[net-next-2.6.git] / arch / x86 / kernel / dumpstack.c

/*
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *  Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs
 */
#include <linux/kallsyms.h>
#include <linux/kprobes.h>
#include <linux/uaccess.h>
#include <linux/utsname.h>
#include <linux/hardirq.h>
#include <linux/kdebug.h>
#include <linux/module.h>
#include <linux/ptrace.h>
#include <linux/ftrace.h>
#include <linux/kexec.h>
#include <linux/bug.h>
#include <linux/nmi.h>
#include <linux/sysfs.h>

#include <asm/stacktrace.h>

#include "dumpstack.h"

int panic_on_unrecovered_nmi;
int panic_on_io_nmi;
unsigned int code_bytes = 64;
int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
static int die_counter;

void printk_address(unsigned long address, int reliable)
{
	printk(" [<%p>] %s%pS\n", (void *) address,
			reliable ? "" : "? ", (void *) address);
}

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static void
print_ftrace_graph_addr(unsigned long addr, void *data,
			const struct stacktrace_ops *ops,
			struct thread_info *tinfo, int *graph)
{
	struct task_struct *task = tinfo->task;
	unsigned long ret_addr;
	int index = task->curr_ret_stack;

	if (addr != (unsigned long)return_to_handler)
		return;

	if (!task->ret_stack || index < *graph)
		return;

	index -= *graph;
	ret_addr = task->ret_stack[index].ret;

	ops->address(data, ret_addr, 1);

	(*graph)++;
}
#else
static inline void
print_ftrace_graph_addr(unsigned long addr, void *data,
			const struct stacktrace_ops *ops,
			struct thread_info *tinfo, int *graph)
{ }
#endif

/*
 * x86-64 can have up to three kernel stacks:
 * process stack
 * interrupt stack
 * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
 */

static inline int valid_stack_ptr(struct thread_info *tinfo,
			void *p, unsigned int size, void *end)
{
	void *t = tinfo;
	if (end) {
		if (p < end && p >= (end-THREAD_SIZE))
			return 1;
		else
			return 0;
	}
	return p > t && p < t + THREAD_SIZE - size;
}

unsigned long
print_context_stack(struct thread_info *tinfo,
		unsigned long *stack, unsigned long bp,
		const struct stacktrace_ops *ops, void *data,
		unsigned long *end, int *graph)
{
	struct stack_frame *frame = (struct stack_frame *)bp;

	while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) {
		unsigned long addr;

		addr = *stack;
		if (__kernel_text_address(addr)) {
			if ((unsigned long) stack == bp + sizeof(long)) {
				ops->address(data, addr, 1);
				frame = frame->next_frame;
				bp = (unsigned long) frame;
			} else {
				ops->address(data, addr, 0);
			}
			print_ftrace_graph_addr(addr, data, ops, tinfo, graph);
		}
		stack++;
	}
	return bp;
}


static void
print_trace_warning_symbol(void *data, char *msg, unsigned long symbol)
{
	printk(data);
	print_symbol(msg, symbol);
	printk("\n");
}

static void print_trace_warning(void *data, char *msg)
{
	printk("%s%s\n", (char *)data, msg);
}

static int print_trace_stack(void *data, char *name)
{
	printk("%s <%s> ", (char *)data, name);
	return 0;
}

/*
 * Print one address/symbol entries per line.
 */
static void print_trace_address(void *data, unsigned long addr, int reliable)
{
	touch_nmi_watchdog();
	printk(data);
	printk_address(addr, reliable);
}

static const struct stacktrace_ops print_trace_ops = {
	.warning = print_trace_warning,
	.warning_symbol = print_trace_warning_symbol,
	.stack = print_trace_stack,
	.address = print_trace_address,
};

void
show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
		unsigned long *stack, unsigned long bp, char *log_lvl)
{
	printk("%sCall Trace:\n", log_lvl);
	dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
}

void show_trace(struct task_struct *task, struct pt_regs *regs,
		unsigned long *stack, unsigned long bp)
{
	show_trace_log_lvl(task, regs, stack, bp, "");
}

void show_stack(struct task_struct *task, unsigned long *sp)
{
	show_stack_log_lvl(task, NULL, sp, 0, "");
}

