xref: /freebsd/sys/rpc/clnt_vc.c (revision d1d015864103b253b3fcb2f72a0da5b0cfeb31b6)
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/sx.h>
71 #include <sys/syslog.h>
72 #include <sys/time.h>
73 #include <sys/uio.h>
74 
75 #include <net/vnet.h>
76 
77 #include <netinet/tcp.h>
78 
79 #include <rpc/rpc.h>
80 #include <rpc/rpc_com.h>
81 #include <rpc/krpc.h>
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 static void clnt_vc_upcallsdone(struct ct_data *);
110 
111 static const char clnt_vc_errstr[] = "%s : %s";
112 static const char clnt_vc_str[] = "clnt_vc_create";
113 static const char clnt_read_vc_str[] = "read_vc";
114 static const char __no_mem_str[] = "out of memory";
115 
116 /*
117  * Create a client handle for a connection.
118  * Default options are set, which the user can change using clnt_control()'s.
119  * The rpc/vc package does buffering similar to stdio, so the client
120  * must pick send and receive buffer sizes, 0 => use the default.
121  * NB: fd is copied into a private area.
122  * NB: The rpch->cl_auth is set null authentication. Caller may wish to
123  * set this something more useful.
124  *
125  * fd should be an open socket
126  */
127 CLIENT *
128 clnt_vc_create(
129 	struct socket *so,		/* open file descriptor */
130 	struct sockaddr *raddr,		/* servers address */
131 	const rpcprog_t prog,		/* program number */
132 	const rpcvers_t vers,		/* version number */
133 	size_t sendsz,			/* buffer recv size */
134 	size_t recvsz,			/* buffer send size */
135 	int intrflag)			/* interruptible */
136 {
137 	CLIENT *cl;			/* client handle */
138 	struct ct_data *ct = NULL;	/* client handle */
139 	struct timeval now;
140 	struct rpc_msg call_msg;
141 	static uint32_t disrupt;
142 	struct __rpc_sockinfo si;
143 	XDR xdrs;
144 	int error, interrupted, one = 1, sleep_flag;
145 	struct sockopt sopt;
146 
147 	if (disrupt == 0)
148 		disrupt = (uint32_t)(long)raddr;
149 
150 	cl = (CLIENT *)mem_alloc(sizeof (*cl));
151 	ct = (struct ct_data *)mem_alloc(sizeof (*ct));
152 
153 	mtx_init(&ct->ct_lock, "ct->ct_lock", NULL, MTX_DEF);
154 	ct->ct_threads = 0;
155 	ct->ct_closing = FALSE;
156 	ct->ct_closed = FALSE;
157 	ct->ct_upcallrefs = 0;
158 
159 	if ((so->so_state & (SS_ISCONNECTED|SS_ISCONFIRMING)) == 0) {
160 		error = soconnect(so, raddr, curthread);
161 		SOCK_LOCK(so);
162 		interrupted = 0;
163 		sleep_flag = PSOCK;
164 		if (intrflag != 0)
165 			sleep_flag |= PCATCH;
166 		while ((so->so_state & SS_ISCONNECTING)
167 		    && so->so_error == 0) {
168 			error = msleep(&so->so_timeo, SOCK_MTX(so),
169 			    sleep_flag, "connec", 0);
170 			if (error) {
171 				if (error == EINTR || error == ERESTART)
172 					interrupted = 1;
173 				break;
174 			}
175 		}
176 		if (error == 0) {
177 			error = so->so_error;
178 			so->so_error = 0;
179 		}
180 		SOCK_UNLOCK(so);
181 		if (error) {
182 			if (!interrupted)
183 				so->so_state &= ~SS_ISCONNECTING;
184 			rpc_createerr.cf_stat = RPC_SYSTEMERROR;
185 			rpc_createerr.cf_error.re_errno = error;
186 			goto err;
187 		}
188 	}
189 
190 	if (!__rpc_socket2sockinfo(so, &si)) {
191 		goto err;
192 	}
193 
194 	if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
195 		bzero(&sopt, sizeof(sopt));
196 		sopt.sopt_dir = SOPT_SET;
197 		sopt.sopt_level = SOL_SOCKET;
198 		sopt.sopt_name = SO_KEEPALIVE;
199 		sopt.sopt_val = &one;
200 		sopt.sopt_valsize = sizeof(one);
201 		sosetopt(so, &sopt);
202 	}
203 
204 	if (so->so_proto->pr_protocol == IPPROTO_TCP) {
205 		bzero(&sopt, sizeof(sopt));
206 		sopt.sopt_dir = SOPT_SET;
207 		sopt.sopt_level = IPPROTO_TCP;
208 		sopt.sopt_name = TCP_NODELAY;
209 		sopt.sopt_val = &one;
210 		sopt.sopt_valsize = sizeof(one);
