xref: /freebsd/sys/rpc/clnt_vc.c (revision 4b330699f819a81d8e34d471225143ffeb321855)
1 /*	$NetBSD: clnt_vc.c,v 1.4 2000/07/14 08:40:42 fvdl Exp $	*/
2 
3 /*-
4  * Copyright (c) 2009, Sun Microsystems, Inc.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions are met:
9  * - Redistributions of source code must retain the above copyright notice,
10  *   this list of conditions and the following disclaimer.
11  * - Redistributions in binary form must reproduce the above copyright notice,
12  *   this list of conditions and the following disclaimer in the documentation
13  *   and/or other materials provided with the distribution.
14  * - Neither the name of Sun Microsystems, Inc. nor the names of its
15  *   contributors may be used to endorse or promote products derived
16  *   from this software without specific prior written permission.
17  *
18  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
19  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
22  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28  * POSSIBILITY OF SUCH DAMAGE.
29  */
30 
31 #if defined(LIBC_SCCS) && !defined(lint)
32 static char *sccsid2 = "@(#)clnt_tcp.c 1.37 87/10/05 Copyr 1984 Sun Micro";
33 static char *sccsid = "@(#)clnt_tcp.c	2.2 88/08/01 4.0 RPCSRC";
34 static char sccsid3[] = "@(#)clnt_vc.c 1.19 89/03/16 Copyr 1988 Sun Micro";
35 #endif
36 #include <sys/cdefs.h>
37 __FBSDID("$FreeBSD$");
38 
39 /*
40  * clnt_tcp.c, Implements a TCP/IP based, client side RPC.
41  *
42  * Copyright (C) 1984, Sun Microsystems, Inc.
43  *
44  * TCP based RPC supports 'batched calls'.
45  * A sequence of calls may be batched-up in a send buffer.  The rpc call
46  * return immediately to the client even though the call was not necessarily
47  * sent.  The batching occurs if the results' xdr routine is NULL (0) AND
48  * the rpc timeout value is zero (see clnt.h, rpc).
49  *
50  * Clients should NOT casually batch calls that in fact return results; that is,
51  * the server side should be aware that a call is batched and not produce any
52  * return message.  Batched calls that produce many result messages can
53  * deadlock (netlock) the client and the server....
54  *
55  * Now go hang yourself.
56  */
57 
58 #include <sys/param.h>
59 #include <sys/systm.h>
60 #include <sys/kernel.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 int	fake_wchan;
112 
113 /*
114  * Create a client handle for a connection.
115  * Default options are set, which the user can change using clnt_control()'s.
116  * The rpc/vc package does buffering similar to stdio, so the client
117  * must pick send and receive buffer sizes, 0 => use the default.
118  * NB: fd is copied into a private area.
119  * NB: The rpch->cl_auth is set null authentication. Caller may wish to
120  * set this something more useful.
121  *
122  * fd should be an open socket
123  */
124 CLIENT *
125 clnt_vc_create(
126 	struct socket *so,		/* open file descriptor */
127 	struct sockaddr *raddr,		/* servers address */
128 	const rpcprog_t prog,		/* program number */
129 	const rpcvers_t vers,		/* version number */
130 	size_t sendsz,			/* buffer recv size */
131 	size_t recvsz,			/* buffer send size */
132 	int intrflag)			/* interruptible */
133 {
134 	CLIENT *cl;			/* client handle */