/*
 * The architecture-independent dump_stack generator
 */
void dump_stack(void)
{
	unsigned long bp = 0;
	unsigned long stack;

#ifdef CONFIG_FRAME_POINTER
	if (!bp)
		get_bp(bp);
#endif

	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
		current->pid, current->comm, print_tainted(),
		init_utsname()->release,
		(int)strcspn(init_utsname()->version, " "),
		init_utsname()->version);
	show_trace(NULL, NULL, &stack, bp);
}
EXPORT_SYMBOL(dump_stack);

static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED;
static int die_owner = -1;
static unsigned int die_nest_count;

unsigned __kprobes long oops_begin(void)
{
	int cpu;
	unsigned long flags;

	/* notify the hw-branch tracer so it may disable tracing and
	   add the last trace to the trace buffer -
	   the earlier this happens, the more useful the trace. */
	trace_hw_branch_oops();

	oops_enter();

	/* racy, but better than risking deadlock. */
	raw_local_irq_save(flags);
	cpu = smp_processor_id();
	if (!__raw_spin_trylock(&die_lock)) {
		if (cpu == die_owner)
			/* nested oops. should stop eventually */;
		else
			__raw_spin_lock(&die_lock);
	}
	die_nest_count++;
	die_owner = cpu;
	console_verbose();
	bust_spinlocks(1);
	return flags;
}

void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr)
{
	if (regs && kexec_should_crash(current))
		crash_kexec(regs);

	bust_spinlocks(0);
	die_owner = -1;
	add_taint(TAINT_DIE);
	die_nest_count--;
	if (!die_nest_count)
		/* Nest count reaches zero, release the lock. */
		__raw_spin_unlock(&die_lock);
	raw_local_irq_restore(flags);
	oops_exit();

	if (!signr)
		return;
	if (in_interrupt())
		panic("Fatal exception in interrupt");
	if (panic_on_oops)
		panic("Fatal exception");
	do_exit(signr);
}

int __kprobes __die(const char *str, struct pt_regs *regs, long err)
{
#ifdef CONFIG_X86_32
	unsigned short ss;
	unsigned long sp;
#endif
	printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
#ifdef CONFIG_PREEMPT
	printk("PREEMPT ");
#endif
#ifdef CONFIG_SMP
	printk("SMP ");
#endif
#ifdef CONFIG_DEBUG_PAGEALLOC
	printk("DEBUG_PAGEALLOC");
#endif
	printk("\n");
	sysfs_printk_last_file();
	if (notify_die(DIE_OOPS, str, regs, err,
			current->thread.trap_no, SIGSEGV) == NOTIFY_STOP)
		return 1;

	show_registers(regs);
#ifdef CONFIG_X86_32
	if (user_mode_vm(regs)) {
		sp = regs->sp;
		ss = regs->ss & 0xffff;
	} else {
		sp = kernel_stack_pointer(regs);
		savesegment(ss, ss);
	}
	printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip);
	print_symbol("%s", regs->ip);
	printk(" SS:ESP %04x:%08lx\n", ss, sp);
#else
	/* Executive summary in case the oops scrolled away */
	printk(KERN_ALERT "RIP ");
	printk_address(regs->ip, 1);
	printk(" RSP <%016lx>\n", regs->sp);
#endif
	return 0;
}

/*
 * This is gone through when something in the kernel has done something bad
 * and is about to be terminated:
 */
void die(const char *str, struct pt_regs *regs, long err)
{
	unsigned long flags = oops_begin();
	int sig = SIGSEGV;

	if (!user_mode_vm(regs))
		report_bug(regs->ip, regs);

	if (__die(str, regs, err))
		sig = 0;
	oops_end(flags, regs, sig);
}

void notrace __kprobes
die_nmi(char *str, struct pt_regs *regs, int do_panic)
{
	unsigned long flags;

	if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP)
		return;

