]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | #ifndef _LINUX_SCHED_H |
2 | #define _LINUX_SCHED_H | |
3 | ||
b7b3c76a DW |
4 | #include <linux/auxvec.h> /* For AT_VECTOR_SIZE */ |
5 | ||
6 | /* | |
7 | * cloning flags: | |
8 | */ | |
9 | #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */ | |
10 | #define CLONE_VM 0x00000100 /* set if VM shared between processes */ | |
11 | #define CLONE_FS 0x00000200 /* set if fs info shared between processes */ | |
12 | #define CLONE_FILES 0x00000400 /* set if open files shared between processes */ | |
13 | #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */ | |
14 | #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */ | |
15 | #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */ | |
16 | #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */ | |
17 | #define CLONE_THREAD 0x00010000 /* Same thread group? */ | |
18 | #define CLONE_NEWNS 0x00020000 /* New namespace group? */ | |
19 | #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */ | |
20 | #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */ | |
21 | #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */ | |
22 | #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */ | |
23 | #define CLONE_DETACHED 0x00400000 /* Unused, ignored */ | |
24 | #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */ | |
25 | #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */ | |
26 | #define CLONE_STOPPED 0x02000000 /* Start in stopped state */ | |
071df104 | 27 | #define CLONE_NEWUTS 0x04000000 /* New utsname group? */ |
25b21cb2 | 28 | #define CLONE_NEWIPC 0x08000000 /* New ipcs */ |
77ec739d | 29 | #define CLONE_NEWUSER 0x10000000 /* New user namespace */ |
169e3674 | 30 | #define CLONE_NEWNET 0x40000000 /* New network namespace */ |
b7b3c76a DW |
31 | |
32 | /* | |
33 | * Scheduling policies | |
34 | */ | |
35 | #define SCHED_NORMAL 0 | |
36 | #define SCHED_FIFO 1 | |
37 | #define SCHED_RR 2 | |
38 | #define SCHED_BATCH 3 | |
0e6aca43 IM |
39 | /* SCHED_ISO: reserved but not implemented yet */ |
40 | #define SCHED_IDLE 5 | |
b7b3c76a | 41 | |
a3b6714e | 42 | #ifdef __KERNEL__ |
b7b3c76a DW |
43 | |
44 | struct sched_param { | |
45 | int sched_priority; | |
46 | }; | |
47 | ||
1da177e4 LT |
48 | #include <asm/param.h> /* for HZ */ |
49 | ||
1da177e4 LT |
50 | #include <linux/capability.h> |
51 | #include <linux/threads.h> | |
52 | #include <linux/kernel.h> | |
53 | #include <linux/types.h> | |
54 | #include <linux/timex.h> | |
55 | #include <linux/jiffies.h> | |
56 | #include <linux/rbtree.h> | |
57 | #include <linux/thread_info.h> | |
58 | #include <linux/cpumask.h> | |
59 | #include <linux/errno.h> | |
60 | #include <linux/nodemask.h> | |
61 | ||
62 | #include <asm/system.h> | |
63 | #include <asm/semaphore.h> | |
64 | #include <asm/page.h> | |
65 | #include <asm/ptrace.h> | |
66 | #include <asm/mmu.h> | |
67 | #include <asm/cputime.h> | |
68 | ||
69 | #include <linux/smp.h> | |
70 | #include <linux/sem.h> | |
71 | #include <linux/signal.h> | |
72 | #include <linux/securebits.h> | |
73 | #include <linux/fs_struct.h> | |
74 | #include <linux/compiler.h> | |
75 | #include <linux/completion.h> | |
76 | #include <linux/pid.h> | |
77 | #include <linux/percpu.h> | |
78 | #include <linux/topology.h> | |
79 | #include <linux/seccomp.h> | |
e56d0903 | 80 | #include <linux/rcupdate.h> |
0771dfef | 81 | #include <linux/futex.h> |
23f78d4a | 82 | #include <linux/rtmutex.h> |
1da177e4 | 83 | |
a3b6714e DW |
84 | #include <linux/time.h> |
85 | #include <linux/param.h> | |
86 | #include <linux/resource.h> | |
87 | #include <linux/timer.h> | |
88 | #include <linux/hrtimer.h> | |
7c3ab738 | 89 | #include <linux/task_io_accounting.h> |
a3b6714e DW |
90 | |
91 | #include <asm/processor.h> | |
36d57ac4 | 92 | |
1da177e4 | 93 | struct exec_domain; |
c87e2837 | 94 | struct futex_pi_state; |
d89d8796 | 95 | struct bio; |
1da177e4 | 96 | |
1da177e4 LT |
97 | /* |
98 | * List of flags we want to share for kernel threads, | |
99 | * if only because they are not used by them anyway. | |
100 | */ | |
101 | #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND) | |
102 | ||
103 | /* | |
104 | * These are the constant used to fake the fixed-point load-average | |
105 | * counting. Some notes: | |
106 | * - 11 bit fractions expand to 22 bits by the multiplies: this gives | |
107 | * a load-average precision of 10 bits integer + 11 bits fractional | |
108 | * - if you want to count load-averages more often, you need more | |
109 | * precision, or rounding will get you. With 2-second counting freq, | |
110 | * the EXP_n values would be 1981, 2034 and 2043 if still using only | |
111 | * 11 bit fractions. | |
112 | */ | |
113 | extern unsigned long avenrun[]; /* Load averages */ | |
114 | ||
115 | #define FSHIFT 11 /* nr of bits of precision */ | |
116 | #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */ | |
0c2043ab | 117 | #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */ |
1da177e4 LT |
118 | #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */ |
119 | #define EXP_5 2014 /* 1/exp(5sec/5min) */ | |
120 | #define EXP_15 2037 /* 1/exp(5sec/15min) */ | |
121 | ||
122 | #define CALC_LOAD(load,exp,n) \ | |
123 | load *= exp; \ | |
124 | load += n*(FIXED_1-exp); \ | |
125 | load >>= FSHIFT; | |
126 | ||
127 | extern unsigned long total_forks; | |
128 | extern int nr_threads; | |
1da177e4 LT |
129 | DECLARE_PER_CPU(unsigned long, process_counts); |
130 | extern int nr_processes(void); | |
131 | extern unsigned long nr_running(void); | |
132 | extern unsigned long nr_uninterruptible(void); | |
db1b1fef | 133 | extern unsigned long nr_active(void); |
1da177e4 | 134 | extern unsigned long nr_iowait(void); |
2dd73a4f | 135 | extern unsigned long weighted_cpuload(const int cpu); |
1da177e4 | 136 | |
43ae34cb IM |
137 | struct seq_file; |
138 | struct cfs_rq; | |
139 | #ifdef CONFIG_SCHED_DEBUG | |
140 | extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m); | |
141 | extern void proc_sched_set_task(struct task_struct *p); | |
142 | extern void | |
5cef9eca | 143 | print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq); |
43ae34cb IM |
144 | #else |
145 | static inline void | |
146 | proc_sched_show_task(struct task_struct *p, struct seq_file *m) | |
147 | { | |
148 | } | |
149 | static inline void proc_sched_set_task(struct task_struct *p) | |
150 | { | |
151 | } | |
152 | static inline void | |
5cef9eca | 153 | print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) |
43ae34cb IM |
154 | { |
155 | } | |
156 | #endif | |
1da177e4 | 157 | |
4a8342d2 LT |
158 | /* |
159 | * Task state bitmask. NOTE! These bits are also | |
160 | * encoded in fs/proc/array.c: get_task_state(). | |
161 | * | |
162 | * We have two separate sets of flags: task->state | |
163 | * is about runnability, while task->exit_state are | |
164 | * about the task exiting. Confusing, but this way | |
165 | * modifying one set can't modify the other one by | |
166 | * mistake. | |
167 | */ | |
1da177e4 LT |
168 | #define TASK_RUNNING 0 |
169 | #define TASK_INTERRUPTIBLE 1 | |
170 | #define TASK_UNINTERRUPTIBLE 2 | |
4a8342d2 LT |
171 | #define TASK_STOPPED 4 |
172 | #define TASK_TRACED 8 | |
173 | /* in tsk->exit_state */ | |
174 | #define EXIT_ZOMBIE 16 | |
175 | #define EXIT_DEAD 32 | |
176 | /* in tsk->state again */ | |
177 | #define TASK_NONINTERACTIVE 64 | |
c394cc9f | 178 | #define TASK_DEAD 128 |
1da177e4 LT |
179 | |
180 | #define __set_task_state(tsk, state_value) \ | |
181 | do { (tsk)->state = (state_value); } while (0) | |
182 | #define set_task_state(tsk, state_value) \ | |
183 | set_mb((tsk)->state, (state_value)) | |
184 | ||
498d0c57 AM |
185 | /* |
186 | * set_current_state() includes a barrier so that the write of current->state | |
187 | * is correctly serialised wrt the caller's subsequent test of whether to | |
188 | * actually sleep: | |
189 | * | |
190 | * set_current_state(TASK_UNINTERRUPTIBLE); | |
191 | * if (do_i_need_to_sleep()) | |
192 | * schedule(); | |
193 | * | |
194 | * If the caller does not need such serialisation then use __set_current_state() | |
195 | */ | |
1da177e4 LT |
196 | #define __set_current_state(state_value) \ |
197 | do { current->state = (state_value); } while (0) | |
198 | #define set_current_state(state_value) \ | |
199 | set_mb(current->state, (state_value)) | |
200 | ||
201 | /* Task command name length */ | |
202 | #define TASK_COMM_LEN 16 | |
203 | ||
1da177e4 LT |
204 | #include <linux/spinlock.h> |
205 | ||
206 | /* | |
207 | * This serializes "schedule()" and also protects | |
208 | * the run-queue from deletions/modifications (but | |
209 | * _adding_ to the beginning of the run-queue has | |
210 | * a separate lock). | |
211 | */ | |
212 | extern rwlock_t tasklist_lock; | |
213 | extern spinlock_t mmlist_lock; | |
214 | ||
36c8b586 | 215 | struct task_struct; |
1da177e4 LT |
216 | |
217 | extern void sched_init(void); | |
218 | extern void sched_init_smp(void); | |
36c8b586 | 219 | extern void init_idle(struct task_struct *idle, int cpu); |
1df21055 | 220 | extern void init_idle_bootup_task(struct task_struct *idle); |
1da177e4 LT |
221 | |
222 | extern cpumask_t nohz_cpu_mask; | |
46cb4b7c SS |
223 | #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ) |
224 | extern int select_nohz_load_balancer(int cpu); | |
225 | #else | |
226 | static inline int select_nohz_load_balancer(int cpu) | |
227 | { | |
228 | return 0; | |
229 | } | |
230 | #endif | |
1da177e4 | 231 | |
e59e2ae2 | 232 | /* |
39bc89fd | 233 | * Only dump TASK_* tasks. (0 for all tasks) |
e59e2ae2 IM |
234 | */ |
235 | extern void show_state_filter(unsigned long state_filter); | |
236 | ||
237 | static inline void show_state(void) | |
238 | { | |
39bc89fd | 239 | show_state_filter(0); |
e59e2ae2 IM |
240 | } |
241 | ||
1da177e4 LT |
242 | extern void show_regs(struct pt_regs *); |
