[BACK]Return to uipc_socket.c CVS log [TXT][DIR] Up to [cvs.NetBSD.org] / src / sys / kern

Annotation of src/sys/kern/uipc_socket.c, Revision 1.198

1.198   ! elad        1: /*     $NetBSD: uipc_socket.c,v 1.197 2009/12/29 03:48:18 elad 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>
1.198   ! elad       66: __KERNEL_RCSID(0, "$NetBSD: uipc_socket.c,v 1.197 2009/12/29 03:48:18 elad 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>
1.37      thorpej    95:
1.179     christos   96: #ifdef COMPAT_50
                     97: #include <compat/sys/time.h>
1.184     christos   98: #include <compat/sys/socket.h>
1.179     christos   99: #endif
                    100:
1.64      thorpej   101: #include <uvm/uvm.h>
                    102:
1.77      thorpej   103: MALLOC_DEFINE(M_SOOPTS, "soopts", "socket options");
                    104: MALLOC_DEFINE(M_SONAME, "soname", "socket name");
1.37      thorpej   105:
1.142     dyoung    106: extern const struct fileops socketops;
                    107:
1.54      lukem     108: extern int     somaxconn;                      /* patchable (XXX sysctl) */
                    109: int            somaxconn = SOMAXCONN;
1.160     ad        110: kmutex_t       *softnet_lock;
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:
1.167     ad        138: #if defined(SOSEND_NO_LOAN) || defined(MULTIPROCESSOR)
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:
1.191     elad      154: static kauth_listener_t socket_listener;
                    155:
1.64      thorpej   156: #define        SOCK_LOAN_CHUNK         65536
                    157:
1.117     yamt      158: static size_t sodopendfree(void);
                    159: static size_t sodopendfreel(void);
1.93      yamt      160:
1.178     pooka     161: static void sysctl_kern_somaxkva_setup(void);
                    162: static struct sysctllog *socket_sysctllog;
                    163:
1.113     thorpej   164: static vsize_t
1.129     yamt      165: sokvareserve(struct socket *so, vsize_t len)
1.80      yamt      166: {
1.98      christos  167:        int error;
1.80      yamt      168:
1.136     ad        169:        mutex_enter(&so_pendfree_lock);
1.80      yamt      170:        while (socurkva + len > somaxkva) {
1.93      yamt      171:                size_t freed;
                    172:
                    173:                /*
                    174:                 * try to do pendfree.
                    175:                 */
                    176:
1.117     yamt      177:                freed = sodopendfreel();
1.93      yamt      178:
                    179:                /*
                    180:                 * if some kva was freed, try again.
                    181:                 */
                    182:
                    183:                if (freed)
1.80      yamt      184:                        continue;
1.93      yamt      185:
1.80      yamt      186:                SOSEND_COUNTER_INCR(&sosend_kvalimit);
1.136     ad        187:                error = cv_wait_sig(&socurkva_cv, &so_pendfree_lock);
1.98      christos  188:                if (error) {
                    189:                        len = 0;
                    190:                        break;
                    191:                }
1.80      yamt      192:        }
1.93      yamt      193:        socurkva += len;
1.136     ad        194:        mutex_exit(&so_pendfree_lock);
1.98      christos  195:        return len;
1.95      yamt      196: }
                    197:
1.113     thorpej   198: static void
1.95      yamt      199: sokvaunreserve(vsize_t len)
                    200: {
                    201:
1.136     ad        202:        mutex_enter(&so_pendfree_lock);
1.95      yamt      203:        socurkva -= len;
1.136     ad        204:        cv_broadcast(&socurkva_cv);
                    205:        mutex_exit(&so_pendfree_lock);
1.95      yamt      206: }
                    207:
                    208: /*
                    209:  * sokvaalloc: allocate kva for loan.
                    210:  */
                    211:
                    212: vaddr_t
                    213: sokvaalloc(vsize_t len, struct socket *so)
                    214: {
                    215:        vaddr_t lva;
                    216:
                    217:        /*
                    218:         * reserve kva.
                    219:         */
                    220:
1.98      christos  221:        if (sokvareserve(so, len) == 0)
                    222:                return 0;
1.93      yamt      223:
                    224:        /*
                    225:         * allocate kva.
                    226:         */
1.80      yamt      227:
1.109     yamt      228:        lva = uvm_km_alloc(kernel_map, len, 0, UVM_KMF_VAONLY | UVM_KMF_WAITVA);
1.95      yamt      229:        if (lva == 0) {
                    230:                sokvaunreserve(len);
1.80      yamt      231:                return (0);
1.95      yamt      232:        }
1.80      yamt      233:
                    234:        return lva;
                    235: }
                    236:
1.93      yamt      237: /*
                    238:  * sokvafree: free kva for loan.
                    239:  */
                    240:
1.80      yamt      241: void
                    242: sokvafree(vaddr_t sva, vsize_t len)
                    243: {
1.93      yamt      244:
                    245:        /*
                    246:         * free kva.
                    247:         */
1.80      yamt      248:
1.109     yamt      249:        uvm_km_free(kernel_map, sva, len, UVM_KMF_VAONLY);
1.93      yamt      250:
                    251:        /*
                    252:         * unreserve kva.
                    253:         */
                    254:
1.95      yamt      255:        sokvaunreserve(len);
1.80      yamt      256: }
                    257:
1.64      thorpej   258: static void
1.134     christos  259: sodoloanfree(struct vm_page **pgs, void *buf, size_t size)
1.64      thorpej   260: {
1.156     yamt      261:        vaddr_t sva, eva;
1.64      thorpej   262:        vsize_t len;
1.156     yamt      263:        int npgs;
                    264:
                    265:        KASSERT(pgs != NULL);
1.64      thorpej   266:
                    267:        eva = round_page((vaddr_t) buf + size);
                    268:        sva = trunc_page((vaddr_t) buf);
                    269:        len = eva - sva;
                    270:        npgs = len >> PAGE_SHIFT;
                    271:
                    272:        pmap_kremove(sva, len);
                    273:        pmap_update(pmap_kernel());
                    274:        uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE);
1.80      yamt      275:        sokvafree(sva, len);
1.64      thorpej   276: }
                    277:
                    278: static size_t
1.152     matt      279: sodopendfree(void)
1.64      thorpej   280: {
1.93      yamt      281:        size_t rv;
1.64      thorpej   282:
1.160     ad        283:        if (__predict_true(so_pendfree == NULL))
                    284:                return 0;
                    285:
1.136     ad        286:        mutex_enter(&so_pendfree_lock);
1.117     yamt      287:        rv = sodopendfreel();
1.136     ad        288:        mutex_exit(&so_pendfree_lock);
1.93      yamt      289:
                    290:        return rv;
                    291: }
                    292:
                    293: /*
                    294:  * sodopendfreel: free mbufs on "pendfree" list.
1.136     ad        295:  * unlock and relock so_pendfree_lock when freeing mbufs.
1.93      yamt      296:  *
1.136     ad        297:  * => called with so_pendfree_lock held.
1.93      yamt      298:  */
                    299:
                    300: static size_t
1.152     matt      301: sodopendfreel(void)
1.93      yamt      302: {
1.137     ad        303:        struct mbuf *m, *next;
1.93      yamt      304:        size_t rv = 0;
                    305:
1.136     ad        306:        KASSERT(mutex_owned(&so_pendfree_lock));
1.64      thorpej   307:
1.137     ad        308:        while (so_pendfree != NULL) {
1.64      thorpej   309:                m = so_pendfree;
1.93      yamt      310:                so_pendfree = NULL;
1.136     ad        311:                mutex_exit(&so_pendfree_lock);
1.93      yamt      312:
                    313:                for (; m != NULL; m = next) {
                    314:                        next = m->m_next;
1.156     yamt      315:                        KASSERT((~m->m_flags & (M_EXT|M_EXT_PAGES)) == 0);
                    316:                        KASSERT(m->m_ext.ext_refcnt == 0);
1.93      yamt      317:
                    318:                        rv += m->m_ext.ext_size;
1.156     yamt      319:                        sodoloanfree(m->m_ext.ext_pgs, m->m_ext.ext_buf,
1.93      yamt      320:                            m->m_ext.ext_size);
1.145     ad        321:                        pool_cache_put(mb_cache, m);
1.93      yamt      322:                }
1.64      thorpej   323:
1.136     ad        324:                mutex_enter(&so_pendfree_lock);
1.64      thorpej   325:        }
                    326:
                    327:        return (rv);
                    328: }
                    329:
1.80      yamt      330: void
1.134     christos  331: soloanfree(struct mbuf *m, void *buf, size_t size, void *arg)
1.64      thorpej   332: {
                    333:
1.156     yamt      334:        KASSERT(m != NULL);
1.64      thorpej   335:
1.93      yamt      336:        /*
                    337:         * postpone freeing mbuf.
                    338:         *
                    339:         * we can't do it in interrupt context
                    340:         * because we need to put kva back to kernel_map.
