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 * $FreeBSD$
38 */
39
40 #ifndef _SYS_PROC_H_
41 #define _SYS_PROC_H_
42
43 #include <sys/callout.h> /* For struct callout. */
44 #include <sys/event.h> /* For struct klist. */
45 #ifdef _KERNEL
46 #include <sys/_eventhandler.h>
47 #endif
48 #include <sys/condvar.h>
49 #ifndef _KERNEL
50 #include <sys/filedesc.h>
51 #endif
52 #include <sys/queue.h>
53 #include <sys/_lock.h>
54 #include <sys/lock_profile.h>
55 #include <sys/_mutex.h>
56 #include <sys/osd.h>
57 #include <sys/priority.h>
58 #include <sys/rtprio.h> /* XXX. */
59 #include <sys/runq.h>
60 #include <sys/resource.h>
61 #include <sys/sigio.h>
62 #include <sys/signal.h>
63 #include <sys/signalvar.h>
64 #ifndef _KERNEL
65 #include <sys/time.h> /* For structs itimerval, timeval. */
66 #else
67 #include <sys/pcpu.h>
68 #include <sys/systm.h>
69 #endif
70 #include <sys/ucontext.h>
71 #include <sys/ucred.h>
72 #include <sys/types.h>
73 #include <sys/_domainset.h>
74
75 #include <machine/proc.h> /* Machine-dependent proc substruct. */
76 #ifdef _KERNEL
77 #include <machine/cpu.h>
78 #endif
79
80 /*
81 * One structure allocated per session.
82 *
83 * List of locks
84 * (m) locked by s_mtx mtx
85 * (e) locked by proctree_lock sx
86 * (c) const until freeing
87 */
88 struct session {
89 u_int s_count; /* Ref cnt; pgrps in session - atomic. */
90 struct proc *s_leader; /* (m + e) Session leader. */
91 struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
92 struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */
93 struct tty *s_ttyp; /* (e) Controlling tty. */
94 pid_t s_sid; /* (c) Session ID. */
95 /* (m) Setlogin() name: */
96 char s_login[roundup(MAXLOGNAME, sizeof(long))];
97 struct mtx s_mtx; /* Mutex to protect members. */
98 };
99
100 /*
101 * One structure allocated per process group.
102 *
103 * List of locks
104 * (m) locked by pg_mtx mtx
105 * (e) locked by proctree_lock sx
106 * (c) const until freeing
107 */
108 struct pgrp {
109 LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
110 LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
111 struct session *pg_session; /* (c) Pointer to session. */
112 struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
113 pid_t pg_id; /* (c) Process group id. */
114 struct mtx pg_mtx; /* Mutex to protect members */
115 int pg_flags; /* (m) PGRP_ flags */
116 };
117
118 #define PGRP_ORPHANED 0x00000001 /* Group is orphaned */
119
120 /*
121 * pargs, used to hold a copy of the command line, if it had a sane length.
122 */
123 struct pargs {
124 u_int ar_ref; /* Reference count. */
125 u_int ar_length; /* Length. */
126 u_char ar_args[1]; /* Arguments. */
127 };
128
129 /*-
130 * Description of a process.
131 *
132 * This structure contains the information needed to manage a thread of
133 * control, known in UN*X as a process; it has references to substructures
134 * containing descriptions of things that the process uses, but may share
135 * with related processes. The process structure and the substructures
136 * are always addressable except for those marked "(CPU)" below,
137 * which might be addressable only on a processor on which the process
138 * is running.
139 *
140 * Below is a key of locks used to protect each member of struct proc. The
141 * lock is indicated by a reference to a specific character in parens in the
142 * associated comment.
143 * * - not yet protected
144 * a - only touched by curproc or parent during fork/wait
145 * b - created at fork, never changes
146 * (exception aiods switch vmspaces, but they are also
147 * marked 'P_SYSTEM' so hopefully it will be left alone)
148 * c - locked by proc mtx
149 * d - locked by allproc_lock lock
150 * e - locked by proctree_lock lock
151 * f - session mtx
152 * g - process group mtx
153 * h - callout_lock mtx
154 * i - by curproc or the master session mtx
155 * j - locked by proc slock
156 * k - only accessed by curthread
157 * k*- only accessed by curthread and from an interrupt
158 * kx- only accessed by curthread and by debugger
159 * l - the attaching proc or attaching proc parent
160 * m - Giant
161 * n - not locked, lazy
162 * o - ktrace lock
163 * q - td_contested lock
164 * r - p_peers lock
165 * s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9)
166 * t - thread lock
167 * u - process stat lock
168 * w - process timer lock
169 * x - created at fork, only changes during single threading in exec
170 * y - created at first aio, doesn't change until exit or exec at which
171 * point we are single-threaded and only curthread changes it
172 * z - zombie threads lock
173 *
174 * If the locking key specifies two identifiers (for example, p_pptr) then
175 * either lock is sufficient for read access, but both locks must be held
176 * for write access.
177 */
178 struct cpuset;
179 struct filecaps;
180 struct filemon;
181 struct kaioinfo;
182 struct kaudit_record;
183 struct kcov_info;
184 struct kdtrace_proc;
185 struct kdtrace_thread;
186 struct mqueue_notifier;
187 struct p_sched;
188 struct proc;
189 struct procdesc;
190 struct racct;
191 struct sbuf;
192 struct sleepqueue;
193 struct socket;
194 struct syscall_args;
195 struct td_sched;
196 struct thread;
197 struct trapframe;
198 struct turnstile;
199 struct vm_map;
200 struct vm_map_entry;
201 struct epoch_tracker;
202
203 /*
204 * XXX: Does this belong in resource.h or resourcevar.h instead?
205 * Resource usage extension. The times in rusage structs in the kernel are
206 * never up to date. The actual times are kept as runtimes and tick counts
207 * (with control info in the "previous" times), and are converted when
208 * userland asks for rusage info. Backwards compatibility prevents putting
209 * this directly in the user-visible rusage struct.