135 	struct ct_data *ct = NULL;	/* client handle */
136 	struct timeval now;
137 	struct rpc_msg call_msg;
138 	static uint32_t disrupt;
139 	struct __rpc_sockinfo si;
140 	XDR xdrs;
141 	int error, interrupted, one = 1, sleep_flag;
142 	struct sockopt sopt;
143 
144 	if (disrupt == 0)
145 		disrupt = (uint32_t)(long)raddr;
146 
147 	cl = (CLIENT *)mem_alloc(sizeof (*cl));
148 	ct = (struct ct_data *)mem_alloc(sizeof (*ct));
149 
150 	mtx_init(&ct->ct_lock, "ct->ct_lock", NULL, MTX_DEF);
151 	ct->ct_threads = 0;
152 	ct->ct_closing = FALSE;
153 	ct->ct_closed = FALSE;
154 	ct->ct_upcallrefs = 0;
155 
156 	if ((so->so_state & (SS_ISCONNECTED|SS_ISCONFIRMING)) == 0) {
157 		error = soconnect(so, raddr, curthread);
158 		SOCK_LOCK(so);
159 		interrupted = 0;
160 		sleep_flag = PSOCK;
161 		if (intrflag != 0)
162 			sleep_flag |= PCATCH;
163 		while ((so->so_state & SS_ISCONNECTING)
164 		    && so->so_error == 0) {
165 			error = msleep(&so->so_timeo, SOCK_MTX(so),
166 			    sleep_flag, "connec", 0);
167 			if (error) {
168 				if (error == EINTR || error == ERESTART)
169 					interrupted = 1;
170 				break;
171 			}
172 		}
173 		if (error == 0) {
174 			error = so->so_error;
175 			so->so_error = 0;
176 		}
177 		SOCK_UNLOCK(so);
178 		if (error) {
179 			if (!interrupted)
180 				so->so_state &= ~SS_ISCONNECTING;
181 			rpc_createerr.cf_stat = RPC_SYSTEMERROR;
182 			rpc_createerr.cf_error.re_errno = error;
183 			goto err;
184 		}
185 	}
186 
187 	if (!__rpc_socket2sockinfo(so, &si)) {
188 		goto err;
189 	}
190 
191 	if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
192 		bzero(&sopt, sizeof(sopt));
193 		sopt.sopt_dir = SOPT_SET;
194 		sopt.sopt_level = SOL_SOCKET;
195 		sopt.sopt_name = SO_KEEPALIVE;
196 		sopt.sopt_val = &one;
197 		sopt.sopt_valsize = sizeof(one);
198 		sosetopt(so, &sopt);
199 	}
200 
201 	if (so->so_proto->pr_protocol == IPPROTO_TCP) {
202 		bzero(&sopt, sizeof(sopt));
203 		sopt.sopt_dir = SOPT_SET;
204 		sopt.sopt_level = IPPROTO_TCP;
205 		sopt.sopt_name = TCP_NODELAY;
206 		sopt.sopt_val = &one;
207 		sopt.sopt_valsize = sizeof(one);
208 		sosetopt(so, &sopt);
209 	}
210 
211 	ct->ct_closeit = FALSE;
212 
213 	/*
214 	 * Set up private data struct
215 	 */
216 	ct->ct_socket = so;
217 	ct->ct_wait.tv_sec = -1;
218 	ct->ct_wait.tv_usec = -1;
219 	memcpy(&ct->ct_addr, raddr, raddr->sa_len);
220 
221 	/*
222 	 * Initialize call message
223 	 */
224 	getmicrotime(&now);
225 	ct->ct_xid = ((uint32_t)++disrupt) ^ __RPC_GETXID(&now);
226 	call_msg.rm_xid = ct->ct_xid;
227 	call_msg.rm_direction = CALL;
228 	call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION;
229 	call_msg.rm_call.cb_prog = (uint32_t)prog;
230 	call_msg.rm_call.cb_vers = (uint32_t)vers;
231 
232 	/*
233 	 * pre-serialize the static part of the call msg and stash it away
234 	 */
235 	xdrmem_create(&xdrs, ct->ct_mcallc, MCALL_MSG_SIZE,
236 	    XDR_ENCODE);
237 	if (! xdr_callhdr(&xdrs, &call_msg)) {
238 		if (ct->ct_closeit) {
239 			soclose(ct->ct_socket);
240 		}
241 		goto err;
242 	}
243 	ct->ct_mpos = XDR_GETPOS(&xdrs);
244 	XDR_DESTROY(&xdrs);
245 	ct->ct_waitchan = "rpcrecv";
246 	ct->ct_waitflag = 0;
247 
248 	/*
249 	 * Create a client handle which uses xdrrec for serialization
250 	 * and authnone for authentication.