243 | ||
244 | /* | |
245 | * TASK is a pointer to the task whose backtrace we want to see (or NULL for current | |
246 | * task), SP is the stack pointer of the first frame that should be shown in the back | |
247 | * trace (or NULL if the entire call-chain of the task should be shown). | |
248 | */ | |
249 | extern void show_stack(struct task_struct *task, unsigned long *sp); | |
250 | ||
251 | void io_schedule(void); | |
252 | long io_schedule_timeout(long timeout); | |
253 | ||
254 | extern void cpu_init (void); | |
255 | extern void trap_init(void); | |
256 | extern void update_process_times(int user); | |
257 | extern void scheduler_tick(void); | |
258 | ||
8446f1d3 | 259 | #ifdef CONFIG_DETECT_SOFTLOCKUP |
6687a97d | 260 | extern void softlockup_tick(void); |
8446f1d3 IM |
261 | extern void spawn_softlockup_task(void); |
262 | extern void touch_softlockup_watchdog(void); | |
04c9167f | 263 | extern void touch_all_softlockup_watchdogs(void); |
8446f1d3 | 264 | #else |
6687a97d | 265 | static inline void softlockup_tick(void) |
8446f1d3 IM |
266 | { |
267 | } | |
268 | static inline void spawn_softlockup_task(void) | |
269 | { | |
270 | } | |
271 | static inline void touch_softlockup_watchdog(void) | |
272 | { | |
273 | } | |
04c9167f JF |
274 | static inline void touch_all_softlockup_watchdogs(void) |
275 | { | |
276 | } | |
8446f1d3 IM |
277 | #endif |
278 | ||
279 | ||
1da177e4 LT |
280 | /* Attach to any functions which should be ignored in wchan output. */ |
281 | #define __sched __attribute__((__section__(".sched.text"))) | |
282 | /* Is this address in the __sched functions? */ | |
283 | extern int in_sched_functions(unsigned long addr); | |
284 | ||
285 | #define MAX_SCHEDULE_TIMEOUT LONG_MAX | |
286 | extern signed long FASTCALL(schedule_timeout(signed long timeout)); | |
64ed93a2 NA |
287 | extern signed long schedule_timeout_interruptible(signed long timeout); |
288 | extern signed long schedule_timeout_uninterruptible(signed long timeout); | |
1da177e4 LT |
289 | asmlinkage void schedule(void); |
290 | ||
ab516013 | 291 | struct nsproxy; |
acce292c | 292 | struct user_namespace; |
1da177e4 LT |
293 | |
294 | /* Maximum number of active map areas.. This is a random (large) number */ | |
295 | #define DEFAULT_MAX_MAP_COUNT 65536 | |
296 | ||
297 | extern int sysctl_max_map_count; | |
298 | ||
299 | #include <linux/aio.h> | |
300 | ||
301 | extern unsigned long | |
302 | arch_get_unmapped_area(struct file *, unsigned long, unsigned long, | |
303 | unsigned long, unsigned long); | |
304 | extern unsigned long | |
305 | arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr, | |
306 | unsigned long len, unsigned long pgoff, | |
307 | unsigned long flags); | |
1363c3cd WW |
308 | extern void arch_unmap_area(struct mm_struct *, unsigned long); |
309 | extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long); | |
1da177e4 | 310 | |
f412ac08 HD |
311 | #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS |
312 | /* | |
313 | * The mm counters are not protected by its page_table_lock, | |
314 | * so must be incremented atomically. | |
315 | */ | |
d3cb4871 CL |
316 | #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value) |
317 | #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member)) | |
318 | #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member) | |
319 | #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member) | |
320 | #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member) | |
321 | typedef atomic_long_t mm_counter_t; | |
f412ac08 HD |
322 | |
323 | #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ | |
324 | /* | |
325 | * The mm counters are protected by its page_table_lock, | |
326 | * so can be incremented directly. | |
327 | */ | |
1da177e4 LT |
328 | #define set_mm_counter(mm, member, value) (mm)->_##member = (value) |
329 | #define get_mm_counter(mm, member) ((mm)->_##member) | |
330 | #define add_mm_counter(mm, member, value) (mm)->_##member += (value) | |
331 | #define inc_mm_counter(mm, member) (mm)->_##member++ | |
332 | #define dec_mm_counter(mm, member) (mm)->_##member-- | |
f412ac08 HD |
333 | typedef unsigned long mm_counter_t; |
334 | ||
335 | #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ | |
4294621f | 336 | |
f412ac08 HD |
337 | #define get_mm_rss(mm) \ |
338 | (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss)) | |
365e9c87 HD |
339 | #define update_hiwater_rss(mm) do { \ |
340 | unsigned long _rss = get_mm_rss(mm); \ | |
341 | if ((mm)->hiwater_rss < _rss) \ | |
342 | (mm)->hiwater_rss = _rss; \ | |
343 | } while (0) | |
344 | #define update_hiwater_vm(mm) do { \ | |
345 | if ((mm)->hiwater_vm < (mm)->total_vm) \ | |
346 | (mm)->hiwater_vm = (mm)->total_vm; \ | |
347 | } while (0) | |
348 | ||
6c5d5238 KH |
349 | extern void set_dumpable(struct mm_struct *mm, int value); |
350 | extern int get_dumpable(struct mm_struct *mm); | |
351 | ||
352 | /* mm flags */ | |
3cb4a0bb | 353 | /* dumpable bits */ |
6c5d5238 KH |
354 | #define MMF_DUMPABLE 0 /* core dump is permitted */ |
355 | #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */ | |
3cb4a0bb KH |
356 | #define MMF_DUMPABLE_BITS 2 |
357 | ||
358 | /* coredump filter bits */ | |
359 | #define MMF_DUMP_ANON_PRIVATE 2 | |
360 | #define MMF_DUMP_ANON_SHARED 3 | |
361 | #define MMF_DUMP_MAPPED_PRIVATE 4 | |
362 | #define MMF_DUMP_MAPPED_SHARED 5 | |
363 | #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS | |
364 | #define MMF_DUMP_FILTER_BITS 4 | |
365 | #define MMF_DUMP_FILTER_MASK \ | |
366 | (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT) | |
367 | #define MMF_DUMP_FILTER_DEFAULT \ | |
368 | ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED)) | |
6c5d5238 | 369 | |
1da177e4 LT |
370 | struct mm_struct { |
371 | struct vm_area_struct * mmap; /* list of VMAs */ | |
372 | struct rb_root mm_rb; | |
373 | struct vm_area_struct * mmap_cache; /* last find_vma result */ | |
374 | unsigned long (*get_unmapped_area) (struct file *filp, | |
375 | unsigned long addr, unsigned long len, | |
376 | unsigned long pgoff, unsigned long flags); | |
1363c3cd | 377 | void (*unmap_area) (struct mm_struct *mm, unsigned long addr); |
0551fbd2 BH |
378 | unsigned long mmap_base; /* base of mmap area */ |
379 | unsigned long task_size; /* size of task vm space */ | |
380 | unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */ | |
1363c3cd | 381 | unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */ |
1da177e4 LT |
382 | pgd_t * pgd; |
383 | atomic_t mm_users; /* How many users with user space? */ | |
384 | atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */ | |
385 | int map_count; /* number of VMAs */ | |
386 | struct rw_semaphore mmap_sem; | |
387 | spinlock_t page_table_lock; /* Protects page tables and some counters */ | |
388 | ||
389 | struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung | |
390 | * together off init_mm.mmlist, and are protected | |
391 | * by mmlist_lock | |
392 | */ | |
393 | ||
f412ac08 HD |
394 | /* Special counters, in some configurations protected by the |
395 | * page_table_lock, in other configurations by being atomic. | |
396 | */ | |
4294621f | 397 | mm_counter_t _file_rss; |
1da177e4 LT |
398 | mm_counter_t _anon_rss; |
399 | ||
f449952b HD |
400 | unsigned long hiwater_rss; /* High-watermark of RSS usage */ |
401 | unsigned long hiwater_vm; /* High-water virtual memory usage */ | |
402 | ||
403 | unsigned long total_vm, locked_vm, shared_vm, exec_vm; | |
404 | unsigned long stack_vm, reserved_vm, def_flags, nr_ptes; | |
405 | unsigned long start_code, end_code, start_data, end_data; | |
406 | unsigned long start_brk, brk, start_stack; | |
407 | unsigned long arg_start, arg_end, env_start, env_end; | |
408 | ||
36d57ac4 | 409 | unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */ |
1da177e4 | 410 | |
1da177e4 LT |
411 | cpumask_t cpu_vm_mask; |
412 | ||
413 | /* Architecture-specific MM context */ | |
414 | mm_context_t context; | |
415 | ||
7602bdf2 AC |
416 | /* Swap token stuff */ |
417 | /* | |
418 | * Last value of global fault stamp as seen by this process. | |
419 | * In other words, this value gives an indication of how long | |
420 | * it has been since this task got the token. | |
421 | * Look at mm/thrash.c | |
422 | */ | |
423 | unsigned int faultstamp; | |
424 | unsigned int token_priority; | |
425 | unsigned int last_interval; | |
1da177e4 | 426 | |
6c5d5238 | 427 | unsigned long flags; /* Must use atomic bitops to access the bits */ |
36de6437 | 428 | |
1da177e4 LT |
429 | /* coredumping support */ |
430 | int core_waiters; | |
431 | struct completion *core_startup_done, core_done; | |
432 | ||
433 | /* aio bits */ | |
434 | rwlock_t ioctx_list_lock; | |
435 | struct kioctx *ioctx_list; | |
1da177e4 LT |
436 | }; |
437 | ||
438 | struct sighand_struct { | |
439 | atomic_t count; | |
440 | struct k_sigaction action[_NSIG]; | |
441 | spinlock_t siglock; | |
b8fceee1 | 442 | wait_queue_head_t signalfd_wqh; |
1da177e4 LT |
443 | }; |
444 | ||
0e464814 | 445 | struct pacct_struct { |
f6ec29a4 KK |
446 | int ac_flag; |
447 | long ac_exitcode; | |
0e464814 | 448 | unsigned long ac_mem; |
77787bfb KK |
449 | cputime_t ac_utime, ac_stime; |
450 | unsigned long ac_minflt, ac_majflt; | |
0e464814 KK |
451 | }; |
452 | ||
1da177e4 LT |
453 | /* |
454 | * NOTE! "signal_struct" does not have it's own | |
455 | * locking, because a shared signal_struct always | |
456 | * implies a shared sighand_struct, so locking | |
457 | * sighand_struct is always a proper superset of | |
458 | * the locking of signal_struct. | |
459 | */ | |
460 | struct signal_struct { | |
461 | atomic_t count; | |
462 | atomic_t live; | |
463 | ||
464 | wait_queue_head_t wait_chldexit; /* for wait4() */ | |
465 | ||
466 | /* current thread group signal load-balancing target: */ | |
