1 /* SPDX-License-Identifier: GPL-2.0 */
5 * This header has combined a lot of unrelated to each other stuff.
6 * The process of splitting its content is in progress while keeping
7 * backward compatibility. That's why it's highly recommended NOT to
8 * include this header inside another header file, especially under
9 * generic or architectural include/ directory.
11 #ifndef _LINUX_KERNEL_H
12 #define _LINUX_KERNEL_H
14 #include <linux/stdarg.h>
15 #include <linux/align.h>
16 #include <linux/limits.h>
17 #include <linux/linkage.h>
18 #include <linux/stddef.h>
19 #include <linux/types.h>
20 #include <linux/compiler.h>
21 #include <linux/container_of.h>
22 #include <linux/bitops.h>
23 #include <linux/kstrtox.h>
24 #include <linux/log2.h>
25 #include <linux/math.h>
26 #include <linux/minmax.h>
27 #include <linux/typecheck.h>
28 #include <linux/panic.h>
29 #include <linux/printk.h>
30 #include <linux/build_bug.h>
31 #include <linux/static_call_types.h>
32 #include <linux/instruction_pointer.h>
33 #include <asm/byteorder.h>
35 #include <uapi/linux/kernel.h>
37 #define STACK_MAGIC 0xdeadbeef
40 * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value
43 * NOTE: @x is not checked for > 0xff; larger values produce odd results.
45 #define REPEAT_BYTE(x) ((~0ul / 0xff) * (x))
47 /* generic data direction definitions */
52 * ARRAY_SIZE - get the number of elements in array @arr
53 * @arr: array to be sized
55 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr))
57 #define PTR_IF(cond, ptr) ((cond) ? (ptr) : NULL)
59 #define u64_to_user_ptr(x) ( \
61 typecheck(u64, (x)); \
62 (void __user *)(uintptr_t)(x); \
67 * lower_48_bits() - return bits 0-47 of a number
68 * @n: the number we're accessing
70 static inline u64 lower_48_bits(u64 n)
72 return n & ((1ull << 48) - 1);
76 * upper_32_bits - return bits 32-63 of a number
77 * @n: the number we're accessing
79 * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress
80 * the "right shift count >= width of type" warning when that quantity is
83 #define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
86 * lower_32_bits - return bits 0-31 of a number
87 * @n: the number we're accessing
89 #define lower_32_bits(n) ((u32)((n) & 0xffffffff))
92 * upper_16_bits - return bits 16-31 of a number
93 * @n: the number we're accessing
95 #define upper_16_bits(n) ((u16)((n) >> 16))
98 * lower_16_bits - return bits 0-15 of a number
99 * @n: the number we're accessing
101 #define lower_16_bits(n) ((u16)((n) & 0xffff))
106 #ifdef CONFIG_PREEMPT_VOLUNTARY_BUILD
108 extern int __cond_resched(void);
109 # define might_resched() __cond_resched()
111 #elif defined(CONFIG_PREEMPT_DYNAMIC) && defined(CONFIG_HAVE_PREEMPT_DYNAMIC_CALL)
113 extern int __cond_resched(void);
115 DECLARE_STATIC_CALL(might_resched, __cond_resched);
117 static __always_inline void might_resched(void)
119 static_call_mod(might_resched)();
122 #elif defined(CONFIG_PREEMPT_DYNAMIC) && defined(CONFIG_HAVE_PREEMPT_DYNAMIC_KEY)
124 extern int dynamic_might_resched(void);
125 # define might_resched() dynamic_might_resched()
129 # define might_resched() do { } while (0)
131 #endif /* CONFIG_PREEMPT_* */
133 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
134 extern void __might_resched(const char *file, int line, unsigned int offsets);
135 extern void __might_sleep(const char *file, int line);
136 extern void __cant_sleep(const char *file, int line, int preempt_offset);
137 extern void __cant_migrate(const char *file, int line);
140 * might_sleep - annotation for functions that can sleep
142 * this macro will print a stack trace if it is executed in an atomic
143 * context (spinlock, irq-handler, ...). Additional sections where blocking is
144 * not allowed can be annotated with non_block_start() and non_block_end()
147 * This is a useful debugging help to be able to catch problems early and not
148 * be bitten later when the calling function happens to sleep when it is not
151 # define might_sleep() \
