Annotation of src/sys/kern/uipc_socket.c, Revision 1.78
1.78 ! matt 1: /* $NetBSD: uipc_socket.c,v 1.77 2003/02/01 06:23:44 thorpej Exp $ */
1.64 thorpej 2:
3: /*-
4: * Copyright (c) 2002 The NetBSD Foundation, Inc.
5: * All rights reserved.
6: *
7: * This code is derived from software contributed to The NetBSD Foundation
8: * by Jason R. Thorpe of Wasabi Systems, Inc.
9: *
10: * Redistribution and use in source and binary forms, with or without
11: * modification, are permitted provided that the following conditions
12: * are met:
13: * 1. Redistributions of source code must retain the above copyright
14: * notice, this list of conditions and the following disclaimer.
15: * 2. Redistributions in binary form must reproduce the above copyright
16: * notice, this list of conditions and the following disclaimer in the
17: * documentation and/or other materials provided with the distribution.
18: * 3. All advertising materials mentioning features or use of this software
19: * must display the following acknowledgement:
20: * This product includes software developed by the NetBSD
21: * Foundation, Inc. and its contributors.
22: * 4. Neither the name of The NetBSD Foundation nor the names of its
23: * contributors may be used to endorse or promote products derived
24: * from this software without specific prior written permission.
25: *
26: * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
27: * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
28: * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
29: * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
30: * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
31: * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
32: * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
33: * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
34: * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
35: * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36: * POSSIBILITY OF SUCH DAMAGE.
37: */
1.16 cgd 38:
1.1 cgd 39: /*
1.15 mycroft 40: * Copyright (c) 1982, 1986, 1988, 1990, 1993
41: * The Regents of the University of California. All rights reserved.
1.1 cgd 42: *
43: * Redistribution and use in source and binary forms, with or without
44: * modification, are permitted provided that the following conditions
45: * are met:
46: * 1. Redistributions of source code must retain the above copyright
47: * notice, this list of conditions and the following disclaimer.
48: * 2. Redistributions in binary form must reproduce the above copyright
49: * notice, this list of conditions and the following disclaimer in the
50: * documentation and/or other materials provided with the distribution.
51: * 3. All advertising materials mentioning features or use of this software
52: * must display the following acknowledgement:
53: * This product includes software developed by the University of
54: * California, Berkeley and its contributors.
55: * 4. Neither the name of the University nor the names of its contributors
56: * may be used to endorse or promote products derived from this software
57: * without specific prior written permission.
58: *
59: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
60: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
61: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
62: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
63: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
64: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
65: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
66: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
67: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
68: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
69: * SUCH DAMAGE.
70: *
1.32 fvdl 71: * @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95
1.1 cgd 72: */
1.59 lukem 73:
74: #include <sys/cdefs.h>
1.78 ! matt 75: __KERNEL_RCSID(0, "$NetBSD: uipc_socket.c,v 1.77 2003/02/01 06:23:44 thorpej Exp $");
1.64 thorpej 76:
77: #include "opt_sock_counters.h"
78: #include "opt_sosend_loan.h"
1.1 cgd 79:
1.9 mycroft 80: #include <sys/param.h>
81: #include <sys/systm.h>
82: #include <sys/proc.h>
83: #include <sys/file.h>
84: #include <sys/malloc.h>
85: #include <sys/mbuf.h>
86: #include <sys/domain.h>
87: #include <sys/kernel.h>
88: #include <sys/protosw.h>
89: #include <sys/socket.h>
90: #include <sys/socketvar.h>
1.21 christos 91: #include <sys/signalvar.h>
1.9 mycroft 92: #include <sys/resourcevar.h>
1.37 thorpej 93: #include <sys/pool.h>
1.72 jdolecek 94: #include <sys/event.h>
1.37 thorpej 95:
1.64 thorpej 96: #include <uvm/uvm.h>
97:
1.54 lukem 98: struct pool socket_pool;
1.77 thorpej 99:
100: MALLOC_DEFINE(M_SOOPTS, "soopts", "socket options");
101: MALLOC_DEFINE(M_SONAME, "soname", "socket name");
1.37 thorpej 102:
1.54 lukem 103: extern int somaxconn; /* patchable (XXX sysctl) */
104: int somaxconn = SOMAXCONN;
1.49 jonathan 105:
1.64 thorpej 106: #ifdef SOSEND_COUNTERS
107: #include <sys/device.h>
108:
109: struct evcnt sosend_loan_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
110: NULL, "sosend", "loan big");
111: struct evcnt sosend_copy_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
112: NULL, "sosend", "copy big");
113: struct evcnt sosend_copy_small = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
114: NULL, "sosend", "copy small");
115: struct evcnt sosend_kvalimit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
116: NULL, "sosend", "kva limit");
117:
118: #define SOSEND_COUNTER_INCR(ev) (ev)->ev_count++
119:
120: #else
121:
122: #define SOSEND_COUNTER_INCR(ev) /* nothing */
123:
124: #endif /* SOSEND_COUNTERS */
125:
1.37 thorpej 126: void
1.54 lukem 127: soinit(void)
1.37 thorpej 128: {
129:
130: pool_init(&socket_pool, sizeof(struct socket), 0, 0, 0,
1.62 thorpej 131: "sockpl", NULL);
1.64 thorpej 132:
133: #ifdef SOSEND_COUNTERS
134: evcnt_attach_static(&sosend_loan_big);
135: evcnt_attach_static(&sosend_copy_big);
136: evcnt_attach_static(&sosend_copy_small);
137: evcnt_attach_static(&sosend_kvalimit);
138: #endif /* SOSEND_COUNTERS */
1.37 thorpej 139: }
1.1 cgd 140:
1.71 thorpej 141: #ifdef SOSEND_NO_LOAN
142: int use_sosend_loan = 0;
143: #else
1.65 thorpej 144: int use_sosend_loan = 1;
145: #endif
1.64 thorpej 146:
147: struct mbuf *so_pendfree;
148:
149: int somaxkva = 16 * 1024 * 1024;
150: int socurkva;
151: int sokvawaiters;
152:
153: #define SOCK_LOAN_THRESH 4096
154: #define SOCK_LOAN_CHUNK 65536
155:
156: static void
1.76 thorpej 157: sodoloanfree(caddr_t buf, size_t size)
1.64 thorpej 158: {
159: struct vm_page **pgs;
160: vaddr_t va, sva, eva;
161: vsize_t len;
162: paddr_t pa;
163: int i, npgs;
164:
165: eva = round_page((vaddr_t) buf + size);
166: sva = trunc_page((vaddr_t) buf);
167: len = eva - sva;
168: npgs = len >> PAGE_SHIFT;
169:
170: pgs = alloca(npgs * sizeof(*pgs));
171:
172: for (i = 0, va = sva; va < eva; i++, va += PAGE_SIZE) {
173: if (pmap_extract(pmap_kernel(), va, &pa) == FALSE)
