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 * @(#)proc.h 8.15 (Berkeley) 5/19/95
37 */
38
39 #ifndef _SYS_PROC_H_
40 #define _SYS_PROC_H_
41
42 #include <sys/callout.h> /* For struct callout. */
43 #include <sys/event.h> /* For struct klist. */
44 #ifdef _KERNEL
45 #include <sys/_eventhandler.h>
46 #endif
47 #include <sys/condvar.h>
48 #ifndef _KERNEL
49 #include <sys/filedesc.h>
50 #endif
51 #include <sys/queue.h>
52 #include <sys/_lock.h>
53 #include <sys/lock_profile.h>
54 #include <sys/_mutex.h>
55 #include <sys/osd.h>
56 #include <sys/priority.h>
57 #include <sys/rtprio.h> /* XXX. */
58 #include <sys/runq.h>
59 #include <sys/resource.h>
60 #include <sys/sigio.h>
61 #include <sys/signal.h>
62 #include <sys/signalvar.h>
63 #ifndef _KERNEL
64 #include <sys/time.h> /* For structs itimerval, timeval. */
65 #else
66 #include <sys/pcpu.h>
67 #include <sys/systm.h>
68 #endif
69 #include <sys/ucontext.h>
70 #include <sys/ucred.h>
71 #include <sys/types.h>
72 #include <sys/_domainset.h>
73
74 #include <machine/proc.h> /* Machine-dependent proc substruct. */
75 #ifdef _KERNEL
76 #include <machine/cpu.h>
77 #endif
78
79 /*
80 * One structure allocated per session.
81 *
82 * List of locks
83 * (m) locked by s_mtx mtx
84 * (e) locked by proctree_lock sx
85 * (c) const until freeing
86 */
87 struct session {
88 u_int s_count; /* Ref cnt; pgrps in session - atomic. */
89 struct proc *s_leader; /* (m + e) Session leader. */
90 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
91 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */
92 struct tty *s_ttyp; /* (e) Controlling tty. */
93 pid_t s_sid; /* (c) Session ID. */
94 /* (m) Setlogin() name: */
95 char s_login[roundup(MAXLOGNAME, sizeof(long))];
96 struct mtx s_mtx; /* Mutex to protect members. */
97 };
98
99 /*
100 * One structure allocated per process group.
101 *
102 * List of locks
103 * (m) locked by pg_mtx mtx
104 * (e) locked by proctree_lock sx
105 * (c) const until freeing
106 */
107 struct pgrp {
108 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
109 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
110 struct session *pg_session; /* (c) Pointer to session. */
111 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
112 pid_t pg_id; /* (c) Process group id. */
113 struct mtx pg_mtx; /* Mutex to protect members */
114 int pg_flags; /* (m) PGRP_ flags */
115 struct sx pg_killsx; /* Mutual exclusion between group member
116 * fork() and killpg() */
117 };
118
119 #define PGRP_ORPHANED 0x00000001 /* Group is orphaned */
120
121 /*
122 * pargs, used to hold a copy of the command line, if it had a sane length.
123 */
124 struct pargs {
125 u_int ar_ref; /* Reference count. */
126 u_int ar_length; /* Length. */
127 u_char ar_args[1]; /* Arguments. */
128 };
129
130 /*-
131 * Description of a process.
132 *
133 * This structure contains the information needed to manage a thread of
134 * control, known in UN*X as a process; it has references to substructures
135 * containing descriptions of things that the process uses, but may share
136 * with related processes. The process structure and the substructures
137 * are always addressable except for those marked "(CPU)" below,
138 * which might be addressable only on a processor on which the process
139 * is running.
140 *
141 * Below is a key of locks used to protect each member of struct proc. The
142 * lock is indicated by a reference to a specific character in parens in the
143 * associated comment.
144 * * - not yet protected
145 * a - only touched by curproc or parent during fork/wait
146 * b - created at fork, never changes
147 * (exception aiods switch vmspaces, but they are also
148 * marked 'P_SYSTEM' so hopefully it will be left alone)
149 * c - locked by proc mtx
150 * d - locked by allproc_lock lock
151 * e - locked by proctree_lock lock
152 * f - session mtx
153 * g - process group mtx
154 * h - callout_lock mtx
155 * i - by curproc or the master session mtx
156 * j - locked by proc slock
157 * k - only accessed by curthread
158 * k*- only accessed by curthread and from an interrupt
159 * kx- only accessed by curthread and by debugger
160 * l - the attaching proc or attaching proc parent
161 * m - Giant
162 * n - not locked, lazy
163 * o - ktrace lock
164 * q - td_contested lock
165 * r - p_peers lock
166 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9)
167 * t - thread lock
168 * u - process stat lock
169 * w - process timer lock
170 * x - created at fork, only changes during single threading in exec
171 * y - created at first aio, doesn't change until exit or exec at which
172 * point we are single-threaded and only curthread changes it
173 * z - zombie threads lock
174 *
175 * If the locking key specifies two identifiers (for example, p_pptr) then
176 * either lock is sufficient for read access, but both locks must be held
177 * for write access.
178 */
179 struct cpuset;
180 struct filecaps;
181 struct filemon;
182 struct kaioinfo;
183 struct kaudit_record;
184 struct kcov_info;
185 struct kdtrace_proc;
186 struct kdtrace_thread;
187 struct kq_timer_cb_data;
188 struct mqueue_notifier;
189 struct p_sched;
190 struct proc;
191 struct procdesc;
192 struct racct;
193 struct sbuf;
194 struct sleepqueue;
195 struct socket;
196 struct syscall_args;
197 struct td_sched;
198 struct thread;
199 struct trapframe;
200 struct turnstile;
201 struct vm_map;
202 struct vm_map_entry;
203 struct epoch_tracker;
204
205 /*
206 * XXX: Does this belong in resource.h or resourcevar.h instead?
