fault.c 25.8 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12
/*
 *  Copyright (C) 1995  Linus Torvalds
 *  Copyright (C) 2001,2002 Andi Kleen, SuSE Labs.
 */

#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/ptrace.h>
13
#include <linux/mmiotrace.h>
Linus Torvalds's avatar
Linus Torvalds committed
14 15 16 17 18 19 20 21
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/tty.h>
#include <linux/vt_kern.h>		/* For unblank_screen() */
#include <linux/compiler.h>
Harvey Harrison's avatar
Harvey Harrison committed
22 23
#include <linux/highmem.h>
#include <linux/bootmem.h>		/* for max_low_pfn */
24
#include <linux/vmalloc.h>
Linus Torvalds's avatar
Linus Torvalds committed
25
#include <linux/module.h>
26
#include <linux/kprobes.h>
27
#include <linux/uaccess.h>
28
#include <linux/kdebug.h>
29
#include <linux/magic.h>
Linus Torvalds's avatar
Linus Torvalds committed
30 31

#include <asm/system.h>
Harvey Harrison's avatar
Harvey Harrison committed
32 33
#include <asm/desc.h>
#include <asm/segment.h>
Linus Torvalds's avatar
Linus Torvalds committed
34 35 36 37 38
#include <asm/pgalloc.h>
#include <asm/smp.h>
#include <asm/tlbflush.h>
#include <asm/proto.h>
#include <asm-generic/sections.h>
39
#include <asm/traps.h>
Linus Torvalds's avatar
Linus Torvalds committed
40

41 42 43 44 45 46 47 48
/*
 * Page fault error code bits
 *	bit 0 == 0 means no page found, 1 means protection fault
 *	bit 1 == 0 means read, 1 means write
 *	bit 2 == 0 means kernel, 1 means user-mode
 *	bit 3 == 1 means use of reserved bit detected
 *	bit 4 == 1 means fault was an instruction fetch
 */
49
#define PF_PROT		(1<<0)
50
#define PF_WRITE	(1<<1)
51 52
#define PF_USER		(1<<2)
#define PF_RSVD		(1<<3)
53 54
#define PF_INSTR	(1<<4)

55
static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr)
56
{
57
#ifdef CONFIG_MMIOTRACE
58 59 60
	if (unlikely(is_kmmio_active()))
		if (kmmio_handler(regs, addr) == 1)
			return -1;
61
#endif
62
	return 0;
63 64
}

65
static inline int notify_page_fault(struct pt_regs *regs)
66
{
67
#ifdef CONFIG_KPROBES
68 69 70
	int ret = 0;

	/* kprobe_running() needs smp_processor_id() */
71
	if (!user_mode_vm(regs)) {
72 73 74 75 76
		preempt_disable();
		if (kprobe_running() && kprobe_fault_handler(regs, 14))
			ret = 1;
		preempt_enable();
	}
77

78 79 80 81
	return ret;
#else
	return 0;
#endif
82
}
83

84 85 86 87 88 89 90 91 92 93 94
/*
 * X86_32
 * Sometimes AMD Athlon/Opteron CPUs report invalid exceptions on prefetch.
 * Check that here and ignore it.
 *
 * X86_64
 * Sometimes the CPU reports invalid exceptions on prefetch.
 * Check that here and ignore it.
 *
 * Opcode checker based on code by Richard Brunner
 */
95 96
static int is_prefetch(struct pt_regs *regs, unsigned long error_code,
			unsigned long addr)
97
{
98
	unsigned char *instr;
Linus Torvalds's avatar
Linus Torvalds committed
99
	int scan_more = 1;
100
	int prefetch = 0;
101
	unsigned char *max_instr;
Linus Torvalds's avatar
Linus Torvalds committed
102

Ingo Molnar's avatar
Ingo Molnar committed
103 104 105 106
	/*
	 * If it was a exec (instruction fetch) fault on NX page, then
	 * do not ignore the fault:
	 */
107
	if (error_code & PF_INSTR)
Linus Torvalds's avatar
Linus Torvalds committed
108
		return 0;
109

110
	instr = (unsigned char *)convert_ip_to_linear(current, regs);
111
	max_instr = instr + 15;
Linus Torvalds's avatar
Linus Torvalds committed
112

113
	if (user_mode(regs) && instr >= (unsigned char *)TASK_SIZE)
Linus Torvalds's avatar
Linus Torvalds committed
114 115
		return 0;

116
	while (scan_more && instr < max_instr) {
Linus Torvalds's avatar
Linus Torvalds committed
117 118 119 120
		unsigned char opcode;
		unsigned char instr_hi;
		unsigned char instr_lo;

121
		if (probe_kernel_address(instr, opcode))
122
			break;
Linus Torvalds's avatar
Linus Torvalds committed
123

124 125
		instr_hi = opcode & 0xf0;
		instr_lo = opcode & 0x0f;
Linus Torvalds's avatar
Linus Torvalds committed
126 127
		instr++;

