1 /* $NetBSD: linux_misc.c,v 1.231 2015/03/14 08:32:08 njoly Exp $ */ 2 3 /*- 4 * Copyright (c) 1995, 1998, 1999, 2008 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 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Linux compatibility module. Try to deal with various Linux system calls. 35 */ 36 37 /* 38 * These functions have been moved to multiarch to allow 39 * selection of which machines include them to be 40 * determined by the individual files.linux_<arch> files. 41 * 42 * Function in multiarch: 43 * linux_sys_break : linux_break.c 44 * linux_sys_alarm : linux_misc_notalpha.c 45 * linux_sys_getresgid : linux_misc_notalpha.c 46 * linux_sys_nice : linux_misc_notalpha.c 47 * linux_sys_readdir : linux_misc_notalpha.c 48 * linux_sys_setresgid : linux_misc_notalpha.c 49 * linux_sys_time : linux_misc_notalpha.c 50 * linux_sys_utime : linux_misc_notalpha.c 51 * linux_sys_waitpid : linux_misc_notalpha.c 52 * linux_sys_old_mmap : linux_oldmmap.c 53 * linux_sys_oldolduname : linux_oldolduname.c 54 * linux_sys_oldselect : linux_oldselect.c 55 * linux_sys_olduname : linux_olduname.c 56 * linux_sys_pipe : linux_pipe.c 57 */ 58 59 #include <sys/cdefs.h> 60 __KERNEL_RCSID(0, "$NetBSD: linux_misc.c,v 1.231 2015/03/14 08:32:08 njoly Exp $"); 61 62 #include <sys/param.h> 63 #include <sys/systm.h> 64 #include <sys/namei.h> 65 #include <sys/proc.h> 66 #include <sys/dirent.h> 67 #include <sys/file.h> 68 #include <sys/stat.h> 69 #include <sys/filedesc.h> 70 #include <sys/ioctl.h> 71 #include <sys/kernel.h> 72 #include <sys/malloc.h> 73 #include <sys/mbuf.h> 74 #include <sys/mman.h> 75 #include <sys/mount.h> 76 #include <sys/poll.h> 77 #include <sys/prot.h> 78 #include <sys/reboot.h> 79 #include <sys/resource.h> 80 #include <sys/resourcevar.h> 81 #include <sys/select.h> 82 #include <sys/signal.h> 83 #include <sys/signalvar.h> 84 #include <sys/socket.h> 85 #include <sys/time.h> 86 #include <sys/times.h> 87 #include <sys/vnode.h> 88 #include <sys/uio.h> 89 #include <sys/wait.h> 90 #include <sys/utsname.h> 91 #include <sys/unistd.h> 92 #include <sys/vfs_syscalls.h> 93 #include <sys/swap.h> /* for SWAP_ON */ 94 #include <sys/sysctl.h> /* for KERN_DOMAINNAME */ 95 #include <sys/kauth.h> 96 97 #include <sys/ptrace.h> 98 #include <machine/ptrace.h> 99 100 #include <sys/syscall.h> 101 #include <sys/syscallargs.h> 102 103 #include <compat/sys/resource.h> 104 105 #include <compat/linux/common/linux_machdep.h> 106 #include <compat/linux/common/linux_types.h> 107 #include <compat/linux/common/linux_signal.h> 108 #include <compat/linux/common/linux_ipc.h> 109 #include <compat/linux/common/linux_sem.h> 110 111 #include <compat/linux/common/linux_fcntl.h> 112 #include <compat/linux/common/linux_mmap.h> 113 #include <compat/linux/common/linux_dirent.h> 114 #include <compat/linux/common/linux_util.h> 115 #include <compat/linux/common/linux_misc.h> 116 #include <compat/linux/common/linux_statfs.h> 117 #include <compat/linux/common/linux_limit.h> 118 #include <compat/linux/common/linux_ptrace.h> 119 #include <compat/linux/common/linux_reboot.h> 120 #include <compat/linux/common/linux_emuldata.h> 121 #include <compat/linux/common/linux_sched.h> 122 123 #include <compat/linux/linux_syscallargs.h> 124 125 const int linux_ptrace_request_map[] = { 126 LINUX_PTRACE_TRACEME, PT_TRACE_ME, 127 LINUX_PTRACE_PEEKTEXT, PT_READ_I, 128 LINUX_PTRACE_PEEKDATA, PT_READ_D, 129 LINUX_PTRACE_POKETEXT, PT_WRITE_I, 130 LINUX_PTRACE_POKEDATA, PT_WRITE_D, 131 LINUX_PTRACE_CONT, PT_CONTINUE, 132 LINUX_PTRACE_KILL, PT_KILL, 133 LINUX_PTRACE_ATTACH, PT_ATTACH, 134 LINUX_PTRACE_DETACH, PT_DETACH, 135 # ifdef PT_STEP 136 LINUX_PTRACE_SINGLESTEP, PT_STEP, 137 # endif 138 LINUX_PTRACE_SYSCALL, PT_SYSCALL, 139 -1 140 }; 141 142 const struct linux_mnttypes linux_fstypes[] = { 143 { MOUNT_FFS, LINUX_DEFAULT_SUPER_MAGIC }, 144 { MOUNT_NFS, LINUX_NFS_SUPER_MAGIC }, 145 { MOUNT_MFS, LINUX_DEFAULT_SUPER_MAGIC }, 146 { MOUNT_MSDOS, LINUX_MSDOS_SUPER_MAGIC }, 147 { MOUNT_LFS, LINUX_DEFAULT_SUPER_MAGIC }, 148 { MOUNT_FDESC, LINUX_DEFAULT_SUPER_MAGIC }, 149 { MOUNT_NULL, LINUX_DEFAULT_SUPER_MAGIC }, 150 { MOUNT_OVERLAY, LINUX_DEFAULT_SUPER_MAGIC }, 151 { MOUNT_UMAP, LINUX_DEFAULT_SUPER_MAGIC }, 152 { MOUNT_KERNFS, LINUX_DEFAULT_SUPER_MAGIC }, 153 { MOUNT_PROCFS, LINUX_PROC_SUPER_MAGIC }, 154 { MOUNT_AFS, LINUX_DEFAULT_SUPER_MAGIC }, 155 { MOUNT_CD9660, LINUX_ISOFS_SUPER_MAGIC }, 156 { MOUNT_UNION, LINUX_DEFAULT_SUPER_MAGIC }, 157 { MOUNT_ADOSFS, LINUX_ADFS_SUPER_MAGIC }, 158 { MOUNT_EXT2FS, LINUX_EXT2_SUPER_MAGIC }, 159 { MOUNT_CFS, LINUX_DEFAULT_SUPER_MAGIC }, 160 { MOUNT_CODA, LINUX_CODA_SUPER_MAGIC }, 161 { MOUNT_FILECORE, LINUX_DEFAULT_SUPER_MAGIC }, 162 { MOUNT_NTFS, LINUX_DEFAULT_SUPER_MAGIC }, 163 { MOUNT_SMBFS, LINUX_SMB_SUPER_MAGIC }, 164 { MOUNT_PTYFS, LINUX_DEVPTS_SUPER_MAGIC }, 165 { MOUNT_TMPFS, LINUX_TMPFS_SUPER_MAGIC } 166 }; 167 const int linux_fstypes_cnt = sizeof(linux_fstypes) / sizeof(linux_fstypes[0]); 168 169 # ifdef DEBUG_LINUX 