474
	if ((pud_val(pud) & _REGION_ENTRY_TYPE_MASK) < _REGION_ENTRY_TYPE_R3)
		return 1;
475
	return (pud_val(pud) & _REGION_ENTRY_ORIGIN) != 0UL;
L
Linus Torvalds 已提交
476 477
}

M
Martin Schwidefsky 已提交
478
static inline int pud_none(pud_t pud)
L
Linus Torvalds 已提交
479
{
M
Martin Schwidefsky 已提交
480 481
	if ((pud_val(pud) & _REGION_ENTRY_TYPE_MASK) < _REGION_ENTRY_TYPE_R3)
		return 0;
482
	return (pud_val(pud) & _REGION_ENTRY_INV) != 0UL;
L
Linus Torvalds 已提交
483 484
}

485 486 487 488 489 490 491
static inline int pud_large(pud_t pud)
{
	if ((pud_val(pud) & _REGION_ENTRY_TYPE_MASK) != _REGION_ENTRY_TYPE_R3)
		return 0;
	return !!(pud_val(pud) & _REGION3_ENTRY_LARGE);
}

M
Martin Schwidefsky 已提交
492
static inline int pud_bad(pud_t pud)
L
Linus Torvalds 已提交
493
{
M
Martin Schwidefsky 已提交
494 495 496 497 498
	/*
	 * With dynamic page table levels the pud can be a region table
	 * entry or a segment table entry. Check for the bit that are
	 * invalid for either table entry.
	 */
499
	unsigned long mask =
M
Martin Schwidefsky 已提交
500
		~_SEGMENT_ENTRY_ORIGIN & ~_REGION_ENTRY_INV &
501 502
		~_REGION_ENTRY_TYPE_MASK & ~_REGION_ENTRY_LENGTH;
	return (pud_val(pud) & mask) != 0;
L
Linus Torvalds 已提交
503 504
}

505
#endif /* CONFIG_64BIT */
506

507
static inline int pmd_present(pmd_t pmd)
L
Linus Torvalds 已提交
508
{
509 510 511
	unsigned long mask = _SEGMENT_ENTRY_INV | _SEGMENT_ENTRY_RO;
	return (pmd_val(pmd) & mask) == _HPAGE_TYPE_NONE ||
	       !(pmd_val(pmd) & _SEGMENT_ENTRY_INV);
L
Linus Torvalds 已提交
512 513
}

514
static inline int pmd_none(pmd_t pmd)
L
Linus Torvalds 已提交
515
{
516 517
	return (pmd_val(pmd) & _SEGMENT_ENTRY_INV) &&
	       !(pmd_val(pmd) & _SEGMENT_ENTRY_RO);
L
Linus Torvalds 已提交
518 519
}

520 521 522 523 524 525 526 527 528
static inline int pmd_large(pmd_t pmd)
{
#ifdef CONFIG_64BIT
	return !!(pmd_val(pmd) & _SEGMENT_ENTRY_LARGE);
#else
	return 0;
#endif
}

529
static inline int pmd_bad(pmd_t pmd)
L
Linus Torvalds 已提交
530
{
531 532
	unsigned long mask = ~_SEGMENT_ENTRY_ORIGIN & ~_SEGMENT_ENTRY_INV;
	return (pmd_val(pmd) & mask) != _SEGMENT_ENTRY;
L
Linus Torvalds 已提交
533 534
}

535 536 537 538
#define __HAVE_ARCH_PMDP_SPLITTING_FLUSH
extern void pmdp_splitting_flush(struct vm_area_struct *vma,
				 unsigned long addr, pmd_t *pmdp);

539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
#define  __HAVE_ARCH_PMDP_SET_ACCESS_FLAGS
extern int pmdp_set_access_flags(struct vm_area_struct *vma,
				 unsigned long address, pmd_t *pmdp,
				 pmd_t entry, int dirty);

#define __HAVE_ARCH_PMDP_CLEAR_YOUNG_FLUSH
extern int pmdp_clear_flush_young(struct vm_area_struct *vma,
				  unsigned long address, pmd_t *pmdp);

#define __HAVE_ARCH_PMD_WRITE
static inline int pmd_write(pmd_t pmd)
{
	return (pmd_val(pmd) & _SEGMENT_ENTRY_RO) == 0;
}

static inline int pmd_young(pmd_t pmd)
{
	return 0;
}

559
static inline int pte_none(pte_t pte)
L
Linus Torvalds 已提交
560
{
M
Martin Schwidefsky 已提交
561
	return (pte_val(pte) & _PAGE_INVALID) && !(pte_val(pte) & _PAGE_SWT);
L
Linus Torvalds 已提交
562 563
}

564
static inline int pte_present(pte_t pte)
L
Linus Torvalds 已提交
565
{
M
Martin Schwidefsky 已提交
566 567 568 569
	unsigned long mask = _PAGE_RO | _PAGE_INVALID | _PAGE_SWT | _PAGE_SWX;
	return (pte_val(pte) & mask) == _PAGE_TYPE_NONE ||
		(!(pte_val(pte) & _PAGE_INVALID) &&
		 !(pte_val(pte) & _PAGE_SWT));
L
Linus Torvalds 已提交
570 571
}

572
static inline int pte_file(pte_t pte)
L
Linus Torvalds 已提交
573
{
M
Martin Schwidefsky 已提交
574 575
	unsigned long mask = _PAGE_RO | _PAGE_INVALID | _PAGE_SWT;
	return (pte_val(pte) & mask) == _PAGE_TYPE_FILE;
L
Linus Torvalds 已提交
576 577
}

