[mirror_ubuntu-artful-kernel.git] / arch / powerpc / mm / fault.c

/*
 *  PowerPC version
 *    Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
 *
 *  Derived from "arch/i386/mm/fault.c"
 *    Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *
 *  Modified by Cort Dougan and Paul Mackerras.
 *
 *  Modified for PPC64 by Dave Engebretsen (engebret@ibm.com)
 *
 *  This program is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU General Public License
 *  as published by the Free Software Foundation; either version
 *  2 of the License, or (at your option) any later version.
 */

#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/ptrace.h>
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/highmem.h>
#include <linux/module.h>
#include <linux/kprobes.h>
#include <linux/kdebug.h>

#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
#include <asm/mmu_context.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/tlbflush.h>
#include <asm/siginfo.h>


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

	/* kprobe_running() needs smp_processor_id() */
	if (!user_mode(regs)) {
		preempt_disable();
		if (kprobe_running() && kprobe_fault_handler(regs, 11))
			ret = 1;
		preempt_enable();
	}

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

/*
 * Check whether the instruction at regs->nip is a store using
 * an update addressing form which will update r1.
 */
static int store_updates_sp(struct pt_regs *regs)
{
	unsigned int inst;

	if (get_user(inst, (unsigned int __user *)regs->nip))
		return 0;
	/* check for 1 in the rA field */
	if (((inst >> 16) & 0x1f) != 1)
		return 0;
	/* check major opcode */
	switch (inst >> 26) {
	case 37:	/* stwu */
	case 39:	/* stbu */
	case 45:	/* sthu */
	case 53:	/* stfsu */
	case 55:	/* stfdu */
		return 1;
	case 62:	/* std or stdu */
		return (inst & 3) == 1;
	case 31:
		/* check minor opcode */
		switch ((inst >> 1) & 0x3ff) {
		case 181:	/* stdux */
		case 183:	/* stwux */
		case 247:	/* stbux */
		case 439:	/* sthux */
		case 695:	/* stfsux */
		case 759:	/* stfdux */
			return 1;
		}
	}
	return 0;
}

#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
static void do_dabr(struct pt_regs *regs, unsigned long address,
		    unsigned long error_code)
{
	siginfo_t info;

	if (notify_die(DIE_DABR_MATCH, "dabr_match", regs, error_code,
			11, SIGSEGV) == NOTIFY_STOP)
		return;

	if (debugger_dabr_match(regs))
		return;

	/* Clear the DABR */
	set_dabr(0);

	/* Deliver the signal to userspace */
	info.si_signo = SIGTRAP;
	info.si_errno = 0;
	info.si_code = TRAP_HWBKPT;
	info.si_addr = (void __user *)address;
	force_sig_info(SIGTRAP, &info, current);
}
#endif /* !(CONFIG_4xx || CONFIG_BOOKE)*/

/*
 * For 600- and 800-family processors, the error_code parameter is DSISR
 * for a data fault, SRR1 for an instruction fault. For 400-family processors
 * the error_code parameter is ESR for a data fault, 0 for an instruction
 * fault.
 * For 64-bit processors, the error_code parameter is
 *  - DSISR for a non-SLB data access fault,
 *  - SRR1 & 0x08000000 for a non-SLB instruction access fault
 *  - 0 any SLB fault.
 *
 * The return value is 0 if the fault was handled, or the signal
 * number if this is a kernel fault that can't be handled here.
 */
int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
			    unsigned long error_code)
{
	struct vm_area_struct * vma;
	struct mm_struct *mm = current->mm;
	siginfo_t info;
	int code = SEGV_MAPERR;
	int is_write = 0, ret;
	int trap = TRAP(regs);
 	int is_exec = trap == 0x400;