210 *
211 * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
212 * Locking for td_rux: (t) for all fields.
213 */
214 struct rusage_ext {
215 uint64_t rux_runtime; /* (cu) Real time. */
216 uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */
217 uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */
218 uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */
219 uint64_t rux_uu; /* (c) Previous user time in usec. */
220 uint64_t rux_su; /* (c) Previous sys time in usec. */
221 uint64_t rux_tu; /* (c) Previous total time in usec. */
222 };
223
224 /*
225 * Kernel runnable context (thread).
226 * This is what is put to sleep and reactivated.
227 * Thread context. Processes may have multiple threads.
228 */
229 struct thread {
230 struct mtx *volatile td_lock; /* replaces sched lock */
231 struct proc *td_proc; /* (*) Associated process. */
232 TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
233 TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */
234 union {
235 TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */
236 struct thread *td_zombie; /* Zombie list linkage */
237 };
238 TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */
239 LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */
240 struct cpuset *td_cpuset; /* (t) CPU affinity mask. */
241 struct domainset_ref td_domain; /* (a) NUMA policy */
242 struct seltd *td_sel; /* Select queue/channel. */
243 struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
244 struct turnstile *td_turnstile; /* (k) Associated turnstile. */
245 struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */
246 struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
247 lwpid_t td_tid; /* (b) Thread ID. */
248 sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
249 #define td_siglist td_sigqueue.sq_signals
250 u_char td_lend_user_pri; /* (t) Lend user pri. */
251 u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */
252
253 /* Cleared during fork1() */
254 #define td_startzero td_flags
255 int td_flags; /* (t) TDF_* flags. */
256 int td_inhibitors; /* (t) Why can not run. */
257 int td_pflags; /* (k) Private thread (TDP_*) flags. */
258 int td_pflags2; /* (k) Private thread (TDP2_*) flags. */
259 int td_dupfd; /* (k) Ret value from fdopen. XXX */
260 int td_sqqueue; /* (t) Sleepqueue queue blocked on. */
261 const void *td_wchan; /* (t) Sleep address. */
262 const char *td_wmesg; /* (t) Reason for sleep. */
263 volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
264 u_char td_tsqueue; /* (t) Turnstile queue blocked on. */
265 short td_locks; /* (k) Debug: count of non-spin locks */
266 short td_rw_rlocks; /* (k) Count of rwlock read locks. */
267 short td_sx_slocks; /* (k) Count of sx shared locks. */
268 short td_lk_slocks; /* (k) Count of lockmgr shared locks. */
269 short td_stopsched; /* (k) Scheduler stopped. */
270 struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */
271 const char *td_lockname; /* (t) Name of lock blocked on. */
272 LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
273 struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
274 int td_intr_nesting_level; /* (k) Interrupt recursion. */
275 int td_pinned; /* (k) Temporary cpu pin count. */
276 struct ucred *td_realucred; /* (k) Reference to credentials. */
277 struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */
278 struct plimit *td_limit; /* (k) Resource limits. */
279 int td_slptick; /* (t) Time at sleep. */
280 int td_blktick; /* (t) Time spent blocked. */
281 int td_swvoltick; /* (t) Time at last SW_VOL switch. */
282 int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */
283 u_int td_cow; /* (*) Number of copy-on-write faults */
284 struct rusage td_ru; /* (t) rusage information. */
285 struct rusage_ext td_rux; /* (t) Internal rusage information. */
286 uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */
287 uint64_t td_runtime; /* (t) How many cpu ticks we've run. */
288 u_int td_pticks; /* (t) Statclock hits for profiling */
289 u_int td_sticks; /* (t) Statclock hits in system mode. */
290 u_int td_iticks; /* (t) Statclock hits in intr mode. */
291 u_int td_uticks; /* (t) Statclock hits in user mode. */
292 int td_intrval; /* (t) Return value for sleepq. */
293 sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
294 volatile u_int td_generation; /* (k) For detection of preemption */
295 stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
296 int td_xsig; /* (c) Signal for ptrace */
297 u_long td_profil_addr; /* (k) Temporary addr until AST. */
298 u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
299 char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
300 struct file *td_fpop; /* (k) file referencing cdev under op */
301 int td_dbgflags; /* (c) Userland debugger flags */
302 siginfo_t td_si; /* (c) For debugger or core file */
303 int td_ng_outbound; /* (k) Thread entered ng from above. */
304 struct osd td_osd; /* (k) Object specific data. */
305 struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
306 pid_t td_dbg_forked; /* (c) Child pid for debugger. */
307 struct vnode *td_vp_reserved;/* (k) Prealloated vnode. */
308 u_int td_no_sleeping; /* (k) Sleeping disabled count. */
309 void *td_su; /* (k) FFS SU private */
310 sbintime_t td_sleeptimo; /* (t) Sleep timeout. */
311 int td_rtcgen; /* (s) rtc_generation of abs. sleep */
312 int td_errno; /* (k) Error from last syscall. */
313 size_t td_vslock_sz; /* (k) amount of vslock-ed space */
314 struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */
315 u_int td_ucredref; /* (k) references on td_realucred */
316 #define td_endzero td_sigmask
317
318 /* Copied during fork1() or create_thread(). */
319 #define td_startcopy td_endzero
320 sigset_t td_sigmask; /* (c) Current signal mask. */
321 u_char td_rqindex; /* (t) Run queue index. */
322 u_char td_base_pri; /* (t) Thread base kernel priority. */
323 u_char td_priority; /* (t) Thread active priority. */
324 u_char td_pri_class; /* (t) Scheduling class. */
325 u_char td_user_pri; /* (t) User pri from estcpu and nice. */
326 u_char td_base_user_pri; /* (t) Base user pri */
327 u_char td_pre_epoch_prio; /* (k) User pri on entry to epoch */
328 uintptr_t td_rb_list; /* (k) Robust list head. */
329 uintptr_t td_rbp_list; /* (k) Robust priv list head. */
330 uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */
331 struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on
332 fork for child tracing. */
333 void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */
334 uint32_t td_sigblock_val; /* (k) fast sigblock value read at
335 td_sigblock_ptr on kern entry */
336 #define td_endcopy td_pcb
337
338 /*
339 * Fields that must be manually set in fork1() or create_thread()
340 * or already have been set in the allocator, constructor, etc.
