printk.c 44.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12
/*
 *  linux/kernel/printk.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 * Modified to make sys_syslog() more flexible: added commands to
 * return the last 4k of kernel messages, regardless of whether
 * they've been read or not.  Added option to suppress kernel printk's
 * to the console.  Added hook for sending the console messages
 * elsewhere, in preparation for a serial line console (someday).
 * Ted Ts'o, 2/11/93.
 * Modified for sysctl support, 1/8/97, Chris Horn.
J
Jesper Juhl 已提交
13
 * Fixed SMP synchronization, 08/08/99, Manfred Spraul
14
 *     manfred@colorfullife.com
L
Linus Torvalds 已提交
15
 * Rewrote bits to get rid of console_lock
16
 *	01Mar01 Andrew Morton
L
Linus Torvalds 已提交
17 18 19 20 21 22 23 24
 */

#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/tty.h>
#include <linux/tty_driver.h>
#include <linux/console.h>
#include <linux/init.h>
R
Randy Dunlap 已提交
25 26
#include <linux/jiffies.h>
#include <linux/nmi.h>
L
Linus Torvalds 已提交
27
#include <linux/module.h>
J
Jan Engelhardt 已提交
28
#include <linux/moduleparam.h>
L
Linus Torvalds 已提交
29 30 31 32 33
#include <linux/interrupt.h>			/* For in_interrupt() */
#include <linux/delay.h>
#include <linux/smp.h>
#include <linux/security.h>
#include <linux/bootmem.h>
34
#include <linux/memblock.h>
L
Linus Torvalds 已提交
35
#include <linux/syscalls.h>
36
#include <linux/kexec.h>
37
#include <linux/kdb.h>
38
#include <linux/ratelimit.h>
39
#include <linux/kmsg_dump.h>
40
#include <linux/syslog.h>
41 42
#include <linux/cpu.h>
#include <linux/notifier.h>
43
#include <linux/rculist.h>
L
Linus Torvalds 已提交
44 45 46

#include <asm/uaccess.h>

47 48 49
#define CREATE_TRACE_POINTS
#include <trace/events/printk.h>

50 51 52
/*
 * Architectures can override it:
 */
53
void asmlinkage __attribute__((weak)) early_printk(const char *fmt, ...)
54 55 56
{
}

L
Linus Torvalds 已提交
57 58 59
#define __LOG_BUF_LEN	(1 << CONFIG_LOG_BUF_SHIFT)

/* printk's without a loglevel use this.. */
60
#define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL
L
Linus Torvalds 已提交
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75

/* We show everything that is MORE important than this.. */
#define MINIMUM_CONSOLE_LOGLEVEL 1 /* Minimum loglevel we let people use */
#define DEFAULT_CONSOLE_LOGLEVEL 7 /* anything MORE serious than KERN_DEBUG */

DECLARE_WAIT_QUEUE_HEAD(log_wait);

int console_printk[4] = {
	DEFAULT_CONSOLE_LOGLEVEL,	/* console_loglevel */
	DEFAULT_MESSAGE_LOGLEVEL,	/* default_message_loglevel */
	MINIMUM_CONSOLE_LOGLEVEL,	/* minimum_console_loglevel */
	DEFAULT_CONSOLE_LOGLEVEL,	/* default_console_loglevel */
};

/*
P
Patrick Pletscher 已提交
76
 * Low level drivers may need that to know if they can schedule in
L
Linus Torvalds 已提交
77 78 79 80 81 82 83 84 85 86
 * their unblank() callback or not. So let's export it.
 */
int oops_in_progress;
EXPORT_SYMBOL(oops_in_progress);

/*
 * console_sem protects the console_drivers list, and also
 * provides serialisation for access to the entire console
 * driver system.
 */
87
static DEFINE_SEMAPHORE(console_sem);
L
Linus Torvalds 已提交
88
struct console *console_drivers;
I
Ingo Molnar 已提交
89 90
EXPORT_SYMBOL_GPL(console_drivers);

L
Linus Torvalds 已提交
91 92 93 94 95 96 97 98
/*
 * This is used for debugging the mess that is the VT code by
 * keeping track if we have the console semaphore held. It's
 * definitely not the perfect debug tool (we don't know if _WE_
 * hold it are racing, but it helps tracking those weird code
 * path in the console code where we end up in places I want
 * locked without the console sempahore held
 */
99
static int console_locked, console_suspended;
L
Linus Torvalds 已提交
100 101 102 103

/*
 * logbuf_lock protects log_buf, log_start, log_end, con_start and logged_chars
 * It is also used in interesting ways to provide interlocking in
104
 * console_unlock();.
L
Linus Torvalds 已提交
105
 */
106
static DEFINE_RAW_SPINLOCK(logbuf_lock);
L
Linus Torvalds 已提交
107

108
#define LOG_BUF_MASK (log_buf_len-1)
L
Linus Torvalds 已提交
109 110 111 112 113 114
#define LOG_BUF(idx) (log_buf[(idx) & LOG_BUF_MASK])

/*
 * The indices into log_buf are not constrained to log_buf_len - they
 * must be masked before subscripting
 */
115 116 117
static unsigned log_start;	/* Index into log_buf: next char to be read by syslog() */
static unsigned con_start;	/* Index into log_buf: next char to be sent to consoles */
static unsigned log_end;	/* Index into log_buf: most-recently-written-char + 1 */
L
Linus Torvalds 已提交
118

119 120 121 122 123
/*
 * If exclusive_console is non-NULL then only this console is to be printed to.
 */
static struct console *exclusive_console;

L
Linus Torvalds 已提交
124 125 126 127 128 129 130 131
/*
 *	Array of consoles built from command line options (console=)
 */
struct console_cmdline
{
	char	name[8];			/* Name of the driver	    */
	int	index;				/* Minor dev. to use	    */
	char	*options;			/* Options for the driver   */
132 133 134
#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
	char	*brl_options;			/* Options for braille driver */
#endif
L
Linus Torvalds 已提交
135 136 137 138 139 140 141
};

#define MAX_CMDLINECONSOLES 8

static struct console_cmdline console_cmdline[MAX_CMDLINECONSOLES];
static int selected_console = -1;
static int preferred_console = -1;
142 143
int console_set_on_cmdline;
EXPORT_SYMBOL(console_set_on_cmdline);
L
Linus Torvalds 已提交
144 145 146 147

/* Flag: console code may call schedule() */
static int console_may_schedule;

M
Matt Mackall 已提交
148 149 150 151 152
#ifdef CONFIG_PRINTK

static char __log_buf[__LOG_BUF_LEN];
static char *log_buf = __log_buf;
static int log_buf_len = __LOG_BUF_LEN;
153
static unsigned logged_chars; /* Number of chars produced since last read+clear operation */
154
static int saved_console_loglevel = -1;
M
Matt Mackall 已提交
155

156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
#ifdef CONFIG_KEXEC
/*
 * This appends the listed symbols to /proc/vmcoreinfo
 *
 * /proc/vmcoreinfo is used by various utiilties, like crash and makedumpfile to
 * obtain access to symbols that are otherwise very difficult to locate.  These
 * symbols are specifically used so that utilities can access and extract the
 * dmesg log from a vmcore file after a crash.
 */
void log_buf_kexec_setup(void)
{
	VMCOREINFO_SYMBOL(log_buf);
	VMCOREINFO_SYMBOL(log_end);
	VMCOREINFO_SYMBOL(log_buf_len);
	VMCOREINFO_SYMBOL(logged_chars);
}
#endif

174 175 176 177
/* requested log_buf_len from kernel cmdline */
static unsigned long __initdata new_log_buf_len;

/* save requested log_buf_len since it's too early to process it */
L
Linus Torvalds 已提交
178 179
static int __init log_buf_len_setup(char *str)
{
180
	unsigned size = memparse(str, &str);
L
Linus Torvalds 已提交
181 182 183

	if (size)
		size = roundup_pow_of_two(size);
184 185 186 187
	if (size > log_buf_len)
		new_log_buf_len = size;

	return 0;
L
Linus Torvalds 已提交
188
}
189 190 191 192 193 194 195 196 197 198 199
early_param("log_buf_len", log_buf_len_setup);

void __init setup_log_buf(int early)
{
	unsigned long flags;
	unsigned start, dest_idx, offset;
	char *new_log_buf;
	int free;

	if (!new_log_buf_len)
		return;
L
Linus Torvalds 已提交
200

201 202 203 204
	if (early) {
		unsigned long mem;

		mem = memblock_alloc(new_log_buf_len, PAGE_SIZE);
T
Tejun Heo 已提交
205
		if (!mem)
206 207 208 209 210 211 212 213 214 215 216 217
			return;
		new_log_buf = __va(mem);
	} else {
		new_log_buf = alloc_bootmem_nopanic(new_log_buf_len);
	}

	if (unlikely(!new_log_buf)) {
		pr_err("log_buf_len: %ld bytes not available\n",
			new_log_buf_len);
		return;
	}