#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
	/*
	 * Fortunately the bit assignments in SRR1 for an instruction
	 * fault and DSISR for a data fault are mostly the same for the
	 * bits we are interested in.  But there are some bits which
	 * indicate errors in DSISR but can validly be set in SRR1.
	 */
	if (trap == 0x400)
		error_code &= 0x48200000;
	else
		is_write = error_code & DSISR_ISSTORE;
#else
	is_write = error_code & ESR_DST;
#endif /* CONFIG_4xx || CONFIG_BOOKE */

	if (notify_page_fault(regs))
		return 0;

	if (trap == 0x300) {
		if (debugger_fault_handler(regs))
			return 0;
	}

	/* On a kernel SLB miss we can only check for a valid exception entry */
	if (!user_mode(regs) && (address >= TASK_SIZE))
		return SIGSEGV;

#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
  	if (error_code & DSISR_DABRMATCH) {
		/* DABR match */
		do_dabr(regs, address, error_code);
		return 0;
	}
#endif /* !(CONFIG_4xx || CONFIG_BOOKE)*/

	if (in_atomic() || mm == NULL) {
		if (!user_mode(regs))
			return SIGSEGV;
		/* in_atomic() in user mode is really bad,
		   as is current->mm == NULL. */
		printk(KERN_EMERG "Page fault in user mode with"
		       "in_atomic() = %d mm = %p\n", in_atomic(), mm);
		printk(KERN_EMERG "NIP = %lx  MSR = %lx\n",
		       regs->nip, regs->msr);
		die("Weird page fault", regs, SIGSEGV);
	}

	/* When running in the kernel we expect faults to occur only to
	 * addresses in user space.  All other faults represent errors in the
	 * kernel and should generate an OOPS.  Unfortunately, in the case of an
	 * erroneous fault occurring in a code path which already holds mmap_sem
	 * we will deadlock attempting to validate the fault against the
	 * address space.  Luckily the kernel only validly references user
	 * space from well defined areas of code, which are listed in the
	 * exceptions table.
	 *
	 * As the vast majority of faults will be valid we will only perform
	 * the source reference check when there is a possibility of a deadlock.
	 * Attempt to lock the address space, if we cannot we then validate the
	 * source.  If this is invalid we can skip the address space check,
	 * thus avoiding the deadlock.
	 */
	if (!down_read_trylock(&mm->mmap_sem)) {
		if (!user_mode(regs) && !search_exception_tables(regs->nip))
			goto bad_area_nosemaphore;

		down_read(&mm->mmap_sem);
	}

	vma = find_vma(mm, address);
	if (!vma)
		goto bad_area;
	if (vma->vm_start <= address)
		goto good_area;
	if (!(vma->vm_flags & VM_GROWSDOWN))
		goto bad_area;

	/*
	 * N.B. The POWER/Open ABI allows programs to access up to
	 * 288 bytes below the stack pointer.
	 * The kernel signal delivery code writes up to about 1.5kB
	 * below the stack pointer (r1) before decrementing it.
	 * The exec code can write slightly over 640kB to the stack
	 * before setting the user r1.  Thus we allow the stack to
	 * expand to 1MB without further checks.
	 */
	if (address + 0x100000 < vma->vm_end) {
		/* get user regs even if this fault is in kernel mode */
		struct pt_regs *uregs = current->thread.regs;
		if (uregs == NULL)
			goto bad_area;

		/*
		 * A user-mode access to an address a long way below
		 * the stack pointer is only valid if the instruction
		 * is one which would update the stack pointer to the
		 * address accessed if the instruction completed,
		 * i.e. either stwu rs,n(r1) or stwux rs,r1,rb
		 * (or the byte, halfword, float or double forms).
		 *
		 * If we don't check this then any write to the area
		 * between the last mapped region and the stack will
		 * expand the stack rather than segfaulting.
		 */
		if (address + 2048 < uregs->gpr[1]
		    && (!user_mode(regs) || !store_updates_sp(regs)))
			goto bad_area;
	}
	if (expand_stack(vma, address))
		goto bad_area;

