at v5.15-rc2 18 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/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 88struct completion; 89struct user; 90 91#ifdef CONFIG_PREEMPT_VOLUNTARY 92 93extern int __cond_resched(void); 94# define might_resched() __cond_resched() 95 96#elif defined(CONFIG_PREEMPT_DYNAMIC) 97 98extern int __cond_resched(void); 99 100DECLARE_STATIC_CALL(might_resched, __cond_resched); 101 102static __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 114extern void ___might_sleep(const char *file, int line, int preempt_offset); 115extern void __might_sleep(const char *file, int line, int preempt_offset); 116extern void __cant_sleep(const char *file, int line, int preempt_offset); 117extern 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__) 188void __might_fault(const char *file, int line); 189#else 190static inline void might_fault(void) { } 191#endif 192 193void do_exit(long error_code) __noreturn; 194void complete_and_exit(struct completion *, long) __noreturn; 195 196extern int num_to_str(char *buf, int size, 197 unsigned long long num, unsigned int width); 198 199/* lib/printf utilities */ 200 201extern __printf(2, 3) int sprintf(char *buf, const char * fmt, ...); 202extern __printf(2, 0) int vsprintf(char *buf, const char *, va_list); 203extern __printf(3, 4) 204int snprintf(char *buf, size_t size, const char *fmt, ...); 205extern __printf(3, 0) 206int vsnprintf(char *buf, size_t size, const char *fmt, va_list args); 207extern __printf(3, 4) 208int scnprintf(char *buf, size_t size, const char *fmt, ...); 209extern __printf(3, 0) 210int vscnprintf(char *buf, size_t size, const char *fmt, va_list args); 211extern __printf(2, 3) __malloc 212char *kasprintf(gfp_t gfp, const char *fmt, ...); 213extern __printf(2, 0) __malloc 214char *kvasprintf(gfp_t gfp, const char *fmt, va_list args); 215extern __printf(2, 0) 216const char *kvasprintf_const(gfp_t gfp, const char *fmt, va_list args); 217 218extern __scanf(2, 3) 219int sscanf(const char *, const char *, ...); 220extern __scanf(2, 0) 221int vsscanf(const char *, const char *, va_list); 222 223extern int no_hash_pointers_enable(char *str); 224 225extern int get_option(char **str, int *pint); 226extern char *get_options(const char *str, int nints, int *ints); 227extern unsigned long long memparse(const char *ptr, char **retptr); 228extern bool parse_option_str(const char *str, const char *option); 229extern char *next_arg(char *args, char **param, char **val); 230 231extern int core_kernel_text(unsigned long addr); 232extern int init_kernel_text(unsigned long addr); 233extern int core_kernel_data(unsigned long addr); 234extern int __kernel_text_address(unsigned long addr); 235extern int kernel_text_address(unsigned long addr); 236extern int func_ptr_is_kernel_text(void *ptr); 237 238extern void bust_spinlocks(int yes); 239 240extern int root_mountflags; 241 242extern 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 */ 248extern enum system_states { 249 SYSTEM_BOOTING, 250 SYSTEM_SCHEDULING, 251 SYSTEM_RUNNING, 252 SYSTEM_HALT, 253 SYSTEM_POWER_OFF, 254 SYSTEM_RESTART, 255 SYSTEM_SUSPEND, 256} system_state; 257 258extern const char hex_asc[]; 259#define hex_asc_lo(x) hex_asc[((x) & 0x0f)] 260#define hex_asc_hi(x) hex_asc[((x) & 0xf0) >> 4] 261 262static inline char *hex_byte_pack(char *buf, u8 byte) 263{ 264 *buf++ = hex_asc_hi(byte); 265 *buf++ = hex_asc_lo(byte); 266 return buf; 267} 268 269extern const char hex_asc_upper[]; 270#define hex_asc_upper_lo(x) hex_asc_upper[((x) & 0x0f)] 271#define hex_asc_upper_hi(x) hex_asc_upper[((x) & 0xf0) >> 4] 272 273static inline char *hex_byte_pack_upper(char *buf, u8 byte) 274{ 275 *buf++ = hex_asc_upper_hi(byte); 276 *buf++ = hex_asc_upper_lo(byte); 277 return buf; 278} 279 280extern int hex_to_bin(char ch); 281extern int __must_check hex2bin(u8 *dst, const char *src, size_t count); 282extern char *bin2hex(char *dst, const void *src, size_t count); 283 284bool mac_pton(const char *s, u8 *mac); 285 286/* 287 * General tracing related utility functions - trace_printk(), 288 * tracing_on/tracing_off and tracing_start()/tracing_stop 289 * 290 * Use tracing_on/tracing_off when you want to quickly turn on or off 291 * tracing. It simply enables or disables the recording of the trace events. 292 * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on 293 * file, which gives a means for the kernel and userspace to interact. 