xref: /freebsd/sys/rpc/clnt_vc.c (revision 9a14aa017b21c292740c00ee098195cd46642730)
1 /*	$NetBSD: clnt_vc.c,v 1.4 2000/07/14 08:40:42 fvdl Exp $	*/
2 
3 /*
4  * Sun RPC is a product of Sun Microsystems, Inc. and is provided for
5  * unrestricted use provided that this legend is included on all tape
6  * media and as a part of the software program in whole or part.  Users
7  * may copy or modify Sun RPC without charge, but are not authorized
8  * to license or distribute it to anyone else except as part of a product or
9  * program developed by the user.
10  *
11  * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE
12  * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR
13  * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE.
14  *
15  * Sun RPC is provided with no support and without any obligation on the
16  * part of Sun Microsystems, Inc. to assist in its use, correction,
17  * modification or enhancement.
18  *
19  * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE
20  * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC
21  * OR ANY PART THEREOF.
22  *
23  * In no event will Sun Microsystems, Inc. be liable for any lost revenue
24  * or profits or other special, indirect and consequential damages, even if
25  * Sun has been advised of the possibility of such damages.
26  *
27  * Sun Microsystems, Inc.
28  * 2550 Garcia Avenue
29  * Mountain View, California  94043
30  */
31 
32 #if defined(LIBC_SCCS) && !defined(lint)
33 static char *sccsid2 = "@(#)clnt_tcp.c 1.37 87/10/05 Copyr 1984 Sun Micro";
34 static char *sccsid = "@(#)clnt_tcp.c	2.2 88/08/01 4.0 RPCSRC";
35 static char sccsid3[] = "@(#)clnt_vc.c 1.19 89/03/16 Copyr 1988 Sun Micro";
36 #endif
37 #include <sys/cdefs.h>
38 __FBSDID("$FreeBSD$");
39 
40 /*
41  * clnt_tcp.c, Implements a TCP/IP based, client side RPC.
42  *
43  * Copyright (C) 1984, Sun Microsystems, Inc.
44  *
45  * TCP based RPC supports 'batched calls'.
46  * A sequence of calls may be batched-up in a send buffer.  The rpc call
47  * return immediately to the client even though the call was not necessarily
48  * sent.  The batching occurs if the results' xdr routine is NULL (0) AND
49  * the rpc timeout value is zero (see clnt.h, rpc).
50  *
51  * Clients should NOT casually batch calls that in fact return results; that is,
52  * the server side should be aware that a call is batched and not produce any
53  * return message.  Batched calls that produce many result messages can
54  * deadlock (netlock) the client and the server....
55  *
56  * Now go hang yourself.
57  */
58 
59 #include <sys/param.h>
60 #include <sys/systm.h>
61 #include <sys/lock.h>
62 #include <sys/malloc.h>
63 #include <sys/mbuf.h>
64 #include <sys/mutex.h>
65 #include <sys/pcpu.h>
66 #include <sys/proc.h>
67 #include <sys/protosw.h>
68 #include <sys/socket.h>
69 #include <sys/socketvar.h>
70 #include <sys/syslog.h>
71 #include <sys/time.h>
72 #include <sys/uio.h>
73 
74 #include <net/vnet.h>
75 
76 #include <netinet/tcp.h>
77 
78 #include <rpc/rpc.h>
79 #include <rpc/rpc_com.h>
80 
81 #define MCALL_MSG_SIZE 24
82 
83 struct cmessage {
84         struct cmsghdr cmsg;
85         struct cmsgcred cmcred;
86 };
87 
88 static enum clnt_stat clnt_vc_call(CLIENT *, struct rpc_callextra *,
89     rpcproc_t, struct mbuf *, struct mbuf **, struct timeval);
90 static void clnt_vc_geterr(CLIENT *, struct rpc_err *);
91 static bool_t clnt_vc_freeres(CLIENT *, xdrproc_t, void *);
92 static void clnt_vc_abort(CLIENT *);
93 static bool_t clnt_vc_control(CLIENT *, u_int, void *);
94 static void clnt_vc_close(CLIENT *);
95 static void clnt_vc_destroy(CLIENT *);
96 static bool_t time_not_ok(struct timeval *);
97 static int clnt_vc_soupcall(struct socket *so, void *arg, int waitflag);
98 
99 static struct clnt_ops clnt_vc_ops = {
100 	.cl_call =	clnt_vc_call,
101 	.cl_abort =	clnt_vc_abort,
102 	.cl_geterr =	clnt_vc_geterr,
103 	.cl_freeres =	clnt_vc_freeres,
104 	.cl_close =	clnt_vc_close,
105 	.cl_destroy =	clnt_vc_destroy,
106 	.cl_control =	clnt_vc_control
107 };
108 
109 /*
110  * A pending RPC request which awaits a reply. Requests which have
111  * received their reply will have cr_xid set to zero and cr_mrep to
112  * the mbuf chain of the reply.
