vmem.c 13.1 KB
Newer Older
1 2 3
// SPDX-License-Identifier: GPL-2.0
#include <linux/sched/task.h>
#include <linux/pgtable.h>
4
#include <linux/kasan.h>
5 6 7
#include <asm/pgalloc.h>
#include <asm/facility.h>
#include <asm/sections.h>
8
#include <asm/physmem_info.h>
9
#include <asm/maccess.h>
10
#include <asm/abs_lowcore.h>
11 12 13
#include "decompressor.h"
#include "boot.h"

14 15
unsigned long __bootdata_preserved(s390_invalid_asce);

16 17 18 19
#ifdef CONFIG_PROC_FS
atomic_long_t __bootdata_preserved(direct_pages_count[PG_DIRECT_MAP_MAX]);
#endif

20 21 22 23
#define init_mm			(*(struct mm_struct *)vmlinux.init_mm_off)
#define swapper_pg_dir		vmlinux.swapper_pg_dir_off
#define invalid_pg_dir		vmlinux.invalid_pg_dir_off

24 25
enum populate_mode {
	POPULATE_NONE,
26
	POPULATE_DIRECT,
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
	POPULATE_ABS_LOWCORE,
#ifdef CONFIG_KASAN
	POPULATE_KASAN_MAP_SHADOW,
	POPULATE_KASAN_ZERO_SHADOW,
	POPULATE_KASAN_SHALLOW
#endif
};

static void pgtable_populate(unsigned long addr, unsigned long end, enum populate_mode mode);

#ifdef CONFIG_KASAN

#define kasan_early_shadow_page	vmlinux.kasan_early_shadow_page_off
#define kasan_early_shadow_pte	((pte_t *)vmlinux.kasan_early_shadow_pte_off)
#define kasan_early_shadow_pmd	((pmd_t *)vmlinux.kasan_early_shadow_pmd_off)
#define kasan_early_shadow_pud	((pud_t *)vmlinux.kasan_early_shadow_pud_off)
#define kasan_early_shadow_p4d	((p4d_t *)vmlinux.kasan_early_shadow_p4d_off)
#define __sha(x)		((unsigned long)kasan_mem_to_shadow((void *)x))

static pte_t pte_z;

48 49 50 51 52 53 54
static inline void kasan_populate(unsigned long start, unsigned long end, enum populate_mode mode)
{
	start = PAGE_ALIGN_DOWN(__sha(start));
	end = PAGE_ALIGN(__sha(end));
	pgtable_populate(start, end, mode);
}

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
static void kasan_populate_shadow(void)
{
	pmd_t pmd_z = __pmd(__pa(kasan_early_shadow_pte) | _SEGMENT_ENTRY);
	pud_t pud_z = __pud(__pa(kasan_early_shadow_pmd) | _REGION3_ENTRY);
	p4d_t p4d_z = __p4d(__pa(kasan_early_shadow_pud) | _REGION2_ENTRY);
	unsigned long untracked_end;
	unsigned long start, end;
	int i;

	pte_z = __pte(__pa(kasan_early_shadow_page) | pgprot_val(PAGE_KERNEL_RO));
	if (!machine.has_nx)
		pte_z = clear_pte_bit(pte_z, __pgprot(_PAGE_NOEXEC));
	crst_table_init((unsigned long *)kasan_early_shadow_p4d, p4d_val(p4d_z));
	crst_table_init((unsigned long *)kasan_early_shadow_pud, pud_val(pud_z));
	crst_table_init((unsigned long *)kasan_early_shadow_pmd, pmd_val(pmd_z));
	memset64((u64 *)kasan_early_shadow_pte, pte_val(pte_z), PTRS_PER_PTE);

