diff --git a/sys/kern/kern_prot.c b/sys/kern/kern_prot.c
index ed15cb566499..1e6073b554e4 100644
--- a/sys/kern/kern_prot.c
+++ b/sys/kern/kern_prot.c
@@ -1,2502 +1,2521 @@
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 1982, 1986, 1989, 1990, 1991, 1993
* The Regents of the University of California.
* (c) UNIX System Laboratories, Inc.
* Copyright (c) 2000-2001 Robert N. M. Watson.
* All rights reserved.
*
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)kern_prot.c 8.6 (Berkeley) 1/21/94
*/
/*
* System calls related to processes and protection
*/
#include <sys/cdefs.h>
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/acct.h>
#include <sys/kdb.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/loginclass.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/ptrace.h>
#include <sys/refcount.h>
#include <sys/sx.h>
#include <sys/priv.h>
#include <sys/proc.h>
#ifdef COMPAT_43
#include <sys/sysent.h>
#endif
#include <sys/sysproto.h>
#include <sys/jail.h>
#include <sys/racct.h>
#include <sys/rctl.h>
#include <sys/resourcevar.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/syscallsubr.h>
#include <sys/sysctl.h>
#ifdef REGRESSION
FEATURE(regression,
"Kernel support for interfaces necessary for regression testing (SECURITY RISK!)");
#endif
#include <security/audit/audit.h>
#include <security/mac/mac_framework.h>
static MALLOC_DEFINE(M_CRED, "cred", "credentials");
SYSCTL_NODE(_security, OID_AUTO, bsd, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"BSD security policy");
static void crfree_final(struct ucred *cr);
static void crsetgroups_locked(struct ucred *cr, int ngrp,
gid_t *groups);
#ifndef _SYS_SYSPROTO_H_
struct getpid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_getpid(struct thread *td, struct getpid_args *uap)
{
struct proc *p = td->td_proc;
td->td_retval[0] = p->p_pid;
#if defined(COMPAT_43)
if (SV_PROC_FLAG(p, SV_AOUT))
td->td_retval[1] = kern_getppid(td);
#endif
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct getppid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_getppid(struct thread *td, struct getppid_args *uap)
{
td->td_retval[0] = kern_getppid(td);
return (0);
}
int
kern_getppid(struct thread *td)
{
struct proc *p = td->td_proc;
return (p->p_oppid);
}
/*
* Get process group ID; note that POSIX getpgrp takes no parameter.
*/
#ifndef _SYS_SYSPROTO_H_
struct getpgrp_args {
int dummy;
};
#endif
int
sys_getpgrp(struct thread *td, struct getpgrp_args *uap)
{
struct proc *p = td->td_proc;
PROC_LOCK(p);
td->td_retval[0] = p->p_pgrp->pg_id;
PROC_UNLOCK(p);
return (0);
}
/* Get an arbitrary pid's process group id */
#ifndef _SYS_SYSPROTO_H_
struct getpgid_args {
pid_t pid;
};
#endif
int
sys_getpgid(struct thread *td, struct getpgid_args *uap)
{
struct proc *p;
int error;
if (uap->pid == 0) {
p = td->td_proc;
PROC_LOCK(p);
} else {
p = pfind(uap->pid);
if (p == NULL)
return (ESRCH);
error = p_cansee(td, p);
if (error) {
PROC_UNLOCK(p);
return (error);
}
}
td->td_retval[0] = p->p_pgrp->pg_id;
PROC_UNLOCK(p);
return (0);
}
/*
* Get an arbitrary pid's session id.
*/
#ifndef _SYS_SYSPROTO_H_
struct getsid_args {
pid_t pid;
};
#endif
int
sys_getsid(struct thread *td, struct getsid_args *uap)
{
return (kern_getsid(td, uap->pid));
}
int
kern_getsid(struct thread *td, pid_t pid)
{
struct proc *p;
int error;
if (pid == 0) {
p = td->td_proc;
PROC_LOCK(p);
} else {
p = pfind(pid);
if (p == NULL)
return (ESRCH);
error = p_cansee(td, p);
if (error) {
PROC_UNLOCK(p);
return (error);
}
}
td->td_retval[0] = p->p_session->s_sid;
PROC_UNLOCK(p);
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct getuid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_getuid(struct thread *td, struct getuid_args *uap)
{
td->td_retval[0] = td->td_ucred->cr_ruid;
#if defined(COMPAT_43)
td->td_retval[1] = td->td_ucred->cr_uid;
#endif
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct geteuid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_geteuid(struct thread *td, struct geteuid_args *uap)
{
td->td_retval[0] = td->td_ucred->cr_uid;
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct getgid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_getgid(struct thread *td, struct getgid_args *uap)
{
td->td_retval[0] = td->td_ucred->cr_rgid;
#if defined(COMPAT_43)
td->td_retval[1] = td->td_ucred->cr_groups[0];
#endif
return (0);
}
/*
* Get effective group ID. The "egid" is groups[0], and could be obtained
* via getgroups. This syscall exists because it is somewhat painful to do
* correctly in a library function.
*/
#ifndef _SYS_SYSPROTO_H_
struct getegid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_getegid(struct thread *td, struct getegid_args *uap)
{
td->td_retval[0] = td->td_ucred->cr_groups[0];
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct getgroups_args {
int gidsetsize;
gid_t *gidset;
};
#endif
int
sys_getgroups(struct thread *td, struct getgroups_args *uap)
{
struct ucred *cred;
int ngrp, error;
cred = td->td_ucred;
ngrp = cred->cr_ngroups;
if (uap->gidsetsize == 0) {
error = 0;
goto out;
}
if (uap->gidsetsize < ngrp)
return (EINVAL);
error = copyout(cred->cr_groups, uap->gidset, ngrp * sizeof(gid_t));
out:
td->td_retval[0] = ngrp;
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setsid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_setsid(struct thread *td, struct setsid_args *uap)
{
struct pgrp *pgrp;
int error;
struct proc *p = td->td_proc;
struct pgrp *newpgrp;
struct session *newsess;
pgrp = NULL;
newpgrp = uma_zalloc(pgrp_zone, M_WAITOK);
newsess = malloc(sizeof(struct session), M_SESSION, M_WAITOK | M_ZERO);
again:
error = 0;
sx_xlock(&proctree_lock);
if (p->p_pgid == p->p_pid || (pgrp = pgfind(p->p_pid)) != NULL) {
if (pgrp != NULL)
PGRP_UNLOCK(pgrp);
error = EPERM;
} else {
error = enterpgrp(p, p->p_pid, newpgrp, newsess);
if (error == ERESTART)
goto again;
MPASS(error == 0);
td->td_retval[0] = p->p_pid;
newpgrp = NULL;
newsess = NULL;
}
sx_xunlock(&proctree_lock);
uma_zfree(pgrp_zone, newpgrp);
free(newsess, M_SESSION);
return (error);
}
/*
* set process group (setpgid/old setpgrp)
*
* caller does setpgid(targpid, targpgid)
*
* pid must be caller or child of caller (ESRCH)
* if a child
* pid must be in same session (EPERM)
* pid can't have done an exec (EACCES)
* if pgid != pid
* there must exist some pid in same session having pgid (EPERM)
* pid must not be session leader (EPERM)
*/
#ifndef _SYS_SYSPROTO_H_
struct setpgid_args {
int pid; /* target process id */
int pgid; /* target pgrp id */
};
#endif
/* ARGSUSED */
int
sys_setpgid(struct thread *td, struct setpgid_args *uap)
{
struct proc *curp = td->td_proc;
struct proc *targp; /* target process */
struct pgrp *pgrp; /* target pgrp */
int error;
struct pgrp *newpgrp;
if (uap->pgid < 0)
return (EINVAL);
newpgrp = uma_zalloc(pgrp_zone, M_WAITOK);
again:
error = 0;
sx_xlock(&proctree_lock);
if (uap->pid != 0 && uap->pid != curp->p_pid) {
if ((targp = pfind(uap->pid)) == NULL) {
error = ESRCH;
goto done;
}
if (!inferior(targp)) {
PROC_UNLOCK(targp);
error = ESRCH;
goto done;
}
if ((error = p_cansee(td, targp))) {
PROC_UNLOCK(targp);
goto done;
}
if (targp->p_pgrp == NULL ||
targp->p_session != curp->p_session) {
PROC_UNLOCK(targp);
error = EPERM;
goto done;
}
if (targp->p_flag & P_EXEC) {
PROC_UNLOCK(targp);
error = EACCES;
goto done;
}
PROC_UNLOCK(targp);
} else
targp = curp;
if (SESS_LEADER(targp)) {
error = EPERM;
goto done;
}
if (uap->pgid == 0)
uap->pgid = targp->p_pid;
if ((pgrp = pgfind(uap->pgid)) == NULL) {
if (uap->pgid == targp->p_pid) {
error = enterpgrp(targp, uap->pgid, newpgrp,
NULL);
if (error == 0)
newpgrp = NULL;
} else
error = EPERM;
} else {
if (pgrp == targp->p_pgrp) {
PGRP_UNLOCK(pgrp);
goto done;
}
if (pgrp->pg_id != targp->p_pid &&
pgrp->pg_session != curp->p_session) {
PGRP_UNLOCK(pgrp);
error = EPERM;
goto done;
}
PGRP_UNLOCK(pgrp);
error = enterthispgrp(targp, pgrp);
}
done:
KASSERT(error == 0 || newpgrp != NULL,
("setpgid failed and newpgrp is NULL"));
if (error == ERESTART)
goto again;
sx_xunlock(&proctree_lock);
uma_zfree(pgrp_zone, newpgrp);
return (error);
}
/*
* Use the clause in B.4.2.2 that allows setuid/setgid to be 4.2/4.3BSD
* compatible. It says that setting the uid/gid to euid/egid is a special
* case of "appropriate privilege". Once the rules are expanded out, this
* basically means that setuid(nnn) sets all three id's, in all permitted
* cases unless _POSIX_SAVED_IDS is enabled. In that case, setuid(getuid())
* does not set the saved id - this is dangerous for traditional BSD
* programs. For this reason, we *really* do not want to set
* _POSIX_SAVED_IDS and do not want to clear POSIX_APPENDIX_B_4_2_2.
*/
#define POSIX_APPENDIX_B_4_2_2
#ifndef _SYS_SYSPROTO_H_
struct setuid_args {
uid_t uid;
};
#endif
/* ARGSUSED */
int
sys_setuid(struct thread *td, struct setuid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
uid_t uid;
struct uidinfo *uip;
int error;
uid = uap->uid;
AUDIT_ARG_UID(uid);
newcred = crget();
uip = uifind(uid);
PROC_LOCK(p);
/*
* Copy credentials so other references do not see our changes.
*/
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setuid(oldcred, uid);
if (error)
goto fail;
#endif
/*
* See if we have "permission" by POSIX 1003.1 rules.
*
* Note that setuid(geteuid()) is a special case of
* "appropriate privileges" in appendix B.4.2.2. We need
* to use this clause to be compatible with traditional BSD
* semantics. Basically, it means that "setuid(xx)" sets all
* three id's (assuming you have privs).
*
* Notes on the logic. We do things in three steps.
* 1: We determine if the euid is going to change, and do EPERM
* right away. We unconditionally change the euid later if this
* test is satisfied, simplifying that part of the logic.
* 2: We determine if the real and/or saved uids are going to
* change. Determined by compile options.
* 3: Change euid last. (after tests in #2 for "appropriate privs")
*/
if (uid != oldcred->cr_ruid && /* allow setuid(getuid()) */
#ifdef _POSIX_SAVED_IDS
uid != oldcred->cr_svuid && /* allow setuid(saved gid) */
#endif
#ifdef POSIX_APPENDIX_B_4_2_2 /* Use BSD-compat clause from B.4.2.2 */
uid != oldcred->cr_uid && /* allow setuid(geteuid()) */
#endif
(error = priv_check_cred(oldcred, PRIV_CRED_SETUID)) != 0)
goto fail;
#ifdef _POSIX_SAVED_IDS
/*
* Do we have "appropriate privileges" (are we root or uid == euid)
* If so, we are changing the real uid and/or saved uid.
*/
if (
#ifdef POSIX_APPENDIX_B_4_2_2 /* Use the clause from B.4.2.2 */
uid == oldcred->cr_uid ||
#endif
/* We are using privs. */
priv_check_cred(oldcred, PRIV_CRED_SETUID) == 0)
#endif
{
/*
* Set the real uid and transfer proc count to new user.
*/
if (uid != oldcred->cr_ruid) {
change_ruid(newcred, uip);
setsugid(p);
}
/*
* Set saved uid
*
* XXX always set saved uid even if not _POSIX_SAVED_IDS, as
* the security of seteuid() depends on it. B.4.2.2 says it
* is important that we should do this.
*/
if (uid != oldcred->cr_svuid) {
change_svuid(newcred, uid);
setsugid(p);
}
}
/*
* In all permitted cases, we are changing the euid.
*/
if (uid != oldcred->cr_uid) {
change_euid(newcred, uip);
setsugid(p);
}
proc_set_cred(p, newcred);
#ifdef RACCT
racct_proc_ucred_changed(p, oldcred, newcred);
crhold(newcred);
#endif
PROC_UNLOCK(p);
#ifdef RCTL
rctl_proc_ucred_changed(p, newcred);
crfree(newcred);
#endif
uifree(uip);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
uifree(uip);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct seteuid_args {
uid_t euid;
};
#endif
/* ARGSUSED */
int
sys_seteuid(struct thread *td, struct seteuid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
uid_t euid;
struct uidinfo *euip;
int error;
euid = uap->euid;
AUDIT_ARG_EUID(euid);
newcred = crget();
euip = uifind(euid);
PROC_LOCK(p);
/*
* Copy credentials so other references do not see our changes.
*/
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_seteuid(oldcred, euid);
if (error)
goto fail;
#endif
if (euid != oldcred->cr_ruid && /* allow seteuid(getuid()) */
euid != oldcred->cr_svuid && /* allow seteuid(saved uid) */
(error = priv_check_cred(oldcred, PRIV_CRED_SETEUID)) != 0)
goto fail;
/*
* Everything's okay, do it.
*/
if (oldcred->cr_uid != euid) {
change_euid(newcred, euip);
setsugid(p);
}
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
uifree(euip);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
uifree(euip);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setgid_args {
gid_t gid;
};
#endif
/* ARGSUSED */
int
sys_setgid(struct thread *td, struct setgid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
gid_t gid;
int error;
gid = uap->gid;
AUDIT_ARG_GID(gid);
newcred = crget();
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setgid(oldcred, gid);
if (error)
goto fail;
#endif
/*
* See if we have "permission" by POSIX 1003.1 rules.
