1 /* SPDX-License-Identifier: GPL-2.0 */ 2 /* 3 * NOTE: 4 * 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. 10 */ 11 #ifndef _LINUX_KERNEL_H 12 #define _LINUX_KERNEL_H 13 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> 34 35 #include <uapi/linux/kernel.h> 36 37 #define STACK_MAGIC 0xdeadbeef 38 39 /** 40 * REPEAT_BYTE - repeat the value @x multiple times as an unsigned long value 41 * @x: value to repeat 42 * 43 * NOTE: @x is not checked for > 0xff; larger values produce odd results. 44 */ 45 #define REPEAT_BYTE(x) ((~0ul / 0xff) * (x)) 46 47 /* generic data direction definitions */ 48 #define READ 0 49 #define WRITE 1 50 51 /** 52 * ARRAY_SIZE - get the number of elements in array @arr 53 * @arr: array to be sized 54 */ 55 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]) + __must_be_array(arr)) 56 57 #define PTR_IF(cond, ptr) ((cond) ? (ptr) : NULL) 58 59 #define u64_to_user_ptr(x) ( \ 60 { \ 61 typecheck(u64, (x)); \ 62 (void __user *)(uintptr_t)(x); \ 63 } \ 64 ) 65 66 /** 67 * upper_32_bits - return bits 32-63 of a number 68 * @n: the number we're accessing 69 * 70 * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress 71 * the "right shift count >= width of type" warning when that quantity is 72 * 32-bits. 73 */ 74 #define upper_32_bits(n) ((u32)(((n) >> 16) >> 16)) 75 76 /** 77 * lower_32_bits - return bits 0-31 of a number 78 * @n: the number we're accessing 79 */ 80 #define lower_32_bits(n) ((u32)((n) & 0xffffffff)) 81 82 /** 83 * upper_16_bits - return bits 16-31 of a number 84 * @n: the number we're accessing 85 */ 86 #define upper_16_bits(n) ((u16)((n) >> 16)) 87 88 /** 89 * lower_16_bits - return bits 0-15 of a number 90 * @n: the number we're accessing 91 */ 92 #define lower_16_bits(n) ((u16)((n) & 0xffff)) 93 94 struct completion; 95 struct user; 96 97 #ifdef CONFIG_PREEMPT_VOLUNTARY_BUILD 98 99 extern int __cond_resched(void); 100 # define might_resched() __cond_resched() 101 102 #elif defined(CONFIG_PREEMPT_DYNAMIC) 103 104 extern int __cond_resched(void); 105 106 DECLARE_STATIC_CALL(might_resched, __cond_resched); 107 108 static __always_inline void might_resched(void) 109 { 110 static_call_mod(might_resched)(); 111 } 112 113 #else 114 115 # define might_resched() do { } while (0) 116 117 #endif /* CONFIG_PREEMPT_* */ 118 119 #ifdef CONFIG_DEBUG_ATOMIC_SLEEP 120 extern void __might_resched(const char *file, int line, unsigned int offsets); 121 extern void __might_sleep(const char *file, int line); 122 extern void __cant_sleep(const char *file, int line, int preempt_offset); 123 extern void __cant_migrate(const char *file, int line); 124 125 /** 126 * might_sleep - annotation for functions that can sleep 127 * 128 * this macro will print a stack trace if it is executed in an atomic 129 * context (spinlock, irq-handler, ...). Additional sections where blocking is 130 * not allowed can be annotated with non_block_start() and non_block_end() 131 * pairs. 132 * 133 * This is a useful debugging help to be able to catch problems early and not 134 * be bitten later when the calling function happens to sleep when it is not 135 * supposed to. 136 */ 137 # define might_sleep() \ 138 do { __might_sleep(__FILE__, __LINE__); might_resched(); } while (0) 139 /** 140 * cant_sleep - annotation for functions that cannot sleep 141 * 142 * this macro will print a stack trace if it is executed with preemption enabled 143 */ 144 # define cant_sleep() \ 145 do { __cant_sleep(__FILE__, __LINE__, 0); } while (0) 146 # define sched_annotate_sleep() (current->task_state_change = 0) 147 148 /** 149 * cant_migrate - annotation for functions that cannot migrate 150 * 151 * Will print a stack trace if executed in code which is migratable 152 */ 153 # define cant_migrate() \ 154 do { \ 155 if (IS_ENABLED(CONFIG_SMP)) \ 156 __cant_migrate(__FILE__, __LINE__); \ 157 } while (0) 158 159 /** 160 * non_block_start - annotate the start of section where sleeping is prohibited 161 * 162 * This is on behalf of the oom reaper, specifically when it is calling the mmu 163 * notifiers. The problem is that if the notifier were to block on, for example, 164 * mutex_lock() and if the process which holds that mutex were to perform a 165 * sleeping memory allocation, the oom reaper is now blocked on completion of 166 * that memory allocation. Other blocking calls like wait_event() pose similar 167 * issues. 