218
	raw_spin_lock_irqsave(&logbuf_lock, flags);
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
	log_buf_len = new_log_buf_len;
	log_buf = new_log_buf;
	new_log_buf_len = 0;
	free = __LOG_BUF_LEN - log_end;

	offset = start = min(con_start, log_start);
	dest_idx = 0;
	while (start != log_end) {
		unsigned log_idx_mask = start & (__LOG_BUF_LEN - 1);

		log_buf[dest_idx] = __log_buf[log_idx_mask];
		start++;
		dest_idx++;
	}
	log_start -= offset;
	con_start -= offset;
	log_end -= offset;
236
	raw_spin_unlock_irqrestore(&logbuf_lock, flags);
237 238 239 240 241

	pr_info("log_buf_len: %d\n", log_buf_len);
	pr_info("early log buf free: %d(%d%%)\n",
		free, (free * 100) / __LOG_BUF_LEN);
}
L
Linus Torvalds 已提交
242

R
Randy Dunlap 已提交
243 244
#ifdef CONFIG_BOOT_PRINTK_DELAY

245
static int boot_delay; /* msecs delay after each printk during bootup */
246
static unsigned long long loops_per_msec;	/* based on boot_delay */
R
Randy Dunlap 已提交
247 248 249 250 251 252 253 254 255 256 257 258

static int __init boot_delay_setup(char *str)
{
	unsigned long lpj;

	lpj = preset_lpj ? preset_lpj : 1000000;	/* some guess */
	loops_per_msec = (unsigned long long)lpj / 1000 * HZ;

	get_option(&str, &boot_delay);
	if (boot_delay > 10 * 1000)
		boot_delay = 0;

259 260 261
	pr_debug("boot_delay: %u, preset_lpj: %ld, lpj: %lu, "
		"HZ: %d, loops_per_msec: %llu\n",
		boot_delay, preset_lpj, lpj, HZ, loops_per_msec);
R
Randy Dunlap 已提交
262 263 264 265 266 267 268 269 270 271 272 273
	return 1;
}
__setup("boot_delay=", boot_delay_setup);

static void boot_delay_msec(void)
{
	unsigned long long k;
	unsigned long timeout;

	if (boot_delay == 0 || system_state != SYSTEM_BOOTING)
		return;

274
	k = (unsigned long long)loops_per_msec * boot_delay;
R
Randy Dunlap 已提交
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295

	timeout = jiffies + msecs_to_jiffies(boot_delay);
	while (k) {
		k--;
		cpu_relax();
		/*
		 * use (volatile) jiffies to prevent
		 * compiler reduction; loop termination via jiffies
		 * is secondary and may or may not happen.
		 */
		if (time_after(jiffies, timeout))
			break;
		touch_nmi_watchdog();
	}
}
#else
static inline void boot_delay_msec(void)
{
}
#endif

296 297 298 299 300 301
#ifdef CONFIG_SECURITY_DMESG_RESTRICT
int dmesg_restrict = 1;
#else
int dmesg_restrict;
#endif

302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323
static int syslog_action_restricted(int type)
{
	if (dmesg_restrict)
		return 1;
	/* Unless restricted, we allow "read all" and "get buffer size" for everybody */
	return type != SYSLOG_ACTION_READ_ALL && type != SYSLOG_ACTION_SIZE_BUFFER;
}

static int check_syslog_permissions(int type, bool from_file)
{
	/*
	 * If this is from /proc/kmsg and we've already opened it, then we've
	 * already done the capabilities checks at open time.
	 */
	if (from_file && type != SYSLOG_ACTION_OPEN)
		return 0;

	if (syslog_action_restricted(type)) {
		if (capable(CAP_SYSLOG))
			return 0;
		/* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */
		if (capable(CAP_SYS_ADMIN)) {
324 325 326 327
			printk_once(KERN_WARNING "%s (%d): "
				 "Attempt to access syslog with CAP_SYS_ADMIN "
				 "but no CAP_SYSLOG (deprecated).\n",
				 current->comm, task_pid_nr(current));
328 329 330 331 332 333 334
			return 0;
		}
		return -EPERM;
	}
	return 0;
}

335
int do_syslog(int type, char __user *buf, int len, bool from_file)
L
Linus Torvalds 已提交
336
{
337
	unsigned i, j, limit, count;
L
Linus Torvalds 已提交
338 339
	int do_clear = 0;
	char c;
340
	int error;
L
Linus Torvalds 已提交
341

342 343 344
	error = check_syslog_permissions(type, from_file);
	if (error)
		goto out;
345 346

	error = security_syslog(type);
L
Linus Torvalds 已提交
347 348 349 350
	if (error)
		return error;

	switch (type) {
351
	case SYSLOG_ACTION_CLOSE:	/* Close log */
L
Linus Torvalds 已提交
352
		break;
353
	case SYSLOG_ACTION_OPEN:	/* Open log */
L
Linus Torvalds 已提交
354
		break;
355
	case SYSLOG_ACTION_READ:	/* Read from log */
L
Linus Torvalds 已提交
356 357 358 359 360 361 362 363 364 365
		error = -EINVAL;
		if (!buf || len < 0)
			goto out;
		error = 0;
		if (!len)
			goto out;
		if (!access_ok(VERIFY_WRITE, buf, len)) {
			error = -EFAULT;
			goto out;
		}
J
Jesper Juhl 已提交
366 367
		error = wait_event_interruptible(log_wait,
							(log_start - log_end));
L
Linus Torvalds 已提交
368 369 370
		if (error)
			goto out;
		i = 0;
371
		raw_spin_lock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
372 373 374
		while (!error && (log_start != log_end) && i < len) {
			c = LOG_BUF(log_start);
			log_start++;
375
			raw_spin_unlock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
376 377 378 379
			error = __put_user(c,buf);
			buf++;
			i++;
			cond_resched();
380
			raw_spin_lock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
381
		}
382
		raw_spin_unlock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
383 384 385
		if (!error)
			error = i;
		break;
386 387
	/* Read/clear last kernel messages */
	case SYSLOG_ACTION_READ_CLEAR:
J
Jesper Juhl 已提交
388
		do_clear = 1;
L
Linus Torvalds 已提交
389
		/* FALL THRU */
390 391
	/* Read last kernel messages */
	case SYSLOG_ACTION_READ_ALL:
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401 402 403 404
		error = -EINVAL;
		if (!buf || len < 0)
			goto out;
		error = 0;
		if (!len)
			goto out;
		if (!access_ok(VERIFY_WRITE, buf, len)) {
			error = -EFAULT;
			goto out;
		}
		count = len;
		if (count > log_buf_len)
			count = log_buf_len;
405
		raw_spin_lock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
406 407 408 409 410 411 412
		if (count > logged_chars)
			count = logged_chars;
		if (do_clear)
			logged_chars = 0;
		limit = log_end;
		/*
		 * __put_user() could sleep, and while we sleep
J
Jesper Juhl 已提交
413
		 * printk() could overwrite the messages
L
Linus Torvalds 已提交
414 415 416
		 * we try to copy to user space. Therefore
		 * the messages are copied in reverse. <manfreds>
		 */
J
Jesper Juhl 已提交
417
		for (i = 0; i < count && !error; i++) {
L
Linus Torvalds 已提交
418 419 420 421
			j = limit-1-i;
			if (j + log_buf_len < log_end)
				break;
			c = LOG_BUF(j);
422
			raw_spin_unlock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
423 424
			error = __put_user(c,&buf[count-1-i]);
			cond_resched();
425
			raw_spin_lock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
426
		}
427
		raw_spin_unlock_irq(&logbuf_lock);
L
Linus Torvalds 已提交
428 429 430
		if (error)
			break;
		error = i;
J
Jesper Juhl 已提交
431
		if (i != count) {
L
Linus Torvalds 已提交
432 433
			int offset = count-error;
			/* buffer overflow during copy, correct user buffer. */
J
Jesper Juhl 已提交
434
			for (i = 0; i < error; i++) {
L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442 443
				if (__get_user(c,&buf[i+offset]) ||
				    __put_user(c,&buf[i])) {
					error = -EFAULT;
					break;
				}
				cond_resched();
			}
		}
		break;
444 445
	/* Clear ring buffer */
	case SYSLOG_ACTION_CLEAR:
L
Linus Torvalds 已提交
446 447
		logged_chars = 0;
		break;
448 449
	/* Disable logging to console */
	case SYSLOG_ACTION_CONSOLE_OFF:
450 451
		if (saved_console_loglevel == -1)
			saved_console_loglevel = console_loglevel;
L
Linus Torvalds 已提交
452 453
		console_loglevel = minimum_console_loglevel;
		break;
454 455
	/* Enable logging to console */
	case SYSLOG_ACTION_CONSOLE_ON:
456 457 458 459
		if (saved_console_loglevel != -1) {
			console_loglevel = saved_console_loglevel;
			saved_console_loglevel = -1;
		}
L
Linus Torvalds 已提交
460
		break;
461 462
	/* Set level of messages printed to console */
	case SYSLOG_ACTION_CONSOLE_LEVEL:
L
Linus Torvalds 已提交
463 464 465 466 467 468
		error = -EINVAL;
		if (len < 1 || len > 8)
			goto out;
		if (len < minimum_console_loglevel)
			len = minimum_console_loglevel;
		console_loglevel = len;
469 470
		/* Implicitly re-enable logging to console */
		saved_console_loglevel = -1;
L
Linus Torvalds 已提交
471 472
		error = 0;
		break;
473 474
	/* Number of chars in the log buffer */
	case SYSLOG_ACTION_SIZE_UNREAD:
L
Linus Torvalds 已提交
475 476
		error = log_end - log_start;
		break;
477 478
	/* Size of the log buffer */
	case SYSLOG_ACTION_SIZE_BUFFER:
L
Linus Torvalds 已提交
479 480 481 482 483 484 485 486 487 488
		error = log_buf_len;
		break;
	default:
		error = -EINVAL;
		break;
	}
out:
	return error;
}