*
* Note that setgid(getegid()) is a special case of
* "appropriate privileges" in appendix B.4.2.2. We need
* to use this clause to be compatible with traditional BSD
* semantics. Basically, it means that "setgid(xx)" sets all
* three id's (assuming you have privs).
*
* For notes on the logic here, see setuid() above.
*/
if (gid != oldcred->cr_rgid && /* allow setgid(getgid()) */
#ifdef _POSIX_SAVED_IDS
gid != oldcred->cr_svgid && /* allow setgid(saved gid) */
#endif
#ifdef POSIX_APPENDIX_B_4_2_2 /* Use BSD-compat clause from B.4.2.2 */
gid != oldcred->cr_groups[0] && /* allow setgid(getegid()) */
#endif
(error = priv_check_cred(oldcred, PRIV_CRED_SETGID)) != 0)
goto fail;
#ifdef _POSIX_SAVED_IDS
/*
* Do we have "appropriate privileges" (are we root or gid == egid)
* If so, we are changing the real uid and saved gid.
*/
if (
#ifdef POSIX_APPENDIX_B_4_2_2 /* use the clause from B.4.2.2 */
gid == oldcred->cr_groups[0] ||
#endif
/* We are using privs. */
priv_check_cred(oldcred, PRIV_CRED_SETGID) == 0)
#endif
{
/*
* Set real gid
*/
if (oldcred->cr_rgid != gid) {
change_rgid(newcred, gid);
setsugid(p);
}
/*
* Set saved gid
*
* XXX always set saved gid even if not _POSIX_SAVED_IDS, as
* the security of setegid() depends on it. B.4.2.2 says it
* is important that we should do this.
*/
if (oldcred->cr_svgid != gid) {
change_svgid(newcred, gid);
setsugid(p);
}
}
/*
* In all cases permitted cases, we are changing the egid.
* Copy credentials so other references do not see our changes.
*/
if (oldcred->cr_groups[0] != gid) {
change_egid(newcred, gid);
setsugid(p);
}
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setegid_args {
gid_t egid;
};
#endif
/* ARGSUSED */
int
sys_setegid(struct thread *td, struct setegid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
gid_t egid;
int error;
egid = uap->egid;
AUDIT_ARG_EGID(egid);
newcred = crget();
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setegid(oldcred, egid);
if (error)
goto fail;
#endif
if (egid != oldcred->cr_rgid && /* allow setegid(getgid()) */
egid != oldcred->cr_svgid && /* allow setegid(saved gid) */
(error = priv_check_cred(oldcred, PRIV_CRED_SETEGID)) != 0)
goto fail;
if (oldcred->cr_groups[0] != egid) {
change_egid(newcred, egid);
setsugid(p);
}
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setgroups_args {
int gidsetsize;
gid_t *gidset;
};
#endif
/* ARGSUSED */
int
sys_setgroups(struct thread *td, struct setgroups_args *uap)
{
gid_t smallgroups[XU_NGROUPS];
gid_t *groups;
int gidsetsize, error;
gidsetsize = uap->gidsetsize;
if (gidsetsize > ngroups_max + 1 || gidsetsize < 0)
return (EINVAL);
if (gidsetsize > XU_NGROUPS)
groups = malloc(gidsetsize * sizeof(gid_t), M_TEMP, M_WAITOK);
else
groups = smallgroups;
error = copyin(uap->gidset, groups, gidsetsize * sizeof(gid_t));
if (error == 0)
error = kern_setgroups(td, gidsetsize, groups);
if (gidsetsize > XU_NGROUPS)
free(groups, M_TEMP);
return (error);
}
int
kern_setgroups(struct thread *td, u_int ngrp, gid_t *groups)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
int error;
MPASS(ngrp <= ngroups_max + 1);
AUDIT_ARG_GROUPSET(groups, ngrp);
newcred = crget();
crextend(newcred, ngrp);
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setgroups(oldcred, ngrp, groups);
if (error)
goto fail;
#endif
error = priv_check_cred(oldcred, PRIV_CRED_SETGROUPS);
if (error)
goto fail;
if (ngrp == 0) {
/*
* setgroups(0, NULL) is a legitimate way of clearing the
* groups vector on non-BSD systems (which generally do not
* have the egid in the groups[0]). We risk security holes
* when running non-BSD software if we do not do the same.
*/
newcred->cr_ngroups = 1;
} else {
crsetgroups_locked(newcred, ngrp, groups);
}
setsugid(p);
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setreuid_args {
uid_t ruid;
uid_t euid;
};
#endif
/* ARGSUSED */
int
sys_setreuid(struct thread *td, struct setreuid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
uid_t euid, ruid;
struct uidinfo *euip, *ruip;
int error;
euid = uap->euid;
ruid = uap->ruid;
AUDIT_ARG_EUID(euid);
AUDIT_ARG_RUID(ruid);
newcred = crget();
euip = uifind(euid);
ruip = uifind(ruid);
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setreuid(oldcred, ruid, euid);
if (error)
goto fail;
#endif
if (((ruid != (uid_t)-1 && ruid != oldcred->cr_ruid &&
ruid != oldcred->cr_svuid) ||
(euid != (uid_t)-1 && euid != oldcred->cr_uid &&
euid != oldcred->cr_ruid && euid != oldcred->cr_svuid)) &&
(error = priv_check_cred(oldcred, PRIV_CRED_SETREUID)) != 0)
goto fail;
if (euid != (uid_t)-1 && oldcred->cr_uid != euid) {
change_euid(newcred, euip);
setsugid(p);
}
if (ruid != (uid_t)-1 && oldcred->cr_ruid != ruid) {
change_ruid(newcred, ruip);
setsugid(p);
}
if ((ruid != (uid_t)-1 || newcred->cr_uid != newcred->cr_ruid) &&
newcred->cr_svuid != newcred->cr_uid) {
change_svuid(newcred, newcred->cr_uid);
setsugid(p);
}
proc_set_cred(p, newcred);
#ifdef RACCT
racct_proc_ucred_changed(p, oldcred, newcred);
crhold(newcred);
#endif
PROC_UNLOCK(p);
#ifdef RCTL
rctl_proc_ucred_changed(p, newcred);
crfree(newcred);
#endif
uifree(ruip);
uifree(euip);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
uifree(ruip);
uifree(euip);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct setregid_args {
gid_t rgid;
gid_t egid;
};
#endif
/* ARGSUSED */
int
sys_setregid(struct thread *td, struct setregid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
gid_t egid, rgid;
int error;
egid = uap->egid;
rgid = uap->rgid;
AUDIT_ARG_EGID(egid);
AUDIT_ARG_RGID(rgid);
newcred = crget();
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setregid(oldcred, rgid, egid);
if (error)
goto fail;
#endif
if (((rgid != (gid_t)-1 && rgid != oldcred->cr_rgid &&
rgid != oldcred->cr_svgid) ||
(egid != (gid_t)-1 && egid != oldcred->cr_groups[0] &&
egid != oldcred->cr_rgid && egid != oldcred->cr_svgid)) &&
(error = priv_check_cred(oldcred, PRIV_CRED_SETREGID)) != 0)
goto fail;
if (egid != (gid_t)-1 && oldcred->cr_groups[0] != egid) {
change_egid(newcred, egid);
setsugid(p);
}
if (rgid != (gid_t)-1 && oldcred->cr_rgid != rgid) {
change_rgid(newcred, rgid);
setsugid(p);
}
if ((rgid != (gid_t)-1 || newcred->cr_groups[0] != newcred->cr_rgid) &&
newcred->cr_svgid != newcred->cr_groups[0]) {
change_svgid(newcred, newcred->cr_groups[0]);
setsugid(p);
}
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
crfree(newcred);
return (error);
}
/*
* setresuid(ruid, euid, suid) is like setreuid except control over the saved
* uid is explicit.
*/
#ifndef _SYS_SYSPROTO_H_
struct setresuid_args {
uid_t ruid;
uid_t euid;
uid_t suid;
};
#endif
/* ARGSUSED */
int
sys_setresuid(struct thread *td, struct setresuid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
uid_t euid, ruid, suid;
struct uidinfo *euip, *ruip;
int error;
euid = uap->euid;
ruid = uap->ruid;
suid = uap->suid;
AUDIT_ARG_EUID(euid);
AUDIT_ARG_RUID(ruid);
AUDIT_ARG_SUID(suid);
newcred = crget();
euip = uifind(euid);
ruip = uifind(ruid);
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setresuid(oldcred, ruid, euid, suid);
if (error)
goto fail;
#endif
if (((ruid != (uid_t)-1 && ruid != oldcred->cr_ruid &&
ruid != oldcred->cr_svuid &&
ruid != oldcred->cr_uid) ||
(euid != (uid_t)-1 && euid != oldcred->cr_ruid &&
euid != oldcred->cr_svuid &&
euid != oldcred->cr_uid) ||
(suid != (uid_t)-1 && suid != oldcred->cr_ruid &&
suid != oldcred->cr_svuid &&
suid != oldcred->cr_uid)) &&
(error = priv_check_cred(oldcred, PRIV_CRED_SETRESUID)) != 0)
goto fail;
if (euid != (uid_t)-1 && oldcred->cr_uid != euid) {
change_euid(newcred, euip);
setsugid(p);
}
if (ruid != (uid_t)-1 && oldcred->cr_ruid != ruid) {
change_ruid(newcred, ruip);
setsugid(p);
}
if (suid != (uid_t)-1 && oldcred->cr_svuid != suid) {
change_svuid(newcred, suid);
setsugid(p);
}
proc_set_cred(p, newcred);
#ifdef RACCT
racct_proc_ucred_changed(p, oldcred, newcred);
crhold(newcred);
#endif
PROC_UNLOCK(p);
#ifdef RCTL
rctl_proc_ucred_changed(p, newcred);
crfree(newcred);
#endif
uifree(ruip);
uifree(euip);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
uifree(ruip);
uifree(euip);
crfree(newcred);
return (error);
}
/*
* setresgid(rgid, egid, sgid) is like setregid except control over the saved
* gid is explicit.
*/
#ifndef _SYS_SYSPROTO_H_
struct setresgid_args {
gid_t rgid;
gid_t egid;
gid_t sgid;
};
#endif
/* ARGSUSED */
int
sys_setresgid(struct thread *td, struct setresgid_args *uap)
{
struct proc *p = td->td_proc;
struct ucred *newcred, *oldcred;
gid_t egid, rgid, sgid;
int error;
egid = uap->egid;
rgid = uap->rgid;
sgid = uap->sgid;
AUDIT_ARG_EGID(egid);
AUDIT_ARG_RGID(rgid);
AUDIT_ARG_SGID(sgid);
newcred = crget();
PROC_LOCK(p);
oldcred = crcopysafe(p, newcred);
#ifdef MAC
error = mac_cred_check_setresgid(oldcred, rgid, egid, sgid);
if (error)
goto fail;
#endif
if (((rgid != (gid_t)-1 && rgid != oldcred->cr_rgid &&
rgid != oldcred->cr_svgid &&
rgid != oldcred->cr_groups[0]) ||
(egid != (gid_t)-1 && egid != oldcred->cr_rgid &&
egid != oldcred->cr_svgid &&
egid != oldcred->cr_groups[0]) ||
(sgid != (gid_t)-1 && sgid != oldcred->cr_rgid &&
sgid != oldcred->cr_svgid &&
sgid != oldcred->cr_groups[0])) &&
(error = priv_check_cred(oldcred, PRIV_CRED_SETRESGID)) != 0)
goto fail;
if (egid != (gid_t)-1 && oldcred->cr_groups[0] != egid) {
change_egid(newcred, egid);
setsugid(p);
}
if (rgid != (gid_t)-1 && oldcred->cr_rgid != rgid) {
change_rgid(newcred, rgid);
setsugid(p);
}
if (sgid != (gid_t)-1 && oldcred->cr_svgid != sgid) {
change_svgid(newcred, sgid);
setsugid(p);
}
proc_set_cred(p, newcred);
PROC_UNLOCK(p);
crfree(oldcred);
return (0);
fail:
PROC_UNLOCK(p);
crfree(newcred);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct getresuid_args {
uid_t *ruid;
uid_t *euid;
uid_t *suid;
};
#endif
/* ARGSUSED */
int
sys_getresuid(struct thread *td, struct getresuid_args *uap)
{
struct ucred *cred;
int error1 = 0, error2 = 0, error3 = 0;
cred = td->td_ucred;
if (uap->ruid)
error1 = copyout(&cred->cr_ruid,
uap->ruid, sizeof(cred->cr_ruid));
if (uap->euid)
error2 = copyout(&cred->cr_uid,
uap->euid, sizeof(cred->cr_uid));
if (uap->suid)
error3 = copyout(&cred->cr_svuid,
uap->suid, sizeof(cred->cr_svuid));
return (error1 ? error1 : error2 ? error2 : error3);
}
#ifndef _SYS_SYSPROTO_H_
struct getresgid_args {
gid_t *rgid;
gid_t *egid;
gid_t *sgid;
};
#endif
/* ARGSUSED */
int
sys_getresgid(struct thread *td, struct getresgid_args *uap)
{
struct ucred *cred;
int error1 = 0, error2 = 0, error3 = 0;
cred = td->td_ucred;
if (uap->rgid)
error1 = copyout(&cred->cr_rgid,
uap->rgid, sizeof(cred->cr_rgid));
if (uap->egid)
error2 = copyout(&cred->cr_groups[0],
uap->egid, sizeof(cred->cr_groups[0]));
if (uap->sgid)
error3 = copyout(&cred->cr_svgid,
uap->sgid, sizeof(cred->cr_svgid));
return (error1 ? error1 : error2 ? error2 : error3);
}
#ifndef _SYS_SYSPROTO_H_
struct issetugid_args {
int dummy;
};
#endif
/* ARGSUSED */
int
sys_issetugid(struct thread *td, struct issetugid_args *uap)
{
struct proc *p = td->td_proc;
/*
* Note: OpenBSD sets a P_SUGIDEXEC flag set at execve() time,
* we use P_SUGID because we consider changing the owners as
* "tainting" as well.
* This is significant for procs that start as root and "become"
* a user without an exec - programs cannot know *everything*
* that libc *might* have put in their data segment.
*/
td->td_retval[0] = (p->p_flag & P_SUGID) ? 1 : 0;
return (0);
}
int
sys___setugid(struct thread *td, struct __setugid_args *uap)
{
#ifdef REGRESSION
struct proc *p;
p = td->td_proc;
switch (uap->flag) {
case 0:
PROC_LOCK(p);
p->p_flag &= ~P_SUGID;
PROC_UNLOCK(p);
return (0);
case 1:
PROC_LOCK(p);
p->p_flag |= P_SUGID;
PROC_UNLOCK(p);
return (0);
default:
return (EINVAL);
}
#else /* !REGRESSION */
return (ENOSYS);
#endif /* REGRESSION */
}
/*
* Check if gid is a member of the group set.
