[linux-2.6-block.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;

	pr_alert("pgd = %p\n", mm->pgd);
	pgd = pgd_offset(mm, addr);
	pr_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)) {
			pr_cont("(bad)");
			break;
		}

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

		if (pud_none(*pud))
			break;

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

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

		if (pmd_none(*pmd))
			break;

		if (pmd_bad(*pmd)) {
			pr_cont("(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);
		pr_cont(", *pte=%08llx", (long long)pte_val(*pte));
#ifndef CONFIG_ARM_LPAE
		pr_cont(", *ppte=%08llx",
		       (long long)pte_val(pte[PTE_HWTABLE_PTRS]));
#endif
		pte_unmap(pte);
	} while(0);

	pr_cont("\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);
	pr_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;
	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;

	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 (faulthandler_disabled() || !mm)
		goto no_context;

	if (user_mode(regs))
		flags |= FAULT_FLAG_USER;
	if (fsr & FSR_WRITE)
		flags |= FAULT_FLAG_WRITE;

	/*
	 * 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 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_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 (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.
 */
#ifndef CONFIG_ARM_LPAE
static int
do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	do_bad_area(addr, fsr, regs);
	return 0;
}
#endif /* CONFIG_ARM_LPAE */

/*
 * 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;

	pr_alert("Unhandled fault: %s (0x%03x) at 0x%08lx\n",
		inf->name, fsr, addr);
	show_pte(current->mm, 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;