	/*
	 * Current memory layout:
	 * +- 0 -------------+	       +- shadow start -+
	 * |1:1 ident mapping|	      /|1/8 of ident map|
	 * |		     |	     / |		|
	 * +-end of ident map+	    /  +----------------+
	 * | ... gap ...     |	   /   |    kasan	|
	 * |		     |	  /    |  zero page	|
	 * +- vmalloc area  -+	 /     |   mapping	|
	 * | vmalloc_size    |	/      | (untracked)	|
	 * +- modules vaddr -+ /       +----------------+
	 * | 2Gb	     |/        |    unmapped	| allocated per module
	 * +- shadow start  -+	       +----------------+
	 * | 1/8 addr space  |	       | zero pg mapping| (untracked)
	 * +- shadow end ----+---------+- shadow end ---+
	 *
	 * Current memory layout (KASAN_VMALLOC):
	 * +- 0 -------------+	       +- shadow start -+
	 * |1:1 ident mapping|	      /|1/8 of ident map|
	 * |		     |	     / |		|
	 * +-end of ident map+	    /  +----------------+
	 * | ... gap ...     |	   /   | kasan zero page| (untracked)
	 * |		     |	  /    | mapping	|
	 * +- vmalloc area  -+	 /     +----------------+
	 * | vmalloc_size    |	/      |shallow populate|
	 * +- modules vaddr -+ /       +----------------+
	 * | 2Gb	     |/        |shallow populate|
	 * +- shadow start  -+	       +----------------+
	 * | 1/8 addr space  |	       | zero pg mapping| (untracked)
	 * +- shadow end ----+---------+- shadow end ---+
	 */

	for_each_physmem_usable_range(i, &start, &end)
105
		kasan_populate(start, end, POPULATE_KASAN_MAP_SHADOW);
106 107 108
	if (IS_ENABLED(CONFIG_KASAN_VMALLOC)) {
		untracked_end = VMALLOC_START;
		/* shallowly populate kasan shadow for vmalloc and modules */
109
		kasan_populate(VMALLOC_START, MODULES_END, POPULATE_KASAN_SHALLOW);
110 111 112 113
	} else {
		untracked_end = MODULES_VADDR;
	}
	/* populate kasan shadow for untracked memory */
114 115
	kasan_populate(ident_map_size, untracked_end, POPULATE_KASAN_ZERO_SHADOW);
	kasan_populate(MODULES_END, _REGION1_SIZE, POPULATE_KASAN_ZERO_SHADOW);
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206
}

static bool kasan_pgd_populate_zero_shadow(pgd_t *pgd, unsigned long addr,
					   unsigned long end, enum populate_mode mode)
{
	if (mode == POPULATE_KASAN_ZERO_SHADOW &&
	    IS_ALIGNED(addr, PGDIR_SIZE) && end - addr >= PGDIR_SIZE) {
		pgd_populate(&init_mm, pgd, kasan_early_shadow_p4d);
		return true;
	}
	return false;
}

static bool kasan_p4d_populate_zero_shadow(p4d_t *p4d, unsigned long addr,
					   unsigned long end, enum populate_mode mode)
{
	if (mode == POPULATE_KASAN_ZERO_SHADOW &&
	    IS_ALIGNED(addr, P4D_SIZE) && end - addr >= P4D_SIZE) {
		p4d_populate(&init_mm, p4d, kasan_early_shadow_pud);
		return true;
	}
	return false;
}

static bool kasan_pud_populate_zero_shadow(pud_t *pud, unsigned long addr,
					   unsigned long end, enum populate_mode mode)
{
	if (mode == POPULATE_KASAN_ZERO_SHADOW &&
	    IS_ALIGNED(addr, PUD_SIZE) && end - addr >= PUD_SIZE) {
		pud_populate(&init_mm, pud, kasan_early_shadow_pmd);
		return true;
	}
	return false;
}

static bool kasan_pmd_populate_zero_shadow(pmd_t *pmd, unsigned long addr,
					   unsigned long end, enum populate_mode mode)
{
	if (mode == POPULATE_KASAN_ZERO_SHADOW &&
	    IS_ALIGNED(addr, PMD_SIZE) && end - addr >= PMD_SIZE) {
		pmd_populate(&init_mm, pmd, kasan_early_shadow_pte);
		return true;
	}
	return false;
}

