fault_32.c 17.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
/*
 *  Copyright (C) 1995  Linus Torvalds
 */

#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>
#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/highmem.h>
20
#include <linux/bootmem.h>		/* for max_low_pfn */
21
#include <linux/vmalloc.h>
L
Linus Torvalds 已提交
22
#include <linux/module.h>
23
#include <linux/kprobes.h>
24
#include <linux/uaccess.h>
25
#include <linux/kdebug.h>
L
Linus Torvalds 已提交
26 27 28

#include <asm/system.h>
#include <asm/desc.h>
29
#include <asm/segment.h>
L
Linus Torvalds 已提交
30

31 32 33 34 35 36 37 38
/*
 * 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
 */
39
#define PF_PROT		(1<<0)
40
#define PF_WRITE	(1<<1)
41 42
#define PF_USER		(1<<2)
#define PF_RSVD		(1<<3)
43 44
#define PF_INSTR	(1<<4)

45
static inline int notify_page_fault(struct pt_regs *regs)
46
{
47
#ifdef CONFIG_KPROBES
48 49 50 51 52 53 54 55 56
	int ret = 0;

	/* kprobe_running() needs smp_processor_id() */
	if (!user_mode_vm(regs)) {
		preempt_disable();
		if (kprobe_running() && kprobe_fault_handler(regs, 14))
			ret = 1;
		preempt_enable();
	}
57

58 59 60 61
	return ret;
#else
	return 0;
#endif
62
}
63

64
/*
65
 * X86_32
L
Linus Torvalds 已提交
66 67
 * Sometimes AMD Athlon/Opteron CPUs report invalid exceptions on prefetch.
 * Check that here and ignore it.
68 69 70 71 72 73
 *
 * X86_64
 * Sometimes the CPU reports invalid exceptions on prefetch.
 * Check that here and ignore it.
 *
 * Opcode checker based on code by Richard Brunner
L
Linus Torvalds 已提交
74
 */
75 76
static int is_prefetch(struct pt_regs *regs, unsigned long addr,
		       unsigned long error_code)
77
{
78
	unsigned char *instr;
L
Linus Torvalds 已提交
79
	int scan_more = 1;
80
	int prefetch = 0;
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96
	unsigned char *max_instr;

#ifdef CONFIG_X86_32
	if (unlikely(boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&
		     boot_cpu_data.x86 >= 6)) {
		/* Catch an obscure case of prefetch inside an NX page. */
		if (nx_enabled && (error_code & PF_INSTR))
			return 0;
	} else {
		return 0;
	}
#else
	/* If it was a exec fault ignore */
	if (error_code & PF_INSTR)
		return 0;
#endif
L
Linus Torvalds 已提交
97

98
	instr = (unsigned char *)convert_ip_to_linear(current, regs);
99 100 101 102 103 104
	max_instr = instr + 15;

	if (user_mode(regs) && instr >= (unsigned char *)TASK_SIZE)
		return 0;

	while (scan_more && instr < max_instr) {
L
Linus Torvalds 已提交
105 106 107 108
		unsigned char opcode;
		unsigned char instr_hi;
		unsigned char instr_lo;

109
		if (probe_kernel_address(instr, opcode))
110
			break;
L
Linus Torvalds 已提交
111

112 113
		instr_hi = opcode & 0xf0;
		instr_lo = opcode & 0x0f;
L
Linus Torvalds 已提交
114 115
		instr++;

116
		switch (instr_hi) {
L
Linus Torvalds 已提交
117 118
		case 0x20:
		case 0x30:
119 120 121 122 123 124
			/*
			 * 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
			 */
L
Linus Torvalds 已提交
125 126
			scan_more = ((instr_lo & 7) == 0x6);
			break;
127 128 129 130 131 132 133 134 135 136 137 138
#ifdef CONFIG_X86_64
		case 0x40:
			/*
			 * 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.
			 */
			scan_more = (!user_mode(regs)) || (regs->cs == __USER_CS);
			break;
#endif
L
Linus Torvalds 已提交
139 140 141
		case 0x60:
			/* 0x64 thru 0x67 are valid prefixes in all modes. */
			scan_more = (instr_lo & 0xC) == 0x4;
142
			break;
L
Linus Torvalds 已提交
143
		case 0xF0:
144
			/* 0xF0, 0xF2, 0xF3 are valid prefixes in all modes. */
L
Linus Torvalds 已提交
145
			scan_more = !instr_lo || (instr_lo>>1) == 1;
146
			break;
L
Linus Torvalds 已提交
147 148 149
		case 0x00:
			/* Prefetch instruction is 0x0F0D or 0x0F18 */
			scan_more = 0;
150

