traps_64.c 28.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
/*
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *  Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs
 *
 *  Pentium III FXSR, SSE support
 *	Gareth Hughes <gareth@valinux.com>, May 2000
 */

/*
 * 'Traps.c' handles hardware traps and faults after we have saved some
 * state in 'entry.S'.
 */
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/ptrace.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/spinlock.h>
#include <linux/interrupt.h>
24
#include <linux/kallsyms.h>
L
Linus Torvalds 已提交
25 26
#include <linux/module.h>
#include <linux/moduleparam.h>
27
#include <linux/nmi.h>
28
#include <linux/kprobes.h>
29
#include <linux/kexec.h>
30
#include <linux/unwind.h>
31
#include <linux/uaccess.h>
32
#include <linux/bug.h>
33
#include <linux/kdebug.h>
34
#include <linux/utsname.h>
L
Linus Torvalds 已提交
35

D
Dave Jiang 已提交
36 37 38 39
#if defined(CONFIG_EDAC)
#include <linux/edac.h>
#endif

L
Linus Torvalds 已提交
40 41 42 43 44 45 46
#include <asm/system.h>
#include <asm/io.h>
#include <asm/atomic.h>
#include <asm/debugreg.h>
#include <asm/desc.h>
#include <asm/i387.h>
#include <asm/processor.h>
47
#include <asm/unwind.h>
L
Linus Torvalds 已提交
48 49 50 51 52
#include <asm/smp.h>
#include <asm/pgalloc.h>
#include <asm/pda.h>
#include <asm/proto.h>
#include <asm/nmi.h>
53
#include <asm/stacktrace.h>
L
Linus Torvalds 已提交
54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

asmlinkage void divide_error(void);
asmlinkage void debug(void);
asmlinkage void nmi(void);
asmlinkage void int3(void);
asmlinkage void overflow(void);
asmlinkage void bounds(void);
asmlinkage void invalid_op(void);
asmlinkage void device_not_available(void);
asmlinkage void double_fault(void);
asmlinkage void coprocessor_segment_overrun(void);
asmlinkage void invalid_TSS(void);
asmlinkage void segment_not_present(void);
asmlinkage void stack_segment(void);
asmlinkage void general_protection(void);
asmlinkage void page_fault(void);
asmlinkage void coprocessor_error(void);
asmlinkage void simd_coprocessor_error(void);
asmlinkage void reserved(void);
asmlinkage void alignment_check(void);
asmlinkage void machine_check(void);
asmlinkage void spurious_interrupt_bug(void);

static inline void conditional_sti(struct pt_regs *regs)
{
	if (regs->eflags & X86_EFLAGS_IF)
		local_irq_enable();
}

83 84 85 86 87 88 89 90 91 92 93
static inline void preempt_conditional_sti(struct pt_regs *regs)
{
	preempt_disable();
	if (regs->eflags & X86_EFLAGS_IF)
		local_irq_enable();
}

static inline void preempt_conditional_cli(struct pt_regs *regs)
{
	if (regs->eflags & X86_EFLAGS_IF)
		local_irq_disable();
94 95
	/* Make sure to not schedule here because we could be running
	   on an exception stack. */
96 97 98
	preempt_enable_no_resched();
}

99
int kstack_depth_to_print = 12;
L
Linus Torvalds 已提交
100 101

#ifdef CONFIG_KALLSYMS
102 103
void printk_address(unsigned long address)
{
L
Linus Torvalds 已提交
104 105 106
	unsigned long offset = 0, symsize;
	const char *symname;
	char *modname;
107
	char *delim = ":";
L
Linus Torvalds 已提交
108 109
	char namebuf[128];

110 111 112 113 114 115 116
	symname = kallsyms_lookup(address, &symsize, &offset,
					&modname, namebuf);
	if (!symname) {
		printk(" [<%016lx>]\n", address);
		return;
	}
	if (!modname)
L
Linus Torvalds 已提交
117
		modname = delim = ""; 		
118 119 120
	printk(" [<%016lx>] %s%s%s%s+0x%lx/0x%lx\n",
		address, delim, modname, delim, symname, offset, symsize);
}
L
Linus Torvalds 已提交
121
#else
122 123 124 125
void printk_address(unsigned long address)
{
	printk(" [<%016lx>]\n", address);
}
L
Linus Torvalds 已提交
126 127
#endif

128
static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack,
129
					unsigned *usedp, char **idp)
130
{
131
	static char ids[][8] = {
132 133 134 135 136
		[DEBUG_STACK - 1] = "#DB",
		[NMI_STACK - 1] = "NMI",
		[DOUBLEFAULT_STACK - 1] = "#DF",
		[STACKFAULT_STACK - 1] = "#SS",
		[MCE_STACK - 1] = "#MC",
137 138 139
#if DEBUG_STKSZ > EXCEPTION_STKSZ
		[N_EXCEPTION_STACKS ... N_EXCEPTION_STACKS + DEBUG_STKSZ / EXCEPTION_STKSZ - 2] = "#DB[?]"
#endif
140 141
	};
	unsigned k;
L
Linus Torvalds 已提交
142