170 #define DPRINTF(a) uprintf a 171 # else 172 #define DPRINTF(a) 173 # endif 174 175 /* Local linux_misc.c functions: */ 176 static void linux_to_bsd_mmap_args(struct sys_mmap_args *, 177 const struct linux_sys_mmap_args *); 178 static int linux_mmap(struct lwp *, const struct linux_sys_mmap_args *, 179 register_t *, off_t); 180 181 182 /* 183 * The information on a terminated (or stopped) process needs 184 * to be converted in order for Linux binaries to get a valid signal 185 * number out of it. 186 */ 187 int 188 bsd_to_linux_wstat(int st) 189 { 190 191 int sig; 192 193 if (WIFSIGNALED(st)) { 194 sig = WTERMSIG(st); 195 if (sig >= 0 && sig < NSIG) 196 st= (st & ~0177) | native_to_linux_signo[sig]; 197 } else if (WIFSTOPPED(st)) { 198 sig = WSTOPSIG(st); 199 if (sig >= 0 && sig < NSIG) 200 st = (st & ~0xff00) | 201 (native_to_linux_signo[sig] << 8); 202 } 203 return st; 204 } 205 206 /* 207 * wait4(2). Passed on to the NetBSD call, surrounded by code to 208 * reserve some space for a NetBSD-style wait status, and converting 209 * it to what Linux wants. 210 */ 211 int 212 linux_sys_wait4(struct lwp *l, const struct linux_sys_wait4_args *uap, register_t *retval) 213 { 214 /* { 215 syscallarg(int) pid; 216 syscallarg(int *) status; 217 syscallarg(int) options; 218 syscallarg(struct rusage50 *) rusage; 219 } */ 220 int error, status, options, linux_options, pid = SCARG(uap, pid); 221 struct rusage50 ru50; 222 struct rusage ru; 223 proc_t *p; 224 225 linux_options = SCARG(uap, options); 226 options = WOPTSCHECKED; 227 if (linux_options & ~(LINUX_WAIT4_KNOWNFLAGS)) 228 return (EINVAL); 229 230 if (linux_options & LINUX_WAIT4_WNOHANG) 231 options |= WNOHANG; 232 if (linux_options & LINUX_WAIT4_WUNTRACED) 233 options |= WUNTRACED; 234 if (linux_options & LINUX_WAIT4_WALL) 235 options |= WALLSIG; 236 if (linux_options & LINUX_WAIT4_WCLONE) 237 options |= WALTSIG; 238 # ifdef DIAGNOSTIC 239 if (linux_options & LINUX_WAIT4_WNOTHREAD) 240 printf("WARNING: %s: linux process %d.%d called " 241 "waitpid with __WNOTHREAD set!", 242 __FILE__, l->l_proc->p_pid, l->l_lid); 243 244 # endif 245 246 error = do_sys_wait(&pid, &status, options, 247 SCARG(uap, rusage) != NULL ? &ru : NULL); 248 249 retval[0] = pid; 250 if (pid == 0) 251 return error; 252 253 p = curproc; 254 mutex_enter(p->p_lock); 255 sigdelset(&p->p_sigpend.sp_set, SIGCHLD); /* XXXAD ksiginfo leak */ 256 mutex_exit(p->p_lock); 257 258 if (SCARG(uap, rusage) != NULL) { 259 rusage_to_rusage50(&ru, &ru50); 260 error = copyout(&ru, SCARG(uap, rusage), sizeof(ru)); 261 } 262 263 if (error == 0 && SCARG(uap, status) != NULL) { 264 status = bsd_to_linux_wstat(status); 265 error = copyout(&status, SCARG(uap, status), sizeof status); 266 } 267 268 return error; 269 } 270 271 /* 272 * Linux brk(2). Like native, but always return the new break value. 273 */ 274 int 275 linux_sys_brk(struct lwp *l, const struct linux_sys_brk_args *uap, register_t *retval) 276 { 277 /* { 278 syscallarg(char *) nsize; 279 } */ 280 struct proc *p = l->l_proc; 281 struct vmspace *vm = p->p_vmspace; 282 struct sys_obreak_args oba; 283 284 SCARG(&oba, nsize) = SCARG(uap, nsize); 285 286 (void) sys_obreak(l, &oba, retval); 287 retval[0] = (register_t)((char *)vm->vm_daddr + ptoa(vm->vm_dsize)); 288 return 0; 289 } 290 291 /* 292 * Implement the fs stat functions. Straightforward. 293 */ 294 int 295 linux_sys_statfs(struct lwp *l, const struct linux_sys_statfs_args *uap, register_t *retval) 296 { 297 /* { 298 syscallarg(const char *) path; 299 syscallarg(struct linux_statfs *) sp; 300 } */ 301 struct statvfs *sb; 302 struct linux_statfs ltmp; 303 int error; 304 305 sb = STATVFSBUF_GET(); 306 error = do_sys_pstatvfs(l, SCARG(uap, path), ST_WAIT, sb); 307 if (error == 0) { 308 bsd_to_linux_statfs(sb, <mp); 309 error = copyout(<mp, SCARG(uap, sp), sizeof ltmp); 310 } 311 STATVFSBUF_PUT(sb); 312 313 return error; 314 } 315 316 int 317 linux_sys_fstatfs(struct lwp *l, const struct linux_sys_fstatfs_args *uap, register_t *retval) 318 { 319 /* { 320 syscallarg(int) fd; 321 syscallarg(struct linux_statfs *) sp; 322 } */ 323 struct statvfs *sb; 324 struct linux_statfs ltmp; 325 int error; 326 327 sb = STATVFSBUF_GET(); 328 error = do_sys_fstatvfs(l, SCARG(uap, fd), ST_WAIT, sb); 329 if (error == 0) { 330 bsd_to_linux_statfs(sb, <mp); 331 error = copyout(<mp, SCARG(uap, sp), sizeof ltmp); 332 } 333 STATVFSBUF_PUT(sb); 334 335 return error; 336 } 337 338 /* 339 * uname(). Just copy the info from the various strings stored in the 340 * kernel, and put it in the Linux utsname structure. That structure 341 * is almost the same as the NetBSD one, only it has fields 65 characters 342 * long, and an extra domainname field. 