Annotation of src/sys/kern/uipc_socket.c, Revision 1.156
1.156 ! yamt 1: /* $NetBSD: uipc_socket.c,v 1.155 2008/03/21 21:55:00 ad Exp $ */
1.64 thorpej 2:
3: /*-
1.154 ad 4: * Copyright (c) 2002, 2007, 2008 The NetBSD Foundation, Inc.
1.64 thorpej 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.
1.85 agc 51: * 3. Neither the name of the University nor the names of its contributors
1.1 cgd 52: * may be used to endorse or promote products derived from this software
53: * without specific prior written permission.
54: *
55: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
56: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
57: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
58: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
59: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
60: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
61: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
62: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
63: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
64: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
65: * SUCH DAMAGE.
66: *
1.32 fvdl 67: * @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95
1.1 cgd 68: */
1.59 lukem 69:
70: #include <sys/cdefs.h>
1.156 ! yamt 71: __KERNEL_RCSID(0, "$NetBSD: uipc_socket.c,v 1.155 2008/03/21 21:55:00 ad Exp $");
1.64 thorpej 72:
73: #include "opt_sock_counters.h"
74: #include "opt_sosend_loan.h"
1.81 martin 75: #include "opt_mbuftrace.h"
1.84 ragge 76: #include "opt_somaxkva.h"
1.1 cgd 77:
1.9 mycroft 78: #include <sys/param.h>
79: #include <sys/systm.h>
80: #include <sys/proc.h>
81: #include <sys/file.h>
1.142 dyoung 82: #include <sys/filedesc.h>
1.9 mycroft 83: #include <sys/malloc.h>
84: #include <sys/mbuf.h>
85: #include <sys/domain.h>
86: #include <sys/kernel.h>
87: #include <sys/protosw.h>
88: #include <sys/socket.h>
89: #include <sys/socketvar.h>
1.21 christos 90: #include <sys/signalvar.h>
1.9 mycroft 91: #include <sys/resourcevar.h>
1.37 thorpej 92: #include <sys/pool.h>
1.72 jdolecek 93: #include <sys/event.h>
1.89 christos 94: #include <sys/poll.h>
1.118 elad 95: #include <sys/kauth.h>
1.136 ad 96: #include <sys/mutex.h>
97: #include <sys/condvar.h>
1.37 thorpej 98:
1.64 thorpej 99: #include <uvm/uvm.h>
100:
1.135 ad 101: POOL_INIT(socket_pool, sizeof(struct socket), 0, 0, 0, "sockpl", NULL,
102: IPL_SOFTNET);
1.77 thorpej 103:
104: MALLOC_DEFINE(M_SOOPTS, "soopts", "socket options");
105: MALLOC_DEFINE(M_SONAME, "soname", "socket name");
1.37 thorpej 106:
1.142 dyoung 107: extern const struct fileops socketops;
108:
1.54 lukem 109: extern int somaxconn; /* patchable (XXX sysctl) */
110: int somaxconn = SOMAXCONN;
1.49 jonathan 111:
1.64 thorpej 112: #ifdef SOSEND_COUNTERS
113: #include <sys/device.h>
114:
1.113 thorpej 115: static struct evcnt sosend_loan_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 116: NULL, "sosend", "loan big");
1.113 thorpej 117: static struct evcnt sosend_copy_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 118: NULL, "sosend", "copy big");
1.113 thorpej 119: static struct evcnt sosend_copy_small = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 120: NULL, "sosend", "copy small");
1.113 thorpej 121: static struct evcnt sosend_kvalimit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 122: NULL, "sosend", "kva limit");
123:
124: #define SOSEND_COUNTER_INCR(ev) (ev)->ev_count++
125:
1.101 matt 126: EVCNT_ATTACH_STATIC(sosend_loan_big);
127: EVCNT_ATTACH_STATIC(sosend_copy_big);
128: EVCNT_ATTACH_STATIC(sosend_copy_small);
129: EVCNT_ATTACH_STATIC(sosend_kvalimit);
1.64 thorpej 130: #else
131:
132: #define SOSEND_COUNTER_INCR(ev) /* nothing */
133:
134: #endif /* SOSEND_COUNTERS */
135:
1.119 yamt 136: static struct callback_entry sokva_reclaimerentry;
1.1 cgd 137:
1.71 thorpej 138: #ifdef SOSEND_NO_LOAN
1.121 yamt 139: int sock_loan_thresh = -1;
1.71 thorpej 140: #else
1.121 yamt 141: int sock_loan_thresh = 4096;
1.65 thorpej 142: #endif
1.64 thorpej 143:
1.136 ad 144: static kmutex_t so_pendfree_lock;
1.113 thorpej 145: static struct mbuf *so_pendfree;
1.64 thorpej 146:
1.84 ragge 147: #ifndef SOMAXKVA
148: #define SOMAXKVA (16 * 1024 * 1024)
149: #endif
150: int somaxkva = SOMAXKVA;
1.113 thorpej 151: static int socurkva;
1.136 ad 152: static kcondvar_t socurkva_cv;
1.64 thorpej 153:
154: #define SOCK_LOAN_CHUNK 65536
155:
1.117 yamt 156: static size_t sodopendfree(void);
157: static size_t sodopendfreel(void);
1.93 yamt 158:
1.113 thorpej 159: static vsize_t
1.129 yamt 160: sokvareserve(struct socket *so, vsize_t len)
1.80 yamt 161: {
1.98 christos 162: int error;
1.80 yamt 163:
1.136 ad 164: mutex_enter(&so_pendfree_lock);
1.80 yamt 165: while (socurkva + len > somaxkva) {
1.93 yamt 166: size_t freed;
167:
168: /*
169: * try to do pendfree.
170: */
171:
1.117 yamt 172: freed = sodopendfreel();
1.93 yamt 173:
174: /*
175: * if some kva was freed, try again.
176: */
177:
178: if (freed)
1.80 yamt 179: continue;
1.93 yamt 180:
1.80 yamt 181: SOSEND_COUNTER_INCR(&sosend_kvalimit);
1.136 ad 182: error = cv_wait_sig(&socurkva_cv, &so_pendfree_lock);
1.98 christos 183: if (error) {
184: len = 0;
185: break;
186: }
1.80 yamt 187: }
1.93 yamt 188: socurkva += len;
1.136 ad 189: mutex_exit(&so_pendfree_lock);
1.98 christos 190: return len;
1.95 yamt 191: }
192:
1.113 thorpej 193: static void
1.95 yamt 194: sokvaunreserve(vsize_t len)
195: {
196:
1.136 ad 197: mutex_enter(&so_pendfree_lock);
1.95 yamt 198: socurkva -= len;
1.136 ad 199: cv_broadcast(&socurkva_cv);
200: mutex_exit(&so_pendfree_lock);
1.95 yamt 201: }
202:
203: /*
204: * sokvaalloc: allocate kva for loan.
205: */
206:
207: vaddr_t
208: sokvaalloc(vsize_t len, struct socket *so)
209: {
210: vaddr_t lva;
211:
212: /*
213: * reserve kva.
214: */
215:
1.98 christos 216: if (sokvareserve(so, len) == 0)
217: return 0;
1.93 yamt 218:
219: /*
220: * allocate kva.
221: */
1.80 yamt 222:
1.109 yamt 223: lva = uvm_km_alloc(kernel_map, len, 0, UVM_KMF_VAONLY | UVM_KMF_WAITVA);
1.95 yamt 224: if (lva == 0) {
225: sokvaunreserve(len);
1.80 yamt 226: return (0);
1.95 yamt 227: }
1.80 yamt 228:
229: return lva;
230: }
231:
1.93 yamt 232: /*
233: * sokvafree: free kva for loan.
234: */
235:
1.80 yamt 236: void
237: sokvafree(vaddr_t sva, vsize_t len)
238: {
1.93 yamt 239:
240: /*
241: * free kva.
242: */
1.80 yamt 243:
1.109 yamt 244: uvm_km_free(kernel_map, sva, len, UVM_KMF_VAONLY);
1.93 yamt 245:
246: /*
247: * unreserve kva.
248: */
249:
1.95 yamt 250: sokvaunreserve(len);
1.80 yamt 251: }
252:
1.64 thorpej 253: static void
1.134 christos 254: sodoloanfree(struct vm_page **pgs, void *buf, size_t size)
1.64 thorpej 255: {
1.156 ! yamt 256: vaddr_t sva, eva;
1.64 thorpej 257: vsize_t len;
1.156 ! yamt 258: int npgs;
! 259:
! 260: KASSERT(pgs != NULL);
1.64 thorpej 261:
262: eva = round_page((vaddr_t) buf + size);
263: sva = trunc_page((vaddr_t) buf);
264: len = eva - sva;
265: npgs = len >> PAGE_SHIFT;
266:
267: pmap_kremove(sva, len);
268: pmap_update(pmap_kernel());
269: uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE);
1.80 yamt 270: sokvafree(sva, len);
1.64 thorpej 271: }
272:
273: static size_t
1.152 matt 274: sodopendfree(void)
1.64 thorpej 275: {
1.93 yamt 276: size_t rv;
1.64 thorpej 277:
1.136 ad 278: mutex_enter(&so_pendfree_lock);
1.117 yamt 279: rv = sodopendfreel();
1.136 ad 280: mutex_exit(&so_pendfree_lock);
1.93 yamt 281:
282: return rv;
283: }
284:
285: /*
286: * sodopendfreel: free mbufs on "pendfree" list.
1.136 ad 287: * unlock and relock so_pendfree_lock when freeing mbufs.
1.93 yamt 288: *
1.136 ad 289: * => called with so_pendfree_lock held.
