[deliverable/linux.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/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;
	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE |
						 (write ? FAULT_FLAG_WRITE : 0);

#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;

retry:
	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, flags);

	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
		return;

	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 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 (flags & FAULT_FLAG_ALLOW_RETRY) {
		if (fault & VM_FAULT_MAJOR) {
			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
						  regs, address);
			tsk->maj_flt++;
		} else {
			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
						  regs, address);
			tsk->min_flt++;
		}
		if (fault & VM_FAULT_RETRY) {
			flags &= ~FAULT_FLAG_ALLOW_RETRY;

			/*
			 * No need to up_read(&mm->mmap_sem) as we would
			 * have already released it in __lock_page_or_retry
			 * in mm/filemap.c.
			 */

			goto retry;
		}
	}

	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>
1da177e4 LT	25	#include <asm/uaccess.h>
1da177e4 LT	26	#include <asm/ptrace.h>
16033d61	27	#include <asm/highmem.h> /* For VMALLOC_END */
c1bf207d	28	#include <linux/kdebug.h>
1da177e4 LT	29
	30	/*
	31	* This routine handles page faults. It determines the address,
	32	* and the problem, and then passes it off to one of the appropriate
	33	* routines.
	34	*/
c1bf207d	35	asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, unsigned long write,
1da177e4 LT	36	unsigned long address)
	37	{
	38	struct vm_area_struct * vma = NULL;
	39	struct task_struct *tsk = current;
	40	struct mm_struct *mm = tsk->mm;
	41	const int field = sizeof(unsigned long) * 2;
	42	siginfo_t info;
83c54070	43	int fault;
43ca4957 KC	44	unsigned int flags = FAULT_FLAG_ALLOW_RETRY \| FAULT_FLAG_KILLABLE \|
43ca4957 KC	45	(write ? FAULT_FLAG_WRITE : 0);
1da177e4 LT	46
1da177e4 LT	47	#if 0
d6f70360	48	printk("Cpu%d[%s:%d:%0lx:%ld:%0lx]\n", raw_smp_processor_id(),
1da177e4 LT	49	current->comm, current->pid, field, address, write,
	50	field, regs->cp0_epc);
	51	#endif
	52
c1bf207d DD	53	#ifdef CONFIG_KPROBES
	54	/*
	55	* This is to notify the fault handler of the kprobes. The
	56	* exception code is redundant as it is also carried in REGS,
	57	* but we pass it anyhow.
	58	*/
	59	if (notify_die(DIE_PAGE_FAULT, "page fault", regs, -1,
	60	(regs->cp0_cause >> 2) & 0x1f, SIGSEGV) == NOTIFY_STOP)
	61	return;
	62	#endif
	63
1da177e4 LT	64	info.si_code = SEGV_MAPERR;
	65
	66	/*
	67	* We fault-in kernel-space virtual memory on-demand. The
	68	* 'reference' page table is init_mm.pgd.
	69	*
	70	* NOTE! We MUST NOT take any locks for this case. We may
	71	* be in an interrupt or a critical region, and should
	72	* only copy the information from the master page table,
	73	* nothing more.
	74	*/
2ca2ebfd DD	75	#ifdef CONFIG_64BIT
	76	# define VMALLOC_FAULT_TARGET no_context
	77	#else
	78	# define VMALLOC_FAULT_TARGET vmalloc_fault
	79	#endif
	80
16033d61	81	if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END))
2ca2ebfd	82	goto VMALLOC_FAULT_TARGET;
656be92f AN	83	#ifdef MODULE_START
656be92f AN	84	if (unlikely(address >= MODULE_START && address < MODULE_END))
2ca2ebfd	85	goto VMALLOC_FAULT_TARGET;
