fault.c 14 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
11
12
13
/*
 *  linux/arch/arm/mm/fault.c
 *
 *  Copyright (C) 1995  Linus Torvalds
 *  Modifications for ARM processor (c) 1995-2004 Russell King
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/module.h>
#include <linux/signal.h>
#include <linux/mm.h>
14
#include <linux/hardirq.h>
Linus Torvalds's avatar
Linus Torvalds committed
15
#include <linux/init.h>
16
#include <linux/kprobes.h>
17
#include <linux/uaccess.h>
18
#include <linux/page-flags.h>
19
#include <linux/sched.h>
Russell King's avatar
Russell King committed
20
#include <linux/highmem.h>
21
#include <linux/perf_event.h>
Linus Torvalds's avatar
Linus Torvalds committed
22

23
#include <asm/exception.h>
Linus Torvalds's avatar
Linus Torvalds committed
24
#include <asm/pgtable.h>
25
26
#include <asm/system_misc.h>
#include <asm/system_info.h>
Linus Torvalds's avatar
Linus Torvalds committed
27
28
29
30
#include <asm/tlbflush.h>

#include "fault.h"

31
#ifdef CONFIG_MMU
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

#ifdef CONFIG_KPROBES
static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
{
	int ret = 0;

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

	return ret;
}
#else
static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
{
	return 0;
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
55
56
57
58
59
60
61
62
63
64
65
66
67
/*
 * This is useful to dump out the page tables associated with
 * 'addr' in mm 'mm'.
 */
void show_pte(struct mm_struct *mm, unsigned long addr)
{
	pgd_t *pgd;

	if (!mm)
		mm = &init_mm;

	printk(KERN_ALERT "pgd = %p\n", mm->pgd);
	pgd = pgd_offset(mm, addr);
68
69
	printk(KERN_ALERT "[%08lx] *pgd=%08llx",
			addr, (long long)pgd_val(*pgd));
Linus Torvalds's avatar
Linus Torvalds committed
70
71

	do {
72
		pud_t *pud;
Linus Torvalds's avatar
Linus Torvalds committed
73
74
75
76
77
78
79
80
81
82
83
		pmd_t *pmd;
		pte_t *pte;

		if (pgd_none(*pgd))
			break;

		if (pgd_bad(*pgd)) {
			printk("(bad)");
			break;
		}

84
85
		pud = pud_offset(pgd, addr);
		if (PTRS_PER_PUD != 1)
86
			printk(", *pud=%08llx", (long long)pud_val(*pud));
87
88
89
90
91
92
93
94
95
96

		if (pud_none(*pud))
			break;

		if (pud_bad(*pud)) {
			printk("(bad)");
			break;
		}

		pmd = pmd_offset(pud, addr);
97
		if (PTRS_PER_PMD != 1)
98
			printk(", *pmd=%08llx", (long long)pmd_val(*pmd));
Linus Torvalds's avatar
Linus Torvalds committed
99
100
101
102
103
104
105
106
107
108

		if (pmd_none(*pmd))
			break;

		if (pmd_bad(*pmd)) {
			printk("(bad)");
			break;
		}

		/* We must not map this if we have highmem enabled */
109
110
111
		if (PageHighMem(pfn_to_page(pmd_val(*pmd) >> PAGE_SHIFT)))
			break;

Linus Torvalds's avatar
Linus Torvalds committed
112
		pte = pte_offset_map(pmd, addr);
113
		printk(", *pte=%08llx", (long long)pte_val(*pte));
114
#ifndef CONFIG_ARM_LPAE
115
116
		printk(", *ppte=%08llx",
		       (long long)pte_val(pte[PTE_HWTABLE_PTRS]));
117
#endif
Linus Torvalds's avatar
Linus Torvalds committed
118
119
120
121
122
		pte_unmap(pte);
	} while(0);

	printk("\n");
}
123
124
125
126
#else					/* CONFIG_MMU */
void show_pte(struct mm_struct *mm, unsigned long addr)
{ }
#endif					/* CONFIG_MMU */
Linus Torvalds's avatar
Linus Torvalds committed
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

/*
 * Oops.  The kernel tried to access some page that wasn't present.
 */
static void
__do_kernel_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
		  struct pt_regs *regs)
{
	/*
	 * Are we prepared to handle this kernel fault?
	 */
	if (fixup_exception(regs))
		return;