151
			if (probe_kernel_address(instr, opcode))
L
Linus Torvalds 已提交
152 153 154
				break;
			prefetch = (instr_lo == 0xF) &&
				(opcode == 0x0D || opcode == 0x18);
155
			break;
L
Linus Torvalds 已提交
156 157 158
		default:
			scan_more = 0;
			break;
159
		}
L
Linus Torvalds 已提交
160 161 162 163
	}
	return prefetch;
}

164
static void force_sig_info_fault(int si_signo, int si_code,
165 166 167 168 169 170 171 172 173 174 175
	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);
}

176
void do_invalid_op(struct pt_regs *, unsigned long);
L
Linus Torvalds 已提交
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
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;
206
	if (!pmd_present(*pmd)) {
207
		set_pmd(pmd, *pmd_k);
208 209
		arch_flush_lazy_mmu_mode();
	} else
210 211 212 213
		BUG_ON(pmd_page(*pmd) != pmd_page(*pmd_k));
	return pmd_k;
}

214 215 216 217 218 219
#ifdef CONFIG_X86_64
static const char errata93_warning[] =
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";
220
#endif
221 222 223 224 225 226 227

/* Workaround for K8 erratum #93 & buggy BIOS.
   BIOS SMM functions are required to use a specific workaround
   to avoid corruption of the 64bit RIP register on C stepping K8.
   A lot of BIOS that didn't get tested properly miss this.
   The OS sees this as a page fault with the upper 32bits of RIP cleared.
   Try to work around it here.
228 229 230
   Note we only handle faults in kernel here.
   Does nothing for X86_32
 */
231 232
static int is_errata93(struct pt_regs *regs, unsigned long address)
{
233
#ifdef CONFIG_X86_64
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
	static int warned;
	if (address != regs->ip)
		return 0;
	if ((address >> 32) != 0)
		return 0;
	address |= 0xffffffffUL << 32;
	if ((address >= (u64)_stext && address <= (u64)_etext) ||
	    (address >= MODULES_VADDR && address <= MODULES_END)) {
		if (!warned) {
			printk(errata93_warning);
			warned = 1;
		}
		regs->ip = address;
		return 1;
	}
249
#endif
250 251
	return 0;
}
252

253

254 255 256 257 258 259 260
/*
 * Handle a fault on the vmalloc or module mapping area
 *
 * This assumes no large pages in there.
 */
static inline int vmalloc_fault(unsigned long address)
{
261
#ifdef CONFIG_X86_32
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279
	unsigned long pgd_paddr;
	pmd_t *pmd_k;
	pte_t *pte_k;
	/*
	 * 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;
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
#else
	pgd_t *pgd, *pgd_ref;
	pud_t *pud, *pud_ref;
	pmd_t *pmd, *pmd_ref;
	pte_t *pte, *pte_ref;

	/* Copy kernel mappings over when needed. This can also
	   happen within a race in page table update. In the later
	   case just flush. */

	pgd = pgd_offset(current->mm ?: &init_mm, address);
	pgd_ref = pgd_offset_k(address);
	if (pgd_none(*pgd_ref))
		return -1;
	if (pgd_none(*pgd))
		set_pgd(pgd, *pgd_ref);
	else
		BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref));

	/* 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;
	if (pud_none(*pud) || pud_page_vaddr(*pud) != pud_page_vaddr(*pud_ref))
		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);
	/* 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))
		BUG();
	return 0;
#endif
325 326
}

327 328
int show_unhandled_signals = 1;

L
Linus Torvalds 已提交
329 330 331 332 333
/*
 * This routine handles page faults.  It determines the address,
 * and the problem, and then passes it off to one of the appropriate
 * routines.
 */
334
void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
L
Linus Torvalds 已提交
335 336 337
{
	struct task_struct *tsk;
	struct mm_struct *mm;
338
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
339
	unsigned long address;
340
	int write, si_code;
N
Nick Piggin 已提交
341
	int fault;
L
Linus Torvalds 已提交
342

P
Peter Zijlstra 已提交
343 344 345 346 347
	/*
	 * We can fault from pretty much anywhere, with unknown IRQ state.
	 */
	trace_hardirqs_fixup();

