tracepoint.c 14.7 KB
Newer Older
1
/*
2
 * Copyright (C) 2008-2014 Mathieu Desnoyers
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 */
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/types.h>
#include <linux/jhash.h>
#include <linux/list.h>
#include <linux/rcupdate.h>
#include <linux/tracepoint.h>
#include <linux/err.h>
#include <linux/slab.h>
27
#include <linux/sched/signal.h>
28
#include <linux/sched/task.h>
29
#include <linux/static_key.h>
30

31 32
extern struct tracepoint * const __start___tracepoints_ptrs[];
extern struct tracepoint * const __stop___tracepoints_ptrs[];
33 34 35 36

/* Set to 1 to enable tracepoint debug output */
static const int tracepoint_debug;

37
#ifdef CONFIG_MODULES
38
/*
39
 * Tracepoint module list mutex protects the local module list.
40
 */
41
static DEFINE_MUTEX(tracepoint_module_list_mutex);
42

43
/* Local list of struct tp_module */
44 45 46
static LIST_HEAD(tracepoint_module_list);
#endif /* CONFIG_MODULES */

47
/*
48 49
 * tracepoints_mutex protects the builtin and module tracepoints.
 * tracepoints_mutex nests inside tracepoint_module_list_mutex.
50
 */
51
static DEFINE_MUTEX(tracepoints_mutex);
52 53 54

/*
 * Note about RCU :
55
 * It is used to delay the free of multiple probes array until a quiescent
56 57
 * state is reached.
 */
58
struct tp_probes {
59
	struct rcu_head rcu;
60
	struct tracepoint_func probes[0];
61
};
62

63
static inline void *allocate_probes(int count)
64
{
65
	struct tp_probes *p  = kmalloc(count * sizeof(struct tracepoint_func)
66 67
			+ sizeof(struct tp_probes), GFP_KERNEL);
	return p == NULL ? NULL : p->probes;
68 69
}

70
static void rcu_free_old_probes(struct rcu_head *head)
71
{
72
	kfree(container_of(head, struct tp_probes, rcu));
73 74
}

75
static inline void release_probes(struct tracepoint_func *old)
76 77 78 79
{
	if (old) {
		struct tp_probes *tp_probes = container_of(old,
			struct tp_probes, probes[0]);
80
		call_rcu_sched(&tp_probes->rcu, rcu_free_old_probes);
81
	}
82 83
}

84
static void debug_print_probes(struct tracepoint_func *funcs)
85 86 87
{
	int i;

88
	if (!tracepoint_debug || !funcs)
89 90
		return;

91 92
	for (i = 0; funcs[i].func; i++)
		printk(KERN_DEBUG "Probe %d : %p\n", i, funcs[i].func);
93 94
}

95 96 97
static struct tracepoint_func *
func_add(struct tracepoint_func **funcs, struct tracepoint_func *tp_func,
	 int prio)
98
{
99
	struct tracepoint_func *old, *new;
100 101
	int nr_probes = 0;
	int pos = -1;
102

103
	if (WARN_ON(!tp_func->func))
104
		return ERR_PTR(-EINVAL);
105

106 107
	debug_print_probes(*funcs);
	old = *funcs;
108 109
	if (old) {
		/* (N -> N+1), (N != 0, 1) probes */
110 111 112 113
		for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
			/* Insert before probes of lower priority */
			if (pos < 0 && old[nr_probes].prio < prio)
				pos = nr_probes;
114 115
			if (old[nr_probes].func == tp_func->func &&
			    old[nr_probes].data == tp_func->data)
116
				return ERR_PTR(-EEXIST);
117
		}
118 119
	}
	/* + 2 : one for new probe, one for NULL func */
120
	new = allocate_probes(nr_probes + 2);
121 122
	if (new == NULL)
		return ERR_PTR(-ENOMEM);
123 124 125 126 127 128 129 130 131 132 133 134 135 136
	if (old) {
		if (pos < 0) {
			pos = nr_probes;
			memcpy(new, old, nr_probes * sizeof(struct tracepoint_func));
		} else {
			/* Copy higher priority probes ahead of the new probe */
			memcpy(new, old, pos * sizeof(struct tracepoint_func));
			/* Copy the rest after it. */
			memcpy(new + pos + 1, old + pos,
			       (nr_probes - pos) * sizeof(struct tracepoint_func));
		}
	} else
		pos = 0;
	new[pos] = *tp_func;
137
	new[nr_probes + 1].func = NULL;
138 139
	*funcs = new;
	debug_print_probes(*funcs);
140 141 142
	return old;
}