174: panic("sodoloanfree: va 0x%lx not mapped", va);
175: pgs[i] = PHYS_TO_VM_PAGE(pa);
176: }
177:
178: pmap_kremove(sva, len);
179: pmap_update(pmap_kernel());
180: uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE);
181: uvm_km_free(kernel_map, sva, len);
182: socurkva -= len;
183: if (sokvawaiters)
184: wakeup(&socurkva);
185: }
186:
187: static size_t
188: sodopendfree(struct socket *so)
189: {
190: struct mbuf *m;
191: size_t rv = 0;
192: int s;
193:
194: s = splvm();
195:
196: for (;;) {
197: m = so_pendfree;
198: if (m == NULL)
199: break;
200: so_pendfree = m->m_next;
201: splx(s);
202:
203: rv += m->m_ext.ext_size;
204: sodoloanfree(m->m_ext.ext_buf, m->m_ext.ext_size);
205: s = splvm();
206: pool_cache_put(&mbpool_cache, m);
207: }
208:
209: for (;;) {
210: m = so->so_pendfree;
211: if (m == NULL)
212: break;
213: so->so_pendfree = m->m_next;
214: splx(s);
215:
216: rv += m->m_ext.ext_size;
217: sodoloanfree(m->m_ext.ext_buf, m->m_ext.ext_size);
218: s = splvm();
219: pool_cache_put(&mbpool_cache, m);
220: }
221:
222: splx(s);
223: return (rv);
224: }
225:
226: static void
1.76 thorpej 227: soloanfree(struct mbuf *m, caddr_t buf, size_t size, void *arg)
1.64 thorpej 228: {
229: struct socket *so = arg;
230: int s;
231:
232: if (m == NULL) {
233: sodoloanfree(buf, size);
234: return;
235: }
236:
237: s = splvm();
238: m->m_next = so->so_pendfree;
239: so->so_pendfree = m;
240: splx(s);
241: if (sokvawaiters)
242: wakeup(&socurkva);
243: }
244:
245: static long
246: sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space)
247: {
248: struct iovec *iov = uio->uio_iov;
249: vaddr_t sva, eva;
250: vsize_t len;
251: struct vm_page **pgs;
252: vaddr_t lva, va;
253: int npgs, s, i, error;
254:
255: if (uio->uio_segflg != UIO_USERSPACE)
256: return (0);
257:
258: if (iov->iov_len < (size_t) space)
259: space = iov->iov_len;
260: if (space > SOCK_LOAN_CHUNK)
261: space = SOCK_LOAN_CHUNK;
262:
263: eva = round_page((vaddr_t) iov->iov_base + space);
264: sva = trunc_page((vaddr_t) iov->iov_base);
265: len = eva - sva;
266: npgs = len >> PAGE_SHIFT;
267:
268: while (socurkva + len > somaxkva) {
269: if (sodopendfree(so))
270: continue;
271: SOSEND_COUNTER_INCR(&sosend_kvalimit);
272: s = splvm();
273: sokvawaiters++;
274: (void) tsleep(&socurkva, PVM, "sokva", 0);
275: sokvawaiters--;
276: splx(s);
277: }
278:
279: lva = uvm_km_valloc_wait(kernel_map, len);
280: if (lva == 0)
281: return (0);
282: socurkva += len;
283:
284: pgs = alloca(npgs * sizeof(*pgs));
285:
286: error = uvm_loan(&uio->uio_procp->p_vmspace->vm_map, sva, len,
287: pgs, UVM_LOAN_TOPAGE);
288: if (error) {
289: uvm_km_free(kernel_map, lva, len);
290: socurkva -= len;
291: return (0);
292: }
293:
294: for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE)
295: pmap_kenter_pa(va, VM_PAGE_TO_PHYS(pgs[i]), VM_PROT_READ);
296: pmap_update(pmap_kernel());
297:
298: lva += (vaddr_t) iov->iov_base & PAGE_MASK;
299:
300: MEXTADD(m, (caddr_t) lva, space, M_MBUF, soloanfree, so);
301:
302: uio->uio_resid -= space;
303: /* uio_offset not updated, not set/used for write(2) */
304: uio->uio_iov->iov_base = (caddr_t) uio->uio_iov->iov_base + space;
305: uio->uio_iov->iov_len -= space;
306: if (uio->uio_iov->iov_len == 0) {
307: uio->uio_iov++;
308: uio->uio_iovcnt--;
309: }
310:
311: return (space);
312: }
313:
1.1 cgd 314: /*
315: * Socket operation routines.
316: * These routines are called by the routines in
317: * sys_socket.c or from a system process, and
318: * implement the semantics of socket operations by
319: * switching out to the protocol specific routines.
320: */
321: /*ARGSUSED*/
1.3 andrew 322: int
1.54 lukem 323: socreate(int dom, struct socket **aso, int type, int proto)
1.1 cgd 324: {
1.54 lukem 325: struct proc *p;
326: struct protosw *prp;
327: struct socket *so;
328: int error, s;
1.1 cgd 329:
1.54 lukem 330: p = curproc; /* XXX */
1.1 cgd 331: if (proto)
332: prp = pffindproto(dom, proto, type);
333: else
334: prp = pffindtype(dom, type);
1.15 mycroft 335: if (prp == 0 || prp->pr_usrreq == 0)
1.1 cgd 336: return (EPROTONOSUPPORT);
337: if (prp->pr_type != type)
338: return (EPROTOTYPE);
1.39 matt 339: s = splsoftnet();
1.37 thorpej 340: so = pool_get(&socket_pool, PR_WAITOK);
1.38 perry 341: memset((caddr_t)so, 0, sizeof(*so));
1.31 thorpej 342: TAILQ_INIT(&so->so_q0);
343: TAILQ_INIT(&so->so_q);
1.1 cgd 344: so->so_type = type;
345: so->so_proto = prp;
1.33 matt 346: so->so_send = sosend;
347: so->so_receive = soreceive;
1.78 ! matt 348: #ifdef MBUFTRACE
! 349: so->so_rcv.sb_mowner = &prp->pr_domain->dom_mowner;
! 350: so->so_snd.sb_mowner = &prp->pr_domain->dom_mowner;
! 351: so->so_mowner = &prp->pr_domain->dom_mowner;
! 352: #endif
1.44 lukem 353: if (p != 0)
354: so->so_uid = p->p_ucred->cr_uid;
1.22 mycroft 355: error = (*prp->pr_usrreq)(so, PRU_ATTACH, (struct mbuf *)0,
356: (struct mbuf *)(long)proto, (struct mbuf *)0, p);
1.1 cgd 357: if (error) {
358: so->so_state |= SS_NOFDREF;
359: sofree(so);
1.39 matt 360: splx(s);
1.1 cgd 361: return (error);
362: }
1.39 matt 363: splx(s);
1.1 cgd 364: *aso = so;
365: return (0);
366: }
367:
1.3 andrew 368: int
1.54 lukem 369: sobind(struct socket *so, struct mbuf *nam, struct proc *p)
1.1 cgd 370: {
1.54 lukem 371: int s, error;
1.1 cgd 372:
1.54 lukem 373: s = splsoftnet();
1.22 mycroft 374: error = (*so->so_proto->pr_usrreq)(so, PRU_BIND, (struct mbuf *)0,
375: nam, (struct mbuf *)0, p);
1.1 cgd 376: splx(s);
377: return (error);
378: }
379:
1.3 andrew 380: int
1.54 lukem 381: solisten(struct socket *so, int backlog)
1.1 cgd 382: {
1.54 lukem 383: int s, error;
1.1 cgd 384:
1.54 lukem 385: s = splsoftnet();
1.22 mycroft 386: error = (*so->so_proto->pr_usrreq)(so, PRU_LISTEN, (struct mbuf *)0,
387: (struct mbuf *)0, (struct mbuf *)0, (struct proc *)0);
1.1 cgd 388: if (error) {
389: splx(s);
390: return (error);
391: }
1.63 matt 392: if (TAILQ_EMPTY(&so->so_q))
1.1 cgd 393: so->so_options |= SO_ACCEPTCONN;
394: if (backlog < 0)
395: backlog = 0;
1.49 jonathan 396: so->so_qlimit = min(backlog, somaxconn);
1.1 cgd 397: splx(s);
398: return (0);
399: }
400:
1.21 christos 401: void
1.54 lukem 402: sofree(struct socket *so)
1.1 cgd 403: {
1.64 thorpej 404: struct mbuf *m;
1.1 cgd 405:
1.43 mycroft 406: if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0)
1.1 cgd 407: return;
1.43 mycroft 408: if (so->so_head) {
409: /*
410: * We must not decommission a socket that's on the accept(2)
411: * queue. If we do, then accept(2) may hang after select(2)
412: * indicated that the listening socket was ready.
413: */
414: if (!soqremque(so, 0))
415: return;
416: }
1.1 cgd 417: sbrelease(&so->so_snd);
418: sorflush(so);
1.64 thorpej 419: while ((m = so->so_pendfree) != NULL) {
420: so->so_pendfree = m->m_next;
421: m->m_next = so_pendfree;
422: so_pendfree = m;
423: }
1.37 thorpej 424: pool_put(&socket_pool, so);
1.1 cgd 425: }
426:
427: /*
428: * Close a socket on last file table reference removal.