                    341:         */
                    342:
1.136     ad        343:        mutex_enter(&so_pendfree_lock);
1.92      yamt      344:        m->m_next = so_pendfree;
                    345:        so_pendfree = m;
1.136     ad        346:        cv_broadcast(&socurkva_cv);
                    347:        mutex_exit(&so_pendfree_lock);
1.64      thorpej   348: }
                    349:
                    350: static long
                    351: sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space)
                    352: {
                    353:        struct iovec *iov = uio->uio_iov;
                    354:        vaddr_t sva, eva;
                    355:        vsize_t len;
1.156     yamt      356:        vaddr_t lva;
                    357:        int npgs, error;
                    358:        vaddr_t va;
                    359:        int i;
1.64      thorpej   360:
1.116     yamt      361:        if (VMSPACE_IS_KERNEL_P(uio->uio_vmspace))
1.64      thorpej   362:                return (0);
                    363:
                    364:        if (iov->iov_len < (size_t) space)
                    365:                space = iov->iov_len;
                    366:        if (space > SOCK_LOAN_CHUNK)
                    367:                space = SOCK_LOAN_CHUNK;
                    368:
                    369:        eva = round_page((vaddr_t) iov->iov_base + space);
                    370:        sva = trunc_page((vaddr_t) iov->iov_base);
                    371:        len = eva - sva;
                    372:        npgs = len >> PAGE_SHIFT;
                    373:
1.79      thorpej   374:        KASSERT(npgs <= M_EXT_MAXPAGES);
                    375:
1.80      yamt      376:        lva = sokvaalloc(len, so);
1.64      thorpej   377:        if (lva == 0)
1.80      yamt      378:                return 0;
1.64      thorpej   379:
1.116     yamt      380:        error = uvm_loan(&uio->uio_vmspace->vm_map, sva, len,
1.79      thorpej   381:            m->m_ext.ext_pgs, UVM_LOAN_TOPAGE);
1.64      thorpej   382:        if (error) {
1.80      yamt      383:                sokvafree(lva, len);
1.64      thorpej   384:                return (0);
                    385:        }
                    386:
                    387:        for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE)
1.79      thorpej   388:                pmap_kenter_pa(va, VM_PAGE_TO_PHYS(m->m_ext.ext_pgs[i]),
1.194     cegger    389:                    VM_PROT_READ, 0);
1.64      thorpej   390:        pmap_update(pmap_kernel());
                    391:
                    392:        lva += (vaddr_t) iov->iov_base & PAGE_MASK;
                    393:
1.134     christos  394:        MEXTADD(m, (void *) lva, space, M_MBUF, soloanfree, so);
1.79      thorpej   395:        m->m_flags |= M_EXT_PAGES | M_EXT_ROMAP;
1.64      thorpej   396:
                    397:        uio->uio_resid -= space;
                    398:        /* uio_offset not updated, not set/used for write(2) */
1.134     christos  399:        uio->uio_iov->iov_base = (char *)uio->uio_iov->iov_base + space;
1.64      thorpej   400:        uio->uio_iov->iov_len -= space;
                    401:        if (uio->uio_iov->iov_len == 0) {
                    402:                uio->uio_iov++;
                    403:                uio->uio_iovcnt--;
                    404:        }
                    405:
                    406:        return (space);
                    407: }
                    408:
1.119     yamt      409: static int
1.129     yamt      410: sokva_reclaim_callback(struct callback_entry *ce, void *obj, void *arg)
1.119     yamt      411: {
                    412:
                    413:        KASSERT(ce == &sokva_reclaimerentry);
                    414:        KASSERT(obj == NULL);
                    415:
                    416:        sodopendfree();
                    417:        if (!vm_map_starved_p(kernel_map)) {
                    418:                return CALLBACK_CHAIN_ABORT;
                    419:        }
                    420:        return CALLBACK_CHAIN_CONTINUE;
                    421: }
                    422:
1.142     dyoung    423: struct mbuf *
1.147     dyoung    424: getsombuf(struct socket *so, int type)
1.142     dyoung    425: {
                    426:        struct mbuf *m;
                    427:
1.147     dyoung    428:        m = m_get(M_WAIT, type);
1.142     dyoung    429:        MCLAIM(m, so->so_mowner);
                    430:        return m;
                    431: }
                    432:
1.191     elad      433: static int
                    434: socket_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie,
                    435:     void *arg0, void *arg1, void *arg2, void *arg3)
                    436: {
                    437:        int result;
                    438:        enum kauth_network_req req;
                    439:
                    440:        result = KAUTH_RESULT_DEFER;
                    441:        req = (enum kauth_network_req)arg0;
                    442:
1.193     elad      443:        if ((action != KAUTH_NETWORK_SOCKET) &&
                    444:            (action != KAUTH_NETWORK_BIND))
1.191     elad      445:                return result;
                    446:
                    447:        switch (req) {
1.193     elad      448:        case KAUTH_REQ_NETWORK_BIND_PORT:
                    449:                result = KAUTH_RESULT_ALLOW;
                    450:                break;
                    451:
1.191     elad      452:        case KAUTH_REQ_NETWORK_SOCKET_DROP: {
                    453:                /* Normal users can only drop their own connections. */
                    454:                struct socket *so = (struct socket *)arg1;
                    455:                uid_t sockuid = so->so_uidinfo->ui_uid;
                    456:
                    457:                if (sockuid == kauth_cred_getuid(cred) ||
                    458:                    sockuid == kauth_cred_geteuid(cred))
                    459:                        result = KAUTH_RESULT_ALLOW;
                    460:
                    461:                break;
                    462:                }
                    463:
                    464:        case KAUTH_REQ_NETWORK_SOCKET_OPEN:
                    465:                /* We allow "raw" routing/bluetooth sockets to anyone. */
                    466:                if ((u_long)arg1 == PF_ROUTE || (u_long)arg1 == PF_BLUETOOTH)
                    467:                        result = KAUTH_RESULT_ALLOW;
                    468:                else {
                    469:                        /* Privileged, let secmodel handle this. */
                    470:                        if ((u_long)arg2 == SOCK_RAW)
                    471:                                break;
                    472:                }
                    473:
                    474:                result = KAUTH_RESULT_ALLOW;
                    475:
                    476:                break;
                    477:
1.192     elad      478:        case KAUTH_REQ_NETWORK_SOCKET_CANSEE:
                    479:                result = KAUTH_RESULT_ALLOW;
                    480:
                    481:                break;
                    482:
1.191     elad      483:        default:
                    484:                break;
                    485:        }
                    486:
                    487:        return result;
                    488: }
                    489:
1.119     yamt      490: void
                    491: soinit(void)
                    492: {
                    493:
1.178     pooka     494:        sysctl_kern_somaxkva_setup();
                    495:
1.148     ad        496:        mutex_init(&so_pendfree_lock, MUTEX_DEFAULT, IPL_VM);
1.160     ad        497:        softnet_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE);
1.136     ad        498:        cv_init(&socurkva_cv, "sokva");
1.166     ad        499:        soinit2();
1.136     ad        500:
1.119     yamt      501:        /* Set the initial adjusted socket buffer size. */
                    502:        if (sb_max_set(sb_max))
                    503:                panic("bad initial sb_max value: %lu", sb_max);
                    504:
                    505:        callback_register(&vm_map_to_kernel(kernel_map)->vmk_reclaim_callback,
                    506:            &sokva_reclaimerentry, NULL, sokva_reclaim_callback);
1.191     elad      507:
                    508:        socket_listener = kauth_listen_scope(KAUTH_SCOPE_NETWORK,
                    509:            socket_listener_cb, NULL);
1.119     yamt      510: }
                    511:
1.1       cgd       512: /*
                    513:  * Socket operation routines.
                    514:  * These routines are called by the routines in
                    515:  * sys_socket.c or from a system process, and
                    516:  * implement the semantics of socket operations by
                    517:  * switching out to the protocol specific routines.
