[deliverable/linux.git] / arch / arm / mm / fault.c

/*
 *  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>
#include <linux/hardirq.h>
#include <linux/init.h>
#include <linux/kprobes.h>
#include <linux/uaccess.h>
#include <linux/page-flags.h>
#include <linux/sched.h>
#include <linux/highmem.h>
#include <linux/perf_event.h>

#include <asm/exception.h>
#include <asm/system.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>

#include "fault.h"

#ifdef CONFIG_MMU

#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

/*
 * 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);
	printk(KERN_ALERT "[%08lx] *pgd=%08llx",
			addr, (long long)pgd_val(*pgd));

	do {
		pud_t *pud;
		pmd_t *pmd;
		pte_t *pte;

		if (pgd_none(*pgd))
			break;

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

		pud = pud_offset(pgd, addr);
		if (PTRS_PER_PUD != 1)
			printk(", *pud=%08llx", (long long)pud_val(*pud));

		if (pud_none(*pud))
			break;

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

		pmd = pmd_offset(pud, addr);
		if (PTRS_PER_PMD != 1)
			printk(", *pmd=%08llx", (long long)pmd_val(*pmd));

		if (pmd_none(*pmd))
			break;

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

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

		pte = pte_offset_map(pmd, addr);
		printk(", *pte=%08llx", (long long)pte_val(*pte));
#ifndef CONFIG_ARM_LPAE
		printk(", *ppte=%08llx",
		       (long long)pte_val(pte[PTE_HWTABLE_PTRS]));
#endif
		pte_unmap(pte);
	} while(0);

	printk("\n");
}
#else					/* CONFIG_MMU */
void show_pte(struct mm_struct *mm, unsigned long addr)
{ }
#endif					/* CONFIG_MMU */

/*
 * 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,
		unsigned int fsr, unsigned int sig, int code,
		struct pt_regs *regs)
{
	struct siginfo si;

#ifdef CONFIG_DEBUG_USER
	if (((user_debug & UDBG_SEGV) && (sig == SIGSEGV)) ||
	    ((user_debug & UDBG_BUS)  && (sig == SIGBUS))) {
		printk(KERN_DEBUG "%s: unhandled page fault (%d) at 0x%08lx, code 0x%03x\n",
		       tsk->comm, sig, addr, fsr);
		show_pte(tsk->mm, addr);
		show_regs(regs);
	}
#endif

	tsk->thread.address = addr;
	tsk->thread.error_code = fsr;
	tsk->thread.trap_no = 14;
	si.si_signo = sig;
	si.si_errno = 0;
	si.si_code = code;
	si.si_addr = (void __user *)addr;
	force_sig_info(sig, &si, tsk);
}

void do_bad_area(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->active_mm;

	/*
	 * If we are in kernel mode at this point, we
	 * have no context to handle this fault with.
	 */
	if (user_mode(regs))
		__do_user_fault(tsk, addr, fsr, SIGSEGV, SEGV_MAPERR, regs);
	else
		__do_kernel_fault(mm, addr, fsr, regs);
}

#ifdef CONFIG_MMU
#define VM_FAULT_BADMAP		0x010000
#define VM_FAULT_BADACCESS	0x020000

/*
 * 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;
	if (fsr & FSR_LNX_PF)
		mask = VM_EXEC;

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

static int __kprobes
__do_page_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
		unsigned int flags, struct task_struct *tsk)
{
	struct vm_area_struct *vma;
	int fault;

	vma = find_vma(mm, addr);
	fault = VM_FAULT_BADMAP;
	if (unlikely(!vma))
		goto out;
	if (unlikely(vma->vm_start > addr))
		goto check_stack;

	/*
	 * Ok, we have a good vm_area for this
	 * memory access, so we can handle it.
	 */
good_area:
	if (access_error(fsr, vma)) {
		fault = VM_FAULT_BADACCESS;
		goto out;
	}

	return handle_mm_fault(mm, vma, addr & PAGE_MASK, flags);

