dumpstack.c 8.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12
/*
 *  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>
13
#include <linux/ftrace.h>
14 15 16 17 18 19 20 21 22
#include <linux/kexec.h>
#include <linux/bug.h>
#include <linux/nmi.h>
#include <linux/sysfs.h>

#include <asm/stacktrace.h>


int panic_on_unrecovered_nmi;
23
int panic_on_io_nmi;
24 25 26 27
unsigned int code_bytes = 64;
int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
static int die_counter;

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
bool in_task_stack(unsigned long *stack, struct task_struct *task,
		   struct stack_info *info)
{
	unsigned long *begin = task_stack_page(task);
	unsigned long *end   = task_stack_page(task) + THREAD_SIZE;

	if (stack < begin || stack >= end)
		return false;

	info->type	= STACK_TYPE_TASK;
	info->begin	= begin;
	info->end	= end;
	info->next_sp	= NULL;

	return true;
}

45
static void printk_stack_address(unsigned long address, int reliable,
46
				 char *log_lvl)
47
{
48
	touch_nmi_watchdog();
49
	printk("%s [<%p>] %s%pB\n",
50
		log_lvl, (void *)address, reliable ? "" : "? ",
51
		(void *)address);
52 53
}

54 55 56 57 58
void printk_address(unsigned long address)
{
	pr_cont(" [<%p>] %pS\n", (void *)address, (void *)address);
}

59 60 61 62 63 64 65 66
/*
 * x86-64 can have up to three kernel stacks:
 * process stack
 * interrupt stack
 * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
 */

unsigned long
67
print_context_stack(struct task_struct *task,
68 69
		unsigned long *stack, unsigned long bp,
		const struct stacktrace_ops *ops, void *data,
70
		struct stack_info *info, int *graph)
71 72 73
{
	struct stack_frame *frame = (struct stack_frame *)bp;

74 75 76 77 78 79 80 81
	/*
	 * If we overflowed the stack into a guard page, jump back to the
	 * bottom of the usable stack.
	 */
	if ((unsigned long)task_stack_page(task) - (unsigned long)stack <
	    PAGE_SIZE)
		stack = (unsigned long *)task_stack_page(task);

82
	while (on_stack(info, stack, sizeof(*stack))) {
83
		unsigned long addr = *stack;
84 85

		if (__kernel_text_address(addr)) {
86 87 88
			unsigned long real_addr;
			int reliable = 0;

89
			if ((unsigned long) stack == bp + sizeof(long)) {
90
				reliable = 1;
91 92 93
				frame = frame->next_frame;
				bp = (unsigned long) frame;
			}
94

95 96 97 98 99 100 101 102 103
			/*
			 * When function graph tracing is enabled for a
			 * function, its return address on the stack is
			 * replaced with the address of an ftrace handler
			 * (return_to_handler).  In that case, before printing
			 * the "real" address, we want to print the handler
			 * address as an "unreliable" hint that function graph
			 * tracing was involved.
			 */
104 105 106
			real_addr = ftrace_graph_ret_addr(task, graph, addr,
							  stack);
			if (real_addr != addr)
107 108 109
				ops->address(data, addr, 0);

			ops->address(data, real_addr, reliable);
110 111 112 113 114
		}
		stack++;
	}
	return bp;
}
115 116 117
EXPORT_SYMBOL_GPL(print_context_stack);

unsigned long
118
print_context_stack_bp(struct task_struct *task,
119 120
		       unsigned long *stack, unsigned long bp,
		       const struct stacktrace_ops *ops, void *data,
121
		       struct stack_info *info, int *graph)
122 123
{
	struct stack_frame *frame = (struct stack_frame *)bp;
124
	unsigned long *retp = &frame->return_address;
125

126
	while (on_stack(info, stack, sizeof(*stack) * 2)) {
127 128
		unsigned long addr = *retp;
		unsigned long real_addr;
129

130 131 132
		if (!__kernel_text_address(addr))
			break;

133
		real_addr = ftrace_graph_ret_addr(task, graph, addr, retp);
134 135
		if (ops->address(data, real_addr, 1))
			break;
136

137
		frame = frame->next_frame;
138
		retp = &frame->return_address;
139
	}
140

141 142 143
	return (unsigned long)frame;
}
EXPORT_SYMBOL_GPL(print_context_stack_bp);
144

145
static int print_trace_stack(void *data, const char *name)
146 147 148 149 150 151 152 153
{
	printk("%s <%s> ", (char *)data, name);
	return 0;
}

/*
 * Print one address/symbol entries per line.
 */
154
static int print_trace_address(void *data, unsigned long addr, int reliable)
155
{
156
	printk_stack_address(addr, reliable, data);
157
	return 0;
158 159 160
}

static const struct stacktrace_ops print_trace_ops = {
161 162
	.stack			= print_trace_stack,
	.address		= print_trace_address,
163
	.walk_stack		= print_context_stack,
164 165 166 167
};

void
show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
168
		unsigned long *stack, unsigned long bp, char *log_lvl)
169 170
{
	printk("%sCall Trace:\n", log_lvl);
171
	dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
172 173 174 175
}

void show_stack(struct task_struct *task, unsigned long *sp)
{
176 177 178 179 180 181 182
	unsigned long bp = 0;

