1 /*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1986, 1989, 1991, 1993
5 * The Regents of the University of California. All rights reserved.
6 * (c) UNIX System Laboratories, Inc.
7 * All or some portions of this file are derived from material licensed
8 * to the University of California by American Telephone and Telegraph
9 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
10 * the permission of UNIX System Laboratories, Inc.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
20 * 3. Neither the name of the University nor the names of its contributors
21 * may be used to endorse or promote products derived from this software
22 * without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 */
36
37 #ifndef _SYS_PROC_H_
38 #define _SYS_PROC_H_
39
40 #include <sys/callout.h> /* For struct callout. */
41 #include <sys/event.h> /* For struct klist. */
42 #ifdef _KERNEL
43 #include <sys/_eventhandler.h>
44 #endif
45 #include <sys/condvar.h>
46 #ifndef _KERNEL
47 #include <sys/filedesc.h>
48 #endif
49 #include <sys/queue.h>
50 #include <sys/_lock.h>
51 #include <sys/lock_profile.h>
52 #include <sys/_mutex.h>
53 #include <sys/osd.h>
54 #include <sys/priority.h>
55 #include <sys/rtprio.h> /* XXX. */
56 #include <sys/runq.h>
57 #include <sys/resource.h>
58 #include <sys/sigio.h>
59 #include <sys/signal.h>
60 #include <sys/signalvar.h>
61 #ifndef _KERNEL
62 #include <sys/time.h> /* For structs itimerval, timeval. */
63 #else
64 #include <sys/pcpu.h>
65 #include <sys/systm.h>
66 #endif
67 #include <sys/ucontext.h>
68 #include <sys/ucred.h>
69 #include <sys/types.h>
70 #include <sys/_domainset.h>
71
72 #include <machine/proc.h> /* Machine-dependent proc substruct. */
73 #ifdef _KERNEL
74 #include <machine/cpu.h>
75 #endif
76
77 /*
78 * One structure allocated per session.
79 *
80 * List of locks
81 * (m) locked by s_mtx mtx
82 * (e) locked by proctree_lock sx
83 * (c) const until freeing
84 */
85 struct session {
86 u_int s_count; /* Ref cnt; pgrps in session - atomic. */
87 struct proc *s_leader; /* (m + e) Session leader. */
88 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
89 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */
90 struct tty *s_ttyp; /* (e) Controlling tty. */
91 pid_t s_sid; /* (c) Session ID. */
92 /* (m) Setlogin() name: */
93 char s_login[roundup(MAXLOGNAME, sizeof(long))];
94 struct mtx s_mtx; /* Mutex to protect members. */
95 };
96
97 /*
98 * One structure allocated per process group.
99 *
100 * List of locks
101 * (m) locked by pg_mtx mtx
102 * (e) locked by proctree_lock sx
103 * (c) const until freeing
104 */
105 struct pgrp {
106 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
107 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
108 struct session *pg_session; /* (c) Pointer to session. */
109 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
110 pid_t pg_id; /* (c) Process group id. */
111 struct mtx pg_mtx; /* Mutex to protect members */
112 int pg_flags; /* (m) PGRP_ flags */
113 struct sx pg_killsx; /* Mutual exclusion between group member
114 * fork() and killpg() */
115 };
116
117 #define PGRP_ORPHANED 0x00000001 /* Group is orphaned */
118
119 /*
120 * pargs, used to hold a copy of the command line, if it had a sane length.
121 */
122 struct pargs {
123 u_int ar_ref; /* Reference count. */
124 u_int ar_length; /* Length. */
125 u_char ar_args[1]; /* Arguments. */
126 };
127
128 /*-
129 * Description of a process.
130 *
131 * This structure contains the information needed to manage a thread of
132 * control, known in UN*X as a process; it has references to substructures
133 * containing descriptions of things that the process uses, but may share
134 * with related processes. The process structure and the substructures
135 * are always addressable except for those marked "(CPU)" below,
136 * which might be addressable only on a processor on which the process
137 * is running.
138 *
139 * Below is a key of locks used to protect each member of struct proc. The
140 * lock is indicated by a reference to a specific character in parens in the
141 * associated comment.
142 * * - not yet protected
143 * a - only touched by curproc or parent during fork/wait
144 * b - created at fork, never changes
145 * (exception aiods switch vmspaces, but they are also
146 * marked 'P_SYSTEM' so hopefully it will be left alone)
147 * c - locked by proc mtx
148 * d - locked by allproc_lock lock
149 * e - locked by proctree_lock lock
150 * f - session mtx
151 * g - process group mtx
152 * h - callout_lock mtx
153 * i - by curproc or the master session mtx
154 * j - locked by proc slock
155 * k - only accessed by curthread
156 * k*- only accessed by curthread and from an interrupt
157 * kx- only accessed by curthread and by debugger
158 * l - the attaching proc or attaching proc parent
159 * n - not locked, lazy
160 * o - ktrace lock
161 * q - td_contested lock
162 * r - p_peers lock
163 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9)
164 * t - thread lock
165 * u - process stat lock
166 * w - process timer lock
167 * x - created at fork, only changes during single threading in exec
168 * y - created at first aio, doesn't change until exit or exec at which
169 * point we are single-threaded and only curthread changes it
170 *
171 * If the locking key specifies two identifiers (for example, p_pptr) then
172 * either lock is sufficient for read access, but both locks must be held
173 * for write access.
174 */
175 struct cpuset;
176 struct filecaps;
177 struct filemon;
178 struct kaioinfo;
179 struct kaudit_record;
180 struct kcov_info;
181 struct kdtrace_proc;
182 struct kdtrace_thread;
183 struct kmsan_td;
184 struct kq_timer_cb_data;
185 struct mqueue_notifier;
186 struct p_sched;
187 struct proc;
188 struct procdesc;
189 struct racct;
190 struct sbuf;
191 struct sleepqueue;
192 struct socket;
193 struct td_sched;
194 struct thread;
195 struct trapframe;
196 struct turnstile;
197 struct vm_map;
198 struct vm_map_entry;
199 struct epoch_tracker;
200
201 struct syscall_args {
202 u_int code;
203 u_int original_code;
204 struct sysent *callp;
205 register_t args[8];
206 };
207
208 /*
209 * XXX: Does this belong in resource.h or resourcevar.h instead?