                    518:  */
                    519: /*ARGSUSED*/
1.3       andrew    520: int
1.160     ad        521: socreate(int dom, struct socket **aso, int type, int proto, struct lwp *l,
                    522:         struct socket *lockso)
1.1       cgd       523: {
1.99      matt      524:        const struct protosw    *prp;
1.54      lukem     525:        struct socket   *so;
1.115     yamt      526:        uid_t           uid;
1.160     ad        527:        int             error;
                    528:        kmutex_t        *lock;
1.1       cgd       529:
1.132     elad      530:        error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_SOCKET,
                    531:            KAUTH_REQ_NETWORK_SOCKET_OPEN, KAUTH_ARG(dom), KAUTH_ARG(type),
                    532:            KAUTH_ARG(proto));
1.140     dyoung    533:        if (error != 0)
                    534:                return error;
1.127     elad      535:
1.1       cgd       536:        if (proto)
                    537:                prp = pffindproto(dom, proto, type);
                    538:        else
                    539:                prp = pffindtype(dom, type);
1.140     dyoung    540:        if (prp == NULL) {
1.120     ginsbach  541:                /* no support for domain */
                    542:                if (pffinddomain(dom) == 0)
1.140     dyoung    543:                        return EAFNOSUPPORT;
1.120     ginsbach  544:                /* no support for socket type */
                    545:                if (proto == 0 && type != 0)
1.140     dyoung    546:                        return EPROTOTYPE;
                    547:                return EPROTONOSUPPORT;
1.120     ginsbach  548:        }
1.140     dyoung    549:        if (prp->pr_usrreq == NULL)
                    550:                return EPROTONOSUPPORT;
1.1       cgd       551:        if (prp->pr_type != type)
1.140     dyoung    552:                return EPROTOTYPE;
1.160     ad        553:
                    554:        so = soget(true);
1.1       cgd       555:        so->so_type = type;
                    556:        so->so_proto = prp;
1.33      matt      557:        so->so_send = sosend;
                    558:        so->so_receive = soreceive;
1.78      matt      559: #ifdef MBUFTRACE
                    560:        so->so_rcv.sb_mowner = &prp->pr_domain->dom_mowner;
                    561:        so->so_snd.sb_mowner = &prp->pr_domain->dom_mowner;
                    562:        so->so_mowner = &prp->pr_domain->dom_mowner;
                    563: #endif
1.138     rmind     564:        uid = kauth_cred_geteuid(l->l_cred);
1.115     yamt      565:        so->so_uidinfo = uid_find(uid);
1.168     yamt      566:        so->so_egid = kauth_cred_getegid(l->l_cred);
                    567:        so->so_cpid = l->l_proc->p_pid;
1.160     ad        568:        if (lockso != NULL) {
                    569:                /* Caller wants us to share a lock. */
                    570:                lock = lockso->so_lock;
                    571:                so->so_lock = lock;
                    572:                mutex_obj_hold(lock);
                    573:                mutex_enter(lock);
                    574:        } else {
                    575:                /* Lock assigned and taken during PRU_ATTACH. */
                    576:        }
1.140     dyoung    577:        error = (*prp->pr_usrreq)(so, PRU_ATTACH, NULL,
                    578:            (struct mbuf *)(long)proto, NULL, l);
1.160     ad        579:        KASSERT(solocked(so));
1.140     dyoung    580:        if (error != 0) {
1.1       cgd       581:                so->so_state |= SS_NOFDREF;
                    582:                sofree(so);
1.140     dyoung    583:                return error;
1.1       cgd       584:        }
1.198   ! elad      585:        so->so_cred = kauth_cred_dup(l->l_cred);
1.160     ad        586:        sounlock(so);
1.1       cgd       587:        *aso = so;
1.140     dyoung    588:        return 0;
1.1       cgd       589: }
                    590:
1.142     dyoung    591: /* On success, write file descriptor to fdout and return zero.  On
                    592:  * failure, return non-zero; *fdout will be undefined.
                    593:  */
                    594: int
                    595: fsocreate(int domain, struct socket **sop, int type, int protocol,
                    596:     struct lwp *l, int *fdout)
                    597: {
                    598:        struct socket   *so;
                    599:        struct file     *fp;
                    600:        int             fd, error;
                    601:
1.155     ad        602:        if ((error = fd_allocfile(&fp, &fd)) != 0)
1.142     dyoung    603:                return (error);
                    604:        fp->f_flag = FREAD|FWRITE;
                    605:        fp->f_type = DTYPE_SOCKET;
                    606:        fp->f_ops = &socketops;
1.160     ad        607:        error = socreate(domain, &so, type, protocol, l, NULL);
1.142     dyoung    608:        if (error != 0) {
1.155     ad        609:                fd_abort(curproc, fp, fd);
1.142     dyoung    610:        } else {
                    611:                if (sop != NULL)
                    612:                        *sop = so;
                    613:                fp->f_data = so;
1.155     ad        614:                fd_affix(curproc, fp, fd);
1.142     dyoung    615:                *fdout = fd;
                    616:        }
                    617:        return error;
                    618: }
                    619:
1.3       andrew    620: int
1.190     dyoung    621: sofamily(const struct socket *so)
                    622: {
                    623:        const struct protosw *pr;
                    624:        const struct domain *dom;
                    625:
                    626:        if ((pr = so->so_proto) == NULL)
                    627:                return AF_UNSPEC;
                    628:        if ((dom = pr->pr_domain) == NULL)
                    629:                return AF_UNSPEC;
                    630:        return dom->dom_family;
                    631: }
                    632:
                    633: int
1.114     christos  634: sobind(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1       cgd       635: {
1.160     ad        636:        int     error;
1.1       cgd       637:
1.160     ad        638:        solock(so);
1.140     dyoung    639:        error = (*so->so_proto->pr_usrreq)(so, PRU_BIND, NULL, nam, NULL, l);
1.160     ad        640:        sounlock(so);
1.140     dyoung    641:        return error;
1.1       cgd       642: }
                    643:
1.3       andrew    644: int
1.150     elad      645: solisten(struct socket *so, int backlog, struct lwp *l)
1.1       cgd       646: {
1.160     ad        647:        int     error;
1.1       cgd       648:
1.160     ad        649:        solock(so);
1.158     ad        650:        if ((so->so_state & (SS_ISCONNECTED | SS_ISCONNECTING |
1.163     ad        651:            SS_ISDISCONNECTING)) != 0) {
                    652:                sounlock(so);
1.158     ad        653:                return (EOPNOTSUPP);
1.163     ad        654:        }
1.140     dyoung    655:        error = (*so->so_proto->pr_usrreq)(so, PRU_LISTEN, NULL,
1.150     elad      656:            NULL, NULL, l);
1.140     dyoung    657:        if (error != 0) {
1.160     ad        658:                sounlock(so);
1.140     dyoung    659:                return error;
1.1       cgd       660:        }
1.63      matt      661:        if (TAILQ_EMPTY(&so->so_q))
1.1       cgd       662:                so->so_options |= SO_ACCEPTCONN;
                    663:        if (backlog < 0)
                    664:                backlog = 0;
1.49      jonathan  665:        so->so_qlimit = min(backlog, somaxconn);
1.160     ad        666:        sounlock(so);
1.140     dyoung    667:        return 0;
1.1       cgd       668: }
                    669:
1.21      christos  670: void
1.54      lukem     671: sofree(struct socket *so)
1.1       cgd       672: {
1.161     ad        673:        u_int refs;
1.1       cgd       674:
1.160     ad        675:        KASSERT(solocked(so));
                    676:
                    677:        if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0) {
                    678:                sounlock(so);
1.1       cgd       679:                return;
1.160     ad        680:        }
1.43      mycroft   681:        if (so->so_head) {
                    682:                /*
                    683:                 * We must not decommission a socket that's on the accept(2)
                    684:                 * queue.  If we do, then accept(2) may hang after select(2)
                    685:                 * indicated that the listening socket was ready.
                    686:                 */
1.160     ad        687:                if (!soqremque(so, 0)) {
                    688:                        sounlock(so);
1.43      mycroft   689:                        return;
1.160     ad        690:                }
1.43      mycroft   691:        }
1.98      christos  692:        if (so->so_rcv.sb_hiwat)
1.110     christos  693:                (void)chgsbsize(so->so_uidinfo, &so->so_rcv.sb_hiwat, 0,
1.98      christos  694:                    RLIM_INFINITY);
                    695:        if (so->so_snd.sb_hiwat)
1.110     christos  696:                (void)chgsbsize(so->so_uidinfo, &so->so_snd.sb_hiwat, 0,
1.98      christos  697:                    RLIM_INFINITY);
                    698:        sbrelease(&so->so_snd, so);
1.160     ad        699:        KASSERT(!cv_has_waiters(&so->so_cv));
                    700:        KASSERT(!cv_has_waiters(&so->so_rcv.sb_cv));
                    701:        KASSERT(!cv_has_waiters(&so->so_snd.sb_cv));
1.1       cgd       702:        sorflush(so);
1.161     ad        703:        refs = so->so_aborting; /* XXX */
1.177     ad        704:        /* Remove acccept filter if one is present. */
1.170     tls       705:        if (so->so_accf != NULL)
1.177     ad        706:                (void)accept_filt_clear(so);
1.160     ad        707:        sounlock(so);
1.161     ad        708:        if (refs == 0)          /* XXX */
                    709:                soput(so);
1.1       cgd       710: }
                    711:
                    712: /*
                    713:  * Close a socket on last file table reference removal.
                    714:  * Initiate disconnect if connected.
                    715:  * Free socket when disconnect complete.
