xref: /netbsd-src/sys/nfs/nfs_subs.c (revision 3b01aba77a7a698587faaae455bbfe740923c1f5)
1 /*	$NetBSD: nfs_subs.c,v 1.95 2001/06/07 01:04:40 lukem Exp $	*/
2 
3 /*
4  * Copyright (c) 1989, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Rick Macklem at The University of Guelph.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. All advertising materials mentioning features or use of this software
19  *    must display the following acknowledgement:
20  *	This product includes software developed by the University of
21  *	California, Berkeley and its contributors.
22  * 4. Neither the name of the University nor the names of its contributors
23  *    may be used to endorse or promote products derived from this software
24  *    without specific prior written permission.
25  *
26  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36  * SUCH DAMAGE.
37  *
38  *	@(#)nfs_subs.c	8.8 (Berkeley) 5/22/95
39  */
40 
41 /*
42  * Copyright 2000 Wasabi Systems, Inc.
43  * All rights reserved.
44  *
45  * Written by Frank van der Linden for Wasabi Systems, Inc.
46  *
47  * Redistribution and use in source and binary forms, with or without
48  * modification, are permitted provided that the following conditions
49  * are met:
50  * 1. Redistributions of source code must retain the above copyright
51  *    notice, this list of conditions and the following disclaimer.
52  * 2. Redistributions in binary form must reproduce the above copyright
53  *    notice, this list of conditions and the following disclaimer in the
54  *    documentation and/or other materials provided with the distribution.
55  * 3. All advertising materials mentioning features or use of this software
56  *    must display the following acknowledgement:
57  *      This product includes software developed for the NetBSD Project by
58  *      Wasabi Systems, Inc.
59  * 4. The name of Wasabi Systems, Inc. may not be used to endorse
60  *    or promote products derived from this software without specific prior
61  *    written permission.
62  *
63  * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
64  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
65  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
66  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL WASABI SYSTEMS, INC
67  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
68  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
69  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
70  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
71  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
72  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
73  * POSSIBILITY OF SUCH DAMAGE.
74  */
75 
76 #include "fs_nfs.h"
77 #include "opt_nfs.h"
78 #include "opt_nfsserver.h"
79 #include "opt_iso.h"
80 #include "opt_inet.h"
81 
82 /*
83  * These functions support the macros and help fiddle mbuf chains for
84  * the nfs op functions. They do things like create the rpc header and
85  * copy data between mbuf chains and uio lists.
86  */
87 #include <sys/param.h>
88 #include <sys/proc.h>
89 #include <sys/systm.h>
90 #include <sys/kernel.h>
91 #include <sys/mount.h>
92 #include <sys/vnode.h>
93 #include <sys/namei.h>
94 #include <sys/mbuf.h>
95 #include <sys/socket.h>
96 #include <sys/stat.h>
97 #include <sys/malloc.h>
98 #include <sys/time.h>
99 #include <sys/dirent.h>
100 
101 #include <uvm/uvm_extern.h>
102 
103 #include <nfs/rpcv2.h>
104 #include <nfs/nfsproto.h>
105 #include <nfs/nfsnode.h>
106 #include <nfs/nfs.h>
107 #include <nfs/xdr_subs.h>
108 #include <nfs/nfsm_subs.h>
109 #include <nfs/nfsmount.h>
110 #include <nfs/nqnfs.h>
111 #include <nfs/nfsrtt.h>
112 #include <nfs/nfs_var.h>
113 
114 #include <miscfs/specfs/specdev.h>
115 
116 #include <netinet/in.h>
117 #ifdef ISO
118 #include <netiso/iso.h>
119 #endif
120 
121 /*
122  * Data items converted to xdr at startup, since they are constant
123  * This is kinda hokey, but may save a little time doing byte swaps
124  */
125 u_int32_t nfs_xdrneg1;
126 u_int32_t rpc_call, rpc_vers, rpc_reply, rpc_msgdenied, rpc_autherr,
127 	rpc_mismatch, rpc_auth_unix, rpc_msgaccepted,
128 	rpc_auth_kerb;
129 u_int32_t nfs_prog, nqnfs_prog, nfs_true, nfs_false;
130 
131 /* And other global data */
132 static u_int32_t nfs_xid = 0;
133 const nfstype nfsv2_type[9] =
134 	{ NFNON, NFREG, NFDIR, NFBLK, NFCHR, NFLNK, NFNON, NFCHR, NFNON };
135 const nfstype nfsv3_type[9] =
136 	{ NFNON, NFREG, NFDIR, NFBLK, NFCHR, NFLNK, NFSOCK, NFFIFO, NFNON };
137 const enum vtype nv2tov_type[8] =
138 	{ VNON, VREG, VDIR, VBLK, VCHR, VLNK, VNON, VNON };
139 const enum vtype nv3tov_type[8] =
140 	{ VNON, VREG, VDIR, VBLK, VCHR, VLNK, VSOCK, VFIFO };
141 int nfs_ticks;
142 
143 /* NFS client/server stats. */
144 struct nfsstats nfsstats;
145 
146 /*
147  * Mapping of old NFS Version 2 RPC numbers to generic numbers.
148  */
149 const int nfsv3_procid[NFS_NPROCS] = {
150 	NFSPROC_NULL,
151 	NFSPROC_GETATTR,
152 	NFSPROC_SETATTR,
153 	NFSPROC_NOOP,
154 	NFSPROC_LOOKUP,
155 	NFSPROC_READLINK,
156 	NFSPROC_READ,
157 	NFSPROC_NOOP,
158 	NFSPROC_WRITE,
159 	NFSPROC_CREATE,
160 	NFSPROC_REMOVE,
161 	NFSPROC_RENAME,
162 	NFSPROC_LINK,
163 	NFSPROC_SYMLINK,
164 	NFSPROC_MKDIR,
165 	NFSPROC_RMDIR,
166 	NFSPROC_READDIR,
167 	NFSPROC_FSSTAT,
168 	NFSPROC_NOOP,
169 	NFSPROC_NOOP,
170 	NFSPROC_NOOP,
171 	NFSPROC_NOOP,
172 	NFSPROC_NOOP,
173 	NFSPROC_NOOP,
174 	NFSPROC_NOOP,
175 	NFSPROC_NOOP
176 };
177 
178 /*
179  * and the reverse mapping from generic to Version 2 procedure numbers
180  */
181 const int nfsv2_procid[NFS_NPROCS] = {
182 	NFSV2PROC_NULL,
183 	NFSV2PROC_GETATTR,
184 	NFSV2PROC_SETATTR,
185 	NFSV2PROC_LOOKUP,
186 	NFSV2PROC_NOOP,
187 	NFSV2PROC_READLINK,
188 	NFSV2PROC_READ,
189 	NFSV2PROC_WRITE,
190 	NFSV2PROC_CREATE,
191 	NFSV2PROC_MKDIR,
192 	NFSV2PROC_SYMLINK,
193 	NFSV2PROC_CREATE,
194 	NFSV2PROC_REMOVE,
195 	NFSV2PROC_RMDIR,
196 	NFSV2PROC_RENAME,
197 	NFSV2PROC_LINK,
198 	NFSV2PROC_READDIR,
199 	NFSV2PROC_NOOP,
200 	NFSV2PROC_STATFS,
201 	NFSV2PROC_NOOP,
202 	NFSV2PROC_NOOP,
203 	NFSV2PROC_NOOP,
204 	NFSV2PROC_NOOP,
205 	NFSV2PROC_NOOP,
206 	NFSV2PROC_NOOP,
207 	NFSV2PROC_NOOP,
208 };
209 
210 /*
211  * Maps errno values to nfs error numbers.
212  * Use NFSERR_IO as the catch all for ones not specifically defined in
213  * RFC 1094.
214  */
215 static const u_char nfsrv_v2errmap[ELAST] = {
216   NFSERR_PERM,	NFSERR_NOENT,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
217   NFSERR_NXIO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
218   NFSERR_IO,	NFSERR_IO,	NFSERR_ACCES,	NFSERR_IO,	NFSERR_IO,
219   NFSERR_IO,	NFSERR_EXIST,	NFSERR_IO,	NFSERR_NODEV,	NFSERR_NOTDIR,
220   NFSERR_ISDIR,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
221   NFSERR_IO,	NFSERR_FBIG,	NFSERR_NOSPC,	NFSERR_IO,	NFSERR_ROFS,
222   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
223   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
224   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
225   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
226   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
227   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
228   NFSERR_IO,	NFSERR_IO,	NFSERR_NAMETOL,	NFSERR_IO,	NFSERR_IO,
229   NFSERR_NOTEMPTY, NFSERR_IO,	NFSERR_IO,	NFSERR_DQUOT,	NFSERR_STALE,
230   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
231   NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,	NFSERR_IO,
232   NFSERR_IO,	NFSERR_IO,
233 };
234 
235 /*
236  * Maps errno values to nfs error numbers.
237  * Although it is not obvious whether or not NFS clients really care if
238  * a returned error value is in the specified list for the procedure, the
239  * safest thing to do is filter them appropriately. For Version 2, the
240  * X/Open XNFS document is the only specification that defines error values
241  * for each RPC (The RFC simply lists all possible error values for all RPCs),
242  * so I have decided to not do this for Version 2.
243  * The first entry is the default error return and the rest are the valid
244  * errors for that RPC in increasing numeric order.