168 */ 169 # define non_block_start() (current->non_block_count++) 170 /** 171 * non_block_end - annotate the end of section where sleeping is prohibited 172 * 173 * Closes a section opened by non_block_start(). 174 */ 175 # define non_block_end() WARN_ON(current->non_block_count-- == 0) 176 #else 177 static inline void __might_resched(const char *file, int line, 178 unsigned int offsets) { } 179 static inline void __might_sleep(const char *file, int line) { } 180 # define might_sleep() do { might_resched(); } while (0) 181 # define cant_sleep() do { } while (0) 182 # define cant_migrate() do { } while (0) 183 # define sched_annotate_sleep() do { } while (0) 184 # define non_block_start() do { } while (0) 185 # define non_block_end() do { } while (0) 186 #endif 187 188 #define might_sleep_if(cond) do { if (cond) might_sleep(); } while (0) 189 190 #if defined(CONFIG_MMU) && \ 191 (defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)) 192 #define might_fault() __might_fault(__FILE__, __LINE__) 193 void __might_fault(const char *file, int line); 194 #else 195 static inline void might_fault(void) { } 196 #endif 197 198 void do_exit(long error_code) __noreturn; 199 void complete_and_exit(struct completion *, long) __noreturn; 200 201 extern int num_to_str(char *buf, int size, 202 unsigned long long num, unsigned int width); 203 204 /* lib/printf utilities */ 205 206 extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...); 207 extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list); 208 extern __printf(3, 4) 209 int snprintf(char *buf, size_t size, const char *fmt, ...); 210 extern __printf(3, 0) 211 int vsnprintf(char *buf, size_t size, const char *fmt, va_list args); 212 extern __printf(3, 4) 213 int scnprintf(char *buf, size_t size, const char *fmt, ...); 214 extern __printf(3, 0) 215 int vscnprintf(char *buf, size_t size, const char *fmt, va_list args); 216 extern __printf(2, 3) __malloc 217 char *kasprintf(gfp_t gfp, const char *fmt, ...); 218 extern __printf(2, 0) __malloc 219 char *kvasprintf(gfp_t gfp, const char *fmt, va_list args); 220 extern __printf(2, 0) 221 const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args); 222 223 extern __scanf(2, 3) 224 int sscanf(const char *, const char *, ...); 225 extern __scanf(2, 0) 226 int vsscanf(const char *, const char *, va_list); 227 228 extern int no_hash_pointers_enable(char *str); 229 230 extern int get_option(char **str, int *pint); 231 extern char *get_options(const char *str, int nints, int *ints); 232 extern unsigned long long memparse(const char *ptr, char **retptr); 233 extern bool parse_option_str(const char *str, const char *option); 234 extern char *next_arg(char *args, char **param, char **val); 235 236 extern int core_kernel_text(unsigned long addr); 237 extern int __kernel_text_address(unsigned long addr); 238 extern int kernel_text_address(unsigned long addr); 239 extern int func_ptr_is_kernel_text(void *ptr); 240 241 extern void bust_spinlocks(int yes); 242 243 extern int root_mountflags; 244 245 extern bool early_boot_irqs_disabled; 246 247 /* 248 * Values used for system_state. Ordering of the states must not be changed 249 * as code checks for <, <=, >, >= STATE. 250 */ 251 extern enum system_states { 252 SYSTEM_BOOTING, 253 SYSTEM_SCHEDULING, 254 SYSTEM_FREEING_INITMEM, 255 SYSTEM_RUNNING, 256 SYSTEM_HALT, 257 SYSTEM_POWER_OFF, 258 SYSTEM_RESTART, 259 SYSTEM_SUSPEND, 260 } system_state; 261 262 extern const char hex_asc[]; 263 #define hex_asc_lo(x) hex_asc[((x) & 0x0f)] 264 #define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4] 265 266 static inline char *hex_byte_pack(char *buf, u8 byte) 267 { 268 *buf++ = hex_asc_hi(byte); 269 *buf++ = hex_asc_lo(byte); 270 return buf; 271 } 272 273 extern const char hex_asc_upper[]; 274 #define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)] 275 #define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4] 276 277 static inline char *hex_byte_pack_upper(char *buf, u8 byte) 278 { 279 *buf++ = hex_asc_upper_hi(byte); 280 *buf++ = hex_asc_upper_lo(byte); 281 return buf; 282 } 283 284 extern int hex_to_bin(char ch); 285 extern int __must_check hex2bin(u8 *dst, const char *src, size_t count); 286 extern char *bin2hex(char *dst, const void *src, size_t count); 287 288 bool mac_pton(const char *s, u8 *mac); 289 290 /* 291 * General tracing related utility functions - trace_printk(), 292 * tracing_on/tracing_off and tracing_start()/tracing_stop 293 * 294 * Use tracing_on/tracing_off when you want to quickly turn on or off 295 * tracing. It simply enables or disables the recording of the trace events. 