                    716:  */
1.3       andrew    717: int
1.54      lukem     718: soclose(struct socket *so)
1.1       cgd       719: {
1.54      lukem     720:        struct socket   *so2;
1.160     ad        721:        int             error;
                    722:        int             error2;
1.1       cgd       723:
1.54      lukem     724:        error = 0;
1.160     ad        725:        solock(so);
1.1       cgd       726:        if (so->so_options & SO_ACCEPTCONN) {
1.172     ad        727:                for (;;) {
                    728:                        if ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) {
1.160     ad        729:                                KASSERT(solocked2(so, so2));
                    730:                                (void) soqremque(so2, 0);
                    731:                                /* soabort drops the lock. */
                    732:                                (void) soabort(so2);
                    733:                                solock(so);
1.172     ad        734:                                continue;
1.160     ad        735:                        }
1.172     ad        736:                        if ((so2 = TAILQ_FIRST(&so->so_q)) != 0) {
1.160     ad        737:                                KASSERT(solocked2(so, so2));
                    738:                                (void) soqremque(so2, 1);
                    739:                                /* soabort drops the lock. */
                    740:                                (void) soabort(so2);
                    741:                                solock(so);
1.172     ad        742:                                continue;
1.160     ad        743:                        }
1.172     ad        744:                        break;
                    745:                }
1.1       cgd       746:        }
                    747:        if (so->so_pcb == 0)
                    748:                goto discard;
                    749:        if (so->so_state & SS_ISCONNECTED) {
                    750:                if ((so->so_state & SS_ISDISCONNECTING) == 0) {
                    751:                        error = sodisconnect(so);
                    752:                        if (error)
                    753:                                goto drop;
                    754:                }
                    755:                if (so->so_options & SO_LINGER) {
1.151     ad        756:                        if ((so->so_state & SS_ISDISCONNECTING) && so->so_nbio)
1.1       cgd       757:                                goto drop;
1.21      christos  758:                        while (so->so_state & SS_ISCONNECTED) {
1.185     yamt      759:                                error = sowait(so, true, so->so_linger * hz);
1.21      christos  760:                                if (error)
1.1       cgd       761:                                        break;
1.21      christos  762:                        }
1.1       cgd       763:                }
                    764:        }
1.54      lukem     765:  drop:
1.1       cgd       766:        if (so->so_pcb) {
1.160     ad        767:                error2 = (*so->so_proto->pr_usrreq)(so, PRU_DETACH,
1.140     dyoung    768:                    NULL, NULL, NULL, NULL);
1.1       cgd       769:                if (error == 0)
                    770:                        error = error2;
                    771:        }
1.54      lukem     772:  discard:
1.1       cgd       773:        if (so->so_state & SS_NOFDREF)
                    774:                panic("soclose: NOFDREF");
1.198   ! elad      775:        kauth_cred_free(so->so_cred);
1.1       cgd       776:        so->so_state |= SS_NOFDREF;
                    777:        sofree(so);
                    778:        return (error);
                    779: }
                    780:
                    781: /*
1.160     ad        782:  * Must be called with the socket locked..  Will return with it unlocked.
1.1       cgd       783:  */
1.3       andrew    784: int
1.54      lukem     785: soabort(struct socket *so)
1.1       cgd       786: {
1.161     ad        787:        u_int refs;
1.139     yamt      788:        int error;
1.160     ad        789:
                    790:        KASSERT(solocked(so));
                    791:        KASSERT(so->so_head == NULL);
1.1       cgd       792:
1.161     ad        793:        so->so_aborting++;              /* XXX */
1.140     dyoung    794:        error = (*so->so_proto->pr_usrreq)(so, PRU_ABORT, NULL,
                    795:            NULL, NULL, NULL);
1.161     ad        796:        refs = --so->so_aborting;       /* XXX */
1.164     drochner  797:        if (error || (refs == 0)) {
1.139     yamt      798:                sofree(so);
1.160     ad        799:        } else {
                    800:                sounlock(so);
1.139     yamt      801:        }
                    802:        return error;
1.1       cgd       803: }
                    804:
1.3       andrew    805: int
1.54      lukem     806: soaccept(struct socket *so, struct mbuf *nam)
1.1       cgd       807: {
1.160     ad        808:        int     error;
                    809:
                    810:        KASSERT(solocked(so));
1.1       cgd       811:
1.54      lukem     812:        error = 0;
1.1       cgd       813:        if ((so->so_state & SS_NOFDREF) == 0)
                    814:                panic("soaccept: !NOFDREF");
                    815:        so->so_state &= ~SS_NOFDREF;
1.55      thorpej   816:        if ((so->so_state & SS_ISDISCONNECTED) == 0 ||
                    817:            (so->so_proto->pr_flags & PR_ABRTACPTDIS) == 0)
1.41      mycroft   818:                error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT,
1.140     dyoung    819:                    NULL, nam, NULL, NULL);
1.41      mycroft   820:        else
1.53      itojun    821:                error = ECONNABORTED;
1.52      itojun    822:
1.1       cgd       823:        return (error);
                    824: }
                    825:
1.3       andrew    826: int
1.114     christos  827: soconnect(struct socket *so, struct mbuf *nam, struct lwp *l)
1.1       cgd       828: {
1.160     ad        829:        int             error;
                    830:
                    831:        KASSERT(solocked(so));
1.1       cgd       832:
                    833:        if (so->so_options & SO_ACCEPTCONN)
                    834:                return (EOPNOTSUPP);
                    835:        /*
                    836:         * If protocol is connection-based, can only connect once.
                    837:         * Otherwise, if connected, try to disconnect first.
                    838:         * This allows user to disconnect by connecting to, e.g.,
                    839:         * a null address.
                    840:         */
                    841:        if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
                    842:            ((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
                    843:            (error = sodisconnect(so))))
                    844:                error = EISCONN;
                    845:        else
                    846:                error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
1.140     dyoung    847:                    NULL, nam, NULL, l);
1.1       cgd       848:        return (error);
                    849: }
                    850:
1.3       andrew    851: int
1.54      lukem     852: soconnect2(struct socket *so1, struct socket *so2)
1.1       cgd       853: {
1.160     ad        854:        int     error;
                    855:
                    856:        KASSERT(solocked2(so1, so2));
1.1       cgd       857:
1.22      mycroft   858:        error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2,
1.140     dyoung    859:            NULL, (struct mbuf *)so2, NULL, NULL);
1.1       cgd       860:        return (error);
                    861: }
                    862:
1.3       andrew    863: int
1.54      lukem     864: sodisconnect(struct socket *so)
1.1       cgd       865: {
1.160     ad        866:        int     error;
                    867:
                    868:        KASSERT(solocked(so));
1.1       cgd       869:
                    870:        if ((so->so_state & SS_ISCONNECTED) == 0) {
                    871:                error = ENOTCONN;
1.160     ad        872:        } else if (so->so_state & SS_ISDISCONNECTING) {
1.1       cgd       873:                error = EALREADY;
1.160     ad        874:        } else {
                    875:                error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
                    876:                    NULL, NULL, NULL, NULL);
1.1       cgd       877:        }
1.117     yamt      878:        sodopendfree();
1.1       cgd       879:        return (error);
                    880: }
                    881:
1.15      mycroft   882: #define        SBLOCKWAIT(f)   (((f) & MSG_DONTWAIT) ? M_NOWAIT : M_WAITOK)
1.1       cgd       883: /*
                    884:  * Send on a socket.
                    885:  * If send must go all at once and message is larger than
                    886:  * send buffering, then hard error.
                    887:  * Lock against other senders.
                    888:  * If must go all at once and not enough room now, then
                    889:  * inform user that this would block and do nothing.
                    890:  * Otherwise, if nonblocking, send as much as possible.
                    891:  * The data to be sent is described by "uio" if nonzero,
                    892:  * otherwise by the mbuf chain "top" (which must be null
                    893:  * if uio is not).  Data provided in mbuf chain must be small
                    894:  * enough to send all at once.
                    895:  *
                    896:  * Returns nonzero on error, timeout or signal; callers
                    897:  * must check for short counts if EINTR/ERESTART are returned.
                    898:  * Data and control buffers are freed on return.
                    899:  */
1.3       andrew    900: int
1.54      lukem     901: sosend(struct socket *so, struct mbuf *addr, struct uio *uio, struct mbuf *top,
1.114     christos  902:        struct mbuf *control, int flags, struct lwp *l)
1.1       cgd       903: {
1.54      lukem     904:        struct mbuf     **mp, *m;
1.114     christos  905:        struct proc     *p;
1.58      jdolecek  906:        long            space, len, resid, clen, mlen;
                    907:        int             error, s, dontroute, atomic;
1.196     dsl       908:        short           wakeup_state = 0;
1.54      lukem     909:
1.114     christos  910:        p = l->l_proc;
1.117     yamt      911:        sodopendfree();
1.160     ad        912:        clen = 0;
1.64      thorpej   913:
1.160     ad        914:        /*
                    915:         * solock() provides atomicity of access.  splsoftnet() prevents
                    916:         * protocol processing soft interrupts from interrupting us and
                    917:         * blocking (expensive).
                    918:         */
                    919:        s = splsoftnet();
                    920:        solock(so);
1.54      lukem     921:        atomic = sosendallatonce(so) || top;
1.1       cgd       922:        if (uio)
                    923:                resid = uio->uio_resid;
                    924:        else
                    925:                resid = top->m_pkthdr.len;
1.7       cgd       926:        /*
                    927:         * In theory resid should be unsigned.