207 * Resource usage extension. The times in rusage structs in the kernel are
208 * never up to date. The actual times are kept as runtimes and tick counts
209 * (with control info in the "previous" times), and are converted when
210 * userland asks for rusage info. Backwards compatibility prevents putting
211 * this directly in the user-visible rusage struct.
212 *
213 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
214 * Locking for td_rux: (t) for all fields.
215 */
216 struct rusage_ext {
217 uint64_t rux_runtime; /* (cu) Real time. */
218 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */
219 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */
220 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */
221 uint64_t rux_uu; /* (c) Previous user time in usec. */
222 uint64_t rux_su; /* (c) Previous sys time in usec. */
223 uint64_t rux_tu; /* (c) Previous total time in usec. */
224 };
225
226 /*
227 * Kernel runnable context (thread).
228 * This is what is put to sleep and reactivated.
229 * Thread context. Processes may have multiple threads.
230 */
231 struct thread {
232 struct mtx *volatile td_lock; /* replaces sched lock */
233 struct proc *td_proc; /* (*) Associated process. */
234 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
235 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */
236 union {
237 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */
238 struct thread *td_zombie; /* Zombie list linkage */
239 };
240 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */
241 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */
242 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */
243 struct domainset_ref td_domain; /* (a) NUMA policy */
244 struct seltd *td_sel; /* Select queue/channel. */
245 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
246 struct turnstile *td_turnstile; /* (k) Associated turnstile. */
247 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */
248 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
249 lwpid_t td_tid; /* (b) Thread ID. */
250 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
251 #define td_siglist td_sigqueue.sq_signals
252 u_char td_lend_user_pri; /* (t) Lend user pri. */
253 u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */
254
255 /* Cleared during fork1(), thread_create(), or kthread_add(). */
256 #define td_startzero td_flags
257 int td_flags; /* (t) TDF_* flags. */
258 int td_inhibitors; /* (t) Why can not run. */
259 int td_pflags; /* (k) Private thread (TDP_*) flags. */
260 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */
261 int td_dupfd; /* (k) Ret value from fdopen. XXX */
262 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */
263 const void *td_wchan; /* (t) Sleep address. */
264 const char *td_wmesg; /* (t) Reason for sleep. */
265 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
266 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */
267 short td_locks; /* (k) Debug: count of non-spin locks */
268 short td_rw_rlocks; /* (k) Count of rwlock read locks. */
269 short td_sx_slocks; /* (k) Count of sx shared locks. */
270 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */
271 short td_stopsched; /* (k) Scheduler stopped. */
272 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */
273 const char *td_lockname; /* (t) Name of lock blocked on. */
274 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
275 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
276 int td_intr_nesting_level; /* (k) Interrupt recursion. */
277 int td_pinned; /* (k) Temporary cpu pin count. */
278 struct ucred *td_realucred; /* (k) Reference to credentials. */
279 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */
280 struct plimit *td_limit; /* (k) Resource limits. */
281 int td_slptick; /* (t) Time at sleep. */
282 int td_blktick; /* (t) Time spent blocked. */
283 int td_swvoltick; /* (t) Time at last SW_VOL switch. */
284 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */
285 u_int td_cow; /* (*) Number of copy-on-write faults */
286 struct rusage td_ru; /* (t) rusage information. */
287 struct rusage_ext td_rux; /* (t) Internal rusage information. */
288 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */
289 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */
290 u_int td_pticks; /* (t) Statclock hits for profiling */
291 u_int td_sticks; /* (t) Statclock hits in system mode. */
292 u_int td_iticks; /* (t) Statclock hits in intr mode. */
293 u_int td_uticks; /* (t) Statclock hits in user mode. */
294 int td_intrval; /* (t) Return value for sleepq. */
295 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
296 volatile u_int td_generation; /* (k) For detection of preemption */
297 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
298 int td_xsig; /* (c) Signal for ptrace */
299 u_long td_profil_addr; /* (k) Temporary addr until AST. */
300 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
301 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
302 struct file *td_fpop; /* (k) file referencing cdev under op */
303 int td_dbgflags; /* (c) Userland debugger flags */
304 siginfo_t td_si; /* (c) For debugger or core file */
305 int td_ng_outbound; /* (k) Thread entered ng from above. */
306 struct osd td_osd; /* (k) Object specific data. */
307 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
308 pid_t td_dbg_forked; /* (c) Child pid for debugger. */
309 struct vnode *td_vp_reserved;/* (k) Preallocated vnode. */
310 u_int td_no_sleeping; /* (k) Sleeping disabled count. */
311 void *td_su; /* (k) FFS SU private */
312 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */
313 int td_rtcgen; /* (s) rtc_generation of abs. sleep */
314 int td_errno; /* (k) Error from last syscall. */
315 size_t td_vslock_sz; /* (k) amount of vslock-ed space */
316 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */
317 u_int td_ucredref; /* (k) references on td_realucred */
318 #define td_endzero td_sigmask
319
320 /* Copied during fork1(), thread_create(), or kthread_add(). */
321 #define td_startcopy td_endzero
322 sigset_t td_sigmask; /* (c) Current signal mask. */
323 u_char td_rqindex; /* (t) Run queue index. */
324 u_char td_base_pri; /* (t) Thread base kernel priority. */
325 u_char td_priority; /* (t) Thread active priority. */
326 u_char td_pri_class; /* (t) Scheduling class. */
327 u_char td_user_pri; /* (t) User pri from estcpu and nice. */
328 u_char td_base_user_pri; /* (t) Base user pri */
329 u_char td_unused_0; /* no longer used field */
330 uintptr_t td_rb_list; /* (k) Robust list head. */
331 uintptr_t td_rbp_list; /* (k) Robust priv list head. */
332 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */
333 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on
334 fork for child tracing. */
335 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */
336 uint32_t td_sigblock_val; /* (k) fast sigblock value read at
337 td_sigblock_ptr on kern entry */
338 #define td_endcopy td_pcb
339
340 /*
341 * Fields that must be manually set in fork1(), thread_create(), kthread_add(),
342 * or already have been set in the allocator, constructor, etc.
