[deliverable/linux.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 <linux/ftrace.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
	sp = (unsigned long) (&regs->sp);
	savesegment(ss, ss);
	if (user_mode(regs)) {
		sp = regs->sp;
		ss = regs->ss & 0xffff;
	}
	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>
b1818748	18	#include <linux/ftrace.h>
878719e8 NH	19
	20	#include <asm/stacktrace.h>
	21
	22	#include "dumpstack.h"
	23
	24	int panic_on_unrecovered_nmi;
5211a242	25	int panic_on_io_nmi;
878719e8 NH	26	unsigned int code_bytes = 64;
	27	int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
	28	static int die_counter;
	29
	30	void printk_address(unsigned long address, int reliable)
	31	{
	32	printk(" [<%p>] %s%pS\n", (void *) address,
	33	reliable ? "" : "? ", (void *) address);
	34	}
	35
7ee991fb SR	36	#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	37	static void
	38	print_ftrace_graph_addr(unsigned long addr, void *data,
	39	const struct stacktrace_ops *ops,
	40	struct thread_info tinfo, int graph)
	41	{
	42	struct task_struct *task = tinfo->task;
	43	unsigned long ret_addr;
	44	int index = task->curr_ret_stack;
	45
	46	if (addr != (unsigned long)return_to_handler)
	47	return;
	48
	49	if (!task->ret_stack \|\| index < *graph)
	50	return;
	51
	52	index -= *graph;
	53	ret_addr = task->ret_stack[index].ret;
	54
	55	ops->address(data, ret_addr, 1);
	56
	57	(*graph)++;
	58	}
	59	#else
	60	static inline void
	61	print_ftrace_graph_addr(unsigned long addr, void *data,
	62	const struct stacktrace_ops *ops,
	63	struct thread_info tinfo, int graph)
	64	{ }
	65	#endif
	66
878719e8 NH	67	/*
	68	* x86-64 can have up to three kernel stacks:
	69	* process stack
	70	* interrupt stack
	71	* severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
	72	*/
	73
	74	static inline int valid_stack_ptr(struct thread_info *tinfo,
	75	void p, unsigned int size, void end)
	76	{
	77	void *t = tinfo;
	78	if (end) {
	79	if (p < end && p >= (end-THREAD_SIZE))
	80	return 1;
	81	else
	82	return 0;
	83	}
	84	return p > t && p < t + THREAD_SIZE - size;
	85	}
	86
	87	unsigned long
	88	print_context_stack(struct thread_info *tinfo,
	89	unsigned long *stack, unsigned long bp,
	90	const struct stacktrace_ops ops, void data,
7ee991fb	91	unsigned long end, int graph)
878719e8 NH	92	{
	93	struct stack_frame frame = (struct stack_frame )bp;
	94
	95	while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) {
	96	unsigned long addr;
	97
	98	addr = *stack;
	99	if (__kernel_text_address(addr)) {
	100	if ((unsigned long) stack == bp + sizeof(long)) {
	101	ops->address(data, addr, 1);
	102	frame = frame->next_frame;
	103	bp = (unsigned long) frame;
	104	} else {
2c344e9d	105	ops->address(data, addr, 0);
878719e8	106	}
7ee991fb	107	print_ftrace_graph_addr(addr, data, ops, tinfo, graph);
878719e8 NH	108	}
	109	stack++;
	110	}
	111	return bp;
	112	}
	113
	114
	115	static void
	116	print_trace_warning_symbol(void data, char msg, unsigned long symbol)
	117	{
	118	printk(data);
	119	print_symbol(msg, symbol);
	120	printk("\n");
	121	}
	122
	123	static void print_trace_warning(void data, char msg)
	124	{
	125	printk("%s%s\n", (char *)data, msg);
	126	}
	127
	128	static int print_trace_stack(void data, char name)
	129	{
	130	printk("%s <%s> ", (char *)data, name);
	131	return 0;
	132	}
	133
	134	/*
	135	* Print one address/symbol entries per line.