294 * Place a tracing_off() in the kernel where you want tracing to end. 295 * From user space, examine the trace, and then echo 1 > tracing_on 296 * to continue tracing. 297 * 298 * tracing_stop/tracing_start has slightly more overhead. It is used 299 * by things like suspend to ram where disabling the recording of the 300 * trace is not enough, but tracing must actually stop because things 301 * like calling smp_processor_id() may crash the system. 302 * 303 * Most likely, you want to use tracing_on/tracing_off. 304 */ 305 306enum ftrace_dump_mode { 307 DUMP_NONE, 308 DUMP_ALL, 309 DUMP_ORIG, 310}; 311 312#ifdef CONFIG_TRACING 313void tracing_on(void); 314void tracing_off(void); 315int tracing_is_on(void); 316void tracing_snapshot(void); 317void tracing_snapshot_alloc(void); 318 319extern void tracing_start(void); 320extern void tracing_stop(void); 321 322static inline __printf(1, 2) 323void ____trace_printk_check_format(const char *fmt, ...) 324{ 325} 326#define __trace_printk_check_format(fmt, args...) \ 327do { \ 328 if (0) \ 329 ____trace_printk_check_format(fmt, ##args); \ 330} while (0) 331 332/** 333 * trace_printk - printf formatting in the ftrace buffer 334 * @fmt: the printf format for printing 335 * 336 * Note: __trace_printk is an internal function for trace_printk() and 337 * the @ip is passed in via the trace_printk() macro. 338 * 339 * This function allows a kernel developer to debug fast path sections 340 * that printk is not appropriate for. By scattering in various 341 * printk like tracing in the code, a developer can quickly see 342 * where problems are occurring. 343 * 344 * This is intended as a debugging tool for the developer only. 345 * Please refrain from leaving trace_printks scattered around in 346 * your code. (Extra memory is used for special buffers that are 347 * allocated when trace_printk() is used.) 348 * 349 * A little optimization trick is done here. If there's only one 350 * argument, there's no need to scan the string for printf formats. 351 * The trace_puts() will suffice. But how can we take advantage of 352 * using trace_puts() when trace_printk() has only one argument? 353 * By stringifying the args and checking the size we can tell 354 * whether or not there are args. __stringify((__VA_ARGS__)) will 355 * turn into "()\0" with a size of 3 when there are no args, anything 356 * else will be bigger. All we need to do is define a string to this, 357 * and then take its size and compare to 3. If it's bigger, use 358 * do_trace_printk() otherwise, optimize it to trace_puts(). Then just 359 * let gcc optimize the rest. 360 */ 361 362#define trace_printk(fmt, ...) \ 363do { \ 364 char _______STR[] = __stringify((__VA_ARGS__)); \ 365 if (sizeof(_______STR) > 3) \ 366 do_trace_printk(fmt, ##__VA_ARGS__); \ 367 else \ 368 trace_puts(fmt); \ 369} while (0) 370 371#define do_trace_printk(fmt, args...) \ 372do { \ 373 static const char *trace_printk_fmt __used \ 374 __section("__trace_printk_fmt") = \ 375 __builtin_constant_p(fmt) ? fmt : NULL; \ 376 \ 377 __trace_printk_check_format(fmt, ##args); \ 378 \ 379 if (__builtin_constant_p(fmt)) \ 380 __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ 381 else \ 382 __trace_printk(_THIS_IP_, fmt, ##args); \ 383} while (0) 384 385extern __printf(2, 3) 386int __trace_bprintk(unsigned long ip, const char *fmt, ...); 387 388extern __printf(2, 3) 389int __trace_printk(unsigned long ip, const char *fmt, ...); 390 391/** 392 * trace_puts - write a string into the ftrace buffer 393 * @str: the string to record 394 * 395 * Note: __trace_bputs is an internal function for trace_puts and 396 * the @ip is passed in via the trace_puts macro. 397 * 398 * This is similar to trace_printk() but is made for those really fast 399 * paths that a developer wants the least amount of "Heisenbug" effects, 400 * where the processing of the print format is still too much. 401 * 402 * This function allows a kernel developer to debug fast path sections 403 * that printk is not appropriate for. By scattering in various 404 * printk like tracing in the code, a developer can quickly see 405 * where problems are occurring. 