343 */ 344 int 345 linux_sys_uname(struct lwp *l, const struct linux_sys_uname_args *uap, register_t *retval) 346 { 347 /* { 348 syscallarg(struct linux_utsname *) up; 349 } */ 350 struct linux_utsname luts; 351 352 strlcpy(luts.l_sysname, linux_sysname, sizeof(luts.l_sysname)); 353 strlcpy(luts.l_nodename, hostname, sizeof(luts.l_nodename)); 354 strlcpy(luts.l_release, linux_release, sizeof(luts.l_release)); 355 strlcpy(luts.l_version, linux_version, sizeof(luts.l_version)); 356 strlcpy(luts.l_machine, LINUX_UNAME_ARCH, sizeof(luts.l_machine)); 357 strlcpy(luts.l_domainname, domainname, sizeof(luts.l_domainname)); 358 359 return copyout(&luts, SCARG(uap, up), sizeof(luts)); 360 } 361 362 /* Used directly on: alpha, mips, ppc, sparc, sparc64 */ 363 /* Used indirectly on: arm, i386, m68k */ 364 365 /* 366 * New type Linux mmap call. 367 * Only called directly on machines with >= 6 free regs. 368 */ 369 int 370 linux_sys_mmap(struct lwp *l, const struct linux_sys_mmap_args *uap, register_t *retval) 371 { 372 /* { 373 syscallarg(unsigned long) addr; 374 syscallarg(size_t) len; 375 syscallarg(int) prot; 376 syscallarg(int) flags; 377 syscallarg(int) fd; 378 syscallarg(linux_off_t) offset; 379 } */ 380 381 if (SCARG(uap, offset) & PAGE_MASK) 382 return EINVAL; 383 384 return linux_mmap(l, uap, retval, SCARG(uap, offset)); 385 } 386 387 /* 388 * Guts of most architectures' mmap64() implementations. This shares 389 * its list of arguments with linux_sys_mmap(). 390 * 391 * The difference in linux_sys_mmap2() is that "offset" is actually 392 * (offset / pagesize), not an absolute byte count. This translation 393 * to pagesize offsets is done inside glibc between the mmap64() call 394 * point, and the actual syscall. 395 */ 396 int 397 linux_sys_mmap2(struct lwp *l, const struct linux_sys_mmap2_args *uap, register_t *retval) 398 { 399 /* { 400 syscallarg(unsigned long) addr; 401 syscallarg(size_t) len; 402 syscallarg(int) prot; 403 syscallarg(int) flags; 404 syscallarg(int) fd; 405 syscallarg(linux_off_t) offset; 406 } */ 407 408 return linux_mmap(l, uap, retval, 409 ((off_t)SCARG(uap, offset)) << PAGE_SHIFT); 410 } 411 412 /* 413 * Massage arguments and call system mmap(2). 414 */ 415 static int 416 linux_mmap(struct lwp *l, const struct linux_sys_mmap_args *uap, register_t *retval, off_t offset) 417 { 418 struct sys_mmap_args cma; 419 int error; 420 size_t mmoff=0; 421 422 linux_to_bsd_mmap_args(&cma, uap); 423 SCARG(&cma, pos) = offset; 424 425 if (SCARG(uap, flags) & LINUX_MAP_GROWSDOWN) { 426 /* 427 * Request for stack-like memory segment. On linux, this 428 * works by mmap()ping (small) segment, which is automatically 429 * extended when page fault happens below the currently 430 * allocated area. We emulate this by allocating (typically 431 * bigger) segment sized at current stack size limit, and 432 * offsetting the requested and returned address accordingly. 433 * Since physical pages are only allocated on-demand, this 434 * is effectively identical. 435 */ 436 rlim_t ssl = l->l_proc->p_rlimit[RLIMIT_STACK].rlim_cur; 437 438 if (SCARG(&cma, len) < ssl) { 439 /* Compute the address offset */ 440 mmoff = round_page(ssl) - SCARG(uap, len); 441 442 if (SCARG(&cma, addr)) 443 SCARG(&cma, addr) = (char *)SCARG(&cma, addr) - mmoff; 444 445 SCARG(&cma, len) = (size_t) ssl; 446 } 447 } 448 449 error = sys_mmap(l, &cma, retval); 450 if (error) 451 return (error); 452 453 /* Shift the returned address for stack-like segment if necessary */ 454 retval[0] += mmoff; 455 456 return (0); 457 } 458 459 static void 460 linux_to_bsd_mmap_args(struct sys_mmap_args *cma, const struct linux_sys_mmap_args *uap) 461 { 462 int flags = MAP_TRYFIXED, fl = SCARG(uap, flags); 463 464 flags |= cvtto_bsd_mask(fl, LINUX_MAP_SHARED, MAP_SHARED); 465 flags |= cvtto_bsd_mask(fl, LINUX_MAP_PRIVATE, MAP_PRIVATE); 466 flags |= cvtto_bsd_mask(fl, LINUX_MAP_FIXED, MAP_FIXED); 467 flags |= cvtto_bsd_mask(fl, LINUX_MAP_ANON, MAP_ANON); 468 flags |= cvtto_bsd_mask(fl, LINUX_MAP_LOCKED, MAP_WIRED); 469 /* XXX XAX ERH: Any other flags here? There are more defined... */ 470 471 SCARG(cma, addr) = (void *)SCARG(uap, addr); 472 SCARG(cma, len) = SCARG(uap, len); 473 SCARG(cma, prot) = SCARG(uap, prot); 474 if (SCARG(cma, prot) & VM_PROT_WRITE) /* XXX */ 475 SCARG(cma, prot) |= VM_PROT_READ; 476 SCARG(cma, flags) = flags; 477 SCARG(cma, fd) = flags & MAP_ANON ? -1 : SCARG(uap, fd); 478 SCARG(cma, PAD) = 0; 479 } 480 481 #define LINUX_MREMAP_MAYMOVE 1 482 #define LINUX_MREMAP_FIXED 2 483 484 int 485 linux_sys_mremap(struct lwp *l, const struct linux_sys_mremap_args *uap, register_t *retval) 486 { 487 /* { 488 syscallarg(void *) old_address; 489 syscallarg(size_t) old_size; 490 syscallarg(size_t) new_size; 491 syscallarg(u_long) flags; 492 } */ 493 494 struct proc *p; 495 struct vm_map *map; 496 vaddr_t oldva; 497 vaddr_t newva; 498 size_t oldsize; 499 size_t newsize; 500 int flags; 501 int uvmflags; 502 int error; 503 504 flags = SCARG(uap, flags); 505 oldva = (vaddr_t)SCARG(uap, old_address); 506 oldsize = round_page(SCARG(uap, old_size)); 507 newsize = round_page(SCARG(uap, new_size)); 508 if ((flags & ~(LINUX_MREMAP_FIXED|LINUX_MREMAP_MAYMOVE)) != 0) { 509 error = EINVAL; 510 goto done; 511 } 512 if ((flags & LINUX_MREMAP_FIXED) != 0) { 513 if ((flags & LINUX_MREMAP_MAYMOVE) == 0) { 514 error = EINVAL; 515 goto done; 516 } 517 #if 0 /* notyet */ 518 newva = SCARG(uap, new_address); 519 uvmflags = MAP_FIXED; 520 #else /* notyet */ 521 error = EOPNOTSUPP; 522 goto done; 523 #endif /* notyet */ 524 } else if ((flags & LINUX_MREMAP_MAYMOVE) != 0) { 525 uvmflags = 0; 526 } else { 527 newva = oldva; 528 uvmflags = MAP_FIXED; 529 } 530 p = l->l_proc; 531 map = &p->p_vmspace->vm_map; 532 error = uvm_mremap(map, oldva, oldsize, map, &newva, newsize, p, 533 uvmflags); 534 535 done: 536 *retval = (error != 0) ? 