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_resched(const char *file, int line, unsigned int offsets); 115 extern void __might_sleep(const char *file, int line); 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__); 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_resched(const char *file, int line, 172 unsigned int offsets) { } 173 static inline void __might_sleep(const char *file, int line) { } 174 # define might_sleep() do { might_resched(); } while (0) 175 # define cant_sleep() do { } while (0) 176 # define cant_migrate() do { } while (0) 177 # define sched_annotate_sleep() do { } while (0) 178 # define non_block_start() do { } while (0) 179 # define non_block_end() do { } while (0) 180 #endif 181 182 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0) 183 184 #if defined(CONFIG_MMU) && \ 185 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)) 186 #define might_fault() __might_fault(__FILE__, __LINE__) 187 void __might_fault(const char *file, int line); 188 #else 189 static inline void might_fault(void) { } 190 #endif 191 192 void do_exit(long error_code) __noreturn; 193 void complete_and_exit(struct completion *, long) __noreturn; 194 195 extern int num_to_str(char *buf, int size, 196 unsigned long long num, unsigned int width); 197 198 /* lib/printf utilities */ 199 200 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...); 201 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list); 202 extern __printf(3, 4) 203 int snprintf(char *buf, size_t size, const char *fmt, ...); 204 extern __printf(3, 0) 205 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args); 206 extern __printf(3, 4) 207 int scnprintf(char *buf, size_t size, const char *fmt, ...); 208 extern __printf(3, 0) 209 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args); 210 extern __printf(2, 3) __malloc 211 char *kasprintf(gfp_t gfp, const char *fmt, ...); 212 extern __printf(2, 0) __malloc 213 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args); 214 extern __printf(2, 0) 215 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args); 216 217 extern __scanf(2, 3) 218 int sscanf(const char *, const char *, ...); 219 extern __scanf(2, 0) 220 int vsscanf(const char *, const char *, va_list); 221 222 extern int no_hash_pointers_enable(char *str); 223 224 extern int get_option(char **str, int *pint); 225 extern char *get_options(const char *str, int nints, int *ints); 226 extern unsigned long long memparse(const char *ptr, char **retptr); 227 extern bool parse_option_str(const char *str, const char *option); 228 extern char *next_arg(char *args, char **param, char **val); 229 230 extern int core_kernel_text(unsigned long addr); 231 extern int init_kernel_text(unsigned long addr); 232 extern int core_kernel_data(unsigned long addr); 233 extern int __kernel_text_address(unsigned long addr); 234 extern int kernel_text_address(unsigned long addr); 235 extern int func_ptr_is_kernel_text(void *ptr); 236 237 extern void bust_spinlocks(int yes); 238 239 extern int root_mountflags; 240 241 extern bool early_boot_irqs_disabled; 242 243 /* 244 * Values used for system_state. Ordering of the states must not be changed 245 * as code checks for <, <=, >, >= STATE. 246 */ 247 extern enum system_states { 248 SYSTEM_BOOTING, 249 SYSTEM_SCHEDULING, 250 SYSTEM_RUNNING, 251 SYSTEM_HALT, 252 SYSTEM_POWER_OFF, 253 SYSTEM_RESTART, 254 SYSTEM_SUSPEND, 255 } system_state; 256 257 extern const char hex_asc[]; 258 #define hex_asc_lo(x) hex_asc[((x) & 0x0f)] 259 #define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4] 260 261 static inline char *hex_byte_pack(char *buf, u8 byte) 262 { 263 *buf++ = hex_asc_hi(byte); 264 *buf++ = hex_asc_lo(byte); 265 return buf; 266 } 267 268 extern const char hex_asc_upper[]; 269 #define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)] 270 #define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4] 271 272 static inline char *hex_byte_pack_upper(char *buf, u8 byte) 273 { 274 *buf++ = hex_asc_upper_hi(byte); 275 *buf++ = hex_asc_upper_lo(byte); 276 return buf; 277 } 278 279 extern int hex_to_bin(char ch); 280 extern int __must_check hex2bin(u8 *dst, const char *src, size_t count); 281 extern char *bin2hex(char *dst, const void *src, size_t count); 282 283 bool mac_pton(const char *s, u8 *mac); 284 285 /* 286 * General tracing related utility functions - trace_printk(), 287 * tracing_on/tracing_off and tracing_start()/tracing_stop 288 * 289 * Use tracing_on/tracing_off when you want to quickly turn on or off 290 * tracing. It simply enables or disables the recording of the trace events. 