[GitHub/mt8127/android_kernel_alcatel_ttab.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/pgtable.h>
#include <asm/system_misc.h>
#include <asm/system_info.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:
	/* Don't allow expansion below FIRST_USER_ADDRESS */
	if (vma->vm_flags & VM_GROWSDOWN &&
	    addr >= FIRST_USER_ADDRESS && !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 (!(fault & VM_FAULT_ERROR) && 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;
			flags |= FAULT_FLAG_TRIED;
			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);

	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	24	#include <asm/pgtable.h>
9f97da78 DH	25	#include <asm/system_misc.h>
9f97da78 DH	26	#include <asm/system_info.h>
1da177e4	27	#include <asm/tlbflush.h>
1da177e4 LT	28
	29	#include "fault.h"
	30
09529f7a	31	#ifdef CONFIG_MMU
25ce1dd7 NP	32
	33	#ifdef CONFIG_KPROBES
	34	static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
	35	{
	36	int ret = 0;
	37
	38	if (!user_mode(regs)) {
	39	/* kprobe_running() needs smp_processor_id() */
	40	preempt_disable();
	41	if (kprobe_running() && kprobe_fault_handler(regs, fsr))
	42	ret = 1;
	43	preempt_enable();
	44	}
	45
	46	return ret;
	47	}
	48	#else
	49	static inline int notify_page_fault(struct pt_regs *regs, unsigned int fsr)
	50	{
	51	return 0;
	52	}
	53	#endif
	54
1da177e4 LT	55	/*
	56	* This is useful to dump out the page tables associated with
	57	* 'addr' in mm 'mm'.
	58	*/
	59	void show_pte(struct mm_struct *mm, unsigned long addr)
	60	{
	61	pgd_t *pgd;
	62
	63	if (!mm)
	64	mm = &init_mm;
	65
	66	printk(KERN_ALERT "pgd = %p\n", mm->pgd);
	67	pgd = pgd_offset(mm, addr);
29a38193 WD	68	printk(KERN_ALERT "[%08lx] *pgd=%08llx",
29a38193 WD	69	addr, (long long)pgd_val(*pgd));
1da177e4 LT	70
1da177e4 LT	71	do {
516295e5	72	pud_t *pud;
1da177e4 LT	73	pmd_t *pmd;
	74	pte_t *pte;
	75
	76	if (pgd_none(*pgd))
	77	break;
	78
	79	if (pgd_bad(*pgd)) {
	80	printk("(bad)");
	81	break;
	82	}
	83
516295e5 RK	84	pud = pud_offset(pgd, addr);
516295e5 RK	85	if (PTRS_PER_PUD != 1)
140d5dc1	86	printk(", pud=%08llx", (long long)pud_val(pud));
516295e5 RK	87
	88	if (pud_none(*pud))
	89	break;
	90
	91	if (pud_bad(*pud)) {
	92	printk("(bad)");
	93	break;
	94	}
	95
	96	pmd = pmd_offset(pud, addr);
da46c79a	97	if (PTRS_PER_PMD != 1)
29a38193	98	printk(", pmd=%08llx", (long long)pmd_val(pmd));
1da177e4 LT	99
	100	if (pmd_none(*pmd))
	101	break;
	102
	103	if (pmd_bad(*pmd)) {
	104	printk("(bad)");
	105	break;
	106	}
	107
1da177e4	108	/* We must not map this if we have highmem enabled */
252d4c27 NP	109	if (PageHighMem(pfn_to_page(pmd_val(*pmd) >> PAGE_SHIFT)))
	110	break;
	111
1da177e4	112	pte = pte_offset_map(pmd, addr);
29a38193	113	printk(", pte=%08llx", (long long)pte_val(pte));
f7b8156d	114	#ifndef CONFIG_ARM_LPAE
29a38193 WD	115	printk(", *ppte=%08llx",
29a38193 WD	116	(long long)pte_val(pte[PTE_HWTABLE_PTRS]));
f7b8156d	117	#endif
1da177e4	118	pte_unmap(pte);
1da177e4 LT	119	} while(0);
	120
	121	printk("\n");
	122	}
09529f7a CM	123	#else /* CONFIG_MMU */
	124	void show_pte(struct mm_struct *mm, unsigned long addr)
	125	{ }
	126	#endif /* CONFIG_MMU */
1da177e4 LT	127
	128	/*
	129	* Oops. The kernel tried to access some page that wasn't present.