143 144 145 146
	/*
	 * Iterate over all exception stacks, and figure out whether
	 * 'stack' is in one of them:
	 */
147
	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
148
		unsigned long end = per_cpu(orig_ist, cpu).ist[k];
149 150 151 152
		/*
		 * Is 'stack' above this exception frame's end?
		 * If yes then skip to the next frame.
		 */
153 154
		if (stack >= end)
			continue;
155 156 157 158
		/*
		 * Is 'stack' above this exception frame's start address?
		 * If yes then we found the right frame.
		 */
159
		if (stack >= end - EXCEPTION_STKSZ) {
160 161 162 163 164 165
			/*
			 * 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:
			 */
166 167 168 169 170 171
			if (*usedp & (1U << k))
				break;
			*usedp |= 1U << k;
			*idp = ids[k];
			return (unsigned long *)end;
		}
172 173 174 175 176
		/*
		 * 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:
		 */
177 178 179 180
#if DEBUG_STKSZ > EXCEPTION_STKSZ
		if (k == DEBUG_STACK - 1 && stack >= end - DEBUG_STKSZ) {
			unsigned j = N_EXCEPTION_STACKS - 1;

181 182 183 184 185
			/*
			 * Black magic. A large debug stack is composed of
			 * multiple exception stack entries, which we
			 * iterate through now. Dont look:
			 */
186 187 188 189 190 191 192 193 194 195 196 197
			do {
				++j;
				end -= EXCEPTION_STKSZ;
				ids[j][4] = '1' + (j - N_EXCEPTION_STACKS);
			} while (stack < end - EXCEPTION_STKSZ);
			if (*usedp & (1U << j))
				break;
			*usedp |= 1U << j;
			*idp = ids[j];
			return (unsigned long *)end;
		}
#endif
L
Linus Torvalds 已提交
198 199
	}
	return NULL;
200
}
L
Linus Torvalds 已提交
201

202 203
#define MSG(txt) ops->warning(data, txt)

L
Linus Torvalds 已提交
204
/*
S
Simon Arlott 已提交
205
 * x86-64 can have up to three kernel stacks: 
L
Linus Torvalds 已提交
206 207
 * process stack
 * interrupt stack
208
 * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
L
Linus Torvalds 已提交
209 210
 */

211 212 213 214 215 216
static inline int valid_stack_ptr(struct thread_info *tinfo, void *p)
{
	void *t = (void *)tinfo;
        return p > t && p < t + THREAD_SIZE - 3;
}

217 218
void dump_trace(struct task_struct *tsk, struct pt_regs *regs,
		unsigned long *stack,
J
Jan Beulich 已提交
219
		const struct stacktrace_ops *ops, void *data)
L
Linus Torvalds 已提交
220
{
221
	const unsigned cpu = get_cpu();
222
	unsigned long *irqstack_end = (unsigned long*)cpu_pda(cpu)->irqstackptr;
223
	unsigned used = 0;
224
	struct thread_info *tinfo;
L
Linus Torvalds 已提交
225

226 227 228
	if (!tsk)
		tsk = current;

229 230 231 232 233
	if (!stack) {
		unsigned long dummy;
		stack = &dummy;
		if (tsk && tsk != current)
			stack = (unsigned long *)tsk->thread.rsp;
234 235
	}

236 237 238 239 240
	/*
	 * Print function call entries within a stack. 'cond' is the
	 * "end of stackframe" condition, that the 'stack++'
	 * iteration will eventually trigger.
	 */
241 242
#define HANDLE_STACK(cond) \
	do while (cond) { \
243
		unsigned long addr = *stack++; \
244 245 246
		/* Use unlocked access here because except for NMIs	\
		   we should be already protected against module unloads */ \
		if (__kernel_text_address(addr)) { \
247 248 249 250 251 252 253 254
			/* \
			 * If the address is either in the text segment of the \
			 * kernel, or in the region which contains vmalloc'ed \
			 * memory, it *may* be the address of a calling \
			 * routine; if so, print it so that someone tracing \
			 * down the cause of the crash will be able to figure \
			 * out the call path that was taken. \
			 */ \
255
			ops->address(data, addr);   \
256 257 258
		} \
	} while (0)

259 260 261 262 263
	/*
	 * Print function call entries in all stacks, starting at the
	 * current stack address. If the stacks consist of nested
	 * exceptions
	 */
264 265
	for (;;) {
		char *id;
266 267 268 269 270
		unsigned long *estack_end;
		estack_end = in_exception_stack(cpu, (unsigned long)stack,
						&used, &id);

		if (estack_end) {
271 272
			if (ops->stack(data, id) < 0)
				break;
273
			HANDLE_STACK (stack < estack_end);
274
			ops->stack(data, "<EOE>");
275 276 277 278 279
			/*
			 * We link to the next stack via the
			 * second-to-last pointer (index -2 to end) in the
			 * exception stack:
			 */
280 281
			stack = (unsigned long *) estack_end[-2];
			continue;
L
Linus Torvalds 已提交
282
		}
283 284 285 286 287 288
		if (irqstack_end) {
			unsigned long *irqstack;
			irqstack = irqstack_end -
				(IRQSTACKSIZE - 64) / sizeof(*irqstack);

			if (stack >= irqstack && stack < irqstack_end) {
289 290
				if (ops->stack(data, "IRQ") < 0)
					break;
291
				HANDLE_STACK (stack < irqstack_end);
292 293 294 295 296
				/*
				 * We link to the next stack (which would be
				 * the process stack normally) the last
				 * pointer (index -1 to end) in the IRQ stack:
				 */
297 298
				stack = (unsigned long *) (irqstack_end[-1]);
				irqstack_end = NULL;
299
				ops->stack(data, "EOI");
300
				continue;
L
Linus Torvalds 已提交
301 302
			}
		}
303
		break;
L
Linus Torvalds 已提交
304
	}
305