1.93 yamt 290: */
291:
292: static size_t
1.152 matt 293: sodopendfreel(void)
1.93 yamt 294: {
1.137 ad 295: struct mbuf *m, *next;
1.93 yamt 296: size_t rv = 0;
297:
1.136 ad 298: KASSERT(mutex_owned(&so_pendfree_lock));
1.64 thorpej 299:
1.137 ad 300: while (so_pendfree != NULL) {
1.64 thorpej 301: m = so_pendfree;
1.93 yamt 302: so_pendfree = NULL;
1.136 ad 303: mutex_exit(&so_pendfree_lock);
1.93 yamt 304:
305: for (; m != NULL; m = next) {
306: next = m->m_next;
1.156 ! yamt 307: KASSERT((~m->m_flags & (M_EXT|M_EXT_PAGES)) == 0);
! 308: KASSERT(m->m_ext.ext_refcnt == 0);
1.93 yamt 309:
310: rv += m->m_ext.ext_size;
1.156 ! yamt 311: sodoloanfree(m->m_ext.ext_pgs, m->m_ext.ext_buf,
1.93 yamt 312: m->m_ext.ext_size);
1.145 ad 313: pool_cache_put(mb_cache, m);
1.93 yamt 314: }
1.64 thorpej 315:
1.136 ad 316: mutex_enter(&so_pendfree_lock);
1.64 thorpej 317: }
318:
319: return (rv);
320: }
321:
1.80 yamt 322: void
1.134 christos 323: soloanfree(struct mbuf *m, void *buf, size_t size, void *arg)
1.64 thorpej 324: {
325:
1.156 ! yamt 326: KASSERT(m != NULL);
1.64 thorpej 327:
1.93 yamt 328: /*
329: * postpone freeing mbuf.
330: *
331: * we can't do it in interrupt context
332: * because we need to put kva back to kernel_map.
333: */
334:
1.136 ad 335: mutex_enter(&so_pendfree_lock);
1.92 yamt 336: m->m_next = so_pendfree;
337: so_pendfree = m;
1.136 ad 338: cv_broadcast(&socurkva_cv);
339: mutex_exit(&so_pendfree_lock);
1.64 thorpej 340: }
341:
342: static long
343: sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space)
344: {
345: struct iovec *iov = uio->uio_iov;
346: vaddr_t sva, eva;
347: vsize_t len;
1.156 ! yamt 348: vaddr_t lva;
! 349: int npgs, error;
! 350: vaddr_t va;
! 351: int i;
1.64 thorpej 352:
1.116 yamt 353: if (VMSPACE_IS_KERNEL_P(uio->uio_vmspace))
1.64 thorpej 354: return (0);
355:
356: if (iov->iov_len < (size_t) space)
357: space = iov->iov_len;
358: if (space > SOCK_LOAN_CHUNK)
359: space = SOCK_LOAN_CHUNK;
360:
361: eva = round_page((vaddr_t) iov->iov_base + space);
362: sva = trunc_page((vaddr_t) iov->iov_base);
363: len = eva - sva;
364: npgs = len >> PAGE_SHIFT;
365:
1.79 thorpej 366: /* XXX KDASSERT */
367: KASSERT(npgs <= M_EXT_MAXPAGES);
368:
1.80 yamt 369: lva = sokvaalloc(len, so);
1.64 thorpej 370: if (lva == 0)
1.80 yamt 371: return 0;
1.64 thorpej 372:
1.116 yamt 373: error = uvm_loan(&uio->uio_vmspace->vm_map, sva, len,
1.79 thorpej 374: m->m_ext.ext_pgs, UVM_LOAN_TOPAGE);
1.64 thorpej 375: if (error) {
1.80 yamt 376: sokvafree(lva, len);
1.64 thorpej 377: return (0);
378: }
379:
380: for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE)
1.79 thorpej 381: pmap_kenter_pa(va, VM_PAGE_TO_PHYS(m->m_ext.ext_pgs[i]),
382: VM_PROT_READ);
1.64 thorpej 383: pmap_update(pmap_kernel());
384:
385: lva += (vaddr_t) iov->iov_base & PAGE_MASK;
386:
1.134 christos 387: MEXTADD(m, (void *) lva, space, M_MBUF, soloanfree, so);
1.79 thorpej 388: m->m_flags |= M_EXT_PAGES | M_EXT_ROMAP;
1.64 thorpej 389:
390: uio->uio_resid -= space;
391: /* uio_offset not updated, not set/used for write(2) */
1.134 christos 392: uio->uio_iov->iov_base = (char *)uio->uio_iov->iov_base + space;
1.64 thorpej 393: uio->uio_iov->iov_len -= space;
394: if (uio->uio_iov->iov_len == 0) {
395: uio->uio_iov++;
396: uio->uio_iovcnt--;
397: }
398:
399: return (space);
400: }
401:
1.119 yamt 402: static int
1.129 yamt 403: sokva_reclaim_callback(struct callback_entry *ce, void *obj, void *arg)
1.119 yamt 404: {
405:
406: KASSERT(ce == &sokva_reclaimerentry);
407: KASSERT(obj == NULL);
408:
409: sodopendfree();
410: if (!vm_map_starved_p(kernel_map)) {
411: return CALLBACK_CHAIN_ABORT;
412: }
413: return CALLBACK_CHAIN_CONTINUE;
414: }
415:
1.142 dyoung 416: struct mbuf *
1.147 dyoung 417: getsombuf(struct socket *so, int type)
1.142 dyoung 418: {
419: struct mbuf *m;
420:
1.147 dyoung 421: m = m_get(M_WAIT, type);
1.142 dyoung 422: MCLAIM(m, so->so_mowner);
423: return m;
424: }
425:
426: struct mbuf *
427: m_intopt(struct socket *so, int val)
428: {
429: struct mbuf *m;
430:
1.147 dyoung 431: m = getsombuf(so, MT_SOOPTS);
1.142 dyoung 432: m->m_len = sizeof(int);
433: *mtod(m, int *) = val;
434: return m;
435: }
436:
1.119 yamt 437: void
438: soinit(void)
439: {
440:
1.148 ad 441: mutex_init(&so_pendfree_lock, MUTEX_DEFAULT, IPL_VM);
1.136 ad 442: cv_init(&socurkva_cv, "sokva");
443:
1.119 yamt 444: /* Set the initial adjusted socket buffer size. */
445: if (sb_max_set(sb_max))
446: panic("bad initial sb_max value: %lu", sb_max);
447:
448: callback_register(&vm_map_to_kernel(kernel_map)->vmk_reclaim_callback,
449: &sokva_reclaimerentry, NULL, sokva_reclaim_callback);
450: }
451:
1.1 cgd 452: /*
453: * Socket operation routines.
454: * These routines are called by the routines in
455: * sys_socket.c or from a system process, and
456: * implement the semantics of socket operations by
457: * switching out to the protocol specific routines.
458: */
459: /*ARGSUSED*/
1.3 andrew 460: int
1.114 christos 461: socreate(int dom, struct socket **aso, int type, int proto, struct lwp *l)
1.1 cgd 462: {
1.99 matt 463: const struct protosw *prp;
1.54 lukem 464: struct socket *so;
1.115 yamt 465: uid_t uid;
1.54 lukem 466: int error, s;
1.1 cgd 467:
1.132 elad 468: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_SOCKET,
469: KAUTH_REQ_NETWORK_SOCKET_OPEN, KAUTH_ARG(dom), KAUTH_ARG(type),
470: KAUTH_ARG(proto));
1.140 dyoung 471: if (error != 0)
472: return error;
1.127 elad 473:
1.1 cgd 474: if (proto)
475: prp = pffindproto(dom, proto, type);
476: else
477: prp = pffindtype(dom, type);
1.140 dyoung 478: if (prp == NULL) {
1.120 ginsbach 479: /* no support for domain */
480: if (pffinddomain(dom) == 0)
1.140 dyoung 481: return EAFNOSUPPORT;
1.120 ginsbach 482: /* no support for socket type */
483: if (proto == 0 && type != 0)
1.140 dyoung 484: return EPROTOTYPE;
485: return EPROTONOSUPPORT;
1.120 ginsbach 486: }
1.140 dyoung 487: if (prp->pr_usrreq == NULL)
488: return EPROTONOSUPPORT;
1.1 cgd 489: if (prp->pr_type != type)
1.140 dyoung 490: return EPROTOTYPE;
1.39 matt 491: s = splsoftnet();
1.37 thorpej 492: so = pool_get(&socket_pool, PR_WAITOK);
1.140 dyoung 493: memset(so, 0, sizeof(*so));
1.31 thorpej 494: TAILQ_INIT(&so->so_q0);
495: TAILQ_INIT(&so->so_q);
1.1 cgd 496: so->so_type = type;
497: so->so_proto = prp;
1.33 matt 498: so->so_send = sosend;
499: so->so_receive = soreceive;
1.78 matt 500: #ifdef MBUFTRACE
501: so->so_rcv.sb_mowner = &prp->pr_domain->dom_mowner;
502: so->so_snd.sb_mowner = &prp->pr_domain->dom_mowner;
503: so->so_mowner = &prp->pr_domain->dom_mowner;
504: #endif
1.143 ad 505: selinit(&so->so_rcv.sb_sel);
506: selinit(&so->so_snd.sb_sel);
1.138 rmind 507: uid = kauth_cred_geteuid(l->l_cred);
1.115 yamt 508: so->so_uidinfo = uid_find(uid);
1.140 dyoung 509: error = (*prp->pr_usrreq)(so, PRU_ATTACH, NULL,
510: (struct mbuf *)(long)proto, NULL, l);
511: if (error != 0) {
1.1 cgd 512: so->so_state |= SS_NOFDREF;
513: sofree(so);
1.39 matt 514: splx(s);
1.140 dyoung 515: return error;
1.1 cgd 516: }
1.39 matt 517: splx(s);
1.1 cgd 518: *aso = so;
1.140 dyoung 519: return 0;
1.1 cgd 520: }
521:
1.142 dyoung 522: /* On success, write file descriptor to fdout and return zero. On
523: * failure, return non-zero; *fdout will be undefined.