113  */
114 struct ct_request {
115 	TAILQ_ENTRY(ct_request) cr_link;
116 	uint32_t		cr_xid;		/* XID of request */
117 	struct mbuf		*cr_mrep;	/* reply received by upcall */
118 	int			cr_error;	/* any error from upcall */
119 	char			cr_verf[MAX_AUTH_BYTES]; /* reply verf */
120 };
121 
122 TAILQ_HEAD(ct_request_list, ct_request);
123 
124 struct ct_data {
125 	struct mtx	ct_lock;
126 	int		ct_threads;	/* number of threads in clnt_vc_call */
127 	bool_t		ct_closing;	/* TRUE if we are closing */
128 	bool_t		ct_closed;	/* TRUE if we are closed */
129 	struct socket	*ct_socket;	/* connection socket */
130 	bool_t		ct_closeit;	/* close it on destroy */
131 	struct timeval	ct_wait;	/* wait interval in milliseconds */
132 	struct sockaddr_storage	ct_addr; /* remote addr */
133 	struct rpc_err	ct_error;
134 	uint32_t	ct_xid;
135 	char		ct_mcallc[MCALL_MSG_SIZE]; /* marshalled callmsg */
136 	size_t		ct_mpos;	/* pos after marshal */
137 	const char	*ct_waitchan;
138 	int		ct_waitflag;
139 	struct mbuf	*ct_record;	/* current reply record */
140 	size_t		ct_record_resid; /* how much left of reply to read */
141 	bool_t		ct_record_eor;	 /* true if reading last fragment */
142 	struct ct_request_list ct_pending;
143 	int		ct_upcallrefs;	/* Ref cnt of upcalls in prog. */
144 };
145 
146 static void clnt_vc_upcallsdone(struct ct_data *);
147 
148 static const char clnt_vc_errstr[] = "%s : %s";
149 static const char clnt_vc_str[] = "clnt_vc_create";
150 static const char clnt_read_vc_str[] = "read_vc";
151 static const char __no_mem_str[] = "out of memory";
152 
153 /*
154  * Create a client handle for a connection.
155  * Default options are set, which the user can change using clnt_control()'s.
156  * The rpc/vc package does buffering similar to stdio, so the client
157  * must pick send and receive buffer sizes, 0 => use the default.
158  * NB: fd is copied into a private area.
159  * NB: The rpch->cl_auth is set null authentication. Caller may wish to
160  * set this something more useful.
161  *
162  * fd should be an open socket
163  */
164 CLIENT *
165 clnt_vc_create(
166 	struct socket *so,		/* open file descriptor */
167 	struct sockaddr *raddr,		/* servers address */
168 	const rpcprog_t prog,		/* program number */
169 	const rpcvers_t vers,		/* version number */
170 	size_t sendsz,			/* buffer recv size */
171 	size_t recvsz,			/* buffer send size */
172 	int intrflag)			/* interruptible */
173 {
174 	CLIENT *cl;			/* client handle */
175 	struct ct_data *ct = NULL;	/* client handle */
176 	struct timeval now;
177 	struct rpc_msg call_msg;
178 	static uint32_t disrupt;
179 	struct __rpc_sockinfo si;
180 	XDR xdrs;
181 	int error, interrupted, one = 1, sleep_flag;
182 	struct sockopt sopt;
183 
184 	if (disrupt == 0)
185 		disrupt = (uint32_t)(long)raddr;
186 
187 	cl = (CLIENT *)mem_alloc(sizeof (*cl));
188 	ct = (struct ct_data *)mem_alloc(sizeof (*ct));
189 
190 	mtx_init(&ct->ct_lock, "ct->ct_lock", NULL, MTX_DEF);
191 	ct->ct_threads = 0;
192 	ct->ct_closing = FALSE;
193 	ct->ct_closed = FALSE;
194 	ct->ct_upcallrefs = 0;
195 
196 	if ((so->so_state & (SS_ISCONNECTED|SS_ISCONFIRMING)) == 0) {
197 		error = soconnect(so, raddr, curthread);
198 		SOCK_LOCK(so);
199 		interrupted = 0;
200 		sleep_flag = PSOCK;
201 		if (intrflag != 0)
202 			sleep_flag |= (PCATCH | PBDRY);
203 		while ((so->so_state & SS_ISCONNECTING)
204 		    && so->so_error == 0) {
205 			error = msleep(&so->so_timeo, SOCK_MTX(so),
206 			    sleep_flag, "connec", 0);
207 			if (error) {
208 				if (error == EINTR || error == ERESTART)
209 					interrupted = 1;
210 				break;
211 			}
212 		}
213 		if (error == 0) {
214 			error = so->so_error;
215 			so->so_error = 0;
216 		}
217 		SOCK_UNLOCK(so);
