xref: /linux-6.15/include/linux/kernel.h (revision eef8abda)
1 /* SPDX-License-Identifier: GPL-2.0 */
2 #ifndef _LINUX_KERNEL_H
3 #define _LINUX_KERNEL_H
4 
5 #include <stdarg.h>
6 #include <linux/limits.h>
7 #include <linux/linkage.h>
8 #include <linux/stddef.h>
9 #include <linux/types.h>
10 #include <linux/compiler.h>
11 #include <linux/bitops.h>
12 #include <linux/log2.h>
13 #include <linux/math.h>
14 #include <linux/minmax.h>
15 #include <linux/typecheck.h>
16 #include <linux/printk.h>
17 #include <linux/build_bug.h>
18 #include <linux/static_call_types.h>
19 #include <asm/byteorder.h>
20 
21 #include <uapi/linux/kernel.h>
22 
23 #define STACK_MAGIC	0xdeadbeef
24 
25 /**
26  * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value
27  * @x: value to repeat
28  *
29  * NOTE: @x is not checked for > 0xff; larger values produce odd results.
30  */
31 #define REPEAT_BYTE(x)	((~0ul / 0xff) * (x))
32 
33 /* @a is a power of 2 value */
34 #define ALIGN(x, a)		__ALIGN_KERNEL((x), (a))
35 #define ALIGN_DOWN(x, a)	__ALIGN_KERNEL((x) - ((a) - 1), (a))
36 #define __ALIGN_MASK(x, mask)	__ALIGN_KERNEL_MASK((x), (mask))
37 #define PTR_ALIGN(p, a)		((typeof(p))ALIGN((unsigned long)(p), (a)))
38 #define PTR_ALIGN_DOWN(p, a)	((typeof(p))ALIGN_DOWN((unsigned long)(p), (a)))
39 #define IS_ALIGNED(x, a)		(((x) & ((typeof(x))(a) - 1)) == 0)
40 
41 /* generic data direction definitions */
42 #define READ			0
43 #define WRITE			1
44 
45 /**
46  * ARRAY_SIZE - get the number of elements in array @arr
47  * @arr: array to be sized
48  */
49 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr))
50 
51 #define u64_to_user_ptr(x) (		\
52 {					\
53 	typecheck(u64, (x));		\
54 	(void __user *)(uintptr_t)(x);	\
55 }					\
56 )
57 
58 #define typeof_member(T, m)	typeof(((T*)0)->m)
59 
60 #define _RET_IP_		(unsigned long)__builtin_return_address(0)
61 #define _THIS_IP_  ({ __label__ __here; __here: (unsigned long)&&__here; })
62 
63 /**
64  * upper_32_bits - return bits 32-63 of a number
65  * @n: the number we're accessing
66  *
67  * A basic shift-right of a 64- or 32-bit quantity.  Use this to suppress
68  * the "right shift count >= width of type" warning when that quantity is
69  * 32-bits.
70  */
71 #define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
72 
73 /**
74  * lower_32_bits - return bits 0-31 of a number
75  * @n: the number we're accessing
76  */
77 #define lower_32_bits(n) ((u32)((n) & 0xffffffff))
78 
79 struct completion;
80 struct pt_regs;
81 struct user;
82 
83 #ifdef CONFIG_PREEMPT_VOLUNTARY
84 
85 extern int __cond_resched(void);
86 # define might_resched() __cond_resched()
87 
88 #elif defined(CONFIG_PREEMPT_DYNAMIC)
89 
90 extern int __cond_resched(void);
91 
92 DECLARE_STATIC_CALL(might_resched, __cond_resched);
93 
94 static __always_inline void might_resched(void)
95 {
96 	static_call_mod(might_resched)();
97 }
98 
99 #else
100 
101 # define might_resched() do { } while (0)
102 
103 #endif /* CONFIG_PREEMPT_* */
104 
105 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
106 extern void ___might_sleep(const char *file, int line, int preempt_offset);
107 extern void __might_sleep(const char *file, int line, int preempt_offset);
108 extern void __cant_sleep(const char *file, int line, int preempt_offset);
109 extern void __cant_migrate(const char *file, int line);
110 
111 /**
112  * might_sleep - annotation for functions that can sleep
113  *
114  * this macro will print a stack trace if it is executed in an atomic
115  * context (spinlock, irq-handler, ...). Additional sections where blocking is
116  * not allowed can be annotated with non_block_start() and non_block_end()
117  * pairs.