128
		switch (instr_hi) {
Linus Torvalds's avatar
Linus Torvalds committed
129 130
		case 0x20:
		case 0x30:
131 132 133 134 135 136
			/*
			 * Values 0x26,0x2E,0x36,0x3E are valid x86 prefixes.
			 * In X86_64 long mode, the CPU will signal invalid
			 * opcode if some of these prefixes are present so
			 * X86_64 will never get here anyway
			 */
Linus Torvalds's avatar
Linus Torvalds committed
137 138
			scan_more = ((instr_lo & 7) == 0x6);
			break;
139
#ifdef CONFIG_X86_64
Linus Torvalds's avatar
Linus Torvalds committed
140
		case 0x40:
141 142 143 144 145 146 147
			/*
			 * In AMD64 long mode 0x40..0x4F are valid REX prefixes
			 * Need to figure out under what instruction mode the
			 * instruction was issued. Could check the LDT for lm,
			 * but for now it's good enough to assume that long
			 * mode only uses well known segments or kernel.
			 */
148
			scan_more = (!user_mode(regs)) || (regs->cs == __USER_CS);
Linus Torvalds's avatar
Linus Torvalds committed
149
			break;
150
#endif
Linus Torvalds's avatar
Linus Torvalds committed
151 152 153
		case 0x60:
			/* 0x64 thru 0x67 are valid prefixes in all modes. */
			scan_more = (instr_lo & 0xC) == 0x4;
154
			break;
Linus Torvalds's avatar
Linus Torvalds committed
155
		case 0xF0:
156
			/* 0xF0, 0xF2, 0xF3 are valid prefixes in all modes. */
Linus Torvalds's avatar
Linus Torvalds committed
157
			scan_more = !instr_lo || (instr_lo>>1) == 1;
158
			break;
Linus Torvalds's avatar
Linus Torvalds committed
159 160 161
		case 0x00:
			/* Prefetch instruction is 0x0F0D or 0x0F18 */
			scan_more = 0;
162

163
			if (probe_kernel_address(instr, opcode))
Linus Torvalds's avatar
Linus Torvalds committed
164 165 166
				break;
			prefetch = (instr_lo == 0xF) &&
				(opcode == 0x0D || opcode == 0x18);
167
			break;
Linus Torvalds's avatar
Linus Torvalds committed
168 169 170
		default:
			scan_more = 0;
			break;
171
		}
Linus Torvalds's avatar
Linus Torvalds committed
172 173 174 175
	}
	return prefetch;
}

176 177 178 179 180 181 182 183 184 185 186 187
static void force_sig_info_fault(int si_signo, int si_code,
	unsigned long address, struct task_struct *tsk)
{
	siginfo_t info;

	info.si_signo = si_signo;
	info.si_errno = 0;
	info.si_code = si_code;
	info.si_addr = (void __user *)address;
	force_sig_info(si_signo, &info, tsk);
}

188
#ifdef CONFIG_X86_64
189 190
static int bad_address(void *p)
{
Linus Torvalds's avatar
Linus Torvalds committed
191
	unsigned long dummy;
192
	return probe_kernel_address((unsigned long *)p, dummy);
193
}
194
#endif
Linus Torvalds's avatar
Linus Torvalds committed
195

196
static void dump_pagetable(unsigned long address)
Linus Torvalds's avatar
Linus Torvalds committed
197
{
198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
#ifdef CONFIG_X86_32
	__typeof__(pte_val(__pte(0))) page;

	page = read_cr3();
	page = ((__typeof__(page) *) __va(page))[address >> PGDIR_SHIFT];
#ifdef CONFIG_X86_PAE
	printk("*pdpt = %016Lx ", page);
	if ((page >> PAGE_SHIFT) < max_low_pfn
	    && page & _PAGE_PRESENT) {
		page &= PAGE_MASK;
		page = ((__typeof__(page) *) __va(page))[(address >> PMD_SHIFT)
		                                         & (PTRS_PER_PMD - 1)];
		printk(KERN_CONT "*pde = %016Lx ", page);
		page &= ~_PAGE_NX;
	}
#else
	printk("*pde = %08lx ", page);
#endif

	/*
	 * We must not directly access the pte in the highpte
	 * case if the page table is located in highmem.
	 * And let's rather not kmap-atomic the pte, just in case
	 * it's allocated already.
	 */
	if ((page >> PAGE_SHIFT) < max_low_pfn
	    && (page & _PAGE_PRESENT)
	    && !(page & _PAGE_PSE)) {
		page &= PAGE_MASK;
		page = ((__typeof__(page) *) __va(page))[(address >> PAGE_SHIFT)
		                                         & (PTRS_PER_PTE - 1)];
		printk("*pte = %0*Lx ", sizeof(page)*2, (u64)page);
	}

	printk("\n");
#else /* CONFIG_X86_64 */
Linus Torvalds's avatar
Linus Torvalds committed
234 235 236 237 238
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

239
	pgd = (pgd_t *)read_cr3();
Linus Torvalds's avatar
Linus Torvalds committed
240