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

static int __kprobes
do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	struct task_struct *tsk;
	struct mm_struct *mm;
	int fault, sig, code;
	int write = fsr & FSR_WRITE;
	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE |
				(write ? FAULT_FLAG_WRITE : 0);

	if (notify_page_fault(regs, fsr))
		return 0;

	tsk = current;
	mm  = tsk->mm;

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

	/*
	 * If we're in an interrupt or have no user
	 * context, we must not take the fault..
	 */
	if (in_atomic() || !mm)
		goto no_context;

	/*
	 * 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;
retry:
		down_read(&mm->mmap_sem);
	} else {
		/*
		 * The above down_read_trylock() might have succeeded in
		 * which case, we'll have missed the might_sleep() from
		 * down_read()
		 */
		might_sleep();
#ifdef CONFIG_DEBUG_VM
		if (!user_mode(regs) &&
		    !search_exception_tables(regs->ARM_pc))
			goto no_context;
#endif
	}

	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.
	 */

	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
	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);

	/*
	 * Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
	 */
	if (likely(!(fault & (VM_FAULT_ERROR | VM_FAULT_BADMAP | VM_FAULT_BADACCESS))))
		return 0;

	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;
	}

	/*
	 * 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;

	if (fault & VM_FAULT_SIGBUS) {
		/*
		 * We had some memory, but were unable to
		 * successfully fix up this page fault.
		 */
		sig = SIGBUS;
		code = BUS_ADRERR;
	} else {
		/*
		 * Something tried to access memory that
		 * isn't in our memory map..
		 */
		sig = SIGSEGV;
		code = fault == VM_FAULT_BADACCESS ?
			SEGV_ACCERR : SEGV_MAPERR;
	}

	__do_user_fault(tsk, addr, fsr, sig, code, regs);
	return 0;

no_context:
	__do_kernel_fault(mm, addr, fsr, regs);
	return 0;
}
#else					/* CONFIG_MMU */
static int
do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	return 0;
}
#endif					/* CONFIG_MMU */

/*
 * 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.
 */
#ifdef CONFIG_MMU
static int __kprobes
do_translation_fault(unsigned long addr, unsigned int fsr,
		     struct pt_regs *regs)
{
	unsigned int index;
	pgd_t *pgd, *pgd_k;
	pud_t *pud, *pud_k;
	pmd_t *pmd, *pmd_k;

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

	if (user_mode(regs))
		goto bad_area;

	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);

	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);

#ifdef CONFIG_ARM_LPAE
	/*
	 * Only one hardware entry per PMD with LPAE.
	 */
	index = 0;
#else
	/*
	 * 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;
#endif
	if (pmd_none(pmd_k[index]))
		goto bad_area;

	copy_pmd(pmd, pmd_k);
	return 0;

bad_area:
	do_bad_area(addr, fsr, regs);
	return 0;
}
#else					/* CONFIG_MMU */
static int
do_translation_fault(unsigned long addr, unsigned int fsr,
		     struct pt_regs *regs)
{
	return 0;
}
#endif					/* CONFIG_MMU */

/*
 * 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)
{
	do_bad_area(addr, fsr, regs);
	return 0;
}

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

struct fsr_info {
	int	(*fn)(unsigned long addr, unsigned int fsr, struct pt_regs *regs);
	int	sig;
	int	code;
	const char *name;
};

/* FSR definition */
#ifdef CONFIG_ARM_LPAE
#include "fsr-3level.c"
#else
#include "fsr-2level.c"
#endif

void __init
hook_fault_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(fsr_info))
		BUG();

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

/*
 * Dispatch a data abort to the relevant handler.
 */
asmlinkage void __exception
do_DataAbort(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
	struct siginfo info;

	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
		return;

	printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
		inf->name, fsr, 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, fsr, 0);
}

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;
}

asmlinkage void __exception
do_PrefetchAbort(unsigned long addr, unsigned int ifsr, struct pt_regs *regs)
{
	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);
}