118  *
119  * This is a useful debugging help to be able to catch problems early and not
120  * be bitten later when the calling function happens to sleep when it is not
121  * supposed to.
122  */
123 # define might_sleep() \
124 	do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0)
125 /**
126  * cant_sleep - annotation for functions that cannot sleep
127  *
128  * this macro will print a stack trace if it is executed with preemption enabled
129  */
130 # define cant_sleep() \
131 	do { __cant_sleep(__FILE__, __LINE__, 0); } while (0)
132 # define sched_annotate_sleep()	(current->task_state_change = 0)
133 
134 /**
135  * cant_migrate - annotation for functions that cannot migrate
136  *
137  * Will print a stack trace if executed in code which is migratable
138  */
139 # define cant_migrate()							\
140 	do {								\
141 		if (IS_ENABLED(CONFIG_SMP))				\
142 			__cant_migrate(__FILE__, __LINE__);		\
143 	} while (0)
144 
145 /**
146  * non_block_start - annotate the start of section where sleeping is prohibited
147  *
148  * This is on behalf of the oom reaper, specifically when it is calling the mmu
149  * notifiers. The problem is that if the notifier were to block on, for example,
150  * mutex_lock() and if the process which holds that mutex were to perform a
151  * sleeping memory allocation, the oom reaper is now blocked on completion of
152  * that memory allocation. Other blocking calls like wait_event() pose similar
153  * issues.
154  */
155 # define non_block_start() (current->non_block_count++)
156 /**
157  * non_block_end - annotate the end of section where sleeping is prohibited
158  *
159  * Closes a section opened by non_block_start().
160  */
161 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
162 #else
163   static inline void ___might_sleep(const char *file, int line,
164 				   int preempt_offset) { }
165   static inline void __might_sleep(const char *file, int line,
166 				   int preempt_offset) { }
167 # define might_sleep() do { might_resched(); } while (0)
168 # define cant_sleep() do { } while (0)
169 # define cant_migrate()		do { } while (0)
170 # define sched_annotate_sleep() do { } while (0)
171 # define non_block_start() do { } while (0)
172 # define non_block_end() do { } while (0)
173 #endif
174 
175 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0)
176 
177 #if defined(CONFIG_MMU) && \
178 	(defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP))
179 #define might_fault() __might_fault(__FILE__, __LINE__)
180 void __might_fault(const char *file, int line);
181 #else
182 static inline void might_fault(void) { }
183 #endif
184 
185 extern struct atomic_notifier_head panic_notifier_list;
186 extern long (*panic_blink)(int state);
187 __printf(1, 2)
188 void panic(const char *fmt, ...) __noreturn __cold;
189 void nmi_panic(struct pt_regs *regs, const char *msg);
190 extern void oops_enter(void);
191 extern void oops_exit(void);
192 extern bool oops_may_print(void);
193 void do_exit(long error_code) __noreturn;
194 void complete_and_exit(struct completion *, long) __noreturn;
195 
196 /* Internal, do not use. */
197 int __must_check _kstrtoul(const char *s, unsigned int base, unsigned long *res);
198 int __must_check _kstrtol(const char *s, unsigned int base, long *res);
199 
200 int __must_check kstrtoull(const char *s, unsigned int base, unsigned long long *res);
201 int __must_check kstrtoll(const char *s, unsigned int base, long long *res);
202 
203 /**
204  * kstrtoul - convert a string to an unsigned long
205  * @s: The start of the string. The string must be null-terminated, and may also
206  *  include a single newline before its terminating null. The first character
207  *  may also be a plus sign, but not a minus sign.
