trace_ksym.c 12.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
/*
 * trace_ksym.c - Kernel Symbol Tracer
 *
 * 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.
 *
 * Copyright (C) IBM Corporation, 2009
 */

#include <linux/kallsyms.h>
#include <linux/uaccess.h>
#include <linux/debugfs.h>
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/fs.h>

#include "trace_output.h"
#include "trace_stat.h"
#include "trace.h"

/* For now, let us restrict the no. of symbols traced simultaneously to number
 * of available hardware breakpoint registers.
 */
#define KSYM_TRACER_MAX HBP_NUM

#define KSYM_TRACER_OP_LEN 3 /* rw- */

39 40 41 42 43 44 45 46 47
struct trace_ksym {
	struct hw_breakpoint	*ksym_hbp;
	unsigned long		ksym_addr;
#ifdef CONFIG_PROFILE_KSYM_TRACER
	unsigned long		counter;
#endif
	struct hlist_node	ksym_hlist;
};

48 49 50 51 52 53 54
static struct trace_array *ksym_trace_array;

static unsigned int ksym_filter_entry_count;
static unsigned int ksym_tracing_enabled;

static HLIST_HEAD(ksym_filter_head);

55 56
static DEFINE_MUTEX(ksym_tracer_mutex);

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
#ifdef CONFIG_PROFILE_KSYM_TRACER

#define MAX_UL_INT 0xffffffff

void ksym_collect_stats(unsigned long hbp_hit_addr)
{
	struct hlist_node *node;
	struct trace_ksym *entry;

	rcu_read_lock();
	hlist_for_each_entry_rcu(entry, node, &ksym_filter_head, ksym_hlist) {
		if ((entry->ksym_addr == hbp_hit_addr) &&
		    (entry->counter <= MAX_UL_INT)) {
			entry->counter++;
			break;
		}
	}
	rcu_read_unlock();
}
#endif /* CONFIG_PROFILE_KSYM_TRACER */

void ksym_hbp_handler(struct hw_breakpoint *hbp, struct pt_regs *regs)
{
	struct ring_buffer_event *event;
	struct trace_array *tr;
82
	struct ksym_trace_entry *entry;
83 84 85 86 87 88 89 90 91 92 93 94 95
	int pc;

	if (!ksym_tracing_enabled)
		return;

	tr = ksym_trace_array;
	pc = preempt_count();

	event = trace_buffer_lock_reserve(tr, TRACE_KSYM,
							sizeof(*entry), 0, pc);
	if (!event)
		return;

96 97 98
	entry		= ring_buffer_event_data(event);
	entry->ip	= instruction_pointer(regs);
	entry->type	= hbp->info.type;
99
	strlcpy(entry->ksym_name, hbp->info.name, KSYM_SYMBOL_LEN);
100 101
	strlcpy(entry->cmd, current->comm, TASK_COMM_LEN);

102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
#ifdef CONFIG_PROFILE_KSYM_TRACER
	ksym_collect_stats(hbp->info.address);
#endif /* CONFIG_PROFILE_KSYM_TRACER */

	trace_buffer_unlock_commit(tr, event, 0, pc);
}

/* Valid access types are represented as
 *
 * rw- : Set Read/Write Access Breakpoint
 * -w- : Set Write Access Breakpoint
 * --- : Clear Breakpoints
 * --x : Set Execution Break points (Not available yet)
 *
 */
117
static int ksym_trace_get_access_type(char *str)
118
{
119
	int access = 0;
120

121 122 123 124 125 126 127 128 129 130 131 132
	if (str[0] == 'r')
		access += 4;
	else if (str[0] != '-')
		return -EINVAL;

	if (str[1] == 'w')
		access += 2;
	else if (str[1] != '-')
		return -EINVAL;

	if (str[2] != '-')
		return -EINVAL;
133 134 135 136 137

	switch (access) {
	case 6:
		access = HW_BREAKPOINT_RW;
		break;
138 139 140
	case 4:
		access = -EINVAL;
		break;
141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
	case 2:
		access = HW_BREAKPOINT_WRITE;
		break;
	}

	return access;
}

/*
 * There can be several possible malformed requests and we attempt to capture
 * all of them. We enumerate some of the rules
 * 1. We will not allow kernel symbols with ':' since it is used as a delimiter.
 *    i.e. multiple ':' symbols disallowed. Possible uses are of the form
 *    <module>:<ksym_name>:<op>.
 * 2. No delimiter symbol ':' in the input string
 * 3. Spurious operator symbols or symbols not in their respective positions
 * 4. <ksym_name>:--- i.e. clear breakpoint request when ksym_name not in file
 * 5. Kernel symbol not a part of /proc/kallsyms
 * 6. Duplicate requests
 */
static int parse_ksym_trace_str(char *input_string, char **ksymname,
							unsigned long *addr)
{
	int ret;