211 		sosetopt(so, &sopt);
212 	}
213 
214 	ct->ct_closeit = FALSE;
215 
216 	/*
217 	 * Set up private data struct
218 	 */
219 	ct->ct_socket = so;
220 	ct->ct_wait.tv_sec = -1;
221 	ct->ct_wait.tv_usec = -1;
222 	memcpy(&ct->ct_addr, raddr, raddr->sa_len);
223 
224 	/*
225 	 * Initialize call message
226 	 */
227 	getmicrotime(&now);
228 	ct->ct_xid = ((uint32_t)++disrupt) ^ __RPC_GETXID(&now);
229 	call_msg.rm_xid = ct->ct_xid;
230 	call_msg.rm_direction = CALL;
231 	call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION;
232 	call_msg.rm_call.cb_prog = (uint32_t)prog;
233 	call_msg.rm_call.cb_vers = (uint32_t)vers;
234 
235 	/*
236 	 * pre-serialize the static part of the call msg and stash it away
237 	 */
238 	xdrmem_create(&xdrs, ct->ct_mcallc, MCALL_MSG_SIZE,
239 	    XDR_ENCODE);
240 	if (! xdr_callhdr(&xdrs, &call_msg)) {
241 		if (ct->ct_closeit) {
242 			soclose(ct->ct_socket);
243 		}
244 		goto err;
245 	}
246 	ct->ct_mpos = XDR_GETPOS(&xdrs);
247 	XDR_DESTROY(&xdrs);
248 	ct->ct_waitchan = "rpcrecv";
249 	ct->ct_waitflag = 0;
250 
251 	/*
252 	 * Create a client handle which uses xdrrec for serialization
253 	 * and authnone for authentication.
254 	 */
255 	sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
256 	recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
257 	error = soreserve(ct->ct_socket, sendsz, recvsz);
258 	if (error != 0) {
259 		if (ct->ct_closeit) {
260 			soclose(ct->ct_socket);
261 		}
262 		goto err;
263 	}
264 	cl->cl_refs = 1;
265 	cl->cl_ops = &clnt_vc_ops;
266 	cl->cl_private = ct;
267 	cl->cl_auth = authnone_create();
268 
269 	SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
270 	soupcall_set(ct->ct_socket, SO_RCV, clnt_vc_soupcall, ct);
271 	SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
272 
273 	ct->ct_record = NULL;
274 	ct->ct_record_resid = 0;
275 	TAILQ_INIT(&ct->ct_pending);
276 	return (cl);
277 
278 err:
279 	if (cl) {
280 		if (ct) {
281 			mtx_destroy(&ct->ct_lock);
282 			mem_free(ct, sizeof (struct ct_data));
283 		}
284 		if (cl)
285 			mem_free(cl, sizeof (CLIENT));
286 	}
287 	return ((CLIENT *)NULL);
288 }
289 
290 static enum clnt_stat
291 clnt_vc_call(
292 	CLIENT		*cl,		/* client handle */
293 	struct rpc_callextra *ext,	/* call metadata */
294 	rpcproc_t	proc,		/* procedure number */
295 	struct mbuf	*args,		/* pointer to args */
296 	struct mbuf	**resultsp,	/* pointer to results */
297 	struct timeval	utimeout)
298 {
299 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
300 	AUTH *auth;
301 	struct rpc_err *errp;
302 	enum clnt_stat stat;
303 	XDR xdrs;
304 	struct rpc_msg reply_msg;
305 	bool_t ok;
306 	int nrefreshes = 2;		/* number of times to refresh cred */
307 	struct timeval timeout;
308 	uint32_t xid;
309 	struct mbuf *mreq = NULL, *results;
310 	struct ct_request *cr;
311 	int error;
312 
313 	cr = malloc(sizeof(struct ct_request), M_RPC, M_WAITOK);
314 
315 	mtx_lock(&ct->ct_lock);
316 
317 	if (ct->ct_closing || ct->ct_closed) {
318 		mtx_unlock(&ct->ct_lock);
319 		free(cr, M_RPC);
320 		return (RPC_CANTSEND);
321 	}
322 	ct->ct_threads++;
323 
324 	if (ext) {
325 		auth = ext->rc_auth;
326 		errp = &ext->rc_err;
327 	} else {
328 		auth = cl->cl_auth;
329 		errp = &ct->ct_error;
330 	}
331 
332 	cr->cr_mrep = NULL;
333 	cr->cr_error = 0;
334 
335 	if (ct->ct_wait.tv_usec == -1) {
336 		timeout = utimeout;	/* use supplied timeout */
337 	} else {
338 		timeout = ct->ct_wait;	/* use default timeout */
339 	}
340 
341 call_again:
342 	mtx_assert(&ct->ct_lock, MA_OWNED);
343 
344 	ct->ct_xid++;
345 	xid = ct->ct_xid;
346 
347 	mtx_unlock(&ct->ct_lock);
348 
349 	/*
350 	 * Leave space to pre-pend the record mark.