36c8b586 | 467 | struct task_struct *curr_target; |
1da177e4 LT |
468 | |
469 | /* shared signal handling: */ | |
470 | struct sigpending shared_pending; | |
471 | ||
472 | /* thread group exit support */ | |
473 | int group_exit_code; | |
474 | /* overloaded: | |
475 | * - notify group_exit_task when ->count is equal to notify_count | |
476 | * - everyone except group_exit_task is stopped during signal delivery | |
477 | * of fatal signals, group_exit_task processes the signal. | |
478 | */ | |
479 | struct task_struct *group_exit_task; | |
480 | int notify_count; | |
481 | ||
482 | /* thread group stop support, overloads group_exit_code too */ | |
483 | int group_stop_count; | |
484 | unsigned int flags; /* see SIGNAL_* flags below */ | |
485 | ||
486 | /* POSIX.1b Interval Timers */ | |
487 | struct list_head posix_timers; | |
488 | ||
489 | /* ITIMER_REAL timer for the process */ | |
2ff678b8 | 490 | struct hrtimer real_timer; |
05cfb614 | 491 | struct task_struct *tsk; |
2ff678b8 | 492 | ktime_t it_real_incr; |
1da177e4 LT |
493 | |
494 | /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */ | |
495 | cputime_t it_prof_expires, it_virt_expires; | |
496 | cputime_t it_prof_incr, it_virt_incr; | |
497 | ||
498 | /* job control IDs */ | |
499 | pid_t pgrp; | |
ab521dc0 | 500 | struct pid *tty_old_pgrp; |
1ec320af CLG |
501 | |
502 | union { | |
503 | pid_t session __deprecated; | |
504 | pid_t __session; | |
505 | }; | |
506 | ||
1da177e4 LT |
507 | /* boolean value for session group leader */ |
508 | int leader; | |
509 | ||
510 | struct tty_struct *tty; /* NULL if no tty */ | |
511 | ||
512 | /* | |
513 | * Cumulative resource counters for dead threads in the group, | |
514 | * and for reaped dead child processes forked by this group. | |
515 | * Live threads maintain their own counters and add to these | |
516 | * in __exit_signal, except for the group leader. | |
517 | */ | |
518 | cputime_t utime, stime, cutime, cstime; | |
519 | unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw; | |
520 | unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt; | |
6eaeeaba | 521 | unsigned long inblock, oublock, cinblock, coublock; |
1da177e4 LT |
522 | |
523 | /* | |
524 | * Cumulative ns of scheduled CPU time for dead threads in the | |
525 | * group, not including a zombie group leader. (This only differs | |
526 | * from jiffies_to_ns(utime + stime) if sched_clock uses something | |
527 | * other than jiffies.) | |
528 | */ | |
41b86e9c | 529 | unsigned long long sum_sched_runtime; |
1da177e4 LT |
530 | |
531 | /* | |
532 | * We don't bother to synchronize most readers of this at all, | |
533 | * because there is no reader checking a limit that actually needs | |
534 | * to get both rlim_cur and rlim_max atomically, and either one | |
535 | * alone is a single word that can safely be read normally. | |
536 | * getrlimit/setrlimit use task_lock(current->group_leader) to | |
537 | * protect this instead of the siglock, because they really | |
538 | * have no need to disable irqs. | |
539 | */ | |
540 | struct rlimit rlim[RLIM_NLIMITS]; | |
541 | ||
542 | struct list_head cpu_timers[3]; | |
543 | ||
544 | /* keep the process-shared keyrings here so that they do the right | |
545 | * thing in threads created with CLONE_THREAD */ | |
546 | #ifdef CONFIG_KEYS | |
547 | struct key *session_keyring; /* keyring inherited over fork */ | |
548 | struct key *process_keyring; /* keyring private to this process */ | |
549 | #endif | |
0e464814 KK |
550 | #ifdef CONFIG_BSD_PROCESS_ACCT |
551 | struct pacct_struct pacct; /* per-process accounting information */ | |
552 | #endif | |
ad4ecbcb | 553 | #ifdef CONFIG_TASKSTATS |
ad4ecbcb SN |
554 | struct taskstats *stats; |
555 | #endif | |
522ed776 MT |
556 | #ifdef CONFIG_AUDIT |
557 | unsigned audit_tty; | |
558 | struct tty_audit_buf *tty_audit_buf; | |
559 | #endif | |
1da177e4 LT |
560 | }; |
561 | ||
4866cde0 NP |
562 | /* Context switch must be unlocked if interrupts are to be enabled */ |
563 | #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW | |
564 | # define __ARCH_WANT_UNLOCKED_CTXSW | |
565 | #endif | |
566 | ||
1da177e4 LT |
567 | /* |
568 | * Bits in flags field of signal_struct. | |
569 | */ | |
570 | #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */ | |
571 | #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */ | |
572 | #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */ | |
573 | #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */ | |
574 | ||
1da177e4 LT |
575 | /* |
576 | * Some day this will be a full-fledged user tracking system.. | |
577 | */ | |
578 | struct user_struct { | |
579 | atomic_t __count; /* reference count */ | |
580 | atomic_t processes; /* How many processes does this user have? */ | |
581 | atomic_t files; /* How many open files does this user have? */ | |
582 | atomic_t sigpending; /* How many pending signals does this user have? */ | |
2d9048e2 | 583 | #ifdef CONFIG_INOTIFY_USER |
0eeca283 RL |
584 | atomic_t inotify_watches; /* How many inotify watches does this user have? */ |
585 | atomic_t inotify_devs; /* How many inotify devs does this user have opened? */ | |
586 | #endif | |
1da177e4 LT |
587 | /* protected by mq_lock */ |
588 | unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */ | |
589 | unsigned long locked_shm; /* How many pages of mlocked shm ? */ | |
590 | ||
591 | #ifdef CONFIG_KEYS | |
592 | struct key *uid_keyring; /* UID specific keyring */ | |
593 | struct key *session_keyring; /* UID's default session keyring */ | |
594 | #endif | |
595 | ||
596 | /* Hash table maintenance information */ | |
735de223 | 597 | struct hlist_node uidhash_node; |
1da177e4 LT |
598 | uid_t uid; |
599 | }; | |
600 | ||
601 | extern struct user_struct *find_user(uid_t); | |
602 | ||
603 | extern struct user_struct root_user; | |
604 | #define INIT_USER (&root_user) | |
605 | ||
1da177e4 LT |
606 | struct backing_dev_info; |
607 | struct reclaim_state; | |
608 | ||
52f17b6c | 609 | #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) |
1da177e4 LT |
610 | struct sched_info { |
611 | /* cumulative counters */ | |
172ba844 BS |
612 | unsigned long pcnt; /* # of times run on this cpu */ |
613 | unsigned long long cpu_time, /* time spent on the cpu */ | |
614 | run_delay; /* time spent waiting on a runqueue */ | |
1da177e4 LT |
615 | |
616 | /* timestamps */ | |
172ba844 BS |
617 | unsigned long long last_arrival,/* when we last ran on a cpu */ |
618 | last_queued; /* when we were last queued to run */ | |
1da177e4 | 619 | }; |
52f17b6c | 620 | #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */ |
1da177e4 | 621 | |
52f17b6c | 622 | #ifdef CONFIG_SCHEDSTATS |
15ad7cdc | 623 | extern const struct file_operations proc_schedstat_operations; |
52f17b6c | 624 | #endif /* CONFIG_SCHEDSTATS */ |
1da177e4 | 625 | |
ca74e92b SN |
626 | #ifdef CONFIG_TASK_DELAY_ACCT |
627 | struct task_delay_info { | |
628 | spinlock_t lock; | |
629 | unsigned int flags; /* Private per-task flags */ | |
630 | ||
631 | /* For each stat XXX, add following, aligned appropriately | |
632 | * | |
633 | * struct timespec XXX_start, XXX_end; | |
634 | * u64 XXX_delay; | |
635 | * u32 XXX_count; | |
636 | * | |
637 | * Atomicity of updates to XXX_delay, XXX_count protected by | |
638 | * single lock above (split into XXX_lock if contention is an issue). | |
639 | */ | |
0ff92245 SN |
640 | |
641 | /* | |
642 | * XXX_count is incremented on every XXX operation, the delay | |
643 | * associated with the operation is added to XXX_delay. | |
644 | * XXX_delay contains the accumulated delay time in nanoseconds. | |
645 | */ | |
646 | struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */ | |
647 | u64 blkio_delay; /* wait for sync block io completion */ | |
648 | u64 swapin_delay; /* wait for swapin block io completion */ | |
649 | u32 blkio_count; /* total count of the number of sync block */ | |
650 | /* io operations performed */ | |
651 | u32 swapin_count; /* total count of the number of swapin block */ | |
652 | /* io operations performed */ | |
ca74e92b | 653 | }; |
52f17b6c CS |
654 | #endif /* CONFIG_TASK_DELAY_ACCT */ |
655 | ||
656 | static inline int sched_info_on(void) | |
657 | { | |
658 | #ifdef CONFIG_SCHEDSTATS | |
659 | return 1; | |
660 | #elif defined(CONFIG_TASK_DELAY_ACCT) | |
661 | extern int delayacct_on; | |
662 | return delayacct_on; | |
663 | #else | |
664 | return 0; | |
ca74e92b | 665 | #endif |
52f17b6c | 666 | } |
ca74e92b | 667 | |
d15bcfdb IM |
668 | enum cpu_idle_type { |
669 | CPU_IDLE, | |
670 | CPU_NOT_IDLE, | |
671 | CPU_NEWLY_IDLE, | |
672 | CPU_MAX_IDLE_TYPES | |
1da177e4 LT |
673 | }; |
674 | ||
675 | /* | |
676 | * sched-domains (multiprocessor balancing) declarations: | |
677 | */ | |
9aa7b369 IM |
678 | |
679 | /* | |
680 | * Increase resolution of nice-level calculations: | |
681 | */ | |
682 | #define SCHED_LOAD_SHIFT 10 | |
683 | #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT) | |
684 | ||
f8700df7 | 685 | #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE |
1da177e4 | 686 | |
2dd73a4f | 687 | #ifdef CONFIG_SMP |
1da177e4 LT |
688 | #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */ |
689 | #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */ | |
690 | #define SD_BALANCE_EXEC 4 /* Balance on exec */ | |
147cbb4b NP |
691 | #define SD_BALANCE_FORK 8 /* Balance on fork, clone */ |
692 | #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */ | |
693 | #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */ | |
694 | #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */ | |
695 | #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */ | |
5c45bf27 | 696 | #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */ |
89c4710e | 697 | #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */ |
08c183f3 | 698 | #define SD_SERIALIZE 1024 /* Only a single load balancing instance */ |