0 : (register_t)newva; 537 return error; 538 } 539 540 #ifdef USRSTACK 541 int 542 linux_sys_mprotect(struct lwp *l, const struct linux_sys_mprotect_args *uap, register_t *retval) 543 { 544 /* { 545 syscallarg(const void *) start; 546 syscallarg(unsigned long) len; 547 syscallarg(int) prot; 548 } */ 549 struct vm_map_entry *entry; 550 struct vm_map *map; 551 struct proc *p; 552 vaddr_t end, start, len, stacklim; 553 int prot, grows; 554 555 start = (vaddr_t)SCARG(uap, start); 556 len = round_page(SCARG(uap, len)); 557 prot = SCARG(uap, prot); 558 grows = prot & (LINUX_PROT_GROWSDOWN | LINUX_PROT_GROWSUP); 559 prot &= ~grows; 560 end = start + len; 561 562 if (start & PAGE_MASK) 563 return EINVAL; 564 if (end < start) 565 return EINVAL; 566 if (end == start) 567 return 0; 568 569 if (prot & ~(PROT_READ | PROT_WRITE | PROT_EXEC)) 570 return EINVAL; 571 if (grows == (LINUX_PROT_GROWSDOWN | LINUX_PROT_GROWSUP)) 572 return EINVAL; 573 574 p = l->l_proc; 575 map = &p->p_vmspace->vm_map; 576 vm_map_lock(map); 577 # ifdef notdef 578 VM_MAP_RANGE_CHECK(map, start, end); 579 # endif 580 if (!uvm_map_lookup_entry(map, start, &entry) || entry->start > start) { 581 vm_map_unlock(map); 582 return ENOMEM; 583 } 584 585 /* 586 * Approximate the behaviour of PROT_GROWS{DOWN,UP}. 587 */ 588 589 stacklim = (vaddr_t)p->p_limit->pl_rlimit[RLIMIT_STACK].rlim_cur; 590 if (grows & LINUX_PROT_GROWSDOWN) { 591 if (USRSTACK - stacklim <= start && start < USRSTACK) { 592 start = USRSTACK - stacklim; 593 } else { 594 start = entry->start; 595 } 596 } else if (grows & LINUX_PROT_GROWSUP) { 597 if (USRSTACK <= end && end < USRSTACK + stacklim) { 598 end = USRSTACK + stacklim; 599 } else { 600 end = entry->end; 601 } 602 } 603 vm_map_unlock(map); 604 return uvm_map_protect(map, start, end, prot, FALSE); 605 } 606 #endif /* USRSTACK */ 607 608 /* 609 * This code is partly stolen from src/lib/libc/compat-43/times.c 610 */ 611 612 #define CONVTCK(r) (r.tv_sec * hz + r.tv_usec / (1000000 / hz)) 613 614 int 615 linux_sys_times(struct lwp *l, const struct linux_sys_times_args *uap, register_t *retval) 616 { 617 /* { 618 syscallarg(struct times *) tms; 619 } */ 620 struct proc *p = l->l_proc; 621 struct timeval t; 622 int error; 623 624 if (SCARG(uap, tms)) { 625 struct linux_tms ltms; 626 struct rusage ru; 627 628 mutex_enter(p->p_lock); 629 calcru(p, &ru.ru_utime, &ru.ru_stime, NULL, NULL); 630 ltms.ltms_utime = CONVTCK(ru.ru_utime); 631 ltms.ltms_stime = CONVTCK(ru.ru_stime); 632 ltms.ltms_cutime = CONVTCK(p->p_stats->p_cru.ru_utime); 633 ltms.ltms_cstime = CONVTCK(p->p_stats->p_cru.ru_stime); 634 mutex_exit(p->p_lock); 635 636 if ((error = copyout(<ms, SCARG(uap, tms), sizeof ltms))) 637 return error; 638 } 639 640 getmicrouptime(&t); 641 642 retval[0] = ((linux_clock_t)(CONVTCK(t))); 643 return 0; 644 } 645 646 #undef CONVTCK 647 648 /* 649 * Linux 'readdir' call. This code is mostly taken from the 650 * SunOS getdents call (see compat/sunos/sunos_misc.c), though 651 * an attempt has been made to keep it a little cleaner (failing 652 * miserably, because of the cruft needed if count 1 is passed). 653 * 654 * The d_off field should contain the offset of the next valid entry, 655 * but in Linux it has the offset of the entry itself. We emulate 656 * that bug here. 657 * 658 * Read in BSD-style entries, convert them, and copy them out. 659 * 660 * Note that this doesn't handle union-mounted filesystems. 661 */ 662 int 663 linux_sys_getdents(struct lwp *l, const struct linux_sys_getdents_args *uap, register_t *retval) 664 { 665 /* { 666 syscallarg(int) fd; 667 syscallarg(struct linux_dirent *) dent; 668 syscallarg(unsigned int) count; 669 } */ 670 struct dirent *bdp; 671 struct vnode *vp; 672 char *inp, *tbuf; /* BSD-format */ 673 int len, reclen; /* BSD-format */ 674 char *outp; /* Linux-format */ 675 int resid, linux_reclen = 0; /* Linux-format */ 676 struct file *fp; 677 struct uio auio; 678 struct iovec aiov; 679 struct linux_dirent idb; 680 off_t off; /* true file offset */ 681 int buflen, error, eofflag, nbytes, oldcall; 682 struct vattr va; 683 off_t *cookiebuf = NULL, *cookie; 684 int ncookies; 685 686 /* fd_getvnode() will use the descriptor for us */ 687 if ((error = fd_getvnode(SCARG(uap, fd), &fp)) != 0) 688 return (error); 689 690 if ((fp->f_flag & FREAD) == 0) { 691 error = EBADF; 692 goto out1; 