341 */
342 struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
343 enum td_states {
344 TDS_INACTIVE = 0x0,
345 TDS_INHIBITED,
346 TDS_CAN_RUN,
347 TDS_RUNQ,
348 TDS_RUNNING
349 } td_state; /* (t) thread state */
350 union {
351 register_t tdu_retval[2];
352 off_t tdu_off;
353 } td_uretoff; /* (k) Syscall aux returns. */
354 #define td_retval td_uretoff.tdu_retval
355 u_int td_cowgen; /* (k) Generation of COW pointers. */
356 /* LP64 hole */
357 struct callout td_slpcallout; /* (h) Callout for sleep. */
358 struct trapframe *td_frame; /* (k) */
359 vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
360 int td_kstack_pages; /* (a) Size of the kstack. */
361 volatile u_int td_critnest; /* (k*) Critical section nest level. */
362 struct mdthread td_md; /* (k) Any machine-dependent fields. */
363 struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
364 struct lpohead td_lprof[2]; /* (a) lock profiling objects. */
365 struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */
366 struct vnet *td_vnet; /* (k) Effective vnet. */
367 const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */
368 struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
369 struct proc *td_rfppwait_p; /* (k) The vforked child */
370 struct vm_page **td_ma; /* (k) uio pages held */
371 int td_ma_cnt; /* (k) size of *td_ma */
372 /* LP64 hole */
373 void *td_emuldata; /* Emulator state data */
374 int td_lastcpu; /* (t) Last cpu we were on. */
375 int td_oncpu; /* (t) Which cpu we are on. */
376 void *td_lkpi_task; /* LinuxKPI task struct pointer */
377 int td_pmcpend;
378 #ifdef EPOCH_TRACE
379 SLIST_HEAD(, epoch_tracker) td_epochs;
380 #endif
381 };
382
383 struct thread0_storage {
384 struct thread t0st_thread;
385 uint64_t t0st_sched[10];
386 };
387
388 struct mtx *thread_lock_block(struct thread *);
389 void thread_lock_block_wait(struct thread *);
390 void thread_lock_set(struct thread *, struct mtx *);
391 void thread_lock_unblock(struct thread *, struct mtx *);
392 #define THREAD_LOCK_ASSERT(td, type) \
393 mtx_assert((td)->td_lock, (type))
394
395 #define THREAD_LOCK_BLOCKED_ASSERT(td, type) \
396 do { \
397 struct mtx *__m = (td)->td_lock; \
398 if (__m != &blocked_lock) \
399 mtx_assert(__m, (type)); \
400 } while (0)
401
402 #ifdef INVARIANTS
403 #define THREAD_LOCKPTR_ASSERT(td, lock) \
404 do { \
405 struct mtx *__m; \
406 __m = (td)->td_lock; \
407 KASSERT(__m == (lock), \
408 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
409 } while (0)
410
411 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \
412 do { \
413 struct mtx *__m; \
414 __m = (td)->td_lock; \
415 KASSERT(__m == (lock) || __m == &blocked_lock, \
416 ("Thread %p lock %p does not match %p", td, __m, (lock))); \
417 } while (0)
418
419 #define TD_LOCKS_INC(td) ((td)->td_locks++)
420 #define TD_LOCKS_DEC(td) do { \
421 KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \
422 ("thread %p owns no locks", (td))); \
423 (td)->td_locks--; \
424 } while (0)
425 #else
426 #define THREAD_LOCKPTR_ASSERT(td, lock)
427 #define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock)
428
429 #define TD_LOCKS_INC(td)
430 #define TD_LOCKS_DEC(td)
431 #endif
432
433 /*
434 * Flags kept in td_flags:
435 * To change these you MUST have the scheduler lock.
436 */
437 #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
438 #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
439 #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */
440 #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
441 #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
442 #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
443 #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */
444 #define TDF_UNUSED80 0x00000080 /* unused. */
445 #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */
446 #define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */
447 #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
448 #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */
449 #define TDF_UNUSED12 0x00001000 /* --available-- */
450 #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */
451 #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
452 #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */
453 #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */
454 #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */
455 #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */
456 #define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */
457 #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
458 #define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */
459 #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */
460 #define TDF_UNUSED23 0x00800000 /* --available-- */
461 #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
462 #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
463 #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
464 #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
465 #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */
466 #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */
467 #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */
468
469 /* Userland debug flags */
470 #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */
471 #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */
472 #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */
473 #define TDB_SCE 0x00000008 /* Thread performs syscall enter */
474 #define TDB_SCX 0x00000010 /* Thread performs syscall exit */
475 #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */
476 #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new
477 process */
478 #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child
479 only) */
480 #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */
481 #define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */
482 #define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */
483 #define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */
484 #define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */
485 #define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */
486
487 /*
488 * "Private" flags kept in td_pflags:
489 * These are only written by curthread and thus need no locking.