251 	 */
252 	sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
253 	recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
254 	error = soreserve(ct->ct_socket, sendsz, recvsz);
255 	if (error != 0) {
256 		if (ct->ct_closeit) {
257 			soclose(ct->ct_socket);
258 		}
259 		goto err;
260 	}
261 	cl->cl_refs = 1;
262 	cl->cl_ops = &clnt_vc_ops;
263 	cl->cl_private = ct;
264 	cl->cl_auth = authnone_create();
265 
266 	SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
267 	soupcall_set(ct->ct_socket, SO_RCV, clnt_vc_soupcall, ct);
268 	SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
269 
270 	ct->ct_record = NULL;
271 	ct->ct_record_resid = 0;
272 	TAILQ_INIT(&ct->ct_pending);
273 	return (cl);
274 
275 err:
276 	mtx_destroy(&ct->ct_lock);
277 	mem_free(ct, sizeof (struct ct_data));
278 	mem_free(cl, sizeof (CLIENT));
279 
280 	return ((CLIENT *)NULL);
281 }
282 
283 static enum clnt_stat
284 clnt_vc_call(
285 	CLIENT		*cl,		/* client handle */
286 	struct rpc_callextra *ext,	/* call metadata */
287 	rpcproc_t	proc,		/* procedure number */
288 	struct mbuf	*args,		/* pointer to args */
289 	struct mbuf	**resultsp,	/* pointer to results */
290 	struct timeval	utimeout)
291 {
292 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
293 	AUTH *auth;
294 	struct rpc_err *errp;
295 	enum clnt_stat stat;
296 	XDR xdrs;
297 	struct rpc_msg reply_msg;
298 	bool_t ok;
299 	int nrefreshes = 2;		/* number of times to refresh cred */
300 	struct timeval timeout;
301 	uint32_t xid;
302 	struct mbuf *mreq = NULL, *results;
303 	struct ct_request *cr;
304 	int error, trycnt;
305 
306 	cr = malloc(sizeof(struct ct_request), M_RPC, M_WAITOK);
307 
308 	mtx_lock(&ct->ct_lock);
309 
310 	if (ct->ct_closing || ct->ct_closed) {
311 		mtx_unlock(&ct->ct_lock);
312 		free(cr, M_RPC);
313 		return (RPC_CANTSEND);
314 	}
315 	ct->ct_threads++;
316 
317 	if (ext) {
318 		auth = ext->rc_auth;
319 		errp = &ext->rc_err;
320 	} else {
321 		auth = cl->cl_auth;
322 		errp = &ct->ct_error;
323 	}
324 
325 	cr->cr_mrep = NULL;
326 	cr->cr_error = 0;
327 
328 	if (ct->ct_wait.tv_usec == -1) {
329 		timeout = utimeout;	/* use supplied timeout */
330 	} else {
331 		timeout = ct->ct_wait;	/* use default timeout */
332 	}
333 
334 	/*
335 	 * After 15sec of looping, allow it to return RPC_CANTSEND, which will
336 	 * cause the clnt_reconnect layer to create a new TCP connection.
337 	 */
338 	trycnt = 15 * hz;
339 call_again:
340 	mtx_assert(&ct->ct_lock, MA_OWNED);
341 	if (ct->ct_closing || ct->ct_closed) {
342 		ct->ct_threads--;
343 		wakeup(ct);
344 		mtx_unlock(&ct->ct_lock);
345 		free(cr, M_RPC);
346 		return (RPC_CANTSEND);
347 	}
348 
349 	ct->ct_xid++;
350 	xid = ct->ct_xid;
351 
352 	mtx_unlock(&ct->ct_lock);
353 
354 	/*
355 	 * Leave space to pre-pend the record mark.
356 	 */
357 	mreq = m_gethdr(M_WAITOK, MT_DATA);
358 	mreq->m_data += sizeof(uint32_t);
359 	KASSERT(ct->ct_mpos + sizeof(uint32_t) <= MHLEN,
360 	    ("RPC header too big"));
361 	bcopy(ct->ct_mcallc, mreq->m_data, ct->ct_mpos);
362 	mreq->m_len = ct->ct_mpos;
363 
364 	/*
365 	 * The XID is the first thing in the request.
366 	 */
367 	*mtod(mreq, uint32_t *) = htonl(xid);
368 
369 	xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
370 
371 	errp->re_status = stat = RPC_SUCCESS;
372 
373 	if ((! XDR_PUTINT32(&xdrs, &proc)) ||
374 	    (! AUTH_MARSHALL(auth, xid, &xdrs,
375 		m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
376 		errp->re_status = stat = RPC_CANTENCODEARGS;
377 		mtx_lock(&ct->ct_lock);
378 		goto out;
379 	}
380 	mreq->m_pkthdr.len = m_length(mreq, NULL);
381 
382 	/*
383 	 * Prepend a record marker containing the packet length.
384 	 */
385 	M_PREPEND(mreq, sizeof(uint32_t), M_WAITOK);
386 	*mtod(mreq, uint32_t *) =
387 		htonl(0x80000000 | (mreq->m_pkthdr.len - sizeof(uint32_t)));
388 
389 	cr->cr_xid = xid;
390 	mtx_lock(&ct->ct_lock);
391 	/*
392 	 * Check to see if the other end has already started to close down
393 	 * the connection. The upcall will have set ct_error.re_status
394 	 * to RPC_CANTRECV if this is the case.
395 	 * If the other end starts to close down the connection after this
396 	 * point, it will be detected later when cr_error is checked,
397 	 * since the request is in the ct_pending queue.
398 	 */
399 	if (ct->ct_error.re_status == RPC_CANTRECV) {
400 		if (errp != &ct->ct_error) {
401 			errp->re_errno = ct->ct_error.re_errno;
402 			errp->re_status = RPC_CANTRECV;
403 		}
404 		stat = RPC_CANTRECV;
405 		goto out;
406 	}
407 	TAILQ_INSERT_TAIL(&ct->ct_pending, cr, cr_link);
408 	mtx_unlock(&ct->ct_lock);
409 
410 	/*
411 	 * sosend consumes mreq.