210 * Resource usage extension. The times in rusage structs in the kernel are
211 * never up to date. The actual times are kept as runtimes and tick counts
212 * (with control info in the "previous" times), and are converted when
213 * userland asks for rusage info. Backwards compatibility prevents putting
214 * this directly in the user-visible rusage struct.
215 *
216 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
217 * Locking for td_rux: (t) for all fields.
218 */
219 struct rusage_ext {
220 uint64_t rux_runtime; /* (cu) Real time. */
221 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */
222 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */
223 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */
224 uint64_t rux_uu; /* (c) Previous user time in usec. */
225 uint64_t rux_su; /* (c) Previous sys time in usec. */
226 uint64_t rux_tu; /* (c) Previous total time in usec. */
227 };
228
229 /*
230 * Kernel runnable context (thread).
231 * This is what is put to sleep and reactivated.
232 * Thread context. Processes may have multiple threads.
233 */
234 struct thread {
235 struct mtx *volatile td_lock; /* replaces sched lock */
236 struct proc *td_proc; /* (*) Associated process. */
237 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
238 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */
239 union {
240 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */
241 struct thread *td_zombie; /* Zombie list linkage */
242 };
243 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */
244 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */
245 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */
246 struct domainset_ref td_domain; /* (a) NUMA policy */
247 struct seltd *td_sel; /* Select queue/channel. */
248 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
249 struct turnstile *td_turnstile; /* (k) Associated turnstile. */
250 void *td_pad1; /* Available */
251 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
252 lwpid_t td_tid; /* (b) Thread ID. */
253 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
254 #define td_siglist td_sigqueue.sq_signals
255 u_char td_lend_user_pri; /* (t) Lend user pri. */
256 u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */
257 u_char td_base_ithread_pri; /* (t) Base ithread pri */
258 struct kmsan_td *td_kmsan; /* (k) KMSAN state */
259
260 /* Cleared during fork1(), thread_create(), or kthread_add(). */
261 #define td_startzero td_flags
262 int td_flags; /* (t) TDF_* flags. */
263 int td_ast; /* (t) TDA_* indicators */
264 int td_inhibitors; /* (t) Why can not run. */
265 int td_pflags; /* (k) Private thread (TDP_*) flags. */
266 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */
267 int td_dupfd; /* (k) Ret value from fdopen. XXX */
268 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */
269 const void *td_wchan; /* (t) Sleep address. */
270 const char *td_wmesg; /* (t) Reason for sleep. */
271 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
272 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */
273 u_char _td_pad0[2]; /* Available. */
274 int td_locks; /* (k) Debug: count of non-spin locks */
275 int td_rw_rlocks; /* (k) Count of rwlock read locks. */
276 int td_sx_slocks; /* (k) Count of sx shared locks. */
277 int td_lk_slocks; /* (k) Count of lockmgr shared locks. */
278 struct lock_object *td_wantedlock; /* (k) Lock we are contending on */
279 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */
280 const char *td_lockname; /* (t) Name of lock blocked on. */
281 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
282 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
283 int td_intr_nesting_level; /* (k) Interrupt recursion. */
284 int td_pinned; /* (k) Temporary cpu pin count. */
285 struct ucred *td_realucred; /* (k) Reference to credentials. */
286 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */
287 struct plimit *td_limit; /* (k) Resource limits. */
288 int td_slptick; /* (t) Time at sleep. */
289 int td_blktick; /* (t) Time spent blocked. */
290 int td_swvoltick; /* (t) Time at last SW_VOL switch. */
291 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */
292 u_int td_cow; /* (*) Number of copy-on-write faults */
293 struct rusage td_ru; /* (t) rusage information. */
294 struct rusage_ext td_rux; /* (t) Internal rusage information. */
295 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */
296 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */
297 u_int td_pticks; /* (t) Statclock hits for profiling */
298 u_int td_sticks; /* (t) Statclock hits in system mode. */
299 u_int td_iticks; /* (t) Statclock hits in intr mode. */
300 u_int td_uticks; /* (t) Statclock hits in user mode. */
301 int td_intrval; /* (t) Return value for sleepq. */
302 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
303 volatile u_int td_generation; /* (k) For detection of preemption */
304 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
305 int td_xsig; /* (c) Signal for ptrace */
306 u_long td_profil_addr; /* (k) Temporary addr until AST. */
307 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
308 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
309 struct file *td_fpop; /* (k) file referencing cdev under op */
310 int td_dbgflags; /* (c) Userland debugger flags */
311 siginfo_t td_si; /* (c) For debugger or core file */
312 int td_ng_outbound; /* (k) Thread entered ng from above. */
313 struct osd td_osd; /* (k) Object specific data. */
314 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
315 pid_t td_dbg_forked; /* (c) Child pid for debugger. */
316 u_int td_no_sleeping; /* (k) Sleeping disabled count. */
317 struct vnode *td_vp_reserved;/* (k) Preallocated vnode. */
318 void *td_su; /* (k) FFS SU private */
319 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */
320 int td_rtcgen; /* (s) rtc_generation of abs. sleep */
321 int td_errno; /* (k) Error from last syscall. */
322 size_t td_vslock_sz; /* (k) amount of vslock-ed space */
323 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */
324 long td_ucredref; /* (k) references on td_realucred */
325 #define td_endzero td_sigmask
326
327 /* Copied during fork1(), thread_create(), or kthread_add(). */
328 #define td_startcopy td_endzero
329 sigset_t td_sigmask; /* (c) Current signal mask. */
330 u_char td_rqindex; /* (t) Run queue index. */
331 u_char td_base_pri; /* (t) Thread base kernel priority. */
332 u_char td_priority; /* (t) Thread active priority. */
333 u_char td_pri_class; /* (t) Scheduling class. */
334 u_char td_user_pri; /* (t) User pri from estcpu and nice. */
335 u_char td_base_user_pri; /* (t) Base user pri */
336 uintptr_t td_rb_list; /* (k) Robust list head. */
337 uintptr_t td_rbp_list; /* (k) Robust priv list head. */
338 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */
339 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on
340 fork for child tracing. */
341 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */
342 uint32_t td_sigblock_val; /* (k) fast sigblock value read at
343 td_sigblock_ptr on kern entry */
344 #define td_endcopy td_pcb
345
346 /*
347 * Fields that must be manually set in fork1(), thread_create(), kthread_add(),
348 * or already have been set in the allocator, constructor, etc.