*/
int
groupmember(gid_t gid, struct ucred *cred)
{
int l;
int h;
int m;
if (cred->cr_groups[0] == gid)
return(1);
/*
* If gid was not our primary group, perform a binary search
* of the supplemental groups. This is possible because we
* sort the groups in crsetgroups().
*/
l = 1;
h = cred->cr_ngroups;
while (l < h) {
m = l + ((h - l) / 2);
if (cred->cr_groups[m] < gid)
l = m + 1;
else
h = m;
}
if ((l < cred->cr_ngroups) && (cred->cr_groups[l] == gid))
return (1);
return (0);
}
/*
* Test the active securelevel against a given level. securelevel_gt()
* implements (securelevel > level). securelevel_ge() implements
* (securelevel >= level). Note that the logic is inverted -- these
* functions return EPERM on "success" and 0 on "failure".
*
* Due to care taken when setting the securelevel, we know that no jail will
* be less secure that its parent (or the physical system), so it is sufficient
* to test the current jail only.
*
* XXXRW: Possibly since this has to do with privilege, it should move to
* kern_priv.c.
*/
int
securelevel_gt(struct ucred *cr, int level)
{
return (cr->cr_prison->pr_securelevel > level ? EPERM : 0);
}
int
securelevel_ge(struct ucred *cr, int level)
{
return (cr->cr_prison->pr_securelevel >= level ? EPERM : 0);
}
/*
* 'see_other_uids' determines whether or not visibility of processes
* and sockets with credentials holding different real uids is possible
* using a variety of system MIBs.
* XXX: data declarations should be together near the beginning of the file.
*/
static int see_other_uids = 1;
SYSCTL_INT(_security_bsd, OID_AUTO, see_other_uids, CTLFLAG_RW,
&see_other_uids, 0,
"Unprivileged processes may see subjects/objects with different real uid");
/*-
* Determine if u1 "can see" the subject specified by u2, according to the
* 'see_other_uids' policy.
* Returns: 0 for permitted, ESRCH otherwise
* Locks: none
* References: *u1 and *u2 must not change during the call
* u1 may equal u2, in which case only one reference is required
*/
int
cr_canseeotheruids(struct ucred *u1, struct ucred *u2)
{
if (!see_other_uids && u1->cr_ruid != u2->cr_ruid) {
if (priv_check_cred(u1, PRIV_SEEOTHERUIDS) != 0)
return (ESRCH);
}
return (0);
}
/*
* 'see_other_gids' determines whether or not visibility of processes
* and sockets with credentials holding different real gids is possible
* using a variety of system MIBs.
* XXX: data declarations should be together near the beginning of the file.
*/
static int see_other_gids = 1;
SYSCTL_INT(_security_bsd, OID_AUTO, see_other_gids, CTLFLAG_RW,
&see_other_gids, 0,
"Unprivileged processes may see subjects/objects with different real gid");
/*
* Determine if u1 can "see" the subject specified by u2, according to the
* 'see_other_gids' policy.
* Returns: 0 for permitted, ESRCH otherwise
* Locks: none
* References: *u1 and *u2 must not change during the call
* u1 may equal u2, in which case only one reference is required
*/
int
cr_canseeothergids(struct ucred *u1, struct ucred *u2)
{
int i, match;
if (!see_other_gids) {
match = 0;
for (i = 0; i < u1->cr_ngroups; i++) {
if (groupmember(u1->cr_groups[i], u2))
match = 1;
if (match)
break;
}
if (!match) {
if (priv_check_cred(u1, PRIV_SEEOTHERGIDS) != 0)
return (ESRCH);
}
}
return (0);
}
/*
* 'see_jail_proc' determines whether or not visibility of processes and
* sockets with credentials holding different jail ids is possible using a
* variety of system MIBs.
*
* XXX: data declarations should be together near the beginning of the file.
*/
static int see_jail_proc = 1;
SYSCTL_INT(_security_bsd, OID_AUTO, see_jail_proc, CTLFLAG_RW,
&see_jail_proc, 0,
"Unprivileged processes may see subjects/objects with different jail ids");
/*-
* Determine if u1 "can see" the subject specified by u2, according to the
* 'see_jail_proc' policy.
* Returns: 0 for permitted, ESRCH otherwise
* Locks: none
* References: *u1 and *u2 must not change during the call
* u1 may equal u2, in which case only one reference is required
*/
int
cr_canseejailproc(struct ucred *u1, struct ucred *u2)
{
if (see_jail_proc || /* Policy deactivated. */
u1->cr_prison == u2->cr_prison || /* Same jail. */
priv_check_cred(u1, PRIV_SEEJAILPROC) == 0) /* Privileged. */
return (0);
return (ESRCH);
}
+/*
+ * Helper for cr_cansee*() functions to abide by system-wide security.bsd.see_*
+ * policies. Determines if u1 "can see" u2 according to these policies.
+ * Returns: 0 for permitted, ESRCH otherwise
+ */
+int
+cr_bsd_visible(struct ucred *u1, struct ucred *u2)
+{
+ int error;
+
+ if ((error = cr_canseeotheruids(u1, u2)))
+ return (error);
+ if ((error = cr_canseeothergids(u1, u2)))
+ return (error);
+ if ((error = cr_canseejailproc(u1, u2)))
+ return (error);
+ return (0);
+}
+
/*-
* Determine if u1 "can see" the subject specified by u2.
* Returns: 0 for permitted, an errno value otherwise
* Locks: none
* References: *u1 and *u2 must not change during the call
* u1 may equal u2, in which case only one reference is required
*/
int
cr_cansee(struct ucred *u1, struct ucred *u2)
{
int error;
if ((error = prison_check(u1, u2)))
return (error);
#ifdef MAC
if ((error = mac_cred_check_visible(u1, u2)))
return (error);
#endif
if ((error = cr_canseeotheruids(u1, u2)))
return (error);
if ((error = cr_canseeothergids(u1, u2)))
return (error);
if ((error = cr_canseejailproc(u1, u2)))
return (error);
return (0);
}
/*-
* Determine if td "can see" the subject specified by p.
* Returns: 0 for permitted, an errno value otherwise
* Locks: Sufficient locks to protect p->p_ucred must be held. td really
* should be curthread.
* References: td and p must be valid for the lifetime of the call
*/
int
p_cansee(struct thread *td, struct proc *p)
{
/* Wrap cr_cansee() for all functionality. */
KASSERT(td == curthread, ("%s: td not curthread", __func__));
PROC_LOCK_ASSERT(p, MA_OWNED);
if (td->td_proc == p)
return (0);
return (cr_cansee(td->td_ucred, p->p_ucred));
}
/*
* 'conservative_signals' prevents the delivery of a broad class of
* signals by unprivileged processes to processes that have changed their
* credentials since the last invocation of execve(). This can prevent
* the leakage of cached information or retained privileges as a result
* of a common class of signal-related vulnerabilities. However, this
* may interfere with some applications that expect to be able to
* deliver these signals to peer processes after having given up
* privilege.
*/
static int conservative_signals = 1;
SYSCTL_INT(_security_bsd, OID_AUTO, conservative_signals, CTLFLAG_RW,
&conservative_signals, 0, "Unprivileged processes prevented from "
"sending certain signals to processes whose credentials have changed");
/*-
* Determine whether cred may deliver the specified signal to proc.
* Returns: 0 for permitted, an errno value otherwise.
* Locks: A lock must be held for proc.
* References: cred and proc must be valid for the lifetime of the call.
*/
int
cr_cansignal(struct ucred *cred, struct proc *proc, int signum)
{
int error;
PROC_LOCK_ASSERT(proc, MA_OWNED);
/*
* Jail semantics limit the scope of signalling to proc in the
* same jail as cred, if cred is in jail.
*/
error = prison_check(cred, proc->p_ucred);
if (error)
return (error);
#ifdef MAC
if ((error = mac_proc_check_signal(cred, proc, signum)))
return (error);
#endif
if ((error = cr_canseeotheruids(cred, proc->p_ucred)))
return (error);
if ((error = cr_canseeothergids(cred, proc->p_ucred)))
return (error);
/*
* UNIX signal semantics depend on the status of the P_SUGID
* bit on the target process. If the bit is set, then additional
* restrictions are placed on the set of available signals.
*/
if (conservative_signals && (proc->p_flag & P_SUGID)) {
switch (signum) {
case 0:
case SIGKILL:
case SIGINT:
case SIGTERM:
case SIGALRM:
case SIGSTOP:
case SIGTTIN:
case SIGTTOU:
case SIGTSTP:
case SIGHUP:
case SIGUSR1:
case SIGUSR2:
/*
* Generally, permit job and terminal control
* signals.
*/
break;
default:
/* Not permitted without privilege. */
error = priv_check_cred(cred, PRIV_SIGNAL_SUGID);
if (error)
return (error);
}
}
/*
* Generally, the target credential's ruid or svuid must match the
* subject credential's ruid or euid.
*/
if (cred->cr_ruid != proc->p_ucred->cr_ruid &&
cred->cr_ruid != proc->p_ucred->cr_svuid &&
cred->cr_uid != proc->p_ucred->cr_ruid &&
cred->cr_uid != proc->p_ucred->cr_svuid) {
error = priv_check_cred(cred, PRIV_SIGNAL_DIFFCRED);
if (error)
return (error);
}
return (0);
}
/*-
* Determine whether td may deliver the specified signal to p.
* Returns: 0 for permitted, an errno value otherwise
* Locks: Sufficient locks to protect various components of td and p
* must be held. td must be curthread, and a lock must be
* held for p.
* References: td and p must be valid for the lifetime of the call
*/
int
p_cansignal(struct thread *td, struct proc *p, int signum)
{
KASSERT(td == curthread, ("%s: td not curthread", __func__));
PROC_LOCK_ASSERT(p, MA_OWNED);
if (td->td_proc == p)
return (0);
/*
* UNIX signalling semantics require that processes in the same
* session always be able to deliver SIGCONT to one another,
* overriding the remaining protections.
*/
/* XXX: This will require an additional lock of some sort. */
if (signum == SIGCONT && td->td_proc->p_session == p->p_session)
return (0);
/*
* Some compat layers use SIGTHR and higher signals for
* communication between different kernel threads of the same
* process, so that they expect that it's always possible to
* deliver them, even for suid applications where cr_cansignal() can
* deny such ability for security consideration. It should be
* pretty safe to do since the only way to create two processes
* with the same p_leader is via rfork(2).
*/
if (td->td_proc->p_leader != NULL && signum >= SIGTHR &&
signum < SIGTHR + 4 && td->td_proc->p_leader == p->p_leader)
return (0);
return (cr_cansignal(td->td_ucred, p, signum));
}
/*-
* Determine whether td may reschedule p.
* Returns: 0 for permitted, an errno value otherwise
* Locks: Sufficient locks to protect various components of td and p
* must be held. td must be curthread, and a lock must
* be held for p.
* References: td and p must be valid for the lifetime of the call
*/
int
p_cansched(struct thread *td, struct proc *p)
{
int error;
KASSERT(td == curthread, ("%s: td not curthread", __func__));
PROC_LOCK_ASSERT(p, MA_OWNED);
if (td->td_proc == p)
return (0);
if ((error = prison_check(td->td_ucred, p->p_ucred)))
return (error);
#ifdef MAC
if ((error = mac_proc_check_sched(td->td_ucred, p)))
return (error);
#endif
if ((error = cr_canseeotheruids(td->td_ucred, p->p_ucred)))
return (error);
if ((error = cr_canseeothergids(td->td_ucred, p->p_ucred)))
return (error);
if (td->td_ucred->cr_ruid != p->p_ucred->cr_ruid &&
td->td_ucred->cr_uid != p->p_ucred->cr_ruid) {
error = priv_check(td, PRIV_SCHED_DIFFCRED);
if (error)
return (error);
}
return (0);
}
/*
* Handle getting or setting the prison's unprivileged_proc_debug
* value.
*/
static int
sysctl_unprivileged_proc_debug(SYSCTL_HANDLER_ARGS)
{
int error, val;
val = prison_allow(req->td->td_ucred, PR_ALLOW_UNPRIV_DEBUG);
error = sysctl_handle_int(oidp, &val, 0, req);
if (error != 0 || req->newptr == NULL)
return (error);
if (val != 0 && val != 1)
return (EINVAL);
prison_set_allow(req->td->td_ucred, PR_ALLOW_UNPRIV_DEBUG, val);
return (0);
}
/*
* The 'unprivileged_proc_debug' flag may be used to disable a variety of
* unprivileged inter-process debugging services, including some procfs
* functionality, ptrace(), and ktrace(). In the past, inter-process
* debugging has been involved in a variety of security problems, and sites
* not requiring the service might choose to disable it when hardening
* systems.
*/
SYSCTL_PROC(_security_bsd, OID_AUTO, unprivileged_proc_debug,
CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_PRISON | CTLFLAG_SECURE |
CTLFLAG_MPSAFE, 0, 0, sysctl_unprivileged_proc_debug, "I",
"Unprivileged processes may use process debugging facilities");
/*-
* Determine whether td may debug p.
* Returns: 0 for permitted, an errno value otherwise
* Locks: Sufficient locks to protect various components of td and p
* must be held. td must be curthread, and a lock must
* be held for p.
* References: td and p must be valid for the lifetime of the call
*/
int
p_candebug(struct thread *td, struct proc *p)
{
int error, grpsubset, i, uidsubset;
KASSERT(td == curthread, ("%s: td not curthread", __func__));
PROC_LOCK_ASSERT(p, MA_OWNED);
if (td->td_proc == p)
return (0);
if ((error = priv_check(td, PRIV_DEBUG_UNPRIV)))
return (error);
if ((error = prison_check(td->td_ucred, p->p_ucred)))
return (error);
#ifdef MAC
if ((error = mac_proc_check_debug(td->td_ucred, p)))
return (error);
#endif
if ((error = cr_canseeotheruids(td->td_ucred, p->p_ucred)))
return (error);
if ((error = cr_canseeothergids(td->td_ucred, p->p_ucred)))
return (error);
/*
* Is p's group set a subset of td's effective group set? This
* includes p's egid, group access list, rgid, and svgid.
