subr_syscall.c revision 240246
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: head/sys/kern/subr_syscall.c 240246 2012-09-08 18:35:15Z attilio $");
46
47#include <sys/capability.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	if (p->p_flag & P_TRACED) {
68		traced = 1;
69		PROC_LOCK(p);
70		td->td_dbgflags &= ~TDB_USERWR;
71		td->td_dbgflags |= TDB_SCE;
72		PROC_UNLOCK(p);
73	} else
74		traced = 0;
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
81	CTR6(KTR_SYSC,
82"syscall: td=%p pid %d %s (%#lx, %#lx, %#lx)",
83	    td, td->td_proc->p_pid, syscallname(p, sa->code),
84	    sa->args[0], sa->args[1], sa->args[2]);
85
86	if (error == 0) {
87		STOPEVENT(p, S_SCE, sa->narg);
88		if (p->p_flag & P_TRACED && p->p_stops & S_PT_SCE) {
89			PROC_LOCK(p);
90			ptracestop((td), SIGTRAP);
91			PROC_UNLOCK(p);
92		}
93		if (td->td_dbgflags & TDB_USERWR) {
94			/*
95			 * Reread syscall number and arguments if
96			 * debugger modified registers or memory.
97			 */
98			error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
99#ifdef KTRACE
100			if (KTRPOINT(td, KTR_SYSCALL))
101				ktrsyscall(sa->code, sa->narg, sa->args);
102#endif
103			if (error != 0)
104				goto retval;
105		}
106
107#ifdef CAPABILITY_MODE
108		/*
109		 * In capability mode, we only allow access to system calls
110		 * flagged with SYF_CAPENABLED.
111		 */
112		if (IN_CAPABILITY_MODE(td) &&
113		    !(sa->callp->sy_flags & SYF_CAPENABLED)) {
114			error = ECAPMODE;
115			goto retval;
116		}
117#endif
118
119		error = syscall_thread_enter(td, sa->callp);
120		if (error != 0)
121			goto retval;
122
123#ifdef KDTRACE_HOOKS
124		/*
125		 * If the systrace module has registered it's probe
126		 * callback and if there is a probe active for the
127		 * syscall 'entry', process the probe.
128		 */
129		if (systrace_probe_func != NULL && sa->callp->sy_entry != 0)
130			(*systrace_probe_func)(sa->callp->sy_entry, sa->code,
131			    sa->callp, sa->args, 0);
132#endif
133
134		AUDIT_SYSCALL_ENTER(sa->code, td);
135		error = (sa->callp->sy_call)(td, sa->args);
136		AUDIT_SYSCALL_EXIT(error, td);
137
138		/* Save the latest error return value. */
139		if ((td->td_pflags & TDP_NERRNO) == 0)
140			td->td_errno = error;
141
142#ifdef KDTRACE_HOOKS
143		/*
144		 * If the systrace module has registered it's probe
145		 * callback and if there is a probe active for the
146		 * syscall 'return', process the probe.
147		 */
148		if (systrace_probe_func != NULL && sa->callp->sy_return != 0)
149			(*systrace_probe_func)(sa->callp->sy_return, sa->code,
150			    sa->callp, NULL, (error) ? -1 : td->td_retval[0]);
151#endif
152		syscall_thread_exit(td, sa->callp);
153		CTR4(KTR_SYSC, "syscall: p=%p error=%d return %#lx %#lx",
154		    p, error, td->td_retval[0], td->td_retval[1]);
155	}
156 retval:
157	if (traced) {
158		PROC_LOCK(p);
159		td->td_dbgflags &= ~TDB_SCE;
160		PROC_UNLOCK(p);
161	}
162	(p->p_sysent->sv_set_syscall_retval)(td, error);
163	return (error);
164}
165
166static inline void
167syscallret(struct thread *td, int error, struct syscall_args *sa __unused)
168{
169	struct proc *p, *p2;
170	int traced;
171
172	p = td->td_proc;
173
174	/*
175	 * Handle reschedule and other end-of-syscall issues
176	 */
177	userret(td, td->td_frame);
178
179	CTR4(KTR_SYSC, "syscall %s exit thread %p pid %d proc %s",
180	    syscallname(p, sa->code), td, td->td_proc->p_pid, td->td_name);
181
182#ifdef KTRACE
183	if (KTRPOINT(td, KTR_SYSRET)) {
184		ktrsysret(sa->code, (td->td_pflags & TDP_NERRNO) == 0 ?
185		    error : td->td_errno, td->td_retval[0]);
186	}
187#endif
188	td->td_pflags &= ~TDP_NERRNO;
189
190	if (p->p_flag & P_TRACED) {
191		traced = 1;
192		PROC_LOCK(p);
193		td->td_dbgflags |= TDB_SCX;
194		PROC_UNLOCK(p);
195	} else
196		traced = 0;
197	/*
198	 * This works because errno is findable through the
199	 * register set.  If we ever support an emulation where this
200	 * is not the case, this code will need to be revisited.
201	 */
202	STOPEVENT(p, S_SCX, sa->code);
203	if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) {
204		PROC_LOCK(p);
205		/*
206		 * If tracing the execed process, trap to the debugger
207		 * so that breakpoints can be set before the program
208		 * executes.  If debugger requested tracing of syscall
209		 * returns, do it now too.
210		 */
211		if (traced &&
212		    ((td->td_dbgflags & (TDB_FORK | TDB_EXEC)) != 0 ||
213		    (p->p_stops & S_PT_SCX) != 0))
214			ptracestop(td, SIGTRAP);
215		td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK);
216		PROC_UNLOCK(p);
217	}
218
219	if (td->td_pflags & TDP_RFPPWAIT) {
220		/*
221		 * Preserve synchronization semantics of vfork.  If
222		 * waiting for child to exec or exit, fork set
223		 * P_PPWAIT on child, and there we sleep on our proc
224		 * (in case of exit).
225		 *
226		 * Do it after the ptracestop() above is finished, to
227		 * not block our debugger until child execs or exits
228		 * to finish vfork wait.
229		 */
230		td->td_pflags &= ~TDP_RFPPWAIT;
231		p2 = td->td_rfppwait_p;
232		PROC_LOCK(p2);
233		while (p2->p_flag & P_PPWAIT)
234			cv_wait(&p2->p_pwait, &p2->p_mtx);
235		PROC_UNLOCK(p2);
236	}
237}
238