Annotation of src/sys/kern/uipc_socket.c, Revision 1.208
1.208 ! para 1: /* $NetBSD: uipc_socket.c,v 1.207 2012/01/25 00:28:36 christos Exp $ */
1.64 thorpej 2:
3: /*-
1.188 ad 4: * Copyright (c) 2002, 2007, 2008, 2009 The NetBSD Foundation, Inc.
1.64 thorpej 5: * All rights reserved.
6: *
7: * This code is derived from software contributed to The NetBSD Foundation
1.188 ad 8: * by Jason R. Thorpe of Wasabi Systems, Inc, and by Andrew Doran.
1.64 thorpej 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: *
19: * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20: * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21: * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22: * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23: * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24: * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25: * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26: * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27: * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28: * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29: * POSSIBILITY OF SUCH DAMAGE.
30: */
1.16 cgd 31:
1.1 cgd 32: /*
1.159 ad 33: * Copyright (c) 2004 The FreeBSD Foundation
34: * Copyright (c) 2004 Robert Watson
1.15 mycroft 35: * Copyright (c) 1982, 1986, 1988, 1990, 1993
36: * The Regents of the University of California. All rights reserved.
1.1 cgd 37: *
38: * Redistribution and use in source and binary forms, with or without
39: * modification, are permitted provided that the following conditions
40: * are met:
41: * 1. Redistributions of source code must retain the above copyright
42: * notice, this list of conditions and the following disclaimer.
43: * 2. Redistributions in binary form must reproduce the above copyright
44: * notice, this list of conditions and the following disclaimer in the
45: * documentation and/or other materials provided with the distribution.
1.85 agc 46: * 3. Neither the name of the University nor the names of its contributors
1.1 cgd 47: * may be used to endorse or promote products derived from this software
48: * without specific prior written permission.
49: *
50: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
51: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
52: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
53: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
54: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
55: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
56: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
57: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
58: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
59: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
60: * SUCH DAMAGE.
61: *
1.32 fvdl 62: * @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95
1.1 cgd 63: */
1.59 lukem 64:
65: #include <sys/cdefs.h>
1.208 ! para 66: __KERNEL_RCSID(0, "$NetBSD: uipc_socket.c,v 1.207 2012/01/25 00:28:36 christos Exp $");
1.64 thorpej 67:
1.179 christos 68: #include "opt_compat_netbsd.h"
1.64 thorpej 69: #include "opt_sock_counters.h"
70: #include "opt_sosend_loan.h"
1.81 martin 71: #include "opt_mbuftrace.h"
1.84 ragge 72: #include "opt_somaxkva.h"
1.167 ad 73: #include "opt_multiprocessor.h" /* XXX */
1.1 cgd 74:
1.9 mycroft 75: #include <sys/param.h>
76: #include <sys/systm.h>
77: #include <sys/proc.h>
78: #include <sys/file.h>
1.142 dyoung 79: #include <sys/filedesc.h>
1.173 plunky 80: #include <sys/kmem.h>
1.9 mycroft 81: #include <sys/mbuf.h>
82: #include <sys/domain.h>
83: #include <sys/kernel.h>
84: #include <sys/protosw.h>
85: #include <sys/socket.h>
86: #include <sys/socketvar.h>
1.21 christos 87: #include <sys/signalvar.h>
1.9 mycroft 88: #include <sys/resourcevar.h>
1.174 pooka 89: #include <sys/uidinfo.h>
1.72 jdolecek 90: #include <sys/event.h>
1.89 christos 91: #include <sys/poll.h>
1.118 elad 92: #include <sys/kauth.h>
1.136 ad 93: #include <sys/mutex.h>
94: #include <sys/condvar.h>
1.205 bouyer 95: #include <sys/kthread.h>
1.37 thorpej 96:
1.179 christos 97: #ifdef COMPAT_50
98: #include <compat/sys/time.h>
1.184 christos 99: #include <compat/sys/socket.h>
1.179 christos 100: #endif
101:
1.202 uebayasi 102: #include <uvm/uvm_extern.h>
103: #include <uvm/uvm_loan.h>
104: #include <uvm/uvm_page.h>
1.64 thorpej 105:
1.77 thorpej 106: MALLOC_DEFINE(M_SOOPTS, "soopts", "socket options");
107: MALLOC_DEFINE(M_SONAME, "soname", "socket name");
1.37 thorpej 108:
1.142 dyoung 109: extern const struct fileops socketops;
110:
1.54 lukem 111: extern int somaxconn; /* patchable (XXX sysctl) */
112: int somaxconn = SOMAXCONN;
1.160 ad 113: kmutex_t *softnet_lock;
1.49 jonathan 114:
1.64 thorpej 115: #ifdef SOSEND_COUNTERS
116: #include <sys/device.h>
117:
1.113 thorpej 118: static struct evcnt sosend_loan_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 119: NULL, "sosend", "loan big");
1.113 thorpej 120: static struct evcnt sosend_copy_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 121: NULL, "sosend", "copy big");
1.113 thorpej 122: static struct evcnt sosend_copy_small = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 123: NULL, "sosend", "copy small");
1.113 thorpej 124: static struct evcnt sosend_kvalimit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
1.64 thorpej 125: NULL, "sosend", "kva limit");
126:
127: #define SOSEND_COUNTER_INCR(ev) (ev)->ev_count++
128:
1.101 matt 129: EVCNT_ATTACH_STATIC(sosend_loan_big);
130: EVCNT_ATTACH_STATIC(sosend_copy_big);
131: EVCNT_ATTACH_STATIC(sosend_copy_small);
132: EVCNT_ATTACH_STATIC(sosend_kvalimit);
1.64 thorpej 133: #else
134:
135: #define SOSEND_COUNTER_INCR(ev) /* nothing */
136:
137: #endif /* SOSEND_COUNTERS */
138:
1.167 ad 139: #if defined(SOSEND_NO_LOAN) || defined(MULTIPROCESSOR)
1.121 yamt 140: int sock_loan_thresh = -1;
1.71 thorpej 141: #else
1.121 yamt 142: int sock_loan_thresh = 4096;
1.65 thorpej 143: #endif
1.64 thorpej 144:
1.136 ad 145: static kmutex_t so_pendfree_lock;
1.205 bouyer 146: static struct mbuf *so_pendfree = NULL;
1.64 thorpej 147:
1.84 ragge 148: #ifndef SOMAXKVA
149: #define SOMAXKVA (16 * 1024 * 1024)
150: #endif
151: int somaxkva = SOMAXKVA;
1.113 thorpej 152: static int socurkva;
1.136 ad 153: static kcondvar_t socurkva_cv;
1.64 thorpej 154:
1.191 elad 155: static kauth_listener_t socket_listener;
156:
1.64 thorpej 157: #define SOCK_LOAN_CHUNK 65536
158:
1.205 bouyer 159: static void sopendfree_thread(void *);
160: static kcondvar_t pendfree_thread_cv;
161: static lwp_t *sopendfree_lwp;
1.93 yamt 162:
1.178 pooka 163: static void sysctl_kern_somaxkva_setup(void);
164: static struct sysctllog *socket_sysctllog;
165:
1.113 thorpej 166: static vsize_t
1.129 yamt 167: sokvareserve(struct socket *so, vsize_t len)
1.80 yamt 168: {
1.98 christos 169: int error;
1.80 yamt 170:
1.136 ad 171: mutex_enter(&so_pendfree_lock);
1.80 yamt 172: while (socurkva + len > somaxkva) {
173: SOSEND_COUNTER_INCR(&sosend_kvalimit);
1.136 ad 174: error = cv_wait_sig(&socurkva_cv, &so_pendfree_lock);
1.98 christos 175: if (error) {
176: len = 0;
177: break;
178: }
1.80 yamt 179: }
1.93 yamt 180: socurkva += len;
1.136 ad 181: mutex_exit(&so_pendfree_lock);
1.98 christos 182: return len;
1.95 yamt 183: }
184:
1.113 thorpej 185: static void
1.95 yamt 186: sokvaunreserve(vsize_t len)
187: {
188:
1.136 ad 189: mutex_enter(&so_pendfree_lock);
1.95 yamt 190: socurkva -= len;
1.136 ad 191: cv_broadcast(&socurkva_cv);
192: mutex_exit(&so_pendfree_lock);
1.95 yamt 193: }
194:
195: /*
196: * sokvaalloc: allocate kva for loan.
197: */
198:
199: vaddr_t
200: sokvaalloc(vsize_t len, struct socket *so)
201: {
202: vaddr_t lva;
203:
204: /*
205: * reserve kva.
206: */
207:
1.98 christos 208: if (sokvareserve(so, len) == 0)
209: return 0;
1.93 yamt 210:
211: /*
212: * allocate kva.
213: */
1.80 yamt 214:
1.109 yamt 215: lva = uvm_km_alloc(kernel_map, len, 0, UVM_KMF_VAONLY | UVM_KMF_WAITVA);
1.95 yamt 216: if (lva == 0) {
217: sokvaunreserve(len);
1.80 yamt 218: return (0);
1.95 yamt 219: }
1.80 yamt 220:
221: return lva;
222: }
223:
1.93 yamt 224: /*
225: * sokvafree: free kva for loan.
226: */
227:
1.80 yamt 228: void
229: sokvafree(vaddr_t sva, vsize_t len)
230: {
1.93 yamt 231:
232: /*
233: * free kva.
234: */
1.80 yamt 235:
1.109 yamt 236: uvm_km_free(kernel_map, sva, len, UVM_KMF_VAONLY);
1.93 yamt 237:
238: /*
239: * unreserve kva.
240: */
241:
1.95 yamt 242: sokvaunreserve(len);
1.80 yamt 243: }
244:
1.64 thorpej 245: static void
1.134 christos 246: sodoloanfree(struct vm_page **pgs, void *buf, size_t size)
1.64 thorpej 247: {
1.156 yamt 248: vaddr_t sva, eva;
1.64 thorpej 249: vsize_t len;
1.156 yamt 250: int npgs;
251:
252: KASSERT(pgs != NULL);
1.64 thorpej 253:
254: eva = round_page((vaddr_t) buf + size);
255: sva = trunc_page((vaddr_t) buf);
256: len = eva - sva;
257: npgs = len >> PAGE_SHIFT;
258:
259: pmap_kremove(sva, len);
260: pmap_update(pmap_kernel());
261: uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE);
1.80 yamt 262: sokvafree(sva, len);
1.64 thorpej 263: }
264:
1.93 yamt 265: /*
1.205 bouyer 266: * sopendfree_thread: free mbufs on "pendfree" list.
1.136 ad 267: * unlock and relock so_pendfree_lock when freeing mbufs.
1.93 yamt 268: */
269:
1.205 bouyer 270: static void
271: sopendfree_thread(void *v)
1.93 yamt 272: {
1.137 ad 273: struct mbuf *m, *next;
1.205 bouyer 274: size_t rv;
1.93 yamt 275:
1.205 bouyer 276: mutex_enter(&so_pendfree_lock);
1.64 thorpej 277:
1.205 bouyer 278: for (;;) {
279: rv = 0;
280: while (so_pendfree != NULL) {
281: m = so_pendfree;
282: so_pendfree = NULL;
283: mutex_exit(&so_pendfree_lock);
284:
285: for (; m != NULL; m = next) {
286: next = m->m_next;
287: KASSERT((~m->m_flags & (M_EXT|M_EXT_PAGES)) == 0);
288: KASSERT(m->m_ext.ext_refcnt == 0);
289:
290: rv += m->m_ext.ext_size;
291: sodoloanfree(m->m_ext.ext_pgs, m->m_ext.ext_buf,
292: m->m_ext.ext_size);
293: pool_cache_put(mb_cache, m);
294: }
1.93 yamt 295:
1.205 bouyer 296: mutex_enter(&so_pendfree_lock);
1.93 yamt 297: }
1.205 bouyer 298: if (rv)
299: cv_broadcast(&socurkva_cv);
300: cv_wait(&pendfree_thread_cv, &so_pendfree_lock);
1.64 thorpej 301: }
1.205 bouyer 302: panic("sopendfree_thread");
303: /* NOTREACHED */
1.64 thorpej 304: }
305:
1.80 yamt 306: void
1.134 christos 307: soloanfree(struct mbuf *m, void *buf, size_t size, void *arg)
1.64 thorpej 308: {
309:
1.156 yamt 310: KASSERT(m != NULL);
1.64 thorpej 311:
1.93 yamt 312: /*
313: * postpone freeing mbuf.
314: *
315: * we can't do it in interrupt context
316: * because we need to put kva back to kernel_map.