5c45bf27 | 699 | |
89c4710e SS |
700 | #define BALANCE_FOR_MC_POWER \ |
701 | (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0) | |
702 | ||
703 | #define BALANCE_FOR_PKG_POWER \ | |
704 | ((sched_mc_power_savings || sched_smt_power_savings) ? \ | |
705 | SD_POWERSAVINGS_BALANCE : 0) | |
706 | ||
707 | #define test_sd_parent(sd, flag) ((sd->parent && \ | |
708 | (sd->parent->flags & flag)) ? 1 : 0) | |
5c45bf27 | 709 | |
1da177e4 LT |
710 | |
711 | struct sched_group { | |
712 | struct sched_group *next; /* Must be a circular list */ | |
713 | cpumask_t cpumask; | |
714 | ||
715 | /* | |
716 | * CPU power of this group, SCHED_LOAD_SCALE being max power for a | |
717 | * single CPU. This is read only (except for setup, hotplug CPU). | |
5517d86b | 718 | * Note : Never change cpu_power without recompute its reciprocal |
1da177e4 | 719 | */ |
5517d86b ED |
720 | unsigned int __cpu_power; |
721 | /* | |
722 | * reciprocal value of cpu_power to avoid expensive divides | |
723 | * (see include/linux/reciprocal_div.h) | |
724 | */ | |
725 | u32 reciprocal_cpu_power; | |
1da177e4 LT |
726 | }; |
727 | ||
728 | struct sched_domain { | |
729 | /* These fields must be setup */ | |
730 | struct sched_domain *parent; /* top domain must be null terminated */ | |
1a848870 | 731 | struct sched_domain *child; /* bottom domain must be null terminated */ |
1da177e4 LT |
732 | struct sched_group *groups; /* the balancing groups of the domain */ |
733 | cpumask_t span; /* span of all CPUs in this domain */ | |
734 | unsigned long min_interval; /* Minimum balance interval ms */ | |
735 | unsigned long max_interval; /* Maximum balance interval ms */ | |
736 | unsigned int busy_factor; /* less balancing by factor if busy */ | |
737 | unsigned int imbalance_pct; /* No balance until over watermark */ | |
1da177e4 | 738 | unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */ |
7897986b NP |
739 | unsigned int busy_idx; |
740 | unsigned int idle_idx; | |
741 | unsigned int newidle_idx; | |
742 | unsigned int wake_idx; | |
147cbb4b | 743 | unsigned int forkexec_idx; |
1da177e4 LT |
744 | int flags; /* See SD_* */ |
745 | ||
746 | /* Runtime fields. */ | |
747 | unsigned long last_balance; /* init to jiffies. units in jiffies */ | |
748 | unsigned int balance_interval; /* initialise to 1. units in ms. */ | |
749 | unsigned int nr_balance_failed; /* initialise to 0 */ | |
750 | ||
751 | #ifdef CONFIG_SCHEDSTATS | |
752 | /* load_balance() stats */ | |
d15bcfdb IM |
753 | unsigned long lb_cnt[CPU_MAX_IDLE_TYPES]; |
754 | unsigned long lb_failed[CPU_MAX_IDLE_TYPES]; | |
755 | unsigned long lb_balanced[CPU_MAX_IDLE_TYPES]; | |
756 | unsigned long lb_imbalance[CPU_MAX_IDLE_TYPES]; | |
757 | unsigned long lb_gained[CPU_MAX_IDLE_TYPES]; | |
758 | unsigned long lb_hot_gained[CPU_MAX_IDLE_TYPES]; | |
759 | unsigned long lb_nobusyg[CPU_MAX_IDLE_TYPES]; | |
760 | unsigned long lb_nobusyq[CPU_MAX_IDLE_TYPES]; | |
1da177e4 LT |
761 | |
762 | /* Active load balancing */ | |
763 | unsigned long alb_cnt; | |
764 | unsigned long alb_failed; | |
765 | unsigned long alb_pushed; | |
766 | ||
68767a0a NP |
767 | /* SD_BALANCE_EXEC stats */ |
768 | unsigned long sbe_cnt; | |
769 | unsigned long sbe_balanced; | |
1da177e4 LT |
770 | unsigned long sbe_pushed; |
771 | ||
68767a0a NP |
772 | /* SD_BALANCE_FORK stats */ |
773 | unsigned long sbf_cnt; | |
774 | unsigned long sbf_balanced; | |
775 | unsigned long sbf_pushed; | |
776 | ||
1da177e4 LT |
777 | /* try_to_wake_up() stats */ |
778 | unsigned long ttwu_wake_remote; | |
779 | unsigned long ttwu_move_affine; | |
780 | unsigned long ttwu_move_balance; | |
781 | #endif | |
782 | }; | |
783 | ||
51888ca2 | 784 | extern int partition_sched_domains(cpumask_t *partition1, |
1a20ff27 | 785 | cpumask_t *partition2); |
198e2f18 | 786 | |
198e2f18 | 787 | #endif /* CONFIG_SMP */ |
1da177e4 | 788 | |
d02c7a8c CK |
789 | /* |
790 | * A runqueue laden with a single nice 0 task scores a weighted_cpuload of | |
791 | * SCHED_LOAD_SCALE. This function returns 1 if any cpu is laden with a | |
792 | * task of nice 0 or enough lower priority tasks to bring up the | |
793 | * weighted_cpuload | |
794 | */ | |
795 | static inline int above_background_load(void) | |
796 | { | |
797 | unsigned long cpu; | |
798 | ||
799 | for_each_online_cpu(cpu) { | |
800 | if (weighted_cpuload(cpu) >= SCHED_LOAD_SCALE) | |
801 | return 1; | |
802 | } | |
803 | return 0; | |
804 | } | |
1da177e4 LT |
805 | |
806 | struct io_context; /* See blkdev.h */ | |
1da177e4 LT |
807 | struct cpuset; |
808 | ||
809 | #define NGROUPS_SMALL 32 | |
810 | #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t))) | |
811 | struct group_info { | |
812 | int ngroups; | |
813 | atomic_t usage; | |
814 | gid_t small_block[NGROUPS_SMALL]; | |
815 | int nblocks; | |
816 | gid_t *blocks[0]; | |
817 | }; | |
818 | ||
819 | /* | |
820 | * get_group_info() must be called with the owning task locked (via task_lock()) | |
821 | * when task != current. The reason being that the vast majority of callers are | |
822 | * looking at current->group_info, which can not be changed except by the | |
823 | * current task. Changing current->group_info requires the task lock, too. | |
824 | */ | |
825 | #define get_group_info(group_info) do { \ | |
826 | atomic_inc(&(group_info)->usage); \ | |
827 | } while (0) | |
828 | ||
829 | #define put_group_info(group_info) do { \ | |
830 | if (atomic_dec_and_test(&(group_info)->usage)) \ | |
831 | groups_free(group_info); \ | |
832 | } while (0) | |
833 | ||
3e30148c DH |
834 | extern struct group_info *groups_alloc(int gidsetsize); |
835 | extern void groups_free(struct group_info *group_info); | |
836 | extern int set_current_groups(struct group_info *group_info); | |
837 | extern int groups_search(struct group_info *group_info, gid_t grp); | |
1da177e4 LT |
838 | /* access the groups "array" with this macro */ |
839 | #define GROUP_AT(gi, i) \ | |
840 | ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK]) | |
841 | ||
383f2835 | 842 | #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK |
36c8b586 | 843 | extern void prefetch_stack(struct task_struct *t); |
383f2835 KC |
844 | #else |
845 | static inline void prefetch_stack(struct task_struct *t) { } | |
846 | #endif | |
1da177e4 LT |
847 | |
848 | struct audit_context; /* See audit.c */ | |
849 | struct mempolicy; | |
b92ce558 | 850 | struct pipe_inode_info; |
4865ecf1 | 851 | struct uts_namespace; |
1da177e4 | 852 | |
20b8a59f IM |
853 | struct rq; |
854 | struct sched_domain; | |
855 | ||
856 | struct sched_class { | |
857 | struct sched_class *next; | |
858 | ||
fd390f6a | 859 | void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup); |
f02231e5 | 860 | void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep); |
4530d7ab | 861 | void (*yield_task) (struct rq *rq); |
20b8a59f IM |
862 | |
863 | void (*check_preempt_curr) (struct rq *rq, struct task_struct *p); | |
864 | ||
fb8d4724 | 865 | struct task_struct * (*pick_next_task) (struct rq *rq); |
31ee529c | 866 | void (*put_prev_task) (struct rq *rq, struct task_struct *p); |
20b8a59f | 867 | |
43010659 | 868 | unsigned long (*load_balance) (struct rq *this_rq, int this_cpu, |
20b8a59f IM |
869 | struct rq *busiest, |
870 | unsigned long max_nr_move, unsigned long max_load_move, | |
871 | struct sched_domain *sd, enum cpu_idle_type idle, | |
a4ac01c3 | 872 | int *all_pinned, int *this_best_prio); |
20b8a59f | 873 | |
83b699ed | 874 | void (*set_curr_task) (struct rq *rq); |
20b8a59f | 875 | void (*task_tick) (struct rq *rq, struct task_struct *p); |
ee0827d8 | 876 | void (*task_new) (struct rq *rq, struct task_struct *p); |
20b8a59f IM |
877 | }; |
878 | ||
879 | struct load_weight { | |
880 | unsigned long weight, inv_weight; | |
881 | }; | |
882 | ||
883 | /* | |
884 | * CFS stats for a schedulable entity (task, task-group etc) | |
885 | * | |
886 | * Current field usage histogram: | |
887 | * | |
888 | * 4 se->block_start | |
889 | * 4 se->run_node | |
890 | * 4 se->sleep_start | |
20b8a59f | 891 | * 6 se->load.weight |
20b8a59f IM |
892 | */ |
893 | struct sched_entity { | |
20b8a59f IM |
894 | struct load_weight load; /* for load-balancing */ |
895 | struct rb_node run_node; | |
896 | unsigned int on_rq; | |
897 | ||
94c18227 IM |
898 | u64 exec_start; |
899 | u64 sum_exec_runtime; | |
e9acbff6 | 900 | u64 vruntime; |
f6cf891c | 901 | u64 prev_sum_exec_runtime; |
94c18227 IM |
902 | |
903 | #ifdef CONFIG_SCHEDSTATS | |
20b8a59f | 904 | u64 wait_start; |
94c18227 | 905 | u64 wait_max; |
94c18227 | 906 | |
20b8a59f | 907 | u64 sleep_start; |
20b8a59f | 908 | u64 sleep_max; |
94c18227 IM |
909 | s64 sum_sleep_runtime; |
910 | ||
911 | u64 block_start; | |
20b8a59f IM |
912 | u64 block_max; |
913 | u64 exec_max; | |
eba1ed4b | 914 | u64 slice_max; |
94c18227 IM |
915 | #endif |
916 | ||
20b8a59f IM |
917 | #ifdef CONFIG_FAIR_GROUP_SCHED |
918 | struct sched_entity *parent; | |
919 | /* rq on which this entity is (to be) queued: */ | |
920 | struct cfs_rq *cfs_rq; | |
921 | /* rq "owned" by this entity/group: */ | |
922 | struct cfs_rq *my_q; | |
923 | #endif | |
924 | }; | |
70b97a7f | 925 | |
1da177e4 LT |
926 | struct task_struct { |
927 | volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */ | |
f7e4217b | 928 | void *stack; |
1da177e4 | 929 | atomic_t usage; |
97dc32cd WC |
930 | unsigned int flags; /* per process flags, defined below */ |
931 | unsigned int ptrace; | |
1da177e4 | 932 | |
36772092 | 933 | int lock_depth; /* BKL lock depth */ |
1da177e4 | 934 | |
2dd73a4f PW |
935 | #ifdef CONFIG_SMP |
936 | #ifdef __ARCH_WANT_UNLOCKED_CTXSW | |
4866cde0 NP |
937 | int oncpu; |
938 | #endif | |
2dd73a4f | 939 | #endif |
50e645a8 | 940 | |
b29739f9 | 941 | int prio, static_prio, normal_prio; |
1da177e4 | 942 | struct list_head run_list; |