351 	 */
352 	mreq = m_gethdr(M_WAITOK, MT_DATA);
353 	mreq->m_data += sizeof(uint32_t);
354 	KASSERT(ct->ct_mpos + sizeof(uint32_t) <= MHLEN,
355 	    ("RPC header too big"));
356 	bcopy(ct->ct_mcallc, mreq->m_data, ct->ct_mpos);
357 	mreq->m_len = ct->ct_mpos;
358 
359 	/*
360 	 * The XID is the first thing in the request.
361 	 */
362 	*mtod(mreq, uint32_t *) = htonl(xid);
363 
364 	xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
365 
366 	errp->re_status = stat = RPC_SUCCESS;
367 
368 	if ((! XDR_PUTINT32(&xdrs, &proc)) ||
369 	    (! AUTH_MARSHALL(auth, xid, &xdrs,
370 		m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
371 		errp->re_status = stat = RPC_CANTENCODEARGS;
372 		mtx_lock(&ct->ct_lock);
373 		goto out;
374 	}
375 	mreq->m_pkthdr.len = m_length(mreq, NULL);
376 
377 	/*
378 	 * Prepend a record marker containing the packet length.
379 	 */
380 	M_PREPEND(mreq, sizeof(uint32_t), M_WAITOK);
381 	*mtod(mreq, uint32_t *) =
382 		htonl(0x80000000 | (mreq->m_pkthdr.len - sizeof(uint32_t)));
383 
384 	cr->cr_xid = xid;
385 	mtx_lock(&ct->ct_lock);
386 	/*
387 	 * Check to see if the other end has already started to close down
388 	 * the connection. The upcall will have set ct_error.re_status
389 	 * to RPC_CANTRECV if this is the case.
390 	 * If the other end starts to close down the connection after this
391 	 * point, it will be detected later when cr_error is checked,
392 	 * since the request is in the ct_pending queue.
393 	 */
394 	if (ct->ct_error.re_status == RPC_CANTRECV) {
395 		if (errp != &ct->ct_error) {
396 			errp->re_errno = ct->ct_error.re_errno;
397 			errp->re_status = RPC_CANTRECV;
398 		}
399 		stat = RPC_CANTRECV;
400 		goto out;
401 	}
402 	TAILQ_INSERT_TAIL(&ct->ct_pending, cr, cr_link);
403 	mtx_unlock(&ct->ct_lock);
404 
405 	/*
406 	 * sosend consumes mreq.
407 	 */
408 	error = sosend(ct->ct_socket, NULL, NULL, mreq, NULL, 0, curthread);
409 	mreq = NULL;
410 	if (error == EMSGSIZE) {
411 		SOCKBUF_LOCK(&ct->ct_socket->so_snd);
412 		sbwait(&ct->ct_socket->so_snd);
413 		SOCKBUF_UNLOCK(&ct->ct_socket->so_snd);
414 		AUTH_VALIDATE(auth, xid, NULL, NULL);
415 		mtx_lock(&ct->ct_lock);
416 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
417 		goto call_again;
418 	}
419 
420 	reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
421 	reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
422 	reply_msg.acpted_rply.ar_verf.oa_length = 0;
423 	reply_msg.acpted_rply.ar_results.where = NULL;
424 	reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
425 
426 	mtx_lock(&ct->ct_lock);
427 	if (error) {
428 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
429 		errp->re_errno = error;
430 		errp->re_status = stat = RPC_CANTSEND;
431 		goto out;
432 	}
433 
434 	/*
435 	 * Check to see if we got an upcall while waiting for the
436 	 * lock. In both these cases, the request has been removed
437 	 * from ct->ct_pending.