524: */
525: int
526: fsocreate(int domain, struct socket **sop, int type, int protocol,
527: struct lwp *l, int *fdout)
528: {
529: struct socket *so;
530: struct file *fp;
531: int fd, error;
532:
1.155 ad 533: if ((error = fd_allocfile(&fp, &fd)) != 0)
1.142 dyoung 534: return (error);
535: fp->f_flag = FREAD|FWRITE;
536: fp->f_type = DTYPE_SOCKET;
537: fp->f_ops = &socketops;
538: error = socreate(domain, &so, type, protocol, l);
539: if (error != 0) {
1.155 ad 540: fd_abort(curproc, fp, fd);
1.142 dyoung 541: } else {
542: if (sop != NULL)
543: *sop = so;
544: fp->f_data = so;
1.155 ad 545: fd_affix(curproc, fp, fd);
1.142 dyoung 546: *fdout = fd;
547: }
548: return error;
549: }
550:
1.3 andrew 551: int
1.114 christos 552: sobind(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1 cgd 553: {
1.54 lukem 554: int s, error;
1.1 cgd 555:
1.54 lukem 556: s = splsoftnet();
1.140 dyoung 557: error = (*so->so_proto->pr_usrreq)(so, PRU_BIND, NULL, nam, NULL, l);
1.1 cgd 558: splx(s);
1.140 dyoung 559: return error;
1.1 cgd 560: }
561:
1.3 andrew 562: int
1.150 elad 563: solisten(struct socket *so, int backlog, struct lwp *l)
1.1 cgd 564: {
1.54 lukem 565: int s, error;
1.1 cgd 566:
1.54 lukem 567: s = splsoftnet();
1.140 dyoung 568: error = (*so->so_proto->pr_usrreq)(so, PRU_LISTEN, NULL,
1.150 elad 569: NULL, NULL, l);
1.140 dyoung 570: if (error != 0) {
1.1 cgd 571: splx(s);
1.140 dyoung 572: return error;
1.1 cgd 573: }
1.63 matt 574: if (TAILQ_EMPTY(&so->so_q))
1.1 cgd 575: so->so_options |= SO_ACCEPTCONN;
576: if (backlog < 0)
577: backlog = 0;
1.49 jonathan 578: so->so_qlimit = min(backlog, somaxconn);
1.1 cgd 579: splx(s);
1.140 dyoung 580: return 0;
1.1 cgd 581: }
582:
1.21 christos 583: void
1.54 lukem 584: sofree(struct socket *so)
1.1 cgd 585: {
586:
1.43 mycroft 587: if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0)
1.1 cgd 588: return;
1.43 mycroft 589: if (so->so_head) {
590: /*
591: * We must not decommission a socket that's on the accept(2)
592: * queue. If we do, then accept(2) may hang after select(2)
593: * indicated that the listening socket was ready.
594: */
595: if (!soqremque(so, 0))
596: return;
597: }
1.98 christos 598: if (so->so_rcv.sb_hiwat)
1.110 christos 599: (void)chgsbsize(so->so_uidinfo, &so->so_rcv.sb_hiwat, 0,
1.98 christos 600: RLIM_INFINITY);
601: if (so->so_snd.sb_hiwat)
1.110 christos 602: (void)chgsbsize(so->so_uidinfo, &so->so_snd.sb_hiwat, 0,
1.98 christos 603: RLIM_INFINITY);
604: sbrelease(&so->so_snd, so);
1.1 cgd 605: sorflush(so);
1.143 ad 606: seldestroy(&so->so_rcv.sb_sel);
607: seldestroy(&so->so_snd.sb_sel);
1.37 thorpej 608: pool_put(&socket_pool, so);
1.1 cgd 609: }
610:
611: /*
612: * Close a socket on last file table reference removal.
613: * Initiate disconnect if connected.
614: * Free socket when disconnect complete.
615: */
1.3 andrew 616: int
1.54 lukem 617: soclose(struct socket *so)
1.1 cgd 618: {
1.54 lukem 619: struct socket *so2;
620: int s, error;
1.1 cgd 621:
1.54 lukem 622: error = 0;
623: s = splsoftnet(); /* conservative */
1.1 cgd 624: if (so->so_options & SO_ACCEPTCONN) {
1.63 matt 625: while ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) {
1.42 mycroft 626: (void) soqremque(so2, 0);
1.41 mycroft 627: (void) soabort(so2);
628: }
1.63 matt 629: while ((so2 = TAILQ_FIRST(&so->so_q)) != 0) {
1.42 mycroft 630: (void) soqremque(so2, 1);
1.41 mycroft 631: (void) soabort(so2);
632: }
1.1 cgd 633: }
634: if (so->so_pcb == 0)
635: goto discard;
636: if (so->so_state & SS_ISCONNECTED) {
637: if ((so->so_state & SS_ISDISCONNECTING) == 0) {
638: error = sodisconnect(so);
639: if (error)
640: goto drop;
641: }
642: if (so->so_options & SO_LINGER) {
1.151 ad 643: if ((so->so_state & SS_ISDISCONNECTING) && so->so_nbio)
1.1 cgd 644: goto drop;
1.21 christos 645: while (so->so_state & SS_ISCONNECTED) {
1.134 christos 646: error = tsleep((void *)&so->so_timeo,
1.21 christos 647: PSOCK | PCATCH, netcls,
1.30 thorpej 648: so->so_linger * hz);
1.21 christos 649: if (error)
1.1 cgd 650: break;
1.21 christos 651: }
1.1 cgd 652: }
653: }
1.54 lukem 654: drop:
1.1 cgd 655: if (so->so_pcb) {
1.22 mycroft 656: int error2 = (*so->so_proto->pr_usrreq)(so, PRU_DETACH,
1.140 dyoung 657: NULL, NULL, NULL, NULL);
1.1 cgd 658: if (error == 0)
659: error = error2;
660: }
1.54 lukem 661: discard:
1.1 cgd 662: if (so->so_state & SS_NOFDREF)
663: panic("soclose: NOFDREF");
664: so->so_state |= SS_NOFDREF;
665: sofree(so);
666: splx(s);
667: return (error);
668: }
669:
670: /*
1.20 mycroft 671: * Must be called at splsoftnet...
1.1 cgd 672: */
1.3 andrew 673: int
1.54 lukem 674: soabort(struct socket *so)
1.1 cgd 675: {
1.139 yamt 676: int error;
1.1 cgd 677:
1.139 yamt 678: KASSERT(so->so_head == NULL);
1.140 dyoung 679: error = (*so->so_proto->pr_usrreq)(so, PRU_ABORT, NULL,
680: NULL, NULL, NULL);
1.139 yamt 681: if (error) {
682: sofree(so);
683: }
684: return error;
1.1 cgd 685: }
686:
1.3 andrew 687: int
1.54 lukem 688: soaccept(struct socket *so, struct mbuf *nam)
1.1 cgd 689: {
1.54 lukem 690: int s, error;
1.1 cgd 691:
1.54 lukem 692: error = 0;
693: s = splsoftnet();
1.1 cgd 694: if ((so->so_state & SS_NOFDREF) == 0)
695: panic("soaccept: !NOFDREF");
696: so->so_state &= ~SS_NOFDREF;
1.55 thorpej 697: if ((so->so_state & SS_ISDISCONNECTED) == 0 ||
698: (so->so_proto->pr_flags & PR_ABRTACPTDIS) == 0)
1.41 mycroft 699: error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT,
1.140 dyoung 700: NULL, nam, NULL, NULL);
1.41 mycroft 701: else
1.53 itojun 702: error = ECONNABORTED;
1.52 itojun 703:
1.1 cgd 704: splx(s);
705: return (error);
706: }
707:
1.3 andrew 708: int
1.114 christos 709: soconnect(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1 cgd 710: {
1.54 lukem 711: int s, error;
1.1 cgd 712:
713: if (so->so_options & SO_ACCEPTCONN)
714: return (EOPNOTSUPP);
1.20 mycroft 715: s = splsoftnet();
1.1 cgd 716: /*
717: * If protocol is connection-based, can only connect once.
718: * Otherwise, if connected, try to disconnect first.
719: * This allows user to disconnect by connecting to, e.g.,
720: * a null address.
721: */
722: if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
723: ((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
724: (error = sodisconnect(so))))
725: error = EISCONN;
726: else
727: error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
1.140 dyoung 728: NULL, nam, NULL, l);
1.1 cgd 729: splx(s);
730: return (error);
731: }
732:
1.3 andrew 733: int
1.54 lukem 734: soconnect2(struct socket *so1, struct socket *so2)
1.1 cgd 735: {
1.54 lukem 736: int s, error;
1.1 cgd 737:
1.54 lukem 738: s = splsoftnet();
1.22 mycroft 739: error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2,
1.140 dyoung 740: NULL, (struct mbuf *)so2, NULL, NULL);
1.1 cgd 741: splx(s);
742: return (error);
743: }
744:
1.3 andrew 745: int
1.54 lukem 746: sodisconnect(struct socket *so)
1.1 cgd 747: {
1.54 lukem 748: int s, error;
1.1 cgd 749:
1.54 lukem 750: s = splsoftnet();
1.1 cgd 751: if ((so->so_state & SS_ISCONNECTED) == 0) {
752: error = ENOTCONN;
753: goto bad;
754: }
755: if (so->so_state & SS_ISDISCONNECTING) {
756: error = EALREADY;
757: goto bad;
758: }
1.22 mycroft 759: error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
1.140 dyoung 760: NULL, NULL, NULL, NULL);
1.54 lukem 761: bad:
1.1 cgd 762: splx(s);
1.117 yamt 763: sodopendfree();
1.1 cgd 764: return (error);
765: }
766:
1.15 mycroft 767: #define SBLOCKWAIT(f) (((f) & MSG_DONTWAIT) ? M_NOWAIT : M_WAITOK)
1.1 cgd 768: /*
769: * Send on a socket.