	130	*/
	131	static void
	132	__do_kernel_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
	133	struct pt_regs *regs)
	134	{
	135	/*
	136	* Are we prepared to handle this kernel fault?
	137	*/
	138	if (fixup_exception(regs))
	139	return;
	140
	141	/*
	142	* No handler, we'll have to terminate things with extreme prejudice.
	143	*/
	144	bust_spinlocks(1);
	145	printk(KERN_ALERT
	146	"Unable to handle kernel %s at virtual address %08lx\n",
	147	(addr < PAGE_SIZE) ? "NULL pointer dereference" :
	148	"paging request", addr);
	149
	150	show_pte(mm, addr);
	151	die("Oops", regs, fsr);
	152	bust_spinlocks(0);
	153	do_exit(SIGKILL);
	154	}
	155
	156	/*
	157	* Something tried to access memory that isn't in our memory map..
	158	* User mode accesses just cause a SIGSEGV
	159	*/
	160	static void
	161	__do_user_fault(struct task_struct *tsk, unsigned long addr,
2d137c24 AM	162	unsigned int fsr, unsigned int sig, int code,
2d137c24 AM	163	struct pt_regs *regs)
1da177e4 LT	164	{
	165	struct siginfo si;
	166
	167	#ifdef CONFIG_DEBUG_USER
f5274c2d JM	168	if (((user_debug & UDBG_SEGV) && (sig == SIGSEGV)) \|\|
f5274c2d JM	169	((user_debug & UDBG_BUS) && (sig == SIGBUS))) {
2d137c24 AM	170	printk(KERN_DEBUG "%s: unhandled page fault (%d) at 0x%08lx, code 0x%03x\n",
2d137c24 AM	171	tsk->comm, sig, addr, fsr);
1da177e4 LT	172	show_pte(tsk->mm, addr);
	173	show_regs(regs);
	174	}
	175	#endif
	176
	177	tsk->thread.address = addr;
	178	tsk->thread.error_code = fsr;
	179	tsk->thread.trap_no = 14;
2d137c24	180	si.si_signo = sig;
1da177e4 LT	181	si.si_errno = 0;
	182	si.si_code = code;
	183	si.si_addr = (void __user *)addr;
2d137c24	184	force_sig_info(sig, &si, tsk);
1da177e4 LT	185	}
1da177e4 LT	186
e5beac37	187	void do_bad_area(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
1da177e4	188	{
e5beac37 RK	189	struct task_struct *tsk = current;
	190	struct mm_struct *mm = tsk->active_mm;
	191
1da177e4 LT	192	/*
	193	* If we are in kernel mode at this point, we
	194	* have no context to handle this fault with.
	195	*/
	196	if (user_mode(regs))
2d137c24	197	__do_user_fault(tsk, addr, fsr, SIGSEGV, SEGV_MAPERR, regs);
1da177e4 LT	198	else
	199	__do_kernel_fault(mm, addr, fsr, regs);
	200	}
	201
09529f7a	202	#ifdef CONFIG_MMU
5c72fc5c NP	203	#define VM_FAULT_BADMAP 0x010000
5c72fc5c NP	204	#define VM_FAULT_BADACCESS 0x020000
1da177e4	205
d374bf14 RK	206	/*
	207	* Check that the permissions on the VMA allow for the fault which occurred.
	208	* If we encountered a write fault, we must have write permission, otherwise
	209	* we allow any permission.