166
	*ksymname = strsep(&input_string, ":");
167 168 169 170
	*addr = kallsyms_lookup_name(*ksymname);

	/* Check for malformed request: (2), (1) and (5) */
	if ((!input_string) ||
171 172 173 174
	    (strlen(input_string) != KSYM_TRACER_OP_LEN) ||
	    (*addr == 0))
		return -EINVAL;;

175 176 177 178 179 180 181 182
	ret = ksym_trace_get_access_type(input_string);

	return ret;
}

int process_new_ksym_entry(char *ksymname, int op, unsigned long addr)
{
	struct trace_ksym *entry;
L
Li Zefan 已提交
183
	int ret = -ENOMEM;
184 185 186 187 188 189 190 191 192 193 194 195 196

	if (ksym_filter_entry_count >= KSYM_TRACER_MAX) {
		printk(KERN_ERR "ksym_tracer: Maximum limit:(%d) reached. No"
		" new requests for tracing can be accepted now.\n",
			KSYM_TRACER_MAX);
		return -ENOSPC;
	}

	entry = kzalloc(sizeof(struct trace_ksym), GFP_KERNEL);
	if (!entry)
		return -ENOMEM;

	entry->ksym_hbp = kzalloc(sizeof(struct hw_breakpoint), GFP_KERNEL);
L
Li Zefan 已提交
197 198 199 200 201 202
	if (!entry->ksym_hbp)
		goto err;

	entry->ksym_hbp->info.name = kstrdup(ksymname, GFP_KERNEL);
	if (!entry->ksym_hbp->info.name)
		goto err;
203 204 205 206 207 208 209 210 211 212 213 214

	entry->ksym_hbp->info.type = op;
	entry->ksym_addr = entry->ksym_hbp->info.address = addr;
#ifdef CONFIG_X86
	entry->ksym_hbp->info.len = HW_BREAKPOINT_LEN_4;
#endif
	entry->ksym_hbp->triggered = (void *)ksym_hbp_handler;

	ret = register_kernel_hw_breakpoint(entry->ksym_hbp);
	if (ret < 0) {
		printk(KERN_INFO "ksym_tracer request failed. Try again"
					" later!!\n");
L
Li Zefan 已提交
215 216
		ret = -EAGAIN;
		goto err;
217 218 219 220
	}
	hlist_add_head_rcu(&(entry->ksym_hlist), &ksym_filter_head);
	ksym_filter_entry_count++;
	return 0;
L
Li Zefan 已提交
221 222 223 224 225 226
err:
	if (entry->ksym_hbp)
		kfree(entry->ksym_hbp->info.name);
	kfree(entry->ksym_hbp);
	kfree(entry);
	return ret;
227 228 229 230 231 232 233
}

static ssize_t ksym_trace_filter_read(struct file *filp, char __user *ubuf,
						size_t count, loff_t *ppos)
{
	struct trace_ksym *entry;
	struct hlist_node *node;
234 235 236 237 238 239 240 241
	struct trace_seq *s;
	ssize_t cnt = 0;
	int ret;

	s = kmalloc(sizeof(*s), GFP_KERNEL);
	if (!s)
		return -ENOMEM;
	trace_seq_init(s);
242 243 244 245

	mutex_lock(&ksym_tracer_mutex);

	hlist_for_each_entry(entry, node, &ksym_filter_head, ksym_hlist) {
246
		ret = trace_seq_printf(s, "%s:", entry->ksym_hbp->info.name);
247
		if (entry->ksym_hbp->info.type == HW_BREAKPOINT_WRITE)
248
			ret = trace_seq_puts(s, "-w-\n");
249
		else if (entry->ksym_hbp->info.type == HW_BREAKPOINT_RW)
250 251
			ret = trace_seq_puts(s, "rw-\n");
		WARN_ON_ONCE(!ret);
252
	}
253 254 255

	cnt = simple_read_from_buffer(ubuf, count, ppos, s->buffer, s->len);

256 257
	mutex_unlock(&ksym_tracer_mutex);