489
SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len)
L
Linus Torvalds 已提交
490
{
491
	return do_syslog(type, buf, len, SYSLOG_FROM_CALL);
L
Linus Torvalds 已提交
492 493
}

494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
#ifdef	CONFIG_KGDB_KDB
/* kdb dmesg command needs access to the syslog buffer.  do_syslog()
 * uses locks so it cannot be used during debugging.  Just tell kdb
 * where the start and end of the physical and logical logs are.  This
 * is equivalent to do_syslog(3).
 */
void kdb_syslog_data(char *syslog_data[4])
{
	syslog_data[0] = log_buf;
	syslog_data[1] = log_buf + log_buf_len;
	syslog_data[2] = log_buf + log_end -
		(logged_chars < log_buf_len ? logged_chars : log_buf_len);
	syslog_data[3] = log_buf + log_end;
}
#endif	/* CONFIG_KGDB_KDB */

L
Linus Torvalds 已提交
510 511 512
/*
 * Call the console drivers on a range of log_buf
 */
513
static void __call_console_drivers(unsigned start, unsigned end)
L
Linus Torvalds 已提交
514 515 516
{
	struct console *con;

517
	for_each_console(con) {
518 519
		if (exclusive_console && con != exclusive_console)
			continue;
520 521 522
		if ((con->flags & CON_ENABLED) && con->write &&
				(cpu_online(smp_processor_id()) ||
				(con->flags & CON_ANYTIME)))
L
Linus Torvalds 已提交
523 524 525 526
			con->write(con, &LOG_BUF(start), end - start);
	}
}

527
static bool __read_mostly ignore_loglevel;
528

529
static int __init ignore_loglevel_setup(char *str)
530 531 532 533
{
	ignore_loglevel = 1;
	printk(KERN_INFO "debug: ignoring loglevel setting.\n");

534
	return 0;
535 536
}

537
early_param("ignore_loglevel", ignore_loglevel_setup);
538
module_param(ignore_loglevel, bool, S_IRUGO | S_IWUSR);
539 540
MODULE_PARM_DESC(ignore_loglevel, "ignore loglevel setting, to"
	"print all kernel messages to the console.");
541

L
Linus Torvalds 已提交
542 543 544
/*
 * Write out chars from start to end - 1 inclusive
 */
545 546
static void _call_console_drivers(unsigned start,
				unsigned end, int msg_log_level)
L
Linus Torvalds 已提交
547
{
548 549
	trace_console(&LOG_BUF(0), start, end, log_buf_len);

550
	if ((msg_log_level < console_loglevel || ignore_loglevel) &&
L
Linus Torvalds 已提交
551 552 553 554 555 556 557 558 559 560 561 562
			console_drivers && start != end) {
		if ((start & LOG_BUF_MASK) > (end & LOG_BUF_MASK)) {
			/* wrapped write */
			__call_console_drivers(start & LOG_BUF_MASK,
						log_buf_len);
			__call_console_drivers(0, end & LOG_BUF_MASK);
		} else {
			__call_console_drivers(start, end);
		}
	}
}

563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
/*
 * Parse the syslog header <[0-9]*>. The decimal value represents 32bit, the
 * lower 3 bit are the log level, the rest are the log facility. In case
 * userspace passes usual userspace syslog messages to /dev/kmsg or
 * /dev/ttyprintk, the log prefix might contain the facility. Printk needs
 * to extract the correct log level for in-kernel processing, and not mangle
 * the original value.
 *
 * If a prefix is found, the length of the prefix is returned. If 'level' is
 * passed, it will be filled in with the log level without a possible facility
 * value. If 'special' is passed, the special printk prefix chars are accepted
 * and returned. If no valid header is found, 0 is returned and the passed
 * variables are not touched.
 */
static size_t log_prefix(const char *p, unsigned int *level, char *special)
{
	unsigned int lev = 0;
	char sp = '\0';
	size_t len;

	if (p[0] != '<' || !p[1])
		return 0;
	if (p[2] == '>') {
		/* usual single digit level number or special char */
		switch (p[1]) {
		case '0' ... '7':
			lev = p[1] - '0';
			break;
		case 'c': /* KERN_CONT */
		case 'd': /* KERN_DEFAULT */
			sp = p[1];
			break;
		default:
			return 0;
		}
		len = 3;
	} else {
		/* multi digit including the level and facility number */
		char *endp = NULL;

		lev = (simple_strtoul(&p[1], &endp, 10) & 7);
		if (endp == NULL || endp[0] != '>')
			return 0;
		len = (endp + 1) - p;
	}

	/* do not accept special char if not asked for */
	if (sp && !special)
		return 0;

	if (special) {
		*special = sp;
		/* return special char, do not touch level */
		if (sp)
			return len;
	}

	if (level)
		*level = lev;
	return len;
}

L
Linus Torvalds 已提交
625 626 627
/*
 * Call the console drivers, asking them to write out
 * log_buf[start] to log_buf[end - 1].
628
 * The console_lock must be held.
L
Linus Torvalds 已提交
629
 */
630
static void call_console_drivers(unsigned start, unsigned end)
L
Linus Torvalds 已提交
631
{
632
	unsigned cur_index, start_print;
L
Linus Torvalds 已提交
633 634
	static int msg_level = -1;

635
	BUG_ON(((int)(start - end)) > 0);
L
Linus Torvalds 已提交
636 637 638 639

	cur_index = start;
	start_print = start;
	while (cur_index != end) {
640 641 642
		if (msg_level < 0 && ((end - cur_index) > 2)) {
			/* strip log prefix */
			cur_index += log_prefix(&LOG_BUF(cur_index), &msg_level, NULL);
L
Linus Torvalds 已提交
643 644 645 646 647
			start_print = cur_index;
		}
		while (cur_index != end) {
			char c = LOG_BUF(cur_index);