291 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on 292 * file, which gives a means for the kernel and userspace to interact. 293 * Place a tracing_off() in the kernel where you want tracing to end. 294 * From user space, examine the trace, and then echo 1 > tracing_on 295 * to continue tracing. 296 * 297 * tracing_stop/tracing_start has slightly more overhead. It is used 298 * by things like suspend to ram where disabling the recording of the 299 * trace is not enough, but tracing must actually stop because things 300 * like calling smp_processor_id() may crash the system. 301 * 302 * Most likely, you want to use tracing_on/tracing_off. 303 */ 304 305 enum ftrace_dump_mode { 306 DUMP_NONE, 307 DUMP_ALL, 308 DUMP_ORIG, 309 }; 310 311 #ifdef CONFIG_TRACING 312 void tracing_on(void); 313 void tracing_off(void); 314 int tracing_is_on(void); 315 void tracing_snapshot(void); 316 void tracing_snapshot_alloc(void); 317 318 extern void tracing_start(void); 319 extern void tracing_stop(void); 320 321 static inline __printf(1, 2) 322 void ____trace_printk_check_format(const char *fmt, ...) 323 { 324 } 325 #define __trace_printk_check_format(fmt, args...) \ 326 do { \ 327 if (0) \ 328 ____trace_printk_check_format(fmt, ##args); \ 329 } while (0) 330 331 /** 332 * trace_printk - printf formatting in the ftrace buffer 333 * @fmt: the printf format for printing 334 * 335 * Note: __trace_printk is an internal function for trace_printk() and 336 * the @ip is passed in via the trace_printk() macro. 337 * 338 * This function allows a kernel developer to debug fast path sections 339 * that printk is not appropriate for. By scattering in various 340 * printk like tracing in the code, a developer can quickly see 341 * where problems are occurring. 342 * 343 * This is intended as a debugging tool for the developer only. 344 * Please refrain from leaving trace_printks scattered around in 345 * your code. (Extra memory is used for special buffers that are 346 * allocated when trace_printk() is used.) 347 * 348 * A little optimization trick is done here. If there's only one 349 * argument, there's no need to scan the string for printf formats. 350 * The trace_puts() will suffice. But how can we take advantage of 351 * using trace_puts() when trace_printk() has only one argument? 352 * By stringifying the args and checking the size we can tell 353 * whether or not there are args. __stringify((__VA_ARGS__)) will 354 * turn into "()\0" with a size of 3 when there are no args, anything 355 * else will be bigger. All we need to do is define a string to this, 356 * and then take its size and compare to 3. If it's bigger, use 357 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just 358 * let gcc optimize the rest. 359 */ 360 361 #define trace_printk(fmt, ...) \ 362 do { \ 363 char _______STR[] = __stringify((__VA_ARGS__)); \ 364 if (sizeof(_______STR) > 3) \ 365 do_trace_printk(fmt, ##__VA_ARGS__); \ 366 else \ 367 trace_puts(fmt); \ 368 } while (0) 369 370 #define do_trace_printk(fmt, args...) \ 371 do { \ 372 static const char *trace_printk_fmt __used \ 373 __section("__trace_printk_fmt") = \ 374 __builtin_constant_p(fmt) ? fmt : NULL; \ 375 \ 376 __trace_printk_check_format(fmt, ##args); \ 377 \ 378 if (__builtin_constant_p(fmt)) \ 379 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ 380 else \ 381 __trace_printk(_THIS_IP_, fmt, ##args); \ 382 } while (0) 383 384 extern __printf(2, 3) 385 int __trace_bprintk(unsigned long ip, const char *fmt, ...); 386 387 extern __printf(2, 3) 388 int __trace_printk(unsigned long ip, const char *fmt, ...); 389 390 /** 391 * trace_puts - write a string into the ftrace buffer 392 * @str: the string to record 393 * 394 * Note: __trace_bputs is an internal function for trace_puts and 395 * the @ip is passed in via the trace_puts macro. 396 * 397 * This is similar to trace_printk() but is made for those really fast 398 * paths that a developer wants the least amount of "Heisenbug" effects, 399 * where the processing of the print format is still too much. 400 * 401 * This function allows a kernel developer to debug fast path sections 402 * that printk is not appropriate for. By scattering in various 403 * printk like tracing in the code, a developer can quickly see 404 * where problems are occurring. 