429: * Initiate disconnect if connected.
430: * Free socket when disconnect complete.
431: */
1.3 andrew 432: int
1.54 lukem 433: soclose(struct socket *so)
1.1 cgd 434: {
1.54 lukem 435: struct socket *so2;
436: int s, error;
1.1 cgd 437:
1.54 lukem 438: error = 0;
439: s = splsoftnet(); /* conservative */
1.1 cgd 440: if (so->so_options & SO_ACCEPTCONN) {
1.63 matt 441: while ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) {
1.42 mycroft 442: (void) soqremque(so2, 0);
1.41 mycroft 443: (void) soabort(so2);
444: }
1.63 matt 445: while ((so2 = TAILQ_FIRST(&so->so_q)) != 0) {
1.42 mycroft 446: (void) soqremque(so2, 1);
1.41 mycroft 447: (void) soabort(so2);
448: }
1.1 cgd 449: }
450: if (so->so_pcb == 0)
451: goto discard;
452: if (so->so_state & SS_ISCONNECTED) {
453: if ((so->so_state & SS_ISDISCONNECTING) == 0) {
454: error = sodisconnect(so);
455: if (error)
456: goto drop;
457: }
458: if (so->so_options & SO_LINGER) {
459: if ((so->so_state & SS_ISDISCONNECTING) &&
460: (so->so_state & SS_NBIO))
461: goto drop;
1.21 christos 462: while (so->so_state & SS_ISCONNECTED) {
463: error = tsleep((caddr_t)&so->so_timeo,
464: PSOCK | PCATCH, netcls,
1.30 thorpej 465: so->so_linger * hz);
1.21 christos 466: if (error)
1.1 cgd 467: break;
1.21 christos 468: }
1.1 cgd 469: }
470: }
1.54 lukem 471: drop:
1.1 cgd 472: if (so->so_pcb) {
1.22 mycroft 473: int error2 = (*so->so_proto->pr_usrreq)(so, PRU_DETACH,
474: (struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0,
475: (struct proc *)0);
1.1 cgd 476: if (error == 0)
477: error = error2;
478: }
1.54 lukem 479: discard:
1.1 cgd 480: if (so->so_state & SS_NOFDREF)
481: panic("soclose: NOFDREF");
482: so->so_state |= SS_NOFDREF;
483: sofree(so);
484: splx(s);
485: return (error);
486: }
487:
488: /*
1.20 mycroft 489: * Must be called at splsoftnet...
1.1 cgd 490: */
1.3 andrew 491: int
1.54 lukem 492: soabort(struct socket *so)
1.1 cgd 493: {
494:
1.22 mycroft 495: return (*so->so_proto->pr_usrreq)(so, PRU_ABORT, (struct mbuf *)0,
496: (struct mbuf *)0, (struct mbuf *)0, (struct proc *)0);
1.1 cgd 497: }
498:
1.3 andrew 499: int
1.54 lukem 500: soaccept(struct socket *so, struct mbuf *nam)
1.1 cgd 501: {
1.54 lukem 502: int s, error;
1.1 cgd 503:
1.54 lukem 504: error = 0;
505: s = splsoftnet();
1.1 cgd 506: if ((so->so_state & SS_NOFDREF) == 0)
507: panic("soaccept: !NOFDREF");
508: so->so_state &= ~SS_NOFDREF;
1.55 thorpej 509: if ((so->so_state & SS_ISDISCONNECTED) == 0 ||
510: (so->so_proto->pr_flags & PR_ABRTACPTDIS) == 0)
1.41 mycroft 511: error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT,
512: (struct mbuf *)0, nam, (struct mbuf *)0, (struct proc *)0);
513: else
1.53 itojun 514: error = ECONNABORTED;
1.52 itojun 515:
1.1 cgd 516: splx(s);
517: return (error);
518: }
519:
1.3 andrew 520: int
1.54 lukem 521: soconnect(struct socket *so, struct mbuf *nam)
1.1 cgd 522: {
1.54 lukem 523: struct proc *p;
524: int s, error;
1.1 cgd 525:
1.54 lukem 526: p = curproc; /* XXX */
1.1 cgd 527: if (so->so_options & SO_ACCEPTCONN)
528: return (EOPNOTSUPP);
1.20 mycroft 529: s = splsoftnet();
1.1 cgd 530: /*
531: * If protocol is connection-based, can only connect once.
532: * Otherwise, if connected, try to disconnect first.
533: * This allows user to disconnect by connecting to, e.g.,
534: * a null address.
535: */
536: if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
537: ((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
538: (error = sodisconnect(so))))
539: error = EISCONN;
540: else
541: error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
1.23 mycroft 542: (struct mbuf *)0, nam, (struct mbuf *)0, p);
1.1 cgd 543: splx(s);
544: return (error);
545: }
546:
1.3 andrew 547: int
1.54 lukem 548: soconnect2(struct socket *so1, struct socket *so2)
1.1 cgd 549: {
1.54 lukem 550: int s, error;
1.1 cgd 551:
1.54 lukem 552: s = splsoftnet();
1.22 mycroft 553: error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2,
554: (struct mbuf *)0, (struct mbuf *)so2, (struct mbuf *)0,
555: (struct proc *)0);
1.1 cgd 556: splx(s);
557: return (error);
558: }
559:
1.3 andrew 560: int
1.54 lukem 561: sodisconnect(struct socket *so)
1.1 cgd 562: {
1.54 lukem 563: int s, error;
1.1 cgd 564:
1.54 lukem 565: s = splsoftnet();
1.1 cgd 566: if ((so->so_state & SS_ISCONNECTED) == 0) {
567: error = ENOTCONN;
568: goto bad;
569: }
570: if (so->so_state & SS_ISDISCONNECTING) {
571: error = EALREADY;
572: goto bad;
573: }
1.22 mycroft 574: error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
575: (struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0,
576: (struct proc *)0);
1.54 lukem 577: bad:
1.1 cgd 578: splx(s);
1.64 thorpej 579: sodopendfree(so);
1.1 cgd 580: return (error);
581: }
582:
1.15 mycroft 583: #define SBLOCKWAIT(f) (((f) & MSG_DONTWAIT) ? M_NOWAIT : M_WAITOK)
1.1 cgd 584: /*
585: * Send on a socket.
586: * If send must go all at once and message is larger than
587: * send buffering, then hard error.
588: * Lock against other senders.
589: * If must go all at once and not enough room now, then
590: * inform user that this would block and do nothing.
591: * Otherwise, if nonblocking, send as much as possible.
592: * The data to be sent is described by "uio" if nonzero,
593: * otherwise by the mbuf chain "top" (which must be null
594: * if uio is not). Data provided in mbuf chain must be small
595: * enough to send all at once.
596: *
597: * Returns nonzero on error, timeout or signal; callers
598: * must check for short counts if EINTR/ERESTART are returned.
599: * Data and control buffers are freed on return.
600: */
1.3 andrew 601: int
1.54 lukem 602: sosend(struct socket *so, struct mbuf *addr, struct uio *uio, struct mbuf *top,
603: struct mbuf *control, int flags)
1.1 cgd 604: {
1.54 lukem 605: struct proc *p;
606: struct mbuf **mp, *m;
1.58 jdolecek 607: long space, len, resid, clen, mlen;
608: int error, s, dontroute, atomic;
1.54 lukem 609:
1.64 thorpej 610: sodopendfree(so);
611:
1.54 lukem 612: p = curproc; /* XXX */
613: clen = 0;
614: atomic = sosendallatonce(so) || top;
1.1 cgd 615: if (uio)
616: resid = uio->uio_resid;
617: else
618: resid = top->m_pkthdr.len;
1.7 cgd 619: /*
620: * In theory resid should be unsigned.
621: * However, space must be signed, as it might be less than 0
622: * if we over-committed, and we must use a signed comparison
623: * of space and resid. On the other hand, a negative resid
624: * causes us to loop sending 0-length segments to the protocol.