                    928:         * However, space must be signed, as it might be less than 0
                    929:         * if we over-committed, and we must use a signed comparison
                    930:         * of space and resid.  On the other hand, a negative resid
                    931:         * causes us to loop sending 0-length segments to the protocol.
                    932:         */
1.29      mycroft   933:        if (resid < 0) {
                    934:                error = EINVAL;
                    935:                goto out;
                    936:        }
1.1       cgd       937:        dontroute =
                    938:            (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
                    939:            (so->so_proto->pr_flags & PR_ATOMIC);
1.165     christos  940:        l->l_ru.ru_msgsnd++;
1.1       cgd       941:        if (control)
                    942:                clen = control->m_len;
1.54      lukem     943:  restart:
1.21      christos  944:        if ((error = sblock(&so->so_snd, SBLOCKWAIT(flags))) != 0)
1.1       cgd       945:                goto out;
                    946:        do {
1.160     ad        947:                if (so->so_state & SS_CANTSENDMORE) {
                    948:                        error = EPIPE;
                    949:                        goto release;
                    950:                }
1.48      thorpej   951:                if (so->so_error) {
                    952:                        error = so->so_error;
                    953:                        so->so_error = 0;
                    954:                        goto release;
                    955:                }
1.1       cgd       956:                if ((so->so_state & SS_ISCONNECTED) == 0) {
                    957:                        if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
                    958:                                if ((so->so_state & SS_ISCONFIRMING) == 0 &&
1.160     ad        959:                                    !(resid == 0 && clen != 0)) {
                    960:                                        error = ENOTCONN;
                    961:                                        goto release;
                    962:                                }
                    963:                        } else if (addr == 0) {
                    964:                                error = EDESTADDRREQ;
                    965:                                goto release;
                    966:                        }
1.1       cgd       967:                }
                    968:                space = sbspace(&so->so_snd);
                    969:                if (flags & MSG_OOB)
                    970:                        space += 1024;
1.21      christos  971:                if ((atomic && resid > so->so_snd.sb_hiwat) ||
1.160     ad        972:                    clen > so->so_snd.sb_hiwat) {
                    973:                        error = EMSGSIZE;
                    974:                        goto release;
                    975:                }
1.96      mycroft   976:                if (space < resid + clen &&
1.1       cgd       977:                    (atomic || space < so->so_snd.sb_lowat || space < clen)) {
1.160     ad        978:                        if (so->so_nbio) {
                    979:                                error = EWOULDBLOCK;
                    980:                                goto release;
                    981:                        }
1.1       cgd       982:                        sbunlock(&so->so_snd);
1.196     dsl       983:                        if (wakeup_state & SS_RESTARTSYS) {
                    984:                                error = ERESTART;
                    985:                                goto out;
                    986:                        }
1.1       cgd       987:                        error = sbwait(&so->so_snd);
                    988:                        if (error)
                    989:                                goto out;
1.196     dsl       990:                        wakeup_state = so->so_state;
1.1       cgd       991:                        goto restart;
                    992:                }
1.196     dsl       993:                wakeup_state = 0;
1.1       cgd       994:                mp = &top;
                    995:                space -= clen;
                    996:                do {
1.45      tv        997:                        if (uio == NULL) {
                    998:                                /*
                    999:                                 * Data is prepackaged in "top".
                   1000:                                 */
                   1001:                                resid = 0;
                   1002:                                if (flags & MSG_EOR)
                   1003:                                        top->m_flags |= M_EOR;
                   1004:                        } else do {
1.160     ad       1005:                                sounlock(so);
                   1006:                                splx(s);
1.144     dyoung   1007:                                if (top == NULL) {
1.78      matt     1008:                                        m = m_gethdr(M_WAIT, MT_DATA);
1.45      tv       1009:                                        mlen = MHLEN;
                   1010:                                        m->m_pkthdr.len = 0;
1.140     dyoung   1011:                                        m->m_pkthdr.rcvif = NULL;
1.45      tv       1012:                                } else {
1.78      matt     1013:                                        m = m_get(M_WAIT, MT_DATA);
1.45      tv       1014:                                        mlen = MLEN;
                   1015:                                }
1.78      matt     1016:                                MCLAIM(m, so->so_snd.sb_mowner);
1.121     yamt     1017:                                if (sock_loan_thresh >= 0 &&
                   1018:                                    uio->uio_iov->iov_len >= sock_loan_thresh &&
                   1019:                                    space >= sock_loan_thresh &&
1.64      thorpej  1020:                                    (len = sosend_loan(so, uio, m,
                   1021:                                                       space)) != 0) {
                   1022:                                        SOSEND_COUNTER_INCR(&sosend_loan_big);
                   1023:                                        space -= len;
                   1024:                                        goto have_data;
                   1025:                                }
1.45      tv       1026:                                if (resid >= MINCLSIZE && space >= MCLBYTES) {
1.64      thorpej  1027:                                        SOSEND_COUNTER_INCR(&sosend_copy_big);
1.78      matt     1028:                                        m_clget(m, M_WAIT);
1.45      tv       1029:                                        if ((m->m_flags & M_EXT) == 0)
                   1030:                                                goto nopages;
                   1031:                                        mlen = MCLBYTES;
                   1032:                                        if (atomic && top == 0) {
1.58      jdolecek 1033:                                                len = lmin(MCLBYTES - max_hdr,
1.54      lukem    1034:                                                    resid);
1.45      tv       1035:                                                m->m_data += max_hdr;
                   1036:                                        } else
1.58      jdolecek 1037:                                                len = lmin(MCLBYTES, resid);
1.45      tv       1038:                                        space -= len;
                   1039:                                } else {
1.64      thorpej  1040:  nopages:
                   1041:                                        SOSEND_COUNTER_INCR(&sosend_copy_small);
1.58      jdolecek 1042:                                        len = lmin(lmin(mlen, resid), space);
1.45      tv       1043:                                        space -= len;
                   1044:                                        /*
                   1045:                                         * For datagram protocols, leave room
                   1046:                                         * for protocol headers in first mbuf.
                   1047:                                         */
                   1048:                                        if (atomic && top == 0 && len < mlen)
                   1049:                                                MH_ALIGN(m, len);
                   1050:                                }
1.144     dyoung   1051:                                error = uiomove(mtod(m, void *), (int)len, uio);
1.64      thorpej  1052:  have_data:
1.45      tv       1053:                                resid = uio->uio_resid;
                   1054:                                m->m_len = len;
                   1055:                                *mp = m;
                   1056:                                top->m_pkthdr.len += len;
1.160     ad       1057:                                s = splsoftnet();
                   1058:                                solock(so);
1.144     dyoung   1059:                                if (error != 0)
1.45      tv       1060:                                        goto release;
                   1061:                                mp = &m->m_next;
                   1062:                                if (resid <= 0) {
                   1063:                                        if (flags & MSG_EOR)
                   1064:                                                top->m_flags |= M_EOR;
                   1065:                                        break;
                   1066:                                }
                   1067:                        } while (space > 0 && atomic);
1.108     perry    1068:
1.160     ad       1069:                        if (so->so_state & SS_CANTSENDMORE) {
                   1070:                                error = EPIPE;
                   1071:                                goto release;
                   1072:                        }
1.45      tv       1073:                        if (dontroute)
                   1074:                                so->so_options |= SO_DONTROUTE;
                   1075:                        if (resid > 0)
                   1076:                                so->so_state |= SS_MORETOCOME;
1.46      sommerfe 1077:                        error = (*so->so_proto->pr_usrreq)(so,
                   1078:                            (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
1.160     ad       1079:                            top, addr, control, curlwp);
1.45      tv       1080:                        if (dontroute)
                   1081:                                so->so_options &= ~SO_DONTROUTE;
                   1082:                        if (resid > 0)
                   1083:                                so->so_state &= ~SS_MORETOCOME;
                   1084:                        clen = 0;
1.144     dyoung   1085:                        control = NULL;
                   1086:                        top = NULL;
1.45      tv       1087:                        mp = &top;
1.144     dyoung   1088:                        if (error != 0)
1.1       cgd      1089:                                goto release;
                   1090:                } while (resid && space > 0);
                   1091:        } while (resid);
                   1092:
1.54      lukem    1093:  release:
1.1       cgd      1094:        sbunlock(&so->so_snd);
1.54      lukem    1095:  out:
1.160     ad       1096:        sounlock(so);
                   1097:        splx(s);
1.1       cgd      1098:        if (top)
                   1099:                m_freem(top);
                   1100:        if (control)
                   1101:                m_freem(control);
                   1102:        return (error);
                   1103: }
                   1104:
                   1105: /*
1.159     ad       1106:  * Following replacement or removal of the first mbuf on the first
                   1107:  * mbuf chain of a socket buffer, push necessary state changes back
                   1108:  * into the socket buffer so that other consumers see the values
                   1109:  * consistently.  'nextrecord' is the callers locally stored value of
                   1110:  * the original value of sb->sb_mb->m_nextpkt which must be restored
                   1111:  * when the lead mbuf changes.  NOTE: 'nextrecord' may be NULL.