490 */
491 #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
492 #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
493 #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
494 #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */
495 #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
496 #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
497 #define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */
498 #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */
499 #define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */
500 #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
501 #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
502 #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */
503 #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
504 #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
505 #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
506 #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
507 #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
508 #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
509 #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
510 #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
511 #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */
512 #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */
513 #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */
514 #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
515 #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */
516 #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */
517 #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */
518 #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */
519 #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */
520 #define TDP_FORKING 0x20000000 /* Thread is being created through fork() */
521 #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */
522 #define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */
523
524 #define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */
525 #define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */
526
527 /*
528 * Reasons that the current thread can not be run yet.
529 * More than one may apply.
530 */
531 #define TDI_SUSPENDED 0x0001 /* On suspension queue. */
532 #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
533 #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */
534 #define TDI_LOCK 0x0008 /* Stopped on a lock. */
535 #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
536
537 #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
538 #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
539 #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
540 #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
541 #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
542 #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
543 #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING)
544 #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ)
545 #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN)
546 #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED)
547 #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
548 #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD)
549
550 #define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \
551 ((td)->td_flags & TDF_SINTR) != 0)
552
553 #define KTDSTATE(td) \
554 (((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \
555 ((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \
556 ((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \
557 ((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \
558 ((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding")
559
560 #define TD_SET_INHIB(td, inhib) do { \
561 (td)->td_state = TDS_INHIBITED; \
562 (td)->td_inhibitors |= (inhib); \
563 } while (0)
564
565 #define TD_CLR_INHIB(td, inhib) do { \
566 if (((td)->td_inhibitors & (inhib)) && \
567 (((td)->td_inhibitors &= ~(inhib)) == 0)) \
568 (td)->td_state = TDS_CAN_RUN; \
569 } while (0)
570
571 #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
572 #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
573 #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
574 #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
575 #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
576 #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
577
578 #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
579 #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
580 #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
581 #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
582 #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
583
584 #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING
585 #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ
586 #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN
587
588 #define TD_SBDRY_INTR(td) \
589 (((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0)
590 #define TD_SBDRY_ERRNO(td) \
591 (((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART)
592
593 /*
594 * Process structure.
595 */
596 struct proc {
597 LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
598 TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */
599 struct mtx p_slock; /* process spin lock */
600 struct ucred *p_ucred; /* (c) Process owner's identity. */
601 struct filedesc *p_fd; /* (b) Open files. */
602 struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
603 struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */
604 struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
605 struct plimit *p_limit; /* (c) Resource limits. */
606 struct callout p_limco; /* (c) Limit callout handle */
607 struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
608
609 int p_flag; /* (c) P_* flags. */
610 int p_flag2; /* (c) P2_* flags. */
611 enum p_states {
612 PRS_NEW = 0, /* In creation */
613 PRS_NORMAL, /* threads can be run. */
614 PRS_ZOMBIE
615 } p_state; /* (j/c) Process status. */
616 pid_t p_pid; /* (b) Process identifier. */
617 LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
618 LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
619 struct proc *p_pptr; /* (c + e) Pointer to parent process. */
620 LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
621 LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
622 struct proc *p_reaper; /* (e) My reaper. */
623 LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants
624 (if I am reaper). */
625 LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of
626 the same reaper. */
627 struct mtx p_mtx; /* (n) Lock for this struct. */
628 struct mtx p_statmtx; /* Lock for the stats */
629 struct mtx p_itimmtx; /* Lock for the virt/prof timers */
630 struct mtx p_profmtx; /* Lock for the profiling */
631 struct ksiginfo *p_ksi; /* Locked by parent proc lock */
632 sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
633 #define p_siglist p_sigqueue.sq_signals
634 pid_t p_oppid; /* (c + e) Real parent pid. */
635
636 /* The following fields are all zeroed upon creation in fork. */
637 #define p_startzero p_vmspace
638 struct vmspace *p_vmspace; /* (b) Address space. */
639 u_int p_swtick; /* (c) Tick when swapped in or out. */
640 u_int p_cowgen; /* (c) Generation of COW pointers. */
641 struct itimerval p_realtimer; /* (c) Alarm timer. */
642 struct rusage p_ru; /* (a) Exit information. */
643 struct rusage_ext p_rux; /* (cu) Internal resource usage. */
644 struct rusage_ext p_crux; /* (c) Internal child resource usage. */
645 int p_profthreads; /* (c) Num threads in addupc_task. */
646 volatile int p_exitthreads; /* (j) Number of threads exiting */
647 int p_traceflag; /* (o) Kernel trace points. */
648 struct vnode *p_tracevp; /* (c + o) Trace to vnode. */
649 struct ucred *p_tracecred; /* (o) Credentials to trace with. */
650 struct vnode *p_textvp; /* (b) Vnode of executable. */
651 u_int p_lock; /* (c) Proclock (prevent swap) count. */
652 struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
653 int p_sigparent; /* (c) Signal to parent on exit. */
654 int p_sig; /* (n) For core dump/debugger XXX. */
655 u_int p_ptevents; /* (c + e) ptrace() event mask. */