143 144
static void *func_remove(struct tracepoint_func **funcs,
		struct tracepoint_func *tp_func)
145 146
{
	int nr_probes = 0, nr_del = 0, i;
147
	struct tracepoint_func *old, *new;
148

149
	old = *funcs;
150

151
	if (!old)
152
		return ERR_PTR(-ENOENT);
153

154
	debug_print_probes(*funcs);
155
	/* (N -> M), (N > 1, M >= 0) probes */
156
	if (tp_func->func) {
157
		for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
158 159
			if (old[nr_probes].func == tp_func->func &&
			     old[nr_probes].data == tp_func->data)
160 161
				nr_del++;
		}
162 163
	}

164 165 166 167
	/*
	 * If probe is NULL, then nr_probes = nr_del = 0, and then the
	 * entire entry will be removed.
	 */
168 169
	if (nr_probes - nr_del == 0) {
		/* N -> 0, (N > 1) */
170 171
		*funcs = NULL;
		debug_print_probes(*funcs);
172 173 174 175 176
		return old;
	} else {
		int j = 0;
		/* N -> M, (N > 1, M > 0) */
		/* + 1 for NULL */
177
		new = allocate_probes(nr_probes - nr_del + 1);
178 179
		if (new == NULL)
			return ERR_PTR(-ENOMEM);
180
		for (i = 0; old[i].func; i++)
181 182
			if (old[i].func != tp_func->func
					|| old[i].data != tp_func->data)
183
				new[j++] = old[i];
184
		new[nr_probes - nr_del].func = NULL;
185
		*funcs = new;
186
	}
187
	debug_print_probes(*funcs);
188 189 190 191
	return old;
}

/*
192
 * Add the probe function to a tracepoint.
193
 */
194
static int tracepoint_add_func(struct tracepoint *tp,
195
			       struct tracepoint_func *func, int prio)
196
{
197
	struct tracepoint_func *old, *tp_funcs;
198
	int ret;
199

200 201 202 203 204
	if (tp->regfunc && !static_key_enabled(&tp->key)) {
		ret = tp->regfunc();
		if (ret < 0)
			return ret;
	}
205

206 207
	tp_funcs = rcu_dereference_protected(tp->funcs,
			lockdep_is_held(&tracepoints_mutex));
208
	old = func_add(&tp_funcs, func, prio);
209
	if (IS_ERR(old)) {
210
		WARN_ON_ONCE(PTR_ERR(old) != -ENOMEM);
211 212
		return PTR_ERR(old);
	}
213

214 215 216 217 218 219 220
	/*
	 * rcu_assign_pointer has a smp_wmb() which makes sure that the new
	 * probe callbacks array is consistent before setting a pointer to it.
	 * This array is referenced by __DO_TRACE from
	 * include/linux/tracepoints.h. A matching smp_read_barrier_depends()
	 * is used.
	 */
221 222 223
	rcu_assign_pointer(tp->funcs, tp_funcs);
	if (!static_key_enabled(&tp->key))
		static_key_slow_inc(&tp->key);
224
	release_probes(old);
225
	return 0;
226 227 228
}