                   1112:  */
                   1113: static void
                   1114: sbsync(struct sockbuf *sb, struct mbuf *nextrecord)
                   1115: {
                   1116:
1.160     ad       1117:        KASSERT(solocked(sb->sb_so));
                   1118:
1.159     ad       1119:        /*
                   1120:         * First, update for the new value of nextrecord.  If necessary,
                   1121:         * make it the first record.
                   1122:         */
                   1123:        if (sb->sb_mb != NULL)
                   1124:                sb->sb_mb->m_nextpkt = nextrecord;
                   1125:        else
                   1126:                sb->sb_mb = nextrecord;
                   1127:
                   1128:         /*
                   1129:          * Now update any dependent socket buffer fields to reflect
                   1130:          * the new state.  This is an inline of SB_EMPTY_FIXUP, with
                   1131:          * the addition of a second clause that takes care of the
                   1132:          * case where sb_mb has been updated, but remains the last
                   1133:          * record.
                   1134:          */
                   1135:         if (sb->sb_mb == NULL) {
                   1136:                 sb->sb_mbtail = NULL;
                   1137:                 sb->sb_lastrecord = NULL;
                   1138:         } else if (sb->sb_mb->m_nextpkt == NULL)
                   1139:                 sb->sb_lastrecord = sb->sb_mb;
                   1140: }
                   1141:
                   1142: /*
1.1       cgd      1143:  * Implement receive operations on a socket.
                   1144:  * We depend on the way that records are added to the sockbuf
                   1145:  * by sbappend*.  In particular, each record (mbufs linked through m_next)
                   1146:  * must begin with an address if the protocol so specifies,
                   1147:  * followed by an optional mbuf or mbufs containing ancillary data,
                   1148:  * and then zero or more mbufs of data.
                   1149:  * In order to avoid blocking network interrupts for the entire time here,
                   1150:  * we splx() while doing the actual copy to user space.
                   1151:  * Although the sockbuf is locked, new data may still be appended,
                   1152:  * and thus we must maintain consistency of the sockbuf during that time.
                   1153:  *
                   1154:  * The caller may receive the data as a single mbuf chain by supplying
                   1155:  * an mbuf **mp0 for use in returning the chain.  The uio is then used
                   1156:  * only for the count in uio_resid.
                   1157:  */
1.3       andrew   1158: int
1.54      lukem    1159: soreceive(struct socket *so, struct mbuf **paddr, struct uio *uio,
                   1160:        struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
1.1       cgd      1161: {
1.116     yamt     1162:        struct lwp *l = curlwp;
1.160     ad       1163:        struct mbuf     *m, **mp, *mt;
1.146     dyoung   1164:        int atomic, flags, len, error, s, offset, moff, type, orig_resid;
1.99      matt     1165:        const struct protosw    *pr;
1.54      lukem    1166:        struct mbuf     *nextrecord;
1.67      he       1167:        int             mbuf_removed = 0;
1.146     dyoung   1168:        const struct domain *dom;
1.196     dsl      1169:        short           wakeup_state = 0;
1.64      thorpej  1170:
1.54      lukem    1171:        pr = so->so_proto;
1.146     dyoung   1172:        atomic = pr->pr_flags & PR_ATOMIC;
                   1173:        dom = pr->pr_domain;
1.1       cgd      1174:        mp = mp0;
1.54      lukem    1175:        type = 0;
                   1176:        orig_resid = uio->uio_resid;
1.102     jonathan 1177:
1.144     dyoung   1178:        if (paddr != NULL)
                   1179:                *paddr = NULL;
                   1180:        if (controlp != NULL)
                   1181:                *controlp = NULL;
                   1182:        if (flagsp != NULL)
1.1       cgd      1183:                flags = *flagsp &~ MSG_EOR;
                   1184:        else
                   1185:                flags = 0;
1.66      enami    1186:
                   1187:        if ((flags & MSG_DONTWAIT) == 0)
1.117     yamt     1188:                sodopendfree();
1.66      enami    1189:
1.1       cgd      1190:        if (flags & MSG_OOB) {
                   1191:                m = m_get(M_WAIT, MT_DATA);
1.160     ad       1192:                solock(so);
1.17      cgd      1193:                error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m,
1.140     dyoung   1194:                    (struct mbuf *)(long)(flags & MSG_PEEK), NULL, l);
1.160     ad       1195:                sounlock(so);
1.1       cgd      1196:                if (error)
                   1197:                        goto bad;
                   1198:                do {
1.134     christos 1199:                        error = uiomove(mtod(m, void *),
1.1       cgd      1200:                            (int) min(uio->uio_resid, m->m_len), uio);
                   1201:                        m = m_free(m);
1.144     dyoung   1202:                } while (uio->uio_resid > 0 && error == 0 && m);
1.54      lukem    1203:  bad:
1.144     dyoung   1204:                if (m != NULL)
1.1       cgd      1205:                        m_freem(m);
1.144     dyoung   1206:                return error;
1.1       cgd      1207:        }
1.144     dyoung   1208:        if (mp != NULL)
1.140     dyoung   1209:                *mp = NULL;
1.160     ad       1210:
                   1211:        /*
                   1212:         * solock() provides atomicity of access.  splsoftnet() prevents
                   1213:         * protocol processing soft interrupts from interrupting us and
                   1214:         * blocking (expensive).
                   1215:         */
                   1216:        s = splsoftnet();
                   1217:        solock(so);
1.1       cgd      1218:        if (so->so_state & SS_ISCONFIRMING && uio->uio_resid)
1.140     dyoung   1219:                (*pr->pr_usrreq)(so, PRU_RCVD, NULL, NULL, NULL, l);
1.1       cgd      1220:
1.54      lukem    1221:  restart:
1.160     ad       1222:        if ((error = sblock(&so->so_rcv, SBLOCKWAIT(flags))) != 0) {
                   1223:                sounlock(so);
                   1224:                splx(s);
1.144     dyoung   1225:                return error;
1.160     ad       1226:        }
1.1       cgd      1227:
                   1228:        m = so->so_rcv.sb_mb;
                   1229:        /*
                   1230:         * If we have less data than requested, block awaiting more
                   1231:         * (subject to any timeout) if:
1.15      mycroft  1232:         *   1. the current count is less than the low water mark,
1.1       cgd      1233:         *   2. MSG_WAITALL is set, and it is possible to do the entire
1.15      mycroft  1234:         *      receive operation at once if we block (resid <= hiwat), or
                   1235:         *   3. MSG_DONTWAIT is not set.
1.1       cgd      1236:         * If MSG_WAITALL is set but resid is larger than the receive buffer,
                   1237:         * we have to do the receive in sections, and thus risk returning
                   1238:         * a short count if a timeout or signal occurs after we start.
                   1239:         */
1.144     dyoung   1240:        if (m == NULL ||
                   1241:            ((flags & MSG_DONTWAIT) == 0 &&
                   1242:             so->so_rcv.sb_cc < uio->uio_resid &&
                   1243:             (so->so_rcv.sb_cc < so->so_rcv.sb_lowat ||
                   1244:              ((flags & MSG_WAITALL) &&
                   1245:               uio->uio_resid <= so->so_rcv.sb_hiwat)) &&
1.146     dyoung   1246:             m->m_nextpkt == NULL && !atomic)) {
1.1       cgd      1247: #ifdef DIAGNOSTIC
1.144     dyoung   1248:                if (m == NULL && so->so_rcv.sb_cc)
1.1       cgd      1249:                        panic("receive 1");
                   1250: #endif
                   1251:                if (so->so_error) {
1.144     dyoung   1252:                        if (m != NULL)
1.15      mycroft  1253:                                goto dontblock;
1.1       cgd      1254:                        error = so->so_error;
                   1255:                        if ((flags & MSG_PEEK) == 0)
                   1256:                                so->so_error = 0;
                   1257:                        goto release;
                   1258:                }
                   1259:                if (so->so_state & SS_CANTRCVMORE) {
1.144     dyoung   1260:                        if (m != NULL)
1.15      mycroft  1261:                                goto dontblock;
1.1       cgd      1262:                        else
                   1263:                                goto release;
                   1264:                }
1.144     dyoung   1265:                for (; m != NULL; m = m->m_next)
1.1       cgd      1266:                        if (m->m_type == MT_OOBDATA  || (m->m_flags & M_EOR)) {
                   1267:                                m = so->so_rcv.sb_mb;
                   1268:                                goto dontblock;
                   1269:                        }
                   1270:                if ((so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) == 0 &&
                   1271:                    (so->so_proto->pr_flags & PR_CONNREQUIRED)) {
                   1272:                        error = ENOTCONN;
                   1273:                        goto release;
                   1274:                }
                   1275:                if (uio->uio_resid == 0)
                   1276:                        goto release;
1.151     ad       1277:                if (so->so_nbio || (flags & MSG_DONTWAIT)) {
1.1       cgd      1278:                        error = EWOULDBLOCK;
                   1279:                        goto release;
                   1280:                }
1.69      thorpej  1281:                SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1");
                   1282:                SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1");
1.1       cgd      1283:                sbunlock(&so->so_rcv);
1.196     dsl      1284:                if (wakeup_state & SS_RESTARTSYS)
                   1285:                        error = ERESTART;
                   1286:                else
                   1287:                        error = sbwait(&so->so_rcv);
1.160     ad       1288:                if (error != 0) {
                   1289:                        sounlock(so);
                   1290:                        splx(s);
1.144     dyoung   1291:                        return error;
1.160     ad       1292:                }
1.196     dsl      1293:                wakeup_state = so->so_state;
1.1       cgd      1294:                goto restart;
                   1295:        }
1.54      lukem    1296:  dontblock:
1.69      thorpej  1297:        /*
                   1298:         * On entry here, m points to the first record of the socket buffer.