208  * @base: The number base to use. The maximum supported base is 16. If base is
209  *  given as 0, then the base of the string is automatically detected with the
210  *  conventional semantics - If it begins with 0x the number will be parsed as a
211  *  hexadecimal (case insensitive), if it otherwise begins with 0, it will be
212  *  parsed as an octal number. Otherwise it will be parsed as a decimal.
213  * @res: Where to write the result of the conversion on success.
214  *
215  * Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error.
216  * Preferred over simple_strtoul(). Return code must be checked.
217 */
218 static inline int __must_check kstrtoul(const char *s, unsigned int base, unsigned long *res)
219 {
220 	/*
221 	 * We want to shortcut function call, but
222 	 * __builtin_types_compatible_p(unsigned long, unsigned long long) = 0.
223 	 */
224 	if (sizeof(unsigned long) == sizeof(unsigned long long) &&
225 	    __alignof__(unsigned long) == __alignof__(unsigned long long))
226 		return kstrtoull(s, base, (unsigned long long *)res);
227 	else
228 		return _kstrtoul(s, base, res);
229 }
230 
231 /**
232  * kstrtol - convert a string to a long
233  * @s: The start of the string. The string must be null-terminated, and may also
234  *  include a single newline before its terminating null. The first character
235  *  may also be a plus sign or a minus sign.
236  * @base: The number base to use. The maximum supported base is 16. If base is
237  *  given as 0, then the base of the string is automatically detected with the
238  *  conventional semantics - If it begins with 0x the number will be parsed as a
239  *  hexadecimal (case insensitive), if it otherwise begins with 0, it will be
240  *  parsed as an octal number. Otherwise it will be parsed as a decimal.
241  * @res: Where to write the result of the conversion on success.
242  *
243  * Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error.
244  * Preferred over simple_strtol(). Return code must be checked.
245  */
246 static inline int __must_check kstrtol(const char *s, unsigned int base, long *res)
247 {
248 	/*
249 	 * We want to shortcut function call, but
250 	 * __builtin_types_compatible_p(long, long long) = 0.
251 	 */
252 	if (sizeof(long) == sizeof(long long) &&
253 	    __alignof__(long) == __alignof__(long long))
254 		return kstrtoll(s, base, (long long *)res);
255 	else
256 		return _kstrtol(s, base, res);
257 }
258 
259 int __must_check kstrtouint(const char *s, unsigned int base, unsigned int *res);
260 int __must_check kstrtoint(const char *s, unsigned int base, int *res);
261 
262 static inline int __must_check kstrtou64(const char *s, unsigned int base, u64 *res)
263 {
264 	return kstrtoull(s, base, res);
265 }
266 
267 static inline int __must_check kstrtos64(const char *s, unsigned int base, s64 *res)
268 {
269 	return kstrtoll(s, base, res);
270 }
271 
272 static inline int __must_check kstrtou32(const char *s, unsigned int base, u32 *res)
273 {
274 	return kstrtouint(s, base, res);
275 }
276 
277 static inline int __must_check kstrtos32(const char *s, unsigned int base, s32 *res)
278 {
279 	return kstrtoint(s, base, res);
280 }
281 
282 int __must_check kstrtou16(const char *s, unsigned int base, u16 *res);
283 int __must_check kstrtos16(const char *s, unsigned int base, s16 *res);
284 int __must_check kstrtou8(const char *s, unsigned int base, u8 *res);
285 int __must_check kstrtos8(const char *s, unsigned int base, s8 *res);