J
Jesper Juhl 已提交
648
			cur_index++;
L
Linus Torvalds 已提交
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
			if (c == '\n') {
				if (msg_level < 0) {
					/*
					 * printk() has already given us loglevel tags in
					 * the buffer.  This code is here in case the
					 * log buffer has wrapped right round and scribbled
					 * on those tags
					 */
					msg_level = default_message_loglevel;
				}
				_call_console_drivers(start_print, cur_index, msg_level);
				msg_level = -1;
				start_print = cur_index;
				break;
			}
		}
	}
	_call_console_drivers(start_print, end, msg_level);
}

static void emit_log_char(char c)
{
	LOG_BUF(log_end) = c;
	log_end++;
	if (log_end - log_start > log_buf_len)
		log_start = log_end - log_buf_len;
	if (log_end - con_start > log_buf_len)
		con_start = log_end - log_buf_len;
	if (logged_chars < log_buf_len)
		logged_chars++;
}

/*
 * Zap console related locks when oopsing. Only zap at most once
 * every 10 seconds, to leave time for slow consoles to print a
 * full oops.
 */
static void zap_locks(void)
{
	static unsigned long oops_timestamp;

	if (time_after_eq(jiffies, oops_timestamp) &&
J
Jesper Juhl 已提交
691
			!time_after(jiffies, oops_timestamp + 30 * HZ))
L
Linus Torvalds 已提交
692 693 694 695
		return;

	oops_timestamp = jiffies;

696
	debug_locks_off();
L
Linus Torvalds 已提交
697
	/* If a crash is occurring, make sure we can't deadlock */
698
	raw_spin_lock_init(&logbuf_lock);
L
Linus Torvalds 已提交
699
	/* And make sure that we print immediately */
700
	sema_init(&console_sem, 1);
L
Linus Torvalds 已提交
701 702 703
}

#if defined(CONFIG_PRINTK_TIME)
704
static bool printk_time = 1;
L
Linus Torvalds 已提交
705
#else
706
static bool printk_time = 0;
L
Linus Torvalds 已提交
707
#endif
708
module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR);
L
Linus Torvalds 已提交
709

710 711 712
static bool always_kmsg_dump;
module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | S_IWUSR);

713 714 715 716 717
/* Check if we have any console registered that can be called early in boot. */
static int have_callable_console(void)
{
	struct console *con;

718
	for_each_console(con)
719 720 721 722 723 724
		if (con->flags & CON_ANYTIME)
			return 1;

	return 0;
}

725 726 727 728
/**
 * printk - print a kernel message
 * @fmt: format string
 *
729
 * This is printk().  It can be called from any context.  We want it to work.
J
Jesper Juhl 已提交
730
 *
731
 * We try to grab the console_lock.  If we succeed, it's easy - we log the output and
L
Linus Torvalds 已提交
732 733
 * call the console drivers.  If we fail to get the semaphore we place the output
 * into the log buffer and return.  The current holder of the console_sem will
734 735
 * notice the new output in console_unlock(); and will send it to the
 * consoles before releasing the lock.
L
Linus Torvalds 已提交
736 737 738 739
 *
 * One effect of this deferred printing is that code which calls printk() and
 * then changes console_loglevel may break. This is because console_loglevel
 * is inspected when the actual printing occurs.
740 741 742
 *
 * See also:
 * printf(3)
743 744
 *
 * See the vsnprintf() documentation for format string extensions over C99.
L
Linus Torvalds 已提交
745
 */
M
Matt Mackall 已提交
746

L
Linus Torvalds 已提交
747 748 749 750 751
asmlinkage int printk(const char *fmt, ...)
{
	va_list args;
	int r;

752 753 754 755 756 757 758 759
#ifdef CONFIG_KGDB_KDB
	if (unlikely(kdb_trap_printk)) {
		va_start(args, fmt);
		r = vkdb_printf(fmt, args);
		va_end(args);
		return r;
	}
#endif
L
Linus Torvalds 已提交
760 761 762 763 764 765 766
	va_start(args, fmt);
	r = vprintk(fmt, args);
	va_end(args);

	return r;
}

767 768 769
/* cpu currently holding logbuf_lock */
static volatile unsigned int printk_cpu = UINT_MAX;

770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
/*
 * Can we actually use the console at this time on this cpu?
 *
 * Console drivers may assume that per-cpu resources have
 * been allocated. So unless they're explicitly marked as
 * being able to cope (CON_ANYTIME) don't call them until
 * this CPU is officially up.
 */
static inline int can_use_console(unsigned int cpu)
{
	return cpu_online(cpu) || have_callable_console();
}

/*
 * Try to get console ownership to actually show the kernel
 * messages from a 'printk'. Return true (and with the
786
 * console_lock held, and 'console_locked' set) if it
787 788 789 790 791 792
 * is successful, false otherwise.
 *
 * This gets called with the 'logbuf_lock' spinlock held and
 * interrupts disabled. It should return with 'lockbuf_lock'
 * released but interrupts still disabled.
 */
793
static int console_trylock_for_printk(unsigned int cpu)
N
Namhyung Kim 已提交
794
	__releases(&logbuf_lock)
795
{
796
	int retval = 0, wake = 0;
797

798
	if (console_trylock()) {
799 800 801 802 803 804 805 806 807 808
		retval = 1;

		/*
		 * If we can't use the console, we need to release
		 * the console semaphore by hand to avoid flushing
		 * the buffer. We need to hold the console semaphore
		 * in order to do this test safely.
		 */
		if (!can_use_console(cpu)) {
			console_locked = 0;
809
			wake = 1;
810 811 812
			retval = 0;
		}
	}
813
	printk_cpu = UINT_MAX;
814 815
	if (wake)
		up(&console_sem);
816
	raw_spin_unlock(&logbuf_lock);
817 818
	return retval;
}
819 820 821
static const char recursion_bug_msg [] =
		KERN_CRIT "BUG: recent printk recursion!\n";
static int recursion_bug;
822
static int new_text_line = 1;
823
static char printk_buf[1024];
824

825 826 827 828 829 830 831 832 833 834 835 836 837 838
int printk_delay_msec __read_mostly;

static inline void printk_delay(void)
{
	if (unlikely(printk_delay_msec)) {
		int m = printk_delay_msec;

		while (m--) {
			mdelay(1);
			touch_nmi_watchdog();
		}
	}
}

L
Linus Torvalds 已提交
839 840
asmlinkage int vprintk(const char *fmt, va_list args)
{
841
	int printed_len = 0;
842
	int current_log_level = default_message_loglevel;
843
	unsigned long flags;
844
	int this_cpu;
L
Linus Torvalds 已提交
845
	char *p;
846 847
	size_t plen;
	char special;
L
Linus Torvalds 已提交
848

R
Randy Dunlap 已提交
849
	boot_delay_msec();
850
	printk_delay();
R
Randy Dunlap 已提交
851

L
Linus Torvalds 已提交
852
	/* This stops the holder of console_sem just where we want him */
853
	local_irq_save(flags);
854 855 856 857 858 859 860 861 862 863 864 865 866
	this_cpu = smp_processor_id();

	/*
	 * Ouch, printk recursed into itself!
	 */
	if (unlikely(printk_cpu == this_cpu)) {
		/*
		 * If a crash is occurring during printk() on this CPU,
		 * then try to get the crash message out but make sure
		 * we can't deadlock. Otherwise just return to avoid the
		 * recursion and return - but flag the recursion so that
		 * it can be printed at the next appropriate moment:
		 */
867
		if (!oops_in_progress && !lockdep_recursing(current)) {
868
			recursion_bug = 1;
869 870 871 872 873
			goto out_restore_irqs;
		}
		zap_locks();
	}

874
	lockdep_off();
875
	raw_spin_lock(&logbuf_lock);
876
	printk_cpu = this_cpu;
L
Linus Torvalds 已提交
877

878 879 880
	if (recursion_bug) {
		recursion_bug = 0;
		strcpy(printk_buf, recursion_bug_msg);
881
		printed_len = strlen(recursion_bug_msg);
882
	}
L
Linus Torvalds 已提交
883
	/* Emit the output into the temporary buffer */
884
	printed_len += vscnprintf(printk_buf + printed_len,
T
Tejun Heo 已提交
885
				  sizeof(printk_buf) - printed_len, fmt, args);
L
Linus Torvalds 已提交
886

887 888
	p = printk_buf;