	/*
	 * We are in trouble anyway, lets at least try
	 * to get a message out.
	 */
	flags = oops_begin();
	printk(KERN_EMERG "%s", str);
	printk(" on CPU%d, ip %08lx, registers:\n",
		smp_processor_id(), regs->ip);
	show_registers(regs);
	oops_end(flags, regs, 0);
	if (do_panic || panic_on_oops)
		panic("Non maskable interrupt");
	nmi_exit();
	local_irq_enable();
	do_exit(SIGBUS);
}

static int __init oops_setup(char *s)
{
	if (!s)
		return -EINVAL;
	if (!strcmp(s, "panic"))
		panic_on_oops = 1;
	return 0;
}
early_param("oops", oops_setup);

static int __init kstack_setup(char *s)
{
	if (!s)
		return -EINVAL;
	kstack_depth_to_print = simple_strtoul(s, NULL, 0);
	return 0;
}
early_param("kstack", kstack_setup);

static int __init code_bytes_setup(char *s)
{
	code_bytes = simple_strtoul(s, NULL, 0);
	if (code_bytes > 8192)
		code_bytes = 8192;

	return 1;
}
__setup("code_bytes=", code_bytes_setup);
Commit	Line	Data
878719e8 NH	1	/*
	2	* Copyright (C) 1991, 1992 Linus Torvalds
	3	* Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs
	4	*/
	5	#include <linux/kallsyms.h>
	6	#include <linux/kprobes.h>
	7	#include <linux/uaccess.h>
	8	#include <linux/utsname.h>
	9	#include <linux/hardirq.h>
	10	#include <linux/kdebug.h>
	11	#include <linux/module.h>
	12	#include <linux/ptrace.h>
712406a6	13	#include <linux/ftrace.h>
878719e8 NH	14	#include <linux/kexec.h>
	15	#include <linux/bug.h>
	16	#include <linux/nmi.h>
	17	#include <linux/sysfs.h>
	18
	19	#include <asm/stacktrace.h>
	20
	21	#include "dumpstack.h"
	22
	23	int panic_on_unrecovered_nmi;
5211a242	24	int panic_on_io_nmi;
878719e8 NH	25	unsigned int code_bytes = 64;
	26	int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
	27	static int die_counter;
	28
	29	void printk_address(unsigned long address, int reliable)
	30	{
	31	printk(" [<%p>] %s%pS\n", (void *) address,
	32	reliable ? "" : "? ", (void *) address);
	33	}
	34
7ee991fb SR	35	#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	36	static void
	37	print_ftrace_graph_addr(unsigned long addr, void *data,
	38	const struct stacktrace_ops *ops,
	39	struct thread_info tinfo, int graph)
	40	{
	41	struct task_struct *task = tinfo->task;
	42	unsigned long ret_addr;
	43	int index = task->curr_ret_stack;
	44
	45	if (addr != (unsigned long)return_to_handler)
	46	return;
	47
	48	if (!task->ret_stack \|\| index < *graph)
	49	return;
	50
	51	index -= *graph;
	52	ret_addr = task->ret_stack[index].ret;
	53
	54	ops->address(data, ret_addr, 1);
	55
	56	(*graph)++;
	57	}
	58	#else
	59	static inline void
	60	print_ftrace_graph_addr(unsigned long addr, void *data,
	61	const struct stacktrace_ops *ops,
	62	struct thread_info tinfo, int graph)
	63	{ }
	64	#endif
	65
878719e8 NH	66	/*
	67	* x86-64 can have up to three kernel stacks:
	68	* process stack
	69	* interrupt stack
	70	* severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
	71	*/
	72
	73	static inline int valid_stack_ptr(struct thread_info *tinfo,
	74	void p, unsigned int size, void end)
	75	{
	76	void *t = tinfo;
	77	if (end) {
	78	if (p < end && p >= (end-THREAD_SIZE))
	79	return 1;
	80	else
	81	return 0;
	82	}
	83	return p > t && p < t + THREAD_SIZE - size;
	84	}
	85
	86	unsigned long
	87	print_context_stack(struct thread_info *tinfo,
	88	unsigned long *stack, unsigned long bp,
	89	const struct stacktrace_ops ops, void data,
7ee991fb	90	unsigned long end, int graph)
878719e8 NH	91	{
	92	struct stack_frame frame = (struct stack_frame )bp;
	93