306
	/*
307
	 * This handles the process stack:
308
	 */
309
	tinfo = task_thread_info(tsk);
310
	HANDLE_STACK (valid_stack_ptr(tinfo, stack));
311
#undef HANDLE_STACK
312
	put_cpu();
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
}
EXPORT_SYMBOL(dump_trace);

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

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

static int print_trace_stack(void *data, char *name)
{
	printk(" <%s> ", name);
	return 0;
}
333

334 335
static void print_trace_address(void *data, unsigned long addr)
{
336
	touch_nmi_watchdog();
337 338 339
	printk_address(addr);
}

J
Jan Beulich 已提交
340
static const struct stacktrace_ops print_trace_ops = {
341 342 343 344 345 346 347 348 349 350 351
	.warning = print_trace_warning,
	.warning_symbol = print_trace_warning_symbol,
	.stack = print_trace_stack,
	.address = print_trace_address,
};

void
show_trace(struct task_struct *tsk, struct pt_regs *regs, unsigned long *stack)
{
	printk("\nCall Trace:\n");
	dump_trace(tsk, regs, stack, &print_trace_ops, NULL);
L
Linus Torvalds 已提交
352 353 354
	printk("\n");
}

355 356
static void
_show_stack(struct task_struct *tsk, struct pt_regs *regs, unsigned long *rsp)
L
Linus Torvalds 已提交
357 358 359
{
	unsigned long *stack;
	int i;
360
	const int cpu = smp_processor_id();
361 362
	unsigned long *irqstack_end = (unsigned long *) (cpu_pda(cpu)->irqstackptr);
	unsigned long *irqstack = (unsigned long *) (cpu_pda(cpu)->irqstackptr - IRQSTACKSIZE);
L
Linus Torvalds 已提交
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385

	// debugging aid: "show_stack(NULL, NULL);" prints the
	// back trace for this cpu.

	if (rsp == NULL) {
		if (tsk)
			rsp = (unsigned long *)tsk->thread.rsp;
		else
			rsp = (unsigned long *)&rsp;
	}

	stack = rsp;
	for(i=0; i < kstack_depth_to_print; i++) {
		if (stack >= irqstack && stack <= irqstack_end) {
			if (stack == irqstack_end) {
				stack = (unsigned long *) (irqstack_end[-1]);
				printk(" <EOI> ");
			}
		} else {
		if (((long) stack & (THREAD_SIZE-1)) == 0)
			break;
		}
		if (i && ((i % 4) == 0))
386 387
			printk("\n");
		printk(" %016lx", *stack++);
388
		touch_nmi_watchdog();
L
Linus Torvalds 已提交
389
	}
390 391 392 393 394 395
	show_trace(tsk, regs, rsp);
}

void show_stack(struct task_struct *tsk, unsigned long * rsp)
{
	_show_stack(tsk, NULL, rsp);
L
Linus Torvalds 已提交
396 397 398 399 400 401 402 403
}

/*
 * The architecture-independent dump_stack generator
 */
void dump_stack(void)
{
	unsigned long dummy;
404 405 406 407 408 409

	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);
410
	show_trace(NULL, NULL, &dummy);
L
Linus Torvalds 已提交
411 412 413 414 415 416 417
}

EXPORT_SYMBOL(dump_stack);

void show_registers(struct pt_regs *regs)
{
	int i;
418
	int in_kernel = !user_mode(regs);
L
Linus Torvalds 已提交
419
	unsigned long rsp;
420
	const int cpu = smp_processor_id();
421
	struct task_struct *cur = cpu_pda(cpu)->pcurrent;
L
Linus Torvalds 已提交
422

423
	rsp = regs->rsp;
L
Linus Torvalds 已提交
424 425 426
	printk("CPU %d ", cpu);
	__show_regs(regs);
	printk("Process %s (pid: %d, threadinfo %p, task %p)\n",
A
Al Viro 已提交
427
		cur->comm, cur->pid, task_thread_info(cur), cur);
L
Linus Torvalds 已提交
428 429 430 431 432 433 434

	/*
	 * When in-kernel, we also print out the stack and code at the
	 * time of the fault..
	 */
	if (in_kernel) {
		printk("Stack: ");
435
		_show_stack(NULL, regs, (unsigned long*)rsp);
L
Linus Torvalds 已提交
436 437

		printk("\nCode: ");
438
		if (regs->rip < PAGE_OFFSET)
L
Linus Torvalds 已提交
439 440
			goto bad;

441
		for (i=0; i<20; i++) {
L
Linus Torvalds 已提交
442
			unsigned char c;
443
			if (__get_user(c, &((unsigned char*)regs->rip)[i])) {
L
Linus Torvalds 已提交
444 445 446 447 448 449 450 451 452 453
bad:
				printk(" Bad RIP value.");
				break;
			}
			printk("%02x ", c);
		}
	}
	printk("\n");
}	

454 455 456 457 458 459 460 461 462
int is_valid_bugaddr(unsigned long rip)
{
	unsigned short ud2;