770: * If send must go all at once and message is larger than
771: * send buffering, then hard error.
772: * Lock against other senders.
773: * If must go all at once and not enough room now, then
774: * inform user that this would block and do nothing.
775: * Otherwise, if nonblocking, send as much as possible.
776: * The data to be sent is described by "uio" if nonzero,
777: * otherwise by the mbuf chain "top" (which must be null
778: * if uio is not). Data provided in mbuf chain must be small
779: * enough to send all at once.
780: *
781: * Returns nonzero on error, timeout or signal; callers
782: * must check for short counts if EINTR/ERESTART are returned.
783: * Data and control buffers are freed on return.
784: */
1.3 andrew 785: int
1.54 lukem 786: sosend(struct socket *so, struct mbuf *addr, struct uio *uio, struct mbuf *top,
1.114 christos 787: struct mbuf *control, int flags, struct lwp *l)
1.1 cgd 788: {
1.54 lukem 789: struct mbuf **mp, *m;
1.114 christos 790: struct proc *p;
1.58 jdolecek 791: long space, len, resid, clen, mlen;
792: int error, s, dontroute, atomic;
1.54 lukem 793:
1.114 christos 794: p = l->l_proc;
1.117 yamt 795: sodopendfree();
1.64 thorpej 796:
1.54 lukem 797: clen = 0;
798: atomic = sosendallatonce(so) || top;
1.1 cgd 799: if (uio)
800: resid = uio->uio_resid;
801: else
802: resid = top->m_pkthdr.len;
1.7 cgd 803: /*
804: * In theory resid should be unsigned.
805: * However, space must be signed, as it might be less than 0
806: * if we over-committed, and we must use a signed comparison
807: * of space and resid. On the other hand, a negative resid
808: * causes us to loop sending 0-length segments to the protocol.
809: */
1.29 mycroft 810: if (resid < 0) {
811: error = EINVAL;
812: goto out;
813: }
1.1 cgd 814: dontroute =
815: (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
816: (so->so_proto->pr_flags & PR_ATOMIC);
1.102 jonathan 817: if (p)
818: p->p_stats->p_ru.ru_msgsnd++;
1.1 cgd 819: if (control)
820: clen = control->m_len;
821: #define snderr(errno) { error = errno; splx(s); goto release; }
822:
1.54 lukem 823: restart:
1.21 christos 824: if ((error = sblock(&so->so_snd, SBLOCKWAIT(flags))) != 0)
1.1 cgd 825: goto out;
826: do {
1.20 mycroft 827: s = splsoftnet();
1.1 cgd 828: if (so->so_state & SS_CANTSENDMORE)
829: snderr(EPIPE);
1.48 thorpej 830: if (so->so_error) {
831: error = so->so_error;
832: so->so_error = 0;
833: splx(s);
834: goto release;
835: }
1.1 cgd 836: if ((so->so_state & SS_ISCONNECTED) == 0) {
837: if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
838: if ((so->so_state & SS_ISCONFIRMING) == 0 &&
839: !(resid == 0 && clen != 0))
840: snderr(ENOTCONN);
841: } else if (addr == 0)
842: snderr(EDESTADDRREQ);
843: }
844: space = sbspace(&so->so_snd);
845: if (flags & MSG_OOB)
846: space += 1024;
1.21 christos 847: if ((atomic && resid > so->so_snd.sb_hiwat) ||
1.11 mycroft 848: clen > so->so_snd.sb_hiwat)
849: snderr(EMSGSIZE);
1.96 mycroft 850: if (space < resid + clen &&
1.1 cgd 851: (atomic || space < so->so_snd.sb_lowat || space < clen)) {
1.151 ad 852: if (so->so_nbio)
1.1 cgd 853: snderr(EWOULDBLOCK);
854: sbunlock(&so->so_snd);
855: error = sbwait(&so->so_snd);
856: splx(s);
857: if (error)
858: goto out;
859: goto restart;
860: }
861: splx(s);
862: mp = ⊤
863: space -= clen;
864: do {
1.45 tv 865: if (uio == NULL) {
866: /*
867: * Data is prepackaged in "top".
868: */
869: resid = 0;
870: if (flags & MSG_EOR)
871: top->m_flags |= M_EOR;
872: } else do {
1.144 dyoung 873: if (top == NULL) {
1.78 matt 874: m = m_gethdr(M_WAIT, MT_DATA);
1.45 tv 875: mlen = MHLEN;
876: m->m_pkthdr.len = 0;
1.140 dyoung 877: m->m_pkthdr.rcvif = NULL;
1.45 tv 878: } else {
1.78 matt 879: m = m_get(M_WAIT, MT_DATA);
1.45 tv 880: mlen = MLEN;
881: }
1.78 matt 882: MCLAIM(m, so->so_snd.sb_mowner);
1.121 yamt 883: if (sock_loan_thresh >= 0 &&
884: uio->uio_iov->iov_len >= sock_loan_thresh &&
885: space >= sock_loan_thresh &&
1.64 thorpej 886: (len = sosend_loan(so, uio, m,
887: space)) != 0) {
888: SOSEND_COUNTER_INCR(&sosend_loan_big);
889: space -= len;
890: goto have_data;
891: }
1.45 tv 892: if (resid >= MINCLSIZE && space >= MCLBYTES) {
1.64 thorpej 893: SOSEND_COUNTER_INCR(&sosend_copy_big);
1.78 matt 894: m_clget(m, M_WAIT);
1.45 tv 895: if ((m->m_flags & M_EXT) == 0)
896: goto nopages;
897: mlen = MCLBYTES;
898: if (atomic && top == 0) {
1.58 jdolecek 899: len = lmin(MCLBYTES - max_hdr,
1.54 lukem 900: resid);
1.45 tv 901: m->m_data += max_hdr;
902: } else
1.58 jdolecek 903: len = lmin(MCLBYTES, resid);
1.45 tv 904: space -= len;
905: } else {
1.64 thorpej 906: nopages:
907: SOSEND_COUNTER_INCR(&sosend_copy_small);
1.58 jdolecek 908: len = lmin(lmin(mlen, resid), space);
1.45 tv 909: space -= len;
910: /*
911: * For datagram protocols, leave room
912: * for protocol headers in first mbuf.
913: */
914: if (atomic && top == 0 && len < mlen)
915: MH_ALIGN(m, len);
916: }
1.144 dyoung 917: error = uiomove(mtod(m, void *), (int)len, uio);
1.64 thorpej 918: have_data:
1.45 tv 919: resid = uio->uio_resid;
920: m->m_len = len;
921: *mp = m;
922: top->m_pkthdr.len += len;
1.144 dyoung 923: if (error != 0)
1.45 tv 924: goto release;
925: mp = &m->m_next;
926: if (resid <= 0) {
927: if (flags & MSG_EOR)
928: top->m_flags |= M_EOR;
929: break;
930: }
931: } while (space > 0 && atomic);
1.108 perry 932:
1.46 sommerfe 933: s = splsoftnet();
934:
935: if (so->so_state & SS_CANTSENDMORE)
936: snderr(EPIPE);
1.45 tv 937:
938: if (dontroute)
939: so->so_options |= SO_DONTROUTE;
940: if (resid > 0)
941: so->so_state |= SS_MORETOCOME;
1.46 sommerfe 942: error = (*so->so_proto->pr_usrreq)(so,
943: (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
1.114 christos 944: top, addr, control, curlwp); /* XXX */
1.45 tv 945: if (dontroute)
946: so->so_options &= ~SO_DONTROUTE;
947: if (resid > 0)
948: so->so_state &= ~SS_MORETOCOME;
1.46 sommerfe 949: splx(s);
950:
1.45 tv 951: clen = 0;
1.144 dyoung 952: control = NULL;
953: top = NULL;
1.45 tv 954: mp = ⊤
1.144 dyoung 955: if (error != 0)
1.1 cgd 956: goto release;
957: } while (resid && space > 0);
958: } while (resid);
959:
1.54 lukem 960: release:
1.1 cgd 961: sbunlock(&so->so_snd);
1.54 lukem 962: out:
1.1 cgd 963: if (top)
964: m_freem(top);
965: if (control)
966: m_freem(control);
967: return (error);
968: }
969:
970: /*
971: * Implement receive operations on a socket.
972: * We depend on the way that records are added to the sockbuf
973: * by sbappend*. In particular, each record (mbufs linked through m_next)
974: * must begin with an address if the protocol so specifies,
975: * followed by an optional mbuf or mbufs containing ancillary data,
976: * and then zero or more mbufs of data.
977: * In order to avoid blocking network interrupts for the entire time here,
978: * we splx() while doing the actual copy to user space.
979: * Although the sockbuf is locked, new data may still be appended,
980: * and thus we must maintain consistency of the sockbuf during that time.
981: *
982: * The caller may receive the data as a single mbuf chain by supplying
983: * an mbuf **mp0 for use in returning the chain. The uio is then used
984: * only for the count in uio_resid.