241
	pgd = __va((unsigned long)pgd & PHYSICAL_PAGE_MASK);
Linus Torvalds's avatar
Linus Torvalds committed
242 243
	pgd += pgd_index(address);
	if (bad_address(pgd)) goto bad;
244
	printk("PGD %lx ", pgd_val(*pgd));
245
	if (!pgd_present(*pgd)) goto ret;
Linus Torvalds's avatar
Linus Torvalds committed
246

247
	pud = pud_offset(pgd, address);
Linus Torvalds's avatar
Linus Torvalds committed
248 249
	if (bad_address(pud)) goto bad;
	printk("PUD %lx ", pud_val(*pud));
250 251
	if (!pud_present(*pud) || pud_large(*pud))
		goto ret;
Linus Torvalds's avatar
Linus Torvalds committed
252 253 254 255

	pmd = pmd_offset(pud, address);
	if (bad_address(pmd)) goto bad;
	printk("PMD %lx ", pmd_val(*pmd));
256
	if (!pmd_present(*pmd) || pmd_large(*pmd)) goto ret;
Linus Torvalds's avatar
Linus Torvalds committed
257 258 259

	pte = pte_offset_kernel(pmd, address);
	if (bad_address(pte)) goto bad;
260
	printk("PTE %lx", pte_val(*pte));
Linus Torvalds's avatar
Linus Torvalds committed
261 262 263 264 265
ret:
	printk("\n");
	return;
bad:
	printk("BAD\n");
266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
#endif
}

#ifdef CONFIG_X86_32
static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
{
	unsigned index = pgd_index(address);
	pgd_t *pgd_k;
	pud_t *pud, *pud_k;
	pmd_t *pmd, *pmd_k;

	pgd += index;
	pgd_k = init_mm.pgd + index;

	if (!pgd_present(*pgd_k))
		return NULL;

	/*
	 * set_pgd(pgd, *pgd_k); here would be useless on PAE
	 * and redundant with the set_pmd() on non-PAE. As would
	 * set_pud.
	 */

	pud = pud_offset(pgd, address);
	pud_k = pud_offset(pgd_k, address);
	if (!pud_present(*pud_k))
		return NULL;

	pmd = pmd_offset(pud, address);
	pmd_k = pmd_offset(pud_k, address);
	if (!pmd_present(*pmd_k))
		return NULL;
	if (!pmd_present(*pmd)) {
		set_pmd(pmd, *pmd_k);
		arch_flush_lazy_mmu_mode();
	} else
		BUG_ON(pmd_page(*pmd) != pmd_page(*pmd_k));
	return pmd_k;
Linus Torvalds's avatar
Linus Torvalds committed
304
}
305
#endif
Linus Torvalds's avatar
Linus Torvalds committed
306

307
#ifdef CONFIG_X86_64
308
static const char errata93_warning[] =
Linus Torvalds's avatar
Linus Torvalds committed
309 310 311 312
KERN_ERR "******* Your BIOS seems to not contain a fix for K8 errata #93\n"
KERN_ERR "******* Working around it, but it may cause SEGVs or burn power.\n"
KERN_ERR "******* Please consider a BIOS update.\n"
KERN_ERR "******* Disabling USB legacy in the BIOS may also help.\n";
313
#endif
Linus Torvalds's avatar
Linus Torvalds committed
314 315 316

/* Workaround for K8 erratum #93 & buggy BIOS.
   BIOS SMM functions are required to use a specific workaround
317 318
   to avoid corruption of the 64bit RIP register on C stepping K8.
   A lot of BIOS that didn't get tested properly miss this.
Linus Torvalds's avatar
Linus Torvalds committed
319 320
   The OS sees this as a page fault with the upper 32bits of RIP cleared.
   Try to work around it here.
321 322 323
   Note we only handle faults in kernel here.
   Does nothing for X86_32
 */
324
static int is_errata93(struct pt_regs *regs, unsigned long address)
Linus Torvalds's avatar
Linus Torvalds committed
325
{
326
#ifdef CONFIG_X86_64
Linus Torvalds's avatar
Linus Torvalds committed
327
	static int warned;
328
	if (address != regs->ip)
Linus Torvalds's avatar
Linus Torvalds committed
329
		return 0;
330
	if ((address >> 32) != 0)
Linus Torvalds's avatar
Linus Torvalds committed
331 332
		return 0;
	address |= 0xffffffffUL << 32;
333 334
	if ((address >= (u64)_stext && address <= (u64)_etext) ||
	    (address >= MODULES_VADDR && address <= MODULES_END)) {
Linus Torvalds's avatar
Linus Torvalds committed
335
		if (!warned) {
336
			printk(errata93_warning);
Linus Torvalds's avatar
Linus Torvalds committed
337 338
			warned = 1;
		}
339
		regs->ip = address;
Linus Torvalds's avatar
Linus Torvalds committed
340 341
		return 1;
	}
342
#endif
Linus Torvalds's avatar
Linus Torvalds committed
343
	return 0;
344
}
Linus Torvalds's avatar
Linus Torvalds committed
345