#ifndef CONFIG_ARM_LPAE
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");
	}

	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");
	}

	return 0;
}

arch_initcall(exceptions_init);
#endif
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/arch/arm/mm/fault.c
	3	*
	4	* Copyright (C) 1995 Linus Torvalds
	5	* Modifications for ARM processor (c) 1995-2004 Russell King
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License version 2 as
	9	* published by the Free Software Foundation.
	10	*/
1da177e4 LT	11	#include <linux/module.h>
1da177e4 LT	12	#include <linux/signal.h>
1da177e4	13	#include <linux/mm.h>
67306da6	14	#include <linux/hardirq.h>
1da177e4	15	#include <linux/init.h>
25ce1dd7	16	#include <linux/kprobes.h>
33fa9b13	17	#include <linux/uaccess.h>
252d4c27	18	#include <linux/page-flags.h>
412bb0a6	19	#include <linux/sched.h>
65cec8e3	20	#include <linux/highmem.h>
7ada189f	21	#include <linux/perf_event.h>
1da177e4	22
5a567d78	23	#include <asm/exception.h>
1da177e4 LT	24	#include <asm/system.h>
	25	#include <asm/pgtable.h>
	26	#include <asm/tlbflush.h>
1da177e4 LT	27
	28	#include "fault.h"
	29
09529f7a	30	#ifdef CONFIG_MMU
25ce1dd7 NP	31
	32	#ifdef CONFIG_KPROBES
	33	static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
	34	{
	35	int ret = 0;
	36
	37	if (!user_mode(regs)) {
	38	/* kprobe_running() needs smp_processor_id() */
	39	preempt_disable();
	40	if (kprobe_running() && kprobe_fault_handler(regs, fsr))
	41	ret = 1;
	42	preempt_enable();
	43	}
	44
	45	return ret;
	46	}
	47	#else
	48	static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
	49	{
	50	return 0;
	51	}
	52	#endif
	53
1da177e4 LT	54	/*
	55	* This is useful to dump out the page tables associated with
	56	* 'addr' in mm 'mm'.
	57	*/
	58	void show_pte(struct mm_struct *mm, unsigned long addr)
	59	{
	60	pgd_t *pgd;
	61
	62	if (!mm)
	63	mm = &init_mm;
	64
	65	printk(KERN_ALERT "pgd = %p\n", mm->pgd);
	66	pgd = pgd_offset(mm, addr);
29a38193 WD	67	printk(KERN_ALERT "[%08lx] *pgd=%08llx",
29a38193 WD	68	addr, (long long)pgd_val(*pgd));
1da177e4 LT	69
1da177e4 LT	70	do {
516295e5	71	pud_t *pud;
1da177e4 LT	72	pmd_t *pmd;
	73	pte_t *pte;
	74
	75	if (pgd_none(*pgd))
	76	break;
	77
	78	if (pgd_bad(*pgd)) {
	79	printk("(bad)");
	80	break;
	81	}
	82
516295e5 RK	83	pud = pud_offset(pgd, addr);
516295e5 RK	84	if (PTRS_PER_PUD != 1)
140d5dc1	85	printk(", pud=%08llx", (long long)pud_val(pud));
516295e5 RK	86
	87	if (pud_none(*pud))
	88	break;
	89
	90	if (pud_bad(*pud)) {
	91	printk("(bad)");
	92	break;
	93	}
	94
	95	pmd = pmd_offset(pud, addr);
da46c79a	96	if (PTRS_PER_PMD != 1)
29a38193	97	printk(", pmd=%08llx", (long long)pmd_val(pmd));
1da177e4 LT	98
	99	if (pmd_none(*pmd))
	100	break;
	101
	102	if (pmd_bad(*pmd)) {
	103	printk("(bad)");
	104	break;
	105	}
	106
1da177e4	107	/* We must not map this if we have highmem enabled */
252d4c27 NP	108	if (PageHighMem(pfn_to_page(pmd_val(*pmd) >> PAGE_SHIFT)))
	109	break;
	110
1da177e4	111	pte = pte_offset_map(pmd, addr);
29a38193	112	printk(", pte=%08llx", (long long)pte_val(pte));
