[net-next-2.6.git] / arch / mips / mm / fault.c

/*
 * This file is subject to the terms and conditions of the GNU General Public
 * License.  See the file "COPYING" in the main directory of this archive
 * for more details.
 *
 * Copyright (C) 1995 - 2000 by Ralf Baechle
 */
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/interrupt.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/smp.h>
#include <linux/module.h>
#include <linux/kprobes.h>
#include <linux/perf_event.h>

#include <asm/branch.h>
#include <asm/mmu_context.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/ptrace.h>
#include <asm/highmem.h>		/* For VMALLOC_END */
#include <linux/kdebug.h>

/*
 * This routine handles page faults.  It determines the address,
 * and the problem, and then passes it off to one of the appropriate
 * routines.
 */
asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, unsigned long write,
			      unsigned long address)
{
	struct vm_area_struct * vma = NULL;
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;
	const int field = sizeof(unsigned long) * 2;
	siginfo_t info;
	int fault;

#if 0
	printk("Cpu%d[%s:%d:%0*lx:%ld:%0*lx]\n", raw_smp_processor_id(),
	       current->comm, current->pid, field, address, write,
	       field, regs->cp0_epc);
#endif

#ifdef CONFIG_KPROBES
	/*
	 * This is to notify the fault handler of the kprobes.  The
	 * exception code is redundant as it is also carried in REGS,
	 * but we pass it anyhow.
	 */
	if (notify_die(DIE_PAGE_FAULT, "page fault", regs, -1,
		       (regs->cp0_cause >> 2) & 0x1f, SIGSEGV) == NOTIFY_STOP)
		return;
#endif

	info.si_code = SEGV_MAPERR;

	/*
	 * We fault-in kernel-space virtual memory on-demand. The
	 * 'reference' page table is init_mm.pgd.
	 *
	 * 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_64BIT
# define VMALLOC_FAULT_TARGET no_context
#else
# define VMALLOC_FAULT_TARGET vmalloc_fault
#endif

	if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END))
		goto VMALLOC_FAULT_TARGET;
#ifdef MODULE_START
	if (unlikely(address >= MODULE_START && address < MODULE_END))
		goto VMALLOC_FAULT_TARGET;
#endif

	/*
	 * If we're in an interrupt or have no user
	 * context, we must not take the fault..
	 */
	if (in_atomic() || !mm)
		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;
	if (expand_stack(vma, address))
		goto bad_area;
/*
 * Ok, we have a good vm_area for this memory access, so
 * we can handle it..
 */
good_area:
	info.si_code = SEGV_ACCERR;

	if (write) {
		if (!(vma->vm_flags & VM_WRITE))
			goto bad_area;
	} else {
		if (kernel_uses_smartmips_rixi) {
			if (address == regs->cp0_epc && !(vma->vm_flags & VM_EXEC)) {
#if 0
				pr_notice("Cpu%d[%s:%d:%0*lx:%ld:%0*lx] XI violation\n",
					  raw_smp_processor_id(),
					  current->comm, current->pid,
					  field, address, write,
					  field, regs->cp0_epc);
#endif
				goto bad_area;
			}
			if (!(vma->vm_flags & VM_READ)) {
#if 0
				pr_notice("Cpu%d[%s:%d:%0*lx:%ld:%0*lx] RI violation\n",
					  raw_smp_processor_id(),
					  current->comm, current->pid,
					  field, address, write,
					  field, regs->cp0_epc);
#endif
				goto bad_area;
			}
		} else {
			if (!(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)))
				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.
	 */
	fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0);
	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
	if (unlikely(fault & VM_FAULT_ERROR)) {
		if (fault & VM_FAULT_OOM)
			goto out_of_memory;
		else if (fault & VM_FAULT_SIGBUS)
			goto do_sigbus;
		BUG();
	}
	if (fault & VM_FAULT_MAJOR) {
		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ,
				1, 0, regs, address);
		tsk->maj_flt++;
	} else {
		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN,
				1, 0, regs, address);
		tsk->min_flt++;
	}

	up_read(&mm->mmap_sem);
	return;