985: */
1.3 andrew 986: int
1.54 lukem 987: soreceive(struct socket *so, struct mbuf **paddr, struct uio *uio,
988: struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
1.1 cgd 989: {
1.116 yamt 990: struct lwp *l = curlwp;
1.54 lukem 991: struct mbuf *m, **mp;
1.146 dyoung 992: int atomic, flags, len, error, s, offset, moff, type, orig_resid;
1.99 matt 993: const struct protosw *pr;
1.54 lukem 994: struct mbuf *nextrecord;
1.67 he 995: int mbuf_removed = 0;
1.146 dyoung 996: const struct domain *dom;
1.64 thorpej 997:
1.54 lukem 998: pr = so->so_proto;
1.146 dyoung 999: atomic = pr->pr_flags & PR_ATOMIC;
1000: dom = pr->pr_domain;
1.1 cgd 1001: mp = mp0;
1.54 lukem 1002: type = 0;
1003: orig_resid = uio->uio_resid;
1.102 jonathan 1004:
1.144 dyoung 1005: if (paddr != NULL)
1006: *paddr = NULL;
1007: if (controlp != NULL)
1008: *controlp = NULL;
1009: if (flagsp != NULL)
1.1 cgd 1010: flags = *flagsp &~ MSG_EOR;
1011: else
1012: flags = 0;
1.66 enami 1013:
1014: if ((flags & MSG_DONTWAIT) == 0)
1.117 yamt 1015: sodopendfree();
1.66 enami 1016:
1.1 cgd 1017: if (flags & MSG_OOB) {
1018: m = m_get(M_WAIT, MT_DATA);
1.17 cgd 1019: error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m,
1.140 dyoung 1020: (struct mbuf *)(long)(flags & MSG_PEEK), NULL, l);
1.1 cgd 1021: if (error)
1022: goto bad;
1023: do {
1.134 christos 1024: error = uiomove(mtod(m, void *),
1.1 cgd 1025: (int) min(uio->uio_resid, m->m_len), uio);
1026: m = m_free(m);
1.144 dyoung 1027: } while (uio->uio_resid > 0 && error == 0 && m);
1.54 lukem 1028: bad:
1.144 dyoung 1029: if (m != NULL)
1.1 cgd 1030: m_freem(m);
1.144 dyoung 1031: return error;
1.1 cgd 1032: }
1.144 dyoung 1033: if (mp != NULL)
1.140 dyoung 1034: *mp = NULL;
1.1 cgd 1035: if (so->so_state & SS_ISCONFIRMING && uio->uio_resid)
1.140 dyoung 1036: (*pr->pr_usrreq)(so, PRU_RCVD, NULL, NULL, NULL, l);
1.1 cgd 1037:
1.54 lukem 1038: restart:
1.21 christos 1039: if ((error = sblock(&so->so_rcv, SBLOCKWAIT(flags))) != 0)
1.144 dyoung 1040: return error;
1.20 mycroft 1041: s = splsoftnet();
1.1 cgd 1042:
1043: m = so->so_rcv.sb_mb;
1044: /*
1045: * If we have less data than requested, block awaiting more
1046: * (subject to any timeout) if:
1.15 mycroft 1047: * 1. the current count is less than the low water mark,
1.1 cgd 1048: * 2. MSG_WAITALL is set, and it is possible to do the entire
1.15 mycroft 1049: * receive operation at once if we block (resid <= hiwat), or
1050: * 3. MSG_DONTWAIT is not set.
1.1 cgd 1051: * If MSG_WAITALL is set but resid is larger than the receive buffer,
1052: * we have to do the receive in sections, and thus risk returning
1053: * a short count if a timeout or signal occurs after we start.
1054: */
1.144 dyoung 1055: if (m == NULL ||
1056: ((flags & MSG_DONTWAIT) == 0 &&
1057: so->so_rcv.sb_cc < uio->uio_resid &&
1058: (so->so_rcv.sb_cc < so->so_rcv.sb_lowat ||
1059: ((flags & MSG_WAITALL) &&
1060: uio->uio_resid <= so->so_rcv.sb_hiwat)) &&
1.146 dyoung 1061: m->m_nextpkt == NULL && !atomic)) {
1.1 cgd 1062: #ifdef DIAGNOSTIC
1.144 dyoung 1063: if (m == NULL && so->so_rcv.sb_cc)
1.1 cgd 1064: panic("receive 1");
1065: #endif
1066: if (so->so_error) {
1.144 dyoung 1067: if (m != NULL)
1.15 mycroft 1068: goto dontblock;
1.1 cgd 1069: error = so->so_error;
1070: if ((flags & MSG_PEEK) == 0)
1071: so->so_error = 0;
1072: goto release;
1073: }
1074: if (so->so_state & SS_CANTRCVMORE) {
1.144 dyoung 1075: if (m != NULL)
1.15 mycroft 1076: goto dontblock;
1.1 cgd 1077: else
1078: goto release;
1079: }
1.144 dyoung 1080: for (; m != NULL; m = m->m_next)
1.1 cgd 1081: if (m->m_type == MT_OOBDATA || (m->m_flags & M_EOR)) {
1082: m = so->so_rcv.sb_mb;
1083: goto dontblock;
1084: }
1085: if ((so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) == 0 &&
1086: (so->so_proto->pr_flags & PR_CONNREQUIRED)) {
1087: error = ENOTCONN;
1088: goto release;
1089: }
1090: if (uio->uio_resid == 0)
1091: goto release;
1.151 ad 1092: if (so->so_nbio || (flags & MSG_DONTWAIT)) {
1.1 cgd 1093: error = EWOULDBLOCK;
1094: goto release;
1095: }
1.69 thorpej 1096: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1");
1097: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1");
1.1 cgd 1098: sbunlock(&so->so_rcv);
1099: error = sbwait(&so->so_rcv);
1100: splx(s);
1.144 dyoung 1101: if (error != 0)
1102: return error;
1.1 cgd 1103: goto restart;
1104: }
1.54 lukem 1105: dontblock:
1.69 thorpej 1106: /*
1107: * On entry here, m points to the first record of the socket buffer.
1108: * While we process the initial mbufs containing address and control
1109: * info, we save a copy of m->m_nextpkt into nextrecord.
1110: */
1.144 dyoung 1111: if (l != NULL)
1.114 christos 1112: l->l_proc->p_stats->p_ru.ru_msgrcv++;
1.69 thorpej 1113: KASSERT(m == so->so_rcv.sb_mb);
1114: SBLASTRECORDCHK(&so->so_rcv, "soreceive 1");
1115: SBLASTMBUFCHK(&so->so_rcv, "soreceive 1");
1.1 cgd 1116: nextrecord = m->m_nextpkt;
1117: if (pr->pr_flags & PR_ADDR) {
1118: #ifdef DIAGNOSTIC
1119: if (m->m_type != MT_SONAME)
1120: panic("receive 1a");
1121: #endif
1.3 andrew 1122: orig_resid = 0;
1.1 cgd 1123: if (flags & MSG_PEEK) {
1124: if (paddr)
1125: *paddr = m_copy(m, 0, m->m_len);
1126: m = m->m_next;
1127: } else {
1128: sbfree(&so->so_rcv, m);
1.67 he 1129: mbuf_removed = 1;
1.144 dyoung 1130: if (paddr != NULL) {
1.1 cgd 1131: *paddr = m;
1132: so->so_rcv.sb_mb = m->m_next;
1.144 dyoung 1133: m->m_next = NULL;
1.1 cgd 1134: m = so->so_rcv.sb_mb;
1135: } else {
1136: MFREE(m, so->so_rcv.sb_mb);
1137: m = so->so_rcv.sb_mb;
1138: }
1139: }
1140: }
1.144 dyoung 1141: while (m != NULL && m->m_type == MT_CONTROL && error == 0) {
1.1 cgd 1142: if (flags & MSG_PEEK) {
1.144 dyoung 1143: if (controlp != NULL)
1.1 cgd 1144: *controlp = m_copy(m, 0, m->m_len);
1145: m = m->m_next;
1146: } else {
1147: sbfree(&so->so_rcv, m);
1.67 he 1148: mbuf_removed = 1;
1.144 dyoung 1149: if (controlp != NULL) {
1.114 christos 1150: if (dom->dom_externalize && l &&
1.1 cgd 1151: mtod(m, struct cmsghdr *)->cmsg_type ==
1152: SCM_RIGHTS)
1.114 christos 1153: error = (*dom->dom_externalize)(m, l);
1.1 cgd 1154: *controlp = m;
1155: so->so_rcv.sb_mb = m->m_next;
1.144 dyoung 1156: m->m_next = NULL;
1.1 cgd 1157: m = so->so_rcv.sb_mb;
1158: } else {
1.106 itojun 1159: /*
1160: * Dispose of any SCM_RIGHTS message that went
1161: * through the read path rather than recv.
1162: */
1.146 dyoung 1163: if (dom->dom_dispose &&
1.106 itojun 1164: mtod(m, struct cmsghdr *)->cmsg_type == SCM_RIGHTS)
1.146 dyoung 1165: (*dom->dom_dispose)(m);
1.1 cgd 1166: MFREE(m, so->so_rcv.sb_mb);
1167: m = so->so_rcv.sb_mb;
1168: }
1169: }
1.144 dyoung 1170: if (controlp != NULL) {
1.3 andrew 1171: orig_resid = 0;
1.1 cgd 1172: controlp = &(*controlp)->m_next;
1.3 andrew 1173: }
1.1 cgd 1174: }
1.69 thorpej 1175:
1176: /*
1177: * If m is non-NULL, we have some data to read. From now on,
1178: * make sure to keep sb_lastrecord consistent when working on
1179: * the last packet on the chain (nextrecord == NULL) and we
1180: * change m->m_nextpkt.
1181: */
1.144 dyoung 1182: if (m != NULL) {
1.69 thorpej 1183: if ((flags & MSG_PEEK) == 0) {
1.1 cgd 1184: m->m_nextpkt = nextrecord;
1.69 thorpej 1185: /*
1186: * If nextrecord == NULL (this is a single chain),
1187: * then sb_lastrecord may not be valid here if m
1188: * was changed earlier.