889 890 891 892 893 894 895 896 897 898 899 900 901 902 903
	/* Read log level and handle special printk prefix */
	plen = log_prefix(p, &current_log_level, &special);
	if (plen) {
		p += plen;

		switch (special) {
		case 'c': /* Strip <c> KERN_CONT, continue line */
			plen = 0;
			break;
		case 'd': /* Strip <d> KERN_DEFAULT, start new line */
			plen = 0;
		default:
			if (!new_text_line) {
				emit_log_char('\n');
				new_text_line = 1;
904 905 906 907
			}
		}
	}

L
Linus Torvalds 已提交
908
	/*
909 910
	 * Copy the output into log_buf. If the caller didn't provide
	 * the appropriate log prefix, we insert them here
L
Linus Torvalds 已提交
911
	 */
912
	for (; *p; p++) {
913 914 915
		if (new_text_line) {
			new_text_line = 0;

916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
			if (plen) {
				/* Copy original log prefix */
				int i;

				for (i = 0; i < plen; i++)
					emit_log_char(printk_buf[i]);
				printed_len += plen;
			} else {
				/* Add log prefix */
				emit_log_char('<');
				emit_log_char(current_log_level + '0');
				emit_log_char('>');
				printed_len += 3;
			}

L
Linus Torvalds 已提交
931
			if (printk_time) {
932
				/* Add the current time stamp */
L
Linus Torvalds 已提交
933 934 935 936 937
				char tbuf[50], *tp;
				unsigned tlen;
				unsigned long long t;
				unsigned long nanosec_rem;

938
				t = cpu_clock(printk_cpu);
L
Linus Torvalds 已提交
939
				nanosec_rem = do_div(t, 1000000000);
940 941 942
				tlen = sprintf(tbuf, "[%5lu.%06lu] ",
						(unsigned long) t,
						nanosec_rem / 1000);
L
Linus Torvalds 已提交
943 944 945

				for (tp = tbuf; tp < tbuf + tlen; tp++)
					emit_log_char(*tp);
946
				printed_len += tlen;
L
Linus Torvalds 已提交
947
			}
948

L
Linus Torvalds 已提交
949 950 951
			if (!*p)
				break;
		}
952

L
Linus Torvalds 已提交
953 954
		emit_log_char(*p);
		if (*p == '\n')
955
			new_text_line = 1;
L
Linus Torvalds 已提交
956 957
	}

958 959 960 961 962 963
	/*
	 * Try to acquire and then immediately release the
	 * console semaphore. The release will do all the
	 * actual magic (print out buffers, wake up klogd,
	 * etc). 
	 *
964
	 * The console_trylock_for_printk() function
965 966 967
	 * will release 'logbuf_lock' regardless of whether it
	 * actually gets the semaphore or not.
	 */
968 969
	if (console_trylock_for_printk(this_cpu))
		console_unlock();
970

971
	lockdep_on();
972
out_restore_irqs:
973
	local_irq_restore(flags);
974

L
Linus Torvalds 已提交
975 976 977 978 979
	return printed_len;
}
EXPORT_SYMBOL(printk);
EXPORT_SYMBOL(vprintk);

M
Matt Mackall 已提交
980 981
#else

982
static void call_console_drivers(unsigned start, unsigned end)
J
Jesper Juhl 已提交
983 984
{
}
M
Matt Mackall 已提交
985 986 987

#endif

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 1015 1016 1017
static int __add_preferred_console(char *name, int idx, char *options,
				   char *brl_options)
{
	struct console_cmdline *c;
	int i;

	/*
	 *	See if this tty is not yet registered, and
	 *	if we have a slot free.
	 */
	for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++)
		if (strcmp(console_cmdline[i].name, name) == 0 &&
			  console_cmdline[i].index == idx) {
				if (!brl_options)
					selected_console = i;
				return 0;
		}
	if (i == MAX_CMDLINECONSOLES)
		return -E2BIG;
	if (!brl_options)
		selected_console = i;
	c = &console_cmdline[i];
	strlcpy(c->name, name, sizeof(c->name));
	c->options = options;
#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
	c->brl_options = brl_options;
#endif
	c->index = idx;
	return 0;
}
1018 1019 1020 1021 1022
/*
 * Set up a list of consoles.  Called from init/main.c
 */
static int __init console_setup(char *str)
{
Y
Yinghai Lu 已提交
1023
	char buf[sizeof(console_cmdline[0].name) + 4]; /* 4 for index */
1024
	char *s, *options, *brl_options = NULL;
1025 1026
	int idx;

1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
	if (!memcmp(str, "brl,", 4)) {
		brl_options = "";
		str += 4;
	} else if (!memcmp(str, "brl=", 4)) {
		brl_options = str + 4;
		str = strchr(brl_options, ',');
		if (!str) {
			printk(KERN_ERR "need port name after brl=\n");
			return 1;
		}
		*(str++) = 0;
	}
#endif

1042 1043 1044 1045
	/*
	 * Decode str into name, index, options.
	 */
	if (str[0] >= '0' && str[0] <= '9') {
Y
Yinghai Lu 已提交
1046 1047
		strcpy(buf, "ttyS");
		strncpy(buf + 4, str, sizeof(buf) - 5);
1048
	} else {
Y
Yinghai Lu 已提交
1049
		strncpy(buf, str, sizeof(buf) - 1);
1050
	}
Y
Yinghai Lu 已提交
1051
	buf[sizeof(buf) - 1] = 0;
1052 1053 1054 1055
	if ((options = strchr(str, ',')) != NULL)
		*(options++) = 0;
#ifdef __sparc__
	if (!strcmp(str, "ttya"))
Y
Yinghai Lu 已提交
1056
		strcpy(buf, "ttyS0");
1057
	if (!strcmp(str, "ttyb"))
Y
Yinghai Lu 已提交
1058
		strcpy(buf, "ttyS1");
1059
#endif
Y
Yinghai Lu 已提交
1060
	for (s = buf; *s; s++)
1061 1062 1063 1064 1065
		if ((*s >= '0' && *s <= '9') || *s == ',')
			break;
	idx = simple_strtoul(s, NULL, 10);
	*s = 0;

1066
	__add_preferred_console(buf, idx, options, brl_options);
1067
	console_set_on_cmdline = 1;
1068 1069 1070 1071
	return 1;
}
__setup("console=", console_setup);

1072 1073
/**
 * add_preferred_console - add a device to the list of preferred consoles.
1074 1075 1076
 * @name: device name
 * @idx: device index
 * @options: options for this console
1077 1078 1079 1080 1081 1082 1083 1084
 *
 * The last preferred console added will be used for kernel messages
 * and stdin/out/err for init.  Normally this is used by console_setup
 * above to handle user-supplied console arguments; however it can also
 * be used by arch-specific code either to override the user or more
 * commonly to provide a default console (ie from PROM variables) when
 * the user has not supplied one.
 */
1085
int add_preferred_console(char *name, int idx, char *options)
1086
{
1087
	return __add_preferred_console(name, idx, options, NULL);
1088 1089
}

1090
int update_console_cmdline(char *name, int idx, char *name_new, int idx_new, char *options)
1091 1092 1093 1094 1095 1096 1097 1098
{
	struct console_cmdline *c;
	int i;

	for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++)
		if (strcmp(console_cmdline[i].name, name) == 0 &&
			  console_cmdline[i].index == idx) {
				c = &console_cmdline[i];
1099
				strlcpy(c->name, name_new, sizeof(c->name));
1100 1101 1102 1103 1104 1105 1106 1107 1108
				c->name[sizeof(c->name) - 1] = 0;
				c->options = options;
				c->index = idx_new;
				return i;
		}
	/* not found */
	return -1;
}

1109
bool console_suspend_enabled = 1;
1110 1111 1112 1113 1114 1115 1116 1117
EXPORT_SYMBOL(console_suspend_enabled);

static int __init console_suspend_disable(char *str)
{
	console_suspend_enabled = 0;
	return 1;
}
__setup("no_console_suspend", console_suspend_disable);
1118 1119 1120 1121
module_param_named(console_suspend, console_suspend_enabled,
		bool, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(console_suspend, "suspend console during suspend"
	" and hibernate operations");
1122

1123 1124 1125 1126 1127 1128 1129
/**
 * suspend_console - suspend the console subsystem
 *
 * This disables printk() while we go into suspend states
 */
void suspend_console(void)
{
1130 1131
	if (!console_suspend_enabled)
		return;
1132
	printk("Suspending console(s) (use no_console_suspend to debug)\n");
1133
	console_lock();
1134
	console_suspended = 1;
1135
	up(&console_sem);
1136 1137 1138 1139
}

void resume_console(void)
{
1140 1141
	if (!console_suspend_enabled)
		return;
1142
	down(&console_sem);
1143
	console_suspended = 0;
1144
	console_unlock();
1145 1146
}