349 */
350 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
351 enum td_states {
352 TDS_INACTIVE = 0x0,
353 TDS_INHIBITED,
354 TDS_CAN_RUN,
355 TDS_RUNQ,
356 TDS_RUNNING
357 } td_state; /* (t) thread state */
358 /* Note: td_state must be accessed using TD_{GET,SET}_STATE(). */
359 union {
360 syscallarg_t tdu_retval[2];
361 off_t tdu_off;
362 } td_uretoff; /* (k) Syscall aux returns. */
363 #define td_retval td_uretoff.tdu_retval
364 u_int td_cowgen; /* (k) Generation of COW pointers. */
365 /* LP64 hole */
366 struct callout td_slpcallout; /* (h) Callout for sleep. */
367 struct trapframe *td_frame; /* (k) */
368 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
369 u_short td_kstack_pages; /* (a) Size of the kstack. */
370 u_short td_kstack_domain; /* (a) Domain backing kstack KVA. */
371 volatile u_int td_critnest; /* (k*) Critical section nest level. */
372 struct mdthread td_md; /* (k) Any machine-dependent fields. */
373 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
374 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */
375 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */
376 struct vnet *td_vnet; /* (k) Effective vnet. */
377 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */
378 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
379 struct proc *td_rfppwait_p; /* (k) The vforked child */
380 struct vm_page **td_ma; /* (k) uio pages held */
381 int td_ma_cnt; /* (k) size of *td_ma */
382 /* LP64 hole */
383 void *td_emuldata; /* Emulator state data */
384 int td_lastcpu; /* (t) Last cpu we were on. */
385 int td_oncpu; /* (t) Which cpu we are on. */
386 void *td_lkpi_task; /* LinuxKPI task struct pointer */
387 int td_pmcpend;
388 void *td_remotereq; /* (c) dbg remote request. */
389 off_t td_ktr_io_lim; /* (k) limit for ktrace file size */
390 #ifdef EPOCH_TRACE
391 SLIST_HEAD(, epoch_tracker) td_epochs;
392 #endif
393 };
394
395 struct thread0_storage {
396 struct thread t0st_thread;
397 uint64_t t0st_sched[10];
398 };
399
400 struct mtx *thread_lock_block(struct thread *);
401 void thread_lock_block_wait(struct thread *);
402 void thread_lock_set(struct thread *, struct mtx *);
403 void thread_lock_unblock(struct thread *, struct mtx *);
404 #define THREAD_LOCK_ASSERT(td, type) \
405 mtx_assert((td)->td_lock, (type))
406
407 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \
408 do { \
409 struct mtx *__m = (td)->td_lock; \
410 if (__m != &blocked_lock) \
411 mtx_assert(__m, (type)); \
412 } while (0)
413
414 #ifdef INVARIANTS
415 #define THREAD_LOCKPTR_ASSERT(td, lock) \
416 do { \
417 struct mtx *__m; \
418 __m = (td)->td_lock; \
419 KASSERT(__m == (lock), \
420 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
421 } while (0)
422
423 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \
424 do { \
425 struct mtx *__m; \
426 __m = (td)->td_lock; \
427 KASSERT(__m == (lock) || __m == &blocked_lock, \
428 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
429 } while (0)
430
431 #define TD_LOCKS_INC(td) ((td)->td_locks++)
432 #define TD_LOCKS_DEC(td) do { \
433 KASSERT(SCHEDULER_STOPPED() || (td)->td_locks > 0, \
434 ("Thread %p owns no locks", (td))); \
435 (td)->td_locks--; \
436 } while (0)
437 #else
438 #define THREAD_LOCKPTR_ASSERT(td, lock)
439 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock)
440
441 #define TD_LOCKS_INC(td)
442 #define TD_LOCKS_DEC(td)
443 #endif
444
445 /*
446 * Flags kept in td_flags:
447 * To change these you MUST have the scheduler lock.
448 */
449 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
450 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
451 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */
452 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
453 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
454 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
455 #define TDF_UNUSED11 0x00000040 /* Available */
456 #define TDF_SIGWAIT 0x00000080 /* Ignore ignored signals */
457 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */
458 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */
459 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
460 #define TDF_UNUSED1 0x00000800 /* Available */
461 #define TDF_UNUSED2 0x00001000 /* Available */
462 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */
463 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
464 #define TDF_UNUSED3 0x00008000 /* Available */
465 #define TDF_UNUSED4 0x00010000 /* Available */
466 #define TDF_UNUSED5 0x00020000 /* Available */
467 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */
468 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */
469 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
470 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */
471 #define TDF_UNUSED12 0x00400000 /* Available */
472 #define TDF_UNUSED6 0x00800000 /* Available */
473 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
474 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
475 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
476 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
477 #define TDF_UNUSED7 0x10000000 /* Available */
478 #define TDF_UNUSED8 0x20000000 /* Available */
479 #define TDF_UNUSED9 0x40000000 /* Available */
480 #define TDF_UNUSED10 0x80000000 /* Available */
481
482 enum {
483 TDA_AST = 0, /* Special: call all non-flagged AST handlers */
484 TDA_OWEUPC,
485 TDA_HWPMC,
486 TDA_VFORK,
487 TDA_ALRM,
488 TDA_PROF,
489 TDA_MAC,
490 TDA_SCHED,
491 TDA_UFS,
492 TDA_GEOM,
493 TDA_KQUEUE,
494 TDA_RACCT,
495 TDA_MOD1, /* For third party use, before signals are */
496 TAD_MOD2, /* processed .. */
497 TDA_PSELECT, /* For discarding temporary signal mask */
498 TDA_SIG,
499 TDA_KTRACE,
500 TDA_SUSPEND,
501 TDA_SIGSUSPEND,
502 TDA_MOD3, /* .. and after */
503 TAD_MOD4,
504 TDA_MAX,
505 };
506 #define TDAI(tda) (1U << (tda))
507 #define td_ast_pending(td, tda) ((td->td_ast & TDAI(tda)) != 0)
508
509 /* Userland debug flags */
510 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */
511 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */
512 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */
513 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */
514 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */
515 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */
516 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new
517 process */
518 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child
519 only) */
520 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */
521 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */
522 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */
523 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */
524 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */
525 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */
526 #define TDB_SSWITCH 0x00004000 /* Suspended in ptracestop */
527 #define TDB_BOUNDARY 0x00008000 /* ptracestop() at boundary */
528 #define TDB_COREDUMPREQ 0x00010000 /* Coredump request */
529 #define TDB_SCREMOTEREQ 0x00020000 /* Remote syscall request */
530
531 /*
532 * "Private" flags kept in td_pflags:
533 * These are only written by curthread and thus need no locking.