N
Nick Piggin 已提交
578 579
static inline int pte_special(pte_t pte)
{
N
Nick Piggin 已提交
580
	return (pte_val(pte) & _PAGE_SPECIAL);
N
Nick Piggin 已提交
581 582
}

M
Martin Schwidefsky 已提交
583
#define __HAVE_ARCH_PTE_SAME
584 585 586 587
static inline int pte_same(pte_t a, pte_t b)
{
	return pte_val(a) == pte_val(b);
}
L
Linus Torvalds 已提交
588

589
static inline pgste_t pgste_get_lock(pte_t *ptep)
590
{
591
	unsigned long new = 0;
592
#ifdef CONFIG_PGSTE
593 594
	unsigned long old;

595
	preempt_disable();
596 597 598 599 600 601 602 603 604
	asm(
		"	lg	%0,%2\n"
		"0:	lgr	%1,%0\n"
		"	nihh	%0,0xff7f\n"	/* clear RCP_PCL_BIT in old */
		"	oihh	%1,0x0080\n"	/* set RCP_PCL_BIT in new */
		"	csg	%0,%1,%2\n"
		"	jl	0b\n"
		: "=&d" (old), "=&d" (new), "=Q" (ptep[PTRS_PER_PTE])
		: "Q" (ptep[PTRS_PER_PTE]) : "cc");
605
#endif
606
	return __pgste(new);
607 608
}

609
static inline void pgste_set_unlock(pte_t *ptep, pgste_t pgste)
610 611
{
#ifdef CONFIG_PGSTE
612 613 614 615 616
	asm(
		"	nihh	%1,0xff7f\n"	/* clear RCP_PCL_BIT */
		"	stg	%1,%0\n"
		: "=Q" (ptep[PTRS_PER_PTE])
		: "d" (pgste_val(pgste)), "Q" (ptep[PTRS_PER_PTE]) : "cc");
617 618 619 620
	preempt_enable();
#endif
}

621
static inline pgste_t pgste_update_all(pte_t *ptep, pgste_t pgste)
622 623
{
#ifdef CONFIG_PGSTE
624
	unsigned long address, bits;
625 626
	unsigned char skey;

M
Martin Schwidefsky 已提交
627 628
	if (!pte_present(*ptep))
		return pgste;
629 630
	address = pte_val(*ptep) & PAGE_MASK;
	skey = page_get_storage_key(address);
631 632
	bits = skey & (_PAGE_CHANGED | _PAGE_REFERENCED);
	/* Clear page changed & referenced bit in the storage key */
633 634 635 636
	if (bits & _PAGE_CHANGED)
		page_set_storage_key(address, skey ^ bits, 1);
	else if (bits)
		page_reset_referenced(address);
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
	/* Transfer page changed & referenced bit to guest bits in pgste */
	pgste_val(pgste) |= bits << 48;		/* RCP_GR_BIT & RCP_GC_BIT */
	/* Get host changed & referenced bits from pgste */
	bits |= (pgste_val(pgste) & (RCP_HR_BIT | RCP_HC_BIT)) >> 52;
	/* Clear host bits in pgste. */
	pgste_val(pgste) &= ~(RCP_HR_BIT | RCP_HC_BIT);
	pgste_val(pgste) &= ~(RCP_ACC_BITS | RCP_FP_BIT);
	/* Copy page access key and fetch protection bit to pgste */
	pgste_val(pgste) |=
		(unsigned long) (skey & (_PAGE_ACC_BITS | _PAGE_FP_BIT)) << 56;
	/* Transfer changed and referenced to kvm user bits */
	pgste_val(pgste) |= bits << 45;		/* KVM_UR_BIT & KVM_UC_BIT */
	/* Transfer changed & referenced to pte sofware bits */
	pte_val(*ptep) |= bits << 1;		/* _PAGE_SWR & _PAGE_SWC */
#endif
	return pgste;

}

static inline pgste_t pgste_update_young(pte_t *ptep, pgste_t pgste)
{
#ifdef CONFIG_PGSTE
	int young;

M
Martin Schwidefsky 已提交
661 662
	if (!pte_present(*ptep))
		return pgste;
663 664 665 666 667 668 669 670 671 672 673 674 675
	young = page_reset_referenced(pte_val(*ptep) & PAGE_MASK);
	/* Transfer page referenced bit to pte software bit (host view) */
	if (young || (pgste_val(pgste) & RCP_HR_BIT))
		pte_val(*ptep) |= _PAGE_SWR;
	/* Clear host referenced bit in pgste. */
	pgste_val(pgste) &= ~RCP_HR_BIT;
	/* Transfer page referenced bit to guest bit in pgste */
	pgste_val(pgste) |= (unsigned long) young << 50; /* set RCP_GR_BIT */
#endif
	return pgste;

}

M
Martin Schwidefsky 已提交
676
static inline void pgste_set_pte(pte_t *ptep, pgste_t pgste, pte_t entry)
677 678
{
#ifdef CONFIG_PGSTE
679
	unsigned long address;
680 681
	unsigned long okey, nkey;

M
Martin Schwidefsky 已提交
682 683 684
	if (!pte_present(entry))
		return;
	address = pte_val(entry) & PAGE_MASK;
685
	okey = nkey = page_get_storage_key(address);
686 687 688 689
	nkey &= ~(_PAGE_ACC_BITS | _PAGE_FP_BIT);
	/* Set page access key and fetch protection bit from pgste */
	nkey |= (pgste_val(pgste) & (RCP_ACC_BITS | RCP_FP_BIT)) >> 56;
	if (okey != nkey)
690
		page_set_storage_key(address, nkey, 1);
691 692 693
#endif
}