245  */
246 static const short nfsv3err_null[] = {
247 	0,
248 	0,
249 };
250 
251 static const short nfsv3err_getattr[] = {
252 	NFSERR_IO,
253 	NFSERR_IO,
254 	NFSERR_STALE,
255 	NFSERR_BADHANDLE,
256 	NFSERR_SERVERFAULT,
257 	0,
258 };
259 
260 static const short nfsv3err_setattr[] = {
261 	NFSERR_IO,
262 	NFSERR_PERM,
263 	NFSERR_IO,
264 	NFSERR_ACCES,
265 	NFSERR_INVAL,
266 	NFSERR_NOSPC,
267 	NFSERR_ROFS,
268 	NFSERR_DQUOT,
269 	NFSERR_STALE,
270 	NFSERR_BADHANDLE,
271 	NFSERR_NOT_SYNC,
272 	NFSERR_SERVERFAULT,
273 	0,
274 };
275 
276 static const short nfsv3err_lookup[] = {
277 	NFSERR_IO,
278 	NFSERR_NOENT,
279 	NFSERR_IO,
280 	NFSERR_ACCES,
281 	NFSERR_NOTDIR,
282 	NFSERR_NAMETOL,
283 	NFSERR_STALE,
284 	NFSERR_BADHANDLE,
285 	NFSERR_SERVERFAULT,
286 	0,
287 };
288 
289 static const short nfsv3err_access[] = {
290 	NFSERR_IO,
291 	NFSERR_IO,
292 	NFSERR_STALE,
293 	NFSERR_BADHANDLE,
294 	NFSERR_SERVERFAULT,
295 	0,
296 };
297 
298 static const short nfsv3err_readlink[] = {
299 	NFSERR_IO,
300 	NFSERR_IO,
301 	NFSERR_ACCES,
302 	NFSERR_INVAL,
303 	NFSERR_STALE,
304 	NFSERR_BADHANDLE,
305 	NFSERR_NOTSUPP,
306 	NFSERR_SERVERFAULT,
307 	0,
308 };
309 
310 static const short nfsv3err_read[] = {
311 	NFSERR_IO,
312 	NFSERR_IO,
313 	NFSERR_NXIO,
314 	NFSERR_ACCES,
315 	NFSERR_INVAL,
316 	NFSERR_STALE,
317 	NFSERR_BADHANDLE,
318 	NFSERR_SERVERFAULT,
319 	NFSERR_JUKEBOX,
320 	0,
321 };
322 
323 static const short nfsv3err_write[] = {
324 	NFSERR_IO,
325 	NFSERR_IO,
326 	NFSERR_ACCES,
327 	NFSERR_INVAL,
328 	NFSERR_FBIG,
329 	NFSERR_NOSPC,
330 	NFSERR_ROFS,
331 	NFSERR_DQUOT,
332 	NFSERR_STALE,
333 	NFSERR_BADHANDLE,
334 	NFSERR_SERVERFAULT,
335 	NFSERR_JUKEBOX,
336 	0,
337 };
338 
339 static const short nfsv3err_create[] = {
340 	NFSERR_IO,
341 	NFSERR_IO,
342 	NFSERR_ACCES,
343 	NFSERR_EXIST,
344 	NFSERR_NOTDIR,
345 	NFSERR_NOSPC,
346 	NFSERR_ROFS,
347 	NFSERR_NAMETOL,
348 	NFSERR_DQUOT,
349 	NFSERR_STALE,
350 	NFSERR_BADHANDLE,
351 	NFSERR_NOTSUPP,
352 	NFSERR_SERVERFAULT,
353 	0,
354 };
355 
356 static const short nfsv3err_mkdir[] = {
357 	NFSERR_IO,
358 	NFSERR_IO,
359 	NFSERR_ACCES,
360 	NFSERR_EXIST,
361 	NFSERR_NOTDIR,
362 	NFSERR_NOSPC,
363 	NFSERR_ROFS,
364 	NFSERR_NAMETOL,
365 	NFSERR_DQUOT,
366 	NFSERR_STALE,
367 	NFSERR_BADHANDLE,
368 	NFSERR_NOTSUPP,
369 	NFSERR_SERVERFAULT,
370 	0,
371 };
372 
373 static const short nfsv3err_symlink[] = {
374 	NFSERR_IO,
375 	NFSERR_IO,
376 	NFSERR_ACCES,
377 	NFSERR_EXIST,
378 	NFSERR_NOTDIR,
379 	NFSERR_NOSPC,
380 	NFSERR_ROFS,
381 	NFSERR_NAMETOL,
382 	NFSERR_DQUOT,
383 	NFSERR_STALE,
384 	NFSERR_BADHANDLE,
385 	NFSERR_NOTSUPP,
386 	NFSERR_SERVERFAULT,
387 	0,
388 };
389 
390 static const short nfsv3err_mknod[] = {
391 	NFSERR_IO,
392 	NFSERR_IO,
393 	NFSERR_ACCES,
394 	NFSERR_EXIST,
395 	NFSERR_NOTDIR,
396 	NFSERR_NOSPC,
397 	NFSERR_ROFS,
398 	NFSERR_NAMETOL,
399 	NFSERR_DQUOT,
400 	NFSERR_STALE,
401 	NFSERR_BADHANDLE,
402 	NFSERR_NOTSUPP,
403 	NFSERR_SERVERFAULT,
404 	NFSERR_BADTYPE,
405 	0,
406 };
407 
408 static const short nfsv3err_remove[] = {
409 	NFSERR_IO,
410 	NFSERR_NOENT,
411 	NFSERR_IO,
412 	NFSERR_ACCES,
413 	NFSERR_NOTDIR,
414 	NFSERR_ROFS,
415 	NFSERR_NAMETOL,
416 	NFSERR_STALE,
417 	NFSERR_BADHANDLE,
418 	NFSERR_SERVERFAULT,
419 	0,
420 };
421 
422 static const short nfsv3err_rmdir[] = {
423 	NFSERR_IO,
424 	NFSERR_NOENT,
425 	NFSERR_IO,
426 	NFSERR_ACCES,
427 	NFSERR_EXIST,
428 	NFSERR_NOTDIR,
429 	NFSERR_INVAL,
430 	NFSERR_ROFS,
431 	NFSERR_NAMETOL,
432 	NFSERR_NOTEMPTY,
433 	NFSERR_STALE,
434 	NFSERR_BADHANDLE,
435 	NFSERR_NOTSUPP,
436 	NFSERR_SERVERFAULT,
437 	0,
438 };
439 
440 static const short nfsv3err_rename[] = {
441 	NFSERR_IO,
442 	NFSERR_NOENT,
443 	NFSERR_IO,
444 	NFSERR_ACCES,
445 	NFSERR_EXIST,
446 	NFSERR_XDEV,
447 	NFSERR_NOTDIR,
448 	NFSERR_ISDIR,
449 	NFSERR_INVAL,
450 	NFSERR_NOSPC,
451 	NFSERR_ROFS,
452 	NFSERR_MLINK,
453 	NFSERR_NAMETOL,
454 	NFSERR_NOTEMPTY,
455 	NFSERR_DQUOT,
456 	NFSERR_STALE,
457 	NFSERR_BADHANDLE,
458 	NFSERR_NOTSUPP,
459 	NFSERR_SERVERFAULT,
460 	0,
461 };
462 
463 static const short nfsv3err_link[] = {
464 	NFSERR_IO,
465 	NFSERR_IO,
466 	NFSERR_ACCES,
467 	NFSERR_EXIST,
468 	NFSERR_XDEV,
469 	NFSERR_NOTDIR,
470 	NFSERR_INVAL,
471 	NFSERR_NOSPC,
472 	NFSERR_ROFS,
473 	NFSERR_MLINK,
474 	NFSERR_NAMETOL,
475 	NFSERR_DQUOT,
476 	NFSERR_STALE,
477 	NFSERR_BADHANDLE,
478 	NFSERR_NOTSUPP,
479 	NFSERR_SERVERFAULT,
480 	0,
481 };
482 
483 static const short nfsv3err_readdir[] = {
484 	NFSERR_IO,
485 	NFSERR_IO,
486 	NFSERR_ACCES,
487 	NFSERR_NOTDIR,
488 	NFSERR_STALE,
489 	NFSERR_BADHANDLE,
490 	NFSERR_BAD_COOKIE,
491 	NFSERR_TOOSMALL,
492 	NFSERR_SERVERFAULT,
493 	0,
494 };
495 
496 static const short nfsv3err_readdirplus[] = {
497 	NFSERR_IO,
498 	NFSERR_IO,
499 	NFSERR_ACCES,
500 	NFSERR_NOTDIR,
501 	NFSERR_STALE,
502 	NFSERR_BADHANDLE,
503 	NFSERR_BAD_COOKIE,
504 	NFSERR_NOTSUPP,
505 	NFSERR_TOOSMALL,
506 	NFSERR_SERVERFAULT,
507 	0,
508 };
509 
510 static const short nfsv3err_fsstat[] = {
511 	NFSERR_IO,
512 	NFSERR_IO,
513 	NFSERR_STALE,
514 	NFSERR_BADHANDLE,
515 	NFSERR_SERVERFAULT,
516 	0,
517 };
518 
519 static const short nfsv3err_fsinfo[] = {
520 	NFSERR_STALE,
521 	NFSERR_STALE,
522 	NFSERR_BADHANDLE,
523 	NFSERR_SERVERFAULT,
524 	0,
525 };
526 
527 static const short nfsv3err_pathconf[] = {
528 	NFSERR_STALE,
529 	NFSERR_STALE,
530 	NFSERR_BADHANDLE,
531 	NFSERR_SERVERFAULT,
532 	0,
533 };
534 
535 static const short nfsv3err_commit[] = {
536 	NFSERR_IO,
537 	NFSERR_IO,
538 	NFSERR_STALE,
539 	NFSERR_BADHANDLE,
540 	NFSERR_SERVERFAULT,
541 	0,
542 };
543 
544 static const short * const nfsrv_v3errmap[] = {
545 	nfsv3err_null,
546 	nfsv3err_getattr,
547 	nfsv3err_setattr,
548 	nfsv3err_lookup,
549 	nfsv3err_access,
550 	nfsv3err_readlink,
551 	nfsv3err_read,
552 	nfsv3err_write,
553 	nfsv3err_create,
554 	nfsv3err_mkdir,
555 	nfsv3err_symlink,
556 	nfsv3err_mknod,
557 	nfsv3err_remove,
558 	nfsv3err_rmdir,
559 	nfsv3err_rename,
560 	nfsv3err_link,
561 	nfsv3err_readdir,
562 	nfsv3err_readdirplus,
563 	nfsv3err_fsstat,
564 	nfsv3err_fsinfo,
565 	nfsv3err_pathconf,
566 	nfsv3err_commit,
567 };
568 
569 extern struct nfsrtt nfsrtt;
570 extern time_t nqnfsstarttime;
571 extern int nqsrv_clockskew;
572 extern int nqsrv_writeslack;
573 extern int nqsrv_maxlease;
574 extern const int nqnfs_piggy[NFS_NPROCS];
575 extern struct nfsnodehashhead *nfsnodehashtbl;
576 extern u_long nfsnodehash;
577 
578 LIST_HEAD(nfsnodehashhead, nfsnode);
579 u_long nfsdirhashmask;
580 
581 int nfs_webnamei __P((struct nameidata *, struct vnode *, struct proc *));
582 
583 /*
584  * Create the header for an rpc request packet
585  * The hsiz is the size of the rest of the nfs request header.
586  * (just used to decide if a cluster is a good idea)
587  */
588 struct mbuf *
589 nfsm_reqh(vp, procid, hsiz, bposp)
590 	struct vnode *vp;
591 	u_long procid;
592 	int hsiz;
593 	caddr_t *bposp;
594 {
595 	struct mbuf *mb;
596 	caddr_t bpos;
597 	struct nfsmount *nmp;
598 #ifndef NFS_V2_ONLY
599 	u_int32_t *tl;
600 	struct mbuf *mb2;
601 	int nqflag;
602 #endif
603 
604 	MGET(mb, M_WAIT, MT_DATA);
605 	if (hsiz >= MINCLSIZE)
606 		MCLGET(mb, M_WAIT);
607 	mb->m_len = 0;
608 	bpos = mtod(mb, caddr_t);
609 
610 	/*
611 	 * For NQNFS, add lease request.
612 	 */
613 	if (vp) {
614 		nmp = VFSTONFS(vp->v_mount);
615 #ifndef NFS_V2_ONLY
616 		if (nmp->nm_flag & NFSMNT_NQNFS) {
617 			nqflag = NQNFS_NEEDLEASE(vp, procid);
618 			if (nqflag) {
619 				nfsm_build(tl, u_int32_t *, 2*NFSX_UNSIGNED);
620 				*tl++ = txdr_unsigned(nqflag);
621 				*tl = txdr_unsigned(nmp->nm_leaseterm);
622 			} else {
623 				nfsm_build(tl, u_int32_t *, NFSX_UNSIGNED);
624 				*tl = 0;
625 			}
626 		}
627 #endif
628 	}
629 	/* Finally, return values */
630 	*bposp = bpos;
631 	return (mb);
632 }
633 
634 /*
635  * Build the RPC header and fill in the authorization info.
636  * The authorization string argument is only used when the credentials
637  * come from outside of the kernel.
638  * Returns the head of the mbuf list.
639  */
640 struct mbuf *
641 nfsm_rpchead(cr, nmflag, procid, auth_type, auth_len, auth_str, verf_len,
642 	verf_str, mrest, mrest_len, mbp, xidp)
643 	struct ucred *cr;
644 	int nmflag;
645 	int procid;
646 	int auth_type;
647 	int auth_len;
648 	char *auth_str;
649 	int verf_len;
650 	char *verf_str;
651 	struct mbuf *mrest;
652 	int mrest_len;
653 	struct mbuf **mbp;
654 	u_int32_t *xidp;
655 {
656 	struct mbuf *mb;
657 	u_int32_t *tl;
658 	caddr_t bpos;
659 	int i;
660 	struct mbuf *mreq, *mb2;
661 	int siz, grpsiz, authsiz;
662 	struct timeval tv;
663 	static u_int32_t base;
664 
665 	authsiz = nfsm_rndup(auth_len);
666 	MGETHDR(mb, M_WAIT, MT_DATA);
667 	if ((authsiz + 10 * NFSX_UNSIGNED) >= MINCLSIZE) {
668 		MCLGET(mb, M_WAIT);
669 	} else if ((authsiz + 10 * NFSX_UNSIGNED) < MHLEN) {
670 		MH_ALIGN(mb, authsiz + 10 * NFSX_UNSIGNED);
671 	} else {
672 		MH_ALIGN(mb, 8 * NFSX_UNSIGNED);
673 	}
674 	mb->m_len = 0;
675 	mreq = mb;
676 	bpos = mtod(mb, caddr_t);
677 
678 	/*
679 	 * First the RPC header.
680 	 */
681 	nfsm_build(tl, u_int32_t *, 8 * NFSX_UNSIGNED);
682 
683 	/*
684 	 * derive initial xid from system time
685 	 * XXX time is invalid if root not yet mounted
686 	 */
687 	if (!base && (rootvp)) {
688 		microtime(&tv);
689 		base = tv.tv_sec << 12;
690 		nfs_xid = base;
691 	}
692 	/*
693 	 * Skip zero xid if it should ever happen.
694 	 */
695 	if (++nfs_xid == 0)
696 		nfs_xid++;
697 
698 	*tl++ = *xidp = txdr_unsigned(nfs_xid);
699 	*tl++ = rpc_call;
700 	*tl++ = rpc_vers;
701 	if (nmflag & NFSMNT_NQNFS) {
702 		*tl++ = txdr_unsigned(NQNFS_PROG);
703 		*tl++ = txdr_unsigned(NQNFS_VER3);
704 	} else {
705 		*tl++ = txdr_unsigned(NFS_PROG);
706 		if (nmflag & NFSMNT_NFSV3)
707 			*tl++ = txdr_unsigned(NFS_VER3);
708 		else
709 			*tl++ = txdr_unsigned(NFS_VER2);
710 	}
711 	if (nmflag & NFSMNT_NFSV3)
712 		*tl++ = txdr_unsigned(procid);
713 	else
714 		*tl++ = txdr_unsigned(nfsv2_procid[procid]);
715 
716 	/*
717 	 * And then the authorization cred.