438 	 */
439 	if (cr->cr_error) {
440 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
441 		errp->re_errno = cr->cr_error;
442 		errp->re_status = stat = RPC_CANTRECV;
443 		goto out;
444 	}
445 	if (cr->cr_mrep) {
446 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
447 		goto got_reply;
448 	}
449 
450 	/*
451 	 * Hack to provide rpc-based message passing
452 	 */
453 	if (timeout.tv_sec == 0 && timeout.tv_usec == 0) {
454 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
455 		errp->re_status = stat = RPC_TIMEDOUT;
456 		goto out;
457 	}
458 
459 	error = msleep(cr, &ct->ct_lock, ct->ct_waitflag, ct->ct_waitchan,
460 	    tvtohz(&timeout));
461 
462 	TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
463 
464 	if (error) {
465 		/*
466 		 * The sleep returned an error so our request is still
467 		 * on the list. Turn the error code into an
468 		 * appropriate client status.
469 		 */
470 		errp->re_errno = error;
471 		switch (error) {
472 		case EINTR:
473 			stat = RPC_INTR;
474 			break;
475 		case EWOULDBLOCK:
476 			stat = RPC_TIMEDOUT;
477 			break;
478 		default:
479 			stat = RPC_CANTRECV;
480 		}
481 		errp->re_status = stat;
482 		goto out;
483 	} else {
484 		/*
485 		 * We were woken up by the upcall.  If the
486 		 * upcall had a receive error, report that,
487 		 * otherwise we have a reply.
488 		 */
489 		if (cr->cr_error) {
490 			errp->re_errno = cr->cr_error;
491 			errp->re_status = stat = RPC_CANTRECV;
492 			goto out;
493 		}
494 	}
495 
496 got_reply:
497 	/*
498 	 * Now decode and validate the response. We need to drop the
499 	 * lock since xdr_replymsg may end up sleeping in malloc.
500 	 */
501 	mtx_unlock(&ct->ct_lock);
502 
503 	if (ext && ext->rc_feedback)
504 		ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
505 
506 	xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
507 	ok = xdr_replymsg(&xdrs, &reply_msg);
508 	cr->cr_mrep = NULL;
509 
510 	if (ok) {
511 		if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
512 		    (reply_msg.acpted_rply.ar_stat == SUCCESS))
513 			errp->re_status = stat = RPC_SUCCESS;
514 		else
515 			stat = _seterr_reply(&reply_msg, errp);
516 
517 		if (stat == RPC_SUCCESS) {
518 			results = xdrmbuf_getall(&xdrs);
519 			if (!AUTH_VALIDATE(auth, xid,
520 				&reply_msg.acpted_rply.ar_verf,
521 				&results)) {
522 				errp->re_status = stat = RPC_AUTHERROR;
523 				errp->re_why = AUTH_INVALIDRESP;
524 			} else {
525 				KASSERT(results,
526 				    ("auth validated but no result"));
527 				*resultsp = results;
528 			}
529 		}		/* end successful completion */
530 		/*
531 		 * If unsuccesful AND error is an authentication error
532 		 * then refresh credentials and try again, else break
533 		 */
534 		else if (stat == RPC_AUTHERROR)
535 			/* maybe our credentials need to be refreshed ... */
536 			if (nrefreshes > 0 &&
537 			    AUTH_REFRESH(auth, &reply_msg)) {
538 				nrefreshes--;
539 				XDR_DESTROY(&xdrs);
540 				mtx_lock(&ct->ct_lock);
541 				goto call_again;
542 			}
543 		/* end of unsuccessful completion */
544 	}	/* end of valid reply message */
545 	else {
546 		errp->re_status = stat = RPC_CANTDECODERES;
547 	}
548 	XDR_DESTROY(&xdrs);
549 	mtx_lock(&ct->ct_lock);
550 out:
551 	mtx_assert(&ct->ct_lock, MA_OWNED);
552 
553 	KASSERT(stat != RPC_SUCCESS || *resultsp,
554 	    ("RPC_SUCCESS without reply"));
555 
556 	if (mreq)
557 		m_freem(mreq);
