priv: Remove some uses of PRIV_ROOT/PRISON_ROOT
[dragonfly.git] / sys / emulation / linux / linux_misc.c
blob12f2e39f11aac4a6c6f70d5b9ee794ba2bf71a32
1 /*-
2 * Copyright (c) 1994-1995 Søren Schmidt
3 * All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer
10 * in this position and unchanged.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 * 3. The name of the author may not be used to endorse or promote products
15 * derived from this software withough specific prior written permission
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 * $FreeBSD: src/sys/compat/linux/linux_misc.c,v 1.85.2.9 2002/09/24 08:11:41 mdodd Exp $
29 * $DragonFly: src/sys/emulation/linux/linux_misc.c,v 1.39 2007/06/26 19:31:03 dillon Exp $
32 #include "opt_compat.h"
34 #include <sys/param.h>
35 #include <sys/systm.h>
36 #include <sys/fcntl.h>
37 #include <sys/imgact_aout.h>
38 #include <sys/kernel.h>
39 #include <sys/kern_syscall.h>
40 #include <sys/lock.h>
41 #include <sys/mman.h>
42 #include <sys/mount.h>
43 #include <sys/poll.h>
44 #include <sys/proc.h>
45 #include <sys/priv.h>
46 #include <sys/nlookup.h>
47 #include <sys/blist.h>
48 #include <sys/reboot.h>
49 #include <sys/resourcevar.h>
50 #include <sys/signalvar.h>
51 #include <sys/signal2.h>
52 #include <sys/stat.h>
53 #include <sys/sysctl.h>
54 #include <sys/sysproto.h>
55 #include <sys/time.h>
56 #include <sys/unistd.h>
57 #include <sys/vmmeter.h>
58 #include <sys/vnode.h>
59 #include <sys/wait.h>
60 #include <sys/thread2.h>
62 #include <vm/vm.h>
63 #include <vm/pmap.h>
64 #include <vm/vm_kern.h>
65 #include <vm/vm_map.h>
66 #include <vm/vm_extern.h>
67 #include <vm/vm_object.h>
68 #include <vm/vm_zone.h>
69 #include <vm/swap_pager.h>
71 #include <machine/frame.h>
72 #include <machine/limits.h>
73 #include <machine/psl.h>
74 #include <machine/sysarch.h>
75 #ifdef __i386__
76 #include <machine/segments.h>
77 #endif
79 #include <sys/sched.h>
81 #include <emulation/linux/linux_sysproto.h>
82 #include <arch_linux/linux.h>
83 #include <arch_linux/linux_proto.h>
84 #include "linux_mib.h"
85 #include "linux_util.h"
87 #define BSD_TO_LINUX_SIGNAL(sig) \
88 (((sig) <= LINUX_SIGTBLSZ) ? bsd_to_linux_signal[_SIG_IDX(sig)] : sig)
90 static unsigned int linux_to_bsd_resource[LINUX_RLIM_NLIMITS] = {
91 RLIMIT_CPU, RLIMIT_FSIZE, RLIMIT_DATA, RLIMIT_STACK,
92 RLIMIT_CORE, RLIMIT_RSS, RLIMIT_NPROC, RLIMIT_NOFILE,
93 RLIMIT_MEMLOCK, -1
96 struct l_sysinfo {
97 l_long uptime; /* Seconds since boot */
98 l_ulong loads[3]; /* 1, 5, and 15 minute load averages */
99 l_ulong totalram; /* Total usable main memory size */
100 l_ulong freeram; /* Available memory size */
101 l_ulong sharedram; /* Amount of shared memory */
102 l_ulong bufferram; /* Memory used by buffers */
103 l_ulong totalswap; /* Total swap space size */
104 l_ulong freeswap; /* swap space still available */
105 l_ushort procs; /* Number of current processes */
106 char _f[22]; /* Pads structure to 64 bytes */
110 sys_linux_sysinfo(struct linux_sysinfo_args *args)
112 struct l_sysinfo sysinfo;
113 vm_object_t object;
114 int i;
115 struct timespec ts;
117 /* Uptime is copied out of print_uptime() in kern_shutdown.c */
118 getnanouptime(&ts);
119 i = 0;
120 if (ts.tv_sec >= 86400) {
121 ts.tv_sec %= 86400;
122 i = 1;
124 if (i || ts.tv_sec >= 3600) {
125 ts.tv_sec %= 3600;
126 i = 1;
128 if (i || ts.tv_sec >= 60) {
129 ts.tv_sec %= 60;
130 i = 1;
132 sysinfo.uptime=ts.tv_sec;
134 /* Use the information from the mib to get our load averages */
135 for (i = 0; i < 3; i++)
136 sysinfo.loads[i] = averunnable.ldavg[i];
138 sysinfo.totalram = Maxmem * PAGE_SIZE;
139 sysinfo.freeram = sysinfo.totalram - vmstats.v_wire_count * PAGE_SIZE;
141 sysinfo.sharedram = 0;
142 for (object = TAILQ_FIRST(&vm_object_list); object != NULL;
143 object = TAILQ_NEXT(object, object_list))
144 if (object->shadow_count > 1)
145 sysinfo.sharedram += object->resident_page_count;
147 sysinfo.sharedram *= PAGE_SIZE;
148 sysinfo.bufferram = 0;
150 if (swapblist == NULL) {
151 sysinfo.totalswap= 0;
152 sysinfo.freeswap = 0;
153 } else {
154 sysinfo.totalswap = swapblist->bl_blocks * 1024;
155 sysinfo.freeswap = swapblist->bl_root->u.bmu_avail * PAGE_SIZE;
158 sysinfo.procs = 20; /* Hack */
160 return copyout(&sysinfo, (caddr_t)args->info, sizeof(sysinfo));
164 sys_linux_alarm(struct linux_alarm_args *args)
166 struct thread *td = curthread;