	pr_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
4ed89f22	66	pr_alert("pgd = %p\n", mm->pgd);
1da177e4	67	pgd = pgd_offset(mm, addr);
4ed89f22	68	pr_alert("[%08lx] *pgd=%08llx",
29a38193	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)) {
4ed89f22	80	pr_cont("(bad)");
1da177e4 LT	81	break;
	82	}
	83
516295e5 RK	84	pud = pud_offset(pgd, addr);
516295e5 RK	85	if (PTRS_PER_PUD != 1)
4ed89f22	86	pr_cont(", pud=%08llx", (long long)pud_val(pud));
516295e5 RK	87
	88	if (pud_none(*pud))
	89	break;
	90
	91	if (pud_bad(*pud)) {
4ed89f22	92	pr_cont("(bad)");
516295e5 RK	93	break;
	94	}
	95
	96	pmd = pmd_offset(pud, addr);
da46c79a	97	if (PTRS_PER_PMD != 1)
4ed89f22	98	pr_cont(", pmd=%08llx", (long long)pmd_val(pmd));
1da177e4 LT	99
	100	if (pmd_none(*pmd))
	101	break;
	102
	103	if (pmd_bad(*pmd)) {
4ed89f22	104	pr_cont("(bad)");
1da177e4 LT	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);
4ed89f22	113	pr_cont(", pte=%08llx", (long long)pte_val(pte));
f7b8156d	114	#ifndef CONFIG_ARM_LPAE
4ed89f22	115	pr_cont(", *ppte=%08llx",
29a38193	116	(long long)pte_val(pte[PTE_HWTABLE_PTRS]));
f7b8156d	117	#endif
1da177e4	118	pte_unmap(pte);
1da177e4 LT	119	} while(0);
1da177e4 LT	120
4ed89f22	121	pr_cont("\n");
1da177e4	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);
4ed89f22 RK	145	pr_alert("Unable to handle kernel %s at virtual address %08lx\n",
	146	(addr < PAGE_SIZE) ? "NULL pointer dereference" :
	147	"paging request", addr);
1da177e4 LT	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	248	check_stack:
9b61a4d1 RK	249	/* Don't allow expansion below FIRST_USER_ADDRESS */
	250	if (vma->vm_flags & VM_GROWSDOWN &&
	251	addr >= FIRST_USER_ADDRESS && !expand_stack(vma, addr))
1da177e4 LT	252	goto good_area;
	253	out:
	254	return fault;
	255	}
	256
785d3cd2	257	static int __kprobes
1da177e4 LT	258	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	259	{
	260	struct task_struct *tsk;
	261	struct mm_struct *mm;
2d137c24	262	int fault, sig, code;
759496ba	263	unsigned int flags = FAULT_FLAG_ALLOW_RETRY \| FAULT_FLAG_KILLABLE;
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	*/
70ffdb93	279	if (faulthandler_disabled() \|\| !mm)
1da177e4 LT	280	goto no_context;
1da177e4 LT	281
759496ba JW	282	if (user_mode(regs))
	283	flags \|= FAULT_FLAG_USER;
	284	if (fsr & FSR_WRITE)
	285	flags \|= FAULT_FLAG_WRITE;
	286
840ff6a4 RK	287	/*
	288	* As per x86, we may deadlock here. However, since the kernel only
	289	* validly references user space from well defined areas of the code,
	290	* we can bug out early if this is from code which shouldn't.
	291	*/
	292	if (!down_read_trylock(&mm->mmap_sem)) {
	293	if (!user_mode(regs) && !search_exception_tables(regs->ARM_pc))
	294	goto no_context;
8878a539	295	retry:
840ff6a4	296	down_read(&mm->mmap_sem);
bf456992 RK	297	} else {
	298	/*
	299	* The above down_read_trylock() might have succeeded in
	300	* which case, we'll have missed the might_sleep() from
	301	* down_read()
	302	*/
	303	might_sleep();
1d212712 ID	304	#ifdef CONFIG_DEBUG_VM
	305	if (!user_mode(regs) &&
	306	!search_exception_tables(regs->ARM_pc))
	307	goto no_context;
	308	#endif
840ff6a4 RK	309	}
840ff6a4 RK	310
8878a539 KC	311	fault = __do_page_fault(mm, addr, fsr, flags, tsk);
	312
	313	/* If we need to retry but a fatal signal is pending, handle the
	314	* signal first. We do not need to release the mmap_sem because
	315	* it would already be released in __lock_page_or_retry in
	316	* mm/filemap.c. */
	317	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
	318	return 0;
	319
	320	/*
	321	* Major/minor page fault accounting is only done on the
	322	* initial attempt. If we go through a retry, it is extremely
	323	* likely that the page will be found in page cache at that point.
	324	*/
1da177e4	325
a8b0ca17	326	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
dff2aa7a	327	if (!(fault & VM_FAULT_ERROR) && flags & FAULT_FLAG_ALLOW_RETRY) {
8878a539 KC	328	if (fault & VM_FAULT_MAJOR) {
	329	tsk->maj_flt++;
	330	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
	331	regs, addr);
	332	} else {
	333	tsk->min_flt++;
	334	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
	335	regs, addr);
	336	}
	337	if (fault & VM_FAULT_RETRY) {
