[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/vt_kern.h>		/* For unblank_screen() */
#include <linux/module.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 */

/*
 * 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 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

	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.
	 */
	if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END))
		goto vmalloc_fault;
#ifdef MODULE_START
	if (unlikely(address >= MODULE_START && address < MODULE_END))
		goto vmalloc_fault;
#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 (!(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);
	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)
		tsk->maj_flt++;
	else
		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;
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;
	}
}
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 LT	19	#include <linux/vt_kern.h> /* For unblank_screen() */
	20	#include <linux/module.h>
	21
	22	#include <asm/branch.h>
	23	#include <asm/mmu_context.h>
	24	#include <asm/system.h>
	25	#include <asm/uaccess.h>
	26	#include <asm/ptrace.h>
16033d61	27	#include <asm/highmem.h> /* For VMALLOC_END */
1da177e4 LT	28
	29	/*
	30	* This routine handles page faults. It determines the address,
	31	* and the problem, and then passes it off to one of the appropriate
	32	* routines.
	33	*/
	34	asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long write,
	35	unsigned long address)
	36	{
	37	struct vm_area_struct * vma = NULL;
	38	struct task_struct *tsk = current;
	39	struct mm_struct *mm = tsk->mm;
	40	const int field = sizeof(unsigned long) * 2;
	41	siginfo_t info;
83c54070	42	int fault;
1da177e4 LT	43
1da177e4 LT	44	#if 0
d6f70360	45	printk("Cpu%d[%s:%d:%0lx:%ld:%0lx]\n", raw_smp_processor_id(),
1da177e4 LT	46	current->comm, current->pid, field, address, write,
	47	field, regs->cp0_epc);
	48	#endif
	49
	50	info.si_code = SEGV_MAPERR;
	51
	52	/*
	53	* We fault-in kernel-space virtual memory on-demand. The
	54	* 'reference' page table is init_mm.pgd.
	55	*
	56	* NOTE! We MUST NOT take any locks for this case. We may
	57	* be in an interrupt or a critical region, and should
	58	* only copy the information from the master page table,
	59	* nothing more.
	60	*/
16033d61	61	if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END))
1da177e4	62	goto vmalloc_fault;
656be92f AN	63	#ifdef MODULE_START
	64	if (unlikely(address >= MODULE_START && address < MODULE_END))
	65	goto vmalloc_fault;
	66	#endif
1da177e4 LT	67
	68	/*
	69	* If we're in an interrupt or have no user
	70	* context, we must not take the fault..
	71	*/
	72	if (in_atomic() \|\| !mm)
	73	goto bad_area_nosemaphore;
	74
	75	down_read(&mm->mmap_sem);
	76	vma = find_vma(mm, address);
	77	if (!vma)
	78	goto bad_area;
	79	if (vma->vm_start <= address)
	80	goto good_area;
	81	if (!(vma->vm_flags & VM_GROWSDOWN))
	82	goto bad_area;
	83	if (expand_stack(vma, address))
	84	goto bad_area;
	85	/*
	86	* Ok, we have a good vm_area for this memory access, so
	87	* we can handle it..
	88	*/
	89	good_area:
	90	info.si_code = SEGV_ACCERR;
	91
	92	if (write) {
	93	if (!(vma->vm_flags & VM_WRITE))
	94	goto bad_area;
	95	} else {
00932ba3	96	if (!(vma->vm_flags & (VM_READ \| VM_WRITE \| VM_EXEC)))
1da177e4 LT	97	goto bad_area;
	98	}
	99
1da177e4 LT	100	/*
	101	* If for any reason at all we couldn't handle the fault,
	102	* make sure we exit gracefully rather than endlessly redo
	103	* the fault.
	104	*/
d06063cc	105	fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0);
83c54070 NP	106	if (unlikely(fault & VM_FAULT_ERROR)) {
	107	if (fault & VM_FAULT_OOM)
	108	goto out_of_memory;
	109	else if (fault & VM_FAULT_SIGBUS)
	110	goto do_sigbus;
1da177e4 LT	111	BUG();
1da177e4 LT	112	}
83c54070 NP	113	if (fault & VM_FAULT_MAJOR)
	114	tsk->maj_flt++;
	115	else
	116	tsk->min_flt++;
1da177e4 LT	117
	118	up_read(&mm->mmap_sem);
	119	return;
	120
	121	/*
	122	* Something tried to access memory that isn't in our memory map..
	123	* Fix it, but check if it's kernel or user first..