286 int __must_check kstrtobool(const char *s, bool *res);
287 
288 int __must_check kstrtoull_from_user(const char __user *s, size_t count, unsigned int base, unsigned long long *res);
289 int __must_check kstrtoll_from_user(const char __user *s, size_t count, unsigned int base, long long *res);
290 int __must_check kstrtoul_from_user(const char __user *s, size_t count, unsigned int base, unsigned long *res);
291 int __must_check kstrtol_from_user(const char __user *s, size_t count, unsigned int base, long *res);
292 int __must_check kstrtouint_from_user(const char __user *s, size_t count, unsigned int base, unsigned int *res);
293 int __must_check kstrtoint_from_user(const char __user *s, size_t count, unsigned int base, int *res);
294 int __must_check kstrtou16_from_user(const char __user *s, size_t count, unsigned int base, u16 *res);
295 int __must_check kstrtos16_from_user(const char __user *s, size_t count, unsigned int base, s16 *res);
296 int __must_check kstrtou8_from_user(const char __user *s, size_t count, unsigned int base, u8 *res);
297 int __must_check kstrtos8_from_user(const char __user *s, size_t count, unsigned int base, s8 *res);
298 int __must_check kstrtobool_from_user(const char __user *s, size_t count, bool *res);
299 
300 static inline int __must_check kstrtou64_from_user(const char __user *s, size_t count, unsigned int base, u64 *res)
301 {
302 	return kstrtoull_from_user(s, count, base, res);
303 }
304 
305 static inline int __must_check kstrtos64_from_user(const char __user *s, size_t count, unsigned int base, s64 *res)
306 {
307 	return kstrtoll_from_user(s, count, base, res);
308 }
309 
310 static inline int __must_check kstrtou32_from_user(const char __user *s, size_t count, unsigned int base, u32 *res)
311 {
312 	return kstrtouint_from_user(s, count, base, res);
313 }
314 
315 static inline int __must_check kstrtos32_from_user(const char __user *s, size_t count, unsigned int base, s32 *res)
316 {
317 	return kstrtoint_from_user(s, count, base, res);
318 }
319 
320 /*
321  * Use kstrto<foo> instead.
322  *
323  * NOTE: simple_strto<foo> does not check for the range overflow and,
324  *	 depending on the input, may give interesting results.
325  *
326  * Use these functions if and only if you cannot use kstrto<foo>, because
327  * the conversion ends on the first non-digit character, which may be far
328  * beyond the supported range. It might be useful to parse the strings like
329  * 10x50 or 12:21 without altering original string or temporary buffer in use.
330  * Keep in mind above caveat.
331  */
332 
333 extern unsigned long simple_strtoul(const char *,char **,unsigned int);
334 extern long simple_strtol(const char *,char **,unsigned int);
335 extern unsigned long long simple_strtoull(const char *,char **,unsigned int);
336 extern long long simple_strtoll(const char *,char **,unsigned int);
337 
338 extern int num_to_str(char *buf, int size,
339 		      unsigned long long num, unsigned int width);
340 
341 /* lib/printf utilities */
342 
343 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...);
344 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list);
345 extern __printf(3, 4)
346 int snprintf(char *buf, size_t size, const char *fmt, ...);
347 extern __printf(3, 0)
348 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args);
349 extern __printf(3, 4)
350 int scnprintf(char *buf, size_t size, const char *fmt, ...);
351 extern __printf(3, 0)