*/
grpsubset = 1;
for (i = 0; i < p->p_ucred->cr_ngroups; i++) {
if (!groupmember(p->p_ucred->cr_groups[i], td->td_ucred)) {
grpsubset = 0;
break;
}
}
grpsubset = grpsubset &&
groupmember(p->p_ucred->cr_rgid, td->td_ucred) &&
groupmember(p->p_ucred->cr_svgid, td->td_ucred);
/*
* Are the uids present in p's credential equal to td's
* effective uid? This includes p's euid, svuid, and ruid.
*/
uidsubset = (td->td_ucred->cr_uid == p->p_ucred->cr_uid &&
td->td_ucred->cr_uid == p->p_ucred->cr_svuid &&
td->td_ucred->cr_uid == p->p_ucred->cr_ruid);
/*
* If p's gids aren't a subset, or the uids aren't a subset,
* or the credential has changed, require appropriate privilege
* for td to debug p.
*/
if (!grpsubset || !uidsubset) {
error = priv_check(td, PRIV_DEBUG_DIFFCRED);
if (error)
return (error);
}
/*
* Has the credential of the process changed since the last exec()?
*/
if ((p->p_flag & P_SUGID) != 0) {
error = priv_check(td, PRIV_DEBUG_SUGID);
if (error)
return (error);
}
/* Can't trace init when securelevel > 0. */
if (p == initproc) {
error = securelevel_gt(td->td_ucred, 0);
if (error)
return (error);
}
/*
* Can't trace a process that's currently exec'ing.
*
* XXX: Note, this is not a security policy decision, it's a
* basic correctness/functionality decision. Therefore, this check
* should be moved to the caller's of p_candebug().
*/
if ((p->p_flag & P_INEXEC) != 0)
return (EBUSY);
/* Denied explicitly */
if ((p->p_flag2 & P2_NOTRACE) != 0) {
error = priv_check(td, PRIV_DEBUG_DENIED);
if (error != 0)
return (error);
}
return (0);
}
/*-
* Determine whether the subject represented by cred can "see" a socket.
* Returns: 0 for permitted, ENOENT otherwise.
*/
int
cr_canseesocket(struct ucred *cred, struct socket *so)
{
int error;
error = prison_check(cred, so->so_cred);
if (error)
return (ENOENT);
#ifdef MAC
error = mac_socket_check_visible(cred, so);
if (error)
return (error);
#endif
if (cr_canseeotheruids(cred, so->so_cred))
return (ENOENT);
if (cr_canseeothergids(cred, so->so_cred))
return (ENOENT);
return (0);
}
/*-
* Determine whether td can wait for the exit of p.
* Returns: 0 for permitted, an errno value otherwise
* Locks: Sufficient locks to protect various components of td and p
* must be held. td must be curthread, and a lock must
* be held for p.
* References: td and p must be valid for the lifetime of the call
*/
int
p_canwait(struct thread *td, struct proc *p)
{
int error;
KASSERT(td == curthread, ("%s: td not curthread", __func__));
PROC_LOCK_ASSERT(p, MA_OWNED);
if ((error = prison_check(td->td_ucred, p->p_ucred)))
return (error);
#ifdef MAC
if ((error = mac_proc_check_wait(td->td_ucred, p)))
return (error);
#endif
#if 0
/* XXXMAC: This could have odd effects on some shells. */
if ((error = cr_canseeotheruids(td->td_ucred, p->p_ucred)))
return (error);
#endif
return (0);
}
/*
* Credential management.
*
* struct ucred objects are rarely allocated but gain and lose references all
* the time (e.g., on struct file alloc/dealloc) turning refcount updates into
* a significant source of cache-line ping ponging. Common cases are worked
* around by modifying thread-local counter instead if the cred to operate on
* matches td_realucred.
*
* The counter is split into 2 parts:
* - cr_users -- total count of all struct proc and struct thread objects
* which have given cred in p_ucred and td_ucred respectively
* - cr_ref -- the actual ref count, only valid if cr_users == 0
*
* If users == 0 then cr_ref behaves similarly to refcount(9), in particular if
* the count reaches 0 the object is freeable.
* If users > 0 and curthread->td_realucred == cred, then updates are performed
* against td_ucredref.
* In other cases updates are performed against cr_ref.
*
* Changing td_realucred into something else decrements cr_users and transfers
* accumulated updates.
*/
struct ucred *
crcowget(struct ucred *cr)
{
mtx_lock(&cr->cr_mtx);
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
cr->cr_users++;
cr->cr_ref++;
mtx_unlock(&cr->cr_mtx);
return (cr);
}
static struct ucred *
crunuse(struct thread *td)
{
struct ucred *cr, *crold;
MPASS(td->td_realucred == td->td_ucred);
cr = td->td_realucred;
mtx_lock(&cr->cr_mtx);
cr->cr_ref += td->td_ucredref;
td->td_ucredref = 0;
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
cr->cr_users--;
if (cr->cr_users == 0) {
KASSERT(cr->cr_ref > 0, ("%s: ref %ld not > 0 on cred %p",
__func__, cr->cr_ref, cr));
crold = cr;
} else {
cr->cr_ref--;
crold = NULL;
}
mtx_unlock(&cr->cr_mtx);
td->td_realucred = NULL;
return (crold);
}
static void
crunusebatch(struct ucred *cr, int users, int ref)
{
KASSERT(users > 0, ("%s: passed users %d not > 0 ; cred %p",
__func__, users, cr));
mtx_lock(&cr->cr_mtx);
KASSERT(cr->cr_users >= users, ("%s: users %d not > %d on cred %p",
__func__, cr->cr_users, users, cr));
cr->cr_users -= users;
cr->cr_ref += ref;
cr->cr_ref -= users;
if (cr->cr_users > 0) {
mtx_unlock(&cr->cr_mtx);
return;
}
KASSERT(cr->cr_ref >= 0, ("%s: ref %ld not >= 0 on cred %p",
__func__, cr->cr_ref, cr));
if (cr->cr_ref > 0) {
mtx_unlock(&cr->cr_mtx);
return;
}
crfree_final(cr);
}
void
crcowfree(struct thread *td)
{
struct ucred *cr;
cr = crunuse(td);
if (cr != NULL)
crfree(cr);
}
struct ucred *
crcowsync(void)
{
struct thread *td;
struct proc *p;
struct ucred *crnew, *crold;
td = curthread;
p = td->td_proc;
PROC_LOCK_ASSERT(p, MA_OWNED);
MPASS(td->td_realucred == td->td_ucred);
if (td->td_realucred == p->p_ucred)
return (NULL);
crnew = crcowget(p->p_ucred);
crold = crunuse(td);
td->td_realucred = crnew;
td->td_ucred = td->td_realucred;
return (crold);
}
/*
* Batching.
*/
void
credbatch_add(struct credbatch *crb, struct thread *td)
{
struct ucred *cr;
MPASS(td->td_realucred != NULL);
MPASS(td->td_realucred == td->td_ucred);
MPASS(TD_GET_STATE(td) == TDS_INACTIVE);
cr = td->td_realucred;
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
if (crb->cred != cr) {
if (crb->users > 0) {
MPASS(crb->cred != NULL);
crunusebatch(crb->cred, crb->users, crb->ref);
crb->users = 0;
crb->ref = 0;
}
}
crb->cred = cr;
crb->users++;
crb->ref += td->td_ucredref;
td->td_ucredref = 0;
td->td_realucred = NULL;
}
void
credbatch_final(struct credbatch *crb)
{
MPASS(crb->cred != NULL);
MPASS(crb->users > 0);
crunusebatch(crb->cred, crb->users, crb->ref);
}
/*
* Allocate a zeroed cred structure.
*/
struct ucred *
crget(void)
{
struct ucred *cr;
cr = malloc(sizeof(*cr), M_CRED, M_WAITOK | M_ZERO);
mtx_init(&cr->cr_mtx, "cred", NULL, MTX_DEF);
cr->cr_ref = 1;
#ifdef AUDIT
audit_cred_init(cr);
#endif
#ifdef MAC
mac_cred_init(cr);
#endif
cr->cr_groups = cr->cr_smallgroups;
cr->cr_agroups =
sizeof(cr->cr_smallgroups) / sizeof(cr->cr_smallgroups[0]);
return (cr);
}
/*
* Claim another reference to a ucred structure.
*/
struct ucred *
crhold(struct ucred *cr)
{
struct thread *td;
td = curthread;
if (__predict_true(td->td_realucred == cr)) {
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
td->td_ucredref++;
return (cr);
}
mtx_lock(&cr->cr_mtx);
cr->cr_ref++;
mtx_unlock(&cr->cr_mtx);
return (cr);
}
/*
* Free a cred structure. Throws away space when ref count gets to 0.
*/
void
crfree(struct ucred *cr)
{
struct thread *td;
td = curthread;
if (__predict_true(td->td_realucred == cr)) {
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
td->td_ucredref--;
return;
}
mtx_lock(&cr->cr_mtx);
KASSERT(cr->cr_users >= 0, ("%s: users %d not >= 0 on cred %p",
__func__, cr->cr_users, cr));
cr->cr_ref--;
if (cr->cr_users > 0) {
mtx_unlock(&cr->cr_mtx);
return;
}
KASSERT(cr->cr_ref >= 0, ("%s: ref %ld not >= 0 on cred %p",
__func__, cr->cr_ref, cr));
if (cr->cr_ref > 0) {
mtx_unlock(&cr->cr_mtx);
return;
}
crfree_final(cr);
}
static void
crfree_final(struct ucred *cr)
{
KASSERT(cr->cr_users == 0, ("%s: users %d not == 0 on cred %p",
__func__, cr->cr_users, cr));
KASSERT(cr->cr_ref == 0, ("%s: ref %ld not == 0 on cred %p",
__func__, cr->cr_ref, cr));
/*
* Some callers of crget(), such as nfs_statfs(), allocate a temporary
* credential, but don't allocate a uidinfo structure.
*/
if (cr->cr_uidinfo != NULL)
uifree(cr->cr_uidinfo);
if (cr->cr_ruidinfo != NULL)
uifree(cr->cr_ruidinfo);
if (cr->cr_prison != NULL)
prison_free(cr->cr_prison);
if (cr->cr_loginclass != NULL)
loginclass_free(cr->cr_loginclass);
#ifdef AUDIT
audit_cred_destroy(cr);
#endif
#ifdef MAC
mac_cred_destroy(cr);
#endif
mtx_destroy(&cr->cr_mtx);
if (cr->cr_groups != cr->cr_smallgroups)
free(cr->cr_groups, M_CRED);
free(cr, M_CRED);
}
/*
* Copy a ucred's contents from a template. Does not block.
*/
void
crcopy(struct ucred *dest, struct ucred *src)
{
KASSERT(dest->cr_ref == 1, ("crcopy of shared ucred"));
bcopy(&src->cr_startcopy, &dest->cr_startcopy,
(unsigned)((caddr_t)&src->cr_endcopy -
(caddr_t)&src->cr_startcopy));
dest->cr_flags = src->cr_flags;
crsetgroups(dest, src->cr_ngroups, src->cr_groups);
uihold(dest->cr_uidinfo);
uihold(dest->cr_ruidinfo);
prison_hold(dest->cr_prison);
loginclass_hold(dest->cr_loginclass);
#ifdef AUDIT
audit_cred_copy(src, dest);
#endif
#ifdef MAC
mac_cred_copy(src, dest);
#endif
}
/*
* Dup cred struct to a new held one.
*/
struct ucred *
crdup(struct ucred *cr)
{
struct ucred *newcr;
newcr = crget();
crcopy(newcr, cr);
return (newcr);
}
/*
* Fill in a struct xucred based on a struct ucred.
*/
void
cru2x(struct ucred *cr, struct xucred *xcr)
{
int ngroups;
bzero(xcr, sizeof(*xcr));
xcr->cr_version = XUCRED_VERSION;
xcr->cr_uid = cr->cr_uid;
ngroups = MIN(cr->cr_ngroups, XU_NGROUPS);
xcr->cr_ngroups = ngroups;
bcopy(cr->cr_groups, xcr->cr_groups,
ngroups * sizeof(*cr->cr_groups));
}
void
cru2xt(struct thread *td, struct xucred *xcr)
{
cru2x(td->td_ucred, xcr);
xcr->cr_pid = td->td_proc->p_pid;
}
/*
* Set initial process credentials.
* Callers are responsible for providing the reference for provided credentials.
*/
void
proc_set_cred_init(struct proc *p, struct ucred *newcred)
{
p->p_ucred = crcowget(newcred);
}
/*
* Change process credentials.
* Callers are responsible for providing the reference for passed credentials
* and for freeing old ones.
*
* Process has to be locked except when it does not have credentials (as it
* should not be visible just yet) or when newcred is NULL (as this can be
* only used when the process is about to be freed, at which point it should
* not be visible anymore).
*/
void
proc_set_cred(struct proc *p, struct ucred *newcred)
{
struct ucred *cr;
cr = p->p_ucred;
MPASS(cr != NULL);
PROC_LOCK_ASSERT(p, MA_OWNED);
KASSERT(newcred->cr_users == 0, ("%s: users %d not 0 on cred %p",
__func__, newcred->cr_users, newcred));
mtx_lock(&cr->cr_mtx);
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
cr->cr_users--;
mtx_unlock(&cr->cr_mtx);
p->p_ucred = newcred;
newcred->cr_users = 1;
PROC_UPDATE_COW(p);
}
void
proc_unset_cred(struct proc *p)
{
struct ucred *cr;
MPASS(p->p_state == PRS_ZOMBIE || p->p_state == PRS_NEW);
cr = p->p_ucred;
p->p_ucred = NULL;
KASSERT(cr->cr_users > 0, ("%s: users %d not > 0 on cred %p",
__func__, cr->cr_users, cr));
mtx_lock(&cr->cr_mtx);
cr->cr_users--;
if (cr->cr_users == 0)
KASSERT(cr->cr_ref > 0, ("%s: ref %ld not > 0 on cred %p",
__func__, cr->cr_ref, cr));
mtx_unlock(&cr->cr_mtx);
crfree(cr);
}
struct ucred *
crcopysafe(struct proc *p, struct ucred *cr)
{
struct ucred *oldcred;
int groups;
PROC_LOCK_ASSERT(p, MA_OWNED);
oldcred = p->p_ucred;
while (cr->cr_agroups < oldcred->cr_agroups) {
groups = oldcred->cr_agroups;
PROC_UNLOCK(p);
crextend(cr, groups);
PROC_LOCK(p);
oldcred = p->p_ucred;
}
crcopy(cr, oldcred);
return (oldcred);
}
/*
* Extend the passed in credential to hold n items.
*/
void
crextend(struct ucred *cr, int n)
{
int cnt;
/* Truncate? */
if (n <= cr->cr_agroups)
return;
/*
* We extend by 2 each time since we're using a power of two
* allocator until we need enough groups to fill a page.