20b8a59f IM |
943 | struct sched_class *sched_class; |
944 | struct sched_entity se; | |
1da177e4 | 945 | |
e107be36 AK |
946 | #ifdef CONFIG_PREEMPT_NOTIFIERS |
947 | /* list of struct preempt_notifier: */ | |
948 | struct hlist_head preempt_notifiers; | |
949 | #endif | |
950 | ||
22e2c507 | 951 | unsigned short ioprio; |
6c5c9341 | 952 | #ifdef CONFIG_BLK_DEV_IO_TRACE |
2056a782 | 953 | unsigned int btrace_seq; |
6c5c9341 | 954 | #endif |
1da177e4 | 955 | |
97dc32cd | 956 | unsigned int policy; |
1da177e4 | 957 | cpumask_t cpus_allowed; |
50e645a8 | 958 | unsigned int time_slice; |
1da177e4 | 959 | |
52f17b6c | 960 | #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) |
1da177e4 LT |
961 | struct sched_info sched_info; |
962 | #endif | |
963 | ||
964 | struct list_head tasks; | |
965 | /* | |
966 | * ptrace_list/ptrace_children forms the list of my children | |
967 | * that were stolen by a ptracer. | |
968 | */ | |
969 | struct list_head ptrace_children; | |
970 | struct list_head ptrace_list; | |
971 | ||
972 | struct mm_struct *mm, *active_mm; | |
973 | ||
974 | /* task state */ | |
975 | struct linux_binfmt *binfmt; | |
97dc32cd | 976 | int exit_state; |
1da177e4 LT |
977 | int exit_code, exit_signal; |
978 | int pdeath_signal; /* The signal sent when the parent dies */ | |
979 | /* ??? */ | |
97dc32cd | 980 | unsigned int personality; |
1da177e4 LT |
981 | unsigned did_exec:1; |
982 | pid_t pid; | |
983 | pid_t tgid; | |
0a425405 AV |
984 | |
985 | #ifdef CONFIG_CC_STACKPROTECTOR | |
986 | /* Canary value for the -fstack-protector gcc feature */ | |
987 | unsigned long stack_canary; | |
988 | #endif | |
1da177e4 LT |
989 | /* |
990 | * pointers to (original) parent process, youngest child, younger sibling, | |
991 | * older sibling, respectively. (p->father can be replaced with | |
992 | * p->parent->pid) | |
993 | */ | |
994 | struct task_struct *real_parent; /* real parent process (when being debugged) */ | |
995 | struct task_struct *parent; /* parent process */ | |
996 | /* | |
997 | * children/sibling forms the list of my children plus the | |
998 | * tasks I'm ptracing. | |
999 | */ | |
1000 | struct list_head children; /* list of my children */ | |
1001 | struct list_head sibling; /* linkage in my parent's children list */ | |
1002 | struct task_struct *group_leader; /* threadgroup leader */ | |
1003 | ||
1004 | /* PID/PID hash table linkage. */ | |
92476d7f | 1005 | struct pid_link pids[PIDTYPE_MAX]; |
47e65328 | 1006 | struct list_head thread_group; |
1da177e4 LT |
1007 | |
1008 | struct completion *vfork_done; /* for vfork() */ | |
1009 | int __user *set_child_tid; /* CLONE_CHILD_SETTID */ | |
1010 | int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */ | |
1011 | ||
97dc32cd | 1012 | unsigned int rt_priority; |
1da177e4 LT |
1013 | cputime_t utime, stime; |
1014 | unsigned long nvcsw, nivcsw; /* context switch counts */ | |
924b42d5 TJ |
1015 | struct timespec start_time; /* monotonic time */ |
1016 | struct timespec real_start_time; /* boot based time */ | |
1da177e4 LT |
1017 | /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */ |
1018 | unsigned long min_flt, maj_flt; | |
1019 | ||
1020 | cputime_t it_prof_expires, it_virt_expires; | |
1021 | unsigned long long it_sched_expires; | |
1022 | struct list_head cpu_timers[3]; | |
1023 | ||
1024 | /* process credentials */ | |
1025 | uid_t uid,euid,suid,fsuid; | |
1026 | gid_t gid,egid,sgid,fsgid; | |
1027 | struct group_info *group_info; | |
1028 | kernel_cap_t cap_effective, cap_inheritable, cap_permitted; | |
1029 | unsigned keep_capabilities:1; | |
1030 | struct user_struct *user; | |
1031 | #ifdef CONFIG_KEYS | |
b5f545c8 | 1032 | struct key *request_key_auth; /* assumed request_key authority */ |
1da177e4 | 1033 | struct key *thread_keyring; /* keyring private to this thread */ |
3e30148c | 1034 | unsigned char jit_keyring; /* default keyring to attach requested keys to */ |
1da177e4 | 1035 | #endif |
e07e23e1 AV |
1036 | /* |
1037 | * fpu_counter contains the number of consecutive context switches | |
1038 | * that the FPU is used. If this is over a threshold, the lazy fpu | |
1039 | * saving becomes unlazy to save the trap. This is an unsigned char | |
1040 | * so that after 256 times the counter wraps and the behavior turns | |
1041 | * lazy again; this to deal with bursty apps that only use FPU for | |
1042 | * a short time | |
1043 | */ | |
1044 | unsigned char fpu_counter; | |
1da177e4 | 1045 | int oomkilladj; /* OOM kill score adjustment (bit shift). */ |
36772092 PBG |
1046 | char comm[TASK_COMM_LEN]; /* executable name excluding path |
1047 | - access with [gs]et_task_comm (which lock | |
1048 | it with task_lock()) | |
1049 | - initialized normally by flush_old_exec */ | |
1da177e4 LT |
1050 | /* file system info */ |
1051 | int link_count, total_link_count; | |
3d5b6fcc | 1052 | #ifdef CONFIG_SYSVIPC |
1da177e4 LT |
1053 | /* ipc stuff */ |
1054 | struct sysv_sem sysvsem; | |
3d5b6fcc | 1055 | #endif |
1da177e4 LT |
1056 | /* CPU-specific state of this task */ |
1057 | struct thread_struct thread; | |
1058 | /* filesystem information */ | |
1059 | struct fs_struct *fs; | |
1060 | /* open file information */ | |
1061 | struct files_struct *files; | |
1651e14e | 1062 | /* namespaces */ |
ab516013 | 1063 | struct nsproxy *nsproxy; |
1da177e4 LT |
1064 | /* signal handlers */ |
1065 | struct signal_struct *signal; | |
1066 | struct sighand_struct *sighand; | |
1067 | ||
1068 | sigset_t blocked, real_blocked; | |
150256d8 | 1069 | sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */ |
1da177e4 LT |
1070 | struct sigpending pending; |
1071 | ||
1072 | unsigned long sas_ss_sp; | |
1073 | size_t sas_ss_size; | |
1074 | int (*notifier)(void *priv); | |
1075 | void *notifier_data; | |
1076 | sigset_t *notifier_mask; | |
1077 | ||
1078 | void *security; | |
1079 | struct audit_context *audit_context; | |
1080 | seccomp_t seccomp; | |
1081 | ||
1082 | /* Thread group tracking */ | |
1083 | u32 parent_exec_id; | |
1084 | u32 self_exec_id; | |
1085 | /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */ | |
1086 | spinlock_t alloc_lock; | |
1da177e4 | 1087 | |
b29739f9 IM |
1088 | /* Protection of the PI data structures: */ |
1089 | spinlock_t pi_lock; | |
1090 | ||
23f78d4a IM |
1091 | #ifdef CONFIG_RT_MUTEXES |
1092 | /* PI waiters blocked on a rt_mutex held by this task */ | |
1093 | struct plist_head pi_waiters; | |
1094 | /* Deadlock detection and priority inheritance handling */ | |
1095 | struct rt_mutex_waiter *pi_blocked_on; | |
23f78d4a IM |
1096 | #endif |
1097 | ||
408894ee IM |
1098 | #ifdef CONFIG_DEBUG_MUTEXES |
1099 | /* mutex deadlock detection */ | |
1100 | struct mutex_waiter *blocked_on; | |
1101 | #endif | |
de30a2b3 IM |
1102 | #ifdef CONFIG_TRACE_IRQFLAGS |
1103 | unsigned int irq_events; | |
1104 | int hardirqs_enabled; | |
1105 | unsigned long hardirq_enable_ip; | |
1106 | unsigned int hardirq_enable_event; | |
1107 | unsigned long hardirq_disable_ip; | |
1108 | unsigned int hardirq_disable_event; | |
1109 | int softirqs_enabled; | |
1110 | unsigned long softirq_disable_ip; | |
1111 | unsigned int softirq_disable_event; | |
1112 | unsigned long softirq_enable_ip; | |
1113 | unsigned int softirq_enable_event; | |
1114 | int hardirq_context; | |
1115 | int softirq_context; | |
1116 | #endif | |
fbb9ce95 IM |
1117 | #ifdef CONFIG_LOCKDEP |
1118 | # define MAX_LOCK_DEPTH 30UL | |
1119 | u64 curr_chain_key; | |
1120 | int lockdep_depth; | |
1121 | struct held_lock held_locks[MAX_LOCK_DEPTH]; | |
1122 | unsigned int lockdep_recursion; | |
1123 | #endif | |
408894ee | 1124 | |
1da177e4 LT |
1125 | /* journalling filesystem info */ |
1126 | void *journal_info; | |
1127 | ||
d89d8796 NB |
1128 | /* stacked block device info */ |
1129 | struct bio *bio_list, **bio_tail; | |
1130 | ||
1da177e4 LT |
1131 | /* VM state */ |
1132 | struct reclaim_state *reclaim_state; | |
1133 | ||
1da177e4 LT |
1134 | struct backing_dev_info *backing_dev_info; |
1135 | ||
1136 | struct io_context *io_context; | |
1137 | ||
1138 | unsigned long ptrace_message; | |
1139 | siginfo_t *last_siginfo; /* For ptrace use. */ | |
1140 | /* | |
1141 | * current io wait handle: wait queue entry to use for io waits | |
1142 | * If this thread is processing aio, this points at the waitqueue | |
1143 | * inside the currently handled kiocb. It may be NULL (i.e. default | |
1144 | * to a stack based synchronous wait) if its doing sync IO. | |
1145 | */ | |
1146 | wait_queue_t *io_wait; | |
4b98d11b | 1147 | #ifdef CONFIG_TASK_XACCT |
1da177e4 LT |
1148 | /* i/o counters(bytes read/written, #syscalls */ |
1149 | u64 rchar, wchar, syscr, syscw; | |
4b98d11b | 1150 | #endif |
7c3ab738 | 1151 | struct task_io_accounting ioac; |
8f0ab514 | 1152 | #if defined(CONFIG_TASK_XACCT) |
1da177e4 LT |
1153 | u64 acct_rss_mem1; /* accumulated rss usage */ |
1154 | u64 acct_vm_mem1; /* accumulated virtual memory usage */ | |
db5fed26 | 1155 | cputime_t acct_stimexpd;/* stime since last update */ |
1da177e4 LT |
1156 | #endif |
1157 | #ifdef CONFIG_NUMA | |
1158 | struct mempolicy *mempolicy; | |
1159 | short il_next; | |
1160 | #endif | |
1161 | #ifdef CONFIG_CPUSETS | |
1162 | struct cpuset *cpuset; | |
1163 | nodemask_t mems_allowed; | |
1164 | int cpuset_mems_generation; | |
825a46af | 1165 | int cpuset_mem_spread_rotor; |
1da177e4 | 1166 | #endif |
0771dfef | 1167 | struct robust_list_head __user *robust_list; |
34f192c6 IM |
1168 | #ifdef CONFIG_COMPAT |
1169 | struct compat_robust_list_head __user *compat_robust_list; | |
1170 | #endif | |
c87e2837 IM |
1171 | struct list_head pi_state_list; |
1172 | struct futex_pi_state *pi_state_cache; | |
0771dfef | 1173 | |
22e2c507 | 1174 | atomic_t fs_excl; /* holding fs exclusive resources */ |
e56d0903 | 1175 | struct rcu_head rcu; |
b92ce558 JA |
1176 | |
1177 | /* | |