348 349 350 351
	tsk = current;
	mm = tsk->mm;
	prefetchw(&mm->mmap_sem);

L
Linus Torvalds 已提交
352
	/* get the address */
353
	address = read_cr2();
L
Linus Torvalds 已提交
354

355
	si_code = SEGV_MAPERR;
L
Linus Torvalds 已提交
356

357 358 359
	if (notify_page_fault(regs))
		return;

L
Linus Torvalds 已提交
360 361 362 363 364 365 366 367 368 369 370
	/*
	 * 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
371
	 * protection error (error_code & 9) == 0.
L
Linus Torvalds 已提交
372
	 */
373
	if (unlikely(address >= TASK_SIZE)) {
374 375
		if (!(error_code & (PF_RSVD|PF_USER|PF_PROT)) &&
		    vmalloc_fault(address) >= 0)
376 377
			return;
		/*
L
Linus Torvalds 已提交
378 379 380 381
		 * Don't take the mm semaphore here. If we fixup a prefetch
		 * fault we could otherwise deadlock.
		 */
		goto bad_area_nosemaphore;
382 383 384 385
	}

	/* It's safe to allow irq's after cr2 has been saved and the vmalloc
	   fault has been handled. */
386
	if (regs->flags & (X86_EFLAGS_IF|VM_MASK))
387
		local_irq_enable();
L
Linus Torvalds 已提交
388 389 390

	/*
	 * If we're in an interrupt, have no user context or are running in an
391
	 * atomic region then we must not take the fault.
L
Linus Torvalds 已提交
392 393 394 395 396 397
	 */
	if (in_atomic() || !mm)
		goto bad_area_nosemaphore;

	/* When running in the kernel we expect faults to occur only to
	 * addresses in user space.  All other faults represent errors in the
S
Simon Arlott 已提交
398
	 * kernel and should generate an OOPS.  Unfortunately, in the case of an
A
Adrian Bunk 已提交
399
	 * erroneous fault occurring in a code path which already holds mmap_sem
L
Linus Torvalds 已提交
400 401 402 403 404 405
	 * 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
S
Simon Arlott 已提交
406
	 * the source reference check when there is a possibility of a deadlock.
L
Linus Torvalds 已提交
407 408 409 410 411
	 * 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.
	 */
	if (!down_read_trylock(&mm->mmap_sem)) {
412
		if ((error_code & PF_USER) == 0 &&
413
		    !search_exception_tables(regs->ip))
L
Linus Torvalds 已提交
414 415 416 417 418 419 420 421 422 423 424
			goto bad_area_nosemaphore;
		down_read(&mm->mmap_sem);
	}

	vma = find_vma(mm, address);
	if (!vma)
		goto bad_area;
	if (vma->vm_start <= address)
		goto good_area;
	if (!(vma->vm_flags & VM_GROWSDOWN))
		goto bad_area;
425
	if (error_code & PF_USER) {
L
Linus Torvalds 已提交
426
		/*
427
		 * Accessing the stack below %sp is always a bug.
428 429
		 * The large cushion allows instructions like enter
		 * and pusha to work.  ("enter $65535,$31" pushes
430
		 * 32 pointers and then decrements %sp by 65535.)
L
Linus Torvalds 已提交
431
		 */
432
		if (address + 65536 + 32 * sizeof(unsigned long) < regs->sp)
L
Linus Torvalds 已提交
433 434 435 436 437 438 439 440 441
			goto bad_area;
	}
	if (expand_stack(vma, address))
		goto bad_area;
/*
 * Ok, we have a good vm_area for this memory access, so
 * we can handle it..
 */
good_area:
442
	si_code = SEGV_ACCERR;
L
Linus Torvalds 已提交
443
	write = 0;
444 445 446 447 448 449 450 451 452 453 454 455
	switch (error_code & (PF_PROT|PF_WRITE)) {
	default:	/* 3: write, present */
		/* fall through */
	case PF_WRITE:		/* write, not present */
		if (!(vma->vm_flags & VM_WRITE))
			goto bad_area;
		write++;
		break;
	case PF_PROT:		/* read, present */
		goto bad_area;
	case 0:			/* read, not present */
		if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))
L
Linus Torvalds 已提交
456 457 458 459 460 461 462 463 464
			goto bad_area;
	}