static bool kasan_pte_populate_zero_shadow(pte_t *pte, enum populate_mode mode)
{
	pte_t entry;

	if (mode == POPULATE_KASAN_ZERO_SHADOW) {
		set_pte(pte, pte_z);
		return true;
	}
	return false;
}
#else

static inline void kasan_populate_shadow(void) {}

static inline bool kasan_pgd_populate_zero_shadow(pgd_t *pgd, unsigned long addr,
						  unsigned long end, enum populate_mode mode)
{
	return false;
}

static inline bool kasan_p4d_populate_zero_shadow(p4d_t *p4d, unsigned long addr,
						  unsigned long end, enum populate_mode mode)
{
	return false;
}

static inline bool kasan_pud_populate_zero_shadow(pud_t *pud, unsigned long addr,
						  unsigned long end, enum populate_mode mode)
{
	return false;
}

static inline bool kasan_pmd_populate_zero_shadow(pmd_t *pmd, unsigned long addr,
						  unsigned long end, enum populate_mode mode)
{
	return false;
}

static bool kasan_pte_populate_zero_shadow(pte_t *pte, enum populate_mode mode)
{
	return false;
}

#endif

207 208 209 210 211 212 213 214
/*
 * Mimic virt_to_kpte() in lack of init_mm symbol. Skip pmd NULL check though.
 */
static inline pte_t *__virt_to_kpte(unsigned long va)
{
	return pte_offset_kernel(pmd_offset(pud_offset(p4d_offset(pgd_offset_k(va), va), va), va), va);
}

215 216
static void *boot_crst_alloc(unsigned long val)
{
217
	unsigned long size = PAGE_SIZE << CRST_ALLOC_ORDER;
218 219
	unsigned long *table;

220 221
	table = (unsigned long *)physmem_alloc_top_down(RR_VMEM, size, size);
	crst_table_init(table, val);
222 223 224 225 226
	return table;
}

static pte_t *boot_pte_alloc(void)
{
227
	static void *pte_leftover;
228 229
	pte_t *pte;

230 231 232 233 234 235 236 237 238 239 240 241
	/*
	 * handling pte_leftovers this way helps to avoid memory fragmentation
	 * during POPULATE_KASAN_MAP_SHADOW when EDAT is off
	 */
	if (!pte_leftover) {
		pte_leftover = (void *)physmem_alloc_top_down(RR_VMEM, PAGE_SIZE, PAGE_SIZE);
		pte = pte_leftover + _PAGE_TABLE_SIZE;
	} else {
		pte = pte_leftover;
		pte_leftover = NULL;
	}

242 243 244 245
	memset64((u64 *)pte, _PAGE_INVALID, PTRS_PER_PTE);
	return pte;
}

246
static unsigned long _pa(unsigned long addr, unsigned long size, enum populate_mode mode)
247 248
{
	switch (mode) {
249 250
	case POPULATE_NONE:
		return -1;
251
	case POPULATE_DIRECT:
252
		return addr;
253 254
	case POPULATE_ABS_LOWCORE:
		return __abs_lowcore_pa(addr);
255 256 257 258 259 260
#ifdef CONFIG_KASAN
	case POPULATE_KASAN_MAP_SHADOW:
		addr = physmem_alloc_top_down(RR_VMEM, size, size);
		memset((void *)addr, 0, size);
		return addr;
#endif
261 262 263 264 265
	default:
		return -1;
	}
}

266 267 268 269 270 271 272 273 274 275 276 277
static bool can_large_pud(pud_t *pu_dir, unsigned long addr, unsigned long end)
{
	return machine.has_edat2 &&
	       IS_ALIGNED(addr, PUD_SIZE) && (end - addr) >= PUD_SIZE;
}

static bool can_large_pmd(pmd_t *pm_dir, unsigned long addr, unsigned long end)
{
	return machine.has_edat1 &&
	       IS_ALIGNED(addr, PMD_SIZE) && (end - addr) >= PMD_SIZE;
}

