dumpstack.c 7.9 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
static void printk_stack_address(unsigned long address, int reliable,
29
				 char *log_lvl)
30
{
31
	touch_nmi_watchdog();
32
	printk("%s [<%p>] %s%pB\n",
33
		log_lvl, (void *)address, reliable ? "" : "? ",
34
		(void *)address);
35 36
}

37 38 39 40 41
void printk_address(unsigned long address)
{
	pr_cont(" [<%p>] %pS\n", (void *)address, (void *)address);
}

42 43 44 45 46 47 48
/*
 * x86-64 can have up to three kernel stacks:
 * process stack
 * interrupt stack
 * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
 */

49
static inline int valid_stack_ptr(struct task_struct *task,
50 51
			void *p, unsigned int size, void *end)
{
52
	void *t = task_stack_page(task);
53 54 55 56 57 58
	if (end) {
		if (p < end && p >= (end-THREAD_SIZE))
			return 1;
		else
			return 0;
	}
59
	return p >= t && p < t + THREAD_SIZE - size;
60 61 62
}

unsigned long
63
print_context_stack(struct task_struct *task,
64 65
		unsigned long *stack, unsigned long bp,
		const struct stacktrace_ops *ops, void *data,
66
		unsigned long *end, int *graph)
67 68 69
{
	struct stack_frame *frame = (struct stack_frame *)bp;

70 71 72 73 74 75 76 77
	/*
	 * 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);

78
	while (valid_stack_ptr(task, stack, sizeof(*stack), end)) {
79
		unsigned long addr = *stack;
80 81

		if (__kernel_text_address(addr)) {
82 83 84
			unsigned long real_addr;
			int reliable = 0;

85
			if ((unsigned long) stack == bp + sizeof(long)) {
86
				reliable = 1;
87 88 89
				frame = frame->next_frame;
				bp = (unsigned long) frame;
			}
90

91 92 93 94 95 96 97 98 99
			/*
			 * 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.
			 */
100 101 102
			real_addr = ftrace_graph_ret_addr(task, graph, addr,
							  stack);
			if (real_addr != addr)
103 104 105
				ops->address(data, addr, 0);

			ops->address(data, real_addr, reliable);
106 107 108 109 110
		}
		stack++;
	}
	return bp;
}
111 112 113
EXPORT_SYMBOL_GPL(print_context_stack);

unsigned long
114
print_context_stack_bp(struct task_struct *task,
115 116 117 118 119
		       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;
120
	unsigned long *retp = &frame->return_address;
121

122 123 124
	while (valid_stack_ptr(task, retp, sizeof(*retp), end)) {
		unsigned long addr = *retp;
		unsigned long real_addr;
125

126 127 128
		if (!__kernel_text_address(addr))
			break;

129
		real_addr = ftrace_graph_ret_addr(task, graph, addr, retp);
130 131
		if (ops->address(data, real_addr, 1))
			break;
132

133
		frame = frame->next_frame;
134
		retp = &frame->return_address;
135
	}
136

137 138 139
	return (unsigned long)frame;
}
EXPORT_SYMBOL_GPL(print_context_stack_bp);
140 141 142 143 144 145 146 147 148 149

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

/*
 * Print one address/symbol entries per line.
 */
150
static int print_trace_address(void *data, unsigned long addr, int reliable)
151
{
152
	printk_stack_address(addr, reliable, data);
153
	return 0;
154 155 156
}

static const struct stacktrace_ops print_trace_ops = {
157 158
	.stack			= print_trace_stack,
	.address		= print_trace_address,
159
	.walk_stack		= print_context_stack,
160 161 162 163
};

void
show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
164
		unsigned long *stack, unsigned long bp, char *log_lvl)
165 166
{
	printk("%sCall Trace:\n", log_lvl);
167
	dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl);
168 169 170 171
}

void show_stack(struct task_struct *task, unsigned long *sp)
{
172 173 174 175 176 177 178 179 180 181 182 183 184
	unsigned long bp = 0;
	unsigned long stack;

