xref: /netbsd-src/sys/kern/kern_exec.c (revision 88fcb00c0357f2d7c1774f86a352637bfda96184)
1 /*	$NetBSD: kern_exec.c,v 1.316 2011/06/06 22:04:34 matt Exp $	*/
2 
3 /*-
4  * Copyright (c) 2008 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
17  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
18  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
19  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
20  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
21  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
22  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
23  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
24  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
25  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
26  * POSSIBILITY OF SUCH DAMAGE.
27  */
28 
29 /*-
30  * Copyright (C) 1993, 1994, 1996 Christopher G. Demetriou
31  * Copyright (C) 1992 Wolfgang Solfrank.
32  * Copyright (C) 1992 TooLs GmbH.
33  * All rights reserved.
34  *
35  * Redistribution and use in source and binary forms, with or without
36  * modification, are permitted provided that the following conditions
37  * are met:
38  * 1. Redistributions of source code must retain the above copyright
39  *    notice, this list of conditions and the following disclaimer.
40  * 2. Redistributions in binary form must reproduce the above copyright
41  *    notice, this list of conditions and the following disclaimer in the
42  *    documentation and/or other materials provided with the distribution.
43  * 3. All advertising materials mentioning features or use of this software
44  *    must display the following acknowledgement:
45  *	This product includes software developed by TooLs GmbH.
46  * 4. The name of TooLs GmbH may not be used to endorse or promote products
47  *    derived from this software without specific prior written permission.
48  *
49  * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR
50  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
51  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
52  * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
53  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
54  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
55  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
56  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
57  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
58  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
59  */
60 
61 #include <sys/cdefs.h>
62 __KERNEL_RCSID(0, "$NetBSD: kern_exec.c,v 1.316 2011/06/06 22:04:34 matt Exp $");
63 
64 #include "opt_ktrace.h"
65 #include "opt_modular.h"
66 #include "opt_syscall_debug.h"
67 #include "veriexec.h"
68 #include "opt_pax.h"
69 #include "opt_sa.h"
70 
71 #include <sys/param.h>
72 #include <sys/systm.h>
73 #include <sys/filedesc.h>
74 #include <sys/kernel.h>
75 #include <sys/proc.h>
76 #include <sys/mount.h>
77 #include <sys/malloc.h>
78 #include <sys/kmem.h>
79 #include <sys/namei.h>
80 #include <sys/vnode.h>
81 #include <sys/file.h>
82 #include <sys/acct.h>
83 #include <sys/exec.h>
84 #include <sys/ktrace.h>
85 #include <sys/uidinfo.h>
86 #include <sys/wait.h>
87 #include <sys/mman.h>
88 #include <sys/ras.h>
89 #include <sys/signalvar.h>
90 #include <sys/stat.h>
91 #include <sys/syscall.h>
92 #include <sys/kauth.h>
93 #include <sys/lwpctl.h>
94 #include <sys/pax.h>
95 #include <sys/cpu.h>
96 #include <sys/module.h>
97 #include <sys/sa.h>
98 #include <sys/savar.h>
99 #include <sys/syscallvar.h>
100 #include <sys/syscallargs.h>
101 #if NVERIEXEC > 0
102 #include <sys/verified_exec.h>
103 #endif /* NVERIEXEC > 0 */
104 #include <sys/sdt.h>
105 
106 #include <uvm/uvm_extern.h>
107 
108 #include <machine/reg.h>
109 
110 #include <compat/common/compat_util.h>
111 
112 static int exec_sigcode_map(struct proc *, const struct emul *);
113 
114 #ifdef DEBUG_EXEC
115 #define DPRINTF(a) printf a
116 #define COPYPRINTF(s, a, b) printf("%s, %d: copyout%s @%p %zu\n", __func__, \
117     __LINE__, (s), (a), (b))
118 #else
119 #define DPRINTF(a)
120 #define COPYPRINTF(s, a, b)
121 #endif /* DEBUG_EXEC */
122 
123 /*
124  * DTrace SDT provider definitions
125  */
126 SDT_PROBE_DEFINE(proc,,,exec,
127 	    "char *", NULL,
128 	    NULL, NULL, NULL, NULL,
129 	    NULL, NULL, NULL, NULL);
130 SDT_PROBE_DEFINE(proc,,,exec_success,
131 	    "char *", NULL,
132 	    NULL, NULL, NULL, NULL,
133 	    NULL, NULL, NULL, NULL);
134 SDT_PROBE_DEFINE(proc,,,exec_failure,
135 	    "int", NULL,
136 	    NULL, NULL, NULL, NULL,
137 	    NULL, NULL, NULL, NULL);
138 
139 /*
140  * Exec function switch:
141  *
142  * Note that each makecmds function is responsible for loading the
143  * exec package with the necessary functions for any exec-type-specific
144  * handling.
145  *
146  * Functions for specific exec types should be defined in their own
147  * header file.
148  */
149 static const struct execsw	**execsw = NULL;
150 static int			nexecs;
151 
152 u_int	exec_maxhdrsz;	 /* must not be static - used by netbsd32 */
153 
154 /* list of dynamically loaded execsw entries */
155 static LIST_HEAD(execlist_head, exec_entry) ex_head =
156     LIST_HEAD_INITIALIZER(ex_head);
157 struct exec_entry {
158 	LIST_ENTRY(exec_entry)	ex_list;
159 	SLIST_ENTRY(exec_entry)	ex_slist;
160 	const struct execsw	*ex_sw;
161 };
162 
163 #ifndef __HAVE_SYSCALL_INTERN
164 void	syscall(void);
165 #endif
166 
167 #ifdef KERN_SA
168 static struct sa_emul saemul_netbsd = {
169 	sizeof(ucontext_t),
170 	sizeof(struct sa_t),
171 	sizeof(struct sa_t *),
172 	NULL,
173 	NULL,
174 	cpu_upcall,
175 	(void (*)(struct lwp *, void *))getucontext_sa,
176 	sa_ucsp
177 };
178 #endif /* KERN_SA */
179 
180 /* NetBSD emul struct */
181 struct emul emul_netbsd = {
182 	.e_name =		"netbsd",
183 	.e_path =		NULL,
184 #ifndef __HAVE_MINIMAL_EMUL
185 	.e_flags =		EMUL_HAS_SYS___syscall,
186 	.e_errno =		NULL,
187 	.e_nosys =		SYS_syscall,
188 	.e_nsysent =		SYS_NSYSENT,
189 #endif
190 	.e_sysent =		sysent,
191 #ifdef SYSCALL_DEBUG
192 	.e_syscallnames =	syscallnames,
193 #else
194 	.e_syscallnames =	NULL,
195 #endif
196 	.e_sendsig =		sendsig,
197 	.e_trapsignal =		trapsignal,
198 	.e_tracesig =		NULL,
199 	.e_sigcode =		NULL,
200 	.e_esigcode =		NULL,
201 	.e_sigobject =		NULL,
202 	.e_setregs =		setregs,
203 	.e_proc_exec =		NULL,
204 	.e_proc_fork =		NULL,
205 	.e_proc_exit =		NULL,
206 	.e_lwp_fork =		NULL,
207 	.e_lwp_exit =		NULL,
208 #ifdef __HAVE_SYSCALL_INTERN
209 	.e_syscall_intern =	syscall_intern,
210 #else
211 	.e_syscall =		syscall,
212 #endif
213 	.e_sysctlovly =		NULL,
214 	.e_fault =		NULL,
215 	.e_vm_default_addr =	uvm_default_mapaddr,
216 	.e_usertrap =		NULL,
217 #ifdef KERN_SA
218 	.e_sa =			&saemul_netbsd,
219 #else
220 	.e_sa =			NULL,
221 #endif
222 	.e_ucsize =		sizeof(ucontext_t),
223 	.e_startlwp =		startlwp
224 };
225 
226 /*
227  * Exec lock. Used to control access to execsw[] structures.
228  * This must not be static so that netbsd32 can access it, too.