343 */
344 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
345 enum td_states {
346 TDS_INACTIVE = 0x0,
347 TDS_INHIBITED,
348 TDS_CAN_RUN,
349 TDS_RUNQ,
350 TDS_RUNNING
351 } td_state; /* (t) thread state */
352 union {
353 register_t tdu_retval[2];
354 off_t tdu_off;
355 } td_uretoff; /* (k) Syscall aux returns. */
356 #define td_retval td_uretoff.tdu_retval
357 u_int td_cowgen; /* (k) Generation of COW pointers. */
358 /* LP64 hole */
359 struct callout td_slpcallout; /* (h) Callout for sleep. */
360 struct trapframe *td_frame; /* (k) */
361 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
362 int td_kstack_pages; /* (a) Size of the kstack. */
363 volatile u_int td_critnest; /* (k*) Critical section nest level. */
364 struct mdthread td_md; /* (k) Any machine-dependent fields. */
365 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
366 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */
367 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */
368 struct vnet *td_vnet; /* (k) Effective vnet. */
369 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */
370 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
371 struct proc *td_rfppwait_p; /* (k) The vforked child */
372 struct vm_page **td_ma; /* (k) uio pages held */
373 int td_ma_cnt; /* (k) size of *td_ma */
374 /* LP64 hole */
375 void *td_emuldata; /* Emulator state data */
376 int td_lastcpu; /* (t) Last cpu we were on. */
377 int td_oncpu; /* (t) Which cpu we are on. */
378 void *td_lkpi_task; /* LinuxKPI task struct pointer */
379 int td_pmcpend;
380 void *td_remotereq; /* (c) dbg remote request. */
381 off_t td_ktr_io_lim; /* (k) limit for ktrace file size */
382 #ifdef EPOCH_TRACE
383 SLIST_HEAD(, epoch_tracker) td_epochs;
384 #endif
385 };
386
387 struct thread0_storage {
388 struct thread t0st_thread;
389 uint64_t t0st_sched[10];
390 };
391
392 struct mtx *thread_lock_block(struct thread *);
393 void thread_lock_block_wait(struct thread *);
394 void thread_lock_set(struct thread *, struct mtx *);
395 void thread_lock_unblock(struct thread *, struct mtx *);
396 #define THREAD_LOCK_ASSERT(td, type) \
397 mtx_assert((td)->td_lock, (type))
398
399 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \
400 do { \
401 struct mtx *__m = (td)->td_lock; \
402 if (__m != &blocked_lock) \
403 mtx_assert(__m, (type)); \
404 } while (0)
405
406 #ifdef INVARIANTS
407 #define THREAD_LOCKPTR_ASSERT(td, lock) \
408 do { \
409 struct mtx *__m; \
410 __m = (td)->td_lock; \
411 KASSERT(__m == (lock), \
412 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
413 } while (0)
414
415 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \
416 do { \
417 struct mtx *__m; \
418 __m = (td)->td_lock; \
419 KASSERT(__m == (lock) || __m == &blocked_lock, \
420 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
421 } while (0)
422
423 #define TD_LOCKS_INC(td) ((td)->td_locks++)
424 #define TD_LOCKS_DEC(td) do { \
425 KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \
426 ("thread %p owns no locks", (td))); \
427 (td)->td_locks--; \
428 } while (0)
429 #else
430 #define THREAD_LOCKPTR_ASSERT(td, lock)
431 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock)
432
433 #define TD_LOCKS_INC(td)
434 #define TD_LOCKS_DEC(td)
435 #endif
436
437 /*
438 * Flags kept in td_flags:
439 * To change these you MUST have the scheduler lock.
440 */
441 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
442 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
443 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */
444 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
445 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
446 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
447 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */
448 #define TDF_SIGWAIT 0x00000080 /* Ignore ignored signals */
449 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */
450 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */
451 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
452 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */
453 #define TDF_KQTICKLED 0x00001000 /* AST drain kqueue taskqueue */
454 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */
455 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
456 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */
457 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */
458 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */
459 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */
460 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */
461 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
462 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */
463 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */
464 #define TDF_UNUSED6 0x00800000 /* Available */
465 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
466 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
467 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
468 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
469 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */
470 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */
471 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */
472
473 /* Userland debug flags */
474 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */
475 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */
476 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */
477 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */
478 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */
479 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */
480 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new
481 process */
482 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child
483 only) */
484 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */
485 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */
486 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */
487 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */
488 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */
489 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */
490 #define TDB_SSWITCH 0x00004000 /* Suspended in ptracestop */
491 #define TDB_BOUNDARY 0x00008000 /* ptracestop() at boundary */
492 #define TDB_COREDUMPREQ 0x00010000 /* Coredump request */
493 #define TDB_SCREMOTEREQ 0x00020000 /* Remote syscall request */
494
495 /*
496 * "Private" flags kept in td_pflags:
497 * These are only written by curthread and thus need no locking.
498 */
499 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
500 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
501 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
502 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */
503 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
504 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
505 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */
506 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */
507 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */
508 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
509 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
510 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */
511 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
512 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
513 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
514 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
515 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
516 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
517 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
518 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
519 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */
520 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */
521 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */
522 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
523 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */
524 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */
525 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */
526 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */
527 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */
528 #define TDP_FORKING 0x20000000 /* Thread is being created through fork() */
529 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */
530 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */
531
532 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */
533 #define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */
534 #define TDP2_ACCT 0x00000004 /* Doing accounting */
535
536 /*
537 * Reasons that the current thread can not be run yet.
538 * More than one may apply.