* Once we're allocating multiple pages, only allocate as many
* as we actually need. The case of processes needing a
* non-power of two number of pages seems more likely than
* a real world process that adds thousands of groups one at a
* time.
*/
if ( n < PAGE_SIZE / sizeof(gid_t) ) {
if (cr->cr_agroups == 0)
cnt = MAX(1, MINALLOCSIZE / sizeof(gid_t));
else
cnt = cr->cr_agroups * 2;
while (cnt < n)
cnt *= 2;
} else
cnt = roundup2(n, PAGE_SIZE / sizeof(gid_t));
/* Free the old array. */
if (cr->cr_groups != cr->cr_smallgroups)
free(cr->cr_groups, M_CRED);
cr->cr_groups = malloc(cnt * sizeof(gid_t), M_CRED, M_WAITOK | M_ZERO);
cr->cr_agroups = cnt;
}
/*
* Copy groups in to a credential, preserving any necessary invariants.
* Currently this includes the sorting of all supplemental gids.
* crextend() must have been called before hand to ensure sufficient
* space is available.
*/
static void
crsetgroups_locked(struct ucred *cr, int ngrp, gid_t *groups)
{
int i;
int j;
gid_t g;
KASSERT(cr->cr_agroups >= ngrp, ("cr_ngroups is too small"));
bcopy(groups, cr->cr_groups, ngrp * sizeof(gid_t));
cr->cr_ngroups = ngrp;
/*
* Sort all groups except cr_groups[0] to allow groupmember to
* perform a binary search.
*
* XXX: If large numbers of groups become common this should
* be replaced with shell sort like linux uses or possibly
* heap sort.
*/
for (i = 2; i < ngrp; i++) {
g = cr->cr_groups[i];
for (j = i-1; j >= 1 && g < cr->cr_groups[j]; j--)
cr->cr_groups[j + 1] = cr->cr_groups[j];
cr->cr_groups[j + 1] = g;
}
}
/*
* Copy groups in to a credential after expanding it if required.
* Truncate the list to (ngroups_max + 1) if it is too large.
*/
void
crsetgroups(struct ucred *cr, int ngrp, gid_t *groups)
{
if (ngrp > ngroups_max + 1)
ngrp = ngroups_max + 1;
crextend(cr, ngrp);
crsetgroups_locked(cr, ngrp, groups);
}
/*
* Get login name, if available.
*/
#ifndef _SYS_SYSPROTO_H_
struct getlogin_args {
char *namebuf;
u_int namelen;
};
#endif
/* ARGSUSED */
int
sys_getlogin(struct thread *td, struct getlogin_args *uap)
{
char login[MAXLOGNAME];
struct proc *p = td->td_proc;
size_t len;
if (uap->namelen > MAXLOGNAME)
uap->namelen = MAXLOGNAME;
PROC_LOCK(p);
SESS_LOCK(p->p_session);
len = strlcpy(login, p->p_session->s_login, uap->namelen) + 1;
SESS_UNLOCK(p->p_session);
PROC_UNLOCK(p);
if (len > uap->namelen)
return (ERANGE);
return (copyout(login, uap->namebuf, len));
}
/*
* Set login name.
*/
#ifndef _SYS_SYSPROTO_H_
struct setlogin_args {
char *namebuf;
};
#endif
/* ARGSUSED */
int
sys_setlogin(struct thread *td, struct setlogin_args *uap)
{
struct proc *p = td->td_proc;
int error;
char logintmp[MAXLOGNAME];
CTASSERT(sizeof(p->p_session->s_login) >= sizeof(logintmp));
error = priv_check(td, PRIV_PROC_SETLOGIN);
if (error)
return (error);
error = copyinstr(uap->namebuf, logintmp, sizeof(logintmp), NULL);
if (error != 0) {
if (error == ENAMETOOLONG)
error = EINVAL;
return (error);
}
AUDIT_ARG_LOGIN(logintmp);
PROC_LOCK(p);
SESS_LOCK(p->p_session);
strcpy(p->p_session->s_login, logintmp);
SESS_UNLOCK(p->p_session);
PROC_UNLOCK(p);
return (0);
}
void
setsugid(struct proc *p)
{
PROC_LOCK_ASSERT(p, MA_OWNED);
p->p_flag |= P_SUGID;
}
/*-
* Change a process's effective uid.
* Side effects: newcred->cr_uid and newcred->cr_uidinfo will be modified.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_euid(struct ucred *newcred, struct uidinfo *euip)
{
newcred->cr_uid = euip->ui_uid;
uihold(euip);
uifree(newcred->cr_uidinfo);
newcred->cr_uidinfo = euip;
}
/*-
* Change a process's effective gid.
* Side effects: newcred->cr_gid will be modified.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_egid(struct ucred *newcred, gid_t egid)
{
newcred->cr_groups[0] = egid;
}
/*-
* Change a process's real uid.
* Side effects: newcred->cr_ruid will be updated, newcred->cr_ruidinfo
* will be updated, and the old and new cr_ruidinfo proc
* counts will be updated.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_ruid(struct ucred *newcred, struct uidinfo *ruip)
{
(void)chgproccnt(newcred->cr_ruidinfo, -1, 0);
newcred->cr_ruid = ruip->ui_uid;
uihold(ruip);
uifree(newcred->cr_ruidinfo);
newcred->cr_ruidinfo = ruip;
(void)chgproccnt(newcred->cr_ruidinfo, 1, 0);
}
/*-
* Change a process's real gid.
* Side effects: newcred->cr_rgid will be updated.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_rgid(struct ucred *newcred, gid_t rgid)
{
newcred->cr_rgid = rgid;
}
/*-
* Change a process's saved uid.
* Side effects: newcred->cr_svuid will be updated.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_svuid(struct ucred *newcred, uid_t svuid)
{
newcred->cr_svuid = svuid;
}
/*-
* Change a process's saved gid.
* Side effects: newcred->cr_svgid will be updated.
* References: newcred must be an exclusive credential reference for the
* duration of the call.
*/
void
change_svgid(struct ucred *newcred, gid_t svgid)
{
newcred->cr_svgid = svgid;
}
bool allow_ptrace = true;
SYSCTL_BOOL(_security_bsd, OID_AUTO, allow_ptrace, CTLFLAG_RWTUN,
&allow_ptrace, 0,
"Deny ptrace(2) use by returning ENOSYS");
diff --git a/sys/sys/proc.h b/sys/sys/proc.h
index 3102cae7add0..8609bbd124ad 100644
--- a/sys/sys/proc.h
+++ b/sys/sys/proc.h
@@ -1,1365 +1,1366 @@
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 1986, 1989, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)proc.h 8.15 (Berkeley) 5/19/95
*/
#ifndef _SYS_PROC_H_
#define _SYS_PROC_H_
#include <sys/callout.h> /* For struct callout. */
#include <sys/event.h> /* For struct klist. */
#ifdef _KERNEL
#include <sys/_eventhandler.h>
#endif
#include <sys/condvar.h>
#ifndef _KERNEL
#include <sys/filedesc.h>
#endif
#include <sys/queue.h>
#include <sys/_lock.h>
#include <sys/lock_profile.h>
#include <sys/_mutex.h>
#include <sys/osd.h>
#include <sys/priority.h>
#include <sys/rtprio.h> /* XXX. */
#include <sys/runq.h>
#include <sys/resource.h>
#include <sys/sigio.h>
#include <sys/signal.h>
#include <sys/signalvar.h>
#ifndef _KERNEL
#include <sys/time.h> /* For structs itimerval, timeval. */
#else
#include <sys/pcpu.h>
#include <sys/systm.h>
#endif
#include <sys/ucontext.h>
#include <sys/ucred.h>
#include <sys/types.h>
#include <sys/_domainset.h>
#include <machine/proc.h> /* Machine-dependent proc substruct. */
#ifdef _KERNEL
#include <machine/cpu.h>
#endif
/*
* One structure allocated per session.
*
* List of locks
* (m) locked by s_mtx mtx
* (e) locked by proctree_lock sx
* (c) const until freeing
*/
struct session {
u_int s_count; /* Ref cnt; pgrps in session - atomic. */
struct proc *s_leader; /* (m + e) Session leader. */
struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */
struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */
struct tty *s_ttyp; /* (e) Controlling tty. */
pid_t s_sid; /* (c) Session ID. */
/* (m) Setlogin() name: */
char s_login[roundup(MAXLOGNAME, sizeof(long))];
struct mtx s_mtx; /* Mutex to protect members. */
};
/*
* One structure allocated per process group.
*
* List of locks
* (m) locked by pg_mtx mtx
* (e) locked by proctree_lock sx
* (c) const until freeing
*/
struct pgrp {
LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */
LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */
struct session *pg_session; /* (c) Pointer to session. */
struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */
pid_t pg_id; /* (c) Process group id. */
struct mtx pg_mtx; /* Mutex to protect members */
int pg_flags; /* (m) PGRP_ flags */
struct sx pg_killsx; /* Mutual exclusion between group member
* fork() and killpg() */
};
#define PGRP_ORPHANED 0x00000001 /* Group is orphaned */
/*
* pargs, used to hold a copy of the command line, if it had a sane length.
*/
struct pargs {
u_int ar_ref; /* Reference count. */
u_int ar_length; /* Length. */
u_char ar_args[1]; /* Arguments. */
};
/*-
* Description of a process.
*
* This structure contains the information needed to manage a thread of
* control, known in UN*X as a process; it has references to substructures
* containing descriptions of things that the process uses, but may share
* with related processes. The process structure and the substructures
* are always addressable except for those marked "(CPU)" below,
* which might be addressable only on a processor on which the process
* is running.
*
* Below is a key of locks used to protect each member of struct proc. The
* lock is indicated by a reference to a specific character in parens in the
* associated comment.
* * - not yet protected
* a - only touched by curproc or parent during fork/wait
* b - created at fork, never changes
* (exception aiods switch vmspaces, but they are also
* marked 'P_SYSTEM' so hopefully it will be left alone)
* c - locked by proc mtx
* d - locked by allproc_lock lock
* e - locked by proctree_lock lock
* f - session mtx
* g - process group mtx
* h - callout_lock mtx
* i - by curproc or the master session mtx
* j - locked by proc slock
* k - only accessed by curthread
* k*- only accessed by curthread and from an interrupt
* kx- only accessed by curthread and by debugger
* l - the attaching proc or attaching proc parent
* n - not locked, lazy
* o - ktrace lock
* q - td_contested lock
* r - p_peers lock
* s - see sleepq_switch(), sleeping_on_old_rtc(), and sleep(9)
* t - thread lock
* u - process stat lock
* w - process timer lock
* x - created at fork, only changes during single threading in exec
* y - created at first aio, doesn't change until exit or exec at which
* point we are single-threaded and only curthread changes it
*
* If the locking key specifies two identifiers (for example, p_pptr) then
* either lock is sufficient for read access, but both locks must be held
* for write access.
*/
struct cpuset;
struct filecaps;
struct filemon;
struct kaioinfo;
struct kaudit_record;
struct kcov_info;
struct kdtrace_proc;
struct kdtrace_thread;
struct kmsan_td;
struct kq_timer_cb_data;
struct mqueue_notifier;
struct p_sched;
struct proc;
struct procdesc;
struct racct;
struct sbuf;
struct sleepqueue;
struct socket;
struct td_sched;
struct thread;
struct trapframe;
struct turnstile;
struct vm_map;
struct vm_map_entry;
struct epoch_tracker;
struct syscall_args {
u_int code;
u_int original_code;
struct sysent *callp;
register_t args[8];
};
/*
* XXX: Does this belong in resource.h or resourcevar.h instead?
* Resource usage extension. The times in rusage structs in the kernel are
* never up to date. The actual times are kept as runtimes and tick counts
* (with control info in the "previous" times), and are converted when
* userland asks for rusage info. Backwards compatibility prevents putting
* this directly in the user-visible rusage struct.
*
* Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux.
* Locking for td_rux: (t) for all fields.
*/
struct rusage_ext {
uint64_t rux_runtime; /* (cu) Real time. */
uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */
uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */
uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */
uint64_t rux_uu; /* (c) Previous user time in usec. */
uint64_t rux_su; /* (c) Previous sys time in usec. */
uint64_t rux_tu; /* (c) Previous total time in usec. */
};
/*
* Kernel runnable context (thread).
* This is what is put to sleep and reactivated.
* Thread context. Processes may have multiple threads.
