subr_syscall.c revision 110190
1/*-
2 * Copyright (C) 1994, David Greenman
3 * Copyright (c) 1990, 1993
4 *	The Regents of the University of California.  All rights reserved.
5 *
6 * This code is derived from software contributed to Berkeley by
7 * the University of Utah, and William Jolitz.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 *    notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 *    notice, this list of conditions and the following disclaimer in the
16 *    documentation and/or other materials provided with the distribution.
17 * 3. All advertising materials mentioning features or use of this software
18 *    must display the following acknowledgement:
19 *	This product includes software developed by the University of
20 *	California, Berkeley and its contributors.
21 * 4. Neither the name of the University nor the names of its contributors
22 *    may be used to endorse or promote products derived from this software
23 *    without specific prior written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * SUCH DAMAGE.
36 *
37 *	from: @(#)trap.c	7.4 (Berkeley) 5/13/91
38 * $FreeBSD: head/sys/kern/subr_trap.c 110190 2003-02-01 12:17:09Z julian $
39 */
40
41#include "opt_mac.h"
42#ifdef __i386__
43#include "opt_npx.h"
44#endif
45
46#include <sys/param.h>
47#include <sys/bus.h>
48#include <sys/kernel.h>
49#include <sys/lock.h>
50#include <sys/mac.h>
51#include <sys/mutex.h>
52#include <sys/proc.h>
53#include <sys/kse.h>
54#include <sys/ktr.h>
55#include <sys/resourcevar.h>
56#include <sys/sched.h>
57#include <sys/signalvar.h>
58#include <sys/systm.h>
59#include <sys/vmmeter.h>
60#include <machine/cpu.h>
61#include <machine/pcb.h>
62
63/*
64 * Define the code needed before returning to user mode, for
65 * trap and syscall.
66 *
67 * MPSAFE
68 */
69void
70userret(td, frame, oticks)
71	struct thread *td;
72	struct trapframe *frame;
73	u_int oticks;
74{
75	struct proc *p = td->td_proc;
76	struct kse *ke = td->td_kse;
77
78	CTR3(KTR_SYSC, "userret: thread %p (pid %d, %s)", td, p->p_pid,
79            p->p_comm);
80#ifdef INVARIANTS
81	/* Check that we called signotify() enough. */
82	mtx_lock(&Giant);
83	PROC_LOCK(p);
84	mtx_lock_spin(&sched_lock);
85	if (SIGPENDING(p) && ((p->p_sflag & PS_NEEDSIGCHK) == 0 ||
86	    (td->td_kse->ke_flags & KEF_ASTPENDING) == 0))
87		printf("failed to set signal flags properly for ast()\n");
88	mtx_unlock_spin(&sched_lock);
89	PROC_UNLOCK(p);
90	mtx_unlock(&Giant);
91#endif
92
93	/*
94	 * Let the scheduler adjust our priority etc.
95	 */
96	sched_userret(td);
97
98	/*
99	 * We need to check to see if we have to exit or wait due to a
100	 * single threading requirement or some other STOP condition.
101	 * Don't bother doing all the work if the stop bits are not set
102	 * at this time.. If we miss it, we miss it.. no big deal.
103	 */
104	if (P_SHOULDSTOP(p)) {
105		PROC_LOCK(p);
106		thread_suspend_check(0);	/* Can suspend or kill */
107		PROC_UNLOCK(p);
108	}
109
110	/*
111	 * Do special thread processing, e.g. upcall tweaking and such.
112	 */
113	if (p->p_flag & P_KSES) {
114		thread_userret(td, frame);
115	}
116
117	/*
118	 * Charge system time if profiling.
119	 *
120	 * XXX should move PS_PROFIL to a place that can obviously be
121	 * accessed safely without sched_lock.
122	 */
123	if (p->p_sflag & PS_PROFIL) {
124		quad_t ticks;
125
126		mtx_lock_spin(&sched_lock);
127		ticks = ke->ke_sticks - oticks;
128		mtx_unlock_spin(&sched_lock);
129		addupc_task(ke, TRAPF_PC(frame), (u_int)ticks * psratio);
130	}
131}
132
133/*
134 * Process an asynchronous software trap.
135 * This is relatively easy.
136 * This function will return with preemption disabled.