412 	 */
413 	error = sosend(ct->ct_socket, NULL, NULL, mreq, NULL, 0, curthread);
414 	mreq = NULL;
415 	if (error == EMSGSIZE || (error == ERESTART &&
416 	    (ct->ct_waitflag & PCATCH) == 0 && trycnt-- > 0)) {
417 		SOCKBUF_LOCK(&ct->ct_socket->so_snd);
418 		sbwait(&ct->ct_socket->so_snd);
419 		SOCKBUF_UNLOCK(&ct->ct_socket->so_snd);
420 		AUTH_VALIDATE(auth, xid, NULL, NULL);
421 		mtx_lock(&ct->ct_lock);
422 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
423 		/* Sleep for 1 clock tick before trying the sosend() again. */
424 		msleep(&fake_wchan, &ct->ct_lock, 0, "rpclpsnd", 1);
425 		goto call_again;
426 	}
427 
428 	reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
429 	reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
430 	reply_msg.acpted_rply.ar_verf.oa_length = 0;
431 	reply_msg.acpted_rply.ar_results.where = NULL;
432 	reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
433 
434 	mtx_lock(&ct->ct_lock);
435 	if (error) {
436 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
437 		errp->re_errno = error;
438 		errp->re_status = stat = RPC_CANTSEND;
439 		goto out;
440 	}
441 
442 	/*
443 	 * Check to see if we got an upcall while waiting for the
444 	 * lock. In both these cases, the request has been removed
445 	 * from ct->ct_pending.
446 	 */
447 	if (cr->cr_error) {
448 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
449 		errp->re_errno = cr->cr_error;
450 		errp->re_status = stat = RPC_CANTRECV;
451 		goto out;
452 	}
453 	if (cr->cr_mrep) {
454 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
455 		goto got_reply;
456 	}
457 
458 	/*
459 	 * Hack to provide rpc-based message passing
460 	 */
461 	if (timeout.tv_sec == 0 && timeout.tv_usec == 0) {
462 		TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
463 		errp->re_status = stat = RPC_TIMEDOUT;
464 		goto out;
465 	}
466 
467 	error = msleep(cr, &ct->ct_lock, ct->ct_waitflag, ct->ct_waitchan,
468 	    tvtohz(&timeout));
469 
470 	TAILQ_REMOVE(&ct->ct_pending, cr, cr_link);
471 
472 	if (error) {
473 		/*
474 		 * The sleep returned an error so our request is still
475 		 * on the list. Turn the error code into an
476 		 * appropriate client status.
477 		 */
478 		errp->re_errno = error;
479 		switch (error) {
480 		case EINTR:
481 			stat = RPC_INTR;
482 			break;
483 		case EWOULDBLOCK:
484 			stat = RPC_TIMEDOUT;
485 			break;
486 		default:
487 			stat = RPC_CANTRECV;
488 		}
489 		errp->re_status = stat;
490 		goto out;
491 	} else {
492 		/*
493 		 * We were woken up by the upcall.  If the
494 		 * upcall had a receive error, report that,
495 		 * otherwise we have a reply.
496 		 */
497 		if (cr->cr_error) {
498 			errp->re_errno = cr->cr_error;
499 			errp->re_status = stat = RPC_CANTRECV;
500 			goto out;
501 		}
502 	}
503 
504 got_reply:
505 	/*
506 	 * Now decode and validate the response. We need to drop the
507 	 * lock since xdr_replymsg may end up sleeping in malloc.