317: */
318:
1.136 ad 319: mutex_enter(&so_pendfree_lock);
1.92 yamt 320: m->m_next = so_pendfree;
321: so_pendfree = m;
1.205 bouyer 322: cv_signal(&pendfree_thread_cv);
1.136 ad 323: mutex_exit(&so_pendfree_lock);
1.64 thorpej 324: }
325:
326: static long
327: sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space)
328: {
329: struct iovec *iov = uio->uio_iov;
330: vaddr_t sva, eva;
331: vsize_t len;
1.156 yamt 332: vaddr_t lva;
333: int npgs, error;
334: vaddr_t va;
335: int i;
1.64 thorpej 336:
1.116 yamt 337: if (VMSPACE_IS_KERNEL_P(uio->uio_vmspace))
1.64 thorpej 338: return (0);
339:
340: if (iov->iov_len < (size_t) space)
341: space = iov->iov_len;
342: if (space > SOCK_LOAN_CHUNK)
343: space = SOCK_LOAN_CHUNK;
344:
345: eva = round_page((vaddr_t) iov->iov_base + space);
346: sva = trunc_page((vaddr_t) iov->iov_base);
347: len = eva - sva;
348: npgs = len >> PAGE_SHIFT;
349:
1.79 thorpej 350: KASSERT(npgs <= M_EXT_MAXPAGES);
351:
1.80 yamt 352: lva = sokvaalloc(len, so);
1.64 thorpej 353: if (lva == 0)
1.80 yamt 354: return 0;
1.64 thorpej 355:
1.116 yamt 356: error = uvm_loan(&uio->uio_vmspace->vm_map, sva, len,
1.79 thorpej 357: m->m_ext.ext_pgs, UVM_LOAN_TOPAGE);
1.64 thorpej 358: if (error) {
1.80 yamt 359: sokvafree(lva, len);
1.64 thorpej 360: return (0);
361: }
362:
363: for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE)
1.79 thorpej 364: pmap_kenter_pa(va, VM_PAGE_TO_PHYS(m->m_ext.ext_pgs[i]),
1.194 cegger 365: VM_PROT_READ, 0);
1.64 thorpej 366: pmap_update(pmap_kernel());
367:
368: lva += (vaddr_t) iov->iov_base & PAGE_MASK;
369:
1.134 christos 370: MEXTADD(m, (void *) lva, space, M_MBUF, soloanfree, so);
1.79 thorpej 371: m->m_flags |= M_EXT_PAGES | M_EXT_ROMAP;
1.64 thorpej 372:
373: uio->uio_resid -= space;
374: /* uio_offset not updated, not set/used for write(2) */
1.134 christos 375: uio->uio_iov->iov_base = (char *)uio->uio_iov->iov_base + space;
1.64 thorpej 376: uio->uio_iov->iov_len -= space;
377: if (uio->uio_iov->iov_len == 0) {
378: uio->uio_iov++;
379: uio->uio_iovcnt--;
380: }
381:
382: return (space);
383: }
384:
1.142 dyoung 385: struct mbuf *
1.147 dyoung 386: getsombuf(struct socket *so, int type)
1.142 dyoung 387: {
388: struct mbuf *m;
389:
1.147 dyoung 390: m = m_get(M_WAIT, type);
1.142 dyoung 391: MCLAIM(m, so->so_mowner);
392: return m;
393: }
394:
1.191 elad 395: static int
396: socket_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie,
397: void *arg0, void *arg1, void *arg2, void *arg3)
398: {
399: int result;
400: enum kauth_network_req req;
401:
402: result = KAUTH_RESULT_DEFER;
403: req = (enum kauth_network_req)arg0;
404:
1.193 elad 405: if ((action != KAUTH_NETWORK_SOCKET) &&
406: (action != KAUTH_NETWORK_BIND))
1.191 elad 407: return result;
408:
409: switch (req) {
1.193 elad 410: case KAUTH_REQ_NETWORK_BIND_PORT:
411: result = KAUTH_RESULT_ALLOW;
412: break;
413:
1.191 elad 414: case KAUTH_REQ_NETWORK_SOCKET_DROP: {
415: /* Normal users can only drop their own connections. */
416: struct socket *so = (struct socket *)arg1;
417:
1.199 elad 418: if (proc_uidmatch(cred, so->so_cred))
1.191 elad 419: result = KAUTH_RESULT_ALLOW;
420:
421: break;
422: }
423:
424: case KAUTH_REQ_NETWORK_SOCKET_OPEN:
425: /* We allow "raw" routing/bluetooth sockets to anyone. */
1.203 matt 426: if ((u_long)arg1 == PF_ROUTE || (u_long)arg1 == PF_OROUTE
427: || (u_long)arg1 == PF_BLUETOOTH) {
1.191 elad 428: result = KAUTH_RESULT_ALLOW;
1.203 matt 429: } else {
1.191 elad 430: /* Privileged, let secmodel handle this. */
431: if ((u_long)arg2 == SOCK_RAW)
432: break;
433: }
434:
435: result = KAUTH_RESULT_ALLOW;
436:
437: break;
438:
1.192 elad 439: case KAUTH_REQ_NETWORK_SOCKET_CANSEE:
440: result = KAUTH_RESULT_ALLOW;
441:
442: break;
443:
1.191 elad 444: default:
445: break;
446: }
447:
448: return result;
449: }
450:
1.119 yamt 451: void
452: soinit(void)
453: {
454:
1.178 pooka 455: sysctl_kern_somaxkva_setup();
456:
1.148 ad 457: mutex_init(&so_pendfree_lock, MUTEX_DEFAULT, IPL_VM);
1.160 ad 458: softnet_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE);
1.136 ad 459: cv_init(&socurkva_cv, "sokva");
1.205 bouyer 460: cv_init(&pendfree_thread_cv, "sopendfr");
1.166 ad 461: soinit2();
1.136 ad 462:
1.119 yamt 463: /* Set the initial adjusted socket buffer size. */
464: if (sb_max_set(sb_max))
465: panic("bad initial sb_max value: %lu", sb_max);
466:
1.191 elad 467: socket_listener = kauth_listen_scope(KAUTH_SCOPE_NETWORK,
468: socket_listener_cb, NULL);
1.119 yamt 469: }
470:
1.205 bouyer 471: void
472: soinit1(void)
473: {
474: int error = kthread_create(PRI_NONE, KTHREAD_MPSAFE, NULL,
475: sopendfree_thread, NULL, &sopendfree_lwp, "sopendfree");
476: if (error)
477: panic("soinit1 %d", error);
478: }
479:
1.1 cgd 480: /*
481: * Socket operation routines.
482: * These routines are called by the routines in
483: * sys_socket.c or from a system process, and
484: * implement the semantics of socket operations by
485: * switching out to the protocol specific routines.
486: */
487: /*ARGSUSED*/
1.3 andrew 488: int
1.160 ad 489: socreate(int dom, struct socket **aso, int type, int proto, struct lwp *l,
490: struct socket *lockso)
1.1 cgd 491: {
1.99 matt 492: const struct protosw *prp;
1.54 lukem 493: struct socket *so;
1.115 yamt 494: uid_t uid;
1.160 ad 495: int error;
496: kmutex_t *lock;
1.1 cgd 497:
1.132 elad 498: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_SOCKET,
499: KAUTH_REQ_NETWORK_SOCKET_OPEN, KAUTH_ARG(dom), KAUTH_ARG(type),
500: KAUTH_ARG(proto));
1.140 dyoung 501: if (error != 0)
502: return error;
1.127 elad 503:
1.1 cgd 504: if (proto)
505: prp = pffindproto(dom, proto, type);
506: else
507: prp = pffindtype(dom, type);
1.140 dyoung 508: if (prp == NULL) {
1.120 ginsbach 509: /* no support for domain */
510: if (pffinddomain(dom) == 0)
1.140 dyoung 511: return EAFNOSUPPORT;
1.120 ginsbach 512: /* no support for socket type */
513: if (proto == 0 && type != 0)
1.140 dyoung 514: return EPROTOTYPE;
515: return EPROTONOSUPPORT;
1.120 ginsbach 516: }
1.140 dyoung 517: if (prp->pr_usrreq == NULL)
518: return EPROTONOSUPPORT;
1.1 cgd 519: if (prp->pr_type != type)
1.140 dyoung 520: return EPROTOTYPE;
1.160 ad 521:
522: so = soget(true);
1.1 cgd 523: so->so_type = type;
524: so->so_proto = prp;
1.33 matt 525: so->so_send = sosend;
526: so->so_receive = soreceive;
1.78 matt 527: #ifdef MBUFTRACE
528: so->so_rcv.sb_mowner = &prp->pr_domain->dom_mowner;
529: so->so_snd.sb_mowner = &prp->pr_domain->dom_mowner;
530: so->so_mowner = &prp->pr_domain->dom_mowner;
531: #endif
1.138 rmind 532: uid = kauth_cred_geteuid(l->l_cred);
1.115 yamt 533: so->so_uidinfo = uid_find(uid);
1.168 yamt 534: so->so_cpid = l->l_proc->p_pid;
1.160 ad 535: if (lockso != NULL) {
536: /* Caller wants us to share a lock. */
537: lock = lockso->so_lock;
538: so->so_lock = lock;
539: mutex_obj_hold(lock);
540: mutex_enter(lock);
541: } else {
542: /* Lock assigned and taken during PRU_ATTACH. */
543: }
1.140 dyoung 544: error = (*prp->pr_usrreq)(so, PRU_ATTACH, NULL,
545: (struct mbuf *)(long)proto, NULL, l);
1.160 ad 546: KASSERT(solocked(so));
1.140 dyoung 547: if (error != 0) {
1.1 cgd 548: so->so_state |= SS_NOFDREF;
549: sofree(so);
1.140 dyoung 550: return error;
1.1 cgd 551: }
1.198 elad 552: so->so_cred = kauth_cred_dup(l->l_cred);
1.160 ad 553: sounlock(so);
1.1 cgd 554: *aso = so;
1.140 dyoung 555: return 0;
1.1 cgd 556: }
557:
1.142 dyoung 558: /* On success, write file descriptor to fdout and return zero. On
559: * failure, return non-zero; *fdout will be undefined.
560: */
561: int
562: fsocreate(int domain, struct socket **sop, int type, int protocol,
563: struct lwp *l, int *fdout)
564: {
565: struct socket *so;
566: struct file *fp;
567: int fd, error;
1.204 christos 568: int flags = type & SOCK_FLAGS_MASK;
1.142 dyoung 569:
1.204 christos 570: type &= ~SOCK_FLAGS_MASK;
1.155 ad 571: if ((error = fd_allocfile(&fp, &fd)) != 0)
1.204 christos 572: return error;
573: fd_set_exclose(l, fd, (flags & SOCK_CLOEXEC) != 0);
1.207 christos 574: fp->f_flag = FREAD|FWRITE|((flags & SOCK_NONBLOCK) ? FNONBLOCK : 0)|
575: ((flags & SOCK_NOSIGPIPE) ? FNOSIGPIPE : 0);
1.142 dyoung 576: fp->f_type = DTYPE_SOCKET;
577: fp->f_ops = &socketops;
1.160 ad 578: error = socreate(domain, &so, type, protocol, l, NULL);
1.142 dyoung 579: if (error != 0) {
1.155 ad 580: fd_abort(curproc, fp, fd);
1.142 dyoung 581: } else {
582: if (sop != NULL)
583: *sop = so;
584: fp->f_data = so;
1.155 ad 585: fd_affix(curproc, fp, fd);
1.142 dyoung 586: *fdout = fd;
587: }
588: return error;
589: }
590:
1.3 andrew 591: int
1.190 dyoung 592: sofamily(const struct socket *so)
593: {
594: const struct protosw *pr;
595: const struct domain *dom;
596:
597: if ((pr = so->so_proto) == NULL)
598: return AF_UNSPEC;
599: if ((dom = pr->pr_domain) == NULL)
600: return AF_UNSPEC;
601: return dom->dom_family;
602: }
603:
604: int
1.114 christos 605: sobind(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1 cgd 606: {
1.160 ad 607: int error;
1.1 cgd 608:
1.160 ad 609: solock(so);
1.140 dyoung 610: error = (*so->so_proto->pr_usrreq)(so, PRU_BIND, NULL, nam, NULL, l);
1.160 ad 611: sounlock(so);
1.140 dyoung 612: return error;
1.1 cgd 613: }
614:
1.3 andrew 615: int
1.150 elad 616: solisten(struct socket *so, int backlog, struct lwp *l)
1.1 cgd 617: {
1.160 ad 618: int error;
1.1 cgd 619:
1.160 ad 620: solock(so);
1.158 ad 621: if ((so->so_state & (SS_ISCONNECTED | SS_ISCONNECTING |
1.163 ad 622: SS_ISDISCONNECTING)) != 0) {
623: sounlock(so);
1.158 ad 624: return (EOPNOTSUPP);
1.163 ad 625: }
1.140 dyoung 626: error = (*so->so_proto->pr_usrreq)(so, PRU_LISTEN, NULL,
1.150 elad 627: NULL, NULL, l);
1.140 dyoung 628: if (error != 0) {
1.160 ad 629: sounlock(so);
1.140 dyoung 630: return error;
1.1 cgd 631: }
1.63 matt 632: if (TAILQ_EMPTY(&so->so_q))
1.1 cgd 633: so->so_options |= SO_ACCEPTCONN;
634: if (backlog < 0)
635: backlog = 0;
1.49 jonathan 636: so->so_qlimit = min(backlog, somaxconn);
1.160 ad 637: sounlock(so);
1.140 dyoung 638: return 0;
1.1 cgd 639: }
640:
1.21 christos 641: void
1.54 lukem 642: sofree(struct socket *so)
1.1 cgd 643: {
1.161 ad 644: u_int refs;
1.1 cgd 645:
1.160 ad 646: KASSERT(solocked(so));
647:
648: if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0) {
649: sounlock(so);
1.1 cgd 650: return;
1.160 ad 651: }
1.43 mycroft 652: if (so->so_head) {
653: /*
654: * We must not decommission a socket that's on the accept(2)
655: * queue. If we do, then accept(2) may hang after select(2)
656: * indicated that the listening socket was ready.
