subr_syscall.c revision 304188
1/*-
2 * Copyright (C) 1994, David Greenman
3 * Copyright (c) 1990, 1993
4 *	The Regents of the University of California.  All rights reserved.
5 * Copyright (C) 2010 Konstantin Belousov <kib@freebsd.org>
6 *
7 * This code is derived from software contributed to Berkeley by
8 * the University of Utah, and William Jolitz.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 3. All advertising materials mentioning features or use of this software
19 *    must display the following acknowledgement:
20 *	This product includes software developed by the University of
21 *	California, Berkeley and its contributors.
22 * 4. Neither the name of the University nor the names of its contributors
23 *    may be used to endorse or promote products derived from this software
24 *    without specific prior written permission.
25 *
26 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 * SUCH DAMAGE.
37 *
38 *	from: @(#)trap.c	7.4 (Berkeley) 5/13/91
39 */
40
41#include "opt_capsicum.h"
42#include "opt_ktrace.h"
43#include "opt_kdtrace.h"
44
45__FBSDID("$FreeBSD: stable/10/sys/kern/subr_syscall.c 304188 2016-08-15 21:10:41Z jhb $");
46
47#include <sys/capsicum.h>
48#include <sys/ktr.h>
49#ifdef KTRACE
50#include <sys/uio.h>
51#include <sys/ktrace.h>
52#endif
53#include <security/audit/audit.h>
54
55static inline int
56syscallenter(struct thread *td, struct syscall_args *sa)
57{
58	struct proc *p;
59	int error, traced;
60
61	PCPU_INC(cnt.v_syscall);
62	p = td->td_proc;
63
64	td->td_pticks = 0;
65	if (td->td_ucred != p->p_ucred)
66		cred_update_thread(td);
67	traced = (p->p_flag & P_TRACED) != 0;
68	if (traced || td->td_dbgflags & TDB_USERWR) {
69		PROC_LOCK(p);
70		td->td_dbgflags &= ~TDB_USERWR;
71		if (traced)
72			td->td_dbgflags |= TDB_SCE;
73		PROC_UNLOCK(p);
74	}
75	error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
76#ifdef KTRACE
77	if (KTRPOINT(td, KTR_SYSCALL))
78		ktrsyscall(sa->code, sa->narg, sa->args);
79#endif
80	KTR_START4(KTR_SYSC, "syscall", syscallname(p, sa->code),
81	    (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "arg0:%p", sa->args[0],
82	    "arg1:%p", sa->args[1], "arg2:%p", sa->args[2]);
83
84	if (error == 0) {
85
86		STOPEVENT(p, S_SCE, sa->narg);
87		if (p->p_flag & P_TRACED) {
88			PROC_LOCK(p);
89			td->td_dbg_sc_code = sa->code;
90			td->td_dbg_sc_narg = sa->narg;
91			if (p->p_ptevents & PTRACE_SCE)
92				ptracestop((td), SIGTRAP);
93			PROC_UNLOCK(p);
94		}
95		if (td->td_dbgflags & TDB_USERWR) {
96			/*
97			 * Reread syscall number and arguments if
98			 * debugger modified registers or memory.
99			 */
100			error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
101			PROC_LOCK(p);
102			td->td_dbg_sc_code = sa->code;
103			td->td_dbg_sc_narg = sa->narg;
104			PROC_UNLOCK(p);
105#ifdef KTRACE
106			if (KTRPOINT(td, KTR_SYSCALL))
107				ktrsyscall(sa->code, sa->narg, sa->args);
108#endif
109			if (error != 0)
110				goto retval;
111		}
112
113#ifdef CAPABILITY_MODE
114		/*
115		 * In capability mode, we only allow access to system calls
116		 * flagged with SYF_CAPENABLED.
117		 */
118		if (IN_CAPABILITY_MODE(td) &&
119		    !(sa->callp->sy_flags & SYF_CAPENABLED)) {
120			error = ECAPMODE;
121			goto retval;
122		}
123#endif
124
125		error = syscall_thread_enter(td, sa->callp);
126		if (error != 0)
127			goto retval;
128
129#ifdef KDTRACE_HOOKS
130		/*
131		 * If the systrace module has registered it's probe
132		 * callback and if there is a probe active for the
133		 * syscall 'entry', process the probe.