/*
229
 * Remove a probe function from a tracepoint.
230 231 232 233
 * Note: only waiting an RCU period after setting elem->call to the empty
 * function insures that the original callback is not used anymore. This insured
 * by preempt_disable around the call site.
 */
234 235
static int tracepoint_remove_func(struct tracepoint *tp,
		struct tracepoint_func *func)
236
{
237
	struct tracepoint_func *old, *tp_funcs;
238

239 240
	tp_funcs = rcu_dereference_protected(tp->funcs,
			lockdep_is_held(&tracepoints_mutex));
241 242
	old = func_remove(&tp_funcs, func);
	if (IS_ERR(old)) {
243
		WARN_ON_ONCE(PTR_ERR(old) != -ENOMEM);
244
		return PTR_ERR(old);
245
	}
246

247 248 249 250
	if (!tp_funcs) {
		/* Removed last function */
		if (tp->unregfunc && static_key_enabled(&tp->key))
			tp->unregfunc();
251

252 253
		if (static_key_enabled(&tp->key))
			static_key_slow_dec(&tp->key);
254
	}
255
	rcu_assign_pointer(tp->funcs, tp_funcs);
256
	release_probes(old);
257
	return 0;
258 259
}

260 261
/**
 * tracepoint_probe_register -  Connect a probe to a tracepoint
262
 * @tp: tracepoint
263
 * @probe: probe handler
264
 * @data: tracepoint data
265
 * @prio: priority of this function over other registered functions
266
 *
267 268 269 270 271
 * Returns 0 if ok, error value on error.
 * Note: if @tp is within a module, the caller is responsible for
 * unregistering the probe before the module is gone. This can be
 * performed either with a tracepoint module going notifier, or from
 * within module exit functions.
272
 */
273 274
int tracepoint_probe_register_prio(struct tracepoint *tp, void *probe,
				   void *data, int prio)
275
{
276 277
	struct tracepoint_func tp_func;
	int ret;
278 279

	mutex_lock(&tracepoints_mutex);
280 281
	tp_func.func = probe;
	tp_func.data = data;
282 283
	tp_func.prio = prio;
	ret = tracepoint_add_func(tp, &tp_func, prio);
284
	mutex_unlock(&tracepoints_mutex);
285
	return ret;
286
}
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
EXPORT_SYMBOL_GPL(tracepoint_probe_register_prio);

/**
 * tracepoint_probe_register -  Connect a probe to a tracepoint
 * @tp: tracepoint
 * @probe: probe handler
 * @data: tracepoint data
 * @prio: priority of this function over other registered functions
 *
 * Returns 0 if ok, error value on error.
 * Note: if @tp is within a module, the caller is responsible for
 * unregistering the probe before the module is gone. This can be
 * performed either with a tracepoint module going notifier, or from
 * within module exit functions.
 */
int tracepoint_probe_register(struct tracepoint *tp, void *probe, void *data)
{
	return tracepoint_probe_register_prio(tp, probe, data, TRACEPOINT_DEFAULT_PRIO);
}
306 307 308 309
EXPORT_SYMBOL_GPL(tracepoint_probe_register);

/**
 * tracepoint_probe_unregister -  Disconnect a probe from a tracepoint
310
 * @tp: tracepoint
311
 * @probe: probe function pointer
312
 * @data: tracepoint data
313
 *
314
 * Returns 0 if ok, error value on error.
315
 */
316
int tracepoint_probe_unregister(struct tracepoint *tp, void *probe, void *data)
317
{
318 319
	struct tracepoint_func tp_func;
	int ret;
320 321

	mutex_lock(&tracepoints_mutex);
322 323 324
	tp_func.func = probe;
	tp_func.data = data;
	ret = tracepoint_remove_func(tp, &tp_func);
325
	mutex_unlock(&tracepoints_mutex);
326
	return ret;
327 328 329
}
EXPORT_SYMBOL_GPL(tracepoint_probe_unregister);