218 		if (error) {
219 			if (!interrupted)
220 				so->so_state &= ~SS_ISCONNECTING;
221 			rpc_createerr.cf_stat = RPC_SYSTEMERROR;
222 			rpc_createerr.cf_error.re_errno = error;
223 			goto err;
224 		}
225 	}
226 
227 	if (!__rpc_socket2sockinfo(so, &si)) {
228 		goto err;
229 	}
230 
231 	if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
232 		bzero(&sopt, sizeof(sopt));
233 		sopt.sopt_dir = SOPT_SET;
234 		sopt.sopt_level = SOL_SOCKET;
235 		sopt.sopt_name = SO_KEEPALIVE;
236 		sopt.sopt_val = &one;
237 		sopt.sopt_valsize = sizeof(one);
238 		sosetopt(so, &sopt);
239 	}
240 
241 	if (so->so_proto->pr_protocol == IPPROTO_TCP) {
242 		bzero(&sopt, sizeof(sopt));
243 		sopt.sopt_dir = SOPT_SET;
244 		sopt.sopt_level = IPPROTO_TCP;
245 		sopt.sopt_name = TCP_NODELAY;
246 		sopt.sopt_val = &one;
247 		sopt.sopt_valsize = sizeof(one);
248 		sosetopt(so, &sopt);
249 	}
250 
251 	ct->ct_closeit = FALSE;
252 
253 	/*
254 	 * Set up private data struct
255 	 */
256 	ct->ct_socket = so;
257 	ct->ct_wait.tv_sec = -1;
258 	ct->ct_wait.tv_usec = -1;
259 	memcpy(&ct->ct_addr, raddr, raddr->sa_len);
260 
261 	/*
262 	 * Initialize call message
263 	 */
264 	getmicrotime(&now);
265 	ct->ct_xid = ((uint32_t)++disrupt) ^ __RPC_GETXID(&now);
266 	call_msg.rm_xid = ct->ct_xid;
267 	call_msg.rm_direction = CALL;
268 	call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION;
269 	call_msg.rm_call.cb_prog = (uint32_t)prog;
270 	call_msg.rm_call.cb_vers = (uint32_t)vers;
271 
272 	/*
273 	 * pre-serialize the static part of the call msg and stash it away
274 	 */
275 	xdrmem_create(&xdrs, ct->ct_mcallc, MCALL_MSG_SIZE,
276 	    XDR_ENCODE);
277 	if (! xdr_callhdr(&xdrs, &call_msg)) {
278 		if (ct->ct_closeit) {
279 			soclose(ct->ct_socket);
280 		}
281 		goto err;
282 	}
283 	ct->ct_mpos = XDR_GETPOS(&xdrs);
284 	XDR_DESTROY(&xdrs);
285 	ct->ct_waitchan = "rpcrecv";
286 	ct->ct_waitflag = 0;
287 
288 	/*
289 	 * Create a client handle which uses xdrrec for serialization
290 	 * and authnone for authentication.
291 	 */
292 	sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
293 	recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
294 	error = soreserve(ct->ct_socket, sendsz, recvsz);
295 	if (error != 0) {
296 		if (ct->ct_closeit) {
297 			soclose(ct->ct_socket);
298 		}
299 		goto err;
300 	}
301 	cl->cl_refs = 1;
302 	cl->cl_ops = &clnt_vc_ops;
303 	cl->cl_private = ct;
304 	cl->cl_auth = authnone_create();
305 
306 	SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
307 	soupcall_set(ct->ct_socket, SO_RCV, clnt_vc_soupcall, ct);
308 	SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
309 
310 	ct->ct_record = NULL;
311 	ct->ct_record_resid = 0;
312 	TAILQ_INIT(&ct->ct_pending);
313 	return (cl);
314 
315 err:
316 	if (cl) {
317 		if (ct) {
318 			mtx_destroy(&ct->ct_lock);
319 			mem_free(ct, sizeof (struct ct_data));
320 		}
321 		if (cl)
322 			mem_free(cl, sizeof (CLIENT));
323 	}
324 	return ((CLIENT *)NULL);
325 }
326 
327 static enum clnt_stat
328 clnt_vc_call(
329 	CLIENT		*cl,		/* client handle */
330 	struct rpc_callextra *ext,	/* call metadata */
331 	rpcproc_t	proc,		/* procedure number */
332 	struct mbuf	*args,		/* pointer to args */
333 	struct mbuf	**resultsp,	/* pointer to results */
334 	struct timeval	utimeout)
335 {
336 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
337 	AUTH *auth;
338 	struct rpc_err *errp;
339 	enum clnt_stat stat;
340 	XDR xdrs;
341 	struct rpc_msg reply_msg;
342 	bool_t ok;
343 	int nrefreshes = 2;		/* number of times to refresh cred */
344 	struct timeval timeout;
345 	uint32_t xid;
346 	struct mbuf *mreq = NULL, *results;
347 	struct ct_request *cr;
348 	int error;
349 