625: */
1.29 mycroft 626: if (resid < 0) {
627: error = EINVAL;
628: goto out;
629: }
1.1 cgd 630: dontroute =
631: (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
632: (so->so_proto->pr_flags & PR_ATOMIC);
1.12 mycroft 633: p->p_stats->p_ru.ru_msgsnd++;
1.1 cgd 634: if (control)
635: clen = control->m_len;
636: #define snderr(errno) { error = errno; splx(s); goto release; }
637:
1.54 lukem 638: restart:
1.21 christos 639: if ((error = sblock(&so->so_snd, SBLOCKWAIT(flags))) != 0)
1.1 cgd 640: goto out;
641: do {
1.20 mycroft 642: s = splsoftnet();
1.1 cgd 643: if (so->so_state & SS_CANTSENDMORE)
644: snderr(EPIPE);
1.48 thorpej 645: if (so->so_error) {
646: error = so->so_error;
647: so->so_error = 0;
648: splx(s);
649: goto release;
650: }
1.1 cgd 651: if ((so->so_state & SS_ISCONNECTED) == 0) {
652: if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
653: if ((so->so_state & SS_ISCONFIRMING) == 0 &&
654: !(resid == 0 && clen != 0))
655: snderr(ENOTCONN);
656: } else if (addr == 0)
657: snderr(EDESTADDRREQ);
658: }
659: space = sbspace(&so->so_snd);
660: if (flags & MSG_OOB)
661: space += 1024;
1.21 christos 662: if ((atomic && resid > so->so_snd.sb_hiwat) ||
1.11 mycroft 663: clen > so->so_snd.sb_hiwat)
664: snderr(EMSGSIZE);
665: if (space < resid + clen && uio &&
1.1 cgd 666: (atomic || space < so->so_snd.sb_lowat || space < clen)) {
667: if (so->so_state & SS_NBIO)
668: snderr(EWOULDBLOCK);
669: sbunlock(&so->so_snd);
670: error = sbwait(&so->so_snd);
671: splx(s);
672: if (error)
673: goto out;
674: goto restart;
675: }
676: splx(s);
677: mp = ⊤
678: space -= clen;
679: do {
1.45 tv 680: if (uio == NULL) {
681: /*
682: * Data is prepackaged in "top".
683: */
684: resid = 0;
685: if (flags & MSG_EOR)
686: top->m_flags |= M_EOR;
687: } else do {
688: if (top == 0) {
1.78 ! matt 689: m = m_gethdr(M_WAIT, MT_DATA);
1.45 tv 690: mlen = MHLEN;
691: m->m_pkthdr.len = 0;
692: m->m_pkthdr.rcvif = (struct ifnet *)0;
693: } else {
1.78 ! matt 694: m = m_get(M_WAIT, MT_DATA);
1.45 tv 695: mlen = MLEN;
696: }
1.78 ! matt 697: MCLAIM(m, so->so_snd.sb_mowner);
1.65 thorpej 698: if (use_sosend_loan &&
699: uio->uio_iov->iov_len >= SOCK_LOAN_THRESH &&
1.64 thorpej 700: space >= SOCK_LOAN_THRESH &&
701: (len = sosend_loan(so, uio, m,
702: space)) != 0) {
703: SOSEND_COUNTER_INCR(&sosend_loan_big);
704: space -= len;
705: goto have_data;
706: }
1.45 tv 707: if (resid >= MINCLSIZE && space >= MCLBYTES) {
1.64 thorpej 708: SOSEND_COUNTER_INCR(&sosend_copy_big);
1.78 ! matt 709: m_clget(m, M_WAIT);
1.45 tv 710: if ((m->m_flags & M_EXT) == 0)
711: goto nopages;
712: mlen = MCLBYTES;
713: if (atomic && top == 0) {
1.58 jdolecek 714: len = lmin(MCLBYTES - max_hdr,
1.54 lukem 715: resid);
1.45 tv 716: m->m_data += max_hdr;
717: } else
1.58 jdolecek 718: len = lmin(MCLBYTES, resid);
1.45 tv 719: space -= len;
720: } else {
1.64 thorpej 721: nopages:
722: SOSEND_COUNTER_INCR(&sosend_copy_small);
1.58 jdolecek 723: len = lmin(lmin(mlen, resid), space);
1.45 tv 724: space -= len;
725: /*
726: * For datagram protocols, leave room
727: * for protocol headers in first mbuf.
728: */
729: if (atomic && top == 0 && len < mlen)
730: MH_ALIGN(m, len);
731: }
1.54 lukem 732: error = uiomove(mtod(m, caddr_t), (int)len,
733: uio);
1.64 thorpej 734: have_data:
1.45 tv 735: resid = uio->uio_resid;
736: m->m_len = len;
737: *mp = m;
738: top->m_pkthdr.len += len;
739: if (error)
740: goto release;
741: mp = &m->m_next;
742: if (resid <= 0) {
743: if (flags & MSG_EOR)
744: top->m_flags |= M_EOR;
745: break;
746: }
747: } while (space > 0 && atomic);
1.46 sommerfe 748:
749: s = splsoftnet();
750:
751: if (so->so_state & SS_CANTSENDMORE)
752: snderr(EPIPE);
1.45 tv 753:
754: if (dontroute)
755: so->so_options |= SO_DONTROUTE;
756: if (resid > 0)
757: so->so_state |= SS_MORETOCOME;
1.46 sommerfe 758: error = (*so->so_proto->pr_usrreq)(so,
759: (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
760: top, addr, control, p);
1.45 tv 761: if (dontroute)
762: so->so_options &= ~SO_DONTROUTE;
763: if (resid > 0)
764: so->so_state &= ~SS_MORETOCOME;
1.46 sommerfe 765: splx(s);
766:
1.45 tv 767: clen = 0;
768: control = 0;
769: top = 0;
770: mp = ⊤
1.1 cgd 771: if (error)
772: goto release;
773: } while (resid && space > 0);
774: } while (resid);
775:
1.54 lukem 776: release:
1.1 cgd 777: sbunlock(&so->so_snd);
1.54 lukem 778: out:
1.1 cgd 779: if (top)
780: m_freem(top);
781: if (control)
782: m_freem(control);
783: return (error);
784: }
785:
786: /*
787: * Implement receive operations on a socket.
788: * We depend on the way that records are added to the sockbuf
789: * by sbappend*. In particular, each record (mbufs linked through m_next)
790: * must begin with an address if the protocol so specifies,
791: * followed by an optional mbuf or mbufs containing ancillary data,
792: * and then zero or more mbufs of data.
793: * In order to avoid blocking network interrupts for the entire time here,
794: * we splx() while doing the actual copy to user space.
795: * Although the sockbuf is locked, new data may still be appended,
796: * and thus we must maintain consistency of the sockbuf during that time.
797: *
798: * The caller may receive the data as a single mbuf chain by supplying
799: * an mbuf **mp0 for use in returning the chain. The uio is then used
800: * only for the count in uio_resid.
