[mirror_ubuntu-hirsute-kernel.git] / arch / unicore32 / mm / fault.c

/*
 * linux/arch/unicore32/mm/fault.c
 *
 * Code specific to PKUnity SoC and UniCore ISA
 *
 * Copyright (C) 2001-2010 GUAN Xue-tao
 *
 * 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/extable.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/signal.h>
#include <linux/io.h>

#include <asm/pgtable.h>
#include <asm/tlbflush.h>

/*
 * Fault status register encodings.  We steal bit 31 for our own purposes.
 */
#define FSR_LNX_PF		(1 << 31)

static inline int fsr_fs(unsigned int fsr)
{
	/* xyabcde will be abcde+xy */
	return (fsr & 31) + ((fsr & (3 << 5)) >> 5);
}

/*
 * 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;

	printk(KERN_ALERT "pgd = %p\n", mm->pgd);
	pgd = pgd_offset(mm, addr);
	printk(KERN_ALERT "[%08lx] *pgd=%08lx", addr, pgd_val(*pgd));

	do {
		pmd_t *pmd;
		pte_t *pte;

		if (pgd_none(*pgd))
			break;

		if (pgd_bad(*pgd)) {
			printk("(bad)");
			break;
		}

		pmd = pmd_offset((pud_t *) pgd, addr);
		if (PTRS_PER_PMD != 1)
			printk(", *pmd=%08lx", pmd_val(*pmd));

		if (pmd_none(*pmd))
			break;

		if (pmd_bad(*pmd)) {
			printk("(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);
		printk(", *pte=%08lx", pte_val(*pte));
		pte_unmap(pte);
	} while (0);

	printk("\n");
}

/*
 * 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);
	printk(KERN_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;

	tsk->thread.address = addr;
	tsk->thread.error_code = fsr;
	tsk->thread.trap_no = 14;
	clear_siginfo(&si);
	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);
}

#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 ^ 0x12))	/* write? */
		mask = VM_WRITE;
	if (fsr & FSR_LNX_PF)
		mask = VM_EXEC;

	return vma->vm_flags & mask ? false : true;
}

static int __do_pf(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;
	}

	/*
	 * 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(vma, addr & PAGE_MASK, flags);
	return fault;

check_stack:
	if (vma->vm_flags & VM_GROWSDOWN && !expand_stack(vma, addr))
		goto good_area;
out:
	return fault;
}

static int do_pf(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;

	tsk = current;
	mm = tsk->mm;

	/*
	 * 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 ^ 0x12))
		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->UCreg_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->UCreg_pc))
			goto no_context;
#endif
	}

	fault = __do_pf(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;

	if (!(fault & VM_FAULT_ERROR) && (flags & FAULT_FLAG_ALLOW_RETRY)) {
		if (fault & VM_FAULT_MAJOR)
			tsk->maj_flt++;
		else
			tsk->min_flt++;
		if (fault & VM_FAULT_RETRY) {
			/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
			* of starvation. */
			flags &= ~FAULT_FLAG_ALLOW_RETRY;
			goto retry;
		}
	}

	up_read(&mm->mmap_sem);

	/*
	 * Handle the "normal" case first - VM_FAULT_MAJOR
	 */
	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;
}

/*
 * 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.
 */
static int do_ifault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	unsigned int index;
	pgd_t *pgd, *pgd_k;
	pmd_t *pmd, *pmd_k;

	if (addr < TASK_SIZE)
		return do_pf(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;

	pmd_k = pmd_offset((pud_t *) pgd_k, addr);
	pmd = pmd_offset((pud_t *) pgd, addr);

	if (pmd_none(*pmd_k))
		goto bad_area;

	set_pmd(pmd, *pmd_k);
	flush_pmd_entry(pmd);
	return 0;

bad_area:
	do_bad_area(addr, fsr, regs);
	return 0;
}

/*
 * This abort handler always returns "fault".
 */
static int do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	return 1;
}

static int do_good(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
{
	unsigned int res1, res2;

	printk("dabt exception but no error!\n");