	94	while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) {
	95	unsigned long addr;
	96
	97	addr = *stack;
	98	if (__kernel_text_address(addr)) {
	99	if ((unsigned long) stack == bp + sizeof(long)) {
	100	ops->address(data, addr, 1);
	101	frame = frame->next_frame;
	102	bp = (unsigned long) frame;
	103	} else {
2c344e9d	104	ops->address(data, addr, 0);
878719e8	105	}
7ee991fb	106	print_ftrace_graph_addr(addr, data, ops, tinfo, graph);
878719e8 NH	107	}
	108	stack++;
	109	}
	110	return bp;
	111	}
	112
	113
	114	static void
	115	print_trace_warning_symbol(void data, char msg, unsigned long symbol)
	116	{
	117	printk(data);
	118	print_symbol(msg, symbol);
	119	printk("\n");
	120	}
	121
	122	static void print_trace_warning(void data, char msg)
	123	{
	124	printk("%s%s\n", (char *)data, msg);
	125	}
	126
	127	static int print_trace_stack(void data, char name)
	128	{
	129	printk("%s <%s> ", (char *)data, name);
	130	return 0;
	131	}
	132
	133	/*
	134	* Print one address/symbol entries per line.
	135	*/
	136	static void print_trace_address(void *data, unsigned long addr, int reliable)
	137	{
	138	touch_nmi_watchdog();
	139	printk(data);
	140	printk_address(addr, reliable);
	141	}
	142
	143	static const struct stacktrace_ops print_trace_ops = {
	144	.warning = print_trace_warning,
	145	.warning_symbol = print_trace_warning_symbol,
	146	.stack = print_trace_stack,
	147	.address = print_trace_address,
	148	};
	149
	150	void
	151	show_trace_log_lvl(struct task_struct task, struct pt_regs regs,
	152	unsigned long stack, unsigned long bp, char log_lvl)
	153	{
	154	printk("%sCall Trace:\n", log_lvl);
	155	dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
	156	}
	157
	158	void show_trace(struct task_struct task, struct pt_regs regs,
	159	unsigned long *stack, unsigned long bp)
	160	{
	161	show_trace_log_lvl(task, regs, stack, bp, "");
	162	}
	163
	164	void show_stack(struct task_struct task, unsigned long sp)
	165	{
	166	show_stack_log_lvl(task, NULL, sp, 0, "");
	167	}
	168
	169	/*
	170	* The architecture-independent dump_stack generator
171	*/
172	void dump_stack(void)
173	{
174	unsigned long bp = 0;
175	unsigned long stack;
176
177	#ifdef CONFIG_FRAME_POINTER
178	if (!bp)
179	get_bp(bp);
180	#endif
181
182	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
183	current->pid, current->comm, print_tainted(),
184	init_utsname()->release,
185	(int)strcspn(init_utsname()->version, " "),
186	init_utsname()->version);
187	show_trace(NULL, NULL, &stack, bp);
188	}
189	EXPORT_SYMBOL(dump_stack);
190
191	static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED;
192	static int die_owner = -1;
193	static unsigned int die_nest_count;
194
195	unsigned __kprobes long oops_begin(void)
196	{
197	int cpu;
198	unsigned long flags;
199
b1818748 MM	200	/* notify the hw-branch tracer so it may disable tracing and
	201	add the last trace to the trace buffer -
	202	the earlier this happens, the more useful the trace. */
	203	trace_hw_branch_oops();
	204
878719e8 NH	205	oops_enter();
	206
	207	/* racy, but better than risking deadlock. */
	208	raw_local_irq_save(flags);
	209	cpu = smp_processor_id();
	210	if (!__raw_spin_trylock(&die_lock)) {
	211	if (cpu == die_owner)
	212	/* nested oops. should stop eventually */;
	213	else
	214	__raw_spin_lock(&die_lock);
	215	}
	216	die_nest_count++;
	217	die_owner = cpu;
	218	console_verbose();
	219	bust_spinlocks(1);
	220	return flags;
	221	}
	222
	223	void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr)