1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166
/**
 * console_cpu_notify - print deferred console messages after CPU hotplug
 * @self: notifier struct
 * @action: CPU hotplug event
 * @hcpu: unused
 *
 * If printk() is called from a CPU that is not online yet, the messages
 * will be spooled but will not show up on the console.  This function is
 * called when a new CPU comes online (or fails to come up), and ensures
 * that any such output gets printed.
 */
static int __cpuinit console_cpu_notify(struct notifier_block *self,
	unsigned long action, void *hcpu)
{
	switch (action) {
	case CPU_ONLINE:
	case CPU_DEAD:
	case CPU_DYING:
	case CPU_DOWN_FAILED:
	case CPU_UP_CANCELED:
1167 1168
		console_lock();
		console_unlock();
1169 1170 1171 1172
	}
	return NOTIFY_OK;
}

L
Linus Torvalds 已提交
1173
/**
1174
 * console_lock - lock the console system for exclusive use.
L
Linus Torvalds 已提交
1175
 *
1176
 * Acquires a lock which guarantees that the caller has
L
Linus Torvalds 已提交
1177 1178 1179 1180
 * exclusive access to the console system and the console_drivers list.
 *
 * Can sleep, returns nothing.
 */
1181
void console_lock(void)
L
Linus Torvalds 已提交
1182
{
1183
	BUG_ON(in_interrupt());
L
Linus Torvalds 已提交
1184
	down(&console_sem);
1185 1186
	if (console_suspended)
		return;
L
Linus Torvalds 已提交
1187 1188 1189
	console_locked = 1;
	console_may_schedule = 1;
}
1190
EXPORT_SYMBOL(console_lock);
L
Linus Torvalds 已提交
1191

1192 1193 1194 1195 1196 1197 1198 1199 1200
/**
 * console_trylock - try to lock the console system for exclusive use.
 *
 * Tried to acquire a lock which guarantees that the caller has
 * exclusive access to the console system and the console_drivers list.
 *
 * returns 1 on success, and 0 on failure to acquire the lock.
 */
int console_trylock(void)
L
Linus Torvalds 已提交
1201 1202
{
	if (down_trylock(&console_sem))
1203
		return 0;
1204 1205
	if (console_suspended) {
		up(&console_sem);
1206
		return 0;
1207
	}
L
Linus Torvalds 已提交
1208 1209
	console_locked = 1;
	console_may_schedule = 0;
1210
	return 1;
L
Linus Torvalds 已提交
1211
}
1212
EXPORT_SYMBOL(console_trylock);
L
Linus Torvalds 已提交
1213 1214 1215 1216 1217 1218

int is_console_locked(void)
{
	return console_locked;
}

1219 1220 1221 1222 1223 1224 1225 1226
/*
 * Delayed printk facility, for scheduler-internal messages:
 */
#define PRINTK_BUF_SIZE		512

#define PRINTK_PENDING_WAKEUP	0x01
#define PRINTK_PENDING_SCHED	0x02

P
Peter Zijlstra 已提交
1227
static DEFINE_PER_CPU(int, printk_pending);
1228
static DEFINE_PER_CPU(char [PRINTK_BUF_SIZE], printk_sched_buf);
P
Peter Zijlstra 已提交
1229 1230

void printk_tick(void)
1231
{
1232
	if (__this_cpu_read(printk_pending)) {
1233 1234 1235 1236 1237 1238 1239
		int pending = __this_cpu_xchg(printk_pending, 0);
		if (pending & PRINTK_PENDING_SCHED) {
			char *buf = __get_cpu_var(printk_sched_buf);
			printk(KERN_WARNING "[sched_delayed] %s", buf);
		}
		if (pending & PRINTK_PENDING_WAKEUP)
			wake_up_interruptible(&log_wait);
P
Peter Zijlstra 已提交
1240 1241 1242 1243 1244
	}
}

int printk_needs_cpu(int cpu)
{
1245
	if (cpu_is_offline(cpu))
1246
		printk_tick();
1247
	return __this_cpu_read(printk_pending);
P
Peter Zijlstra 已提交
1248 1249 1250 1251 1252
}

void wake_up_klogd(void)
{
	if (waitqueue_active(&log_wait))
1253
		this_cpu_or(printk_pending, PRINTK_PENDING_WAKEUP);
1254 1255
}

L
Linus Torvalds 已提交
1256
/**
1257
 * console_unlock - unlock the console system
L
Linus Torvalds 已提交
1258
 *
1259
 * Releases the console_lock which the caller holds on the console system
L
Linus Torvalds 已提交
1260 1261
 * and the console driver list.
 *
1262 1263 1264
 * While the console_lock was held, console output may have been buffered
 * by printk().  If this is the case, console_unlock(); emits
 * the output prior to releasing the lock.
L
Linus Torvalds 已提交
1265 1266 1267
 *
 * If there is output waiting for klogd, we wake it up.
 *
1268
 * console_unlock(); may be called from any context.
L
Linus Torvalds 已提交
1269
 */
1270
void console_unlock(void)
L
Linus Torvalds 已提交
1271 1272
{
	unsigned long flags;
1273
	unsigned _con_start, _log_end;
1274
	unsigned wake_klogd = 0, retry = 0;
L
Linus Torvalds 已提交
1275

1276
	if (console_suspended) {
1277
		up(&console_sem);
1278 1279
		return;
	}
1280 1281 1282

	console_may_schedule = 0;

1283
again:
L
Linus Torvalds 已提交
1284
	for ( ; ; ) {
1285
		raw_spin_lock_irqsave(&logbuf_lock, flags);
L
Linus Torvalds 已提交
1286 1287 1288 1289 1290 1291
		wake_klogd |= log_start - log_end;
		if (con_start == log_end)
			break;			/* Nothing to print */
		_con_start = con_start;
		_log_end = log_end;
		con_start = log_end;		/* Flush */
1292
		raw_spin_unlock(&logbuf_lock);
1293
		stop_critical_timings();	/* don't trace print latency */
L
Linus Torvalds 已提交
1294
		call_console_drivers(_con_start, _log_end);
1295
		start_critical_timings();
L
Linus Torvalds 已提交
1296 1297 1298
		local_irq_restore(flags);
	}
	console_locked = 0;
1299 1300 1301 1302 1303

	/* Release the exclusive_console once it is used */
	if (unlikely(exclusive_console))
		exclusive_console = NULL;

1304
	raw_spin_unlock(&logbuf_lock);
1305

1306
	up(&console_sem);
1307 1308 1309 1310 1311 1312 1313

	/*
	 * Someone could have filled up the buffer again, so re-check if there's
	 * something to flush. In case we cannot trylock the console_sem again,
	 * there's a new owner and the console_unlock() from them will do the
	 * flush, no worries.
	 */
1314
	raw_spin_lock(&logbuf_lock);
1315 1316
	if (con_start != log_end)
		retry = 1;
P
Peter Zijlstra 已提交
1317 1318
	raw_spin_unlock_irqrestore(&logbuf_lock, flags);

1319 1320 1321
	if (retry && console_trylock())
		goto again;

1322 1323
	if (wake_klogd)
		wake_up_klogd();
L
Linus Torvalds 已提交
1324
}
1325
EXPORT_SYMBOL(console_unlock);
L
Linus Torvalds 已提交
1326

1327 1328
/**
 * console_conditional_schedule - yield the CPU if required
L
Linus Torvalds 已提交
1329 1330 1331 1332 1333
 *
 * If the console code is currently allowed to sleep, and
 * if this CPU should yield the CPU to another task, do
 * so here.
 *
1334
 * Must be called within console_lock();.
L
Linus Torvalds 已提交
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354
 */
void __sched console_conditional_schedule(void)
{
	if (console_may_schedule)
		cond_resched();
}
EXPORT_SYMBOL(console_conditional_schedule);

void console_unblank(void)
{
	struct console *c;