558 	if (cr->cr_mrep)
559 		m_freem(cr->cr_mrep);
560 
561 	ct->ct_threads--;
562 	if (ct->ct_closing)
563 		wakeup(ct);
564 
565 	mtx_unlock(&ct->ct_lock);
566 
567 	if (auth && stat != RPC_SUCCESS)
568 		AUTH_VALIDATE(auth, xid, NULL, NULL);
569 
570 	free(cr, M_RPC);
571 
572 	return (stat);
573 }
574 
575 static void
576 clnt_vc_geterr(CLIENT *cl, struct rpc_err *errp)
577 {
578 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
579 
580 	*errp = ct->ct_error;
581 }
582 
583 static bool_t
584 clnt_vc_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
585 {
586 	XDR xdrs;
587 	bool_t dummy;
588 
589 	xdrs.x_op = XDR_FREE;
590 	dummy = (*xdr_res)(&xdrs, res_ptr);
591 
592 	return (dummy);
593 }
594 
595 /*ARGSUSED*/
596 static void
597 clnt_vc_abort(CLIENT *cl)
598 {
599 }
600 
601 static bool_t
602 clnt_vc_control(CLIENT *cl, u_int request, void *info)
603 {
604 	struct ct_data *ct = (struct ct_data *)cl->cl_private;
605 	void *infop = info;
606 	SVCXPRT *xprt;
607 
608 	mtx_lock(&ct->ct_lock);
609 
610 	switch (request) {
611 	case CLSET_FD_CLOSE:
612 		ct->ct_closeit = TRUE;
613 		mtx_unlock(&ct->ct_lock);
614 		return (TRUE);
615 	case CLSET_FD_NCLOSE:
616 		ct->ct_closeit = FALSE;
617 		mtx_unlock(&ct->ct_lock);
618 		return (TRUE);
619 	default:
620 		break;
621 	}
622 
623 	/* for other requests which use info */
624 	if (info == NULL) {
625 		mtx_unlock(&ct->ct_lock);
626 		return (FALSE);
627 	}
628 	switch (request) {
629 	case CLSET_TIMEOUT:
630 		if (time_not_ok((struct timeval *)info)) {
631 			mtx_unlock(&ct->ct_lock);
632 			return (FALSE);
633 		}
634 		ct->ct_wait = *(struct timeval *)infop;
635 		break;
636 	case CLGET_TIMEOUT:
637 		*(struct timeval *)infop = ct->ct_wait;
638 		break;
639 	case CLGET_SERVER_ADDR:
640 		(void) memcpy(info, &ct->ct_addr, (size_t)ct->ct_addr.ss_len);
641 		break;
642 	case CLGET_SVC_ADDR:
643 		/*
644 		 * Slightly different semantics to userland - we use
645 		 * sockaddr instead of netbuf.
646 		 */
647 		memcpy(info, &ct->ct_addr, ct->ct_addr.ss_len);
648 		break;
649 	case CLSET_SVC_ADDR:		/* set to new address */
650 		mtx_unlock(&ct->ct_lock);
651 		return (FALSE);
652 	case CLGET_XID:
653 		*(uint32_t *)info = ct->ct_xid;
654 		break;
655 	case CLSET_XID:
656 		/* This will set the xid of the NEXT call */
657 		/* decrement by 1 as clnt_vc_call() increments once */
658 		ct->ct_xid = *(uint32_t *)info - 1;
659 		break;
660 	case CLGET_VERS:
661 		/*
662 		 * This RELIES on the information that, in the call body,
663 		 * the version number field is the fifth field from the
664 		 * begining of the RPC header. MUST be changed if the
665 		 * call_struct is changed
666 		 */
667 		*(uint32_t *)info =
668 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
669 		    4 * BYTES_PER_XDR_UNIT));
670 		break;
671 
672 	case CLSET_VERS:
673 		*(uint32_t *)(void *)(ct->ct_mcallc +
674 		    4 * BYTES_PER_XDR_UNIT) =
675 		    htonl(*(uint32_t *)info);
676 		break;
677 
678 	case CLGET_PROG:
679 		/*
680 		 * This RELIES on the information that, in the call body,
681 		 * the program number field is the fourth field from the
682 		 * begining of the RPC header. MUST be changed if the
683 		 * call_struct is changed
684 		 */
685 		*(uint32_t *)info =
686 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
687 		    3 * BYTES_PER_XDR_UNIT));
688 		break;
689 
690 	case CLSET_PROG:
691 		*(uint32_t *)(void *)(ct->ct_mcallc +
692 		    3 * BYTES_PER_XDR_UNIT) =
693 		    htonl(*(uint32_t *)info);
694 		break;
695 
696 	case CLSET_WAITCHAN:
697 		ct->ct_waitchan = (const char *)info;
698 		break;
699 
700 	case CLGET_WAITCHAN:
701 		*(const char **) info = ct->ct_waitchan;
702 		break;
703 
704 	case CLSET_INTERRUPTIBLE:
705 		if (*(int *) info)
706 			ct->ct_waitflag = PCATCH;
707 		else
708 			ct->ct_waitflag = 0;
709 		break;
710 
711 	case CLGET_INTERRUPTIBLE:
712 		if (ct->ct_waitflag)
713 			*(int *) info = TRUE;
714 		else
715 			*(int *) info = FALSE;
716 		break;
717 
718 	case CLSET_BACKCHANNEL:
719 		xprt = (SVCXPRT *)info;
720 		if (ct->ct_backchannelxprt == NULL) {
721 			xprt->xp_p2 = ct;
722 			ct->ct_backchannelxprt = xprt;
723 		}
724 		break;
725 
726 	default:
727 		mtx_unlock(&ct->ct_lock);
728 		return (FALSE);
729 	}
730 
731 	mtx_unlock(&ct->ct_lock);
732 	return (TRUE);
733 }
734 
735 static void
736 clnt_vc_close(CLIENT *cl)
737 {
738 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
739 	struct ct_request *cr;
740 
741 	mtx_lock(&ct->ct_lock);
742 
743 	if (ct->ct_closed) {
744 		mtx_unlock(&ct->ct_lock);
745 		return;
746 	}
747 
748 	if (ct->ct_closing) {
749 		while (ct->ct_closing)
750 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
751 		KASSERT(ct->ct_closed, ("client should be closed"));
752 		mtx_unlock(&ct->ct_lock);
753 		return;
754 	}
755 
756 	if (ct->ct_socket) {
757 		ct->ct_closing = TRUE;
758 		mtx_unlock(&ct->ct_lock);
759 
760 		SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
761 		soupcall_clear(ct->ct_socket, SO_RCV);
762 		clnt_vc_upcallsdone(ct);
763 		SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
764 
765 		/*
766 		 * Abort any pending requests and wait until everyone
767 		 * has finished with clnt_vc_call.
768 		 */
769 		mtx_lock(&ct->ct_lock);
770 		TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
771 			cr->cr_xid = 0;
772 			cr->cr_error = ESHUTDOWN;
773 			wakeup(cr);
774 		}
775 
776 		while (ct->ct_threads)
777 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
778 	}
779 
780 	ct->ct_closing = FALSE;
781 	ct->ct_closed = TRUE;
782 	mtx_unlock(&ct->ct_lock);
783 	wakeup(ct);
784 }
785 
786 static void
787 clnt_vc_destroy(CLIENT *cl)
788 {
789 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
790 	struct socket *so = NULL;
791 	SVCXPRT *xprt;
792 
793 	clnt_vc_close(cl);
794 
795 	mtx_lock(&ct->ct_lock);
796 	xprt = ct->ct_backchannelxprt;
797 	ct->ct_backchannelxprt = NULL;
798 	if (xprt != NULL) {
799 		mtx_unlock(&ct->ct_lock);	/* To avoid a LOR. */
800 		sx_xlock(&xprt->xp_lock);
801 		mtx_lock(&ct->ct_lock);
802 		xprt->xp_p2 = NULL;
803 		xprt_unregister(xprt);
804 	}
805 
806 	if (ct->ct_socket) {
807 		if (ct->ct_closeit) {
808 			so = ct->ct_socket;
809 		}
810 	}
811 
812 	mtx_unlock(&ct->ct_lock);
813 	if (xprt != NULL) {
814 		sx_xunlock(&xprt->xp_lock);
815 		SVC_RELEASE(xprt);
816 	}
817 
818 	mtx_destroy(&ct->ct_lock);
819 	if (so) {
820 		soshutdown(so, SHUT_WR);
821 		soclose(so);
822 	}
823 	mem_free(ct, sizeof(struct ct_data));
824 	if (cl->cl_netid && cl->cl_netid[0])
825 		mem_free(cl->cl_netid, strlen(cl->cl_netid) +1);
826 	if (cl->cl_tp && cl->cl_tp[0])
827 		mem_free(cl->cl_tp, strlen(cl->cl_tp) +1);
828 	mem_free(cl, sizeof(CLIENT));
829 }
830 
831 /*
832  * Make sure that the time is not garbage.   -1 value is disallowed.