	338	/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
	339	* of starvation. */
	340	flags &= ~FAULT_FLAG_ALLOW_RETRY;
45cac65b	341	flags \|= FAULT_FLAG_TRIED;
8878a539 KC	342	goto retry;
	343	}
	344	}
	345
	346	up_read(&mm->mmap_sem);
7ada189f	347
1da177e4	348	/*
ff2afb9d	349	* Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
1da177e4	350	*/
5c72fc5c	351	if (likely(!(fault & (VM_FAULT_ERROR \| VM_FAULT_BADMAP \| VM_FAULT_BADACCESS))))
1da177e4 LT	352	return 0;
1da177e4 LT	353
87134102 JW	354	/*
	355	* If we are in kernel mode at this point, we
	356	* have no context to handle this fault with.
	357	*/
	358	if (!user_mode(regs))
	359	goto no_context;
	360
b42c6344 RK	361	if (fault & VM_FAULT_OOM) {
	362	/*
	363	* We ran out of memory, call the OOM killer, and return to
	364	* userspace (which will retry the fault, or kill us if we
	365	* got oom-killed)
	366	*/
	367	pagefault_out_of_memory();
	368	return 0;
	369	}
	370
83c54070	371	if (fault & VM_FAULT_SIGBUS) {
2d137c24	372	/*
	373	* We had some memory, but were unable to
	374	* successfully fix up this page fault.
	375	*/
	376	sig = SIGBUS;
	377	code = BUS_ADRERR;
83c54070	378	} else {
2d137c24	379	/*
	380	* Something tried to access memory that
	381	* isn't in our memory map..
	382	*/
	383	sig = SIGSEGV;
	384	code = fault == VM_FAULT_BADACCESS ?
	385	SEGV_ACCERR : SEGV_MAPERR;
1da177e4	386	}
1da177e4	387
2d137c24	388	__do_user_fault(tsk, addr, fsr, sig, code, regs);
2d137c24	389	return 0;
1da177e4 LT	390
	391	no_context:
	392	__do_kernel_fault(mm, addr, fsr, regs);
	393	return 0;
	394	}
09529f7a CM	395	#else /* CONFIG_MMU */
	396	static int
	397	do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	398	{
	399	return 0;
	400	}
	401	#endif /* CONFIG_MMU */
1da177e4 LT	402
	403	/*
	404	* First Level Translation Fault Handler
	405	*
	406	* We enter here because the first level page table doesn't contain
	407	* a valid entry for the address.
	408	*
	409	* If the address is in kernel space (>= TASK_SIZE), then we are
	410	* probably faulting in the vmalloc() area.
	411	*
	412	* If the init_task's first level page tables contains the relevant
	413	* entry, we copy the it to this task. If not, we send the process
	414	* a signal, fixup the exception, or oops the kernel.
	415	*
	416	* NOTE! We MUST NOT take any locks for this case. We may be in an
	417	* interrupt or a critical region, and should only copy the information
	418	* from the master page table, nothing more.
	419	*/
09529f7a	420	#ifdef CONFIG_MMU
785d3cd2	421	static int __kprobes
1da177e4 LT	422	do_translation_fault(unsigned long addr, unsigned int fsr,
	423	struct pt_regs *regs)
	424	{
1da177e4 LT	425	unsigned int index;
1da177e4 LT	426	pgd_t pgd, pgd_k;
516295e5	427	pud_t pud, pud_k;
1da177e4 LT	428	pmd_t pmd, pmd_k;
	429
	430	if (addr < TASK_SIZE)
	431	return do_page_fault(addr, fsr, regs);
	432
5e27fb78 A	433	if (user_mode(regs))
	434	goto bad_area;
	435
1da177e4 LT	436	index = pgd_index(addr);
1da177e4 LT	437
1da177e4 LT	438	pgd = cpu_get_pgd() + index;
	439	pgd_k = init_mm.pgd + index;
	440
	441	if (pgd_none(*pgd_k))
	442	goto bad_area;
1da177e4 LT	443	if (!pgd_present(*pgd))
	444	set_pgd(pgd, *pgd_k);
	445
516295e5 RK	446	pud = pud_offset(pgd, addr);
	447	pud_k = pud_offset(pgd_k, addr);
	448
	449	if (pud_none(*pud_k))
	450	goto bad_area;
	451	if (!pud_present(*pud))
	452	set_pud(pud, *pud_k);
	453
	454	pmd = pmd_offset(pud, addr);
	455	pmd_k = pmd_offset(pud_k, addr);
1da177e4	456
f7b8156d CM	457	#ifdef CONFIG_ARM_LPAE
	458	/*
	459	* Only one hardware entry per PMD with LPAE.
	460	*/
	461	index = 0;
	462	#else
33a9c41b KS	463	/*
	464	* On ARM one Linux PGD entry contains two hardware entries (see page
	465	* tables layout in pgtable.h). We normally guarantee that we always
	466	* fill both L1 entries. But create_mapping() doesn't follow the rule.
	467	* It can create inidividual L1 entries, so here we have to call
	468	* pmd_none() check for the entry really corresponded to address, not
	469	* for the first of pair.
	470	*/
	471	index = (addr >> SECTION_SHIFT) & 1;
f7b8156d	472	#endif
33a9c41b	473	if (pmd_none(pmd_k[index]))
1da177e4 LT	474	goto bad_area;
	475
	476	copy_pmd(pmd, pmd_k);
	477	return 0;
	478
	479	bad_area:
e5beac37	480	do_bad_area(addr, fsr, regs);