346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
/*
 * Work around K8 erratum #100 K8 in compat mode occasionally jumps to illegal
 * addresses >4GB.  We catch this in the page fault handler because these
 * addresses are not reachable. Just detect this case and return.  Any code
 * segment in LDT is compatibility mode.
 */
static int is_errata100(struct pt_regs *regs, unsigned long address)
{
#ifdef CONFIG_X86_64
	if ((regs->cs == __USER32_CS || (regs->cs & (1<<2))) &&
	    (address >> 32))
		return 1;
#endif
	return 0;
}

362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
{
#ifdef CONFIG_X86_F00F_BUG
	unsigned long nr;
	/*
	 * Pentium F0 0F C7 C8 bug workaround.
	 */
	if (boot_cpu_data.f00f_bug) {
		nr = (address - idt_descr.address) >> 3;

		if (nr == 6) {
			do_invalid_op(regs, 0);
			return 1;
		}
	}
#endif
	return 0;
}

381 382 383
static void show_fault_oops(struct pt_regs *regs, unsigned long error_code,
			    unsigned long address)
{
384 385 386
#ifdef CONFIG_X86_32
	if (!oops_may_print())
		return;
387
#endif
388 389 390

#ifdef CONFIG_X86_PAE
	if (error_code & PF_INSTR) {
391
		unsigned int level;
392 393 394 395 396
		pte_t *pte = lookup_address(address, &level);

		if (pte && pte_present(*pte) && !pte_exec(*pte))
			printk(KERN_CRIT "kernel tried to execute "
				"NX-protected page - exploit attempt? "
397
				"(uid: %d)\n", current_uid());
398 399 400
	}
#endif

401
	printk(KERN_ALERT "BUG: unable to handle kernel ");
402
	if (address < PAGE_SIZE)
403
		printk(KERN_CONT "NULL pointer dereference");
404
	else
405
		printk(KERN_CONT "paging request");
406
	printk(KERN_CONT " at %p\n", (void *) address);
407
	printk(KERN_ALERT "IP:");
408 409 410 411
	printk_address(regs->ip, 1);
	dump_pagetable(address);
}

412
#ifdef CONFIG_X86_64
413 414
static noinline void pgtable_bad(struct pt_regs *regs,
			 unsigned long error_code, unsigned long address)
Linus Torvalds's avatar
Linus Torvalds committed
415
{
416
	unsigned long flags = oops_begin();
417
	int sig = SIGKILL;
418
	struct task_struct *tsk = current;
419

Linus Torvalds's avatar
Linus Torvalds committed
420
	printk(KERN_ALERT "%s: Corrupted page table at address %lx\n",
421
	       tsk->comm, address);
Linus Torvalds's avatar
Linus Torvalds committed
422
	dump_pagetable(address);
423 424 425
	tsk->thread.cr2 = address;
	tsk->thread.trap_no = 14;
	tsk->thread.error_code = error_code;
426
	if (__die("Bad pagetable", regs, error_code))
427 428
		sig = 0;
	oops_end(flags, regs, sig);
Linus Torvalds's avatar
Linus Torvalds committed
429
}
430
#endif
Linus Torvalds's avatar
Linus Torvalds committed
431

432 433 434 435
static noinline void no_context(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	struct task_struct *tsk = current;
436 437
	unsigned long *stackend;

438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473
#ifdef CONFIG_X86_64
	unsigned long flags;
	int sig;
#endif

	/* Are we prepared to handle this kernel fault?  */
	if (fixup_exception(regs))
		return;

	/*
	 * X86_32
	 * Valid to do another page fault here, because if this fault
	 * had been triggered by is_prefetch fixup_exception would have
	 * handled it.
	 *
	 * X86_64
	 * Hall of shame of CPU/BIOS bugs.
	 */
	if (is_prefetch(regs, error_code, address))
		return;

	if (is_errata93(regs, address))
		return;

	/*
	 * Oops. The kernel tried to access some bad page. We'll have to
	 * terminate things with extreme prejudice.
	 */
#ifdef CONFIG_X86_32
	bust_spinlocks(1);
#else
	flags = oops_begin();
#endif

	show_fault_oops(regs, error_code, address);

474 475 476 477
 	stackend = end_of_stack(tsk);
	if (*stackend != STACK_END_MAGIC)
		printk(KERN_ALERT "Thread overran stack, or stack corrupted\n");

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621
	tsk->thread.cr2 = address;
	tsk->thread.trap_no = 14;
	tsk->thread.error_code = error_code;

#ifdef CONFIG_X86_32
	die("Oops", regs, error_code);
	bust_spinlocks(0);
	do_exit(SIGKILL);
#else
	sig = SIGKILL;
	if (__die("Oops", regs, error_code))
		sig = 0;
	/* Executive summary in case the body of the oops scrolled away */
	printk(KERN_EMERG "CR2: %016lx\n", address);
	oops_end(flags, regs, sig);
#endif
}

static void __bad_area_nosemaphore(struct pt_regs *regs,
			unsigned long error_code, unsigned long address,
			int si_code)
{
	struct task_struct *tsk = current;