539 */
540 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */
541 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
542 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */
543 #define TDI_LOCK 0x0008 /* Stopped on a lock. */
544 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
545
546 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
547 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
548 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
549 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
550 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
551 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
552 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING)
553 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ)
554 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN)
555 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED)
556 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
557 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD)
558
559 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \
560 ((td)->td_flags & TDF_SINTR) != 0)
561
562 #define KTDSTATE(td) \
563 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \
564 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \
565 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \
566 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \
567 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding")
568
569 #define TD_SET_INHIB(td, inhib) do { \
570 (td)->td_state = TDS_INHIBITED; \
571 (td)->td_inhibitors |= (inhib); \
572 } while (0)
573
574 #define TD_CLR_INHIB(td, inhib) do { \
575 if (((td)->td_inhibitors & (inhib)) && \
576 (((td)->td_inhibitors &= ~(inhib)) == 0)) \
577 (td)->td_state = TDS_CAN_RUN; \
578 } while (0)
579
580 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
581 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
582 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
583 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
584 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
585 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
586
587 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
588 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
589 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
590 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
591 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
592
593 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING
594 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ
595 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN
596
597 #define TD_SBDRY_INTR(td) \
598 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0)
599 #define TD_SBDRY_ERRNO(td) \
600 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART)
601
602 /*
603 * Process structure.
604 */
605 struct proc {
606 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
607 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */
608 struct mtx p_slock; /* process spin lock */
609 struct ucred *p_ucred; /* (c) Process owner's identity. */
610 struct filedesc *p_fd; /* (b) Open files. */
611 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
612 struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */
613 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
614 struct plimit *p_limit; /* (c) Resource limits. */
615 struct callout p_limco; /* (c) Limit callout handle */
616 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
617
618 int p_flag; /* (c) P_* flags. */
619 int p_flag2; /* (c) P2_* flags. */
620 enum p_states {
621 PRS_NEW = 0, /* In creation */
622 PRS_NORMAL, /* threads can be run. */
623 PRS_ZOMBIE
624 } p_state; /* (j/c) Process status. */
625 pid_t p_pid; /* (b) Process identifier. */
626 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
627 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
628 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
629 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
630 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
631 struct proc *p_reaper; /* (e) My reaper. */
632 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants
633 (if I am reaper). */
634 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of
635 the same reaper. */
636 struct mtx p_mtx; /* (n) Lock for this struct. */
637 struct mtx p_statmtx; /* Lock for the stats */
638 struct mtx p_itimmtx; /* Lock for the virt/prof timers */
639 struct mtx p_profmtx; /* Lock for the profiling */
640 struct ksiginfo *p_ksi; /* Locked by parent proc lock */
641 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
642 #define p_siglist p_sigqueue.sq_signals
643 pid_t p_oppid; /* (c + e) Real parent pid. */
644
645 /* The following fields are all zeroed upon creation in fork. */
646 #define p_startzero p_vmspace
647 struct vmspace *p_vmspace; /* (b) Address space. */
648 u_int p_swtick; /* (c) Tick when swapped in or out. */
649 u_int p_cowgen; /* (c) Generation of COW pointers. */
650 struct itimerval p_realtimer; /* (c) Alarm timer. */
651 struct rusage p_ru; /* (a) Exit information. */
652 struct rusage_ext p_rux; /* (cu) Internal resource usage. */
653 struct rusage_ext p_crux; /* (c) Internal child resource usage. */
654 int p_profthreads; /* (c) Num threads in addupc_task. */
655 volatile int p_exitthreads; /* (j) Number of threads exiting */
656 int p_traceflag; /* (o) Kernel trace points. */
657 struct ktr_io_params *p_ktrioparms; /* (c + o) Params for ktrace. */
658 void *p_pad0;
659 struct vnode *p_textvp; /* (b) Vnode of executable. */
660 u_int p_lock; /* (c) Proclock (prevent swap) count. */
661 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
662 int p_sigparent; /* (c) Signal to parent on exit. */
663 int p_sig; /* (n) For core dump/debugger XXX. */
664 u_int p_ptevents; /* (c + e) ptrace() event mask. */
665 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */
666 struct thread *p_singlethread;/* (c + j) If single threading this is it */
667 int p_suspcount; /* (j) Num threads in suspended mode. */
668 struct thread *p_xthread; /* (c) Trap thread */
669 int p_boundary_count;/* (j) Num threads at user boundary */
670 int p_pendingcnt; /* how many signals are pending */
671 struct itimers *p_itimers; /* (c) POSIX interval timers. */
672 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */
673 u_int p_treeflag; /* (e) P_TREE flags */
674 int p_pendingexits; /* (c) Count of pending thread exits. */
675 struct filemon *p_filemon; /* (c) filemon-specific data. */
676 int p_pdeathsig; /* (c) Signal from parent on exit. */
677 /* End area that is zeroed on creation. */
678 #define p_endzero p_magic
679
680 /* The following fields are all copied upon creation in fork. */
681 #define p_startcopy p_endzero
682 u_int p_magic; /* (b) Magic number. */
683 int p_osrel; /* (x) osreldate for the
684 binary (from ELF note, if any) */
685 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */
686 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */
687 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
688 struct pargs *p_args; /* (c) Process arguments. */
689 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */
690 signed char p_nice; /* (c) Process "nice" value. */
691 int p_fibnum; /* in this routing domain XXX MRT */
692 pid_t p_reapsubtree; /* (e) Pid of the direct child of the
693 reaper which spawned
694 our subtree. */
695 uint16_t p_elf_machine; /* (x) ELF machine type */
696 uint64_t p_elf_flags; /* (x) ELF flags */
697 /* End area that is copied on creation. */
698 #define p_endcopy p_xexit
699
700 u_int p_xexit; /* (c) Exit code. */
701 u_int p_xsig; /* (c) Stop/kill sig. */
702 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
703 struct knlist *p_klist; /* (c) Knotes attached to this proc. */
704 int p_numthreads; /* (c) Number of threads. */
705 struct mdproc p_md; /* Any machine-dependent fields. */
706 struct callout p_itcallout; /* (h + c) Interval timer callout. */
707 u_short p_acflag; /* (c) Accounting flags. */
708 struct proc *p_peers; /* (r) */
709 struct proc *p_leader; /* (b) */
710 void *p_emuldata; /* (c) Emulator state data. */
711 struct label *p_label; /* (*) Proc (not subject) MAC label. */
712 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
713 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
714 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */
715 struct cv p_pwait; /* (*) wait cv for exit/exec. */
716 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */
717 struct racct *p_racct; /* (b) Resource accounting. */
718 int p_throttled; /* (c) Flag for racct pcpu throttling */
719 /*
720 * An orphan is the child that has been re-parented to the
721 * debugger as a result of attaching to it. Need to keep
722 * track of them for parent to be able to collect the exit
723 * status of what used to be children.