258 259 260
	kfree(s);

	return cnt;
261 262
}

263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281
static void __ksym_trace_reset(void)
{
	struct trace_ksym *entry;
	struct hlist_node *node, *node1;

	mutex_lock(&ksym_tracer_mutex);
	hlist_for_each_entry_safe(entry, node, node1, &ksym_filter_head,
								ksym_hlist) {
		unregister_kernel_hw_breakpoint(entry->ksym_hbp);
		ksym_filter_entry_count--;
		hlist_del_rcu(&(entry->ksym_hlist));
		synchronize_rcu();
		kfree(entry->ksym_hbp->info.name);
		kfree(entry->ksym_hbp);
		kfree(entry);
	}
	mutex_unlock(&ksym_tracer_mutex);
}

282 283 284 285 286 287 288 289 290 291
static ssize_t ksym_trace_filter_write(struct file *file,
					const char __user *buffer,
						size_t count, loff_t *ppos)
{
	struct trace_ksym *entry;
	struct hlist_node *node;
	char *input_string, *ksymname = NULL;
	unsigned long ksym_addr = 0;
	int ret, op, changed = 0;

292
	input_string = kzalloc(count + 1, GFP_KERNEL);
293 294 295 296 297 298 299
	if (!input_string)
		return -ENOMEM;

	if (copy_from_user(input_string, buffer, count)) {
		kfree(input_string);
		return -EFAULT;
	}
300
	input_string[count] = '\0';
301

302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
	strstrip(input_string);

	/*
	 * Clear all breakpoints if:
	 * 1: echo > ksym_trace_filter
	 * 2: echo 0 > ksym_trace_filter
	 * 3: echo "*:---" > ksym_trace_filter
	 */
	if (!input_string[0] || !strcmp(input_string, "0") ||
	    !strcmp(input_string, "*:---")) {
		__ksym_trace_reset();
		kfree(input_string);
		return count;
	}

317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
	ret = op = parse_ksym_trace_str(input_string, &ksymname, &ksym_addr);
	if (ret < 0) {
		kfree(input_string);
		return ret;
	}

	mutex_lock(&ksym_tracer_mutex);

	ret = -EINVAL;
	hlist_for_each_entry(entry, node, &ksym_filter_head, ksym_hlist) {
		if (entry->ksym_addr == ksym_addr) {
			/* Check for malformed request: (6) */
			if (entry->ksym_hbp->info.type != op)
				changed = 1;
			else
L
Li Zefan 已提交
332
				goto out;
333 334 335 336 337 338 339 340
			break;
		}
	}
	if (changed) {
		unregister_kernel_hw_breakpoint(entry->ksym_hbp);
		entry->ksym_hbp->info.type = op;
		if (op > 0) {
			ret = register_kernel_hw_breakpoint(entry->ksym_hbp);
L
Li Zefan 已提交
341 342 343
			if (ret == 0)
				goto out;
		}
344 345 346
		ksym_filter_entry_count--;
		hlist_del_rcu(&(entry->ksym_hlist));
		synchronize_rcu();
L
Li Zefan 已提交
347
		kfree(entry->ksym_hbp->info.name);
348 349
		kfree(entry->ksym_hbp);
		kfree(entry);
350
		ret = 0;
L
Li Zefan 已提交
351
		goto out;
352 353 354
	} else {
		/* Check for malformed request: (4) */
		if (op == 0)
L
Li Zefan 已提交
355
			goto out;
356 357
		ret = process_new_ksym_entry(ksymname, op, ksym_addr);
	}
L
Li Zefan 已提交
358 359
out:
	mutex_unlock(&ksym_tracer_mutex);
360 361 362

	kfree(input_string);

L
Li Zefan 已提交
363 364
	if (!ret)
		ret = count;
365 366 367 368 369 370 371 372 373 374 375 376
	return ret;
}

static const struct file_operations ksym_tracing_fops = {
	.open		= tracing_open_generic,
	.read		= ksym_trace_filter_read,
	.write		= ksym_trace_filter_write,
};

static void ksym_trace_reset(struct trace_array *tr)
{
	ksym_tracing_enabled = 0;
377
	__ksym_trace_reset();
378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
}

static int ksym_trace_init(struct trace_array *tr)
{
	int cpu, ret = 0;

	for_each_online_cpu(cpu)
		tracing_reset(tr, cpu);
	ksym_tracing_enabled = 1;
	ksym_trace_array = tr;

	return ret;
}

static void ksym_trace_print_header(struct seq_file *m)
{
	seq_puts(m,
395 396
		 "#       TASK-PID   CPU#      Symbol                    "
		 "Type    Function\n");
397
	seq_puts(m,
398 399
		 "#          |        |          |                       "
		 " |         |\n");
400 401 402 403 404 405
}

static enum print_line_t ksym_trace_output(struct trace_iterator *iter)
{
	struct trace_entry *entry = iter->ent;
	struct trace_seq *s = &iter->seq;
406
	struct ksym_trace_entry *field;
407 408 409 410 411 412 413 414
	char str[KSYM_SYMBOL_LEN];
	int ret;

	if (entry->type != TRACE_KSYM)
		return TRACE_TYPE_UNHANDLED;

	trace_assign_type(field, entry);