352 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args);
353 extern __printf(2, 3) __malloc
354 char *kasprintf(gfp_t gfp, const char *fmt, ...);
355 extern __printf(2, 0) __malloc
356 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args);
357 extern __printf(2, 0)
358 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args);
359 
360 extern __scanf(2, 3)
361 int sscanf(const char *, const char *, ...);
362 extern __scanf(2, 0)
363 int vsscanf(const char *, const char *, va_list);
364 
365 extern int get_option(char **str, int *pint);
366 extern char *get_options(const char *str, int nints, int *ints);
367 extern unsigned long long memparse(const char *ptr, char **retptr);
368 extern bool parse_option_str(const char *str, const char *option);
369 extern char *next_arg(char *args, char **param, char **val);
370 
371 extern int core_kernel_text(unsigned long addr);
372 extern int init_kernel_text(unsigned long addr);
373 extern int core_kernel_data(unsigned long addr);
374 extern int __kernel_text_address(unsigned long addr);
375 extern int kernel_text_address(unsigned long addr);
376 extern int func_ptr_is_kernel_text(void *ptr);
377 
378 #ifdef CONFIG_SMP
379 extern unsigned int sysctl_oops_all_cpu_backtrace;
380 #else
381 #define sysctl_oops_all_cpu_backtrace 0
382 #endif /* CONFIG_SMP */
383 
384 extern void bust_spinlocks(int yes);
385 extern int panic_timeout;
386 extern unsigned long panic_print;
387 extern int panic_on_oops;
388 extern int panic_on_unrecovered_nmi;
389 extern int panic_on_io_nmi;
390 extern int panic_on_warn;
391 extern unsigned long panic_on_taint;
392 extern bool panic_on_taint_nousertaint;
393 extern int sysctl_panic_on_rcu_stall;
394 extern int sysctl_max_rcu_stall_to_panic;
395 extern int sysctl_panic_on_stackoverflow;
396 
397 extern bool crash_kexec_post_notifiers;
398 
399 /*
400  * panic_cpu is used for synchronizing panic() and crash_kexec() execution. It
401  * holds a CPU number which is executing panic() currently. A value of
402  * PANIC_CPU_INVALID means no CPU has entered panic() or crash_kexec().
403  */
404 extern atomic_t panic_cpu;
405 #define PANIC_CPU_INVALID	-1
406 
407 /*
408  * Only to be used by arch init code. If the user over-wrote the default
409  * CONFIG_PANIC_TIMEOUT, honor it.
410  */
411 static inline void set_arch_panic_timeout(int timeout, int arch_default_timeout)
412 {
413 	if (panic_timeout == arch_default_timeout)
414 		panic_timeout = timeout;
415 }
416 extern const char *print_tainted(void);
417 enum lockdep_ok {
418 	LOCKDEP_STILL_OK,
419 	LOCKDEP_NOW_UNRELIABLE
420 };
421 extern void add_taint(unsigned flag, enum lockdep_ok);
422 extern int test_taint(unsigned flag);
423 extern unsigned long get_taint(void);
424 extern int root_mountflags;
425 
426 extern bool early_boot_irqs_disabled;
427 
428 /*
429  * Values used for system_state. Ordering of the states must not be changed
430  * as code checks for <, <=, >, >= STATE.
431  */
432 extern enum system_states {
433 	SYSTEM_BOOTING,
434 	SYSTEM_SCHEDULING,
435 	SYSTEM_RUNNING,
436 	SYSTEM_HALT,
437 	SYSTEM_POWER_OFF,
438 	SYSTEM_RESTART,
439 	SYSTEM_SUSPEND,
440 } system_state;
441 
442 /* This cannot be an enum because some may be used in assembly source. */
443 #define TAINT_PROPRIETARY_MODULE	0
444 #define TAINT_FORCED_MODULE		1
445 #define TAINT_CPU_OUT_OF_SPEC		2
446 #define TAINT_FORCED_RMMOD		3
447 #define TAINT_MACHINE_CHECK		4
448 #define TAINT_BAD_PAGE			5
449 #define TAINT_USER			6
