1 /* $NetBSD: linux_misc.c,v 1.64 2000/02/17 12:34:26 abs Exp $ */ 2 3 /*- 4 * Copyright (c) 1995, 1998, 1999 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Frank van der Linden and Eric Haszlakiewicz; by Jason R. Thorpe 9 * of the Numerical Aerospace Simulation Facility, NASA Ames Research Center. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. All advertising materials mentioning features or use of this software 20 * must display the following acknowledgement: 21 * This product includes software developed by the NetBSD 22 * Foundation, Inc. and its contributors. 23 * 4. Neither the name of The NetBSD Foundation nor the names of its 24 * contributors may be used to endorse or promote products derived 25 * from this software without specific prior written permission. 26 * 27 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 28 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 29 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 30 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 31 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 32 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 33 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 34 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 35 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 36 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 37 * POSSIBILITY OF SUCH DAMAGE. 38 */ 39 40 /* 41 * Linux compatibility module. Try to deal with various Linux system calls. 42 */ 43 44 /* 45 * These functions have been moved to multiarch to allow 46 * selection of which machines include them to be 47 * determined by the individual files.linux_<arch> files. 48 * 49 * Function in multiarch: 50 * linux_sys_break : linux_break.c 51 * linux_sys_alarm : linux_misc_notalpha.c 52 * linux_sys_getresgid : linux_misc_notalpha.c 53 * linux_sys_nice : linux_misc_notalpha.c 54 * linux_sys_readdir : linux_misc_notalpha.c 55 * linux_sys_setresgid : linux_misc_notalpha.c 56 * linux_sys_time : linux_misc_notalpha.c 57 * linux_sys_utime : linux_misc_notalpha.c 58 * linux_sys_waitpid : linux_misc_notalpha.c 59 * linux_sys_old_mmap : linux_oldmmap.c 60 * linux_sys_oldolduname : linux_oldolduname.c 61 * linux_sys_oldselect : linux_oldselect.c 62 * linux_sys_olduname : linux_olduname.c 63 * linux_sys_pipe : linux_pipe.c 64 */ 65 66 #include <sys/param.h> 67 #include <sys/systm.h> 68 #include <sys/namei.h> 69 #include <sys/proc.h> 70 #include <sys/dirent.h> 71 #include <sys/file.h> 72 #include <sys/stat.h> 73 #include <sys/filedesc.h> 74 #include <sys/ioctl.h> 75 #include <sys/kernel.h> 76 #include <sys/malloc.h> 77 #include <sys/mbuf.h> 78 #include <sys/mman.h> 79 #include <sys/mount.h> 80 #include <sys/ptrace.h> 81 #include <sys/resource.h> 82 #include <sys/resourcevar.h> 83 #include <sys/signal.h> 84 #include <sys/signalvar.h> 85 #include <sys/socket.h> 86 #include <sys/time.h> 87 #include <sys/times.h> 88 #include <sys/vnode.h> 89 #include <sys/uio.h> 90 #include <sys/wait.h> 91 #include <sys/utsname.h> 92 #include <sys/unistd.h> 93 94 #include <sys/syscallargs.h> 95 96 #include <vm/vm.h> 97 #include <vm/vm_param.h> 98 99 #include <compat/linux/common/linux_types.h> 100 #include <compat/linux/common/linux_signal.h> 101 102 #include <compat/linux/linux_syscallargs.h> 103 104 #include <compat/linux/common/linux_fcntl.h> 105 #include <compat/linux/common/linux_mmap.h> 106 #include <compat/linux/common/linux_dirent.h> 107 #include <compat/linux/common/linux_util.h> 108 #include <compat/linux/common/linux_misc.h> 109 #include <compat/linux/common/linux_ptrace.h> 110 111 int linux_ptrace_request_map[] = { 112 LINUX_PTRACE_TRACEME, PT_TRACE_ME, 113 LINUX_PTRACE_PEEKTEXT, PT_READ_I, 114 