1.159     ad       1299:         * From this point onward, we maintain 'nextrecord' as a cache of the
                   1300:         * pointer to the next record in the socket buffer.  We must keep the
                   1301:         * various socket buffer pointers and local stack versions of the
                   1302:         * pointers in sync, pushing out modifications before dropping the
1.160     ad       1303:         * socket lock, and re-reading them when picking it up.
1.159     ad       1304:         *
                   1305:         * Otherwise, we will race with the network stack appending new data
                   1306:         * or records onto the socket buffer by using inconsistent/stale
                   1307:         * versions of the field, possibly resulting in socket buffer
                   1308:         * corruption.
                   1309:         *
                   1310:         * By holding the high-level sblock(), we prevent simultaneous
                   1311:         * readers from pulling off the front of the socket buffer.
1.69      thorpej  1312:         */
1.144     dyoung   1313:        if (l != NULL)
1.157     ad       1314:                l->l_ru.ru_msgrcv++;
1.69      thorpej  1315:        KASSERT(m == so->so_rcv.sb_mb);
                   1316:        SBLASTRECORDCHK(&so->so_rcv, "soreceive 1");
                   1317:        SBLASTMBUFCHK(&so->so_rcv, "soreceive 1");
1.1       cgd      1318:        nextrecord = m->m_nextpkt;
                   1319:        if (pr->pr_flags & PR_ADDR) {
                   1320: #ifdef DIAGNOSTIC
                   1321:                if (m->m_type != MT_SONAME)
                   1322:                        panic("receive 1a");
                   1323: #endif
1.3       andrew   1324:                orig_resid = 0;
1.1       cgd      1325:                if (flags & MSG_PEEK) {
                   1326:                        if (paddr)
                   1327:                                *paddr = m_copy(m, 0, m->m_len);
                   1328:                        m = m->m_next;
                   1329:                } else {
                   1330:                        sbfree(&so->so_rcv, m);
1.67      he       1331:                        mbuf_removed = 1;
1.144     dyoung   1332:                        if (paddr != NULL) {
1.1       cgd      1333:                                *paddr = m;
                   1334:                                so->so_rcv.sb_mb = m->m_next;
1.144     dyoung   1335:                                m->m_next = NULL;
1.1       cgd      1336:                                m = so->so_rcv.sb_mb;
                   1337:                        } else {
                   1338:                                MFREE(m, so->so_rcv.sb_mb);
                   1339:                                m = so->so_rcv.sb_mb;
                   1340:                        }
1.159     ad       1341:                        sbsync(&so->so_rcv, nextrecord);
1.1       cgd      1342:                }
                   1343:        }
1.159     ad       1344:
                   1345:        /*
                   1346:         * Process one or more MT_CONTROL mbufs present before any data mbufs
                   1347:         * in the first mbuf chain on the socket buffer.  If MSG_PEEK, we
                   1348:         * just copy the data; if !MSG_PEEK, we call into the protocol to
                   1349:         * perform externalization (or freeing if controlp == NULL).
                   1350:         */
                   1351:        if (__predict_false(m != NULL && m->m_type == MT_CONTROL)) {
                   1352:                struct mbuf *cm = NULL, *cmn;
                   1353:                struct mbuf **cme = &cm;
                   1354:
                   1355:                do {
                   1356:                        if (flags & MSG_PEEK) {
                   1357:                                if (controlp != NULL) {
                   1358:                                        *controlp = m_copy(m, 0, m->m_len);
                   1359:                                        controlp = &(*controlp)->m_next;
                   1360:                                }
                   1361:                                m = m->m_next;
                   1362:                        } else {
                   1363:                                sbfree(&so->so_rcv, m);
1.1       cgd      1364:                                so->so_rcv.sb_mb = m->m_next;
1.144     dyoung   1365:                                m->m_next = NULL;
1.159     ad       1366:                                *cme = m;
                   1367:                                cme = &(*cme)->m_next;
1.1       cgd      1368:                                m = so->so_rcv.sb_mb;
1.159     ad       1369:                        }
                   1370:                } while (m != NULL && m->m_type == MT_CONTROL);
                   1371:                if ((flags & MSG_PEEK) == 0)
                   1372:                        sbsync(&so->so_rcv, nextrecord);
                   1373:                for (; cm != NULL; cm = cmn) {
                   1374:                        cmn = cm->m_next;
                   1375:                        cm->m_next = NULL;
                   1376:                        type = mtod(cm, struct cmsghdr *)->cmsg_type;
                   1377:                        if (controlp != NULL) {
                   1378:                                if (dom->dom_externalize != NULL &&
                   1379:                                    type == SCM_RIGHTS) {
1.160     ad       1380:                                        sounlock(so);
1.159     ad       1381:                                        splx(s);
                   1382:                                        error = (*dom->dom_externalize)(cm, l);
                   1383:                                        s = splsoftnet();
1.160     ad       1384:                                        solock(so);
1.159     ad       1385:                                }
                   1386:                                *controlp = cm;
                   1387:                                while (*controlp != NULL)
                   1388:                                        controlp = &(*controlp)->m_next;
1.1       cgd      1389:                        } else {
1.106     itojun   1390:                                /*
                   1391:                                 * Dispose of any SCM_RIGHTS message that went
                   1392:                                 * through the read path rather than recv.
                   1393:                                 */
1.159     ad       1394:                                if (dom->dom_dispose != NULL &&
                   1395:                                    type == SCM_RIGHTS) {
1.160     ad       1396:                                        sounlock(so);
1.159     ad       1397:                                        (*dom->dom_dispose)(cm);
1.160     ad       1398:                                        solock(so);
1.159     ad       1399:                                }
                   1400:                                m_freem(cm);
1.1       cgd      1401:                        }
                   1402:                }
1.159     ad       1403:                if (m != NULL)
                   1404:                        nextrecord = so->so_rcv.sb_mb->m_nextpkt;
                   1405:                else
                   1406:                        nextrecord = so->so_rcv.sb_mb;
                   1407:                orig_resid = 0;
1.1       cgd      1408:        }
1.69      thorpej  1409:
1.159     ad       1410:        /* If m is non-NULL, we have some data to read. */
                   1411:        if (__predict_true(m != NULL)) {
1.1       cgd      1412:                type = m->m_type;
                   1413:                if (type == MT_OOBDATA)
                   1414:                        flags |= MSG_OOB;
                   1415:        }
1.69      thorpej  1416:        SBLASTRECORDCHK(&so->so_rcv, "soreceive 2");
                   1417:        SBLASTMBUFCHK(&so->so_rcv, "soreceive 2");
                   1418:
1.1       cgd      1419:        moff = 0;
                   1420:        offset = 0;
1.144     dyoung   1421:        while (m != NULL && uio->uio_resid > 0 && error == 0) {
1.1       cgd      1422:                if (m->m_type == MT_OOBDATA) {
                   1423:                        if (type != MT_OOBDATA)
                   1424:                                break;
                   1425:                } else if (type == MT_OOBDATA)
                   1426:                        break;
                   1427: #ifdef DIAGNOSTIC
                   1428:                else if (m->m_type != MT_DATA && m->m_type != MT_HEADER)
                   1429:                        panic("receive 3");
                   1430: #endif
                   1431:                so->so_state &= ~SS_RCVATMARK;
1.196     dsl      1432:                wakeup_state = 0;
1.1       cgd      1433:                len = uio->uio_resid;
                   1434:                if (so->so_oobmark && len > so->so_oobmark - offset)
                   1435:                        len = so->so_oobmark - offset;
                   1436:                if (len > m->m_len - moff)
                   1437:                        len = m->m_len - moff;
                   1438:                /*
                   1439:                 * If mp is set, just pass back the mbufs.
                   1440:                 * Otherwise copy them out via the uio, then free.
