dumpstack_64.c 8.1 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/hardirq.h>
#include <linux/kdebug.h>
#include <linux/module.h>
#include <linux/ptrace.h>
#include <linux/kexec.h>
13
#include <linux/sysfs.h>
14 15 16 17 18 19
#include <linux/bug.h>
#include <linux/nmi.h>

#include <asm/stacktrace.h>


20 21
#define N_EXCEPTION_STACKS_END \
		(N_EXCEPTION_STACKS + DEBUG_STKSZ/EXCEPTION_STKSZ - 2)
22 23

static char x86_stack_ids[][8] = {
24 25 26 27
		[ DEBUG_STACK-1			]	= "#DB",
		[ NMI_STACK-1			]	= "NMI",
		[ DOUBLEFAULT_STACK-1		]	= "#DF",
		[ MCE_STACK-1			]	= "#MC",
28
#if DEBUG_STKSZ > EXCEPTION_STKSZ
29 30
		[ N_EXCEPTION_STACKS ...
		  N_EXCEPTION_STACKS_END	]	= "#DB[?]"
31
#endif
32
};
33 34

static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack,
35
					 unsigned *usedp, char **idp)
36
{
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
	unsigned k;

	/*
	 * Iterate over all exception stacks, and figure out whether
	 * 'stack' is in one of them:
	 */
	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
		unsigned long end = per_cpu(orig_ist, cpu).ist[k];
		/*
		 * Is 'stack' above this exception frame's end?
		 * If yes then skip to the next frame.
		 */
		if (stack >= end)
			continue;
		/*
		 * Is 'stack' above this exception frame's start address?
		 * If yes then we found the right frame.
		 */
		if (stack >= end - EXCEPTION_STKSZ) {
			/*
			 * Make sure we only iterate through an exception
			 * stack once. If it comes up for the second time
			 * then there's something wrong going on - just
			 * break out and return NULL:
			 */
			if (*usedp & (1U << k))
				break;
			*usedp |= 1U << k;
65
			*idp = x86_stack_ids[k];
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84
			return (unsigned long *)end;
		}
		/*
		 * If this is a debug stack, and if it has a larger size than
		 * the usual exception stacks, then 'stack' might still
		 * be within the lower portion of the debug stack:
		 */
#if DEBUG_STKSZ > EXCEPTION_STKSZ
		if (k == DEBUG_STACK - 1 && stack >= end - DEBUG_STKSZ) {
			unsigned j = N_EXCEPTION_STACKS - 1;

			/*
			 * Black magic. A large debug stack is composed of
			 * multiple exception stack entries, which we
			 * iterate through now. Dont look:
			 */
			do {
				++j;
				end -= EXCEPTION_STKSZ;
85 86
				x86_stack_ids[j][4] = '1' +
						(j - N_EXCEPTION_STACKS);
87 88 89 90
			} while (stack < end - EXCEPTION_STKSZ);
			if (*usedp & (1U << j))
				break;
			*usedp |= 1U << j;
91
			*idp = x86_stack_ids[j];
92 93 94 95 96 97 98
			return (unsigned long *)end;
		}
#endif
	}
	return NULL;
}

99 100 101 102 103 104 105
static inline int
in_irq_stack(unsigned long *stack, unsigned long *irq_stack,
	     unsigned long *irq_stack_end)
{
	return (stack >= irq_stack && stack < irq_stack_end);
}

106 107 108 109 110 111 112 113 114 115 116
static const unsigned long irq_stack_size =
	(IRQ_STACK_SIZE - 64) / sizeof(unsigned long);

enum stack_type {
	STACK_IS_UNKNOWN,
	STACK_IS_NORMAL,
	STACK_IS_EXCEPTION,
	STACK_IS_IRQ,
};

static enum stack_type
117
analyze_stack(int cpu, struct task_struct *task, unsigned long *stack,
118 119
	      unsigned long **stack_end, unsigned long *irq_stack,
	      unsigned *used, char **id)
120 121 122 123 124 125 126 127
{
	unsigned long addr;

	addr = ((unsigned long)stack & (~(THREAD_SIZE - 1)));
	if ((unsigned long)task_stack_page(task) == addr)
		return STACK_IS_NORMAL;