	210	*/
	211	static inline bool access_error(unsigned int fsr, struct vm_area_struct *vma)
	212	{
	213	unsigned int mask = VM_READ \| VM_WRITE \| VM_EXEC;
	214
	215	if (fsr & FSR_WRITE)
	216	mask = VM_WRITE;
df297bf6 RK	217	if (fsr & FSR_LNX_PF)
df297bf6 RK	218	mask = VM_EXEC;
d374bf14 RK	219
	220	return vma->vm_flags & mask ? false : true;
	221	}
	222
	223	static int __kprobes
1da177e4	224	__do_page_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
8878a539	225	unsigned int flags, struct task_struct *tsk)
1da177e4 LT	226	{
1da177e4 LT	227	struct vm_area_struct *vma;
d374bf14	228	int fault;
1da177e4 LT	229
	230	vma = find_vma(mm, addr);
	231	fault = VM_FAULT_BADMAP;
d374bf14	232	if (unlikely(!vma))
1da177e4	233	goto out;
d374bf14	234	if (unlikely(vma->vm_start > addr))
1da177e4 LT	235	goto check_stack;
	236
	237	/*
	238	* Ok, we have a good vm_area for this
	239	* memory access, so we can handle it.
	240	*/
	241	good_area:
d374bf14 RK	242	if (access_error(fsr, vma)) {
d374bf14 RK	243	fault = VM_FAULT_BADACCESS;
1da177e4	244	goto out;
d374bf14	245	}
1da177e4	246
8878a539	247	return handle_mm_fault(mm, vma, addr & PAGE_MASK, flags);
1da177e4	248
1da177e4	249	check_stack:
9b61a4d1 RK	250	/* Don't allow expansion below FIRST_USER_ADDRESS */
	251	if (vma->vm_flags & VM_GROWSDOWN &&
	252	addr >= FIRST_USER_ADDRESS && !expand_stack(vma, addr))
1da177e4 LT	253	goto good_area;
	254	out:
	255	return fault;
	256	}
	257
785d3cd2	258	static int __kprobes
1da177e4 LT	259	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	260	{
	261	struct task_struct *tsk;
	262	struct mm_struct *mm;
2d137c24	263	int fault, sig, code;
8878a539 KC	264	int write = fsr & FSR_WRITE;
	265	unsigned int flags = FAULT_FLAG_ALLOW_RETRY \| FAULT_FLAG_KILLABLE \|
	266	(write ? FAULT_FLAG_WRITE : 0);
1da177e4	267
25ce1dd7 NP	268	if (notify_page_fault(regs, fsr))
	269	return 0;
	270
1da177e4 LT	271	tsk = current;
	272	mm = tsk->mm;
	273
02fe2845 RK	274	/* Enable interrupts if they were enabled in the parent context. */
	275	if (interrupts_enabled(regs))
	276	local_irq_enable();
	277
1da177e4 LT	278	/*
	279	* If we're in an interrupt or have no user
	280	* context, we must not take the fault..
	281	*/
6edaf68a	282	if (in_atomic() \|\| !mm)
1da177e4 LT	283	goto no_context;
1da177e4 LT	284
840ff6a4 RK	285	/*
	286	* As per x86, we may deadlock here. However, since the kernel only
	287	* validly references user space from well defined areas of the code,
	288	* we can bug out early if this is from code which shouldn't.
	289	*/
	290	if (!down_read_trylock(&mm->mmap_sem)) {
	291	if (!user_mode(regs) && !search_exception_tables(regs->ARM_pc))
	292	goto no_context;
8878a539	293	retry:
840ff6a4	294	down_read(&mm->mmap_sem);
bf456992 RK	295	} else {
	296	/*
	297	* The above down_read_trylock() might have succeeded in
	298	* which case, we'll have missed the might_sleep() from
	299	* down_read()
	300	*/
	301	might_sleep();
1d212712 ID	302	#ifdef CONFIG_DEBUG_VM
	303	if (!user_mode(regs) &&
	304	!search_exception_tables(regs->ARM_pc))
	305	goto no_context;
	306	#endif
840ff6a4 RK	307	}
840ff6a4 RK	308
8878a539 KC	309	fault = __do_page_fault(mm, addr, fsr, flags, tsk);
	310
	311	/* If we need to retry but a fatal signal is pending, handle the
	312	* signal first. We do not need to release the mmap_sem because
	313	* it would already be released in __lock_page_or_retry in
	314	* mm/filemap.c. */
	315	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
	316	return 0;
	317
	318	/*
	319	* Major/minor page fault accounting is only done on the
	320	* initial attempt. If we go through a retry, it is extremely
	321	* likely that the page will be found in page cache at that point.