f7b8156d	113	#ifndef CONFIG_ARM_LPAE
29a38193 WD	114	printk(", *ppte=%08llx",
29a38193 WD	115	(long long)pte_val(pte[PTE_HWTABLE_PTRS]));
f7b8156d	116	#endif
1da177e4	117	pte_unmap(pte);
1da177e4 LT	118	} while(0);
	119
	120	printk("\n");
	121	}
09529f7a CM	122	#else /* CONFIG_MMU */
	123	void show_pte(struct mm_struct *mm, unsigned long addr)
	124	{ }
	125	#endif /* CONFIG_MMU */
1da177e4 LT	126
	127	/*
	128	* Oops. The kernel tried to access some page that wasn't present.
	129	*/
	130	static void
	131	__do_kernel_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
	132	struct pt_regs *regs)
	133	{
	134	/*
	135	* Are we prepared to handle this kernel fault?
	136	*/
	137	if (fixup_exception(regs))
	138	return;
	139
	140	/*
	141	* No handler, we'll have to terminate things with extreme prejudice.
	142	*/
	143	bust_spinlocks(1);
	144	printk(KERN_ALERT
	145	"Unable to handle kernel %s at virtual address %08lx\n",
	146	(addr < PAGE_SIZE) ? "NULL pointer dereference" :
	147	"paging request", addr);
	148
	149	show_pte(mm, addr);
	150	die("Oops", regs, fsr);
	151	bust_spinlocks(0);
	152	do_exit(SIGKILL);
	153	}
	154
	155	/*
	156	* Something tried to access memory that isn't in our memory map..
	157	* User mode accesses just cause a SIGSEGV
	158	*/
	159	static void
	160	__do_user_fault(struct task_struct *tsk, unsigned long addr,
2d137c24	161	unsigned int fsr, unsigned int sig, int code,
2d137c24	162	struct pt_regs *regs)
1da177e4 LT	163	{
	164	struct siginfo si;
	165
	166	#ifdef CONFIG_DEBUG_USER
f5274c2d JM	167	if (((user_debug & UDBG_SEGV) && (sig == SIGSEGV)) \|\|
f5274c2d JM	168	((user_debug & UDBG_BUS) && (sig == SIGBUS))) {
2d137c24	169	printk(KERN_DEBUG "%s: unhandled page fault (%d) at 0x%08lx, code 0x%03x\n",
2d137c24	170	tsk->comm, sig, addr, fsr);
1da177e4 LT	171	show_pte(tsk->mm, addr);
	172	show_regs(regs);
	173	}
	174	#endif
	175
	176	tsk->thread.address = addr;
	177	tsk->thread.error_code = fsr;
	178	tsk->thread.trap_no = 14;
2d137c24	179	si.si_signo = sig;
1da177e4 LT	180	si.si_errno = 0;
	181	si.si_code = code;
	182	si.si_addr = (void __user *)addr;
2d137c24	183	force_sig_info(sig, &si, tsk);
1da177e4 LT	184	}
1da177e4 LT	185
e5beac37	186	void do_bad_area(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
1da177e4	187	{
e5beac37 RK	188	struct task_struct *tsk = current;
	189	struct mm_struct *mm = tsk->active_mm;
	190
1da177e4 LT	191	/*
	192	* If we are in kernel mode at this point, we
	193	* have no context to handle this fault with.
	194	*/
	195	if (user_mode(regs))
2d137c24	196	__do_user_fault(tsk, addr, fsr, SIGSEGV, SEGV_MAPERR, regs);
1da177e4 LT	197	else
	198	__do_kernel_fault(mm, addr, fsr, regs);
	199	}
	200
09529f7a	201	#ifdef CONFIG_MMU
5c72fc5c NP	202	#define VM_FAULT_BADMAP 0x010000
5c72fc5c NP	203	#define VM_FAULT_BADACCESS 0x020000
1da177e4	204
d374bf14 RK	205	/*
	206	* Check that the permissions on the VMA allow for the fault which occurred.
	207	* If we encountered a write fault, we must have write permission, otherwise
	208	* we allow any permission.
	209	*/
