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 44__FBSDID("$FreeBSD: stable/11/sys/kern/subr_syscall.c 304188 2016-08-15 21:10:41Z jhb $"); 45 46#include <sys/capsicum.h> 47#include <sys/ktr.h> 48#ifdef KTRACE 49#include <sys/uio.h> 50#include <sys/ktrace.h> 51#endif 52#include <security/audit/audit.h> 53 54static inline int 55syscallenter(struct thread *td, struct syscall_args *sa) 56{ 57 struct proc *p; 58 int error, traced; 59 60 PCPU_INC(cnt.v_syscall); 61 p = td->td_proc; 62 63 td->td_pticks = 0; 64 if (td->td_cowgen != p->p_cowgen) 65 thread_cow_update(td); 66 traced = (p->p_flag & P_TRACED) != 0; 67 if (traced || td->td_dbgflags & TDB_USERWR) { 68 PROC_LOCK(p); 69 td->td_dbgflags &= ~TDB_USERWR; 70 if (traced) 71 td->td_dbgflags |= TDB_SCE; 72 PROC_UNLOCK(p); 73 } 74 error = (p->p_sysent->sv_fetch_syscall_args)(td, sa); 75#ifdef KTRACE 76 if (KTRPOINT(td, KTR_SYSCALL)) 77 ktrsyscall(sa->code, sa->narg, sa->args); 78#endif 79 KTR_START4(KTR_SYSC, "syscall", syscallname(p, sa->code), 80 (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "arg0:%p", sa->args[0], 81 "arg1:%p", sa->args[1], "arg2:%p", sa->args[2]); 82 83 if (error == 0) { 84 85 STOPEVENT(p, S_SCE, sa->narg); 86 if (p->p_flag & P_TRACED) { 87 PROC_LOCK(p); 88 td->td_dbg_sc_code = sa->code; 89 td->td_dbg_sc_narg = sa->narg; 90 if (p->p_ptevents & PTRACE_SCE) 91 ptracestop((td), SIGTRAP); 92 PROC_UNLOCK(p); 93 } 94 if (td->td_dbgflags & TDB_USERWR) { 95 /* 96 * Reread syscall number and arguments if 97 * debugger modified registers or memory. 98 */ 99 error = (p->p_sysent->sv_fetch_syscall_args)(td, sa); 100 PROC_LOCK(p); 101 td->td_dbg_sc_code = sa->code; 102 td->td_dbg_sc_narg = sa->narg; 103 PROC_UNLOCK(p); 104#ifdef KTRACE 105 if (KTRPOINT(td, KTR_SYSCALL)) 106 ktrsyscall(sa->code, sa->narg, sa->args); 107#endif 108 if (error != 0) 109 goto retval; 110 } 111 112#ifdef CAPABILITY_MODE 113 /* 114 * In capability mode, we only allow access to system calls 115 * flagged with SYF_CAPENABLED. 116 */ 117 if (IN_CAPABILITY_MODE(td) && 118 !(sa->callp->sy_flags & SYF_CAPENABLED)) { 119 error = ECAPMODE; 120 goto retval; 121 } 122#endif 123 124 error = syscall_thread_enter(td, sa->callp); 125 if (error != 0) 126 goto retval; 127 128#ifdef KDTRACE_HOOKS 129 /* Give the syscall:::entry DTrace probe a chance to fire. */ 130 if (systrace_probe_func != NULL && sa->callp->sy_entry != 0) 131 (*systrace_probe_func)(sa, SYSTRACE_ENTRY, 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 /* Give the syscall:::return DTrace probe a chance to fire. */ 144 if (systrace_probe_func != NULL && sa->callp->sy_return != 0) 145 (*systrace_probe_func)(sa, SYSTRACE_RETURN, 146 error ? -1 : td->td_retval[0]); 147#endif 148 syscall_thread_exit(td, sa->callp); 149 } 150 retval: 151 KTR_STOP4(KTR_SYSC, "syscall", syscallname(p, sa->code), 152 (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "error:%d", error, 153 "retval0:%#lx", td->td_retval[0], "retval1:%#lx", 154 td->td_retval[1]); 155 if (traced) { 156 PROC_LOCK(p); 157 td->td_dbgflags &= ~TDB_SCE; 158 PROC_UNLOCK(p); 159 } 160 (p->p_sysent->sv_set_syscall_retval)(td, error); 161 return (error); 162} 163 164static inline void 165syscallret(struct thread *td, int error, struct syscall_args *sa) 166{ 167 struct proc *p, *p2; 168 int traced; 169 170 KASSERT((td->td_pflags & TDP_FORKING) == 0, 171 ("fork() did not clear TDP_FORKING upon completion")); 172 173 p = td->td_proc; 174 175 /* 176 * Handle reschedule and other end-of-syscall issues 177 */ 178 userret(td, td->td_frame); 179 180#ifdef KTRACE 181 if (KTRPOINT(td, KTR_SYSRET)) { 182 ktrsysret(sa->code, (td->td_pflags & TDP_NERRNO) == 0 ? 183 error : td->td_errno, td->td_retval[0]); 184 } 185#endif 186 td->td_pflags &= ~TDP_NERRNO; 187 188 if (p->p_flag & P_TRACED) { 189 traced = 1; 190 PROC_LOCK(p); 191 td->td_dbgflags |= TDB_SCX; 192 PROC_UNLOCK(p); 193 } else 194 traced = 0; 195 /* 196 * This works because errno is findable through the 197 * register set. If we ever support an emulation where this 198 * is not the case, this code will need to be revisited. 199 */ 200 STOPEVENT(p, S_SCX, sa->code); 201 if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) { 202 PROC_LOCK(p); 203 /* 204 * If tracing the execed process, trap to the debugger 205 * so that breakpoints can be set before the program 206 * executes. If debugger requested tracing of syscall 207 * returns, do it now too. 208 */ 209 if (traced && 210 ((td->td_dbgflags & (TDB_FORK | TDB_EXEC)) != 0 || 211 (p->p_ptevents & PTRACE_SCX) != 0)) 212 ptracestop(td, SIGTRAP); 213 td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK); 214 PROC_UNLOCK(p); 215 } 216 217 if (td->td_pflags & TDP_RFPPWAIT) { 218 /* 219 * Preserve synchronization semantics of vfork. If 220 * waiting for child to exec or exit, fork set 221 * P_PPWAIT on child, and there we sleep on our proc 222 * (in case of exit). 223 * 224 * Do it after the ptracestop() above is finished, to 225 * not block our debugger until child execs or exits 226 * to finish vfork wait. 227 */ 228 td->td_pflags &= ~TDP_RFPPWAIT; 229 p2 = td->td_rfppwait_p; 230again: 231 PROC_LOCK(p2); 232 while (p2->p_flag & P_PPWAIT) { 233 PROC_LOCK(p); 234 if (thread_suspend_check_needed()) { 235 PROC_UNLOCK(p2); 236 thread_suspend_check(0); 237 PROC_UNLOCK(p); 238 goto again; 239 } else { 240 PROC_UNLOCK(p); 241 } 242 cv_timedwait(&p2->p_pwait, &p2->p_mtx, hz); 243 } 244 PROC_UNLOCK(p2); 245 } 246} 247