	/*
	 * Stack frames below this one aren't interesting.  Don't show them
	 * if we're printing for %current.
	 */
	if (!sp && (!task || task == current)) {
		sp = &stack;
		bp = stack_frame(current, NULL);
	}

	show_stack_log_lvl(task, NULL, sp, bp, "");
185 186
}

187 188 189 190 191
void show_stack_regs(struct pt_regs *regs)
{
	show_stack_log_lvl(current, regs, (unsigned long *)regs->sp, regs->bp, "");
}

192
static arch_spinlock_t die_lock = __ARCH_SPIN_LOCK_UNLOCKED;
193 194 195
static int die_owner = -1;
static unsigned int die_nest_count;

196
unsigned long oops_begin(void)
197 198 199 200 201 202 203 204 205
{
	int cpu;
	unsigned long flags;

	oops_enter();

	/* racy, but better than risking deadlock. */
	raw_local_irq_save(flags);
	cpu = smp_processor_id();
206
	if (!arch_spin_trylock(&die_lock)) {
207 208 209
		if (cpu == die_owner)
			/* nested oops. should stop eventually */;
		else
210
			arch_spin_lock(&die_lock);
211 212 213 214 215 216 217
	}
	die_nest_count++;
	die_owner = cpu;
	console_verbose();
	bust_spinlocks(1);
	return flags;
}
218
EXPORT_SYMBOL_GPL(oops_begin);
219
NOKPROBE_SYMBOL(oops_begin);
220

221 222
void __noreturn rewind_stack_do_exit(int signr);

223
void oops_end(unsigned long flags, struct pt_regs *regs, int signr)
224 225 226 227 228 229
{
	if (regs && kexec_should_crash(current))
		crash_kexec(regs);

	bust_spinlocks(0);
	die_owner = -1;
230
	add_taint(TAINT_DIE, LOCKDEP_NOW_UNRELIABLE);
231 232 233
	die_nest_count--;
	if (!die_nest_count)
		/* Nest count reaches zero, release the lock. */
234
		arch_spin_unlock(&die_lock);
235 236 237 238 239 240 241 242 243
	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");
244 245 246 247 248 249 250

	/*
	 * 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);
251
}
252
NOKPROBE_SYMBOL(oops_end);
253

254
int __die(const char *str, struct pt_regs *regs, long err)
255 256 257 258 259
{
#ifdef CONFIG_X86_32
	unsigned short ss;
	unsigned long sp;
#endif
260
	printk(KERN_DEFAULT
261 262 263 264 265 266
	       "%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"           : "");

267
	if (notify_die(DIE_OOPS, str, regs, err,
268
			current->thread.trap_nr, SIGSEGV) == NOTIFY_STOP)
269 270
		return 1;

271
	print_modules();
272
	show_regs(regs);
273
#ifdef CONFIG_X86_32
274
	if (user_mode(regs)) {
275 276
		sp = regs->sp;
		ss = regs->ss & 0xffff;
277 278 279
	} else {
		sp = kernel_stack_pointer(regs);
		savesegment(ss, ss);
280 281 282 283 284 285 286
	}
	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 ");
287
	printk_address(regs->ip);
288 289 290 291
	printk(" RSP <%016lx>\n", regs->sp);
#endif
	return 0;
}
292
NOKPROBE_SYMBOL(__die);
293 294 295 296 297 298 299 300 301 302

/*
 * 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;

303
	if (!user_mode(regs))
304 305 306 307 308 309 310 311 312
		report_bug(regs->ip, regs);

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

static int __init kstack_setup(char *s)
{
313 314 315
	ssize_t ret;
	unsigned long val;

316 317
	if (!s)
		return -EINVAL;
318 319 320 321 322

	ret = kstrtoul(s, 0, &val);
	if (ret)
		return ret;
	kstack_depth_to_print = val;
323 324 325 326 327 328
	return 0;
}
early_param("kstack", kstack_setup);

static int __init code_bytes_setup(char *s)
{
329 330 331 332 333 334 335 336 337 338 339
	ssize_t ret;
	unsigned long val;

	if (!s)
		return -EINVAL;

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

	code_bytes = val;
340 341 342 343 344 345
	if (code_bytes > 8192)
		code_bytes = 8192;

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