450 #define TAINT_DIE			7
451 #define TAINT_OVERRIDDEN_ACPI_TABLE	8
452 #define TAINT_WARN			9
453 #define TAINT_CRAP			10
454 #define TAINT_FIRMWARE_WORKAROUND	11
455 #define TAINT_OOT_MODULE		12
456 #define TAINT_UNSIGNED_MODULE		13
457 #define TAINT_SOFTLOCKUP		14
458 #define TAINT_LIVEPATCH			15
459 #define TAINT_AUX			16
460 #define TAINT_RANDSTRUCT		17
461 #define TAINT_FLAGS_COUNT		18
462 #define TAINT_FLAGS_MAX			((1UL << TAINT_FLAGS_COUNT) - 1)
463 
464 struct taint_flag {
465 	char c_true;	/* character printed when tainted */
466 	char c_false;	/* character printed when not tainted */
467 	bool module;	/* also show as a per-module taint flag */
468 };
469 
470 extern const struct taint_flag taint_flags[TAINT_FLAGS_COUNT];
471 
472 extern const char hex_asc[];
473 #define hex_asc_lo(x)	hex_asc[((x) & 0x0f)]
474 #define hex_asc_hi(x)	hex_asc[((x) & 0xf0) >> 4]
475 
476 static inline char *hex_byte_pack(char *buf, u8 byte)
477 {
478 	*buf++ = hex_asc_hi(byte);
479 	*buf++ = hex_asc_lo(byte);
480 	return buf;
481 }
482 
483 extern const char hex_asc_upper[];
484 #define hex_asc_upper_lo(x)	hex_asc_upper[((x) & 0x0f)]
485 #define hex_asc_upper_hi(x)	hex_asc_upper[((x) & 0xf0) >> 4]
486 
487 static inline char *hex_byte_pack_upper(char *buf, u8 byte)
488 {
489 	*buf++ = hex_asc_upper_hi(byte);
490 	*buf++ = hex_asc_upper_lo(byte);
491 	return buf;
492 }
493 
494 extern int hex_to_bin(char ch);
495 extern int __must_check hex2bin(u8 *dst, const char *src, size_t count);
496 extern char *bin2hex(char *dst, const void *src, size_t count);
497 
498 bool mac_pton(const char *s, u8 *mac);
499 
500 /*
501  * General tracing related utility functions - trace_printk(),
502  * tracing_on/tracing_off and tracing_start()/tracing_stop
503  *
504  * Use tracing_on/tracing_off when you want to quickly turn on or off
505  * tracing. It simply enables or disables the recording of the trace events.
506  * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
507  * file, which gives a means for the kernel and userspace to interact.
508  * Place a tracing_off() in the kernel where you want tracing to end.
509  * From user space, examine the trace, and then echo 1 > tracing_on
510  * to continue tracing.
511  *
512  * tracing_stop/tracing_start has slightly more overhead. It is used
513  * by things like suspend to ram where disabling the recording of the
514  * trace is not enough, but tracing must actually stop because things
515  * like calling smp_processor_id() may crash the system.
516  *
517  * Most likely, you want to use tracing_on/tracing_off.
518  */
519 
520 enum ftrace_dump_mode {
521 	DUMP_NONE,
522 	DUMP_ALL,
523 	DUMP_ORIG,
524 };
525 
526 #ifdef CONFIG_TRACING
527 void tracing_on(void);
528 void tracing_off(void);
529 int tracing_is_on(void);
530 void tracing_snapshot(void);
531 void tracing_snapshot_alloc(void);
532 
533 extern void tracing_start(void);
534 extern void tracing_stop(void);
535 
536 static inline __printf(1, 2)
537 void ____trace_printk_check_format(const char *fmt, ...)
538 {
539 }
540 #define __trace_printk_check_format(fmt, args...)			\
541 do {									\
542 	if (0)								\
543 		____trace_printk_check_format(fmt, ##args);		\
544 } while (0)
545 
546 /**
547  * trace_printk - printf formatting in the ftrace buffer
548  * @fmt: the printf format for printing
549  *
550  * Note: __trace_printk is an internal function for trace_printk() and
551  *       the @ip is passed in via the trace_printk() macro.