718 	 */
719 	*tl++ = txdr_unsigned(auth_type);
720 	*tl = txdr_unsigned(authsiz);
721 	switch (auth_type) {
722 	case RPCAUTH_UNIX:
723 		nfsm_build(tl, u_int32_t *, auth_len);
724 		*tl++ = 0;		/* stamp ?? */
725 		*tl++ = 0;		/* NULL hostname */
726 		*tl++ = txdr_unsigned(cr->cr_uid);
727 		*tl++ = txdr_unsigned(cr->cr_gid);
728 		grpsiz = (auth_len >> 2) - 5;
729 		*tl++ = txdr_unsigned(grpsiz);
730 		for (i = 0; i < grpsiz; i++)
731 			*tl++ = txdr_unsigned(cr->cr_groups[i]);
732 		break;
733 	case RPCAUTH_KERB4:
734 		siz = auth_len;
735 		while (siz > 0) {
736 			if (M_TRAILINGSPACE(mb) == 0) {
737 				MGET(mb2, M_WAIT, MT_DATA);
738 				if (siz >= MINCLSIZE)
739 					MCLGET(mb2, M_WAIT);
740 				mb->m_next = mb2;
741 				mb = mb2;
742 				mb->m_len = 0;
743 				bpos = mtod(mb, caddr_t);
744 			}
745 			i = min(siz, M_TRAILINGSPACE(mb));
746 			memcpy(bpos, auth_str, i);
747 			mb->m_len += i;
748 			auth_str += i;
749 			bpos += i;
750 			siz -= i;
751 		}
752 		if ((siz = (nfsm_rndup(auth_len) - auth_len)) > 0) {
753 			for (i = 0; i < siz; i++)
754 				*bpos++ = '\0';
755 			mb->m_len += siz;
756 		}
757 		break;
758 	};
759 
760 	/*
761 	 * And the verifier...
762 	 */
763 	nfsm_build(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
764 	if (verf_str) {
765 		*tl++ = txdr_unsigned(RPCAUTH_KERB4);
766 		*tl = txdr_unsigned(verf_len);
767 		siz = verf_len;
768 		while (siz > 0) {
769 			if (M_TRAILINGSPACE(mb) == 0) {
770 				MGET(mb2, M_WAIT, MT_DATA);
771 				if (siz >= MINCLSIZE)
772 					MCLGET(mb2, M_WAIT);
773 				mb->m_next = mb2;
774 				mb = mb2;
775 				mb->m_len = 0;
776 				bpos = mtod(mb, caddr_t);
777 			}
778 			i = min(siz, M_TRAILINGSPACE(mb));
779 			memcpy(bpos, verf_str, i);
780 			mb->m_len += i;
781 			verf_str += i;
782 			bpos += i;
783 			siz -= i;
784 		}
785 		if ((siz = (nfsm_rndup(verf_len) - verf_len)) > 0) {
786 			for (i = 0; i < siz; i++)
787 				*bpos++ = '\0';
788 			mb->m_len += siz;
789 		}
790 	} else {
791 		*tl++ = txdr_unsigned(RPCAUTH_NULL);
792 		*tl = 0;
793 	}
794 	mb->m_next = mrest;
795 	mreq->m_pkthdr.len = authsiz + 10 * NFSX_UNSIGNED + mrest_len;
796 	mreq->m_pkthdr.rcvif = (struct ifnet *)0;
797 	*mbp = mb;
798 	return (mreq);
799 }
800 
801 /*
802  * copies mbuf chain to the uio scatter/gather list
803  */
804 int
805 nfsm_mbuftouio(mrep, uiop, siz, dpos)
806 	struct mbuf **mrep;
807 	struct uio *uiop;
808 	int siz;
809 	caddr_t *dpos;
810 {
811 	char *mbufcp, *uiocp;
812 	int xfer, left, len;
813 	struct mbuf *mp;
814 	long uiosiz, rem;
815 	int error = 0;
816 
817 	mp = *mrep;
818 	mbufcp = *dpos;
819 	len = mtod(mp, caddr_t)+mp->m_len-mbufcp;
820 	rem = nfsm_rndup(siz)-siz;
821 	while (siz > 0) {
822 		if (uiop->uio_iovcnt <= 0 || uiop->uio_iov == NULL)
823 			return (EFBIG);
824 		left = uiop->uio_iov->iov_len;
825 		uiocp = uiop->uio_iov->iov_base;
826 		if (left > siz)
827 			left = siz;
828 		uiosiz = left;
829 		while (left > 0) {
830 			while (len == 0) {
831 				mp = mp->m_next;
832 				if (mp == NULL)
833 					return (EBADRPC);
834 				mbufcp = mtod(mp, caddr_t);
835 				len = mp->m_len;
836 			}
837 			xfer = (left > len) ? len : left;
838 #ifdef notdef
839 			/* Not Yet.. */
840 			if (uiop->uio_iov->iov_op != NULL)
841 				(*(uiop->uio_iov->iov_op))
842 				(mbufcp, uiocp, xfer);
843 			else
844 #endif
845 			if (uiop->uio_segflg == UIO_SYSSPACE)
846 				memcpy(uiocp, mbufcp, xfer);
847 			else
848 				copyout(mbufcp, uiocp, xfer);
849 			left -= xfer;
850 			len -= xfer;
851 			mbufcp += xfer;
852 			uiocp += xfer;
853 			uiop->uio_offset += xfer;
854 			uiop->uio_resid -= xfer;
855 		}
856 		if (uiop->uio_iov->iov_len <= siz) {
857 			uiop->uio_iovcnt--;
858 			uiop->uio_iov++;
859 		} else {
860 			uiop->uio_iov->iov_base =
861 			    (caddr_t)uiop->uio_iov->iov_base + uiosiz;
862 			uiop->uio_iov->iov_len -= uiosiz;
863 		}
864 		siz -= uiosiz;
865 	}
866 	*dpos = mbufcp;
867 	*mrep = mp;
868 	if (rem > 0) {
869 		if (len < rem)
870 			error = nfs_adv(mrep, dpos, rem, len);
871 		else
872 			*dpos += rem;
873 	}
874 	return (error);
875 }
876 
877 /*
878  * copies a uio scatter/gather list to an mbuf chain.
879  * NOTE: can ony handle iovcnt == 1
880  */
881 int
882 nfsm_uiotombuf(uiop, mq, siz, bpos)
883 	struct uio *uiop;
884 	struct mbuf **mq;
885 	int siz;
886 	caddr_t *bpos;
887 {
888 	char *uiocp;
889 	struct mbuf *mp, *mp2;
890 	int xfer, left, mlen;
891 	int uiosiz, clflg, rem;
892 	char *cp;
893 
894 #ifdef DIAGNOSTIC
895 	if (uiop->uio_iovcnt != 1)
896 		panic("nfsm_uiotombuf: iovcnt != 1");
897 #endif
898 
899 	if (siz > MLEN)		/* or should it >= MCLBYTES ?? */
900 		clflg = 1;
901 	else
902 		clflg = 0;
903 	rem = nfsm_rndup(siz)-siz;
904 	mp = mp2 = *mq;
905 	while (siz > 0) {
906 		left = uiop->uio_iov->iov_len;
907 		uiocp = uiop->uio_iov->iov_base;
908 		if (left > siz)
909 			left = siz;
910 		uiosiz = left;
911 		while (left > 0) {
912 			mlen = M_TRAILINGSPACE(mp);
913 			if (mlen == 0) {
914 				MGET(mp, M_WAIT, MT_DATA);
915 				if (clflg)
916 					MCLGET(mp, M_WAIT);
917 				mp->m_len = 0;
918 				mp2->m_next = mp;
919 				mp2 = mp;
920 				mlen = M_TRAILINGSPACE(mp);
921 			}
922 			xfer = (left > mlen) ? mlen : left;
923 #ifdef notdef
924 			/* Not Yet.. */
925 			if (uiop->uio_iov->iov_op != NULL)
926 				(*(uiop->uio_iov->iov_op))
927 				(uiocp, mtod(mp, caddr_t)+mp->m_len, xfer);
928 			else
929 #endif
930 			if (uiop->uio_segflg == UIO_SYSSPACE)
931 				memcpy(mtod(mp, caddr_t)+mp->m_len, uiocp, xfer);
932 			else
933 				copyin(uiocp, mtod(mp, caddr_t)+mp->m_len, xfer);
934 			mp->m_len += xfer;
935 			left -= xfer;
936 			uiocp += xfer;
937 			uiop->uio_offset += xfer;
938 			uiop->uio_resid -= xfer;
939 		}
940 		uiop->uio_iov->iov_base = (caddr_t)uiop->uio_iov->iov_base +
941 		    uiosiz;
942 		uiop->uio_iov->iov_len -= uiosiz;
943 		siz -= uiosiz;
944 	}
945 	if (rem > 0) {
946 		if (rem > M_TRAILINGSPACE(mp)) {
947 			MGET(mp, M_WAIT, MT_DATA);
948 			mp->m_len = 0;
949 			mp2->m_next = mp;
950 		}
951 		cp = mtod(mp, caddr_t)+mp->m_len;
952 		for (left = 0; left < rem; left++)
953 			*cp++ = '\0';
954 		mp->m_len += rem;
955 		*bpos = cp;
956 	} else
957 		*bpos = mtod(mp, caddr_t)+mp->m_len;
958 	*mq = mp;
959 	return (0);
960 }
961 
962 /*
963  * Get at least "siz" bytes of correctly aligned data.
964  * When called the mbuf pointers are not necessarily correct,
965  * dsosp points to what ought to be in m_data and left contains
966  * what ought to be in m_len.
967  * This is used by the macros nfsm_dissect and nfsm_dissecton for tough
968  * cases. (The macros use the vars. dpos and dpos2)
969  */
970 int
971 nfsm_disct(mdp, dposp, siz, left, cp2)
972 	struct mbuf **mdp;
973 	caddr_t *dposp;
974 	int siz;
975 	int left;
976 	caddr_t *cp2;
977 {
978 	struct mbuf *m1, *m2;
979 	struct mbuf *havebuf = NULL;
980 	caddr_t src = *dposp;
981 	caddr_t dst;
982 	int len;
983 
984 #ifdef DEBUG
985 	if (left < 0)
986 		panic("nfsm_disct: left < 0");
987 #endif
988 	m1 = *mdp;
989 	/*
990 	 * Skip through the mbuf chain looking for an mbuf with
991 	 * some data. If the first mbuf found has enough data
992 	 * and it is correctly aligned return it.
993 	 */
994 	while (left == 0) {
995 		havebuf = m1;
996 		*mdp = m1 = m1->m_next;
997 		if (m1 == NULL)
998 			return (EBADRPC);
999 		src = mtod(m1, caddr_t);
1000 		left = m1->m_len;
1001 		/*
1002 		 * If we start a new mbuf and it is big enough
1003 		 * and correctly aligned just return it, don't
1004 		 * do any pull up.
1005 		 */
1006 		if (left >= siz && nfsm_aligned(src)) {
1007 			*cp2 = src;
1008 			*dposp = src + siz;
1009 			return (0);
1010 		}
1011 	}
1012 	if (m1->m_flags & M_EXT) {
1013 		if (havebuf) {
1014 			/* If the first mbuf with data has external data
1015 			 * and there is a previous empty mbuf use it
1016 			 * to move the data into.
1017 			 */
1018 			m2 = m1;
1019 			*mdp = m1 = havebuf;
1020 			if (m1->m_flags & M_EXT) {
1021 				MEXTREMOVE(m1);
1022 			}
1023 		} else {
1024 			/*
1025 			 * If the first mbuf has a external data
1026 			 * and there is no previous empty mbuf
1027 			 * allocate a new mbuf and move the external
1028 			 * data to the new mbuf. Also make the first
1029 			 * mbuf look empty.
1030 			 */
1031 			m2 = m_get(M_WAIT, MT_DATA);
1032 			m2->m_ext = m1->m_ext;
1033 			m2->m_data = src;
1034 			m2->m_len = left;
1035 			MCLADDREFERENCE(m1, m2);
1036 			MEXTREMOVE(m1);
1037 			m2->m_next = m1->m_next;
1038 			m1->m_next = m2;
1039 		}
1040 		m1->m_len = 0;
1041 		dst = m1->m_dat;
1042 	} else {
1043 		/*
1044 		 * If the first mbuf has no external data
1045 		 * move the data to the front of the mbuf.
1046 		 */
1047 		if ((dst = m1->m_dat) != src)
1048 			memmove(dst, src, left);
1049 		dst += left;
1050 		m1->m_len = left;
1051 		m2 = m1->m_next;
1052 	}
1053 	m1->m_flags &= ~M_PKTHDR;
1054 	*cp2 = m1->m_data = m1->m_dat;   /* data is at beginning of buffer */
1055 	*dposp = mtod(m1, caddr_t) + siz;
1056 	/*
1057 	 * Loop through mbufs pulling data up into first mbuf until
1058 	 * the first mbuf is full or there is no more data to
1059 	 * pullup.
1060 	 */
1061 	while ((len = (MLEN - m1->m_len)) != 0 && m2) {
1062 		if ((len = min(len, m2->m_len)) != 0)
1063 			memcpy(dst, m2->m_data, len);
1064 		m1->m_len += len;
1065 		dst += len;
1066 		m2->m_data += len;
1067 		m2->m_len -= len;
1068 		m2 = m2->m_next;
1069 	}
1070 	if (m1->m_len < siz)
1071 		return (EBADRPC);
1072 	return (0);
1073 }
1074 
1075 /*
1076  * Advance the position in the mbuf chain.