*/
struct thread {
struct mtx *volatile td_lock; /* replaces sched lock */
struct proc *td_proc; /* (*) Associated process. */
TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */
TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */
union {
TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */
struct thread *td_zombie; /* Zombie list linkage */
};
TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */
LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */
struct cpuset *td_cpuset; /* (t) CPU affinity mask. */
struct domainset_ref td_domain; /* (a) NUMA policy */
struct seltd *td_sel; /* Select queue/channel. */
struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */
struct turnstile *td_turnstile; /* (k) Associated turnstile. */
struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */
struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */
lwpid_t td_tid; /* (b) Thread ID. */
sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */
#define td_siglist td_sigqueue.sq_signals
u_char td_lend_user_pri; /* (t) Lend user pri. */
u_char td_allocdomain; /* (b) NUMA domain backing this struct thread. */
u_char td_base_ithread_pri; /* (t) Base ithread pri */
struct kmsan_td *td_kmsan; /* (k) KMSAN state */
/* Cleared during fork1(), thread_create(), or kthread_add(). */
#define td_startzero td_flags
int td_flags; /* (t) TDF_* flags. */
int td_ast; /* (t) TDA_* indicators */
int td_inhibitors; /* (t) Why can not run. */
int td_pflags; /* (k) Private thread (TDP_*) flags. */
int td_pflags2; /* (k) Private thread (TDP2_*) flags. */
int td_dupfd; /* (k) Ret value from fdopen. XXX */
int td_sqqueue; /* (t) Sleepqueue queue blocked on. */
const void *td_wchan; /* (t) Sleep address. */
const char *td_wmesg; /* (t) Reason for sleep. */
volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */
u_char td_tsqueue; /* (t) Turnstile queue blocked on. */
u_char td_stopsched; /* (k) Scheduler stopped. */
int td_locks; /* (k) Debug: count of non-spin locks */
int td_rw_rlocks; /* (k) Count of rwlock read locks. */
int td_sx_slocks; /* (k) Count of sx shared locks. */
int td_lk_slocks; /* (k) Count of lockmgr shared locks. */
struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */
const char *td_lockname; /* (t) Name of lock blocked on. */
LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */
struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */
int td_intr_nesting_level; /* (k) Interrupt recursion. */
int td_pinned; /* (k) Temporary cpu pin count. */
struct ucred *td_realucred; /* (k) Reference to credentials. */
struct ucred *td_ucred; /* (k) Used credentials, temporarily switchable. */
struct plimit *td_limit; /* (k) Resource limits. */
int td_slptick; /* (t) Time at sleep. */
int td_blktick; /* (t) Time spent blocked. */
int td_swvoltick; /* (t) Time at last SW_VOL switch. */
int td_swinvoltick; /* (t) Time at last SW_INVOL switch. */
u_int td_cow; /* (*) Number of copy-on-write faults */
struct rusage td_ru; /* (t) rusage information. */
struct rusage_ext td_rux; /* (t) Internal rusage information. */
uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */
uint64_t td_runtime; /* (t) How many cpu ticks we've run. */
u_int td_pticks; /* (t) Statclock hits for profiling */
u_int td_sticks; /* (t) Statclock hits in system mode. */
u_int td_iticks; /* (t) Statclock hits in intr mode. */
u_int td_uticks; /* (t) Statclock hits in user mode. */
int td_intrval; /* (t) Return value for sleepq. */
sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */
volatile u_int td_generation; /* (k) For detection of preemption */
stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */
int td_xsig; /* (c) Signal for ptrace */
u_long td_profil_addr; /* (k) Temporary addr until AST. */
u_int td_profil_ticks; /* (k) Temporary ticks until AST. */
char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */
struct file *td_fpop; /* (k) file referencing cdev under op */
int td_dbgflags; /* (c) Userland debugger flags */
siginfo_t td_si; /* (c) For debugger or core file */
int td_ng_outbound; /* (k) Thread entered ng from above. */
struct osd td_osd; /* (k) Object specific data. */
struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
pid_t td_dbg_forked; /* (c) Child pid for debugger. */
struct vnode *td_vp_reserved;/* (k) Preallocated vnode. */
u_int td_no_sleeping; /* (k) Sleeping disabled count. */
void *td_su; /* (k) FFS SU private */
sbintime_t td_sleeptimo; /* (t) Sleep timeout. */
int td_rtcgen; /* (s) rtc_generation of abs. sleep */
int td_errno; /* (k) Error from last syscall. */
size_t td_vslock_sz; /* (k) amount of vslock-ed space */
struct kcov_info *td_kcov_info; /* (*) Kernel code coverage data */
long td_ucredref; /* (k) references on td_realucred */
#define td_endzero td_sigmask
/* Copied during fork1(), thread_create(), or kthread_add(). */
#define td_startcopy td_endzero
sigset_t td_sigmask; /* (c) Current signal mask. */
u_char td_rqindex; /* (t) Run queue index. */
u_char td_base_pri; /* (t) Thread base kernel priority. */
u_char td_priority; /* (t) Thread active priority. */
u_char td_pri_class; /* (t) Scheduling class. */
u_char td_user_pri; /* (t) User pri from estcpu and nice. */
u_char td_base_user_pri; /* (t) Base user pri */
uintptr_t td_rb_list; /* (k) Robust list head. */
uintptr_t td_rbp_list; /* (k) Robust priv list head. */
uintptr_t td_rb_inact; /* (k) Current in-action mutex loc. */
struct syscall_args td_sa; /* (kx) Syscall parameters. Copied on
fork for child tracing. */
void *td_sigblock_ptr; /* (k) uptr for fast sigblock. */
uint32_t td_sigblock_val; /* (k) fast sigblock value read at
td_sigblock_ptr on kern entry */
#define td_endcopy td_pcb
/*
* Fields that must be manually set in fork1(), thread_create(), kthread_add(),
* or already have been set in the allocator, constructor, etc.
*/
struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */
enum td_states {
TDS_INACTIVE = 0x0,
TDS_INHIBITED,
TDS_CAN_RUN,
TDS_RUNQ,
TDS_RUNNING
} td_state; /* (t) thread state */
/* Note: td_state must be accessed using TD_{GET,SET}_STATE(). */
union {
syscallarg_t tdu_retval[2];
off_t tdu_off;
} td_uretoff; /* (k) Syscall aux returns. */
#define td_retval td_uretoff.tdu_retval
u_int td_cowgen; /* (k) Generation of COW pointers. */
/* LP64 hole */
struct callout td_slpcallout; /* (h) Callout for sleep. */
struct trapframe *td_frame; /* (k) */
vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */
int td_kstack_pages; /* (a) Size of the kstack. */
volatile u_int td_critnest; /* (k*) Critical section nest level. */
struct mdthread td_md; /* (k) Any machine-dependent fields. */
struct kaudit_record *td_ar; /* (k) Active audit record, if any. */
struct lpohead td_lprof[2]; /* (a) lock profiling objects. */
struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */
struct vnet *td_vnet; /* (k) Effective vnet. */
const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */
struct trapframe *td_intr_frame;/* (k) Frame of the current irq */
struct proc *td_rfppwait_p; /* (k) The vforked child */
struct vm_page **td_ma; /* (k) uio pages held */
int td_ma_cnt; /* (k) size of *td_ma */
/* LP64 hole */
void *td_emuldata; /* Emulator state data */
int td_lastcpu; /* (t) Last cpu we were on. */
int td_oncpu; /* (t) Which cpu we are on. */
void *td_lkpi_task; /* LinuxKPI task struct pointer */
int td_pmcpend;
void *td_remotereq; /* (c) dbg remote request. */
off_t td_ktr_io_lim; /* (k) limit for ktrace file size */
#ifdef EPOCH_TRACE
SLIST_HEAD(, epoch_tracker) td_epochs;
#endif
};
struct thread0_storage {
struct thread t0st_thread;
uint64_t t0st_sched[10];
};
struct mtx *thread_lock_block(struct thread *);
void thread_lock_block_wait(struct thread *);
void thread_lock_set(struct thread *, struct mtx *);
void thread_lock_unblock(struct thread *, struct mtx *);
#define THREAD_LOCK_ASSERT(td, type) \
mtx_assert((td)->td_lock, (type))
#define THREAD_LOCK_BLOCKED_ASSERT(td, type) \
do { \
struct mtx *__m = (td)->td_lock; \
if (__m != &blocked_lock) \
mtx_assert(__m, (type)); \
} while (0)
#ifdef INVARIANTS
#define THREAD_LOCKPTR_ASSERT(td, lock) \
do { \
struct mtx *__m; \
__m = (td)->td_lock; \
KASSERT(__m == (lock), \
("Thread %p lock %p does not match %p", td, __m, (lock))); \
} while (0)
#define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock) \
do { \
struct mtx *__m; \
__m = (td)->td_lock; \
KASSERT(__m == (lock) || __m == &blocked_lock, \
("Thread %p lock %p does not match %p", td, __m, (lock))); \
} while (0)
#define TD_LOCKS_INC(td) ((td)->td_locks++)
#define TD_LOCKS_DEC(td) do { \
KASSERT(SCHEDULER_STOPPED_TD(td) || (td)->td_locks > 0, \
("Thread %p owns no locks", (td))); \
(td)->td_locks--; \
} while (0)
#else
#define THREAD_LOCKPTR_ASSERT(td, lock)
#define THREAD_LOCKPTR_BLOCKED_ASSERT(td, lock)
#define TD_LOCKS_INC(td)
#define TD_LOCKS_DEC(td)
#endif
/*
* Flags kept in td_flags:
* To change these you MUST have the scheduler lock.
*/
#define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */
#define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */
#define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */
#define TDF_SINTR 0x00000008 /* Sleep is interruptible. */
#define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */
#define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */
#define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */
#define TDF_SIGWAIT 0x00000080 /* Ignore ignored signals */
#define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */
#define TDF_ALLPROCSUSP 0x00000200 /* suspended by SINGLE_ALLPROC */
#define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */
#define TDF_UNUSED1 0x00000800 /* Available */
#define TDF_UNUSED2 0x00001000 /* Available */
#define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */
#define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */
#define TDF_UNUSED3 0x00008000 /* Available */
#define TDF_UNUSED4 0x00010000 /* Available */
#define TDF_UNUSED5 0x00020000 /* Available */
#define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */
#define TDF_SERESTART 0x00080000 /* ERESTART on stop attempts. */
#define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */
#define TDF_SEINTR 0x00200000 /* EINTR on stop attempts. */
#define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */
#define TDF_UNUSED6 0x00800000 /* Available */
#define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */
#define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */
#define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */
#define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */
#define TDF_UNUSED7 0x10000000 /* Available */
#define TDF_UNUSED8 0x20000000 /* Available */
#define TDF_UNUSED9 0x40000000 /* Available */
#define TDF_UNUSED10 0x80000000 /* Available */
enum {
TDA_AST = 0, /* Special: call all non-flagged AST handlers */
TDA_OWEUPC,
TDA_HWPMC,
TDA_VFORK,
TDA_ALRM,
TDA_PROF,
TDA_MAC,
TDA_SCHED,
TDA_UFS,
TDA_GEOM,
TDA_KQUEUE,
TDA_RACCT,
TDA_MOD1, /* For third party use, before signals are */
TAD_MOD2, /* processed .. */
TDA_SIG,
TDA_KTRACE,
TDA_SUSPEND,
TDA_SIGSUSPEND,
TDA_MOD3, /* .. and after */
TAD_MOD4,
TDA_MAX,
};
#define TDAI(tda) (1U << (tda))
#define td_ast_pending(td, tda) ((td->td_ast & TDAI(tda)) != 0)
/* Userland debug flags */
#define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */
#define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */
#define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */
#define TDB_SCE 0x00000008 /* Thread performs syscall enter */
#define TDB_SCX 0x00000010 /* Thread performs syscall exit */
#define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */
#define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new
process */
#define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child
only) */
#define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */
#define TDB_BORN 0x00000200 /* New LWP indicator for ptrace() */
#define TDB_EXIT 0x00000400 /* Exiting LWP indicator for ptrace() */
#define TDB_VFORK 0x00000800 /* vfork indicator for ptrace() */
#define TDB_FSTP 0x00001000 /* The thread is PT_ATTACH leader */
#define TDB_STEP 0x00002000 /* (x86) PSL_T set for PT_STEP */
#define TDB_SSWITCH 0x00004000 /* Suspended in ptracestop */
#define TDB_BOUNDARY 0x00008000 /* ptracestop() at boundary */
#define TDB_COREDUMPREQ 0x00010000 /* Coredump request */
#define TDB_SCREMOTEREQ 0x00020000 /* Remote syscall request */
/*
* "Private" flags kept in td_pflags:
* These are only written by curthread and thus need no locking.
*/
#define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */
#define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */
#define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */
#define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */
#define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */
#define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */
#define TDP_DEADLKTREAT 0x00000040 /* Lock acquisition - deadlock treatment. */
#define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */
#define TDP_SIGFASTBLOCK 0x00000100 /* Fast sigblock active */
#define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */
#define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */
#define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */
#define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */
#define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */
#define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */
#define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */
#define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */
#define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */
#define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */
#define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */
#define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */
#define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */
#define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */
#define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */
#define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */
#define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */
#define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */
#define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */
#define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */
#define TDP_INTCPCALLOUT 0x20000000 /* used by netinet/tcp_timer.c */
#define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */
#define TDP_SIGFASTPENDING 0x80000000 /* Pending signal due to sigfastblock */
#define TDP2_SBPAGES 0x00000001 /* Owns sbusy on some pages */
#define TDP2_COMPAT32RB 0x00000002 /* compat32 ABI for robust lists */
#define TDP2_ACCT 0x00000004 /* Doing accounting */
/*
* Reasons that the current thread can not be run yet.
* More than one may apply.
*/
#define TDI_SUSPENDED 0x0001 /* On suspension queue. */
#define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */
#define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */
#define TDI_LOCK 0x0008 /* Stopped on a lock. */
#define TDI_IWAIT 0x0010 /* Awaiting interrupt. */
#define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING)
#define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL)
#define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED)
#define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED)
#define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK)
#define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT)
#ifdef _KERNEL
#define TD_GET_STATE(td) atomic_load_int(&(td)->td_state)
#else
#define TD_GET_STATE(td) ((td)->td_state)
#endif
#define TD_IS_RUNNING(td) (TD_GET_STATE(td) == TDS_RUNNING)
#define TD_ON_RUNQ(td) (TD_GET_STATE(td) == TDS_RUNQ)
#define TD_CAN_RUN(td) (TD_GET_STATE(td) == TDS_CAN_RUN)
#define TD_IS_INHIBITED(td) (TD_GET_STATE(td) == TDS_INHIBITED)
#define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED)
#define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD)
#define TD_CAN_ABORT(td) (TD_ON_SLEEPQ((td)) && \
((td)->td_flags & TDF_SINTR) != 0)
#define KTDSTATE(td) \
(((td)->td_inhibitors & TDI_SLEEPING) != 0 ? "sleep" : \
((td)->td_inhibitors & TDI_SUSPENDED) != 0 ? "suspended" : \
((td)->td_inhibitors & TDI_SWAPPED) != 0 ? "swapped" : \
((td)->td_inhibitors & TDI_LOCK) != 0 ? "blocked" : \
((td)->td_inhibitors & TDI_IWAIT) != 0 ? "iwait" : "yielding")
#define TD_SET_INHIB(td, inhib) do { \
TD_SET_STATE(td, TDS_INHIBITED); \
(td)->td_inhibitors |= (inhib); \
} while (0)
#define TD_CLR_INHIB(td, inhib) do { \
if (((td)->td_inhibitors & (inhib)) && \
(((td)->td_inhibitors &= ~(inhib)) == 0)) \
TD_SET_STATE(td, TDS_CAN_RUN); \
} while (0)
#define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING)
#define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED)
#define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK)
#define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED)
#define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT)
#define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING)
#define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING)
#define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED)
#define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK)
#define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED)
#define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT)
#ifdef _KERNEL
#define TD_SET_STATE(td, state) atomic_store_int(&(td)->td_state, state)
#else
#define TD_SET_STATE(td, state) (td)->td_state = state
#endif
#define TD_SET_RUNNING(td) TD_SET_STATE(td, TDS_RUNNING)
#define TD_SET_RUNQ(td) TD_SET_STATE(td, TDS_RUNQ)
#define TD_SET_CAN_RUN(td) TD_SET_STATE(td, TDS_CAN_RUN)
#define TD_SBDRY_INTR(td) \
(((td)->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 0)
#define TD_SBDRY_ERRNO(td) \
(((td)->td_flags & TDF_SEINTR) != 0 ? EINTR : ERESTART)
/*
* Process structure.