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

	return ud2 == 0x0b0f;
}
L
Linus Torvalds 已提交
463

464
#ifdef CONFIG_BUG
L
Linus Torvalds 已提交
465 466 467 468
void out_of_line_bug(void)
{ 
	BUG(); 
} 
469
EXPORT_SYMBOL(out_of_line_bug);
470
#endif
L
Linus Torvalds 已提交
471

A
Andi Kleen 已提交
472
static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED;
L
Linus Torvalds 已提交
473
static int die_owner = -1;
474
static unsigned int die_nest_count;
L
Linus Torvalds 已提交
475

476
unsigned __kprobes long oops_begin(void)
L
Linus Torvalds 已提交
477
{
A
Andrew Morton 已提交
478
	int cpu;
479 480
	unsigned long flags;

481 482
	oops_enter();

483
	/* racy, but better than risking deadlock. */
A
Andi Kleen 已提交
484
	raw_local_irq_save(flags);
A
Andrew Morton 已提交
485
	cpu = smp_processor_id();
A
Andi Kleen 已提交
486
	if (!__raw_spin_trylock(&die_lock)) {
L
Linus Torvalds 已提交
487 488 489
		if (cpu == die_owner) 
			/* nested oops. should stop eventually */;
		else
A
Andi Kleen 已提交
490
			__raw_spin_lock(&die_lock);
L
Linus Torvalds 已提交
491
	}
492
	die_nest_count++;
493
	die_owner = cpu;
L
Linus Torvalds 已提交
494
	console_verbose();
495 496
	bust_spinlocks(1);
	return flags;
L
Linus Torvalds 已提交
497 498
}

499
void __kprobes oops_end(unsigned long flags)
L
Linus Torvalds 已提交
500 501
{ 
	die_owner = -1;
502
	bust_spinlocks(0);
503
	die_nest_count--;
A
Andi Kleen 已提交
504
	if (!die_nest_count)
505
		/* Nest count reaches zero, release the lock. */
A
Andi Kleen 已提交
506 507
		__raw_spin_unlock(&die_lock);
	raw_local_irq_restore(flags);
L
Linus Torvalds 已提交
508
	if (panic_on_oops)
509
		panic("Fatal exception");
510
	oops_exit();
511
}
L
Linus Torvalds 已提交
512

513
void __kprobes __die(const char * str, struct pt_regs * regs, long err)
L
Linus Torvalds 已提交
514 515 516 517 518 519 520 521 522 523 524 525 526
{
	static int die_counter;
	printk(KERN_EMERG "%s: %04lx [%u] ", 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");
527
	notify_die(DIE_OOPS, str, regs, err, current->thread.trap_no, SIGSEGV);
L
Linus Torvalds 已提交
528
	show_registers(regs);
529
	add_taint(TAINT_DIE);
L
Linus Torvalds 已提交
530 531 532 533
	/* Executive summary in case the oops scrolled away */
	printk(KERN_ALERT "RIP ");
	printk_address(regs->rip); 
	printk(" RSP <%016lx>\n", regs->rsp); 
534 535
	if (kexec_should_crash(current))
		crash_kexec(regs);
L
Linus Torvalds 已提交
536 537 538 539
}

void die(const char * str, struct pt_regs * regs, long err)
{
540 541
	unsigned long flags = oops_begin();

542
	if (!user_mode(regs))
543
		report_bug(regs->rip, regs);
544

L
Linus Torvalds 已提交
545
	__die(str, regs, err);
546
	oops_end(flags);
L
Linus Torvalds 已提交
547 548 549
	do_exit(SIGSEGV); 
}

550
void __kprobes die_nmi(char *str, struct pt_regs *regs, int do_panic)
L
Linus Torvalds 已提交
551
{
552 553
	unsigned long flags = oops_begin();

L
Linus Torvalds 已提交
554 555 556 557
	/*
	 * We are in trouble anyway, lets at least try
	 * to get a message out.
	 */
558
	printk(str, smp_processor_id());
L
Linus Torvalds 已提交
559
	show_registers(regs);
560 561
	if (kexec_should_crash(current))
		crash_kexec(regs);
562 563
	if (do_panic || panic_on_oops)
		panic("Non maskable interrupt");
564
	oops_end(flags);
565 566
	nmi_exit();
	local_irq_enable();
L
Linus Torvalds 已提交
567 568 569
	do_exit(SIGSEGV);
}

570 571 572
static void __kprobes do_trap(int trapnr, int signr, char *str,
			      struct pt_regs * regs, long error_code,
			      siginfo_t *info)
L
Linus Torvalds 已提交
573
{
574 575 576
	struct task_struct *tsk = current;

	if (user_mode(regs)) {
577 578 579 580 581 582 583 584 585 586 587 588 589
		/*
		 * We want error_code and trap_no set for userspace
		 * faults and kernelspace faults which result in
		 * die(), but not kernelspace faults which are fixed
		 * up.  die() gives the process no chance to handle
		 * the signal and notice the kernel fault information,
		 * so that won't result in polluting the information
		 * about previously queued, but not yet delivered,
		 * faults.  See also do_general_protection below.
		 */
		tsk->thread.error_code = error_code;
		tsk->thread.trap_no = trapnr;