LINUX_PTRACE_PEEKDATA, PT_READ_D, 115 LINUX_PTRACE_POKETEXT, PT_WRITE_I, 116 LINUX_PTRACE_POKEDATA, PT_WRITE_D, 117 LINUX_PTRACE_CONT, PT_CONTINUE, 118 LINUX_PTRACE_KILL, PT_KILL, 119 LINUX_PTRACE_ATTACH, PT_ATTACH, 120 LINUX_PTRACE_DETACH, PT_DETACH, 121 -1 122 }; 123 124 /* Local linux_misc.c functions: */ 125 static void bsd_to_linux_statfs __P((struct statfs *, struct linux_statfs *)); 126 127 /* 128 * The information on a terminated (or stopped) process needs 129 * to be converted in order for Linux binaries to get a valid signal 130 * number out of it. 131 */ 132 void 133 bsd_to_linux_wstat(st) 134 int *st; 135 { 136 137 int sig; 138 139 if (WIFSIGNALED(*st)) { 140 sig = WTERMSIG(*st); 141 if (sig >= 0 && sig < NSIG) 142 *st= (*st& ~0177) | native_to_linux_sig[sig]; 143 } else if (WIFSTOPPED(*st)) { 144 sig = WSTOPSIG(*st); 145 if (sig >= 0 && sig < NSIG) 146 *st = (*st & ~0xff00) | (native_to_linux_sig[sig] << 8); 147 } 148 } 149 150 /* 151 * This is very much the same as waitpid() 152 */ 153 int 154 linux_sys_wait4(p, v, retval) 155 struct proc *p; 156 void *v; 157 register_t *retval; 158 { 159 struct linux_sys_wait4_args /* { 160 syscallarg(int) pid; 161 syscallarg(int *) status; 162 syscallarg(int) options; 163 syscallarg(struct rusage *) rusage; 164 } */ *uap = v; 165 struct sys_wait4_args w4a; 166 int error, *status, tstat, options, linux_options; 167 caddr_t sg; 168 169 if (SCARG(uap, status) != NULL) { 170 sg = stackgap_init(p->p_emul); 171 status = (int *) stackgap_alloc(&sg, sizeof *status); 172 } else 173 status = NULL; 174 175 linux_options = SCARG(uap, options); 176 options = 0; 177 if (linux_options & 178 ~(LINUX_WAIT4_WNOHANG|LINUX_WAIT4_WUNTRACED|LINUX_WAIT4_WCLONE)) 179 return (EINVAL); 180 181 if (linux_options & LINUX_WAIT4_WNOHANG) 182 options |= WNOHANG; 183 if (linux_options & LINUX_WAIT4_WUNTRACED) 184 options |= WUNTRACED; 185 if (linux_options & LINUX_WAIT4_WCLONE) 186 options |= WALTSIG; 187 188 SCARG(&w4a, pid) = SCARG(uap, pid); 189 SCARG(&w4a, status) = status; 190 SCARG(&w4a, options) = options; 191 SCARG(&w4a, rusage) = SCARG(uap, rusage); 192 193 if ((error = sys_wait4(p, &w4a, retval))) 194 return error; 195 196 sigdelset(&p->p_siglist, SIGCHLD); 197 198 if (status != NULL) { 199 if ((error = copyin(status, &tstat, sizeof tstat))) 200 return error; 201 202 bsd_to_linux_wstat(&tstat); 203 return copyout(&tstat, SCARG(uap, status), sizeof tstat); 204 } 205 206 return 0; 207 } 208 209 /* 210 * Linux brk(2). The check if the new address is >= the old one is 211 * done in the kernel in Linux. NetBSD does it in the library. 212 */ 213 int 214 linux_sys_brk(p, v, retval) 215 struct proc *p; 216 void *v; 217 register_t *retval; 218 { 219 struct linux_sys_brk_args /* { 220 syscallarg(char *) nsize; 221 } */ *uap = v; 222 char *nbrk = SCARG(uap, nsize); 223 struct sys_obreak_args oba; 224 struct vmspace *vm = p->p_vmspace; 225 caddr_t oldbrk; 226 227 oldbrk = vm->vm_daddr + ctob(vm->vm_dsize); 228 /* 229 * XXX inconsistent.. Linux always returns at least the old 230 * brk value, but it will be page-aligned if this fails, 231 * and possibly not page aligned if it succeeds (the user 232 * supplied pointer is returned). 233 */ 234 SCARG(&oba, nsize) = nbrk; 235 236 if ((caddr_t) nbrk > vm->vm_daddr && sys_obreak(p, &oba, retval) == 0) 237 retval[0] = (register_t)nbrk; 238 else 239 retval[0] = (register_t)oldbrk; 240 241 return 0; 242 } 243 244 /* 245 * Convert BSD statfs structure to Linux statfs structure. 246 * The Linux structure has less fields, and it also wants 247 * the length of a name in a dir entry in a field, which 248 * we fake (probably the wrong way). 