350 	cr = malloc(sizeof(struct ct_request), M_RPC, M_WAITOK);
351 
352 	mtx_lock(&ct->ct_lock);
353 
354 	if (ct->ct_closing || ct->ct_closed) {
355 		mtx_unlock(&ct->ct_lock);
356 		free(cr, M_RPC);
357 		return (RPC_CANTSEND);
358 	}
359 	ct->ct_threads++;
360 
361 	if (ext) {
362 		auth = ext->rc_auth;
363 		errp = &ext->rc_err;
364 	} else {
365 		auth = cl->cl_auth;
366 		errp = &ct->ct_error;
367 	}
368 
369 	cr->cr_mrep = NULL;
370 	cr->cr_error = 0;
371 
372 	if (ct->ct_wait.tv_usec == -1) {
373 		timeout = utimeout;	/* use supplied timeout */
374 	} else {
375 		timeout = ct->ct_wait;	/* use default timeout */
376 	}
377 
378 call_again:
379 	mtx_assert(&ct->ct_lock, MA_OWNED);
380 
381 	ct->ct_xid++;
382 	xid = ct->ct_xid;
383 
384 	mtx_unlock(&ct->ct_lock);
385 
386 	/*
387 	 * Leave space to pre-pend the record mark.
388 	 */
389 	MGETHDR(mreq, M_WAIT, MT_DATA);
390 	mreq->m_data += sizeof(uint32_t);
391 	KASSERT(ct->ct_mpos + sizeof(uint32_t) <= MHLEN,
392 	    ("RPC header too big"));
393 	bcopy(ct->ct_mcallc, mreq->m_data, ct->ct_mpos);
394 	mreq->m_len = ct->ct_mpos;
395 
396 	/*
397 	 * The XID is the first thing in the request.
398 	 */
399 	*mtod(mreq, uint32_t *) = htonl(xid);
400 
401 	xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
402 
403 	errp->re_status = stat = RPC_SUCCESS;
404 
405 	if ((! XDR_PUTINT32(&xdrs, &proc)) ||
406 	    (! AUTH_MARSHALL(auth, xid, &xdrs,
407 		m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
408 		errp->re_status = stat = RPC_CANTENCODEARGS;
409 		mtx_lock(&ct->ct_lock);
410 		goto out;
411 	}
412 	mreq->m_pkthdr.len = m_length(mreq, NULL);
413 
414 	/*
415 	 * Prepend a record marker containing the packet length.
416 	 */
417 	M_PREPEND(mreq, sizeof(uint32_t), M_WAIT);
418 	*mtod(mreq, uint32_t *) =
419 		htonl(0x80000000 | (mreq->m_pkthdr.len - sizeof(uint32_t)));
420 
421 	cr->cr_xid = xid;
422 	mtx_lock(&ct->ct_lock);
423 	/*
424 	 * Check to see if the other end has already started to close down
425 	 * the connection. The upcall will have set ct_error.re_status
426 	 * to RPC_CANTRECV if this is the case.
427 	 * If the other end starts to close down the connection after this
428 	 * point, it will be detected later when cr_error is checked,
429 	 * since the request is in the ct_pending queue.
430 	 */
431 	if (ct->ct_error.re_status == RPC_CANTRECV) {
432 		if (errp != &ct->ct_error) {
433 			errp->re_errno = ct->ct_error.re_errno;
434 			errp->re_status = RPC_CANTRECV;
435 		}
436 		stat = RPC_CANTRECV;
437 		goto out;
438 	}
439 	TAILQ_INSERT_TAIL(&ct->ct_pending, cr, cr_link);
440 	mtx_unlock(&ct->ct_lock);
441 
442 	/*
443 	 * sosend consumes mreq.
444 	 */
445 	error = sosend(ct->ct_socket, NULL, NULL, mreq, NULL, 0, curthread);
446 	mreq = NULL;
447 	if (error == EMSGSIZE) {
448 		SOCKBUF_LOCK(&ct->ct_socket->so_snd);
449 		sbwait(&ct->ct_socket->so_snd);
450 		SOCKBUF_UNLOCK(&ct->ct_socket->so_snd);
451 		AUTH_VALIDATE(auth, xid, NULL, NULL);
452 		mtx_lock(&ct->ct_lock);
453 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
454 		goto call_again;
455 	}
456 
457 	reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
458 	reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
459 	reply_msg.acpted_rply.ar_verf.oa_length = 0;
460 	reply_msg.acpted_rply.ar_results.where = NULL;
461 	reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
462 
463 	mtx_lock(&ct->ct_lock);
464 	if (error) {
465 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
466 		errp->re_errno = error;
467 		errp->re_status = stat = RPC_CANTSEND;
468 		goto out;
469 	}
470 
471 	/*
472 	 * Check to see if we got an upcall while waiting for the
473 	 * lock. In both these cases, the request has been removed
474 	 * from ct->ct_pending.