	__asm__ __volatile__(
			"mff %0,f0\n"
			"mff %1,f1\n"
			: "=r"(res1), "=r"(res2)
			:
			: "memory");

	printk(KERN_EMERG "r0 :%08x  r1 :%08x\n", res1, res2);
	panic("shut up\n");
	return 0;
}

static struct fsr_info {
	int (*fn) (unsigned long addr, unsigned int fsr, struct pt_regs *regs);
	int sig;
	int code;
	const char *name;
} fsr_info[] = {
	/*
	 * The following are the standard Unicore-I and UniCore-II aborts.
	 */
	{ do_good,	SIGBUS,  0,		"no error"		},
	{ do_bad,	SIGBUS,  BUS_ADRALN,	"alignment exception"	},
	{ do_bad,	SIGBUS,  BUS_OBJERR,	"external exception"	},
	{ do_bad,	SIGBUS,  0,		"burst operation"	},
	{ do_bad,	SIGBUS,  0,		"unknown 00100"		},
	{ do_ifault,	SIGSEGV, SEGV_MAPERR,	"2nd level pt non-exist"},
	{ do_bad,	SIGBUS,  0,		"2nd lvl large pt non-exist" },
	{ do_bad,	SIGBUS,  0,		"invalid pte"		},
	{ do_pf,	SIGSEGV, SEGV_MAPERR,	"page miss"		},
	{ do_bad,	SIGBUS,  0,		"middle page miss"	},
	{ do_bad,	SIGBUS,	 0,		"large page miss"	},
	{ do_pf,	SIGSEGV, SEGV_MAPERR,	"super page (section) miss" },
	{ do_bad,	SIGBUS,  0,		"unknown 01100"		},
	{ do_bad,	SIGBUS,  0,		"unknown 01101"		},
	{ do_bad,	SIGBUS,  0,		"unknown 01110"		},
	{ do_bad,	SIGBUS,  0,		"unknown 01111"		},
	{ do_bad,	SIGBUS,  0,		"addr: up 3G or IO"	},
	{ do_pf,	SIGSEGV, SEGV_ACCERR,	"read unreadable addr"	},
	{ do_pf,	SIGSEGV, SEGV_ACCERR,	"write unwriteable addr"},
	{ do_pf,	SIGSEGV, SEGV_ACCERR,	"exec unexecutable addr"},
	{ do_bad,	SIGBUS,  0,		"unknown 10100"		},
	{ do_bad,	SIGBUS,  0,		"unknown 10101"		},
	{ do_bad,	SIGBUS,  0,		"unknown 10110"		},
	{ do_bad,	SIGBUS,  0,		"unknown 10111"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11000"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11001"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11010"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11011"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11100"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11101"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11110"		},
	{ do_bad,	SIGBUS,  0,		"unknown 11111"		}
};

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 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;

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

	clear_siginfo(&info);
	info.si_signo = inf->sig;
	info.si_errno = 0;
	info.si_code = inf->code;
	info.si_addr = (void __user *)addr;
	uc32_notify_die("", regs, &info, fsr, 0);
}

asmlinkage void do_PrefetchAbort(unsigned long addr,
			unsigned int ifsr, struct pt_regs *regs)
{
	const struct fsr_info *inf = fsr_info + fsr_fs(ifsr);
	struct siginfo info;

	if (!inf->fn(addr, ifsr | FSR_LNX_PF, regs))
		return;

	printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
	       inf->name, ifsr, addr);