	/*
	 * No handler, we'll have to terminate things with extreme prejudice.
	 */
	bust_spinlocks(1);
	printk(KERN_ALERT
		"Unable to handle kernel %s at virtual address %08lx\n",
		(addr < PAGE_SIZE) ? "NULL pointer dereference" :
		"paging request", addr);

	show_pte(mm, addr);
	die("Oops", regs, fsr);
	bust_spinlocks(0);
	do_exit(SIGKILL);
}

/*
 * Something tried to access memory that isn't in our memory map..
 * User mode accesses just cause a SIGSEGV
 */
static void
__do_user_fault(struct task_struct *tsk, unsigned long addr,
162
163
		unsigned int fsr, unsigned int sig, int code,
		struct pt_regs *regs)
Linus Torvalds's avatar
Linus Torvalds committed
164
165
166
167
168
{
	struct siginfo si;

#ifdef CONFIG_DEBUG_USER
	if (user_debug & UDBG_SEGV) {
169
170
		printk(KERN_DEBUG "%s: unhandled page fault (%d) at 0x%08lx, code 0x%03x\n",
		       tsk->comm, sig, addr, fsr);
Linus Torvalds's avatar
Linus Torvalds committed
171
172
173
174
175
176
177
178
		show_pte(tsk->mm, addr);
		show_regs(regs);
	}
#endif

	tsk->thread.address = addr;
	tsk->thread.error_code = fsr;
	tsk->thread.trap_no = 14;
179
	si.si_signo = sig;
Linus Torvalds's avatar
Linus Torvalds committed
180
181
182
	si.si_errno = 0;
	si.si_code = code;
	si.si_addr = (void __user *)addr;
183
	force_sig_info(sig, &si, tsk);
Linus Torvalds's avatar
Linus Torvalds committed
184
185
}

186
void do_bad_area(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
Linus Torvalds's avatar
Linus Torvalds committed
187
{
188
189
190
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->active_mm;

Linus Torvalds's avatar
Linus Torvalds committed
191
192
193
194
195
	/*
	 * If we are in kernel mode at this point, we
	 * have no context to handle this fault with.
	 */
	if (user_mode(regs))
196
		__do_user_fault(tsk, addr, fsr, SIGSEGV, SEGV_MAPERR, regs);
Linus Torvalds's avatar
Linus Torvalds committed
197
198
199
200
	else
		__do_kernel_fault(mm, addr, fsr, regs);
}

201
#ifdef CONFIG_MMU
202
203
#define VM_FAULT_BADMAP		0x010000
#define VM_FAULT_BADACCESS	0x020000
Linus Torvalds's avatar
Linus Torvalds committed
204

205
206
207
208
209
210
211
212
213
214
215
/*
 * Check that the permissions on the VMA allow for the fault which occurred.
 * If we encountered a write fault, we must have write permission, otherwise
 * we allow any permission.
 */
static inline bool access_error(unsigned int fsr, struct vm_area_struct *vma)
{
	unsigned int mask = VM_READ | VM_WRITE | VM_EXEC;

	if (fsr & FSR_WRITE)
		mask = VM_WRITE;
216
217
	if (fsr & FSR_LNX_PF)
		mask = VM_EXEC;
218
219
220
221
222

	return vma->vm_flags & mask ? false : true;
}

static int __kprobes
Linus Torvalds's avatar
Linus Torvalds committed
223
__do_page_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
224
		unsigned int flags, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
225
226
{
	struct vm_area_struct *vma;
227
	int fault;
Linus Torvalds's avatar
Linus Torvalds committed
228
229
230

	vma = find_vma(mm, addr);
	fault = VM_FAULT_BADMAP;
231
	if (unlikely(!vma))
Linus Torvalds's avatar
Linus Torvalds committed
232
		goto out;
233
	if (unlikely(vma->vm_start > addr))
Linus Torvalds's avatar
Linus Torvalds committed
234
235
236
237
238
239
240
		goto check_stack;

	/*
	 * Ok, we have a good vm_area for this
	 * memory access, so we can handle it.
	 */
good_area:
241
242
	if (access_error(fsr, vma)) {
		fault = VM_FAULT_BADACCESS;
Linus Torvalds's avatar
Linus Torvalds committed
243
		goto out;
244
	}
Linus Torvalds's avatar
Linus Torvalds committed
245

246
	return handle_mm_fault(mm, vma, addr & PAGE_MASK, flags);
Linus Torvalds's avatar
Linus Torvalds committed
247
248
249
250
251
252
253
254

check_stack:
	if (vma->vm_flags & VM_GROWSDOWN && !expand_stack(vma, addr))
		goto good_area;
out:
	return fault;
}