229  */
230 krwlock_t exec_lock;
231 
232 static kmutex_t sigobject_lock;
233 
234 static void *
235 exec_pool_alloc(struct pool *pp, int flags)
236 {
237 
238 	return (void *)uvm_km_alloc(kernel_map, NCARGS, 0,
239 	    UVM_KMF_PAGEABLE | UVM_KMF_WAITVA);
240 }
241 
242 static void
243 exec_pool_free(struct pool *pp, void *addr)
244 {
245 
246 	uvm_km_free(kernel_map, (vaddr_t)addr, NCARGS, UVM_KMF_PAGEABLE);
247 }
248 
249 static struct pool exec_pool;
250 
251 static struct pool_allocator exec_palloc = {
252 	.pa_alloc = exec_pool_alloc,
253 	.pa_free = exec_pool_free,
254 	.pa_pagesz = NCARGS
255 };
256 
257 /*
258  * check exec:
259  * given an "executable" described in the exec package's namei info,
260  * see what we can do with it.
261  *
262  * ON ENTRY:
263  *	exec package with appropriate namei info
264  *	lwp pointer of exec'ing lwp
265  *	NO SELF-LOCKED VNODES
266  *
267  * ON EXIT:
268  *	error:	nothing held, etc.  exec header still allocated.
269  *	ok:	filled exec package, executable's vnode (unlocked).
270  *
271  * EXEC SWITCH ENTRY:
272  * 	Locked vnode to check, exec package, proc.
273  *
274  * EXEC SWITCH EXIT:
275  *	ok:	return 0, filled exec package, executable's vnode (unlocked).
276  *	error:	destructive:
277  *			everything deallocated execept exec header.
278  *		non-destructive:
279  *			error code, executable's vnode (unlocked),
280  *			exec header unmodified.
281  */
282 int
283 /*ARGSUSED*/
284 check_exec(struct lwp *l, struct exec_package *epp, struct pathbuf *pb)
285 {
286 	int		error, i;
287 	struct vnode	*vp;
288 	struct nameidata nd;
289 	size_t		resid;
290 
291 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | TRYEMULROOT, pb);
292 
293 	/* first get the vnode */
294 	if ((error = namei(&nd)) != 0)
295 		return error;
296 	epp->ep_vp = vp = nd.ni_vp;
297 	/* this cannot overflow as both are size PATH_MAX */
298 	strcpy(epp->ep_resolvedname, nd.ni_pnbuf);
299 
300 #ifdef DIAGNOSTIC
301 	/* paranoia (take this out once namei stuff stabilizes) */
302 	memset(nd.ni_pnbuf, '~', PATH_MAX);
303 #endif
304 
305 	/* check access and type */
306 	if (vp->v_type != VREG) {
307 		error = EACCES;
308 		goto bad1;
309 	}
310 	if ((error = VOP_ACCESS(vp, VEXEC, l->l_cred)) != 0)
311 		goto bad1;
312 
313 	/* get attributes */
314 	if ((error = VOP_GETATTR(vp, epp->ep_vap, l->l_cred)) != 0)
315 		goto bad1;
316 
317 	/* Check mount point */
318 	if (vp->v_mount->mnt_flag & MNT_NOEXEC) {
319 		error = EACCES;
320 		goto bad1;
321 	}
322 	if (vp->v_mount->mnt_flag & MNT_NOSUID)
323 		epp->ep_vap->va_mode &= ~(S_ISUID | S_ISGID);
324 
325 	/* try to open it */
326 	if ((error = VOP_OPEN(vp, FREAD, l->l_cred)) != 0)
327 		goto bad1;
328 
329 	/* unlock vp, since we need it unlocked from here on out. */
330 	VOP_UNLOCK(vp);
331 
332 #if NVERIEXEC > 0
333 	error = veriexec_verify(l, vp, epp->ep_resolvedname,
334 	    epp->ep_flags & EXEC_INDIR ? VERIEXEC_INDIRECT : VERIEXEC_DIRECT,
335 	    NULL);
336 	if (error)
337 		goto bad2;
338 #endif /* NVERIEXEC > 0 */
339 
340 #ifdef PAX_SEGVGUARD
341 	error = pax_segvguard(l, vp, epp->ep_resolvedname, false);
342 	if (error)
343 		goto bad2;
344 #endif /* PAX_SEGVGUARD */
345 
346 	/* now we have the file, get the exec header */
347 	error = vn_rdwr(UIO_READ, vp, epp->ep_hdr, epp->ep_hdrlen, 0,
348 			UIO_SYSSPACE, 0, l->l_cred, &resid, NULL);
349 	if (error)
350 		goto bad2;
351 	epp->ep_hdrvalid = epp->ep_hdrlen - resid;
352 
353 	/*
354 	 * Set up default address space limits.  Can be overridden
355 	 * by individual exec packages.
356 	 *
357 	 * XXX probably should be all done in the exec packages.
358 	 */
359 	epp->ep_vm_minaddr = VM_MIN_ADDRESS;
360 	epp->ep_vm_maxaddr = VM_MAXUSER_ADDRESS;
361 	/*
362 	 * set up the vmcmds for creation of the process
363 	 * address space
364 	 */
365 	error = ENOEXEC;
366 	for (i = 0; i < nexecs; i++) {
367 		int newerror;
368 
369 		epp->ep_esch = execsw[i];
370 		newerror = (*execsw[i]->es_makecmds)(l, epp);
371 
372 		if (!newerror) {
373 			/* Seems ok: check that entry point is sane */
374 			if (epp->ep_entry > VM_MAXUSER_ADDRESS) {
375 				error = ENOEXEC;
376 				break;
377 			}
378 
379 			/* check limits */
380 			if ((epp->ep_tsize > MAXTSIZ) ||
381 			    (epp->ep_dsize > (u_quad_t)l->l_proc->p_rlimit
382 						    [RLIMIT_DATA].rlim_cur)) {
383 				error = ENOMEM;
384 				break;
385 			}
386 			return 0;
387 		}
388 
389 		if (epp->ep_emul_root != NULL) {
390 			vrele(epp->ep_emul_root);
391 			epp->ep_emul_root = NULL;
392 		}
393 		if (epp->ep_interp != NULL) {
394 			vrele(epp->ep_interp);
395 			epp->ep_interp = NULL;
396 		}
397 
398 		/* make sure the first "interesting" error code is saved. */
399 		if (error == ENOEXEC)
400 			error = newerror;
401 
402 		if (epp->ep_flags & EXEC_DESTR)
403 			/* Error from "#!" code, tidied up by recursive call */
404 			return error;
405 	}
406 
407 	/* not found, error */
408 
409 	/*
410 	 * free any vmspace-creation commands,
411 	 * and release their references
412 	 */
413 	kill_vmcmds(&epp->ep_vmcmds);
414 
415 bad2:
416 	/*
417 	 * close and release the vnode, restore the old one, free the
418 	 * pathname buf, and punt.
419 	 */
420 	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
421 	VOP_CLOSE(vp, FREAD, l->l_cred);
422 	vput(vp);
423 	return error;
424 
425 bad1:
426 	/*
427 	 * free the namei pathname buffer, and put the vnode
428 	 * (which we don't yet have open).
429 	 */
430 	vput(vp);				/* was still locked */
431 	return error;
432 }
433 
434 #ifdef __MACHINE_STACK_GROWS_UP
435 #define STACK_PTHREADSPACE NBPG
436 #else
437 #define STACK_PTHREADSPACE 0
438 #endif
439 
440 static int
441 execve_fetch_element(char * const *array, size_t index, char **value)
442 {
443 	return copyin(array + index, value, sizeof(*value));
444 }
445 
446 /*
447  * exec system call
448  */
449 /* ARGSUSED */
450 int
451 sys_execve(struct lwp *l, const struct sys_execve_args *uap, register_t *retval)
452 {
453 	/* {
454 		syscallarg(const char *)	path;
455 		syscallarg(char * const *)	argp;
456 		syscallarg(char * const *)	envp;
457 	} */
458 
459 	return execve1(l, SCARG(uap, path), SCARG(uap, argp),
460 	    SCARG(uap, envp), execve_fetch_element);
461 }
462 
463 /*
464  * Load modules to try and execute an image that we do not understand.