1da177e4 LT	481	return 0;
1da177e4 LT	482	}
09529f7a CM	483	#else /* CONFIG_MMU */
	484	static int
	485	do_translation_fault(unsigned long addr, unsigned int fsr,
	486	struct pt_regs *regs)
	487	{
	488	return 0;
	489	}
	490	#endif /* CONFIG_MMU */
1da177e4 LT	491
	492	/*
	493	* Some section permission faults need to be handled gracefully.
	494	* They can happen due to a __{get,put}_user during an oops.
	495	*/
809e660f	496	#ifndef CONFIG_ARM_LPAE
1da177e4 LT	497	static int
	498	do_sect_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	499	{
e5beac37	500	do_bad_area(addr, fsr, regs);
1da177e4 LT	501	return 0;
1da177e4 LT	502	}
809e660f	503	#endif /* CONFIG_ARM_LPAE */
1da177e4 LT	504
	505	/*
	506	* This abort handler always returns "fault".
	507	*/
	508	static int
	509	do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	510	{
	511	return 1;
	512	}
	513
136848d4	514	struct fsr_info {
1da177e4 LT	515	int (fn)(unsigned long addr, unsigned int fsr, struct pt_regs regs);
1da177e4 LT	516	int sig;
cfb0810e	517	int code;
1da177e4	518	const char *name;
1da177e4 LT	519	};
1da177e4 LT	520
136848d4	521	/* FSR definition */
f7b8156d CM	522	#ifdef CONFIG_ARM_LPAE
	523	#include "fsr-3level.c"
	524	#else
136848d4	525	#include "fsr-2level.c"
f7b8156d	526	#endif
136848d4	527
1da177e4 LT	528	void __init
1da177e4 LT	529	hook_fault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
6338a6aa	530	int sig, int code, const char *name)
1da177e4	531	{
6338a6aa KS	532	if (nr < 0 \|\| nr >= ARRAY_SIZE(fsr_info))
	533	BUG();
	534
	535	fsr_info[nr].fn = fn;
	536	fsr_info[nr].sig = sig;
	537	fsr_info[nr].code = code;
	538	fsr_info[nr].name = name;
1da177e4 LT	539	}
	540
	541	/*
	542	* Dispatch a data abort to the relevant handler.
	543	*/
7ab3f8d5	544	asmlinkage void __exception
1da177e4 LT	545	do_DataAbort(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
1da177e4 LT	546	{
c88d6aa7	547	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
cfb0810e	548	struct siginfo info;
1da177e4	549
df297bf6	550	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
1da177e4 LT	551	return;
1da177e4 LT	552
4ed89f22	553	pr_alert("Unhandled fault: %s (0x%03x) at 0x%08lx\n",
1da177e4	554	inf->name, fsr, addr);
6d021b72	555	show_pte(current->mm, addr);
cfb0810e RK	556
	557	info.si_signo = inf->sig;
	558	info.si_errno = 0;
	559	info.si_code = inf->code;
	560	info.si_addr = (void __user *)addr;
1eeb66a1	561	arm_notify_die("", regs, &info, fsr, 0);
1da177e4 LT	562	}
1da177e4 LT	563
3a4b5dca WD	564	void __init
	565	hook_ifault_code(int nr, int (fn)(unsigned long, unsigned int, struct pt_regs ),
	566	int sig, int code, const char *name)
	567	{
	568	if (nr < 0 \|\| nr >= ARRAY_SIZE(ifsr_info))
	569	BUG();
	570
	571	ifsr_info[nr].fn = fn;
	572	ifsr_info[nr].sig = sig;
	573	ifsr_info[nr].code = code;
	574	ifsr_info[nr].name = name;
	575	}
	576
7ab3f8d5	577	asmlinkage void __exception
4fb28474	578	do_PrefetchAbort(unsigned long addr, unsigned int ifsr, struct pt_regs *regs)
1da177e4	579	{
d25ef8b8 KS	580	const struct fsr_info *inf = ifsr_info + fsr_fs(ifsr);
	581	struct siginfo info;
	582
	583	if (!inf->fn(addr, ifsr \| FSR_LNX_PF, regs))
	584	return;
	585
4ed89f22	586	pr_alert("Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
d25ef8b8 KS	587	inf->name, ifsr, addr);
	588
	589	info.si_signo = inf->sig;
	590	info.si_errno = 0;
	591	info.si_code = inf->code;
	592	info.si_addr = (void __user *)addr;
	593	arm_notify_die("", regs, &info, ifsr, 0);
1da177e4 LT	594	}
1da177e4 LT	595
f7b8156d	596	#ifndef CONFIG_ARM_LPAE
993bf4ec KS	597	static int __init exceptions_init(void)
	598	{
	599	if (cpu_architecture() >= CPU_ARCH_ARMv6) {
	600	hook_fault_code(4, do_translation_fault, SIGSEGV, SEGV_MAPERR,
	601	"I-cache maintenance fault");
	602	}
	603
b8ab5397 KS	604	if (cpu_architecture() >= CPU_ARCH_ARMv7) {
	605	/*
	606	* TODO: Access flag faults introduced in ARMv6K.
	607	* Runtime check for 'K' extension is needed
	608	*/
	609	hook_fault_code(3, do_bad, SIGSEGV, SEGV_MAPERR,
	610	"section access flag fault");
	611	hook_fault_code(6, do_bad, SIGSEGV, SEGV_MAPERR,
	612	"section access flag fault");
	613	}
	614
993bf4ec KS	615	return 0;
	616	}
	617
	618	arch_initcall(exceptions_init);
f7b8156d	619	#endif