406 * 407 * This is intended as a debugging tool for the developer only. 408 * Please refrain from leaving trace_puts scattered around in 409 * your code. (Extra memory is used for special buffers that are 410 * allocated when trace_puts() is used.) 411 * 412 * Returns: 0 if nothing was written, positive # if string was. 413 * (1 when __trace_bputs is used, strlen(str) when __trace_puts is used) 414 */ 415 416#define trace_puts(str) ({ \ 417 static const char *trace_printk_fmt __used \ 418 __section("__trace_printk_fmt") = \ 419 __builtin_constant_p(str) ? str : NULL; \ 420 \ 421 if (__builtin_constant_p(str)) \ 422 __trace_bputs(_THIS_IP_, trace_printk_fmt); \ 423 else \ 424 __trace_puts(_THIS_IP_, str, strlen(str)); \ 425}) 426extern int __trace_bputs(unsigned long ip, const char *str); 427extern int __trace_puts(unsigned long ip, const char *str, int size); 428 429extern void trace_dump_stack(int skip); 430 431/* 432 * The double __builtin_constant_p is because gcc will give us an error 433 * if we try to allocate the static variable to fmt if it is not a 434 * constant. Even with the outer if statement. 435 */ 436#define ftrace_vprintk(fmt, vargs) \ 437do { \ 438 if (__builtin_constant_p(fmt)) { \ 439 static const char *trace_printk_fmt __used \ 440 __section("__trace_printk_fmt") = \ 441 __builtin_constant_p(fmt) ? fmt : NULL; \ 442 \ 443 __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ 444 } else \ 445 __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ 446} while (0) 447 448extern __printf(2, 0) int 449__ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); 450 451extern __printf(2, 0) int 452__ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); 453 454extern void ftrace_dump(enum ftrace_dump_mode oops_dump_mode); 455#else 456static inline void tracing_start(void) { } 457static inline void tracing_stop(void) { } 458static inline void trace_dump_stack(int skip) { } 459 460static inline void tracing_on(void) { } 461static inline void tracing_off(void) { } 462static inline int tracing_is_on(void) { return 0; } 463static inline void tracing_snapshot(void) { } 464static inline void tracing_snapshot_alloc(void) { } 465 466static inline __printf(1, 2) 467int trace_printk(const char *fmt, ...) 468{ 469 return 0; 470} 471static __printf(1, 0) inline int 472ftrace_vprintk(const char *fmt, va_list ap) 473{ 474 return 0; 475} 476static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { } 477#endif /* CONFIG_TRACING */ 478 479/* This counts to 12. Any more, it will return 13th argument. */ 480#define __COUNT_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n 481#define COUNT_ARGS(X...) __COUNT_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0) 482 483#define __CONCAT(a, b) a ## b 484#define CONCATENATE(a, b) __CONCAT(a, b) 485 486/** 487 * container_of - cast a member of a structure out to the containing structure 488 * @ptr: the pointer to the member. 489 * @type: the type of the container struct this is embedded in. 490 * @member: the name of the member within the struct. 491 * 492 */ 493#define container_of(ptr, type, member) ({ \ 494 void *__mptr = (void *)(ptr); \ 495 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 496 !__same_type(*(ptr), void), \ 497 "pointer type mismatch in container_of()"); \ 498 ((type *)(__mptr - offsetof(type, member))); }) 499 500/** 501 * container_of_safe - cast a member of a structure out to the containing structure 502 * @ptr: the pointer to the member. 503 * @type: the type of the container struct this is embedded in. 504 * @member: the name of the member within the struct. 505 * 506 * If IS_ERR_OR_NULL(ptr), ptr is returned unchanged. 507 */ 508#define container_of_safe(ptr, type, member) ({ \ 509 void *__mptr = (void *)(ptr); \ 510 BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 511 !__same_type(*(ptr), void), \ 512 "pointer type mismatch in container_of()"); \ 513 IS_ERR_OR_NULL(__mptr) ? ERR_CAST(__mptr) : \ 514 ((type *)(__mptr - offsetof(type, member))); }) 515 516/* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ 517#ifdef CONFIG_FTRACE_MCOUNT_RECORD 518# define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD 519#endif 520 521/* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */ 522#define VERIFY_OCTAL_PERMISSIONS(perms) \ 523 (BUILD_BUG_ON_ZERO((perms) < 0) + \ 524 BUILD_BUG_ON_ZERO((perms) > 0777) + \ 525 /* USER_READABLE >= GROUP_READABLE >= OTHER_READABLE */ \ 526 BUILD_BUG_ON_ZERO((((perms) >> 6) & 4) < (((perms) >> 3) & 4)) + \ 527 BUILD_BUG_ON_ZERO((((perms) >> 3) & 4) < ((perms) & 4)) + \ 528 /* USER_WRITABLE >= GROUP_WRITABLE */ \ 529 BUILD_BUG_ON_ZERO((((perms) >> 6) & 2) < (((perms) >> 3) & 2)) + \ 530 /* OTHER_WRITABLE? Generally considered a bad idea. */ \ 531 BUILD_BUG_ON_ZERO((perms) & 2) + \ 532 (perms)) 533#endif