465  * If no execsw entries are present, we load those likely to be needed
466  * in order to run native images only.  Otherwise, we autoload all
467  * possible modules that could let us run the binary.  XXX lame
468  */
469 static void
470 exec_autoload(void)
471 {
472 #ifdef MODULAR
473 	static const char * const native[] = {
474 		"exec_elf32",
475 		"exec_elf64",
476 		"exec_script",
477 		NULL
478 	};
479 	static const char * const compat[] = {
480 		"exec_elf32",
481 		"exec_elf64",
482 		"exec_script",
483 		"exec_aout",
484 		"exec_coff",
485 		"exec_ecoff",
486 		"compat_aoutm68k",
487 		"compat_freebsd",
488 		"compat_ibcs2",
489 		"compat_linux",
490 		"compat_linux32",
491 		"compat_netbsd32",
492 		"compat_sunos",
493 		"compat_sunos32",
494 		"compat_svr4",
495 		"compat_svr4_32",
496 		"compat_ultrix",
497 		NULL
498 	};
499 	char const * const *list;
500 	int i;
501 
502 	list = (nexecs == 0 ? native : compat);
503 	for (i = 0; list[i] != NULL; i++) {
504 		if (module_autoload(list[i], MODULE_CLASS_MISC) != 0) {
505 		    	continue;
506 		}
507 	   	yield();
508 	}
509 #endif
510 }
511 
512 int
513 execve1(struct lwp *l, const char *path, char * const *args,
514     char * const *envs, execve_fetch_element_t fetch_element)
515 {
516 	int			error;
517 	struct exec_package	pack;
518 	struct pathbuf		*pb;
519 	struct vattr		attr;
520 	struct proc		*p;
521 	char			*argp;
522 	char			*dp, *sp;
523 	long			argc, envc;
524 	size_t			i, len;
525 	char			*stack;
526 	struct ps_strings	arginfo;
527 	struct ps_strings32	arginfo32;
528 	void			*aip;
529 	struct vmspace		*vm;
530 	struct exec_fakearg	*tmpfap;
531 	int			szsigcode;
532 	struct exec_vmcmd	*base_vcp;
533 	int			oldlwpflags;
534 	ksiginfo_t		ksi;
535 	ksiginfoq_t		kq;
536 	const char		*pathstring;
537 	char			*resolvedpathbuf;
538 	const char		*commandname;
539 	u_int			modgen;
540 	size_t			ps_strings_sz;
541 
542 	p = l->l_proc;
543  	modgen = 0;
544 
545 	SDT_PROBE(proc,,,exec, path, 0, 0, 0, 0);
546 
547 	/*
548 	 * Check if we have exceeded our number of processes limit.
549 	 * This is so that we handle the case where a root daemon
550 	 * forked, ran setuid to become the desired user and is trying
551 	 * to exec. The obvious place to do the reference counting check
552 	 * is setuid(), but we don't do the reference counting check there
553 	 * like other OS's do because then all the programs that use setuid()
554 	 * must be modified to check the return code of setuid() and exit().
555 	 * It is dangerous to make setuid() fail, because it fails open and
556 	 * the program will continue to run as root. If we make it succeed
557 	 * and return an error code, again we are not enforcing the limit.
558 	 * The best place to enforce the limit is here, when the process tries
559 	 * to execute a new image, because eventually the process will need
560 	 * to call exec in order to do something useful.
561 	 */
562  retry:
563 	if ((p->p_flag & PK_SUGID) && kauth_authorize_generic(l->l_cred,
564 	    KAUTH_GENERIC_ISSUSER, NULL) != 0 && chgproccnt(kauth_cred_getuid(
565 	    l->l_cred), 0) > p->p_rlimit[RLIMIT_NPROC].rlim_cur)
566 		return EAGAIN;
567 
568 	oldlwpflags = l->l_flag & (LW_SA | LW_SA_UPCALL);
569 	if (l->l_flag & LW_SA) {
570 		lwp_lock(l);
571 		l->l_flag &= ~(LW_SA | LW_SA_UPCALL);
572 		lwp_unlock(l);
573 	}
574 
575 	/*
576 	 * Drain existing references and forbid new ones.  The process
577 	 * should be left alone until we're done here.  This is necessary
578 	 * to avoid race conditions - e.g. in ptrace() - that might allow
579 	 * a local user to illicitly obtain elevated privileges.
580 	 */
581 	rw_enter(&p->p_reflock, RW_WRITER);
582 
583 	base_vcp = NULL;
584 	/*
585 	 * Init the namei data to point the file user's program name.
586 	 * This is done here rather than in check_exec(), so that it's
587 	 * possible to override this settings if any of makecmd/probe
588 	 * functions call check_exec() recursively - for example,
589 	 * see exec_script_makecmds().
590 	 */
591 	error = pathbuf_copyin(path, &pb);
592 	if (error) {
593 		DPRINTF(("%s: pathbuf_copyin path @%p %d\n", __func__,
594 		    path, error));
595 		goto clrflg;
596 	}
597 	pathstring = pathbuf_stringcopy_get(pb);
598 	resolvedpathbuf = PNBUF_GET();
599 #ifdef DIAGNOSTIC
600 	strcpy(resolvedpathbuf, "/wrong");
601 #endif
602 
603 	/*
604 	 * initialize the fields of the exec package.
605 	 */
606 	pack.ep_name = path;
607 	pack.ep_kname = pathstring;
608 	pack.ep_resolvedname = resolvedpathbuf;
609 	pack.ep_hdr = kmem_alloc(exec_maxhdrsz, KM_SLEEP);
610 	pack.ep_hdrlen = exec_maxhdrsz;
611 	pack.ep_hdrvalid = 0;
612 	pack.ep_emul_arg = NULL;
613 	pack.ep_vmcmds.evs_cnt = 0;
614 	pack.ep_vmcmds.evs_used = 0;
615 	pack.ep_vap = &attr;
616 	pack.ep_flags = 0;
617 	pack.ep_emul_root = NULL;
618 	pack.ep_interp = NULL;
619 	pack.ep_esch = NULL;
620 	pack.ep_pax_flags = 0;
621 
622 	rw_enter(&exec_lock, RW_READER);
623 
624 	/* see if we can run it. */
625 	if ((error = check_exec(l, &pack, pb)) != 0) {
626 		if (error != ENOENT) {
627 			DPRINTF(("%s: check exec failed %d\n",
628 			    __func__, error));
629 		}
630 		goto freehdr;
631 	}
632 
633 	/* XXX -- THE FOLLOWING SECTION NEEDS MAJOR CLEANUP */
634 
635 	/* allocate an argument buffer */
636 	argp = pool_get(&exec_pool, PR_WAITOK);
637 	KASSERT(argp != NULL);
638 	dp = argp;
639 	argc = 0;
640 
641 	/* copy the fake args list, if there's one, freeing it as we go */
642 	if (pack.ep_flags & EXEC_HASARGL) {
643 		tmpfap = pack.ep_fa;
644 		while (tmpfap->fa_arg != NULL) {
645 			const char *cp;
646 
647 			cp = tmpfap->fa_arg;
648 			while (*cp)
649 				*dp++ = *cp++;
650 			*dp++ = '\0';
651 			ktrexecarg(tmpfap->fa_arg, cp - tmpfap->fa_arg);
652 
653 			kmem_free(tmpfap->fa_arg, tmpfap->fa_len);