693 } 694 695 vp = (struct vnode *)fp->f_data; 696 if (vp->v_type != VDIR) { 697 error = ENOTDIR; 698 goto out1; 699 } 700 701 vn_lock(vp, LK_SHARED | LK_RETRY); 702 error = VOP_GETATTR(vp, &va, l->l_cred); 703 VOP_UNLOCK(vp); 704 if (error) 705 goto out1; 706 707 nbytes = SCARG(uap, count); 708 if (nbytes == 1) { /* emulating old, broken behaviour */ 709 nbytes = sizeof (idb); 710 buflen = max(va.va_blocksize, nbytes); 711 oldcall = 1; 712 } else { 713 buflen = min(MAXBSIZE, nbytes); 714 if (buflen < va.va_blocksize) 715 buflen = va.va_blocksize; 716 oldcall = 0; 717 } 718 tbuf = malloc(buflen, M_TEMP, M_WAITOK); 719 720 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 721 off = fp->f_offset; 722 again: 723 aiov.iov_base = tbuf; 724 aiov.iov_len = buflen; 725 auio.uio_iov = &aiov; 726 auio.uio_iovcnt = 1; 727 auio.uio_rw = UIO_READ; 728 auio.uio_resid = buflen; 729 auio.uio_offset = off; 730 UIO_SETUP_SYSSPACE(&auio); 731 /* 732 * First we read into the malloc'ed buffer, then 733 * we massage it into user space, one record at a time. 734 */ 735 error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, &cookiebuf, 736 &ncookies); 737 if (error) 738 goto out; 739 740 inp = tbuf; 741 outp = (void *)SCARG(uap, dent); 742 resid = nbytes; 743 if ((len = buflen - auio.uio_resid) == 0) 744 goto eof; 745 746 for (cookie = cookiebuf; len > 0; len -= reclen) { 747 bdp = (struct dirent *)inp; 748 reclen = bdp->d_reclen; 749 if (reclen & 3) 750 panic("linux_readdir"); 751 if (bdp->d_fileno == 0) { 752 inp += reclen; /* it is a hole; squish it out */ 753 if (cookie) 754 off = *cookie++; 755 else 756 off += reclen; 757 continue; 758 } 759 linux_reclen = LINUX_RECLEN(&idb, bdp->d_namlen); 760 if (reclen > len || resid < linux_reclen) { 761 /* entry too big for buffer, so just stop */ 762 outp++; 763 break; 764 } 765 /* 766 * Massage in place to make a Linux-shaped dirent (otherwise 767 * we have to worry about touching user memory outside of 768 * the copyout() call). 769 */ 770 idb.d_ino = bdp->d_fileno; 771 /* 772 * The old readdir() call misuses the offset and reclen fields. 773 */ 774 if (oldcall) { 775 idb.d_off = (linux_off_t)linux_reclen; 776 idb.d_reclen = (u_short)bdp->d_namlen; 777 } else { 778 if (sizeof (idb.d_off) <= 4 && (off >> 32) != 0) { 779 compat_offseterr(vp, "linux_getdents"); 780 error = EINVAL; 781 goto out; 782 } 783 idb.d_off = (linux_off_t)off; 784 idb.d_reclen = (u_short)linux_reclen; 785 /* Linux puts d_type at the end of each record */ 786 *((char *)&idb + idb.d_reclen - 1) = bdp->d_type; 787 } 788 strcpy(idb.d_name, bdp->d_name); 789 if ((error = copyout((void *)&idb, outp, linux_reclen))) 790 goto out; 791 /* advance past this real entry */ 792 inp += reclen; 793 if (cookie) 794 off = *cookie++; /* each entry points to itself */ 795 else 796 off += reclen; 797 /* advance output past Linux-shaped entry */ 798 outp += linux_reclen; 799 resid -= linux_reclen; 800 if (oldcall) 801 break; 802 } 803 804 /* if we squished out the whole block, try again */ 805 if (outp == (void *)SCARG(uap, dent)) { 806 if (cookiebuf) 807 free(cookiebuf, M_TEMP); 808 cookiebuf = NULL; 809 goto again; 810 } 811 fp->f_offset = off; /* update the vnode offset */ 812 813 if (oldcall) 814 nbytes = resid + linux_reclen; 815 816 eof: 817 *retval = nbytes - resid; 818 out: 819 VOP_UNLOCK(vp); 820 if (cookiebuf) 821 free(cookiebuf, M_TEMP); 822 free(tbuf, M_TEMP); 823 out1: 824 fd_putfile(SCARG(uap, fd)); 825 return error; 826 } 827 828 /* 829 * Even when just using registers to pass arguments to syscalls you can 830 * have 5 of them on the i386. So this newer version of select() does 831 * this. 832 */ 833 int 834 linux_sys_select(struct lwp *l, const struct linux_sys_select_args *uap, register_t *retval) 835 { 836 /* { 837 syscallarg(int) nfds; 838 syscallarg(fd_set *) readfds; 839 syscallarg(fd_set *) writefds; 840 syscallarg(fd_set *) exceptfds; 841 syscallarg(struct timeval50 *) timeout; 842 } */ 843 844 return linux_select1(l, retval, SCARG(uap, nfds), SCARG(uap, readfds), 845 SCARG(uap, writefds), SCARG(uap, exceptfds), 846 (struct linux_timeval *)SCARG(uap, timeout)); 847 } 848 849 /* 850 * Common code for the old and new versions of select(). A couple of 851 * things are important: 852 * 1) return the amount of time left in the 'timeout' parameter 853 * 2) select never returns ERESTART on Linux, always return EINTR 854 */ 855 int 856 linux_select1(struct lwp *l, register_t *retval, int nfds, fd_set *readfds, 857 fd_set *writefds, fd_set *exceptfds, struct linux_timeval *timeout) 858 { 859 struct timespec ts0, ts1, uts, *ts = NULL; 860 struct linux_timeval ltv; 861 int error; 862 863 /* 864 * Store current time for computation of the amount of 865 * time left. 866 */ 867 if (timeout) { 868 if ((error = copyin(timeout, <v, sizeof(ltv)))) 869 return error; 870 uts.tv_sec = ltv.tv_sec; 871 uts.tv_nsec = ltv.tv_usec * 1000; 872 if (itimespecfix(&uts)) { 873 /* 874 * The timeval was invalid. Convert it to something 875 * valid that will act as it does under Linux. 