255
static int __kprobes
Linus Torvalds's avatar
Linus Torvalds committed
256
257
258
259
do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	struct task_struct *tsk;
	struct mm_struct *mm;
260
	int fault, sig, code;
261
262
263
	int write = fsr & FSR_WRITE;
	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE |
				(write ? FAULT_FLAG_WRITE : 0);
Linus Torvalds's avatar
Linus Torvalds committed
264

265
266
267
	if (notify_page_fault(regs, fsr))
		return 0;

Linus Torvalds's avatar
Linus Torvalds committed
268
269
270
	tsk = current;
	mm  = tsk->mm;

271
272
273
274
	/* Enable interrupts if they were enabled in the parent context. */
	if (interrupts_enabled(regs))
		local_irq_enable();

Linus Torvalds's avatar
Linus Torvalds committed
275
276
277
278
	/*
	 * If we're in an interrupt or have no user
	 * context, we must not take the fault..
	 */
279
	if (in_atomic() || !mm)
Linus Torvalds's avatar
Linus Torvalds committed
280
281
		goto no_context;

282
283
284
285
286
287
288
289
	/*
	 * As per x86, we may deadlock here.  However, since the kernel only
	 * validly references user space from well defined areas of the code,
	 * we can bug out early if this is from code which shouldn't.
	 */
	if (!down_read_trylock(&mm->mmap_sem)) {
		if (!user_mode(regs) && !search_exception_tables(regs->ARM_pc))
			goto no_context;
290
retry:
291
		down_read(&mm->mmap_sem);
292
293
294
295
296
297
298
	} else {
		/*
		 * The above down_read_trylock() might have succeeded in
		 * which case, we'll have missed the might_sleep() from
		 * down_read()
		 */
		might_sleep();
299
300
301
302
303
#ifdef CONFIG_DEBUG_VM
		if (!user_mode(regs) &&
		    !search_exception_tables(regs->ARM_pc))
			goto no_context;
#endif
304
305
	}

306
307
308
309
310
311
312
313
314
315
316
317
318
319
	fault = __do_page_fault(mm, addr, fsr, flags, tsk);

	/* If we need to retry but a fatal signal is pending, handle the
	 * signal first. We do not need to release the mmap_sem because
	 * it would already be released in __lock_page_or_retry in
	 * mm/filemap.c. */
	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
		return 0;

	/*
	 * Major/minor page fault accounting is only done on the
	 * initial attempt. If we go through a retry, it is extremely
	 * likely that the page will be found in page cache at that point.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
320

321
	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
	if (flags & FAULT_FLAG_ALLOW_RETRY) {
		if (fault & VM_FAULT_MAJOR) {
			tsk->maj_flt++;
			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
					regs, addr);
		} else {
			tsk->min_flt++;
			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
					regs, addr);
		}
		if (fault & VM_FAULT_RETRY) {
			/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
			* of starvation. */
			flags &= ~FAULT_FLAG_ALLOW_RETRY;
			goto retry;
		}
	}

	up_read(&mm->mmap_sem);
341

Linus Torvalds's avatar
Linus Torvalds committed
342
	/*
343
	 * Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
Linus Torvalds's avatar
Linus Torvalds committed
344
	 */
345
	if (likely(!(fault & (VM_FAULT_ERROR | VM_FAULT_BADMAP | VM_FAULT_BADACCESS))))
Linus Torvalds's avatar
Linus Torvalds committed
346
347
		return 0;

348
349
350
351
352
353
354
355
356
357
	if (fault & VM_FAULT_OOM) {
		/*
		 * We ran out of memory, call the OOM killer, and return to
		 * userspace (which will retry the fault, or kill us if we
		 * got oom-killed)
		 */
		pagefault_out_of_memory();
		return 0;
	}

Linus Torvalds's avatar
Linus Torvalds committed
358
359
360
361
362
363
364
	/*
	 * If we are in kernel mode at this point, we
	 * have no context to handle this fault with.
	 */
	if (!user_mode(regs))
		goto no_context;

Nick Piggin's avatar
Nick Piggin committed
365
	if (fault & VM_FAULT_SIGBUS) {
366
367
368
369
370
371
		/*
		 * We had some memory, but were unable to
		 * successfully fix up this page fault.
		 */
		sig = SIGBUS;
		code = BUS_ADRERR;
Nick Piggin's avatar
Nick Piggin committed
372
	} else {
373
374
375
376
377
378
379
		/*
		 * Something tried to access memory that
		 * isn't in our memory map..
		 */
		sig = SIGSEGV;
		code = fault == VM_FAULT_BADACCESS ?
			SEGV_ACCERR : SEGV_MAPERR;
Linus Torvalds's avatar
Linus Torvalds committed
380
381
	}