656 struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */
657 struct thread *p_singlethread;/* (c + j) If single threading this is it */
658 int p_suspcount; /* (j) Num threads in suspended mode. */
659 struct thread *p_xthread; /* (c) Trap thread */
660 int p_boundary_count;/* (j) Num threads at user boundary */
661 int p_pendingcnt; /* how many signals are pending */
662 struct itimers *p_itimers; /* (c) POSIX interval timers. */
663 struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */
664 u_int p_treeflag; /* (e) P_TREE flags */
665 int p_pendingexits; /* (c) Count of pending thread exits. */
666 struct filemon *p_filemon; /* (c) filemon-specific data. */
667 int p_pdeathsig; /* (c) Signal from parent on exit. */
668 /* End area that is zeroed on creation. */
669 #define p_endzero p_magic
670
671 /* The following fields are all copied upon creation in fork. */
672 #define p_startcopy p_endzero
673 u_int p_magic; /* (b) Magic number. */
674 int p_osrel; /* (x) osreldate for the
675 binary (from ELF note, if any) */
676 uint32_t p_fctl0; /* (x) ABI feature control, ELF note */
677 char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */
678 struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
679 struct pargs *p_args; /* (c) Process arguments. */
680 rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */
681 signed char p_nice; /* (c) Process "nice" value. */
682 int p_fibnum; /* in this routing domain XXX MRT */
683 pid_t p_reapsubtree; /* (e) Pid of the direct child of the
684 reaper which spawned
685 our subtree. */
686 uint16_t p_elf_machine; /* (x) ELF machine type */
687 uint64_t p_elf_flags; /* (x) ELF flags */
688 /* End area that is copied on creation. */
689 #define p_endcopy p_xexit
690
691 u_int p_xexit; /* (c) Exit code. */
692 u_int p_xsig; /* (c) Stop/kill sig. */
693 struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
694 struct knlist *p_klist; /* (c) Knotes attached to this proc. */
695 int p_numthreads; /* (c) Number of threads. */
696 struct mdproc p_md; /* Any machine-dependent fields. */
697 struct callout p_itcallout; /* (h + c) Interval timer callout. */
698 u_short p_acflag; /* (c) Accounting flags. */
699 struct proc *p_peers; /* (r) */
700 struct proc *p_leader; /* (b) */
701 void *p_emuldata; /* (c) Emulator state data. */
702 struct label *p_label; /* (*) Proc (not subject) MAC label. */
703 STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
704 LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
705 struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */
706 struct cv p_pwait; /* (*) wait cv for exit/exec. */
707 uint64_t p_prev_runtime; /* (c) Resource usage accounting. */
708 struct racct *p_racct; /* (b) Resource accounting. */
709 int p_throttled; /* (c) Flag for racct pcpu throttling */
710 /*
711 * An orphan is the child that has been re-parented to the
712 * debugger as a result of attaching to it. Need to keep
713 * track of them for parent to be able to collect the exit
714 * status of what used to be children.
715 */
716 LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */
717 LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */
718 };
719
720 #define p_session p_pgrp->pg_session
721 #define p_pgid p_pgrp->pg_id
722
723 #define NOCPU (-1) /* For when we aren't on a CPU. */
724 #define NOCPU_OLD (255)
725 #define MAXCPU_OLD (254)
726
727 #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock)
728 #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock)
729 #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type))
730
731 #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx)
732 #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx)
733 #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type))
734
735 #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx)
736 #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx)
737 #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type))
738
739 #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx)
740 #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx)
741 #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type))
742
743 /* These flags are kept in p_flag. */
744 #define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory
745 lock. */
746 #define P_CONTROLT 0x00000002 /* Has a controlling terminal. */
747 #define P_KPROC 0x00000004 /* Kernel process. */
748 #define P_UNUSED3 0x00000008 /* --available-- */
749 #define P_PPWAIT 0x00000010 /* Parent is waiting for child to
750 exec/exit. */
751 #define P_PROFIL 0x00000020 /* Has started profiling. */
752 #define P_STOPPROF 0x00000040 /* Has thread requesting to stop
753 profiling. */
754 #define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup
755 shortcuts) */
756 #define P_SUGID 0x00000100 /* Had set id privileges since last
757 exec. */
758 #define P_SYSTEM 0x00000200 /* System proc: no sigs, stats or
759 swapping. */
760 #define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit,
761 not wait. */
762 #define P_TRACED 0x00000800 /* Debugged process being traced. */
763 #define P_WAITED 0x00001000 /* Someone is waiting for us. */
764 #define P_WEXIT 0x00002000 /* Working on exiting. */
765 #define P_EXEC 0x00004000 /* Process called exec. */
766 #define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary
767 ASAP. */
768 #define P_CONTINUED 0x00010000 /* Proc has continued from a stopped
769 state. */
770 #define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */
771 #define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */
772 #define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to
773 user). */
774 #define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */
775 #define P_SIGEVENT 0x00200000 /* Process pending signals changed. */
776 #define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user
777 boundary. */
778 #define P_HWPMC 0x00800000 /* Process is using HWPMCs */
779 #define P_JAILED 0x01000000 /* Process is in jail. */
780 #define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */
781 #define P_INEXEC 0x04000000 /* Process is in execve(). */
782 #define P_STATCHILD 0x08000000 /* Child process stopped or exited. */
783 #define P_INMEM 0x10000000 /* Loaded into memory. */
784 #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */
785 #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */
786 #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */
787
788 #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
789 #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
790 #define P_KILLED(p) ((p)->p_flag & P_WKILLED)
791
792 /* These flags are kept in p_flag2. */
793 #define P2_INHERIT_PROTECTED 0x00000001 /* New children get
794 P_PROTECTED. */
795 #define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or
796 coredumps. */
797 #define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on
798 exec(2). */
799 #define P2_AST_SU 0x00000008 /* Handles SU ast for
800 kthreads. */
801 #define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not
802 yet handled. */
803 #define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */
804 #define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */
805 #define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */
806 #define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk
807 area. */
808 #define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied
809 PROT_MAX. */
810 #define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied
811 PROT_MAX. */
812 #define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for
813 MAP_STACK */
814 #define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled
815 after exec */
816
817 /* Flags protected by proctree_lock, kept in p_treeflags. */
818 #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */
819 #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan
820 list */
821 #define P_TREE_REAPER 0x00000004 /* Reaper of subtree */
822 #define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */
823
824 /*
825 * These were process status values (p_stat), now they are only used in
826 * legacy conversion code.