	210	static inline bool access_error(unsigned int fsr, struct vm_area_struct *vma)
	211	{
	212	unsigned int mask = VM_READ \| VM_WRITE \| VM_EXEC;
	213
	214	if (fsr & FSR_WRITE)
	215	mask = VM_WRITE;
df297bf6 RK	216	if (fsr & FSR_LNX_PF)
df297bf6 RK	217	mask = VM_EXEC;
d374bf14 RK	218
	219	return vma->vm_flags & mask ? false : true;
	220	}
	221
	222	static int __kprobes
1da177e4	223	__do_page_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
8878a539	224	unsigned int flags, struct task_struct *tsk)
1da177e4 LT	225	{
1da177e4 LT	226	struct vm_area_struct *vma;
d374bf14	227	int fault;
1da177e4 LT	228
	229	vma = find_vma(mm, addr);
	230	fault = VM_FAULT_BADMAP;
d374bf14	231	if (unlikely(!vma))
1da177e4	232	goto out;
d374bf14	233	if (unlikely(vma->vm_start > addr))
1da177e4 LT	234	goto check_stack;
	235
	236	/*
	237	* Ok, we have a good vm_area for this
	238	* memory access, so we can handle it.
	239	*/
	240	good_area:
d374bf14 RK	241	if (access_error(fsr, vma)) {
d374bf14 RK	242	fault = VM_FAULT_BADACCESS;
1da177e4	243	goto out;
d374bf14	244	}
1da177e4	245
8878a539	246	return handle_mm_fault(mm, vma, addr & PAGE_MASK, flags);
1da177e4	247
1da177e4 LT	248	check_stack:
	249	if (vma->vm_flags & VM_GROWSDOWN && !expand_stack(vma, addr))
	250	goto good_area;
	251	out:
	252	return fault;
	253	}
	254
785d3cd2	255	static int __kprobes
1da177e4 LT	256	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	257	{
	258	struct task_struct *tsk;
	259	struct mm_struct *mm;
2d137c24	260	int fault, sig, code;
8878a539 KC	261	int write = fsr & FSR_WRITE;
	262	unsigned int flags = FAULT_FLAG_ALLOW_RETRY \| FAULT_FLAG_KILLABLE \|
	263	(write ? FAULT_FLAG_WRITE : 0);
1da177e4	264
25ce1dd7 NP	265	if (notify_page_fault(regs, fsr))
	266	return 0;
	267
1da177e4 LT	268	tsk = current;
	269	mm = tsk->mm;
	270
02fe2845 RK	271	/* Enable interrupts if they were enabled in the parent context. */
	272	if (interrupts_enabled(regs))
	273	local_irq_enable();
	274
1da177e4 LT	275	/*
	276	* If we're in an interrupt or have no user
	277	* context, we must not take the fault..
	278	*/
6edaf68a	279	if (in_atomic() \|\| !mm)
1da177e4 LT	280	goto no_context;
1da177e4 LT	281
840ff6a4 RK	282	/*
	283	* As per x86, we may deadlock here. However, since the kernel only
	284	* validly references user space from well defined areas of the code,
	285	* we can bug out early if this is from code which shouldn't.
	286	*/
	287	if (!down_read_trylock(&mm->mmap_sem)) {
	288	if (!user_mode(regs) && !search_exception_tables(regs->ARM_pc))
	289	goto no_context;
8878a539	290	retry:
840ff6a4	291	down_read(&mm->mmap_sem);
bf456992 RK	292	} else {
	293	/*
	294	* The above down_read_trylock() might have succeeded in
	295	* which case, we'll have missed the might_sleep() from
	296	* down_read()
	297	*/
	298	might_sleep();
1d212712 ID	299	#ifdef CONFIG_DEBUG_VM
	300	if (!user_mode(regs) &&
	301	!search_exception_tables(regs->ARM_pc))
	302	goto no_context;
	303	#endif
840ff6a4 RK	304	}
840ff6a4 RK	305
8878a539 KC	306	fault = __do_page_fault(mm, addr, fsr, flags, tsk);
	307