	322	*/
1da177e4	323
a8b0ca17	324	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
dff2aa7a	325	if (!(fault & VM_FAULT_ERROR) && flags & FAULT_FLAG_ALLOW_RETRY) {
8878a539 KC	326	if (fault & VM_FAULT_MAJOR) {
	327	tsk->maj_flt++;
	328	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
	329	regs, addr);
	330	} else {
	331	tsk->min_flt++;
	332	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
	333	regs, addr);
	334	}
	335	if (fault & VM_FAULT_RETRY) {
	336	/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
	337	* of starvation. */
	338	flags &= ~FAULT_FLAG_ALLOW_RETRY;
45cac65b	339	flags \|= FAULT_FLAG_TRIED;
8878a539 KC	340	goto retry;
	341	}
	342	}
	343
	344	up_read(&mm->mmap_sem);
7ada189f	345
1da177e4	346	/*
ff2afb9d	347	* Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
1da177e4	348	*/
5c72fc5c	349	if (likely(!(fault & (VM_FAULT_ERROR \| VM_FAULT_BADMAP \| VM_FAULT_BADACCESS))))
1da177e4 LT	350	return 0;
1da177e4 LT	351
b42c6344 RK	352	if (fault & VM_FAULT_OOM) {
	353	/*
	354	* We ran out of memory, call the OOM killer, and return to
	355	* userspace (which will retry the fault, or kill us if we
	356	* got oom-killed)
	357	*/
	358	pagefault_out_of_memory();
	359	return 0;
	360	}
	361
1da177e4 LT	362	/*
	363	* If we are in kernel mode at this point, we
	364	* have no context to handle this fault with.
	365	*/
	366	if (!user_mode(regs))
	367	goto no_context;
	368
83c54070	369	if (fault & VM_FAULT_SIGBUS) {
2d137c24 AM	370	/*
	371	* We had some memory, but were unable to
	372	* successfully fix up this page fault.
	373	*/
	374	sig = SIGBUS;
	375	code = BUS_ADRERR;
83c54070	376	} else {
2d137c24 AM	377	/*
	378	* Something tried to access memory that
	379	* isn't in our memory map..
	380	*/
	381	sig = SIGSEGV;
	382	code = fault == VM_FAULT_BADACCESS ?
	383	SEGV_ACCERR : SEGV_MAPERR;
1da177e4	384	}
1da177e4	385
2d137c24 AM	386	__do_user_fault(tsk, addr, fsr, sig, code, regs);
2d137c24 AM	387	return 0;
1da177e4 LT	388
	389	no_context:
	390	__do_kernel_fault(mm, addr, fsr, regs);
	391	return 0;
	392	}
09529f7a CM	393	#else /* CONFIG_MMU */
	394	static int
	395	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	396	{
	397	return 0;
	398	}
	399	#endif /* CONFIG_MMU */
1da177e4 LT	400
	401	/*
	402	* First Level Translation Fault Handler
	403	*
	404	* We enter here because the first level page table doesn't contain
	405	* a valid entry for the address.
	406	*
	407	* If the address is in kernel space (>= TASK_SIZE), then we are
	408	* probably faulting in the vmalloc() area.
	409	*
	410	* If the init_task's first level page tables contains the relevant
	411	* entry, we copy the it to this task. If not, we send the process
	412	* a signal, fixup the exception, or oops the kernel.