	124	*/
	125	bad_area:
	126	up_read(&mm->mmap_sem);
	127
	128	bad_area_nosemaphore:
	129	/* User mode accesses just cause a SIGSEGV */
	130	if (user_mode(regs)) {
	131	tsk->thread.cp0_badvaddr = address;
	132	tsk->thread.error_code = write;
	133	#if 0
	134	printk("do_page_fault() #2: sending SIGSEGV to %s for "
	135	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	136	tsk->comm,
	137	write ? "write access to" : "read access from",
	138	field, address,
	139	field, (unsigned long) regs->cp0_epc,
	140	field, (unsigned long) regs->regs[31]);
	141	#endif
	142	info.si_signo = SIGSEGV;
	143	info.si_errno = 0;
	144	/* info.si_code has been set above */
fe00f943	145	info.si_addr = (void __user *) address;
1da177e4 LT	146	force_sig_info(SIGSEGV, &info, tsk);
	147	return;
	148	}
	149
	150	no_context:
	151	/* Are we prepared to handle this kernel fault? */
	152	if (fixup_exception(regs)) {
	153	current->thread.cp0_baduaddr = address;
	154	return;
	155	}
	156
	157	/*
	158	* Oops. The kernel tried to access some bad page. We'll have to
	159	* terminate things with extreme prejudice.
	160	*/
1da177e4 LT	161	bust_spinlocks(1);
	162
	163	printk(KERN_ALERT "CPU %d Unable to handle kernel paging request at "
	164	"virtual address %0lx, epc == %0lx, ra == %0*lx\n",
d6f70360	165	raw_smp_processor_id(), field, address, field, regs->cp0_epc,
1da177e4 LT	166	field, regs->regs[31]);
	167	die("Oops", regs);
	168
1da177e4	169	out_of_memory:
c7c1e384 RB	170	/*
	171	* We ran out of memory, call the OOM killer, and return the userspace
	172	* (which will retry the fault, or kill us if we got oom-killed).
	173	*/
a887b4da	174	up_read(&mm->mmap_sem);
c7c1e384 RB	175	pagefault_out_of_memory();
c7c1e384 RB	176	return;
1da177e4 LT	177
	178	do_sigbus:
	179	up_read(&mm->mmap_sem);
	180
	181	/* Kernel mode? Handle exceptions or die */
	182	if (!user_mode(regs))
	183	goto no_context;
41c594ab	184	else
1da177e4 LT	185	/*
	186	* Send a sigbus, regardless of whether we were in kernel
	187	* or user mode.
	188	*/
41c594ab RB	189	#if 0
	190	printk("do_page_fault() #3: sending SIGBUS to %s for "
	191	"invalid %s\n%0lx (epc == %0lx, ra == %0*lx)\n",
	192	tsk->comm,
	193	write ? "write access to" : "read access from",
	194	field, address,
	195	field, (unsigned long) regs->cp0_epc,
	196	field, (unsigned long) regs->regs[31]);
	197	#endif
1da177e4 LT	198	tsk->thread.cp0_badvaddr = address;
	199	info.si_signo = SIGBUS;
	200	info.si_errno = 0;
	201	info.si_code = BUS_ADRERR;
fe00f943	202	info.si_addr = (void __user *) address;
1da177e4 LT	203	force_sig_info(SIGBUS, &info, tsk);
	204
	205	return;
1da177e4 LT	206	vmalloc_fault:
	207	{
	208	/*
	209	* Synchronize this task's top level page-table
	210	* with the 'reference' page table.
	211	*
	212	* Do _not_ use "tsk" here. We might be inside
	213	* an interrupt in the middle of a task switch..
	214	*/
	215	int offset = __pgd_offset(address);
	216	pgd_t pgd, pgd_k;
c6e8b587	217	pud_t pud, pud_k;
1da177e4 LT	218	pmd_t pmd, pmd_k;
	219	pte_t *pte_k;
	220
d6f70360	221	pgd = (pgd_t *) pgd_current[raw_smp_processor_id()] + offset;
1da177e4 LT	222	pgd_k = init_mm.pgd + offset;
	223
	224	if (!pgd_present(*pgd_k))
	225	goto no_context;
	226	set_pgd(pgd, *pgd_k);
	227
c6e8b587 RB	228	pud = pud_offset(pgd, address);
	229	pud_k = pud_offset(pgd_k, address);
	230	if (!pud_present(*pud_k))
	231	goto no_context;
	232
	233	pmd = pmd_offset(pud, address);
	234	pmd_k = pmd_offset(pud_k, address);
1da177e4 LT	235	if (!pmd_present(*pmd_k))
	236	goto no_context;
	237	set_pmd(pmd, *pmd_k);
	238
	239	pte_k = pte_offset_kernel(pmd_k, address);
	240	if (!pte_present(*pte_k))
	241	goto no_context;
	242	return;
	243	}
	244	}