	clear_siginfo(&info);
	info.si_signo = inf->sig;
	info.si_errno = 0;
	info.si_code = inf->code;
	info.si_addr = (void __user *)addr;
	uc32_notify_die("", regs, &info, ifsr, 0);
}
Commit	Line	Data
56372b0b G	1	/*
	2	* linux/arch/unicore32/mm/fault.c
	3	*
	4	* Code specific to PKUnity SoC and UniCore ISA
	5	*
	6	* Copyright (C) 2001-2010 GUAN Xue-tao
	7	*
	8	* This program is free software; you can redistribute it and/or modify
	9	* it under the terms of the GNU General Public License version 2 as
	10	* published by the Free Software Foundation.
	11	*/
0f91f3f6	12	#include <linux/extable.h>
56372b0b G	13	#include <linux/signal.h>
	14	#include <linux/mm.h>
	15	#include <linux/hardirq.h>
	16	#include <linux/init.h>
	17	#include <linux/kprobes.h>
	18	#include <linux/uaccess.h>
	19	#include <linux/page-flags.h>
3f07c014	20	#include <linux/sched/signal.h>
56372b0b G	21	#include <linux/io.h>
56372b0b G	22
56372b0b G	23	#include <asm/pgtable.h>
	24	#include <asm/tlbflush.h>
	25
	26	/*
	27	* Fault status register encodings. We steal bit 31 for our own purposes.
	28	*/
	29	#define FSR_LNX_PF (1 << 31)
	30
	31	static inline int fsr_fs(unsigned int fsr)
	32	{
	33	/* xyabcde will be abcde+xy */
	34	return (fsr & 31) + ((fsr & (3 << 5)) >> 5);
	35	}
	36
	37	/*
	38	* This is useful to dump out the page tables associated with
	39	* 'addr' in mm 'mm'.
	40	*/
	41	void show_pte(struct mm_struct *mm, unsigned long addr)
	42	{
	43	pgd_t *pgd;
	44
	45	if (!mm)
	46	mm = &init_mm;
	47
	48	printk(KERN_ALERT "pgd = %p\n", mm->pgd);
	49	pgd = pgd_offset(mm, addr);
	50	printk(KERN_ALERT "[%08lx] pgd=%08lx", addr, pgd_val(pgd));
	51
	52	do {
	53	pmd_t *pmd;
	54	pte_t *pte;
	55
	56	if (pgd_none(*pgd))
	57	break;
	58
	59	if (pgd_bad(*pgd)) {
	60	printk("(bad)");
	61	break;
	62	}
	63
	64	pmd = pmd_offset((pud_t *) pgd, addr);
	65	if (PTRS_PER_PMD != 1)
	66	printk(", pmd=%08lx", pmd_val(pmd));
	67
	68	if (pmd_none(*pmd))
	69	break;
	70
	71	if (pmd_bad(*pmd)) {
	72	printk("(bad)");
	73	break;
	74	}
	75
	76	/* We must not map this if we have highmem enabled */
	77	if (PageHighMem(pfn_to_page(pmd_val(*pmd) >> PAGE_SHIFT)))
	78	break;
	79
	80	pte = pte_offset_map(pmd, addr);
	81	printk(", pte=%08lx", pte_val(pte));
	82	pte_unmap(pte);
	83	} while (0);
	84
	85	printk("\n");
	86	}
87
88	/*
89	* Oops. The kernel tried to access some page that wasn't present.
90	*/
91	static void __do_kernel_fault(struct mm_struct *mm, unsigned long addr,
92	unsigned int fsr, struct pt_regs *regs)
93	{
94	/*
95	* Are we prepared to handle this kernel fault?
96	*/
97	if (fixup_exception(regs))
98	return;
99
100	/*
101	* No handler, we'll have to terminate things with extreme prejudice.
102	*/
103	bust_spinlocks(1);
104	printk(KERN_ALERT
105	"Unable to handle kernel %s at virtual address %08lx\n",
106	(addr < PAGE_SIZE) ? "NULL pointer dereference" :
107	"paging request", addr);
108
109	show_pte(mm, addr);
110	die("Oops", regs, fsr);
111	bust_spinlocks(0);
112	do_exit(SIGKILL);
113	}
114
115	/*
116	* Something tried to access memory that isn't in our memory map..
117	* User mode accesses just cause a SIGSEGV
118	*/
119	static void __do_user_fault(struct task_struct *tsk, unsigned long addr,
120	unsigned int fsr, unsigned int sig, int code,
121	struct pt_regs *regs)
122	{
123	struct siginfo si;
124
125	tsk->thread.address = addr;
126	tsk->thread.error_code = fsr;
127	tsk->thread.trap_no = 14;
3eb0f519	128	clear_siginfo(&si);
56372b0b G	129	si.si_signo = sig;
	130	si.si_errno = 0;
	131	si.si_code = code;
	132	si.si_addr = (void __user *)addr;
	133	force_sig_info(sig, &si, tsk);
	134	}
	135