694 695 696 697
/**
 * struct gmap_struct - guest address space
 * @mm: pointer to the parent mm_struct
 * @table: pointer to the page directory
698
 * @asce: address space control element for gmap page table
699 700 701 702 703 704
 * @crst_list: list of all crst tables used in the guest address space
 */
struct gmap {
	struct list_head list;
	struct mm_struct *mm;
	unsigned long *table;
705
	unsigned long asce;
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735
	struct list_head crst_list;
};

/**
 * struct gmap_rmap - reverse mapping for segment table entries
 * @next: pointer to the next gmap_rmap structure in the list
 * @entry: pointer to a segment table entry
 */
struct gmap_rmap {
	struct list_head list;
	unsigned long *entry;
};

/**
 * struct gmap_pgtable - gmap information attached to a page table
 * @vmaddr: address of the 1MB segment in the process virtual memory
 * @mapper: list of segment table entries maping a page table
 */
struct gmap_pgtable {
	unsigned long vmaddr;
	struct list_head mapper;
};

struct gmap *gmap_alloc(struct mm_struct *mm);
void gmap_free(struct gmap *gmap);
void gmap_enable(struct gmap *gmap);
void gmap_disable(struct gmap *gmap);
int gmap_map_segment(struct gmap *gmap, unsigned long from,
		     unsigned long to, unsigned long length);
int gmap_unmap_segment(struct gmap *gmap, unsigned long to, unsigned long len);
736
unsigned long __gmap_fault(unsigned long address, struct gmap *);
737
unsigned long gmap_fault(unsigned long address, struct gmap *);
738
void gmap_discard(unsigned long from, unsigned long to, struct gmap *);
739

740 741 742 743 744 745 746 747 748 749 750 751
/*
 * Certain architectures need to do special things when PTEs
 * within a page table are directly modified.  Thus, the following
 * hook is made available.
 */
static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
			      pte_t *ptep, pte_t entry)
{
	pgste_t pgste;

	if (mm_has_pgste(mm)) {
		pgste = pgste_get_lock(ptep);
M
Martin Schwidefsky 已提交
752
		pgste_set_pte(ptep, pgste, entry);
753 754 755 756 757 758
		*ptep = entry;
		pgste_set_unlock(ptep, pgste);
	} else
		*ptep = entry;
}

L
Linus Torvalds 已提交
759 760 761 762
/*
 * query functions pte_write/pte_dirty/pte_young only work if
 * pte_present() is true. Undefined behaviour if not..
 */
763
static inline int pte_write(pte_t pte)
L
Linus Torvalds 已提交
764 765 766 767
{
	return (pte_val(pte) & _PAGE_RO) == 0;
}

768
static inline int pte_dirty(pte_t pte)
L
Linus Torvalds 已提交
769
{
770 771 772 773
#ifdef CONFIG_PGSTE
	if (pte_val(pte) & _PAGE_SWC)
		return 1;
#endif
L
Linus Torvalds 已提交
774 775 776
	return 0;
}

777
static inline int pte_young(pte_t pte)
L
Linus Torvalds 已提交
778
{
779 780 781 782
#ifdef CONFIG_PGSTE
	if (pte_val(pte) & _PAGE_SWR)
		return 1;
#endif
L
Linus Torvalds 已提交
783 784 785 786 787 788 789
	return 0;
}

/*
 * pgd/pmd/pte modification functions
 */

790
static inline void pgd_clear(pgd_t *pgd)
791
{
792
#ifdef CONFIG_64BIT
M
Martin Schwidefsky 已提交
793 794
	if ((pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R2)
		pgd_val(*pgd) = _REGION2_ENTRY_EMPTY;
795
#endif
796 797
}

798
static inline void pud_clear(pud_t *pud)
L
Linus Torvalds 已提交
799
{
800
#ifdef CONFIG_64BIT
M
Martin Schwidefsky 已提交
801 802
	if ((pud_val(*pud) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3)
		pud_val(*pud) = _REGION3_ENTRY_EMPTY;
803
#endif
L
Linus Torvalds 已提交
804 805
}

806
static inline void pmd_clear(pmd_t *pmdp)
L
Linus Torvalds 已提交
807
{
808
	pmd_val(*pmdp) = _SEGMENT_ENTRY_EMPTY;
L
Linus Torvalds 已提交
809 810
}

811
static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
L
Linus Torvalds 已提交
812
{
813
	pte_val(*ptep) = _PAGE_TYPE_EMPTY;
L
Linus Torvalds 已提交
814 815 816 817 818 819
}

/*
 * The following pte modification functions only work if
 * pte_present() is true. Undefined behaviour if not..
 */
820
static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
L
Linus Torvalds 已提交
821
{
822
	pte_val(pte) &= _PAGE_CHG_MASK;
L
Linus Torvalds 已提交
823 824 825 826
	pte_val(pte) |= pgprot_val(newprot);
	return pte;
}

827
static inline pte_t pte_wrprotect(pte_t pte)
L
Linus Torvalds 已提交
828
{
829
	/* Do not clobber _PAGE_TYPE_NONE pages!  */
L
Linus Torvalds 已提交
830 831 832 833 834
	if (!(pte_val(pte) & _PAGE_INVALID))
		pte_val(pte) |= _PAGE_RO;
	return pte;
}