167 struct proc *p = td->td_proc;
168 struct itimerval it, old_it;
169 struct timeval tv;
171 KKASSERT(p);
173 #ifdef DEBUG
174 if (ldebug(alarm))
175 kprintf(ARGS(alarm, "%u"), args->secs);
176 #endif
178 if (args->secs > 100000000)
179 return EINVAL;
181 it.it_value.tv_sec = (long)args->secs;
182 it.it_value.tv_usec = 0;
183 it.it_interval.tv_sec = 0;
184 it.it_interval.tv_usec = 0;
185 crit_enter();
186 old_it = p->p_realtimer;
187 getmicrouptime(&tv);
188 if (timevalisset(&old_it.it_value))
189 callout_stop(&p->p_ithandle);
190 if (it.it_value.tv_sec != 0) {
191 callout_reset(&p->p_ithandle, tvtohz_high(&it.it_value),
192 realitexpire, p);
193 timevaladd(&it.it_value, &tv);
195 p->p_realtimer = it;
196 crit_exit();
197 if (timevalcmp(&old_it.it_value, &tv, >)) {
198 timevalsub(&old_it.it_value, &tv);
199 if (old_it.it_value.tv_usec != 0)
200 old_it.it_value.tv_sec++;
201 args->sysmsg_result = old_it.it_value.tv_sec;
203 return 0;
207 sys_linux_brk(struct linux_brk_args *args)
209 struct thread *td = curthread;
210 struct proc *p = td->td_proc;
211 struct vmspace *vm;
212 vm_offset_t new, old;
213 struct obreak_args bsd_args;
215 KKASSERT(p);
216 vm = p->p_vmspace;
217 #ifdef DEBUG
218 if (ldebug(brk))
219 kprintf(ARGS(brk, "%p"), (void *)args->dsend);
220 #endif
221 old = (vm_offset_t)vm->vm_daddr + ctob(vm->vm_dsize);
222 new = (vm_offset_t)args->dsend;
223 bsd_args.sysmsg_result = 0;
224 bsd_args.nsize = (char *) new;
225 bsd_args.sysmsg_result = 0;
226 if (((caddr_t)new > vm->vm_daddr) && !sys_obreak(&bsd_args))
227 args->sysmsg_result = (long)new;
228 else
229 args->sysmsg_result = (long)old;
231 return 0;
235 sys_linux_uselib(struct linux_uselib_args *args)
237 struct thread *td = curthread;
238 struct proc *p;
239 struct nlookupdata nd;
240 struct vnode *vp;
241 struct exec *a_out;
242 struct vattr attr;
243 vm_offset_t vmaddr;
244 unsigned long file_offset;
245 vm_offset_t buffer;
246 unsigned long bss_size;
247 int error;
248 int locked;
249 char *path;
251 KKASSERT(td->td_proc);
252 p = td->td_proc;
254 error = linux_copyin_path(args->library, &path, LINUX_PATH_EXISTS);
255 if (error)
256 return (error);
257 #ifdef DEBUG
258 if (ldebug(uselib))
259 kprintf(ARGS(uselib, "%s"), path);
260 #endif
262 a_out = NULL;
263 locked = 0;
264 vp = NULL;
266 error = nlookup_init(&nd, path, UIO_SYSSPACE, NLC_FOLLOW);
267 nd.nl_flags |= NLC_EXEC;
268 if (error == 0)
269 error = nlookup(&nd);
270 if (error == 0)
271 error = cache_vget(&nd.nl_nch, nd.nl_cred, LK_EXCLUSIVE, &vp);
272 if (error)
273 goto cleanup;
275 * From here on down, we have a locked vnode that must be unlocked.
277 locked = 1;
279 /* Writable? */
280 if (vp->v_writecount) {
281 error = ETXTBSY;
282 goto cleanup;
285 /* Executable? */
286 error = VOP_GETATTR(vp, &attr);
287 if (error)
288 goto cleanup;
290 if ((vp->v_mount->mnt_flag & MNT_NOEXEC) ||
291 ((attr.va_mode & 0111) == 0) || (attr.va_type != VREG)) {
292 error = ENOEXEC;
293 goto cleanup;
296 /* Sensible size? */
297 if (attr.va_size == 0) {
298 error = ENOEXEC;
299 goto cleanup;
302 error = VOP_OPEN(vp, FREAD, p->p_ucred, NULL);
303 if (error)
304 goto cleanup;
307 * Lock no longer needed
309 vn_unlock(vp);
310 locked = 0;
312 /* Pull in executable header into kernel_map */
313 error = vm_mmap(&kernel_map, (vm_offset_t *)&a_out, PAGE_SIZE,
314 VM_PROT_READ, VM_PROT_READ, 0, (caddr_t)vp, 0);
315 if (error)
316 goto cleanup;
318 /* Is it a Linux binary ? */
319 if (((a_out->a_magic >> 16) & 0xff) != 0x64) {
320 error = ENOEXEC;
321 goto cleanup;
325 * While we are here, we should REALLY do some more checks
328 /* Set file/virtual offset based on a.out variant. */
329 switch ((int)(a_out->a_magic & 0xffff)) {
330 case 0413: /* ZMAGIC */
331 file_offset = 1024;
332 break;
333 case 0314: /* QMAGIC */
334 file_offset = 0;
335 break;
336 default:
337 error = ENOEXEC;
338 goto cleanup;
341 bss_size = round_page(a_out->a_bss);
343 /* Check various fields in header for validity/bounds. */
344 if (a_out->a_text & PAGE_MASK || a_out->a_data & PAGE_MASK) {
345 error = ENOEXEC;
346 goto cleanup;
349 /* text + data can't exceed file size */
350 if (a_out->a_data + a_out->a_text > attr.va_size) {
351 error = EFAULT;
352 goto cleanup;
356 * text/data/bss must not exceed limits
357 * XXX - this is not complete. it should check current usage PLUS
358 * the resources needed by this library.