	224	{
	225	if (regs && kexec_should_crash(current))
	226	crash_kexec(regs);
	227
	228	bust_spinlocks(0);
	229	die_owner = -1;
	230	add_taint(TAINT_DIE);
	231	die_nest_count--;
	232	if (!die_nest_count)
	233	/* Nest count reaches zero, release the lock. */
	234	__raw_spin_unlock(&die_lock);
	235	raw_local_irq_restore(flags);
	236	oops_exit();
	237
	238	if (!signr)
	239	return;
	240	if (in_interrupt())
	241	panic("Fatal exception in interrupt");
	242	if (panic_on_oops)
	243	panic("Fatal exception");
	244	do_exit(signr);
	245	}
	246
	247	int __kprobes __die(const char str, struct pt_regs regs, long err)
	248	{
	249	#ifdef CONFIG_X86_32
	250	unsigned short ss;
	251	unsigned long sp;
	252	#endif
	253	printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
	254	#ifdef CONFIG_PREEMPT
	255	printk("PREEMPT ");
	256	#endif
	257	#ifdef CONFIG_SMP
	258	printk("SMP ");
	259	#endif
	260	#ifdef CONFIG_DEBUG_PAGEALLOC
	261	printk("DEBUG_PAGEALLOC");
	262	#endif
	263	printk("\n");
	264	sysfs_printk_last_file();
	265	if (notify_die(DIE_OOPS, str, regs, err,
	266	current->thread.trap_no, SIGSEGV) == NOTIFY_STOP)
	267	return 1;
	268
269	show_registers(regs);
270	#ifdef CONFIG_X86_32
a343c75d	271	if (user_mode_vm(regs)) {
878719e8 NH	272	sp = regs->sp;
878719e8 NH	273	ss = regs->ss & 0xffff;
a343c75d PA	274	} else {
	275	sp = kernel_stack_pointer(regs);
	276	savesegment(ss, ss);
878719e8 NH	277	}
	278	printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip);
	279	print_symbol("%s", regs->ip);
	280	printk(" SS:ESP %04x:%08lx\n", ss, sp);
	281	#else
	282	/* Executive summary in case the oops scrolled away */
	283	printk(KERN_ALERT "RIP ");
	284	printk_address(regs->ip, 1);
	285	printk(" RSP <%016lx>\n", regs->sp);
	286	#endif
	287	return 0;
	288	}
	289
	290	/*
	291	* This is gone through when something in the kernel has done something bad
	292	* and is about to be terminated:
	293	*/
	294	void die(const char str, struct pt_regs regs, long err)
	295	{
	296	unsigned long flags = oops_begin();
	297	int sig = SIGSEGV;
	298
	299	if (!user_mode_vm(regs))
	300	report_bug(regs->ip, regs);
	301
	302	if (__die(str, regs, err))
	303	sig = 0;
	304	oops_end(flags, regs, sig);
	305	}
	306
	307	void notrace __kprobes
	308	die_nmi(char str, struct pt_regs regs, int do_panic)
	309	{
	310	unsigned long flags;
	311
	312	if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP)
	313	return;
	314
	315	/*
	316	* We are in trouble anyway, lets at least try
	317	* to get a message out.
	318	*/
	319	flags = oops_begin();
	320	printk(KERN_EMERG "%s", str);
	321	printk(" on CPU%d, ip %08lx, registers:\n",
	322	smp_processor_id(), regs->ip);
	323	show_registers(regs);
	324	oops_end(flags, regs, 0);
	325	if (do_panic \|\| panic_on_oops)
	326	panic("Non maskable interrupt");
	327	nmi_exit();
	328	local_irq_enable();
	329	do_exit(SIGBUS);
	330	}
	331
	332	static int __init oops_setup(char *s)
	333	{
	334	if (!s)
	335	return -EINVAL;
	336	if (!strcmp(s, "panic"))
	337	panic_on_oops = 1;
	338	return 0;
	339	}
	340	early_param("oops", oops_setup);
341
342	static int __init kstack_setup(char *s)
343	{
344	if (!s)
345	return -EINVAL;
346	kstack_depth_to_print = simple_strtoul(s, NULL, 0);
347	return 0;
348	}
349	early_param("kstack", kstack_setup);
350
351	static int __init code_bytes_setup(char *s)
352	{
353	code_bytes = simple_strtoul(s, NULL, 0);
354	if (code_bytes > 8192)
355	code_bytes = 8192;
356
357	return 1;
358	}
359	__setup("code_bytes=", code_bytes_setup);