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/bitops.h> 13 #include <linux/kstrtox.h> 14 #include <linux/log2.h> 15 #include <linux/math.h> 16 #include <linux/minmax.h> 17 #include <linux/typecheck.h> 18 #include <linux/panic.h> 19 #include <linux/printk.h> 20 #include <linux/build_bug.h> 21 #include <linux/static_call_types.h> 22 #include <asm/byteorder.h> 23 24 #include <uapi/linux/kernel.h> 25 26 #define STACK_MAGIC 0xdeadbeef 27 28 /** 29 * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value 30 * @x: value to repeat 31 * 32 * NOTE: @x is not checked for > 0xff; larger values produce odd results. 33 */ 34 #define REPEAT_BYTE(x) ((~0ul / 0xff) * (x)) 35 36 /* generic data direction definitions */ 37 #define READ 0 38 #define WRITE 1 39 40 /** 41 * ARRAY_SIZE - get the number of elements in array @arr 42 * @arr: array to be sized 43 */ 44 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr)) 45 46 #define PTR_IF(cond, ptr) ((cond) ? (ptr) : NULL) 47 48 #define u64_to_user_ptr(x) ( \ 49 { \ 50 typecheck(u64, (x)); \ 51 (void __user *)(uintptr_t)(x); \ 52 } \ 53 ) 54 55 #define typeof_member(T, m) typeof(((T*)0)->m) 56 57 #define _RET_IP_ (unsigned long)__builtin_return_address(0) 58 #define _THIS_IP_ ({ __label__ __here; __here: (unsigned long)&&__here; }) 59 60 /** 61 * upper_32_bits - return bits 32-63 of a number 62 * @n: the number we're accessing 63 * 64 * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress 65 * the "right shift count >= width of type" warning when that quantity is 66 * 32-bits. 67 */ 68 #define upper_32_bits(n) ((u32)(((n) >> 16) >> 16)) 69 70 /** 71 * lower_32_bits - return bits 0-31 of a number 72 * @n: the number we're accessing 73 */ 74 #define lower_32_bits(n) ((u32)((n) & 0xffffffff)) 75 76 /** 77 * upper_16_bits - return bits 16-31 of a number 78 * @n: the number we're accessing 79 */ 80 #define upper_16_bits(n) ((u16)((n) >> 16)) 81 82 /** 83 * lower_16_bits - return bits 0-15 of a number 84 * @n: the number we're accessing 85 */ 86 #define lower_16_bits(n) ((u16)((n) & 0xffff)) 87 88 struct completion; 89 struct user; 90 91 #ifdef CONFIG_PREEMPT_VOLUNTARY 92 93 extern int __cond_resched(void); 94 # define might_resched() __cond_resched() 95 96 #elif defined(CONFIG_PREEMPT_DYNAMIC) 97 98 extern int __cond_resched(void); 99 100 DECLARE_STATIC_CALL(might_resched, __cond_resched); 101 102 static __always_inline void might_resched(void) 103 { 104 static_call_mod(might_resched)(); 105 } 106 107 #else 108 109 # define might_resched() do { } while (0) 110 111 #endif /* CONFIG_PREEMPT_* */ 112 113 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP 114 extern void ___might_sleep(const char *file, int line, int preempt_offset); 115 extern void __might_sleep(const char *file, int line, int preempt_offset); 116 extern void __cant_sleep(const char *file, int line, int preempt_offset); 117 extern void __cant_migrate(const char *file, int line); 118 119 /** 120 * might_sleep - annotation for functions that can sleep 121 * 122 * this macro will print a stack trace if it is executed in an atomic 123 * context (spinlock, irq-handler, ...). Additional sections where blocking is 124 * not allowed can be annotated with non_block_start() and non_block_end() 125 * pairs. 126 * 127 * This is a useful debugging help to be able to catch problems early and not 128 * be bitten later when the calling function happens to sleep when it is not 129 * supposed to. 130 */ 131 # define might_sleep() \ 132 do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0) 133 /** 134 * cant_sleep - annotation for functions that cannot sleep 135 * 136 * this macro will print a stack trace if it is executed with preemption enabled 137 */ 138 # define cant_sleep() \ 139 do { __cant_sleep(__FILE__, __LINE__, 0); } while (0) 140 # define sched_annotate_sleep() (current->task_state_change = 0) 141 142 /** 143 * cant_migrate - annotation for functions that cannot migrate 144 * 145 * Will print a stack trace if executed in code which is migratable 146 */ 147 # define cant_migrate() \ 148 do { \ 149 if (IS_ENABLED(CONFIG_SMP)) \ 150 __cant_migrate(__FILE__, __LINE__); \ 151 } while (0) 152 153 /** 154 * non_block_start - annotate the start of section where sleeping is prohibited 155 * 156 * This is on behalf of the oom reaper, specifically when it is calling the mmu 157 * notifiers. The problem is that if the notifier were to block on, for example, 158 * mutex_lock() and if the process which holds that mutex were to perform a 159 * sleeping memory allocation, the oom reaper is now blocked on completion of 160 * that memory allocation. Other blocking calls like wait_event() pose similar 161 * issues. 