508 	 */
509 	mtx_unlock(&ct->ct_lock);
510 
511 	if (ext && ext->rc_feedback)
512 		ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
513 
514 	xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
515 	ok = xdr_replymsg(&xdrs, &reply_msg);
516 	cr->cr_mrep = NULL;
517 
518 	if (ok) {
519 		if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
520 		    (reply_msg.acpted_rply.ar_stat == SUCCESS))
521 			errp->re_status = stat = RPC_SUCCESS;
522 		else
523 			stat = _seterr_reply(&reply_msg, errp);
524 
525 		if (stat == RPC_SUCCESS) {
526 			results = xdrmbuf_getall(&xdrs);
527 			if (!AUTH_VALIDATE(auth, xid,
528 				&reply_msg.acpted_rply.ar_verf,
529 				&results)) {
530 				errp->re_status = stat = RPC_AUTHERROR;
531 				errp->re_why = AUTH_INVALIDRESP;
532 			} else {
533 				KASSERT(results,
534 				    ("auth validated but no result"));
535 				*resultsp = results;
536 			}
537 		}		/* end successful completion */
538 		/*
539 		 * If unsuccessful AND error is an authentication error
540 		 * then refresh credentials and try again, else break
541 		 */
542 		else if (stat == RPC_AUTHERROR)
543 			/* maybe our credentials need to be refreshed ... */
544 			if (nrefreshes > 0 &&
545 			    AUTH_REFRESH(auth, &reply_msg)) {
546 				nrefreshes--;
547 				XDR_DESTROY(&xdrs);
548 				mtx_lock(&ct->ct_lock);
549 				goto call_again;
550 			}
551 		/* end of unsuccessful completion */
552 	}	/* end of valid reply message */
553 	else {
554 		errp->re_status = stat = RPC_CANTDECODERES;
555 	}
556 	XDR_DESTROY(&xdrs);
557 	mtx_lock(&ct->ct_lock);
558 out:
559 	mtx_assert(&ct->ct_lock, MA_OWNED);
560 
561 	KASSERT(stat != RPC_SUCCESS || *resultsp,
562 	    ("RPC_SUCCESS without reply"));
563 
564 	if (mreq)
565 		m_freem(mreq);
566 	if (cr->cr_mrep)
567 		m_freem(cr->cr_mrep);
568 
569 	ct->ct_threads--;
570 	if (ct->ct_closing)
571 		wakeup(ct);
572 
573 	mtx_unlock(&ct->ct_lock);
574 
575 	if (auth && stat != RPC_SUCCESS)
576 		AUTH_VALIDATE(auth, xid, NULL, NULL);
577 
578 	free(cr, M_RPC);
579 
580 	return (stat);
581 }
582 
583 static void
584 clnt_vc_geterr(CLIENT *cl, struct rpc_err *errp)
585 {
586 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
587 
588 	*errp = ct->ct_error;
589 }
590 
591 static bool_t
592 clnt_vc_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
593 {
594 	XDR xdrs;
595 	bool_t dummy;
596 
597 	xdrs.x_op = XDR_FREE;
598 	dummy = (*xdr_res)(&xdrs, res_ptr);
599 
600 	return (dummy);
601 }
602 
603 /*ARGSUSED*/
604 static void
605 clnt_vc_abort(CLIENT *cl)
606 {
607 }
608 
609 static bool_t
610 clnt_vc_control(CLIENT *cl, u_int request, void *info)
611 {
612 	struct ct_data *ct = (struct ct_data *)cl->cl_private;
613 	void *infop = info;
614 	SVCXPRT *xprt;
615 
616 	mtx_lock(&ct->ct_lock);
617 
618 	switch (request) {
619 	case CLSET_FD_CLOSE:
620 		ct->ct_closeit = TRUE;
621 		mtx_unlock(&ct->ct_lock);
622 		return (TRUE);
623 	case CLSET_FD_NCLOSE:
624 		ct->ct_closeit = FALSE;
625 		mtx_unlock(&ct->ct_lock);
626 		return (TRUE);
627 	default:
628 		break;
629 	}
630 
631 	/* for other requests which use info */
632 	if (info == NULL) {
633 		mtx_unlock(&ct->ct_lock);
634 		return (FALSE);
635 	}
636 	switch (request) {
637 	case CLSET_TIMEOUT:
638 		if (time_not_ok((struct timeval *)info)) {
639 			mtx_unlock(&ct->ct_lock);
640 			return (FALSE);
641 		}
642 		ct->ct_wait = *(struct timeval *)infop;
643 		break;
644 	case CLGET_TIMEOUT:
645 		*(struct timeval *)infop = ct->ct_wait;
646 		break;
647 	case CLGET_SERVER_ADDR:
648 		(void) memcpy(info, &ct->ct_addr, (size_t)ct->ct_addr.ss_len);
649 		break;
650 	case CLGET_SVC_ADDR:
651 		/*
652 		 * Slightly different semantics to userland - we use
653 		 * sockaddr instead of netbuf.