152 do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0)
154 * cant_sleep - annotation for functions that cannot sleep
156 * this macro will print a stack trace if it is executed with preemption enabled
158 # define cant_sleep() \
159 do { __cant_sleep(__FILE__, __LINE__, 0); } while (0)
160 # define sched_annotate_sleep() (current->task_state_change = 0)
163 * cant_migrate - annotation for functions that cannot migrate
165 * Will print a stack trace if executed in code which is migratable
167 # define cant_migrate() \
169 if (IS_ENABLED(CONFIG_SMP)) \
170 __cant_migrate(__FILE__, __LINE__); \
174 * non_block_start - annotate the start of section where sleeping is prohibited
176 * This is on behalf of the oom reaper, specifically when it is calling the mmu
177 * notifiers. The problem is that if the notifier were to block on, for example,
178 * mutex_lock() and if the process which holds that mutex were to perform a
179 * sleeping memory allocation, the oom reaper is now blocked on completion of
180 * that memory allocation. Other blocking calls like wait_event() pose similar
183 # define non_block_start() (current->non_block_count++)
185 * non_block_end - annotate the end of section where sleeping is prohibited
187 * Closes a section opened by non_block_start().
189 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
191 static inline void __might_resched(const char *file, int line,
192 unsigned int offsets) { }
193 static inline void __might_sleep(const char *file, int line) { }
194 # define might_sleep() do { might_resched(); } while (0)
195 # define cant_sleep() do { } while (0)
196 # define cant_migrate() do { } while (0)
197 # define sched_annotate_sleep() do { } while (0)
198 # define non_block_start() do { } while (0)
199 # define non_block_end() do { } while (0)
202 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0)
204 #if defined(CONFIG_MMU) && \
205 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP))
206 #define might_fault() __might_fault(__FILE__, __LINE__)
207 void __might_fault(const char *file, int line);
209 static inline void might_fault(void) { }
212 void do_exit(long error_code) __noreturn;
214 extern int num_to_str(char *buf, int size,
215 unsigned long long num, unsigned int width);
217 /* lib/printf utilities */
219 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...);
220 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list);
221 extern __printf(3, 4)
222 int snprintf(char *buf, size_t size, const char *fmt, ...);
223 extern __printf(3, 0)
224 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args);
225 extern __printf(3, 4)
226 int scnprintf(char *buf, size_t size, const char *fmt, ...);
227 extern __printf(3, 0)
228 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args);
229 extern __printf(2, 3) __malloc
230 char *kasprintf(gfp_t gfp, const char *fmt, ...);
231 extern __printf(2, 0) __malloc
232 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args);
233 extern __printf(2, 0)
234 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args);
237 int sscanf(const char *, const char *, ...);
239 int vsscanf(const char *, const char *, va_list);
241 extern int no_hash_pointers_enable(char *str);
243 extern int get_option(char **str, int *pint);
244 extern char *get_options(const char *str, int nints, int *ints);
245 extern unsigned long long memparse(const char *ptr, char **retptr);
246 extern bool parse_option_str(const char *str, const char *option);
247 extern char *next_arg(char *args, char **param, char **val);
249 extern int core_kernel_text(unsigned long addr);
250 extern int __kernel_text_address(unsigned long addr);
251 extern int kernel_text_address(unsigned long addr);
252 extern int func_ptr_is_kernel_text(void *ptr);
254 extern void bust_spinlocks(int yes);
256 extern int root_mountflags;
258 extern bool early_boot_irqs_disabled;
261 * Values used for system_state. Ordering of the states must not be changed
262 * as code checks for <, <=, >, >= STATE.