876 */ 877 uts.tv_sec += uts.tv_nsec / 1000000000; 878 uts.tv_nsec %= 1000000000; 879 if (uts.tv_nsec < 0) { 880 uts.tv_sec -= 1; 881 uts.tv_nsec += 1000000000; 882 } 883 if (uts.tv_sec < 0) 884 timespecclear(&uts); 885 } 886 ts = &uts; 887 nanotime(&ts0); 888 } 889 890 error = selcommon(retval, nfds, readfds, writefds, exceptfds, ts, NULL); 891 892 if (error) { 893 /* 894 * See fs/select.c in the Linux kernel. Without this, 895 * Maelstrom doesn't work. 896 */ 897 if (error == ERESTART) 898 error = EINTR; 899 return error; 900 } 901 902 if (timeout) { 903 if (*retval) { 904 /* 905 * Compute how much time was left of the timeout, 906 * by subtracting the current time and the time 907 * before we started the call, and subtracting 908 * that result from the user-supplied value. 909 */ 910 nanotime(&ts1); 911 timespecsub(&ts1, &ts0, &ts1); 912 timespecsub(&uts, &ts1, &uts); 913 if (uts.tv_sec < 0) 914 timespecclear(&uts); 915 } else 916 timespecclear(&uts); 917 ltv.tv_sec = uts.tv_sec; 918 ltv.tv_usec = uts.tv_nsec / 1000; 919 if ((error = copyout(<v, timeout, sizeof(ltv)))) 920 return error; 921 } 922 923 return 0; 924 } 925 926 int 927 linux_sys_ppoll(struct lwp *l, 928 const struct linux_sys_ppoll_args *uap, register_t *retval) 929 { 930 /* { 931 syscallarg(struct pollfd *) fds; 932 syscallarg(u_int) nfds; 933 syscallarg(struct linux_timespec *) timeout; 934 syscallarg(linux_sigset_t *) sigset; 935 } */ 936 struct linux_timespec lts0, *lts; 937 struct timespec ts0, *ts = NULL; 938 linux_sigset_t lsigmask0, *lsigmask; 939 sigset_t sigmask0, *sigmask = NULL; 940 int error; 941 942 lts = SCARG(uap, timeout); 943 if (lts) { 944 if ((error = copyin(lts, <s0, sizeof(lts0))) != 0) 945 return error; 946 linux_to_native_timespec(&ts0, <s0); 947 ts = &ts0; 948 } 949 950 lsigmask = SCARG(uap, sigset); 951 if (lsigmask) { 952 if ((error = copyin(lsigmask, &lsigmask0, sizeof(lsigmask0)))) 953 return error; 954 linux_to_native_sigset(&sigmask0, &lsigmask0); 955 sigmask = &sigmask0; 956 } 957 958 return pollcommon(retval, SCARG(uap, fds), SCARG(uap, nfds), 959 ts, sigmask); 960 } 961 962 /* 963 * Set the 'personality' (emulation mode) for the current process. Only 964 * accept the Linux personality here (0). This call is needed because 965 * the Linux ELF crt0 issues it in an ugly kludge to make sure that 966 * ELF binaries run in Linux mode, not SVR4 mode. 967 */ 968 int 969 linux_sys_personality(struct lwp *l, const struct linux_sys_personality_args *uap, register_t *retval) 970 { 971 /* { 972 syscallarg(unsigned long) per; 973 } */ 974 struct linux_emuldata *led; 975 int per; 976 977 per = SCARG(uap, per); 978 led = l->l_emuldata; 979 if (per == LINUX_PER_QUERY) { 980 retval[0] = led->led_personality; 981 return 0; 982 } 983 984 switch (per & LINUX_PER_MASK) { 985 case LINUX_PER_LINUX: 986 case LINUX_PER_LINUX32: 987 led->led_personality = per; 988 break; 989 990 default: 991 return EINVAL; 992 } 993 994 retval[0] = per; 995 return 0; 996 } 997 998 /* 999 * We have nonexistent fsuid equal to uid. 1000 * If modification is requested, refuse. 1001 */ 1002 int 1003 linux_sys_setfsuid(struct lwp *l, const struct linux_sys_setfsuid_args *uap, register_t *retval) 1004 { 1005 /* { 1006 syscallarg(uid_t) uid; 1007 } */ 1008 uid_t uid; 1009 1010 uid = SCARG(uap, uid); 1011 if (kauth_cred_getuid(l->l_cred) != uid) 1012 return sys_nosys(l, uap, retval); 1013 1014 *retval = uid; 1015 return 0; 1016 } 1017 1018 int 1019 linux_sys_setfsgid(struct lwp *l, const struct linux_sys_setfsgid_args *uap, register_t *retval) 1020 { 1021 /* { 1022 syscallarg(gid_t) gid; 1023 } */ 1024 gid_t gid; 1025 1026 gid = SCARG(uap, gid); 1027 if (kauth_cred_getgid(l->l_cred) != gid) 1028 return sys_nosys(l, uap, retval); 1029 1030 *retval = gid; 1031 return 0; 1032 } 1033 1034 int 1035 linux_sys_setresuid(struct lwp *l, const struct linux_sys_setresuid_args *uap, register_t *retval) 1036 { 1037 /* { 1038 syscallarg(uid_t) ruid; 1039 syscallarg(uid_t) euid; 1040 syscallarg(uid_t) suid; 1041 } */ 1042 1043 /* 1044 * Note: These checks are a little different than the NetBSD 1045 * setreuid(2) call performs. This precisely follows the 1046 * behavior of the Linux kernel. 