360 if (a_out->a_text > maxtsiz ||
361 a_out->a_data + bss_size > p->p_rlimit[RLIMIT_DATA].rlim_cur) {
362 error = ENOMEM;
363 goto cleanup;
366 /* prevent more writers */
367 vp->v_flag |= VTEXT;
370 * Check if file_offset page aligned. Currently we cannot handle
371 * misalinged file offsets, and so we read in the entire image
372 * (what a waste).
374 if (file_offset & PAGE_MASK) {
375 #ifdef DEBUG
376 kprintf("uselib: Non page aligned binary %lu\n", file_offset);
377 #endif
378 /* Map text+data read/write/execute */
380 /* a_entry is the load address and is page aligned */
381 vmaddr = trunc_page(a_out->a_entry);
383 /* get anon user mapping, read+write+execute */
384 error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0,
385 &vmaddr, a_out->a_text + a_out->a_data,
386 FALSE,
387 VM_MAPTYPE_NORMAL,
388 VM_PROT_ALL, VM_PROT_ALL,
390 if (error)
391 goto cleanup;
393 /* map file into kernel_map */
394 error = vm_mmap(&kernel_map, &buffer,
395 round_page(a_out->a_text + a_out->a_data + file_offset),
396 VM_PROT_READ, VM_PROT_READ, 0, (caddr_t)vp,
397 trunc_page(file_offset));
398 if (error)
399 goto cleanup;
401 /* copy from kernel VM space to user space */
402 error = copyout((caddr_t)(uintptr_t)(buffer + file_offset),
403 (caddr_t)vmaddr, a_out->a_text + a_out->a_data);
405 /* release temporary kernel space */
406 vm_map_remove(&kernel_map, buffer, buffer +
407 round_page(a_out->a_text + a_out->a_data + file_offset));
409 if (error)
410 goto cleanup;
411 } else {
412 #ifdef DEBUG
413 kprintf("uselib: Page aligned binary %lu\n", file_offset);
414 #endif
416 * for QMAGIC, a_entry is 20 bytes beyond the load address
417 * to skip the executable header
419 vmaddr = trunc_page(a_out->a_entry);
422 * Map it all into the process's space as a single
423 * copy-on-write "data" segment.
425 error = vm_mmap(&p->p_vmspace->vm_map, &vmaddr,
426 a_out->a_text + a_out->a_data, VM_PROT_ALL, VM_PROT_ALL,
427 MAP_PRIVATE | MAP_FIXED, (caddr_t)vp, file_offset);
428 if (error)
429 goto cleanup;
431 #ifdef DEBUG
432 kprintf("mem=%08lx = %08lx %08lx\n", (long)vmaddr, ((long*)vmaddr)[0],
433 ((long*)vmaddr)[1]);
434 #endif
435 if (bss_size != 0) {
436 /* Calculate BSS start address */
437 vmaddr = trunc_page(a_out->a_entry) + a_out->a_text +
438 a_out->a_data;
440 /* allocate some 'anon' space */
441 error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0,
442 &vmaddr, bss_size,
443 FALSE,
444 VM_MAPTYPE_NORMAL,
445 VM_PROT_ALL, VM_PROT_ALL,
447 if (error)
448 goto cleanup;
451 cleanup:
452 /* Unlock/release vnode */
453 if (vp) {
454 if (locked)
455 vn_unlock(vp);
456 vrele(vp);
458 /* Release the kernel mapping. */
459 if (a_out) {
460 vm_map_remove(&kernel_map, (vm_offset_t)a_out,
461 (vm_offset_t)a_out + PAGE_SIZE);
463 nlookup_done(&nd);
464 linux_free_path(&path);
465 return (error);
469 sys_linux_select(struct linux_select_args *args)
471 struct select_args bsa;
472 struct timeval tv0, tv1, utv, *tvp;
473 caddr_t sg;
474 int error;
476 #ifdef DEBUG
477 if (ldebug(select))
478 kprintf(ARGS(select, "%d, %p, %p, %p, %p"), args->nfds,
479 (void *)args->readfds, (void *)args->writefds,
480 (void *)args->exceptfds, (void *)args->timeout);
481 #endif
483 error = 0;
484 bsa.sysmsg_result = 0;
485 bsa.nd = args->nfds;
486 bsa.in = args->readfds;
487 bsa.ou = args->writefds;
488 bsa.ex = args->exceptfds;
489 bsa.tv = (struct timeval *)args->timeout;
492 * Store current time for computation of the amount of
493 * time left.