552  *
553  * This function allows a kernel developer to debug fast path sections
554  * that printk is not appropriate for. By scattering in various
555  * printk like tracing in the code, a developer can quickly see
556  * where problems are occurring.
557  *
558  * This is intended as a debugging tool for the developer only.
559  * Please refrain from leaving trace_printks scattered around in
560  * your code. (Extra memory is used for special buffers that are
561  * allocated when trace_printk() is used.)
562  *
563  * A little optimization trick is done here. If there's only one
564  * argument, there's no need to scan the string for printf formats.
565  * The trace_puts() will suffice. But how can we take advantage of
566  * using trace_puts() when trace_printk() has only one argument?
567  * By stringifying the args and checking the size we can tell
568  * whether or not there are args. __stringify((__VA_ARGS__)) will
569  * turn into "()\0" with a size of 3 when there are no args, anything
570  * else will be bigger. All we need to do is define a string to this,
571  * and then take its size and compare to 3. If it's bigger, use
572  * do_trace_printk() otherwise, optimize it to trace_puts(). Then just
573  * let gcc optimize the rest.
574  */
575 
576 #define trace_printk(fmt, ...)				\
577 do {							\
578 	char _______STR[] = __stringify((__VA_ARGS__));	\
579 	if (sizeof(_______STR) > 3)			\
580 		do_trace_printk(fmt, ##__VA_ARGS__);	\
581 	else						\
582 		trace_puts(fmt);			\
583 } while (0)
584 
585 #define do_trace_printk(fmt, args...)					\
586 do {									\
587 	static const char *trace_printk_fmt __used			\
588 		__section("__trace_printk_fmt") =			\
589 		__builtin_constant_p(fmt) ? fmt : NULL;			\
590 									\
591 	__trace_printk_check_format(fmt, ##args);			\
592 									\
593 	if (__builtin_constant_p(fmt))					\
594 		__trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args);	\
595 	else								\
596 		__trace_printk(_THIS_IP_, fmt, ##args);			\
597 } while (0)
598 
599 extern __printf(2, 3)
600 int __trace_bprintk(unsigned long ip, const char *fmt, ...);
601 
602 extern __printf(2, 3)
603 int __trace_printk(unsigned long ip, const char *fmt, ...);
604 
605 /**
606  * trace_puts - write a string into the ftrace buffer
607  * @str: the string to record
608  *
609  * Note: __trace_bputs is an internal function for trace_puts and
610  *       the @ip is passed in via the trace_puts macro.
611  *
612  * This is similar to trace_printk() but is made for those really fast
613  * paths that a developer wants the least amount of "Heisenbug" effects,
614  * where the processing of the print format is still too much.
615  *
616  * This function allows a kernel developer to debug fast path sections
617  * that printk is not appropriate for. By scattering in various
618  * printk like tracing in the code, a developer can quickly see
619  * where problems are occurring.
620  *
621  * This is intended as a debugging tool for the developer only.
622  * Please refrain from leaving trace_puts scattered around in
623  * your code. (Extra memory is used for special buffers that are
624  * allocated when trace_puts() is used.)
625  *
626  * Returns: 0 if nothing was written, positive # if string was.
627  *  (1 when __trace_bputs is used, strlen(str) when __trace_puts is used)
628  */
629 
630 #define trace_puts(str) ({						\
631 	static const char *trace_printk_fmt __used			\
632 		__section("__trace_printk_fmt") =			\
633 		__builtin_constant_p(str) ? str : NULL;			\
634 									\
635 	if (__builtin_constant_p(str))					\
636 		__trace_bputs(_THIS_IP_, trace_printk_fmt);		\
637 	else								\
638 		__trace_puts(_THIS_IP_, str, strlen(str));		\
639 })
640 extern int __trace_bputs(unsigned long ip, const char *str);
641 extern int __trace_puts(unsigned long ip, const char *str, int size);
642 
643 extern void trace_dump_stack(int skip);
644 
645 /*
646  * The double __builtin_constant_p is because gcc will give us an error
647  * if we try to allocate the static variable to fmt if it is not a
648  * constant. Even with the outer if statement.