1178 | * cache last used pipe for splice | |
1179 | */ | |
1180 | struct pipe_inode_info *splice_pipe; | |
ca74e92b SN |
1181 | #ifdef CONFIG_TASK_DELAY_ACCT |
1182 | struct task_delay_info *delays; | |
f4f154fd AM |
1183 | #endif |
1184 | #ifdef CONFIG_FAULT_INJECTION | |
1185 | int make_it_fail; | |
ca74e92b | 1186 | #endif |
1da177e4 LT |
1187 | }; |
1188 | ||
e05606d3 IM |
1189 | /* |
1190 | * Priority of a process goes from 0..MAX_PRIO-1, valid RT | |
1191 | * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH | |
1192 | * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority | |
1193 | * values are inverted: lower p->prio value means higher priority. | |
1194 | * | |
1195 | * The MAX_USER_RT_PRIO value allows the actual maximum | |
1196 | * RT priority to be separate from the value exported to | |
1197 | * user-space. This allows kernel threads to set their | |
1198 | * priority to a value higher than any user task. Note: | |
1199 | * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO. | |
1200 | */ | |
1201 | ||
1202 | #define MAX_USER_RT_PRIO 100 | |
1203 | #define MAX_RT_PRIO MAX_USER_RT_PRIO | |
1204 | ||
1205 | #define MAX_PRIO (MAX_RT_PRIO + 40) | |
1206 | #define DEFAULT_PRIO (MAX_RT_PRIO + 20) | |
1207 | ||
1208 | static inline int rt_prio(int prio) | |
1209 | { | |
1210 | if (unlikely(prio < MAX_RT_PRIO)) | |
1211 | return 1; | |
1212 | return 0; | |
1213 | } | |
1214 | ||
1215 | static inline int rt_task(struct task_struct *p) | |
1216 | { | |
1217 | return rt_prio(p->prio); | |
1218 | } | |
1219 | ||
1da177e4 LT |
1220 | static inline pid_t process_group(struct task_struct *tsk) |
1221 | { | |
1222 | return tsk->signal->pgrp; | |
1223 | } | |
1224 | ||
1ec320af CLG |
1225 | static inline pid_t signal_session(struct signal_struct *sig) |
1226 | { | |
1227 | return sig->__session; | |
1228 | } | |
1229 | ||
937949d9 CLG |
1230 | static inline pid_t process_session(struct task_struct *tsk) |
1231 | { | |
1ec320af CLG |
1232 | return signal_session(tsk->signal); |
1233 | } | |
1234 | ||
1235 | static inline void set_signal_session(struct signal_struct *sig, pid_t session) | |
1236 | { | |
1237 | sig->__session = session; | |
937949d9 CLG |
1238 | } |
1239 | ||
22c935f4 EB |
1240 | static inline struct pid *task_pid(struct task_struct *task) |
1241 | { | |
1242 | return task->pids[PIDTYPE_PID].pid; | |
1243 | } | |
1244 | ||
1245 | static inline struct pid *task_tgid(struct task_struct *task) | |
1246 | { | |
1247 | return task->group_leader->pids[PIDTYPE_PID].pid; | |
1248 | } | |
1249 | ||
1250 | static inline struct pid *task_pgrp(struct task_struct *task) | |
1251 | { | |
1252 | return task->group_leader->pids[PIDTYPE_PGID].pid; | |
1253 | } | |
1254 | ||
1255 | static inline struct pid *task_session(struct task_struct *task) | |
1256 | { | |
1257 | return task->group_leader->pids[PIDTYPE_SID].pid; | |
1258 | } | |
1259 | ||
1da177e4 LT |
1260 | /** |
1261 | * pid_alive - check that a task structure is not stale | |
1262 | * @p: Task structure to be checked. | |
1263 | * | |
1264 | * Test if a process is not yet dead (at most zombie state) | |
1265 | * If pid_alive fails, then pointers within the task structure | |
1266 | * can be stale and must not be dereferenced. | |
1267 | */ | |
1268 | static inline int pid_alive(struct task_struct *p) | |
1269 | { | |
92476d7f | 1270 | return p->pids[PIDTYPE_PID].pid != NULL; |
1da177e4 LT |
1271 | } |
1272 | ||
f400e198 | 1273 | /** |
3260259f HK |
1274 | * is_init - check if a task structure is init |
1275 | * @tsk: Task structure to be checked. | |
1276 | * | |
1277 | * Check if a task structure is the first user space task the kernel created. | |
f400e198 SB |
1278 | */ |
1279 | static inline int is_init(struct task_struct *tsk) | |
1280 | { | |
1281 | return tsk->pid == 1; | |
1282 | } | |
1283 | ||
9ec52099 CLG |
1284 | extern struct pid *cad_pid; |
1285 | ||
1da177e4 | 1286 | extern void free_task(struct task_struct *tsk); |
1da177e4 | 1287 | #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0) |
e56d0903 | 1288 | |
158d9ebd | 1289 | extern void __put_task_struct(struct task_struct *t); |
e56d0903 IM |
1290 | |
1291 | static inline void put_task_struct(struct task_struct *t) | |
1292 | { | |
1293 | if (atomic_dec_and_test(&t->usage)) | |
8c7904a0 | 1294 | __put_task_struct(t); |
e56d0903 | 1295 | } |
1da177e4 LT |
1296 | |
1297 | /* | |
1298 | * Per process flags | |
1299 | */ | |
1300 | #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */ | |
1301 | /* Not implemented yet, only for 486*/ | |
1302 | #define PF_STARTING 0x00000002 /* being created */ | |
1303 | #define PF_EXITING 0x00000004 /* getting shut down */ | |
778e9a9c | 1304 | #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ |
1da177e4 LT |
1305 | #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */ |
1306 | #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */ | |
1307 | #define PF_DUMPCORE 0x00000200 /* dumped core */ | |
1308 | #define PF_SIGNALED 0x00000400 /* killed by a signal */ | |
1309 | #define PF_MEMALLOC 0x00000800 /* Allocating memory */ | |
1310 | #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */ | |
1311 | #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */ | |
1da177e4 LT |
1312 | #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */ |
1313 | #define PF_FROZEN 0x00010000 /* frozen for system suspend */ | |
1314 | #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */ | |
1315 | #define PF_KSWAPD 0x00040000 /* I am kswapd */ | |
1316 | #define PF_SWAPOFF 0x00080000 /* I am in swapoff */ | |
1317 | #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */ | |
b31dc66a JA |
1318 | #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */ |
1319 | #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */ | |
1320 | #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */ | |
1321 | #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */ | |
1322 | #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */ | |
c61afb18 | 1323 | #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */ |
61a87122 | 1324 | #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */ |
ba96a0c8 | 1325 | #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */ |
1da177e4 LT |
1326 | |
1327 | /* | |
1328 | * Only the _current_ task can read/write to tsk->flags, but other | |
1329 | * tasks can access tsk->flags in readonly mode for example | |
1330 | * with tsk_used_math (like during threaded core dumping). | |
1331 | * There is however an exception to this rule during ptrace | |
1332 | * or during fork: the ptracer task is allowed to write to the | |
1333 | * child->flags of its traced child (same goes for fork, the parent | |
1334 | * can write to the child->flags), because we're guaranteed the | |
1335 | * child is not running and in turn not changing child->flags | |
1336 | * at the same time the parent does it. | |
1337 | */ | |
1338 | #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0) | |
1339 | #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0) | |
1340 | #define clear_used_math() clear_stopped_child_used_math(current) | |
1341 | #define set_used_math() set_stopped_child_used_math(current) | |
1342 | #define conditional_stopped_child_used_math(condition, child) \ | |
1343 | do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0) | |
1344 | #define conditional_used_math(condition) \ | |
1345 | conditional_stopped_child_used_math(condition, current) | |
1346 | #define copy_to_stopped_child_used_math(child) \ | |
1347 | do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0) | |
1348 | /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */ | |
1349 | #define tsk_used_math(p) ((p)->flags & PF_USED_MATH) | |
1350 | #define used_math() tsk_used_math(current) | |
1351 | ||
1352 | #ifdef CONFIG_SMP | |
36c8b586 | 1353 | extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask); |
1da177e4 | 1354 | #else |
36c8b586 | 1355 | static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask) |
1da177e4 | 1356 | { |
4098f991 | 1357 | if (!cpu_isset(0, new_mask)) |
1da177e4 LT |
1358 | return -EINVAL; |
1359 | return 0; | |
1360 | } | |
1361 | #endif | |
1362 | ||
1363 | extern unsigned long long sched_clock(void); | |
e436d800 IM |
1364 | |
1365 | /* | |
1366 | * For kernel-internal use: high-speed (but slightly incorrect) per-cpu | |
1367 | * clock constructed from sched_clock(): | |
1368 | */ | |
1369 | extern unsigned long long cpu_clock(int cpu); | |
1370 | ||
36c8b586 | 1371 | extern unsigned long long |
41b86e9c | 1372 | task_sched_runtime(struct task_struct *task); |
1da177e4 LT |
1373 | |
1374 | /* sched_exec is called by processes performing an exec */ | |
1375 | #ifdef CONFIG_SMP | |
1376 | extern void sched_exec(void); | |
1377 | #else | |
1378 | #define sched_exec() {} | |
1379 | #endif | |
1380 | ||
2aa44d05 IM |
1381 | extern void sched_clock_idle_sleep_event(void); |
1382 | extern void sched_clock_idle_wakeup_event(u64 delta_ns); | |
bb29ab26 | 1383 | |
1da177e4 LT |
1384 | #ifdef CONFIG_HOTPLUG_CPU |
1385 | extern void idle_task_exit(void); | |
1386 | #else | |
1387 | static inline void idle_task_exit(void) {} | |
1388 | #endif | |
1389 | ||
1390 | extern void sched_idle_next(void); | |
b29739f9 | 1391 | |
2bd8e6d4 | 1392 | #ifdef CONFIG_SCHED_DEBUG |
21805085 | 1393 | extern unsigned int sysctl_sched_latency; |
172ac3db | 1394 | extern unsigned int sysctl_sched_min_granularity; |
bf0f6f24 IM |
1395 | extern unsigned int sysctl_sched_wakeup_granularity; |
1396 | extern unsigned int sysctl_sched_batch_wakeup_granularity; | |
1397 | extern unsigned int sysctl_sched_stat_granularity; | |
1398 | extern unsigned int sysctl_sched_runtime_limit; | |
1399 | extern unsigned int sysctl_sched_child_runs_first; | |
1400 | extern unsigned int sysctl_sched_features; | |
2bd8e6d4 IM |
1401 | #endif |
1402 | ||