833  * Note this is different from time_not_ok in clnt_dg.c
834  */
835 static bool_t
836 time_not_ok(struct timeval *t)
837 {
838 	return (t->tv_sec <= -1 || t->tv_sec > 100000000 ||
839 		t->tv_usec <= -1 || t->tv_usec > 1000000);
840 }
841 
842 int
843 clnt_vc_soupcall(struct socket *so, void *arg, int waitflag)
844 {
845 	struct ct_data *ct = (struct ct_data *) arg;
846 	struct uio uio;
847 	struct mbuf *m, *m2;
848 	struct ct_request *cr;
849 	int error, rcvflag, foundreq;
850 	uint32_t xid_plus_direction[2], header;
851 	bool_t do_read;
852 	SVCXPRT *xprt;
853 	struct cf_conn *cd;
854 
855 	CTASSERT(sizeof(xid_plus_direction) == 2 * sizeof(uint32_t));
856 	ct->ct_upcallrefs++;
857 	uio.uio_td = curthread;
858 	do {
859 		/*
860 		 * If ct_record_resid is zero, we are waiting for a
861 		 * record mark.
862 		 */
863 		if (ct->ct_record_resid == 0) {
864 
865 			/*
866 			 * Make sure there is either a whole record
867 			 * mark in the buffer or there is some other
868 			 * error condition
869 			 */
870 			do_read = FALSE;
871 			if (so->so_rcv.sb_cc >= sizeof(uint32_t)
872 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
873 			    || so->so_error)
874 				do_read = TRUE;
875 
876 			if (!do_read)
877 				break;
878 
879 			SOCKBUF_UNLOCK(&so->so_rcv);
880 			uio.uio_resid = sizeof(uint32_t);
881 			m = NULL;
882 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
883 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
884 			SOCKBUF_LOCK(&so->so_rcv);
885 
886 			if (error == EWOULDBLOCK)
887 				break;
888 
889 			/*
890 			 * If there was an error, wake up all pending
891 			 * requests.
892 			 */
893 			if (error || uio.uio_resid > 0) {
894 			wakeup_all:
895 				mtx_lock(&ct->ct_lock);
896 				if (!error) {
897 					/*
898 					 * We must have got EOF trying
899 					 * to read from the stream.
900 					 */
901 					error = ECONNRESET;
902 				}
903 				ct->ct_error.re_status = RPC_CANTRECV;
904 				ct->ct_error.re_errno = error;
905 				TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
906 					cr->cr_error = error;
907 					wakeup(cr);
908 				}
909 				mtx_unlock(&ct->ct_lock);
910 				break;
911 			}
912 			m_copydata(m, 0, sizeof(uint32_t), (char *)&header);
913 			header = ntohl(header);
914 			ct->ct_record = NULL;
915 			ct->ct_record_resid = header & 0x7fffffff;
916 			ct->ct_record_eor = ((header & 0x80000000) != 0);
917 			m_freem(m);
918 		} else {
919 			/*
920 			 * Wait until the socket has the whole record
921 			 * buffered.
922 			 */
923 			do_read = FALSE;
924 			if (so->so_rcv.sb_cc >= ct->ct_record_resid
925 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
926 			    || so->so_error)
927 				do_read = TRUE;
928 
929 			if (!do_read)
930 				break;
931 
932 			/*
933 			 * We have the record mark. Read as much as
934 			 * the socket has buffered up to the end of
935 			 * this record.
936 			 */
937 			SOCKBUF_UNLOCK(&so->so_rcv);
938 			uio.uio_resid = ct->ct_record_resid;
939 			m = NULL;
940 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
941 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
942 			SOCKBUF_LOCK(&so->so_rcv);
943 
944 			if (error == EWOULDBLOCK)
945 				break;
946 
947 			if (error || uio.uio_resid == ct->ct_record_resid)
948 				goto wakeup_all;
949 
950 			/*
951 			 * If we have part of the record already,
952 			 * chain this bit onto the end.