835
static inline pte_t pte_mkwrite(pte_t pte)
L
Linus Torvalds 已提交
836 837 838 839 840
{
	pte_val(pte) &= ~_PAGE_RO;
	return pte;
}

841
static inline pte_t pte_mkclean(pte_t pte)
L
Linus Torvalds 已提交
842
{
843 844 845
#ifdef CONFIG_PGSTE
	pte_val(pte) &= ~_PAGE_SWC;
#endif
L
Linus Torvalds 已提交
846 847 848
	return pte;
}

849
static inline pte_t pte_mkdirty(pte_t pte)
L
Linus Torvalds 已提交
850 851 852 853
{
	return pte;
}

854
static inline pte_t pte_mkold(pte_t pte)
L
Linus Torvalds 已提交
855
{
856 857 858
#ifdef CONFIG_PGSTE
	pte_val(pte) &= ~_PAGE_SWR;
#endif
L
Linus Torvalds 已提交
859 860 861
	return pte;
}

862
static inline pte_t pte_mkyoung(pte_t pte)
L
Linus Torvalds 已提交
863 864 865 866
{
	return pte;
}

N
Nick Piggin 已提交
867 868
static inline pte_t pte_mkspecial(pte_t pte)
{
N
Nick Piggin 已提交
869
	pte_val(pte) |= _PAGE_SPECIAL;
N
Nick Piggin 已提交
870 871 872
	return pte;
}

873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
#ifdef CONFIG_HUGETLB_PAGE
static inline pte_t pte_mkhuge(pte_t pte)
{
	/*
	 * PROT_NONE needs to be remapped from the pte type to the ste type.
	 * The HW invalid bit is also different for pte and ste. The pte
	 * invalid bit happens to be the same as the ste _SEGMENT_ENTRY_LARGE
	 * bit, so we don't have to clear it.
	 */
	if (pte_val(pte) & _PAGE_INVALID) {
		if (pte_val(pte) & _PAGE_SWT)
			pte_val(pte) |= _HPAGE_TYPE_NONE;
		pte_val(pte) |= _SEGMENT_ENTRY_INV;
	}
	/*
	 * Clear SW pte bits SWT and SWX, there are no SW bits in a segment
	 * table entry.
	 */
	pte_val(pte) &= ~(_PAGE_SWT | _PAGE_SWX);
	/*
	 * Also set the change-override bit because we don't need dirty bit
	 * tracking for hugetlbfs pages.
	 */
	pte_val(pte) |= (_SEGMENT_ENTRY_LARGE | _SEGMENT_ENTRY_CO);
	return pte;
}
#endif

901
/*
902
 * Get (and clear) the user dirty bit for a pte.
903
 */
904 905
static inline int ptep_test_and_clear_user_dirty(struct mm_struct *mm,
						 pte_t *ptep)
906
{
907 908 909 910 911 912 913 914 915 916
	pgste_t pgste;
	int dirty = 0;

	if (mm_has_pgste(mm)) {
		pgste = pgste_get_lock(ptep);
		pgste = pgste_update_all(ptep, pgste);
		dirty = !!(pgste_val(pgste) & KVM_UC_BIT);
		pgste_val(pgste) &= ~KVM_UC_BIT;
		pgste_set_unlock(ptep, pgste);
		return dirty;
917 918 919
	}
	return dirty;
}
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938

/*
 * Get (and clear) the user referenced bit for a pte.
 */
static inline int ptep_test_and_clear_user_young(struct mm_struct *mm,
						 pte_t *ptep)
{
	pgste_t pgste;
	int young = 0;

	if (mm_has_pgste(mm)) {
		pgste = pgste_get_lock(ptep);
		pgste = pgste_update_young(ptep, pgste);
		young = !!(pgste_val(pgste) & KVM_UR_BIT);
		pgste_val(pgste) &= ~KVM_UR_BIT;
		pgste_set_unlock(ptep, pgste);
	}
	return young;
}
939

M
Martin Schwidefsky 已提交
940 941 942
#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
					    unsigned long addr, pte_t *ptep)
L
Linus Torvalds 已提交
943
{
944 945 946 947 948 949 950 951 952 953 954
	pgste_t pgste;
	pte_t pte;

	if (mm_has_pgste(vma->vm_mm)) {
		pgste = pgste_get_lock(ptep);
		pgste = pgste_update_young(ptep, pgste);
		pte = *ptep;
		*ptep = pte_mkold(pte);
		pgste_set_unlock(ptep, pgste);
		return pte_young(pte);
	}
L
Linus Torvalds 已提交
955 956 957
	return 0;
}

M
Martin Schwidefsky 已提交
958 959 960
#define __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH
static inline int ptep_clear_flush_young(struct vm_area_struct *vma,
					 unsigned long address, pte_t *ptep)
L
Linus Torvalds 已提交
961
{
962 963 964 965 966
	/* No need to flush TLB
	 * On s390 reference bits are in storage key and never in TLB
	 * With virtualization we handle the reference bit, without we
	 * we can simply return */
	return ptep_test_and_clear_young(vma, address, ptep);
L
Linus Torvalds 已提交
967 968
}

