at v5.16-rc6 16 kB view raw
1/* SPDX-License-Identifier: GPL-2.0 */ 2#ifndef _LINUX_KERNEL_H 3#define _LINUX_KERNEL_H 4 5#include <linux/stdarg.h> 6#include <linux/align.h> 7#include <linux/limits.h> 8#include <linux/linkage.h> 9#include <linux/stddef.h> 10#include <linux/types.h> 11#include <linux/compiler.h> 12#include <linux/container_of.h> 13#include <linux/bitops.h> 14#include <linux/kstrtox.h> 15#include <linux/log2.h> 16#include <linux/math.h> 17#include <linux/minmax.h> 18#include <linux/typecheck.h> 19#include <linux/panic.h> 20#include <linux/printk.h> 21#include <linux/build_bug.h> 22#include <linux/static_call_types.h> 23#include <linux/instruction_pointer.h> 24#include <asm/byteorder.h> 25 26#include <uapi/linux/kernel.h> 27 28#define STACK_MAGIC 0xdeadbeef 29 30/** 31 * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value 32 * @x: value to repeat 33 * 34 * NOTE: @x is not checked for > 0xff; larger values produce odd results. 35 */ 36#define REPEAT_BYTE(x) ((~0ul / 0xff) * (x)) 37 38/* generic data direction definitions */ 39#define READ 0 40#define WRITE 1 41 42/** 43 * ARRAY_SIZE - get the number of elements in array @arr 44 * @arr: array to be sized 45 */ 46#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr)) 47 48#define PTR_IF(cond, ptr) ((cond) ? (ptr) : NULL) 49 50#define u64_to_user_ptr(x) ( \ 51{ \ 52 typecheck(u64, (x)); \ 53 (void __user *)(uintptr_t)(x); \ 54} \ 55) 56 57/** 58 * upper_32_bits - return bits 32-63 of a number 59 * @n: the number we're accessing 60 * 61 * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress 62 * the "right shift count >= width of type" warning when that quantity is 63 * 32-bits. 64 */ 65#define upper_32_bits(n) ((u32)(((n) >> 16) >> 16)) 66 67/** 68 * lower_32_bits - return bits 0-31 of a number 69 * @n: the number we're accessing 70 */ 71#define lower_32_bits(n) ((u32)((n) & 0xffffffff)) 72 73/** 74 * upper_16_bits - return bits 16-31 of a number 75 * @n: the number we're accessing 76 */ 77#define upper_16_bits(n) ((u16)((n) >> 16)) 78 79/** 80 * lower_16_bits - return bits 0-15 of a number 81 * @n: the number we're accessing 82 */ 83#define lower_16_bits(n) ((u16)((n) & 0xffff)) 84 85struct completion; 86struct user; 87 88#ifdef CONFIG_PREEMPT_VOLUNTARY_BUILD 89 90extern int __cond_resched(void); 91# define might_resched() __cond_resched() 92 93#elif defined(CONFIG_PREEMPT_DYNAMIC) 94 95extern int __cond_resched(void); 96 97DECLARE_STATIC_CALL(might_resched, __cond_resched); 98 99static __always_inline void might_resched(void) 100{ 101 static_call_mod(might_resched)(); 102} 103 104#else 105 106# define might_resched() do { } while (0) 107 108#endif /* CONFIG_PREEMPT_* */ 109 110#ifdef CONFIG_DEBUG_ATOMIC_SLEEP 111extern void __might_resched(const char *file, int line, unsigned int offsets); 112extern void __might_sleep(const char *file, int line); 113extern void __cant_sleep(const char *file, int line, int preempt_offset); 114extern void __cant_migrate(const char *file, int line); 115 116/** 117 * might_sleep - annotation for functions that can sleep 118 * 119 * this macro will print a stack trace if it is executed in an atomic 120 * context (spinlock, irq-handler, ...). Additional sections where blocking is 121 * not allowed can be annotated with non_block_start() and non_block_end() 122 * pairs. 123 * 124 * This is a useful debugging help to be able to catch problems early and not 125 * be bitten later when the calling function happens to sleep when it is not 126 * supposed to. 127 */ 128# define might_sleep() \ 129 do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0) 130/** 131 * cant_sleep - annotation for functions that cannot sleep 132 * 133 * this macro will print a stack trace if it is executed with preemption enabled 134 */ 135# define cant_sleep() \ 136 do { __cant_sleep(__FILE__, __LINE__, 0); } while (0) 137# define sched_annotate_sleep() (current->task_state_change = 0) 138 139/** 140 * cant_migrate - annotation for functions that cannot migrate 141 * 142 * Will print a stack trace if executed in code which is migratable 143 */ 144# define cant_migrate() \ 145 do { \ 146 if (IS_ENABLED(CONFIG_SMP)) \ 147 __cant_migrate(__FILE__, __LINE__); \ 148 } while (0) 149 150/** 151 * non_block_start - annotate the start of section where sleeping is prohibited 152 * 153 * This is on behalf of the oom reaper, specifically when it is calling the mmu 154 * notifiers. The problem is that if the notifier were to block on, for example, 155 * mutex_lock() and if the process which holds that mutex were to perform a 156 * sleeping memory allocation, the oom reaper is now blocked on completion of 157 * that memory allocation. Other blocking calls like wait_event() pose similar 158 * issues. 