495 if (args->timeout) {
496 if ((error = copyin((caddr_t)args->timeout, &utv,
497 sizeof(utv))))
498 goto select_out;
499 #ifdef DEBUG
500 if (ldebug(select))
501 kprintf(LMSG("incoming timeout (%ld/%ld)"),
502 utv.tv_sec, utv.tv_usec);
503 #endif
505 if (itimerfix(&utv)) {
507 * The timeval was invalid. Convert it to something
508 * valid that will act as it does under Linux.
510 sg = stackgap_init();
511 tvp = stackgap_alloc(&sg, sizeof(utv));
512 utv.tv_sec += utv.tv_usec / 1000000;
513 utv.tv_usec %= 1000000;
514 if (utv.tv_usec < 0) {
515 utv.tv_sec -= 1;
516 utv.tv_usec += 1000000;
518 if (utv.tv_sec < 0)
519 timevalclear(&utv);
520 if ((error = copyout(&utv, tvp, sizeof(utv))))
521 goto select_out;
522 bsa.tv = tvp;
524 microtime(&tv0);
527 error = sys_select(&bsa);
528 args->sysmsg_result = bsa.sysmsg_result;
529 #ifdef DEBUG
530 if (ldebug(select))
531 kprintf(LMSG("real select returns %d"), error);
532 #endif
533 if (error) {
535 * See fs/select.c in the Linux kernel. Without this,
536 * Maelstrom doesn't work.
538 if (error == ERESTART)
539 error = EINTR;
540 goto select_out;
543 if (args->timeout) {
544 if (args->sysmsg_result) {
546 * Compute how much time was left of the timeout,
547 * by subtracting the current time and the time
548 * before we started the call, and subtracting
549 * that result from the user-supplied value.
551 microtime(&tv1);
552 timevalsub(&tv1, &tv0);
553 timevalsub(&utv, &tv1);
554 if (utv.tv_sec < 0)
555 timevalclear(&utv);
556 } else
557 timevalclear(&utv);
558 #ifdef DEBUG
559 if (ldebug(select))
560 kprintf(LMSG("outgoing timeout (%ld/%ld)"),
561 utv.tv_sec, utv.tv_usec);
562 #endif
563 if ((error = copyout(&utv, (caddr_t)args->timeout,
564 sizeof(utv))))
565 goto select_out;
568 select_out:
569 #ifdef DEBUG
570 if (ldebug(select))
571 kprintf(LMSG("select_out -> %d"), error);
572 #endif
573 return error;
576 int
577 sys_linux_mremap(struct linux_mremap_args *args)
579 struct munmap_args bsd_args;
580 int error = 0;
582 #ifdef DEBUG
583 if (ldebug(mremap))
584 kprintf(ARGS(mremap, "%p, %08lx, %08lx, %08lx"),
585 (void *)args->addr,
586 (unsigned long)args->old_len,
587 (unsigned long)args->new_len,
588 (unsigned long)args->flags);
589 #endif
590 args->new_len = round_page(args->new_len);
591 args->old_len = round_page(args->old_len);
593 if (args->new_len > args->old_len) {
594 args->sysmsg_result = 0;
595 return ENOMEM;
598 if (args->new_len < args->old_len) {
599 bsd_args.sysmsg_result = 0;
600 bsd_args.addr = (caddr_t)(args->addr + args->new_len);
601 bsd_args.len = args->old_len - args->new_len;
602 error = sys_munmap(&bsd_args);
605 args->sysmsg_resultp = error ? NULL : (void *)args->addr;
606 return error;
609 #define LINUX_MS_ASYNC 0x0001
610 #define LINUX_MS_INVALIDATE 0x0002
611 #define LINUX_MS_SYNC 0x0004
614 sys_linux_msync(struct linux_msync_args *args)
616 struct msync_args bsd_args;
617 int error;
619 bsd_args.addr = (caddr_t)args->addr;
620 bsd_args.len = args->len;
621 bsd_args.flags = args->fl & ~LINUX_MS_SYNC;
622 bsd_args.sysmsg_result = 0;
624 error = sys_msync(&bsd_args);
625 args->sysmsg_result = bsd_args.sysmsg_result;
626 return(error);
630 sys_linux_time(struct linux_time_args *args)
632 struct timeval tv;
633 l_time_t tm;
634 int error;
636 #ifdef DEBUG
637 if (ldebug(time))
638 kprintf(ARGS(time, "*"));
639 #endif
641 microtime(&tv);
642 tm = tv.tv_sec;
643 if (args->tm && (error = copyout(&tm, (caddr_t)args->tm, sizeof(tm))))
644 return error;
645 args->sysmsg_lresult = tm;
646 return 0;
649 struct l_times_argv {
650 l_long tms_utime;
651 l_long tms_stime;
652 l_long tms_cutime;
653 l_long tms_cstime;
656 #define CLK_TCK 100 /* Linux uses 100 */
658 #define CONVTCK(r) (r.tv_sec * CLK_TCK + r.tv_usec / (1000000 / CLK_TCK))
661 sys_linux_times(struct linux_times_args *args)
663 struct thread *td = curthread;
664 struct proc *p = td->td_proc;
665 struct timeval tv;
666 struct l_times_argv tms;
667 struct rusage ru;
668 int error;
670 KKASSERT(p);
671 #ifdef DEBUG
672 if (ldebug(times))
673 kprintf(ARGS(times, "*"));
674 #endif
676 calcru_proc(p, &ru);
678 tms.tms_utime = CONVTCK(ru.ru_utime);
679 tms.tms_stime = CONVTCK(ru.ru_stime);
681 tms.tms_cutime = CONVTCK(p->p_cru.ru_utime);
682 tms.tms_cstime = CONVTCK(p->p_cru.ru_stime);
684 if ((error = copyout(&tms, (caddr_t)args->buf, sizeof(tms))))
685 return error;
687 microuptime(&tv);
688 args->sysmsg_result = (int)CONVTCK(tv);
689 return 0;
693 sys_linux_newuname(struct linux_newuname_args *args)