801: */
1.3 andrew 802: int
1.54 lukem 803: soreceive(struct socket *so, struct mbuf **paddr, struct uio *uio,
804: struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
1.1 cgd 805: {
1.54 lukem 806: struct mbuf *m, **mp;
807: int flags, len, error, s, offset, moff, type, orig_resid;
808: struct protosw *pr;
809: struct mbuf *nextrecord;
1.67 he 810: int mbuf_removed = 0;
1.64 thorpej 811:
1.54 lukem 812: pr = so->so_proto;
1.1 cgd 813: mp = mp0;
1.54 lukem 814: type = 0;
815: orig_resid = uio->uio_resid;
1.1 cgd 816: if (paddr)
817: *paddr = 0;
818: if (controlp)
819: *controlp = 0;
820: if (flagsp)
821: flags = *flagsp &~ MSG_EOR;
822: else
823: flags = 0;
1.66 enami 824:
825: if ((flags & MSG_DONTWAIT) == 0)
826: sodopendfree(so);
827:
1.1 cgd 828: if (flags & MSG_OOB) {
829: m = m_get(M_WAIT, MT_DATA);
1.17 cgd 830: error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m,
1.22 mycroft 831: (struct mbuf *)(long)(flags & MSG_PEEK), (struct mbuf *)0,
832: (struct proc *)0);
1.1 cgd 833: if (error)
834: goto bad;
835: do {
836: error = uiomove(mtod(m, caddr_t),
837: (int) min(uio->uio_resid, m->m_len), uio);
838: m = m_free(m);
839: } while (uio->uio_resid && error == 0 && m);
1.54 lukem 840: bad:
1.1 cgd 841: if (m)
842: m_freem(m);
843: return (error);
844: }
845: if (mp)
846: *mp = (struct mbuf *)0;
847: if (so->so_state & SS_ISCONFIRMING && uio->uio_resid)
1.22 mycroft 848: (*pr->pr_usrreq)(so, PRU_RCVD, (struct mbuf *)0,
849: (struct mbuf *)0, (struct mbuf *)0, (struct proc *)0);
1.1 cgd 850:
1.54 lukem 851: restart:
1.21 christos 852: if ((error = sblock(&so->so_rcv, SBLOCKWAIT(flags))) != 0)
1.1 cgd 853: return (error);
1.20 mycroft 854: s = splsoftnet();
1.1 cgd 855:
856: m = so->so_rcv.sb_mb;
857: /*
858: * If we have less data than requested, block awaiting more
859: * (subject to any timeout) if:
1.15 mycroft 860: * 1. the current count is less than the low water mark,
1.1 cgd 861: * 2. MSG_WAITALL is set, and it is possible to do the entire
1.15 mycroft 862: * receive operation at once if we block (resid <= hiwat), or
863: * 3. MSG_DONTWAIT is not set.
1.1 cgd 864: * If MSG_WAITALL is set but resid is larger than the receive buffer,
865: * we have to do the receive in sections, and thus risk returning
866: * a short count if a timeout or signal occurs after we start.
867: */
1.21 christos 868: if (m == 0 || (((flags & MSG_DONTWAIT) == 0 &&
1.15 mycroft 869: so->so_rcv.sb_cc < uio->uio_resid) &&
1.1 cgd 870: (so->so_rcv.sb_cc < so->so_rcv.sb_lowat ||
871: ((flags & MSG_WAITALL) && uio->uio_resid <= so->so_rcv.sb_hiwat)) &&
1.21 christos 872: m->m_nextpkt == 0 && (pr->pr_flags & PR_ATOMIC) == 0)) {
1.1 cgd 873: #ifdef DIAGNOSTIC
874: if (m == 0 && so->so_rcv.sb_cc)
875: panic("receive 1");
876: #endif
877: if (so->so_error) {
878: if (m)
1.15 mycroft 879: goto dontblock;
1.1 cgd 880: error = so->so_error;
881: if ((flags & MSG_PEEK) == 0)
882: so->so_error = 0;
883: goto release;
884: }
885: if (so->so_state & SS_CANTRCVMORE) {
886: if (m)
1.15 mycroft 887: goto dontblock;
1.1 cgd 888: else
889: goto release;
890: }
891: for (; m; m = m->m_next)
892: if (m->m_type == MT_OOBDATA || (m->m_flags & M_EOR)) {
893: m = so->so_rcv.sb_mb;
894: goto dontblock;
895: }
896: if ((so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) == 0 &&
897: (so->so_proto->pr_flags & PR_CONNREQUIRED)) {
898: error = ENOTCONN;
899: goto release;
900: }
901: if (uio->uio_resid == 0)
902: goto release;
1.15 mycroft 903: if ((so->so_state & SS_NBIO) || (flags & MSG_DONTWAIT)) {
1.1 cgd 904: error = EWOULDBLOCK;
905: goto release;
906: }
1.69 thorpej 907: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1");
908: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1");
1.1 cgd 909: sbunlock(&so->so_rcv);
910: error = sbwait(&so->so_rcv);
911: splx(s);
912: if (error)
913: return (error);
914: goto restart;
915: }
1.54 lukem 916: dontblock:
1.69 thorpej 917: /*
918: * On entry here, m points to the first record of the socket buffer.
919: * While we process the initial mbufs containing address and control
920: * info, we save a copy of m->m_nextpkt into nextrecord.
921: */
1.15 mycroft 922: #ifdef notyet /* XXXX */
923: if (uio->uio_procp)
924: uio->uio_procp->p_stats->p_ru.ru_msgrcv++;
925: #endif
1.69 thorpej 926: KASSERT(m == so->so_rcv.sb_mb);
927: SBLASTRECORDCHK(&so->so_rcv, "soreceive 1");
928: SBLASTMBUFCHK(&so->so_rcv, "soreceive 1");
1.1 cgd 929: nextrecord = m->m_nextpkt;
930: if (pr->pr_flags & PR_ADDR) {
931: #ifdef DIAGNOSTIC
932: if (m->m_type != MT_SONAME)
933: panic("receive 1a");
934: #endif
1.3 andrew 935: orig_resid = 0;
1.1 cgd 936: if (flags & MSG_PEEK) {
937: if (paddr)
938: *paddr = m_copy(m, 0, m->m_len);
939: m = m->m_next;
940: } else {
941: sbfree(&so->so_rcv, m);
1.67 he 942: mbuf_removed = 1;
1.1 cgd 943: if (paddr) {
944: *paddr = m;
945: so->so_rcv.sb_mb = m->m_next;
946: m->m_next = 0;
947: m = so->so_rcv.sb_mb;
948: } else {
949: MFREE(m, so->so_rcv.sb_mb);
950: m = so->so_rcv.sb_mb;
951: }
952: }
953: }
954: while (m && m->m_type == MT_CONTROL && error == 0) {
955: if (flags & MSG_PEEK) {
956: if (controlp)
957: *controlp = m_copy(m, 0, m->m_len);
958: m = m->m_next;
959: } else {
960: sbfree(&so->so_rcv, m);
1.67 he 961: mbuf_removed = 1;
1.1 cgd 962: if (controlp) {
963: if (pr->pr_domain->dom_externalize &&
964: mtod(m, struct cmsghdr *)->cmsg_type ==
965: SCM_RIGHTS)
1.45 tv 966: error = (*pr->pr_domain->dom_externalize)(m);
1.1 cgd 967: *controlp = m;
968: so->so_rcv.sb_mb = m->m_next;
969: m->m_next = 0;
970: m = so->so_rcv.sb_mb;
971: } else {
972: MFREE(m, so->so_rcv.sb_mb);
973: m = so->so_rcv.sb_mb;
974: }
975: }
1.3 andrew 976: if (controlp) {
977: orig_resid = 0;
1.1 cgd 978: controlp = &(*controlp)->m_next;
1.3 andrew 979: }
1.1 cgd 980: }
1.69 thorpej 981:
982: /*
983: * If m is non-NULL, we have some data to read. From now on,
984: * make sure to keep sb_lastrecord consistent when working on
985: * the last packet on the chain (nextrecord == NULL) and we
986: * change m->m_nextpkt.
987: */
1.1 cgd 988: if (m) {
1.69 thorpej 989: if ((flags & MSG_PEEK) == 0) {
1.1 cgd 990: m->m_nextpkt = nextrecord;
1.69 thorpej 991: /*
992: * If nextrecord == NULL (this is a single chain),
993: * then sb_lastrecord may not be valid here if m
994: * was changed earlier.