/*
 * Something tried to access memory that isn't in our memory map..
 * Fix it, but check if it's kernel or user first..
 */
bad_area:
	up_read(&mm->mmap_sem);

bad_area_nosemaphore:
	/* User mode accesses just cause a SIGSEGV */
	if (user_mode(regs)) {
		tsk->thread.cp0_badvaddr = address;
		tsk->thread.error_code = write;
#if 0
		printk("do_page_fault() #2: sending SIGSEGV to %s for "
		       "invalid %s\n%0*lx (epc == %0*lx, ra == %0*lx)\n",
		       tsk->comm,
		       write ? "write access to" : "read access from",
		       field, address,
		       field, (unsigned long) regs->cp0_epc,
		       field, (unsigned long) regs->regs[31]);
#endif
		info.si_signo = SIGSEGV;
		info.si_errno = 0;
		/* info.si_code has been set above */
		info.si_addr = (void __user *) address;
		force_sig_info(SIGSEGV, &info, tsk);
		return;
	}

no_context:
	/* Are we prepared to handle this kernel fault?  */
	if (fixup_exception(regs)) {
		current->thread.cp0_baduaddr = address;
		return;
	}

	/*
	 * Oops. The kernel tried to access some bad page. We'll have to
	 * terminate things with extreme prejudice.
	 */
	bust_spinlocks(1);

	printk(KERN_ALERT "CPU %d Unable to handle kernel paging request at "
	       "virtual address %0*lx, epc == %0*lx, ra == %0*lx\n",
	       raw_smp_processor_id(), field, address, field, regs->cp0_epc,
	       field,  regs->regs[31]);
	die("Oops", regs);

out_of_memory:
	/*
	 * We ran out of memory, call the OOM killer, and return the userspace
	 * (which will retry the fault, or kill us if we got oom-killed).
	 */
	up_read(&mm->mmap_sem);
	pagefault_out_of_memory();
	return;

do_sigbus:
	up_read(&mm->mmap_sem);

	/* Kernel mode? Handle exceptions or die */
	if (!user_mode(regs))
		goto no_context;
	else
	/*
	 * Send a sigbus, regardless of whether we were in kernel
	 * or user mode.
	 */
#if 0
		printk("do_page_fault() #3: sending SIGBUS to %s for "
		       "invalid %s\n%0*lx (epc == %0*lx, ra == %0*lx)\n",
		       tsk->comm,
		       write ? "write access to" : "read access from",
		       field, address,
		       field, (unsigned long) regs->cp0_epc,
		       field, (unsigned long) regs->regs[31]);
#endif
	tsk->thread.cp0_badvaddr = address;
	info.si_signo = SIGBUS;
	info.si_errno = 0;
	info.si_code = BUS_ADRERR;
	info.si_addr = (void __user *) address;
	force_sig_info(SIGBUS, &info, tsk);

	return;
#ifndef CONFIG_64BIT
vmalloc_fault:
	{
		/*
		 * Synchronize this task's top level page-table
		 * with the 'reference' page table.
		 *
		 * Do _not_ use "tsk" here. We might be inside
		 * an interrupt in the middle of a task switch..
		 */
		int offset = __pgd_offset(address);
		pgd_t *pgd, *pgd_k;
		pud_t *pud, *pud_k;
		pmd_t *pmd, *pmd_k;
		pte_t *pte_k;

		pgd = (pgd_t *) pgd_current[raw_smp_processor_id()] + offset;
		pgd_k = init_mm.pgd + offset;