590 591
		if (show_unhandled_signals && unhandled_signal(tsk, signr) &&
		    printk_ratelimit())
L
Linus Torvalds 已提交
592 593 594
			printk(KERN_INFO
			       "%s[%d] trap %s rip:%lx rsp:%lx error:%lx\n",
			       tsk->comm, tsk->pid, str,
595
			       regs->rip, regs->rsp, error_code); 
L
Linus Torvalds 已提交
596 597 598 599 600 601 602 603 604 605 606 607 608

		if (info)
			force_sig_info(signr, info, tsk);
		else
			force_sig(signr, tsk);
		return;
	}


	/* kernel trap */ 
	{	     
		const struct exception_table_entry *fixup;
		fixup = search_exception_tables(regs->rip);
609
		if (fixup)
L
Linus Torvalds 已提交
610
			regs->rip = fixup->fixup;
611 612 613
		else {
			tsk->thread.error_code = error_code;
			tsk->thread.trap_no = trapnr;
L
Linus Torvalds 已提交
614
			die(str, regs, error_code);
615
		}
L
Linus Torvalds 已提交
616 617 618 619 620 621 622 623 624 625
		return;
	}
}

#define DO_ERROR(trapnr, signr, str, name) \
asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
{ \
	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
							== NOTIFY_STOP) \
		return; \
626
	conditional_sti(regs);						\
L
Linus Torvalds 已提交
627 628 629 630 631 632 633 634 635 636 637 638 639 640
	do_trap(trapnr, signr, str, regs, error_code, NULL); \
}

#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \
asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
{ \
	siginfo_t info; \
	info.si_signo = signr; \
	info.si_errno = 0; \
	info.si_code = sicode; \
	info.si_addr = (void __user *)siaddr; \
	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
							== NOTIFY_STOP) \
		return; \
641
	conditional_sti(regs);						\
L
Linus Torvalds 已提交
642 643 644 645 646 647
	do_trap(trapnr, signr, str, regs, error_code, &info); \
}

DO_ERROR_INFO( 0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->rip)
DO_ERROR( 4, SIGSEGV, "overflow", overflow)
DO_ERROR( 5, SIGSEGV, "bounds", bounds)
648
DO_ERROR_INFO( 6, SIGILL,  "invalid opcode", invalid_op, ILL_ILLOPN, regs->rip)
L
Linus Torvalds 已提交
649 650 651 652 653 654
DO_ERROR( 7, SIGSEGV, "device not available", device_not_available)
DO_ERROR( 9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun)
DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS)
DO_ERROR(11, SIGBUS,  "segment not present", segment_not_present)
DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, 0)
DO_ERROR(18, SIGSEGV, "reserved", reserved)
655 656 657 658 659 660 661 662 663 664 665

/* Runs on IST stack */
asmlinkage void do_stack_segment(struct pt_regs *regs, long error_code)
{
	if (notify_die(DIE_TRAP, "stack segment", regs, error_code,
			12, SIGBUS) == NOTIFY_STOP)
		return;
	preempt_conditional_sti(regs);
	do_trap(12, SIGBUS, "stack segment", regs, error_code, NULL);
	preempt_conditional_cli(regs);
}
666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682

asmlinkage void do_double_fault(struct pt_regs * regs, long error_code)
{
	static const char str[] = "double fault";
	struct task_struct *tsk = current;

	/* Return not checked because double check cannot be ignored */
	notify_die(DIE_TRAP, str, regs, error_code, 8, SIGSEGV);

	tsk->thread.error_code = error_code;
	tsk->thread.trap_no = 8;

	/* This is always a kernel trap and never fixable (and thus must
	   never return). */
	for (;;)
		die(str, regs, error_code);
}
L
Linus Torvalds 已提交
683

684 685
asmlinkage void __kprobes do_general_protection(struct pt_regs * regs,
						long error_code)
L
Linus Torvalds 已提交
686
{
687 688
	struct task_struct *tsk = current;

L
Linus Torvalds 已提交
689 690
	conditional_sti(regs);

691
	if (user_mode(regs)) {
692 693 694
		tsk->thread.error_code = error_code;
		tsk->thread.trap_no = 13;

695 696
		if (show_unhandled_signals && unhandled_signal(tsk, SIGSEGV) &&
		    printk_ratelimit())
L
Linus Torvalds 已提交
697 698 699
			printk(KERN_INFO
		       "%s[%d] general protection rip:%lx rsp:%lx error:%lx\n",
			       tsk->comm, tsk->pid,
700
			       regs->rip, regs->rsp, error_code); 
L
Linus Torvalds 已提交
701 702 703 704 705 706 707 708 709 710 711 712 713

		force_sig(SIGSEGV, tsk);
		return;
	} 

	/* kernel gp */
	{
		const struct exception_table_entry *fixup;
		fixup = search_exception_tables(regs->rip);
		if (fixup) {
			regs->rip = fixup->fixup;
			return;
		}
714 715 716

		tsk->thread.error_code = error_code;
		tsk->thread.trap_no = 13;
L
Linus Torvalds 已提交
717 718 719 720 721 722 723
		if (notify_die(DIE_GPF, "general protection fault", regs,
					error_code, 13, SIGSEGV) == NOTIFY_STOP)
			return;
		die("general protection fault", regs, error_code);
	}
}

724 725
static __kprobes void
mem_parity_error(unsigned char reason, struct pt_regs * regs)
L
Linus Torvalds 已提交
726
{
727 728
	printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x.\n",
		reason);
729
	printk(KERN_EMERG "You have some hardware problem, likely on the PCI bus.\n");
730