162 */ 163 # define non_block_start() (current->non_block_count++) 164 /** 165 * non_block_end - annotate the end of section where sleeping is prohibited 166 * 167 * Closes a section opened by non_block_start(). 168 */ 169 # define non_block_end() WARN_ON(current->non_block_count-- == 0) 170 #else 171 static inline void ___might_sleep(const char *file, int line, 172 int preempt_offset) { } 173 static inline void __might_sleep(const char *file, int line, 174 int preempt_offset) { } 175 # define might_sleep() do { might_resched(); } while (0) 176 # define cant_sleep() do { } while (0) 177 # define cant_migrate() do { } while (0) 178 # define sched_annotate_sleep() do { } while (0) 179 # define non_block_start() do { } while (0) 180 # define non_block_end() do { } while (0) 181 #endif 182 183 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0) 184 185 #if defined(CONFIG_MMU) && \ 186 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)) 187 #define might_fault() __might_fault(__FILE__, __LINE__) 188 void __might_fault(const char *file, int line); 189 #else 190 static inline void might_fault(void) { } 191 #endif 192 193 void do_exit(long error_code) __noreturn; 194 void complete_and_exit(struct completion *, long) __noreturn; 195 196 extern int num_to_str(char *buf, int size, 197 unsigned long long num, unsigned int width); 198 199 /* lib/printf utilities */ 200 201 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...); 202 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list); 203 extern __printf(3, 4) 204 int snprintf(char *buf, size_t size, const char *fmt, ...); 205 extern __printf(3, 0) 206 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args); 207 extern __printf(3, 4) 208 int scnprintf(char *buf, size_t size, const char *fmt, ...); 209 extern __printf(3, 0) 210 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args); 211 extern __printf(2, 3) __malloc 212 char *kasprintf(gfp_t gfp, const char *fmt, ...); 213 extern __printf(2, 0) __malloc 214 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args); 215 extern __printf(2, 0) 216 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args); 217 218 extern __scanf(2, 3) 219 int sscanf(const char *, const char *, ...); 220 extern __scanf(2, 0) 221 int vsscanf(const char *, const char *, va_list); 222 223 extern int no_hash_pointers_enable(char *str); 224 225 extern int get_option(char **str, int *pint); 226 extern char *get_options(const char *str, int nints, int *ints); 227 extern unsigned long long memparse(const char *ptr, char **retptr); 228 extern bool parse_option_str(const char *str, const char *option); 229 extern char *next_arg(char *args, char **param, char **val); 230 231 extern int core_kernel_text(unsigned long addr); 232 extern int init_kernel_text(unsigned long addr); 233 extern int core_kernel_data(unsigned long addr); 234 extern int __kernel_text_address(unsigned long addr); 235 extern int kernel_text_address(unsigned long addr); 236 extern int func_ptr_is_kernel_text(void *ptr); 237 238 extern void bust_spinlocks(int yes); 239 240 extern int root_mountflags; 241 242 extern bool early_boot_irqs_disabled; 243 244 /* 245 * Values used for system_state. Ordering of the states must not be changed 246 * as code checks for <, <=, >, >= STATE. 247 */ 248 extern enum system_states { 249 SYSTEM_BOOTING, 250 SYSTEM_SCHEDULING, 251 SYSTEM_FREEING_INITMEM, 252 SYSTEM_RUNNING, 253 SYSTEM_HALT, 254 SYSTEM_POWER_OFF, 255 SYSTEM_RESTART, 256 SYSTEM_SUSPEND, 257 } system_state; 258 259 extern const char hex_asc[]; 260 #define hex_asc_lo(x) hex_asc[((x) & 0x0f)] 261 #define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4] 262 263 static inline char *hex_byte_pack(char *buf, u8 byte) 264 { 265 *buf++ = hex_asc_hi(byte); 266 *buf++ = hex_asc_lo(byte); 267 return buf; 268 } 269 270 extern const char hex_asc_upper[]; 271 #define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)] 272 #define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4] 273 274 static inline char *hex_byte_pack_upper(char *buf, u8 byte) 275 { 276 *buf++ = hex_asc_upper_hi(byte); 277 *buf++ = hex_asc_upper_lo(byte); 278 return buf; 279 } 280 281 extern int hex_to_bin(char ch); 282 extern int __must_check hex2bin(u8 *dst, const char *src, size_t count); 283 extern char *bin2hex(char *dst, const void *src, size_t count); 284 285 bool mac_pton(const char *s, u8 *mac); 286 287 /* 288 * General tracing related utility functions - trace_printk(), 289 * tracing_on/tracing_off and tracing_start()/tracing_stop 290 * 291 * Use tracing_on/tracing_off when you want to quickly turn on or off 292 * tracing. It simply enables or disables the recording of the trace events. 