	/*
	 * Stack frames below this one aren't interesting.  Don't show them
	 * if we're printing for %current.
	 */
	if (!sp && (!task || task == current)) {
183 184
		sp = get_stack_pointer(current, NULL);
		bp = (unsigned long)get_frame_pointer(current, NULL);
185 186 187
	}

	show_stack_log_lvl(task, NULL, sp, bp, "");
188 189
}

190 191
void show_stack_regs(struct pt_regs *regs)
{
192
	show_stack_log_lvl(current, regs, NULL, 0, "");
193 194
}

195
static arch_spinlock_t die_lock = __ARCH_SPIN_LOCK_UNLOCKED;
196 197 198
static int die_owner = -1;
static unsigned int die_nest_count;

199
unsigned long oops_begin(void)
200 201 202 203 204 205 206 207 208
{
	int cpu;
	unsigned long flags;

	oops_enter();

	/* racy, but better than risking deadlock. */
	raw_local_irq_save(flags);
	cpu = smp_processor_id();
209
	if (!arch_spin_trylock(&die_lock)) {
210 211 212
		if (cpu == die_owner)
			/* nested oops. should stop eventually */;
		else
213
			arch_spin_lock(&die_lock);
214 215 216 217 218 219 220
	}
	die_nest_count++;
	die_owner = cpu;
	console_verbose();
	bust_spinlocks(1);
	return flags;
}
221
EXPORT_SYMBOL_GPL(oops_begin);
222
NOKPROBE_SYMBOL(oops_begin);
223

224 225
void __noreturn rewind_stack_do_exit(int signr);

226
void oops_end(unsigned long flags, struct pt_regs *regs, int signr)
227 228 229 230 231 232
{
	if (regs && kexec_should_crash(current))
		crash_kexec(regs);

	bust_spinlocks(0);
	die_owner = -1;
233
	add_taint(TAINT_DIE, LOCKDEP_NOW_UNRELIABLE);
234 235 236
	die_nest_count--;
	if (!die_nest_count)
		/* Nest count reaches zero, release the lock. */
237
		arch_spin_unlock(&die_lock);
238 239 240 241 242 243 244 245 246
	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");
247 248 249 250 251 252 253

	/*
	 * We're not going to return, but we might be on an IST stack or
	 * have very little stack space left.  Rewind the stack and kill
	 * the task.
	 */
	rewind_stack_do_exit(signr);
254
}
255
NOKPROBE_SYMBOL(oops_end);
256

257
int __die(const char *str, struct pt_regs *regs, long err)
258 259 260 261 262
{
#ifdef CONFIG_X86_32
	unsigned short ss;
	unsigned long sp;
#endif
263
	printk(KERN_DEFAULT
264 265 266 267 268 269
	       "%s: %04lx [#%d]%s%s%s%s\n", str, err & 0xffff, ++die_counter,
	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT"         : "",
	       IS_ENABLED(CONFIG_SMP)     ? " SMP"             : "",
	       debug_pagealloc_enabled()  ? " DEBUG_PAGEALLOC" : "",
	       IS_ENABLED(CONFIG_KASAN)   ? " KASAN"           : "");

270
	if (notify_die(DIE_OOPS, str, regs, err,
271
			current->thread.trap_nr, SIGSEGV) == NOTIFY_STOP)
272 273
		return 1;

274
	print_modules();
275
	show_regs(regs);
276
#ifdef CONFIG_X86_32
277
	if (user_mode(regs)) {
278 279
		sp = regs->sp;
		ss = regs->ss & 0xffff;
280 281 282
	} else {
		sp = kernel_stack_pointer(regs);
		savesegment(ss, ss);
283 284 285 286 287 288 289
	}
	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 ");
290
	printk_address(regs->ip);
291 292 293 294
	printk(" RSP <%016lx>\n", regs->sp);
#endif
	return 0;
}
295
NOKPROBE_SYMBOL(__die);
296 297 298 299 300 301 302 303 304 305

/*
 * 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;

306
	if (!user_mode(regs))
307 308 309 310 311 312 313 314 315
		report_bug(regs->ip, regs);

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

static int __init kstack_setup(char *s)
{
316 317 318
	ssize_t ret;
	unsigned long val;

319 320
	if (!s)
		return -EINVAL;
321 322 323 324 325

	ret = kstrtoul(s, 0, &val);
	if (ret)
		return ret;
	kstack_depth_to_print = val;
326 327 328 329 330 331
	return 0;
}
early_param("kstack", kstack_setup);

static int __init code_bytes_setup(char *s)
{
332 333 334 335 336 337 338 339 340 341 342
	ssize_t ret;
	unsigned long val;

	if (!s)
		return -EINVAL;

	ret = kstrtoul(s, 0, &val);
	if (ret)
		return ret;

	code_bytes = val;
343 344 345 346 347 348
	if (code_bytes > 8192)
		code_bytes = 8192;

	return 1;
}
__setup("code_bytes=", code_bytes_setup);