	413	*
	414	* NOTE! We MUST NOT take any locks for this case. We may be in an
	415	* interrupt or a critical region, and should only copy the information
	416	* from the master page table, nothing more.
	417	*/
09529f7a	418	#ifdef CONFIG_MMU
785d3cd2	419	static int __kprobes
1da177e4 LT	420	do_translation_fault(unsigned long addr, unsigned int fsr,
	421	struct pt_regs *regs)
	422	{
1da177e4 LT	423	unsigned int index;
1da177e4 LT	424	pgd_t pgd, pgd_k;
516295e5	425	pud_t pud, pud_k;
1da177e4 LT	426	pmd_t pmd, pmd_k;
	427
	428	if (addr < TASK_SIZE)
	429	return do_page_fault(addr, fsr, regs);
	430
5e27fb78 A	431	if (user_mode(regs))
	432	goto bad_area;
	433
1da177e4 LT	434	index = pgd_index(addr);
1da177e4 LT	435
1da177e4 LT	436	pgd = cpu_get_pgd() + index;
	437	pgd_k = init_mm.pgd + index;
	438
	439	if (pgd_none(*pgd_k))
	440	goto bad_area;
1da177e4 LT	441	if (!pgd_present(*pgd))
	442	set_pgd(pgd, *pgd_k);
	443
516295e5 RK	444	pud = pud_offset(pgd, addr);
	445	pud_k = pud_offset(pgd_k, addr);
	446
	447	if (pud_none(*pud_k))
	448	goto bad_area;
	449	if (!pud_present(*pud))
	450	set_pud(pud, *pud_k);
	451
	452	pmd = pmd_offset(pud, addr);
	453	pmd_k = pmd_offset(pud_k, addr);
1da177e4	454
f7b8156d CM	455	#ifdef CONFIG_ARM_LPAE
	456	/*
	457	* Only one hardware entry per PMD with LPAE.
	458	*/
	459	index = 0;
	460	#else
33a9c41b KS	461	/*
	462	* On ARM one Linux PGD entry contains two hardware entries (see page
	463	* tables layout in pgtable.h). We normally guarantee that we always
	464	* fill both L1 entries. But create_mapping() doesn't follow the rule.
	465	* It can create inidividual L1 entries, so here we have to call
	466	* pmd_none() check for the entry really corresponded to address, not
	467	* for the first of pair.
	468	*/
	469	index = (addr >> SECTION_SHIFT) & 1;
f7b8156d	470	#endif
33a9c41b	471	if (pmd_none(pmd_k[index]))
1da177e4 LT	472	goto bad_area;
	473
	474	copy_pmd(pmd, pmd_k);
	475	return 0;
	476
	477	bad_area:
e5beac37	478	do_bad_area(addr, fsr, regs);
1da177e4 LT	479	return 0;
1da177e4 LT	480	}
09529f7a CM	481	#else /* CONFIG_MMU */
	482	static int
	483	do_translation_fault(unsigned long addr, unsigned int fsr,
	484	struct pt_regs *regs)
	485	{
	486	return 0;
	487	}
	488	#endif /* CONFIG_MMU */
1da177e4 LT	489
	490	/*
	491	* Some section permission faults need to be handled gracefully.
	492	* They can happen due to a __{get,put}_user during an oops.
	493	*/
	494	static int
	495	do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	496	{
e5beac37	497	do_bad_area(addr, fsr, regs);
1da177e4 LT	498	return 0;
	499	}
	500
	501	/*
	502	* This abort handler always returns "fault".