654 			tmpfap++; argc++;
655 		}
656 		kmem_free(pack.ep_fa, pack.ep_fa_len);
657 		pack.ep_flags &= ~EXEC_HASARGL;
658 	}
659 
660 	/* Now get argv & environment */
661 	if (args == NULL) {
662 		DPRINTF(("%s: null args\n", __func__));
663 		error = EINVAL;
664 		goto bad;
665 	}
666 	/* 'i' will index the argp/envp element to be retrieved */
667 	i = 0;
668 	if (pack.ep_flags & EXEC_SKIPARG)
669 		i++;
670 
671 	while (1) {
672 		len = argp + ARG_MAX - dp;
673 		if ((error = (*fetch_element)(args, i, &sp)) != 0) {
674 			DPRINTF(("%s: fetch_element args %d\n",
675 			    __func__, error));
676 			goto bad;
677 		}
678 		if (!sp)
679 			break;
680 		if ((error = copyinstr(sp, dp, len, &len)) != 0) {
681 			DPRINTF(("%s: copyinstr args %d\n", __func__, error));
682 			if (error == ENAMETOOLONG)
683 				error = E2BIG;
684 			goto bad;
685 		}
686 		ktrexecarg(dp, len - 1);
687 		dp += len;
688 		i++;
689 		argc++;
690 	}
691 
692 	envc = 0;
693 	/* environment need not be there */
694 	if (envs != NULL) {
695 		i = 0;
696 		while (1) {
697 			len = argp + ARG_MAX - dp;
698 			if ((error = (*fetch_element)(envs, i, &sp)) != 0) {
699 				DPRINTF(("%s: fetch_element env %d\n",
700 				    __func__, error));
701 				goto bad;
702 			}
703 			if (!sp)
704 				break;
705 			if ((error = copyinstr(sp, dp, len, &len)) != 0) {
706 				DPRINTF(("%s: copyinstr env %d\n",
707 				    __func__, error));
708 				if (error == ENAMETOOLONG)
709 					error = E2BIG;
710 				goto bad;
711 			}
712 			ktrexecenv(dp, len - 1);
713 			dp += len;
714 			i++;
715 			envc++;
716 		}
717 	}
718 
719 	dp = (char *) ALIGN(dp);
720 
721 	szsigcode = pack.ep_esch->es_emul->e_esigcode -
722 	    pack.ep_esch->es_emul->e_sigcode;
723 
724 #ifdef __MACHINE_STACK_GROWS_UP
725 /* See big comment lower down */
726 #define	RTLD_GAP	32
727 #else
728 #define	RTLD_GAP	0
729 #endif
730 
731 	/* Now check if args & environ fit into new stack */
732 	if (pack.ep_flags & EXEC_32) {
733 		aip = &arginfo32;
734 		ps_strings_sz = sizeof(struct ps_strings32);
735 		len = ((argc + envc + 2 + pack.ep_esch->es_arglen) *
736 		    sizeof(int) + sizeof(int) + dp + RTLD_GAP +
737 		    szsigcode + ps_strings_sz + STACK_PTHREADSPACE)
738 		    - argp;
739 	} else {
740 		aip = &arginfo;
741 		ps_strings_sz = sizeof(struct ps_strings);
742 		len = ((argc + envc + 2 + pack.ep_esch->es_arglen) *
743 		    sizeof(char *) + sizeof(int) + dp + RTLD_GAP +
744 		    szsigcode + ps_strings_sz + STACK_PTHREADSPACE)
745 		    - argp;
746 	}
747 
748 #ifdef PAX_ASLR
749 	if (pax_aslr_active(l))
750 		len += (arc4random() % PAGE_SIZE);
751 #endif /* PAX_ASLR */
752 
753 #ifdef STACKLALIGN	/* arm, etc. */
754 	len = STACKALIGN(len);	/* make the stack "safely" aligned */
755 #else
756 	len = ALIGN(len);	/* make the stack "safely" aligned */
757 #endif
758 
759 	if (len > pack.ep_ssize) { /* in effect, compare to initial limit */
760 		DPRINTF(("%s: stack limit exceeded %zu\n", __func__, len));
761 		error = ENOMEM;
762 		goto bad;
763 	}
764 
765 	/* Get rid of other LWPs. */
766 	if (p->p_sa || p->p_nlwps > 1) {
767 		mutex_enter(p->p_lock);
768 		exit_lwps(l);
769 		mutex_exit(p->p_lock);
770 	}
771 	KDASSERT(p->p_nlwps == 1);
772 
773 	/* Destroy any lwpctl info. */
774 	if (p->p_lwpctl != NULL)
775 		lwp_ctl_exit();
776 
777 #ifdef KERN_SA
778 	/* Release any SA state. */
779 	if (p->p_sa)
780 		sa_release(p);
781 #endif /* KERN_SA */
782 
783 	/* Remove POSIX timers */
784 	timers_free(p, TIMERS_POSIX);
785 
786 	/* adjust "active stack depth" for process VSZ */
787 	pack.ep_ssize = len;	/* maybe should go elsewhere, but... */
788 
789 	/*
790 	 * Do whatever is necessary to prepare the address space
791 	 * for remapping.  Note that this might replace the current
792 	 * vmspace with another!
793 	 */
794 	uvmspace_exec(l, pack.ep_vm_minaddr, pack.ep_vm_maxaddr);
795 
796 	/* record proc's vnode, for use by procfs and others */
797         if (p->p_textvp)
798                 vrele(p->p_textvp);
799 	vref(pack.ep_vp);
800 	p->p_textvp = pack.ep_vp;
801 
802 	/* Now map address space */
803 	vm = p->p_vmspace;
804 	vm->vm_taddr = (void *)pack.ep_taddr;
805 	vm->vm_tsize = btoc(pack.ep_tsize);
806 	vm->vm_daddr = (void*)pack.ep_daddr;
807 	vm->vm_dsize = btoc(pack.ep_dsize);
808 	vm->vm_ssize = btoc(pack.ep_ssize);
809 	vm->vm_issize = 0;
810 	vm->vm_maxsaddr = (void *)pack.ep_maxsaddr;
811 	vm->vm_minsaddr = (void *)pack.ep_minsaddr;
812 
813 #ifdef PAX_ASLR
814 	pax_aslr_init(l, vm);
815 #endif /* PAX_ASLR */
816 
817 	/* create the new process's VM space by running the vmcmds */
818 #ifdef DIAGNOSTIC
819 	if (pack.ep_vmcmds.evs_used == 0)
820 		panic("%s: no vmcmds", __func__);
821 #endif
822 	for (i = 0; i < pack.ep_vmcmds.evs_used && !error; i++) {
823 		struct exec_vmcmd *vcp;
824 
825 		vcp = &pack.ep_vmcmds.evs_cmds[i];
826 		if (vcp->ev_flags & VMCMD_RELATIVE) {
827 #ifdef DIAGNOSTIC
828 			if (base_vcp == NULL)
829 				panic("%s: relative vmcmd with no base",
830 				    __func__);
831 			if (vcp->ev_flags & VMCMD_BASE)
832 				panic("%s: illegal base & relative vmcmd",
833 				    __func__);
834 #endif
835 			vcp->ev_addr += base_vcp->ev_addr;
836 		}
837 		error = (*vcp->ev_proc)(l, vcp);
838 #ifdef DEBUG_EXEC
839 		if (error) {
840 			size_t j;
841 			struct exec_vmcmd *vp = &pack.ep_vmcmds.evs_cmds[0];
842 			uprintf("vmcmds %zu/%u, error %d\n", i,
843 			    pack.ep_vmcmds.evs_used, error);
844 			for (j = 0; j <= i; j++)
845 				uprintf("vmcmd[%zu] = vmcmd_map_%s %#"
846 				    PRIxVADDR"/%#"PRIxVSIZE" fd@%#"
847 				    PRIxVSIZE" prot=0%o flags=%d\n", j,
848 				    vp[j].ev_proc == vmcmd_map_pagedvn ?
849 				    "pagedvn" :
850 				    vp[j].ev_proc == vmcmd_map_readvn ?
851 				    "readvn" :
852 				    vp[j].ev_proc == vmcmd_map_zero ?