827 */
828 #define SIDL 1 /* Process being created by fork. */
829 #define SRUN 2 /* Currently runnable. */
830 #define SSLEEP 3 /* Sleeping on an address. */
831 #define SSTOP 4 /* Process debugging or suspension. */
832 #define SZOMB 5 /* Awaiting collection by parent. */
833 #define SWAIT 6 /* Waiting for interrupt. */
834 #define SLOCK 7 /* Blocked on a lock. */
835
836 #define P_MAGIC 0xbeefface
837
838 #ifdef _KERNEL
839
840 /* Types and flags for mi_switch(). */
841 #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */
842 #define SWT_NONE 0 /* Unspecified switch. */
843 #define SWT_PREEMPT 1 /* Switching due to preemption. */
844 #define SWT_OWEPREEMPT 2 /* Switching due to owepreempt. */
845 #define SWT_TURNSTILE 3 /* Turnstile contention. */
846 #define SWT_SLEEPQ 4 /* Sleepq wait. */
847 #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */
848 #define SWT_RELINQUISH 6 /* yield call. */
849 #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */
850 #define SWT_IDLE 8 /* Switching from the idle thread. */
851 #define SWT_IWAIT 9 /* Waiting for interrupts. */
852 #define SWT_SUSPEND 10 /* Thread suspended. */
853 #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */
854 #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */
855 #define SWT_COUNT 13 /* Number of switch types. */
856 /* Flags */
857 #define SW_VOL 0x0100 /* Voluntary switch. */
858 #define SW_INVOL 0x0200 /* Involuntary switch. */
859 #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */
860
861 /* How values for thread_single(). */
862 #define SINGLE_NO_EXIT 0
863 #define SINGLE_EXIT 1
864 #define SINGLE_BOUNDARY 2
865 #define SINGLE_ALLPROC 3
866
867 #ifdef MALLOC_DECLARE
868 MALLOC_DECLARE(M_PARGS);
869 MALLOC_DECLARE(M_SESSION);
870 MALLOC_DECLARE(M_SUBPROC);
871 #endif
872
873 #define FOREACH_PROC_IN_SYSTEM(p) \
874 LIST_FOREACH((p), &allproc, p_list)
875 #define FOREACH_THREAD_IN_PROC(p, td) \
876 TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
877
878 #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
879
880 /*
881 * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
882 * in a pid_t, as it is used to represent "no process group".
883 */
884 #define PID_MAX 99999
885 #define NO_PID 100000
886 #define THREAD0_TID NO_PID
887 extern pid_t pid_max;
888
889 #define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
890
891 /* Lock and unlock a process. */
892 #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
893 #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
894 #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
895 #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
896 #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
897
898 /* Lock and unlock a process group. */
899 #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
900 #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
901 #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
902 #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
903
904 #define PGRP_LOCK_PGSIGNAL(pg) do { \
905 if ((pg) != NULL) \
906 PGRP_LOCK(pg); \
907 } while (0)
908 #define PGRP_UNLOCK_PGSIGNAL(pg) do { \
909 if ((pg) != NULL) \
910 PGRP_UNLOCK(pg); \
911 } while (0)
912
913 /* Lock and unlock a session. */
914 #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
915 #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
916 #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
917 #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
918
919 /*
920 * Non-zero p_lock ensures that:
921 * - exit1() is not performed until p_lock reaches zero;
922 * - the process' threads stack are not swapped out if they are currently
923 * not (P_INMEM).
924 *
925 * PHOLD() asserts that the process (except the current process) is
926 * not exiting, increments p_lock and swaps threads stacks into memory,
927 * if needed.
928 * _PHOLD() is same as PHOLD(), it takes the process locked.
929 * _PHOLD_LITE() also takes the process locked, but comparing with
930 * _PHOLD(), it only guarantees that exit1() is not executed,
931 * faultin() is not called.
932 */
933 #define PHOLD(p) do { \
934 PROC_LOCK(p); \
935 _PHOLD(p); \
936 PROC_UNLOCK(p); \
937 } while (0)
938 #define _PHOLD(p) do { \
939 PROC_LOCK_ASSERT((p), MA_OWNED); \
940 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
941 ("PHOLD of exiting process %p", p)); \
942 (p)->p_lock++; \
943 if (((p)->p_flag & P_INMEM) == 0) \
944 faultin((p)); \
945 } while (0)
946 #define _PHOLD_LITE(p) do { \
947 PROC_LOCK_ASSERT((p), MA_OWNED); \
948 KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
949 ("PHOLD of exiting process %p", p)); \
950 (p)->p_lock++; \
951 } while (0)
952 #define PROC_ASSERT_HELD(p) do { \
953 KASSERT((p)->p_lock > 0, ("process %p not held", p)); \
954 } while (0)
955
956 #define PRELE(p) do { \
957 PROC_LOCK((p)); \
958 _PRELE((p)); \
959 PROC_UNLOCK((p)); \
960 } while (0)
961 #define _PRELE(p) do { \
962 PROC_LOCK_ASSERT((p), MA_OWNED); \
963 PROC_ASSERT_HELD(p); \
964 (--(p)->p_lock); \
965 if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
966 wakeup(&(p)->p_lock); \
967 } while (0)
968 #define PROC_ASSERT_NOT_HELD(p) do { \
969 KASSERT((p)->p_lock == 0, ("process %p held", p)); \
970 } while (0)
971
972 #define PROC_UPDATE_COW(p) do { \
973 PROC_LOCK_ASSERT((p), MA_OWNED); \
974 (p)->p_cowgen++; \
975 } while (0)
976
977 /* Check whether a thread is safe to be swapped out. */
978 #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP)
979
980 /* Control whether or not it is safe for curthread to sleep. */
981 #define THREAD_NO_SLEEPING() do { \
982 curthread->td_no_sleeping++; \
983 MPASS(curthread->td_no_sleeping > 0); \
984 } while (0)