	503	*/
	504	static int
	505	do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	506	{
	507	return 1;
	508	}
	509
136848d4	510	struct fsr_info {
1da177e4 LT	511	int (fn)(unsigned long addr, unsigned int fsr, struct pt_regs regs);
1da177e4 LT	512	int sig;
cfb0810e	513	int code;
1da177e4	514	const char *name;
1da177e4 LT	515	};
1da177e4 LT	516
136848d4	517	/* FSR definition */
f7b8156d CM	518	#ifdef CONFIG_ARM_LPAE
	519	#include "fsr-3level.c"
	520	#else
136848d4	521	#include "fsr-2level.c"
f7b8156d	522	#endif
136848d4	523
1da177e4 LT	524	void __init
1da177e4 LT	525	hook_fault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
6338a6aa	526	int sig, int code, const char *name)
1da177e4	527	{
6338a6aa KS	528	if (nr < 0 \|\| nr >= ARRAY_SIZE(fsr_info))
	529	BUG();
	530
	531	fsr_info[nr].fn = fn;
	532	fsr_info[nr].sig = sig;
	533	fsr_info[nr].code = code;
	534	fsr_info[nr].name = name;
1da177e4 LT	535	}
	536
	537	/*
	538	* Dispatch a data abort to the relevant handler.
	539	*/
7ab3f8d5	540	asmlinkage void __exception
1da177e4 LT	541	do_DataAbort(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
1da177e4 LT	542	{
c88d6aa7	543	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
cfb0810e	544	struct siginfo info;
1da177e4	545
df297bf6	546	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
1da177e4 LT	547	return;
	548
	549	printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
	550	inf->name, fsr, addr);
cfb0810e RK	551
	552	info.si_signo = inf->sig;
	553	info.si_errno = 0;
	554	info.si_code = inf->code;
	555	info.si_addr = (void __user *)addr;
1eeb66a1	556	arm_notify_die("", regs, &info, fsr, 0);
1da177e4 LT	557	}
1da177e4 LT	558
3a4b5dca WD	559	void __init
	560	hook_ifault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
	561	int sig, int code, const char *name)
	562	{
	563	if (nr < 0 \|\| nr >= ARRAY_SIZE(ifsr_info))
	564	BUG();
	565
	566	ifsr_info[nr].fn = fn;
	567	ifsr_info[nr].sig = sig;
	568	ifsr_info[nr].code = code;
	569	ifsr_info[nr].name = name;
	570	}
	571
7ab3f8d5	572	asmlinkage void __exception
4fb28474	573	do_PrefetchAbort(unsigned long addr, unsigned int ifsr, struct pt_regs *regs)
1da177e4	574	{
d25ef8b8 KS	575	const struct fsr_info *inf = ifsr_info + fsr_fs(ifsr);
	576	struct siginfo info;
	577
	578	if (!inf->fn(addr, ifsr \| FSR_LNX_PF, regs))
	579	return;
	580
	581	printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
	582	inf->name, ifsr, addr);
	583
	584	info.si_signo = inf->sig;
	585	info.si_errno = 0;
	586	info.si_code = inf->code;
	587	info.si_addr = (void __user *)addr;
	588	arm_notify_die("", regs, &info, ifsr, 0);
1da177e4 LT	589	}
1da177e4 LT	590
f7b8156d	591	#ifndef CONFIG_ARM_LPAE
993bf4ec KS	592	static int __init exceptions_init(void)
	593	{
	594	if (cpu_architecture() >= CPU_ARCH_ARMv6) {
	595	hook_fault_code(4, do_translation_fault, SIGSEGV, SEGV_MAPERR,
	596	"I-cache maintenance fault");
	597	}
	598
b8ab5397 KS	599	if (cpu_architecture() >= CPU_ARCH_ARMv7) {
	600	/*
	601	* TODO: Access flag faults introduced in ARMv6K.
	602	* Runtime check for 'K' extension is needed
	603	*/
	604	hook_fault_code(3, do_bad, SIGSEGV, SEGV_MAPERR,
	605	"section access flag fault");
	606	hook_fault_code(6, do_bad, SIGSEGV, SEGV_MAPERR,
	607	"section access flag fault");
	608	}
	609
993bf4ec KS	610	return 0;
	611	}
	612
	613	arch_initcall(exceptions_init);
f7b8156d	614	#endif