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/hex.h>
24 #include <linux/kstrtox.h>
25 #include <linux/log2.h>
26 #include <linux/math.h>
27 #include <linux/minmax.h>
28 #include <linux/typecheck.h>
29 #include <linux/panic.h>
30 #include <linux/printk.h>
31 #include <linux/build_bug.h>
32 #include <linux/static_call_types.h>
33 #include <linux/instruction_pointer.h>
34 #include <asm/byteorder.h>
36 #include <uapi/linux/kernel.h>
38 #define STACK_MAGIC 0xdeadbeef
41 * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value
44 * NOTE: @x is not checked for > 0xff; larger values produce odd results.
46 #define REPEAT_BYTE(x) ((~0ul / 0xff) * (x))
48 /* generic data direction definitions */
53 * ARRAY_SIZE - get the number of elements in array @arr
54 * @arr: array to be sized
56 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr))
58 #define PTR_IF(cond, ptr) ((cond) ? (ptr) : NULL)
60 #define u64_to_user_ptr(x) ( \
62 typecheck(u64, (x)); \
63 (void __user *)(uintptr_t)(x); \
68 * upper_32_bits - return bits 32-63 of a number
69 * @n: the number we're accessing
71 * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress
72 * the "right shift count >= width of type" warning when that quantity is
75 #define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
78 * lower_32_bits - return bits 0-31 of a number
79 * @n: the number we're accessing
81 #define lower_32_bits(n) ((u32)((n) & 0xffffffff))
84 * upper_16_bits - return bits 16-31 of a number
85 * @n: the number we're accessing
87 #define upper_16_bits(n) ((u16)((n) >> 16))
90 * lower_16_bits - return bits 0-15 of a number
91 * @n: the number we're accessing
93 #define lower_16_bits(n) ((u16)((n) & 0xffff))
98 #ifdef CONFIG_PREEMPT_VOLUNTARY_BUILD
100 extern int __cond_resched(void);
101 # define might_resched() __cond_resched()
103 #elif defined(CONFIG_PREEMPT_DYNAMIC) && defined(CONFIG_HAVE_PREEMPT_DYNAMIC_CALL)
105 extern int __cond_resched(void);
107 DECLARE_STATIC_CALL(might_resched, __cond_resched);
109 static __always_inline void might_resched(void)
111 static_call_mod(might_resched)();
114 #elif defined(CONFIG_PREEMPT_DYNAMIC) && defined(CONFIG_HAVE_PREEMPT_DYNAMIC_KEY)
116 extern int dynamic_might_resched(void);
117 # define might_resched() dynamic_might_resched()
121 # define might_resched() do { } while (0)
123 #endif /* CONFIG_PREEMPT_* */
125 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
126 extern void __might_resched(const char *file, int line, unsigned int offsets);
127 extern void __might_sleep(const char *file, int line);
128 extern void __cant_sleep(const char *file, int line, int preempt_offset);
129 extern void __cant_migrate(const char *file, int line);
132 * might_sleep - annotation for functions that can sleep
134 * this macro will print a stack trace if it is executed in an atomic
135 * context (spinlock, irq-handler, ...). Additional sections where blocking is
136 * not allowed can be annotated with non_block_start() and non_block_end()
139 * This is a useful debugging help to be able to catch problems early and not
140 * be bitten later when the calling function happens to sleep when it is not
143 # define might_sleep() \
144 do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0)
146 * cant_sleep - annotation for functions that cannot sleep
148 * this macro will print a stack trace if it is executed with preemption enabled
150 # define cant_sleep() \
151 do { __cant_sleep(__FILE__, __LINE__, 0); } while (0)
152 # define sched_annotate_sleep() (current->task_state_change = 0)
155 * cant_migrate - annotation for functions that cannot migrate
157 * Will print a stack trace if executed in code which is migratable
159 # define cant_migrate() \
161 if (IS_ENABLED(CONFIG_SMP)) \
162 __cant_migrate(__FILE__, __LINE__); \
166 * non_block_start - annotate the start of section where sleeping is prohibited
168 * This is on behalf of the oom reaper, specifically when it is calling the mmu
169 * notifiers. The problem is that if the notifier were to block on, for example,
170 * mutex_lock() and if the process which holds that mutex were to perform a
171 * sleeping memory allocation, the oom reaper is now blocked on completion of
172 * that memory allocation. Other blocking calls like wait_event() pose similar
175 # define non_block_start() (current->non_block_count++)
177 * non_block_end - annotate the end of section where sleeping is prohibited
179 * Closes a section opened by non_block_start().