	308	/* If we need to retry but a fatal signal is pending, handle the
	309	* signal first. We do not need to release the mmap_sem because
	310	* it would already be released in __lock_page_or_retry in
	311	* mm/filemap.c. */
	312	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
	313	return 0;
	314
	315	/*
	316	* Major/minor page fault accounting is only done on the
	317	* initial attempt. If we go through a retry, it is extremely
	318	* likely that the page will be found in page cache at that point.
	319	*/
1da177e4	320
a8b0ca17	321	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
8878a539 KC	322	if (flags & FAULT_FLAG_ALLOW_RETRY) {
	323	if (fault & VM_FAULT_MAJOR) {
	324	tsk->maj_flt++;
	325	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
	326	regs, addr);
	327	} else {
	328	tsk->min_flt++;
	329	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
	330	regs, addr);
	331	}
	332	if (fault & VM_FAULT_RETRY) {
	333	/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
	334	* of starvation. */
	335	flags &= ~FAULT_FLAG_ALLOW_RETRY;
	336	goto retry;
	337	}
	338	}
	339
	340	up_read(&mm->mmap_sem);
7ada189f	341
1da177e4	342	/*
ff2afb9d	343	* Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
1da177e4	344	*/
5c72fc5c	345	if (likely(!(fault & (VM_FAULT_ERROR \| VM_FAULT_BADMAP \| VM_FAULT_BADACCESS))))
1da177e4 LT	346	return 0;
1da177e4 LT	347
b42c6344 RK	348	if (fault & VM_FAULT_OOM) {
	349	/*
	350	* We ran out of memory, call the OOM killer, and return to
	351	* userspace (which will retry the fault, or kill us if we
	352	* got oom-killed)
	353	*/
	354	pagefault_out_of_memory();
	355	return 0;
	356	}
	357
1da177e4 LT	358	/*
	359	* If we are in kernel mode at this point, we
	360	* have no context to handle this fault with.
	361	*/
	362	if (!user_mode(regs))
	363	goto no_context;
	364
83c54070	365	if (fault & VM_FAULT_SIGBUS) {
2d137c24	366	/*
	367	* We had some memory, but were unable to
	368	* successfully fix up this page fault.
	369	*/
	370	sig = SIGBUS;
	371	code = BUS_ADRERR;
83c54070	372	} else {
2d137c24	373	/*
	374	* Something tried to access memory that
	375	* isn't in our memory map..
	376	*/
	377	sig = SIGSEGV;
	378	code = fault == VM_FAULT_BADACCESS ?
	379	SEGV_ACCERR : SEGV_MAPERR;
1da177e4	380	}
1da177e4	381
2d137c24	382	__do_user_fault(tsk, addr, fsr, sig, code, regs);
2d137c24	383	return 0;
1da177e4 LT	384
	385	no_context:
	386	__do_kernel_fault(mm, addr, fsr, regs);
	387	return 0;
	388	}
09529f7a CM	389	#else /* CONFIG_MMU */
	390	static int
	391	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	392	{
	393	return 0;
	394	}
	395	#endif /* CONFIG_MMU */
1da177e4 LT	396
	397	/*
	398	* First Level Translation Fault Handler
	399	*
	400	* We enter here because the first level page table doesn't contain
	401	* a valid entry for the address.
	402	*
	403	* If the address is in kernel space (>= TASK_SIZE), then we are
	404	* probably faulting in the vmalloc() area.
	405	*
	406	* If the init_task's first level page tables contains the relevant
	407	* entry, we copy the it to this task. If not, we send the process