657: */
1.160 ad 658: if (!soqremque(so, 0)) {
659: sounlock(so);
1.43 mycroft 660: return;
1.160 ad 661: }
1.43 mycroft 662: }
1.98 christos 663: if (so->so_rcv.sb_hiwat)
1.110 christos 664: (void)chgsbsize(so->so_uidinfo, &so->so_rcv.sb_hiwat, 0,
1.98 christos 665: RLIM_INFINITY);
666: if (so->so_snd.sb_hiwat)
1.110 christos 667: (void)chgsbsize(so->so_uidinfo, &so->so_snd.sb_hiwat, 0,
1.98 christos 668: RLIM_INFINITY);
669: sbrelease(&so->so_snd, so);
1.160 ad 670: KASSERT(!cv_has_waiters(&so->so_cv));
671: KASSERT(!cv_has_waiters(&so->so_rcv.sb_cv));
672: KASSERT(!cv_has_waiters(&so->so_snd.sb_cv));
1.1 cgd 673: sorflush(so);
1.161 ad 674: refs = so->so_aborting; /* XXX */
1.177 ad 675: /* Remove acccept filter if one is present. */
1.170 tls 676: if (so->so_accf != NULL)
1.177 ad 677: (void)accept_filt_clear(so);
1.160 ad 678: sounlock(so);
1.161 ad 679: if (refs == 0) /* XXX */
680: soput(so);
1.1 cgd 681: }
682:
683: /*
684: * Close a socket on last file table reference removal.
685: * Initiate disconnect if connected.
686: * Free socket when disconnect complete.
687: */
1.3 andrew 688: int
1.54 lukem 689: soclose(struct socket *so)
1.1 cgd 690: {
1.54 lukem 691: struct socket *so2;
1.160 ad 692: int error;
693: int error2;
1.1 cgd 694:
1.54 lukem 695: error = 0;
1.160 ad 696: solock(so);
1.1 cgd 697: if (so->so_options & SO_ACCEPTCONN) {
1.172 ad 698: for (;;) {
699: if ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) {
1.160 ad 700: KASSERT(solocked2(so, so2));
701: (void) soqremque(so2, 0);
702: /* soabort drops the lock. */
703: (void) soabort(so2);
704: solock(so);
1.172 ad 705: continue;
1.160 ad 706: }
1.172 ad 707: if ((so2 = TAILQ_FIRST(&so->so_q)) != 0) {
1.160 ad 708: KASSERT(solocked2(so, so2));
709: (void) soqremque(so2, 1);
710: /* soabort drops the lock. */
711: (void) soabort(so2);
712: solock(so);
1.172 ad 713: continue;
1.160 ad 714: }
1.172 ad 715: break;
716: }
1.1 cgd 717: }
718: if (so->so_pcb == 0)
719: goto discard;
720: if (so->so_state & SS_ISCONNECTED) {
721: if ((so->so_state & SS_ISDISCONNECTING) == 0) {
722: error = sodisconnect(so);
723: if (error)
724: goto drop;
725: }
726: if (so->so_options & SO_LINGER) {
1.206 christos 727: if ((so->so_state & (SS_ISDISCONNECTING|SS_NBIO)) ==
728: (SS_ISDISCONNECTING|SS_NBIO))
1.1 cgd 729: goto drop;
1.21 christos 730: while (so->so_state & SS_ISCONNECTED) {
1.185 yamt 731: error = sowait(so, true, so->so_linger * hz);
1.21 christos 732: if (error)
1.1 cgd 733: break;
1.21 christos 734: }
1.1 cgd 735: }
736: }
1.54 lukem 737: drop:
1.1 cgd 738: if (so->so_pcb) {
1.160 ad 739: error2 = (*so->so_proto->pr_usrreq)(so, PRU_DETACH,
1.140 dyoung 740: NULL, NULL, NULL, NULL);
1.1 cgd 741: if (error == 0)
742: error = error2;
743: }
1.54 lukem 744: discard:
1.1 cgd 745: if (so->so_state & SS_NOFDREF)
746: panic("soclose: NOFDREF");
1.198 elad 747: kauth_cred_free(so->so_cred);
1.1 cgd 748: so->so_state |= SS_NOFDREF;
749: sofree(so);
750: return (error);
751: }
752:
753: /*
1.160 ad 754: * Must be called with the socket locked.. Will return with it unlocked.
1.1 cgd 755: */
1.3 andrew 756: int
1.54 lukem 757: soabort(struct socket *so)
1.1 cgd 758: {
1.161 ad 759: u_int refs;
1.139 yamt 760: int error;
1.160 ad 761:
762: KASSERT(solocked(so));
763: KASSERT(so->so_head == NULL);
1.1 cgd 764:
1.161 ad 765: so->so_aborting++; /* XXX */
1.140 dyoung 766: error = (*so->so_proto->pr_usrreq)(so, PRU_ABORT, NULL,
767: NULL, NULL, NULL);
1.161 ad 768: refs = --so->so_aborting; /* XXX */
1.164 drochner 769: if (error || (refs == 0)) {
1.139 yamt 770: sofree(so);
1.160 ad 771: } else {
772: sounlock(so);
1.139 yamt 773: }
774: return error;
1.1 cgd 775: }
776:
1.3 andrew 777: int
1.54 lukem 778: soaccept(struct socket *so, struct mbuf *nam)
1.1 cgd 779: {
1.160 ad 780: int error;
781:
782: KASSERT(solocked(so));
1.1 cgd 783:
1.54 lukem 784: error = 0;
1.1 cgd 785: if ((so->so_state & SS_NOFDREF) == 0)
786: panic("soaccept: !NOFDREF");
787: so->so_state &= ~SS_NOFDREF;
1.55 thorpej 788: if ((so->so_state & SS_ISDISCONNECTED) == 0 ||
789: (so->so_proto->pr_flags & PR_ABRTACPTDIS) == 0)
1.41 mycroft 790: error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT,
1.140 dyoung 791: NULL, nam, NULL, NULL);
1.41 mycroft 792: else
1.53 itojun 793: error = ECONNABORTED;
1.52 itojun 794:
1.1 cgd 795: return (error);
796: }
797:
1.3 andrew 798: int
1.114 christos 799: soconnect(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1 cgd 800: {
1.160 ad 801: int error;
802:
803: KASSERT(solocked(so));
1.1 cgd 804:
805: if (so->so_options & SO_ACCEPTCONN)
806: return (EOPNOTSUPP);
807: /*
808: * If protocol is connection-based, can only connect once.
809: * Otherwise, if connected, try to disconnect first.
810: * This allows user to disconnect by connecting to, e.g.,
811: * a null address.
812: */
813: if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
814: ((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
815: (error = sodisconnect(so))))
816: error = EISCONN;
817: else
818: error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
1.140 dyoung 819: NULL, nam, NULL, l);
1.1 cgd 820: return (error);
821: }
822:
1.3 andrew 823: int
1.54 lukem 824: soconnect2(struct socket *so1, struct socket *so2)
1.1 cgd 825: {
1.160 ad 826: int error;
827:
828: KASSERT(solocked2(so1, so2));
1.1 cgd 829:
1.22 mycroft 830: error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2,
1.140 dyoung 831: NULL, (struct mbuf *)so2, NULL, NULL);
1.1 cgd 832: return (error);
833: }
834:
1.3 andrew 835: int
1.54 lukem 836: sodisconnect(struct socket *so)
1.1 cgd 837: {
1.160 ad 838: int error;
839:
840: KASSERT(solocked(so));
1.1 cgd 841:
842: if ((so->so_state & SS_ISCONNECTED) == 0) {
843: error = ENOTCONN;
1.160 ad 844: } else if (so->so_state & SS_ISDISCONNECTING) {
1.1 cgd 845: error = EALREADY;
1.160 ad 846: } else {
847: error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
848: NULL, NULL, NULL, NULL);
1.1 cgd 849: }
850: return (error);
851: }
852:
1.15 mycroft 853: #define SBLOCKWAIT(f) (((f) & MSG_DONTWAIT) ? M_NOWAIT : M_WAITOK)
1.1 cgd 854: /*
855: * Send on a socket.
856: * If send must go all at once and message is larger than
857: * send buffering, then hard error.
858: * Lock against other senders.
859: * If must go all at once and not enough room now, then
860: * inform user that this would block and do nothing.
861: * Otherwise, if nonblocking, send as much as possible.
862: * The data to be sent is described by "uio" if nonzero,
863: * otherwise by the mbuf chain "top" (which must be null
864: * if uio is not). Data provided in mbuf chain must be small
865: * enough to send all at once.
866: *
867: * Returns nonzero on error, timeout or signal; callers
868: * must check for short counts if EINTR/ERESTART are returned.
869: * Data and control buffers are freed on return.
870: */
1.3 andrew 871: int
1.54 lukem 872: sosend(struct socket *so, struct mbuf *addr, struct uio *uio, struct mbuf *top,
1.114 christos 873: struct mbuf *control, int flags, struct lwp *l)
1.1 cgd 874: {
1.54 lukem 875: struct mbuf **mp, *m;
1.114 christos 876: struct proc *p;
1.58 jdolecek 877: long space, len, resid, clen, mlen;
878: int error, s, dontroute, atomic;
1.196 dsl 879: short wakeup_state = 0;
1.54 lukem 880:
1.114 christos 881: p = l->l_proc;
1.160 ad 882: clen = 0;
1.64 thorpej 883:
1.160 ad 884: /*
885: * solock() provides atomicity of access. splsoftnet() prevents
886: * protocol processing soft interrupts from interrupting us and
887: * blocking (expensive).
888: */
889: s = splsoftnet();
890: solock(so);
1.54 lukem 891: atomic = sosendallatonce(so) || top;
1.1 cgd 892: if (uio)
893: resid = uio->uio_resid;
894: else
895: resid = top->m_pkthdr.len;
1.7 cgd 896: /*
897: * In theory resid should be unsigned.
898: * However, space must be signed, as it might be less than 0
899: * if we over-committed, and we must use a signed comparison
900: * of space and resid. On the other hand, a negative resid
901: * causes us to loop sending 0-length segments to the protocol.