534 */
535 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
536 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
537 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
538 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */
539 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
540 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
541 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */
542 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */
543 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */
544 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
545 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
546 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */
547 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
548 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
549 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
550 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
551 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
552 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
553 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
554 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
555 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */
556 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */
557 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */
558 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
559 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */
560 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */
561 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */
562 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */
563 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */
564 #define TDP_EFIRT 0x20000000 /* In firmware (EFI RT) call */
565 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */
566 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */
567
568 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */
569 #define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */
570 #define TDP2_ACCT 0x00000004 /* Doing accounting */
571 #define TDP2_SAN_QUIET 0x00000008 /* Disable warnings from K(A|M)SAN */
572
573 /*
574 * Reasons that the current thread can not be run yet.
575 * More than one may apply.
576 */
577 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */
578 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
579 #define TDI_LOCK 0x0008 /* Stopped on a lock. */
580 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
581
582 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
583 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
584 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
585 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
586 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
587 #ifdef _KERNEL
588 #define TD_GET_STATE(td) atomic_load_int(&(td)->td_state)
589 #else
590 #define TD_GET_STATE(td) ((td)->td_state)
591 #endif
592 #define TD_IS_RUNNING(td) (TD_GET_STATE(td) == TDS_RUNNING)
593 #define TD_ON_RUNQ(td) (TD_GET_STATE(td) == TDS_RUNQ)
594 #define TD_CAN_RUN(td) (TD_GET_STATE(td) == TDS_CAN_RUN)
595 #define TD_IS_INHIBITED(td) (TD_GET_STATE(td) == TDS_INHIBITED)
596 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
597 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD)
598
599 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \
600 ((td)->td_flags & TDF_SINTR) != 0)
601
602 #define KTDSTATE(td) \
603 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \
604 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \
605 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \
606 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding")
607
608 #define TD_SET_INHIB(td, inhib) do { \
609 TD_SET_STATE(td, TDS_INHIBITED); \
610 (td)->td_inhibitors |= (inhib); \
611 } while (0)
612
613 #define TD_CLR_INHIB(td, inhib) do { \
614 if (((td)->td_inhibitors & (inhib)) && \
615 (((td)->td_inhibitors &= ~(inhib)) == 0)) \
616 TD_SET_STATE(td, TDS_CAN_RUN); \
617 } while (0)
618
619 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
620 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
621 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
622 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
623 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
624
625 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
626 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
627 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
628 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
629
630 #ifdef _KERNEL
631 #define TD_SET_STATE(td, state) atomic_store_int(&(td)->td_state, state)
632 #else
633 #define TD_SET_STATE(td, state) (td)->td_state = state
634 #endif
635 #define TD_SET_RUNNING(td) TD_SET_STATE(td, TDS_RUNNING)
636 #define TD_SET_RUNQ(td) TD_SET_STATE(td, TDS_RUNQ)
637 #define TD_SET_CAN_RUN(td) TD_SET_STATE(td, TDS_CAN_RUN)
638
639
640 #define TD_SBDRY_INTR(td) \
641 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0)
642 #define TD_SBDRY_ERRNO(td) \
643 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART)
644
645 /*
646 * Process structure.
647 */
648 struct proc {
649 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
650 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */
651 struct mtx p_slock; /* process spin lock */
652 struct ucred *p_ucred; /* (c) Process owner's identity. */
653 struct filedesc *p_fd; /* (b) Open files. */
654 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
655 struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */
656 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
657 struct plimit *p_limit; /* (c) Resource limits. */
658 struct callout p_limco; /* (c) Limit callout handle */
659 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
660
661 int p_flag; /* (c) P_* flags. */
662 int p_flag2; /* (c) P2_* flags. */
663 enum p_states {
664 PRS_NEW = 0, /* In creation */
665 PRS_NORMAL, /* threads can be run. */
666 PRS_ZOMBIE
667 } p_state; /* (j/c) Process status. */
668 pid_t p_pid; /* (b) Process identifier. */
669 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
670 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
671 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
672 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
673 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
674 struct proc *p_reaper; /* (e) My reaper. */
675 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants
676 (if I am reaper). */
677 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of
678 the same reaper. */
679 struct mtx p_mtx; /* (n) Lock for this struct. */
680 struct mtx p_statmtx; /* Lock for the stats */
681 struct mtx p_itimmtx; /* Lock for the virt/prof timers */
682 struct mtx p_profmtx; /* Lock for the profiling */
683 struct ksiginfo *p_ksi; /* Locked by parent proc lock */
684 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
685 #define p_siglist p_sigqueue.sq_signals
686 pid_t p_oppid; /* (c + e) Real parent pid. */
687
688 /* The following fields are all zeroed upon creation in fork. */
689 #define p_startzero p_vmspace
690 struct vmspace *p_vmspace; /* (b) Address space. */
691 u_int p_swtick; /* (c) Tick when swapped in or out. */
692 u_int p_cowgen; /* (c) Generation of COW pointers. */
693 struct itimerval p_realtimer; /* (c) Alarm timer. */
694 struct rusage p_ru; /* (a) Exit information. */
695 struct rusage_ext p_rux; /* (cu) Internal resource usage. */
696 struct rusage_ext p_crux; /* (c) Internal child resource usage. */
697 int p_profthreads; /* (c) Num threads in addupc_task. */
698 volatile int p_exitthreads; /* (j) Number of threads exiting */
699 int p_traceflag; /* (o) Kernel trace points. */
700 struct ktr_io_params *p_ktrioparms; /* (c + o) Params for ktrace. */
701 struct vnode *p_textvp; /* (b) Vnode of executable. */
702 struct vnode *p_textdvp; /* (b) Dir containing textvp. */
703 char *p_binname; /* (b) Binary hardlink name. */
704 u_int p_lock; /* (c) Prevent exit. */
705 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
706 int p_sigparent; /* (c) Signal to parent on exit. */
707 int p_sig; /* (n) For core dump/debugger XXX. */
708 u_int p_ptevents; /* (c + e) ptrace() event mask. */
709 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */
710 struct thread *p_singlethread;/* (c + j) If single threading this is it */
711 int p_suspcount; /* (j) Num threads in suspended mode. */
712 struct thread *p_xthread; /* (c) Trap thread */
713 int p_boundary_count;/* (j) Num threads at user boundary */
714 int p_pendingcnt; /* (c) how many signals are pending */