1077  */
1078 int
1079 nfs_adv(mdp, dposp, offs, left)
1080 	struct mbuf **mdp;
1081 	caddr_t *dposp;
1082 	int offs;
1083 	int left;
1084 {
1085 	struct mbuf *m;
1086 	int s;
1087 
1088 	m = *mdp;
1089 	s = left;
1090 	while (s < offs) {
1091 		offs -= s;
1092 		m = m->m_next;
1093 		if (m == NULL)
1094 			return (EBADRPC);
1095 		s = m->m_len;
1096 	}
1097 	*mdp = m;
1098 	*dposp = mtod(m, caddr_t)+offs;
1099 	return (0);
1100 }
1101 
1102 /*
1103  * Copy a string into mbufs for the hard cases...
1104  */
1105 int
1106 nfsm_strtmbuf(mb, bpos, cp, siz)
1107 	struct mbuf **mb;
1108 	char **bpos;
1109 	const char *cp;
1110 	long siz;
1111 {
1112 	struct mbuf *m1 = NULL, *m2;
1113 	long left, xfer, len, tlen;
1114 	u_int32_t *tl;
1115 	int putsize;
1116 
1117 	putsize = 1;
1118 	m2 = *mb;
1119 	left = M_TRAILINGSPACE(m2);
1120 	if (left > 0) {
1121 		tl = ((u_int32_t *)(*bpos));
1122 		*tl++ = txdr_unsigned(siz);
1123 		putsize = 0;
1124 		left -= NFSX_UNSIGNED;
1125 		m2->m_len += NFSX_UNSIGNED;
1126 		if (left > 0) {
1127 			memcpy((caddr_t) tl, cp, left);
1128 			siz -= left;
1129 			cp += left;
1130 			m2->m_len += left;
1131 			left = 0;
1132 		}
1133 	}
1134 	/* Loop around adding mbufs */
1135 	while (siz > 0) {
1136 		MGET(m1, M_WAIT, MT_DATA);
1137 		if (siz > MLEN)
1138 			MCLGET(m1, M_WAIT);
1139 		m1->m_len = NFSMSIZ(m1);
1140 		m2->m_next = m1;
1141 		m2 = m1;
1142 		tl = mtod(m1, u_int32_t *);
1143 		tlen = 0;
1144 		if (putsize) {
1145 			*tl++ = txdr_unsigned(siz);
1146 			m1->m_len -= NFSX_UNSIGNED;
1147 			tlen = NFSX_UNSIGNED;
1148 			putsize = 0;
1149 		}
1150 		if (siz < m1->m_len) {
1151 			len = nfsm_rndup(siz);
1152 			xfer = siz;
1153 			if (xfer < len)
1154 				*(tl+(xfer>>2)) = 0;
1155 		} else {
1156 			xfer = len = m1->m_len;
1157 		}
1158 		memcpy((caddr_t) tl, cp, xfer);
1159 		m1->m_len = len+tlen;
1160 		siz -= xfer;
1161 		cp += xfer;
1162 	}
1163 	*mb = m1;
1164 	*bpos = mtod(m1, caddr_t)+m1->m_len;
1165 	return (0);
1166 }
1167 
1168 /*
1169  * Directory caching routines. They work as follows:
1170  * - a cache is maintained per VDIR nfsnode.
1171  * - for each offset cookie that is exported to userspace, and can
1172  *   thus be thrown back at us as an offset to VOP_READDIR, store
1173  *   information in the cache.
1174  * - cached are:
1175  *   - cookie itself
1176  *   - blocknumber (essentially just a search key in the buffer cache)
1177  *   - entry number in block.
1178  *   - offset cookie of block in which this entry is stored
1179  *   - 32 bit cookie if NFSMNT_XLATECOOKIE is used.
1180  * - entries are looked up in a hash table
1181  * - also maintained is an LRU list of entries, used to determine
1182  *   which ones to delete if the cache grows too large.
1183  * - if 32 <-> 64 translation mode is requested for a filesystem,
1184  *   the cache also functions as a translation table
1185  * - in the translation case, invalidating the cache does not mean
1186  *   flushing it, but just marking entries as invalid, except for
1187  *   the <64bit cookie, 32bitcookie> pair which is still valid, to
1188  *   still be able to use the cache as a translation table.
1189  * - 32 bit cookies are uniquely created by combining the hash table
1190  *   entry value, and one generation count per hash table entry,
1191  *   incremented each time an entry is appended to the chain.
1192  * - the cache is invalidated each time a direcory is modified
1193  * - sanity checks are also done; if an entry in a block turns
1194  *   out not to have a matching cookie, the cache is invalidated
1195  *   and a new block starting from the wanted offset is fetched from
1196  *   the server.
1197  * - directory entries as read from the server are extended to contain
1198  *   the 64bit and, optionally, the 32bit cookies, for sanity checking
1199  *   the cache and exporting them to userspace through the cookie
1200  *   argument to VOP_READDIR.
1201  */
1202 
1203 u_long
1204 nfs_dirhash(off)
1205 	off_t off;
1206 {
1207 	int i;
1208 	char *cp = (char *)&off;
1209 	u_long sum = 0L;
1210 
1211 	for (i = 0 ; i < sizeof (off); i++)
1212 		sum += *cp++;
1213 
1214 	return sum;
1215 }
1216 
1217 void
1218 nfs_initdircache(vp)
1219 	struct vnode *vp;
1220 {
1221 	struct nfsnode *np = VTONFS(vp);
1222 	struct nfsmount *nmp = VFSTONFS(vp->v_mount);
1223 
1224 	np->n_dircachesize = 0;
1225 	np->n_dblkno = 1;
1226 	np->n_dircache = hashinit(NFS_DIRHASHSIZ, HASH_LIST, M_NFSDIROFF,
1227 	    M_WAITOK, &nfsdirhashmask);
1228 	TAILQ_INIT(&np->n_dirchain);
1229 	if (nmp->nm_flag & NFSMNT_XLATECOOKIE) {
1230 		MALLOC(np->n_dirgens, unsigned *,
1231 		    NFS_DIRHASHSIZ * sizeof (unsigned), M_NFSDIROFF,
1232 		    M_WAITOK);
1233 		memset((caddr_t)np->n_dirgens, 0,
1234 		    NFS_DIRHASHSIZ * sizeof (unsigned));
1235 	}
1236 }
1237 
1238 static struct nfsdircache dzero = {0, 0, {0, 0}, {0, 0}, 0, 0, 0};
1239 
1240 struct nfsdircache *
1241 nfs_searchdircache(vp, off, do32, hashent)
1242 	struct vnode *vp;
1243 	off_t off;
1244 	int do32;
1245 	int *hashent;
1246 {
1247 	struct nfsdirhashhead *ndhp;
1248 	struct nfsdircache *ndp = NULL;
1249 	struct nfsnode *np = VTONFS(vp);
1250 	unsigned ent;
1251 
1252 	/*
1253 	 * Zero is always a valid cookie.
1254 	 */
1255 	if (off == 0)
1256 		return &dzero;
1257 
1258 	/*
1259 	 * We use a 32bit cookie as search key, directly reconstruct
1260 	 * the hashentry. Else use the hashfunction.
1261 	 */
1262 	if (do32) {
1263 		ent = (u_int32_t)off >> 24;
1264 		if (ent >= NFS_DIRHASHSIZ)
1265 			return NULL;
1266 		ndhp = &np->n_dircache[ent];
1267 	} else {
1268 		ndhp = NFSDIRHASH(np, off);
1269 	}
1270 
1271 	if (hashent)
1272 		*hashent = (int)(ndhp - np->n_dircache);
1273 	if (do32) {
1274 		for (ndp = ndhp->lh_first; ndp; ndp = ndp->dc_hash.le_next) {
1275 			if (ndp->dc_cookie32 == (u_int32_t)off) {
1276 				/*
1277 				 * An invalidated entry will become the
1278 				 * start of a new block fetched from
1279 				 * the server.
1280 				 */
1281 				if (ndp->dc_blkno == -1) {
1282 					ndp->dc_blkcookie = ndp->dc_cookie;
1283 					ndp->dc_blkno = np->n_dblkno++;
1284 					ndp->dc_entry = 0;
1285 				}
1286 				break;
1287 			}
1288 		}
1289 	} else {
1290 		for (ndp = ndhp->lh_first; ndp; ndp = ndp->dc_hash.le_next)
1291 			if (ndp->dc_cookie == off)
1292 				break;
1293 	}
1294 	return ndp;
1295 }
1296 
1297 
1298 struct nfsdircache *
1299 nfs_enterdircache(vp, off, blkoff, en, blkno)
1300 	struct vnode *vp;
1301 	off_t off, blkoff;
1302 	daddr_t blkno;
1303 	int en;
1304 {
1305 	struct nfsnode *np = VTONFS(vp);
1306 	struct nfsdirhashhead *ndhp;
1307 	struct nfsdircache *ndp = NULL, *first;
1308 	struct nfsmount *nmp = VFSTONFS(vp->v_mount);
1309 	int hashent, gen, overwrite;
1310 
1311 	if (!np->n_dircache)
1312 		/*
1313 		 * XXX would like to do this in nfs_nget but vtype
1314 		 * isn't known at that time.
1315 		 */
1316 		nfs_initdircache(vp);
1317 
1318 	/*
1319 	 * XXX refuse entries for offset 0. amd(8) erroneously sets
1320 	 * cookie 0 for the '.' entry, making this necessary. This
1321 	 * isn't so bad, as 0 is a special case anyway.
1322 	 */
1323 	if (off == 0)
1324 		return &dzero;
1325 
1326 	ndp = nfs_searchdircache(vp, off, 0, &hashent);
1327 
1328 	if (ndp && ndp->dc_blkno != -1) {
1329 		/*
1330 		 * Overwriting an old entry. Check if it's the same.
1331 		 * If so, just return. If not, remove the old entry.
1332 		 */
1333 		if (ndp->dc_blkcookie == blkoff && ndp->dc_entry == en)
1334 			return ndp;
1335 		TAILQ_REMOVE(&np->n_dirchain, ndp, dc_chain);
1336 		LIST_REMOVE(ndp, dc_hash);
1337 		FREE(ndp, M_NFSDIROFF);
1338 		ndp = 0;
1339 	}
1340 
1341 	ndhp = &np->n_dircache[hashent];
1342 
1343 	if (!ndp) {
1344 		MALLOC(ndp, struct nfsdircache *, sizeof (*ndp), M_NFSDIROFF,
1345 		    M_WAITOK);
1346 		overwrite = 0;
1347 		if (nmp->nm_flag & NFSMNT_XLATECOOKIE) {
1348 			/*
1349 			 * We're allocating a new entry, so bump the
1350 			 * generation number.
1351 			 */
1352 			gen = ++np->n_dirgens[hashent];
1353 			if (gen == 0) {
1354 				np->n_dirgens[hashent]++;
1355 				gen++;
1356 			}
1357 			ndp->dc_cookie32 = (hashent << 24) | (gen & 0xffffff);
1358 		}
1359 	} else
1360 		overwrite = 1;
1361 
1362 	/*
1363 	 * If the entry number is 0, we are at the start of a new block, so
1364 	 * allocate a new blocknumber.
1365 	 */
1366 	if (en == 0)
1367 		ndp->dc_blkno = np->n_dblkno++;
1368 	else
1369 		ndp->dc_blkno = blkno;
1370 
1371 	ndp->dc_cookie = off;
1372 	ndp->dc_blkcookie = blkoff;
1373 	ndp->dc_entry = en;
1374 
1375 	if (overwrite)
1376 		return ndp;
1377 
1378 	/*
1379 	 * If the maximum directory cookie cache size has been reached
1380 	 * for this node, take one off the front. The idea is that
1381 	 * directories are typically read front-to-back once, so that
1382 	 * the oldest entries can be thrown away without much performance
1383 	 * loss.
1384 	 */
1385 	if (np->n_dircachesize == NFS_MAXDIRCACHE) {
1386 		first = np->n_dirchain.tqh_first;
1387 		TAILQ_REMOVE(&np->n_dirchain, first, dc_chain);
1388 		LIST_REMOVE(first, dc_hash);
1389 		FREE(first, M_NFSDIROFF);
1390 	} else
1391 		np->n_dircachesize++;
1392 
1393 	LIST_INSERT_HEAD(ndhp, ndp, dc_hash);
1394 	TAILQ_INSERT_TAIL(&np->n_dirchain, ndp, dc_chain);
1395 	return ndp;
1396 }
1397 
1398 void
1399 nfs_invaldircache(vp, forcefree)
1400 	struct vnode *vp;
1401 	int forcefree;
1402 {
1403 	struct nfsnode *np = VTONFS(vp);
1404 	struct nfsdircache *ndp = NULL;
1405 	struct nfsmount *nmp = VFSTONFS(vp->v_mount);
1406 
1407 #ifdef DIAGNOSTIC
1408 	if (vp->v_type != VDIR)
1409 		panic("nfs: invaldircache: not dir");
1410 #endif
1411 
1412 	if (!np->n_dircache)
1413 		return;
1414 
1415 	if (!(nmp->nm_flag & NFSMNT_XLATECOOKIE) || forcefree) {
1416 		while ((ndp = np->n_dirchain.tqh_first)) {
1417 			TAILQ_REMOVE(&np->n_dirchain, ndp, dc_chain);
1418 			LIST_REMOVE(ndp, dc_hash);
1419 			FREE(ndp, M_NFSDIROFF);
1420 		}
1421 		np->n_dircachesize = 0;
1422 		if (forcefree && np->n_dirgens) {
1423 			FREE(np->n_dirgens, M_NFSDIROFF);
1424 		}
1425 	} else {
1426 		for (ndp = np->n_dirchain.tqh_first; ndp;
1427 		    ndp = ndp->dc_chain.tqe_next)
1428 			ndp->dc_blkno = -1;
1429 	}
1430 
1431 	np->n_dblkno = 1;
1432 }
1433 
1434 /*
1435  * Called once before VFS init to initialize shared and
1436  * server-specific data structures.