902: */
1.29 mycroft 903: if (resid < 0) {
904: error = EINVAL;
905: goto out;
906: }
1.1 cgd 907: dontroute =
908: (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
909: (so->so_proto->pr_flags & PR_ATOMIC);
1.165 christos 910: l->l_ru.ru_msgsnd++;
1.1 cgd 911: if (control)
912: clen = control->m_len;
1.54 lukem 913: restart:
1.21 christos 914: if ((error = sblock(&so->so_snd, SBLOCKWAIT(flags))) != 0)
1.1 cgd 915: goto out;
916: do {
1.160 ad 917: if (so->so_state & SS_CANTSENDMORE) {
918: error = EPIPE;
919: goto release;
920: }
1.48 thorpej 921: if (so->so_error) {
922: error = so->so_error;
923: so->so_error = 0;
924: goto release;
925: }
1.1 cgd 926: if ((so->so_state & SS_ISCONNECTED) == 0) {
927: if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
928: if ((so->so_state & SS_ISCONFIRMING) == 0 &&
1.160 ad 929: !(resid == 0 && clen != 0)) {
930: error = ENOTCONN;
931: goto release;
932: }
933: } else if (addr == 0) {
934: error = EDESTADDRREQ;
935: goto release;
936: }
1.1 cgd 937: }
938: space = sbspace(&so->so_snd);
939: if (flags & MSG_OOB)
940: space += 1024;
1.21 christos 941: if ((atomic && resid > so->so_snd.sb_hiwat) ||
1.160 ad 942: clen > so->so_snd.sb_hiwat) {
943: error = EMSGSIZE;
944: goto release;
945: }
1.96 mycroft 946: if (space < resid + clen &&
1.1 cgd 947: (atomic || space < so->so_snd.sb_lowat || space < clen)) {
1.206 christos 948: if ((so->so_state & SS_NBIO) || (flags & MSG_NBIO)) {
1.160 ad 949: error = EWOULDBLOCK;
950: goto release;
951: }
1.1 cgd 952: sbunlock(&so->so_snd);
1.196 dsl 953: if (wakeup_state & SS_RESTARTSYS) {
954: error = ERESTART;
955: goto out;
956: }
1.1 cgd 957: error = sbwait(&so->so_snd);
958: if (error)
959: goto out;
1.196 dsl 960: wakeup_state = so->so_state;
1.1 cgd 961: goto restart;
962: }
1.196 dsl 963: wakeup_state = 0;
1.1 cgd 964: mp = ⊤
965: space -= clen;
966: do {
1.45 tv 967: if (uio == NULL) {
968: /*
969: * Data is prepackaged in "top".
970: */
971: resid = 0;
972: if (flags & MSG_EOR)
973: top->m_flags |= M_EOR;
974: } else do {
1.160 ad 975: sounlock(so);
976: splx(s);
1.144 dyoung 977: if (top == NULL) {
1.78 matt 978: m = m_gethdr(M_WAIT, MT_DATA);
1.45 tv 979: mlen = MHLEN;
980: m->m_pkthdr.len = 0;
1.140 dyoung 981: m->m_pkthdr.rcvif = NULL;
1.45 tv 982: } else {
1.78 matt 983: m = m_get(M_WAIT, MT_DATA);
1.45 tv 984: mlen = MLEN;
985: }
1.78 matt 986: MCLAIM(m, so->so_snd.sb_mowner);
1.121 yamt 987: if (sock_loan_thresh >= 0 &&
988: uio->uio_iov->iov_len >= sock_loan_thresh &&
989: space >= sock_loan_thresh &&
1.64 thorpej 990: (len = sosend_loan(so, uio, m,
991: space)) != 0) {
992: SOSEND_COUNTER_INCR(&sosend_loan_big);
993: space -= len;
994: goto have_data;
995: }
1.45 tv 996: if (resid >= MINCLSIZE && space >= MCLBYTES) {
1.64 thorpej 997: SOSEND_COUNTER_INCR(&sosend_copy_big);
1.201 oki 998: m_clget(m, M_DONTWAIT);
1.45 tv 999: if ((m->m_flags & M_EXT) == 0)
1000: goto nopages;
1001: mlen = MCLBYTES;
1002: if (atomic && top == 0) {
1.58 jdolecek 1003: len = lmin(MCLBYTES - max_hdr,
1.54 lukem 1004: resid);
1.45 tv 1005: m->m_data += max_hdr;
1006: } else
1.58 jdolecek 1007: len = lmin(MCLBYTES, resid);
1.45 tv 1008: space -= len;
1009: } else {
1.64 thorpej 1010: nopages:
1011: SOSEND_COUNTER_INCR(&sosend_copy_small);
1.58 jdolecek 1012: len = lmin(lmin(mlen, resid), space);
1.45 tv 1013: space -= len;
1014: /*
1015: * For datagram protocols, leave room
1016: * for protocol headers in first mbuf.
1017: */
1018: if (atomic && top == 0 && len < mlen)
1019: MH_ALIGN(m, len);
1020: }
1.144 dyoung 1021: error = uiomove(mtod(m, void *), (int)len, uio);
1.64 thorpej 1022: have_data:
1.45 tv 1023: resid = uio->uio_resid;
1024: m->m_len = len;
1025: *mp = m;
1026: top->m_pkthdr.len += len;
1.160 ad 1027: s = splsoftnet();
1028: solock(so);
1.144 dyoung 1029: if (error != 0)
1.45 tv 1030: goto release;
1031: mp = &m->m_next;
1032: if (resid <= 0) {
1033: if (flags & MSG_EOR)
1034: top->m_flags |= M_EOR;
1035: break;
1036: }
1037: } while (space > 0 && atomic);
1.108 perry 1038:
1.160 ad 1039: if (so->so_state & SS_CANTSENDMORE) {
1040: error = EPIPE;
1041: goto release;
1042: }
1.45 tv 1043: if (dontroute)
1044: so->so_options |= SO_DONTROUTE;
1045: if (resid > 0)
1046: so->so_state |= SS_MORETOCOME;
1.46 sommerfe 1047: error = (*so->so_proto->pr_usrreq)(so,
1048: (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
1.160 ad 1049: top, addr, control, curlwp);
1.45 tv 1050: if (dontroute)
1051: so->so_options &= ~SO_DONTROUTE;
1052: if (resid > 0)
1053: so->so_state &= ~SS_MORETOCOME;
1054: clen = 0;
1.144 dyoung 1055: control = NULL;
1056: top = NULL;
1.45 tv 1057: mp = ⊤
1.144 dyoung 1058: if (error != 0)
1.1 cgd 1059: goto release;
1060: } while (resid && space > 0);
1061: } while (resid);
1062:
1.54 lukem 1063: release:
1.1 cgd 1064: sbunlock(&so->so_snd);
1.54 lukem 1065: out:
1.160 ad 1066: sounlock(so);
1067: splx(s);
1.1 cgd 1068: if (top)
1069: m_freem(top);
1070: if (control)
1071: m_freem(control);
1072: return (error);
1073: }
1074:
1075: /*
1.159 ad 1076: * Following replacement or removal of the first mbuf on the first
1077: * mbuf chain of a socket buffer, push necessary state changes back
1078: * into the socket buffer so that other consumers see the values
1079: * consistently. 'nextrecord' is the callers locally stored value of
1080: * the original value of sb->sb_mb->m_nextpkt which must be restored
1081: * when the lead mbuf changes. NOTE: 'nextrecord' may be NULL.
1082: */
1083: static void
1084: sbsync(struct sockbuf *sb, struct mbuf *nextrecord)
1085: {
1086:
1.160 ad 1087: KASSERT(solocked(sb->sb_so));
1088:
1.159 ad 1089: /*
1090: * First, update for the new value of nextrecord. If necessary,
1091: * make it the first record.
1092: */
1093: if (sb->sb_mb != NULL)
1094: sb->sb_mb->m_nextpkt = nextrecord;
1095: else
1096: sb->sb_mb = nextrecord;
1097:
1098: /*
1099: * Now update any dependent socket buffer fields to reflect
1100: * the new state. This is an inline of SB_EMPTY_FIXUP, with
1101: * the addition of a second clause that takes care of the
1102: * case where sb_mb has been updated, but remains the last
1103: * record.
1104: */
1105: if (sb->sb_mb == NULL) {
1106: sb->sb_mbtail = NULL;
1107: sb->sb_lastrecord = NULL;
1108: } else if (sb->sb_mb->m_nextpkt == NULL)
1109: sb->sb_lastrecord = sb->sb_mb;
1110: }
1111:
1112: /*
1.1 cgd 1113: * Implement receive operations on a socket.
1114: * We depend on the way that records are added to the sockbuf
1115: * by sbappend*. In particular, each record (mbufs linked through m_next)
1116: * must begin with an address if the protocol so specifies,
1117: * followed by an optional mbuf or mbufs containing ancillary data,
1118: * and then zero or more mbufs of data.
1119: * In order to avoid blocking network interrupts for the entire time here,
1120: * we splx() while doing the actual copy to user space.
1121: * Although the sockbuf is locked, new data may still be appended,
1122: * and thus we must maintain consistency of the sockbuf during that time.
1123: *
1124: * The caller may receive the data as a single mbuf chain by supplying
1125: * an mbuf **mp0 for use in returning the chain. The uio is then used
1126: * only for the count in uio_resid.
1127: */
1.3 andrew 1128: int
1.54 lukem 1129: soreceive(struct socket *so, struct mbuf **paddr, struct uio *uio,
1130: struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
1.1 cgd 1131: {
1.116 yamt 1132: struct lwp *l = curlwp;
1.160 ad 1133: struct mbuf *m, **mp, *mt;
1.146 dyoung 1134: int atomic, flags, len, error, s, offset, moff, type, orig_resid;
1.99 matt 1135: const struct protosw *pr;
1.54 lukem 1136: struct mbuf *nextrecord;
1.67 he 1137: int mbuf_removed = 0;
1.146 dyoung 1138: const struct domain *dom;
1.196 dsl 1139: short wakeup_state = 0;
1.64 thorpej 1140:
1.54 lukem 1141: pr = so->so_proto;
1.146 dyoung 1142: atomic = pr->pr_flags & PR_ATOMIC;
1143: dom = pr->pr_domain;
1.1 cgd 1144: mp = mp0;
1.54 lukem 1145: type = 0;
1146: orig_resid = uio->uio_resid;
1.102 jonathan 1147:
1.144 dyoung 1148: if (paddr != NULL)
1149: *paddr = NULL;
1150: if (controlp != NULL)
1151: *controlp = NULL;
1152: if (flagsp != NULL)
1.1 cgd 1153: flags = *flagsp &~ MSG_EOR;
1154: else
1155: flags = 0;
1.66 enami 1156:
1.1 cgd 1157: if (flags & MSG_OOB) {
1158: m = m_get(M_WAIT, MT_DATA);
1.160 ad 1159: solock(so);
1.17 cgd 1160: error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m,
1.140 dyoung 1161: (struct mbuf *)(long)(flags & MSG_PEEK), NULL, l);
1.160 ad 1162: sounlock(so);
1.1 cgd 1163: if (error)
1164: goto bad;
1165: do {
1.134 christos 1166: error = uiomove(mtod(m, void *),
1.1 cgd 1167: (int) min(uio->uio_resid, m->m_len), uio);
1168: m = m_free(m);
1.144 dyoung 1169: } while (uio->uio_resid > 0 && error == 0 && m);
1.54 lukem 1170: bad:
1.144 dyoung 1171: if (m != NULL)
1.1 cgd 1172: m_freem(m);
1.144 dyoung 1173: return error;
1.1 cgd 1174: }
1.144 dyoung 1175: if (mp != NULL)
1.140 dyoung 1176: *mp = NULL;
1.160 ad 1177:
1178: /*
1179: * solock() provides atomicity of access. splsoftnet() prevents
1180: * protocol processing soft interrupts from interrupting us and
1181: * blocking (expensive).
1182: */
1183: s = splsoftnet();
1184: solock(so);
1.1 cgd 1185: if (so->so_state & SS_ISCONFIRMING && uio->uio_resid)
1.140 dyoung 1186: (*pr->pr_usrreq)(so, PRU_RCVD, NULL, NULL, NULL, l);
1.1 cgd 1187:
1.54 lukem 1188: restart:
1.160 ad 1189: if ((error = sblock(&so->so_rcv, SBLOCKWAIT(flags))) != 0) {
1190: sounlock(so);
1191: splx(s);
1.144 dyoung 1192: return error;
1.160 ad 1193: }
1.1 cgd 1194:
1195: m = so->so_rcv.sb_mb;
1196: /*
1197: * If we have less data than requested, block awaiting more
1198: * (subject to any timeout) if:
1.15 mycroft 1199: * 1. the current count is less than the low water mark,
1.1 cgd 1200: * 2. MSG_WAITALL is set, and it is possible to do the entire
1.15 mycroft 1201: * receive operation at once if we block (resid <= hiwat), or
1202: * 3. MSG_DONTWAIT is not set.
1.1 cgd 1203: * If MSG_WAITALL is set but resid is larger than the receive buffer,
1204: * we have to do the receive in sections, and thus risk returning
1205: * a short count if a timeout or signal occurs after we start.