278 279
static void pgtable_pte_populate(pmd_t *pmd, unsigned long addr, unsigned long end,
				 enum populate_mode mode)
280
{
281
	unsigned long pages = 0;
282 283 284 285 286
	pte_t *pte, entry;

	pte = pte_offset_kernel(pmd, addr);
	for (; addr < end; addr += PAGE_SIZE, pte++) {
		if (pte_none(*pte)) {
287 288 289
			if (kasan_pte_populate_zero_shadow(pte, mode))
				continue;
			entry = __pte(_pa(addr, PAGE_SIZE, mode));
290 291 292
			entry = set_pte_bit(entry, PAGE_KERNEL);
			if (!machine.has_nx)
				entry = clear_pte_bit(entry, __pgprot(_PAGE_NOEXEC));
293
			set_pte(pte, entry);
294
			pages++;
295 296
		}
	}
297 298
	if (mode == POPULATE_DIRECT)
		update_page_count(PG_DIRECT_MAP_4K, pages);
299 300
}

301 302
static void pgtable_pmd_populate(pud_t *pud, unsigned long addr, unsigned long end,
				 enum populate_mode mode)
303
{
304
	unsigned long next, pages = 0;
305 306 307 308 309 310 311
	pmd_t *pmd, entry;
	pte_t *pte;

	pmd = pmd_offset(pud, addr);
	for (; addr < end; addr = next, pmd++) {
		next = pmd_addr_end(addr, end);
		if (pmd_none(*pmd)) {
312 313
			if (kasan_pmd_populate_zero_shadow(pmd, addr, next, mode))
				continue;
314
			if (can_large_pmd(pmd, addr, next)) {
315
				entry = __pmd(_pa(addr, _SEGMENT_SIZE, mode));
316 317 318
				entry = set_pmd_bit(entry, SEGMENT_KERNEL);
				if (!machine.has_nx)
					entry = clear_pmd_bit(entry, __pgprot(_SEGMENT_ENTRY_NOEXEC));
319
				set_pmd(pmd, entry);
320
				pages++;
321 322 323 324 325 326 327
				continue;
			}
			pte = boot_pte_alloc();
			pmd_populate(&init_mm, pmd, pte);
		} else if (pmd_large(*pmd)) {
			continue;
		}
328
		pgtable_pte_populate(pmd, addr, next, mode);
329
	}
330 331
	if (mode == POPULATE_DIRECT)
		update_page_count(PG_DIRECT_MAP_1M, pages);
332 333
}

334 335
static void pgtable_pud_populate(p4d_t *p4d, unsigned long addr, unsigned long end,
				 enum populate_mode mode)
336
{
337
	unsigned long next, pages = 0;
338 339 340 341 342 343 344
	pud_t *pud, entry;
	pmd_t *pmd;

	pud = pud_offset(p4d, addr);
	for (; addr < end; addr = next, pud++) {
		next = pud_addr_end(addr, end);
		if (pud_none(*pud)) {
345 346
			if (kasan_pud_populate_zero_shadow(pud, addr, next, mode))
				continue;
347
			if (can_large_pud(pud, addr, next)) {
348
				entry = __pud(_pa(addr, _REGION3_SIZE, mode));
349 350 351
				entry = set_pud_bit(entry, REGION3_KERNEL);
				if (!machine.has_nx)
					entry = clear_pud_bit(entry, __pgprot(_REGION_ENTRY_NOEXEC));
352
				set_pud(pud, entry);
353
				pages++;
354 355 356 357 358 359 360
				continue;
			}
			pmd = boot_crst_alloc(_SEGMENT_ENTRY_EMPTY);
			pud_populate(&init_mm, pud, pmd);
		} else if (pud_large(*pud)) {
			continue;
		}
361
		pgtable_pmd_populate(pud, addr, next, mode);
362
	}
363 364
	if (mode == POPULATE_DIRECT)
		update_page_count(PG_DIRECT_MAP_2G, pages);
365 366
}