159 */ 160# define non_block_start() (current->non_block_count++) 161/** 162 * non_block_end - annotate the end of section where sleeping is prohibited 163 * 164 * Closes a section opened by non_block_start(). 165 */ 166# define non_block_end() WARN_ON(current->non_block_count-- == 0) 167#else 168 static inline void __might_resched(const char *file, int line, 169 unsigned int offsets) { } 170static inline void __might_sleep(const char *file, int line) { } 171# define might_sleep() do { might_resched(); } while (0) 172# define cant_sleep() do { } while (0) 173# define cant_migrate() do { } while (0) 174# define sched_annotate_sleep() do { } while (0) 175# define non_block_start() do { } while (0) 176# define non_block_end() do { } while (0) 177#endif 178 179#define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0) 180 181#if defined(CONFIG_MMU) && \ 182 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)) 183#define might_fault() __might_fault(__FILE__, __LINE__) 184void __might_fault(const char *file, int line); 185#else 186static inline void might_fault(void) { } 187#endif 188 189void do_exit(long error_code) __noreturn; 190void complete_and_exit(struct completion *, long) __noreturn; 191 192extern int num_to_str(char *buf, int size, 193 unsigned long long num, unsigned int width); 194 195/* lib/printf utilities */ 196 197extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...); 198extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list); 199extern __printf(3, 4) 200int snprintf(char *buf, size_t size, const char *fmt, ...); 201extern __printf(3, 0) 202int vsnprintf(char *buf, size_t size, const char *fmt, va_list args); 203extern __printf(3, 4) 204int scnprintf(char *buf, size_t size, const char *fmt, ...); 205extern __printf(3, 0) 206int vscnprintf(char *buf, size_t size, const char *fmt, va_list args); 207extern __printf(2, 3) __malloc 208char *kasprintf(gfp_t gfp, const char *fmt, ...); 209extern __printf(2, 0) __malloc 210char *kvasprintf(gfp_t gfp, const char *fmt, va_list args); 211extern __printf(2, 0) 212const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args); 213 214extern __scanf(2, 3) 215int sscanf(const char *, const char *, ...); 216extern __scanf(2, 0) 217int vsscanf(const char *, const char *, va_list); 218 219extern int no_hash_pointers_enable(char *str); 220 221extern int get_option(char **str, int *pint); 222extern char *get_options(const char *str, int nints, int *ints); 223extern unsigned long long memparse(const char *ptr, char **retptr); 224extern bool parse_option_str(const char *str, const char *option); 225extern char *next_arg(char *args, char **param, char **val); 226 227extern int core_kernel_text(unsigned long addr); 228extern int __kernel_text_address(unsigned long addr); 229extern int kernel_text_address(unsigned long addr); 230extern int func_ptr_is_kernel_text(void *ptr); 231 232extern void bust_spinlocks(int yes); 233 234extern int root_mountflags; 235 236extern bool early_boot_irqs_disabled; 237 238/* 239 * Values used for system_state. Ordering of the states must not be changed 240 * as code checks for <, <=, >, >= STATE. 241 */ 242extern enum system_states { 243 SYSTEM_BOOTING, 244 SYSTEM_SCHEDULING, 245 SYSTEM_FREEING_INITMEM, 246 SYSTEM_RUNNING, 247 SYSTEM_HALT, 248 SYSTEM_POWER_OFF, 249 SYSTEM_RESTART, 250 SYSTEM_SUSPEND, 251} system_state; 252 253extern const char hex_asc[]; 254#define hex_asc_lo(x) hex_asc[((x) & 0x0f)] 255#define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4] 256 257static inline char *hex_byte_pack(char *buf, u8 byte) 258{ 259 *buf++ = hex_asc_hi(byte); 260 *buf++ = hex_asc_lo(byte); 261 return buf; 262} 263 264extern const char hex_asc_upper[]; 265#define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)] 