249 */ 250 static void 251 bsd_to_linux_statfs(bsp, lsp) 252 struct statfs *bsp; 253 struct linux_statfs *lsp; 254 { 255 256 lsp->l_ftype = bsp->f_type; 257 lsp->l_fbsize = bsp->f_bsize; 258 lsp->l_fblocks = bsp->f_blocks; 259 lsp->l_fbfree = bsp->f_bfree; 260 lsp->l_fbavail = bsp->f_bavail; 261 lsp->l_ffiles = bsp->f_files; 262 lsp->l_fffree = bsp->f_ffree; 263 lsp->l_ffsid.val[0] = bsp->f_fsid.val[0]; 264 lsp->l_ffsid.val[1] = bsp->f_fsid.val[1]; 265 lsp->l_fnamelen = MAXNAMLEN; /* XXX */ 266 } 267 268 /* 269 * Implement the fs stat functions. Straightforward. 270 */ 271 int 272 linux_sys_statfs(p, v, retval) 273 struct proc *p; 274 void *v; 275 register_t *retval; 276 { 277 struct linux_sys_statfs_args /* { 278 syscallarg(const char *) path; 279 syscallarg(struct linux_statfs *) sp; 280 } */ *uap = v; 281 struct statfs btmp, *bsp; 282 struct linux_statfs ltmp; 283 struct sys_statfs_args bsa; 284 caddr_t sg; 285 int error; 286 287 sg = stackgap_init(p->p_emul); 288 bsp = (struct statfs *) stackgap_alloc(&sg, sizeof (struct statfs)); 289 290 LINUX_CHECK_ALT_EXIST(p, &sg, SCARG(uap, path)); 291 292 SCARG(&bsa, path) = SCARG(uap, path); 293 SCARG(&bsa, buf) = bsp; 294 295 if ((error = sys_statfs(p, &bsa, retval))) 296 return error; 297 298 if ((error = copyin((caddr_t) bsp, (caddr_t) &btmp, sizeof btmp))) 299 return error; 300 301 bsd_to_linux_statfs(&btmp, <mp); 302 303 return copyout((caddr_t) <mp, (caddr_t) SCARG(uap, sp), sizeof ltmp); 304 } 305 306 int 307 linux_sys_fstatfs(p, v, retval) 308 struct proc *p; 309 void *v; 310 register_t *retval; 311 { 312 struct linux_sys_fstatfs_args /* { 313 syscallarg(int) fd; 314 syscallarg(struct linux_statfs *) sp; 315 } */ *uap = v; 316 struct statfs btmp, *bsp; 317 struct linux_statfs ltmp; 318 struct sys_fstatfs_args bsa; 319 caddr_t sg; 320 int error; 321 322 sg = stackgap_init(p->p_emul); 323 bsp = (struct statfs *) stackgap_alloc(&sg, sizeof (struct statfs)); 324 325 SCARG(&bsa, fd) = SCARG(uap, fd); 326 SCARG(&bsa, buf) = bsp; 327 328 if ((error = sys_fstatfs(p, &bsa, retval))) 329 return error; 330 331 if ((error = copyin((caddr_t) bsp, (caddr_t) &btmp, sizeof btmp))) 332 return error; 333 334 bsd_to_linux_statfs(&btmp, <mp); 335 336 return copyout((caddr_t) <mp, (caddr_t) SCARG(uap, sp), sizeof ltmp); 337 } 338 339 /* 340 * uname(). Just copy the info from the various strings stored in the 341 * kernel, and put it in the Linux utsname structure. That structure 342 * is almost the same as the NetBSD one, only it has fields 65 characters 343 * long, and an extra domainname field. 344 */ 345 int 346 linux_sys_uname(p, v, retval) 347 struct proc *p; 348 void *v; 349 register_t *retval; 350 { 351 struct linux_sys_uname_args /* { 352 syscallarg(struct linux_utsname *) up; 353 } */ *uap = v; 354 extern char ostype[], hostname[], osrelease[], version[], machine[], 355 domainname[]; 356 struct linux_utsname luts; 357 int len; 358 char *cp; 359 360 strncpy(luts.l_sysname, ostype, sizeof(luts.l_sysname)); 361 strncpy(luts.l_nodename, hostname, sizeof(luts.l_nodename)); 362 strncpy(luts.l_release, osrelease, sizeof(luts.l_release)); 363 strncpy(luts.l_version, version, sizeof(luts.l_version)); 364 strncpy(luts.l_machine, machine, sizeof(luts.l_machine)); 365 strncpy(luts.l_domainname, domainname, sizeof(luts.l_domainname)); 366 367 /* This part taken from the the uname() in libc */ 368 len = sizeof(luts.l_version); 369 for (cp = luts.l_version; len--; ++cp) { 370 if (*cp == '\n' || *cp == '\t') { 371 if (len > 1) 372 *cp = ' '; 373 else 374 *cp = '\0'; 375 } 376 } 377 378 return copyout(&luts, SCARG(uap, up), sizeof(luts)); 379 } 380 381 /* Used directly on: alpha, mips, ppc, sparc, sparc64 */ 382 /* Used indirectly on: arm, i386, m68k */ 383 384 /* 385 * New type Linux mmap call. 386 * Only called directly on