969
static inline void __ptep_ipte(unsigned long address, pte_t *ptep)
L
Linus Torvalds 已提交
970
{
971
	if (!(pte_val(*ptep) & _PAGE_INVALID)) {
972
#ifndef CONFIG_64BIT
973
		/* pto must point to the start of the segment table */
L
Linus Torvalds 已提交
974
		pte_t *pto = (pte_t *) (((unsigned long) ptep) & 0x7ffffc00);
975 976 977 978
#else
		/* ipte in zarch mode can do the math */
		pte_t *pto = ptep;
#endif
979 980 981 982
		asm volatile(
			"	ipte	%2,%3"
			: "=m" (*ptep) : "m" (*ptep),
			  "a" (pto), "a" (address));
L
Linus Torvalds 已提交
983
	}
984 985
}

M
Martin Schwidefsky 已提交
986 987 988 989 990 991 992 993 994 995 996 997 998 999
/*
 * This is hard to understand. ptep_get_and_clear and ptep_clear_flush
 * both clear the TLB for the unmapped pte. The reason is that
 * ptep_get_and_clear is used in common code (e.g. change_pte_range)
 * to modify an active pte. The sequence is
 *   1) ptep_get_and_clear
 *   2) set_pte_at
 *   3) flush_tlb_range
 * On s390 the tlb needs to get flushed with the modification of the pte
 * if the pte is active. The only way how this can be implemented is to
 * have ptep_get_and_clear do the tlb flush. In exchange flush_tlb_range
 * is a nop.
 */
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
static inline pte_t ptep_get_and_clear(struct mm_struct *mm,
				       unsigned long address, pte_t *ptep)
{
	pgste_t pgste;
	pte_t pte;

	mm->context.flush_mm = 1;
	if (mm_has_pgste(mm))
		pgste = pgste_get_lock(ptep);

	pte = *ptep;
	if (!mm_exclusive(mm))
		__ptep_ipte(address, ptep);
	pte_val(*ptep) = _PAGE_TYPE_EMPTY;

	if (mm_has_pgste(mm)) {
		pgste = pgste_update_all(&pte, pgste);
		pgste_set_unlock(ptep, pgste);
	}
	return pte;
}

#define __HAVE_ARCH_PTEP_MODIFY_PROT_TRANSACTION
static inline pte_t ptep_modify_prot_start(struct mm_struct *mm,
					   unsigned long address,
					   pte_t *ptep)
{
	pte_t pte;

	mm->context.flush_mm = 1;
	if (mm_has_pgste(mm))
		pgste_get_lock(ptep);

	pte = *ptep;
	if (!mm_exclusive(mm))
		__ptep_ipte(address, ptep);
	return pte;
}

static inline void ptep_modify_prot_commit(struct mm_struct *mm,
					   unsigned long address,
					   pte_t *ptep, pte_t pte)
{
	*ptep = pte;
	if (mm_has_pgste(mm))
		pgste_set_unlock(ptep, *(pgste_t *)(ptep + PTRS_PER_PTE));
}
M
Martin Schwidefsky 已提交
1047 1048

#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
M
Martin Schwidefsky 已提交
1049 1050 1051
static inline pte_t ptep_clear_flush(struct vm_area_struct *vma,
				     unsigned long address, pte_t *ptep)
{
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
	pgste_t pgste;
	pte_t pte;

	if (mm_has_pgste(vma->vm_mm))
		pgste = pgste_get_lock(ptep);

	pte = *ptep;
	__ptep_ipte(address, ptep);
	pte_val(*ptep) = _PAGE_TYPE_EMPTY;

	if (mm_has_pgste(vma->vm_mm)) {
		pgste = pgste_update_all(&pte, pgste);
		pgste_set_unlock(ptep, pgste);
	}
L
Linus Torvalds 已提交
1066 1067 1068
	return pte;
}

M
Martin Schwidefsky 已提交
1069 1070 1071 1072 1073 1074 1075 1076 1077
/*
 * The batched pte unmap code uses ptep_get_and_clear_full to clear the
 * ptes. Here an optimization is possible. tlb_gather_mmu flushes all
 * tlbs of an mm if it can guarantee that the ptes of the mm_struct
 * cannot be accessed while the batched unmap is running. In this case
 * full==1 and a simple pte_clear is enough. See tlb.h.
 */
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm,
1078
					    unsigned long address,
M
Martin Schwidefsky 已提交
1079
					    pte_t *ptep, int full)
L
Linus Torvalds 已提交
1080
{
1081 1082 1083 1084 1085
	pgste_t pgste;
	pte_t pte;

	if (mm_has_pgste(mm))
		pgste = pgste_get_lock(ptep);
M
Martin Schwidefsky 已提交
1086

1087 1088 1089 1090 1091 1092 1093 1094 1095
	pte = *ptep;
	if (!full)
		__ptep_ipte(address, ptep);
	pte_val(*ptep) = _PAGE_TYPE_EMPTY;

	if (mm_has_pgste(mm)) {
		pgste = pgste_update_all(&pte, pgste);
		pgste_set_unlock(ptep, pgste);
	}
M
Martin Schwidefsky 已提交
1096
	return pte;
L
Linus Torvalds 已提交
1097 1098
}

M
Martin Schwidefsky 已提交
1099
#define __HAVE_ARCH_PTEP_SET_WRPROTECT
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
static inline pte_t ptep_set_wrprotect(struct mm_struct *mm,
				       unsigned long address, pte_t *ptep)
{
	pgste_t pgste;
	pte_t pte = *ptep;

	if (pte_write(pte)) {
		mm->context.flush_mm = 1;
		if (mm_has_pgste(mm))
			pgste = pgste_get_lock(ptep);

		if (!mm_exclusive(mm))
			__ptep_ipte(address, ptep);
		*ptep = pte_wrprotect(pte);

		if (mm_has_pgste(mm))
			pgste_set_unlock(ptep, pgste);
	}
	return pte;
}
M
Martin Schwidefsky 已提交
1120 1121