*/
struct proc {
LIST_ENTRY(proc) p_list; /* (d) List of all processes. */
TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */
struct mtx p_slock; /* process spin lock */
struct ucred *p_ucred; /* (c) Process owner's identity. */
struct filedesc *p_fd; /* (b) Open files. */
struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */
struct pwddesc *p_pd; /* (b) Cwd, chroot, jail, umask */
struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */
struct plimit *p_limit; /* (c) Resource limits. */
struct callout p_limco; /* (c) Limit callout handle */
struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */
int p_flag; /* (c) P_* flags. */
int p_flag2; /* (c) P2_* flags. */
enum p_states {
PRS_NEW = 0, /* In creation */
PRS_NORMAL, /* threads can be run. */
PRS_ZOMBIE
} p_state; /* (j/c) Process status. */
pid_t p_pid; /* (b) Process identifier. */
LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */
LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */
struct proc *p_pptr; /* (c + e) Pointer to parent process. */
LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */
LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */
struct proc *p_reaper; /* (e) My reaper. */
LIST_HEAD(, proc) p_reaplist; /* (e) List of my descendants
(if I am reaper). */
LIST_ENTRY(proc) p_reapsibling; /* (e) List of siblings - descendants of
the same reaper. */
struct mtx p_mtx; /* (n) Lock for this struct. */
struct mtx p_statmtx; /* Lock for the stats */
struct mtx p_itimmtx; /* Lock for the virt/prof timers */
struct mtx p_profmtx; /* Lock for the profiling */
struct ksiginfo *p_ksi; /* Locked by parent proc lock */
sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */
#define p_siglist p_sigqueue.sq_signals
pid_t p_oppid; /* (c + e) Real parent pid. */
/* The following fields are all zeroed upon creation in fork. */
#define p_startzero p_vmspace
struct vmspace *p_vmspace; /* (b) Address space. */
u_int p_swtick; /* (c) Tick when swapped in or out. */
u_int p_cowgen; /* (c) Generation of COW pointers. */
struct itimerval p_realtimer; /* (c) Alarm timer. */
struct rusage p_ru; /* (a) Exit information. */
struct rusage_ext p_rux; /* (cu) Internal resource usage. */
struct rusage_ext p_crux; /* (c) Internal child resource usage. */
int p_profthreads; /* (c) Num threads in addupc_task. */
volatile int p_exitthreads; /* (j) Number of threads exiting */
int p_traceflag; /* (o) Kernel trace points. */
struct ktr_io_params *p_ktrioparms; /* (c + o) Params for ktrace. */
struct vnode *p_textvp; /* (b) Vnode of executable. */
struct vnode *p_textdvp; /* (b) Dir containing textvp. */
char *p_binname; /* (b) Binary hardlink name. */
u_int p_lock; /* (c) Proclock (prevent swap) count. */
struct sigiolst p_sigiolst; /* (c) List of sigio sources. */
int p_sigparent; /* (c) Signal to parent on exit. */
int p_sig; /* (n) For core dump/debugger XXX. */
u_int p_ptevents; /* (c + e) ptrace() event mask. */
struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */
struct thread *p_singlethread;/* (c + j) If single threading this is it */
int p_suspcount; /* (j) Num threads in suspended mode. */
struct thread *p_xthread; /* (c) Trap thread */
int p_boundary_count;/* (j) Num threads at user boundary */
int p_pendingcnt; /* how many signals are pending */
struct itimers *p_itimers; /* (c) POSIX interval timers. */
struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */
u_int p_treeflag; /* (e) P_TREE flags */
int p_pendingexits; /* (c) Count of pending thread exits. */
struct filemon *p_filemon; /* (c) filemon-specific data. */
int p_pdeathsig; /* (c) Signal from parent on exit. */
/* End area that is zeroed on creation. */
#define p_endzero p_magic
/* The following fields are all copied upon creation in fork. */
#define p_startcopy p_endzero
u_int p_magic; /* (b) Magic number. */
int p_osrel; /* (x) osreldate for the
binary (from ELF note, if any) */
uint32_t p_fctl0; /* (x) ABI feature control, ELF note */
char p_comm[MAXCOMLEN + 1]; /* (x) Process name. */
struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */
struct pargs *p_args; /* (c) Process arguments. */
rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */
signed char p_nice; /* (c) Process "nice" value. */
int p_fibnum; /* in this routing domain XXX MRT */
pid_t p_reapsubtree; /* (e) Pid of the direct child of the
reaper which spawned
our subtree. */
uint64_t p_elf_flags; /* (x) ELF flags */
void *p_elf_brandinfo; /* (x) Elf_Brandinfo, NULL for
non ELF binaries. */
sbintime_t p_umtx_min_timeout;
/* End area that is copied on creation. */
#define p_endcopy p_xexit
u_int p_xexit; /* (c) Exit code. */
u_int p_xsig; /* (c) Stop/kill sig. */
struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */
struct knlist *p_klist; /* (c) Knotes attached to this proc. */
int p_numthreads; /* (c) Number of threads. */
struct mdproc p_md; /* Any machine-dependent fields. */
struct callout p_itcallout; /* (h + c) Interval timer callout. */
u_short p_acflag; /* (c) Accounting flags. */
struct proc *p_peers; /* (r) */
struct proc *p_leader; /* (b) */
void *p_emuldata; /* (c) Emulator state data. */
struct label *p_label; /* (*) Proc (not subject) MAC label. */
STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */
LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/
struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */
struct cv p_pwait; /* (*) wait cv for exit/exec. */
uint64_t p_prev_runtime; /* (c) Resource usage accounting. */
struct racct *p_racct; /* (b) Resource accounting. */
int p_throttled; /* (c) Flag for racct pcpu throttling */
/*
* An orphan is the child that has been re-parented to the
* debugger as a result of attaching to it. Need to keep
* track of them for parent to be able to collect the exit
* status of what used to be children.
*/
LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */
LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */
TAILQ_HEAD(, kq_timer_cb_data) p_kqtim_stop; /* (c) */
LIST_ENTRY(proc) p_jaillist; /* (d) Jail process linkage. */
};
#define p_session p_pgrp->pg_session
#define p_pgid p_pgrp->pg_id
#define NOCPU (-1) /* For when we aren't on a CPU. */
#define NOCPU_OLD (255)
#define MAXCPU_OLD (254)
#define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock)
#define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock)
#define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type))
#define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx)
#define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx)
#define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type))
#define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx)
#define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx)
#define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type))
#define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx)
#define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx)
#define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type))
/* These flags are kept in p_flag. */
#define P_ADVLOCK 0x00000001 /* Process may hold a POSIX advisory
lock. */
#define P_CONTROLT 0x00000002 /* Has a controlling terminal. */
#define P_KPROC 0x00000004 /* Kernel process. */
#define P_UNUSED3 0x00000008 /* --available-- */
#define P_PPWAIT 0x00000010 /* Parent is waiting for child to
exec/exit. */
#define P_PROFIL 0x00000020 /* Has started profiling. */
#define P_STOPPROF 0x00000040 /* Has thread requesting to stop
profiling. */
#define P_HADTHREADS 0x00000080 /* Has had threads (no cleanup
shortcuts) */
#define P_SUGID 0x00000100 /* Had set id privileges since last
exec. */
#define P_SYSTEM 0x00000200 /* System proc: no sigs, stats or
swapping. */
#define P_SINGLE_EXIT 0x00000400 /* Threads suspending should exit,
not wait. */
#define P_TRACED 0x00000800 /* Debugged process being traced. */
#define P_WAITED 0x00001000 /* Someone is waiting for us. */
#define P_WEXIT 0x00002000 /* Working on exiting. */
#define P_EXEC 0x00004000 /* Process called exec. */
#define P_WKILLED 0x00008000 /* Killed, go to kernel/user boundary
ASAP. */
#define P_CONTINUED 0x00010000 /* Proc has continued from a stopped
state. */
#define P_STOPPED_SIG 0x00020000 /* Stopped due to SIGSTOP/SIGTSTP. */
#define P_STOPPED_TRACE 0x00040000 /* Stopped because of tracing. */
#define P_STOPPED_SINGLE 0x00080000 /* Only 1 thread can continue (not to
user). */
#define P_PROTECTED 0x00100000 /* Do not kill on memory overcommit. */
#define P_SIGEVENT 0x00200000 /* Process pending signals changed. */
#define P_SINGLE_BOUNDARY 0x00400000 /* Threads should suspend at user
boundary. */
#define P_HWPMC 0x00800000 /* Process is using HWPMCs */
#define P_JAILED 0x01000000 /* Process is in jail. */
#define P_TOTAL_STOP 0x02000000 /* Stopped in stop_all_proc. */
#define P_INEXEC 0x04000000 /* Process is in execve(). */
#define P_STATCHILD 0x08000000 /* Child process stopped or exited. */
#define P_INMEM 0x10000000 /* Loaded into memory. */
#define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */
#define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */
#define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */
#define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE)
#define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED)
#define P_KILLED(p) ((p)->p_flag & P_WKILLED)
/* These flags are kept in p_flag2. */
#define P2_INHERIT_PROTECTED 0x00000001 /* New children get
P_PROTECTED. */
#define P2_NOTRACE 0x00000002 /* No ptrace(2) attach or
coredumps. */
#define P2_NOTRACE_EXEC 0x00000004 /* Keep P2_NOPTRACE on
exec(2). */
#define P2_AST_SU 0x00000008 /* Handles SU ast for
kthreads. */
#define P2_PTRACE_FSTP 0x00000010 /* SIGSTOP from PT_ATTACH not
yet handled. */
#define P2_TRAPCAP 0x00000020 /* SIGTRAP on ENOTCAPABLE */
#define P2_ASLR_ENABLE 0x00000040 /* Force enable ASLR. */
#define P2_ASLR_DISABLE 0x00000080 /* Force disable ASLR. */
#define P2_ASLR_IGNSTART 0x00000100 /* Enable ASLR to consume sbrk
area. */
#define P2_PROTMAX_ENABLE 0x00000200 /* Force enable implied
PROT_MAX. */
#define P2_PROTMAX_DISABLE 0x00000400 /* Force disable implied
PROT_MAX. */
#define P2_STKGAP_DISABLE 0x00000800 /* Disable stack gap for
MAP_STACK */
#define P2_STKGAP_DISABLE_EXEC 0x00001000 /* Stack gap disabled
after exec */
#define P2_ITSTOPPED 0x00002000
#define P2_PTRACEREQ 0x00004000 /* Active ptrace req */
#define P2_NO_NEW_PRIVS 0x00008000 /* Ignore setuid */
#define P2_WXORX_DISABLE 0x00010000 /* WX mappings enabled */
#define P2_WXORX_ENABLE_EXEC 0x00020000 /* WXORX enabled after exec */
#define P2_WEXIT 0x00040000 /* exit just started, no
external thread_single() is
permitted */
#define P2_REAPKILLED 0x00080000
#define P2_MEMBAR_PRIVE 0x00100000 /* membar private expedited
registered */
#define P2_MEMBAR_PRIVE_SYNCORE 0x00200000 /* membar private expedited
sync core registered */
#define P2_MEMBAR_GLOBE 0x00400000 /* membar global expedited
registered */
/* Flags protected by proctree_lock, kept in p_treeflags. */
#define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */
#define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan
list */
#define P_TREE_REAPER 0x00000004 /* Reaper of subtree */
#define P_TREE_GRPEXITED 0x00000008 /* exit1() done with job ctl */
/*
* These were process status values (p_stat), now they are only used in
* legacy conversion code.
*/
#define SIDL 1 /* Process being created by fork. */
#define SRUN 2 /* Currently runnable. */
#define SSLEEP 3 /* Sleeping on an address. */
#define SSTOP 4 /* Process debugging or suspension. */
#define SZOMB 5 /* Awaiting collection by parent. */
#define SWAIT 6 /* Waiting for interrupt. */
#define SLOCK 7 /* Blocked on a lock. */
#define P_MAGIC 0xbeefface
#ifdef _KERNEL
/* Types and flags for mi_switch(9). */
#define SW_TYPE_MASK 0xff /* First 8 bits are switch type */
#define SWT_OWEPREEMPT 1 /* Switching due to owepreempt. */
#define SWT_TURNSTILE 2 /* Turnstile contention. */
#define SWT_SLEEPQ 3 /* Sleepq wait. */
#define SWT_RELINQUISH 4 /* yield call. */
#define SWT_NEEDRESCHED 5 /* NEEDRESCHED was set. */
#define SWT_IDLE 6 /* Switching from the idle thread. */
#define SWT_IWAIT 7 /* Waiting for interrupts. */
#define SWT_SUSPEND 8 /* Thread suspended. */
#define SWT_REMOTEPREEMPT 9 /* Remote processor preempted. */
#define SWT_REMOTEWAKEIDLE 10 /* Remote processor preempted idle. */
#define SWT_BIND 11 /* Thread bound to a new CPU. */
#define SWT_COUNT 12 /* Number of switch types. */
/* Flags */
#define SW_VOL 0x0100 /* Voluntary switch. */
#define SW_INVOL 0x0200 /* Involuntary switch. */
#define SW_PREEMPT 0x0400 /* The invol switch is a preemption */
/* How values for thread_single(). */
#define SINGLE_NO_EXIT 0
#define SINGLE_EXIT 1
#define SINGLE_BOUNDARY 2
#define SINGLE_ALLPROC 3
#ifdef MALLOC_DECLARE
MALLOC_DECLARE(M_PARGS);
MALLOC_DECLARE(M_SESSION);
MALLOC_DECLARE(M_SUBPROC);
#endif
#define FOREACH_PROC_IN_SYSTEM(p) \
LIST_FOREACH((p), &allproc, p_list)
#define FOREACH_THREAD_IN_PROC(p, td) \
TAILQ_FOREACH((td), &(p)->p_threads, td_plist)
#define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads)
/*
* We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit
* in a pid_t, as it is used to represent "no process group".