415
	ret = trace_seq_printf(s, "%11s-%-5d [%03d] %-30s ", field->cmd,
416 417 418 419
				entry->pid, iter->cpu, field->ksym_name);
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

420
	switch (field->type) {
421 422 423 424 425 426 427 428 429 430 431 432 433 434
	case HW_BREAKPOINT_WRITE:
		ret = trace_seq_printf(s, " W  ");
		break;
	case HW_BREAKPOINT_RW:
		ret = trace_seq_printf(s, " RW ");
		break;
	default:
		return TRACE_TYPE_PARTIAL_LINE;
	}

	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	sprint_symbol(str, field->ip);
435
	ret = trace_seq_printf(s, "%s\n", str);
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	return TRACE_TYPE_HANDLED;
}

struct tracer ksym_tracer __read_mostly =
{
	.name		= "ksym_tracer",
	.init		= ksym_trace_init,
	.reset		= ksym_trace_reset,
#ifdef CONFIG_FTRACE_SELFTEST
	.selftest	= trace_selftest_startup_ksym,
#endif
	.print_header   = ksym_trace_print_header,
	.print_line	= ksym_trace_output
};

__init static int init_ksym_trace(void)
{
	struct dentry *d_tracer;
	struct dentry *entry;

	d_tracer = tracing_init_dentry();
	ksym_filter_entry_count = 0;

	entry = debugfs_create_file("ksym_trace_filter", 0644, d_tracer,
				    NULL, &ksym_tracing_fops);
	if (!entry)
		pr_warning("Could not create debugfs "
			   "'ksym_trace_filter' file\n");

	return register_tracer(&ksym_tracer);
}
device_initcall(init_ksym_trace);


#ifdef CONFIG_PROFILE_KSYM_TRACER
static int ksym_tracer_stat_headers(struct seq_file *m)
{
476 477 478 479
	seq_puts(m, "  Access Type ");
	seq_puts(m, "  Symbol                                       Counter\n");
	seq_puts(m, "  ----------- ");
	seq_puts(m, "  ------                                       -------\n");
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
	return 0;
}

static int ksym_tracer_stat_show(struct seq_file *m, void *v)
{
	struct hlist_node *stat = v;
	struct trace_ksym *entry;
	int access_type = 0;
	char fn_name[KSYM_NAME_LEN];

	entry = hlist_entry(stat, struct trace_ksym, ksym_hlist);

	if (entry->ksym_hbp)
		access_type = entry->ksym_hbp->info.type;

	switch (access_type) {
	case HW_BREAKPOINT_WRITE:
497
		seq_puts(m, "  W           ");
498 499
		break;
	case HW_BREAKPOINT_RW:
500
		seq_puts(m, "  RW          ");
501 502
		break;
	default:
503
		seq_puts(m, "  NA          ");
504 505 506
	}

	if (lookup_symbol_name(entry->ksym_addr, fn_name) >= 0)
507
		seq_printf(m, "  %-36s", fn_name);
508
	else
509 510
		seq_printf(m, "  %-36s", "<NA>");
	seq_printf(m, " %15lu\n", entry->counter);
511 512 513 514 515 516

	return 0;
}

static void *ksym_tracer_stat_start(struct tracer_stat *trace)
{
517
	return ksym_filter_head.first;
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
}

static void *
ksym_tracer_stat_next(void *v, int idx)
{
	struct hlist_node *stat = v;

	return stat->next;
}

static struct tracer_stat ksym_tracer_stats = {
	.name = "ksym_tracer",
	.stat_start = ksym_tracer_stat_start,
	.stat_next = ksym_tracer_stat_next,
	.stat_headers = ksym_tracer_stat_headers,
	.stat_show = ksym_tracer_stat_show
};

__init static int ksym_tracer_stat_init(void)
{
	int ret;

	ret = register_stat_tracer(&ksym_tracer_stats);
	if (ret) {
		printk(KERN_WARNING "Warning: could not register "
				    "ksym tracer stats\n");
		return 1;
	}

	return 0;
}
fs_initcall(ksym_tracer_stat_init);
#endif /* CONFIG_PROFILE_KSYM_TRACER */