#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
static inline int ptep_set_access_flags(struct vm_area_struct *vma,
					unsigned long address, pte_t *ptep,
					pte_t entry, int dirty)
{
	pgste_t pgste;

	if (pte_same(*ptep, entry))
		return 0;
	if (mm_has_pgste(vma->vm_mm))
		pgste = pgste_get_lock(ptep);

	__ptep_ipte(address, ptep);
	*ptep = entry;

	if (mm_has_pgste(vma->vm_mm))
		pgste_set_unlock(ptep, pgste);
	return 1;
}
L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151

/*
 * Conversion functions: convert a page and protection to a page entry,
 * and a page entry and page directory to the page they refer to.
 */
static inline pte_t mk_pte_phys(unsigned long physpage, pgprot_t pgprot)
{
	pte_t __pte;
	pte_val(__pte) = physpage + pgprot_val(pgprot);
	return __pte;
}

1152 1153
static inline pte_t mk_pte(struct page *page, pgprot_t pgprot)
{
1154
	unsigned long physpage = page_to_phys(page);
L
Linus Torvalds 已提交
1155

1156 1157 1158
	return mk_pte_phys(physpage, pgprot);
}

M
Martin Schwidefsky 已提交
1159 1160 1161 1162
#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
#define pud_index(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
#define pte_index(address) (((address) >> PAGE_SHIFT) & (PTRS_PER_PTE-1))
L
Linus Torvalds 已提交
1163

M
Martin Schwidefsky 已提交
1164 1165
#define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address))
#define pgd_offset_k(address) pgd_offset(&init_mm, address)
L
Linus Torvalds 已提交
1166

1167
#ifndef CONFIG_64BIT
L
Linus Torvalds 已提交
1168

M
Martin Schwidefsky 已提交
1169 1170 1171
#define pmd_deref(pmd) (pmd_val(pmd) & _SEGMENT_ENTRY_ORIGIN)
#define pud_deref(pmd) ({ BUG(); 0UL; })
#define pgd_deref(pmd) ({ BUG(); 0UL; })
1172

M
Martin Schwidefsky 已提交
1173 1174
#define pud_offset(pgd, address) ((pud_t *) pgd)
#define pmd_offset(pud, address) ((pmd_t *) pud + pmd_index(address))
L
Linus Torvalds 已提交
1175

1176
#else /* CONFIG_64BIT */
L
Linus Torvalds 已提交
1177

M
Martin Schwidefsky 已提交
1178 1179
#define pmd_deref(pmd) (pmd_val(pmd) & _SEGMENT_ENTRY_ORIGIN)
#define pud_deref(pud) (pud_val(pud) & _REGION_ENTRY_ORIGIN)
1180
#define pgd_deref(pgd) (pgd_val(pgd) & _REGION_ENTRY_ORIGIN)
L
Linus Torvalds 已提交
1181

1182 1183
static inline pud_t *pud_offset(pgd_t *pgd, unsigned long address)
{
M
Martin Schwidefsky 已提交
1184 1185 1186
	pud_t *pud = (pud_t *) pgd;
	if ((pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R2)
		pud = (pud_t *) pgd_deref(*pgd);
1187 1188
	return pud  + pud_index(address);
}
L
Linus Torvalds 已提交
1189

M
Martin Schwidefsky 已提交
1190
static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address)
L
Linus Torvalds 已提交
1191
{
M
Martin Schwidefsky 已提交
1192 1193 1194
	pmd_t *pmd = (pmd_t *) pud;
	if ((pud_val(*pud) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3)
		pmd = (pmd_t *) pud_deref(*pud);
M
Martin Schwidefsky 已提交
1195
	return pmd + pmd_index(address);
L
Linus Torvalds 已提交
1196 1197
}

1198
#endif /* CONFIG_64BIT */
L
Linus Torvalds 已提交
1199

M
Martin Schwidefsky 已提交
1200 1201 1202
#define pfn_pte(pfn,pgprot) mk_pte_phys(__pa((pfn) << PAGE_SHIFT),(pgprot))
#define pte_pfn(x) (pte_val(x) >> PAGE_SHIFT)
#define pte_page(x) pfn_to_page(pte_pfn(x))
L
Linus Torvalds 已提交
1203

M
Martin Schwidefsky 已提交
1204
#define pmd_page(pmd) pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT)
L
Linus Torvalds 已提交
1205

M
Martin Schwidefsky 已提交
1206 1207 1208
/* Find an entry in the lowest level page table.. */
#define pte_offset(pmd, addr) ((pte_t *) pmd_deref(*(pmd)) + pte_index(addr))
#define pte_offset_kernel(pmd, address) pte_offset(pmd,address)
L
Linus Torvalds 已提交
1209 1210 1211
#define pte_offset_map(pmd, address) pte_offset_kernel(pmd, address)
#define pte_unmap(pte) do { } while (0)

1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227
static inline void __pmd_idte(unsigned long address, pmd_t *pmdp)
{
	unsigned long sto = (unsigned long) pmdp -
			    pmd_index(address) * sizeof(pmd_t);

	if (!(pmd_val(*pmdp) & _SEGMENT_ENTRY_INV)) {
		asm volatile(
			"	.insn	rrf,0xb98e0000,%2,%3,0,0"
			: "=m" (*pmdp)
			: "m" (*pmdp), "a" (sto),
			  "a" ((address & HPAGE_MASK))
			: "cc"
		);
	}
}