good_area:
	code = SEGV_ACCERR;
#if defined(CONFIG_6xx)
	if (error_code & 0x95700000)
		/* an error such as lwarx to I/O controller space,
		   address matching DABR, eciwx, etc. */
		goto bad_area;
#endif /* CONFIG_6xx */
#if defined(CONFIG_8xx)
        /* The MPC8xx seems to always set 0x80000000, which is
         * "undefined".  Of those that can be set, this is the only
         * one which seems bad.
         */
	if (error_code & 0x10000000)
                /* Guarded storage error. */
		goto bad_area;
#endif /* CONFIG_8xx */

	if (is_exec) {
#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
		/* protection fault */
		if (error_code & DSISR_PROTFAULT)
			goto bad_area;
		/*
		 * Allow execution from readable areas if the MMU does not
		 * provide separate controls over reading and executing.
		 */
		if (!(vma->vm_flags & VM_EXEC) &&
		    (cpu_has_feature(CPU_FTR_NOEXECUTE) ||
		     !(vma->vm_flags & (VM_READ | VM_WRITE))))
			goto bad_area;
#else
		pte_t *ptep;
		pmd_t *pmdp;

		/* Since 4xx/Book-E supports per-page execute permission,
		 * we lazily flush dcache to icache. */
		ptep = NULL;
		if (get_pteptr(mm, address, &ptep, &pmdp)) {
			spinlock_t *ptl = pte_lockptr(mm, pmdp);
			spin_lock(ptl);
			if (pte_present(*ptep)) {
				struct page *page = pte_page(*ptep);

				if (!test_bit(PG_arch_1, &page->flags)) {
					flush_dcache_icache_page(page);
					set_bit(PG_arch_1, &page->flags);
				}
				pte_update(ptep, 0, _PAGE_HWEXEC);
				_tlbie(address);
				pte_unmap_unlock(ptep, ptl);
				up_read(&mm->mmap_sem);
				return 0;
			}
			pte_unmap_unlock(ptep, ptl);
		}
#endif
	/* a write */
	} else if (is_write) {
		if (!(vma->vm_flags & VM_WRITE))
			goto bad_area;
	/* a read */
	} else {
		/* protection fault */
		if (error_code & 0x08000000)
			goto bad_area;
		if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))
			goto bad_area;
	}

	/*
	 * If for any reason at all we couldn't handle the fault,
	 * make sure we exit gracefully rather than endlessly redo
	 * the fault.
	 */
 survive:
	ret = handle_mm_fault(mm, vma, address, is_write);
	if (unlikely(ret & VM_FAULT_ERROR)) {
		if (ret & VM_FAULT_OOM)
			goto out_of_memory;
		else if (ret & VM_FAULT_SIGBUS)
			goto do_sigbus;
		BUG();
	}
	if (ret & VM_FAULT_MAJOR)
		current->maj_flt++;
	else
		current->min_flt++;
	up_read(&mm->mmap_sem);
	return 0;

bad_area:
	up_read(&mm->mmap_sem);

bad_area_nosemaphore:
	/* User mode accesses cause a SIGSEGV */
	if (user_mode(regs)) {
		_exception(SIGSEGV, regs, code, address);
		return 0;
	}

	if (is_exec && (error_code & DSISR_PROTFAULT)
	    && printk_ratelimit())
		printk(KERN_CRIT "kernel tried to execute NX-protected"
		       " page (%lx) - exploit attempt? (uid: %d)\n",
		       address, current->uid);

	return SIGSEGV;