1206: */
1.144 dyoung 1207: if (m == NULL ||
1208: ((flags & MSG_DONTWAIT) == 0 &&
1209: so->so_rcv.sb_cc < uio->uio_resid &&
1210: (so->so_rcv.sb_cc < so->so_rcv.sb_lowat ||
1211: ((flags & MSG_WAITALL) &&
1212: uio->uio_resid <= so->so_rcv.sb_hiwat)) &&
1.146 dyoung 1213: m->m_nextpkt == NULL && !atomic)) {
1.1 cgd 1214: #ifdef DIAGNOSTIC
1.144 dyoung 1215: if (m == NULL && so->so_rcv.sb_cc)
1.1 cgd 1216: panic("receive 1");
1217: #endif
1218: if (so->so_error) {
1.144 dyoung 1219: if (m != NULL)
1.15 mycroft 1220: goto dontblock;
1.1 cgd 1221: error = so->so_error;
1222: if ((flags & MSG_PEEK) == 0)
1223: so->so_error = 0;
1224: goto release;
1225: }
1226: if (so->so_state & SS_CANTRCVMORE) {
1.144 dyoung 1227: if (m != NULL)
1.15 mycroft 1228: goto dontblock;
1.1 cgd 1229: else
1230: goto release;
1231: }
1.144 dyoung 1232: for (; m != NULL; m = m->m_next)
1.1 cgd 1233: if (m->m_type == MT_OOBDATA || (m->m_flags & M_EOR)) {
1234: m = so->so_rcv.sb_mb;
1235: goto dontblock;
1236: }
1237: if ((so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) == 0 &&
1238: (so->so_proto->pr_flags & PR_CONNREQUIRED)) {
1239: error = ENOTCONN;
1240: goto release;
1241: }
1242: if (uio->uio_resid == 0)
1243: goto release;
1.206 christos 1244: if ((so->so_state & SS_NBIO) ||
1245: (flags & (MSG_DONTWAIT|MSG_NBIO))) {
1.1 cgd 1246: error = EWOULDBLOCK;
1247: goto release;
1248: }
1.69 thorpej 1249: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1");
1250: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1");
1.1 cgd 1251: sbunlock(&so->so_rcv);
1.196 dsl 1252: if (wakeup_state & SS_RESTARTSYS)
1253: error = ERESTART;
1254: else
1255: error = sbwait(&so->so_rcv);
1.160 ad 1256: if (error != 0) {
1257: sounlock(so);
1258: splx(s);
1.144 dyoung 1259: return error;
1.160 ad 1260: }
1.196 dsl 1261: wakeup_state = so->so_state;
1.1 cgd 1262: goto restart;
1263: }
1.54 lukem 1264: dontblock:
1.69 thorpej 1265: /*
1266: * On entry here, m points to the first record of the socket buffer.
1.159 ad 1267: * From this point onward, we maintain 'nextrecord' as a cache of the
1268: * pointer to the next record in the socket buffer. We must keep the
1269: * various socket buffer pointers and local stack versions of the
1270: * pointers in sync, pushing out modifications before dropping the
1.160 ad 1271: * socket lock, and re-reading them when picking it up.
1.159 ad 1272: *
1273: * Otherwise, we will race with the network stack appending new data
1274: * or records onto the socket buffer by using inconsistent/stale
1275: * versions of the field, possibly resulting in socket buffer
1276: * corruption.
1277: *
1278: * By holding the high-level sblock(), we prevent simultaneous
1279: * readers from pulling off the front of the socket buffer.
1.69 thorpej 1280: */
1.144 dyoung 1281: if (l != NULL)
1.157 ad 1282: l->l_ru.ru_msgrcv++;
1.69 thorpej 1283: KASSERT(m == so->so_rcv.sb_mb);
1284: SBLASTRECORDCHK(&so->so_rcv, "soreceive 1");
1285: SBLASTMBUFCHK(&so->so_rcv, "soreceive 1");
1.1 cgd 1286: nextrecord = m->m_nextpkt;
1287: if (pr->pr_flags & PR_ADDR) {
1288: #ifdef DIAGNOSTIC
1289: if (m->m_type != MT_SONAME)
1290: panic("receive 1a");
1291: #endif
1.3 andrew 1292: orig_resid = 0;
1.1 cgd 1293: if (flags & MSG_PEEK) {
1294: if (paddr)
1295: *paddr = m_copy(m, 0, m->m_len);
1296: m = m->m_next;
1297: } else {
1298: sbfree(&so->so_rcv, m);
1.67 he 1299: mbuf_removed = 1;
1.144 dyoung 1300: if (paddr != NULL) {
1.1 cgd 1301: *paddr = m;
1302: so->so_rcv.sb_mb = m->m_next;
1.144 dyoung 1303: m->m_next = NULL;
1.1 cgd 1304: m = so->so_rcv.sb_mb;
1305: } else {
1306: MFREE(m, so->so_rcv.sb_mb);
1307: m = so->so_rcv.sb_mb;
1308: }
1.159 ad 1309: sbsync(&so->so_rcv, nextrecord);
1.1 cgd 1310: }
1311: }
1.159 ad 1312:
1313: /*
1314: * Process one or more MT_CONTROL mbufs present before any data mbufs
1315: * in the first mbuf chain on the socket buffer. If MSG_PEEK, we
1316: * just copy the data; if !MSG_PEEK, we call into the protocol to
1317: * perform externalization (or freeing if controlp == NULL).
1318: */
1319: if (__predict_false(m != NULL && m->m_type == MT_CONTROL)) {
1320: struct mbuf *cm = NULL, *cmn;
1321: struct mbuf **cme = &cm;
1322:
1323: do {
1324: if (flags & MSG_PEEK) {
1325: if (controlp != NULL) {
1326: *controlp = m_copy(m, 0, m->m_len);
1327: controlp = &(*controlp)->m_next;
1328: }
1329: m = m->m_next;
1330: } else {
1331: sbfree(&so->so_rcv, m);
1.1 cgd 1332: so->so_rcv.sb_mb = m->m_next;
1.144 dyoung 1333: m->m_next = NULL;
1.159 ad 1334: *cme = m;
1335: cme = &(*cme)->m_next;
1.1 cgd 1336: m = so->so_rcv.sb_mb;
1.159 ad 1337: }
1338: } while (m != NULL && m->m_type == MT_CONTROL);
1339: if ((flags & MSG_PEEK) == 0)
1340: sbsync(&so->so_rcv, nextrecord);
1341: for (; cm != NULL; cm = cmn) {
1342: cmn = cm->m_next;
1343: cm->m_next = NULL;
1344: type = mtod(cm, struct cmsghdr *)->cmsg_type;
1345: if (controlp != NULL) {
1346: if (dom->dom_externalize != NULL &&
1347: type == SCM_RIGHTS) {
1.160 ad 1348: sounlock(so);
1.159 ad 1349: splx(s);
1.204 christos 1350: error = (*dom->dom_externalize)(cm, l,
1351: (flags & MSG_CMSG_CLOEXEC) ?
1352: O_CLOEXEC : 0);
1.159 ad 1353: s = splsoftnet();
1.160 ad 1354: solock(so);
1.159 ad 1355: }
1356: *controlp = cm;
1357: while (*controlp != NULL)
1358: controlp = &(*controlp)->m_next;
1.1 cgd 1359: } else {
1.106 itojun 1360: /*
1361: * Dispose of any SCM_RIGHTS message that went
1362: * through the read path rather than recv.
1363: */
1.159 ad 1364: if (dom->dom_dispose != NULL &&
1365: type == SCM_RIGHTS) {
1.160 ad 1366: sounlock(so);
1.159 ad 1367: (*dom->dom_dispose)(cm);
1.160 ad 1368: solock(so);
1.159 ad 1369: }
1370: m_freem(cm);
1.1 cgd 1371: }
1372: }
1.159 ad 1373: if (m != NULL)
1374: nextrecord = so->so_rcv.sb_mb->m_nextpkt;
1375: else
1376: nextrecord = so->so_rcv.sb_mb;
1377: orig_resid = 0;
1.1 cgd 1378: }
1.69 thorpej 1379:
1.159 ad 1380: /* If m is non-NULL, we have some data to read. */
1381: if (__predict_true(m != NULL)) {
1.1 cgd 1382: type = m->m_type;
1383: if (type == MT_OOBDATA)
1384: flags |= MSG_OOB;
1385: }
1.69 thorpej 1386: SBLASTRECORDCHK(&so->so_rcv, "soreceive 2");
1387: SBLASTMBUFCHK(&so->so_rcv, "soreceive 2");
1388:
1.1 cgd 1389: moff = 0;
1390: offset = 0;
1.144 dyoung 1391: while (m != NULL && uio->uio_resid > 0 && error == 0) {
1.1 cgd 1392: if (m->m_type == MT_OOBDATA) {
1393: if (type != MT_OOBDATA)
1394: break;
1395: } else if (type == MT_OOBDATA)
1396: break;
1397: #ifdef DIAGNOSTIC
1398: else if (m->m_type != MT_DATA && m->m_type != MT_HEADER)
1399: panic("receive 3");
1400: #endif
1401: so->so_state &= ~SS_RCVATMARK;
1.196 dsl 1402: wakeup_state = 0;
1.1 cgd 1403: len = uio->uio_resid;
1404: if (so->so_oobmark && len > so->so_oobmark - offset)
1405: len = so->so_oobmark - offset;
1406: if (len > m->m_len - moff)
1407: len = m->m_len - moff;
1408: /*
1409: * If mp is set, just pass back the mbufs.
1410: * Otherwise copy them out via the uio, then free.
1411: * Sockbuf must be consistent here (points to current mbuf,
1412: * it points to next record) when we drop priority;
1413: * we must note any additions to the sockbuf when we
1414: * block interrupts again.
1415: */
1.144 dyoung 1416: if (mp == NULL) {
1.69 thorpej 1417: SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove");
1418: SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove");
1.160 ad 1419: sounlock(so);
1.1 cgd 1420: splx(s);
1.134 christos 1421: error = uiomove(mtod(m, char *) + moff, (int)len, uio);
1.20 mycroft 1422: s = splsoftnet();
1.160 ad 1423: solock(so);
1.144 dyoung 1424: if (error != 0) {
1.67 he 1425: /*
1426: * If any part of the record has been removed
1427: * (such as the MT_SONAME mbuf, which will
1428: * happen when PR_ADDR, and thus also
1429: * PR_ATOMIC, is set), then drop the entire
1430: * record to maintain the atomicity of the
1431: * receive operation.
1432: *
1433: * This avoids a later panic("receive 1a")
1434: * when compiled with DIAGNOSTIC.
1435: */
1.146 dyoung 1436: if (m && mbuf_removed && atomic)
1.67 he 1437: (void) sbdroprecord(&so->so_rcv);
1438:
1.57 jdolecek 1439: goto release;
1.67 he 1440: }
1.1 cgd 1441: } else
1442: uio->uio_resid -= len;
1443: if (len == m->m_len - moff) {
1444: if (m->m_flags & M_EOR)
1445: flags |= MSG_EOR;
1446: if (flags & MSG_PEEK) {
1447: m = m->m_next;
1448: moff = 0;
1449: } else {
1450: nextrecord = m->m_nextpkt;
1451: sbfree(&so->so_rcv, m);
1452: if (mp) {
1453: *mp = m;
1454: mp = &m->m_next;
1455: so->so_rcv.sb_mb = m = m->m_next;
1.140 dyoung 1456: *mp = NULL;
1.1 cgd 1457: } else {
1458: MFREE(m, so->so_rcv.sb_mb);
1459: m = so->so_rcv.sb_mb;
1460: }
1.69 thorpej 1461: /*
1462: * If m != NULL, we also know that
1463: * so->so_rcv.sb_mb != NULL.
1464: */
1465: KASSERT(so->so_rcv.sb_mb == m);
1466: if (m) {
1.1 cgd 1467: m->m_nextpkt = nextrecord;
1.69 thorpej 1468: if (nextrecord == NULL)
1469: so->so_rcv.sb_lastrecord = m;
1470: } else {
1471: so->so_rcv.sb_mb = nextrecord;
1.70 thorpej 1472: SB_EMPTY_FIXUP(&so->so_rcv);
1.69 thorpej 1473: }
1474: SBLASTRECORDCHK(&so->so_rcv, "soreceive 3");
1475: SBLASTMBUFCHK(&so->so_rcv, "soreceive 3");
1.1 cgd 1476: }
1.144 dyoung 1477: } else if (flags & MSG_PEEK)
1478: moff += len;
1479: else {
1.160 ad 1480: if (mp != NULL) {
1481: mt = m_copym(m, 0, len, M_NOWAIT);
1482: if (__predict_false(mt == NULL)) {
1483: sounlock(so);
1484: mt = m_copym(m, 0, len, M_WAIT);
1485: solock(so);
1486: }
1487: *mp = mt;
1488: }
1.144 dyoung 1489: m->m_data += len;
1490: m->m_len -= len;
1491: so->so_rcv.sb_cc -= len;
1.1 cgd 1492: }
1493: if (so->so_oobmark) {
1494: if ((flags & MSG_PEEK) == 0) {
1495: so->so_oobmark -= len;
1496: if (so->so_oobmark == 0) {
1497: so->so_state |= SS_RCVATMARK;
1498: break;
1499: }
1.7 cgd 1500: } else {
1.1 cgd 1501: offset += len;
1.7 cgd 1502: if (offset == so->so_oobmark)
1503: break;
1504: }
1.1 cgd 1505: }
1506: if (flags & MSG_EOR)
1507: break;
1508: /*
1509: * If the MSG_WAITALL flag is set (for non-atomic socket),
1510: * we must not quit until "uio->uio_resid == 0" or an error
1511: * termination. If a signal/timeout occurs, return
1512: * with a short count but without error.