1047 */ 1048 1049 return do_setresuid(l, SCARG(uap, ruid), SCARG(uap, euid), 1050 SCARG(uap, suid), 1051 ID_R_EQ_R | ID_R_EQ_E | ID_R_EQ_S | 1052 ID_E_EQ_R | ID_E_EQ_E | ID_E_EQ_S | 1053 ID_S_EQ_R | ID_S_EQ_E | ID_S_EQ_S ); 1054 } 1055 1056 int 1057 linux_sys_getresuid(struct lwp *l, const struct linux_sys_getresuid_args *uap, register_t *retval) 1058 { 1059 /* { 1060 syscallarg(uid_t *) ruid; 1061 syscallarg(uid_t *) euid; 1062 syscallarg(uid_t *) suid; 1063 } */ 1064 kauth_cred_t pc = l->l_cred; 1065 int error; 1066 uid_t uid; 1067 1068 /* 1069 * Linux copies these values out to userspace like so: 1070 * 1071 * 1. Copy out ruid. 1072 * 2. If that succeeds, copy out euid. 1073 * 3. If both of those succeed, copy out suid. 1074 */ 1075 uid = kauth_cred_getuid(pc); 1076 if ((error = copyout(&uid, SCARG(uap, ruid), sizeof(uid_t))) != 0) 1077 return (error); 1078 1079 uid = kauth_cred_geteuid(pc); 1080 if ((error = copyout(&uid, SCARG(uap, euid), sizeof(uid_t))) != 0) 1081 return (error); 1082 1083 uid = kauth_cred_getsvuid(pc); 1084 1085 return (copyout(&uid, SCARG(uap, suid), sizeof(uid_t))); 1086 } 1087 1088 int 1089 linux_sys_ptrace(struct lwp *l, const struct linux_sys_ptrace_args *uap, register_t *retval) 1090 { 1091 /* { 1092 i386, m68k, powerpc: T=int 1093 alpha, amd64: T=long 1094 syscallarg(T) request; 1095 syscallarg(T) pid; 1096 syscallarg(T) addr; 1097 syscallarg(T) data; 1098 } */ 1099 const int *ptr; 1100 int request; 1101 int error; 1102 1103 ptr = linux_ptrace_request_map; 1104 request = SCARG(uap, request); 1105 while (*ptr != -1) 1106 if (*ptr++ == request) { 1107 struct sys_ptrace_args pta; 1108 1109 SCARG(&pta, req) = *ptr; 1110 SCARG(&pta, pid) = SCARG(uap, pid); 1111 SCARG(&pta, addr) = (void *)SCARG(uap, addr); 1112 SCARG(&pta, data) = SCARG(uap, data); 1113 1114 /* 1115 * Linux ptrace(PTRACE_CONT, pid, 0, 0) means actually 1116 * to continue where the process left off previously. 1117 * The same thing is achieved by addr == (void *) 1 1118 * on NetBSD, so rewrite 'addr' appropriately. 1119 */ 1120 if (request == LINUX_PTRACE_CONT && SCARG(uap, addr)==0) 1121 SCARG(&pta, addr) = (void *) 1; 1122 1123 error = sysent[SYS_ptrace].sy_call(l, &pta, retval); 1124 if (error) 1125 return error; 1126 switch (request) { 1127 case LINUX_PTRACE_PEEKTEXT: 1128 case LINUX_PTRACE_PEEKDATA: 1129 error = copyout (retval, 1130 (void *)SCARG(uap, data), 1131 sizeof *retval); 1132 *retval = SCARG(uap, data); 1133 break; 1134 default: 1135 break; 1136 } 1137 return error; 1138 } 1139 else 1140 ptr++; 1141 1142 return LINUX_SYS_PTRACE_ARCH(l, uap, retval); 1143 } 1144 1145 int 1146 linux_sys_reboot(struct lwp *l, const struct linux_sys_reboot_args *uap, register_t *retval) 1147 { 1148 /* { 1149 syscallarg(int) magic1; 1150 syscallarg(int) magic2; 1151 syscallarg(int) cmd; 1152 syscallarg(void *) arg; 1153 } */ 1154 struct sys_reboot_args /* { 1155 syscallarg(int) opt; 1156 syscallarg(char *) bootstr; 1157 } */ sra; 1158 int error; 1159 1160 if ((error = kauth_authorize_system(l->l_cred, 1161 KAUTH_SYSTEM_REBOOT, 0, NULL, NULL, NULL)) != 0) 1162 return(error); 1163 1164 if (SCARG(uap, magic1) != LINUX_REBOOT_MAGIC1) 1165 return(EINVAL); 1166 if (SCARG(uap, magic2) != LINUX_REBOOT_MAGIC2 && 1167 SCARG(uap, magic2) != LINUX_REBOOT_MAGIC2A && 1168 SCARG(uap, magic2) != LINUX_REBOOT_MAGIC2B) 1169 return(EINVAL); 1170 1171 switch ((unsigned long)SCARG(uap, cmd)) { 1172 case LINUX_REBOOT_CMD_RESTART: 1173 SCARG(&sra, opt) = RB_AUTOBOOT; 1174 break; 1175 case LINUX_REBOOT_CMD_HALT: 1176 SCARG(&sra, opt) = RB_HALT; 1177 break; 1178 case LINUX_REBOOT_CMD_POWER_OFF: 1179 SCARG(&sra, opt) = RB_HALT|RB_POWERDOWN; 1180 break; 1181 case LINUX_REBOOT_CMD_RESTART2: 1182 /* Reboot with an argument. */ 1183 SCARG(&sra, opt) = RB_AUTOBOOT|RB_STRING; 1184 SCARG(&sra, bootstr) = SCARG(uap, arg); 1185 break; 1186 case LINUX_REBOOT_CMD_CAD_ON: 1187 return(EINVAL); /* We don't implement ctrl-alt-delete */ 1188 case LINUX_REBOOT_CMD_CAD_OFF: 1189 return(0); 1190 default: 1191 return(EINVAL); 1192 } 1193 1194 return(sys_reboot(l, &sra, retval)); 1195 } 1196 1197 /* 1198 * Copy of compat_12_sys_swapon(). 1199 */ 1200 int 1201 linux_sys_swapon(struct lwp *l, const struct linux_sys_swapon_args *uap, register_t *retval) 1202 { 1203 /* { 1204 syscallarg(const char *) name; 1205 } */ 1206 struct sys_swapctl_args ua; 1207 1208 SCARG(&ua, cmd) = SWAP_ON; 1209 SCARG(&ua, arg) = (void *)__UNCONST(SCARG(uap, name)); 1210 SCARG(&ua, misc) = 0; /* priority */ 1211 return (sys_swapctl(l, &ua, retval)); 1212 } 1213 1214 /* 1215 * Stop swapping to the file or block device specified by path. 1216 */ 1217 int 1218 linux_sys_swapoff(struct lwp *l, const struct linux_sys_swapoff_args *uap, register_t *retval) 1219 { 1220 /* { 1221 syscallarg(const char *) path; 1222 } */ 1223 struct sys_swapctl_args ua; 1224 1225 SCARG(&ua, cmd) = SWAP_OFF; 1226 SCARG(&ua, arg) = __UNCONST(SCARG(uap, path)); /*XXXUNCONST*/ 1227 return (sys_swapctl(l, &ua, retval)); 1228 } 1229 1230 /* 1231 * Copy of compat_09_sys_setdomainname() 1232 */ 1233 /* ARGSUSED */ 1234 int 1235 linux_sys_setdomainname(struct lwp *l, const struct linux_sys_setdomainname_args *uap, register_t *retval) 1236 { 1237 /* { 1238 syscallarg(char *) domainname; 1239 syscallarg(int) len; 1240 } */ 1241 int name[2]; 1242 1243 name[0] = CTL_KERN; 1244 name[1] = KERN_DOMAINNAME; 1245 return (old_sysctl(&name[0], 2, 0, 0, SCARG(uap, domainname), 1246 SCARG(uap, len), l)); 1247 } 1248 1249 /* 1250 * sysinfo() 1251 */ 1252 /* ARGSUSED */ 1253 int 1254 linux_sys_sysinfo(struct lwp *l, const struct linux_sys_sysinfo_args *uap, register_t *retval) 1255 { 1256 /* { 1257 