695 struct thread *td = curthread;
696 struct l_new_utsname utsname;
697 char *osrelease, *osname;
699 #ifdef DEBUG
700 if (ldebug(newuname))
701 kprintf(ARGS(newuname, "*"));
702 #endif
704 osname = linux_get_osname(td);
705 osrelease = linux_get_osrelease(td);
707 bzero(&utsname, sizeof(utsname));
708 strncpy(utsname.sysname, osname, LINUX_MAX_UTSNAME-1);
709 strncpy(utsname.nodename, hostname, LINUX_MAX_UTSNAME-1);
710 strncpy(utsname.release, osrelease, LINUX_MAX_UTSNAME-1);
711 strncpy(utsname.version, version, LINUX_MAX_UTSNAME-1);
712 strncpy(utsname.machine, machine, LINUX_MAX_UTSNAME-1);
713 strncpy(utsname.domainname, domainname, LINUX_MAX_UTSNAME-1);
715 return (copyout(&utsname, (caddr_t)args->buf, sizeof(utsname)));
718 #if defined(__i386__)
719 struct l_utimbuf {
720 l_time_t l_actime;
721 l_time_t l_modtime;
725 sys_linux_utime(struct linux_utime_args *args)
727 struct timeval tv[2];
728 struct l_utimbuf lut;
729 struct nlookupdata nd;
730 char *path;
731 int error;
733 error = linux_copyin_path(args->fname, &path, LINUX_PATH_EXISTS);
734 if (error)
735 return (error);
736 #ifdef DEBUG
737 if (ldebug(utime))
738 kprintf(ARGS(utime, "%s, *"), path);
739 #endif
741 if (args->times) {
742 error = copyin(args->times, &lut, sizeof(lut));
743 if (error)
744 goto cleanup;
745 tv[0].tv_sec = lut.l_actime;
746 tv[0].tv_usec = 0;
747 tv[1].tv_sec = lut.l_modtime;
748 tv[1].tv_usec = 0;
750 error = nlookup_init(&nd, path, UIO_SYSSPACE, NLC_FOLLOW);
751 if (error == 0)
752 error = kern_utimes(&nd, args->times ? tv : NULL);
753 nlookup_done(&nd);
754 cleanup:
755 linux_free_path(&path);
756 return (error);
758 #endif /* __i386__ */
760 #define __WCLONE 0x80000000
763 sys_linux_waitpid(struct linux_waitpid_args *args)
765 int error, options, status;
767 #ifdef DEBUG
768 if (ldebug(waitpid))
769 kprintf(ARGS(waitpid, "%d, %p, %d"),
770 args->pid, (void *)args->status, args->options);
771 #endif
772 options = args->options & (WNOHANG | WUNTRACED);
773 /* WLINUXCLONE should be equal to __WCLONE, but we make sure */
774 if (args->options & __WCLONE)
775 options |= WLINUXCLONE;
777 error = kern_wait(args->pid, args->status ? &status : NULL, options,
778 NULL, &args->sysmsg_result);
780 if (error == 0 && args->status) {
781 status &= 0xffff;
782 if (WIFSIGNALED(status))
783 status = (status & 0xffffff80) |
784 BSD_TO_LINUX_SIGNAL(WTERMSIG(status));
785 else if (WIFSTOPPED(status))
786 status = (status & 0xffff00ff) |
787 (BSD_TO_LINUX_SIGNAL(WSTOPSIG(status)) << 8);
788 error = copyout(&status, args->status, sizeof(status));
791 return (error);
795 sys_linux_wait4(struct linux_wait4_args *args)
797 struct thread *td = curthread;
798 struct lwp *lp = td->td_lwp;
799 struct rusage rusage;
800 int error, options, status;
802 KKASSERT(lp);
804 #ifdef DEBUG
805 if (ldebug(wait4))
806 kprintf(ARGS(wait4, "%d, %p, %d, %p"),
807 args->pid, (void *)args->status, args->options,
808 (void *)args->rusage);
809 #endif
810 options = args->options & (WNOHANG | WUNTRACED);
811 /* WLINUXCLONE should be equal to __WCLONE, but we make sure */
812 if (args->options & __WCLONE)
813 options |= WLINUXCLONE;
815 error = kern_wait(args->pid, args->status ? &status : NULL, options,
816 args->rusage ? &rusage : NULL, &args->sysmsg_result);
818 if (error == 0)
819 lwp_delsig(lp, SIGCHLD);
821 if (error == 0 && args->status) {
822 status &= 0xffff;
823 if (WIFSIGNALED(status))
824 status = (status & 0xffffff80) |
825 BSD_TO_LINUX_SIGNAL(WTERMSIG(status));
826 else if (WIFSTOPPED(status))
827 status = (status & 0xffff00ff) |
828 (BSD_TO_LINUX_SIGNAL(WSTOPSIG(status)) << 8);
829 error = copyout(&status, args->status, sizeof(status));
831 if (error == 0 && args->rusage)
832 error = copyout(&rusage, args->rusage, sizeof(rusage));
834 return (error);
838 sys_linux_mknod(struct linux_mknod_args *args)
840 struct nlookupdata nd;
841 char *path;
842 int error;
844 error = linux_copyin_path(args->path, &path, LINUX_PATH_CREATE);
845 if (error)
846 return (error);
847 #ifdef DEBUG
848 if (ldebug(mknod))
849 kprintf(ARGS(mknod, "%s, %d, %d"),
850 path, args->mode, args->dev);
851 #endif
852 error = nlookup_init(&nd, path, UIO_SYSSPACE, 0);
853 if (error == 0) {
854 if (args->mode & S_IFIFO) {
855 error = kern_mkfifo(&nd, args->mode);
856 } else {
857 error = kern_mknod(&nd, args->mode,
858 umajor(args->dev),
859 uminor(args->dev));
862 nlookup_done(&nd);
864 linux_free_path(&path);
865 return(error);
869 * UGH! This is just about the dumbest idea I've ever heard!!