656be92f	86	#endif
1da177e4 LT	87
	88	/*
	89	* If we're in an interrupt or have no user
	90	* context, we must not take the fault..
	91	*/
	92	if (in_atomic() \|\| !mm)
	93	goto bad_area_nosemaphore;
	94
43ca4957	95	retry:
1da177e4 LT	96	down_read(&mm->mmap_sem);
	97	vma = find_vma(mm, address);
	98	if (!vma)
	99	goto bad_area;
	100	if (vma->vm_start <= address)
	101	goto good_area;
	102	if (!(vma->vm_flags & VM_GROWSDOWN))
	103	goto bad_area;
	104	if (expand_stack(vma, address))
	105	goto bad_area;
	106	/*
	107	* Ok, we have a good vm_area for this memory access, so
	108	* we can handle it..
	109	*/
	110	good_area:
	111	info.si_code = SEGV_ACCERR;
	112
	113	if (write) {
	114	if (!(vma->vm_flags & VM_WRITE))
	115	goto bad_area;
	116	} else {
6dd9344c DD	117	if (kernel_uses_smartmips_rixi) {
	118	if (address == regs->cp0_epc && !(vma->vm_flags & VM_EXEC)) {
	119	#if 0
	120	pr_notice("Cpu%d[%s:%d:%0lx:%ld:%0lx] XI violation\n",
	121	raw_smp_processor_id(),
	122	current->comm, current->pid,
	123	field, address, write,
	124	field, regs->cp0_epc);
	125	#endif
	126	goto bad_area;
	127	}
	128	if (!(vma->vm_flags & VM_READ)) {
	129	#if 0
	130	pr_notice("Cpu%d[%s:%d:%0lx:%ld:%0lx] RI violation\n",
	131	raw_smp_processor_id(),
	132	current->comm, current->pid,
	133	field, address, write,
	134	field, regs->cp0_epc);
	135	#endif
	136	goto bad_area;
	137	}
	138	} else {
	139	if (!(vma->vm_flags & (VM_READ \| VM_WRITE \| VM_EXEC)))
	140	goto bad_area;
	141	}
1da177e4 LT	142	}
1da177e4 LT	143
1da177e4 LT	144	/*
	145	* If for any reason at all we couldn't handle the fault,
	146	* make sure we exit gracefully rather than endlessly redo
	147	* the fault.
	148	*/
43ca4957 KC	149	fault = handle_mm_fault(mm, vma, address, flags);
	150
	151	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
	152	return;
	153
a8b0ca17	154	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
83c54070 NP	155	if (unlikely(fault & VM_FAULT_ERROR)) {
	156	if (fault & VM_FAULT_OOM)
	157	goto out_of_memory;
	158	else if (fault & VM_FAULT_SIGBUS)
	159	goto do_sigbus;
1da177e4 LT	160	BUG();
1da177e4 LT	161	}
43ca4957 KC	162	if (flags & FAULT_FLAG_ALLOW_RETRY) {
	163	if (fault & VM_FAULT_MAJOR) {
	164	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
	165	regs, address);
	166	tsk->maj_flt++;
	167	} else {
	168	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
	169	regs, address);
	170	tsk->min_flt++;
	171	}
	172	if (fault & VM_FAULT_RETRY) {
	173	flags &= ~FAULT_FLAG_ALLOW_RETRY;
	174
	175	/*
	176	* No need to up_read(&mm->mmap_sem) as we would
	177	* have already released it in __lock_page_or_retry
	178	* in mm/filemap.c.
	179	*/
	180
	181	goto retry;
	182	}
7f788d2d	183	}
1da177e4 LT	184
	185	up_read(&mm->mmap_sem);
	186	return;
	187
	188	/*
	189	* Something tried to access memory that isn't in our memory map..
	190	* Fix it, but check if it's kernel or user first..
	191	*/
	192	bad_area:
	193	up_read(&mm->mmap_sem);
	194
	195	bad_area_nosemaphore:
	196	/* User mode accesses just cause a SIGSEGV */
	197	if (user_mode(regs)) {
	198	tsk->thread.cp0_badvaddr = address;
	199	tsk->thread.error_code = write;
	200	#if 0
	201	printk("do_page_fault() #2: sending SIGSEGV to %s for "
	202	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	203	tsk->comm,
	204	write ? "write access to" : "read access from",