715 struct itimers *p_itimers; /* (c) POSIX interval timers. */
716 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */
717 u_int p_treeflag; /* (e) P_TREE flags */
718 int p_pendingexits; /* (c) Count of pending thread exits. */
719 struct filemon *p_filemon; /* (c) filemon-specific data. */
720 int p_pdeathsig; /* (c) Signal from parent on exit. */
721 /* End area that is zeroed on creation. */
722 #define p_endzero p_magic
723
724 /* The following fields are all copied upon creation in fork. */
725 #define p_startcopy p_endzero
726 u_int p_magic; /* (b) Magic number. */
727 int p_osrel; /* (x) osreldate for the
728 binary (from ELF note, if any) */
729 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */
730 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */
731 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
732 struct pargs *p_args; /* (c) Process arguments. */
733 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */
734 signed char p_nice; /* (c) Process "nice" value. */
735 int p_fibnum; /* in this routing domain XXX MRT */
736 pid_t p_reapsubtree; /* (e) Pid of the direct child of the
737 reaper which spawned
738 our subtree. */
739 uint64_t p_elf_flags; /* (x) ELF flags */
740 void *p_elf_brandinfo; /* (x) Elf_Brandinfo, NULL for
741 non ELF binaries. */
742 sbintime_t p_umtx_min_timeout;
743 /* End area that is copied on creation. */
744 #define p_endcopy p_xexit
745
746 u_int p_xexit; /* (c) Exit code. */
747 u_int p_xsig; /* (c) Stop/kill sig. */
748 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
749 struct knlist *p_klist; /* (c) Knotes attached to this proc. */
750 int p_numthreads; /* (c) Number of threads. */
751 struct mdproc p_md; /* Any machine-dependent fields. */
752 struct callout p_itcallout; /* (h + c) Interval timer callout. */
753 u_short p_acflag; /* (c) Accounting flags. */
754 struct proc *p_peers; /* (r) */
755 struct proc *p_leader; /* (b) */
756 void *p_emuldata; /* (c) Emulator state data. */
757 struct label *p_label; /* (*) Proc (not subject) MAC label. */
758 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
759 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
760 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */
761 struct cv p_pwait; /* (*) wait cv for exit/exec. */
762 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */
763 struct racct *p_racct; /* (b) Resource accounting. */
764 int p_throttled; /* (c) Flag for racct pcpu throttling */
765 /*
766 * An orphan is the child that has been re-parented to the
767 * debugger as a result of attaching to it. Need to keep
768 * track of them for parent to be able to collect the exit
769 * status of what used to be children.
770 */
771 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */
772 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */
773
774 TAILQ_HEAD(, kq_timer_cb_data) p_kqtim_stop; /* (c) */
775 LIST_ENTRY(proc) p_jaillist; /* (d) Jail process linkage. */
776 };
777
778 #define p_session p_pgrp->pg_session
779 #define p_pgid p_pgrp->pg_id
780
781 #define NOCPU (-1) /* For when we aren't on a CPU. */
782 #define NOCPU_OLD (255)
783 #define MAXCPU_OLD (254)
784
785 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock)
786 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock)
787 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type))
788
789 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx)
790 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx)
791 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type))
792
793 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx)
794 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx)
795 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type))
796
797 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx)
798 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx)
799 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type))
800
801 /* These flags are kept in p_flag. */
802 #define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory
803 lock. */
804 #define P_CONTROLT 0x00000002 /* Has a controlling terminal. */
805 #define P_KPROC 0x00000004 /* Kernel process. */
806 #define P_UNUSED3 0x00000008 /* --available-- */
807 #define P_PPWAIT 0x00000010 /* Parent is waiting for child to
808 exec/exit. */
809 #define P_PROFIL 0x00000020 /* Has started profiling. */
810 #define P_STOPPROF 0x00000040 /* Has thread requesting to stop
811 profiling. */
812 #define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup
813 shortcuts) */
814 #define P_SUGID 0x00000100 /* Had set id privileges since last
815 exec. */
816 #define P_SYSTEM 0x00000200 /* System proc: no sigs or stats. */
817 #define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit,
818 not wait. */
819 #define P_TRACED 0x00000800 /* Debugged process being traced. */
820 #define P_WAITED 0x00001000 /* Someone is waiting for us. */
821 #define P_WEXIT 0x00002000 /* Working on exiting. */
822 #define P_EXEC 0x00004000 /* Process called exec. */
823 #define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary
824 ASAP. */
825 #define P_CONTINUED 0x00010000 /* Proc has continued from a stopped
826 state. */
827 #define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */
828 #define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */
829 #define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to
830 user). */
831 #define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */
832 #define P_SIGEVENT 0x00200000 /* Process pending signals changed. */
833 #define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user
834 boundary. */
835 #define P_HWPMC 0x00800000 /* Process is using HWPMCs */
836 #define P_JAILED 0x01000000 /* Process is in jail. */
837 #define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */
838 #define P_INEXEC 0x04000000 /* Process is in execve(). */
839 #define P_STATCHILD 0x08000000 /* Child process stopped or exited. */
840 #define P_INMEM 0x10000000 /* Loaded into memory, always set. */
841 #define P_UNUSED1 0x20000000 /* --available-- */
842 #define P_UNUSED2 0x40000000 /* --available-- */
843 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */
844
845 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
846 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
847 #define P_KILLED(p) ((p)->p_flag & P_WKILLED)
848
849 /* These flags are kept in p_flag2. */
850 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get
851 P_PROTECTED. */
852 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or
853 coredumps. */
854 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on
855 exec(2). */
856 #define P2_AST_SU 0x00000008 /* Handles SU ast for
857 kthreads. */
858 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not
859 yet handled. */
860 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */
861 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */
862 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */
863 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk
864 area. */
865 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied
866 PROT_MAX. */
867 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied
868 PROT_MAX. */
869 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for
870 MAP_STACK */
871 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled
872 after exec */
873 #define P2_ITSTOPPED 0x00002000 /* itimers stopped */
874 #define P2_PTRACEREQ 0x00004000 /* Active ptrace req */
875 #define P2_NO_NEW_PRIVS 0x00008000 /* Ignore setuid */
876 #define P2_WXORX_DISABLE 0x00010000 /* WX mappings enabled */
877 #define P2_WXORX_ENABLE_EXEC 0x00020000 /* WXORX enabled after exec */
878 #define P2_WEXIT 0x00040000 /* exit just started, no
879 external thread_single() is
880 permitted */
881 #define P2_REAPKILLED 0x00080000 /* REAP_KILL pass touched me */
882 #define P2_MEMBAR_PRIVE 0x00100000 /* membar private expedited
883 registered */
884 #define P2_MEMBAR_PRIVE_SYNCORE 0x00200000 /* membar private expedited
885 sync core registered */
886 #define P2_MEMBAR_GLOBE 0x00400000 /* membar global expedited
887 registered */
888
889 /* Flags protected by proctree_lock, kept in p_treeflags. */
890 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */
891 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan
892 list */
893 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */
894 #define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */
895
896 /*
897 * These were process status values (p_stat), now they are only used in
898 * legacy conversion code.