	136	void do_bad_area(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	137	{
	138	struct task_struct *tsk = current;
	139	struct mm_struct *mm = tsk->active_mm;
	140
	141	/*
	142	* If we are in kernel mode at this point, we
	143	* have no context to handle this fault with.
	144	*/
	145	if (user_mode(regs))
	146	__do_user_fault(tsk, addr, fsr, SIGSEGV, SEGV_MAPERR, regs);
	147	else
	148	__do_kernel_fault(mm, addr, fsr, regs);
	149	}
	150
	151	#define VM_FAULT_BADMAP 0x010000
	152	#define VM_FAULT_BADACCESS 0x020000
	153
	154	/*
	155	* Check that the permissions on the VMA allow for the fault which occurred.
	156	* If we encountered a write fault, we must have write permission, otherwise
	157	* we allow any permission.
	158	*/
	159	static inline bool access_error(unsigned int fsr, struct vm_area_struct *vma)
	160	{
	161	unsigned int mask = VM_READ \| VM_WRITE \| VM_EXEC;
	162
	163	if (!(fsr ^ 0x12)) /* write? */
	164	mask = VM_WRITE;
	165	if (fsr & FSR_LNX_PF)
	166	mask = VM_EXEC;
	167
	168	return vma->vm_flags & mask ? false : true;
	169	}
	170
	171	static int __do_pf(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
f3f09d5a	172	unsigned int flags, struct task_struct *tsk)
56372b0b G	173	{
	174	struct vm_area_struct *vma;
	175	int fault;
	176
	177	vma = find_vma(mm, addr);
	178	fault = VM_FAULT_BADMAP;
	179	if (unlikely(!vma))
	180	goto out;
	181	if (unlikely(vma->vm_start > addr))
	182	goto check_stack;
	183
	184	/*
	185	* Ok, we have a good vm_area for this
	186	* memory access, so we can handle it.
	187	*/
	188	good_area:
	189	if (access_error(fsr, vma)) {
	190	fault = VM_FAULT_BADACCESS;
	191	goto out;
	192	}
	193
	194	/*
	195	* If for any reason at all we couldn't handle the fault, make
	196	* sure we exit gracefully rather than endlessly redo the fault.
	197	*/
dcddffd4	198	fault = handle_mm_fault(vma, addr & PAGE_MASK, flags);
56372b0b G	199	return fault;
	200
	201	check_stack:
	202	if (vma->vm_flags & VM_GROWSDOWN && !expand_stack(vma, addr))
	203	goto good_area;
	204	out:
	205	return fault;
	206	}
	207
	208	static int do_pf(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	209	{
	210	struct task_struct *tsk;
	211	struct mm_struct *mm;
	212	int fault, sig, code;
759496ba	213	unsigned int flags = FAULT_FLAG_ALLOW_RETRY \| FAULT_FLAG_KILLABLE;
56372b0b G	214
	215	tsk = current;
	216	mm = tsk->mm;
	217
	218	/*
	219	* If we're in an interrupt or have no user
	220	* context, we must not take the fault..
	221	*/
70ffdb93	222	if (faulthandler_disabled() \|\| !mm)
56372b0b G	223	goto no_context;
56372b0b G	224
759496ba JW	225	if (user_mode(regs))
	226	flags \|= FAULT_FLAG_USER;
	227	if (!(fsr ^ 0x12))
	228	flags \|= FAULT_FLAG_WRITE;
	229
56372b0b G	230	/*
	231	* As per x86, we may deadlock here. However, since the kernel only
	232	* validly references user space from well defined areas of the code,
	233	* we can bug out early if this is from code which shouldn't.
	234	*/
	235	if (!down_read_trylock(&mm->mmap_sem)) {
	236	if (!user_mode(regs)
	237	&& !search_exception_tables(regs->UCreg_pc))
	238	goto no_context;
f3f09d5a	239	retry:
56372b0b G	240	down_read(&mm->mmap_sem);
	241	} else {
	242	/*
	243	* The above down_read_trylock() might have succeeded in
	244	* which case, we'll have missed the might_sleep() from
	245	* down_read()
	246	*/
	247	might_sleep();
	248	#ifdef CONFIG_DEBUG_VM
	249	if (!user_mode(regs) &&
	250	!search_exception_tables(regs->UCreg_pc))
	251	goto no_context;
	252	#endif
	253	}
	254
f3f09d5a KC	255	fault = __do_pf(mm, addr, fsr, flags, tsk);
	256
	257	/* If we need to retry but a fatal signal is pending, handle the
