kern_syscall.c revision 1.4.6.2
1 /* $NetBSD: kern_syscall.c,v 1.4.6.2 2010/08/11 22:54:41 yamt Exp $ */ 2 3 /*- 4 * Copyright (c) 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software developed for The NetBSD Foundation 8 * by Andrew Doran. 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 * 19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 * POSSIBILITY OF SUCH DAMAGE. 30 */ 31 32 #include <sys/cdefs.h> 33 __KERNEL_RCSID(0, "$NetBSD: kern_syscall.c,v 1.4.6.2 2010/08/11 22:54:41 yamt Exp $"); 34 35 #include "opt_modular.h" 36 37 /* XXX To get syscall prototypes. */ 38 #define SYSVSHM 39 #define SYSVSEM 40 #define SYSVMSG 41 42 #include <sys/param.h> 43 #include <sys/module.h> 44 #include <sys/sched.h> 45 #include <sys/syscall.h> 46 #include <sys/syscallargs.h> 47 #include <sys/syscallvar.h> 48 #include <sys/xcall.h> 49 50 int 51 sys_nomodule(struct lwp *l, const void *v, register_t *retval) 52 { 53 #ifdef MODULAR 54 static struct { 55 u_int al_code; 56 const char *al_module; 57 } const autoload[] = { 58 { SYS_aio_cancel, "aio" }, 59 { SYS_aio_error, "aio" }, 60 { SYS_aio_fsync, "aio" }, 61 { SYS_aio_read, "aio" }, 62 { SYS_aio_return, "aio" }, 63 { SYS___aio_suspend50, "aio" }, 64 { SYS_aio_write, "aio" }, 65 { SYS_lio_listio, "aio" }, 66 { SYS_mq_open, "mqueue" }, 67 { SYS_mq_close, "mqueue" }, 68 { SYS_mq_unlink, "mqueue" }, 69 { SYS_mq_getattr, "mqueue" }, 70 { SYS_mq_setattr, "mqueue" }, 71 { SYS_mq_notify, "mqueue" }, 72 { SYS_mq_send, "mqueue" }, 73 { SYS_mq_receive, "mqueue" }, 74 { SYS___mq_timedsend50, "mqueue" }, 75 { SYS___mq_timedreceive50, "mqueue" }, 76 { SYS_compat_43_fstat43, "compat" }, 77 { SYS_compat_43_lstat43, "compat" }, 78 { SYS_compat_43_oaccept, "compat" }, 79 { SYS_compat_43_ocreat, "compat" }, 80 { SYS_compat_43_oftruncate, "compat" }, 81 { SYS_compat_43_ogetdirentries, "compat" }, 82 { SYS_compat_43_ogetdtablesize, "compat" }, 83 { SYS_compat_43_ogethostid, "compat" }, 84 { SYS_compat_43_ogethostname, "compat" }, 85 { SYS_compat_43_ogetkerninfo, "compat" }, 86 { SYS_compat_43_ogetpagesize, "compat" }, 87 { SYS_compat_43_ogetpeername, "compat" }, 88 { SYS_compat_43_ogetrlimit, "compat" }, 89 { SYS_compat_43_ogetsockname, "compat" }, 90 { SYS_compat_43_okillpg, "compat" }, 91 { SYS_compat_43_olseek, "compat" }, 92 { SYS_compat_43_ommap, "compat" }, 93 { SYS_compat_43_oquota, "compat" }, 94 { SYS_compat_43_orecv, "compat" }, 95 { SYS_compat_43_orecvfrom, "compat" }, 96 { SYS_compat_43_orecvmsg, "compat" }, 97 { SYS_compat_43_osend, "compat" }, 98 { SYS_compat_43_osendmsg, "compat" }, 99 { SYS_compat_43_osethostid, "compat" }, 100 { SYS_compat_43_osethostname, "compat" }, 101 { SYS_compat_43_osetrlimit, "compat" }, 102 { SYS_compat_43_osigblock, "compat" }, 103 { SYS_compat_43_osigsetmask, "compat" }, 104 { SYS_compat_43_osigstack, "compat" }, 105 { SYS_compat_43_osigvec, "compat" }, 106 { SYS_compat_43_otruncate, "compat" }, 107 { SYS_compat_43_owait, "compat" }, 108 { SYS_compat_43_stat43, "compat" }, 109 { SYS_compat_09_ogetdomainname, "compat" }, 110 { SYS_compat_09_osetdomainname, "compat" }, 111 { SYS_compat_09_ouname, "compat" }, 112 #ifndef _LP64 113 { SYS_compat_10_omsgsys, "compat" }, 114 { SYS_compat_10_osemsys, "compat" }, 115 { SYS_compat_10_oshmsys, "compat" }, 116 #endif 117 { SYS_compat_12_fstat12, "compat" }, 118 { SYS_compat_12_getdirentries, "compat" }, 119 { SYS_compat_12_lstat12, "compat" }, 120 { SYS_compat_12_msync, "compat" }, 121 { SYS_compat_12_oreboot, "compat" }, 122 { SYS_compat_12_oswapon, "compat" }, 123 { SYS_compat_12_stat12, "compat" }, 124 { SYS_compat_13_sigaction13, "compat" }, 125 { SYS_compat_13_sigaltstack13, "compat" }, 126 { SYS_compat_13_sigpending13, "compat" }, 127 { SYS_compat_13_sigprocmask13, "compat" }, 128 { SYS_compat_13_sigreturn13, "compat" }, 129 { SYS_compat_13_sigsuspend13, "compat" }, 130 { SYS_compat_14___semctl, "compat" }, 131 { SYS_compat_14_msgctl, "compat" }, 132 { SYS_compat_14_shmctl, "compat" }, 133 { SYS_compat_16___sigaction14, "compat" }, 134 { SYS_compat_16___sigreturn14, "compat" }, 135 { SYS_compat_20_fhstatfs, "compat" }, 136 { SYS_compat_20_fstatfs, "compat" }, 137 { SYS_compat_20_getfsstat, "compat" }, 138 { SYS_compat_20_statfs, "compat" }, 139 { SYS_compat_30___fhstat30, "compat" }, 140 { SYS_compat_30___fstat13, "compat" }, 141 { SYS_compat_30___lstat13, "compat" }, 142 { SYS_compat_30___stat13, "compat" }, 143 { SYS_compat_30_fhopen, "compat" }, 144 { SYS_compat_30_fhstat, "compat" }, 145 { SYS_compat_30_fhstatvfs1, "compat" }, 146 { SYS_compat_30_getdents, "compat" }, 147 { SYS_compat_30_getfh, "compat" }, 148 { SYS_compat_30_socket, "compat" }, 149 { SYS_compat_40_mount, "compat" }, 150 { SYS_compat_50_wait4, "compat" }, 151 { SYS_compat_50_mknod, "compat" }, 152 { SYS_compat_50_setitimer, "compat" }, 153 { SYS_compat_50_getitimer, "compat" }, 154 { SYS_compat_50_select, "compat" }, 155 { SYS_compat_50_gettimeofday, "compat" }, 156 { SYS_compat_50_getrusage, "compat" }, 157 { SYS_compat_50_settimeofday, "compat" }, 158 { SYS_compat_50_utimes, "compat" }, 159 { SYS_compat_50_adjtime, "compat" }, 160 { SYS_compat_50_lfs_segwait, "compat" }, 161 { SYS_compat_50_futimes, "compat" }, 162 { SYS_compat_50_clock_gettime, "compat" }, 163 { SYS_compat_50_clock_settime, "compat" }, 164 { SYS_compat_50_clock_getres, "compat" }, 165 { SYS_compat_50_timer_settime, "compat" }, 166 { SYS_compat_50_timer_gettime, "compat" }, 167 { SYS_compat_50_nanosleep, "compat" }, 168 { SYS_compat_50___sigtimedwait, "compat" }, 169 { SYS_compat_50_mq_timedsend, "compat" }, 170 { SYS_compat_50_mq_timedreceive, "compat" }, 171 { SYS_compat_50_lutimes, "compat" }, 172 { SYS_compat_50_____semctl13, "compat" }, 173 { SYS_compat_50___msgctl13, "compat" }, 174 { SYS_compat_50___shmctl13, "compat" }, 175 { SYS_compat_50__lwp_park, "compat" }, 176 { SYS_compat_50_kevent, "compat" }, 177 { SYS_compat_50_pselect, "compat" }, 178 { SYS_compat_50_pollts, "compat" }, 179 { SYS_compat_50___stat30, "compat" }, 180 { SYS_compat_50___fstat30, "compat" }, 181 { SYS_compat_50___lstat30, "compat" }, 182 { SYS_compat_50___ntp_gettime30, "compat" }, 183 { SYS_compat_50___fhstat40, "compat" }, 184 { SYS_compat_50_aio_suspend, "compat" }, 185 { SYS__ksem_init, "ksem" }, 186 { SYS__ksem_open, "ksem" }, 187 { SYS__ksem_unlink, "ksem" }, 188 { SYS__ksem_close, "ksem" }, 189 { SYS__ksem_post, "ksem" }, 190 { SYS__ksem_wait, "ksem" }, 191 { SYS__ksem_trywait, "ksem" }, 192 { SYS__ksem_getvalue, "ksem" }, 193 { SYS__ksem_destroy, "ksem" }, 194 { SYS_nfssvc, "nfsserver" }, 195 }; 196 const struct sysent *sy; 197 const struct emul *em; 198 int code, i; 199 200 /* 201 * Restart the syscall if we interrupted a module unload that 202 * failed. Acquiring module_lock delays us until any unload 203 * has been completed or rolled back. 204 */ 205 mutex_enter(&module_lock); 206 sy = l->l_sysent; 207 if (sy->sy_call != sys_nomodule) { 208 mutex_exit(&module_lock); 209 return ERESTART; 210 } 211 /* 212 * Try to autoload a module to satisfy the request. If it 213 * works, retry the request. 