899 */
900 #define SIDL 1 /* Process being created by fork. */
901 #define SRUN 2 /* Currently runnable. */
902 #define SSLEEP 3 /* Sleeping on an address. */
903 #define SSTOP 4 /* Process debugging or suspension. */
904 #define SZOMB 5 /* Awaiting collection by parent. */
905 #define SWAIT 6 /* Waiting for interrupt. */
906 #define SLOCK 7 /* Blocked on a lock. */
907
908 #define P_MAGIC 0xbeefface
909
910 #ifdef _KERNEL
911
912 /* Types and flags for mi_switch(9). */
913 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */
914 #define SWT_OWEPREEMPT 1 /* Switching due to owepreempt. */
915 #define SWT_TURNSTILE 2 /* Turnstile contention. */
916 #define SWT_SLEEPQ 3 /* Sleepq wait. */
917 #define SWT_RELINQUISH 4 /* yield call. */
918 #define SWT_NEEDRESCHED 5 /* NEEDRESCHED was set. */
919 #define SWT_IDLE 6 /* Switching from the idle thread. */
920 #define SWT_IWAIT 7 /* Waiting for interrupts. */
921 #define SWT_SUSPEND 8 /* Thread suspended. */
922 #define SWT_REMOTEPREEMPT 9 /* Remote processor preempted. */
923 #define SWT_REMOTEWAKEIDLE 10 /* Remote processor preempted idle. */
924 #define SWT_BIND 11 /* Thread bound to a new CPU. */
925 #define SWT_COUNT 12 /* Number of switch types. */
926 /* Flags */
927 #define SW_VOL 0x0100 /* Voluntary switch. */
928 #define SW_INVOL 0x0200 /* Involuntary switch. */
929 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */
930
931 /* How values for thread_single(). */
932 #define SINGLE_NO_EXIT 0
933 #define SINGLE_EXIT 1
934 #define SINGLE_BOUNDARY 2
935 #define SINGLE_ALLPROC 3
936
937 #define FOREACH_PROC_IN_SYSTEM(p) \
938 LIST_FOREACH((p), &allproc, p_list)
939 #define FOREACH_THREAD_IN_PROC(p, td) \
940 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
941
942 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
943
944 /*
945 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
946 * in a pid_t, as it is used to represent "no process group".
947 */
948 #define PID_MAX 99999
949 #define NO_PID (PID_MAX + 1)
950 #define THREAD0_TID NO_PID
951 extern pid_t pid_max;
952
953 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
954
955 /* Lock and unlock a process. */
956 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
957 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
958 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
959 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
960 #define PROC_WAIT_UNLOCKED(p) mtx_wait_unlocked(&(p)->p_mtx)
961 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
962
963 /* Lock and unlock a process group. */
964 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
965 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
966 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
967 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
968
969 #define PGRP_LOCK_PGSIGNAL(pg) do { \
970 if ((pg) != NULL) \
971 PGRP_LOCK(pg); \
972 } while (0)
973 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \
974 if ((pg) != NULL) \
975 PGRP_UNLOCK(pg); \
976 } while (0)
977
978 /* Lock and unlock a session. */
979 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
980 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
981 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
982 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
983
984 /*
985 * A non-zero p_lock prevents the process from exiting; it will sleep in exit1()
986 * until the count reaches zero.
987 *
988 * PHOLD() asserts that the process (except the current process) is
989 * not exiting and increments p_lock.
990 * _PHOLD() is same as PHOLD(), it takes the process locked.