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

		pud = pud_offset(pgd, address);
		pud_k = pud_offset(pgd_k, address);
		if (!pud_present(*pud_k))
			goto no_context;

		pmd = pmd_offset(pud, address);
		pmd_k = pmd_offset(pud_k, address);
		if (!pmd_present(*pmd_k))
			goto no_context;
		set_pmd(pmd, *pmd_k);

		pte_k = pte_offset_kernel(pmd_k, address);
		if (!pte_present(*pte_k))
			goto no_context;
		return;
	}
#endif
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* This file is subject to the terms and conditions of the GNU General Public
	3	* License. See the file "COPYING" in the main directory of this archive
	4	* for more details.
	5	*
	6	* Copyright (C) 1995 - 2000 by Ralf Baechle
	7	*/
	8	#include <linux/signal.h>
	9	#include <linux/sched.h>
	10	#include <linux/interrupt.h>
	11	#include <linux/kernel.h>
	12	#include <linux/errno.h>
	13	#include <linux/string.h>
	14	#include <linux/types.h>
	15	#include <linux/ptrace.h>
	16	#include <linux/mman.h>
	17	#include <linux/mm.h>
	18	#include <linux/smp.h>
1da177e4	19	#include <linux/module.h>
c1bf207d	20	#include <linux/kprobes.h>
7f788d2d	21	#include <linux/perf_event.h>
1da177e4 LT	22
	23	#include <asm/branch.h>
	24	#include <asm/mmu_context.h>
	25	#include <asm/system.h>
	26	#include <asm/uaccess.h>
	27	#include <asm/ptrace.h>
16033d61	28	#include <asm/highmem.h> /* For VMALLOC_END */
c1bf207d	29	#include <linux/kdebug.h>
1da177e4 LT	30
	31	/*
	32	* This routine handles page faults. It determines the address,
	33	* and the problem, and then passes it off to one of the appropriate
	34	* routines.
	35	*/
c1bf207d	36	asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, unsigned long write,
1da177e4 LT	37	unsigned long address)
	38	{
	39	struct vm_area_struct * vma = NULL;
	40	struct task_struct *tsk = current;
	41	struct mm_struct *mm = tsk->mm;
	42	const int field = sizeof(unsigned long) * 2;
	43	siginfo_t info;
83c54070	44	int fault;
1da177e4 LT	45
1da177e4 LT	46	#if 0
d6f70360	47	printk("Cpu%d[%s:%d:%0lx:%ld:%0lx]\n", raw_smp_processor_id(),
1da177e4 LT	48	current->comm, current->pid, field, address, write,
	49	field, regs->cp0_epc);
	50	#endif
	51
c1bf207d DD	52	#ifdef CONFIG_KPROBES
	53	/*
	54	* This is to notify the fault handler of the kprobes. The
	55	* exception code is redundant as it is also carried in REGS,
	56	* but we pass it anyhow.
	57	*/
	58	if (notify_die(DIE_PAGE_FAULT, "page fault", regs, -1,
	59	(regs->cp0_cause >> 2) & 0x1f, SIGSEGV) == NOTIFY_STOP)
	60	return;
	61	#endif
	62
1da177e4 LT	63	info.si_code = SEGV_MAPERR;
	64
	65	/*
	66	* We fault-in kernel-space virtual memory on-demand. The
	67	* 'reference' page table is init_mm.pgd.
	68	*
	69	* NOTE! We MUST NOT take any locks for this case. We may
	70	* be in an interrupt or a critical region, and should
	71	* only copy the information from the master page table,
	72	* nothing more.
	73	*/
2ca2ebfd DD	74	#ifdef CONFIG_64BIT
	75	# define VMALLOC_FAULT_TARGET no_context
	76	#else
	77	# define VMALLOC_FAULT_TARGET vmalloc_fault
	78	#endif
	79
16033d61	80	if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END))
2ca2ebfd	81	goto VMALLOC_FAULT_TARGET;
656be92f AN	82	#ifdef MODULE_START