382
383
	__do_user_fault(tsk, addr, fsr, sig, code, regs);
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
384
385
386
387
388

no_context:
	__do_kernel_fault(mm, addr, fsr, regs);
	return 0;
}
389
390
391
392
393
394
395
#else					/* CONFIG_MMU */
static int
do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	return 0;
}
#endif					/* CONFIG_MMU */
Linus Torvalds's avatar
Linus Torvalds committed
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413

/*
 * First Level Translation Fault Handler
 *
 * We enter here because the first level page table doesn't contain
 * a valid entry for the address.
 *
 * If the address is in kernel space (>= TASK_SIZE), then we are
 * probably faulting in the vmalloc() area.
 *
 * If the init_task's first level page tables contains the relevant
 * entry, we copy the it to this task.  If not, we send the process
 * a signal, fixup the exception, or oops the kernel.
 *
 * 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.
 */
414
#ifdef CONFIG_MMU
415
static int __kprobes
Linus Torvalds's avatar
Linus Torvalds committed
416
417
418
419
420
do_translation_fault(unsigned long addr, unsigned int fsr,
		     struct pt_regs *regs)
{
	unsigned int index;
	pgd_t *pgd, *pgd_k;
421
	pud_t *pud, *pud_k;
Linus Torvalds's avatar
Linus Torvalds committed
422
423
424
425
426
	pmd_t *pmd, *pmd_k;

	if (addr < TASK_SIZE)
		return do_page_fault(addr, fsr, regs);

427
428
429
	if (user_mode(regs))
		goto bad_area;

Linus Torvalds's avatar
Linus Torvalds committed
430
431
432
433
434
435
436
437
438
439
440
441
442
	index = pgd_index(addr);

	/*
	 * FIXME: CP15 C1 is write only on ARMv3 architectures.
	 */
	pgd = cpu_get_pgd() + index;
	pgd_k = init_mm.pgd + index;

	if (pgd_none(*pgd_k))
		goto bad_area;
	if (!pgd_present(*pgd))
		set_pgd(pgd, *pgd_k);

443
444
445
446
447
448
449
450
451
452
	pud = pud_offset(pgd, addr);
	pud_k = pud_offset(pgd_k, addr);

	if (pud_none(*pud_k))
		goto bad_area;
	if (!pud_present(*pud))
		set_pud(pud, *pud_k);

	pmd = pmd_offset(pud, addr);
	pmd_k = pmd_offset(pud_k, addr);
Linus Torvalds's avatar
Linus Torvalds committed
453

454
455
456
457
458
459
#ifdef CONFIG_ARM_LPAE
	/*
	 * Only one hardware entry per PMD with LPAE.
	 */
	index = 0;
#else
460
461
462
463
464
465
466
467
468
	/*
	 * On ARM one Linux PGD entry contains two hardware entries (see page
	 * tables layout in pgtable.h). We normally guarantee that we always
	 * fill both L1 entries. But create_mapping() doesn't follow the rule.
	 * It can create inidividual L1 entries, so here we have to call
	 * pmd_none() check for the entry really corresponded to address, not
	 * for the first of pair.
	 */
	index = (addr >> SECTION_SHIFT) & 1;
469
#endif
470
	if (pmd_none(pmd_k[index]))
Linus Torvalds's avatar
Linus Torvalds committed
471
472
473
474
475
476
		goto bad_area;

	copy_pmd(pmd, pmd_k);
	return 0;

bad_area:
477
	do_bad_area(addr, fsr, regs);
Linus Torvalds's avatar
Linus Torvalds committed
478
479
	return 0;
}
480
481
482
483
484
485
486
487
#else					/* CONFIG_MMU */
static int
do_translation_fault(unsigned long addr, unsigned int fsr,
		     struct pt_regs *regs)
{
	return 0;
}
#endif					/* CONFIG_MMU */
Linus Torvalds's avatar
Linus Torvalds committed
488
489
490
491
492
493
494
495

/*
 * Some section permission faults need to be handled gracefully.
 * They can happen due to a __{get,put}_user during an oops.
 */
static int
do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
496
	do_bad_area(addr, fsr, regs);
Linus Torvalds's avatar
Linus Torvalds committed
497
498
499
500
501
502
503
504
505
506
507
508
	return 0;
}