1189: */
1190: if (nextrecord == NULL) {
1191: KASSERT(so->so_rcv.sb_mb == m);
1192: so->so_rcv.sb_lastrecord = m;
1193: }
1194: }
1.1 cgd 1195: type = m->m_type;
1196: if (type == MT_OOBDATA)
1197: flags |= MSG_OOB;
1.69 thorpej 1198: } else {
1199: if ((flags & MSG_PEEK) == 0) {
1200: KASSERT(so->so_rcv.sb_mb == m);
1201: so->so_rcv.sb_mb = nextrecord;
1.70 thorpej 1202: SB_EMPTY_FIXUP(&so->so_rcv);
1.69 thorpej 1203: }
1.1 cgd 1204: }
1.69 thorpej 1205: SBLASTRECORDCHK(&so->so_rcv, "soreceive 2");
1206: SBLASTMBUFCHK(&so->so_rcv, "soreceive 2");
1207:
1.1 cgd 1208: moff = 0;
1209: offset = 0;
1.144 dyoung 1210: while (m != NULL && uio->uio_resid > 0 && error == 0) {
1.1 cgd 1211: if (m->m_type == MT_OOBDATA) {
1212: if (type != MT_OOBDATA)
1213: break;
1214: } else if (type == MT_OOBDATA)
1215: break;
1216: #ifdef DIAGNOSTIC
1217: else if (m->m_type != MT_DATA && m->m_type != MT_HEADER)
1218: panic("receive 3");
1219: #endif
1220: so->so_state &= ~SS_RCVATMARK;
1221: len = uio->uio_resid;
1222: if (so->so_oobmark && len > so->so_oobmark - offset)
1223: len = so->so_oobmark - offset;
1224: if (len > m->m_len - moff)
1225: len = m->m_len - moff;
1226: /*
1227: * If mp is set, just pass back the mbufs.
1228: * Otherwise copy them out via the uio, then free.
1229: * Sockbuf must be consistent here (points to current mbuf,
1230: * it points to next record) when we drop priority;
1231: * we must note any additions to the sockbuf when we
1232: * block interrupts again.
1233: */
1.144 dyoung 1234: if (mp == NULL) {
1.69 thorpej 1235: SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove");
1236: SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove");
1.1 cgd 1237: splx(s);
1.134 christos 1238: error = uiomove(mtod(m, char *) + moff, (int)len, uio);
1.20 mycroft 1239: s = splsoftnet();
1.144 dyoung 1240: if (error != 0) {
1.67 he 1241: /*
1242: * If any part of the record has been removed
1243: * (such as the MT_SONAME mbuf, which will
1244: * happen when PR_ADDR, and thus also
1245: * PR_ATOMIC, is set), then drop the entire
1246: * record to maintain the atomicity of the
1247: * receive operation.
1248: *
1249: * This avoids a later panic("receive 1a")
1250: * when compiled with DIAGNOSTIC.
1251: */
1.146 dyoung 1252: if (m && mbuf_removed && atomic)
1.67 he 1253: (void) sbdroprecord(&so->so_rcv);
1254:
1.57 jdolecek 1255: goto release;
1.67 he 1256: }
1.1 cgd 1257: } else
1258: uio->uio_resid -= len;
1259: if (len == m->m_len - moff) {
1260: if (m->m_flags & M_EOR)
1261: flags |= MSG_EOR;
1262: if (flags & MSG_PEEK) {
1263: m = m->m_next;
1264: moff = 0;
1265: } else {
1266: nextrecord = m->m_nextpkt;
1267: sbfree(&so->so_rcv, m);
1268: if (mp) {
1269: *mp = m;
1270: mp = &m->m_next;
1271: so->so_rcv.sb_mb = m = m->m_next;
1.140 dyoung 1272: *mp = NULL;
1.1 cgd 1273: } else {
1274: MFREE(m, so->so_rcv.sb_mb);
1275: m = so->so_rcv.sb_mb;
1276: }
1.69 thorpej 1277: /*
1278: * If m != NULL, we also know that
1279: * so->so_rcv.sb_mb != NULL.
1280: */
1281: KASSERT(so->so_rcv.sb_mb == m);
1282: if (m) {
1.1 cgd 1283: m->m_nextpkt = nextrecord;
1.69 thorpej 1284: if (nextrecord == NULL)
1285: so->so_rcv.sb_lastrecord = m;
1286: } else {
1287: so->so_rcv.sb_mb = nextrecord;
1.70 thorpej 1288: SB_EMPTY_FIXUP(&so->so_rcv);
1.69 thorpej 1289: }
1290: SBLASTRECORDCHK(&so->so_rcv, "soreceive 3");
1291: SBLASTMBUFCHK(&so->so_rcv, "soreceive 3");
1.1 cgd 1292: }
1.144 dyoung 1293: } else if (flags & MSG_PEEK)
1294: moff += len;
1295: else {
1296: if (mp != NULL)
1297: *mp = m_copym(m, 0, len, M_WAIT);
1298: m->m_data += len;
1299: m->m_len -= len;
1300: so->so_rcv.sb_cc -= len;
1.1 cgd 1301: }
1302: if (so->so_oobmark) {
1303: if ((flags & MSG_PEEK) == 0) {
1304: so->so_oobmark -= len;
1305: if (so->so_oobmark == 0) {
1306: so->so_state |= SS_RCVATMARK;
1307: break;
1308: }
1.7 cgd 1309: } else {
1.1 cgd 1310: offset += len;
1.7 cgd 1311: if (offset == so->so_oobmark)
1312: break;
1313: }
1.1 cgd 1314: }
1315: if (flags & MSG_EOR)
1316: break;
1317: /*
1318: * If the MSG_WAITALL flag is set (for non-atomic socket),
1319: * we must not quit until "uio->uio_resid == 0" or an error
1320: * termination. If a signal/timeout occurs, return
1321: * with a short count but without error.
1322: * Keep sockbuf locked against other readers.
1323: */
1.144 dyoung 1324: while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 &&
1.3 andrew 1325: !sosendallatonce(so) && !nextrecord) {
1.1 cgd 1326: if (so->so_error || so->so_state & SS_CANTRCVMORE)
1327: break;
1.68 matt 1328: /*
1329: * If we are peeking and the socket receive buffer is
1330: * full, stop since we can't get more data to peek at.
1331: */
1332: if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0)
1333: break;
1334: /*
1335: * If we've drained the socket buffer, tell the
1336: * protocol in case it needs to do something to
1337: * get it filled again.
1338: */
1339: if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb)
1340: (*pr->pr_usrreq)(so, PRU_RCVD,
1.140 dyoung 1341: NULL, (struct mbuf *)(long)flags, NULL, l);
1.69 thorpej 1342: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2");
1343: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2");
1.1 cgd 1344: error = sbwait(&so->so_rcv);
1.144 dyoung 1345: if (error != 0) {
1.1 cgd 1346: sbunlock(&so->so_rcv);
1347: splx(s);
1.144 dyoung 1348: return 0;
1.1 cgd 1349: }
1.21 christos 1350: if ((m = so->so_rcv.sb_mb) != NULL)
1.1 cgd 1351: nextrecord = m->m_nextpkt;
1352: }
1353: }
1.3 andrew 1354:
1.146 dyoung 1355: if (m && atomic) {
1.3 andrew 1356: flags |= MSG_TRUNC;
1357: if ((flags & MSG_PEEK) == 0)
1358: (void) sbdroprecord(&so->so_rcv);
1359: }
1.1 cgd 1360: if ((flags & MSG_PEEK) == 0) {
1.144 dyoung 1361: if (m == NULL) {
1.69 thorpej 1362: /*
1.70 thorpej 1363: * First part is an inline SB_EMPTY_FIXUP(). Second
1.69 thorpej 1364: * part makes sure sb_lastrecord is up-to-date if
1365: * there is still data in the socket buffer.
1366: */
1.1 cgd 1367: so->so_rcv.sb_mb = nextrecord;
1.69 thorpej 1368: if (so->so_rcv.sb_mb == NULL) {
1369: so->so_rcv.sb_mbtail = NULL;
1370: so->so_rcv.sb_lastrecord = NULL;
1371: } else if (nextrecord->m_nextpkt == NULL)
1372: so->so_rcv.sb_lastrecord = nextrecord;
1373: }
1374: SBLASTRECORDCHK(&so->so_rcv, "soreceive 4");
1375: SBLASTMBUFCHK(&so->so_rcv, "soreceive 4");
1.1 cgd 1376: if (pr->pr_flags & PR_WANTRCVD && so->so_pcb)
1.140 dyoung 1377: (*pr->pr_usrreq)(so, PRU_RCVD, NULL,
1378: (struct mbuf *)(long)flags, NULL, l);
1.1 cgd 1379: }
1.3 andrew 1380: if (orig_resid == uio->uio_resid && orig_resid &&
1381: (flags & MSG_EOR) == 0 && (so->so_state & SS_CANTRCVMORE) == 0) {
1382: sbunlock(&so->so_rcv);
1383: splx(s);
1384: goto restart;
1385: }
1.108 perry 1386:
1.144 dyoung 1387: if (flagsp != NULL)
1.1 cgd 1388: *flagsp |= flags;
1.54 lukem 1389: release:
1.1 cgd 1390: sbunlock(&so->so_rcv);
1391: splx(s);
1.144 dyoung 1392: return error;
1.1 cgd 1393: }
1394:
1.14 mycroft 1395: int
1.54 lukem 1396: soshutdown(struct socket *so, int how)
1.1 cgd 1397: {
1.99 matt 1398: const struct protosw *pr;
1.34 kleink 1399:
1.54 lukem 1400: pr = so->so_proto;
1.34 kleink 1401: if (!(how == SHUT_RD || how == SHUT_WR || how == SHUT_RDWR))
1402: return (EINVAL);
1.1 cgd 1403:
1.34 kleink 1404: if (how == SHUT_RD || how == SHUT_RDWR)
1.1 cgd 1405: sorflush(so);
1.34 kleink 1406: if (how == SHUT_WR || how == SHUT_RDWR)
1.140 dyoung 1407: return (*pr->pr_usrreq)(so, PRU_SHUTDOWN, NULL,
1408: NULL, NULL, NULL);
1.144 dyoung 1409: return 0;
1.1 cgd 1410: }
1411:
1.14 mycroft 1412: void
1.54 lukem 1413: sorflush(struct socket *so)
1.1 cgd 1414: {
1.54 lukem 1415: struct sockbuf *sb, asb;
1.99 matt 1416: const struct protosw *pr;
1.54 lukem 1417: int s;
1.1 cgd 1418:
1.54 lukem 1419: sb = &so->so_rcv;
1420: pr = so->so_proto;
1.1 cgd 1421: sb->sb_flags |= SB_NOINTR;
1.15 mycroft 1422: (void) sblock(sb, M_WAITOK);
1.56 thorpej 1423: s = splnet();
1.1 cgd 1424: socantrcvmore(so);
1425: sbunlock(sb);
1426: asb = *sb;
1.86 wrstuden 1427: /*
1428: * Clear most of the sockbuf structure, but leave some of the
1429: * fields valid.
