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)
32static char *sccsid2 = "@(#)clnt_tcp.c 1.37 87/10/05 Copyr 1984 Sun Micro";
33static char *sccsid = "@(#)clnt_tcp.c	2.2 88/08/01 4.0 RPCSRC";
34static char sccsid3[] = "@(#)clnt_vc.c 1.19 89/03/16 Copyr 1988 Sun Micro";
35#endif
36#include <sys/cdefs.h>
37__FBSDID("$FreeBSD: stable/10/sys/rpc/clnt_vc.c 318675 2017-05-22 19:57:20Z rmacklem $");
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
83struct cmessage {
84        struct cmsghdr cmsg;
85        struct cmsgcred cmcred;
86};
87
88static enum clnt_stat clnt_vc_call(CLIENT *, struct rpc_callextra *,
89    rpcproc_t, struct mbuf *, struct mbuf **, struct timeval);
90static void clnt_vc_geterr(CLIENT *, struct rpc_err *);
91static bool_t clnt_vc_freeres(CLIENT *, xdrproc_t, void *);
92static void clnt_vc_abort(CLIENT *);
93static bool_t clnt_vc_control(CLIENT *, u_int, void *);
94static void clnt_vc_close(CLIENT *);
95static void clnt_vc_destroy(CLIENT *);
96static bool_t time_not_ok(struct timeval *);
97static int clnt_vc_soupcall(struct socket *so, void *arg, int waitflag);
98
99static 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
109static void clnt_vc_upcallsdone(struct ct_data *);
110
111static 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 */
124CLIENT *
125clnt_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
275err:
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
283static enum clnt_stat
284clnt_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;
339call_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
504got_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 unsuccesful 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);
558out:
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
583static void
584clnt_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
591static bool_t
592clnt_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*/
604static void
605clnt_vc_abort(CLIENT *cl)
606{
607}
608
609static bool_t
610clnt_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		 * begining 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		 * begining 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
743static void
744clnt_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
794static void
795clnt_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 */
839static bool_t
840time_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
846int
847clnt_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 (so->so_rcv.sb_cc >= 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 (so->so_rcv.sb_cc >= 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 */
1070static void
1071clnt_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