995: */
996: if (nextrecord == NULL) {
997: KASSERT(so->so_rcv.sb_mb == m);
998: so->so_rcv.sb_lastrecord = m;
999: }
1000: }
1.1 cgd 1001: type = m->m_type;
1002: if (type == MT_OOBDATA)
1003: flags |= MSG_OOB;
1.69 thorpej 1004: } else {
1005: if ((flags & MSG_PEEK) == 0) {
1006: KASSERT(so->so_rcv.sb_mb == m);
1007: so->so_rcv.sb_mb = nextrecord;
1.70 thorpej 1008: SB_EMPTY_FIXUP(&so->so_rcv);
1.69 thorpej 1009: }
1.1 cgd 1010: }
1.69 thorpej 1011: SBLASTRECORDCHK(&so->so_rcv, "soreceive 2");
1012: SBLASTMBUFCHK(&so->so_rcv, "soreceive 2");
1013:
1.1 cgd 1014: moff = 0;
1015: offset = 0;
1016: while (m && uio->uio_resid > 0 && error == 0) {
1017: if (m->m_type == MT_OOBDATA) {
1018: if (type != MT_OOBDATA)
1019: break;
1020: } else if (type == MT_OOBDATA)
1021: break;
1022: #ifdef DIAGNOSTIC
1023: else if (m->m_type != MT_DATA && m->m_type != MT_HEADER)
1024: panic("receive 3");
1025: #endif
1026: so->so_state &= ~SS_RCVATMARK;
1027: len = uio->uio_resid;
1028: if (so->so_oobmark && len > so->so_oobmark - offset)
1029: len = so->so_oobmark - offset;
1030: if (len > m->m_len - moff)
1031: len = m->m_len - moff;
1032: /*
1033: * If mp is set, just pass back the mbufs.
1034: * Otherwise copy them out via the uio, then free.
1035: * Sockbuf must be consistent here (points to current mbuf,
1036: * it points to next record) when we drop priority;
1037: * we must note any additions to the sockbuf when we
1038: * block interrupts again.
1039: */
1040: if (mp == 0) {
1.69 thorpej 1041: SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove");
1042: SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove");
1.1 cgd 1043: splx(s);
1044: error = uiomove(mtod(m, caddr_t) + moff, (int)len, uio);
1.20 mycroft 1045: s = splsoftnet();
1.67 he 1046: if (error) {
1047: /*
1048: * If any part of the record has been removed
1049: * (such as the MT_SONAME mbuf, which will
1050: * happen when PR_ADDR, and thus also
1051: * PR_ATOMIC, is set), then drop the entire
1052: * record to maintain the atomicity of the
1053: * receive operation.
1054: *
1055: * This avoids a later panic("receive 1a")
1056: * when compiled with DIAGNOSTIC.
1057: */
1058: if (m && mbuf_removed
1059: && (pr->pr_flags & PR_ATOMIC))
1060: (void) sbdroprecord(&so->so_rcv);
1061:
1.57 jdolecek 1062: goto release;
1.67 he 1063: }
1.1 cgd 1064: } else
1065: uio->uio_resid -= len;
1066: if (len == m->m_len - moff) {
1067: if (m->m_flags & M_EOR)
1068: flags |= MSG_EOR;
1069: if (flags & MSG_PEEK) {
1070: m = m->m_next;
1071: moff = 0;
1072: } else {
1073: nextrecord = m->m_nextpkt;
1074: sbfree(&so->so_rcv, m);
1075: if (mp) {
1076: *mp = m;
1077: mp = &m->m_next;
1078: so->so_rcv.sb_mb = m = m->m_next;
1079: *mp = (struct mbuf *)0;
1080: } else {
1081: MFREE(m, so->so_rcv.sb_mb);
1082: m = so->so_rcv.sb_mb;
1083: }
1.69 thorpej 1084: /*
1085: * If m != NULL, we also know that
1086: * so->so_rcv.sb_mb != NULL.
1087: */
1088: KASSERT(so->so_rcv.sb_mb == m);
1089: if (m) {
1.1 cgd 1090: m->m_nextpkt = nextrecord;
1.69 thorpej 1091: if (nextrecord == NULL)
1092: so->so_rcv.sb_lastrecord = m;
1093: } else {
1094: so->so_rcv.sb_mb = nextrecord;
1.70 thorpej 1095: SB_EMPTY_FIXUP(&so->so_rcv);
1.69 thorpej 1096: }
1097: SBLASTRECORDCHK(&so->so_rcv, "soreceive 3");
1098: SBLASTMBUFCHK(&so->so_rcv, "soreceive 3");
1.1 cgd 1099: }
1100: } else {
1101: if (flags & MSG_PEEK)
1102: moff += len;
1103: else {
1104: if (mp)
1105: *mp = m_copym(m, 0, len, M_WAIT);
1106: m->m_data += len;
1107: m->m_len -= len;
1108: so->so_rcv.sb_cc -= len;
1109: }
1110: }
1111: if (so->so_oobmark) {
1112: if ((flags & MSG_PEEK) == 0) {
1113: so->so_oobmark -= len;
1114: if (so->so_oobmark == 0) {
1115: so->so_state |= SS_RCVATMARK;
1116: break;
1117: }
1.7 cgd 1118: } else {
1.1 cgd 1119: offset += len;
1.7 cgd 1120: if (offset == so->so_oobmark)
1121: break;
1122: }
1.1 cgd 1123: }
1124: if (flags & MSG_EOR)
1125: break;
1126: /*
1127: * If the MSG_WAITALL flag is set (for non-atomic socket),
1128: * we must not quit until "uio->uio_resid == 0" or an error
1129: * termination. If a signal/timeout occurs, return
1130: * with a short count but without error.
1131: * Keep sockbuf locked against other readers.
1132: */
1133: while (flags & MSG_WAITALL && m == 0 && uio->uio_resid > 0 &&
1.3 andrew 1134: !sosendallatonce(so) && !nextrecord) {
1.1 cgd 1135: if (so->so_error || so->so_state & SS_CANTRCVMORE)
1136: break;
1.68 matt 1137: /*
1138: * If we are peeking and the socket receive buffer is
1139: * full, stop since we can't get more data to peek at.
1140: */
1141: if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0)
1142: break;
1143: /*
1144: * If we've drained the socket buffer, tell the
1145: * protocol in case it needs to do something to
1146: * get it filled again.
1147: */
1148: if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb)
1149: (*pr->pr_usrreq)(so, PRU_RCVD,
1150: (struct mbuf *)0,
1151: (struct mbuf *)(long)flags,
1152: (struct mbuf *)0,
1153: (struct proc *)0);
1.69 thorpej 1154: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2");
1155: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2");
1.1 cgd 1156: error = sbwait(&so->so_rcv);
1157: if (error) {
1158: sbunlock(&so->so_rcv);
1159: splx(s);
1160: return (0);
1161: }
1.21 christos 1162: if ((m = so->so_rcv.sb_mb) != NULL)
1.1 cgd 1163: nextrecord = m->m_nextpkt;
1164: }
1165: }
1.3 andrew 1166:
1167: if (m && pr->pr_flags & PR_ATOMIC) {
1168: flags |= MSG_TRUNC;
1169: if ((flags & MSG_PEEK) == 0)
1170: (void) sbdroprecord(&so->so_rcv);
1171: }
1.1 cgd 1172: if ((flags & MSG_PEEK) == 0) {
1.69 thorpej 1173: if (m == 0) {
1174: /*
1.70 thorpej 1175: * First part is an inline SB_EMPTY_FIXUP(). Second
1.69 thorpej 1176: * part makes sure sb_lastrecord is up-to-date if
1177: * there is still data in the socket buffer.