1437  */
1438 void
1439 nfs_init()
1440 {
1441 	nfsrtt.pos = 0;
1442 	rpc_vers = txdr_unsigned(RPC_VER2);
1443 	rpc_call = txdr_unsigned(RPC_CALL);
1444 	rpc_reply = txdr_unsigned(RPC_REPLY);
1445 	rpc_msgdenied = txdr_unsigned(RPC_MSGDENIED);
1446 	rpc_msgaccepted = txdr_unsigned(RPC_MSGACCEPTED);
1447 	rpc_mismatch = txdr_unsigned(RPC_MISMATCH);
1448 	rpc_autherr = txdr_unsigned(RPC_AUTHERR);
1449 	rpc_auth_unix = txdr_unsigned(RPCAUTH_UNIX);
1450 	rpc_auth_kerb = txdr_unsigned(RPCAUTH_KERB4);
1451 	nfs_prog = txdr_unsigned(NFS_PROG);
1452 	nqnfs_prog = txdr_unsigned(NQNFS_PROG);
1453 	nfs_true = txdr_unsigned(TRUE);
1454 	nfs_false = txdr_unsigned(FALSE);
1455 	nfs_xdrneg1 = txdr_unsigned(-1);
1456 	nfs_ticks = (hz * NFS_TICKINTVL + 500) / 1000;
1457 	if (nfs_ticks < 1)
1458 		nfs_ticks = 1;
1459 #ifdef NFSSERVER
1460 	nfsrv_init(0);			/* Init server data structures */
1461 	nfsrv_initcache();		/* Init the server request cache */
1462 #endif /* NFSSERVER */
1463 
1464 #if defined(NFSSERVER) || !defined(NFS_V2_ONLY)
1465 	/*
1466 	 * Initialize the nqnfs data structures.
1467 	 */
1468 	if (nqnfsstarttime == 0) {
1469 		nqnfsstarttime = boottime.tv_sec + nqsrv_maxlease
1470 			+ nqsrv_clockskew + nqsrv_writeslack;
1471 		NQLOADNOVRAM(nqnfsstarttime);
1472 		CIRCLEQ_INIT(&nqtimerhead);
1473 		nqfhhashtbl = hashinit(NQLCHSZ, HASH_LIST, M_NQLEASE,
1474 		    M_WAITOK, &nqfhhash);
1475 	}
1476 #endif
1477 
1478 	/*
1479 	 * Initialize reply list and start timer
1480 	 */
1481 	TAILQ_INIT(&nfs_reqq);
1482 	nfs_timer(NULL);
1483 }
1484 
1485 #ifdef NFS
1486 /*
1487  * Called once at VFS init to initialize client-specific data structures.
1488  */
1489 void
1490 nfs_vfs_init()
1491 {
1492 	int i;
1493 
1494 	/* Ensure async daemons disabled */
1495 	for (i = 0; i < NFS_MAXASYNCDAEMON; i++) {
1496 		nfs_iodwant[i] = (struct proc *)0;
1497 		nfs_iodmount[i] = (struct nfsmount *)0;
1498 	}
1499 	nfs_nhinit();			/* Init the nfsnode table */
1500 }
1501 
1502 void
1503 nfs_vfs_done()
1504 {
1505 	nfs_nhdone();
1506 }
1507 
1508 /*
1509  * Attribute cache routines.
1510  * nfs_loadattrcache() - loads or updates the cache contents from attributes
1511  *	that are on the mbuf list
1512  * nfs_getattrcache() - returns valid attributes if found in cache, returns
1513  *	error otherwise
1514  */
1515 
1516 /*
1517  * Load the attribute cache (that lives in the nfsnode entry) with
1518  * the values on the mbuf list and
1519  * Iff vap not NULL
1520  *    copy the attributes to *vaper
1521  */
1522 int
1523 nfsm_loadattrcache(vpp, mdp, dposp, vaper)
1524 	struct vnode **vpp;
1525 	struct mbuf **mdp;
1526 	caddr_t *dposp;
1527 	struct vattr *vaper;
1528 {
1529 	int32_t t1;
1530 	caddr_t cp2;
1531 	int error = 0;
1532 	struct mbuf *md;
1533 	int v3 = NFS_ISV3(*vpp);
1534 
1535 	md = *mdp;
1536 	t1 = (mtod(md, caddr_t) + md->m_len) - *dposp;
1537 	error = nfsm_disct(mdp, dposp, NFSX_FATTR(v3), t1, &cp2);
1538 	if (error)
1539 		return (error);
1540 	return nfs_loadattrcache(vpp, (struct nfs_fattr *)cp2, vaper);
1541 }
1542 
1543 int
1544 nfs_loadattrcache(vpp, fp, vaper)
1545 	struct vnode **vpp;
1546 	struct nfs_fattr *fp;
1547 	struct vattr *vaper;
1548 {
1549 	struct vnode *vp = *vpp;
1550 	struct vattr *vap;
1551 	int v3 = NFS_ISV3(vp);
1552 	enum vtype vtyp;
1553 	u_short vmode;
1554 	struct timespec mtime;
1555 	struct vnode *nvp;
1556 	int32_t rdev;
1557 	struct nfsnode *np;
1558 	extern int (**spec_nfsv2nodeop_p) __P((void *));
1559 
1560 	if (v3) {
1561 		vtyp = nfsv3tov_type(fp->fa_type);
1562 		vmode = fxdr_unsigned(u_short, fp->fa_mode);
1563 		rdev = makedev(fxdr_unsigned(u_int32_t, fp->fa3_rdev.specdata1),
1564 			fxdr_unsigned(u_int32_t, fp->fa3_rdev.specdata2));
1565 		fxdr_nfsv3time(&fp->fa3_mtime, &mtime);
1566 	} else {
1567 		vtyp = nfsv2tov_type(fp->fa_type);
1568 		vmode = fxdr_unsigned(u_short, fp->fa_mode);
1569 		if (vtyp == VNON || vtyp == VREG)
1570 			vtyp = IFTOVT(vmode);
1571 		rdev = fxdr_unsigned(int32_t, fp->fa2_rdev);
1572 		fxdr_nfsv2time(&fp->fa2_mtime, &mtime);
1573 
1574 		/*
1575 		 * Really ugly NFSv2 kludge.
1576 		 */
1577 		if (vtyp == VCHR && rdev == 0xffffffff)
1578 			vtyp = VFIFO;
1579 	}
1580 
1581 	/*
1582 	 * If v_type == VNON it is a new node, so fill in the v_type,
1583 	 * n_mtime fields. Check to see if it represents a special
1584 	 * device, and if so, check for a possible alias. Once the
1585 	 * correct vnode has been obtained, fill in the rest of the
1586 	 * information.
1587 	 */
1588 	np = VTONFS(vp);
1589 	if (vp->v_type == VNON) {
1590 		vp->v_type = vtyp;
1591 		if (vp->v_type == VFIFO) {
1592 			extern int (**fifo_nfsv2nodeop_p) __P((void *));
1593 			vp->v_op = fifo_nfsv2nodeop_p;
1594 		}
1595 		if (vp->v_type == VCHR || vp->v_type == VBLK) {
1596 			vp->v_op = spec_nfsv2nodeop_p;
1597 			nvp = checkalias(vp, (dev_t)rdev, vp->v_mount);
1598 			if (nvp) {
1599 				/*
1600 				 * Discard unneeded vnode, but save its nfsnode.
1601 				 * Since the nfsnode does not have a lock, its
1602 				 * vnode lock has to be carried over.
1603 				 */
1604 				/*
1605 				 * XXX is the old node sure to be locked here?
1606 				 */
1607 				KASSERT(lockstatus(&vp->v_lock) ==
1608 				    LK_EXCLUSIVE);
1609 				nvp->v_data = vp->v_data;
1610 				vp->v_data = NULL;
1611 				VOP_UNLOCK(vp, 0);
1612 				vp->v_op = spec_vnodeop_p;
1613 				vrele(vp);
1614 				vgone(vp);
1615 				lockmgr(&nvp->v_lock, LK_EXCLUSIVE,
1616 				    &nvp->v_interlock);
1617 				/*
1618 				 * Reinitialize aliased node.
1619 				 */
1620 				np->n_vnode = nvp;
1621 				*vpp = vp = nvp;
1622 			}
1623 		}
1624 		np->n_mtime = mtime.tv_sec;
1625 	}
1626 	vap = np->n_vattr;
1627 	vap->va_type = vtyp;
1628 	vap->va_mode = vmode & ALLPERMS;
1629 	vap->va_rdev = (dev_t)rdev;
1630 	vap->va_mtime = mtime;
1631 	vap->va_fsid = vp->v_mount->mnt_stat.f_fsid.val[0];
1632 	switch (vtyp) {
1633 	case VDIR:
1634 		vap->va_blocksize = NFS_DIRFRAGSIZ;
1635 		break;
1636 	case VBLK:
1637 		vap->va_blocksize = BLKDEV_IOSIZE;
1638 		break;
1639 	case VCHR:
1640 		vap->va_blocksize = MAXBSIZE;
1641 		break;
1642 	default:
1643 		vap->va_blocksize = v3 ? vp->v_mount->mnt_stat.f_iosize :
1644 		    fxdr_unsigned(int32_t, fp->fa2_blocksize);
1645 		break;
1646 	}
1647 	if (v3) {
1648 		vap->va_nlink = fxdr_unsigned(u_short, fp->fa_nlink);
1649 		vap->va_uid = fxdr_unsigned(uid_t, fp->fa_uid);
1650 		vap->va_gid = fxdr_unsigned(gid_t, fp->fa_gid);
1651 		vap->va_size = fxdr_hyper(&fp->fa3_size);
1652 		vap->va_bytes = fxdr_hyper(&fp->fa3_used);
1653 		vap->va_fileid = fxdr_unsigned(int32_t,
1654 		    fp->fa3_fileid.nfsuquad[1]);
1655 		fxdr_nfsv3time(&fp->fa3_atime, &vap->va_atime);
1656 		fxdr_nfsv3time(&fp->fa3_ctime, &vap->va_ctime);
1657 		vap->va_flags = 0;
1658 		vap->va_filerev = 0;
1659 	} else {
1660 		vap->va_nlink = fxdr_unsigned(u_short, fp->fa_nlink);
1661 		vap->va_uid = fxdr_unsigned(uid_t, fp->fa_uid);
1662 		vap->va_gid = fxdr_unsigned(gid_t, fp->fa_gid);
1663 		vap->va_size = fxdr_unsigned(u_int32_t, fp->fa2_size);
1664 		vap->va_bytes = fxdr_unsigned(int32_t, fp->fa2_blocks)
1665 		    * NFS_FABLKSIZE;
1666 		vap->va_fileid = fxdr_unsigned(int32_t, fp->fa2_fileid);
1667 		fxdr_nfsv2time(&fp->fa2_atime, &vap->va_atime);
1668 		vap->va_flags = 0;
1669 		vap->va_ctime.tv_sec = fxdr_unsigned(u_int32_t,
1670 		    fp->fa2_ctime.nfsv2_sec);
1671 		vap->va_ctime.tv_nsec = 0;
1672 		vap->va_gen = fxdr_unsigned(u_int32_t,fp->fa2_ctime.nfsv2_usec);
1673 		vap->va_filerev = 0;
1674 	}
1675 	if (vap->va_size != np->n_size) {
1676 		if ((np->n_flag & NMODIFIED) && vap->va_size < np->n_size) {
1677 			vap->va_size = np->n_size;
1678 		} else {
1679 			np->n_size = vap->va_size;
1680 			if (vap->va_type == VREG) {
1681 				uvm_vnp_setsize(vp, np->n_size);
1682 			}
1683 		}
1684 	}
1685 	np->n_attrstamp = time.tv_sec;
1686 	if (vaper != NULL) {
1687 		memcpy((caddr_t)vaper, (caddr_t)vap, sizeof(*vap));
1688 		if (np->n_flag & NCHG) {
1689 			if (np->n_flag & NACC)
1690 				vaper->va_atime = np->n_atim;
1691 			if (np->n_flag & NUPD)
1692 				vaper->va_mtime = np->n_mtim;
1693 		}
1694 	}
1695 	return (0);
1696 }
1697 
1698 /*
1699  * Check the time stamp
1700  * If the cache is valid, copy contents to *vap and return 0
1701  * otherwise return an error
1702  */
1703 int
1704 nfs_getattrcache(vp, vaper)
1705 	struct vnode *vp;
1706 	struct vattr *vaper;
1707 {
1708 	struct nfsnode *np = VTONFS(vp);
1709 	struct vattr *vap;
1710 
1711 	if ((time.tv_sec - np->n_attrstamp) >= NFS_ATTRTIMEO(np)) {
1712 		nfsstats.attrcache_misses++;
1713 		return (ENOENT);
1714 	}
1715 	nfsstats.attrcache_hits++;
1716 	vap = np->n_vattr;
1717 	if (vap->va_size != np->n_size) {
1718 		if (vap->va_type == VREG) {
1719 			if (np->n_flag & NMODIFIED) {
1720 				if (vap->va_size < np->n_size)
1721 					vap->va_size = np->n_size;
1722 				else
1723 					np->n_size = vap->va_size;
1724 			} else
1725 				np->n_size = vap->va_size;
1726 			uvm_vnp_setsize(vp, np->n_size);
1727 		} else
1728 			np->n_size = vap->va_size;
1729 	}
1730 	memcpy((caddr_t)vaper, (caddr_t)vap, sizeof(struct vattr));
1731 	if (np->n_flag & NCHG) {
1732 		if (np->n_flag & NACC)
1733 			vaper->va_atime = np->n_atim;
1734 		if (np->n_flag & NUPD)
1735 			vaper->va_mtime = np->n_mtim;
1736 	}
1737 	return (0);
1738 }
1739 
1740 /*
1741  * Heuristic to see if the server XDR encodes directory cookies or not.