	/* User mode accesses just cause a SIGSEGV */
	if (error_code & PF_USER) {
		/*
		 * It's possible to have interrupts off here.
		 */
		local_irq_enable();

		/*
		 * Valid to do another page fault here because this one came
		 * from user space.
		 */
		if (is_prefetch(regs, error_code, address))
			return;

		if (is_errata100(regs, address))
			return;

		if (show_unhandled_signals && unhandled_signal(tsk, SIGSEGV) &&
		    printk_ratelimit()) {
			printk(
			"%s%s[%d]: segfault at %lx ip %p sp %p error %lx",
			task_pid_nr(tsk) > 1 ? KERN_INFO : KERN_EMERG,
			tsk->comm, task_pid_nr(tsk), address,
			(void *) regs->ip, (void *) regs->sp, error_code);
			print_vma_addr(" in ", regs->ip);
			printk("\n");
		}

		tsk->thread.cr2 = address;
		/* Kernel addresses are always protection faults */
		tsk->thread.error_code = error_code | (address >= TASK_SIZE);
		tsk->thread.trap_no = 14;
		force_sig_info_fault(SIGSEGV, si_code, address, tsk);
		return;
	}

	if (is_f00f_bug(regs, address))
		return;

	no_context(regs, error_code, address);
}

static noinline void bad_area_nosemaphore(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	__bad_area_nosemaphore(regs, error_code, address, SEGV_MAPERR);
}

static void __bad_area(struct pt_regs *regs,
			unsigned long error_code, unsigned long address,
			int si_code)
{
	struct mm_struct *mm = current->mm;

	/*
	 * Something tried to access memory that isn't in our memory map..
	 * Fix it, but check if it's kernel or user first..
	 */
	up_read(&mm->mmap_sem);

	__bad_area_nosemaphore(regs, error_code, address, si_code);
}

static noinline void bad_area(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	__bad_area(regs, error_code, address, SEGV_MAPERR);
}

static noinline void bad_area_access_error(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	__bad_area(regs, error_code, address, SEGV_ACCERR);
}

/* TODO: fixup for "mm-invoke-oom-killer-from-page-fault.patch" */
static void out_of_memory(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	/*
	 * We ran out of memory, call the OOM killer, and return the userspace
	 * (which will retry the fault, or kill us if we got oom-killed).
	 */
	up_read(&current->mm->mmap_sem);
	pagefault_out_of_memory();
}

static void do_sigbus(struct pt_regs *regs,
			unsigned long error_code, unsigned long address)
{
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;

	up_read(&mm->mmap_sem);

	/* Kernel mode? Handle exceptions or die */
	if (!(error_code & PF_USER))
		no_context(regs, error_code, address);
#ifdef CONFIG_X86_32
	/* User space => ok to do another page fault */
	if (is_prefetch(regs, error_code, address))
		return;
#endif
	tsk->thread.cr2 = address;
	tsk->thread.error_code = error_code;
	tsk->thread.trap_no = 14;
	force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk);
}

static noinline void mm_fault_error(struct pt_regs *regs,
		unsigned long error_code, unsigned long address, unsigned int fault)
{
	if (fault & VM_FAULT_OOM)
		out_of_memory(regs, error_code, address);
	else if (fault & VM_FAULT_SIGBUS)
		do_sigbus(regs, error_code, address);
	else
		BUG();
}

622 623 624 625 626 627 628 629 630 631
static int spurious_fault_check(unsigned long error_code, pte_t *pte)
{
	if ((error_code & PF_WRITE) && !pte_write(*pte))
		return 0;
	if ((error_code & PF_INSTR) && !pte_exec(*pte))
		return 0;

	return 1;
}

632 633 634 635 636 637 638 639 640
/*
 * Handle a spurious fault caused by a stale TLB entry.  This allows
 * us to lazily refresh the TLB when increasing the permissions of a
 * kernel page (RO -> RW or NX -> X).  Doing it eagerly is very
 * expensive since that implies doing a full cross-processor TLB
 * flush, even if no stale TLB entries exist on other processors.
 * There are no security implications to leaving a stale TLB when
 * increasing the permissions on a page.
 */
641 642
static noinline int spurious_fault(unsigned long error_code,
				unsigned long address)
643 644 645 646 647
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
648
	int ret;
649 650 651 652 653 654 655 656 657 658 659 660 661

	/* Reserved-bit violation or user access to kernel space? */
	if (error_code & (PF_USER | PF_RSVD))
		return 0;

	pgd = init_mm.pgd + pgd_index(address);
	if (!pgd_present(*pgd))
		return 0;

	pud = pud_offset(pgd, address);
	if (!pud_present(*pud))
		return 0;

662 663 664
	if (pud_large(*pud))
		return spurious_fault_check(error_code, (pte_t *) pud);

665 666 667 668
	pmd = pmd_offset(pud, address);
	if (!pmd_present(*pmd))
		return 0;

669 670 671
	if (pmd_large(*pmd))
		return spurious_fault_check(error_code, (pte_t *) pmd);

672 673 674 675
	pte = pte_offset_kernel(pmd, address);
	if (!pte_present(*pte))
		return 0;