181 # define non_block_end() WARN_ON(current->non_block_count-- == 0)
183 static inline void __might_resched(const char *file, int line,
184 unsigned int offsets) { }
185 static inline void __might_sleep(const char *file, int line) { }
186 # define might_sleep() do { might_resched(); } while (0)
187 # define cant_sleep() do { } while (0)
188 # define cant_migrate() do { } while (0)
189 # define sched_annotate_sleep() do { } while (0)
190 # define non_block_start() do { } while (0)
191 # define non_block_end() do { } while (0)
194 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0)
196 #if defined(CONFIG_MMU) && \
197 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP))
198 #define might_fault() __might_fault(__FILE__, __LINE__)
199 void __might_fault(const char *file, int line);
201 static inline void might_fault(void) { }
204 void do_exit(long error_code) __noreturn;
206 extern int num_to_str(char *buf, int size,
207 unsigned long long num, unsigned int width);
209 /* lib/printf utilities */
211 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...);
212 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list);
213 extern __printf(3, 4)
214 int snprintf(char *buf, size_t size, const char *fmt, ...);
215 extern __printf(3, 0)
216 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args);
217 extern __printf(3, 4)
218 int scnprintf(char *buf, size_t size, const char *fmt, ...);
219 extern __printf(3, 0)
220 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args);
221 extern __printf(2, 3) __malloc
222 char *kasprintf(gfp_t gfp, const char *fmt, ...);
223 extern __printf(2, 0) __malloc
224 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args);
225 extern __printf(2, 0)
226 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args);
229 int sscanf(const char *, const char *, ...);
231 int vsscanf(const char *, const char *, va_list);
233 extern int no_hash_pointers_enable(char *str);
235 extern int get_option(char **str, int *pint);
236 extern char *get_options(const char *str, int nints, int *ints);
237 extern unsigned long long memparse(const char *ptr, char **retptr);
238 extern bool parse_option_str(const char *str, const char *option);
239 extern char *next_arg(char *args, char **param, char **val);
241 extern int core_kernel_text(unsigned long addr);
242 extern int __kernel_text_address(unsigned long addr);
243 extern int kernel_text_address(unsigned long addr);
244 extern int func_ptr_is_kernel_text(void *ptr);
246 extern void bust_spinlocks(int yes);
248 extern int root_mountflags;
250 extern bool early_boot_irqs_disabled;
253 * Values used for system_state. Ordering of the states must not be changed
254 * as code checks for <, <=, >, >= STATE.
256 extern enum system_states {
259 SYSTEM_FREEING_INITMEM,
268 * General tracing related utility functions - trace_printk(),
269 * tracing_on/tracing_off and tracing_start()/tracing_stop
271 * Use tracing_on/tracing_off when you want to quickly turn on or off
272 * tracing. It simply enables or disables the recording of the trace events.
273 * This also corresponds to the user space /sys/kernel/tracing/tracing_on
274 * file, which gives a means for the kernel and userspace to interact.
275 * Place a tracing_off() in the kernel where you want tracing to end.
276 * From user space, examine the trace, and then echo 1 > tracing_on
277 * to continue tracing.
279 * tracing_stop/tracing_start has slightly more overhead. It is used
280 * by things like suspend to ram where disabling the recording of the
281 * trace is not enough, but tracing must actually stop because things
282 * like calling smp_processor_id() may crash the system.
284 * Most likely, you want to use tracing_on/tracing_off.
287 enum ftrace_dump_mode {
293 #ifdef CONFIG_TRACING
294 void tracing_on(void);
295 void tracing_off(void);
296 int tracing_is_on(void);
297 void tracing_snapshot(void);
298 void tracing_snapshot_alloc(void);
300 extern void tracing_start(void);
301 extern void tracing_stop(void);
303 static inline __printf(1, 2)
304 void ____trace_printk_check_format(const char *fmt, ...)
307 #define __trace_printk_check_format(fmt, args...) \
310 ____trace_printk_check_format(fmt, ##args); \
314 * trace_printk - printf formatting in the ftrace buffer
315 * @fmt: the printf format for printing
317 * Note: __trace_printk is an internal function for trace_printk() and
318 * the @ip is passed in via the trace_printk() macro.
320 * This function allows a kernel developer to debug fast path sections
321 * that printk is not appropriate for. By scattering in various
322 * printk like tracing in the code, a developer can quickly see
323 * where problems are occurring.
325 * This is intended as a debugging tool for the developer only.
326 * Please refrain from leaving trace_printks scattered around in
327 * your code. (Extra memory is used for special buffers that are
328 * allocated when trace_printk() is used.)
330 * A little optimization trick is done here. If there's only one
331 * argument, there's no need to scan the string for printf formats.
332 * The trace_puts() will suffice. But how can we take advantage of
333 * using trace_puts() when trace_printk() has only one argument?