1228
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
1229 1230 1231 1232 1233

#define SEGMENT_NONE	__pgprot(_HPAGE_TYPE_NONE)
#define SEGMENT_RO	__pgprot(_HPAGE_TYPE_RO)
#define SEGMENT_RW	__pgprot(_HPAGE_TYPE_RW)

1234 1235 1236 1237 1238 1239
#define __HAVE_ARCH_PGTABLE_DEPOSIT
extern void pgtable_trans_huge_deposit(struct mm_struct *mm, pgtable_t pgtable);

#define __HAVE_ARCH_PGTABLE_WITHDRAW
extern pgtable_t pgtable_trans_huge_withdraw(struct mm_struct *mm);

1240 1241 1242 1243
static inline int pmd_trans_splitting(pmd_t pmd)
{
	return pmd_val(pmd) & _SEGMENT_ENTRY_SPLIT;
}
1244 1245 1246 1247 1248 1249 1250 1251 1252

static inline void set_pmd_at(struct mm_struct *mm, unsigned long addr,
			      pmd_t *pmdp, pmd_t entry)
{
	*pmdp = entry;
}

static inline unsigned long massage_pgprot_pmd(pgprot_t pgprot)
{
1253 1254 1255 1256 1257 1258 1259 1260 1261
	/*
	 * pgprot is PAGE_NONE, PAGE_RO, or PAGE_RW (see __Pxxx / __Sxxx)
	 * Convert to segment table entry format.
	 */
	if (pgprot_val(pgprot) == pgprot_val(PAGE_NONE))
		return pgprot_val(SEGMENT_NONE);
	if (pgprot_val(pgprot) == pgprot_val(PAGE_RO))
		return pgprot_val(SEGMENT_RO);
	return pgprot_val(SEGMENT_RW);
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278
}

static inline pmd_t pmd_modify(pmd_t pmd, pgprot_t newprot)
{
	pmd_val(pmd) &= _SEGMENT_CHG_MASK;
	pmd_val(pmd) |= massage_pgprot_pmd(newprot);
	return pmd;
}

static inline pmd_t pmd_mkhuge(pmd_t pmd)
{
	pmd_val(pmd) |= _SEGMENT_ENTRY_LARGE;
	return pmd;
}

static inline pmd_t pmd_mkwrite(pmd_t pmd)
{
1279 1280 1281
	/* Do not clobber _HPAGE_TYPE_NONE pages! */
	if (!(pmd_val(pmd) & _SEGMENT_ENTRY_INV))
		pmd_val(pmd) &= ~_SEGMENT_ENTRY_RO;
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
	return pmd;
}

static inline pmd_t pmd_wrprotect(pmd_t pmd)
{
	pmd_val(pmd) |= _SEGMENT_ENTRY_RO;
	return pmd;
}

static inline pmd_t pmd_mkdirty(pmd_t pmd)
{
	/* No dirty bit in the segment table entry. */
	return pmd;
}

static inline pmd_t pmd_mkold(pmd_t pmd)
{
	/* No referenced bit in the segment table entry. */
	return pmd;
}

static inline pmd_t pmd_mkyoung(pmd_t pmd)
{
	/* No referenced bit in the segment table entry. */
	return pmd;
}

#define __HAVE_ARCH_PMDP_TEST_AND_CLEAR_YOUNG
static inline int pmdp_test_and_clear_young(struct vm_area_struct *vma,
					    unsigned long address, pmd_t *pmdp)
{
	unsigned long pmd_addr = pmd_val(*pmdp) & HPAGE_MASK;
	long tmp, rc;
	int counter;

	rc = 0;
	if (MACHINE_HAS_RRBM) {
		counter = PTRS_PER_PTE >> 6;
		asm volatile(
			"0:	.insn	rre,0xb9ae0000,%0,%3\n"	/* rrbm */
			"	ogr	%1,%0\n"
			"	la	%3,0(%4,%3)\n"
			"	brct	%2,0b\n"
			: "=&d" (tmp), "+&d" (rc), "+d" (counter),
			  "+a" (pmd_addr)
			: "a" (64 * 4096UL) : "cc");
		rc = !!rc;
	} else {
		counter = PTRS_PER_PTE;
		asm volatile(
			"0:	rrbe	0,%2\n"
			"	la	%2,0(%3,%2)\n"
			"	brc	12,1f\n"
			"	lhi	%0,1\n"
			"1:	brct	%1,0b\n"
			: "+d" (rc), "+d" (counter), "+a" (pmd_addr)
			: "a" (4096UL) : "cc");
	}
	return rc;
}

#define __HAVE_ARCH_PMDP_GET_AND_CLEAR
static inline pmd_t pmdp_get_and_clear(struct mm_struct *mm,
				       unsigned long address, pmd_t *pmdp)
{
	pmd_t pmd = *pmdp;

	__pmd_idte(address, pmdp);
	pmd_clear(pmdp);
	return pmd;
}

#define __HAVE_ARCH_PMDP_CLEAR_FLUSH
static inline pmd_t pmdp_clear_flush(struct vm_area_struct *vma,
				     unsigned long address, pmd_t *pmdp)
{
	return pmdp_get_and_clear(vma->vm_mm, address, pmdp);
}