853 				    "zero" : "*unknown*",
854 				    vp[j].ev_addr, vp[j].ev_len,
855 				    vp[j].ev_offset, vp[j].ev_prot,
856 				    vp[j].ev_flags);
857 		}
858 #endif /* DEBUG_EXEC */
859 		if (vcp->ev_flags & VMCMD_BASE)
860 			base_vcp = vcp;
861 	}
862 
863 	/* free the vmspace-creation commands, and release their references */
864 	kill_vmcmds(&pack.ep_vmcmds);
865 
866 	vn_lock(pack.ep_vp, LK_EXCLUSIVE | LK_RETRY);
867 	VOP_CLOSE(pack.ep_vp, FREAD, l->l_cred);
868 	vput(pack.ep_vp);
869 
870 	/* if an error happened, deallocate and punt */
871 	if (error) {
872 		DPRINTF(("%s: vmcmd %zu failed: %d\n", __func__, i - 1, error));
873 		goto exec_abort;
874 	}
875 
876 	/* remember information about the process */
877 	arginfo.ps_nargvstr = argc;
878 	arginfo.ps_nenvstr = envc;
879 
880 	/* set command name & other accounting info */
881 	commandname = strrchr(pack.ep_resolvedname, '/');
882 	if (commandname != NULL) {
883 		commandname++;
884 	} else {
885 		commandname = pack.ep_resolvedname;
886 	}
887 	i = min(strlen(commandname), MAXCOMLEN);
888 	(void)memcpy(p->p_comm, commandname, i);
889 	p->p_comm[i] = '\0';
890 
891 	dp = PNBUF_GET();
892 	/*
893 	 * If the path starts with /, we don't need to do any work.
894 	 * This handles the majority of the cases.
895 	 * In the future perhaps we could canonicalize it?
896 	 */
897 	if (pathstring[0] == '/')
898 		(void)strlcpy(pack.ep_path = dp, pathstring, MAXPATHLEN);
899 #ifdef notyet
900 	/*
901 	 * Although this works most of the time [since the entry was just
902 	 * entered in the cache] we don't use it because it theoretically
903 	 * can fail and it is not the cleanest interface, because there
904 	 * could be races. When the namei cache is re-written, this can
905 	 * be changed to use the appropriate function.
906 	 */
907 	else if (!(error = vnode_to_path(dp, MAXPATHLEN, p->p_textvp, l, p)))
908 		pack.ep_path = dp;
909 #endif
910 	else {
911 #ifdef notyet
912 		printf("Cannot get path for pid %d [%s] (error %d)",
913 		    (int)p->p_pid, p->p_comm, error);
914 #endif
915 		pack.ep_path = NULL;
916 		PNBUF_PUT(dp);
917 	}
918 
919 	stack = (char *)STACK_ALLOC(STACK_GROW(vm->vm_minsaddr,
920 		STACK_PTHREADSPACE + ps_strings_sz + szsigcode),
921 		len - (ps_strings_sz + szsigcode));
922 
923 #ifdef __MACHINE_STACK_GROWS_UP
924 	/*
925 	 * The copyargs call always copies into lower addresses
926 	 * first, moving towards higher addresses, starting with
927 	 * the stack pointer that we give.  When the stack grows
928 	 * down, this puts argc/argv/envp very shallow on the
929 	 * stack, right at the first user stack pointer.
930 	 * When the stack grows up, the situation is reversed.
931 	 *
932 	 * Normally, this is no big deal.  But the ld_elf.so _rtld()
933 	 * function expects to be called with a single pointer to
934 	 * a region that has a few words it can stash values into,
935 	 * followed by argc/argv/envp.  When the stack grows down,
936 	 * it's easy to decrement the stack pointer a little bit to
937 	 * allocate the space for these few words and pass the new
938 	 * stack pointer to _rtld.  When the stack grows up, however,
939 	 * a few words before argc is part of the signal trampoline, XXX
940 	 * so we have a problem.
941 	 *
942 	 * Instead of changing how _rtld works, we take the easy way
943 	 * out and steal 32 bytes before we call copyargs.
944 	 * This extra space was allowed for when 'len' was calculated.
945 	 */
946 	stack += RTLD_GAP;
947 #endif /* __MACHINE_STACK_GROWS_UP */
948 
949 	/* Now copy argc, args & environ to new stack */
950 	error = (*pack.ep_esch->es_copyargs)(l, &pack, &arginfo, &stack, argp);
951 	if (pack.ep_path) {
952 		PNBUF_PUT(pack.ep_path);
953 		pack.ep_path = NULL;
954 	}
955 	if (error) {
956 		DPRINTF(("%s: copyargs failed %d\n", __func__, error));
957 		goto exec_abort;
958 	}
959 	/* Move the stack back to original point */
960 	stack = (char *)STACK_GROW(vm->vm_minsaddr, len);
961 
962 	/* fill process ps_strings info */
963 	p->p_psstrp = (vaddr_t)STACK_ALLOC(STACK_GROW(vm->vm_minsaddr,
964 	    STACK_PTHREADSPACE), ps_strings_sz);
965 
966 	if (pack.ep_flags & EXEC_32) {
967 		arginfo32.ps_argvstr = (vaddr_t)arginfo.ps_argvstr;
968 		arginfo32.ps_nargvstr = arginfo.ps_nargvstr;
969 		arginfo32.ps_envstr = (vaddr_t)arginfo.ps_envstr;
970 		arginfo32.ps_nenvstr = arginfo.ps_nenvstr;
971 	}
972 
973 	/* copy out the process's ps_strings structure */
974 	if ((error = copyout(aip, (void *)p->p_psstrp, ps_strings_sz)) != 0) {
975 		DPRINTF(("%s: ps_strings copyout %p->%p size %zu failed\n",
976 		    __func__, aip, (void *)p->p_psstrp, ps_strings_sz));
977 		goto exec_abort;
978 	}
979 
980 	cwdexec(p);
981 	fd_closeexec();		/* handle close on exec */
982 
983 	if (__predict_false(ktrace_on))
984 		fd_ktrexecfd();
985 
986 	execsigs(p);		/* reset catched signals */
987 
988 	l->l_ctxlink = NULL;	/* reset ucontext link */
989 
990 
991 	p->p_acflag &= ~AFORK;
992 	mutex_enter(p->p_lock);
993 	p->p_flag |= PK_EXEC;
994 	mutex_exit(p->p_lock);
995 
996 	/*
997 	 * Stop profiling.
998 	 */
999 	if ((p->p_stflag & PST_PROFIL) != 0) {
1000 		mutex_spin_enter(&p->p_stmutex);
1001 		stopprofclock(p);
1002 		mutex_spin_exit(&p->p_stmutex);
1003 	}
1004 
1005 	/*
1006 	 * It's OK to test PL_PPWAIT unlocked here, as other LWPs have
1007 	 * exited and exec()/exit() are the only places it will be cleared.
1008 	 */
1009 	if ((p->p_lflag & PL_PPWAIT) != 0) {
1010 		mutex_enter(proc_lock);
1011 		l->l_lwpctl = NULL; /* was on loan from blocked parent */
1012 		p->p_lflag &= ~PL_PPWAIT;
1013 		cv_broadcast(&p->p_pptr->p_waitcv);
1014 		mutex_exit(proc_lock);
1015 	}
1016 
1017 	/*
1018 	 * Deal with set[ug]id.  MNT_NOSUID has already been used to disable
1019 	 * s[ug]id.  It's OK to check for PSL_TRACED here as we have blocked
1020 	 * out additional references on the process for the moment.
1021 	 */
1022 	if ((p->p_slflag & PSL_TRACED) == 0 &&
1023 
1024 	    (((attr.va_mode & S_ISUID) != 0 &&
1025 	      kauth_cred_geteuid(l->l_cred) != attr.va_uid) ||
1026 
1027 	     ((attr.va_mode & S_ISGID) != 0 &&
1028 	      kauth_cred_getegid(l->l_cred) != attr.va_gid))) {
1029 		/*
1030 		 * Mark the process as SUGID before we do
1031 		 * anything that might block.
1032 		 */
1033 		proc_crmod_enter();
1034 		proc_crmod_leave(NULL, NULL, true);
1035 
1036 		/* Make sure file descriptors 0..2 are in use. */
1037 		if ((error = fd_checkstd()) != 0) {
1038 			DPRINTF(("%s: fdcheckstd failed %d\n",
1039 			    __func__, error));
1040 			goto exec_abort;
1041 		}
1042 
1043 		/*
1044 		 * Copy the credential so other references don't see our
1045 		 * changes.
1046 		 */
1047 		l->l_cred = kauth_cred_copy(l->l_cred);
1048 #ifdef KTRACE
1049 		/*
1050 		 * If the persistent trace flag isn't set, turn off.