137 */
138void
139ast(struct trapframe *framep)
140{
141	struct thread *td;
142	struct proc *p;
143	struct kse *ke;
144	struct ksegrp *kg;
145	struct rlimit *rlim;
146	u_int prticks, sticks;
147	int sflag;
148	int flags;
149	int sig;
150#if defined(DEV_NPX) && !defined(SMP)
151	int ucode;
152#endif
153
154	td = curthread;
155	p = td->td_proc;
156	kg = td->td_ksegrp;
157
158	CTR3(KTR_SYSC, "ast: thread %p (pid %d, %s)", td, p->p_pid,
159            p->p_comm);
160	KASSERT(TRAPF_USERMODE(framep), ("ast in kernel mode"));
161#ifdef WITNESS
162	if (witness_list(td))
163		panic("Returning to user mode with mutex(s) held");
164#endif
165	mtx_assert(&Giant, MA_NOTOWNED);
166	mtx_assert(&sched_lock, MA_NOTOWNED);
167	td->td_frame = framep;
168
169	/*
170	 * This updates the p_sflag's for the checks below in one
171	 * "atomic" operation with turning off the astpending flag.
172	 * If another AST is triggered while we are handling the
173	 * AST's saved in sflag, the astpending flag will be set and
174	 * ast() will be called again.
175	 */
176	mtx_lock_spin(&sched_lock);
177	ke = td->td_kse;
178	sticks = ke->ke_sticks;
179	flags = ke->ke_flags;
180	sflag = p->p_sflag;
181	p->p_sflag &= ~(PS_ALRMPEND | PS_NEEDSIGCHK | PS_PROFPEND | PS_XCPU);
182#ifdef MAC
183	p->p_sflag &= ~PS_MACPEND;
184#endif
185	ke->ke_flags &= ~(KEF_ASTPENDING | KEF_NEEDRESCHED | KEF_OWEUPC);
186	cnt.v_soft++;
187	prticks = 0;
188	if (flags & KEF_OWEUPC && sflag & PS_PROFIL) {
189		prticks = p->p_stats->p_prof.pr_ticks;
190		p->p_stats->p_prof.pr_ticks = 0;
191	}
192	mtx_unlock_spin(&sched_lock);
193	/*
194	 * XXXKSE While the fact that we owe a user profiling
195	 * tick is stored per KSE in this code, the statistics
196	 * themselves are still stored per process.
197	 * This should probably change, by which I mean that
198	 * possibly the location of both might change.
199	 */
200
201	if (td->td_ucred != p->p_ucred)
202		cred_update_thread(td);
203	if (flags & KEF_OWEUPC && sflag & PS_PROFIL)
204		addupc_task(ke, p->p_stats->p_prof.pr_addr, prticks);
205	if (sflag & PS_ALRMPEND) {
206		PROC_LOCK(p);
207		psignal(p, SIGVTALRM);
208		PROC_UNLOCK(p);
209	}
210#if defined(DEV_NPX) && !defined(SMP)
211	if (PCPU_GET(curpcb)->pcb_flags & PCB_NPXTRAP) {
212		atomic_clear_int(&PCPU_GET(curpcb)->pcb_flags,
213		    PCB_NPXTRAP);
214		ucode = npxtrap();
215		if (ucode != -1) {
216			trapsignal(p, SIGFPE, ucode);
217		}
218	}
219#endif
220	if (sflag & PS_PROFPEND) {
221		PROC_LOCK(p);
222		psignal(p, SIGPROF);
223		PROC_UNLOCK(p);
224	}
225	if (sflag & PS_XCPU) {
226		PROC_LOCK(p);
227		rlim = &p->p_rlimit[RLIMIT_CPU];
228		if (p->p_runtime.sec >= rlim->rlim_max)
229			killproc(p, "exceeded maximum CPU limit");
230		else {
231			psignal(p, SIGXCPU);
232			mtx_lock_spin(&sched_lock);
233			if (p->p_cpulimit < rlim->rlim_max)
234				p->p_cpulimit += 5;
235			mtx_unlock_spin(&sched_lock);
236		}
237		PROC_UNLOCK(p);
238	}
239#ifdef MAC
240	if (sflag & PS_MACPEND)
241		mac_thread_userret(td);
242#endif
243	if (flags & KEF_NEEDRESCHED) {
244		mtx_lock_spin(&sched_lock);
245		sched_prio(td, kg->kg_user_pri);
246		p->p_stats->p_ru.ru_nivcsw++;
247		mi_switch();
248		mtx_unlock_spin(&sched_lock);
249	}
250	if (sflag & PS_NEEDSIGCHK) {
251		PROC_LOCK(p);
252		while ((sig = cursig(td)) != 0)
253			postsig(sig);
254		PROC_UNLOCK(p);
255	}
256
257	userret(td, framep, sticks);
258#ifdef DIAGNOSTIC
259	cred_free_thread(td);
260#endif
261	mtx_assert(&Giant, MA_NOTOWNED);
262}
263