676 677 678 679 680 681 682 683 684 685 686
	ret = spurious_fault_check(error_code, pte);
	if (!ret)
		return 0;

	/*
	 * Make sure we have permissions in PMD
	 * If not, then there's a bug in the page tables.
	 */
	ret = spurious_fault_check(error_code, (pte_t *) pmd);
	WARN_ONCE(!ret, "PMD has incorrect permission bits\n");
	return ret;
687 688
}

Linus Torvalds's avatar
Linus Torvalds committed
689
/*
690 691 692 693
 * X86_32
 * Handle a fault on the vmalloc or module mapping area
 *
 * X86_64
694
 * Handle a fault on the vmalloc area
695 696
 *
 * This assumes no large pages in there.
Linus Torvalds's avatar
Linus Torvalds committed
697
 */
698
static noinline int vmalloc_fault(unsigned long address)
Linus Torvalds's avatar
Linus Torvalds committed
699
{
700 701 702 703
#ifdef CONFIG_X86_32
	unsigned long pgd_paddr;
	pmd_t *pmd_k;
	pte_t *pte_k;
704 705 706 707 708

	/* Make sure we are in vmalloc area */
	if (!(address >= VMALLOC_START && address < VMALLOC_END))
		return -1;

709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724
	/*
	 * Synchronize this task's top level page-table
	 * with the 'reference' page table.
	 *
	 * Do _not_ use "current" here. We might be inside
	 * an interrupt in the middle of a task switch..
	 */
	pgd_paddr = read_cr3();
	pmd_k = vmalloc_sync_one(__va(pgd_paddr), address);
	if (!pmd_k)
		return -1;
	pte_k = pte_offset_kernel(pmd_k, address);
	if (!pte_present(*pte_k))
		return -1;
	return 0;
#else
Linus Torvalds's avatar
Linus Torvalds committed
725 726 727 728 729
	pgd_t *pgd, *pgd_ref;
	pud_t *pud, *pud_ref;
	pmd_t *pmd, *pmd_ref;
	pte_t *pte, *pte_ref;

730 731 732 733
	/* Make sure we are in vmalloc area */
	if (!(address >= VMALLOC_START && address < VMALLOC_END))
		return -1;

Linus Torvalds's avatar
Linus Torvalds committed
734 735 736 737
	/* Copy kernel mappings over when needed. This can also
	   happen within a race in page table update. In the later
	   case just flush. */

738
	pgd = pgd_offset(current->active_mm, address);
Linus Torvalds's avatar
Linus Torvalds committed
739 740 741 742 743
	pgd_ref = pgd_offset_k(address);
	if (pgd_none(*pgd_ref))
		return -1;
	if (pgd_none(*pgd))
		set_pgd(pgd, *pgd_ref);
744
	else
745
		BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref));
Linus Torvalds's avatar
Linus Torvalds committed
746 747 748 749 750 751 752 753

	/* Below here mismatches are bugs because these lower tables
	   are shared */

	pud = pud_offset(pgd, address);
	pud_ref = pud_offset(pgd_ref, address);
	if (pud_none(*pud_ref))
		return -1;
754
	if (pud_none(*pud) || pud_page_vaddr(*pud) != pud_page_vaddr(*pud_ref))
Linus Torvalds's avatar
Linus Torvalds committed
755 756 757 758 759 760 761 762 763 764 765
		BUG();
	pmd = pmd_offset(pud, address);
	pmd_ref = pmd_offset(pud_ref, address);
	if (pmd_none(*pmd_ref))
		return -1;
	if (pmd_none(*pmd) || pmd_page(*pmd) != pmd_page(*pmd_ref))
		BUG();
	pte_ref = pte_offset_kernel(pmd_ref, address);
	if (!pte_present(*pte_ref))
		return -1;
	pte = pte_offset_kernel(pmd, address);
766 767 768 769
	/* Don't use pte_page here, because the mappings can point
	   outside mem_map, and the NUMA hash lookup cannot handle
	   that. */
	if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref))
Linus Torvalds's avatar
Linus Torvalds committed
770 771
		BUG();
	return 0;
772
#endif
Linus Torvalds's avatar
Linus Torvalds committed
773 774
}

775
int show_unhandled_signals = 1;
Linus Torvalds's avatar
Linus Torvalds committed
776

777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
static inline int access_error(unsigned long error_code, int write,
				struct vm_area_struct *vma)
{
	if (write) {
		/* write, present and write, not present */
		if (unlikely(!(vma->vm_flags & VM_WRITE)))
			return 1;
	} else if (unlikely(error_code & PF_PROT)) {
		/* read, present */
		return 1;
	} else {
		/* read, not present */
		if (unlikely(!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE))))
			return 1;
	}

	return 0;
}

796 797 798 799 800 801 802 803 804
static int fault_in_kernel_space(unsigned long address)
{
#ifdef CONFIG_X86_32
	return address >= TASK_SIZE;
#else /* !CONFIG_X86_32 */
	return address >= TASK_SIZE64;
#endif /* CONFIG_X86_32 */
}