	/*
	 * console_unblank can no longer be called in interrupt context unless
	 * oops_in_progress is set to 1..
	 */
	if (oops_in_progress) {
		if (down_trylock(&console_sem) != 0)
			return;
	} else
1355
		console_lock();
L
Linus Torvalds 已提交
1356 1357 1358

	console_locked = 1;
	console_may_schedule = 0;
1359
	for_each_console(c)
L
Linus Torvalds 已提交
1360 1361
		if ((c->flags & CON_ENABLED) && c->unblank)
			c->unblank();
1362
	console_unlock();
L
Linus Torvalds 已提交
1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
}

/*
 * Return the console tty driver structure and its associated index
 */
struct tty_driver *console_device(int *index)
{
	struct console *c;
	struct tty_driver *driver = NULL;

1373
	console_lock();
1374
	for_each_console(c) {
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380
		if (!c->device)
			continue;
		driver = c->device(c, index);
		if (driver)
			break;
	}
1381
	console_unlock();
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391
	return driver;
}

/*
 * Prevent further output on the passed console device so that (for example)
 * serial drivers can disable console output before suspending a port, and can
 * re-enable output afterwards.
 */
void console_stop(struct console *console)
{
1392
	console_lock();
L
Linus Torvalds 已提交
1393
	console->flags &= ~CON_ENABLED;
1394
	console_unlock();
L
Linus Torvalds 已提交
1395 1396 1397 1398 1399
}
EXPORT_SYMBOL(console_stop);

void console_start(struct console *console)
{
1400
	console_lock();
L
Linus Torvalds 已提交
1401
	console->flags |= CON_ENABLED;
1402
	console_unlock();
L
Linus Torvalds 已提交
1403 1404 1405
}
EXPORT_SYMBOL(console_start);

1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
static int __read_mostly keep_bootcon;

static int __init keep_bootcon_setup(char *str)
{
	keep_bootcon = 1;
	printk(KERN_INFO "debug: skip boot console de-registration.\n");

	return 0;
}

early_param("keep_bootcon", keep_bootcon_setup);

L
Linus Torvalds 已提交
1418 1419 1420 1421 1422
/*
 * The console driver calls this routine during kernel initialization
 * to register the console printing procedure with printk() and to
 * print any messages that were printed by the kernel before the
 * console driver was initialized.
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
 *
 * This can happen pretty early during the boot process (because of
 * early_printk) - sometimes before setup_arch() completes - be careful
 * of what kernel features are used - they may not be initialised yet.
 *
 * There are two types of consoles - bootconsoles (early_printk) and
 * "real" consoles (everything which is not a bootconsole) which are
 * handled differently.
 *  - Any number of bootconsoles can be registered at any time.
 *  - As soon as a "real" console is registered, all bootconsoles
 *    will be unregistered automatically.
 *  - Once a "real" console is registered, any attempt to register a
 *    bootconsoles will be rejected
L
Linus Torvalds 已提交
1436
 */
1437
void register_console(struct console *newcon)
L
Linus Torvalds 已提交
1438
{
J
Jesper Juhl 已提交
1439
	int i;
L
Linus Torvalds 已提交
1440
	unsigned long flags;
1441
	struct console *bcon = NULL;
L
Linus Torvalds 已提交
1442

1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
	/*
	 * before we register a new CON_BOOT console, make sure we don't
	 * already have a valid console
	 */
	if (console_drivers && newcon->flags & CON_BOOT) {
		/* find the last or real console */
		for_each_console(bcon) {
			if (!(bcon->flags & CON_BOOT)) {
				printk(KERN_INFO "Too late to register bootconsole %s%d\n",
					newcon->name, newcon->index);
				return;
			}
		}
1456 1457
	}

1458 1459 1460 1461
	if (console_drivers && console_drivers->flags & CON_BOOT)
		bcon = console_drivers;

	if (preferred_console < 0 || bcon || !console_drivers)
L
Linus Torvalds 已提交
1462 1463
		preferred_console = selected_console;

1464 1465
	if (newcon->early_setup)
		newcon->early_setup();
1466

L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472
	/*
	 *	See if we want to use this console driver. If we
	 *	didn't select a console we take the first one
	 *	that registers here.
	 */
	if (preferred_console < 0) {
1473 1474 1475 1476 1477 1478 1479
		if (newcon->index < 0)
			newcon->index = 0;
		if (newcon->setup == NULL ||
		    newcon->setup(newcon, NULL) == 0) {
			newcon->flags |= CON_ENABLED;
			if (newcon->device) {
				newcon->flags |= CON_CONSDEV;
1480 1481
				preferred_console = 0;
			}
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488
		}
	}

	/*
	 *	See if this console matches one we selected on
	 *	the command line.
	 */
J
Jesper Juhl 已提交
1489 1490
	for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0];
			i++) {
1491
		if (strcmp(console_cmdline[i].name, newcon->name) != 0)
L
Linus Torvalds 已提交
1492
			continue;
1493 1494
		if (newcon->index >= 0 &&
		    newcon->index != console_cmdline[i].index)
L
Linus Torvalds 已提交
1495
			continue;
1496 1497
		if (newcon->index < 0)
			newcon->index = console_cmdline[i].index;
1498 1499
#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
		if (console_cmdline[i].brl_options) {
1500 1501
			newcon->flags |= CON_BRL;
			braille_register_console(newcon,
1502 1503 1504 1505 1506 1507
					console_cmdline[i].index,
					console_cmdline[i].options,
					console_cmdline[i].brl_options);
			return;
		}
#endif
1508 1509
		if (newcon->setup &&
		    newcon->setup(newcon, console_cmdline[i].options) != 0)
L
Linus Torvalds 已提交
1510
			break;
1511 1512
		newcon->flags |= CON_ENABLED;
		newcon->index = console_cmdline[i].index;
1513
		if (i == selected_console) {
1514
			newcon->flags |= CON_CONSDEV;
1515 1516
			preferred_console = selected_console;
		}
L
Linus Torvalds 已提交
1517 1518 1519
		break;
	}

1520
	if (!(newcon->flags & CON_ENABLED))
L
Linus Torvalds 已提交
1521 1522
		return;

1523 1524 1525 1526 1527 1528 1529
	/*
	 * If we have a bootconsole, and are switching to a real console,
	 * don't print everything out again, since when the boot console, and
	 * the real console are the same physical device, it's annoying to
	 * see the beginning boot messages twice
	 */
	if (bcon && ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV))
1530
		newcon->flags &= ~CON_PRINTBUFFER;
L
Linus Torvalds 已提交
1531 1532 1533 1534 1535

	/*
	 *	Put this console in the list - keep the
	 *	preferred driver at the head of the list.
	 */
1536
	console_lock();
1537 1538 1539 1540 1541
	if ((newcon->flags & CON_CONSDEV) || console_drivers == NULL) {
		newcon->next = console_drivers;
		console_drivers = newcon;
		if (newcon->next)
			newcon->next->flags &= ~CON_CONSDEV;
L
Linus Torvalds 已提交
1542
	} else {
1543 1544
		newcon->next = console_drivers->next;
		console_drivers->next = newcon;
L
Linus Torvalds 已提交
1545
	}
1546
	if (newcon->flags & CON_PRINTBUFFER) {
L
Linus Torvalds 已提交
1547
		/*
1548
		 * console_unlock(); will print out the buffered messages
L
Linus Torvalds 已提交
1549 1550
		 * for us.
		 */
1551
		raw_spin_lock_irqsave(&logbuf_lock, flags);
L
Linus Torvalds 已提交
1552
		con_start = log_start;
1553
		raw_spin_unlock_irqrestore(&logbuf_lock, flags);
1554 1555 1556 1557 1558 1559
		/*
		 * We're about to replay the log buffer.  Only do this to the
		 * just-registered console to avoid excessive message spam to
		 * the already-registered consoles.
		 */
		exclusive_console = newcon;
L
Linus Torvalds 已提交
1560
	}
1561
	console_unlock();
1562
	console_sysfs_notify();
1563 1564 1565 1566 1567 1568 1569 1570

	/*
	 * By unregistering the bootconsoles after we enable the real console
	 * we get the "console xxx enabled" message on all the consoles -
	 * boot consoles, real consoles, etc - this is to ensure that end
	 * users know there might be something in the kernel's log buffer that
	 * went to the bootconsole (that they do not see on the real console)
	 */
1571 1572 1573
	if (bcon &&
	    ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV) &&
	    !keep_bootcon) {
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
		/* we need to iterate through twice, to make sure we print
		 * everything out, before we unregister the console(s)
		 */
		printk(KERN_INFO "console [%s%d] enabled, bootconsole disabled\n",
			newcon->name, newcon->index);
		for_each_console(bcon)
			if (bcon->flags & CON_BOOT)
				unregister_console(bcon);
	} else {
		printk(KERN_INFO "%sconsole [%s%d] enabled\n",
			(newcon->flags & CON_BOOT) ? "boot" : "" ,
			newcon->name, newcon->index);
	}
L
Linus Torvalds 已提交
1587 1588 1589
}
EXPORT_SYMBOL(register_console);