1403 | extern unsigned int sysctl_sched_compat_yield; | |
bf0f6f24 | 1404 | |
b29739f9 | 1405 | #ifdef CONFIG_RT_MUTEXES |
36c8b586 IM |
1406 | extern int rt_mutex_getprio(struct task_struct *p); |
1407 | extern void rt_mutex_setprio(struct task_struct *p, int prio); | |
1408 | extern void rt_mutex_adjust_pi(struct task_struct *p); | |
b29739f9 | 1409 | #else |
36c8b586 | 1410 | static inline int rt_mutex_getprio(struct task_struct *p) |
b29739f9 IM |
1411 | { |
1412 | return p->normal_prio; | |
1413 | } | |
95e02ca9 | 1414 | # define rt_mutex_adjust_pi(p) do { } while (0) |
b29739f9 IM |
1415 | #endif |
1416 | ||
36c8b586 IM |
1417 | extern void set_user_nice(struct task_struct *p, long nice); |
1418 | extern int task_prio(const struct task_struct *p); | |
1419 | extern int task_nice(const struct task_struct *p); | |
1420 | extern int can_nice(const struct task_struct *p, const int nice); | |
1421 | extern int task_curr(const struct task_struct *p); | |
1da177e4 LT |
1422 | extern int idle_cpu(int cpu); |
1423 | extern int sched_setscheduler(struct task_struct *, int, struct sched_param *); | |
36c8b586 IM |
1424 | extern struct task_struct *idle_task(int cpu); |
1425 | extern struct task_struct *curr_task(int cpu); | |
1426 | extern void set_curr_task(int cpu, struct task_struct *p); | |
1da177e4 LT |
1427 | |
1428 | void yield(void); | |
1429 | ||
1430 | /* | |
1431 | * The default (Linux) execution domain. | |
1432 | */ | |
1433 | extern struct exec_domain default_exec_domain; | |
1434 | ||
1435 | union thread_union { | |
1436 | struct thread_info thread_info; | |
1437 | unsigned long stack[THREAD_SIZE/sizeof(long)]; | |
1438 | }; | |
1439 | ||
1440 | #ifndef __HAVE_ARCH_KSTACK_END | |
1441 | static inline int kstack_end(void *addr) | |
1442 | { | |
1443 | /* Reliable end of stack detection: | |
1444 | * Some APM bios versions misalign the stack | |
1445 | */ | |
1446 | return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*))); | |
1447 | } | |
1448 | #endif | |
1449 | ||
1450 | extern union thread_union init_thread_union; | |
1451 | extern struct task_struct init_task; | |
1452 | ||
1453 | extern struct mm_struct init_mm; | |
1454 | ||
1455 | #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr) | |
1456 | extern struct task_struct *find_task_by_pid_type(int type, int pid); | |
1da177e4 LT |
1457 | extern void __set_special_pids(pid_t session, pid_t pgrp); |
1458 | ||
1459 | /* per-UID process charging. */ | |
acce292c | 1460 | extern struct user_struct * alloc_uid(struct user_namespace *, uid_t); |
1da177e4 LT |
1461 | static inline struct user_struct *get_uid(struct user_struct *u) |
1462 | { | |
1463 | atomic_inc(&u->__count); | |
1464 | return u; | |
1465 | } | |
1466 | extern void free_uid(struct user_struct *); | |
1467 | extern void switch_uid(struct user_struct *); | |
28f300d2 | 1468 | extern void release_uids(struct user_namespace *ns); |
1da177e4 LT |
1469 | |
1470 | #include <asm/current.h> | |
1471 | ||
3171a030 | 1472 | extern void do_timer(unsigned long ticks); |
1da177e4 LT |
1473 | |
1474 | extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state)); | |
1475 | extern int FASTCALL(wake_up_process(struct task_struct * tsk)); | |
1476 | extern void FASTCALL(wake_up_new_task(struct task_struct * tsk, | |
1477 | unsigned long clone_flags)); | |
1478 | #ifdef CONFIG_SMP | |
1479 | extern void kick_process(struct task_struct *tsk); | |
1480 | #else | |
1481 | static inline void kick_process(struct task_struct *tsk) { } | |
1482 | #endif | |
ad46c2c4 IM |
1483 | extern void sched_fork(struct task_struct *p, int clone_flags); |
1484 | extern void sched_dead(struct task_struct *p); | |
1da177e4 LT |
1485 | |
1486 | extern int in_group_p(gid_t); | |
1487 | extern int in_egroup_p(gid_t); | |
1488 | ||
1489 | extern void proc_caches_init(void); | |
1490 | extern void flush_signals(struct task_struct *); | |
10ab825b | 1491 | extern void ignore_signals(struct task_struct *); |
1da177e4 LT |
1492 | extern void flush_signal_handlers(struct task_struct *, int force_default); |
1493 | extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info); | |
1494 | ||
1495 | static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info) | |
1496 | { | |
1497 | unsigned long flags; | |
1498 | int ret; | |
1499 | ||
1500 | spin_lock_irqsave(&tsk->sighand->siglock, flags); | |
1501 | ret = dequeue_signal(tsk, mask, info); | |
1502 | spin_unlock_irqrestore(&tsk->sighand->siglock, flags); | |
1503 | ||
1504 | return ret; | |
1505 | } | |
1506 | ||
1507 | extern void block_all_signals(int (*notifier)(void *priv), void *priv, | |
1508 | sigset_t *mask); | |
1509 | extern void unblock_all_signals(void); | |
1510 | extern void release_task(struct task_struct * p); | |
1511 | extern int send_sig_info(int, struct siginfo *, struct task_struct *); | |
1512 | extern int send_group_sig_info(int, struct siginfo *, struct task_struct *); | |
1513 | extern int force_sigsegv(int, struct task_struct *); | |
1514 | extern int force_sig_info(int, struct siginfo *, struct task_struct *); | |
c4b92fc1 EB |
1515 | extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp); |
1516 | extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp); | |
1517 | extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid); | |
2425c08b | 1518 | extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32); |
c4b92fc1 EB |
1519 | extern int kill_pgrp(struct pid *pid, int sig, int priv); |
1520 | extern int kill_pid(struct pid *pid, int sig, int priv); | |
c3de4b38 | 1521 | extern int kill_proc_info(int, struct siginfo *, pid_t); |
1da177e4 LT |
1522 | extern void do_notify_parent(struct task_struct *, int); |
1523 | extern void force_sig(int, struct task_struct *); | |
1524 | extern void force_sig_specific(int, struct task_struct *); | |
1525 | extern int send_sig(int, struct task_struct *, int); | |
1526 | extern void zap_other_threads(struct task_struct *p); | |
1da177e4 LT |
1527 | extern int kill_proc(pid_t, int, int); |
1528 | extern struct sigqueue *sigqueue_alloc(void); | |
1529 | extern void sigqueue_free(struct sigqueue *); | |
1530 | extern int send_sigqueue(int, struct sigqueue *, struct task_struct *); | |
1531 | extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *); | |
9ac95f2f | 1532 | extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *); |
1da177e4 LT |
1533 | extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long); |
1534 | ||
9ec52099 CLG |
1535 | static inline int kill_cad_pid(int sig, int priv) |
1536 | { | |
1537 | return kill_pid(cad_pid, sig, priv); | |
1538 | } | |
1539 | ||
1da177e4 LT |
1540 | /* These can be the second arg to send_sig_info/send_group_sig_info. */ |
1541 | #define SEND_SIG_NOINFO ((struct siginfo *) 0) | |
1542 | #define SEND_SIG_PRIV ((struct siginfo *) 1) | |
1543 | #define SEND_SIG_FORCED ((struct siginfo *) 2) | |
1544 | ||
621d3121 ON |
1545 | static inline int is_si_special(const struct siginfo *info) |
1546 | { | |
1547 | return info <= SEND_SIG_FORCED; | |
1548 | } | |
1549 | ||
1da177e4 LT |
1550 | /* True if we are on the alternate signal stack. */ |
1551 | ||
1552 | static inline int on_sig_stack(unsigned long sp) | |
1553 | { | |
1554 | return (sp - current->sas_ss_sp < current->sas_ss_size); | |
1555 | } | |
1556 | ||
1557 | static inline int sas_ss_flags(unsigned long sp) | |
1558 | { | |
1559 | return (current->sas_ss_size == 0 ? SS_DISABLE | |
1560 | : on_sig_stack(sp) ? SS_ONSTACK : 0); | |
1561 | } | |
1562 | ||
1da177e4 LT |
1563 | /* |
1564 | * Routines for handling mm_structs | |
1565 | */ | |
1566 | extern struct mm_struct * mm_alloc(void); | |
1567 | ||
1568 | /* mmdrop drops the mm and the page tables */ | |
1569 | extern void FASTCALL(__mmdrop(struct mm_struct *)); | |
1570 | static inline void mmdrop(struct mm_struct * mm) | |
1571 | { | |
6fb43d7b | 1572 | if (unlikely(atomic_dec_and_test(&mm->mm_count))) |
1da177e4 LT |
1573 | __mmdrop(mm); |
1574 | } | |
1575 | ||
1576 | /* mmput gets rid of the mappings and all user-space */ | |
1577 | extern void mmput(struct mm_struct *); | |
1578 | /* Grab a reference to a task's mm, if it is not already going away */ | |
1579 | extern struct mm_struct *get_task_mm(struct task_struct *task); | |
1580 | /* Remove the current tasks stale references to the old mm_struct */ | |
1581 | extern void mm_release(struct task_struct *, struct mm_struct *); | |
1582 | ||
1583 | extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *); | |
1584 | extern void flush_thread(void); | |
1585 | extern void exit_thread(void); | |
1586 | ||
1da177e4 | 1587 | extern void exit_files(struct task_struct *); |
6b3934ef | 1588 | extern void __cleanup_signal(struct signal_struct *); |
a7e5328a | 1589 | extern void __cleanup_sighand(struct sighand_struct *); |
1da177e4 LT |
1590 | extern void exit_itimers(struct signal_struct *); |
1591 | ||
1592 | extern NORET_TYPE void do_group_exit(int); | |
1593 | ||
1da177e4 LT |
1594 | extern void daemonize(const char *, ...); |
1595 | extern int allow_signal(int); | |
1596 | extern int disallow_signal(int); | |
1da177e4 LT |
1597 | |
1598 | extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *); | |
1599 | extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *); | |
36c8b586 | 1600 | struct task_struct *fork_idle(int); |
1da177e4 LT |
1601 | |
1602 | extern void set_task_comm(struct task_struct *tsk, char *from); | |
1603 | extern void get_task_comm(char *to, struct task_struct *tsk); | |
1604 | ||
1605 | #ifdef CONFIG_SMP | |
36c8b586 | 1606 | extern void wait_task_inactive(struct task_struct * p); |
1da177e4 LT |
1607 | #else |
1608 | #define wait_task_inactive(p) do { } while (0) | |
1609 | #endif | |
1610 | ||
1611 | #define remove_parent(p) list_del_init(&(p)->sibling) | |
8fafabd8 | 1612 | #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children) |