654 		 */
655 		memcpy(info, &ct->ct_addr, ct->ct_addr.ss_len);
656 		break;
657 	case CLSET_SVC_ADDR:		/* set to new address */
658 		mtx_unlock(&ct->ct_lock);
659 		return (FALSE);
660 	case CLGET_XID:
661 		*(uint32_t *)info = ct->ct_xid;
662 		break;
663 	case CLSET_XID:
664 		/* This will set the xid of the NEXT call */
665 		/* decrement by 1 as clnt_vc_call() increments once */
666 		ct->ct_xid = *(uint32_t *)info - 1;
667 		break;
668 	case CLGET_VERS:
669 		/*
670 		 * This RELIES on the information that, in the call body,
671 		 * the version number field is the fifth field from the
672 		 * beginning of the RPC header. MUST be changed if the
673 		 * call_struct is changed
674 		 */
675 		*(uint32_t *)info =
676 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
677 		    4 * BYTES_PER_XDR_UNIT));
678 		break;
679 
680 	case CLSET_VERS:
681 		*(uint32_t *)(void *)(ct->ct_mcallc +
682 		    4 * BYTES_PER_XDR_UNIT) =
683 		    htonl(*(uint32_t *)info);
684 		break;
685 
686 	case CLGET_PROG:
687 		/*
688 		 * This RELIES on the information that, in the call body,
689 		 * the program number field is the fourth field from the
690 		 * beginning of the RPC header. MUST be changed if the
691 		 * call_struct is changed
692 		 */
693 		*(uint32_t *)info =
694 		    ntohl(*(uint32_t *)(void *)(ct->ct_mcallc +
695 		    3 * BYTES_PER_XDR_UNIT));
696 		break;
697 
698 	case CLSET_PROG:
699 		*(uint32_t *)(void *)(ct->ct_mcallc +
700 		    3 * BYTES_PER_XDR_UNIT) =
701 		    htonl(*(uint32_t *)info);
702 		break;
703 
704 	case CLSET_WAITCHAN:
705 		ct->ct_waitchan = (const char *)info;
706 		break;
707 
708 	case CLGET_WAITCHAN:
709 		*(const char **) info = ct->ct_waitchan;
710 		break;
711 
712 	case CLSET_INTERRUPTIBLE:
713 		if (*(int *) info)
714 			ct->ct_waitflag = PCATCH;
715 		else
716 			ct->ct_waitflag = 0;
717 		break;
718 
719 	case CLGET_INTERRUPTIBLE:
720 		if (ct->ct_waitflag)
721 			*(int *) info = TRUE;
722 		else
723 			*(int *) info = FALSE;
724 		break;
725 
726 	case CLSET_BACKCHANNEL:
727 		xprt = (SVCXPRT *)info;
728 		if (ct->ct_backchannelxprt == NULL) {
729 			xprt->xp_p2 = ct;
730 			ct->ct_backchannelxprt = xprt;
731 		}
732 		break;
733 
734 	default:
735 		mtx_unlock(&ct->ct_lock);
736 		return (FALSE);
737 	}
738 
739 	mtx_unlock(&ct->ct_lock);
740 	return (TRUE);
741 }
742 
743 static void
744 clnt_vc_close(CLIENT *cl)
745 {
746 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
747 	struct ct_request *cr;
748 
749 	mtx_lock(&ct->ct_lock);
750 
751 	if (ct->ct_closed) {
752 		mtx_unlock(&ct->ct_lock);
753 		return;
754 	}
755 
756 	if (ct->ct_closing) {
757 		while (ct->ct_closing)
758 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
759 		KASSERT(ct->ct_closed, ("client should be closed"));
760 		mtx_unlock(&ct->ct_lock);
761 		return;
762 	}
763 
764 	if (ct->ct_socket) {
765 		ct->ct_closing = TRUE;
766 		mtx_unlock(&ct->ct_lock);
767 
768 		SOCKBUF_LOCK(&ct->ct_socket->so_rcv);
769 		soupcall_clear(ct->ct_socket, SO_RCV);
770 		clnt_vc_upcallsdone(ct);
771 		SOCKBUF_UNLOCK(&ct->ct_socket->so_rcv);
772 
773 		/*
774 		 * Abort any pending requests and wait until everyone
775 		 * has finished with clnt_vc_call.
776 		 */
777 		mtx_lock(&ct->ct_lock);
778 		TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
779 			cr->cr_xid = 0;
780 			cr->cr_error = ESHUTDOWN;
781 			wakeup(cr);
782 		}
783 
784 		while (ct->ct_threads)
785 			msleep(ct, &ct->ct_lock, 0, "rpcclose", 0);
786 	}
787 
788 	ct->ct_closing = FALSE;
789 	ct->ct_closed = TRUE;
790 	mtx_unlock(&ct->ct_lock);
791 	wakeup(ct);
792 }
793 
794 static void
795 clnt_vc_destroy(CLIENT *cl)
796 {
797 	struct ct_data *ct = (struct ct_data *) cl->cl_private;
798 	struct socket *so = NULL;
799 	SVCXPRT *xprt;
800 
801 	clnt_vc_close(cl);
802 
803 	mtx_lock(&ct->ct_lock);
804 	xprt = ct->ct_backchannelxprt;
805 	ct->ct_backchannelxprt = NULL;
806 	if (xprt != NULL) {
807 		mtx_unlock(&ct->ct_lock);	/* To avoid a LOR. */
808 		sx_xlock(&xprt->xp_lock);
809 		mtx_lock(&ct->ct_lock);
810 		xprt->xp_p2 = NULL;
811 		sx_xunlock(&xprt->xp_lock);
812 	}
813 
814 	if (ct->ct_socket) {
815 		if (ct->ct_closeit) {
816 			so = ct->ct_socket;
817 		}
818 	}
819 
820 	mtx_unlock(&ct->ct_lock);
821 
822 	mtx_destroy(&ct->ct_lock);
823 	if (so) {
824 		soshutdown(so, SHUT_WR);
825 		soclose(so);
826 	}
827 	mem_free(ct, sizeof(struct ct_data));
828 	if (cl->cl_netid && cl->cl_netid[0])
829 		mem_free(cl->cl_netid, strlen(cl->cl_netid) +1);
830 	if (cl->cl_tp && cl->cl_tp[0])
831 		mem_free(cl->cl_tp, strlen(cl->cl_tp) +1);
832 	mem_free(cl, sizeof(CLIENT));
833 }
834 
835 /*
836  * Make sure that the time is not garbage.   -1 value is disallowed.