/*
 * We ran out of memory, or some other thing happened to us that made
 * us unable to handle the page fault gracefully.
 */
out_of_memory:
	up_read(&mm->mmap_sem);
	if (is_global_init(current)) {
		yield();
		down_read(&mm->mmap_sem);
		goto survive;
	}
	printk("VM: killing process %s\n", current->comm);
	if (user_mode(regs))
		do_group_exit(SIGKILL);
	return SIGKILL;

do_sigbus:
	up_read(&mm->mmap_sem);
	if (user_mode(regs)) {
		info.si_signo = SIGBUS;
		info.si_errno = 0;
		info.si_code = BUS_ADRERR;
		info.si_addr = (void __user *)address;
		force_sig_info(SIGBUS, &info, current);
		return 0;
	}
	return SIGBUS;
}

/*
 * bad_page_fault is called when we have a bad access from the kernel.
 * It is called from the DSI and ISI handlers in head.S and from some
 * of the procedures in traps.c.
 */
void bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
{
	const struct exception_table_entry *entry;

	/* Are we prepared to handle this fault?  */
	if ((entry = search_exception_tables(regs->nip)) != NULL) {
		regs->nip = entry->fixup;
		return;
	}

	/* kernel has accessed a bad area */

	switch (regs->trap) {
	case 0x300:
	case 0x380:
		printk(KERN_ALERT "Unable to handle kernel paging request for "
			"data at address 0x%08lx\n", regs->dar);
		break;
	case 0x400:
	case 0x480:
		printk(KERN_ALERT "Unable to handle kernel paging request for "
			"instruction fetch\n");
		break;
	default:
		printk(KERN_ALERT "Unable to handle kernel paging request for "
			"unknown fault\n");
		break;
	}
	printk(KERN_ALERT "Faulting instruction address: 0x%08lx\n",
		regs->nip);