 survive:
	/*
	 * If for any reason at all we couldn't handle the fault,
	 * make sure we exit gracefully rather than endlessly redo
	 * the fault.
	 */
N
Nick Piggin 已提交
465 466 467
	fault = handle_mm_fault(mm, vma, address, write);
	if (unlikely(fault & VM_FAULT_ERROR)) {
		if (fault & VM_FAULT_OOM)
L
Linus Torvalds 已提交
468
			goto out_of_memory;
N
Nick Piggin 已提交
469 470 471
		else if (fault & VM_FAULT_SIGBUS)
			goto do_sigbus;
		BUG();
L
Linus Torvalds 已提交
472
	}
N
Nick Piggin 已提交
473 474 475 476
	if (fault & VM_FAULT_MAJOR)
		tsk->maj_flt++;
	else
		tsk->min_flt++;
L
Linus Torvalds 已提交
477 478 479 480

	/*
	 * Did it hit the DOS screen memory VA from vm86 mode?
	 */
481
	if (regs->flags & VM_MASK) {
L
Linus Torvalds 已提交
482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
		unsigned long bit = (address - 0xA0000) >> PAGE_SHIFT;
		if (bit < 32)
			tsk->thread.screen_bitmap |= 1 << bit;
	}
	up_read(&mm->mmap_sem);
	return;

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

bad_area_nosemaphore:
	/* User mode accesses just cause a SIGSEGV */
498
	if (error_code & PF_USER) {
499 500 501 502 503
		/*
		 * It's possible to have interrupts off here.
		 */
		local_irq_enable();

504 505
		/*
		 * Valid to do another page fault here because this one came
L
Linus Torvalds 已提交
506 507 508 509 510
		 * from user space.
		 */
		if (is_prefetch(regs, address, error_code))
			return;

511 512
		if (show_unhandled_signals && unhandled_signal(tsk, SIGSEGV) &&
		    printk_ratelimit()) {
513 514
			printk(
#ifdef CONFIG_X86_32
515
			"%s%s[%d]: segfault at %lx ip %08lx sp %08lx error %lx",
516
#else
517
			"%s%s[%d]: segfault at %lx ip %lx sp %lx error %lx",
518 519 520 521
#endif
			task_pid_nr(tsk) > 1 ? KERN_INFO : KERN_EMERG,
			tsk->comm, task_pid_nr(tsk), address, regs->ip,
			regs->sp, error_code);
522 523
			print_vma_addr(" in ", regs->ip);
			printk("\n");
524
		}
L
Linus Torvalds 已提交
525 526 527 528
		tsk->thread.cr2 = address;
		/* Kernel addresses are always protection faults */
		tsk->thread.error_code = error_code | (address >= TASK_SIZE);
		tsk->thread.trap_no = 14;
529
		force_sig_info_fault(SIGSEGV, si_code, address, tsk);
L
Linus Torvalds 已提交
530 531 532 533 534 535 536 537 538
		return;
	}

#ifdef CONFIG_X86_F00F_BUG
	/*
	 * Pentium F0 0F C7 C8 bug workaround.
	 */
	if (boot_cpu_data.f00f_bug) {
		unsigned long nr;
539

L
Linus Torvalds 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552 553
		nr = (address - idt_descr.address) >> 3;

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

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

554
	/*
L
Linus Torvalds 已提交
555
	 * Valid to do another page fault here, because if this fault
556
	 * had been triggered by is_prefetch fixup_exception would have
L
Linus Torvalds 已提交
557 558
	 * handled it.
	 */
559 560
	if (is_prefetch(regs, address, error_code))
		return;
L
Linus Torvalds 已提交
561

562 563 564
	if (is_errata93(regs, address))
		return;

L
Linus Torvalds 已提交
565 566 567 568 569 570 571
/*
 * Oops. The kernel tried to access some bad page. We'll have to
 * terminate things with extreme prejudice.
 */

	bust_spinlocks(1);

572
	if (oops_may_print()) {
573 574 575
		__typeof__(pte_val(__pte(0))) page;