	*stack_end = in_exception_stack(cpu, (unsigned long)stack,
128
					used, id);
129 130 131
	if (*stack_end)
		return STACK_IS_EXCEPTION;

132 133
	if (!irq_stack)
		return STACK_IS_NORMAL;
134

135 136
	*stack_end = irq_stack;
	irq_stack = irq_stack - irq_stack_size;
137 138 139 140 141 142 143

	if (in_irq_stack(stack, irq_stack, *stack_end))
		return STACK_IS_IRQ;

	return STACK_IS_UNKNOWN;
}

144 145 146 147 148 149 150
/*
 * x86-64 can have up to three kernel stacks:
 * process stack
 * interrupt stack
 * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
 */

151 152
void dump_trace(struct task_struct *task, struct pt_regs *regs,
		unsigned long *stack, unsigned long bp,
153 154 155 156
		const struct stacktrace_ops *ops, void *data)
{
	const unsigned cpu = get_cpu();
	struct thread_info *tinfo;
157
	unsigned long *irq_stack = (unsigned long *)per_cpu(irq_stack_ptr, cpu);
158
	unsigned long dummy;
159
	unsigned used = 0;
160 161
	int graph = 0;
	int done = 0;
162 163 164 165 166

	if (!task)
		task = current;

	if (!stack) {
167 168
		if (regs)
			stack = (unsigned long *)regs->sp;
169
		else if (task != current)
170
			stack = (unsigned long *)task->thread.sp;
171 172
		else
			stack = &dummy;
173 174
	}

175 176
	if (!bp)
		bp = stack_frame(task, regs);
177 178 179 180 181 182
	/*
	 * Print function call entries in all stacks, starting at the
	 * current stack address. If the stacks consist of nested
	 * exceptions
	 */
	tinfo = task_thread_info(task);
183 184 185
	while (!done) {
		unsigned long *stack_end;
		enum stack_type stype;
186 187
		char *id;

188 189
		stype = analyze_stack(cpu, task, stack, &stack_end,
				      irq_stack, &used, &id);
190 191 192 193 194 195 196 197 198 199 200 201

		/* Default finish unless specified to continue */
		done = 1;

		switch (stype) {

		/* Break out early if we are on the thread stack */
		case STACK_IS_NORMAL:
			break;

		case STACK_IS_EXCEPTION:

202 203 204
			if (ops->stack(data, id) < 0)
				break;

205
			bp = ops->walk_stack(tinfo, stack, bp, ops,
206
					     data, stack_end, &graph);
207 208 209 210 211 212
			ops->stack(data, "<EOE>");
			/*
			 * We link to the next stack via the
			 * second-to-last pointer (index -2 to end) in the
			 * exception stack:
			 */
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
			stack = (unsigned long *) stack_end[-2];
			done = 0;
			break;

		case STACK_IS_IRQ:

			if (ops->stack(data, "IRQ") < 0)
				break;
			bp = ops->walk_stack(tinfo, stack, bp,
				     ops, data, stack_end, &graph);
			/*
			 * We link to the next stack (which would be
			 * the process stack normally) the last
			 * pointer (index -1 to end) in the IRQ stack:
			 */
			stack = (unsigned long *) (stack_end[-1]);
229
			irq_stack = NULL;
230 231 232 233 234 235 236
			ops->stack(data, "EOI");
			done = 0;
			break;

		case STACK_IS_UNKNOWN:
			ops->stack(data, "UNK");
			break;
237 238 239 240 241 242
		}
	}