293 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on 294 * file, which gives a means for the kernel and userspace to interact. 295 * Place a tracing_off() in the kernel where you want tracing to end. 296 * From user space, examine the trace, and then echo 1 > tracing_on 297 * to continue tracing. 298 * 299 * tracing_stop/tracing_start has slightly more overhead. It is used 300 * by things like suspend to ram where disabling the recording of the 301 * trace is not enough, but tracing must actually stop because things 302 * like calling smp_processor_id() may crash the system. 303 * 304 * Most likely, you want to use tracing_on/tracing_off. 305 */ 306 307 enum ftrace_dump_mode { 308 DUMP_NONE, 309 DUMP_ALL, 310 DUMP_ORIG, 311 }; 312 313 #ifdef CONFIG_TRACING 314 void tracing_on(void); 315 void tracing_off(void); 316 int tracing_is_on(void); 317 void tracing_snapshot(void); 318 void tracing_snapshot_alloc(void); 319 320 extern void tracing_start(void); 321 extern void tracing_stop(void); 322 323 static inline __printf(1, 2) 324 void ____trace_printk_check_format(const char *fmt, ...) 325 { 326 } 327 #define __trace_printk_check_format(fmt, args...) \ 328 do { \ 329 if (0) \ 330 ____trace_printk_check_format(fmt, ##args); \ 331 } while (0) 332 333 /** 334 * trace_printk - printf formatting in the ftrace buffer 335 * @fmt: the printf format for printing 336 * 337 * Note: __trace_printk is an internal function for trace_printk() and 338 * the @ip is passed in via the trace_printk() macro. 339 * 340 * This function allows a kernel developer to debug fast path sections 341 * that printk is not appropriate for. By scattering in various 342 * printk like tracing in the code, a developer can quickly see 343 * where problems are occurring. 344 * 345 * This is intended as a debugging tool for the developer only. 346 * Please refrain from leaving trace_printks scattered around in 347 * your code. (Extra memory is used for special buffers that are 348 * allocated when trace_printk() is used.) 349 * 350 * A little optimization trick is done here. If there's only one 351 * argument, there's no need to scan the string for printf formats. 352 * The trace_puts() will suffice. But how can we take advantage of 353 * using trace_puts() when trace_printk() has only one argument? 354 * By stringifying the args and checking the size we can tell 355 * whether or not there are args. __stringify((__VA_ARGS__)) will 356 * turn into "()\0" with a size of 3 when there are no args, anything 357 * else will be bigger. All we need to do is define a string to this, 358 * and then take its size and compare to 3. If it's bigger, use 359 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just 360 * let gcc optimize the rest. 361 */ 362 363 #define trace_printk(fmt, ...) \ 364 do { \ 365 char _______STR[] = __stringify((__VA_ARGS__)); \ 366 if (sizeof(_______STR) > 3) \ 367 do_trace_printk(fmt, ##__VA_ARGS__); \ 368 else \ 369 trace_puts(fmt); \ 370 } while (0) 371 372 #define do_trace_printk(fmt, args...) \ 373 do { \ 374 static const char *trace_printk_fmt __used \ 375 __section("__trace_printk_fmt") = \ 376 __builtin_constant_p(fmt) ? fmt : NULL; \ 377 \ 378 __trace_printk_check_format(fmt, ##args); \ 379 \ 380 if (__builtin_constant_p(fmt)) \ 381 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ 382 else \ 383 __trace_printk(_THIS_IP_, fmt, ##args); \ 384 } while (0) 385 386 extern __printf(2, 3) 387 int __trace_bprintk(unsigned long ip, const char *fmt, ...); 388 389 extern __printf(2, 3) 390 int __trace_printk(unsigned long ip, const char *fmt, ...); 391 392 /** 393 * trace_puts - write a string into the ftrace buffer 394 * @str: the string to record 395 * 396 * Note: __trace_bputs is an internal function for trace_puts and 397 * the @ip is passed in via the trace_puts macro. 398 * 399 * This is similar to trace_printk() but is made for those really fast 400 * paths that a developer wants the least amount of "Heisenbug" effects, 401 * where the processing of the print format is still too much. 402 * 403 * This function allows a kernel developer to debug fast path sections 404 * that printk is not appropriate for. By scattering in various 405 * printk like tracing in the code, a developer can quickly see 406 * where problems are occurring. 