	205	field, address,
	206	field, (unsigned long) regs->cp0_epc,
	207	field, (unsigned long) regs->regs[31]);
	208	#endif
	209	info.si_signo = SIGSEGV;
	210	info.si_errno = 0;
	211	/* info.si_code has been set above */
fe00f943	212	info.si_addr = (void __user *) address;
1da177e4 LT	213	force_sig_info(SIGSEGV, &info, tsk);
	214	return;
	215	}
	216
	217	no_context:
	218	/* Are we prepared to handle this kernel fault? */
	219	if (fixup_exception(regs)) {
	220	current->thread.cp0_baduaddr = address;
	221	return;
	222	}
	223
	224	/*
	225	* Oops. The kernel tried to access some bad page. We'll have to
	226	* terminate things with extreme prejudice.
	227	*/
1da177e4 LT	228	bust_spinlocks(1);
	229
	230	printk(KERN_ALERT "CPU %d Unable to handle kernel paging request at "
	231	"virtual address %0lx, epc == %0lx, ra == %0*lx\n",
d6f70360	232	raw_smp_processor_id(), field, address, field, regs->cp0_epc,
1da177e4 LT	233	field, regs->regs[31]);
	234	die("Oops", regs);
	235
1da177e4	236	out_of_memory:
c7c1e384 RB	237	/*
	238	* We ran out of memory, call the OOM killer, and return the userspace
	239	* (which will retry the fault, or kill us if we got oom-killed).
	240	*/
a887b4da	241	up_read(&mm->mmap_sem);
c7c1e384 RB	242	pagefault_out_of_memory();
c7c1e384 RB	243	return;
1da177e4 LT	244
	245	do_sigbus:
	246	up_read(&mm->mmap_sem);
	247
	248	/* Kernel mode? Handle exceptions or die */
	249	if (!user_mode(regs))
	250	goto no_context;
41c594ab	251	else
1da177e4 LT	252	/*
	253	* Send a sigbus, regardless of whether we were in kernel
	254	* or user mode.
	255	*/
41c594ab RB	256	#if 0
	257	printk("do_page_fault() #3: sending SIGBUS to %s for "
	258	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	259	tsk->comm,
	260	write ? "write access to" : "read access from",
	261	field, address,
	262	field, (unsigned long) regs->cp0_epc,
	263	field, (unsigned long) regs->regs[31]);
	264	#endif
1da177e4 LT	265	tsk->thread.cp0_badvaddr = address;
	266	info.si_signo = SIGBUS;
	267	info.si_errno = 0;
	268	info.si_code = BUS_ADRERR;
fe00f943	269	info.si_addr = (void __user *) address;
1da177e4 LT	270	force_sig_info(SIGBUS, &info, tsk);
	271
	272	return;
2ca2ebfd	273	#ifndef CONFIG_64BIT
1da177e4 LT	274	vmalloc_fault:
	275	{
	276	/*
	277	* Synchronize this task's top level page-table
	278	* with the 'reference' page table.
	279	*
	280	* Do _not_ use "tsk" here. We might be inside
	281	* an interrupt in the middle of a task switch..
	282	*/
	283	int offset = __pgd_offset(address);
	284	pgd_t pgd, pgd_k;
c6e8b587	285	pud_t pud, pud_k;
1da177e4 LT	286	pmd_t pmd, pmd_k;
	287	pte_t *pte_k;
	288
d6f70360	289	pgd = (pgd_t *) pgd_current[raw_smp_processor_id()] + offset;
1da177e4 LT	290	pgd_k = init_mm.pgd + offset;
	291
	292	if (!pgd_present(*pgd_k))
	293	goto no_context;
	294	set_pgd(pgd, *pgd_k);
	295
c6e8b587 RB	296	pud = pud_offset(pgd, address);
	297	pud_k = pud_offset(pgd_k, address);
	298	if (!pud_present(*pud_k))
	299	goto no_context;
	300
	301	pmd = pmd_offset(pud, address);
	302	pmd_k = pmd_offset(pud_k, address);
1da177e4 LT	303	if (!pmd_present(*pmd_k))
	304	goto no_context;
	305	set_pmd(pmd, *pmd_k);
	306
	307	pte_k = pte_offset_kernel(pmd_k, address);
	308	if (!pte_present(*pte_k))
	309	goto no_context;
	310	return;
	311	}
2ca2ebfd	312	#endif
1da177e4	313	}