296 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on 297 * file, which gives a means for the kernel and userspace to interact. 298 * Place a tracing_off() in the kernel where you want tracing to end. 299 * From user space, examine the trace, and then echo 1 > tracing_on 300 * to continue tracing. 301 * 302 * tracing_stop/tracing_start has slightly more overhead. It is used 303 * by things like suspend to ram where disabling the recording of the 304 * trace is not enough, but tracing must actually stop because things 305 * like calling smp_processor_id() may crash the system. 306 * 307 * Most likely, you want to use tracing_on/tracing_off. 308 */ 309 310 enum ftrace_dump_mode { 311 DUMP_NONE, 312 DUMP_ALL, 313 DUMP_ORIG, 314 }; 315 316 #ifdef CONFIG_TRACING 317 void tracing_on(void); 318 void tracing_off(void); 319 int tracing_is_on(void); 320 void tracing_snapshot(void); 321 void tracing_snapshot_alloc(void); 322 323 extern void tracing_start(void); 324 extern void tracing_stop(void); 325 326 static inline __printf(1, 2) 327 void ____trace_printk_check_format(const char *fmt, ...) 328 { 329 } 330 #define __trace_printk_check_format(fmt, args...) \ 331 do { \ 332 if (0) \ 333 ____trace_printk_check_format(fmt, ##args); \ 334 } while (0) 335 336 /** 337 * trace_printk - printf formatting in the ftrace buffer 338 * @fmt: the printf format for printing 339 * 340 * Note: __trace_printk is an internal function for trace_printk() and 341 * the @ip is passed in via the trace_printk() macro. 342 * 343 * This function allows a kernel developer to debug fast path sections 344 * that printk is not appropriate for. By scattering in various 345 * printk like tracing in the code, a developer can quickly see 346 * where problems are occurring. 347 * 348 * This is intended as a debugging tool for the developer only. 349 * Please refrain from leaving trace_printks scattered around in 350 * your code. (Extra memory is used for special buffers that are 351 * allocated when trace_printk() is used.) 352 * 353 * A little optimization trick is done here. If there's only one 354 * argument, there's no need to scan the string for printf formats. 355 * The trace_puts() will suffice. But how can we take advantage of 356 * using trace_puts() when trace_printk() has only one argument? 357 * By stringifying the args and checking the size we can tell 358 * whether or not there are args. __stringify((__VA_ARGS__)) will 359 * turn into "()\0" with a size of 3 when there are no args, anything 360 * else will be bigger. All we need to do is define a string to this, 361 * and then take its size and compare to 3. If it's bigger, use 362 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just 363 * let gcc optimize the rest. 364 */ 365 366 #define trace_printk(fmt, ...) \ 367 do { \ 368 char _______STR[] = __stringify((__VA_ARGS__)); \ 369 if (sizeof(_______STR) > 3) \ 370 do_trace_printk(fmt, ##__VA_ARGS__); \ 371 else \ 372 trace_puts(fmt); \ 373 } while (0) 374 375 #define do_trace_printk(fmt, args...) \ 376 do { \ 377 static const char *trace_printk_fmt __used \ 378 __section("__trace_printk_fmt") = \ 379 __builtin_constant_p(fmt) ? fmt : NULL; \ 380 \ 381 __trace_printk_check_format(fmt, ##args); \ 382 \ 383 if (__builtin_constant_p(fmt)) \ 384 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ 385 else \ 386 __trace_printk(_THIS_IP_, fmt, ##args); \ 387 } while (0) 388 389 extern __printf(2, 3) 390 int __trace_bprintk(unsigned long ip, const char *fmt, ...); 391 392 extern __printf(2, 3) 393 int __trace_printk(unsigned long ip, const char *fmt, ...); 394 395 /** 396 * trace_puts - write a string into the ftrace buffer 397 * @str: the string to record 398 * 399 * Note: __trace_bputs is an internal function for trace_puts and 400 * the @ip is passed in via the trace_puts macro. 