D
Dave Jiang 已提交
731 732 733 734 735 736 737
#if defined(CONFIG_EDAC)
	if(edac_handler_set()) {
		edac_atomic_assert_error();
		return;
	}
#endif

738
	if (panic_on_unrecovered_nmi)
739 740 741
		panic("NMI: Not continuing");

	printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
L
Linus Torvalds 已提交
742 743 744 745 746 747

	/* Clear and disable the memory parity error line. */
	reason = (reason & 0xf) | 4;
	outb(reason, 0x61);
}

748 749
static __kprobes void
io_check_error(unsigned char reason, struct pt_regs * regs)
L
Linus Torvalds 已提交
750 751 752 753 754 755 756 757 758 759 760 761
{
	printk("NMI: IOCK error (debug interrupt?)\n");
	show_registers(regs);

	/* Re-enable the IOCK line, wait for a few seconds */
	reason = (reason & 0xf) | 8;
	outb(reason, 0x61);
	mdelay(2000);
	reason &= ~8;
	outb(reason, 0x61);
}

762 763
static __kprobes void
unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
764 765 766 767
{
	printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x.\n",
		reason);
	printk(KERN_EMERG "Do you have a strange power saving mode enabled?\n");
768 769

	if (panic_on_unrecovered_nmi)
770
		panic("NMI: Not continuing");
771

772
	printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
L
Linus Torvalds 已提交
773 774
}

775 776
/* Runs on IST stack. This code must keep interrupts off all the time.
   Nested NMIs are prevented by the CPU. */
777
asmlinkage __kprobes void default_do_nmi(struct pt_regs *regs)
L
Linus Torvalds 已提交
778 779
{
	unsigned char reason = 0;
A
Ashok Raj 已提交
780 781 782
	int cpu;

	cpu = smp_processor_id();
L
Linus Torvalds 已提交
783 784

	/* Only the BSP gets external NMIs from the system.  */
A
Ashok Raj 已提交
785
	if (!cpu)
L
Linus Torvalds 已提交
786 787 788
		reason = get_nmi_reason();

	if (!(reason & 0xc0)) {
789
		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
L
Linus Torvalds 已提交
790 791 792 793 794 795
								== NOTIFY_STOP)
			return;
		/*
		 * Ok, so this is none of the documented NMI sources,
		 * so it must be the NMI watchdog.
		 */
796
		if (nmi_watchdog_tick(regs,reason))
L
Linus Torvalds 已提交
797
			return;
798 799 800
		if (!do_nmi_callback(regs,cpu))
			unknown_nmi_error(reason, regs);

L
Linus Torvalds 已提交
801 802
		return;
	}
803
	if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
L
Linus Torvalds 已提交
804 805 806 807 808 809 810 811 812 813
		return; 

	/* AK: following checks seem to be broken on modern chipsets. FIXME */

	if (reason & 0x80)
		mem_parity_error(reason, regs);
	if (reason & 0x40)
		io_check_error(reason, regs);
}

814
/* runs on IST stack. */
815
asmlinkage void __kprobes do_int3(struct pt_regs * regs, long error_code)
L
Linus Torvalds 已提交
816
{
P
Peter Zijlstra 已提交
817 818
	trace_hardirqs_fixup();

L
Linus Torvalds 已提交
819 820 821
	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP) == NOTIFY_STOP) {
		return;
	}
822
	preempt_conditional_sti(regs);
L
Linus Torvalds 已提交
823
	do_trap(3, SIGTRAP, "int3", regs, error_code, NULL);
824
	preempt_conditional_cli(regs);
L
Linus Torvalds 已提交
825 826
}

827 828 829
/* Help handler running on IST stack to switch back to user stack
   for scheduling or signal handling. The actual stack switch is done in
   entry.S */
830
asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs)
831 832 833 834 835 836
{
	struct pt_regs *regs = eregs;
	/* Did already sync */
	if (eregs == (struct pt_regs *)eregs->rsp)
		;
	/* Exception from user space */
837
	else if (user_mode(eregs))
A
Al Viro 已提交
838
		regs = task_pt_regs(current);
839 840 841 842 843 844 845 846 847
	/* Exception from kernel and interrupts are enabled. Move to
 	   kernel process stack. */
	else if (eregs->eflags & X86_EFLAGS_IF)
		regs = (struct pt_regs *)(eregs->rsp -= sizeof(struct pt_regs));
	if (eregs != regs)
		*regs = *eregs;
	return regs;
}

L
Linus Torvalds 已提交
848
/* runs on IST stack. */
849 850
asmlinkage void __kprobes do_debug(struct pt_regs * regs,
				   unsigned long error_code)
L
Linus Torvalds 已提交
851 852 853 854 855
{
	unsigned long condition;
	struct task_struct *tsk = current;
	siginfo_t info;

856 857
	trace_hardirqs_fixup();

858
	get_debugreg(condition, 6);
L
Linus Torvalds 已提交
859 860

	if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code,
861
						SIGTRAP) == NOTIFY_STOP)
862
		return;
863

864
	preempt_conditional_sti(regs);
L
Linus Torvalds 已提交
865 866 867 868 869 870 871 872 873 874 875