1430: */
1431: memset(&sb->sb_startzero, 0,
1432: sizeof(*sb) - offsetof(struct sockbuf, sb_startzero));
1.1 cgd 1433: splx(s);
1434: if (pr->pr_flags & PR_RIGHTS && pr->pr_domain->dom_dispose)
1435: (*pr->pr_domain->dom_dispose)(asb.sb_mb);
1.98 christos 1436: sbrelease(&asb, so);
1.1 cgd 1437: }
1438:
1.142 dyoung 1439: static int
1440: sosetopt1(struct socket *so, int level, int optname, struct mbuf *m)
1.1 cgd 1441: {
1.142 dyoung 1442: int optval, val;
1.130 christos 1443: struct linger *l;
1.141 yamt 1444: struct sockbuf *sb;
1.142 dyoung 1445: struct timeval *tv;
1446:
1447: switch (optname) {
1448:
1449: case SO_LINGER:
1450: if (m == NULL || m->m_len != sizeof(struct linger))
1451: return EINVAL;
1452: l = mtod(m, struct linger *);
1453: if (l->l_linger < 0 || l->l_linger > USHRT_MAX ||
1454: l->l_linger > (INT_MAX / hz))
1455: return EDOM;
1456: so->so_linger = l->l_linger;
1457: if (l->l_onoff)
1458: so->so_options |= SO_LINGER;
1459: else
1460: so->so_options &= ~SO_LINGER;
1461: break;
1.1 cgd 1462:
1.142 dyoung 1463: case SO_DEBUG:
1464: case SO_KEEPALIVE:
1465: case SO_DONTROUTE:
1466: case SO_USELOOPBACK:
1467: case SO_BROADCAST:
1468: case SO_REUSEADDR:
1469: case SO_REUSEPORT:
1470: case SO_OOBINLINE:
1471: case SO_TIMESTAMP:
1472: if (m == NULL || m->m_len < sizeof(int))
1473: return EINVAL;
1474: if (*mtod(m, int *))
1475: so->so_options |= optname;
1476: else
1477: so->so_options &= ~optname;
1478: break;
1479:
1480: case SO_SNDBUF:
1481: case SO_RCVBUF:
1482: case SO_SNDLOWAT:
1483: case SO_RCVLOWAT:
1484: if (m == NULL || m->m_len < sizeof(int))
1485: return EINVAL;
1.1 cgd 1486:
1.142 dyoung 1487: /*
1488: * Values < 1 make no sense for any of these
1489: * options, so disallow them.
1490: */
1491: optval = *mtod(m, int *);
1492: if (optval < 1)
1493: return EINVAL;
1.1 cgd 1494:
1.142 dyoung 1495: switch (optname) {
1.1 cgd 1496:
1497: case SO_SNDBUF:
1498: case SO_RCVBUF:
1.142 dyoung 1499: sb = (optname == SO_SNDBUF) ?
1500: &so->so_snd : &so->so_rcv;
1501: if (sbreserve(sb, (u_long)optval, so) == 0)
1502: return ENOBUFS;
1503: sb->sb_flags &= ~SB_AUTOSIZE;
1504: break;
1505:
1506: /*
1507: * Make sure the low-water is never greater than
1508: * the high-water.
1509: */
1.1 cgd 1510: case SO_SNDLOWAT:
1.142 dyoung 1511: so->so_snd.sb_lowat =
1512: (optval > so->so_snd.sb_hiwat) ?
1513: so->so_snd.sb_hiwat : optval;
1514: break;
1.1 cgd 1515: case SO_RCVLOWAT:
1.142 dyoung 1516: so->so_rcv.sb_lowat =
1517: (optval > so->so_rcv.sb_hiwat) ?
1518: so->so_rcv.sb_hiwat : optval;
1519: break;
1520: }
1521: break;
1.28 thorpej 1522:
1.142 dyoung 1523: case SO_SNDTIMEO:
1524: case SO_RCVTIMEO:
1525: if (m == NULL || m->m_len < sizeof(*tv))
1526: return EINVAL;
1527: tv = mtod(m, struct timeval *);
1528: if (tv->tv_sec > (INT_MAX - tv->tv_usec / tick) / hz)
1529: return EDOM;
1530: val = tv->tv_sec * hz + tv->tv_usec / tick;
1531: if (val == 0 && tv->tv_usec != 0)
1532: val = 1;
1.28 thorpej 1533:
1.142 dyoung 1534: switch (optname) {
1.28 thorpej 1535:
1.142 dyoung 1536: case SO_SNDTIMEO:
1537: so->so_snd.sb_timeo = val;
1.1 cgd 1538: break;
1539: case SO_RCVTIMEO:
1.142 dyoung 1540: so->so_rcv.sb_timeo = val;
1541: break;
1542: }
1543: break;
1.1 cgd 1544:
1.142 dyoung 1545: default:
1546: return ENOPROTOOPT;
1547: }
1548: return 0;
1549: }
1.1 cgd 1550:
1.142 dyoung 1551: int
1552: sosetopt(struct socket *so, int level, int optname, struct mbuf *m)
1553: {
1554: int error, prerr;
1.1 cgd 1555:
1.142 dyoung 1556: if (level == SOL_SOCKET)
1557: error = sosetopt1(so, level, optname, m);
1558: else
1559: error = ENOPROTOOPT;
1.1 cgd 1560:
1.142 dyoung 1561: if ((error == 0 || error == ENOPROTOOPT) &&
1562: so->so_proto != NULL && so->so_proto->pr_ctloutput != NULL) {
1563: /* give the protocol stack a shot */
1564: prerr = (*so->so_proto->pr_ctloutput)(PRCO_SETOPT, so, level,
1565: optname, &m);
1566: if (prerr == 0)
1567: error = 0;
1568: else if (prerr != ENOPROTOOPT)
1569: error = prerr;
1570: } else if (m != NULL)
1571: (void)m_free(m);
1572: return error;
1.1 cgd 1573: }
1574:
1.14 mycroft 1575: int
1.54 lukem 1576: sogetopt(struct socket *so, int level, int optname, struct mbuf **mp)
1.1 cgd 1577: {
1.54 lukem 1578: struct mbuf *m;
1.1 cgd 1579:
1580: if (level != SOL_SOCKET) {
1581: if (so->so_proto && so->so_proto->pr_ctloutput) {
1582: return ((*so->so_proto->pr_ctloutput)
1583: (PRCO_GETOPT, so, level, optname, mp));
1584: } else
1585: return (ENOPROTOOPT);
1586: } else {
1587: m = m_get(M_WAIT, MT_SOOPTS);
1.36 perry 1588: m->m_len = sizeof(int);
1.1 cgd 1589:
1590: switch (optname) {
1591:
1592: case SO_LINGER:
1.36 perry 1593: m->m_len = sizeof(struct linger);
1.1 cgd 1594: mtod(m, struct linger *)->l_onoff =
1.131 christos 1595: (so->so_options & SO_LINGER) ? 1 : 0;
1.1 cgd 1596: mtod(m, struct linger *)->l_linger = so->so_linger;
1597: break;
1598:
1599: case SO_USELOOPBACK:
1600: case SO_DONTROUTE:
1601: case SO_DEBUG:
1602: case SO_KEEPALIVE:
1603: case SO_REUSEADDR:
1.15 mycroft 1604: case SO_REUSEPORT:
1.1 cgd 1605: case SO_BROADCAST:
1606: case SO_OOBINLINE:
1.26 thorpej 1607: case SO_TIMESTAMP:
1.131 christos 1608: *mtod(m, int *) = (so->so_options & optname) ? 1 : 0;
1.1 cgd 1609: break;
1610:
1611: case SO_TYPE:
1612: *mtod(m, int *) = so->so_type;
1613: break;
1614:
1615: case SO_ERROR:
1616: *mtod(m, int *) = so->so_error;
1617: so->so_error = 0;
1618: break;
1619:
1620: case SO_SNDBUF:
1621: *mtod(m, int *) = so->so_snd.sb_hiwat;
1622: break;
1623:
1624: case SO_RCVBUF:
1625: *mtod(m, int *) = so->so_rcv.sb_hiwat;
1626: break;
1627:
1628: case SO_SNDLOWAT:
1629: *mtod(m, int *) = so->so_snd.sb_lowat;
1630: break;
1631:
1632: case SO_RCVLOWAT:
1633: *mtod(m, int *) = so->so_rcv.sb_lowat;
1634: break;
1635:
1636: case SO_SNDTIMEO:
1637: case SO_RCVTIMEO:
1638: {
1639: int val = (optname == SO_SNDTIMEO ?