	408	* a signal, fixup the exception, or oops the kernel.
	409	*
	410	* NOTE! We MUST NOT take any locks for this case. We may be in an
	411	* interrupt or a critical region, and should only copy the information
	412	* from the master page table, nothing more.
	413	*/
09529f7a	414	#ifdef CONFIG_MMU
785d3cd2	415	static int __kprobes
1da177e4 LT	416	do_translation_fault(unsigned long addr, unsigned int fsr,
	417	struct pt_regs *regs)
	418	{
1da177e4 LT	419	unsigned int index;
1da177e4 LT	420	pgd_t pgd, pgd_k;
516295e5	421	pud_t pud, pud_k;
1da177e4 LT	422	pmd_t pmd, pmd_k;
	423
	424	if (addr < TASK_SIZE)
	425	return do_page_fault(addr, fsr, regs);
	426
5e27fb78 A	427	if (user_mode(regs))
	428	goto bad_area;
	429
1da177e4 LT	430	index = pgd_index(addr);
	431
	432	/*
	433	* FIXME: CP15 C1 is write only on ARMv3 architectures.
	434	*/
	435	pgd = cpu_get_pgd() + index;
	436	pgd_k = init_mm.pgd + index;
	437
	438	if (pgd_none(*pgd_k))
	439	goto bad_area;
1da177e4 LT	440	if (!pgd_present(*pgd))
	441	set_pgd(pgd, *pgd_k);
	442
516295e5 RK	443	pud = pud_offset(pgd, addr);
	444	pud_k = pud_offset(pgd_k, addr);
	445
	446	if (pud_none(*pud_k))
	447	goto bad_area;
	448	if (!pud_present(*pud))
	449	set_pud(pud, *pud_k);
	450
	451	pmd = pmd_offset(pud, addr);
	452	pmd_k = pmd_offset(pud_k, addr);
1da177e4	453
f7b8156d CM	454	#ifdef CONFIG_ARM_LPAE
	455	/*
	456	* Only one hardware entry per PMD with LPAE.
	457	*/
	458	index = 0;
	459	#else
33a9c41b KS	460	/*
	461	* On ARM one Linux PGD entry contains two hardware entries (see page
	462	* tables layout in pgtable.h). We normally guarantee that we always
	463	* fill both L1 entries. But create_mapping() doesn't follow the rule.
	464	* It can create inidividual L1 entries, so here we have to call
	465	* pmd_none() check for the entry really corresponded to address, not
	466	* for the first of pair.
	467	*/
	468	index = (addr >> SECTION_SHIFT) & 1;
f7b8156d	469	#endif
33a9c41b	470	if (pmd_none(pmd_k[index]))
1da177e4 LT	471	goto bad_area;
	472
	473	copy_pmd(pmd, pmd_k);
	474	return 0;
	475
	476	bad_area:
e5beac37	477	do_bad_area(addr, fsr, regs);
1da177e4 LT	478	return 0;
1da177e4 LT	479	}
09529f7a CM	480	#else /* CONFIG_MMU */
	481	static int
	482	do_translation_fault(unsigned long addr, unsigned int fsr,
	483	struct pt_regs *regs)
	484	{
	485	return 0;
	486	}
	487	#endif /* CONFIG_MMU */
1da177e4 LT	488
	489	/*
	490	* Some section permission faults need to be handled gracefully.
	491	* They can happen due to a __{get,put}_user during an oops.
	492	*/
	493	static int
	494	do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	495	{
e5beac37	496	do_bad_area(addr, fsr, regs);
1da177e4 LT	497	return 0;
	498	}
	499
	500	/*
	501	* This abort handler always returns "fault".
	502	*/
	503	static int
	504	do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	505	{
	506	return 1;
	507	}
	508
136848d4	509	struct fsr_info {
1da177e4 LT	510	int (fn)(unsigned long addr, unsigned int fsr, struct pt_regs regs);
1da177e4 LT	511	int sig;
cfb0810e	512	int code;
1da177e4	513	const char *name;