machines with >= 6 free regs. 387 */ 388 int 389 linux_sys_mmap(p, v, retval) 390 struct proc *p; 391 void *v; 392 register_t *retval; 393 { 394 struct linux_sys_mmap_args /* { 395 syscallarg(unsigned long) addr; 396 syscallarg(size_t) len; 397 syscallarg(int) prot; 398 syscallarg(int) flags; 399 syscallarg(int) fd; 400 syscallarg(off_t) offset; 401 } */ *uap = v; 402 struct sys_mmap_args cma; 403 int flags; 404 405 flags = 0; 406 flags |= cvtto_bsd_mask(SCARG(uap,flags), LINUX_MAP_SHARED, MAP_SHARED); 407 flags |= cvtto_bsd_mask(SCARG(uap,flags), LINUX_MAP_PRIVATE, MAP_PRIVATE); 408 flags |= cvtto_bsd_mask(SCARG(uap,flags), LINUX_MAP_FIXED, MAP_FIXED); 409 flags |= cvtto_bsd_mask(SCARG(uap,flags), LINUX_MAP_ANON, MAP_ANON); 410 /* XXX XAX ERH: Any other flags here? There are more defined... */ 411 412 SCARG(&cma,addr) = (void *)SCARG(uap, addr); 413 SCARG(&cma,len) = SCARG(uap, len); 414 SCARG(&cma,prot) = SCARG(uap, prot); 415 if (SCARG(&cma,prot) & VM_PROT_WRITE) /* XXX */ 416 SCARG(&cma,prot) |= VM_PROT_READ; 417 SCARG(&cma,flags) = flags; 418 SCARG(&cma,fd) = SCARG(uap, fd); 419 SCARG(&cma,pad) = 0; 420 SCARG(&cma,pos) = SCARG(uap, offset); 421 422 return sys_mmap(p, &cma, retval); 423 } 424 425 int 426 linux_sys_mremap(p, v, retval) 427 struct proc *p; 428 void *v; 429 register_t *retval; 430 { 431 struct linux_sys_mremap_args /* { 432 syscallarg(void *) old_address; 433 syscallarg(size_t) old_size; 434 syscallarg(size_t) new_size; 435 syscallarg(u_long) flags; 436 } */ *uap = v; 437 struct sys_munmap_args mua; 438 size_t old_size, new_size; 439 int error; 440 441 old_size = round_page(SCARG(uap, old_size)); 442 new_size = round_page(SCARG(uap, new_size)); 443 444 /* 445 * Growing mapped region. 446 */ 447 if (new_size > old_size) { 448 /* 449 * XXX Implement me. What we probably want to do is 450 * XXX dig out the guts of the old mapping, mmap that 451 * XXX object again with the new size, then munmap 452 * XXX the old mapping. 453 */ 454 *retval = 0; 455 return (ENOMEM); 456 } 457 458 /* 459 * Shrinking mapped region. 460 */ 461 if (new_size < old_size) { 462 SCARG(&mua, addr) = (caddr_t)SCARG(uap, old_address) + 463 new_size; 464 SCARG(&mua, len) = old_size - new_size; 465 error = sys_munmap(p, &mua, retval); 466 *retval = error ? 0 : (register_t)SCARG(uap, old_address); 467 return (error); 468 } 469 470 /* 471 * No change. 472 */ 473 *retval = (register_t)SCARG(uap, old_address); 474 return (0); 475 } 476 477 int 478 linux_sys_msync(p, v, retval) 479 struct proc *p; 480 void *v; 481 register_t *retval; 482 { 483 struct linux_sys_msync_args /* { 484 syscallarg(caddr_t) addr; 485 syscallarg(int) len; 486 syscallarg(int) fl; 487 } */ *uap = v; 488 489 struct sys___msync13_args bma; 490 491 /* flags are ignored */ 492 SCARG(&bma, addr) = SCARG(uap, addr); 493 SCARG(&bma, len) = SCARG(uap, len); 494 SCARG(&bma, flags) = SCARG(uap, fl); 495 496 return sys___msync13(p, &bma, retval); 497 } 498 499 /* 500 * This code is partly stolen from src/lib/libc/compat-43/times.c 501 * XXX - CLK_TCK isn't declared in /sys, just in <time.h>, done here 502 */ 503 504 #define CLK_TCK 100 505 #define CONVTCK(r) (r.tv_sec * CLK_TCK + r.tv_usec / (1000000 / CLK_TCK)) 506 507 int 508 linux_sys_times(p, v, retval) 509 struct proc *p; 510 void *v; 511 register_t *retval; 512 { 513 struct linux_sys_times_args /* { 514 syscallarg(struct times *) tms; 515 } */ *uap = v; 516 struct timeval t; 517 struct linux_tms ltms; 518 struct rusage ru; 519 int error, s; 520 521 calcru(p, &ru.ru_utime, &ru.ru_stime, NULL); 522 ltms.ltms_utime = CONVTCK(ru.ru_utime); 523 ltms.ltms_stime = CONVTCK(ru.ru_stime); 524 525 ltms.ltms_cutime = CONVTCK(p->p_stats->p_cru.ru_utime); 526 ltms.ltms_cstime = CONVTCK(p->p_stats->p_cru.ru_stime); 527 528 if ((error = copyout(<ms, SCARG(uap, tms), sizeof ltms))) 529 return error; 530 531 s = splclock(); 532 timersub(&time, &boottime, &t); 533 splx(s); 534 535 retval[0] = ((linux_clock_t)(CONVTCK(t))); 536 return 0; 537 } 538 539 /* 540 * Linux 'readdir' call. This code is mostly taken from the 541 * SunOS getdents call (see compat/sunos/sunos_misc.c), though 542 * an attempt has been made to keep it a little cleaner (failing 543 * miserably, because of the cruft needed if count 1 is passed). 