1513: * Keep sockbuf locked against other readers.
1514: */
1.144 dyoung 1515: while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 &&
1.3 andrew 1516: !sosendallatonce(so) && !nextrecord) {
1.1 cgd 1517: if (so->so_error || so->so_state & SS_CANTRCVMORE)
1518: break;
1.68 matt 1519: /*
1520: * If we are peeking and the socket receive buffer is
1521: * full, stop since we can't get more data to peek at.
1522: */
1523: if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0)
1524: break;
1525: /*
1526: * If we've drained the socket buffer, tell the
1527: * protocol in case it needs to do something to
1528: * get it filled again.
1529: */
1530: if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb)
1531: (*pr->pr_usrreq)(so, PRU_RCVD,
1.140 dyoung 1532: NULL, (struct mbuf *)(long)flags, NULL, l);
1.69 thorpej 1533: SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2");
1534: SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2");
1.196 dsl 1535: if (wakeup_state & SS_RESTARTSYS)
1536: error = ERESTART;
1537: else
1538: error = sbwait(&so->so_rcv);
1.144 dyoung 1539: if (error != 0) {
1.1 cgd 1540: sbunlock(&so->so_rcv);
1.160 ad 1541: sounlock(so);
1.1 cgd 1542: splx(s);
1.144 dyoung 1543: return 0;
1.1 cgd 1544: }
1.21 christos 1545: if ((m = so->so_rcv.sb_mb) != NULL)
1.1 cgd 1546: nextrecord = m->m_nextpkt;
1.196 dsl 1547: wakeup_state = so->so_state;
1.1 cgd 1548: }
1549: }
1.3 andrew 1550:
1.146 dyoung 1551: if (m && atomic) {
1.3 andrew 1552: flags |= MSG_TRUNC;
1553: if ((flags & MSG_PEEK) == 0)
1554: (void) sbdroprecord(&so->so_rcv);
1555: }
1.1 cgd 1556: if ((flags & MSG_PEEK) == 0) {
1.144 dyoung 1557: if (m == NULL) {
1.69 thorpej 1558: /*
1.70 thorpej 1559: * First part is an inline SB_EMPTY_FIXUP(). Second
1.69 thorpej 1560: * part makes sure sb_lastrecord is up-to-date if
1561: * there is still data in the socket buffer.
1562: */
1.1 cgd 1563: so->so_rcv.sb_mb = nextrecord;
1.69 thorpej 1564: if (so->so_rcv.sb_mb == NULL) {
1565: so->so_rcv.sb_mbtail = NULL;
1566: so->so_rcv.sb_lastrecord = NULL;
1567: } else if (nextrecord->m_nextpkt == NULL)
1568: so->so_rcv.sb_lastrecord = nextrecord;
1569: }
1570: SBLASTRECORDCHK(&so->so_rcv, "soreceive 4");
1571: SBLASTMBUFCHK(&so->so_rcv, "soreceive 4");
1.1 cgd 1572: if (pr->pr_flags & PR_WANTRCVD && so->so_pcb)
1.140 dyoung 1573: (*pr->pr_usrreq)(so, PRU_RCVD, NULL,
1574: (struct mbuf *)(long)flags, NULL, l);
1.1 cgd 1575: }
1.3 andrew 1576: if (orig_resid == uio->uio_resid && orig_resid &&
1577: (flags & MSG_EOR) == 0 && (so->so_state & SS_CANTRCVMORE) == 0) {
1578: sbunlock(&so->so_rcv);
1579: goto restart;
1580: }
1.108 perry 1581:
1.144 dyoung 1582: if (flagsp != NULL)
1.1 cgd 1583: *flagsp |= flags;
1.54 lukem 1584: release:
1.1 cgd 1585: sbunlock(&so->so_rcv);
1.160 ad 1586: sounlock(so);
1.1 cgd 1587: splx(s);
1.144 dyoung 1588: return error;
1.1 cgd 1589: }
1590:
1.14 mycroft 1591: int
1.54 lukem 1592: soshutdown(struct socket *so, int how)
1.1 cgd 1593: {
1.99 matt 1594: const struct protosw *pr;
1.160 ad 1595: int error;
1596:
1597: KASSERT(solocked(so));
1.34 kleink 1598:
1.54 lukem 1599: pr = so->so_proto;
1.34 kleink 1600: if (!(how == SHUT_RD || how == SHUT_WR || how == SHUT_RDWR))
1601: return (EINVAL);
1.1 cgd 1602:
1.160 ad 1603: if (how == SHUT_RD || how == SHUT_RDWR) {
1.1 cgd 1604: sorflush(so);
1.160 ad 1605: error = 0;
1606: }
1.34 kleink 1607: if (how == SHUT_WR || how == SHUT_RDWR)
1.160 ad 1608: error = (*pr->pr_usrreq)(so, PRU_SHUTDOWN, NULL,
1.140 dyoung 1609: NULL, NULL, NULL);
1.160 ad 1610:
1611: return error;
1.1 cgd 1612: }
1613:
1.195 dsl 1614: void
1.196 dsl 1615: sorestart(struct socket *so)
1.188 ad 1616: {
1.196 dsl 1617: /*
1618: * An application has called close() on an fd on which another
1619: * of its threads has called a socket system call.
1620: * Mark this and wake everyone up, and code that would block again
1621: * instead returns ERESTART.
1622: * On system call re-entry the fd is validated and EBADF returned.
1623: * Any other fd will block again on the 2nd syscall.
1624: */
1.188 ad 1625: solock(so);
1.196 dsl 1626: so->so_state |= SS_RESTARTSYS;
1.188 ad 1627: cv_broadcast(&so->so_cv);
1.196 dsl 1628: cv_broadcast(&so->so_snd.sb_cv);
1629: cv_broadcast(&so->so_rcv.sb_cv);
1.188 ad 1630: sounlock(so);
1631: }
1632:
1.14 mycroft 1633: void
1.54 lukem 1634: sorflush(struct socket *so)
1.1 cgd 1635: {
1.54 lukem 1636: struct sockbuf *sb, asb;
1.99 matt 1637: const struct protosw *pr;
1.160 ad 1638:
1639: KASSERT(solocked(so));
1.1 cgd 1640:
1.54 lukem 1641: sb = &so->so_rcv;
1642: pr = so->so_proto;
1.160 ad 1643: socantrcvmore(so);
1.1 cgd 1644: sb->sb_flags |= SB_NOINTR;
1.160 ad 1645: (void )sblock(sb, M_WAITOK);
1.1 cgd 1646: sbunlock(sb);
1647: asb = *sb;
1.86 wrstuden 1648: /*
1649: * Clear most of the sockbuf structure, but leave some of the
1650: * fields valid.
1651: */
1652: memset(&sb->sb_startzero, 0,
1653: sizeof(*sb) - offsetof(struct sockbuf, sb_startzero));
1.160 ad 1654: if (pr->pr_flags & PR_RIGHTS && pr->pr_domain->dom_dispose) {
1655: sounlock(so);
1.1 cgd 1656: (*pr->pr_domain->dom_dispose)(asb.sb_mb);
1.160 ad 1657: solock(so);
1658: }
1.98 christos 1659: sbrelease(&asb, so);
1.1 cgd 1660: }
1661:
1.171 plunky 1662: /*
1663: * internal set SOL_SOCKET options
1664: */
1.142 dyoung 1665: static int
1.171 plunky 1666: sosetopt1(struct socket *so, const struct sockopt *sopt)
1.1 cgd 1667: {
1.182 christos 1668: int error = EINVAL, optval, opt;
1.171 plunky 1669: struct linger l;
1670: struct timeval tv;
1.142 dyoung 1671:
1.179 christos 1672: switch ((opt = sopt->sopt_name)) {
1.142 dyoung 1673:
1.170 tls 1674: case SO_ACCEPTFILTER:
1.177 ad 1675: error = accept_filt_setopt(so, sopt);
1676: KASSERT(solocked(so));
1.170 tls 1677: break;
1678:
1.171 plunky 1679: case SO_LINGER:
1680: error = sockopt_get(sopt, &l, sizeof(l));
1.177 ad 1681: solock(so);
1.171 plunky 1682: if (error)
1.177 ad 1683: break;
1.171 plunky 1684: if (l.l_linger < 0 || l.l_linger > USHRT_MAX ||
1.177 ad 1685: l.l_linger > (INT_MAX / hz)) {
1686: error = EDOM;
1687: break;
1688: }
1.171 plunky 1689: so->so_linger = l.l_linger;
1690: if (l.l_onoff)
1691: so->so_options |= SO_LINGER;
1692: else
1693: so->so_options &= ~SO_LINGER;
1.177 ad 1694: break;
1.1 cgd 1695:
1.142 dyoung 1696: case SO_DEBUG:
1697: case SO_KEEPALIVE:
1698: case SO_DONTROUTE:
1699: case SO_USELOOPBACK:
1700: case SO_BROADCAST:
1701: case SO_REUSEADDR:
1702: case SO_REUSEPORT:
1703: case SO_OOBINLINE:
1704: case SO_TIMESTAMP:
1.207 christos 1705: case SO_NOSIGPIPE:
1.184 christos 1706: #ifdef SO_OTIMESTAMP
1707: case SO_OTIMESTAMP:
1708: #endif
1.171 plunky 1709: error = sockopt_getint(sopt, &optval);
1.177 ad 1710: solock(so);
1.171 plunky 1711: if (error)
1.177 ad 1712: break;
1.171 plunky 1713: if (optval)
1.179 christos 1714: so->so_options |= opt;
1.142 dyoung 1715: else
1.179 christos 1716: so->so_options &= ~opt;
1.142 dyoung 1717: break;
1718:
1719: case SO_SNDBUF:
1720: case SO_RCVBUF:
1721: case SO_SNDLOWAT:
1722: case SO_RCVLOWAT:
1.171 plunky 1723: error = sockopt_getint(sopt, &optval);
1.177 ad 1724: solock(so);
1.171 plunky 1725: if (error)
1.177 ad 1726: break;
1.1 cgd 1727:
1.142 dyoung 1728: /*
1729: * Values < 1 make no sense for any of these
1730: * options, so disallow them.
1731: */
1.177 ad 1732: if (optval < 1) {
1733: error = EINVAL;
1734: break;
1735: }
1.1 cgd 1736:
1.179 christos 1737: switch (opt) {
1.171 plunky 1738: case SO_SNDBUF:
1.177 ad 1739: if (sbreserve(&so->so_snd, (u_long)optval, so) == 0) {
1740: error = ENOBUFS;
1741: break;
1742: }
1.171 plunky 1743: so->so_snd.sb_flags &= ~SB_AUTOSIZE;
1744: break;
1.1 cgd 1745:
1746: case SO_RCVBUF:
1.177 ad 1747: if (sbreserve(&so->so_rcv, (u_long)optval, so) == 0) {
1748: error = ENOBUFS;
1749: break;
1750: }
1.171 plunky 1751: so->so_rcv.sb_flags &= ~SB_AUTOSIZE;
1.142 dyoung 1752: break;
1753:
1754: /*
1755: * Make sure the low-water is never greater than
1756: * the high-water.