1da177e4 LT	514	};
1da177e4 LT	515
136848d4	516	/* FSR definition */
f7b8156d CM	517	#ifdef CONFIG_ARM_LPAE
	518	#include "fsr-3level.c"
	519	#else
136848d4	520	#include "fsr-2level.c"
f7b8156d	521	#endif
136848d4	522
1da177e4 LT	523	void __init
1da177e4 LT	524	hook_fault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
6338a6aa	525	int sig, int code, const char *name)
1da177e4	526	{
6338a6aa KS	527	if (nr < 0 \|\| nr >= ARRAY_SIZE(fsr_info))
	528	BUG();
	529
	530	fsr_info[nr].fn = fn;
	531	fsr_info[nr].sig = sig;
	532	fsr_info[nr].code = code;
	533	fsr_info[nr].name = name;
1da177e4 LT	534	}
	535
	536	/*
	537	* Dispatch a data abort to the relevant handler.
	538	*/
7ab3f8d5	539	asmlinkage void __exception
1da177e4 LT	540	do_DataAbort(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
1da177e4 LT	541	{
c88d6aa7	542	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
cfb0810e	543	struct siginfo info;
1da177e4	544
df297bf6	545	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
1da177e4 LT	546	return;
	547
	548	printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
	549	inf->name, fsr, addr);
cfb0810e RK	550
	551	info.si_signo = inf->sig;
	552	info.si_errno = 0;
	553	info.si_code = inf->code;
	554	info.si_addr = (void __user *)addr;
1eeb66a1	555	arm_notify_die("", regs, &info, fsr, 0);
1da177e4 LT	556	}
1da177e4 LT	557
3a4b5dca WD	558	void __init
	559	hook_ifault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
	560	int sig, int code, const char *name)
	561	{
	562	if (nr < 0 \|\| nr >= ARRAY_SIZE(ifsr_info))
	563	BUG();
	564
	565	ifsr_info[nr].fn = fn;
	566	ifsr_info[nr].sig = sig;
	567	ifsr_info[nr].code = code;
	568	ifsr_info[nr].name = name;
	569	}
	570
7ab3f8d5	571	asmlinkage void __exception
4fb28474	572	do_PrefetchAbort(unsigned long addr, unsigned int ifsr, struct pt_regs *regs)
1da177e4	573	{
d25ef8b8 KS	574	const struct fsr_info *inf = ifsr_info + fsr_fs(ifsr);
	575	struct siginfo info;
	576
	577	if (!inf->fn(addr, ifsr \| FSR_LNX_PF, regs))
	578	return;
	579
	580	printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
	581	inf->name, ifsr, addr);
	582
	583	info.si_signo = inf->sig;
	584	info.si_errno = 0;
	585	info.si_code = inf->code;
	586	info.si_addr = (void __user *)addr;
	587	arm_notify_die("", regs, &info, ifsr, 0);
1da177e4 LT	588	}
1da177e4 LT	589
f7b8156d	590	#ifndef CONFIG_ARM_LPAE
993bf4ec KS	591	static int __init exceptions_init(void)
	592	{
	593	if (cpu_architecture() >= CPU_ARCH_ARMv6) {
	594	hook_fault_code(4, do_translation_fault, SIGSEGV, SEGV_MAPERR,
	595	"I-cache maintenance fault");
	596	}
	597
b8ab5397 KS	598	if (cpu_architecture() >= CPU_ARCH_ARMv7) {
	599	/*
	600	* TODO: Access flag faults introduced in ARMv6K.
	601	* Runtime check for 'K' extension is needed
	602	*/
	603	hook_fault_code(3, do_bad, SIGSEGV, SEGV_MAPERR,
	604	"section access flag fault");
	605	hook_fault_code(6, do_bad, SIGSEGV, SEGV_MAPERR,
	606	"section access flag fault");
	607	}
	608
993bf4ec KS	609	return 0;
	610	}
	611
	612	arch_initcall(exceptions_init);
f7b8156d	613	#endif