	/* Mask out spurious debug traps due to lazy DR7 setting */
	if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) {
		if (!tsk->thread.debugreg7) { 
			goto clear_dr7;
		}
	}

	tsk->thread.debugreg6 = condition;

	/* Mask out spurious TF errors due to lazy TF clearing */
876
	if (condition & DR_STEP) {
L
Linus Torvalds 已提交
877 878 879 880 881 882 883 884 885
		/*
		 * The TF error should be masked out only if the current
		 * process is not traced and if the TRAP flag has been set
		 * previously by a tracing process (condition detected by
		 * the PT_DTRACE flag); remember that the i386 TRAP flag
		 * can be modified by the process itself in user mode,
		 * allowing programs to debug themselves without the ptrace()
		 * interface.
		 */
886
                if (!user_mode(regs))
L
Linus Torvalds 已提交
887
                       goto clear_TF_reenable;
888 889 890 891 892 893 894 895
		/*
		 * Was the TF flag set by a debugger? If so, clear it now,
		 * so that register information is correct.
		 */
		if (tsk->ptrace & PT_DTRACE) {
			regs->eflags &= ~TF_MASK;
			tsk->ptrace &= ~PT_DTRACE;
		}
L
Linus Torvalds 已提交
896 897 898 899 900 901 902 903
	}

	/* Ok, finally something we can handle */
	tsk->thread.trap_no = 1;
	tsk->thread.error_code = error_code;
	info.si_signo = SIGTRAP;
	info.si_errno = 0;
	info.si_code = TRAP_BRKPT;
904 905
	info.si_addr = user_mode(regs) ? (void __user *)regs->rip : NULL;
	force_sig_info(SIGTRAP, &info, tsk);
L
Linus Torvalds 已提交
906 907

clear_dr7:
908
	set_debugreg(0UL, 7);
909
	preempt_conditional_cli(regs);
910
	return;
L
Linus Torvalds 已提交
911 912 913 914

clear_TF_reenable:
	set_tsk_thread_flag(tsk, TIF_SINGLESTEP);
	regs->eflags &= ~TF_MASK;
915
	preempt_conditional_cli(regs);
L
Linus Torvalds 已提交
916 917
}

918
static int kernel_math_error(struct pt_regs *regs, const char *str, int trapnr)
L
Linus Torvalds 已提交
919 920 921 922 923 924 925
{
	const struct exception_table_entry *fixup;
	fixup = search_exception_tables(regs->rip);
	if (fixup) {
		regs->rip = fixup->fixup;
		return 1;
	}
926
	notify_die(DIE_GPF, str, regs, 0, trapnr, SIGFPE);
927
	/* Illegal floating point operation in the kernel */
928
	current->thread.trap_no = trapnr;
L
Linus Torvalds 已提交
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
	die(str, regs, 0);
	return 0;
}

/*
 * Note that we play around with the 'TS' bit in an attempt to get
 * the correct behaviour even in the presence of the asynchronous
 * IRQ13 behaviour
 */
asmlinkage void do_coprocessor_error(struct pt_regs *regs)
{
	void __user *rip = (void __user *)(regs->rip);
	struct task_struct * task;
	siginfo_t info;
	unsigned short cwd, swd;

	conditional_sti(regs);
946
	if (!user_mode(regs) &&
947
	    kernel_math_error(regs, "kernel x87 math error", 16))
L
Linus Torvalds 已提交
948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972
		return;

	/*
	 * Save the info for the exception handler and clear the error.
	 */
	task = current;
	save_init_fpu(task);
	task->thread.trap_no = 16;
	task->thread.error_code = 0;
	info.si_signo = SIGFPE;
	info.si_errno = 0;
	info.si_code = __SI_FAULT;
	info.si_addr = rip;
	/*
	 * (~cwd & swd) will mask out exceptions that are not set to unmasked
	 * status.  0x3f is the exception bits in these regs, 0x200 is the
	 * C1 reg you need in case of a stack fault, 0x040 is the stack
	 * fault bit.  We should only be taking one exception at a time,
	 * so if this combination doesn't produce any single exception,
	 * then we have a bad program that isn't synchronizing its FPU usage
	 * and it will suffer the consequences since we won't be able to
	 * fully reproduce the context of the exception
	 */
	cwd = get_fpu_cwd(task);
	swd = get_fpu_swd(task);
973
	switch (swd & ~cwd & 0x3f) {
L
Linus Torvalds 已提交
974 975 976 977
		case 0x000:
		default:
			break;
		case 0x001: /* Invalid Op */
978 979 980 981 982
			/*
			 * swd & 0x240 == 0x040: Stack Underflow
			 * swd & 0x240 == 0x240: Stack Overflow
			 * User must clear the SF bit (0x40) if set
			 */
L
Linus Torvalds 已提交
983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
			info.si_code = FPE_FLTINV;
			break;
		case 0x002: /* Denormalize */
		case 0x010: /* Underflow */
			info.si_code = FPE_FLTUND;
			break;
		case 0x004: /* Zero Divide */
			info.si_code = FPE_FLTDIV;
			break;
		case 0x008: /* Overflow */
			info.si_code = FPE_FLTOVF;
			break;
		case 0x020: /* Precision */
			info.si_code = FPE_FLTRES;
			break;
	}
	force_sig_info(SIGFPE, &info, task);
}

asmlinkage void bad_intr(void)
{
	printk("bad interrupt"); 
}

asmlinkage void do_simd_coprocessor_error(struct pt_regs *regs)
{
	void __user *rip = (void __user *)(regs->rip);
	struct task_struct * task;
	siginfo_t info;
	unsigned short mxcsr;

	conditional_sti(regs);
1015
	if (!user_mode(regs) &&
1016
        	kernel_math_error(regs, "kernel simd math error", 19))
L
Linus Torvalds 已提交
1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
		return;