1742  * it is not supposed to, but a lot of servers may do this. Also, since
1743  * most/all servers will implement V2 as well, it is expected that they
1744  * may return just 32 bits worth of cookie information, so we need to
1745  * find out in which 32 bits this information is available. We do this
1746  * to avoid trouble with emulated binaries that can't handle 64 bit
1747  * directory offsets.
1748  */
1749 
1750 void
1751 nfs_cookieheuristic(vp, flagp, p, cred)
1752 	struct vnode *vp;
1753 	int *flagp;
1754 	struct proc *p;
1755 	struct ucred *cred;
1756 {
1757 	struct uio auio;
1758 	struct iovec aiov;
1759 	caddr_t buf, cp;
1760 	struct dirent *dp;
1761 	off_t *cookies = NULL, *cop;
1762 	int error, eof, nc, len;
1763 
1764 	MALLOC(buf, caddr_t, NFS_DIRFRAGSIZ, M_TEMP, M_WAITOK);
1765 
1766 	aiov.iov_base = buf;
1767 	aiov.iov_len = NFS_DIRFRAGSIZ;
1768 	auio.uio_iov = &aiov;
1769 	auio.uio_iovcnt = 1;
1770 	auio.uio_rw = UIO_READ;
1771 	auio.uio_segflg = UIO_SYSSPACE;
1772 	auio.uio_procp = p;
1773 	auio.uio_resid = NFS_DIRFRAGSIZ;
1774 	auio.uio_offset = 0;
1775 
1776 	error = VOP_READDIR(vp, &auio, cred, &eof, &cookies, &nc);
1777 
1778 	len = NFS_DIRFRAGSIZ - auio.uio_resid;
1779 	if (error || len == 0) {
1780 		FREE(buf, M_TEMP);
1781 		if (cookies)
1782 			free(cookies, M_TEMP);
1783 		return;
1784 	}
1785 
1786 	/*
1787 	 * Find the first valid entry and look at its offset cookie.
1788 	 */
1789 
1790 	cp = buf;
1791 	for (cop = cookies; len > 0; len -= dp->d_reclen) {
1792 		dp = (struct dirent *)cp;
1793 		if (dp->d_fileno != 0 && len >= dp->d_reclen) {
1794 			if ((*cop >> 32) != 0 && (*cop & 0xffffffffLL) == 0) {
1795 				*flagp |= NFSMNT_SWAPCOOKIE;
1796 				nfs_invaldircache(vp, 0);
1797 				nfs_vinvalbuf(vp, 0, cred, p, 1);
1798 			}
1799 			break;
1800 		}
1801 		cop++;
1802 		cp += dp->d_reclen;
1803 	}
1804 
1805 	FREE(buf, M_TEMP);
1806 	free(cookies, M_TEMP);
1807 }
1808 #endif /* NFS */
1809 
1810 /*
1811  * Set up nameidata for a lookup() call and do it.
1812  *
1813  * If pubflag is set, this call is done for a lookup operation on the
1814  * public filehandle. In that case we allow crossing mountpoints and
1815  * absolute pathnames. However, the caller is expected to check that
1816  * the lookup result is within the public fs, and deny access if
1817  * it is not.
1818  */
1819 int
1820 nfs_namei(ndp, fhp, len, slp, nam, mdp, dposp, retdirp, p, kerbflag, pubflag)
1821 	struct nameidata *ndp;
1822 	fhandle_t *fhp;
1823 	int len;
1824 	struct nfssvc_sock *slp;
1825 	struct mbuf *nam;
1826 	struct mbuf **mdp;
1827 	caddr_t *dposp;
1828 	struct vnode **retdirp;
1829 	struct proc *p;
1830 	int kerbflag, pubflag;
1831 {
1832 	int i, rem;
1833 	struct mbuf *md;
1834 	char *fromcp, *tocp, *cp;
1835 	struct iovec aiov;
1836 	struct uio auio;
1837 	struct vnode *dp;
1838 	int error, rdonly, linklen;
1839 	struct componentname *cnp = &ndp->ni_cnd;
1840 
1841 	*retdirp = (struct vnode *)0;
1842 
1843 	if ((len + 1) > MAXPATHLEN)
1844 		return (ENAMETOOLONG);
1845 	cnp->cn_pnbuf = PNBUF_GET();
1846 
1847 	/*
1848 	 * Copy the name from the mbuf list to ndp->ni_pnbuf
1849 	 * and set the various ndp fields appropriately.
1850 	 */
1851 	fromcp = *dposp;
1852 	tocp = cnp->cn_pnbuf;
1853 	md = *mdp;
1854 	rem = mtod(md, caddr_t) + md->m_len - fromcp;
1855 	for (i = 0; i < len; i++) {
1856 		while (rem == 0) {
1857 			md = md->m_next;
1858 			if (md == NULL) {
1859 				error = EBADRPC;
1860 				goto out;
1861 			}
1862 			fromcp = mtod(md, caddr_t);
1863 			rem = md->m_len;
1864 		}
1865 		if (*fromcp == '\0' || (!pubflag && *fromcp == '/')) {
1866 			error = EACCES;
1867 			goto out;
1868 		}
1869 		*tocp++ = *fromcp++;
1870 		rem--;
1871 	}
1872 	*tocp = '\0';
1873 	*mdp = md;
1874 	*dposp = fromcp;
1875 	len = nfsm_rndup(len)-len;
1876 	if (len > 0) {
1877 		if (rem >= len)
1878 			*dposp += len;
1879 		else if ((error = nfs_adv(mdp, dposp, len, rem)) != 0)
1880 			goto out;
1881 	}
1882 
1883 	/*
1884 	 * Extract and set starting directory.
1885 	 */
1886 	error = nfsrv_fhtovp(fhp, FALSE, &dp, ndp->ni_cnd.cn_cred, slp,
1887 	    nam, &rdonly, kerbflag, pubflag);
1888 	if (error)
1889 		goto out;
1890 	if (dp->v_type != VDIR) {
1891 		vrele(dp);
1892 		error = ENOTDIR;
1893 		goto out;
1894 	}
1895 
1896 	if (rdonly)
1897 		cnp->cn_flags |= RDONLY;
1898 
1899 	*retdirp = dp;
1900 
1901 	if (pubflag) {
1902 		/*
1903 		 * Oh joy. For WebNFS, handle those pesky '%' escapes,
1904 		 * and the 'native path' indicator.
1905 		 */
1906 		cp = PNBUF_GET();
1907 		fromcp = cnp->cn_pnbuf;
1908 		tocp = cp;
1909 		if ((unsigned char)*fromcp >= WEBNFS_SPECCHAR_START) {
1910 			switch ((unsigned char)*fromcp) {
1911 			case WEBNFS_NATIVE_CHAR:
1912 				/*
1913 				 * 'Native' path for us is the same
1914 				 * as a path according to the NFS spec,
1915 				 * just skip the escape char.
1916 				 */
1917 				fromcp++;
1918 				break;
1919 			/*
1920 			 * More may be added in the future, range 0x80-0xff
1921 			 */
1922 			default:
1923 				error = EIO;
1924 				FREE(cp, M_NAMEI);
1925 				goto out;
1926 			}
1927 		}
1928 		/*
1929 		 * Translate the '%' escapes, URL-style.
1930 		 */
1931 		while (*fromcp != '\0') {
1932 			if (*fromcp == WEBNFS_ESC_CHAR) {
1933 				if (fromcp[1] != '\0' && fromcp[2] != '\0') {
1934 					fromcp++;
1935 					*tocp++ = HEXSTRTOI(fromcp);
1936 					fromcp += 2;
1937 					continue;
1938 				} else {
1939 					error = ENOENT;
1940 					FREE(cp, M_NAMEI);
1941 					goto out;
1942 				}
1943 			} else
1944 				*tocp++ = *fromcp++;
1945 		}
1946 		*tocp = '\0';
1947 		PNBUF_PUT(cnp->cn_pnbuf);
1948 		cnp->cn_pnbuf = cp;
1949 	}
1950 
1951 	ndp->ni_pathlen = (tocp - cnp->cn_pnbuf) + 1;
1952 	ndp->ni_segflg = UIO_SYSSPACE;
1953 
1954 	if (pubflag) {
1955 		ndp->ni_rootdir = rootvnode;
1956 		ndp->ni_loopcnt = 0;
1957 		if (cnp->cn_pnbuf[0] == '/')
1958 			dp = rootvnode;
1959 	} else {
1960 		cnp->cn_flags |= NOCROSSMOUNT;
1961 	}
1962 
1963 	cnp->cn_proc = p;
1964 	VREF(dp);
1965 
1966     for (;;) {
1967 	cnp->cn_nameptr = cnp->cn_pnbuf;
1968 	ndp->ni_startdir = dp;
1969 	/*
1970 	 * And call lookup() to do the real work
1971 	 */
1972 	error = lookup(ndp);
1973 	if (error) {
1974 		PNBUF_PUT(cnp->cn_pnbuf);
1975 		return (error);
1976 	}
1977 	/*
1978 	 * Check for encountering a symbolic link
1979 	 */
1980 	if ((cnp->cn_flags & ISSYMLINK) == 0) {
1981 		if (cnp->cn_flags & (SAVENAME | SAVESTART))
1982 			cnp->cn_flags |= HASBUF;
1983 		else
1984 			PNBUF_PUT(cnp->cn_pnbuf);
1985 		return (0);
1986 	} else {
1987 		if ((cnp->cn_flags & LOCKPARENT) && (cnp->cn_flags & ISLASTCN))
1988 			VOP_UNLOCK(ndp->ni_dvp, 0);
1989 		if (!pubflag) {
1990 			error = EINVAL;
1991 			break;
1992 		}
1993 
1994 		if (ndp->ni_loopcnt++ >= MAXSYMLINKS) {
1995 			error = ELOOP;
1996 			break;
1997 		}
1998 		if (ndp->ni_vp->v_mount->mnt_flag & MNT_SYMPERM) {
1999 			error = VOP_ACCESS(ndp->ni_vp, VEXEC, cnp->cn_cred,
2000 			    cnp->cn_proc);
2001 			if (error != 0)
2002 				break;
2003 		}
2004 		if (ndp->ni_pathlen > 1)
2005 			cp = PNBUF_GET();
2006 		else
2007 			cp = cnp->cn_pnbuf;
2008 		aiov.iov_base = cp;
2009 		aiov.iov_len = MAXPATHLEN;
2010 		auio.uio_iov = &aiov;
2011 		auio.uio_iovcnt = 1;
2012 		auio.uio_offset = 0;
2013 		auio.uio_rw = UIO_READ;
2014 		auio.uio_segflg = UIO_SYSSPACE;
2015 		auio.uio_procp = (struct proc *)0;
2016 		auio.uio_resid = MAXPATHLEN;
2017 		error = VOP_READLINK(ndp->ni_vp, &auio, cnp->cn_cred);
2018 		if (error) {
2019 		badlink:
2020 			if (ndp->ni_pathlen > 1)
2021 				PNBUF_PUT(cp);
2022 			break;
2023 		}
2024 		linklen = MAXPATHLEN - auio.uio_resid;
2025 		if (linklen == 0) {
2026 			error = ENOENT;
2027 			goto badlink;
2028 		}
2029 		if (linklen + ndp->ni_pathlen >= MAXPATHLEN) {
2030 			error = ENAMETOOLONG;
2031 			goto badlink;
2032 		}
2033 		if (ndp->ni_pathlen > 1) {
2034 			memcpy(cp + linklen, ndp->ni_next, ndp->ni_pathlen);
2035 			PNBUF_PUT(cnp->cn_pnbuf);
2036 			cnp->cn_pnbuf = cp;
2037 		} else
2038 			cnp->cn_pnbuf[linklen] = '\0';
2039 		ndp->ni_pathlen += linklen;
2040 		vput(ndp->ni_vp);
2041 		dp = ndp->ni_dvp;
2042 		/*
2043 		 * Check if root directory should replace current directory.