656be92f AN	83	if (unlikely(address >= MODULE_START && address < MODULE_END))
2ca2ebfd	84	goto VMALLOC_FAULT_TARGET;
656be92f	85	#endif
1da177e4 LT	86
	87	/*
	88	* If we're in an interrupt or have no user
	89	* context, we must not take the fault..
	90	*/
	91	if (in_atomic() \|\| !mm)
	92	goto bad_area_nosemaphore;
	93
	94	down_read(&mm->mmap_sem);
	95	vma = find_vma(mm, address);
	96	if (!vma)
	97	goto bad_area;
	98	if (vma->vm_start <= address)
	99	goto good_area;
	100	if (!(vma->vm_flags & VM_GROWSDOWN))
	101	goto bad_area;
	102	if (expand_stack(vma, address))
	103	goto bad_area;
	104	/*
	105	* Ok, we have a good vm_area for this memory access, so
	106	* we can handle it..
	107	*/
	108	good_area:
	109	info.si_code = SEGV_ACCERR;
	110
	111	if (write) {
	112	if (!(vma->vm_flags & VM_WRITE))
	113	goto bad_area;
	114	} else {
6dd9344c DD	115	if (kernel_uses_smartmips_rixi) {
	116	if (address == regs->cp0_epc && !(vma->vm_flags & VM_EXEC)) {
	117	#if 0
	118	pr_notice("Cpu%d[%s:%d:%0lx:%ld:%0lx] XI violation\n",
	119	raw_smp_processor_id(),
	120	current->comm, current->pid,
	121	field, address, write,
	122	field, regs->cp0_epc);
	123	#endif
	124	goto bad_area;
	125	}
	126	if (!(vma->vm_flags & VM_READ)) {
	127	#if 0
	128	pr_notice("Cpu%d[%s:%d:%0lx:%ld:%0lx] RI violation\n",
	129	raw_smp_processor_id(),
	130	current->comm, current->pid,
	131	field, address, write,
	132	field, regs->cp0_epc);
	133	#endif
	134	goto bad_area;
	135	}
	136	} else {
	137	if (!(vma->vm_flags & (VM_READ \| VM_WRITE \| VM_EXEC)))
	138	goto bad_area;
	139	}
1da177e4 LT	140	}
1da177e4 LT	141
1da177e4 LT	142	/*
	143	* If for any reason at all we couldn't handle the fault,
	144	* make sure we exit gracefully rather than endlessly redo
	145	* the fault.
	146	*/
d06063cc	147	fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0);
7f788d2d	148	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
83c54070 NP	149	if (unlikely(fault & VM_FAULT_ERROR)) {
	150	if (fault & VM_FAULT_OOM)
	151	goto out_of_memory;
	152	else if (fault & VM_FAULT_SIGBUS)
	153	goto do_sigbus;
1da177e4 LT	154	BUG();
1da177e4 LT	155	}
7f788d2d DCZ	156	if (fault & VM_FAULT_MAJOR) {
	157	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ,
	158	1, 0, regs, address);
83c54070	159	tsk->maj_flt++;
7f788d2d DCZ	160	} else {
	161	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN,
	162	1, 0, regs, address);
83c54070	163	tsk->min_flt++;
7f788d2d	164	}
1da177e4 LT	165
	166	up_read(&mm->mmap_sem);
	167	return;
	168
	169	/*
	170	* Something tried to access memory that isn't in our memory map..
	171	* Fix it, but check if it's kernel or user first..
	172	*/
	173	bad_area:
	174	up_read(&mm->mmap_sem);
	175
	176	bad_area_nosemaphore:
	177	/* User mode accesses just cause a SIGSEGV */
	178	if (user_mode(regs)) {
	179	tsk->thread.cp0_badvaddr = address;
	180	tsk->thread.error_code = write;
	181	#if 0
	182	printk("do_page_fault() #2: sending SIGSEGV to %s for "
	183	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	184	tsk->comm,
	185	write ? "write access to" : "read access from",
	186	field, address,
	187	field, (unsigned long) regs->cp0_epc,
	188	field, (unsigned long) regs->regs[31]);