334 * By stringifying the args and checking the size we can tell
335 * whether or not there are args. __stringify((__VA_ARGS__)) will
336 * turn into "()\0" with a size of 3 when there are no args, anything
337 * else will be bigger. All we need to do is define a string to this,
338 * and then take its size and compare to 3. If it's bigger, use
339 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just
340 * let gcc optimize the rest.
343 #define trace_printk(fmt, ...) \
345 char _______STR[] = __stringify((__VA_ARGS__)); \
346 if (sizeof(_______STR) > 3) \
347 do_trace_printk(fmt, ##__VA_ARGS__); \
352 #define do_trace_printk(fmt, args...) \
354 static const char *trace_printk_fmt __used \
355 __section("__trace_printk_fmt") = \
356 __builtin_constant_p(fmt) ? fmt : NULL; \
358 __trace_printk_check_format(fmt, ##args); \
360 if (__builtin_constant_p(fmt)) \
361 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \
363 __trace_printk(_THIS_IP_, fmt, ##args); \
366 extern __printf(2, 3)
367 int __trace_bprintk(unsigned long ip, const char *fmt, ...);
369 extern __printf(2, 3)
370 int __trace_printk(unsigned long ip, const char *fmt, ...);
373 * trace_puts - write a string into the ftrace buffer
374 * @str: the string to record
376 * Note: __trace_bputs is an internal function for trace_puts and
377 * the @ip is passed in via the trace_puts macro.
379 * This is similar to trace_printk() but is made for those really fast
380 * paths that a developer wants the least amount of "Heisenbug" effects,
381 * where the processing of the print format is still too much.
383 * This function allows a kernel developer to debug fast path sections
384 * that printk is not appropriate for. By scattering in various
385 * printk like tracing in the code, a developer can quickly see
386 * where problems are occurring.
388 * This is intended as a debugging tool for the developer only.
389 * Please refrain from leaving trace_puts scattered around in
390 * your code. (Extra memory is used for special buffers that are
391 * allocated when trace_puts() is used.)
393 * Returns: 0 if nothing was written, positive # if string was.
394 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used)
397 #define trace_puts(str) ({ \
398 static const char *trace_printk_fmt __used \
399 __section("__trace_printk_fmt") = \
400 __builtin_constant_p(str) ? str : NULL; \
402 if (__builtin_constant_p(str)) \
403 __trace_bputs(_THIS_IP_, trace_printk_fmt); \
405 __trace_puts(_THIS_IP_, str, strlen(str)); \
407 extern int __trace_bputs(unsigned long ip, const char *str);
408 extern int __trace_puts(unsigned long ip, const char *str, int size);
410 extern void trace_dump_stack(int skip);
413 * The double __builtin_constant_p is because gcc will give us an error
414 * if we try to allocate the static variable to fmt if it is not a
415 * constant. Even with the outer if statement.
417 #define ftrace_vprintk(fmt, vargs) \
419 if (__builtin_constant_p(fmt)) { \
420 static const char *trace_printk_fmt __used \
421 __section("__trace_printk_fmt") = \
422 __builtin_constant_p(fmt) ? fmt : NULL; \
424 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \
426 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \
429 extern __printf(2, 0) int
430 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap);
432 extern __printf(2, 0) int
433 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap);
435 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode);
437 static inline void tracing_start(void) { }
438 static inline void tracing_stop(void) { }
439 static inline void trace_dump_stack(int skip) { }
441 static inline void tracing_on(void) { }
442 static inline void tracing_off(void) { }
443 static inline int tracing_is_on(void) { return 0; }
444 static inline void tracing_snapshot(void) { }
445 static inline void tracing_snapshot_alloc(void) { }
447 static inline __printf(1, 2)
448 int trace_printk(const char *fmt, ...)
452 static __printf(1, 0) inline int
453 ftrace_vprintk(const char *fmt, va_list ap)
457 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
458 #endif /* CONFIG_TRACING */
460 /* This counts to 12. Any more, it will return 13th argument. */
461 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n
462 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0)
464 #define __CONCAT(a, b) a ## b
465 #define CONCATENATE(a, b) __CONCAT(a, b)
467 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */
468 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
469 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD
472 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */
473 #define VERIFY_OCTAL_PERMISSIONS(perms) \
474 (BUILD_BUG_ON_ZERO((perms) < 0) + \
475 BUILD_BUG_ON_ZERO((perms) > 0777) + \
476 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \
477 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \
478 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \
479 /* USER_WRITABLE >= GROUP_WRITABLE */ \
480 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \
481 /* OTHER_WRITABLE? Generally considered a bad idea. */ \
482 BUILD_BUG_ON_ZERO((perms) & 2) + \