J
Jesper Juhl 已提交
1590
int unregister_console(struct console *console)
L
Linus Torvalds 已提交
1591
{
J
Jesper Juhl 已提交
1592
        struct console *a, *b;
L
Linus Torvalds 已提交
1593 1594
	int res = 1;

1595 1596 1597 1598 1599
#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
	if (console->flags & CON_BRL)
		return braille_unregister_console(console);
#endif

1600
	console_lock();
L
Linus Torvalds 已提交
1601 1602 1603
	if (console_drivers == console) {
		console_drivers=console->next;
		res = 0;
1604
	} else if (console_drivers) {
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610
		for (a=console_drivers->next, b=console_drivers ;
		     a; b=a, a=b->next) {
			if (a == console) {
				b->next = a->next;
				res = 0;
				break;
J
Jesper Juhl 已提交
1611
			}
L
Linus Torvalds 已提交
1612 1613
		}
	}
J
Jesper Juhl 已提交
1614

1615
	/*
1616 1617
	 * If this isn't the last console and it has CON_CONSDEV set, we
	 * need to set it on the next preferred console.
L
Linus Torvalds 已提交
1618
	 */
1619
	if (console_drivers != NULL && console->flags & CON_CONSDEV)
1620
		console_drivers->flags |= CON_CONSDEV;
L
Linus Torvalds 已提交
1621

1622
	console_unlock();
1623
	console_sysfs_notify();
L
Linus Torvalds 已提交
1624 1625 1626
	return res;
}
EXPORT_SYMBOL(unregister_console);
M
Matt Mackall 已提交
1627

1628
static int __init printk_late_init(void)
1629
{
1630 1631 1632
	struct console *con;

	for_each_console(con) {
1633
		if (!keep_bootcon && con->flags & CON_BOOT) {
1634
			printk(KERN_INFO "turn off boot console %s%d\n",
1635
				con->name, con->index);
1636
			unregister_console(con);
1637
		}
1638
	}
1639
	hotcpu_notifier(console_cpu_notify, 0);
1640 1641
	return 0;
}
1642
late_initcall(printk_late_init);
1643

1644
#if defined CONFIG_PRINTK
D
Dave Young 已提交
1645

1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665
int printk_sched(const char *fmt, ...)
{
	unsigned long flags;
	va_list args;
	char *buf;
	int r;

	local_irq_save(flags);
	buf = __get_cpu_var(printk_sched_buf);

	va_start(args, fmt);
	r = vsnprintf(buf, PRINTK_BUF_SIZE, fmt, args);
	va_end(args);

	__this_cpu_or(printk_pending, PRINTK_PENDING_SCHED);
	local_irq_restore(flags);

	return r;
}

L
Linus Torvalds 已提交
1666 1667 1668
/*
 * printk rate limiting, lifted from the networking subsystem.
 *
1669 1670
 * This enforces a rate limit: not more than 10 kernel messages
 * every 5s to make a denial-of-service attack impossible.
L
Linus Torvalds 已提交
1671
 */
1672 1673
DEFINE_RATELIMIT_STATE(printk_ratelimit_state, 5 * HZ, 10);

1674
int __printk_ratelimit(const char *func)
L
Linus Torvalds 已提交
1675
{
1676
	return ___ratelimit(&printk_ratelimit_state, func);
L
Linus Torvalds 已提交
1677
}
1678
EXPORT_SYMBOL(__printk_ratelimit);
1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691

/**
 * printk_timed_ratelimit - caller-controlled printk ratelimiting
 * @caller_jiffies: pointer to caller's state
 * @interval_msecs: minimum interval between prints
 *
 * printk_timed_ratelimit() returns true if more than @interval_msecs
 * milliseconds have elapsed since the last time printk_timed_ratelimit()
 * returned true.
 */
bool printk_timed_ratelimit(unsigned long *caller_jiffies,
			unsigned int interval_msecs)
{
1692 1693 1694 1695 1696
	if (*caller_jiffies == 0
			|| !time_in_range(jiffies, *caller_jiffies,
					*caller_jiffies
					+ msecs_to_jiffies(interval_msecs))) {
		*caller_jiffies = jiffies;
1697 1698 1699 1700 1701
		return true;
	}
	return false;
}
EXPORT_SYMBOL(printk_timed_ratelimit);
1702 1703 1704 1705 1706 1707

static DEFINE_SPINLOCK(dump_list_lock);
static LIST_HEAD(dump_list);

/**
 * kmsg_dump_register - register a kernel log dumper.
R
Randy Dunlap 已提交
1708
 * @dumper: pointer to the kmsg_dumper structure
1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
 *
 * Adds a kernel log dumper to the system. The dump callback in the
 * structure will be called when the kernel oopses or panics and must be
 * set. Returns zero on success and %-EINVAL or %-EBUSY otherwise.
 */
int kmsg_dump_register(struct kmsg_dumper *dumper)
{
	unsigned long flags;
	int err = -EBUSY;

	/* The dump callback needs to be set */
	if (!dumper->dump)
		return -EINVAL;

	spin_lock_irqsave(&dump_list_lock, flags);
	/* Don't allow registering multiple times */
	if (!dumper->registered) {
		dumper->registered = 1;
1727
		list_add_tail_rcu(&dumper->list, &dump_list);
1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
		err = 0;
	}
	spin_unlock_irqrestore(&dump_list_lock, flags);

	return err;
}
EXPORT_SYMBOL_GPL(kmsg_dump_register);

/**
 * kmsg_dump_unregister - unregister a kmsg dumper.
R
Randy Dunlap 已提交
1738
 * @dumper: pointer to the kmsg_dumper structure
1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750
 *
 * Removes a dump device from the system. Returns zero on success and
 * %-EINVAL otherwise.
 */
int kmsg_dump_unregister(struct kmsg_dumper *dumper)
{
	unsigned long flags;
	int err = -EINVAL;

	spin_lock_irqsave(&dump_list_lock, flags);
	if (dumper->registered) {
		dumper->registered = 0;
1751
		list_del_rcu(&dumper->list);
1752 1753 1754
		err = 0;
	}
	spin_unlock_irqrestore(&dump_list_lock, flags);
1755
	synchronize_rcu();
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776

	return err;
}
EXPORT_SYMBOL_GPL(kmsg_dump_unregister);

/**
 * kmsg_dump - dump kernel log to kernel message dumpers.
 * @reason: the reason (oops, panic etc) for dumping
 *
 * Iterate through each of the dump devices and call the oops/panic
 * callbacks with the log buffer.
 */
void kmsg_dump(enum kmsg_dump_reason reason)
{
	unsigned long end;
	unsigned chars;
	struct kmsg_dumper *dumper;
	const char *s1, *s2;
	unsigned long l1, l2;
	unsigned long flags;

1777 1778 1779
	if ((reason > KMSG_DUMP_OOPS) && !always_kmsg_dump)
		return;

1780 1781 1782
	/* Theoretically, the log could move on after we do this, but
	   there's not a lot we can do about that. The new messages
	   will overwrite the start of what we dump. */
1783
	raw_spin_lock_irqsave(&logbuf_lock, flags);
1784 1785
	end = log_end & LOG_BUF_MASK;
	chars = logged_chars;
1786
	raw_spin_unlock_irqrestore(&logbuf_lock, flags);
1787

1788 1789 1790
	if (chars > end) {
		s1 = log_buf + log_buf_len - chars + end;
		l1 = chars - end;
1791 1792 1793 1794 1795 1796 1797

		s2 = log_buf;
		l2 = end;
	} else {
		s1 = "";
		l1 = 0;

1798 1799
		s2 = log_buf + end - chars;
		l2 = chars;
1800 1801
	}

1802 1803
	rcu_read_lock();
	list_for_each_entry_rcu(dumper, &dump_list, list)
1804
		dumper->dump(dumper, reason, s1, l1, s2, l2);
1805
	rcu_read_unlock();
1806
}
1807
#endif