724 */
725 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */
726 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */
727
728 TAILQ_HEAD(, kq_timer_cb_data) p_kqtim_stop; /* (c) */
729 struct vnode *p_textdvp; /* (b) Dir containing textvp. */
730 char *p_binname; /* (b) Binary hardlink name. */
731 void *p_elf_brandinfo; /* (x) Elf_Brandinfo, NULL for
732 non ELF binaries. */
733 sbintime_t p_umtx_min_timeout;
734 };
735
736 #define p_session p_pgrp->pg_session
737 #define p_pgid p_pgrp->pg_id
738
739 #define NOCPU (-1) /* For when we aren't on a CPU. */
740 #define NOCPU_OLD (255)
741 #define MAXCPU_OLD (254)
742
743 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock)
744 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock)
745 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type))
746
747 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx)
748 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx)
749 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type))
750
751 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx)
752 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx)
753 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type))
754
755 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx)
756 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx)
757 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type))
758
759 /* These flags are kept in p_flag. */
760 #define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory
761 lock. */
762 #define P_CONTROLT 0x00000002 /* Has a controlling terminal. */
763 #define P_KPROC 0x00000004 /* Kernel process. */
764 #define P_UNUSED3 0x00000008 /* --available-- */
765 #define P_PPWAIT 0x00000010 /* Parent is waiting for child to
766 exec/exit. */
767 #define P_PROFIL 0x00000020 /* Has started profiling. */
768 #define P_STOPPROF 0x00000040 /* Has thread requesting to stop
769 profiling. */
770 #define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup
771 shortcuts) */
772 #define P_SUGID 0x00000100 /* Had set id privileges since last
773 exec. */
774 #define P_SYSTEM 0x00000200 /* System proc: no sigs, stats or
775 swapping. */
776 #define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit,
777 not wait. */
778 #define P_TRACED 0x00000800 /* Debugged process being traced. */
779 #define P_WAITED 0x00001000 /* Someone is waiting for us. */
780 #define P_WEXIT 0x00002000 /* Working on exiting. */
781 #define P_EXEC 0x00004000 /* Process called exec. */
782 #define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary
783 ASAP. */
784 #define P_CONTINUED 0x00010000 /* Proc has continued from a stopped
785 state. */
786 #define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */
787 #define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */
788 #define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to
789 user). */
790 #define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */
791 #define P_SIGEVENT 0x00200000 /* Process pending signals changed. */
792 #define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user
793 boundary. */
794 #define P_HWPMC 0x00800000 /* Process is using HWPMCs */
795 #define P_JAILED 0x01000000 /* Process is in jail. */
796 #define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */
797 #define P_INEXEC 0x04000000 /* Process is in execve(). */
798 #define P_STATCHILD 0x08000000 /* Child process stopped or exited. */
799 #define P_INMEM 0x10000000 /* Loaded into memory. */
800 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */
801 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */
802 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */
803
804 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
805 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
806 #define P_KILLED(p) ((p)->p_flag & P_WKILLED)
807
808 /* These flags are kept in p_flag2. */
809 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get
810 P_PROTECTED. */
811 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or
812 coredumps. */
813 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on
814 exec(2). */
815 #define P2_AST_SU 0x00000008 /* Handles SU ast for
816 kthreads. */
817 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not
818 yet handled. */
819 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */
820 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */
821 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */
822 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk
823 area. */
824 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied
825 PROT_MAX. */
826 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied
827 PROT_MAX. */
828 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for
829 MAP_STACK */
830 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled
831 after exec */
832 #define P2_ITSTOPPED 0x00002000
833 #define P2_PTRACEREQ 0x00004000 /* Active ptrace req */
834 #define P2_NO_NEW_PRIVS 0x00008000 /* Ignore setuid */
835 #define P2_WXORX_DISABLE 0x00010000 /* WX mappings enabled */
836 #define P2_WXORX_ENABLE_EXEC 0x00020000 /* WXORX enabled after exec */
837 #define P2_WEXIT 0x00040000 /* exit just started, no
838 external thread_single() is
839 permitted */
840 #define P2_REAPKILLED 0x00080000
841 #define P2_MEMBAR_PRIVE 0x00100000 /* membar private expedited
842 registered */
843 #define P2_MEMBAR_PRIVE_SYNCORE 0x00200000 /* membar private expedited
844 sync core registered */
845 #define P2_MEMBAR_GLOBE 0x00400000 /* membar global expedited
846 registered */
847
848 /* Flags protected by proctree_lock, kept in p_treeflags. */
849 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */
850 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan
851 list */
852 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */
853 #define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */
854
855 /*
856 * These were process status values (p_stat), now they are only used in
857 * legacy conversion code.