134		 */
135		if (systrace_probe_func != NULL && sa->callp->sy_entry != 0)
136			(*systrace_probe_func)(sa->callp->sy_entry, sa->code,
137			    sa->callp, sa->args, 0);
138#endif
139
140		AUDIT_SYSCALL_ENTER(sa->code, td);
141		error = (sa->callp->sy_call)(td, sa->args);
142		AUDIT_SYSCALL_EXIT(error, td);
143
144		/* Save the latest error return value. */
145		if ((td->td_pflags & TDP_NERRNO) == 0)
146			td->td_errno = error;
147
148#ifdef KDTRACE_HOOKS
149		/*
150		 * If the systrace module has registered it's probe
151		 * callback and if there is a probe active for the
152		 * syscall 'return', process the probe.
153		 */
154		if (systrace_probe_func != NULL && sa->callp->sy_return != 0)
155			(*systrace_probe_func)(sa->callp->sy_return, sa->code,
156			    sa->callp, NULL, (error) ? -1 : td->td_retval[0]);
157#endif
158		syscall_thread_exit(td, sa->callp);
159	}
160 retval:
161	KTR_STOP4(KTR_SYSC, "syscall", syscallname(p, sa->code),
162	    (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "error:%d", error,
163	    "retval0:%#lx", td->td_retval[0], "retval1:%#lx",
164	    td->td_retval[1]);
165	if (traced) {
166		PROC_LOCK(p);
167		td->td_dbgflags &= ~TDB_SCE;
168		PROC_UNLOCK(p);
169	}
170	(p->p_sysent->sv_set_syscall_retval)(td, error);
171	return (error);
172}
173
174static inline void
175syscallret(struct thread *td, int error, struct syscall_args *sa __unused)
176{
177	struct proc *p, *p2;
178	int traced;
179
180	p = td->td_proc;
181
182	/*
183	 * Handle reschedule and other end-of-syscall issues
184	 */
185	userret(td, td->td_frame);
186
187#ifdef KTRACE
188	if (KTRPOINT(td, KTR_SYSRET)) {
189		ktrsysret(sa->code, (td->td_pflags & TDP_NERRNO) == 0 ?
190		    error : td->td_errno, td->td_retval[0]);
191	}
192#endif
193	td->td_pflags &= ~TDP_NERRNO;
194
195	if (p->p_flag & P_TRACED) {
196		traced = 1;
197		PROC_LOCK(p);
198		td->td_dbgflags |= TDB_SCX;
199		PROC_UNLOCK(p);
200	} else
201		traced = 0;
202	/*
203	 * This works because errno is findable through the
204	 * register set.  If we ever support an emulation where this
205	 * is not the case, this code will need to be revisited.
206	 */
207	STOPEVENT(p, S_SCX, sa->code);
208	if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) {
209		PROC_LOCK(p);
210		/*
211		 * If tracing the execed process, trap to the debugger
212		 * so that breakpoints can be set before the program
213		 * executes.  If debugger requested tracing of syscall
214		 * returns, do it now too.
215		 */
216		if (traced &&
217		    ((td->td_dbgflags & (TDB_FORK | TDB_EXEC)) != 0 ||
218		    (p->p_ptevents & PTRACE_SCX) != 0))
219			ptracestop(td, SIGTRAP);
220		td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK);
221		PROC_UNLOCK(p);
222	}
223
224	if (td->td_pflags & TDP_RFPPWAIT) {
225		/*
226		 * Preserve synchronization semantics of vfork.  If
227		 * waiting for child to exec or exit, fork set
228		 * P_PPWAIT on child, and there we sleep on our proc
229		 * (in case of exit).
230		 *
231		 * Do it after the ptracestop() above is finished, to
232		 * not block our debugger until child execs or exits
233		 * to finish vfork wait.
234		 */
235		td->td_pflags &= ~TDP_RFPPWAIT;
236		p2 = td->td_rfppwait_p;
237again:
238		PROC_LOCK(p2);
239		while (p2->p_flag & P_PPWAIT) {
240			PROC_LOCK(p);
241			if (thread_suspend_check_needed()) {
242				PROC_UNLOCK(p2);
243				thread_suspend_check(0);
244				PROC_UNLOCK(p);
245				goto again;
246			} else {
247				PROC_UNLOCK(p);
248			}
249			cv_timedwait(&p2->p_pwait, &p2->p_mtx, hz);
250		}
251		PROC_UNLOCK(p2);
252	}
253}
254