1640: so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
1641:
1642: m->m_len = sizeof(struct timeval);
1643: mtod(m, struct timeval *)->tv_sec = val / hz;
1644: mtod(m, struct timeval *)->tv_usec =
1.27 kleink 1645: (val % hz) * tick;
1.1 cgd 1646: break;
1647: }
1648:
1.107 darrenr 1649: case SO_OVERFLOWED:
1650: *mtod(m, int *) = so->so_rcv.sb_overflowed;
1651: break;
1652:
1.1 cgd 1653: default:
1654: (void)m_free(m);
1655: return (ENOPROTOOPT);
1656: }
1657: *mp = m;
1658: return (0);
1659: }
1660: }
1661:
1.14 mycroft 1662: void
1.54 lukem 1663: sohasoutofband(struct socket *so)
1.1 cgd 1664: {
1.153 rmind 1665:
1.90 christos 1666: fownsignal(so->so_pgid, SIGURG, POLL_PRI, POLLPRI|POLLRDBAND, so);
1.153 rmind 1667: selnotify(&so->so_rcv.sb_sel, POLLPRI | POLLRDBAND, 0);
1.1 cgd 1668: }
1.72 jdolecek 1669:
1670: static void
1671: filt_sordetach(struct knote *kn)
1672: {
1673: struct socket *so;
1674:
1.155 ad 1675: so = ((file_t *)kn->kn_obj)->f_data;
1.73 christos 1676: SLIST_REMOVE(&so->so_rcv.sb_sel.sel_klist, kn, knote, kn_selnext);
1677: if (SLIST_EMPTY(&so->so_rcv.sb_sel.sel_klist))
1.72 jdolecek 1678: so->so_rcv.sb_flags &= ~SB_KNOTE;
1679: }
1680:
1681: /*ARGSUSED*/
1682: static int
1.129 yamt 1683: filt_soread(struct knote *kn, long hint)
1.72 jdolecek 1684: {
1685: struct socket *so;
1686:
1.155 ad 1687: so = ((file_t *)kn->kn_obj)->f_data;
1.72 jdolecek 1688: kn->kn_data = so->so_rcv.sb_cc;
1689: if (so->so_state & SS_CANTRCVMORE) {
1.108 perry 1690: kn->kn_flags |= EV_EOF;
1.72 jdolecek 1691: kn->kn_fflags = so->so_error;
1692: return (1);
1693: }
1694: if (so->so_error) /* temporary udp error */
1695: return (1);
1696: if (kn->kn_sfflags & NOTE_LOWAT)
1697: return (kn->kn_data >= kn->kn_sdata);
1698: return (kn->kn_data >= so->so_rcv.sb_lowat);
1699: }
1700:
1701: static void
1702: filt_sowdetach(struct knote *kn)
1703: {
1704: struct socket *so;
1705:
1.155 ad 1706: so = ((file_t *)kn->kn_obj)->f_data;
1.73 christos 1707: SLIST_REMOVE(&so->so_snd.sb_sel.sel_klist, kn, knote, kn_selnext);
1708: if (SLIST_EMPTY(&so->so_snd.sb_sel.sel_klist))
1.72 jdolecek 1709: so->so_snd.sb_flags &= ~SB_KNOTE;
1710: }
1711:
1712: /*ARGSUSED*/
1713: static int
1.129 yamt 1714: filt_sowrite(struct knote *kn, long hint)
1.72 jdolecek 1715: {
1716: struct socket *so;
1717:
1.155 ad 1718: so = ((file_t *)kn->kn_obj)->f_data;
1.72 jdolecek 1719: kn->kn_data = sbspace(&so->so_snd);
1720: if (so->so_state & SS_CANTSENDMORE) {
1.108 perry 1721: kn->kn_flags |= EV_EOF;
1.72 jdolecek 1722: kn->kn_fflags = so->so_error;
1723: return (1);
1724: }
1725: if (so->so_error) /* temporary udp error */
1726: return (1);
1727: if (((so->so_state & SS_ISCONNECTED) == 0) &&
1728: (so->so_proto->pr_flags & PR_CONNREQUIRED))
1729: return (0);
1730: if (kn->kn_sfflags & NOTE_LOWAT)
1731: return (kn->kn_data >= kn->kn_sdata);
1732: return (kn->kn_data >= so->so_snd.sb_lowat);
1733: }
1734:
1735: /*ARGSUSED*/
1736: static int
1.129 yamt 1737: filt_solisten(struct knote *kn, long hint)
1.72 jdolecek 1738: {
1739: struct socket *so;
1740:
1.155 ad 1741: so = ((file_t *)kn->kn_obj)->f_data;
1.72 jdolecek 1742:
1743: /*
1744: * Set kn_data to number of incoming connections, not
1745: * counting partial (incomplete) connections.
1.108 perry 1746: */
1.72 jdolecek 1747: kn->kn_data = so->so_qlen;
1748: return (kn->kn_data > 0);
1749: }
1750:
1751: static const struct filterops solisten_filtops =
1752: { 1, NULL, filt_sordetach, filt_solisten };
1753: static const struct filterops soread_filtops =
1754: { 1, NULL, filt_sordetach, filt_soread };
1755: static const struct filterops sowrite_filtops =
1756: { 1, NULL, filt_sowdetach, filt_sowrite };
1757:
1758: int
1.129 yamt 1759: soo_kqfilter(struct file *fp, struct knote *kn)
1.72 jdolecek 1760: {
1761: struct socket *so;
1762: struct sockbuf *sb;
1763:
1.155 ad 1764: so = ((file_t *)kn->kn_obj)->f_data;
1.72 jdolecek 1765: switch (kn->kn_filter) {
1766: case EVFILT_READ:
1767: if (so->so_options & SO_ACCEPTCONN)
1768: kn->kn_fop = &solisten_filtops;
1769: else
1770: kn->kn_fop = &soread_filtops;
1771: sb = &so->so_rcv;
1772: break;
1773: case EVFILT_WRITE:
1774: kn->kn_fop = &sowrite_filtops;
1775: sb = &so->so_snd;
1776: break;
1777: default:
1.149 pooka 1778: return (EINVAL);
1.72 jdolecek 1779: }
1.73 christos 1780: SLIST_INSERT_HEAD(&sb->sb_sel.sel_klist, kn, kn_selnext);
1.72 jdolecek 1781: sb->sb_flags |= SB_KNOTE;
1782: return (0);
1783: }
1784:
1.154 ad 1785: static int
1786: sodopoll(struct socket *so, int events)
1787: {
1788: int revents;
1789:
1790: revents = 0;
1791:
1792: if (events & (POLLIN | POLLRDNORM))
1793: if (soreadable(so))
1794: revents |= events & (POLLIN | POLLRDNORM);
1795:
1796: if (events & (POLLOUT | POLLWRNORM))
1797: if (sowritable(so))
1798: revents |= events & (POLLOUT | POLLWRNORM);
1799:
1800: if (events & (POLLPRI | POLLRDBAND))
1801: if (so->so_oobmark || (so->so_state & SS_RCVATMARK))
1802: revents |= events & (POLLPRI | POLLRDBAND);
1803:
1804: return revents;
1805: }
1806:
1807: int
1808: sopoll(struct socket *so, int events)
1809: {
1810: int revents = 0;
1811: int s;
1812:
1813: if ((revents = sodopoll(so, events)) != 0)
1814: return revents;
1815:
1816: KERNEL_LOCK(1, curlwp);
1817: s = splsoftnet();
1818:
1819: if ((revents = sodopoll(so, events)) == 0) {
1820: if (events & (POLLIN | POLLPRI | POLLRDNORM | POLLRDBAND)) {
1821: selrecord(curlwp, &so->so_rcv.sb_sel);
1822: so->so_rcv.sb_flags |= SB_SEL;
1823: }
1824:
1825: if (events & (POLLOUT | POLLWRNORM)) {
1826: selrecord(curlwp, &so->so_snd.sb_sel);
1827: so->so_snd.sb_flags |= SB_SEL;
1828: }
1829: }
1830:
1831: splx(s);
1832: KERNEL_UNLOCK_ONE(curlwp);
1833:
1834: return revents;
1835: }
1836:
1837:
1.94 yamt 1838: #include <sys/sysctl.h>
1839:
1840: static int sysctl_kern_somaxkva(SYSCTLFN_PROTO);
1841:
1842: /*
1843: * sysctl helper routine for kern.somaxkva. ensures that the given
1844: * value is not too small.
1845: * (XXX should we maybe make sure it's not too large as well?)
1846: */
1847: static int
1848: sysctl_kern_somaxkva(SYSCTLFN_ARGS)
1849: {
1850: int error, new_somaxkva;
1851: struct sysctlnode node;
1852:
1853: new_somaxkva = somaxkva;
1854: node = *rnode;
1855: node.sysctl_data = &new_somaxkva;
1856: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1857: if (error || newp == NULL)
1858: return (error);
1859:
1860: if (new_somaxkva < (16 * 1024 * 1024)) /* sanity */
1861: return (EINVAL);
1862:
1.136 ad 1863: mutex_enter(&so_pendfree_lock);
1.94 yamt 1864: somaxkva = new_somaxkva;
1.136 ad 1865: cv_broadcast(&socurkva_cv);
1866: mutex_exit(&so_pendfree_lock);
1.94 yamt 1867:
1868: return (error);
1869: }
1870:
1871: SYSCTL_SETUP(sysctl_kern_somaxkva_setup, "sysctl kern.somaxkva setup")
1872: {
1873:
1.97 atatat 1874: sysctl_createv(clog, 0, NULL, NULL,
1875: CTLFLAG_PERMANENT,
1876: CTLTYPE_NODE, "kern", NULL,
1877: NULL, 0, NULL, 0,
1878: CTL_KERN, CTL_EOL);
1879:
1880: sysctl_createv(clog, 0, NULL, NULL,
1881: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.103 atatat 1882: CTLTYPE_INT, "somaxkva",
1883: SYSCTL_DESCR("Maximum amount of kernel memory to be "
1884: "used for socket buffers"),
1.94 yamt 1885: sysctl_kern_somaxkva, 0, NULL, 0,
1886: CTL_KERN, KERN_SOMAXKVA, CTL_EOL);
1887: }
CVSweb <webmaster@jp.NetBSD.org>