	die("Kernel access of bad area", regs, sig);
}
Commit	Line	Data
14cf11af	1	/*
14cf11af PM	2	* PowerPC version
	3	* Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
	4	*
	5	* Derived from "arch/i386/mm/fault.c"
	6	* Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
	7	*
	8	* Modified by Cort Dougan and Paul Mackerras.
	9	*
	10	* Modified for PPC64 by Dave Engebretsen (engebret@ibm.com)
	11	*
	12	* This program is free software; you can redistribute it and/or
	13	* modify it under the terms of the GNU General Public License
	14	* as published by the Free Software Foundation; either version
	15	* 2 of the License, or (at your option) any later version.
	16	*/
	17
14cf11af PM	18	#include <linux/signal.h>
	19	#include <linux/sched.h>
	20	#include <linux/kernel.h>
	21	#include <linux/errno.h>
	22	#include <linux/string.h>
	23	#include <linux/types.h>
	24	#include <linux/ptrace.h>
	25	#include <linux/mman.h>
	26	#include <linux/mm.h>
	27	#include <linux/interrupt.h>
	28	#include <linux/highmem.h>
	29	#include <linux/module.h>
	30	#include <linux/kprobes.h>
1eeb66a1	31	#include <linux/kdebug.h>
14cf11af PM	32
	33	#include <asm/page.h>
	34	#include <asm/pgtable.h>
	35	#include <asm/mmu.h>
	36	#include <asm/mmu_context.h>
	37	#include <asm/system.h>
	38	#include <asm/uaccess.h>
	39	#include <asm/tlbflush.h>
14cf11af PM	40	#include <asm/siginfo.h>
14cf11af PM	41
4f9e87c0	42
9f90b997 CH	43	#ifdef CONFIG_KPROBES
9f90b997 CH	44	static inline int notify_page_fault(struct pt_regs *regs)
4f9e87c0	45	{
9f90b997 CH	46	int ret = 0;
	47
	48	/* kprobe_running() needs smp_processor_id() */
	49	if (!user_mode(regs)) {
	50	preempt_disable();
	51	if (kprobe_running() && kprobe_fault_handler(regs, 11))
	52	ret = 1;
	53	preempt_enable();
	54	}
4f9e87c0	55
9f90b997	56	return ret;
4f9e87c0 AK	57	}
4f9e87c0 AK	58	#else
9f90b997	59	static inline int notify_page_fault(struct pt_regs *regs)
4f9e87c0	60	{
9f90b997	61	return 0;
4f9e87c0 AK	62	}
	63	#endif
	64
14cf11af PM	65	/*
	66	* Check whether the instruction at regs->nip is a store using
	67	* an update addressing form which will update r1.
	68	*/
	69	static int store_updates_sp(struct pt_regs *regs)
	70	{
	71	unsigned int inst;
	72
	73	if (get_user(inst, (unsigned int __user *)regs->nip))
	74	return 0;
	75	/* check for 1 in the rA field */
	76	if (((inst >> 16) & 0x1f) != 1)
	77	return 0;
	78	/* check major opcode */
	79	switch (inst >> 26) {
	80	case 37: /* stwu */
	81	case 39: /* stbu */
	82	case 45: /* sthu */
	83	case 53: /* stfsu */
	84	case 55: /* stfdu */
	85	return 1;
	86	case 62: /* std or stdu */
	87	return (inst & 3) == 1;
	88	case 31:
	89	/* check minor opcode */
	90	switch ((inst >> 1) & 0x3ff) {
	91	case 181: /* stdux */
	92	case 183: /* stwux */
	93	case 247: /* stbux */
	94	case 439: /* sthux */
	95	case 695: /* stfsux */
	96	case 759: /* stfdux */
	97	return 1;
	98	}
	99	}
	100	return 0;
	101	}
	102
cffb09ce	103	#if !(defined(CONFIG_4xx) \|\| defined(CONFIG_BOOKE))
bce6c5fd AB	104	static void do_dabr(struct pt_regs *regs, unsigned long address,
bce6c5fd AB	105	unsigned long error_code)
14cf11af PM	106	{
	107	siginfo_t info;
	108
	109	if (notify_die(DIE_DABR_MATCH, "dabr_match", regs, error_code,
	110	11, SIGSEGV) == NOTIFY_STOP)
	111	return;
	112
	113	if (debugger_dabr_match(regs))
	114	return;
	115
	116	/* Clear the DABR */
	117	set_dabr(0);
	118
	119	/* Deliver the signal to userspace */
	120	info.si_signo = SIGTRAP;
	121	info.si_errno = 0;
	122	info.si_code = TRAP_HWBKPT;
bce6c5fd	123	info.si_addr = (void __user *)address;