1757: */
1.1 cgd 1758: case SO_SNDLOWAT:
1.171 plunky 1759: if (optval > so->so_snd.sb_hiwat)
1760: optval = so->so_snd.sb_hiwat;
1761:
1762: so->so_snd.sb_lowat = optval;
1.142 dyoung 1763: break;
1.171 plunky 1764:
1.1 cgd 1765: case SO_RCVLOWAT:
1.171 plunky 1766: if (optval > so->so_rcv.sb_hiwat)
1767: optval = so->so_rcv.sb_hiwat;
1768:
1769: so->so_rcv.sb_lowat = optval;
1.142 dyoung 1770: break;
1771: }
1772: break;
1.28 thorpej 1773:
1.179 christos 1774: #ifdef COMPAT_50
1775: case SO_OSNDTIMEO:
1776: case SO_ORCVTIMEO: {
1777: struct timeval50 otv;
1778: error = sockopt_get(sopt, &otv, sizeof(otv));
1.186 pooka 1779: if (error) {
1780: solock(so);
1.183 christos 1781: break;
1.186 pooka 1782: }
1.179 christos 1783: timeval50_to_timeval(&otv, &tv);
1784: opt = opt == SO_OSNDTIMEO ? SO_SNDTIMEO : SO_RCVTIMEO;
1.182 christos 1785: error = 0;
1.179 christos 1786: /*FALLTHROUGH*/
1787: }
1788: #endif /* COMPAT_50 */
1789:
1.142 dyoung 1790: case SO_SNDTIMEO:
1791: case SO_RCVTIMEO:
1.182 christos 1792: if (error)
1.179 christos 1793: error = sockopt_get(sopt, &tv, sizeof(tv));
1.177 ad 1794: solock(so);
1.171 plunky 1795: if (error)
1.177 ad 1796: break;
1.171 plunky 1797:
1.177 ad 1798: if (tv.tv_sec > (INT_MAX - tv.tv_usec / tick) / hz) {
1799: error = EDOM;
1800: break;
1801: }
1.28 thorpej 1802:
1.171 plunky 1803: optval = tv.tv_sec * hz + tv.tv_usec / tick;
1804: if (optval == 0 && tv.tv_usec != 0)
1805: optval = 1;
1.28 thorpej 1806:
1.179 christos 1807: switch (opt) {
1.142 dyoung 1808: case SO_SNDTIMEO:
1.171 plunky 1809: so->so_snd.sb_timeo = optval;
1.1 cgd 1810: break;
1811: case SO_RCVTIMEO:
1.171 plunky 1812: so->so_rcv.sb_timeo = optval;
1.142 dyoung 1813: break;
1814: }
1815: break;
1.1 cgd 1816:
1.142 dyoung 1817: default:
1.177 ad 1818: solock(so);
1819: error = ENOPROTOOPT;
1820: break;
1.142 dyoung 1821: }
1.177 ad 1822: KASSERT(solocked(so));
1823: return error;
1.142 dyoung 1824: }
1.1 cgd 1825:
1.142 dyoung 1826: int
1.171 plunky 1827: sosetopt(struct socket *so, struct sockopt *sopt)
1.142 dyoung 1828: {
1829: int error, prerr;
1.1 cgd 1830:
1.177 ad 1831: if (sopt->sopt_level == SOL_SOCKET) {
1.171 plunky 1832: error = sosetopt1(so, sopt);
1.177 ad 1833: KASSERT(solocked(so));
1834: } else {
1.142 dyoung 1835: error = ENOPROTOOPT;
1.177 ad 1836: solock(so);
1837: }
1.1 cgd 1838:
1.142 dyoung 1839: if ((error == 0 || error == ENOPROTOOPT) &&
1840: so->so_proto != NULL && so->so_proto->pr_ctloutput != NULL) {
1841: /* give the protocol stack a shot */
1.171 plunky 1842: prerr = (*so->so_proto->pr_ctloutput)(PRCO_SETOPT, so, sopt);
1.142 dyoung 1843: if (prerr == 0)
1844: error = 0;
1845: else if (prerr != ENOPROTOOPT)
1846: error = prerr;
1.171 plunky 1847: }
1.160 ad 1848: sounlock(so);
1.142 dyoung 1849: return error;
1.1 cgd 1850: }
1851:
1.171 plunky 1852: /*
1853: * so_setsockopt() is a wrapper providing a sockopt structure for sosetopt()
1854: */
1855: int
1856: so_setsockopt(struct lwp *l, struct socket *so, int level, int name,
1857: const void *val, size_t valsize)
1858: {
1859: struct sockopt sopt;
1860: int error;
1861:
1862: KASSERT(valsize == 0 || val != NULL);
1863:
1864: sockopt_init(&sopt, level, name, valsize);
1865: sockopt_set(&sopt, val, valsize);
1866:
1867: error = sosetopt(so, &sopt);
1868:
1869: sockopt_destroy(&sopt);
1870:
1871: return error;
1872: }
1873:
1874: /*
1875: * internal get SOL_SOCKET options
1876: */
1877: static int
1878: sogetopt1(struct socket *so, struct sockopt *sopt)
1879: {
1.179 christos 1880: int error, optval, opt;
1.171 plunky 1881: struct linger l;
1882: struct timeval tv;
1883:
1.179 christos 1884: switch ((opt = sopt->sopt_name)) {
1.171 plunky 1885:
1886: case SO_ACCEPTFILTER:
1.177 ad 1887: error = accept_filt_getopt(so, sopt);
1.171 plunky 1888: break;
1889:
1890: case SO_LINGER:
1891: l.l_onoff = (so->so_options & SO_LINGER) ? 1 : 0;
1892: l.l_linger = so->so_linger;
1893:
1894: error = sockopt_set(sopt, &l, sizeof(l));
1895: break;
1896:
1897: case SO_USELOOPBACK:
1898: case SO_DONTROUTE:
1899: case SO_DEBUG:
1900: case SO_KEEPALIVE:
1901: case SO_REUSEADDR:
1902: case SO_REUSEPORT:
1903: case SO_BROADCAST:
1904: case SO_OOBINLINE:
1905: case SO_TIMESTAMP:
1.207 christos 1906: case SO_NOSIGPIPE:
1.184 christos 1907: #ifdef SO_OTIMESTAMP
1908: case SO_OTIMESTAMP:
1909: #endif
1.179 christos 1910: error = sockopt_setint(sopt, (so->so_options & opt) ? 1 : 0);
1.171 plunky 1911: break;
1912:
1913: case SO_TYPE:
1914: error = sockopt_setint(sopt, so->so_type);
1915: break;
1916:
1917: case SO_ERROR:
1918: error = sockopt_setint(sopt, so->so_error);
1919: so->so_error = 0;
1920: break;
1921:
1922: case SO_SNDBUF:
1923: error = sockopt_setint(sopt, so->so_snd.sb_hiwat);
1924: break;
1925:
1926: case SO_RCVBUF:
1927: error = sockopt_setint(sopt, so->so_rcv.sb_hiwat);
1928: break;
1929:
1930: case SO_SNDLOWAT:
1931: error = sockopt_setint(sopt, so->so_snd.sb_lowat);
1932: break;
1933:
1934: case SO_RCVLOWAT:
1935: error = sockopt_setint(sopt, so->so_rcv.sb_lowat);
1936: break;
1937:
1.179 christos 1938: #ifdef COMPAT_50
1939: case SO_OSNDTIMEO:
1940: case SO_ORCVTIMEO: {
1941: struct timeval50 otv;
1942:
1943: optval = (opt == SO_OSNDTIMEO ?
1944: so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
1945:
1946: otv.tv_sec = optval / hz;
1947: otv.tv_usec = (optval % hz) * tick;
1948:
1949: error = sockopt_set(sopt, &otv, sizeof(otv));
1950: break;
1951: }
1952: #endif /* COMPAT_50 */
1953:
1.171 plunky 1954: case SO_SNDTIMEO:
1955: case SO_RCVTIMEO:
1.179 christos 1956: optval = (opt == SO_SNDTIMEO ?
1.171 plunky 1957: so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
1958:
1959: tv.tv_sec = optval / hz;
1960: tv.tv_usec = (optval % hz) * tick;
1961:
1962: error = sockopt_set(sopt, &tv, sizeof(tv));
1963: break;
1964:
1965: case SO_OVERFLOWED:
1966: error = sockopt_setint(sopt, so->so_rcv.sb_overflowed);
1967: break;
1968:
1969: default:
1970: error = ENOPROTOOPT;
1971: break;
1972: }
1973:
1974: return (error);
1975: }
1976:
1.14 mycroft 1977: int
1.171 plunky 1978: sogetopt(struct socket *so, struct sockopt *sopt)
1.1 cgd 1979: {
1.160 ad 1980: int error;
1.1 cgd 1981:
1.160 ad 1982: solock(so);
1.171 plunky 1983: if (sopt->sopt_level != SOL_SOCKET) {
1.1 cgd 1984: if (so->so_proto && so->so_proto->pr_ctloutput) {
1.160 ad 1985: error = ((*so->so_proto->pr_ctloutput)
1.171 plunky 1986: (PRCO_GETOPT, so, sopt));
1.1 cgd 1987: } else
1.160 ad 1988: error = (ENOPROTOOPT);
1.1 cgd 1989: } else {
1.171 plunky 1990: error = sogetopt1(so, sopt);
1991: }
1992: sounlock(so);
1993: return (error);
1994: }
1995:
1996: /*
1997: * alloc sockopt data buffer buffer
1998: * - will be released at destroy
1999: */
1.176 plunky 2000: static int
2001: sockopt_alloc(struct sockopt *sopt, size_t len, km_flag_t kmflag)
1.171 plunky 2002: {
2003:
2004: KASSERT(sopt->sopt_size == 0);
2005:
1.176 plunky 2006: if (len > sizeof(sopt->sopt_buf)) {
2007: sopt->sopt_data = kmem_zalloc(len, kmflag);
2008: if (sopt->sopt_data == NULL)
2009: return ENOMEM;
2010: } else
1.171 plunky 2011: sopt->sopt_data = sopt->sopt_buf;
2012:
2013: sopt->sopt_size = len;
1.176 plunky 2014: return 0;
1.171 plunky 2015: }
2016:
2017: /*
2018: * initialise sockopt storage
1.176 plunky 2019: * - MAY sleep during allocation
1.171 plunky 2020: */
2021: void
2022: sockopt_init(struct sockopt *sopt, int level, int name, size_t size)
2023: {
1.1 cgd 2024:
1.171 plunky 2025: memset(sopt, 0, sizeof(*sopt));
1.1 cgd 2026:
1.171 plunky 2027: sopt->sopt_level = level;
2028: sopt->sopt_name = name;
1.176 plunky 2029: (void)sockopt_alloc(sopt, size, KM_SLEEP);
1.171 plunky 2030: }
2031:
2032: /*
2033: * destroy sockopt storage
2034: * - will release any held memory references
2035: */
2036: void
2037: sockopt_destroy(struct sockopt *sopt)
2038: {
2039:
2040: if (sopt->sopt_data != sopt->sopt_buf)
1.173 plunky 2041: kmem_free(sopt->sopt_data, sopt->sopt_size);
1.171 plunky 2042:
2043: memset(sopt, 0, sizeof(*sopt));
2044: }
2045:
2046: /*
2047: * set sockopt value
2048: * - value is copied into sockopt
1.176 plunky 2049: * - memory is allocated when necessary, will not sleep
1.171 plunky 2050: */
2051: int
2052: sockopt_set(struct sockopt *sopt, const void *buf, size_t len)
2053: {
1.176 plunky 2054: int error;
1.171 plunky 2055:
1.176 plunky 2056: if (sopt->sopt_size == 0) {
2057: error = sockopt_alloc(sopt, len, KM_NOSLEEP);
2058: if (error)
2059: return error;
2060: }
1.171 plunky 2061:
2062: KASSERT(sopt->sopt_size == len);
2063: memcpy(sopt->sopt_data, buf, len);
2064: return 0;
2065: }
2066:
2067: /*
2068: * common case of set sockopt integer value
2069: */
2070: int
2071: sockopt_setint(struct sockopt *sopt, int val)
2072: {
2073:
2074: return sockopt_set(sopt, &val, sizeof(int));
2075: }
2076:
2077: /*
2078: * get sockopt value
2079: * - correct size must be given
2080: */
2081: int
2082: sockopt_get(const struct sockopt *sopt, void *buf, size_t len)
2083: {
1.170 tls 2084:
1.171 plunky 2085: if (sopt->sopt_size != len)
2086: return EINVAL;
1.1 cgd 2087:
1.171 plunky 2088: memcpy(buf, sopt->sopt_data, len);
2089: return 0;
2090: }
1.1 cgd 2091:
1.171 plunky 2092: /*
2093: * common case of get sockopt integer value
2094: */
2095: int
2096: sockopt_getint(const struct sockopt *sopt, int *valp)
2097: {
1.1 cgd 2098:
1.171 plunky 2099: return sockopt_get(sopt, valp, sizeof(int));
2100: }
1.1 cgd 2101:
1.171 plunky 2102: /*
2103: * set sockopt value from mbuf
2104: * - ONLY for legacy code