858 */
859 #define SIDL 1 /* Process being created by fork. */
860 #define SRUN 2 /* Currently runnable. */
861 #define SSLEEP 3 /* Sleeping on an address. */
862 #define SSTOP 4 /* Process debugging or suspension. */
863 #define SZOMB 5 /* Awaiting collection by parent. */
864 #define SWAIT 6 /* Waiting for interrupt. */
865 #define SLOCK 7 /* Blocked on a lock. */
866
867 #define P_MAGIC 0xbeefface
868
869 #ifdef _KERNEL
870
871 /* Types and flags for mi_switch(). */
872 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */
873 #define SWT_NONE 0 /* Unspecified switch. */
874 #define SWT_PREEMPT 1 /* Switching due to preemption. */
875 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */
876 #define SWT_TURNSTILE 3 /* Turnstile contention. */
877 #define SWT_SLEEPQ 4 /* Sleepq wait. */
878 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */
879 #define SWT_RELINQUISH 6 /* yield call. */
880 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */
881 #define SWT_IDLE 8 /* Switching from the idle thread. */
882 #define SWT_IWAIT 9 /* Waiting for interrupts. */
883 #define SWT_SUSPEND 10 /* Thread suspended. */
884 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */
885 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */
886 #define SWT_COUNT 13 /* Number of switch types. */
887 /* Flags */
888 #define SW_VOL 0x0100 /* Voluntary switch. */
889 #define SW_INVOL 0x0200 /* Involuntary switch. */
890 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */
891
892 /* How values for thread_single(). */
893 #define SINGLE_NO_EXIT 0
894 #define SINGLE_EXIT 1
895 #define SINGLE_BOUNDARY 2
896 #define SINGLE_ALLPROC 3
897
898 #ifdef MALLOC_DECLARE
899 MALLOC_DECLARE(M_PARGS);
900 MALLOC_DECLARE(M_SESSION);
901 MALLOC_DECLARE(M_SUBPROC);
902 #endif
903
904 #define FOREACH_PROC_IN_SYSTEM(p) \
905 LIST_FOREACH((p), &allproc, p_list)
906 #define FOREACH_THREAD_IN_PROC(p, td) \
907 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
908
909 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
910
911 /*
912 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
913 * in a pid_t, as it is used to represent "no process group".
914 */
915 #define PID_MAX 99999
916 #define NO_PID 100000
917 #define THREAD0_TID NO_PID
918 extern pid_t pid_max;
919
920 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
921
922 /* Lock and unlock a process. */
923 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
924 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
925 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
926 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
927 #define PROC_WAIT_UNLOCKED(p) mtx_wait_unlocked(&(p)->p_mtx)
928 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
929
930 /* Lock and unlock a process group. */
931 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
932 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
933 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
934 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
935
936 #define PGRP_LOCK_PGSIGNAL(pg) do { \
937 if ((pg) != NULL) \
938 PGRP_LOCK(pg); \
939 } while (0)
940 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \
941 if ((pg) != NULL) \
942 PGRP_UNLOCK(pg); \
943 } while (0)
944
945 /* Lock and unlock a session. */
946 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
947 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
948 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
949 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
950
951 /*
952 * Non-zero p_lock ensures that:
953 * - exit1() is not performed until p_lock reaches zero;
954 * - the process' threads stack are not swapped out if they are currently
955 * not (P_INMEM).
956 *
957 * PHOLD() asserts that the process (except the current process) is
958 * not exiting, increments p_lock and swaps threads stacks into memory,
959 * if needed.
960 * _PHOLD() is same as PHOLD(), it takes the process locked.
961 * _PHOLD_LITE() also takes the process locked, but comparing with
962 * _PHOLD(), it only guarantees that exit1() is not executed,
963 * faultin() is not called.
964 */
965 #define PHOLD(p) do { \
966 PROC_LOCK(p); \
967 _PHOLD(p); \
968 PROC_UNLOCK(p); \
969 } while (0)
970 #define _PHOLD(p) do { \
971 PROC_LOCK_ASSERT((p), MA_OWNED); \
972 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
973 ("PHOLD of exiting process %p", p)); \
974 (p)->p_lock++; \
975 if (((p)->p_flag & P_INMEM) == 0) \
976 faultin((p)); \
977 } while (0)
978 #define _PHOLD_LITE(p) do { \
979 PROC_LOCK_ASSERT((p), MA_OWNED); \
980 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
981 ("PHOLD of exiting process %p", p)); \
982 (p)->p_lock++; \
983 } while (0)
984 #define PROC_ASSERT_HELD(p) do { \
985 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \
986 } while (0)
987
988 #define PRELE(p) do { \
989 PROC_LOCK((p)); \
990 _PRELE((p)); \
991 PROC_UNLOCK((p)); \
992 } while (0)
993 #define _PRELE(p) do { \
994 PROC_LOCK_ASSERT((p), MA_OWNED); \
995 PROC_ASSERT_HELD(p); \
996 (--(p)->p_lock); \
997 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
998 wakeup(&(p)->p_lock); \
999 } while (0)
1000 #define PROC_ASSERT_NOT_HELD(p) do { \
1001 KASSERT((p)->p_lock == 0, ("process %p held", p)); \
1002 } while (0)
1003
1004 #define PROC_UPDATE_COW(p) do { \
1005 PROC_LOCK_ASSERT((p), MA_OWNED); \
1006 (p)->p_cowgen++; \
1007 } while (0)
1008
1009 /* Check whether a thread is safe to be swapped out. */
1010 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP)
1011
1012 /* Control whether or not it is safe for curthread to sleep. */
1013 #define THREAD_NO_SLEEPING() do { \
1014 curthread->td_no_sleeping++; \
1015 MPASS(curthread->td_no_sleeping > 0); \
1016 } while (0)
1017
1018 #define THREAD_SLEEPING_OK() do { \
1019 MPASS(curthread->td_no_sleeping > 0); \
1020 curthread->td_no_sleeping--; \