14cf11af PM	124	force_sig_info(SIGTRAP, &info, current);
14cf11af PM	125	}
cffb09ce	126	#endif /* !(CONFIG_4xx \|\| CONFIG_BOOKE)*/
14cf11af PM	127
	128	/*
	129	* For 600- and 800-family processors, the error_code parameter is DSISR
	130	* for a data fault, SRR1 for an instruction fault. For 400-family processors
	131	* the error_code parameter is ESR for a data fault, 0 for an instruction
	132	* fault.
	133	* For 64-bit processors, the error_code parameter is
	134	* - DSISR for a non-SLB data access fault,
	135	* - SRR1 & 0x08000000 for a non-SLB instruction access fault
	136	* - 0 any SLB fault.
	137	*
	138	* The return value is 0 if the fault was handled, or the signal
	139	* number if this is a kernel fault that can't be handled here.
	140	*/
	141	int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
	142	unsigned long error_code)
	143	{
	144	struct vm_area_struct * vma;
	145	struct mm_struct *mm = current->mm;
	146	siginfo_t info;
	147	int code = SEGV_MAPERR;
83c54070	148	int is_write = 0, ret;
14cf11af PM	149	int trap = TRAP(regs);
	150	int is_exec = trap == 0x400;
	151
	152	#if !(defined(CONFIG_4xx) \|\| defined(CONFIG_BOOKE))
	153	/*
	154	* Fortunately the bit assignments in SRR1 for an instruction
	155	* fault and DSISR for a data fault are mostly the same for the
	156	* bits we are interested in. But there are some bits which
	157	* indicate errors in DSISR but can validly be set in SRR1.
	158	*/
	159	if (trap == 0x400)
	160	error_code &= 0x48200000;
	161	else
	162	is_write = error_code & DSISR_ISSTORE;
	163	#else
	164	is_write = error_code & ESR_DST;
	165	#endif /* CONFIG_4xx \|\| CONFIG_BOOKE */
	166
9f90b997	167	if (notify_page_fault(regs))
14cf11af PM	168	return 0;
	169
	170	if (trap == 0x300) {
	171	if (debugger_fault_handler(regs))
	172	return 0;
	173	}
	174
	175	/* On a kernel SLB miss we can only check for a valid exception entry */
	176	if (!user_mode(regs) && (address >= TASK_SIZE))
	177	return SIGSEGV;
	178
	179	#if !(defined(CONFIG_4xx) \|\| defined(CONFIG_BOOKE))
	180	if (error_code & DSISR_DABRMATCH) {
	181	/* DABR match */
bce6c5fd	182	do_dabr(regs, address, error_code);
14cf11af PM	183	return 0;
	184	}
	185	#endif /* !(CONFIG_4xx \|\| CONFIG_BOOKE)*/
	186
	187	if (in_atomic() \|\| mm == NULL) {
	188	if (!user_mode(regs))
	189	return SIGSEGV;
	190	/* in_atomic() in user mode is really bad,
	191	as is current->mm == NULL. */
	192	printk(KERN_EMERG "Page fault in user mode with"
	193	"in_atomic() = %d mm = %p\n", in_atomic(), mm);
	194	printk(KERN_EMERG "NIP = %lx MSR = %lx\n",
	195	regs->nip, regs->msr);
	196	die("Weird page fault", regs, SIGSEGV);
	197	}
	198
	199	/* When running in the kernel we expect faults to occur only to
	200	* addresses in user space. All other faults represent errors in the
fc5266ea AB	201	* kernel and should generate an OOPS. Unfortunately, in the case of an
fc5266ea AB	202	* erroneous fault occurring in a code path which already holds mmap_sem
14cf11af PM	203	* we will deadlock attempting to validate the fault against the
	204	* address space. Luckily the kernel only validly references user
	205	* space from well defined areas of code, which are listed in the
	206	* exceptions table.
	207	*
	208	* As the vast majority of faults will be valid we will only perform
fc5266ea	209	* the source reference check when there is a possibility of a deadlock.
14cf11af PM	210	* Attempt to lock the address space, if we cannot we then validate the
	211	* source. If this is invalid we can skip the address space check,