syscallarg(struct linux_sysinfo *) arg; 1258 } */ 1259 struct linux_sysinfo si; 1260 struct loadavg *la; 1261 1262 si.uptime = time_uptime; 1263 la = &averunnable; 1264 si.loads[0] = la->ldavg[0] * LINUX_SYSINFO_LOADS_SCALE / la->fscale; 1265 si.loads[1] = la->ldavg[1] * LINUX_SYSINFO_LOADS_SCALE / la->fscale; 1266 si.loads[2] = la->ldavg[2] * LINUX_SYSINFO_LOADS_SCALE / la->fscale; 1267 si.totalram = ctob((u_long)physmem); 1268 si.freeram = (u_long)uvmexp.free * uvmexp.pagesize; 1269 si.sharedram = 0; /* XXX */ 1270 si.bufferram = (u_long)uvmexp.filepages * uvmexp.pagesize; 1271 si.totalswap = (u_long)uvmexp.swpages * uvmexp.pagesize; 1272 si.freeswap = 1273 (u_long)(uvmexp.swpages - uvmexp.swpginuse) * uvmexp.pagesize; 1274 si.procs = nprocs; 1275 1276 /* The following are only present in newer Linux kernels. */ 1277 si.totalbig = 0; 1278 si.freebig = 0; 1279 si.mem_unit = 1; 1280 1281 return (copyout(&si, SCARG(uap, arg), sizeof si)); 1282 } 1283 1284 int 1285 linux_sys_getrlimit(struct lwp *l, const struct linux_sys_getrlimit_args *uap, register_t *retval) 1286 { 1287 /* { 1288 syscallarg(int) which; 1289 # ifdef LINUX_LARGEFILE64 1290 syscallarg(struct rlimit *) rlp; 1291 # else 1292 syscallarg(struct orlimit *) rlp; 1293 # endif 1294 } */ 1295 # ifdef LINUX_LARGEFILE64 1296 struct rlimit orl; 1297 # else 1298 struct orlimit orl; 1299 # endif 1300 int which; 1301 1302 which = linux_to_bsd_limit(SCARG(uap, which)); 1303 if (which < 0) 1304 return -which; 1305 1306 bsd_to_linux_rlimit(&orl, &l->l_proc->p_rlimit[which]); 1307 1308 return copyout(&orl, SCARG(uap, rlp), sizeof(orl)); 1309 } 1310 1311 int 1312 linux_sys_setrlimit(struct lwp *l, const struct linux_sys_setrlimit_args *uap, register_t *retval) 1313 { 1314 /* { 1315 syscallarg(int) which; 1316 # ifdef LINUX_LARGEFILE64 1317 syscallarg(struct rlimit *) rlp; 1318 # else 1319 syscallarg(struct orlimit *) rlp; 1320 # endif 1321 } */ 1322 struct rlimit rl; 1323 # ifdef LINUX_LARGEFILE64 1324 struct rlimit orl; 1325 # else 1326 struct orlimit orl; 1327 # endif 1328 int error; 1329 int which; 1330 1331 if ((error = copyin(SCARG(uap, rlp), &orl, sizeof(orl))) != 0) 1332 return error; 1333 1334 which = linux_to_bsd_limit(SCARG(uap, which)); 1335 if (which < 0) 1336 return -which; 1337 1338 linux_to_bsd_rlimit(&rl, &orl); 1339 return dosetrlimit(l, l->l_proc, which, &rl); 1340 } 1341 1342 # if !defined(__mips__) && !defined(__amd64__) 1343 /* XXX: this doesn't look 100% common, at least mips doesn't have it */ 1344 int 1345 linux_sys_ugetrlimit(struct lwp *l, const struct linux_sys_ugetrlimit_args *uap, register_t *retval) 1346 { 1347 return linux_sys_getrlimit(l, (const void *)uap, retval); 1348 } 1349 # endif 1350 1351 /* 1352 * This gets called for unsupported syscalls. The difference to sys_nosys() 1353 * is that process does not get SIGSYS, the call just returns with ENOSYS. 1354 * This is the way Linux does it and glibc depends on this behaviour. 1355 */ 1356 int 1357 linux_sys_nosys(struct lwp *l, const void *v, register_t *retval) 1358 { 1359 return (ENOSYS); 1360 } 1361 1362 int 1363 linux_sys_getpriority(struct lwp *l, const struct linux_sys_getpriority_args *uap, register_t *retval) 1364 { 1365 /* { 1366 syscallarg(int) which; 1367 syscallarg(int) who; 1368 } */ 1369 struct sys_getpriority_args bsa; 1370 int error; 1371 1372 SCARG(&bsa, which) = SCARG(uap, which); 1373 SCARG(&bsa, who) = SCARG(uap, who); 1374 1375 if ((error = sys_getpriority(l, &bsa, retval))) 1376 return error; 1377 1378 *retval = NZERO - *retval; 1379 1380 return 0; 1381 } 1382 1383 int 1384 linux_do_sys_utimensat(struct lwp *l, int fd, const char *path, struct timespec *tsp, int flags, register_t *retval) 1385 { 1386 int follow, error; 1387 1388 follow = (flags & LINUX_AT_SYMLINK_NOFOLLOW) ? NOFOLLOW : FOLLOW; 1389 1390 if (path == NULL && fd != AT_FDCWD) { 1391 file_t *fp; 1392 1393 /* fd_getvnode() will use the descriptor for us */ 1394 if ((error = fd_getvnode(fd, &fp)) != 0) 1395 return error; 1396 error = do_sys_utimensat(l, AT_FDCWD, fp->f_data, NULL, 0, 1397 tsp, UIO_SYSSPACE); 1398 fd_putfile(fd); 1399 return error; 1400 } 1401 1402 return do_sys_utimensat(l, fd, NULL, path, follow, tsp, UIO_SYSSPACE); 1403 } 1404 1405 int 1406 linux_sys_utimensat(struct lwp *l, const struct linux_sys_utimensat_args *uap, 1407 register_t *retval) 1408 { 1409 /* { 1410 syscallarg(int) fd; 1411 syscallarg(const char *) path; 1412 syscallarg(const struct linux_timespec *) times; 1413 syscallarg(int) flag; 1414 } */ 1415 int error; 1416 struct linux_timespec lts[2]; 1417 struct timespec *tsp = NULL, ts[2]; 1418 1419 if (SCARG(uap, times)) { 1420 error = copyin(SCARG(uap, times), <s, sizeof(lts)); 1421 if (error != 0) 1422 return error; 1423 linux_to_native_timespec(&ts[0], <s[0]); 1424 linux_to_native_timespec(&ts[1], <s[1]); 1425 tsp = ts; 1426 } 1427 1428 return linux_do_sys_utimensat(l, SCARG(uap, fd), SCARG(uap, path), 1429 tsp, SCARG(uap, flag), retval); 1430 } 1431