991 */
992 #define PHOLD(p) do { \
993 PROC_LOCK(p); \
994 _PHOLD(p); \
995 PROC_UNLOCK(p); \
996 } while (0)
997 #define _PHOLD(p) do { \
998 PROC_LOCK_ASSERT((p), MA_OWNED); \
999 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
1000 ("PHOLD of exiting process %p", p)); \
1001 (p)->p_lock++; \
1002 } while (0)
1003 #define PROC_ASSERT_HELD(p) do { \
1004 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \
1005 } while (0)
1006
1007 #define PRELE(p) do { \
1008 PROC_LOCK((p)); \
1009 _PRELE((p)); \
1010 PROC_UNLOCK((p)); \
1011 } while (0)
1012 #define _PRELE(p) do { \
1013 PROC_LOCK_ASSERT((p), MA_OWNED); \
1014 PROC_ASSERT_HELD(p); \
1015 (--(p)->p_lock); \
1016 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
1017 wakeup(&(p)->p_lock); \
1018 } while (0)
1019 #define PROC_ASSERT_NOT_HELD(p) do { \
1020 KASSERT((p)->p_lock == 0, ("process %p held", p)); \
1021 } while (0)
1022
1023 #define PROC_UPDATE_COW(p) do { \
1024 struct proc *_p = (p); \
1025 PROC_LOCK_ASSERT((_p), MA_OWNED); \
1026 atomic_store_int(&_p->p_cowgen, _p->p_cowgen + 1); \
1027 } while (0)
1028
1029 #define PROC_COW_CHANGECOUNT(td, p) ({ \
1030 struct thread *_td = (td); \
1031 struct proc *_p = (p); \
1032 MPASS(_td == curthread); \
1033 PROC_LOCK_ASSERT(_p, MA_OWNED); \
1034 _p->p_cowgen - _td->td_cowgen; \
1035 })
1036
1037 /* Control whether or not it is safe for curthread to sleep. */
1038 #define THREAD_NO_SLEEPING() do { \
1039 curthread->td_no_sleeping++; \
1040 MPASS(curthread->td_no_sleeping > 0); \
1041 } while (0)
1042
1043 #define THREAD_SLEEPING_OK() do { \
1044 MPASS(curthread->td_no_sleeping > 0); \
1045 curthread->td_no_sleeping--; \
1046 } while (0)
1047
1048 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0)
1049
1050 #define THREAD_CONTENDS_ON_LOCK(lo) do { \
1051 MPASS(curthread->td_wantedlock == NULL); \
1052 curthread->td_wantedlock = lo; \
1053 } while (0)
1054
1055 #define THREAD_CONTENTION_DONE(lo) do { \
1056 MPASS(curthread->td_wantedlock == lo); \
1057 curthread->td_wantedlock = NULL; \
1058 } while (0)
1059
1060 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
1061 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)])
1062 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
1063 extern struct sx *pidhashtbl_lock;
1064 extern u_long pidhash;
1065 extern u_long pidhashlock;
1066
1067 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
1068 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
1069 extern u_long pgrphash;
1070
1071 extern struct sx allproc_lock;
1072 extern int allproc_gen;
1073 extern struct sx proctree_lock;
1074 extern struct mtx ppeers_lock;
1075 extern struct mtx procid_lock;
1076 extern struct proc proc0; /* Process slot for swapper. */
1077 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */
1078 #define thread0 (thread0_st.t0st_thread)
1079 extern struct vmspace vmspace0; /* VM space for proc0. */
1080 extern int hogticks; /* Limit on kernel cpu hogs. */
1081 extern int lastpid;
1082 extern int nprocs, maxproc; /* Current and max number of procs. */
1083 extern int maxprocperuid; /* Max procs per uid. */
1084 extern u_long ps_arg_cache_limit;
1085
1086 LIST_HEAD(proclist, proc);
1087 TAILQ_HEAD(procqueue, proc);
1088 TAILQ_HEAD(threadqueue, thread);
1089 extern struct proclist allproc; /* List of all processes. */
1090 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
1091
1092 extern struct uma_zone *proc_zone;
1093 extern struct uma_zone *pgrp_zone;
1094
1095 struct proc *pfind(pid_t); /* Find process by id. */
1096 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */
1097 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */
1098 struct pgrp *pgfind(pid_t); /* Find process group by id. */
1099 void pidhash_slockall(void); /* Shared lock all pid hash lists. */
1100 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */
1101
1102 struct fork_req {
1103 int fr_flags;
1104 int fr_pages;
1105 int *fr_pidp;
1106 struct proc **fr_procp;
1107 int *fr_pd_fd;
1108 int fr_pd_flags;
1109 struct filecaps *fr_pd_fcaps;
1110 int fr_flags2;
1111 #define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */
1112 #define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */
1113 #define FR2_KPROC 0x00000004 /* Create a kernel process */
1114 };
1115
1116 /*
1117 * pget() flags.
1118 */
1119 #define PGET_HOLD 0x00001 /* Hold the process. */
1120 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */
1121 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */
1122 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */
1123 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */
1124 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */
1125 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */
1126
1127 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
1128
1129 int pget(pid_t pid, int flags, struct proc **pp);
1130
1131 /* ast_register() flags */
1132 #define ASTR_ASTF_REQUIRED 0x0001 /* td_ast TDAI(TDA_X) flag set is
1133 required for call */
1134 #define ASTR_TDP 0x0002 /* td_pflags flag set is required */
1135 #define ASTR_KCLEAR 0x0004 /* call me on ast_kclear() */
1136 #define ASTR_UNCOND 0x0008 /* call me always */
1137
1138 void ast(struct trapframe *framep);
1139 void ast_kclear(struct thread *td);
1140 void ast_register(int ast, int ast_flags, int tdp,
1141 void (*f)(struct thread *td, int asts));
1142 void ast_deregister(int tda);
1143 void ast_sched_locked(struct thread *td, int tda);
1144 void ast_sched_mask(struct thread *td, int ast);
1145 void ast_sched(struct thread *td, int tda);
1146 void ast_unsched_locked(struct thread *td, int tda);
1147
1148 struct thread *choosethread(void);
1149 int cr_bsd_visible(struct ucred *u1, struct ucred *u2);
1150 int cr_cansee(struct ucred *u1, struct ucred *u2);
1151 int cr_canseesocket(struct ucred *cred, struct socket *so);
1152 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
1153 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
1154 struct session *sess);
1155 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
1156 int fork1(struct thread *, struct fork_req *);
1157 void fork_exit(void (*)(void *, struct trapframe *), void *,
1158 struct trapframe *);
1159 void fork_return(struct thread *, struct trapframe *);
1160 int inferior(struct proc *p);
1161 void itimer_proc_continue(struct proc *p);
1162 void kqtimer_proc_continue(struct proc *p);
1163 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry,
1164 int *resident_count, bool *super);
1165 void kern_yield(int);
1166 void killjobc(void);
1167 int leavepgrp(struct proc *p);
1168 int maybe_preempt(struct thread *td);
1169 void maybe_yield(void);
1170 void mi_switch(int flags);
1171 int p_candebug(struct thread *td, struct proc *p);