264 extern enum system_states {
267 SYSTEM_FREEING_INITMEM,
275 extern const char hex_asc[];
276 #define hex_asc_lo(x) hex_asc[((x) & 0x0f)]
277 #define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4]
279 static inline char *hex_byte_pack(char *buf, u8 byte)
281 *buf++ = hex_asc_hi(byte);
282 *buf++ = hex_asc_lo(byte);
286 extern const char hex_asc_upper[];
287 #define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)]
288 #define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4]
290 static inline char *hex_byte_pack_upper(char *buf, u8 byte)
292 *buf++ = hex_asc_upper_hi(byte);
293 *buf++ = hex_asc_upper_lo(byte);
297 extern int hex_to_bin(char ch);
298 extern int __must_check hex2bin(u8 *dst, const char *src, size_t count);
299 extern char *bin2hex(char *dst, const void *src, size_t count);
301 bool mac_pton(const char *s, u8 *mac);
304 * General tracing related utility functions - trace_printk(),
305 * tracing_on/tracing_off and tracing_start()/tracing_stop
307 * Use tracing_on/tracing_off when you want to quickly turn on or off
308 * tracing. It simply enables or disables the recording of the trace events.
309 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
310 * file, which gives a means for the kernel and userspace to interact.
311 * Place a tracing_off() in the kernel where you want tracing to end.
312 * From user space, examine the trace, and then echo 1 > tracing_on
313 * to continue tracing.
315 * tracing_stop/tracing_start has slightly more overhead. It is used
316 * by things like suspend to ram where disabling the recording of the
317 * trace is not enough, but tracing must actually stop because things
318 * like calling smp_processor_id() may crash the system.
320 * Most likely, you want to use tracing_on/tracing_off.
323 enum ftrace_dump_mode {
329 #ifdef CONFIG_TRACING
330 void tracing_on(void);
331 void tracing_off(void);
332 int tracing_is_on(void);
333 void tracing_snapshot(void);
334 void tracing_snapshot_alloc(void);
336 extern void tracing_start(void);
337 extern void tracing_stop(void);
339 static inline __printf(1, 2)
340 void ____trace_printk_check_format(const char *fmt, ...)
343 #define __trace_printk_check_format(fmt, args...) \
346 ____trace_printk_check_format(fmt, ##args); \
350 * trace_printk - printf formatting in the ftrace buffer
351 * @fmt: the printf format for printing
353 * Note: __trace_printk is an internal function for trace_printk() and
354 * the @ip is passed in via the trace_printk() macro.
356 * This function allows a kernel developer to debug fast path sections
357 * that printk is not appropriate for. By scattering in various
358 * printk like tracing in the code, a developer can quickly see
359 * where problems are occurring.
361 * This is intended as a debugging tool for the developer only.
362 * Please refrain from leaving trace_printks scattered around in
363 * your code. (Extra memory is used for special buffers that are
364 * allocated when trace_printk() is used.)
366 * A little optimization trick is done here. If there's only one
367 * argument, there's no need to scan the string for printf formats.
368 * The trace_puts() will suffice. But how can we take advantage of
369 * using trace_puts() when trace_printk() has only one argument?
370 * By stringifying the args and checking the size we can tell
371 * whether or not there are args. __stringify((__VA_ARGS__)) will
372 * turn into "()\0" with a size of 3 when there are no args, anything
373 * else will be bigger. All we need to do is define a string to this,
374 * and then take its size and compare to 3. If it's bigger, use
375 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just
376 * let gcc optimize the rest.