544 * 545 * The d_off field should contain the offset of the next valid entry, 546 * but in Linux it has the offset of the entry itself. We emulate 547 * that bug here. 548 * 549 * Read in BSD-style entries, convert them, and copy them out. 550 * 551 * Note that this doesn't handle union-mounted filesystems. 552 */ 553 int 554 linux_sys_getdents(p, v, retval) 555 struct proc *p; 556 void *v; 557 register_t *retval; 558 { 559 struct linux_sys_getdents_args /* { 560 syscallarg(int) fd; 561 syscallarg(struct linux_dirent *) dent; 562 syscallarg(unsigned int) count; 563 } */ *uap = v; 564 register struct dirent *bdp; 565 struct vnode *vp; 566 caddr_t inp, buf; /* BSD-format */ 567 int len, reclen; /* BSD-format */ 568 caddr_t outp; /* Linux-format */ 569 int resid, linux_reclen = 0; /* Linux-format */ 570 struct file *fp; 571 struct uio auio; 572 struct iovec aiov; 573 struct linux_dirent idb; 574 off_t off; /* true file offset */ 575 int buflen, error, eofflag, nbytes, oldcall; 576 struct vattr va; 577 off_t *cookiebuf = NULL, *cookie; 578 int ncookies; 579 580 /* getvnode() will use the descriptor for us */ 581 if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0) 582 return (error); 583 584 if ((fp->f_flag & FREAD) == 0) { 585 error = EBADF; 586 goto out1; 587 } 588 589 vp = (struct vnode *)fp->f_data; 590 if (vp->v_type != VDIR) { 591 error = EINVAL; 592 goto out1; 593 } 594 595 if ((error = VOP_GETATTR(vp, &va, p->p_ucred, p))) 596 goto out1; 597 598 nbytes = SCARG(uap, count); 599 if (nbytes == 1) { /* emulating old, broken behaviour */ 600 nbytes = sizeof (struct linux_dirent); 601 buflen = max(va.va_blocksize, nbytes); 602 oldcall = 1; 603 } else { 604 buflen = min(MAXBSIZE, nbytes); 605 if (buflen < va.va_blocksize) 606 buflen = va.va_blocksize; 607 oldcall = 0; 608 } 609 buf = malloc(buflen, M_TEMP, M_WAITOK); 610 611 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 612 off = fp->f_offset; 613 again: 614 aiov.iov_base = buf; 615 aiov.iov_len = buflen; 616 auio.uio_iov = &aiov; 617 auio.uio_iovcnt = 1; 618 auio.uio_rw = UIO_READ; 619 auio.uio_segflg = UIO_SYSSPACE; 620 auio.uio_procp = p; 621 auio.uio_resid = buflen; 622 auio.uio_offset = off; 623 /* 624 * First we read into the malloc'ed buffer, then 625 * we massage it into user space, one record at a time. 626 */ 627 error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, &cookiebuf, 628 &ncookies); 629 if (error) 630 goto out; 631 632 inp = buf; 633 outp = (caddr_t)SCARG(uap, dent); 634 resid = nbytes; 635 if ((len = buflen - auio.uio_resid) == 0) 636 goto eof; 637 638 for (cookie = cookiebuf; len > 0; len -= reclen) { 639 bdp = (struct dirent *)inp; 640 reclen = bdp->d_reclen; 641 if (reclen & 3) 642 panic("linux_readdir"); 643 if (bdp->d_fileno == 0) { 644 inp += reclen; /* it is a hole; squish it out */ 645 off = *cookie++; 646 continue; 647 } 648 linux_reclen = LINUX_RECLEN(&idb, bdp->d_namlen); 649 if (reclen > len || resid < linux_reclen) { 650 /* entry too big for buffer, so just stop */ 651 outp++; 652 break; 653 } 654 /* 655 * Massage in place to make a Linux-shaped dirent (otherwise 656 * we have to worry about touching user memory outside of 657 * the copyout() call). 