1178: */
1.1 cgd 1179: so->so_rcv.sb_mb = nextrecord;
1.69 thorpej 1180: if (so->so_rcv.sb_mb == NULL) {
1181: so->so_rcv.sb_mbtail = NULL;
1182: so->so_rcv.sb_lastrecord = NULL;
1183: } else if (nextrecord->m_nextpkt == NULL)
1184: so->so_rcv.sb_lastrecord = nextrecord;
1185: }
1186: SBLASTRECORDCHK(&so->so_rcv, "soreceive 4");
1187: SBLASTMBUFCHK(&so->so_rcv, "soreceive 4");
1.1 cgd 1188: if (pr->pr_flags & PR_WANTRCVD && so->so_pcb)
1.22 mycroft 1189: (*pr->pr_usrreq)(so, PRU_RCVD, (struct mbuf *)0,
1190: (struct mbuf *)(long)flags, (struct mbuf *)0,
1191: (struct proc *)0);
1.1 cgd 1192: }
1.3 andrew 1193: if (orig_resid == uio->uio_resid && orig_resid &&
1194: (flags & MSG_EOR) == 0 && (so->so_state & SS_CANTRCVMORE) == 0) {
1195: sbunlock(&so->so_rcv);
1196: splx(s);
1197: goto restart;
1198: }
1199:
1.1 cgd 1200: if (flagsp)
1201: *flagsp |= flags;
1.54 lukem 1202: release:
1.1 cgd 1203: sbunlock(&so->so_rcv);
1204: splx(s);
1205: return (error);
1206: }
1207:
1.14 mycroft 1208: int
1.54 lukem 1209: soshutdown(struct socket *so, int how)
1.1 cgd 1210: {
1.54 lukem 1211: struct protosw *pr;
1.34 kleink 1212:
1.54 lukem 1213: pr = so->so_proto;
1.34 kleink 1214: if (!(how == SHUT_RD || how == SHUT_WR || how == SHUT_RDWR))
1215: return (EINVAL);
1.1 cgd 1216:
1.34 kleink 1217: if (how == SHUT_RD || how == SHUT_RDWR)
1.1 cgd 1218: sorflush(so);
1.34 kleink 1219: if (how == SHUT_WR || how == SHUT_RDWR)
1.22 mycroft 1220: return (*pr->pr_usrreq)(so, PRU_SHUTDOWN, (struct mbuf *)0,
1221: (struct mbuf *)0, (struct mbuf *)0, (struct proc *)0);
1.1 cgd 1222: return (0);
1223: }
1224:
1.14 mycroft 1225: void
1.54 lukem 1226: sorflush(struct socket *so)
1.1 cgd 1227: {
1.54 lukem 1228: struct sockbuf *sb, asb;
1229: struct protosw *pr;
1230: int s;
1.1 cgd 1231:
1.54 lukem 1232: sb = &so->so_rcv;
1233: pr = so->so_proto;
1.1 cgd 1234: sb->sb_flags |= SB_NOINTR;
1.15 mycroft 1235: (void) sblock(sb, M_WAITOK);
1.56 thorpej 1236: s = splnet();
1.1 cgd 1237: socantrcvmore(so);
1238: sbunlock(sb);
1239: asb = *sb;
1.38 perry 1240: memset((caddr_t)sb, 0, sizeof(*sb));
1.1 cgd 1241: splx(s);
1242: if (pr->pr_flags & PR_RIGHTS && pr->pr_domain->dom_dispose)
1243: (*pr->pr_domain->dom_dispose)(asb.sb_mb);
1244: sbrelease(&asb);
1245: }
1246:
1.14 mycroft 1247: int
1.54 lukem 1248: sosetopt(struct socket *so, int level, int optname, struct mbuf *m0)
1.1 cgd 1249: {
1.54 lukem 1250: int error;
1251: struct mbuf *m;
1.1 cgd 1252:
1.54 lukem 1253: error = 0;
1254: m = m0;
1.1 cgd 1255: if (level != SOL_SOCKET) {
1256: if (so->so_proto && so->so_proto->pr_ctloutput)
1257: return ((*so->so_proto->pr_ctloutput)
1258: (PRCO_SETOPT, so, level, optname, &m0));
1259: error = ENOPROTOOPT;
1260: } else {
1261: switch (optname) {
1262:
1263: case SO_LINGER:
1.36 perry 1264: if (m == NULL || m->m_len != sizeof(struct linger)) {
1.1 cgd 1265: error = EINVAL;
1266: goto bad;
1267: }
1268: so->so_linger = mtod(m, struct linger *)->l_linger;
1269: /* fall thru... */
1270:
1271: case SO_DEBUG:
1272: case SO_KEEPALIVE:
1273: case SO_DONTROUTE:
1274: case SO_USELOOPBACK:
1275: case SO_BROADCAST:
1276: case SO_REUSEADDR:
1.15 mycroft 1277: case SO_REUSEPORT:
1.1 cgd 1278: case SO_OOBINLINE:
1.26 thorpej 1279: case SO_TIMESTAMP:
1.36 perry 1280: if (m == NULL || m->m_len < sizeof(int)) {
1.1 cgd 1281: error = EINVAL;
1282: goto bad;
1283: }
1284: if (*mtod(m, int *))
1285: so->so_options |= optname;
1286: else
1287: so->so_options &= ~optname;
1288: break;
1289:
1290: case SO_SNDBUF:
1291: case SO_RCVBUF:
1292: case SO_SNDLOWAT:
1293: case SO_RCVLOWAT:
1.28 thorpej 1294: {
1295: int optval;
1296:
1.36 perry 1297: if (m == NULL || m->m_len < sizeof(int)) {
1.1 cgd 1298: error = EINVAL;
1299: goto bad;
1300: }
1.28 thorpej 1301:
1302: /*
1303: * Values < 1 make no sense for any of these
1304: * options, so disallow them.
1305: */
1306: optval = *mtod(m, int *);
1307: if (optval < 1) {
1308: error = EINVAL;
1309: goto bad;
1310: }
1311:
1.1 cgd 1312: switch (optname) {
1313:
1314: case SO_SNDBUF:
1315: case SO_RCVBUF:
1316: if (sbreserve(optname == SO_SNDBUF ?
1317: &so->so_snd : &so->so_rcv,
1.28 thorpej 1318: (u_long) optval) == 0) {
1.1 cgd 1319: error = ENOBUFS;
1320: goto bad;
1321: }
1322: break;
1323:
1.28 thorpej 1324: /*
1325: * Make sure the low-water is never greater than
1326: * the high-water.
1327: */
1.1 cgd 1328: case SO_SNDLOWAT:
1.28 thorpej 1329: so->so_snd.sb_lowat =
1330: (optval > so->so_snd.sb_hiwat) ?
1331: so->so_snd.sb_hiwat : optval;
1.1 cgd 1332: break;
1333: case SO_RCVLOWAT:
1.28 thorpej 1334: so->so_rcv.sb_lowat =
1335: (optval > so->so_rcv.sb_hiwat) ?
1336: so->so_rcv.sb_hiwat : optval;
1.1 cgd 1337: break;
1338: }
1339: break;
1.28 thorpej 1340: }
1.1 cgd 1341:
1342: case SO_SNDTIMEO:
1343: case SO_RCVTIMEO:
1344: {
1345: struct timeval *tv;
1346: short val;
1347:
1.36 perry 1348: if (m == NULL || m->m_len < sizeof(*tv)) {
1.1 cgd 1349: error = EINVAL;
1350: goto bad;
1351: }
1352: tv = mtod(m, struct timeval *);
1.75 itojun 1353: if (tv->tv_sec > (SHRT_MAX - tv->tv_usec / tick) / hz) {
1.1 cgd 1354: error = EDOM;
1355: goto bad;
1356: }
1357: val = tv->tv_sec * hz + tv->tv_usec / tick;
1.74 itojun 1358: if (val == 0 && tv->tv_usec != 0)
1359: val = 1;
1.1 cgd 1360:
1361: switch (optname) {
1362:
1363: case SO_SNDTIMEO:
1364: so->so_snd.sb_timeo = val;
1365: break;
1366: case SO_RCVTIMEO:
1367: so->so_rcv.sb_timeo = val;
1368: break;
1369: }
1370: break;
1371: }
1372:
1373: default:
1374: error = ENOPROTOOPT;
1375: break;
1376: }
1.15 mycroft 1377: if (error == 0 && so->so_proto && so->so_proto->pr_ctloutput) {
1378: (void) ((*so->so_proto->pr_ctloutput)
1379: (PRCO_SETOPT, so, level, optname, &m0));
1380: m = NULL; /* freed by protocol */
1381: }
1.1 cgd 1382: }
1.54 lukem 1383: bad:
1.1 cgd 1384: if (m)
1385: (void) m_free(m);
1386: return (error);
1387: }
1388:
1.14 mycroft 1389: int
1.54 lukem 1390: sogetopt(struct socket *so, int level, int optname, struct mbuf **mp)
1.1 cgd 1391: {
1.54 lukem 1392: struct mbuf *m;
1.1 cgd 1393:
1394: if (level != SOL_SOCKET) {
1395: if (so->so_proto && so->so_proto->pr_ctloutput) {
1396: return ((*so->so_proto->pr_ctloutput)
1397: (PRCO_GETOPT, so, level, optname, mp));
1398: } else
1399: return (ENOPROTOOPT);
1400: } else {
1401: m = m_get(M_WAIT, MT_SOOPTS);
1.36 perry 1402: m->m_len = sizeof(int);
1.1 cgd 1403:
1404: switch (optname) {
1405:
1406: case SO_LINGER:
1.36 perry 1407: m->m_len = sizeof(struct linger);
1.1 cgd 1408: mtod(m, struct linger *)->l_onoff =
1409: so->so_options & SO_LINGER;
1410: mtod(m, struct linger *)->l_linger = so->so_linger;
1411: break;
1412:
1413: case SO_USELOOPBACK:
1414: case SO_DONTROUTE:
1415: case SO_DEBUG:
1416: case SO_KEEPALIVE:
1417: case SO_REUSEADDR:
1.15 mycroft 1418: case SO_REUSEPORT:
1.1 cgd 1419: case SO_BROADCAST:
1420: case SO_OOBINLINE:
1.26 thorpej 1421: case SO_TIMESTAMP:
1.1 cgd 1422: *mtod(m, int *) = so->so_options & optname;
1423: break;
1424:
1425: case SO_TYPE:
1426: *mtod(m, int *) = so->so_type;
1427: break;
1428:
1429: case SO_ERROR:
1430: *mtod(m, int *) = so->so_error;
1431: so->so_error = 0;
1432: break;
1433:
1434: case SO_SNDBUF:
1435: *mtod(m, int *) = so->so_snd.sb_hiwat;
1436: break;
1437:
1438: case SO_RCVBUF:
1439: *mtod(m, int *) = so->so_rcv.sb_hiwat;
1440: break;
1441:
1442: case SO_SNDLOWAT:
1443: *mtod(m, int *) = so->so_snd.sb_lowat;
1444: break;
1445:
1446: case SO_RCVLOWAT:
1447: *mtod(m, int *) = so->so_rcv.sb_lowat;
1448: break;
1449:
1450: case SO_SNDTIMEO:
1451: case SO_RCVTIMEO:
1452: {
1453: int val = (optname == SO_SNDTIMEO ?