872 sys_linux_personality(struct linux_personality_args *args)
874 #ifdef DEBUG
875 if (ldebug(personality))
876 kprintf(ARGS(personality, "%d"), args->per);
877 #endif
878 if (args->per != 0)
879 return EINVAL;
881 /* Yes Jim, it's still a Linux... */
882 args->sysmsg_result = 0;
883 return 0;
887 * Wrappers for get/setitimer for debugging..
890 sys_linux_setitimer(struct linux_setitimer_args *args)
892 struct setitimer_args bsa;
893 struct itimerval foo;
894 int error;
896 #ifdef DEBUG
897 if (ldebug(setitimer))
898 kprintf(ARGS(setitimer, "%p, %p"),
899 (void *)args->itv, (void *)args->oitv);
900 #endif
901 bsa.which = args->which;
902 bsa.itv = (struct itimerval *)args->itv;
903 bsa.oitv = (struct itimerval *)args->oitv;
904 bsa.sysmsg_result = 0;
905 if (args->itv) {
906 if ((error = copyin((caddr_t)args->itv, &foo, sizeof(foo))))
907 return error;
908 #ifdef DEBUG
909 if (ldebug(setitimer)) {
910 kprintf("setitimer: value: sec: %ld, usec: %ld\n",
911 foo.it_value.tv_sec, foo.it_value.tv_usec);
912 kprintf("setitimer: interval: sec: %ld, usec: %ld\n",
913 foo.it_interval.tv_sec, foo.it_interval.tv_usec);
915 #endif
917 error = sys_setitimer(&bsa);
918 args->sysmsg_result = bsa.sysmsg_result;
919 return(error);
923 sys_linux_getitimer(struct linux_getitimer_args *args)
925 struct getitimer_args bsa;
926 int error;
927 #ifdef DEBUG
928 if (ldebug(getitimer))
929 kprintf(ARGS(getitimer, "%p"), (void *)args->itv);
930 #endif
931 bsa.which = args->which;
932 bsa.itv = (struct itimerval *)args->itv;
933 bsa.sysmsg_result = 0;
934 error = sys_getitimer(&bsa);
935 args->sysmsg_result = bsa.sysmsg_result;
936 return(error);
940 sys_linux_nice(struct linux_nice_args *args)
942 struct setpriority_args bsd_args;
943 int error;
945 bsd_args.which = PRIO_PROCESS;
946 bsd_args.who = 0; /* current process */
947 bsd_args.prio = args->inc;
948 bsd_args.sysmsg_result = 0;
949 error = sys_setpriority(&bsd_args);
950 args->sysmsg_result = bsd_args.sysmsg_result;
951 return(error);
955 sys_linux_setgroups(struct linux_setgroups_args *args)
957 struct thread *td = curthread;
958 struct proc *p = td->td_proc;
959 struct ucred *newcred, *oldcred;
960 l_gid_t linux_gidset[NGROUPS];
961 gid_t *bsd_gidset;
962 int ngrp, error;
964 KKASSERT(p);
966 ngrp = args->gidsetsize;
967 oldcred = p->p_ucred;
970 * cr_groups[0] holds egid. Setting the whole set from
971 * the supplied set will cause egid to be changed too.
972 * Keep cr_groups[0] unchanged to prevent that.
975 if ((error = priv_check_cred(oldcred, PRIV_CRED_SETGROUPS, 0)) != 0)
976 return (error);
978 if (ngrp >= NGROUPS)
979 return (EINVAL);
981 newcred = crdup(oldcred);
982 if (ngrp > 0) {
983 error = copyin((caddr_t)args->grouplist, linux_gidset,
984 ngrp * sizeof(l_gid_t));
985 if (error)
986 return (error);
988 newcred->cr_ngroups = ngrp + 1;
990 bsd_gidset = newcred->cr_groups;
991 ngrp--;
992 while (ngrp >= 0) {
993 bsd_gidset[ngrp + 1] = linux_gidset[ngrp];
994 ngrp--;
996 } else {
997 newcred->cr_ngroups = 1;
1000 setsugid();
1001 p->p_ucred = newcred;
1002 crfree(oldcred);
1003 return (0);
1007 sys_linux_getgroups(struct linux_getgroups_args *args)
1009 struct thread *td = curthread;
1010 struct proc *p = td->td_proc;
1011 struct ucred *cred;
1012 l_gid_t linux_gidset[NGROUPS];
1013 gid_t *bsd_gidset;
1014 int bsd_gidsetsz, ngrp, error;
1016 KKASSERT(p);
1018 cred = p->p_ucred;
1019 bsd_gidset = cred->cr_groups;
1020 bsd_gidsetsz = cred->cr_ngroups - 1;
1023 * cr_groups[0] holds egid. Returning the whole set
1024 * here will cause a duplicate. Exclude cr_groups[0]
1025 * to prevent that.