*/
#define PID_MAX 99999
#define NO_PID 100000
#define THREAD0_TID NO_PID
extern pid_t pid_max;
#define SESS_LEADER(p) ((p)->p_session->s_leader == (p))
/* Lock and unlock a process. */
#define PROC_LOCK(p) mtx_lock(&(p)->p_mtx)
#define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx)
#define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx)
#define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx)
#define PROC_WAIT_UNLOCKED(p) mtx_wait_unlocked(&(p)->p_mtx)
#define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type))
/* Lock and unlock a process group. */
#define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx)
#define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx)
#define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx)
#define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type))
#define PGRP_LOCK_PGSIGNAL(pg) do { \
if ((pg) != NULL) \
PGRP_LOCK(pg); \
} while (0)
#define PGRP_UNLOCK_PGSIGNAL(pg) do { \
if ((pg) != NULL) \
PGRP_UNLOCK(pg); \
} while (0)
/* Lock and unlock a session. */
#define SESS_LOCK(s) mtx_lock(&(s)->s_mtx)
#define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx)
#define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx)
#define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type))
/*
* Non-zero p_lock ensures that:
* - exit1() is not performed until p_lock reaches zero;
* - the process' threads stack are not swapped out if they are currently
* not (P_INMEM).
*
* PHOLD() asserts that the process (except the current process) is
* not exiting, increments p_lock and swaps threads stacks into memory,
* if needed.
* _PHOLD() is same as PHOLD(), it takes the process locked.
* _PHOLD_LITE() also takes the process locked, but comparing with
* _PHOLD(), it only guarantees that exit1() is not executed,
* faultin() is not called.
*/
#define PHOLD(p) do { \
PROC_LOCK(p); \
_PHOLD(p); \
PROC_UNLOCK(p); \
} while (0)
#define _PHOLD(p) do { \
PROC_LOCK_ASSERT((p), MA_OWNED); \
KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
("PHOLD of exiting process %p", p)); \
(p)->p_lock++; \
if (((p)->p_flag & P_INMEM) == 0) \
faultin((p)); \
} while (0)
#define _PHOLD_LITE(p) do { \
PROC_LOCK_ASSERT((p), MA_OWNED); \
KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \
("PHOLD of exiting process %p", p)); \
(p)->p_lock++; \
} while (0)
#define PROC_ASSERT_HELD(p) do { \
KASSERT((p)->p_lock > 0, ("process %p not held", p)); \
} while (0)
#define PRELE(p) do { \
PROC_LOCK((p)); \
_PRELE((p)); \
PROC_UNLOCK((p)); \
} while (0)
#define _PRELE(p) do { \
PROC_LOCK_ASSERT((p), MA_OWNED); \
PROC_ASSERT_HELD(p); \
(--(p)->p_lock); \
if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \
wakeup(&(p)->p_lock); \
} while (0)
#define PROC_ASSERT_NOT_HELD(p) do { \
KASSERT((p)->p_lock == 0, ("process %p held", p)); \
} while (0)
#define PROC_UPDATE_COW(p) do { \
struct proc *_p = (p); \
PROC_LOCK_ASSERT((_p), MA_OWNED); \
atomic_store_int(&_p->p_cowgen, _p->p_cowgen + 1); \
} while (0)
#define PROC_COW_CHANGECOUNT(td, p) ({ \
struct thread *_td = (td); \
struct proc *_p = (p); \
MPASS(_td == curthread); \
PROC_LOCK_ASSERT(_p, MA_OWNED); \
_p->p_cowgen - _td->td_cowgen; \
})
/* Check whether a thread is safe to be swapped out. */
#define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP)
/* Control whether or not it is safe for curthread to sleep. */
#define THREAD_NO_SLEEPING() do { \
curthread->td_no_sleeping++; \
MPASS(curthread->td_no_sleeping > 0); \
} while (0)
#define THREAD_SLEEPING_OK() do { \
MPASS(curthread->td_no_sleeping > 0); \
curthread->td_no_sleeping--; \
} while (0)
#define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0)
#define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash])
#define PIDHASHLOCK(pid) (&pidhashtbl_lock[((pid) & pidhashlock)])
extern LIST_HEAD(pidhashhead, proc) *pidhashtbl;
extern struct sx *pidhashtbl_lock;
extern u_long pidhash;
extern u_long pidhashlock;
#define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash])
extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl;
extern u_long pgrphash;
extern struct sx allproc_lock;
extern int allproc_gen;
extern struct sx proctree_lock;
extern struct mtx ppeers_lock;
extern struct mtx procid_lock;
extern struct proc proc0; /* Process slot for swapper. */
extern struct thread0_storage thread0_st; /* Primary thread in proc0. */
#define thread0 (thread0_st.t0st_thread)
extern struct vmspace vmspace0; /* VM space for proc0. */
extern int hogticks; /* Limit on kernel cpu hogs. */
extern int lastpid;
extern int nprocs, maxproc; /* Current and max number of procs. */
extern int maxprocperuid; /* Max procs per uid. */
extern u_long ps_arg_cache_limit;
LIST_HEAD(proclist, proc);
TAILQ_HEAD(procqueue, proc);
TAILQ_HEAD(threadqueue, thread);
extern struct proclist allproc; /* List of all processes. */
extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */
extern struct uma_zone *proc_zone;
extern struct uma_zone *pgrp_zone;
struct proc *pfind(pid_t); /* Find process by id. */
struct proc *pfind_any(pid_t); /* Find (zombie) process by id. */
struct proc *pfind_any_locked(pid_t pid); /* Find process by id, locked. */
struct pgrp *pgfind(pid_t); /* Find process group by id. */
void pidhash_slockall(void); /* Shared lock all pid hash lists. */
void pidhash_sunlockall(void); /* Shared unlock all pid hash lists. */
struct fork_req {
int fr_flags;
int fr_pages;
int *fr_pidp;
struct proc **fr_procp;
int *fr_pd_fd;
int fr_pd_flags;
struct filecaps *fr_pd_fcaps;
int fr_flags2;
#define FR2_DROPSIG_CAUGHT 0x00000001 /* Drop caught non-DFL signals */
#define FR2_SHARE_PATHS 0x00000002 /* Invert sense of RFFDG for paths */
#define FR2_KPROC 0x00000004 /* Create a kernel process */
};
/*
* pget() flags.
*/
#define PGET_HOLD 0x00001 /* Hold the process. */
#define PGET_CANSEE 0x00002 /* Check against p_cansee(). */
#define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */
#define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */
#define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */
#define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */
#define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */
#define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT)
int pget(pid_t pid, int flags, struct proc **pp);
/* ast_register() flags */
#define ASTR_ASTF_REQUIRED 0x0001 /* td_ast TDAI(TDA_X) flag set is
required for call */
#define ASTR_TDP 0x0002 /* td_pflags flag set is required */
#define ASTR_KCLEAR 0x0004 /* call me on ast_kclear() */
#define ASTR_UNCOND 0x0008 /* call me always */
void ast(struct trapframe *framep);
void ast_kclear(struct thread *td);
void ast_register(int ast, int ast_flags, int tdp,
void (*f)(struct thread *td, int asts));
void ast_deregister(int tda);
void ast_sched_locked(struct thread *td, int tda);
void ast_sched_mask(struct thread *td, int ast);
void ast_sched(struct thread *td, int tda);
void ast_unsched_locked(struct thread *td, int tda);
struct thread *choosethread(void);
+int cr_bsd_visible(struct ucred *u1, struct ucred *u2);
int cr_cansee(struct ucred *u1, struct ucred *u2);
int cr_canseesocket(struct ucred *cred, struct socket *so);
int cr_canseeothergids(struct ucred *u1, struct ucred *u2);
int cr_canseeotheruids(struct ucred *u1, struct ucred *u2);
int cr_canseejailproc(struct ucred *u1, struct ucred *u2);
int cr_cansignal(struct ucred *cred, struct proc *proc, int signum);
int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp,
struct session *sess);
int enterthispgrp(struct proc *p, struct pgrp *pgrp);
void faultin(struct proc *p);
int fork1(struct thread *, struct fork_req *);
void fork_exit(void (*)(void *, struct trapframe *), void *,
struct trapframe *);
void fork_return(struct thread *, struct trapframe *);
int inferior(struct proc *p);
void itimer_proc_continue(struct proc *p);
void kqtimer_proc_continue(struct proc *p);
void kern_proc_vmmap_resident(struct vm_map *map, struct vm_map_entry *entry,
int *resident_count, bool *super);
void kern_yield(int);
void kick_proc0(void);
void killjobc(void);
int leavepgrp(struct proc *p);
int maybe_preempt(struct thread *td);
void maybe_yield(void);
void mi_switch(int flags);
int p_candebug(struct thread *td, struct proc *p);
int p_cansee(struct thread *td, struct proc *p);
int p_cansched(struct thread *td, struct proc *p);
int p_cansignal(struct thread *td, struct proc *p, int signum);
int p_canwait(struct thread *td, struct proc *p);
struct pargs *pargs_alloc(int len);
void pargs_drop(struct pargs *pa);
void pargs_hold(struct pargs *pa);
void proc_add_orphan(struct proc *child, struct proc *parent);
int proc_get_binpath(struct proc *p, char *binname, char **fullpath,
char **freepath);
int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb);
int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb);
int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb);
void procinit(void);
int proc_iterate(int (*cb)(struct proc *, void *), void *cbarg);
void proc_linkup0(struct proc *p, struct thread *td);
void proc_linkup(struct proc *p, struct thread *td);
struct proc *proc_realparent(struct proc *child);
void proc_reap(struct thread *td, struct proc *p, int *status, int options);
void proc_reparent(struct proc *child, struct proc *newparent, bool set_oppid);
void proc_set_p2_wexit(struct proc *p);
void proc_set_traced(struct proc *p, bool stop);
void proc_wkilled(struct proc *p);
struct pstats *pstats_alloc(void);
void pstats_fork(struct pstats *src, struct pstats *dst);
void pstats_free(struct pstats *ps);
void proc_clear_orphan(struct proc *p);
void reaper_abandon_children(struct proc *p, bool exiting);
int securelevel_ge(struct ucred *cr, int level);
int securelevel_gt(struct ucred *cr, int level);
void sess_hold(struct session *);
void sess_release(struct session *);
int setrunnable(struct thread *, int);
void setsugid(struct proc *p);
bool should_yield(void);
int sigonstack(size_t sp);
void stopevent(struct proc *, u_int, u_int);
struct thread *tdfind(lwpid_t, pid_t);
void threadinit(void);
void tidhash_add(struct thread *);
void tidhash_remove(struct thread *);
void cpu_idle(int);
int cpu_idle_wakeup(int);
extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */
void cpu_switch(struct thread *, struct thread *, struct mtx *);
void cpu_sync_core(void);
void cpu_throw(struct thread *, struct thread *) __dead2;
bool curproc_sigkilled(void);
void userret(struct thread *, struct trapframe *);
void cpu_exit(struct thread *);
void exit1(struct thread *, int, int) __dead2;
void cpu_copy_thread(struct thread *td, struct thread *td0);
bool cpu_exec_vmspace_reuse(struct proc *p, struct vm_map *map);
int cpu_fetch_syscall_args(struct thread *td);
void cpu_fork(struct thread *, struct proc *, struct thread *, int);
void cpu_fork_kthread_handler(struct thread *, void (*)(void *), void *);
int cpu_procctl(struct thread *td, int idtype, id_t id, int com,
void *data);
void cpu_set_syscall_retval(struct thread *, int);
void cpu_set_upcall(struct thread *, void (*)(void *), void *,
stack_t *);
int cpu_set_user_tls(struct thread *, void *tls_base);
void cpu_thread_alloc(struct thread *);
void cpu_thread_clean(struct thread *);
void cpu_thread_exit(struct thread *);
void cpu_thread_free(struct thread *);
void cpu_thread_swapin(struct thread *);
void cpu_thread_swapout(struct thread *);
struct thread *thread_alloc(int pages);
int thread_alloc_stack(struct thread *, int pages);
int thread_check_susp(struct thread *td, bool sleep);
void thread_cow_get_proc(struct thread *newtd, struct proc *p);
void thread_cow_get(struct thread *newtd, struct thread *td);
void thread_cow_free(struct thread *td);
void thread_cow_update(struct thread *td);
void thread_cow_synced(struct thread *td);
int thread_create(struct thread *td, struct rtprio *rtp,
int (*initialize_thread)(struct thread *, void *), void *thunk);
void thread_exit(void) __dead2;
void thread_free(struct thread *td);
void thread_link(struct thread *td, struct proc *p);
void thread_reap_barrier(void);
int thread_single(struct proc *p, int how);
void thread_single_end(struct proc *p, int how);
void thread_stash(struct thread *td);
void thread_stopped(struct proc *p);
void childproc_stopped(struct proc *child, int reason);
void childproc_continued(struct proc *child);
void childproc_exited(struct proc *child);
void thread_run_flash(struct thread *td);
int thread_suspend_check(int how);
bool thread_suspend_check_needed(void);
void thread_suspend_switch(struct thread *, struct proc *p);
void thread_suspend_one(struct thread *td);
void thread_unlink(struct thread *td);
void thread_unsuspend(struct proc *p);
void thread_wait(struct proc *p);
bool stop_all_proc_block(void);
void stop_all_proc_unblock(void);
void stop_all_proc(void);
void resume_all_proc(void);
static __inline int
curthread_pflags_set(int flags)
{
struct thread *td;
int save;
td = curthread;
save = ~flags | (td->td_pflags & flags);
td->td_pflags |= flags;
return (save);
}
static __inline void
curthread_pflags_restore(int save)
{
curthread->td_pflags &= save;
}
static __inline int
curthread_pflags2_set(int flags)
{
struct thread *td;
int save;
td = curthread;
save = ~flags | (td->td_pflags2 & flags);
td->td_pflags2 |= flags;
return (save);
}
static __inline void
curthread_pflags2_restore(int save)
{
curthread->td_pflags2 &= save;
}
static __inline __pure2 struct td_sched *
td_get_sched(struct thread *td)
{
return ((struct td_sched *)&td[1]);
}
#define PROC_ID_PID 0
#define PROC_ID_GROUP 1
#define PROC_ID_SESSION 2
#define PROC_ID_REAP 3
void proc_id_set(int type, pid_t id);
void proc_id_set_cond(int type, pid_t id);
void proc_id_clear(int type, pid_t id);
EVENTHANDLER_LIST_DECLARE(process_ctor);
EVENTHANDLER_LIST_DECLARE(process_dtor);
EVENTHANDLER_LIST_DECLARE(process_init);
EVENTHANDLER_LIST_DECLARE(process_fini);
EVENTHANDLER_LIST_DECLARE(process_exit);
EVENTHANDLER_LIST_DECLARE(process_fork);
EVENTHANDLER_LIST_DECLARE(process_exec);
EVENTHANDLER_LIST_DECLARE(thread_ctor);
EVENTHANDLER_LIST_DECLARE(thread_dtor);
EVENTHANDLER_LIST_DECLARE(thread_init);
#endif /* _KERNEL */
#endif /* !_SYS_PROC_H_ */