1454: so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
1455:
1456: m->m_len = sizeof(struct timeval);
1457: mtod(m, struct timeval *)->tv_sec = val / hz;
1458: mtod(m, struct timeval *)->tv_usec =
1.27 kleink 1459: (val % hz) * tick;
1.1 cgd 1460: break;
1461: }
1462:
1463: default:
1464: (void)m_free(m);
1465: return (ENOPROTOOPT);
1466: }
1467: *mp = m;
1468: return (0);
1469: }
1470: }
1471:
1.14 mycroft 1472: void
1.54 lukem 1473: sohasoutofband(struct socket *so)
1.1 cgd 1474: {
1475: struct proc *p;
1476:
1477: if (so->so_pgid < 0)
1478: gsignal(-so->so_pgid, SIGURG);
1479: else if (so->so_pgid > 0 && (p = pfind(so->so_pgid)) != 0)
1480: psignal(p, SIGURG);
1.2 cgd 1481: selwakeup(&so->so_rcv.sb_sel);
1.1 cgd 1482: }
1.72 jdolecek 1483:
1484: static void
1485: filt_sordetach(struct knote *kn)
1486: {
1487: struct socket *so;
1488:
1489: so = (struct socket *)kn->kn_fp->f_data;
1.73 christos 1490: SLIST_REMOVE(&so->so_rcv.sb_sel.sel_klist, kn, knote, kn_selnext);
1491: if (SLIST_EMPTY(&so->so_rcv.sb_sel.sel_klist))
1.72 jdolecek 1492: so->so_rcv.sb_flags &= ~SB_KNOTE;
1493: }
1494:
1495: /*ARGSUSED*/
1496: static int
1497: filt_soread(struct knote *kn, long hint)
1498: {
1499: struct socket *so;
1500:
1501: so = (struct socket *)kn->kn_fp->f_data;
1502: kn->kn_data = so->so_rcv.sb_cc;
1503: if (so->so_state & SS_CANTRCVMORE) {
1504: kn->kn_flags |= EV_EOF;
1505: kn->kn_fflags = so->so_error;
1506: return (1);
1507: }
1508: if (so->so_error) /* temporary udp error */
1509: return (1);
1510: if (kn->kn_sfflags & NOTE_LOWAT)
1511: return (kn->kn_data >= kn->kn_sdata);
1512: return (kn->kn_data >= so->so_rcv.sb_lowat);
1513: }
1514:
1515: static void
1516: filt_sowdetach(struct knote *kn)
1517: {
1518: struct socket *so;
1519:
1520: so = (struct socket *)kn->kn_fp->f_data;
1.73 christos 1521: SLIST_REMOVE(&so->so_snd.sb_sel.sel_klist, kn, knote, kn_selnext);
1522: if (SLIST_EMPTY(&so->so_snd.sb_sel.sel_klist))
1.72 jdolecek 1523: so->so_snd.sb_flags &= ~SB_KNOTE;
1524: }
1525:
1526: /*ARGSUSED*/
1527: static int
1528: filt_sowrite(struct knote *kn, long hint)
1529: {
1530: struct socket *so;
1531:
1532: so = (struct socket *)kn->kn_fp->f_data;
1533: kn->kn_data = sbspace(&so->so_snd);
1534: if (so->so_state & SS_CANTSENDMORE) {
1535: kn->kn_flags |= EV_EOF;
1536: kn->kn_fflags = so->so_error;
1537: return (1);
1538: }
1539: if (so->so_error) /* temporary udp error */
1540: return (1);
1541: if (((so->so_state & SS_ISCONNECTED) == 0) &&
1542: (so->so_proto->pr_flags & PR_CONNREQUIRED))
1543: return (0);
1544: if (kn->kn_sfflags & NOTE_LOWAT)
1545: return (kn->kn_data >= kn->kn_sdata);
1546: return (kn->kn_data >= so->so_snd.sb_lowat);
1547: }
1548:
1549: /*ARGSUSED*/
1550: static int
1551: filt_solisten(struct knote *kn, long hint)
1552: {
1553: struct socket *so;
1554:
1555: so = (struct socket *)kn->kn_fp->f_data;
1556:
1557: /*
1558: * Set kn_data to number of incoming connections, not
1559: * counting partial (incomplete) connections.
1560: */
1561: kn->kn_data = so->so_qlen;
1562: return (kn->kn_data > 0);
1563: }
1564:
1565: static const struct filterops solisten_filtops =
1566: { 1, NULL, filt_sordetach, filt_solisten };
1567: static const struct filterops soread_filtops =
1568: { 1, NULL, filt_sordetach, filt_soread };
1569: static const struct filterops sowrite_filtops =
1570: { 1, NULL, filt_sowdetach, filt_sowrite };
1571:
1572: int
1573: soo_kqfilter(struct file *fp, struct knote *kn)
1574: {
1575: struct socket *so;
1576: struct sockbuf *sb;
1577:
1578: so = (struct socket *)kn->kn_fp->f_data;
1579: switch (kn->kn_filter) {
1580: case EVFILT_READ:
1581: if (so->so_options & SO_ACCEPTCONN)
1582: kn->kn_fop = &solisten_filtops;
1583: else
1584: kn->kn_fop = &soread_filtops;
1585: sb = &so->so_rcv;
1586: break;
1587: case EVFILT_WRITE:
1588: kn->kn_fop = &sowrite_filtops;
1589: sb = &so->so_snd;
1590: break;
1591: default:
1592: return (1);
1593: }
1.73 christos 1594: SLIST_INSERT_HEAD(&sb->sb_sel.sel_klist, kn, kn_selnext);
1.72 jdolecek 1595: sb->sb_flags |= SB_KNOTE;
1596: return (0);
1597: }
1598:
CVSweb <webmaster@jp.NetBSD.org>
![[BACK]](/icons/back.gif){kind=icon}
Return to uipc_socket.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [cvs.NetBSD.org] / src / sys / kern