953 			 */
954 			if (ct->ct_record)
955 				m_last(ct->ct_record)->m_next = m;
956 			else
957 				ct->ct_record = m;
958 
959 			ct->ct_record_resid = uio.uio_resid;
960 
961 			/*
962 			 * If we have the entire record, see if we can
963 			 * match it to a request.
964 			 */
965 			if (ct->ct_record_resid == 0
966 			    && ct->ct_record_eor) {
967 				/*
968 				 * The XID is in the first uint32_t of
969 				 * the reply and the message direction
970 				 * is the second one.
971 				 */
972 				if (ct->ct_record->m_len <
973 				    sizeof(xid_plus_direction) &&
974 				    m_length(ct->ct_record, NULL) <
975 				    sizeof(xid_plus_direction)) {
976 					m_freem(ct->ct_record);
977 					break;
978 				}
979 				m_copydata(ct->ct_record, 0,
980 				    sizeof(xid_plus_direction),
981 				    (char *)xid_plus_direction);
982 				xid_plus_direction[0] =
983 				    ntohl(xid_plus_direction[0]);
984 				xid_plus_direction[1] =
985 				    ntohl(xid_plus_direction[1]);
986 				/* Check message direction. */
987 				if (xid_plus_direction[1] == CALL) {
988 					/* This is a backchannel request. */
989 					mtx_lock(&ct->ct_lock);
990 					xprt = ct->ct_backchannelxprt;
991 					if (xprt == NULL) {
992 						mtx_unlock(&ct->ct_lock);
993 						/* Just throw it away. */
994 						m_freem(ct->ct_record);
995 						ct->ct_record = NULL;
996 					} else {
997 						cd = (struct cf_conn *)
998 						    xprt->xp_p1;
999 						m2 = cd->mreq;
1000 						/*
1001 						 * The requests are chained
1002 						 * in the m_nextpkt list.
1003 						 */
1004 						while (m2 != NULL &&
1005 						    m2->m_nextpkt != NULL)
1006 							/* Find end of list. */
1007 							m2 = m2->m_nextpkt;
1008 						if (m2 != NULL)
1009 							m2->m_nextpkt =
1010 							    ct->ct_record;
1011 						else
1012 							cd->mreq =
1013 							    ct->ct_record;
1014 						ct->ct_record->m_nextpkt =
1015 						    NULL;
1016 						ct->ct_record = NULL;
1017 						xprt_active(xprt);
1018 						mtx_unlock(&ct->ct_lock);
1019 					}
1020 				} else {
1021 					mtx_lock(&ct->ct_lock);
1022 					foundreq = 0;
1023 					TAILQ_FOREACH(cr, &ct->ct_pending,
1024 					    cr_link) {
1025 						if (cr->cr_xid ==
1026 						    xid_plus_direction[0]) {
1027 							/*
1028 							 * This one
1029 							 * matches. We leave
1030 							 * the reply mbuf in
1031 							 * cr->cr_mrep. Set
1032 							 * the XID to zero so
1033 							 * that we will ignore
1034 							 * any duplicated
1035 							 * replies.
1036 							 */
1037 							cr->cr_xid = 0;
1038 							cr->cr_mrep =
1039 							    ct->ct_record;
1040 							cr->cr_error = 0;
1041 							foundreq = 1;
1042 							wakeup(cr);
1043 							break;
1044 						}
1045 					}
1046 					mtx_unlock(&ct->ct_lock);
1047 
1048 					if (!foundreq)
1049 						m_freem(ct->ct_record);
1050 					ct->ct_record = NULL;
1051 				}
1052 			}
1053 		}
1054 	} while (m);
1055 	ct->ct_upcallrefs--;
1056 	if (ct->ct_upcallrefs < 0)
1057 		panic("rpcvc upcall refcnt");
1058 	if (ct->ct_upcallrefs == 0)
1059 		wakeup(&ct->ct_upcallrefs);
1060 	return (SU_OK);
1061 }
1062 
1063 /*
1064  * Wait for all upcalls in progress to complete.
1065  */
1066 static void
1067 clnt_vc_upcallsdone(struct ct_data *ct)
1068 {
1069 
1070 	SOCKBUF_LOCK_ASSERT(&ct->ct_socket->so_rcv);
1071 
1072 	while (ct->ct_upcallrefs > 0)
1073 		(void) msleep(&ct->ct_upcallrefs,
1074 		    SOCKBUF_MTX(&ct->ct_socket->so_rcv), 0, "rpcvcup", 0);
1075 }
1076