	258	* signal first. We do not need to release the mmap_sem because
	259	* it would already be released in __lock_page_or_retry in
	260	* mm/filemap.c. */
	261	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
	262	return 0;
	263
	264	if (!(fault & VM_FAULT_ERROR) && (flags & FAULT_FLAG_ALLOW_RETRY)) {
	265	if (fault & VM_FAULT_MAJOR)
	266	tsk->maj_flt++;
	267	else
	268	tsk->min_flt++;
	269	if (fault & VM_FAULT_RETRY) {
	270	/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
	271	* of starvation. */
	272	flags &= ~FAULT_FLAG_ALLOW_RETRY;
	273	goto retry;
	274	}
	275	}
	276
56372b0b G	277	up_read(&mm->mmap_sem);
	278
	279	/*
0e8fb931	280	* Handle the "normal" case first - VM_FAULT_MAJOR
56372b0b G	281	*/
	282	if (likely(!(fault &
	283	(VM_FAULT_ERROR \| VM_FAULT_BADMAP \| VM_FAULT_BADACCESS))))
	284	return 0;
	285
87134102 JW	286	/*
	287	* If we are in kernel mode at this point, we
	288	* have no context to handle this fault with.
	289	*/
	290	if (!user_mode(regs))
	291	goto no_context;
	292
56372b0b G	293	if (fault & VM_FAULT_OOM) {
	294	/*
	295	* We ran out of memory, call the OOM killer, and return to
	296	* userspace (which will retry the fault, or kill us if we
	297	* got oom-killed)
	298	*/
	299	pagefault_out_of_memory();
	300	return 0;
	301	}
	302
56372b0b G	303	if (fault & VM_FAULT_SIGBUS) {
	304	/*
	305	* We had some memory, but were unable to
	306	* successfully fix up this page fault.
	307	*/
	308	sig = SIGBUS;
	309	code = BUS_ADRERR;
	310	} else {
	311	/*
	312	* Something tried to access memory that
	313	* isn't in our memory map..
	314	*/
	315	sig = SIGSEGV;
	316	code = fault == VM_FAULT_BADACCESS ? SEGV_ACCERR : SEGV_MAPERR;
	317	}
	318
	319	__do_user_fault(tsk, addr, fsr, sig, code, regs);
	320	return 0;
	321
	322	no_context:
	323	__do_kernel_fault(mm, addr, fsr, regs);
	324	return 0;
	325	}
	326
	327	/*
	328	* First Level Translation Fault Handler
	329	*
	330	* We enter here because the first level page table doesn't contain
	331	* a valid entry for the address.
	332	*
	333	* If the address is in kernel space (>= TASK_SIZE), then we are
	334	* probably faulting in the vmalloc() area.
	335	*
	336	* If the init_task's first level page tables contains the relevant
	337	* entry, we copy the it to this task. If not, we send the process
	338	* a signal, fixup the exception, or oops the kernel.
	339	*
	340	* NOTE! We MUST NOT take any locks for this case. We may be in an
	341	* interrupt or a critical region, and should only copy the information
	342	* from the master page table, nothing more.
	343	*/
	344	static int do_ifault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
	345	{
	346	unsigned int index;
	347	pgd_t pgd, pgd_k;
	348	pmd_t pmd, pmd_k;
	349
	350	if (addr < TASK_SIZE)
	351	return do_pf(addr, fsr, regs);
	352
	353	if (user_mode(regs))
	354	goto bad_area;
	355
	356	index = pgd_index(addr);
	357
	358	pgd = cpu_get_pgd() + index;
	359	pgd_k = init_mm.pgd + index;
	360
	361	if (pgd_none(*pgd_k))
	362	goto bad_area;
	363
	364	pmd_k = pmd_offset((pud_t *) pgd_k, addr);
	365	pmd = pmd_offset((pud_t *) pgd, addr);
	366
367	if (pmd_none(*pmd_k))
368	goto bad_area;
369
370	set_pmd(pmd, *pmd_k);
371	flush_pmd_entry(pmd);
372	return 0;
373
374	bad_area:
375	do_bad_area(addr, fsr, regs);
376	return 0;
377	}
378
379	/*
380	* This abort handler always returns "fault".
381	*/
382	static int do_bad(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
383	{
384	return 1;
385	}
386
387	static int do_good(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
388	{
389	unsigned int res1, res2;
390
391	printk("dabt exception but no error!\n");
392
393	__asm__ __volatile__(