/*
 * This abort handler always returns "fault".
 */
static int
do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	return 1;
}

509
struct fsr_info {
Linus Torvalds's avatar
Linus Torvalds committed
510
511
	int	(*fn)(unsigned long addr, unsigned int fsr, struct pt_regs *regs);
	int	sig;
512
	int	code;
Linus Torvalds's avatar
Linus Torvalds committed
513
514
515
	const char *name;
};

516
/* FSR definition */
517
518
519
#ifdef CONFIG_ARM_LPAE
#include "fsr-3level.c"
#else
520
#include "fsr-2level.c"
521
#endif
522

Linus Torvalds's avatar
Linus Torvalds committed
523
524
void __init
hook_fault_code(int nr, int (*fn)(unsigned long, unsigned int, struct pt_regs *),
525
		int sig, int code, const char *name)
Linus Torvalds's avatar
Linus Torvalds committed
526
{
527
528
529
530
531
532
533
	if (nr < 0 || nr >= ARRAY_SIZE(fsr_info))
		BUG();

	fsr_info[nr].fn   = fn;
	fsr_info[nr].sig  = sig;
	fsr_info[nr].code = code;
	fsr_info[nr].name = name;
Linus Torvalds's avatar
Linus Torvalds committed
534
535
536
537
538
}

/*
 * Dispatch a data abort to the relevant handler.
 */
539
asmlinkage void __exception
Linus Torvalds's avatar
Linus Torvalds committed
540
541
do_DataAbort(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
542
	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
543
	struct siginfo info;
Linus Torvalds's avatar
Linus Torvalds committed
544

545
	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
Linus Torvalds's avatar
Linus Torvalds committed
546
547
548
549
		return;

	printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
		inf->name, fsr, addr);
550
551
552
553
554

	info.si_signo = inf->sig;
	info.si_errno = 0;
	info.si_code  = inf->code;
	info.si_addr  = (void __user *)addr;
555
	arm_notify_die("", regs, &info, fsr, 0);
Linus Torvalds's avatar
Linus Torvalds committed
556
557
}

558
559
560
561
562
563
564
565
566
567
568
569
570
void __init
hook_ifault_code(int nr, int (*fn)(unsigned long, unsigned int, struct pt_regs *),
		 int sig, int code, const char *name)
{
	if (nr < 0 || nr >= ARRAY_SIZE(ifsr_info))
		BUG();

	ifsr_info[nr].fn   = fn;
	ifsr_info[nr].sig  = sig;
	ifsr_info[nr].code = code;
	ifsr_info[nr].name = name;
}

571
asmlinkage void __exception
572
do_PrefetchAbort(unsigned long addr, unsigned int ifsr, struct pt_regs *regs)
Linus Torvalds's avatar
Linus Torvalds committed
573
{
574
575
576
577
578
579
580
581
582
583
584
585
586
587
	const struct fsr_info *inf = ifsr_info + fsr_fs(ifsr);
	struct siginfo info;

	if (!inf->fn(addr, ifsr | FSR_LNX_PF, regs))
		return;

	printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
		inf->name, ifsr, addr);

	info.si_signo = inf->sig;
	info.si_errno = 0;
	info.si_code  = inf->code;
	info.si_addr  = (void __user *)addr;
	arm_notify_die("", regs, &info, ifsr, 0);
Linus Torvalds's avatar
Linus Torvalds committed
588
589
}

590
#ifndef CONFIG_ARM_LPAE
591
592
593
594
595
596
597
static int __init exceptions_init(void)
{
	if (cpu_architecture() >= CPU_ARCH_ARMv6) {
		hook_fault_code(4, do_translation_fault, SIGSEGV, SEGV_MAPERR,
				"I-cache maintenance fault");
	}

598
599
600
601
602
603
604
605
606
607
608
	if (cpu_architecture() >= CPU_ARCH_ARMv7) {
		/*
		 * TODO: Access flag faults introduced in ARMv6K.
		 * Runtime check for 'K' extension is needed
		 */
		hook_fault_code(3, do_bad, SIGSEGV, SEGV_MAPERR,
				"section access flag fault");
		hook_fault_code(6, do_bad, SIGSEGV, SEGV_MAPERR,
				"section access flag fault");
	}

609
610
611
612
	return 0;
}

arch_initcall(exceptions_init);
613
#endif