330
#ifdef CONFIG_MODULES
331 332
bool trace_module_has_bad_taint(struct module *mod)
{
333 334
	return mod->taints & ~((1 << TAINT_OOT_MODULE) | (1 << TAINT_CRAP) |
			       (1 << TAINT_UNSIGNED_MODULE));
335 336
}

337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
static BLOCKING_NOTIFIER_HEAD(tracepoint_notify_list);

/**
 * register_tracepoint_notifier - register tracepoint coming/going notifier
 * @nb: notifier block
 *
 * Notifiers registered with this function are called on module
 * coming/going with the tracepoint_module_list_mutex held.
 * The notifier block callback should expect a "struct tp_module" data
 * pointer.
 */
int register_tracepoint_module_notifier(struct notifier_block *nb)
{
	struct tp_module *tp_mod;
	int ret;

	mutex_lock(&tracepoint_module_list_mutex);
	ret = blocking_notifier_chain_register(&tracepoint_notify_list, nb);
	if (ret)
		goto end;
	list_for_each_entry(tp_mod, &tracepoint_module_list, list)
		(void) nb->notifier_call(nb, MODULE_STATE_COMING, tp_mod);
end:
	mutex_unlock(&tracepoint_module_list_mutex);
	return ret;
}
EXPORT_SYMBOL_GPL(register_tracepoint_module_notifier);

/**
 * unregister_tracepoint_notifier - unregister tracepoint coming/going notifier
 * @nb: notifier block
 *
 * The notifier block callback should expect a "struct tp_module" data
 * pointer.
 */
int unregister_tracepoint_module_notifier(struct notifier_block *nb)
{
	struct tp_module *tp_mod;
	int ret;

	mutex_lock(&tracepoint_module_list_mutex);
	ret = blocking_notifier_chain_unregister(&tracepoint_notify_list, nb);
	if (ret)
		goto end;
	list_for_each_entry(tp_mod, &tracepoint_module_list, list)
		(void) nb->notifier_call(nb, MODULE_STATE_GOING, tp_mod);
end:
	mutex_unlock(&tracepoint_module_list_mutex);
	return ret;

}
EXPORT_SYMBOL_GPL(unregister_tracepoint_module_notifier);

/*
 * Ensure the tracer unregistered the module's probes before the module
 * teardown is performed. Prevents leaks of probe and data pointers.
 */
static void tp_module_going_check_quiescent(struct tracepoint * const *begin,
		struct tracepoint * const *end)
{
	struct tracepoint * const *iter;

	if (!begin)
		return;
	for (iter = begin; iter < end; iter++)
		WARN_ON_ONCE((*iter)->funcs);
}

405 406
static int tracepoint_module_coming(struct module *mod)
{
407
	struct tp_module *tp_mod;
408 409
	int ret = 0;

410 411 412
	if (!mod->num_tracepoints)
		return 0;

413
	/*
414 415
	 * We skip modules that taint the kernel, especially those with different
	 * module headers (for forced load), to make sure we don't cause a crash.
416
	 * Staging, out-of-tree, and unsigned GPL modules are fine.
417
	 */
418
	if (trace_module_has_bad_taint(mod))
419
		return 0;
420
	mutex_lock(&tracepoint_module_list_mutex);
421 422 423 424 425
	tp_mod = kmalloc(sizeof(struct tp_module), GFP_KERNEL);
	if (!tp_mod) {
		ret = -ENOMEM;
		goto end;
	}
426
	tp_mod->mod = mod;
427
	list_add_tail(&tp_mod->list, &tracepoint_module_list);
428 429
	blocking_notifier_call_chain(&tracepoint_notify_list,
			MODULE_STATE_COMING, tp_mod);
430
end:
431
	mutex_unlock(&tracepoint_module_list_mutex);
432 433 434
	return ret;
}

435
static void tracepoint_module_going(struct module *mod)
436
{
437
	struct tp_module *tp_mod;
438