1051 		 */
1052 		if (p->p_tracep) {
1053 			mutex_enter(&ktrace_lock);
1054 			if (!(p->p_traceflag & KTRFAC_PERSISTENT))
1055 				ktrderef(p);
1056 			mutex_exit(&ktrace_lock);
1057 		}
1058 #endif
1059 		if (attr.va_mode & S_ISUID)
1060 			kauth_cred_seteuid(l->l_cred, attr.va_uid);
1061 		if (attr.va_mode & S_ISGID)
1062 			kauth_cred_setegid(l->l_cred, attr.va_gid);
1063 	} else {
1064 		if (kauth_cred_geteuid(l->l_cred) ==
1065 		    kauth_cred_getuid(l->l_cred) &&
1066 		    kauth_cred_getegid(l->l_cred) ==
1067 		    kauth_cred_getgid(l->l_cred))
1068 			p->p_flag &= ~PK_SUGID;
1069 	}
1070 
1071 	/*
1072 	 * Copy the credential so other references don't see our changes.
1073 	 * Test to see if this is necessary first, since in the common case
1074 	 * we won't need a private reference.
1075 	 */
1076 	if (kauth_cred_geteuid(l->l_cred) != kauth_cred_getsvuid(l->l_cred) ||
1077 	    kauth_cred_getegid(l->l_cred) != kauth_cred_getsvgid(l->l_cred)) {
1078 		l->l_cred = kauth_cred_copy(l->l_cred);
1079 		kauth_cred_setsvuid(l->l_cred, kauth_cred_geteuid(l->l_cred));
1080 		kauth_cred_setsvgid(l->l_cred, kauth_cred_getegid(l->l_cred));
1081 	}
1082 
1083 	/* Update the master credentials. */
1084 	if (l->l_cred != p->p_cred) {
1085 		kauth_cred_t ocred;
1086 
1087 		kauth_cred_hold(l->l_cred);
1088 		mutex_enter(p->p_lock);
1089 		ocred = p->p_cred;
1090 		p->p_cred = l->l_cred;
1091 		mutex_exit(p->p_lock);
1092 		kauth_cred_free(ocred);
1093 	}
1094 
1095 #if defined(__HAVE_RAS)
1096 	/*
1097 	 * Remove all RASs from the address space.
1098 	 */
1099 	ras_purgeall();
1100 #endif
1101 
1102 	doexechooks(p);
1103 
1104 	/* setup new registers and do misc. setup. */
1105 	(*pack.ep_esch->es_emul->e_setregs)(l, &pack, (vaddr_t)stack);
1106 	if (pack.ep_esch->es_setregs)
1107 		(*pack.ep_esch->es_setregs)(l, &pack, (vaddr_t)stack);
1108 
1109 	/* Provide a consistent LWP private setting */
1110 	(void)lwp_setprivate(l, NULL);
1111 
1112 	/* Discard all PCU state; need to start fresh */
1113 	pcu_discard_all(l);
1114 
1115 	/* map the process's signal trampoline code */
1116 	if ((error = exec_sigcode_map(p, pack.ep_esch->es_emul)) != 0) {
1117 		DPRINTF(("%s: map sigcode failed %d\n", __func__, error));
1118 		goto exec_abort;
1119 	}
1120 
1121 	pool_put(&exec_pool, argp);
1122 
1123 	/* notify others that we exec'd */
1124 	KNOTE(&p->p_klist, NOTE_EXEC);
1125 
1126 	kmem_free(pack.ep_hdr, pack.ep_hdrlen);
1127 
1128 	SDT_PROBE(proc,,,exec_success, path, 0, 0, 0, 0);
1129 
1130 	/* The emulation root will usually have been found when we looked
1131 	 * for the elf interpreter (or similar), if not look now. */
1132 	if (pack.ep_esch->es_emul->e_path != NULL && pack.ep_emul_root == NULL)
1133 		emul_find_root(l, &pack);
1134 
1135 	/* Any old emulation root got removed by fdcloseexec */
1136 	rw_enter(&p->p_cwdi->cwdi_lock, RW_WRITER);
1137 	p->p_cwdi->cwdi_edir = pack.ep_emul_root;
1138 	rw_exit(&p->p_cwdi->cwdi_lock);
1139 	pack.ep_emul_root = NULL;
1140 	if (pack.ep_interp != NULL)
1141 		vrele(pack.ep_interp);
1142 
1143 	/*
1144 	 * Call emulation specific exec hook. This can setup per-process
1145 	 * p->p_emuldata or do any other per-process stuff an emulation needs.
1146 	 *
1147 	 * If we are executing process of different emulation than the
1148 	 * original forked process, call e_proc_exit() of the old emulation
1149 	 * first, then e_proc_exec() of new emulation. If the emulation is
1150 	 * same, the exec hook code should deallocate any old emulation
1151 	 * resources held previously by this process.
1152 	 */
1153 	if (p->p_emul && p->p_emul->e_proc_exit
1154 	    && p->p_emul != pack.ep_esch->es_emul)
1155 		(*p->p_emul->e_proc_exit)(p);
1156 
1157 	/*
1158 	 * This is now LWP 1.
1159 	 */
1160 	mutex_enter(p->p_lock);
1161 	p->p_nlwpid = 1;
1162 	l->l_lid = 1;
1163 	mutex_exit(p->p_lock);
1164 
1165 	/*
1166 	 * Call exec hook. Emulation code may NOT store reference to anything
1167 	 * from &pack.
1168 	 */
1169         if (pack.ep_esch->es_emul->e_proc_exec)
1170                 (*pack.ep_esch->es_emul->e_proc_exec)(p, &pack);
1171 
1172 	/* update p_emul, the old value is no longer needed */
1173 	p->p_emul = pack.ep_esch->es_emul;
1174 
1175 	/* ...and the same for p_execsw */
1176 	p->p_execsw = pack.ep_esch;
1177 
1178 #ifdef __HAVE_SYSCALL_INTERN
1179 	(*p->p_emul->e_syscall_intern)(p);
1180 #endif
1181 	ktremul();
1182 
1183 	/* Allow new references from the debugger/procfs. */
1184 	rw_exit(&p->p_reflock);
1185 	rw_exit(&exec_lock);
1186 
1187 	mutex_enter(proc_lock);
1188 
1189 	if ((p->p_slflag & (PSL_TRACED|PSL_SYSCALL)) == PSL_TRACED) {
1190 		KSI_INIT_EMPTY(&ksi);
1191 		ksi.ksi_signo = SIGTRAP;
1192 		ksi.ksi_lid = l->l_lid;
1193 		kpsignal(p, &ksi, NULL);
1194 	}
1195 
1196 	if (p->p_sflag & PS_STOPEXEC) {
1197 		KERNEL_UNLOCK_ALL(l, &l->l_biglocks);
1198 		p->p_pptr->p_nstopchild++;
1199 		p->p_pptr->p_waited = 0;
1200 		mutex_enter(p->p_lock);
1201 		ksiginfo_queue_init(&kq);
1202 		sigclearall(p, &contsigmask, &kq);
1203 		lwp_lock(l);
1204 		l->l_stat = LSSTOP;
1205 		p->p_stat = SSTOP;
1206 		p->p_nrlwps--;
1207 		lwp_unlock(l);
1208 		mutex_exit(p->p_lock);
1209 		mutex_exit(proc_lock);
1210 		lwp_lock(l);
1211 		mi_switch(l);
1212 		ksiginfo_queue_drain(&kq);
1213 		KERNEL_LOCK(l->l_biglocks, l);
1214 	} else {
1215 		mutex_exit(proc_lock);
1216 	}
1217 
1218 	pathbuf_stringcopy_put(pb, pathstring);
1219 	pathbuf_destroy(pb);
1220 	PNBUF_PUT(resolvedpathbuf);
1221 	return (EJUSTRETURN);
1222 
1223  bad:
1224 	/* free the vmspace-creation commands, and release their references */
1225 	kill_vmcmds(&pack.ep_vmcmds);
1226 	/* kill any opened file descriptor, if necessary */
1227 	if (pack.ep_flags & EXEC_HASFD) {
1228 		pack.ep_flags &= ~EXEC_HASFD;
1229 		fd_close(pack.ep_fd);
1230 	}
1231 	/* close and put the exec'd file */
1232 	vn_lock(pack.ep_vp, LK_EXCLUSIVE | LK_RETRY);
1233 	VOP_CLOSE(pack.ep_vp, FREAD, l->l_cred);
1234 	vput(pack.ep_vp);
1235 	pool_put(&exec_pool, argp);
1236 
1237  freehdr:
1238 	kmem_free(pack.ep_hdr, pack.ep_hdrlen);
1239 	if (pack.ep_emul_root != NULL)
1240 		vrele(pack.ep_emul_root);
1241 	if (pack.ep_interp != NULL)
1242 		vrele(pack.ep_interp);
1243 
1244 	rw_exit(&exec_lock);
1245 
1246 	pathbuf_stringcopy_put(pb, pathstring);
1247 	pathbuf_destroy(pb);
1248 	PNBUF_PUT(resolvedpathbuf);
1249 
1250  clrflg:
1251 	lwp_lock(l);
1252 	l->l_flag |= oldlwpflags;
1253 	lwp_unlock(l);
1254 	rw_exit(&p->p_reflock);
1255 
1256 	if (modgen != module_gen && error == ENOEXEC) {
1257 		modgen = module_gen;
1258 		exec_autoload();
1259 		goto retry;
1260 	}
1261 
1262 	SDT_PROBE(proc,,,exec_failure, error, 0, 0, 0, 0);
1263 	return error;
1264 
1265  exec_abort:
1266 	SDT_PROBE(proc,,,exec_failure, error, 0, 0, 0, 0);
1267 	rw_exit(&p->p_reflock);
1268 	rw_exit(&exec_lock);
1269 
1270 	pathbuf_stringcopy_put(pb, pathstring);
1271 	pathbuf_destroy(pb);
1272 	PNBUF_PUT(resolvedpathbuf);
1273 
1274 	/*
1275 	 * the old process doesn't exist anymore.  exit gracefully.