Linus Torvalds's avatar
Linus Torvalds committed
805 806 807 808 809
/*
 * This routine handles page faults.  It determines the address,
 * and the problem, and then passes it off to one of the appropriate
 * routines.
 */
810 811 812 813
#ifdef CONFIG_X86_64
asmlinkage
#endif
void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
Linus Torvalds's avatar
Linus Torvalds committed
814
{
815
	unsigned long address;
Linus Torvalds's avatar
Linus Torvalds committed
816 817
	struct task_struct *tsk;
	struct mm_struct *mm;
818
	struct vm_area_struct *vma;
819
	int write;
820
	int fault;
Linus Torvalds's avatar
Linus Torvalds committed
821

822 823 824 825
	tsk = current;
	mm = tsk->mm;
	prefetchw(&mm->mmap_sem);

Linus Torvalds's avatar
Linus Torvalds committed
826
	/* get the address */
827
	address = read_cr2();
Linus Torvalds's avatar
Linus Torvalds committed
828

829
	if (unlikely(kmmio_fault(regs, address)))
830
		return;
Linus Torvalds's avatar
Linus Torvalds committed
831 832 833 834 835 836 837 838 839 840 841 842

	/*
	 * We fault-in kernel-space virtual memory on-demand. The
	 * 'reference' page table is init_mm.pgd.
	 *
	 * NOTE! We MUST NOT take any locks for this case. We may
	 * be in an interrupt or a critical region, and should
	 * only copy the information from the master page table,
	 * nothing more.
	 *
	 * This verifies that the fault happens in kernel space
	 * (error_code & 4) == 0, and that the fault was not a
843
	 * protection error (error_code & 9) == 0.
Linus Torvalds's avatar
Linus Torvalds committed
844
	 */
845
	if (unlikely(fault_in_kernel_space(address))) {
846 847 848
		if (!(error_code & (PF_RSVD|PF_USER|PF_PROT)) &&
		    vmalloc_fault(address) >= 0)
			return;
849 850

		/* Can handle a stale RO->RW TLB */
851
		if (spurious_fault(error_code, address))
852 853
			return;

854 855 856
		/* kprobes don't want to hook the spurious faults. */
		if (notify_page_fault(regs))
			return;
857 858 859 860
		/*
		 * Don't take the mm semaphore here. If we fixup a prefetch
		 * fault we could otherwise deadlock.
		 */
861 862
		bad_area_nosemaphore(regs, error_code, address);
		return;
863 864
	}

865
	/* kprobes don't want to hook the spurious faults. */
866
	if (unlikely(notify_page_fault(regs)))
867
		return;
868
	/*
869 870 871 872 873
	 * It's safe to allow irq's after cr2 has been saved and the
	 * vmalloc fault has been handled.
	 *
	 * User-mode registers count as a user access even for any
	 * potential system fault or CPU buglet.
874
	 */
875 876 877 878
	if (user_mode_vm(regs)) {
		local_irq_enable();
		error_code |= PF_USER;
	} else if (regs->flags & X86_EFLAGS_IF)
879 880
		local_irq_enable();

881
#ifdef CONFIG_X86_64
882
	if (unlikely(error_code & PF_RSVD))
883
		pgtable_bad(regs, error_code, address);
884
#endif
Linus Torvalds's avatar
Linus Torvalds committed
885 886

	/*
887 888
	 * If we're in an interrupt, have no user context or are running in an
	 * atomic region then we must not take the fault.
Linus Torvalds's avatar
Linus Torvalds committed
889
	 */
890 891 892 893
	if (unlikely(in_atomic() || !mm)) {
		bad_area_nosemaphore(regs, error_code, address);
		return;
	}
Linus Torvalds's avatar
Linus Torvalds committed
894

895 896
	/*
	 * When running in the kernel we expect faults to occur only to
Linus Torvalds's avatar
Linus Torvalds committed
897
	 * addresses in user space.  All other faults represent errors in the
Simon Arlott's avatar
Simon Arlott committed
898
	 * kernel and should generate an OOPS.  Unfortunately, in the case of an
899
	 * erroneous fault occurring in a code path which already holds mmap_sem
Linus Torvalds's avatar
Linus Torvalds committed
900 901 902 903 904 905
	 * we will deadlock attempting to validate the fault against the
	 * address space.  Luckily the kernel only validly references user
	 * space from well defined areas of code, which are listed in the
	 * exceptions table.
	 *
	 * As the vast majority of faults will be valid we will only perform
Simon Arlott's avatar
Simon Arlott committed
906
	 * the source reference check when there is a possibility of a deadlock.
Linus Torvalds's avatar
Linus Torvalds committed
907 908 909 910
	 * Attempt to lock the address space, if we cannot we then validate the
	 * source.  If this is invalid we can skip the address space check,
	 * thus avoiding the deadlock.
	 */
911
	if (unlikely(!down_read_trylock(&mm->mmap_sem))) {
912
		if ((error_code & PF_USER) == 0 &&
913 914 915 916
		    !search_exception_tables(regs->ip)) {
			bad_area_nosemaphore(regs, error_code, address);
			return;
		}
Linus Torvalds's avatar
Linus Torvalds committed
917
		down_read(&mm->mmap_sem);
918 919 920 921 922 923
	} else {
		/*
		 * The above down_read_trylock() might have succeeded in which
		 * case we'll have missed the might_sleep() from down_read().
		 */
		might_sleep();
Linus Torvalds's avatar
Linus Torvalds committed
924 925 926
	}