475 	 */
476 	if (cr->cr_error) {
477 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
478 		errp->re_errno = cr->cr_error;
479 		errp->re_status = stat = RPC_CANTRECV;
480 		goto out;
481 	}
482 	if (cr->cr_mrep) {
483 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
484 		goto got_reply;
485 	}
486 
487 	/*
488 	 * Hack to provide rpc-based message passing
489 	 */
490 	if (timeout.tv_sec == 0 && timeout.tv_usec == 0) {
491 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
492 		errp->re_status = stat = RPC_TIMEDOUT;
493 		goto out;
494 	}
495 
496 	error = msleep(cr, &ct->ct_lock, ct->ct_waitflag, ct->ct_waitchan,
497 	    tvtohz(&timeout));
498 
499 	TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
500 
501 	if (error) {
502 		/*
503 		 * The sleep returned an error so our request is still
504 		 * on the list. Turn the error code into an
505 		 * appropriate client status.
506 		 */
507 		errp->re_errno = error;
508 		switch (error) {
509 		case EINTR:
510 		case ERESTART:
511 			stat = RPC_INTR;
512 			break;
513 		case EWOULDBLOCK:
514 			stat = RPC_TIMEDOUT;
515 			break;
516 		default:
517 			stat = RPC_CANTRECV;
518 		}
519 		errp->re_status = stat;
520 		goto out;
521 	} else {
522 		/*
523 		 * We were woken up by the upcall.  If the
524 		 * upcall had a receive error, report that,
525 		 * otherwise we have a reply.
526 		 */
527 		if (cr->cr_error) {
528 			errp->re_errno = cr->cr_error;
529 			errp->re_status = stat = RPC_CANTRECV;
530 			goto out;
531 		}
532 	}
533 
534 got_reply:
535 	/*
536 	 * Now decode and validate the response. We need to drop the
537 	 * lock since xdr_replymsg may end up sleeping in malloc.
538 	 */
539 	mtx_unlock(&ct->ct_lock);
540 
541 	if (ext && ext->rc_feedback)
542 		ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
543 
544 	xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
545 	ok = xdr_replymsg(&xdrs, &reply_msg);
546 	cr->cr_mrep = NULL;
547 
548 	if (ok) {
549 		if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
550 		    (reply_msg.acpted_rply.ar_stat == SUCCESS))
551 			errp->re_status = stat = RPC_SUCCESS;
552 		else
553 			stat = _seterr_reply(&reply_msg, errp);
554 
555 		if (stat == RPC_SUCCESS) {
556 			results = xdrmbuf_getall(&xdrs);
557 			if (!AUTH_VALIDATE(auth, xid,
558 				&reply_msg.acpted_rply.ar_verf,
559 				&results)) {
560 				errp->re_status = stat = RPC_AUTHERROR;
561 				errp->re_why = AUTH_INVALIDRESP;
562 			} else {
563 				KASSERT(results,
564 				    ("auth validated but no result"));
565 				*resultsp = results;
566 			}
567 		}		/* end successful completion */
568 		/*
569 		 * If unsuccesful AND error is an authentication error
570 		 * then refresh credentials and try again, else break
571 		 */
572 		else if (stat == RPC_AUTHERROR)
573 			/* maybe our credentials need to be refreshed ... */
574 			if (nrefreshes > 0 &&
575 			    AUTH_REFRESH(auth, &reply_msg)) {
576 				nrefreshes--;
577 				XDR_DESTROY(&xdrs);
578 				mtx_lock(&ct->ct_lock);
579 				goto call_again;
580 			}
581 		/* end of unsuccessful completion */
582 	}	/* end of valid reply message */
583 	else {
584 		errp->re_status = stat = RPC_CANTDECODERES;
585 	}
586 	XDR_DESTROY(&xdrs);
587 	mtx_lock(&ct->ct_lock);
588 out:
589 	mtx_assert(&ct->ct_lock, MA_OWNED);
590 
591 	KASSERT(stat != RPC_SUCCESS || *resultsp,
592 	    ("RPC_SUCCESS without reply"));
593 
594 	if (mreq)
595 		m_freem(mreq);
596 	if (cr->cr_mrep)
597 		m_freem(cr->cr_mrep);
598 
599 	ct->ct_threads--;
600 	if (ct->ct_closing)
601 		wakeup(ct);
602 
603 	mtx_unlock(&ct->ct_lock);
604 
605 	if (auth && stat != RPC_SUCCESS)
606 		AUTH_VALIDATE(auth, xid, NULL, NULL);