658 */ 659 idb.d_ino = (linux_ino_t)bdp->d_fileno; 660 /* 661 * The old readdir() call misuses the offset and reclen fields. 662 */ 663 if (oldcall) { 664 idb.d_off = (linux_off_t)linux_reclen; 665 idb.d_reclen = (u_short)bdp->d_namlen; 666 } else { 667 if (sizeof (linux_off_t) < 4 && (off >> 32) != 0) { 668 compat_offseterr(vp, "linux_getdents"); 669 error = EINVAL; 670 goto out; 671 } 672 idb.d_off = (linux_off_t)off; 673 idb.d_reclen = (u_short)linux_reclen; 674 } 675 strcpy(idb.d_name, bdp->d_name); 676 if ((error = copyout((caddr_t)&idb, outp, linux_reclen))) 677 goto out; 678 /* advance past this real entry */ 679 inp += reclen; 680 off = *cookie++; /* each entry points to itself */ 681 /* advance output past Linux-shaped entry */ 682 outp += linux_reclen; 683 resid -= linux_reclen; 684 if (oldcall) 685 break; 686 } 687 688 /* if we squished out the whole block, try again */ 689 if (outp == (caddr_t)SCARG(uap, dent)) 690 goto again; 691 fp->f_offset = off; /* update the vnode offset */ 692 693 if (oldcall) 694 nbytes = resid + linux_reclen; 695 696 eof: 697 *retval = nbytes - resid; 698 out: 699 VOP_UNLOCK(vp, 0); 700 if (cookiebuf) 701 free(cookiebuf, M_TEMP); 702 free(buf, M_TEMP); 703 out1: 704 FILE_UNUSE(fp, p); 705 return error; 706 } 707 708 /* 709 * Even when just using registers to pass arguments to syscalls you can 710 * have 5 of them on the i386. So this newer version of select() does 711 * this. 712 */ 713 int 714 linux_sys_select(p, v, retval) 715 struct proc *p; 716 void *v; 717 register_t *retval; 718 { 719 struct linux_sys_select_args /* { 720 syscallarg(int) nfds; 721 syscallarg(fd_set *) readfds; 722 syscallarg(fd_set *) writefds; 723 syscallarg(fd_set *) exceptfds; 724 syscallarg(struct timeval *) timeout; 725 } */ *uap = v; 726 727 return linux_select1(p, retval, SCARG(uap, nfds), SCARG(uap, readfds), 728 SCARG(uap, writefds), SCARG(uap, exceptfds), SCARG(uap, timeout)); 729 } 730 731 /* 732 * Common code for the old and new versions of select(). A couple of 733 * things are important: 734 * 1) return the amount of time left in the 'timeout' parameter 735 * 2) select never returns ERESTART on Linux, always return EINTR 736 */ 737 int 738 linux_select1(p, retval, nfds, readfds, writefds, exceptfds, timeout) 739 struct proc *p; 740 register_t *retval; 741 int nfds; 742 fd_set *readfds, *writefds, *exceptfds; 743 struct timeval *timeout; 744 { 745 struct sys_select_args bsa; 746 struct timeval tv0, tv1, utv, *tvp; 747 caddr_t sg; 748 int error; 749 750 SCARG(&bsa, nd) = nfds; 751 SCARG(&bsa, in) = readfds; 752 SCARG(&bsa, ou) = writefds; 753 SCARG(&bsa, ex) = exceptfds; 754 SCARG(&bsa, tv) = timeout; 755 756 /* 757 * Store current time for computation of the amount of 758 * time left. 759 */ 760 if (timeout) { 761 if ((error = copyin(timeout, &utv, sizeof(utv)))) 762 return error; 763 if (itimerfix(&utv)) { 764 /* 765 * The timeval was invalid. Convert it to something 766 * valid that will act as it does under Linux. 767 */ 768 sg = stackgap_init(p->p_emul); 769 tvp = stackgap_alloc(&sg, sizeof(utv)); 770 utv.tv_sec += utv.tv_usec / 1000000; 771 utv.tv_usec %= 1000000; 772 if (utv.tv_usec < 0) { 773 utv.tv_sec -= 1; 774 utv.tv_usec += 1000000; 775 } 776 if (utv.tv_sec < 0) 777 timerclear(&utv); 778 if ((error = copyout(&utv, tvp, sizeof(utv)))) 779 return error; 780 SCARG(&bsa, tv) = tvp; 781 } 782 microtime(&tv0); 783 } 784 785 error = sys_select(p, &bsa, retval); 786 if (error) { 787 /* 788 * See fs/select.c in the Linux kernel. Without this, 789 * Maelstrom doesn't work. 790 */ 791 if (error == ERESTART) 792 error = EINTR; 793 return error; 794 } 795 796 if (timeout) { 797 if (*retval) { 798 /* 799 * Compute how much time was left of the timeout, 800 * by subtracting the current time and the time 801 * before we started the call, and subtracting 802 * that result from the user-supplied value. 