1021 } while (0)
1022
1023 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0)
1024
1025 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
1026 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)])
1027 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
1028 extern struct sx *pidhashtbl_lock;
1029 extern u_long pidhash;
1030 extern u_long pidhashlock;
1031
1032 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
1033 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
1034 extern u_long pgrphash;
1035
1036 extern struct sx allproc_lock;
1037 extern int allproc_gen;
1038 extern struct sx proctree_lock;
1039 extern struct mtx ppeers_lock;
1040 extern struct mtx procid_lock;
1041 extern struct proc proc0; /* Process slot for swapper. */
1042 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */
1043 #define thread0 (thread0_st.t0st_thread)
1044 extern struct vmspace vmspace0; /* VM space for proc0. */
1045 extern int hogticks; /* Limit on kernel cpu hogs. */
1046 extern int lastpid;
1047 extern int nprocs, maxproc; /* Current and max number of procs. */
1048 extern int maxprocperuid; /* Max procs per uid. */
1049 extern u_long ps_arg_cache_limit;
1050
1051 LIST_HEAD(proclist, proc);
1052 TAILQ_HEAD(procqueue, proc);
1053 TAILQ_HEAD(threadqueue, thread);
1054 extern struct proclist allproc; /* List of all processes. */
1055 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
1056
1057 extern struct uma_zone *proc_zone;
1058 extern struct uma_zone *pgrp_zone;
1059
1060 struct proc *pfind(pid_t); /* Find process by id. */
1061 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */
1062 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */
1063 struct pgrp *pgfind(pid_t); /* Find process group by id. */
1064 void pidhash_slockall(void); /* Shared lock all pid hash lists. */
1065 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */
1066
1067 struct fork_req {
1068 int fr_flags;
1069 int fr_pages;
1070 int *fr_pidp;
1071 struct proc **fr_procp;
1072 int *fr_pd_fd;
1073 int fr_pd_flags;
1074 struct filecaps *fr_pd_fcaps;
1075 int fr_flags2;
1076 #define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */
1077 #define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */
1078 #define FR2_KPROC 0x00000004 /* Create a kernel process */
1079 };
1080
1081 /*
1082 * pget() flags.
1083 */
1084 #define PGET_HOLD 0x00001 /* Hold the process. */
1085 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */
1086 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */
1087 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */
1088 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */
1089 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */
1090 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */
1091
1092 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
1093
1094 int pget(pid_t pid, int flags, struct proc **pp);
1095
1096 void ast(struct trapframe *framep);
1097 struct thread *choosethread(void);
1098 int cr_bsd_visible(struct ucred *u1, struct ucred *u2);
1099 int cr_cansee(struct ucred *u1, struct ucred *u2);
1100 int cr_canseesocket(struct ucred *cred, struct socket *so);
1101 int cr_canseeothergids(struct ucred *u1, struct ucred *u2);
1102 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2);
1103 int cr_canseejailproc(struct ucred *u1, struct ucred *u2);
1104 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
1105 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
1106 struct session *sess);
1107 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
1108 void faultin(struct proc *p);
1109 int fork1(struct thread *, struct fork_req *);
1110 void fork_rfppwait(struct thread *);
1111 void fork_exit(void (*)(void *, struct trapframe *), void *,
1112 struct trapframe *);
1113 void fork_return(struct thread *, struct trapframe *);
1114 int inferior(struct proc *p);
1115 void itimer_proc_continue(struct proc *p);
1116 void kqtimer_proc_continue(struct proc *p);
1117 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry,
1118 int *resident_count, bool *super);
1119 void kern_yield(int);
1120 void kick_proc0(void);
1121 void killjobc(void);
1122 int leavepgrp(struct proc *p);
1123 int maybe_preempt(struct thread *td);
1124 void maybe_yield(void);
1125 void mi_switch(int flags);
1126 int p_candebug(struct thread *td, struct proc *p);
1127 int p_cansee(struct thread *td, struct proc *p);
1128 int p_cansched(struct thread *td, struct proc *p);
1129 int p_cansignal(struct thread *td, struct proc *p, int signum);
1130 int p_canwait(struct thread *td, struct proc *p);
1131 struct pargs *pargs_alloc(int len);
1132 void pargs_drop(struct pargs *pa);
1133 void pargs_hold(struct pargs *pa);
1134 void proc_add_orphan(struct proc *child, struct proc *parent);
1135 int proc_get_binpath(struct proc *p, char *binname, char **fullpath,
1136 char **freepath);
1137 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
1138 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
1139 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
1140 void procinit(void);
1141 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg);
1142 void proc_linkup0(struct proc *p, struct thread *td);
1143 void proc_linkup(struct proc *p, struct thread *td);
1144 struct proc *proc_realparent(struct proc *child);
1145 void proc_reap(struct thread *td, struct proc *p, int *status, int options);
1146 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid);
1147 void proc_set_p2_wexit(struct proc *p);
1148 void proc_set_traced(struct proc *p, bool stop);
1149 void proc_wkilled(struct proc *p);
1150 struct pstats *pstats_alloc(void);
1151 void pstats_fork(struct pstats *src, struct pstats *dst);
1152 void pstats_free(struct pstats *ps);
1153 void proc_clear_orphan(struct proc *p);
1154 void reaper_abandon_children(struct proc *p, bool exiting);