266#define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4] 267 268static inline char *hex_byte_pack_upper(char *buf, u8 byte) 269{ 270 *buf++ = hex_asc_upper_hi(byte); 271 *buf++ = hex_asc_upper_lo(byte); 272 return buf; 273} 274 275extern int hex_to_bin(char ch); 276extern int __must_check hex2bin(u8 *dst, const char *src, size_t count); 277extern char *bin2hex(char *dst, const void *src, size_t count); 278 279bool mac_pton(const char *s, u8 *mac); 280 281/* 282 * General tracing related utility functions - trace_printk(), 283 * tracing_on/tracing_off and tracing_start()/tracing_stop 284 * 285 * Use tracing_on/tracing_off when you want to quickly turn on or off 286 * tracing. It simply enables or disables the recording of the trace events. 287 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on 288 * file, which gives a means for the kernel and userspace to interact. 289 * Place a tracing_off() in the kernel where you want tracing to end. 290 * From user space, examine the trace, and then echo 1 > tracing_on 291 * to continue tracing. 292 * 293 * tracing_stop/tracing_start has slightly more overhead. It is used 294 * by things like suspend to ram where disabling the recording of the 295 * trace is not enough, but tracing must actually stop because things 296 * like calling smp_processor_id() may crash the system. 297 * 298 * Most likely, you want to use tracing_on/tracing_off. 299 */ 300 301enum ftrace_dump_mode { 302 DUMP_NONE, 303 DUMP_ALL, 304 DUMP_ORIG, 305}; 306 307#ifdef CONFIG_TRACING 308void tracing_on(void); 309void tracing_off(void); 310int tracing_is_on(void); 311void tracing_snapshot(void); 312void tracing_snapshot_alloc(void); 313 314extern void tracing_start(void); 315extern void tracing_stop(void); 316 317static inline __printf(1, 2) 318void ____trace_printk_check_format(const char *fmt, ...) 319{ 320} 321#define __trace_printk_check_format(fmt, args...) \ 322do { \ 323 if (0) \ 324 ____trace_printk_check_format(fmt, ##args); \ 325} while (0) 326 327/** 328 * trace_printk - printf formatting in the ftrace buffer 329 * @fmt: the printf format for printing 330 * 331 * Note: __trace_printk is an internal function for trace_printk() and 332 * the @ip is passed in via the trace_printk() macro. 333 * 334 * This function allows a kernel developer to debug fast path sections 335 * that printk is not appropriate for. By scattering in various 336 * printk like tracing in the code, a developer can quickly see 337 * where problems are occurring. 338 * 339 * This is intended as a debugging tool for the developer only. 340 * Please refrain from leaving trace_printks scattered around in 341 * your code. (Extra memory is used for special buffers that are 342 * allocated when trace_printk() is used.) 343 * 344 * A little optimization trick is done here. If there's only one 345 * argument, there's no need to scan the string for printf formats. 346 * The trace_puts() will suffice. But how can we take advantage of 347 * using trace_puts() when trace_printk() has only one argument? 348 * By stringifying the args and checking the size we can tell 349 * whether or not there are args. __stringify((__VA_ARGS__)) will 350 * turn into "()\0" with a size of 3 when there are no args, anything 351 * else will be bigger. All we need to do is define a string to this, 352 * and then take its size and compare to 3. If it's bigger, use 353 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just 354 * let gcc optimize the rest. 355 */ 356 357#define trace_printk(fmt, ...) \ 358do { \ 359 char _______STR[] = __stringify((__VA_ARGS__)); \ 360 if (sizeof(_______STR) > 3) \ 361 do_trace_printk(fmt, ##__VA_ARGS__); \ 362 else \ 363 trace_puts(fmt); \ 364} while (0) 365 366#define do_trace_printk(fmt, args...) \ 367do { \ 368 static const char *trace_printk_fmt __used \ 369 __section("__trace_printk_fmt") = \ 370 __builtin_constant_p(fmt) ? fmt : NULL; \ 371 \ 372 __trace_printk_check_format(fmt, ##args); \ 373 \ 374 if (__builtin_constant_p(fmt)) \ 375 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ 376 else \ 377 __trace_printk(_THIS_IP_, fmt, ##args); \ 378} while (0) 379 380extern __printf(2, 3) 381int __trace_bprintk(unsigned long ip, const char *fmt, ...); 382 383extern __printf(2, 3) 384int __trace_printk(unsigned long ip, const char *fmt, ...); 385 386/** 387 * trace_puts - write a string into the ftrace buffer 388 * @str: the string to record 389 * 390 * Note: __trace_bputs is an internal function for trace_puts and 391 * the @ip is passed in via the trace_puts macro. 