1276 	 * get rid of the (new) address space we have created, if any, get rid
1277 	 * of our namei data and vnode, and exit noting failure
1278 	 */
1279 	uvm_deallocate(&vm->vm_map, VM_MIN_ADDRESS,
1280 		VM_MAXUSER_ADDRESS - VM_MIN_ADDRESS);
1281 	if (pack.ep_emul_arg)
1282 		free(pack.ep_emul_arg, M_TEMP);
1283 	pool_put(&exec_pool, argp);
1284 	kmem_free(pack.ep_hdr, pack.ep_hdrlen);
1285 	if (pack.ep_emul_root != NULL)
1286 		vrele(pack.ep_emul_root);
1287 	if (pack.ep_interp != NULL)
1288 		vrele(pack.ep_interp);
1289 
1290 	/* Acquire the sched-state mutex (exit1() will release it). */
1291 	mutex_enter(p->p_lock);
1292 	exit1(l, W_EXITCODE(error, SIGABRT));
1293 
1294 	/* NOTREACHED */
1295 	return 0;
1296 }
1297 
1298 int
1299 copyargs(struct lwp *l, struct exec_package *pack, struct ps_strings *arginfo,
1300     char **stackp, void *argp)
1301 {
1302 	char	**cpp, *dp, *sp;
1303 	size_t	len;
1304 	void	*nullp;
1305 	long	argc, envc;
1306 	int	error;
1307 
1308 	cpp = (char **)*stackp;
1309 	nullp = NULL;
1310 	argc = arginfo->ps_nargvstr;
1311 	envc = arginfo->ps_nenvstr;
1312 	if ((error = copyout(&argc, cpp++, sizeof(argc))) != 0) {
1313 		COPYPRINTF("", cpp - 1, sizeof(argc));
1314 		return error;
1315 	}
1316 
1317 	dp = (char *) (cpp + argc + envc + 2 + pack->ep_esch->es_arglen);
1318 	sp = argp;
1319 
1320 	/* XXX don't copy them out, remap them! */
1321 	arginfo->ps_argvstr = cpp; /* remember location of argv for later */
1322 
1323 	for (; --argc >= 0; sp += len, dp += len) {
1324 		if ((error = copyout(&dp, cpp++, sizeof(dp))) != 0) {
1325 			COPYPRINTF("", cpp - 1, sizeof(dp));
1326 			return error;
1327 		}
1328 		if ((error = copyoutstr(sp, dp, ARG_MAX, &len)) != 0) {
1329 			COPYPRINTF("str", dp, (size_t)ARG_MAX);
1330 			return error;
1331 		}
1332 	}
1333 
1334 	if ((error = copyout(&nullp, cpp++, sizeof(nullp))) != 0) {
1335 		COPYPRINTF("", cpp - 1, sizeof(nullp));
1336 		return error;
1337 	}
1338 
1339 	arginfo->ps_envstr = cpp; /* remember location of envp for later */
1340 
1341 	for (; --envc >= 0; sp += len, dp += len) {
1342 		if ((error = copyout(&dp, cpp++, sizeof(dp))) != 0) {
1343 			COPYPRINTF("", cpp - 1, sizeof(dp));
1344 			return error;
1345 		}
1346 		if ((error = copyoutstr(sp, dp, ARG_MAX, &len)) != 0) {
1347 			COPYPRINTF("str", dp, (size_t)ARG_MAX);
1348 			return error;
1349 		}
1350 	}
1351 
1352 	if ((error = copyout(&nullp, cpp++, sizeof(nullp))) != 0) {
1353 		COPYPRINTF("", cpp - 1, sizeof(nullp));
1354 		return error;
1355 	}
1356 
1357 	*stackp = (char *)cpp;
1358 	return 0;
1359 }
1360 
1361 
1362 /*
1363  * Add execsw[] entries.
1364  */
1365 int
1366 exec_add(struct execsw *esp, int count)
1367 {
1368 	struct exec_entry	*it;
1369 	int			i;
1370 
1371 	if (count == 0) {
1372 		return 0;
1373 	}
1374 
1375 	/* Check for duplicates. */
1376 	rw_enter(&exec_lock, RW_WRITER);
1377 	for (i = 0; i < count; i++) {
1378 		LIST_FOREACH(it, &ex_head, ex_list) {
1379 			/* assume unique (makecmds, probe_func, emulation) */
1380 			if (it->ex_sw->es_makecmds == esp[i].es_makecmds &&
1381 			    it->ex_sw->u.elf_probe_func ==
1382 			    esp[i].u.elf_probe_func &&
1383 			    it->ex_sw->es_emul == esp[i].es_emul) {
1384 				rw_exit(&exec_lock);
1385 				return EEXIST;
1386 			}
1387 		}
1388 	}
1389 
1390 	/* Allocate new entries. */
1391 	for (i = 0; i < count; i++) {
1392 		it = kmem_alloc(sizeof(*it), KM_SLEEP);
1393 		it->ex_sw = &esp[i];
1394 		LIST_INSERT_HEAD(&ex_head, it, ex_list);
1395 	}
1396 
1397 	/* update execsw[] */
1398 	exec_init(0);
1399 	rw_exit(&exec_lock);
1400 	return 0;
1401 }
1402 
1403 /*
1404  * Remove execsw[] entry.
1405  */
1406 int
1407 exec_remove(struct execsw *esp, int count)
1408 {
1409 	struct exec_entry	*it, *next;
1410 	int			i;
1411 	const struct proclist_desc *pd;
1412 	proc_t			*p;
1413 
1414 	if (count == 0) {
1415 		return 0;
1416 	}
1417 
1418 	/* Abort if any are busy. */
1419 	rw_enter(&exec_lock, RW_WRITER);
1420 	for (i = 0; i < count; i++) {
1421 		mutex_enter(proc_lock);
1422 		for (pd = proclists; pd->pd_list != NULL; pd++) {
1423 			PROCLIST_FOREACH(p, pd->pd_list) {
1424 				if (p->p_execsw == &esp[i]) {
1425 					mutex_exit(proc_lock);
1426 					rw_exit(&exec_lock);
1427 					return EBUSY;
1428 				}
1429 			}
1430 		}
1431 		mutex_exit(proc_lock);
1432 	}
1433 
1434 	/* None are busy, so remove them all. */
1435 	for (i = 0; i < count; i++) {
1436 		for (it = LIST_FIRST(&ex_head); it != NULL; it = next) {
1437 			next = LIST_NEXT(it, ex_list);
1438 			if (it->ex_sw == &esp[i]) {
1439 				LIST_REMOVE(it, ex_list);
1440 				kmem_free(it, sizeof(*it));
1441 				break;
1442 			}
1443 		}
1444 	}
1445 
1446 	/* update execsw[] */
1447 	exec_init(0);
1448 	rw_exit(&exec_lock);
1449 	return 0;
1450 }
1451 
1452 /*
1453  * Initialize exec structures. If init_boot is true, also does necessary
1454  * one-time initialization (it's called from main() that way).