	vma = find_vma(mm, address);
927 928 929 930 931
	if (unlikely(!vma)) {
		bad_area(regs, error_code, address);
		return;
	}
	if (likely(vma->vm_start <= address))
Linus Torvalds's avatar
Linus Torvalds committed
932
		goto good_area;
933 934 935 936
	if (unlikely(!(vma->vm_flags & VM_GROWSDOWN))) {
		bad_area(regs, error_code, address);
		return;
	}
937
	if (error_code & PF_USER) {
938 939 940 941 942
		/*
		 * Accessing the stack below %sp is always a bug.
		 * The large cushion allows instructions like enter
		 * and pusha to work.  ("enter $65535,$31" pushes
		 * 32 pointers and then decrements %sp by 65535.)
943
		 */
944 945 946 947
		if (unlikely(address + 65536 + 32 * sizeof(unsigned long) < regs->sp)) {
			bad_area(regs, error_code, address);
			return;
		}
Linus Torvalds's avatar
Linus Torvalds committed
948
	}
949 950 951 952 953 954 955 956 957
	if (unlikely(expand_stack(vma, address))) {
		bad_area(regs, error_code, address);
		return;
	}

	/*
	 * Ok, we have a good vm_area for this memory access, so
	 * we can handle it..
	 */
Linus Torvalds's avatar
Linus Torvalds committed
958
good_area:
959 960 961 962
	write = error_code & PF_WRITE;
	if (unlikely(access_error(error_code, write, vma))) {
		bad_area_access_error(regs, error_code, address);
		return;
Linus Torvalds's avatar
Linus Torvalds committed
963 964 965 966 967 968 969
	}

	/*
	 * If for any reason at all we couldn't handle the fault,
	 * make sure we exit gracefully rather than endlessly redo
	 * the fault.
	 */
Nick Piggin's avatar
Nick Piggin committed
970 971
	fault = handle_mm_fault(mm, vma, address, write);
	if (unlikely(fault & VM_FAULT_ERROR)) {
972 973
		mm_fault_error(regs, error_code, address, fault);
		return;
Linus Torvalds's avatar
Linus Torvalds committed
974
	}
Nick Piggin's avatar
Nick Piggin committed
975 976 977 978
	if (fault & VM_FAULT_MAJOR)
		tsk->maj_flt++;
	else
		tsk->min_flt++;
979 980 981 982 983 984 985 986 987 988 989

#ifdef CONFIG_X86_32
	/*
	 * Did it hit the DOS screen memory VA from vm86 mode?
	 */
	if (v8086_mode(regs)) {
		unsigned long bit = (address - 0xA0000) >> PAGE_SHIFT;
		if (bit < 32)
			tsk->thread.screen_bitmap |= 1 << bit;
	}
#endif
Linus Torvalds's avatar
Linus Torvalds committed
990 991
	up_read(&mm->mmap_sem);
}
992

993
DEFINE_SPINLOCK(pgd_lock);
994
LIST_HEAD(pgd_list);
995 996 997

void vmalloc_sync_all(void)
{
998 999
	unsigned long address;

1000
#ifdef CONFIG_X86_32
1001 1002 1003
	if (SHARED_KERNEL_PMD)
		return;

1004 1005 1006
	for (address = VMALLOC_START & PMD_MASK;
	     address >= TASK_SIZE && address < FIXADDR_TOP;
	     address += PMD_SIZE) {
1007 1008 1009 1010 1011 1012 1013 1014
		unsigned long flags;
		struct page *page;

		spin_lock_irqsave(&pgd_lock, flags);
		list_for_each_entry(page, &pgd_list, lru) {
			if (!vmalloc_sync_one(page_address(page),
					      address))
				break;
1015
		}
1016
		spin_unlock_irqrestore(&pgd_lock, flags);
1017 1018
	}
#else /* CONFIG_X86_64 */
1019 1020
	for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END;
	     address += PGDIR_SIZE) {
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
		const pgd_t *pgd_ref = pgd_offset_k(address);
		unsigned long flags;
		struct page *page;

		if (pgd_none(*pgd_ref))
			continue;
		spin_lock_irqsave(&pgd_lock, flags);
		list_for_each_entry(page, &pgd_list, lru) {
			pgd_t *pgd;
			pgd = (pgd_t *)page_address(page) + pgd_index(address);
			if (pgd_none(*pgd))
				set_pgd(pgd, *pgd_ref);
			else
				BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref));
1035
		}
1036
		spin_unlock_irqrestore(&pgd_lock, flags);
1037
	}
1038
#endif
1039
}