	/*
	 * This handles the process stack:
	 */
243
	bp = ops->walk_stack(tinfo, stack, bp, ops, data, NULL, &graph);
244 245 246 247
	put_cpu();
}
EXPORT_SYMBOL(dump_trace);

248
void
249
show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
250
		   unsigned long *sp, unsigned long bp, char *log_lvl)
251
{
252 253
	unsigned long *irq_stack_end;
	unsigned long *irq_stack;
254
	unsigned long *stack;
255
	int cpu;
256
	int i;
257 258 259 260 261 262

	preempt_disable();
	cpu = smp_processor_id();

	irq_stack_end	= (unsigned long *)(per_cpu(irq_stack_ptr, cpu));
	irq_stack	= (unsigned long *)(per_cpu(irq_stack_ptr, cpu) - IRQ_STACK_SIZE);
263 264

	/*
265 266
	 * Debugging aid: "show_stack(NULL, NULL);" prints the
	 * back trace for this cpu:
267 268 269 270 271 272 273 274 275 276
	 */
	if (sp == NULL) {
		if (task)
			sp = (unsigned long *)task->thread.sp;
		else
			sp = (unsigned long *)&sp;
	}

	stack = sp;
	for (i = 0; i < kstack_depth_to_print; i++) {
277 278 279
		if (stack >= irq_stack && stack <= irq_stack_end) {
			if (stack == irq_stack_end) {
				stack = (unsigned long *) (irq_stack_end[-1]);
280
				pr_cont(" <EOI> ");
281 282
			}
		} else {
283
		if (kstack_end(stack))
284 285
			break;
		}
286 287 288 289 290 291
		if ((i % STACKSLOTS_PER_LINE) == 0) {
			if (i != 0)
				pr_cont("\n");
			printk("%s %016lx", log_lvl, *stack++);
		} else
			pr_cont(" %016lx", *stack++);
292 293
		touch_nmi_watchdog();
	}
294 295
	preempt_enable();

296
	pr_cont("\n");
297
	show_trace_log_lvl(task, regs, sp, bp, log_lvl);
298 299
}

300
void show_regs(struct pt_regs *regs)
301 302 303 304 305
{
	int i;
	unsigned long sp;

	sp = regs->sp;
306
	show_regs_print_info(KERN_DEFAULT);
307 308 309 310 311 312 313 314 315 316 317 318
	__show_regs(regs, 1);

	/*
	 * When in-kernel, we also print out the stack and code at the
	 * time of the fault..
	 */
	if (!user_mode(regs)) {
		unsigned int code_prologue = code_bytes * 43 / 64;
		unsigned int code_len = code_bytes;
		unsigned char c;
		u8 *ip;

319
		printk(KERN_DEFAULT "Stack:\n");
320
		show_stack_log_lvl(NULL, regs, (unsigned long *)sp,
321
				   0, KERN_DEFAULT);
322

323
		printk(KERN_DEFAULT "Code: ");
324 325 326

		ip = (u8 *)regs->ip - code_prologue;
		if (ip < (u8 *)PAGE_OFFSET || probe_kernel_address(ip, c)) {
327
			/* try starting at IP */
328 329 330 331 332 333
			ip = (u8 *)regs->ip;
			code_len = code_len - code_prologue + 1;
		}
		for (i = 0; i < code_len; i++, ip++) {
			if (ip < (u8 *)PAGE_OFFSET ||
					probe_kernel_address(ip, c)) {
334
				pr_cont(" Bad RIP value.");
335 336 337
				break;
			}
			if (ip == (u8 *)regs->ip)
338
				pr_cont("<%02x> ", c);
339
			else
340
				pr_cont("%02x ", c);
341 342
		}
	}
343
	pr_cont("\n");
344 345 346 347 348 349 350 351 352 353 354
}

int is_valid_bugaddr(unsigned long ip)
{
	unsigned short ud2;

	if (__copy_from_user(&ud2, (const void __user *) ip, sizeof(ud2)))
		return 0;

	return ud2 == 0x0b0f;
}