607 
608 	free(cr, M_RPC);
609 
610 	return (stat);
611 }
612 
613 static void
614 clnt_vc_geterr(CLIENT *cl, struct rpc_err *errp)
615 {
616 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
617 
618 	*errp = ct->ct_error;
619 }
620 
621 static bool_t
622 clnt_vc_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
623 {
624 	XDR xdrs;
625 	bool_t dummy;
626 
627 	xdrs.x_op = XDR_FREE;
628 	dummy = (*xdr_res)(&xdrs, res_ptr);
629 
630 	return (dummy);
631 }
632 
633 /*ARGSUSED*/
634 static void
635 clnt_vc_abort(CLIENT *cl)
636 {
637 }
638 
639 static bool_t
640 clnt_vc_control(CLIENT *cl, u_int request, void *info)
641 {
642 	struct ct_data *ct = (struct ct_data *)cl->cl_private;
643 	void *infop = info;
644 
645 	mtx_lock(&ct->ct_lock);
646 
647 	switch (request) {
648 	case CLSET_FD_CLOSE:
649 		ct->ct_closeit = TRUE;
650 		mtx_unlock(&ct->ct_lock);
651 		return (TRUE);
652 	case CLSET_FD_NCLOSE:
653 		ct->ct_closeit = FALSE;
654 		mtx_unlock(&ct->ct_lock);
655 		return (TRUE);
656 	default:
657 		break;
658 	}
659 
660 	/* for other requests which use info */
661 	if (info == NULL) {
662 		mtx_unlock(&ct->ct_lock);
663 		return (FALSE);
664 	}
665 	switch (request) {
666 	case CLSET_TIMEOUT:
667 		if (time_not_ok((struct timeval *)info)) {
668 			mtx_unlock(&ct->ct_lock);
669 			return (FALSE);
670 		}
671 		ct->ct_wait = *(struct timeval *)infop;
672 		break;
673 	case CLGET_TIMEOUT:
674 		*(struct timeval *)infop = ct->ct_wait;
675 		break;
676 	case CLGET_SERVER_ADDR:
677 		(void) memcpy(info, &ct->ct_addr, (size_t)ct->ct_addr.ss_len);
678 		break;
679 	case CLGET_SVC_ADDR:
680 		/*
681 		 * Slightly different semantics to userland - we use
682 		 * sockaddr instead of netbuf.
683 		 */
684 		memcpy(info, &ct->ct_addr, ct->ct_addr.ss_len);
685 		break;
686 	case CLSET_SVC_ADDR:		/* set to new address */
687 		mtx_unlock(&ct->ct_lock);
688 		return (FALSE);
689 	case CLGET_XID:
690 		*(uint32_t *)info = ct->ct_xid;
691 		break;
692 	case CLSET_XID:
693 		/* This will set the xid of the NEXT call */
694 		/* decrement by 1 as clnt_vc_call() increments once */
695 		ct->ct_xid = *(uint32_t *)info - 1;
696 		break;
697 	case CLGET_VERS:
698 		/*
699 		 * This RELIES on the information that, in the call body,
700 		 * the version number field is the fifth field from the
701 		 * begining of the RPC header. MUST be changed if the
702 		 * call_struct is changed
703 		 */
704 		*(uint32_t *)info =
705 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
706 		    4 * BYTES_PER_XDR_UNIT));
707 		break;
708 
709 	case CLSET_VERS:
710 		*(uint32_t *)(void *)(ct->ct_mcallc +
711 		    4 * BYTES_PER_XDR_UNIT) =
712 		    htonl(*(uint32_t *)info);
713 		break;
714 
715 	case CLGET_PROG:
716 		/*
717 		 * This RELIES on the information that, in the call body,
718 		 * the program number field is the fourth field from the
719 		 * begining of the RPC header. MUST be changed if the
720 		 * call_struct is changed
721 		 */
722 		*(uint32_t *)info =
723 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
724 		    3 * BYTES_PER_XDR_UNIT));
725 		break;
726 
727 	case CLSET_PROG:
728 		*(uint32_t *)(void *)(ct->ct_mcallc +
729 		    3 * BYTES_PER_XDR_UNIT) =
730 		    htonl(*(uint32_t *)info);
731 		break;
732 
733 	case CLSET_WAITCHAN:
734 		ct->ct_waitchan = (const char *)info;
735 		break;
736 
737 	case CLGET_WAITCHAN:
738 		*(const char **) info = ct->ct_waitchan;
739 		break;
740 
741 	case CLSET_INTERRUPTIBLE:
742 		if (*(int *) info)
743 			ct->ct_waitflag = PCATCH | PBDRY;
744 		else
745 			ct->ct_waitflag = 0;
746 		break;
747 
748 	case CLGET_INTERRUPTIBLE:
749 		if (ct->ct_waitflag)
750 			*(int *) info = TRUE;
751 		else
752 			*(int *) info = FALSE;
753 		break;
754 
755 	default:
756 		mtx_unlock(&ct->ct_lock);
757 		return (FALSE);
758 	}
759 
760 	mtx_unlock(&ct->ct_lock);
761 	return (TRUE);
762 }
763 
764 static void
765 clnt_vc_close(CLIENT *cl)
766 {
767 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
768 	struct ct_request *cr;
769 
770 	mtx_lock(&ct->ct_lock);
771 
772 	if (ct->ct_closed) {
773 		mtx_unlock(&ct->ct_lock);
774 		return;
775 	}
776 
777 	if (ct->ct_closing) {
778 		while (ct->ct_closing)
779 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
780 		KASSERT(ct->ct_closed, ("client should be closed"));
781 		mtx_unlock(&ct->ct_lock);
782 		return;
783 	}
784 
785 	if (ct->ct_socket) {
786 		ct->ct_closing = TRUE;
787 		mtx_unlock(&ct->ct_lock);
788 
789 		SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
790 		soupcall_clear(ct->ct_socket, SO_RCV);
791 		clnt_vc_upcallsdone(ct);
792 		SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
793 
794 		/*
795 		 * Abort any pending requests and wait until everyone
796 		 * has finished with clnt_vc_call.
797 		 */
798 		mtx_lock(&ct->ct_lock);
799 		TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
800 			cr->cr_xid = 0;
801 			cr->cr_error = ESHUTDOWN;
802 			wakeup(cr);
803 		}
804 
805 		while (ct->ct_threads)
806 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
807 	}
808 
809 	ct->ct_closing = FALSE;
810 	ct->ct_closed = TRUE;
811 	mtx_unlock(&ct->ct_lock);
812 	wakeup(ct);
813 }
814 
815 static void
816 clnt_vc_destroy(CLIENT *cl)
817 {
818 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
819 	struct socket *so = NULL;
820 
821 	clnt_vc_close(cl);
822 
823 	mtx_lock(&ct->ct_lock);
824 
825 	if (ct->ct_socket) {
826 		if (ct->ct_closeit) {
827 			so = ct->ct_socket;
828 		}
829 	}
830 
831 	mtx_unlock(&ct->ct_lock);
832 
833 	mtx_destroy(&ct->ct_lock);
834 	if (so) {
835 		soshutdown(so, SHUT_WR);
836 		soclose(so);
837 	}
838 	mem_free(ct, sizeof(struct ct_data));
839 	mem_free(cl, sizeof(CLIENT));
840 }
841 
842 /*
843  * Make sure that the time is not garbage.   -1 value is disallowed.
844  * Note this is different from time_not_ok in clnt_dg.c
845  */
846 static bool_t
847 time_not_ok(struct timeval *t)
848 {
849 	return (t->tv_sec <= -1 || t->tv_sec > 100000000 ||
850 		t->tv_usec <= -1 || t->tv_usec > 1000000);
851 }
852 
853 int
854 clnt_vc_soupcall(struct socket *so, void *arg, int waitflag)
855 {
856 	struct ct_data *ct = (struct ct_data *) arg;
857 	struct uio uio;
858 	struct mbuf *m;
859 	struct ct_request *cr;
860 	int error, rcvflag, foundreq;
861 	uint32_t xid, header;
862 	bool_t do_read;
863 
864 	ct->ct_upcallrefs++;
865 	uio.uio_td = curthread;
866 	do {
867 		/*
868 		 * If ct_record_resid is zero, we are waiting for a
869 		 * record mark.
870 		 */
871 		if (ct->ct_record_resid == 0) {
872 
873 			/*
874 			 * Make sure there is either a whole record
875 			 * mark in the buffer or there is some other
876 			 * error condition
877 			 */
878 			do_read = FALSE;
879 			if (so->so_rcv.sb_cc >= sizeof(uint32_t)
880 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
881 			    || so->so_error)
882 				do_read = TRUE;
883 
884 			if (!do_read)
885 				break;
886 
887 			SOCKBUF_UNLOCK(&so->so_rcv);
888 			uio.uio_resid = sizeof(uint32_t);
889 			m = NULL;
890 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
891 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
892 			SOCKBUF_LOCK(&so->so_rcv);
893 
894 			if (error == EWOULDBLOCK)
895 				break;
896 
897 			/*
898 			 * If there was an error, wake up all pending
899 			 * requests.