1028 if ((ngrp = args->gidsetsize) == 0) {
1029 args->sysmsg_result = bsd_gidsetsz;
1030 return (0);
1033 if (ngrp < bsd_gidsetsz)
1034 return (EINVAL);
1036 ngrp = 0;
1037 while (ngrp < bsd_gidsetsz) {
1038 linux_gidset[ngrp] = bsd_gidset[ngrp + 1];
1039 ngrp++;
1042 if ((error = copyout(linux_gidset, (caddr_t)args->grouplist,
1043 ngrp * sizeof(l_gid_t))))
1044 return (error);
1046 args->sysmsg_result = ngrp;
1047 return (0);
1051 sys_linux_setrlimit(struct linux_setrlimit_args *args)
1053 struct l_rlimit linux_rlim;
1054 struct rlimit rlim;
1055 u_int which;
1056 int error;
1058 #ifdef DEBUG
1059 if (ldebug(setrlimit))
1060 kprintf(ARGS(setrlimit, "%d, %p"),
1061 args->resource, (void *)args->rlim);
1062 #endif
1063 if (args->resource >= LINUX_RLIM_NLIMITS)
1064 return (EINVAL);
1065 which = linux_to_bsd_resource[args->resource];
1066 if (which == -1)
1067 return (EINVAL);
1069 error = copyin(args->rlim, &linux_rlim, sizeof(linux_rlim));
1070 if (error)
1071 return (error);
1072 rlim.rlim_cur = (rlim_t)linux_rlim.rlim_cur;
1073 rlim.rlim_max = (rlim_t)linux_rlim.rlim_max;
1075 error = kern_setrlimit(which, &rlim);
1077 return(error);
1081 sys_linux_old_getrlimit(struct linux_old_getrlimit_args *args)
1083 struct l_rlimit linux_rlim;
1084 struct rlimit rlim;
1085 u_int which;
1086 int error;
1088 #ifdef DEBUG
1089 if (ldebug(old_getrlimit))
1090 kprintf(ARGS(old_getrlimit, "%d, %p"),
1091 args->resource, (void *)args->rlim);
1092 #endif
1093 if (args->resource >= LINUX_RLIM_NLIMITS)
1094 return (EINVAL);
1095 which = linux_to_bsd_resource[args->resource];
1096 if (which == -1)
1097 return (EINVAL);
1099 error = kern_getrlimit(which, &rlim);
1101 if (error == 0) {
1102 linux_rlim.rlim_cur = (l_ulong)rlim.rlim_cur;
1103 if (linux_rlim.rlim_cur == ULONG_MAX)
1104 linux_rlim.rlim_cur = LONG_MAX;
1105 linux_rlim.rlim_max = (l_ulong)rlim.rlim_max;
1106 if (linux_rlim.rlim_max == ULONG_MAX)
1107 linux_rlim.rlim_max = LONG_MAX;
1108 error = copyout(&linux_rlim, args->rlim, sizeof(linux_rlim));
1110 return (error);
1114 sys_linux_getrlimit(struct linux_getrlimit_args *args)
1116 struct l_rlimit linux_rlim;
1117 struct rlimit rlim;
1118 u_int which;
1119 int error;
1121 #ifdef DEBUG
1122 if (ldebug(getrlimit))
1123 kprintf(ARGS(getrlimit, "%d, %p"),
1124 args->resource, (void *)args->rlim);
1125 #endif
1126 if (args->resource >= LINUX_RLIM_NLIMITS)
1127 return (EINVAL);
1128 which = linux_to_bsd_resource[args->resource];
1129 if (which == -1)
1130 return (EINVAL);
1132 error = kern_getrlimit(which, &rlim);
1134 if (error == 0) {
1135 linux_rlim.rlim_cur = (l_ulong)rlim.rlim_cur;
1136 linux_rlim.rlim_max = (l_ulong)rlim.rlim_max;
1137 error = copyout(&linux_rlim, args->rlim, sizeof(linux_rlim));
1139 return (error);
1143 sys_linux_sched_setscheduler(struct linux_sched_setscheduler_args *args)
1145 struct sched_setscheduler_args bsd;
1146 int error;
1148 #ifdef DEBUG
1149 if (ldebug(sched_setscheduler))
1150 kprintf(ARGS(sched_setscheduler, "%d, %d, %p"),
1151 args->pid, args->policy, (const void *)args->param);
1152 #endif
1154 switch (args->policy) {
1155 case LINUX_SCHED_OTHER:
1156 bsd.policy = SCHED_OTHER;
1157 break;
1158 case LINUX_SCHED_FIFO:
1159 bsd.policy = SCHED_FIFO;
1160 break;
1161 case LINUX_SCHED_RR:
1162 bsd.policy = SCHED_RR;
1163 break;
1164 default:
1165 return EINVAL;
1168 bsd.pid = args->pid;
1169 bsd.param = (struct sched_param *)args->param;
1170 bsd.sysmsg_result = 0;
1172 error = sys_sched_setscheduler(&bsd);
1173 args->sysmsg_result = bsd.sysmsg_result;
1174 return(error);
1178 sys_linux_sched_getscheduler(struct linux_sched_getscheduler_args *args)
1180 struct sched_getscheduler_args bsd;
1181 int error;
1183 #ifdef DEBUG
1184 if (ldebug(sched_getscheduler))
1185 kprintf(ARGS(sched_getscheduler, "%d"), args->pid);