803 */ 804 microtime(&tv1); 805 timersub(&tv1, &tv0, &tv1); 806 timersub(&utv, &tv1, &utv); 807 if (utv.tv_sec < 0) 808 timerclear(&utv); 809 } else 810 timerclear(&utv); 811 if ((error = copyout(&utv, timeout, sizeof(utv)))) 812 return error; 813 } 814 815 return 0; 816 } 817 818 /* 819 * Get the process group of a certain process. Look it up 820 * and return the value. 821 */ 822 int 823 linux_sys_getpgid(p, v, retval) 824 struct proc *p; 825 void *v; 826 register_t *retval; 827 { 828 struct linux_sys_getpgid_args /* { 829 syscallarg(int) pid; 830 } */ *uap = v; 831 struct proc *targp; 832 833 if (SCARG(uap, pid) != 0 && SCARG(uap, pid) != p->p_pid) { 834 if ((targp = pfind(SCARG(uap, pid))) == 0) 835 return ESRCH; 836 } 837 else 838 targp = p; 839 840 retval[0] = targp->p_pgid; 841 return 0; 842 } 843 844 /* 845 * Set the 'personality' (emulation mode) for the current process. Only 846 * accept the Linux personality here (0). This call is needed because 847 * the Linux ELF crt0 issues it in an ugly kludge to make sure that 848 * ELF binaries run in Linux mode, not SVR4 mode. 849 */ 850 int 851 linux_sys_personality(p, v, retval) 852 struct proc *p; 853 void *v; 854 register_t *retval; 855 { 856 struct linux_sys_personality_args /* { 857 syscallarg(int) per; 858 } */ *uap = v; 859 860 if (SCARG(uap, per) != 0) 861 return EINVAL; 862 retval[0] = 0; 863 return 0; 864 } 865 866 /* 867 * The calls are here because of type conversions. 868 */ 869 int 870 linux_sys_setreuid(p, v, retval) 871 struct proc *p; 872 void *v; 873 register_t *retval; 874 { 875 struct linux_sys_setreuid_args /* { 876 syscallarg(int) ruid; 877 syscallarg(int) euid; 878 } */ *uap = v; 879 struct sys_setreuid_args bsa; 880 881 SCARG(&bsa, ruid) = ((linux_uid_t)SCARG(uap, ruid) == (linux_uid_t)-1) ? 882 (uid_t)-1 : SCARG(uap, ruid); 883 SCARG(&bsa, euid) = ((linux_uid_t)SCARG(uap, euid) == (linux_uid_t)-1) ? 884 (uid_t)-1 : SCARG(uap, euid); 885 886 return sys_setreuid(p, &bsa, retval); 887 } 888 889 int 890 linux_sys_setregid(p, v, retval) 891 struct proc *p; 892 void *v; 893 register_t *retval; 894 { 895 struct linux_sys_setregid_args /* { 896 syscallarg(int) rgid; 897 syscallarg(int) egid; 898 } */ *uap = v; 899 struct sys_setregid_args bsa; 900 901 SCARG(&bsa, rgid) = ((linux_gid_t)SCARG(uap, rgid) == (linux_gid_t)-1) ? 902 (uid_t)-1 : SCARG(uap, rgid); 903 SCARG(&bsa, egid) = ((linux_gid_t)SCARG(uap, egid) == (linux_gid_t)-1) ? 904 (uid_t)-1 : SCARG(uap, egid); 905 906 return sys_setregid(p, &bsa, retval); 907 } 908 909 /* 910 * We have nonexistent fsuid equal to uid. 911 * If modification is requested, refuse. 912 */ 913 int 914 linux_sys_setfsuid(p, v, retval) 915 struct proc *p; 916 void *v; 917 register_t *retval; 918 { 919 struct linux_sys_setfsuid_args /* { 920 syscallarg(uid_t) uid; 921 } */ *uap = v; 922 uid_t uid; 923 924 uid = SCARG(uap, uid); 925 if (p->p_cred->p_ruid != uid) 926 return sys_nosys(p, v, retval); 927 else 928 return (0); 929 } 930 931 int 932 linux_sys_getfsuid(p, v, retval) 933 struct proc *p; 934 void *v; 935 register_t *retval; 936 { 937 return sys_getuid(p, v, retval); 938 } 939 940 int 941 linux_sys___sysctl(p, v, retval) 942 struct proc *p; 943 void *v; 944 register_t *retval; 945 { 946 struct linux_sys___sysctl_args /* { 947 syscallarg(struct linux___sysctl *) lsp; 948 } */ *uap = v; 949 struct linux___sysctl ls; 950 struct sys___sysctl_args bsa; 951 int error; 952 953 if ((error = copyin(SCARG(uap, lsp), &ls, sizeof ls))) 954 return error; 955 SCARG(&bsa, name) = ls.name; 956 