1da177e4 | 1613 | |
5e85d4ab | 1614 | #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks) |
1da177e4 LT |
1615 | |
1616 | #define for_each_process(p) \ | |
1617 | for (p = &init_task ; (p = next_task(p)) != &init_task ; ) | |
1618 | ||
1619 | /* | |
1620 | * Careful: do_each_thread/while_each_thread is a double loop so | |
1621 | * 'break' will not work as expected - use goto instead. | |
1622 | */ | |
1623 | #define do_each_thread(g, t) \ | |
1624 | for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do | |
1625 | ||
1626 | #define while_each_thread(g, t) \ | |
1627 | while ((t = next_thread(t)) != g) | |
1628 | ||
de12a787 EB |
1629 | /* de_thread depends on thread_group_leader not being a pid based check */ |
1630 | #define thread_group_leader(p) (p == p->group_leader) | |
1da177e4 | 1631 | |
0804ef4b EB |
1632 | /* Do to the insanities of de_thread it is possible for a process |
1633 | * to have the pid of the thread group leader without actually being | |
1634 | * the thread group leader. For iteration through the pids in proc | |
1635 | * all we care about is that we have a task with the appropriate | |
1636 | * pid, we don't actually care if we have the right task. | |
1637 | */ | |
1638 | static inline int has_group_leader_pid(struct task_struct *p) | |
1639 | { | |
1640 | return p->pid == p->tgid; | |
1641 | } | |
1642 | ||
36c8b586 | 1643 | static inline struct task_struct *next_thread(const struct task_struct *p) |
47e65328 ON |
1644 | { |
1645 | return list_entry(rcu_dereference(p->thread_group.next), | |
36c8b586 | 1646 | struct task_struct, thread_group); |
47e65328 ON |
1647 | } |
1648 | ||
36c8b586 | 1649 | static inline int thread_group_empty(struct task_struct *p) |
1da177e4 | 1650 | { |
47e65328 | 1651 | return list_empty(&p->thread_group); |
1da177e4 LT |
1652 | } |
1653 | ||
1654 | #define delay_group_leader(p) \ | |
1655 | (thread_group_leader(p) && !thread_group_empty(p)) | |
1656 | ||
1da177e4 | 1657 | /* |
260ea101 | 1658 | * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring |
22e2c507 | 1659 | * subscriptions and synchronises with wait4(). Also used in procfs. Also |
053199ed | 1660 | * pins the final release of task.io_context. Also protects ->cpuset. |
1da177e4 LT |
1661 | * |
1662 | * Nests both inside and outside of read_lock(&tasklist_lock). | |
1663 | * It must not be nested with write_lock_irq(&tasklist_lock), | |
1664 | * neither inside nor outside. | |
1665 | */ | |
1666 | static inline void task_lock(struct task_struct *p) | |
1667 | { | |
1668 | spin_lock(&p->alloc_lock); | |
1669 | } | |
1670 | ||
1671 | static inline void task_unlock(struct task_struct *p) | |
1672 | { | |
1673 | spin_unlock(&p->alloc_lock); | |
1674 | } | |
1675 | ||
f63ee72e ON |
1676 | extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk, |
1677 | unsigned long *flags); | |
1678 | ||
1679 | static inline void unlock_task_sighand(struct task_struct *tsk, | |
1680 | unsigned long *flags) | |
1681 | { | |
1682 | spin_unlock_irqrestore(&tsk->sighand->siglock, *flags); | |
1683 | } | |
1684 | ||
f037360f AV |
1685 | #ifndef __HAVE_THREAD_FUNCTIONS |
1686 | ||
f7e4217b RZ |
1687 | #define task_thread_info(task) ((struct thread_info *)(task)->stack) |
1688 | #define task_stack_page(task) ((task)->stack) | |
a1261f54 | 1689 | |
10ebffde AV |
1690 | static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org) |
1691 | { | |
1692 | *task_thread_info(p) = *task_thread_info(org); | |
1693 | task_thread_info(p)->task = p; | |
1694 | } | |
1695 | ||
1696 | static inline unsigned long *end_of_stack(struct task_struct *p) | |
1697 | { | |
f7e4217b | 1698 | return (unsigned long *)(task_thread_info(p) + 1); |
10ebffde AV |
1699 | } |
1700 | ||
f037360f AV |
1701 | #endif |
1702 | ||
1da177e4 LT |
1703 | /* set thread flags in other task's structures |
1704 | * - see asm/thread_info.h for TIF_xxxx flags available | |
1705 | */ | |
1706 | static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1707 | { | |
a1261f54 | 1708 | set_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1709 | } |
1710 | ||
1711 | static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1712 | { | |
a1261f54 | 1713 | clear_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1714 | } |
1715 | ||
1716 | static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1717 | { | |
a1261f54 | 1718 | return test_and_set_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1719 | } |
1720 | ||
1721 | static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1722 | { | |
a1261f54 | 1723 | return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1724 | } |
1725 | ||
1726 | static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1727 | { | |
a1261f54 | 1728 | return test_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1729 | } |
1730 | ||
1731 | static inline void set_tsk_need_resched(struct task_struct *tsk) | |
1732 | { | |
1733 | set_tsk_thread_flag(tsk,TIF_NEED_RESCHED); | |
1734 | } | |
1735 | ||
1736 | static inline void clear_tsk_need_resched(struct task_struct *tsk) | |
1737 | { | |
1738 | clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED); | |
1739 | } | |
1740 | ||
1741 | static inline int signal_pending(struct task_struct *p) | |
1742 | { | |
1743 | return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING)); | |
1744 | } | |
1745 | ||
1746 | static inline int need_resched(void) | |
1747 | { | |
1748 | return unlikely(test_thread_flag(TIF_NEED_RESCHED)); | |
1749 | } | |
1750 | ||
1751 | /* | |
1752 | * cond_resched() and cond_resched_lock(): latency reduction via | |
1753 | * explicit rescheduling in places that are safe. The return | |
1754 | * value indicates whether a reschedule was done in fact. | |
1755 | * cond_resched_lock() will drop the spinlock before scheduling, | |
1756 | * cond_resched_softirq() will enable bhs before scheduling. | |
1757 | */ | |
1758 | extern int cond_resched(void); | |
1759 | extern int cond_resched_lock(spinlock_t * lock); | |
1760 | extern int cond_resched_softirq(void); | |
1761 | ||
1762 | /* | |
1763 | * Does a critical section need to be broken due to another | |
1764 | * task waiting?: | |
1765 | */ | |
1766 | #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP) | |
1767 | # define need_lockbreak(lock) ((lock)->break_lock) | |
1768 | #else | |
1769 | # define need_lockbreak(lock) 0 | |
1770 | #endif | |
1771 | ||
1772 | /* | |
1773 | * Does a critical section need to be broken due to another | |
1774 | * task waiting or preemption being signalled: | |
1775 | */ | |
1776 | static inline int lock_need_resched(spinlock_t *lock) | |
1777 | { | |
1778 | if (need_lockbreak(lock) || need_resched()) | |
1779 | return 1; | |
1780 | return 0; | |
1781 | } | |
1782 | ||
7bb44ade RM |
1783 | /* |
1784 | * Reevaluate whether the task has signals pending delivery. | |
1785 | * Wake the task if so. | |
1786 | * This is required every time the blocked sigset_t changes. | |
1787 | * callers must hold sighand->siglock. | |
1788 | */ | |
1789 | extern void recalc_sigpending_and_wake(struct task_struct *t); | |
1da177e4 LT |
1790 | extern void recalc_sigpending(void); |
1791 | ||
1792 | extern void signal_wake_up(struct task_struct *t, int resume_stopped); | |
1793 | ||
1794 | /* | |
1795 | * Wrappers for p->thread_info->cpu access. No-op on UP. | |
1796 | */ | |
1797 | #ifdef CONFIG_SMP | |
1798 | ||
1799 | static inline unsigned int task_cpu(const struct task_struct *p) | |
1800 | { | |
a1261f54 | 1801 | return task_thread_info(p)->cpu; |
1da177e4 LT |
1802 | } |
1803 | ||
c65cc870 | 1804 | extern void set_task_cpu(struct task_struct *p, unsigned int cpu); |
1da177e4 LT |
1805 | |
1806 | #else | |
1807 | ||
1808 | static inline unsigned int task_cpu(const struct task_struct *p) | |
1809 | { | |
1810 | return 0; | |
1811 | } | |
1812 | ||
1813 | static inline void set_task_cpu(struct task_struct *p, unsigned int cpu) | |
1814 | { | |
1815 | } | |
1816 | ||
1817 | #endif /* CONFIG_SMP */ | |
1818 | ||
1819 | #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT | |
1820 | extern void arch_pick_mmap_layout(struct mm_struct *mm); | |
1821 | #else | |
1822 | static inline void arch_pick_mmap_layout(struct mm_struct *mm) | |
1823 | { | |
1824 | mm->mmap_base = TASK_UNMAPPED_BASE; | |
1825 | mm->get_unmapped_area = arch_get_unmapped_area; | |
1826 | mm->unmap_area = arch_unmap_area; | |
1827 | } | |
1828 | #endif | |
1829 | ||
1830 | extern long sched_setaffinity(pid_t pid, cpumask_t new_mask); | |
1831 | extern long sched_getaffinity(pid_t pid, cpumask_t *mask); | |
1832 | ||
5c45bf27 | 1833 | extern int sched_mc_power_savings, sched_smt_power_savings; |
5c45bf27 | 1834 | |
1da177e4 LT |
1835 | extern void normalize_rt_tasks(void); |
1836 | ||
4b98d11b AD |
1837 | #ifdef CONFIG_TASK_XACCT |
1838 | static inline void add_rchar(struct task_struct *tsk, ssize_t amt) | |
1839 | { | |
1840 | tsk->rchar += amt; | |
1841 | } | |
1842 | ||
1843 | static inline void add_wchar(struct task_struct *tsk, ssize_t amt) | |
1844 | { | |
1845 | tsk->wchar += amt; | |
1846 | } | |
1847 | ||
1848 | static inline void inc_syscr(struct task_struct *tsk) | |
1849 | { | |
1850 | tsk->syscr++; | |
1851 | } | |
1852 | ||
1853 | static inline void inc_syscw(struct task_struct *tsk) | |
1854 | { | |
1855 | tsk->syscw++; | |
1856 | } | |
1857 | #else | |
1858 | static inline void add_rchar(struct task_struct *tsk, ssize_t amt) | |
1859 | { | |
1860 | } | |
1861 | ||
1862 | static inline void add_wchar(struct task_struct *tsk, ssize_t amt) | |
1863 | { | |
1864 | } | |
1865 | ||
1866 | static inline void inc_syscr(struct task_struct *tsk) | |
1867 | { | |
1868 | } | |
1869 | ||
1870 | static inline void inc_syscw(struct task_struct *tsk) | |
1871 | { | |
1872 | } | |
1873 | #endif | |
1874 | ||
1da177e4 LT |
1875 | #endif /* __KERNEL__ */ |
1876 | ||
1877 | #endif |