837  * Note this is different from time_not_ok in clnt_dg.c
838  */
839 static bool_t
840 time_not_ok(struct timeval *t)
841 {
842 	return (t->tv_sec <= -1 || t->tv_sec > 100000000 ||
843 		t->tv_usec <= -1 || t->tv_usec > 1000000);
844 }
845 
846 int
847 clnt_vc_soupcall(struct socket *so, void *arg, int waitflag)
848 {
849 	struct ct_data *ct = (struct ct_data *) arg;
850 	struct uio uio;
851 	struct mbuf *m, *m2;
852 	struct ct_request *cr;
853 	int error, rcvflag, foundreq;
854 	uint32_t xid_plus_direction[2], header;
855 	bool_t do_read;
856 	SVCXPRT *xprt;
857 	struct cf_conn *cd;
858 
859 	CTASSERT(sizeof(xid_plus_direction) == 2 * sizeof(uint32_t));
860 	ct->ct_upcallrefs++;
861 	uio.uio_td = curthread;
862 	do {
863 		/*
864 		 * If ct_record_resid is zero, we are waiting for a
865 		 * record mark.
866 		 */
867 		if (ct->ct_record_resid == 0) {
868 
869 			/*
870 			 * Make sure there is either a whole record
871 			 * mark in the buffer or there is some other
872 			 * error condition
873 			 */
874 			do_read = FALSE;
875 			if (sbavail(&so->so_rcv) >= sizeof(uint32_t)
876 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
877 			    || so->so_error)
878 				do_read = TRUE;
879 
880 			if (!do_read)
881 				break;
882 
883 			SOCKBUF_UNLOCK(&so->so_rcv);
884 			uio.uio_resid = sizeof(uint32_t);
885 			m = NULL;
886 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
887 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
888 			SOCKBUF_LOCK(&so->so_rcv);
889 
890 			if (error == EWOULDBLOCK)
891 				break;
892 
893 			/*
894 			 * If there was an error, wake up all pending
895 			 * requests.
896 			 */
897 			if (error || uio.uio_resid > 0) {
898 			wakeup_all:
899 				mtx_lock(&ct->ct_lock);
900 				if (!error) {
901 					/*
902 					 * We must have got EOF trying
903 					 * to read from the stream.
904 					 */
905 					error = ECONNRESET;
906 				}
907 				ct->ct_error.re_status = RPC_CANTRECV;
908 				ct->ct_error.re_errno = error;
909 				TAILQ_FOREACH(cr, &ct->ct_pending, cr_link) {
910 					cr->cr_error = error;
911 					wakeup(cr);
912 				}
913 				mtx_unlock(&ct->ct_lock);
914 				break;
915 			}
916 			m_copydata(m, 0, sizeof(uint32_t), (char *)&header);
917 			header = ntohl(header);
918 			ct->ct_record = NULL;
919 			ct->ct_record_resid = header & 0x7fffffff;
920 			ct->ct_record_eor = ((header & 0x80000000) != 0);
921 			m_freem(m);
922 		} else {
923 			/*
924 			 * Wait until the socket has the whole record
925 			 * buffered.
926 			 */
927 			do_read = FALSE;
928 			if (sbavail(&so->so_rcv) >= ct->ct_record_resid
929 			    || (so->so_rcv.sb_state & SBS_CANTRCVMORE)
930 			    || so->so_error)
931 				do_read = TRUE;
932 
933 			if (!do_read)
934 				break;
935 
936 			/*
937 			 * We have the record mark. Read as much as
938 			 * the socket has buffered up to the end of
939 			 * this record.
940 			 */
941 			SOCKBUF_UNLOCK(&so->so_rcv);
942 			uio.uio_resid = ct->ct_record_resid;
943 			m = NULL;
944 			rcvflag = MSG_DONTWAIT | MSG_SOCALLBCK;
945 			error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
946 			SOCKBUF_LOCK(&so->so_rcv);
947 
948 			if (error == EWOULDBLOCK)
949 				break;
950 
951 			if (error || uio.uio_resid == ct->ct_record_resid)
952 				goto wakeup_all;
953 
954 			/*
955 			 * If we have part of the record already,
956 			 * chain this bit onto the end.