1172 int p_cansee(struct thread *td, struct proc *p);
1173 int p_cansched(struct thread *td, struct proc *p);
1174 int p_cansignal(struct thread *td, struct proc *p, int signum);
1175 int p_canwait(struct thread *td, struct proc *p);
1176 struct pargs *pargs_alloc(int len);
1177 void pargs_drop(struct pargs *pa);
1178 void pargs_hold(struct pargs *pa);
1179 void proc_add_orphan(struct proc *child, struct proc *parent);
1180 int proc_get_binpath(struct proc *p, char *binname, char **fullpath,
1181 char **freepath);
1182 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
1183 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
1184 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
1185 void procinit(void);
1186 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg);
1187 void proc_linkup0(struct proc *p, struct thread *td);
1188 void proc_linkup(struct proc *p, struct thread *td);
1189 struct proc *proc_realparent(struct proc *child);
1190 void proc_reap(struct thread *td, struct proc *p, int *status, int options);
1191 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid);
1192 void proc_set_p2_wexit(struct proc *p);
1193 void proc_set_traced(struct proc *p, bool stop);
1194 void proc_wkilled(struct proc *p);
1195 struct pstats *pstats_alloc(void);
1196 void pstats_fork(struct pstats *src, struct pstats *dst);
1197 void pstats_free(struct pstats *ps);
1198 void proc_clear_orphan(struct proc *p);
1199 void reaper_abandon_children(struct proc *p, bool exiting);
1200 int securelevel_ge(struct ucred *cr, int level);
1201 int securelevel_gt(struct ucred *cr, int level);
1202 void sess_hold(struct session *);
1203 void sess_release(struct session *);
1204 void setrunnable(struct thread *, int);
1205 void setsugid(struct proc *p);
1206 bool should_yield(void);
1207 int sigonstack(size_t sp);
1208 void stopevent(struct proc *, u_int, u_int);
1209 struct thread *tdfind(lwpid_t, pid_t);
1210 void threadinit(void);
1211 void tidhash_add(struct thread *);
1212 void tidhash_remove(struct thread *);
1213 void cpu_idle(int);
1214 int cpu_idle_wakeup(int);
1215 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */
1216 void cpu_switch(struct thread *, struct thread *, struct mtx *);
1217 void cpu_sync_core(void);
1218 void cpu_throw(struct thread *, struct thread *) __dead2;
1219 bool curproc_sigkilled(void);
1220 void userret(struct thread *, struct trapframe *);
1221
1222 void cpu_exit(struct thread *);
1223 void exit1(struct thread *, int, int) __dead2;
1224 void cpu_copy_thread(struct thread *td, struct thread *td0);
1225 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map);
1226 int cpu_fetch_syscall_args(struct thread *td);
1227 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
1228 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *);
1229 int cpu_procctl(struct thread *td, int idtype, id_t id, int com,
1230 void *data);
1231 void cpu_set_syscall_retval(struct thread *, int);
1232 int cpu_set_upcall(struct thread *, void (*)(void *), void *,
1233 stack_t *);
1234 int cpu_set_user_tls(struct thread *, void *tls_base);
1235 void cpu_thread_alloc(struct thread *);
1236 void cpu_thread_clean(struct thread *);
1237 void cpu_thread_exit(struct thread *);
1238 void cpu_thread_free(struct thread *);
1239 struct thread *thread_alloc(int pages);
1240 int thread_check_susp(struct thread *td, bool sleep);
1241 void thread_cow_get_proc(struct thread *newtd, struct proc *p);
1242 void thread_cow_get(struct thread *newtd, struct thread *td);
1243 void thread_cow_free(struct thread *td);
1244 void thread_cow_update(struct thread *td);
1245 void thread_cow_synced(struct thread *td);
1246 int thread_create(struct thread *td, struct rtprio *rtp,
1247 int (*initialize_thread)(struct thread *, void *), void *thunk);
1248 void thread_exit(void) __dead2;
1249 void thread_free(struct thread *td);
1250 void thread_link(struct thread *td, struct proc *p);
1251 void thread_reap_barrier(void);
1252 int thread_recycle(struct thread *, int pages);
1253 int thread_single(struct proc *p, int how);
1254 void thread_single_end(struct proc *p, int how);
1255 void thread_stash(struct thread *td);
1256 void thread_stopped(struct proc *p);
1257 void childproc_stopped(struct proc *child, int reason);
1258 void childproc_continued(struct proc *child);
1259 void childproc_exited(struct proc *child);
1260 void thread_run_flash(struct thread *td);
1261 int thread_suspend_check(int how);
1262 bool thread_suspend_check_needed(void);
1263 void thread_suspend_switch(struct thread *, struct proc *p);
1264 void thread_suspend_one(struct thread *td);
1265 void thread_unlink(struct thread *td);
1266 void thread_unsuspend(struct proc *p);
1267 void thread_wait(struct proc *p);
1268
1269 bool stop_all_proc_block(void);
1270 void stop_all_proc_unblock(void);
1271 void stop_all_proc(void);
1272 void resume_all_proc(void);
1273
1274 static __inline int
curthread_pflags_set(int flags)1275 curthread_pflags_set(int flags)
1276 {
1277 struct thread *td;
1278 int save;
1279
1280 td = curthread;
1281 save = ~flags | (td->td_pflags & flags);
1282 td->td_pflags |= flags;
1283 return (save);
1284 }
1285
1286 static __inline void
curthread_pflags_restore(int save)1287 curthread_pflags_restore(int save)
1288 {
1289
1290 curthread->td_pflags &= save;
1291 }
1292
1293 static __inline int
curthread_pflags2_set(int flags)1294 curthread_pflags2_set(int flags)
1295 {
1296 struct thread *td;
1297 int save;
1298
1299 td = curthread;
1300 save = ~flags | (td->td_pflags2 & flags);
1301 td->td_pflags2 |= flags;
1302 return (save);
1303 }
1304
1305 static __inline void
curthread_pflags2_restore(int save)1306 curthread_pflags2_restore(int save)
1307 {
1308
1309 curthread->td_pflags2 &= save;
1310 }
1311
1312 static __inline __pure2 struct td_sched *
td_get_sched(struct thread * td)1313 td_get_sched(struct thread *td)
1314 {
1315
1316 return ((struct td_sched *)&td[1]);
1317 }
1318
1319 #define PROC_ID_PID 0
1320 #define PROC_ID_GROUP 1
1321 #define PROC_ID_SESSION 2
1322 #define PROC_ID_REAP 3
1323
1324 void proc_id_set(int type, pid_t id);
1325 void proc_id_set_cond(int type, pid_t id);
1326 void proc_id_clear(int type, pid_t id);
1327
1328 EVENTHANDLER_LIST_DECLARE(process_ctor);
1329 EVENTHANDLER_LIST_DECLARE(process_dtor);
1330 EVENTHANDLER_LIST_DECLARE(process_init);
1331 EVENTHANDLER_LIST_DECLARE(process_fini);
1332 EVENTHANDLER_LIST_DECLARE(process_exit);
1333 EVENTHANDLER_LIST_DECLARE(process_fork);
1334 EVENTHANDLER_LIST_DECLARE(process_exec);
1335
1336 EVENTHANDLER_LIST_DECLARE(thread_ctor);
1337 EVENTHANDLER_LIST_DECLARE(thread_dtor);
1338 EVENTHANDLER_LIST_DECLARE(thread_init);
1339
1340 #endif /* _KERNEL */
1341
1342 #endif /* !_SYS_PROC_H_ */
1343