985
986 #define THREAD_SLEEPING_OK() do { \
987 MPASS(curthread->td_no_sleeping > 0); \
988 curthread->td_no_sleeping--; \
989 } while (0)
990
991 #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0)
992
993 #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
994 #define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)])
995 extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
996 extern struct sx *pidhashtbl_lock;
997 extern u_long pidhash;
998 extern u_long pidhashlock;
999
1000 #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
1001 extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
1002 extern u_long pgrphash;
1003
1004 extern struct sx allproc_lock;
1005 extern int allproc_gen;
1006 extern struct sx proctree_lock;
1007 extern struct mtx ppeers_lock;
1008 extern struct mtx procid_lock;
1009 extern struct proc proc0; /* Process slot for swapper. */
1010 extern struct thread0_storage thread0_st; /* Primary thread in proc0. */
1011 #define thread0 (thread0_st.t0st_thread)
1012 extern struct vmspace vmspace0; /* VM space for proc0. */
1013 extern int hogticks; /* Limit on kernel cpu hogs. */
1014 extern int lastpid;
1015 extern int nprocs, maxproc; /* Current and max number of procs. */
1016 extern int maxprocperuid; /* Max procs per uid. */
1017 extern u_long ps_arg_cache_limit;
1018
1019 LIST_HEAD(proclist, proc);
1020 TAILQ_HEAD(procqueue, proc);
1021 TAILQ_HEAD(threadqueue, thread);
1022 extern struct proclist allproc; /* List of all processes. */
1023 extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
1024
1025 extern struct uma_zone *proc_zone;
1026 extern struct uma_zone *pgrp_zone;
1027
1028 struct proc *pfind(pid_t); /* Find process by id. */
1029 struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */
1030 struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */
1031 struct pgrp *pgfind(pid_t); /* Find process group by id. */
1032 void pidhash_slockall(void); /* Shared lock all pid hash lists. */
1033 void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */
1034
1035 struct fork_req {
1036 int fr_flags;
1037 int fr_pages;
1038 int *fr_pidp;
1039 struct proc **fr_procp;
1040 int *fr_pd_fd;
1041 int fr_pd_flags;
1042 struct filecaps *fr_pd_fcaps;
1043 int fr_flags2;
1044 #define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */
1045 #define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */
1046 };
1047
1048 /*
1049 * pget() flags.
1050 */
1051 #define PGET_HOLD 0x00001 /* Hold the process. */
1052 #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */
1053 #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */
1054 #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */
1055 #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */
1056 #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */
1057 #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */
1058
1059 #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
1060
1061 int pget(pid_t pid, int flags, struct proc **pp);
1062
1063 void ast(struct trapframe *framep);
1064 struct thread *choosethread(void);
1065 int cr_cansee(struct ucred *u1, struct ucred *u2);
1066 int cr_canseesocket(struct ucred *cred, struct socket *so);
1067 int cr_canseeothergids(struct ucred *u1, struct ucred *u2);
1068 int cr_canseeotheruids(struct ucred *u1, struct ucred *u2);
1069 int cr_canseejailproc(struct ucred *u1, struct ucred *u2);
1070 int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
1071 int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
1072 struct session *sess);
1073 int enterthispgrp(struct proc *p, struct pgrp *pgrp);
1074 void faultin(struct proc *p);
1075 int fork1(struct thread *, struct fork_req *);
1076 void fork_rfppwait(struct thread *);
1077 void fork_exit(void (*)(void *, struct trapframe *), void *,
1078 struct trapframe *);
1079 void fork_return(struct thread *, struct trapframe *);
1080 int inferior(struct proc *p);
1081 void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry,
1082 int *resident_count, bool *super);
1083 void kern_yield(int);
1084 void kick_proc0(void);
1085 void killjobc(void);
1086 int leavepgrp(struct proc *p);
1087 int maybe_preempt(struct thread *td);
1088 void maybe_yield(void);
1089 void mi_switch(int flags);
1090 int p_candebug(struct thread *td, struct proc *p);
1091 int p_cansee(struct thread *td, struct proc *p);
1092 int p_cansched(struct thread *td, struct proc *p);
1093 int p_cansignal(struct thread *td, struct proc *p, int signum);
1094 int p_canwait(struct thread *td, struct proc *p);
1095 struct pargs *pargs_alloc(int len);
1096 void pargs_drop(struct pargs *pa);
1097 void pargs_hold(struct pargs *pa);
1098 int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
1099 int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
1100 int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
1101 void procinit(void);
1102 int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg);
1103 void proc_linkup0(struct proc *p, struct thread *td);
1104 void proc_linkup(struct proc *p, struct thread *td);
1105 struct proc *proc_realparent(struct proc *child);
1106 void proc_reap(struct thread *td, struct proc *p, int *status, int options);
1107 void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid);
1108 void proc_add_orphan(struct proc *child, struct proc *parent);
1109 void proc_set_traced(struct proc *p, bool stop);
1110 void proc_wkilled(struct proc *p);
1111 struct pstats *pstats_alloc(void);
1112 void pstats_fork(struct pstats *src, struct pstats *dst);
1113 void pstats_free(struct pstats *ps);
1114 void proc_clear_orphan(struct proc *p);
1115 void reaper_abandon_children(struct proc *p, bool exiting);