957 			 */
958 			if (ct->ct_record)
959 				m_last(ct->ct_record)->m_next = m;
960 			else
961 				ct->ct_record = m;
962 
963 			ct->ct_record_resid = uio.uio_resid;
964 
965 			/*
966 			 * If we have the entire record, see if we can
967 			 * match it to a request.
968 			 */
969 			if (ct->ct_record_resid == 0
970 			    && ct->ct_record_eor) {
971 				/*
972 				 * The XID is in the first uint32_t of
973 				 * the reply and the message direction
974 				 * is the second one.
975 				 */
976 				if (ct->ct_record->m_len <
977 				    sizeof(xid_plus_direction) &&
978 				    m_length(ct->ct_record, NULL) <
979 				    sizeof(xid_plus_direction)) {
980 					m_freem(ct->ct_record);
981 					break;
982 				}
983 				m_copydata(ct->ct_record, 0,
984 				    sizeof(xid_plus_direction),
985 				    (char *)xid_plus_direction);
986 				xid_plus_direction[0] =
987 				    ntohl(xid_plus_direction[0]);
988 				xid_plus_direction[1] =
989 				    ntohl(xid_plus_direction[1]);
990 				/* Check message direction. */
991 				if (xid_plus_direction[1] == CALL) {
992 					/* This is a backchannel request. */
993 					mtx_lock(&ct->ct_lock);
994 					xprt = ct->ct_backchannelxprt;
995 					if (xprt == NULL) {
996 						mtx_unlock(&ct->ct_lock);
997 						/* Just throw it away. */
998 						m_freem(ct->ct_record);
999 						ct->ct_record = NULL;
1000 					} else {
1001 						cd = (struct cf_conn *)
1002 						    xprt->xp_p1;
1003 						m2 = cd->mreq;
1004 						/*
1005 						 * The requests are chained
1006 						 * in the m_nextpkt list.
1007 						 */
1008 						while (m2 != NULL &&
1009 						    m2->m_nextpkt != NULL)
1010 							/* Find end of list. */
1011 							m2 = m2->m_nextpkt;
1012 						if (m2 != NULL)
1013 							m2->m_nextpkt =
1014 							    ct->ct_record;
1015 						else
1016 							cd->mreq =
1017 							    ct->ct_record;
1018 						ct->ct_record->m_nextpkt =
1019 						    NULL;
1020 						ct->ct_record = NULL;
1021 						xprt_active(xprt);
1022 						mtx_unlock(&ct->ct_lock);
1023 					}
1024 				} else {
1025 					mtx_lock(&ct->ct_lock);
1026 					foundreq = 0;
1027 					TAILQ_FOREACH(cr, &ct->ct_pending,
1028 					    cr_link) {
1029 						if (cr->cr_xid ==
1030 						    xid_plus_direction[0]) {
1031 							/*
1032 							 * This one
1033 							 * matches. We leave
1034 							 * the reply mbuf in
1035 							 * cr->cr_mrep. Set
1036 							 * the XID to zero so
1037 							 * that we will ignore
1038 							 * any duplicated
1039 							 * replies.
1040 							 */
1041 							cr->cr_xid = 0;
1042 							cr->cr_mrep =
1043 							    ct->ct_record;
1044 							cr->cr_error = 0;
1045 							foundreq = 1;
1046 							wakeup(cr);
1047 							break;
1048 						}
1049 					}
1050 					mtx_unlock(&ct->ct_lock);
1051 
1052 					if (!foundreq)
1053 						m_freem(ct->ct_record);
1054 					ct->ct_record = NULL;
1055 				}
1056 			}
1057 		}
1058 	} while (m);
1059 	ct->ct_upcallrefs--;
1060 	if (ct->ct_upcallrefs < 0)
1061 		panic("rpcvc upcall refcnt");
1062 	if (ct->ct_upcallrefs == 0)
1063 		wakeup(&ct->ct_upcallrefs);
1064 	return (SU_OK);
1065 }
1066 
1067 /*
1068  * Wait for all upcalls in progress to complete.
1069  */
1070 static void
1071 clnt_vc_upcallsdone(struct ct_data *ct)
1072 {
1073 
1074 	SOCKBUF_LOCK_ASSERT(&ct->ct_socket->so_rcv);
1075 
1076 	while (ct->ct_upcallrefs > 0)
1077 		(void) msleep(&ct->ct_upcallrefs,
1078 		    SOCKBUF_MTX(&ct->ct_socket->so_rcv), 0, "rpcvcup", 0);
1079 }
1080