2044 		 */
2045 		if (cnp->cn_pnbuf[0] == '/') {
2046 			vrele(dp);
2047 			dp = ndp->ni_rootdir;
2048 			VREF(dp);
2049 		}
2050 	}
2051    }
2052 	vrele(ndp->ni_dvp);
2053 	vput(ndp->ni_vp);
2054 	ndp->ni_vp = NULL;
2055 out:
2056 	PNBUF_PUT(cnp->cn_pnbuf);
2057 	return (error);
2058 }
2059 
2060 /*
2061  * A fiddled version of m_adj() that ensures null fill to a long
2062  * boundary and only trims off the back end
2063  */
2064 void
2065 nfsm_adj(mp, len, nul)
2066 	struct mbuf *mp;
2067 	int len;
2068 	int nul;
2069 {
2070 	struct mbuf *m;
2071 	int count, i;
2072 	char *cp;
2073 
2074 	/*
2075 	 * Trim from tail.  Scan the mbuf chain,
2076 	 * calculating its length and finding the last mbuf.
2077 	 * If the adjustment only affects this mbuf, then just
2078 	 * adjust and return.  Otherwise, rescan and truncate
2079 	 * after the remaining size.
2080 	 */
2081 	count = 0;
2082 	m = mp;
2083 	for (;;) {
2084 		count += m->m_len;
2085 		if (m->m_next == (struct mbuf *)0)
2086 			break;
2087 		m = m->m_next;
2088 	}
2089 	if (m->m_len > len) {
2090 		m->m_len -= len;
2091 		if (nul > 0) {
2092 			cp = mtod(m, caddr_t)+m->m_len-nul;
2093 			for (i = 0; i < nul; i++)
2094 				*cp++ = '\0';
2095 		}
2096 		return;
2097 	}
2098 	count -= len;
2099 	if (count < 0)
2100 		count = 0;
2101 	/*
2102 	 * Correct length for chain is "count".
2103 	 * Find the mbuf with last data, adjust its length,
2104 	 * and toss data from remaining mbufs on chain.
2105 	 */
2106 	for (m = mp; m; m = m->m_next) {
2107 		if (m->m_len >= count) {
2108 			m->m_len = count;
2109 			if (nul > 0) {
2110 				cp = mtod(m, caddr_t)+m->m_len-nul;
2111 				for (i = 0; i < nul; i++)
2112 					*cp++ = '\0';
2113 			}
2114 			break;
2115 		}
2116 		count -= m->m_len;
2117 	}
2118 	for (m = m->m_next;m;m = m->m_next)
2119 		m->m_len = 0;
2120 }
2121 
2122 /*
2123  * Make these functions instead of macros, so that the kernel text size
2124  * doesn't get too big...
2125  */
2126 void
2127 nfsm_srvwcc(nfsd, before_ret, before_vap, after_ret, after_vap, mbp, bposp)
2128 	struct nfsrv_descript *nfsd;
2129 	int before_ret;
2130 	struct vattr *before_vap;
2131 	int after_ret;
2132 	struct vattr *after_vap;
2133 	struct mbuf **mbp;
2134 	char **bposp;
2135 {
2136 	struct mbuf *mb = *mbp, *mb2;
2137 	char *bpos = *bposp;
2138 	u_int32_t *tl;
2139 
2140 	if (before_ret) {
2141 		nfsm_build(tl, u_int32_t *, NFSX_UNSIGNED);
2142 		*tl = nfs_false;
2143 	} else {
2144 		nfsm_build(tl, u_int32_t *, 7 * NFSX_UNSIGNED);
2145 		*tl++ = nfs_true;
2146 		txdr_hyper(before_vap->va_size, tl);
2147 		tl += 2;
2148 		txdr_nfsv3time(&(before_vap->va_mtime), tl);
2149 		tl += 2;
2150 		txdr_nfsv3time(&(before_vap->va_ctime), tl);
2151 	}
2152 	*bposp = bpos;
2153 	*mbp = mb;
2154 	nfsm_srvpostopattr(nfsd, after_ret, after_vap, mbp, bposp);
2155 }
2156 
2157 void
2158 nfsm_srvpostopattr(nfsd, after_ret, after_vap, mbp, bposp)
2159 	struct nfsrv_descript *nfsd;
2160 	int after_ret;
2161 	struct vattr *after_vap;
2162 	struct mbuf **mbp;
2163 	char **bposp;
2164 {
2165 	struct mbuf *mb = *mbp, *mb2;
2166 	char *bpos = *bposp;
2167 	u_int32_t *tl;
2168 	struct nfs_fattr *fp;
2169 
2170 	if (after_ret) {
2171 		nfsm_build(tl, u_int32_t *, NFSX_UNSIGNED);
2172 		*tl = nfs_false;
2173 	} else {
2174 		nfsm_build(tl, u_int32_t *, NFSX_UNSIGNED + NFSX_V3FATTR);
2175 		*tl++ = nfs_true;
2176 		fp = (struct nfs_fattr *)tl;
2177 		nfsm_srvfattr(nfsd, after_vap, fp);
2178 	}
2179 	*mbp = mb;
2180 	*bposp = bpos;
2181 }
2182 
2183 void
2184 nfsm_srvfattr(nfsd, vap, fp)
2185 	struct nfsrv_descript *nfsd;
2186 	struct vattr *vap;
2187 	struct nfs_fattr *fp;
2188 {
2189 
2190 	fp->fa_nlink = txdr_unsigned(vap->va_nlink);
2191 	fp->fa_uid = txdr_unsigned(vap->va_uid);
2192 	fp->fa_gid = txdr_unsigned(vap->va_gid);
2193 	if (nfsd->nd_flag & ND_NFSV3) {
2194 		fp->fa_type = vtonfsv3_type(vap->va_type);
2195 		fp->fa_mode = vtonfsv3_mode(vap->va_mode);
2196 		txdr_hyper(vap->va_size, &fp->fa3_size);
2197 		txdr_hyper(vap->va_bytes, &fp->fa3_used);
2198 		fp->fa3_rdev.specdata1 = txdr_unsigned(major(vap->va_rdev));
2199 		fp->fa3_rdev.specdata2 = txdr_unsigned(minor(vap->va_rdev));
2200 		fp->fa3_fsid.nfsuquad[0] = 0;
2201 		fp->fa3_fsid.nfsuquad[1] = txdr_unsigned(vap->va_fsid);
2202 		fp->fa3_fileid.nfsuquad[0] = 0;
2203 		fp->fa3_fileid.nfsuquad[1] = txdr_unsigned(vap->va_fileid);
2204 		txdr_nfsv3time(&vap->va_atime, &fp->fa3_atime);
2205 		txdr_nfsv3time(&vap->va_mtime, &fp->fa3_mtime);
2206 		txdr_nfsv3time(&vap->va_ctime, &fp->fa3_ctime);
2207 	} else {
2208 		fp->fa_type = vtonfsv2_type(vap->va_type);
2209 		fp->fa_mode = vtonfsv2_mode(vap->va_type, vap->va_mode);
2210 		fp->fa2_size = txdr_unsigned(vap->va_size);
2211 		fp->fa2_blocksize = txdr_unsigned(vap->va_blocksize);
2212 		if (vap->va_type == VFIFO)
2213 			fp->fa2_rdev = 0xffffffff;
2214 		else
2215 			fp->fa2_rdev = txdr_unsigned(vap->va_rdev);
2216 		fp->fa2_blocks = txdr_unsigned(vap->va_bytes / NFS_FABLKSIZE);
2217 		fp->fa2_fsid = txdr_unsigned(vap->va_fsid);
2218 		fp->fa2_fileid = txdr_unsigned(vap->va_fileid);
2219 		txdr_nfsv2time(&vap->va_atime, &fp->fa2_atime);
2220 		txdr_nfsv2time(&vap->va_mtime, &fp->fa2_mtime);
2221 		txdr_nfsv2time(&vap->va_ctime, &fp->fa2_ctime);
2222 	}
2223 }
2224 
2225 /*
2226  * nfsrv_fhtovp() - convert a fh to a vnode ptr (optionally locked)
2227  * 	- look up fsid in mount list (if not found ret error)
2228  *	- get vp and export rights by calling VFS_FHTOVP()
2229  *	- if cred->cr_uid == 0 or MNT_EXPORTANON set it to credanon
2230  *	- if not lockflag unlock it with VOP_UNLOCK()
2231  */
2232 int
2233 nfsrv_fhtovp(fhp, lockflag, vpp, cred, slp, nam, rdonlyp, kerbflag, pubflag)
2234 	fhandle_t *fhp;
2235 	int lockflag;
2236 	struct vnode **vpp;
2237 	struct ucred *cred;
2238 	struct nfssvc_sock *slp;
2239 	struct mbuf *nam;
2240 	int *rdonlyp;
2241 	int kerbflag;
2242 {
2243 	struct mount *mp;
2244 	int i;
2245 	struct ucred *credanon;
2246 	int error, exflags;
2247 	struct sockaddr_in *saddr;
2248 
2249 	*vpp = (struct vnode *)0;
2250 
2251 	if (nfs_ispublicfh(fhp)) {
2252 		if (!pubflag || !nfs_pub.np_valid)
2253 			return (ESTALE);
2254 		fhp = &nfs_pub.np_handle;
2255 	}
2256 
2257 	mp = vfs_getvfs(&fhp->fh_fsid);
2258 	if (!mp)
2259 		return (ESTALE);
2260 	error = VFS_CHECKEXP(mp, nam, &exflags, &credanon);
2261 	if (error)
2262 		return (error);
2263 	error = VFS_FHTOVP(mp, &fhp->fh_fid, vpp);
2264 	if (error)
2265 		return (error);
2266 
2267 	if (!(exflags & (MNT_EXNORESPORT|MNT_EXPUBLIC))) {
2268 		saddr = mtod(nam, struct sockaddr_in *);
2269 		if ((saddr->sin_family == AF_INET) &&
2270 		    ntohs(saddr->sin_port) >= IPPORT_RESERVED) {
2271 			vput(*vpp);
2272 			return (NFSERR_AUTHERR | AUTH_TOOWEAK);
2273 		}
2274 #ifdef INET6
2275 		if ((saddr->sin_family == AF_INET6) &&
2276 		    ntohs(saddr->sin_port) >= IPV6PORT_RESERVED) {
2277 			vput(*vpp);
2278 			return (NFSERR_AUTHERR | AUTH_TOOWEAK);
2279 		}
2280 #endif
2281 	}
2282 	/*
2283 	 * Check/setup credentials.
2284 	 */
2285 	if (exflags & MNT_EXKERB) {
2286 		if (!kerbflag) {
2287 			vput(*vpp);
2288 			return (NFSERR_AUTHERR | AUTH_TOOWEAK);
2289 		}
2290 	} else if (kerbflag) {
2291 		vput(*vpp);
2292 		return (NFSERR_AUTHERR | AUTH_TOOWEAK);
2293 	} else if (cred->cr_uid == 0 || (exflags & MNT_EXPORTANON)) {
2294 		cred->cr_uid = credanon->cr_uid;
2295 		cred->cr_gid = credanon->cr_gid;
2296 		for (i = 0; i < credanon->cr_ngroups && i < NGROUPS; i++)
2297 			cred->cr_groups[i] = credanon->cr_groups[i];
2298 		cred->cr_ngroups = i;
2299 	}
2300 	if (exflags & MNT_EXRDONLY)
2301 		*rdonlyp = 1;
2302 	else
2303 		*rdonlyp = 0;
2304 	if (!lockflag)
2305 		VOP_UNLOCK(*vpp, 0);
2306 	return (0);
2307 }
2308 
2309 /*
2310  * WebNFS: check if a filehandle is a public filehandle. For v3, this
2311  * means a length of 0, for v2 it means all zeroes. nfsm_srvmtofh has
2312  * transformed this to all zeroes in both cases, so check for it.
2313  */
2314 int
2315 nfs_ispublicfh(fhp)
2316 	fhandle_t *fhp;
2317 {
2318 	char *cp = (char *)fhp;
2319 	int i;
2320 
2321 	for (i = 0; i < NFSX_V3FH; i++)
2322 		if (*cp++ != 0)
2323 			return (FALSE);
2324 	return (TRUE);
2325 }
2326 
2327 /*
2328  * This function compares two net addresses by family and returns TRUE
2329  * if they are the same host.
2330  * If there is any doubt, return FALSE.
2331  * The AF_INET family is handled as a special case so that address mbufs
2332  * don't need to be saved to store "struct in_addr", which is only 4 bytes.