	/*
	 * Save the info for the exception handler and clear the error.
	 */
	task = current;
	save_init_fpu(task);
	task->thread.trap_no = 19;
	task->thread.error_code = 0;
	info.si_signo = SIGFPE;
	info.si_errno = 0;
	info.si_code = __SI_FAULT;
	info.si_addr = rip;
	/*
	 * The SIMD FPU exceptions are handled a little differently, as there
	 * is only a single status/control register.  Thus, to determine which
	 * unmasked exception was caught we must mask the exception mask bits
	 * at 0x1f80, and then use these to mask the exception bits at 0x3f.
	 */
	mxcsr = get_fpu_mxcsr(task);
	switch (~((mxcsr & 0x1f80) >> 7) & (mxcsr & 0x3f)) {
		case 0x000:
		default:
			break;
		case 0x001: /* Invalid Op */
			info.si_code = FPE_FLTINV;
			break;
		case 0x002: /* Denormalize */
		case 0x010: /* Underflow */
			info.si_code = FPE_FLTUND;
			break;
		case 0x004: /* Zero Divide */
			info.si_code = FPE_FLTDIV;
			break;
		case 0x008: /* Overflow */
			info.si_code = FPE_FLTOVF;
			break;
		case 0x020: /* Precision */
			info.si_code = FPE_FLTRES;
			break;
	}
	force_sig_info(SIGFPE, &info, task);
}

asmlinkage void do_spurious_interrupt_bug(struct pt_regs * regs)
{
}

asmlinkage void __attribute__((weak)) smp_thermal_interrupt(void)
1066 1067 1068 1069
{
}

asmlinkage void __attribute__((weak)) mce_threshold_interrupt(void)
L
Linus Torvalds 已提交
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
{
}

/*
 *  'math_state_restore()' saves the current math information in the
 * old math state array, and gets the new ones from the current task
 *
 * Careful.. There are problems with IBM-designed IRQ13 behaviour.
 * Don't touch unless you *really* know how it works.
 */
asmlinkage void math_state_restore(void)
{
	struct task_struct *me = current;
	clts();			/* Allow maths ops (or we recurse) */

	if (!used_math())
		init_fpu(me);
	restore_fpu_checking(&me->thread.i387.fxsave);
A
Al Viro 已提交
1088
	task_thread_info(me)->status |= TS_USEDFPU;
1089
	me->fpu_counter++;
L
Linus Torvalds 已提交
1090 1091 1092 1093 1094 1095 1096
}

void __init trap_init(void)
{
	set_intr_gate(0,&divide_error);
	set_intr_gate_ist(1,&debug,DEBUG_STACK);
	set_intr_gate_ist(2,&nmi,NMI_STACK);
1097
 	set_system_gate_ist(3,&int3,DEBUG_STACK); /* int3 can be called from all */
1098 1099
	set_system_gate(4,&overflow);	/* int4 can be called from all */
	set_intr_gate(5,&bounds);
L
Linus Torvalds 已提交
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
	set_intr_gate(6,&invalid_op);
	set_intr_gate(7,&device_not_available);
	set_intr_gate_ist(8,&double_fault, DOUBLEFAULT_STACK);
	set_intr_gate(9,&coprocessor_segment_overrun);
	set_intr_gate(10,&invalid_TSS);
	set_intr_gate(11,&segment_not_present);
	set_intr_gate_ist(12,&stack_segment,STACKFAULT_STACK);
	set_intr_gate(13,&general_protection);
	set_intr_gate(14,&page_fault);
	set_intr_gate(15,&spurious_interrupt_bug);
	set_intr_gate(16,&coprocessor_error);
	set_intr_gate(17,&alignment_check);
#ifdef CONFIG_X86_MCE
	set_intr_gate_ist(18,&machine_check, MCE_STACK); 
#endif
	set_intr_gate(19,&simd_coprocessor_error);

#ifdef CONFIG_IA32_EMULATION
	set_system_gate(IA32_SYSCALL_VECTOR, ia32_syscall);
#endif
       
	/*
	 * Should be a barrier for any external CPU state.
	 */
	cpu_init();
}


1128
static int __init oops_setup(char *s)
L
Linus Torvalds 已提交
1129
{ 
1130 1131 1132 1133 1134
	if (!s)
		return -EINVAL;
	if (!strcmp(s, "panic"))
		panic_on_oops = 1;
	return 0;
L
Linus Torvalds 已提交
1135
} 
1136
early_param("oops", oops_setup);
L
Linus Torvalds 已提交
1137 1138 1139

static int __init kstack_setup(char *s)
{
1140 1141
	if (!s)
		return -EINVAL;
L
Linus Torvalds 已提交
1142
	kstack_depth_to_print = simple_strtoul(s,NULL,0);
1143
	return 0;
L
Linus Torvalds 已提交
1144
}
1145
early_param("kstack", kstack_setup);