1455  * Once system is multiuser, this should be called with exec_lock held,
1456  * i.e. via exec_{add|remove}().
1457  */
1458 int
1459 exec_init(int init_boot)
1460 {
1461 	const struct execsw 	**sw;
1462 	struct exec_entry	*ex;
1463 	SLIST_HEAD(,exec_entry)	first;
1464 	SLIST_HEAD(,exec_entry)	any;
1465 	SLIST_HEAD(,exec_entry)	last;
1466 	int			i, sz;
1467 
1468 	if (init_boot) {
1469 		/* do one-time initializations */
1470 		rw_init(&exec_lock);
1471 		mutex_init(&sigobject_lock, MUTEX_DEFAULT, IPL_NONE);
1472 		pool_init(&exec_pool, NCARGS, 0, 0, PR_NOALIGN|PR_NOTOUCH,
1473 		    "execargs", &exec_palloc, IPL_NONE);
1474 		pool_sethardlimit(&exec_pool, maxexec, "should not happen", 0);
1475 	} else {
1476 		KASSERT(rw_write_held(&exec_lock));
1477 	}
1478 
1479 	/* Sort each entry onto the appropriate queue. */
1480 	SLIST_INIT(&first);
1481 	SLIST_INIT(&any);
1482 	SLIST_INIT(&last);
1483 	sz = 0;
1484 	LIST_FOREACH(ex, &ex_head, ex_list) {
1485 		switch(ex->ex_sw->es_prio) {
1486 		case EXECSW_PRIO_FIRST:
1487 			SLIST_INSERT_HEAD(&first, ex, ex_slist);
1488 			break;
1489 		case EXECSW_PRIO_ANY:
1490 			SLIST_INSERT_HEAD(&any, ex, ex_slist);
1491 			break;
1492 		case EXECSW_PRIO_LAST:
1493 			SLIST_INSERT_HEAD(&last, ex, ex_slist);
1494 			break;
1495 		default:
1496 			panic("%s", __func__);
1497 			break;
1498 		}
1499 		sz++;
1500 	}
1501 
1502 	/*
1503 	 * Create new execsw[].  Ensure we do not try a zero-sized
1504 	 * allocation.
1505 	 */
1506 	sw = kmem_alloc(sz * sizeof(struct execsw *) + 1, KM_SLEEP);
1507 	i = 0;
1508 	SLIST_FOREACH(ex, &first, ex_slist) {
1509 		sw[i++] = ex->ex_sw;
1510 	}
1511 	SLIST_FOREACH(ex, &any, ex_slist) {
1512 		sw[i++] = ex->ex_sw;
1513 	}
1514 	SLIST_FOREACH(ex, &last, ex_slist) {
1515 		sw[i++] = ex->ex_sw;
1516 	}
1517 
1518 	/* Replace old execsw[] and free used memory. */
1519 	if (execsw != NULL) {
1520 		kmem_free(__UNCONST(execsw),
1521 		    nexecs * sizeof(struct execsw *) + 1);
1522 	}
1523 	execsw = sw;
1524 	nexecs = sz;
1525 
1526 	/* Figure out the maximum size of an exec header. */
1527 	exec_maxhdrsz = sizeof(int);
1528 	for (i = 0; i < nexecs; i++) {
1529 		if (execsw[i]->es_hdrsz > exec_maxhdrsz)
1530 			exec_maxhdrsz = execsw[i]->es_hdrsz;
1531 	}
1532 
1533 	return 0;
1534 }
1535 
1536 static int
1537 exec_sigcode_map(struct proc *p, const struct emul *e)
1538 {
1539 	vaddr_t va;
1540 	vsize_t sz;
1541 	int error;
1542 	struct uvm_object *uobj;
1543 
1544 	sz = (vaddr_t)e->e_esigcode - (vaddr_t)e->e_sigcode;
1545 
1546 	if (e->e_sigobject == NULL || sz == 0) {
1547 		return 0;
1548 	}
1549 
1550 	/*
1551 	 * If we don't have a sigobject for this emulation, create one.
1552 	 *
1553 	 * sigobject is an anonymous memory object (just like SYSV shared
1554 	 * memory) that we keep a permanent reference to and that we map
1555 	 * in all processes that need this sigcode. The creation is simple,
1556 	 * we create an object, add a permanent reference to it, map it in
1557 	 * kernel space, copy out the sigcode to it and unmap it.
1558 	 * We map it with PROT_READ|PROT_EXEC into the process just
1559 	 * the way sys_mmap() would map it.
1560 	 */
1561 
1562 	uobj = *e->e_sigobject;
1563 	if (uobj == NULL) {
1564 		mutex_enter(&sigobject_lock);
1565 		if ((uobj = *e->e_sigobject) == NULL) {
1566 			uobj = uao_create(sz, 0);
1567 			(*uobj->pgops->pgo_reference)(uobj);
1568 			va = vm_map_min(kernel_map);
1569 			if ((error = uvm_map(kernel_map, &va, round_page(sz),
1570 			    uobj, 0, 0,
1571 			    UVM_MAPFLAG(UVM_PROT_RW, UVM_PROT_RW,
1572 			    UVM_INH_SHARE, UVM_ADV_RANDOM, 0)))) {
1573 				printf("kernel mapping failed %d\n", error);
1574 				(*uobj->pgops->pgo_detach)(uobj);
1575 				mutex_exit(&sigobject_lock);
1576 				return (error);
1577 			}
1578 			memcpy((void *)va, e->e_sigcode, sz);
1579 #ifdef PMAP_NEED_PROCWR
1580 			pmap_procwr(&proc0, va, sz);
1581 #endif
1582 			uvm_unmap(kernel_map, va, va + round_page(sz));
1583 			*e->e_sigobject = uobj;
1584 		}
1585 		mutex_exit(&sigobject_lock);
1586 	}
1587 
1588 	/* Just a hint to uvm_map where to put it. */
1589 	va = e->e_vm_default_addr(p, (vaddr_t)p->p_vmspace->vm_daddr,
1590 	    round_page(sz));
1591 
1592 #ifdef __alpha__
1593 	/*
1594 	 * Tru64 puts /sbin/loader at the end of user virtual memory,
1595 	 * which causes the above calculation to put the sigcode at
1596 	 * an invalid address.  Put it just below the text instead.
1597 	 */
1598 	if (va == (vaddr_t)vm_map_max(&p->p_vmspace->vm_map)) {
1599 		va = (vaddr_t)p->p_vmspace->vm_taddr - round_page(sz);
1600 	}
1601 #endif
1602 
1603 	(*uobj->pgops->pgo_reference)(uobj);
1604 	error = uvm_map(&p->p_vmspace->vm_map, &va, round_page(sz),
1605 			uobj, 0, 0,
1606 			UVM_MAPFLAG(UVM_PROT_RX, UVM_PROT_RX, UVM_INH_SHARE,
1607 				    UVM_ADV_RANDOM, 0));
1608 	if (error) {
1609 		DPRINTF(("%s, %d: map %p "
1610 		    "uvm_map %#"PRIxVSIZE"@%#"PRIxVADDR" failed %d\n",
1611 		    __func__, __LINE__, &p->p_vmspace->vm_map, round_page(sz),
1612 		    va, error));
1613 		(*uobj->pgops->pgo_detach)(uobj);
1614 		return (error);
1615 	}
1616 	p->p_sigctx.ps_sigcode = (void *)va;
1617 	return (0);
1618 }
1619