900 			 */
901 			if (error || uio.uio_resid > 0) {
902 			wakeup_all:
903 				mtx_lock(&ct->ct_lock);
904 				if (!error) {
905 					/*
906 					 * We must have got EOF trying
907 					 * to read from the stream.
908 					 */
909 					error = ECONNRESET;
910 				}
911 				ct->ct_error.re_status = RPC_CANTRECV;
912 				ct->ct_error.re_errno = error;
913 				TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
914 					cr->cr_error = error;
915 					wakeup(cr);
916 				}
917 				mtx_unlock(&ct->ct_lock);
918 				break;
919 			}
920 			m_copydata(m, 0, sizeof(uint32_t), (char *)&header);
921 			header = ntohl(header);
922 			ct->ct_record = NULL;
923 			ct->ct_record_resid = header & 0x7fffffff;
924 			ct->ct_record_eor = ((header & 0x80000000) != 0);
925 			m_freem(m);
926 		} else {
927 			/*
928 			 * Wait until the socket has the whole record
929 			 * buffered.
930 			 */
931 			do_read = FALSE;
932 			if (so->so_rcv.sb_cc >= ct->ct_record_resid
933 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
934 			    || so->so_error)
935 				do_read = TRUE;
936 
937 			if (!do_read)
938 				break;
939 
940 			/*
941 			 * We have the record mark. Read as much as
942 			 * the socket has buffered up to the end of
943 			 * this record.
944 			 */
945 			SOCKBUF_UNLOCK(&so->so_rcv);
946 			uio.uio_resid = ct->ct_record_resid;
947 			m = NULL;
948 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
949 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
950 			SOCKBUF_LOCK(&so->so_rcv);
951 
952 			if (error == EWOULDBLOCK)
953 				break;
954 
955 			if (error || uio.uio_resid == ct->ct_record_resid)
956 				goto wakeup_all;
957 
958 			/*
959 			 * If we have part of the record already,
960 			 * chain this bit onto the end.
961 			 */
962 			if (ct->ct_record)
963 				m_last(ct->ct_record)->m_next = m;
964 			else
965 				ct->ct_record = m;
966 
967 			ct->ct_record_resid = uio.uio_resid;
968 
969 			/*
970 			 * If we have the entire record, see if we can
971 			 * match it to a request.
972 			 */
973 			if (ct->ct_record_resid == 0
974 			    && ct->ct_record_eor) {
975 				/*
976 				 * The XID is in the first uint32_t of
977 				 * the reply.
978 				 */
979 				if (ct->ct_record->m_len < sizeof(xid) &&
980 				    m_length(ct->ct_record, NULL) <
981 				    sizeof(xid)) {
982 					m_freem(ct->ct_record);
983 					break;
984 				}
985 				m_copydata(ct->ct_record, 0, sizeof(xid),
986 				    (char *)&xid);
987 				xid = ntohl(xid);
988 
989 				mtx_lock(&ct->ct_lock);
990 				foundreq = 0;
991 				TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
992 					if (cr->cr_xid == xid) {
993 						/*
994 						 * This one
995 						 * matches. We leave
996 						 * the reply mbuf in
997 						 * cr->cr_mrep. Set
998 						 * the XID to zero so
999 						 * that we will ignore
1000 						 * any duplicaed
1001 						 * replies.
1002 						 */
1003 						cr->cr_xid = 0;
1004 						cr->cr_mrep = ct->ct_record;
1005 						cr->cr_error = 0;
1006 						foundreq = 1;
1007 						wakeup(cr);
1008 						break;
1009 					}
1010 				}
1011 				mtx_unlock(&ct->ct_lock);
1012 
1013 				if (!foundreq)
1014 					m_freem(ct->ct_record);
1015 				ct->ct_record = NULL;
1016 			}
1017 		}
1018 	} while (m);
1019 	ct->ct_upcallrefs--;
1020 	if (ct->ct_upcallrefs < 0)
1021 		panic("rpcvc upcall refcnt");
1022 	if (ct->ct_upcallrefs == 0)
1023 		wakeup(&ct->ct_upcallrefs);
1024 	return (SU_OK);
1025 }
1026 
1027 /*
1028  * Wait for all upcalls in progress to complete.
1029  */
1030 static void
1031 clnt_vc_upcallsdone(struct ct_data *ct)
1032 {
1033 
1034 	SOCKBUF_LOCK_ASSERT(&ct->ct_socket->so_rcv);
1035 
1036 	while (ct->ct_upcallrefs > 0)
1037 		(void) msleep(&ct->ct_upcallrefs,
1038 		    SOCKBUF_MTX(&ct->ct_socket->so_rcv), 0, "rpcvcup", 0);
1039 }
1040