	212	* thus avoiding the deadlock.
	213	*/
	214	if (!down_read_trylock(&mm->mmap_sem)) {
	215	if (!user_mode(regs) && !search_exception_tables(regs->nip))
	216	goto bad_area_nosemaphore;
	217
	218	down_read(&mm->mmap_sem);
	219	}
	220
	221	vma = find_vma(mm, address);
	222	if (!vma)
	223	goto bad_area;
	224	if (vma->vm_start <= address)
	225	goto good_area;
	226	if (!(vma->vm_flags & VM_GROWSDOWN))
	227	goto bad_area;
	228
	229	/*
	230	* N.B. The POWER/Open ABI allows programs to access up to
	231	* 288 bytes below the stack pointer.
	232	* The kernel signal delivery code writes up to about 1.5kB
	233	* below the stack pointer (r1) before decrementing it.
	234	* The exec code can write slightly over 640kB to the stack
	235	* before setting the user r1. Thus we allow the stack to
	236	* expand to 1MB without further checks.
	237	*/
	238	if (address + 0x100000 < vma->vm_end) {
	239	/* get user regs even if this fault is in kernel mode */
	240	struct pt_regs *uregs = current->thread.regs;
	241	if (uregs == NULL)
	242	goto bad_area;
	243
	244	/*
	245	* A user-mode access to an address a long way below
	246	* the stack pointer is only valid if the instruction
	247	* is one which would update the stack pointer to the
	248	* address accessed if the instruction completed,
	249	* i.e. either stwu rs,n(r1) or stwux rs,r1,rb
	250	* (or the byte, halfword, float or double forms).
	251	*
	252	* If we don't check this then any write to the area
	253	* between the last mapped region and the stack will
	254	* expand the stack rather than segfaulting.
	255	*/
	256	if (address + 2048 < uregs->gpr[1]
	257	&& (!user_mode(regs) \|\| !store_updates_sp(regs)))
	258	goto bad_area;
	259	}
	260	if (expand_stack(vma, address))
	261	goto bad_area;
	262
	263	good_area:
	264	code = SEGV_ACCERR;
	265	#if defined(CONFIG_6xx)
	266	if (error_code & 0x95700000)
	267	/* an error such as lwarx to I/O controller space,
	268	address matching DABR, eciwx, etc. */
	269	goto bad_area;
	270	#endif /* CONFIG_6xx */
	271	#if defined(CONFIG_8xx)
	272	/* The MPC8xx seems to always set 0x80000000, which is
	273	* "undefined". Of those that can be set, this is the only
274	* one which seems bad.
275	*/
276	if (error_code & 0x10000000)
277	/* Guarded storage error. */
278	goto bad_area;
279	#endif /* CONFIG_8xx */
280
281	if (is_exec) {
9ba4ace3	282	#if !(defined(CONFIG_4xx) \|\| defined(CONFIG_BOOKE))
14cf11af PM	283	/* protection fault */
	284	if (error_code & DSISR_PROTFAULT)
	285	goto bad_area;
08ae6cc1 PM	286	/*
	287	* Allow execution from readable areas if the MMU does not
	288	* provide separate controls over reading and executing.
	289	*/
	290	if (!(vma->vm_flags & VM_EXEC) &&
	291	(cpu_has_feature(CPU_FTR_NOEXECUTE) \|\|
	292	!(vma->vm_flags & (VM_READ \| VM_WRITE))))
14cf11af	293	goto bad_area;
9ba4ace3	294	#else
14cf11af	295	pte_t *ptep;
bab70a4a	296	pmd_t *pmdp;
14cf11af PM	297
	298	/* Since 4xx/Book-E supports per-page execute permission,
	299	* we lazily flush dcache to icache. */
	300	ptep = NULL;
bab70a4a ES	301	if (get_pteptr(mm, address, &ptep, &pmdp)) {
	302	spinlock_t *ptl = pte_lockptr(mm, pmdp);
	303	spin_lock(ptl);
	304	if (pte_present(*ptep)) {
	305	struct page page = pte_page(ptep);
14cf11af	306
bab70a4a ES	307	if (!test_bit(PG_arch_1, &page->flags)) {
	308	flush_dcache_icache_page(page);
	309	set_bit(PG_arch_1, &page->flags);
	310	}
	311	pte_update(ptep, 0, _PAGE_HWEXEC);
	312	_tlbie(address);
	313	pte_unmap_unlock(ptep, ptl);
	314	up_read(&mm->mmap_sem);