SCARG(&bsa, namelen) = ls.namelen; 957 SCARG(&bsa, old) = ls.old; 958 SCARG(&bsa, oldlenp) = ls.oldlenp; 959 SCARG(&bsa, new) = ls.new; 960 SCARG(&bsa, newlen) = ls.newlen; 961 962 return sys___sysctl(p, &bsa, retval); 963 } 964 965 int 966 linux_sys_setresuid(p, v, retval) 967 struct proc *p; 968 void *v; 969 register_t *retval; 970 { 971 struct linux_sys_setresuid_args /* { 972 syscallarg(uid_t) ruid; 973 syscallarg(uid_t) euid; 974 syscallarg(uid_t) suid; 975 } */ *uap = v; 976 struct pcred *pc = p->p_cred; 977 uid_t ruid, euid, suid; 978 int error; 979 980 ruid = SCARG(uap, ruid); 981 euid = SCARG(uap, euid); 982 suid = SCARG(uap, suid); 983 984 /* 985 * Note: These checks are a little different than the NetBSD 986 * setreuid(2) call performs. This precisely follows the 987 * behavior of the Linux kernel. 988 */ 989 if (ruid != (uid_t)-1 && 990 ruid != pc->p_ruid && 991 ruid != pc->pc_ucred->cr_uid && 992 ruid != pc->p_svuid && 993 (error = suser(pc->pc_ucred, &p->p_acflag))) 994 return (error); 995 996 if (euid != (uid_t)-1 && 997 euid != pc->p_ruid && 998 euid != pc->pc_ucred->cr_uid && 999 euid != pc->p_svuid && 1000 (error = suser(pc->pc_ucred, &p->p_acflag))) 1001 return (error); 1002 1003 if (suid != (uid_t)-1 && 1004 suid != pc->p_ruid && 1005 suid != pc->pc_ucred->cr_uid && 1006 suid != pc->p_svuid && 1007 (error = suser(pc->pc_ucred, &p->p_acflag))) 1008 return (error); 1009 1010 /* 1011 * Now assign the new real, effective, and saved UIDs. 1012 * Note that Linux, unlike NetBSD in setreuid(2), does not 1013 * set the saved UID in this call unless the user specifies 1014 * it. 1015 */ 1016 if (ruid != (uid_t)-1) { 1017 (void)chgproccnt(pc->p_ruid, -1); 1018 (void)chgproccnt(ruid, 1); 1019 pc->p_ruid = ruid; 1020 } 1021 1022 if (euid != (uid_t)-1) { 1023 pc->pc_ucred = crcopy(pc->pc_ucred); 1024 pc->pc_ucred->cr_uid = euid; 1025 } 1026 1027 if (suid != (uid_t)-1) 1028 pc->p_svuid = suid; 1029 1030 if (ruid != (uid_t)-1 && euid != (uid_t)-1 && suid != (uid_t)-1) 1031 p->p_flag |= P_SUGID; 1032 return (0); 1033 } 1034 1035 int 1036 linux_sys_getresuid(p, v, retval) 1037 struct proc *p; 1038 void *v; 1039 register_t *retval; 1040 { 1041 struct linux_sys_getresuid_args /* { 1042 syscallarg(uid_t *) ruid; 1043 syscallarg(uid_t *) euid; 1044 syscallarg(uid_t *) suid; 1045 } */ *uap = v; 1046 struct pcred *pc = p->p_cred; 1047 int error; 1048 1049 /* 1050 * Linux copies these values out to userspace like so: 1051 * 1052 * 1. Copy out ruid. 1053 * 2. If that succeeds, copy out euid. 1054 * 3. If both of those succeed, copy out suid. 1055 */ 1056 if ((error = copyout(&pc->p_ruid, SCARG(uap, ruid), 1057 sizeof(uid_t))) != 0) 1058 return (error); 1059 1060 if ((error = copyout(&pc->pc_ucred->cr_uid, SCARG(uap, euid), 1061 sizeof(uid_t))) != 0) 1062 return (error); 1063 1064 return (copyout(&pc->p_svuid, SCARG(uap, suid), sizeof(uid_t))); 1065 } 1066 1067 int 1068 linux_sys_ptrace(p, v, retval) 1069 struct proc *p; 1070 void *v; 1071 register_t *retval; 1072 { 1073 struct linux_sys_ptrace_args /* { 1074 syscallarg(int) request; 1075 syscallarg(int) pid; 1076 syscallarg(int) addr; 1077 syscallarg(int) data; 1078 } */ *uap = v; 1079 int *ptr, request; 1080 1081 ptr = linux_ptrace_request_map; 1082 request = SCARG(uap, request); 1083 while (*ptr != -1) 1084 if (*ptr++ == request) { 1085 struct sys_ptrace_args pta; 1086 caddr_t sg; 1087 1088 sg = stackgap_init(p->p_emul); 1089 1090 SCARG(&pta, req) = *ptr; 1091 SCARG(&pta, pid) = SCARG(uap, pid); 1092 SCARG(&pta, addr) = (caddr_t)SCARG(uap, addr); 1093 SCARG(&pta, data) = SCARG(uap, data); 1094 1095 return sys_ptrace(p, &pta, retval); 1096 } 1097 else 1098 ptr++; 1099 1100 return LINUX_SYS_PTRACE_ARCH(p, uap, retval); 1101 } 1102