1116 int securelevel_ge(struct ucred *cr, int level);
1117 int securelevel_gt(struct ucred *cr, int level);
1118 void sess_hold(struct session *);
1119 void sess_release(struct session *);
1120 int setrunnable(struct thread *, int);
1121 void setsugid(struct proc *p);
1122 int should_yield(void);
1123 int sigonstack(size_t sp);
1124 void stopevent(struct proc *, u_int, u_int);
1125 struct thread *tdfind(lwpid_t, pid_t);
1126 void threadinit(void);
1127 void tidhash_add(struct thread *);
1128 void tidhash_remove(struct thread *);
1129 void cpu_idle(int);
1130 int cpu_idle_wakeup(int);
1131 extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */
1132 void cpu_switch(struct thread *, struct thread *, struct mtx *);
1133 void cpu_throw(struct thread *, struct thread *) __dead2;
1134 void unsleep(struct thread *);
1135 void userret(struct thread *, struct trapframe *);
1136
1137 void cpu_exit(struct thread *);
1138 void exit1(struct thread *, int, int) __dead2;
1139 void cpu_copy_thread(struct thread *td, struct thread *td0);
1140 bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map);
1141 int cpu_fetch_syscall_args(struct thread *td);
1142 void cpu_fork(struct thread *, struct proc *, struct thread *, int);
1143 void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *);
1144 int cpu_procctl(struct thread *td, int idtype, id_t id, int com,
1145 void *data);
1146 void cpu_set_syscall_retval(struct thread *, int);
1147 void cpu_set_upcall(struct thread *, void (*)(void *), void *,
1148 stack_t *);
1149 int cpu_set_user_tls(struct thread *, void *tls_base);
1150 void cpu_thread_alloc(struct thread *);
1151 void cpu_thread_clean(struct thread *);
1152 void cpu_thread_exit(struct thread *);
1153 void cpu_thread_free(struct thread *);
1154 void cpu_thread_swapin(struct thread *);
1155 void cpu_thread_swapout(struct thread *);
1156 struct thread *thread_alloc(int pages);
1157 int thread_alloc_stack(struct thread *, int pages);
1158 int thread_check_susp(struct thread *td, bool sleep);
1159 void thread_cow_get_proc(struct thread *newtd, struct proc *p);
1160 void thread_cow_get(struct thread *newtd, struct thread *td);
1161 void thread_cow_free(struct thread *td);
1162 void thread_cow_update(struct thread *td);
1163 int thread_create(struct thread *td, struct rtprio *rtp,
1164 int (*initialize_thread)(struct thread *, void *), void *thunk);
1165 void thread_exit(void) __dead2;
1166 void thread_free(struct thread *td);
1167 void thread_link(struct thread *td, struct proc *p);
1168 int thread_single(struct proc *p, int how);
1169 void thread_single_end(struct proc *p, int how);
1170 void thread_stash(struct thread *td);
1171 void thread_stopped(struct proc *p);
1172 void childproc_stopped(struct proc *child, int reason);
1173 void childproc_continued(struct proc *child);
1174 void childproc_exited(struct proc *child);
1175 int thread_suspend_check(int how);
1176 bool thread_suspend_check_needed(void);
1177 void thread_suspend_switch(struct thread *, struct proc *p);
1178 void thread_suspend_one(struct thread *td);
1179 void thread_unlink(struct thread *td);
1180 void thread_unsuspend(struct proc *p);
1181 void thread_wait(struct proc *p);
1182
1183 void stop_all_proc(void);
1184 void resume_all_proc(void);
1185
1186 static __inline int
curthread_pflags_set(int flags)1187 curthread_pflags_set(int flags)
1188 {
1189 struct thread *td;
1190 int save;
1191
1192 td = curthread;
1193 save = ~flags | (td->td_pflags & flags);
1194 td->td_pflags |= flags;
1195 return (save);
1196 }
1197
1198 static __inline void
curthread_pflags_restore(int save)1199 curthread_pflags_restore(int save)
1200 {
1201
1202 curthread->td_pflags &= save;
1203 }
1204
1205 static __inline int
curthread_pflags2_set(int flags)1206 curthread_pflags2_set(int flags)
1207 {
1208 struct thread *td;
1209 int save;
1210
1211 td = curthread;
1212 save = ~flags | (td->td_pflags2 & flags);
1213 td->td_pflags2 |= flags;
1214 return (save);
1215 }
1216
1217 static __inline void
curthread_pflags2_restore(int save)1218 curthread_pflags2_restore(int save)
1219 {
1220
1221 curthread->td_pflags2 &= save;
1222 }
1223
1224 static __inline bool
kstack_contains(struct thread * td,vm_offset_t va,size_t len)1225 kstack_contains(struct thread *td, vm_offset_t va, size_t len)
1226 {
1227 return (va >= td->td_kstack && va + len >= va &&
1228 va + len <= td->td_kstack + td->td_kstack_pages * PAGE_SIZE);
1229 }
1230
1231 static __inline __pure2 struct td_sched *
td_get_sched(struct thread * td)1232 td_get_sched(struct thread *td)
1233 {
1234
1235 return ((struct td_sched *)&td[1]);
1236 }
1237
1238 extern void (*softdep_ast_cleanup)(struct thread *);
1239 static __inline void
td_softdep_cleanup(struct thread * td)1240 td_softdep_cleanup(struct thread *td)
1241 {
1242
1243 if (td->td_su != NULL && softdep_ast_cleanup != NULL)
1244 softdep_ast_cleanup(td);
1245 }
1246
1247 #define PROC_ID_PID 0
1248 #define PROC_ID_GROUP 1
1249 #define PROC_ID_SESSION 2
1250 #define PROC_ID_REAP 3
1251
1252 void proc_id_set(int type, pid_t id);
1253 void proc_id_set_cond(int type, pid_t id);
1254 void proc_id_clear(int type, pid_t id);
1255
1256 EVENTHANDLER_LIST_DECLARE(process_ctor);
1257 EVENTHANDLER_LIST_DECLARE(process_dtor);
1258 EVENTHANDLER_LIST_DECLARE(process_init);
1259 EVENTHANDLER_LIST_DECLARE(process_fini);
1260 EVENTHANDLER_LIST_DECLARE(process_exit);
1261 EVENTHANDLER_LIST_DECLARE(process_fork);
1262 EVENTHANDLER_LIST_DECLARE(process_exec);
1263
1264 EVENTHANDLER_LIST_DECLARE(thread_ctor);
1265 EVENTHANDLER_LIST_DECLARE(thread_dtor);
1266 EVENTHANDLER_LIST_DECLARE(thread_init);
1267
1268 #endif /* _KERNEL */
1269
1270 #endif /* !_SYS_PROC_H_ */
1271