1186 #endif
1188 bsd.sysmsg_result = 0;
1189 bsd.pid = args->pid;
1190 error = sys_sched_getscheduler(&bsd);
1191 args->sysmsg_result = bsd.sysmsg_result;
1193 switch (args->sysmsg_result) {
1194 case SCHED_OTHER:
1195 args->sysmsg_result = LINUX_SCHED_OTHER;
1196 break;
1197 case SCHED_FIFO:
1198 args->sysmsg_result = LINUX_SCHED_FIFO;
1199 break;
1200 case SCHED_RR:
1201 args->sysmsg_result = LINUX_SCHED_RR;
1202 break;
1204 return error;
1208 sys_linux_sched_get_priority_max(struct linux_sched_get_priority_max_args *args)
1210 struct sched_get_priority_max_args bsd;
1211 int error;
1213 #ifdef DEBUG
1214 if (ldebug(sched_get_priority_max))
1215 kprintf(ARGS(sched_get_priority_max, "%d"), args->policy);
1216 #endif
1218 switch (args->policy) {
1219 case LINUX_SCHED_OTHER:
1220 bsd.policy = SCHED_OTHER;
1221 break;
1222 case LINUX_SCHED_FIFO:
1223 bsd.policy = SCHED_FIFO;
1224 break;
1225 case LINUX_SCHED_RR:
1226 bsd.policy = SCHED_RR;
1227 break;
1228 default:
1229 return EINVAL;
1231 bsd.sysmsg_result = 0;
1233 error = sys_sched_get_priority_max(&bsd);
1234 args->sysmsg_result = bsd.sysmsg_result;
1235 return(error);
1239 sys_linux_sched_get_priority_min(struct linux_sched_get_priority_min_args *args)
1241 struct sched_get_priority_min_args bsd;
1242 int error;
1244 #ifdef DEBUG
1245 if (ldebug(sched_get_priority_min))
1246 kprintf(ARGS(sched_get_priority_min, "%d"), args->policy);
1247 #endif
1249 switch (args->policy) {
1250 case LINUX_SCHED_OTHER:
1251 bsd.policy = SCHED_OTHER;
1252 break;
1253 case LINUX_SCHED_FIFO:
1254 bsd.policy = SCHED_FIFO;
1255 break;
1256 case LINUX_SCHED_RR:
1257 bsd.policy = SCHED_RR;
1258 break;
1259 default:
1260 return EINVAL;
1262 bsd.sysmsg_result = 0;
1264 error = sys_sched_get_priority_min(&bsd);
1265 args->sysmsg_result = bsd.sysmsg_result;
1266 return(error);
1269 #define REBOOT_CAD_ON 0x89abcdef
1270 #define REBOOT_CAD_OFF 0
1271 #define REBOOT_HALT 0xcdef0123
1274 sys_linux_reboot(struct linux_reboot_args *args)
1276 struct reboot_args bsd_args;
1277 int error;
1279 #ifdef DEBUG
1280 if (ldebug(reboot))
1281 kprintf(ARGS(reboot, "0x%x"), args->cmd);
1282 #endif
1283 if (args->cmd == REBOOT_CAD_ON || args->cmd == REBOOT_CAD_OFF)
1284 return (0);
1285 bsd_args.opt = (args->cmd == REBOOT_HALT) ? RB_HALT : 0;
1286 bsd_args.sysmsg_result = 0;
1288 error = sys_reboot(&bsd_args);
1289 args->sysmsg_result = bsd_args.sysmsg_result;
1290 return(error);
1294 * The FreeBSD native getpid(2), getgid(2) and getuid(2) also modify
1295 * p->p_retval[1] when COMPAT_43 or COMPAT_SUNOS is defined. This
1296 * globbers registers that are assumed to be preserved. The following
1297 * lightweight syscalls fixes this. See also linux_getgid16() and
1298 * linux_getuid16() in linux_uid16.c.
1300 * linux_getpid() - MP SAFE
1301 * linux_getgid() - MP SAFE
1302 * linux_getuid() - MP SAFE
1306 sys_linux_getpid(struct linux_getpid_args *args)
1308 struct thread *td = curthread;
1309 struct proc *p = td->td_proc;
1311 KKASSERT(p);
1313 args->sysmsg_result = p->p_pid;
1314 return (0);
1318 sys_linux_getgid(struct linux_getgid_args *args)
1320 struct thread *td = curthread;
1321 struct proc *p = td->td_proc;
1323 KKASSERT(p);
1325 args->sysmsg_result = p->p_ucred->cr_rgid;
1326 return (0);
1330 sys_linux_getuid(struct linux_getuid_args *args)
1332 struct thread *td = curthread;
1333 struct proc *p = td->td_proc;
1335 KKASSERT(p);
1337 args->sysmsg_result = p->p_ucred->cr_ruid;
1338 return (0);
1342 sys_linux_getsid(struct linux_getsid_args *args)
1344 struct getsid_args bsd;
1345 int error;
1347 bsd.sysmsg_result = 0;
1348 bsd.pid = args->pid;
1349 error = sys_getsid(&bsd);
1350 args->sysmsg_result = bsd.sysmsg_result;
1351 return(error);
1355 linux_nosys(struct nosys_args *args)
1357 /* XXX */
1358 return (ENOSYS);