#ifdef CONFIG_X86_PAE
576
		if (error_code & PF_INSTR) {
577 578 579 580 581 582 583
			pte_t *pte = lookup_address(address);

			if (pte && pte_present(*pte) && !pte_exec_kernel(*pte))
				printk(KERN_CRIT "kernel tried to execute "
					"NX-protected page - exploit attempt? "
					"(uid: %d)\n", current->uid);
		}
584
#endif
585 586 587 588 589 590
		if (address < PAGE_SIZE)
			printk(KERN_ALERT "BUG: unable to handle kernel NULL "
					"pointer dereference");
		else
			printk(KERN_ALERT "BUG: unable to handle kernel paging"
					" request");
591
		printk(" at virtual address %08lx\n", address);
592
		printk(KERN_ALERT "printing ip: %08lx ", regs->ip);
593 594 595 596

		page = read_cr3();
		page = ((__typeof__(page) *) __va(page))[address >> PGDIR_SHIFT];
#ifdef CONFIG_X86_PAE
P
Pavel Emelyanov 已提交
597
		printk("*pdpt = %016Lx ", page);
598 599 600 601 602
		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)];
603
			printk(KERN_CONT "*pde = %016Lx ", page);
604 605 606
			page &= ~_PAGE_NX;
		}
#else
P
Pavel Emelyanov 已提交
607
		printk("*pde = %08lx ", page);
L
Linus Torvalds 已提交
608
#endif
609 610 611 612 613 614 615 616

		/*
		 * 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
617 618
		    && (page & _PAGE_PRESENT)
		    && !(page & _PAGE_PSE)) {
619 620 621
			page &= PAGE_MASK;
			page = ((__typeof__(page) *) __va(page))[(address >> PAGE_SHIFT)
			                                         & (PTRS_PER_PTE - 1)];
P
Pavel Emelyanov 已提交
622
			printk("*pte = %0*Lx ", sizeof(page)*2, (u64)page);
623
		}
P
Pavel Emelyanov 已提交
624 625

		printk("\n");
626 627
	}

628 629 630
	tsk->thread.cr2 = address;
	tsk->thread.trap_no = 14;
	tsk->thread.error_code = error_code;
L
Linus Torvalds 已提交
631 632 633 634 635 636 637 638 639 640
	die("Oops", regs, error_code);
	bust_spinlocks(0);
	do_exit(SIGKILL);

/*
 * We ran out of memory, or some other thing happened to us that made
 * us unable to handle the page fault gracefully.
 */
out_of_memory:
	up_read(&mm->mmap_sem);
641
	if (is_global_init(tsk)) {
L
Linus Torvalds 已提交
642 643 644 645 646
		yield();
		down_read(&mm->mmap_sem);
		goto survive;
	}
	printk("VM: killing process %s\n", tsk->comm);
647
	if (error_code & PF_USER)
648
		do_group_exit(SIGKILL);
L
Linus Torvalds 已提交
649 650 651 652 653 654
	goto no_context;

do_sigbus:
	up_read(&mm->mmap_sem);

	/* Kernel mode? Handle exceptions or die */
655
	if (!(error_code & PF_USER))
L
Linus Torvalds 已提交
656 657 658 659 660 661 662 663 664
		goto no_context;

	/* User space => ok to do another page fault */
	if (is_prefetch(regs, address, error_code))
		return;

	tsk->thread.cr2 = address;
	tsk->thread.error_code = error_code;
	tsk->thread.trap_no = 14;
665
	force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk);
666
}
L
Linus Torvalds 已提交
667

668 669 670 671 672 673 674 675 676 677 678
void vmalloc_sync_all(void)
{
	/*
	 * Note that races in the updates of insync and start aren't
	 * problematic: insync can only get set bits added, and updates to
	 * start are only improving performance (without affecting correctness
	 * if undone).
	 */
	static DECLARE_BITMAP(insync, PTRS_PER_PGD);
	static unsigned long start = TASK_SIZE;
	unsigned long address;
L
Linus Torvalds 已提交
679

680 681 682
	if (SHARED_KERNEL_PMD)
		return;

683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702
	BUILD_BUG_ON(TASK_SIZE & ~PGDIR_MASK);
	for (address = start; address >= TASK_SIZE; address += PGDIR_SIZE) {
		if (!test_bit(pgd_index(address), insync)) {
			unsigned long flags;
			struct page *page;

			spin_lock_irqsave(&pgd_lock, flags);
			for (page = pgd_list; page; page =
					(struct page *)page->index)
				if (!vmalloc_sync_one(page_address(page),
								address)) {
					BUG_ON(page != pgd_list);
					break;
				}
			spin_unlock_irqrestore(&pgd_lock, flags);
			if (!page)
				set_bit(pgd_index(address), insync);
		}
		if (address == start && test_bit(pgd_index(address), insync))
			start = address + PGDIR_SIZE;
L
Linus Torvalds 已提交
703 704
	}
}