	136	*/
	137	static void print_trace_address(void *data, unsigned long addr, int reliable)
	138	{
	139	touch_nmi_watchdog();
	140	printk(data);
	141	printk_address(addr, reliable);
	142	}
	143
	144	static const struct stacktrace_ops print_trace_ops = {
	145	.warning = print_trace_warning,
	146	.warning_symbol = print_trace_warning_symbol,
	147	.stack = print_trace_stack,
	148	.address = print_trace_address,
	149	};
	150
	151	void
	152	show_trace_log_lvl(struct task_struct task, struct pt_regs regs,
	153	unsigned long stack, unsigned long bp, char log_lvl)
	154	{
	155	printk("%sCall Trace:\n", log_lvl);
	156	dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
	157	}
	158
	159	void show_trace(struct task_struct task, struct pt_regs regs,
	160	unsigned long *stack, unsigned long bp)
	161	{
	162	show_trace_log_lvl(task, regs, stack, bp, "");
	163	}
	164
	165	void show_stack(struct task_struct task, unsigned long sp)
	166	{
	167	show_stack_log_lvl(task, NULL, sp, 0, "");
	168	}
	169
	170	/*
	171	* The architecture-independent dump_stack generator
172	*/
173	void dump_stack(void)
174	{
175	unsigned long bp = 0;
176	unsigned long stack;
177
178	#ifdef CONFIG_FRAME_POINTER
179	if (!bp)
180	get_bp(bp);
181	#endif
182
183	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
184	current->pid, current->comm, print_tainted(),
185	init_utsname()->release,
186	(int)strcspn(init_utsname()->version, " "),
187	init_utsname()->version);
188	show_trace(NULL, NULL, &stack, bp);
189	}
190	EXPORT_SYMBOL(dump_stack);
191
192	static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED;
193	static int die_owner = -1;
194	static unsigned int die_nest_count;
195
196	unsigned __kprobes long oops_begin(void)
197	{
198	int cpu;
199	unsigned long flags;
200
b1818748 MM	201	/* notify the hw-branch tracer so it may disable tracing and
	202	add the last trace to the trace buffer -
	203	the earlier this happens, the more useful the trace. */
	204	trace_hw_branch_oops();
	205
878719e8 NH	206	oops_enter();
	207
	208	/* racy, but better than risking deadlock. */
	209	raw_local_irq_save(flags);
	210	cpu = smp_processor_id();
	211	if (!__raw_spin_trylock(&die_lock)) {
	212	if (cpu == die_owner)
	213	/* nested oops. should stop eventually */;
	214	else
	215	__raw_spin_lock(&die_lock);
	216	}
	217	die_nest_count++;
	218	die_owner = cpu;
	219	console_verbose();
	220	bust_spinlocks(1);
	221	return flags;
	222	}
	223
	224	void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr)
	225	{
	226	if (regs && kexec_should_crash(current))
	227	crash_kexec(regs);
	228
	229	bust_spinlocks(0);
	230	die_owner = -1;
	231	add_taint(TAINT_DIE);
	232	die_nest_count--;
	233	if (!die_nest_count)
	234	/* Nest count reaches zero, release the lock. */
	235	__raw_spin_unlock(&die_lock);
	236	raw_local_irq_restore(flags);
	237	oops_exit();
	238
	239	if (!signr)
	240	return;
	241	if (in_interrupt())
	242	panic("Fatal exception in interrupt");
	243	if (panic_on_oops)
	244	panic("Fatal exception");
	245	do_exit(signr);
	246	}
	247
	248	int __kprobes __die(const char str, struct pt_regs regs, long err)
	249	{
	250	#ifdef CONFIG_X86_32
	251	unsigned short ss;
	252	unsigned long sp;
	253	#endif
	254	printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
	255	#ifdef CONFIG_PREEMPT
	256	printk("PREEMPT ");
	257	#endif
	258	#ifdef CONFIG_SMP
	259	printk("SMP ");
	260	#endif
	261	#ifdef CONFIG_DEBUG_PAGEALLOC
	262	printk("DEBUG_PAGEALLOC");
	263	#endif
	264	printk("\n");
	265	sysfs_printk_last_file();
	266	if (notify_die(DIE_OOPS, str, regs, err,
	267	current->thread.trap_no, SIGSEGV) == NOTIFY_STOP)
	268	return 1;
	269
270	show_registers(regs);
271	#ifdef CONFIG_X86_32
272	sp = (unsigned long) (&regs->sp);
273	savesegment(ss, ss);
274	if (user_mode(regs)) {
275	sp = regs->sp;
276	ss = regs->ss & 0xffff;
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);