439
	if (!mod->num_tracepoints)
440
		return;
441

442 443
	mutex_lock(&tracepoint_module_list_mutex);
	list_for_each_entry(tp_mod, &tracepoint_module_list, list) {
444
		if (tp_mod->mod == mod) {
445 446 447 448 449 450 451 452 453 454
			blocking_notifier_call_chain(&tracepoint_notify_list,
					MODULE_STATE_GOING, tp_mod);
			list_del(&tp_mod->list);
			kfree(tp_mod);
			/*
			 * Called the going notifier before checking for
			 * quiescence.
			 */
			tp_module_going_check_quiescent(mod->tracepoints_ptrs,
				mod->tracepoints_ptrs + mod->num_tracepoints);
455 456 457 458 459 460 461 462 463
			break;
		}
	}
	/*
	 * In the case of modules that were tainted at "coming", we'll simply
	 * walk through the list without finding it. We cannot use the "tainted"
	 * flag on "going", in case a module taints the kernel only after being
	 * loaded.
	 */
464
	mutex_unlock(&tracepoint_module_list_mutex);
465
}
466

467 468
static int tracepoint_module_notify(struct notifier_block *self,
		unsigned long val, void *data)
469 470
{
	struct module *mod = data;
471
	int ret = 0;
472 473 474

	switch (val) {
	case MODULE_STATE_COMING:
475 476 477 478
		ret = tracepoint_module_coming(mod);
		break;
	case MODULE_STATE_LIVE:
		break;
479
	case MODULE_STATE_GOING:
480 481 482
		tracepoint_module_going(mod);
		break;
	case MODULE_STATE_UNFORMED:
483 484
		break;
	}
485
	return ret;
486 487
}

488
static struct notifier_block tracepoint_module_nb = {
489 490 491 492
	.notifier_call = tracepoint_module_notify,
	.priority = 0,
};

493
static __init int init_tracepoints(void)
494
{
495 496 497
	int ret;

	ret = register_module_notifier(&tracepoint_module_nb);
498
	if (ret)
499
		pr_warn("Failed to register tracepoint module enter notifier\n");
500

501
	return ret;
502 503
}
__initcall(init_tracepoints);
504
#endif /* CONFIG_MODULES */
505

506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
static void for_each_tracepoint_range(struct tracepoint * const *begin,
		struct tracepoint * const *end,
		void (*fct)(struct tracepoint *tp, void *priv),
		void *priv)
{
	struct tracepoint * const *iter;

	if (!begin)
		return;
	for (iter = begin; iter < end; iter++)
		fct(*iter, priv);
}

/**
 * for_each_kernel_tracepoint - iteration on all kernel tracepoints
 * @fct: callback
 * @priv: private data
 */
void for_each_kernel_tracepoint(void (*fct)(struct tracepoint *tp, void *priv),
		void *priv)
{
	for_each_tracepoint_range(__start___tracepoints_ptrs,
		__stop___tracepoints_ptrs, fct, priv);
}
EXPORT_SYMBOL_GPL(for_each_kernel_tracepoint);

532
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
533

534
/* NB: reg/unreg are called while guarded with the tracepoints_mutex */
535 536
static int sys_tracepoint_refcount;

537
int syscall_regfunc(void)
538
{
539
	struct task_struct *p, *t;
540 541

	if (!sys_tracepoint_refcount) {
542 543
		read_lock(&tasklist_lock);
		for_each_process_thread(p, t) {
544
			set_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
545 546
		}
		read_unlock(&tasklist_lock);
547 548
	}
	sys_tracepoint_refcount++;
549 550

	return 0;
551 552 553 554
}

void syscall_unregfunc(void)
{
555
	struct task_struct *p, *t;
556 557 558

	sys_tracepoint_refcount--;
	if (!sys_tracepoint_refcount) {
559 560
		read_lock(&tasklist_lock);
		for_each_process_thread(p, t) {
561
			clear_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
562 563
		}
		read_unlock(&tasklist_lock);
564 565
	}
}
566
#endif