407 * 408 * This is intended as a debugging tool for the developer only. 409 * Please refrain from leaving trace_puts scattered around in 410 * your code. (Extra memory is used for special buffers that are 411 * allocated when trace_puts() is used.) 412 * 413 * Returns: 0 if nothing was written, positive # if string was. 414 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used) 415 */ 416 417 #define trace_puts(str) ({ \ 418 static const char *trace_printk_fmt __used \ 419 __section("__trace_printk_fmt") = \ 420 __builtin_constant_p(str) ? str : NULL; \ 421 \ 422 if (__builtin_constant_p(str)) \ 423 __trace_bputs(_THIS_IP_, trace_printk_fmt); \ 424 else \ 425 __trace_puts(_THIS_IP_, str, strlen(str)); \ 426 }) 427 extern int __trace_bputs(unsigned long ip, const char *str); 428 extern int __trace_puts(unsigned long ip, const char *str, int size); 429 430 extern void trace_dump_stack(int skip); 431 432 /* 433 * The double __builtin_constant_p is because gcc will give us an error 434 * if we try to allocate the static variable to fmt if it is not a 435 * constant. Even with the outer if statement. 436 */ 437 #define ftrace_vprintk(fmt, vargs) \ 438 do { \ 439 if (__builtin_constant_p(fmt)) { \ 440 static const char *trace_printk_fmt __used \ 441 __section("__trace_printk_fmt") = \ 442 __builtin_constant_p(fmt) ? fmt : NULL; \ 443 \ 444 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ 445 } else \ 446 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ 447 } while (0) 448 449 extern __printf(2, 0) int 450 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); 451 452 extern __printf(2, 0) int 453 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); 454 455 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode); 456 #else 457 static inline void tracing_start(void) { } 458 static inline void tracing_stop(void) { } 459 static inline void trace_dump_stack(int skip) { } 460 461 static inline void tracing_on(void) { } 462 static inline void tracing_off(void) { } 463 static inline int tracing_is_on(void) { return 0; } 464 static inline void tracing_snapshot(void) { } 465 static inline void tracing_snapshot_alloc(void) { } 466 467 static inline __printf(1, 2) 468 int trace_printk(const char *fmt, ...) 469 { 470 return 0; 471 } 472 static __printf(1, 0) inline int 473 ftrace_vprintk(const char *fmt, va_list ap) 474 { 475 return 0; 476 } 477 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { } 478 #endif /* CONFIG_TRACING */ 479 480 /* This counts to 12. Any more, it will return 13th argument. */ 481 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n 482 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0) 483 484 #define __CONCAT(a, b) a ## b 485 #define CONCATENATE(a, b) __CONCAT(a, b) 486 487 /** 488 * container_of - cast a member of a structure out to the containing structure 489 * @ptr: the pointer to the member. 490 * @type: the type of the container struct this is embedded in. 491 * @member: the name of the member within the struct. 492 * 493 */ 494 #define container_of(ptr, type, member) ({ \ 495 void *__mptr = (void *)(ptr); \ 496 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 497 !__same_type(*(ptr), void), \ 498 "pointer type mismatch in container_of()"); \ 499 ((type *)(__mptr - offsetof(type, member))); }) 500 501 /** 502 * container_of_safe - cast a member of a structure out to the containing structure 503 * @ptr: the pointer to the member. 504 * @type: the type of the container struct this is embedded in. 505 * @member: the name of the member within the struct. 506 * 507 * If IS_ERR_OR_NULL(ptr), ptr is returned unchanged. 508 */ 509 #define container_of_safe(ptr, type, member) ({ \ 510 void *__mptr = (void *)(ptr); \ 511 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 512 !__same_type(*(ptr), void), \ 513 "pointer type mismatch in container_of()"); \ 514 IS_ERR_OR_NULL(__mptr) ? ERR_CAST(__mptr) : \ 515 ((type *)(__mptr - offsetof(type, member))); }) 516 517 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ 518 #ifdef CONFIG_FTRACE_MCOUNT_RECORD 519 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD 520 #endif 521 522 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */ 523 #define VERIFY_OCTAL_PERMISSIONS(perms) \ 524 (BUILD_BUG_ON_ZERO((perms) < 0) + \ 525 BUILD_BUG_ON_ZERO((perms) > 0777) + \ 526 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \ 527 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \ 528 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \ 529 /* USER_WRITABLE >= GROUP_WRITABLE */ \ 530 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \ 531 /* OTHER_WRITABLE? Generally considered a bad idea. */ \ 532 BUILD_BUG_ON_ZERO((perms) & 2) + \ 533 (perms)) 534 #endif 535