214 */ 215 em = l->l_proc->p_emul; 216 if (em == &emul_netbsd) { 217 code = sy - em->e_sysent; 218 for (i = 0; i < __arraycount(autoload); i++) { 219 if (autoload[i].al_code != code) { 220 continue; 221 } 222 if (module_autoload(autoload[i].al_module, 223 MODULE_CLASS_ANY) != 0 || 224 sy->sy_call == sys_nomodule) { 225 break; 226 } 227 mutex_exit(&module_lock); 228 return ERESTART; 229 } 230 } 231 mutex_exit(&module_lock); 232 #endif /* MODULAR */ 233 234 return sys_nosys(l, v, retval); 235 } 236 237 int 238 syscall_establish(const struct emul *em, const struct syscall_package *sp) 239 { 240 struct sysent *sy; 241 int i; 242 243 KASSERT(mutex_owned(&module_lock)); 244 245 if (em == NULL) { 246 em = &emul_netbsd; 247 } 248 sy = em->e_sysent; 249 250 /* 251 * Ensure that all preconditions are valid, since this is 252 * an all or nothing deal. Once a system call is entered, 253 * it can become busy and we could be unable to remove it 254 * on error. 255 */ 256 for (i = 0; sp[i].sp_call != NULL; i++) { 257 if (sy[sp[i].sp_code].sy_call != sys_nomodule) { 258 #ifdef DIAGNOSTIC 259 printf("syscall %d is busy\n", sp[i].sp_code); 260 #endif 261 return EBUSY; 262 } 263 } 264 /* Everything looks good, patch them in. */ 265 for (i = 0; sp[i].sp_call != NULL; i++) { 266 sy[sp[i].sp_code].sy_call = sp[i].sp_call; 267 } 268 269 return 0; 270 } 271 272 int 273 syscall_disestablish(const struct emul *em, const struct syscall_package *sp) 274 { 275 struct sysent *sy; 276 uint64_t where; 277 lwp_t *l; 278 int i; 279 280 KASSERT(mutex_owned(&module_lock)); 281 282 if (em == NULL) { 283 em = &emul_netbsd; 284 } 285 sy = em->e_sysent; 286 287 /* 288 * First, patch the system calls to sys_nomodule to gate further 289 * activity. 290 */ 291 for (i = 0; sp[i].sp_call != NULL; i++) { 292 KASSERT(sy[sp[i].sp_code].sy_call == sp[i].sp_call); 293 sy[sp[i].sp_code].sy_call = sys_nomodule; 294 } 295 296 /* 297 * Run a cross call to cycle through all CPUs. This does two 298 * things: lock activity provides a barrier and makes our update 299 * of sy_call visible to all CPUs, and upon return we can be sure 300 * that we see pertinent values of l_sysent posted by remote CPUs. 301 */ 302 where = xc_broadcast(0, (xcfunc_t)nullop, NULL, NULL); 303 xc_wait(where); 304 305 /* 306 * Now it's safe to check l_sysent. Run through all LWPs and see 307 * if anyone is still using the system call. 308 */ 309 for (i = 0; sp[i].sp_call != NULL; i++) { 310 mutex_enter(proc_lock); 311 LIST_FOREACH(l, &alllwp, l_list) { 312 if (l->l_sysent == &sy[sp[i].sp_code]) { 313 break; 314 } 315 } 316 mutex_exit(proc_lock); 317 if (l == NULL) { 318 continue; 319 } 320 /* 321 * We lose: one or more calls are still in use. Put back 322 * the old entrypoints and act like nothing happened. 323 * When we drop module_lock, any system calls held in 324 * sys_nomodule() will be restarted. 325 */ 326 for (i = 0; sp[i].sp_call != NULL; i++) { 327 sy[sp[i].sp_code].sy_call = sp[i].sp_call; 328 } 329 return EBUSY; 330 } 331 332 return 0; 333 } 334
Indexes created Sun Sep 28 16:09:52 GMT 2025