	315	return 0;
14cf11af	316	}
bab70a4a	317	pte_unmap_unlock(ptep, ptl);
14cf11af	318	}
14cf11af PM	319	#endif
	320	/* a write */
	321	} else if (is_write) {
	322	if (!(vma->vm_flags & VM_WRITE))
	323	goto bad_area;
	324	/* a read */
	325	} else {
	326	/* protection fault */
	327	if (error_code & 0x08000000)
	328	goto bad_area;
df67b3da	329	if (!(vma->vm_flags & (VM_READ \| VM_EXEC \| VM_WRITE)))
14cf11af PM	330	goto bad_area;
	331	}
	332
	333	/*
	334	* If for any reason at all we couldn't handle the fault,
	335	* make sure we exit gracefully rather than endlessly redo
	336	* the fault.
	337	*/
	338	survive:
83c54070 NP	339	ret = handle_mm_fault(mm, vma, address, is_write);
	340	if (unlikely(ret & VM_FAULT_ERROR)) {
	341	if (ret & VM_FAULT_OOM)
	342	goto out_of_memory;
	343	else if (ret & VM_FAULT_SIGBUS)
	344	goto do_sigbus;
14cf11af PM	345	BUG();
14cf11af PM	346	}
83c54070 NP	347	if (ret & VM_FAULT_MAJOR)
	348	current->maj_flt++;
	349	else
	350	current->min_flt++;
14cf11af PM	351	up_read(&mm->mmap_sem);
	352	return 0;
	353
	354	bad_area:
	355	up_read(&mm->mmap_sem);
	356
	357	bad_area_nosemaphore:
	358	/* User mode accesses cause a SIGSEGV */
	359	if (user_mode(regs)) {
	360	_exception(SIGSEGV, regs, code, address);
	361	return 0;
	362	}
	363
	364	if (is_exec && (error_code & DSISR_PROTFAULT)
	365	&& printk_ratelimit())
	366	printk(KERN_CRIT "kernel tried to execute NX-protected"
	367	" page (%lx) - exploit attempt? (uid: %d)\n",
	368	address, current->uid);
	369
	370	return SIGSEGV;
	371
	372	/*
	373	* We ran out of memory, or some other thing happened to us that made
	374	* us unable to handle the page fault gracefully.
	375	*/
	376	out_of_memory:
	377	up_read(&mm->mmap_sem);
b460cbc5	378	if (is_global_init(current)) {
14cf11af PM	379	yield();
	380	down_read(&mm->mmap_sem);
	381	goto survive;
	382	}
	383	printk("VM: killing process %s\n", current->comm);
	384	if (user_mode(regs))
effe24bd	385	do_group_exit(SIGKILL);
14cf11af PM	386	return SIGKILL;
	387
	388	do_sigbus:
	389	up_read(&mm->mmap_sem);
	390	if (user_mode(regs)) {
	391	info.si_signo = SIGBUS;
	392	info.si_errno = 0;
	393	info.si_code = BUS_ADRERR;
	394	info.si_addr = (void __user *)address;
	395	force_sig_info(SIGBUS, &info, current);
	396	return 0;
	397	}
	398	return SIGBUS;
	399	}
	400
	401	/*
	402	* bad_page_fault is called when we have a bad access from the kernel.
	403	* It is called from the DSI and ISI handlers in head.S and from some
	404	* of the procedures in traps.c.
	405	*/
	406	void bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
	407	{
	408	const struct exception_table_entry *entry;
	409
	410	/* Are we prepared to handle this fault? */
	411	if ((entry = search_exception_tables(regs->nip)) != NULL) {
	412	regs->nip = entry->fixup;
	413	return;
	414	}
	415
	416	/* kernel has accessed a bad area */
723925b7	417
723925b7	418	switch (regs->trap) {
a416dd8d ME	419	case 0x300:
	420	case 0x380:
	421	printk(KERN_ALERT "Unable to handle kernel paging request for "
	422	"data at address 0x%08lx\n", regs->dar);
	423	break;
	424	case 0x400:
	425	case 0x480:
	426	printk(KERN_ALERT "Unable to handle kernel paging request for "
	427	"instruction fetch\n");
	428	break;
	429	default:
	430	printk(KERN_ALERT "Unable to handle kernel paging request for "
	431	"unknown fault\n");
	432	break;
723925b7 OJ	433	}
	434	printk(KERN_ALERT "Faulting instruction address: 0x%08lx\n",
	435	regs->nip);
	436
14cf11af PM	437	die("Kernel access of bad area", regs, sig);
14cf11af PM	438	}