	189	#endif
	190	info.si_signo = SIGSEGV;
	191	info.si_errno = 0;
	192	/* info.si_code has been set above */
fe00f943	193	info.si_addr = (void __user *) address;
1da177e4 LT	194	force_sig_info(SIGSEGV, &info, tsk);
	195	return;
	196	}
	197
	198	no_context:
	199	/* Are we prepared to handle this kernel fault? */
	200	if (fixup_exception(regs)) {
	201	current->thread.cp0_baduaddr = address;
	202	return;
	203	}
	204
	205	/*
	206	* Oops. The kernel tried to access some bad page. We'll have to
	207	* terminate things with extreme prejudice.
	208	*/
1da177e4 LT	209	bust_spinlocks(1);
	210
	211	printk(KERN_ALERT "CPU %d Unable to handle kernel paging request at "
	212	"virtual address %0lx, epc == %0lx, ra == %0*lx\n",
d6f70360	213	raw_smp_processor_id(), field, address, field, regs->cp0_epc,
1da177e4 LT	214	field, regs->regs[31]);
	215	die("Oops", regs);
	216
1da177e4	217	out_of_memory:
c7c1e384 RB	218	/*
	219	* We ran out of memory, call the OOM killer, and return the userspace
	220	* (which will retry the fault, or kill us if we got oom-killed).
	221	*/
a887b4da	222	up_read(&mm->mmap_sem);
c7c1e384 RB	223	pagefault_out_of_memory();
c7c1e384 RB	224	return;
1da177e4 LT	225
	226	do_sigbus:
	227	up_read(&mm->mmap_sem);
	228
	229	/* Kernel mode? Handle exceptions or die */
	230	if (!user_mode(regs))
	231	goto no_context;
41c594ab	232	else
1da177e4 LT	233	/*
	234	* Send a sigbus, regardless of whether we were in kernel
	235	* or user mode.
	236	*/
41c594ab RB	237	#if 0
	238	printk("do_page_fault() #3: sending SIGBUS to %s for "
	239	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	240	tsk->comm,
	241	write ? "write access to" : "read access from",
	242	field, address,
	243	field, (unsigned long) regs->cp0_epc,
	244	field, (unsigned long) regs->regs[31]);
	245	#endif
1da177e4 LT	246	tsk->thread.cp0_badvaddr = address;
	247	info.si_signo = SIGBUS;
	248	info.si_errno = 0;
	249	info.si_code = BUS_ADRERR;
fe00f943	250	info.si_addr = (void __user *) address;
1da177e4 LT	251	force_sig_info(SIGBUS, &info, tsk);
	252
	253	return;
2ca2ebfd	254	#ifndef CONFIG_64BIT
1da177e4 LT	255	vmalloc_fault:
	256	{
	257	/*
	258	* Synchronize this task's top level page-table
	259	* with the 'reference' page table.
	260	*
	261	* Do _not_ use "tsk" here. We might be inside
	262	* an interrupt in the middle of a task switch..
	263	*/
	264	int offset = __pgd_offset(address);
	265	pgd_t pgd, pgd_k;
c6e8b587	266	pud_t pud, pud_k;
1da177e4 LT	267	pmd_t pmd, pmd_k;
	268	pte_t *pte_k;
	269
d6f70360	270	pgd = (pgd_t *) pgd_current[raw_smp_processor_id()] + offset;
1da177e4 LT	271	pgd_k = init_mm.pgd + offset;
	272
	273	if (!pgd_present(*pgd_k))
	274	goto no_context;
	275	set_pgd(pgd, *pgd_k);
	276
c6e8b587 RB	277	pud = pud_offset(pgd, address);
	278	pud_k = pud_offset(pgd_k, address);
	279	if (!pud_present(*pud_k))
	280	goto no_context;
	281
	282	pmd = pmd_offset(pud, address);
	283	pmd_k = pmd_offset(pud_k, address);
1da177e4 LT	284	if (!pmd_present(*pmd_k))
	285	goto no_context;
	286	set_pmd(pmd, *pmd_k);
	287
	288	pte_k = pte_offset_kernel(pmd_k, address);
	289	if (!pte_present(*pte_k))
	290	goto no_context;
	291	return;
	292	}
2ca2ebfd	293	#endif
1da177e4	294	}