394	"mff %0,f0\n"
395	"mff %1,f1\n"
396	: "=r"(res1), "=r"(res2)
397	:
398	: "memory");
399
400	printk(KERN_EMERG "r0 :%08x r1 :%08x\n", res1, res2);
401	panic("shut up\n");
402	return 0;
403	}
404
405	static struct fsr_info {
406	int (fn) (unsigned long addr, unsigned int fsr, struct pt_regs regs);
407	int sig;
408	int code;
409	const char *name;
410	} fsr_info[] = {
411	/*
412	* The following are the standard Unicore-I and UniCore-II aborts.
413	*/
414	{ do_good, SIGBUS, 0, "no error" },
415	{ do_bad, SIGBUS, BUS_ADRALN, "alignment exception" },
416	{ do_bad, SIGBUS, BUS_OBJERR, "external exception" },
417	{ do_bad, SIGBUS, 0, "burst operation" },
418	{ do_bad, SIGBUS, 0, "unknown 00100" },
419	{ do_ifault, SIGSEGV, SEGV_MAPERR, "2nd level pt non-exist"},
420	{ do_bad, SIGBUS, 0, "2nd lvl large pt non-exist" },
421	{ do_bad, SIGBUS, 0, "invalid pte" },
422	{ do_pf, SIGSEGV, SEGV_MAPERR, "page miss" },
423	{ do_bad, SIGBUS, 0, "middle page miss" },
424	{ do_bad, SIGBUS, 0, "large page miss" },
425	{ do_pf, SIGSEGV, SEGV_MAPERR, "super page (section) miss" },
426	{ do_bad, SIGBUS, 0, "unknown 01100" },
427	{ do_bad, SIGBUS, 0, "unknown 01101" },
428	{ do_bad, SIGBUS, 0, "unknown 01110" },
429	{ do_bad, SIGBUS, 0, "unknown 01111" },
430	{ do_bad, SIGBUS, 0, "addr: up 3G or IO" },
431	{ do_pf, SIGSEGV, SEGV_ACCERR, "read unreadable addr" },
432	{ do_pf, SIGSEGV, SEGV_ACCERR, "write unwriteable addr"},
433	{ do_pf, SIGSEGV, SEGV_ACCERR, "exec unexecutable addr"},
434	{ do_bad, SIGBUS, 0, "unknown 10100" },
435	{ do_bad, SIGBUS, 0, "unknown 10101" },
436	{ do_bad, SIGBUS, 0, "unknown 10110" },
437	{ do_bad, SIGBUS, 0, "unknown 10111" },
438	{ do_bad, SIGBUS, 0, "unknown 11000" },
439	{ do_bad, SIGBUS, 0, "unknown 11001" },
440	{ do_bad, SIGBUS, 0, "unknown 11010" },
441	{ do_bad, SIGBUS, 0, "unknown 11011" },
442	{ do_bad, SIGBUS, 0, "unknown 11100" },
443	{ do_bad, SIGBUS, 0, "unknown 11101" },
444	{ do_bad, SIGBUS, 0, "unknown 11110" },
445	{ do_bad, SIGBUS, 0, "unknown 11111" }
446	};
447
448	void __init hook_fault_code(int nr,
449	int (fn) (unsigned long, unsigned int, struct pt_regs ),
450	int sig, int code, const char *name)
451	{
452	if (nr < 0 \|\| nr >= ARRAY_SIZE(fsr_info))
453	BUG();
454
455	fsr_info[nr].fn = fn;
456	fsr_info[nr].sig = sig;
457	fsr_info[nr].code = code;
458	fsr_info[nr].name = name;
459	}
460
461	/*
462	* Dispatch a data abort to the relevant handler.
463	*/
464	asmlinkage void do_DataAbort(unsigned long addr, unsigned int fsr,
465	struct pt_regs *regs)
466	{
467	const struct fsr_info *inf = fsr_info + fsr_fs(fsr);
468	struct siginfo info;
469
470	if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
471	return;
472
473	printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
474	inf->name, fsr, addr);
475
3eb0f519	476	clear_siginfo(&info);
56372b0b G	477	info.si_signo = inf->sig;
	478	info.si_errno = 0;
	479	info.si_code = inf->code;
	480	info.si_addr = (void __user *)addr;
	481	uc32_notify_die("", regs, &info, fsr, 0);
	482	}
	483
	484	asmlinkage void do_PrefetchAbort(unsigned long addr,
	485	unsigned int ifsr, struct pt_regs *regs)
	486	{
	487	const struct fsr_info *inf = fsr_info + fsr_fs(ifsr);
	488	struct siginfo info;
	489
	490	if (!inf->fn(addr, ifsr \| FSR_LNX_PF, regs))
	491	return;
	492
	493	printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at 0x%08lx\n",
	494	inf->name, ifsr, addr);
	495
3eb0f519	496	clear_siginfo(&info);
56372b0b G	497	info.si_signo = inf->sig;
	498	info.si_errno = 0;
	499	info.si_code = inf->code;
	500	info.si_addr = (void __user *)addr;
	501	uc32_notify_die("", regs, &info, ifsr, 0);
	502	}