#define __HAVE_ARCH_PMDP_INVALIDATE
static inline void pmdp_invalidate(struct vm_area_struct *vma,
				   unsigned long address, pmd_t *pmdp)
{
	__pmd_idte(address, pmdp);
}

static inline pmd_t mk_pmd_phys(unsigned long physpage, pgprot_t pgprot)
{
	pmd_t __pmd;
	pmd_val(__pmd) = physpage + massage_pgprot_pmd(pgprot);
	return __pmd;
}

#define pfn_pmd(pfn, pgprot)	mk_pmd_phys(__pa((pfn) << PAGE_SHIFT), (pgprot))
#define mk_pmd(page, pgprot)	pfn_pmd(page_to_pfn(page), (pgprot))

static inline int pmd_trans_huge(pmd_t pmd)
{
	return pmd_val(pmd) & _SEGMENT_ENTRY_LARGE;
}

static inline int has_transparent_hugepage(void)
{
	return MACHINE_HAS_HPAGE ? 1 : 0;
}

static inline unsigned long pmd_pfn(pmd_t pmd)
{
G
Gerald Schaefer 已提交
1390
	return pmd_val(pmd) >> PAGE_SHIFT;
1391
}
1392 1393
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */

L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428
/*
 * 31 bit swap entry format:
 * A page-table entry has some bits we have to treat in a special way.
 * Bits 0, 20 and bit 23 have to be zero, otherwise an specification
 * exception will occur instead of a page translation exception. The
 * specifiation exception has the bad habit not to store necessary
 * information in the lowcore.
 * Bit 21 and bit 22 are the page invalid bit and the page protection
 * bit. We set both to indicate a swapped page.
 * Bit 30 and 31 are used to distinguish the different page types. For
 * a swapped page these bits need to be zero.
 * This leaves the bits 1-19 and bits 24-29 to store type and offset.
 * We use the 5 bits from 25-29 for the type and the 20 bits from 1-19
 * plus 24 for the offset.
 * 0|     offset        |0110|o|type |00|
 * 0 0000000001111111111 2222 2 22222 33
 * 0 1234567890123456789 0123 4 56789 01
 *
 * 64 bit swap entry format:
 * A page-table entry has some bits we have to treat in a special way.
 * Bits 52 and bit 55 have to be zero, otherwise an specification
 * exception will occur instead of a page translation exception. The
 * specifiation exception has the bad habit not to store necessary
 * information in the lowcore.
 * Bit 53 and bit 54 are the page invalid bit and the page protection
 * bit. We set both to indicate a swapped page.
 * Bit 62 and 63 are used to distinguish the different page types. For
 * a swapped page these bits need to be zero.
 * This leaves the bits 0-51 and bits 56-61 to store type and offset.
 * We use the 5 bits from 57-61 for the type and the 53 bits from 0-51
 * plus 56 for the offset.
 * |                      offset                        |0110|o|type |00|
 *  0000000000111111111122222222223333333333444444444455 5555 5 55566 66
 *  0123456789012345678901234567890123456789012345678901 2345 6 78901 23
 */
1429
#ifndef CONFIG_64BIT
L
Linus Torvalds 已提交
1430 1431 1432 1433
#define __SWP_OFFSET_MASK (~0UL >> 12)
#else
#define __SWP_OFFSET_MASK (~0UL >> 11)
#endif
1434
static inline pte_t mk_swap_pte(unsigned long type, unsigned long offset)
L
Linus Torvalds 已提交
1435 1436 1437
{
	pte_t pte;
	offset &= __SWP_OFFSET_MASK;
1438
	pte_val(pte) = _PAGE_TYPE_SWAP | ((type & 0x1f) << 2) |
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449
		((offset & 1UL) << 7) | ((offset & ~1UL) << 11);
	return pte;
}

#define __swp_type(entry)	(((entry).val >> 2) & 0x1f)
#define __swp_offset(entry)	(((entry).val >> 11) | (((entry).val >> 7) & 1))
#define __swp_entry(type,offset) ((swp_entry_t) { pte_val(mk_swap_pte((type),(offset))) })

#define __pte_to_swp_entry(pte)	((swp_entry_t) { pte_val(pte) })
#define __swp_entry_to_pte(x)	((pte_t) { (x).val })

1450
#ifndef CONFIG_64BIT
L
Linus Torvalds 已提交
1451
# define PTE_FILE_MAX_BITS	26
1452
#else /* CONFIG_64BIT */
L
Linus Torvalds 已提交
1453
# define PTE_FILE_MAX_BITS	59
1454
#endif /* CONFIG_64BIT */
L
Linus Torvalds 已提交
1455 1456 1457 1458 1459 1460

#define pte_to_pgoff(__pte) \
	((((__pte).pte >> 12) << 7) + (((__pte).pte >> 1) & 0x7f))

#define pgoff_to_pte(__off) \
	((pte_t) { ((((__off) & 0x7f) << 1) + (((__off) >> 7) << 12)) \
1461
		   | _PAGE_TYPE_FILE })
L
Linus Torvalds 已提交
1462 1463 1464 1465 1466

#endif /* !__ASSEMBLY__ */

#define kern_addr_valid(addr)   (1)

1467 1468
extern int vmem_add_mapping(unsigned long start, unsigned long size);
extern int vmem_remove_mapping(unsigned long start, unsigned long size);
1469
extern int s390_enable_sie(void);
H
Heiko Carstens 已提交
1470

L
Linus Torvalds 已提交
1471 1472 1473 1474 1475 1476 1477 1478
/*
 * No page table caches to initialise
 */
#define pgtable_cache_init()	do { } while (0)

#include <asm-generic/pgtable.h>

#endif /* _S390_PAGE_H */