392 * 393 * This is similar to trace_printk() but is made for those really fast 394 * paths that a developer wants the least amount of "Heisenbug" effects, 395 * where the processing of the print format is still too much. 396 * 397 * This function allows a kernel developer to debug fast path sections 398 * that printk is not appropriate for. By scattering in various 399 * printk like tracing in the code, a developer can quickly see 400 * where problems are occurring. 401 * 402 * This is intended as a debugging tool for the developer only. 403 * Please refrain from leaving trace_puts scattered around in 404 * your code. (Extra memory is used for special buffers that are 405 * allocated when trace_puts() is used.) 406 * 407 * Returns: 0 if nothing was written, positive # if string was. 408 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used) 409 */ 410 411#define trace_puts(str) ({ \ 412 static const char *trace_printk_fmt __used \ 413 __section("__trace_printk_fmt") = \ 414 __builtin_constant_p(str) ? str : NULL; \ 415 \ 416 if (__builtin_constant_p(str)) \ 417 __trace_bputs(_THIS_IP_, trace_printk_fmt); \ 418 else \ 419 __trace_puts(_THIS_IP_, str, strlen(str)); \ 420}) 421extern int __trace_bputs(unsigned long ip, const char *str); 422extern int __trace_puts(unsigned long ip, const char *str, int size); 423 424extern void trace_dump_stack(int skip); 425 426/* 427 * The double __builtin_constant_p is because gcc will give us an error 428 * if we try to allocate the static variable to fmt if it is not a 429 * constant. Even with the outer if statement. 430 */ 431#define ftrace_vprintk(fmt, vargs) \ 432do { \ 433 if (__builtin_constant_p(fmt)) { \ 434 static const char *trace_printk_fmt __used \ 435 __section("__trace_printk_fmt") = \ 436 __builtin_constant_p(fmt) ? fmt : NULL; \ 437 \ 438 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ 439 } else \ 440 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ 441} while (0) 442 443extern __printf(2, 0) int 444__ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); 445 446extern __printf(2, 0) int 447__ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); 448 449extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode); 450#else 451static inline void tracing_start(void) { } 452static inline void tracing_stop(void) { } 453static inline void trace_dump_stack(int skip) { } 454 455static inline void tracing_on(void) { } 456static inline void tracing_off(void) { } 457static inline int tracing_is_on(void) { return 0; } 458static inline void tracing_snapshot(void) { } 459static inline void tracing_snapshot_alloc(void) { } 460 461static inline __printf(1, 2) 462int trace_printk(const char *fmt, ...) 463{ 464 return 0; 465} 466static __printf(1, 0) inline int 467ftrace_vprintk(const char *fmt, va_list ap) 468{ 469 return 0; 470} 471static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { } 472#endif /* CONFIG_TRACING */ 473 474/* This counts to 12. Any more, it will return 13th argument. */ 475#define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n 476#define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0) 477 478#define __CONCAT(a, b) a ## b 479#define CONCATENATE(a, b) __CONCAT(a, b) 480 481/* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ 482#ifdef CONFIG_FTRACE_MCOUNT_RECORD 483# define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD 484#endif 485 486/* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */ 487#define VERIFY_OCTAL_PERMISSIONS(perms) \ 488 (BUILD_BUG_ON_ZERO((perms) < 0) + \ 489 BUILD_BUG_ON_ZERO((perms) > 0777) + \ 490 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \ 491 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \ 492 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \ 493 /* USER_WRITABLE >= GROUP_WRITABLE */ \ 494 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \ 495 /* OTHER_WRITABLE? Generally considered a bad idea. */ \ 496 BUILD_BUG_ON_ZERO((perms) & 2) + \ 497 (perms)) 498#endif