2105: * - mbuf is released by sockopt
1.176 plunky 2106: * - will not sleep
1.171 plunky 2107: */
2108: int
2109: sockopt_setmbuf(struct sockopt *sopt, struct mbuf *m)
2110: {
2111: size_t len;
1.176 plunky 2112: int error;
1.1 cgd 2113:
1.171 plunky 2114: len = m_length(m);
1.1 cgd 2115:
1.176 plunky 2116: if (sopt->sopt_size == 0) {
2117: error = sockopt_alloc(sopt, len, KM_NOSLEEP);
2118: if (error)
2119: return error;
2120: }
1.1 cgd 2121:
1.171 plunky 2122: KASSERT(sopt->sopt_size == len);
2123: m_copydata(m, 0, len, sopt->sopt_data);
2124: m_freem(m);
1.1 cgd 2125:
1.171 plunky 2126: return 0;
2127: }
1.1 cgd 2128:
1.171 plunky 2129: /*
2130: * get sockopt value into mbuf
2131: * - ONLY for legacy code
2132: * - mbuf to be released by the caller
1.176 plunky 2133: * - will not sleep
1.171 plunky 2134: */
2135: struct mbuf *
2136: sockopt_getmbuf(const struct sockopt *sopt)
2137: {
2138: struct mbuf *m;
1.107 darrenr 2139:
1.176 plunky 2140: if (sopt->sopt_size > MCLBYTES)
2141: return NULL;
2142:
2143: m = m_get(M_DONTWAIT, MT_SOOPTS);
1.171 plunky 2144: if (m == NULL)
2145: return NULL;
2146:
1.176 plunky 2147: if (sopt->sopt_size > MLEN) {
2148: MCLGET(m, M_DONTWAIT);
2149: if ((m->m_flags & M_EXT) == 0) {
2150: m_free(m);
2151: return NULL;
2152: }
1.1 cgd 2153: }
1.176 plunky 2154:
2155: memcpy(mtod(m, void *), sopt->sopt_data, sopt->sopt_size);
2156: m->m_len = sopt->sopt_size;
1.160 ad 2157:
1.171 plunky 2158: return m;
1.1 cgd 2159: }
2160:
1.14 mycroft 2161: void
1.54 lukem 2162: sohasoutofband(struct socket *so)
1.1 cgd 2163: {
1.153 rmind 2164:
1.90 christos 2165: fownsignal(so->so_pgid, SIGURG, POLL_PRI, POLLPRI|POLLRDBAND, so);
1.189 ad 2166: selnotify(&so->so_rcv.sb_sel, POLLPRI | POLLRDBAND, NOTE_SUBMIT);
1.1 cgd 2167: }
1.72 jdolecek 2168:
2169: static void
2170: filt_sordetach(struct knote *kn)
2171: {
2172: struct socket *so;
2173:
1.155 ad 2174: so = ((file_t *)kn->kn_obj)->f_data;
1.160 ad 2175: solock(so);
1.73 christos 2176: SLIST_REMOVE(&so->so_rcv.sb_sel.sel_klist, kn, knote, kn_selnext);
2177: if (SLIST_EMPTY(&so->so_rcv.sb_sel.sel_klist))
1.72 jdolecek 2178: so->so_rcv.sb_flags &= ~SB_KNOTE;
1.160 ad 2179: sounlock(so);
1.72 jdolecek 2180: }
2181:
2182: /*ARGSUSED*/
2183: static int
1.129 yamt 2184: filt_soread(struct knote *kn, long hint)
1.72 jdolecek 2185: {
2186: struct socket *so;
1.160 ad 2187: int rv;
1.72 jdolecek 2188:
1.155 ad 2189: so = ((file_t *)kn->kn_obj)->f_data;
1.160 ad 2190: if (hint != NOTE_SUBMIT)
2191: solock(so);
1.72 jdolecek 2192: kn->kn_data = so->so_rcv.sb_cc;
2193: if (so->so_state & SS_CANTRCVMORE) {
1.108 perry 2194: kn->kn_flags |= EV_EOF;
1.72 jdolecek 2195: kn->kn_fflags = so->so_error;
1.160 ad 2196: rv = 1;
2197: } else if (so->so_error) /* temporary udp error */
2198: rv = 1;
2199: else if (kn->kn_sfflags & NOTE_LOWAT)
2200: rv = (kn->kn_data >= kn->kn_sdata);
2201: else
2202: rv = (kn->kn_data >= so->so_rcv.sb_lowat);
2203: if (hint != NOTE_SUBMIT)
2204: sounlock(so);
2205: return rv;
1.72 jdolecek 2206: }
2207:
2208: static void
2209: filt_sowdetach(struct knote *kn)
2210: {
2211: struct socket *so;
2212:
1.155 ad 2213: so = ((file_t *)kn->kn_obj)->f_data;
1.160 ad 2214: solock(so);
1.73 christos 2215: SLIST_REMOVE(&so->so_snd.sb_sel.sel_klist, kn, knote, kn_selnext);
2216: if (SLIST_EMPTY(&so->so_snd.sb_sel.sel_klist))
1.72 jdolecek 2217: so->so_snd.sb_flags &= ~SB_KNOTE;
1.160 ad 2218: sounlock(so);
1.72 jdolecek 2219: }
2220:
2221: /*ARGSUSED*/
2222: static int
1.129 yamt 2223: filt_sowrite(struct knote *kn, long hint)
1.72 jdolecek 2224: {
2225: struct socket *so;
1.160 ad 2226: int rv;
1.72 jdolecek 2227:
1.155 ad 2228: so = ((file_t *)kn->kn_obj)->f_data;
1.160 ad 2229: if (hint != NOTE_SUBMIT)
2230: solock(so);
1.72 jdolecek 2231: kn->kn_data = sbspace(&so->so_snd);
2232: if (so->so_state & SS_CANTSENDMORE) {
1.108 perry 2233: kn->kn_flags |= EV_EOF;
1.72 jdolecek 2234: kn->kn_fflags = so->so_error;
1.160 ad 2235: rv = 1;
2236: } else if (so->so_error) /* temporary udp error */
2237: rv = 1;
2238: else if (((so->so_state & SS_ISCONNECTED) == 0) &&
1.72 jdolecek 2239: (so->so_proto->pr_flags & PR_CONNREQUIRED))
1.160 ad 2240: rv = 0;
2241: else if (kn->kn_sfflags & NOTE_LOWAT)
2242: rv = (kn->kn_data >= kn->kn_sdata);
2243: else
2244: rv = (kn->kn_data >= so->so_snd.sb_lowat);
2245: if (hint != NOTE_SUBMIT)
2246: sounlock(so);
2247: return rv;
1.72 jdolecek 2248: }
2249:
2250: /*ARGSUSED*/
2251: static int
1.129 yamt 2252: filt_solisten(struct knote *kn, long hint)
1.72 jdolecek 2253: {
2254: struct socket *so;
1.160 ad 2255: int rv;
1.72 jdolecek 2256:
1.155 ad 2257: so = ((file_t *)kn->kn_obj)->f_data;
1.72 jdolecek 2258:
2259: /*
2260: * Set kn_data to number of incoming connections, not
2261: * counting partial (incomplete) connections.
1.108 perry 2262: */
1.160 ad 2263: if (hint != NOTE_SUBMIT)
2264: solock(so);
1.72 jdolecek 2265: kn->kn_data = so->so_qlen;
1.160 ad 2266: rv = (kn->kn_data > 0);
2267: if (hint != NOTE_SUBMIT)
2268: sounlock(so);
2269: return rv;
1.72 jdolecek 2270: }
2271:
2272: static const struct filterops solisten_filtops =
2273: { 1, NULL, filt_sordetach, filt_solisten };
2274: static const struct filterops soread_filtops =
2275: { 1, NULL, filt_sordetach, filt_soread };
2276: static const struct filterops sowrite_filtops =
2277: { 1, NULL, filt_sowdetach, filt_sowrite };
2278:
2279: int
1.129 yamt 2280: soo_kqfilter(struct file *fp, struct knote *kn)
1.72 jdolecek 2281: {
2282: struct socket *so;
2283: struct sockbuf *sb;
2284:
1.155 ad 2285: so = ((file_t *)kn->kn_obj)->f_data;
1.160 ad 2286: solock(so);
1.72 jdolecek 2287: switch (kn->kn_filter) {
2288: case EVFILT_READ:
2289: if (so->so_options & SO_ACCEPTCONN)
2290: kn->kn_fop = &solisten_filtops;
2291: else
2292: kn->kn_fop = &soread_filtops;
2293: sb = &so->so_rcv;
2294: break;
2295: case EVFILT_WRITE:
2296: kn->kn_fop = &sowrite_filtops;
2297: sb = &so->so_snd;
2298: break;
2299: default:
1.160 ad 2300: sounlock(so);
1.149 pooka 2301: return (EINVAL);
1.72 jdolecek 2302: }
1.73 christos 2303: SLIST_INSERT_HEAD(&sb->sb_sel.sel_klist, kn, kn_selnext);
1.72 jdolecek 2304: sb->sb_flags |= SB_KNOTE;
1.160 ad 2305: sounlock(so);
1.72 jdolecek 2306: return (0);
2307: }
2308:
1.154 ad 2309: static int
2310: sodopoll(struct socket *so, int events)
2311: {
2312: int revents;
2313:
2314: revents = 0;
2315:
2316: if (events & (POLLIN | POLLRDNORM))
2317: if (soreadable(so))
2318: revents |= events & (POLLIN | POLLRDNORM);
2319:
2320: if (events & (POLLOUT | POLLWRNORM))
2321: if (sowritable(so))
2322: revents |= events & (POLLOUT | POLLWRNORM);
2323:
2324: if (events & (POLLPRI | POLLRDBAND))
2325: if (so->so_oobmark || (so->so_state & SS_RCVATMARK))
2326: revents |= events & (POLLPRI | POLLRDBAND);
2327:
2328: return revents;
2329: }
2330:
2331: int
2332: sopoll(struct socket *so, int events)
2333: {
2334: int revents = 0;
2335:
1.160 ad 2336: #ifndef DIAGNOSTIC
2337: /*
2338: * Do a quick, unlocked check in expectation that the socket
2339: * will be ready for I/O. Don't do this check if DIAGNOSTIC,
2340: * as the solocked() assertions will fail.
2341: */
1.154 ad 2342: if ((revents = sodopoll(so, events)) != 0)
2343: return revents;
1.160 ad 2344: #endif
1.154 ad 2345:
1.160 ad 2346: solock(so);
1.154 ad 2347: if ((revents = sodopoll(so, events)) == 0) {
2348: if (events & (POLLIN | POLLPRI | POLLRDNORM | POLLRDBAND)) {
2349: selrecord(curlwp, &so->so_rcv.sb_sel);
1.160 ad 2350: so->so_rcv.sb_flags |= SB_NOTIFY;
1.154 ad 2351: }
2352:
2353: if (events & (POLLOUT | POLLWRNORM)) {
2354: selrecord(curlwp, &so->so_snd.sb_sel);
1.160 ad 2355: so->so_snd.sb_flags |= SB_NOTIFY;
1.154 ad 2356: }
2357: }
1.160 ad 2358: sounlock(so);
1.154 ad 2359:
2360: return revents;
2361: }
2362:
2363:
1.94 yamt 2364: #include <sys/sysctl.h>
2365:
2366: static int sysctl_kern_somaxkva(SYSCTLFN_PROTO);
2367:
2368: /*
2369: * sysctl helper routine for kern.somaxkva. ensures that the given
2370: * value is not too small.
2371: * (XXX should we maybe make sure it's not too large as well?)
2372: */
2373: static int
2374: sysctl_kern_somaxkva(SYSCTLFN_ARGS)
2375: {
2376: int error, new_somaxkva;
2377: struct sysctlnode node;
2378:
2379: new_somaxkva = somaxkva;
2380: node = *rnode;
2381: node.sysctl_data = &new_somaxkva;
2382: error = sysctl_lookup(SYSCTLFN_CALL(&node));
2383: if (error || newp == NULL)
2384: return (error);
2385:
2386: if (new_somaxkva < (16 * 1024 * 1024)) /* sanity */
2387: return (EINVAL);
2388:
1.136 ad 2389: mutex_enter(&so_pendfree_lock);
1.94 yamt 2390: somaxkva = new_somaxkva;
1.136 ad 2391: cv_broadcast(&socurkva_cv);
2392: mutex_exit(&so_pendfree_lock);
1.94 yamt 2393:
2394: return (error);
2395: }
2396:
1.178 pooka 2397: static void
1.187 cegger 2398: sysctl_kern_somaxkva_setup(void)
1.94 yamt 2399: {
2400:
1.178 pooka 2401: KASSERT(socket_sysctllog == NULL);
2402: sysctl_createv(&socket_sysctllog, 0, NULL, NULL,
1.97 atatat 2403: CTLFLAG_PERMANENT,
2404: CTLTYPE_NODE, "kern", NULL,
2405: NULL, 0, NULL, 0,
2406: CTL_KERN, CTL_EOL);
2407:
1.178 pooka 2408: sysctl_createv(&socket_sysctllog, 0, NULL, NULL,
1.97 atatat 2409: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.103 atatat 2410: CTLTYPE_INT, "somaxkva",
2411: SYSCTL_DESCR("Maximum amount of kernel memory to be "
2412: "used for socket buffers"),
1.94 yamt 2413: sysctl_kern_somaxkva, 0, NULL, 0,
2414: CTL_KERN, KERN_SOMAXKVA, CTL_EOL);
2415: }
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