379 #define trace_printk(fmt, ...) \
381 char _______STR[] = __stringify((__VA_ARGS__)); \
382 if (sizeof(_______STR) > 3) \
383 do_trace_printk(fmt, ##__VA_ARGS__); \
388 #define do_trace_printk(fmt, args...) \
390 static const char *trace_printk_fmt __used \
391 __section("__trace_printk_fmt") = \
392 __builtin_constant_p(fmt) ? fmt : NULL; \
394 __trace_printk_check_format(fmt, ##args); \
396 if (__builtin_constant_p(fmt)) \
397 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \
399 __trace_printk(_THIS_IP_, fmt, ##args); \
402 extern __printf(2, 3)
403 int __trace_bprintk(unsigned long ip, const char *fmt, ...);
405 extern __printf(2, 3)
406 int __trace_printk(unsigned long ip, const char *fmt, ...);
409 * trace_puts - write a string into the ftrace buffer
410 * @str: the string to record
412 * Note: __trace_bputs is an internal function for trace_puts and
413 * the @ip is passed in via the trace_puts macro.
415 * This is similar to trace_printk() but is made for those really fast
416 * paths that a developer wants the least amount of "Heisenbug" effects,
417 * where the processing of the print format is still too much.
419 * This function allows a kernel developer to debug fast path sections
420 * that printk is not appropriate for. By scattering in various
421 * printk like tracing in the code, a developer can quickly see
422 * where problems are occurring.
424 * This is intended as a debugging tool for the developer only.
425 * Please refrain from leaving trace_puts scattered around in
426 * your code. (Extra memory is used for special buffers that are
427 * allocated when trace_puts() is used.)
429 * Returns: 0 if nothing was written, positive # if string was.
430 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used)
433 #define trace_puts(str) ({ \
434 static const char *trace_printk_fmt __used \
435 __section("__trace_printk_fmt") = \
436 __builtin_constant_p(str) ? str : NULL; \
438 if (__builtin_constant_p(str)) \
439 __trace_bputs(_THIS_IP_, trace_printk_fmt); \
441 __trace_puts(_THIS_IP_, str, strlen(str)); \
443 extern int __trace_bputs(unsigned long ip, const char *str);
444 extern int __trace_puts(unsigned long ip, const char *str, int size);
446 extern void trace_dump_stack(int skip);
449 * The double __builtin_constant_p is because gcc will give us an error
450 * if we try to allocate the static variable to fmt if it is not a
451 * constant. Even with the outer if statement.
453 #define ftrace_vprintk(fmt, vargs) \
455 if (__builtin_constant_p(fmt)) { \
456 static const char *trace_printk_fmt __used \
457 __section("__trace_printk_fmt") = \
458 __builtin_constant_p(fmt) ? fmt : NULL; \
460 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \
462 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \
465 extern __printf(2, 0) int
466 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap);
468 extern __printf(2, 0) int
469 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap);
471 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode);
473 static inline void tracing_start(void) { }
474 static inline void tracing_stop(void) { }
475 static inline void trace_dump_stack(int skip) { }
477 static inline void tracing_on(void) { }
478 static inline void tracing_off(void) { }
479 static inline int tracing_is_on(void) { return 0; }
480 static inline void tracing_snapshot(void) { }
481 static inline void tracing_snapshot_alloc(void) { }
483 static inline __printf(1, 2)
484 int trace_printk(const char *fmt, ...)
488 static __printf(1, 0) inline int
489 ftrace_vprintk(const char *fmt, va_list ap)
493 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
494 #endif /* CONFIG_TRACING */
496 /* This counts to 12. Any more, it will return 13th argument. */
497 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n
498 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0)
500 #define __CONCAT(a, b) a ## b
501 #define CONCATENATE(a, b) __CONCAT(a, b)
503 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */
504 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
505 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD
508 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */
509 #define VERIFY_OCTAL_PERMISSIONS(perms) \
510 (BUILD_BUG_ON_ZERO((perms) < 0) + \
511 BUILD_BUG_ON_ZERO((perms) > 0777) + \
512 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \
513 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \
514 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \
515 /* USER_WRITABLE >= GROUP_WRITABLE */ \
516 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \
517 /* OTHER_WRITABLE? Generally considered a bad idea. */ \
518 BUILD_BUG_ON_ZERO((perms) & 2) + \