1155 int securelevel_ge(struct ucred *cr, int level);
1156 int securelevel_gt(struct ucred *cr, int level);
1157 void sess_hold(struct session *);
1158 void sess_release(struct session *);
1159 int setrunnable(struct thread *, int);
1160 void setsugid(struct proc *p);
1161 bool should_yield(void);
1162 int sigonstack(size_t sp);
1163 void stopevent(struct proc *, u_int, u_int);
1164 struct thread *tdfind(lwpid_t, pid_t);
1165 void threadinit(void);
1166 void tidhash_add(struct thread *);
1167 void tidhash_remove(struct thread *);
1168 void cpu_idle(int);
1169 int cpu_idle_wakeup(int);
1170 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */
1171 void cpu_switch(struct thread *, struct thread *, struct mtx *);
1172 void cpu_sync_core(void);
1173 void cpu_throw(struct thread *, struct thread *) __dead2;
1174 bool curproc_sigkilled(void);
1175 void userret(struct thread *, struct trapframe *);
1176
1177 void cpu_exit(struct thread *);
1178 void exit1(struct thread *, int, int) __dead2;
1179 void cpu_copy_thread(struct thread *td, struct thread *td0);
1180 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map);
1181 int cpu_fetch_syscall_args(struct thread *td);
1182 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
1183 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *);
1184 int cpu_procctl(struct thread *td, int idtype, id_t id, int com,
1185 void *data);
1186 void cpu_set_syscall_retval(struct thread *, int);
1187 int cpu_set_upcall(struct thread *, void (*)(void *), void *,
1188 stack_t *);
1189 int cpu_set_user_tls(struct thread *, void *tls_base);
1190 void cpu_thread_alloc(struct thread *);
1191 void cpu_thread_clean(struct thread *);
1192 void cpu_thread_exit(struct thread *);
1193 void cpu_thread_free(struct thread *);
1194 void cpu_thread_swapin(struct thread *);
1195 void cpu_thread_swapout(struct thread *);
1196 struct thread *thread_alloc(int pages);
1197 int thread_alloc_stack(struct thread *, int pages);
1198 int thread_check_susp(struct thread *td, bool sleep);
1199 void thread_cow_get_proc(struct thread *newtd, struct proc *p);
1200 void thread_cow_get(struct thread *newtd, struct thread *td);
1201 void thread_cow_free(struct thread *td);
1202 void thread_cow_update(struct thread *td);
1203 int thread_create(struct thread *td, struct rtprio *rtp,
1204 int (*initialize_thread)(struct thread *, void *), void *thunk);
1205 void thread_exit(void) __dead2;
1206 void thread_free(struct thread *td);
1207 void thread_link(struct thread *td, struct proc *p);
1208 void thread_reap_barrier(void);
1209 int thread_single(struct proc *p, int how);
1210 void thread_single_end(struct proc *p, int how);
1211 void thread_stash(struct thread *td);
1212 void thread_stopped(struct proc *p);
1213 void childproc_stopped(struct proc *child, int reason);
1214 void childproc_continued(struct proc *child);
1215 void childproc_exited(struct proc *child);
1216 void thread_run_flash(struct thread *td);
1217 int thread_suspend_check(int how);
1218 bool thread_suspend_check_needed(void);
1219 void thread_suspend_switch(struct thread *, struct proc *p);
1220 void thread_suspend_one(struct thread *td);
1221 void thread_unlink(struct thread *td);
1222 void thread_unsuspend(struct proc *p);
1223 void thread_wait(struct proc *p);
1224
1225 bool stop_all_proc_block(void);
1226 void stop_all_proc_unblock(void);
1227 void stop_all_proc(void);
1228 void resume_all_proc(void);
1229
1230 static __inline int
curthread_pflags_set(int flags)1231 curthread_pflags_set(int flags)
1232 {
1233 struct thread *td;
1234 int save;
1235
1236 td = curthread;
1237 save = ~flags | (td->td_pflags & flags);
1238 td->td_pflags |= flags;
1239 return (save);
1240 }
1241
1242 static __inline void
curthread_pflags_restore(int save)1243 curthread_pflags_restore(int save)
1244 {
1245
1246 curthread->td_pflags &= save;
1247 }
1248
1249 static __inline int
curthread_pflags2_set(int flags)1250 curthread_pflags2_set(int flags)
1251 {
1252 struct thread *td;
1253 int save;
1254
1255 td = curthread;
1256 save = ~flags | (td->td_pflags2 & flags);
1257 td->td_pflags2 |= flags;
1258 return (save);
1259 }
1260
1261 static __inline void
curthread_pflags2_restore(int save)1262 curthread_pflags2_restore(int save)
1263 {
1264
1265 curthread->td_pflags2 &= save;
1266 }
1267
1268 static __inline bool
kstack_contains(struct thread * td,vm_offset_t va,size_t len)1269 kstack_contains(struct thread *td, vm_offset_t va, size_t len)
1270 {
1271 return (va >= td->td_kstack && va + len >= va &&
1272 va + len <= td->td_kstack + td->td_kstack_pages * PAGE_SIZE);
1273 }
1274
1275 static __inline __pure2 struct td_sched *
td_get_sched(struct thread * td)1276 td_get_sched(struct thread *td)
1277 {
1278
1279 return ((struct td_sched *)&td[1]);
1280 }
1281
1282 extern void (*softdep_ast_cleanup)(struct thread *);
1283 static __inline void
td_softdep_cleanup(struct thread * td)1284 td_softdep_cleanup(struct thread *td)
1285 {
1286
1287 if (td->td_su != NULL && softdep_ast_cleanup != NULL)
1288 softdep_ast_cleanup(td);
1289 }
1290
1291 #define PROC_ID_PID 0
1292 #define PROC_ID_GROUP 1
1293 #define PROC_ID_SESSION 2
1294 #define PROC_ID_REAP 3
1295
1296 void proc_id_set(int type, pid_t id);
1297 void proc_id_set_cond(int type, pid_t id);
1298 void proc_id_clear(int type, pid_t id);
1299
1300 EVENTHANDLER_LIST_DECLARE(process_ctor);
1301 EVENTHANDLER_LIST_DECLARE(process_dtor);
1302 EVENTHANDLER_LIST_DECLARE(process_init);
1303 EVENTHANDLER_LIST_DECLARE(process_fini);
1304 EVENTHANDLER_LIST_DECLARE(process_exit);
1305 EVENTHANDLER_LIST_DECLARE(process_fork);
1306 EVENTHANDLER_LIST_DECLARE(process_exec);
1307
1308 EVENTHANDLER_LIST_DECLARE(thread_ctor);
1309 EVENTHANDLER_LIST_DECLARE(thread_dtor);
1310 EVENTHANDLER_LIST_DECLARE(thread_init);
1311
1312 #endif /* _KERNEL */
1313
1314 #endif /* !_SYS_PROC_H_ */
1315