405 * 406 * This is intended as a debugging tool for the developer only. 407 * Please refrain from leaving trace_puts scattered around in 408 * your code. (Extra memory is used for special buffers that are 409 * allocated when trace_puts() is used.) 410 * 411 * Returns: 0 if nothing was written, positive # if string was. 412 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used) 413 */ 414 415 #define trace_puts(str) ({ \ 416 static const char *trace_printk_fmt __used \ 417 __section("__trace_printk_fmt") = \ 418 __builtin_constant_p(str) ? str : NULL; \ 419 \ 420 if (__builtin_constant_p(str)) \ 421 __trace_bputs(_THIS_IP_, trace_printk_fmt); \ 422 else \ 423 __trace_puts(_THIS_IP_, str, strlen(str)); \ 424 }) 425 extern int __trace_bputs(unsigned long ip, const char *str); 426 extern int __trace_puts(unsigned long ip, const char *str, int size); 427 428 extern void trace_dump_stack(int skip); 429 430 /* 431 * The double __builtin_constant_p is because gcc will give us an error 432 * if we try to allocate the static variable to fmt if it is not a 433 * constant. Even with the outer if statement. 434 */ 435 #define ftrace_vprintk(fmt, vargs) \ 436 do { \ 437 if (__builtin_constant_p(fmt)) { \ 438 static const char *trace_printk_fmt __used \ 439 __section("__trace_printk_fmt") = \ 440 __builtin_constant_p(fmt) ? fmt : NULL; \ 441 \ 442 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ 443 } else \ 444 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ 445 } while (0) 446 447 extern __printf(2, 0) int 448 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); 449 450 extern __printf(2, 0) int 451 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); 452 453 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode); 454 #else 455 static inline void tracing_start(void) { } 456 static inline void tracing_stop(void) { } 457 static inline void trace_dump_stack(int skip) { } 458 459 static inline void tracing_on(void) { } 460 static inline void tracing_off(void) { } 461 static inline int tracing_is_on(void) { return 0; } 462 static inline void tracing_snapshot(void) { } 463 static inline void tracing_snapshot_alloc(void) { } 464 465 static inline __printf(1, 2) 466 int trace_printk(const char *fmt, ...) 467 { 468 return 0; 469 } 470 static __printf(1, 0) inline int 471 ftrace_vprintk(const char *fmt, va_list ap) 472 { 473 return 0; 474 } 475 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { } 476 #endif /* CONFIG_TRACING */ 477 478 /* This counts to 12. Any more, it will return 13th argument. */ 479 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n 480 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0) 481 482 #define __CONCAT(a, b) a ## b 483 #define CONCATENATE(a, b) __CONCAT(a, b) 484 485 /** 486 * container_of - cast a member of a structure out to the containing structure 487 * @ptr: the pointer to the member. 488 * @type: the type of the container struct this is embedded in. 489 * @member: the name of the member within the struct. 490 * 491 */ 492 #define container_of(ptr, type, member) ({ \ 493 void *__mptr = (void *)(ptr); \ 494 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 495 !__same_type(*(ptr), void), \ 496 "pointer type mismatch in container_of()"); \ 497 ((type *)(__mptr - offsetof(type, member))); }) 498 499 /** 500 * container_of_safe - cast a member of a structure out to the containing structure 501 * @ptr: the pointer to the member. 502 * @type: the type of the container struct this is embedded in. 503 * @member: the name of the member within the struct. 504 * 505 * If IS_ERR_OR_NULL(ptr), ptr is returned unchanged. 506 */ 507 #define container_of_safe(ptr, type, member) ({ \ 508 void *__mptr = (void *)(ptr); \ 509 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 510 !__same_type(*(ptr), void), \ 511 "pointer type mismatch in container_of()"); \ 512 IS_ERR_OR_NULL(__mptr) ? ERR_CAST(__mptr) : \ 513 ((type *)(__mptr - offsetof(type, member))); }) 514 515 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ 516 #ifdef CONFIG_FTRACE_MCOUNT_RECORD 517 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD 518 #endif 519 520 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */ 521 #define VERIFY_OCTAL_PERMISSIONS(perms) \ 522 (BUILD_BUG_ON_ZERO((perms) < 0) + \ 523 BUILD_BUG_ON_ZERO((perms) > 0777) + \ 524 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \ 525 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \ 526 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \ 527 /* USER_WRITABLE >= GROUP_WRITABLE */ \ 528 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \ 529 /* OTHER_WRITABLE? Generally considered a bad idea. */ \ 530 BUILD_BUG_ON_ZERO((perms) & 2) + \ 531 (perms)) 532 #endif 533