367 368
static void pgtable_p4d_populate(pgd_t *pgd, unsigned long addr, unsigned long end,
				 enum populate_mode mode)
369 370 371 372 373 374 375 376 377
{
	unsigned long next;
	p4d_t *p4d;
	pud_t *pud;

	p4d = p4d_offset(pgd, addr);
	for (; addr < end; addr = next, p4d++) {
		next = p4d_addr_end(addr, end);
		if (p4d_none(*p4d)) {
378 379
			if (kasan_p4d_populate_zero_shadow(p4d, addr, next, mode))
				continue;
380 381 382
			pud = boot_crst_alloc(_REGION3_ENTRY_EMPTY);
			p4d_populate(&init_mm, p4d, pud);
		}
383
		pgtable_pud_populate(p4d, addr, next, mode);
384 385 386
	}
}

387
static void pgtable_populate(unsigned long addr, unsigned long end, enum populate_mode mode)
388 389 390 391 392 393 394 395 396
{
	unsigned long next;
	pgd_t *pgd;
	p4d_t *p4d;

	pgd = pgd_offset(&init_mm, addr);
	for (; addr < end; addr = next, pgd++) {
		next = pgd_addr_end(addr, end);
		if (pgd_none(*pgd)) {
397 398
			if (kasan_pgd_populate_zero_shadow(pgd, addr, next, mode))
				continue;
399 400 401
			p4d = boot_crst_alloc(_REGION2_ENTRY_EMPTY);
			pgd_populate(&init_mm, pgd, p4d);
		}
402 403 404 405
#ifdef CONFIG_KASAN
		if (mode == POPULATE_KASAN_SHALLOW)
			continue;
#endif
406
		pgtable_p4d_populate(pgd, addr, next, mode);
407 408 409
	}
}

410
void setup_vmem(unsigned long asce_limit)
411
{
412
	unsigned long start, end;
413 414
	unsigned long asce_type;
	unsigned long asce_bits;
415
	int i;
416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433

	if (asce_limit == _REGION1_SIZE) {
		asce_type = _REGION2_ENTRY_EMPTY;
		asce_bits = _ASCE_TYPE_REGION2 | _ASCE_TABLE_LENGTH;
	} else {
		asce_type = _REGION3_ENTRY_EMPTY;
		asce_bits = _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;
	}
	s390_invalid_asce = invalid_pg_dir | _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;

	crst_table_init((unsigned long *)swapper_pg_dir, asce_type);
	crst_table_init((unsigned long *)invalid_pg_dir, _REGION3_ENTRY_EMPTY);

	/*
	 * To allow prefixing the lowcore must be mapped with 4KB pages.
	 * To prevent creation of a large page at address 0 first map
	 * the lowcore and create the identity mapping only afterwards.
	 */
434
	pgtable_populate(0, sizeof(struct lowcore), POPULATE_DIRECT);
435
	for_each_physmem_usable_range(i, &start, &end)
436
		pgtable_populate(start, end, POPULATE_DIRECT);
437 438
	pgtable_populate(__abs_lowcore, __abs_lowcore + sizeof(struct lowcore),
			 POPULATE_ABS_LOWCORE);
439 440 441
	pgtable_populate(__memcpy_real_area, __memcpy_real_area + PAGE_SIZE,
			 POPULATE_NONE);
	memcpy_real_ptep = __virt_to_kpte(__memcpy_real_area);
442

443 444
	kasan_populate_shadow();

445 446 447 448 449 450 451 452 453
	S390_lowcore.kernel_asce = swapper_pg_dir | asce_bits;
	S390_lowcore.user_asce = s390_invalid_asce;

	__ctl_load(S390_lowcore.kernel_asce, 1, 1);
	__ctl_load(S390_lowcore.user_asce, 7, 7);
	__ctl_load(S390_lowcore.kernel_asce, 13, 13);

	init_mm.context.asce = S390_lowcore.kernel_asce;
}