2333  */
2334 int
2335 netaddr_match(family, haddr, nam)
2336 	int family;
2337 	union nethostaddr *haddr;
2338 	struct mbuf *nam;
2339 {
2340 	struct sockaddr_in *inetaddr;
2341 
2342 	switch (family) {
2343 	case AF_INET:
2344 		inetaddr = mtod(nam, struct sockaddr_in *);
2345 		if (inetaddr->sin_family == AF_INET &&
2346 		    inetaddr->sin_addr.s_addr == haddr->had_inetaddr)
2347 			return (1);
2348 		break;
2349 #ifdef INET6
2350 	case AF_INET6:
2351 	    {
2352 		struct sockaddr_in6 *sin6_1, *sin6_2;
2353 
2354 		sin6_1 = mtod(nam, struct sockaddr_in6 *);
2355 		sin6_2 = mtod(haddr->had_nam, struct sockaddr_in6 *);
2356 		if (sin6_1->sin6_family == AF_INET6 &&
2357 		    IN6_ARE_ADDR_EQUAL(&sin6_1->sin6_addr, &sin6_2->sin6_addr))
2358 			return 1;
2359 	    }
2360 #endif
2361 #ifdef ISO
2362 	case AF_ISO:
2363 	    {
2364 		struct sockaddr_iso *isoaddr1, *isoaddr2;
2365 
2366 		isoaddr1 = mtod(nam, struct sockaddr_iso *);
2367 		isoaddr2 = mtod(haddr->had_nam, struct sockaddr_iso *);
2368 		if (isoaddr1->siso_family == AF_ISO &&
2369 		    isoaddr1->siso_nlen > 0 &&
2370 		    isoaddr1->siso_nlen == isoaddr2->siso_nlen &&
2371 		    SAME_ISOADDR(isoaddr1, isoaddr2))
2372 			return (1);
2373 		break;
2374 	    }
2375 #endif	/* ISO */
2376 	default:
2377 		break;
2378 	};
2379 	return (0);
2380 }
2381 
2382 /*
2383  * The write verifier has changed (probably due to a server reboot), so all
2384  * B_NEEDCOMMIT blocks will have to be written again. Since they are on the
2385  * dirty block list as B_DELWRI, all this takes is clearing the B_NEEDCOMMIT
2386  * flag. Once done the new write verifier can be set for the mount point.
2387  */
2388 void
2389 nfs_clearcommit(mp)
2390 	struct mount *mp;
2391 {
2392 	struct vnode *vp;
2393 	struct nfsnode *np;
2394 	struct vm_page *pg;
2395 	int s;
2396 
2397 	s = splbio();
2398 	LIST_FOREACH(vp, &mp->mnt_vnodelist, v_mntvnodes) {
2399 		KASSERT(vp->v_mount == mp);
2400 		if (vp->v_type == VNON)
2401 			continue;
2402 		np = VTONFS(vp);
2403 		np->n_pushlo = np->n_pushhi = np->n_pushedlo =
2404 		    np->n_pushedhi = 0;
2405 		np->n_commitflags &=
2406 		    ~(NFS_COMMIT_PUSH_VALID | NFS_COMMIT_PUSHED_VALID);
2407 		simple_lock(&vp->v_uvm.u_obj.vmobjlock);
2408 		TAILQ_FOREACH(pg, &vp->v_uvm.u_obj.memq, listq) {
2409 			pg->flags &= ~PG_NEEDCOMMIT;
2410 		}
2411 		simple_unlock(&vp->v_uvm.u_obj.vmobjlock);
2412 	}
2413 	splx(s);
2414 }
2415 
2416 void
2417 nfs_merge_commit_ranges(vp)
2418 	struct vnode *vp;
2419 {
2420 	struct nfsnode *np = VTONFS(vp);
2421 
2422 	if (!(np->n_commitflags & NFS_COMMIT_PUSHED_VALID)) {
2423 		np->n_pushedlo = np->n_pushlo;
2424 		np->n_pushedhi = np->n_pushhi;
2425 		np->n_commitflags |= NFS_COMMIT_PUSHED_VALID;
2426 	} else {
2427 		if (np->n_pushlo < np->n_pushedlo)
2428 			np->n_pushedlo = np->n_pushlo;
2429 		if (np->n_pushhi > np->n_pushedhi)
2430 			np->n_pushedhi = np->n_pushhi;
2431 	}
2432 
2433 	np->n_pushlo = np->n_pushhi = 0;
2434 	np->n_commitflags &= ~NFS_COMMIT_PUSH_VALID;
2435 
2436 #ifdef fvdl_debug
2437 	printf("merge: committed: %u - %u\n", (unsigned)np->n_pushedlo,
2438 	    (unsigned)np->n_pushedhi);
2439 #endif
2440 }
2441 
2442 int
2443 nfs_in_committed_range(vp, off, len)
2444 	struct vnode *vp;
2445 	off_t off, len;
2446 {
2447 	struct nfsnode *np = VTONFS(vp);
2448 	off_t lo, hi;
2449 
2450 	if (!(np->n_commitflags & NFS_COMMIT_PUSHED_VALID))
2451 		return 0;
2452 	lo = off;
2453 	hi = lo + len;
2454 
2455 	return (lo >= np->n_pushedlo && hi <= np->n_pushedhi);
2456 }
2457 
2458 int
2459 nfs_in_tobecommitted_range(vp, off, len)
2460 	struct vnode *vp;
2461 	off_t off, len;
2462 {
2463 	struct nfsnode *np = VTONFS(vp);
2464 	off_t lo, hi;
2465 
2466 	if (!(np->n_commitflags & NFS_COMMIT_PUSH_VALID))
2467 		return 0;
2468 	lo = off;
2469 	hi = lo + len;
2470 
2471 	return (lo >= np->n_pushlo && hi <= np->n_pushhi);
2472 }
2473 
2474 void
2475 nfs_add_committed_range(vp, off, len)
2476 	struct vnode *vp;
2477 	off_t off, len;
2478 {
2479 	struct nfsnode *np = VTONFS(vp);
2480 	off_t lo, hi;
2481 
2482 	lo = off;
2483 	hi = lo + len;
2484 
2485 	if (!(np->n_commitflags & NFS_COMMIT_PUSHED_VALID)) {
2486 		np->n_pushedlo = lo;
2487 		np->n_pushedhi = hi;
2488 		np->n_commitflags |= NFS_COMMIT_PUSHED_VALID;
2489 	} else {
2490 		if (hi > np->n_pushedhi)
2491 			np->n_pushedhi = hi;
2492 		if (lo < np->n_pushedlo)
2493 			np->n_pushedlo = lo;
2494 	}
2495 #ifdef fvdl_debug
2496 	printf("add: committed: %u - %u\n", (unsigned)np->n_pushedlo,
2497 	    (unsigned)np->n_pushedhi);
2498 #endif
2499 }
2500 
2501 void
2502 nfs_del_committed_range(vp, off, len)
2503 	struct vnode *vp;
2504 	off_t off, len;
2505 {
2506 	struct nfsnode *np = VTONFS(vp);
2507 	off_t lo, hi;
2508 
2509 	if (!(np->n_commitflags & NFS_COMMIT_PUSHED_VALID))
2510 		return;
2511 
2512 	lo = off;
2513 	hi = lo + len;
2514 
2515 	if (lo > np->n_pushedhi || hi < np->n_pushedlo)
2516 		return;
2517 	if (lo <= np->n_pushedlo)
2518 		np->n_pushedlo = hi;
2519 	else if (hi >= np->n_pushedhi)
2520 		np->n_pushedhi = lo;
2521 	else {
2522 		/*
2523 		 * XXX There's only one range. If the deleted range
2524 		 * is in the middle, pick the largest of the
2525 		 * contiguous ranges that it leaves.
2526 		 */
2527 		if ((np->n_pushedlo - lo) > (hi - np->n_pushedhi))
2528 			np->n_pushedhi = lo;
2529 		else
2530 			np->n_pushedlo = hi;
2531 	}
2532 #ifdef fvdl_debug
2533 	printf("del: committed: %u - %u\n", (unsigned)np->n_pushedlo,
2534 	    (unsigned)np->n_pushedhi);
2535 #endif
2536 }
2537 
2538 void
2539 nfs_add_tobecommitted_range(vp, off, len)
2540 	struct vnode *vp;
2541 	off_t off, len;
2542 {
2543 	struct nfsnode *np = VTONFS(vp);
2544 	off_t lo, hi;
2545 
2546 	lo = off;
2547 	hi = lo + len;
2548 
2549 	if (!(np->n_commitflags & NFS_COMMIT_PUSH_VALID)) {
2550 		np->n_pushlo = lo;
2551 		np->n_pushhi = hi;
2552 		np->n_commitflags |= NFS_COMMIT_PUSH_VALID;
2553 	} else {
2554 		if (lo < np->n_pushlo)
2555 			np->n_pushlo = lo;
2556 		if (hi > np->n_pushhi)
2557 			np->n_pushhi = hi;
2558 	}
2559 #ifdef fvdl_debug
2560 	printf("add: tobecommitted: %u - %u\n", (unsigned)np->n_pushlo,
2561 	    (unsigned)np->n_pushhi);
2562 #endif
2563 }
2564 
2565 void
2566 nfs_del_tobecommitted_range(vp, off, len)
2567 	struct vnode *vp;
2568 	off_t off, len;
2569 {
2570 	struct nfsnode *np = VTONFS(vp);
2571 	off_t lo, hi;
2572 
2573 	if (!(np->n_commitflags & NFS_COMMIT_PUSH_VALID))
2574 		return;
2575 
2576 	lo = off;
2577 	hi = lo + len;
2578 
2579 	if (lo > np->n_pushhi || hi < np->n_pushlo)
2580 		return;
2581 
2582 	if (lo <= np->n_pushlo)
2583 		np->n_pushlo = hi;
2584 	else if (hi >= np->n_pushhi)
2585 		np->n_pushhi = lo;
2586 	else {
2587 		/*
2588 		 * XXX There's only one range. If the deleted range
2589 		 * is in the middle, pick the largest of the
2590 		 * contiguous ranges that it leaves.
2591 		 */
2592 		if ((np->n_pushlo - lo) > (hi - np->n_pushhi))
2593 			np->n_pushhi = lo;
2594 		else
2595 			np->n_pushlo = hi;
2596 	}
2597 #ifdef fvdl_debug
2598 	printf("del: tobecommitted: %u - %u\n", (unsigned)np->n_pushlo,
2599 	    (unsigned)np->n_pushhi);
2600 #endif
2601 }
2602 
2603 /*
2604  * Map errnos to NFS error numbers. For Version 3 also filter out error
2605  * numbers not specified for the associated procedure.
2606  */
2607 int
2608 nfsrv_errmap(nd, err)
2609 	struct nfsrv_descript *nd;
2610 	int err;
2611 {
2612 	const short *defaulterrp, *errp;
2613 
2614 	if (nd->nd_flag & ND_NFSV3) {
2615 	    if (nd->nd_procnum <= NFSPROC_COMMIT) {
2616 		errp = defaulterrp = nfsrv_v3errmap[nd->nd_procnum];
2617 		while (*++errp) {
2618 			if (*errp == err)
2619 				return (err);
2620 			else if (*errp > err)
2621 				break;
2622 		}
2623 		return ((int)*defaulterrp);
2624 	    } else
2625 		return (err & 0xffff);
2626 	}
2627 	if (err <= ELAST)
2628 		return ((int)nfsrv_v2errmap[err - 1]);
2629 	return (NFSERR_IO);
2630 }
2631 
2632 /*
2633  * Sort the group list in increasing numerical order.
2634  * (Insertion sort by Chris Torek, who was grossed out by the bubble sort
2635  *  that used to be here.)
2636  */
2637 void
2638 nfsrvw_sort(list, num)
2639         gid_t *list;
2640         int num;
2641 {
2642 	int i, j;
2643 	gid_t v;
2644 
2645 	/* Insertion sort. */
2646 	for (i = 1; i < num; i++) {
2647 		v = list[i];
2648 		/* find correct slot for value v, moving others up */
2649 		for (j = i; --j >= 0 && v < list[j];)
2650 			list[j + 1] = list[j];
2651 		list[j + 1] = v;
2652 	}
2653 }
2654 
2655 /*
2656  * copy credentials making sure that the result can be compared with memcmp().
2657  */
2658 void
2659 nfsrv_setcred(incred, outcred)
2660 	struct ucred *incred, *outcred;
2661 {
2662 	int i;
2663 
2664 	memset((caddr_t)outcred, 0, sizeof (struct ucred));
2665 	outcred->cr_ref = 1;
2666 	outcred->cr_uid = incred->cr_uid;
2667 	outcred->cr_gid = incred->cr_gid;
2668 	outcred->cr_ngroups = incred->cr_ngroups;
2669 	for (i = 0; i < incred->cr_ngroups; i++)
2670 		outcred->cr_groups[i] = incred->cr_groups[i];
2671 	nfsrvw_sort(outcred->cr_groups, outcred->cr_ngroups);
2672 }
2673