649  */
650 #define ftrace_vprintk(fmt, vargs)					\
651 do {									\
652 	if (__builtin_constant_p(fmt)) {				\
653 		static const char *trace_printk_fmt __used		\
654 		  __section("__trace_printk_fmt") =			\
655 			__builtin_constant_p(fmt) ? fmt : NULL;		\
656 									\
657 		__ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs);	\
658 	} else								\
659 		__ftrace_vprintk(_THIS_IP_, fmt, vargs);		\
660 } while (0)
661 
662 extern __printf(2, 0) int
663 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap);
664 
665 extern __printf(2, 0) int
666 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap);
667 
668 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode);
669 #else
670 static inline void tracing_start(void) { }
671 static inline void tracing_stop(void) { }
672 static inline void trace_dump_stack(int skip) { }
673 
674 static inline void tracing_on(void) { }
675 static inline void tracing_off(void) { }
676 static inline int tracing_is_on(void) { return 0; }
677 static inline void tracing_snapshot(void) { }
678 static inline void tracing_snapshot_alloc(void) { }
679 
680 static inline __printf(1, 2)
681 int trace_printk(const char *fmt, ...)
682 {
683 	return 0;
684 }
685 static __printf(1, 0) inline int
686 ftrace_vprintk(const char *fmt, va_list ap)
687 {
688 	return 0;
689 }
690 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
691 #endif /* CONFIG_TRACING */
692 
693 /* This counts to 12. Any more, it will return 13th argument. */
694 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n
695 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0)
696 
697 #define __CONCAT(a, b) a ## b
698 #define CONCATENATE(a, b) __CONCAT(a, b)
699 
700 /**
701  * container_of - cast a member of a structure out to the containing structure
702  * @ptr:	the pointer to the member.
703  * @type:	the type of the container struct this is embedded in.
704  * @member:	the name of the member within the struct.
705  *
706  */
707 #define container_of(ptr, type, member) ({				\
708 	void *__mptr = (void *)(ptr);					\
709 	BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) &&	\
710 			 !__same_type(*(ptr), void),			\
711 			 "pointer type mismatch in container_of()");	\
712 	((type *)(__mptr - offsetof(type, member))); })
713 
714 /**
715  * container_of_safe - cast a member of a structure out to the containing structure
716  * @ptr:	the pointer to the member.
717  * @type:	the type of the container struct this is embedded in.
718  * @member:	the name of the member within the struct.
719  *
720  * If IS_ERR_OR_NULL(ptr), ptr is returned unchanged.
721  */
722 #define container_of_safe(ptr, type, member) ({				\
723 	void *__mptr = (void *)(ptr);					\
724 	BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) &&	\
725 			 !__same_type(*(ptr), void),			\
726 			 "pointer type mismatch in container_of()");	\
727 	IS_ERR_OR_NULL(__mptr) ? ERR_CAST(__mptr) :			\
728 		((type *)(__mptr - offsetof(type, member))); })
729 
730 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */
731 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
732 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD
733 #endif
734 
735 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */
736 #define VERIFY_OCTAL_PERMISSIONS(perms)						\
737 	(BUILD_BUG_ON_ZERO((perms) < 0) +					\
738 	 BUILD_BUG_ON_ZERO((perms) > 0777) +					\
739 	 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */		\
740 	 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) +	\
741 	 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) +		\
742 	 /* USER_WRITABLE >= GROUP_WRITABLE */					\
743 	 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) +	\
744 	 /* OTHER_WRITABLE?  Generally considered a bad idea. */		\
745 	 BUILD_BUG_ON_ZERO((perms) & 2) +					\
746 	 (perms))
747 #endif
748