401 * 402 * This is similar to trace_printk() but is made for those really fast 403 * paths that a developer wants the least amount of "Heisenbug" effects, 404 * where the processing of the print format is still too much. 405 * 406 * This function allows a kernel developer to debug fast path sections 407 * that printk is not appropriate for. By scattering in various 408 * printk like tracing in the code, a developer can quickly see 409 * where problems are occurring. 410 * 411 * This is intended as a debugging tool for the developer only. 412 * Please refrain from leaving trace_puts scattered around in 413 * your code. (Extra memory is used for special buffers that are 414 * allocated when trace_puts() is used.) 415 * 416 * Returns: 0 if nothing was written, positive # if string was. 417 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used) 418 */ 419 420 #define trace_puts(str) ({ \ 421 static const char *trace_printk_fmt __used \ 422 __section("__trace_printk_fmt") = \ 423 __builtin_constant_p(str) ? str : NULL; \ 424 \ 425 if (__builtin_constant_p(str)) \ 426 __trace_bputs(_THIS_IP_, trace_printk_fmt); \ 427 else \ 428 __trace_puts(_THIS_IP_, str, strlen(str)); \ 429 }) 430 extern int __trace_bputs(unsigned long ip, const char *str); 431 extern int __trace_puts(unsigned long ip, const char *str, int size); 432 433 extern void trace_dump_stack(int skip); 434 435 /* 436 * The double __builtin_constant_p is because gcc will give us an error 437 * if we try to allocate the static variable to fmt if it is not a 438 * constant. Even with the outer if statement. 439 */ 440 #define ftrace_vprintk(fmt, vargs) \ 441 do { \ 442 if (__builtin_constant_p(fmt)) { \ 443 static const char *trace_printk_fmt __used \ 444 __section("__trace_printk_fmt") = \ 445 __builtin_constant_p(fmt) ? fmt : NULL; \ 446 \ 447 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ 448 } else \ 449 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ 450 } while (0) 451 452 extern __printf(2, 0) int 453 __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); 454 455 extern __printf(2, 0) int 456 __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); 457 458 extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode); 459 #else 460 static inline void tracing_start(void) { } 461 static inline void tracing_stop(void) { } 462 static inline void trace_dump_stack(int skip) { } 463 464 static inline void tracing_on(void) { } 465 static inline void tracing_off(void) { } 466 static inline int tracing_is_on(void) { return 0; } 467 static inline void tracing_snapshot(void) { } 468 static inline void tracing_snapshot_alloc(void) { } 469 470 static inline __printf(1, 2) 471 int trace_printk(const char *fmt, ...) 472 { 473 return 0; 474 } 475 static __printf(1, 0) inline int 476 ftrace_vprintk(const char *fmt, va_list ap) 477 { 478 return 0; 479 } 480 static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { } 481 #endif /* CONFIG_TRACING */ 482 483 /* This counts to 12. Any more, it will return 13th argument. */ 484 #define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n 485 #define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0) 486 487 #define __CONCAT(a, b) a ## b 488 #define CONCATENATE(a, b) __CONCAT(a, b) 489 490 /* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ 491 #ifdef CONFIG_FTRACE_MCOUNT_RECORD 492 # define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD 493 #endif 494 495 /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */ 496 #define VERIFY_OCTAL_PERMISSIONS(perms) \ 497 (BUILD_BUG_ON_ZERO((perms) < 0) + \ 498 BUILD_BUG_ON_ZERO((perms) > 0777) + \ 499 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \ 500 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \ 501 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \ 502 /* USER_WRITABLE >= GROUP_WRITABLE */ \ 503 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \ 504 /* OTHER_WRITABLE? Generally considered a bad idea. */ \ 505 BUILD_BUG_ON_ZERO((perms) & 2) + \ 506 (perms)) 507 #endif 508