                   1441:                 * Sockbuf must be consistent here (points to current mbuf,
                   1442:                 * it points to next record) when we drop priority;
                   1443:                 * we must note any additions to the sockbuf when we
                   1444:                 * block interrupts again.
                   1445:                 */
1.144     dyoung   1446:                if (mp == NULL) {
1.69      thorpej  1447:                        SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove");
                   1448:                        SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove");
1.160     ad       1449:                        sounlock(so);
1.1       cgd      1450:                        splx(s);
1.134     christos 1451:                        error = uiomove(mtod(m, char *) + moff, (int)len, uio);
1.20      mycroft  1452:                        s = splsoftnet();
1.160     ad       1453:                        solock(so);
1.144     dyoung   1454:                        if (error != 0) {
1.67      he       1455:                                /*
                   1456:                                 * If any part of the record has been removed
                   1457:                                 * (such as the MT_SONAME mbuf, which will
                   1458:                                 * happen when PR_ADDR, and thus also
                   1459:                                 * PR_ATOMIC, is set), then drop the entire
                   1460:                                 * record to maintain the atomicity of the
                   1461:                                 * receive operation.
                   1462:                                 *
                   1463:                                 * This avoids a later panic("receive 1a")
                   1464:                                 * when compiled with DIAGNOSTIC.
                   1465:                                 */
1.146     dyoung   1466:                                if (m && mbuf_removed && atomic)
1.67      he       1467:                                        (void) sbdroprecord(&so->so_rcv);
                   1468:
1.57      jdolecek 1469:                                goto release;
1.67      he       1470:                        }
1.1       cgd      1471:                } else
                   1472:                        uio->uio_resid -= len;
                   1473:                if (len == m->m_len - moff) {
                   1474:                        if (m->m_flags & M_EOR)
                   1475:                                flags |= MSG_EOR;
                   1476:                        if (flags & MSG_PEEK) {
                   1477:                                m = m->m_next;
                   1478:                                moff = 0;
                   1479:                        } else {
                   1480:                                nextrecord = m->m_nextpkt;
                   1481:                                sbfree(&so->so_rcv, m);
                   1482:                                if (mp) {
                   1483:                                        *mp = m;
                   1484:                                        mp = &m->m_next;
                   1485:                                        so->so_rcv.sb_mb = m = m->m_next;
1.140     dyoung   1486:                                        *mp = NULL;
1.1       cgd      1487:                                } else {
                   1488:                                        MFREE(m, so->so_rcv.sb_mb);
                   1489:                                        m = so->so_rcv.sb_mb;
                   1490:                                }
1.69      thorpej  1491:                                /*
                   1492:                                 * If m != NULL, we also know that
                   1493:                                 * so->so_rcv.sb_mb != NULL.
                   1494:                                 */
                   1495:                                KASSERT(so->so_rcv.sb_mb == m);
                   1496:                                if (m) {
1.1       cgd      1497:                                        m->m_nextpkt = nextrecord;
1.69      thorpej  1498:                                        if (nextrecord == NULL)
                   1499:                                                so->so_rcv.sb_lastrecord = m;
                   1500:                                } else {
                   1501:                                        so->so_rcv.sb_mb = nextrecord;
1.70      thorpej  1502:                                        SB_EMPTY_FIXUP(&so->so_rcv);
1.69      thorpej  1503:                                }
                   1504:                                SBLASTRECORDCHK(&so->so_rcv, "soreceive 3");
                   1505:                                SBLASTMBUFCHK(&so->so_rcv, "soreceive 3");
1.1       cgd      1506:                        }
1.144     dyoung   1507:                } else if (flags & MSG_PEEK)
                   1508:                        moff += len;
                   1509:                else {
1.160     ad       1510:                        if (mp != NULL) {
                   1511:                                mt = m_copym(m, 0, len, M_NOWAIT);
                   1512:                                if (__predict_false(mt == NULL)) {
                   1513:                                        sounlock(so);
                   1514:                                        mt = m_copym(m, 0, len, M_WAIT);
                   1515:                                        solock(so);
                   1516:                                }
                   1517:                                *mp = mt;
                   1518:                        }
1.144     dyoung   1519:                        m->m_data += len;
                   1520:                        m->m_len -= len;
                   1521:                        so->so_rcv.sb_cc -= len;
1.1       cgd      1522:                }
                   1523:                if (so->so_oobmark) {
                   1524:                        if ((flags & MSG_PEEK) == 0) {
                   1525:                                so->so_oobmark -= len;
                   1526:                                if (so->so_oobmark == 0) {
                   1527:                                        so->so_state |= SS_RCVATMARK;
                   1528:                                        break;
                   1529:                                }
1.7       cgd      1530:                        } else {
1.1       cgd      1531:                                offset += len;
1.7       cgd      1532:                                if (offset == so->so_oobmark)
                   1533:                                        break;
                   1534:                        }
1.1       cgd      1535:                }
                   1536:                if (flags & MSG_EOR)
                   1537:                        break;
                   1538:                /*
                   1539:                 * If the MSG_WAITALL flag is set (for non-atomic socket),
                   1540:                 * we must not quit until "uio->uio_resid == 0" or an error
                   1541:                 * termination.  If a signal/timeout occurs, return
                   1542:                 * with a short count but without error.
                   1543:                 * Keep sockbuf locked against other readers.
                   1544:                 */
1.144     dyoung   1545:                while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 &&
1.3       andrew   1546:                    !sosendallatonce(so) && !nextrecord) {
1.1       cgd      1547:                        if (so->so_error || so->so_state & SS_CANTRCVMORE)
                   1548:                                break;
1.68      matt     1549:                        /*
                   1550:                         * If we are peeking and the socket receive buffer is
                   1551:                         * full, stop since we can't get more data to peek at.
                   1552:                         */
                   1553:                        if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0)
                   1554:                                break;
                   1555:                        /*
                   1556:                         * If we've drained the socket buffer, tell the
                   1557:                         * protocol in case it needs to do something to
                   1558:                         * get it filled again.
                   1559:                         */
                   1560:                        if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb)
                   1561:                                (*pr->pr_usrreq)(so, PRU_RCVD,
1.140     dyoung   1562:                                    NULL, (struct mbuf *)(long)flags, NULL, l);
1.69      thorpej  1563:                        SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2");
                   1564:                        SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2");
1.196     dsl      1565:                        if (wakeup_state & SS_RESTARTSYS)
                   1566:                                error = ERESTART;
                   1567:                        else
                   1568:                                error = sbwait(&so->so_rcv);
1.144     dyoung   1569:                        if (error != 0) {
1.1       cgd      1570:                                sbunlock(&so->so_rcv);
1.160     ad       1571:                                sounlock(so);
1.1       cgd      1572:                                splx(s);
1.144     dyoung   1573:                                return 0;
1.1       cgd      1574:                        }
1.21      christos 1575:                        if ((m = so->so_rcv.sb_mb) != NULL)
1.1       cgd      1576:                                nextrecord = m->m_nextpkt;
1.196     dsl      1577:                        wakeup_state = so->so_state;
1.1       cgd      1578:                }
                   1579:        }
1.3       andrew   1580:
1.146     dyoung   1581:        if (m && atomic) {
1.3       andrew   1582:                flags |= MSG_TRUNC;
                   1583:                if ((flags & MSG_PEEK) == 0)
                   1584:                        (void) sbdroprecord(&so->so_rcv);
                   1585:        }
1.1       cgd      1586:        if ((flags & MSG_PEEK) == 0) {
1.144     dyoung   1587:                if (m == NULL) {
1.69      thorpej  1588:                        /*
1.70      thorpej  1589:                         * First part is an inline SB_EMPTY_FIXUP().  Second
1.69      thorpej  1590:                         * part makes sure sb_lastrecord is up-to-date if
                   1591:                         * there is still data in the socket buffer.
                   1592:                         */
1.1       cgd      1593:                        so->so_rcv.sb_mb = nextrecord;
1.69      thorpej  1594:                        if (so->so_rcv.sb_mb == NULL) {
                   1595:                                so->so_rcv.sb_mbtail = NULL;
                   1596:                                so->so_rcv.sb_lastrecord = NULL;
                   1597:                        } else if (nextrecord->m_nextpkt == NULL)
                   1598:                                so->so_rcv.sb_lastrecord = nextrecord;
                   1599:                }
                   1600:                SBLASTRECORDCHK(&so->so_rcv, "soreceive 4");
                   1601:                SBLASTMBUFCHK(&so->so_rcv, "soreceive 4");
1.1       cgd      1602:                if (pr->pr_flags & PR_WANTRCVD && so->so_pcb)
1.140     dyoung   1603:                        (*pr->pr_usrreq)(so, PRU_RCVD, NULL,
                   1604:                            (struct mbuf *)(long)flags, NULL, l);
1.1       cgd      1605:        }
1.3       andrew   1606:        if (orig_resid == uio->uio_resid && orig_resid &&
                   1607:            (flags & MSG_EOR) == 0 && (so->so_state & SS_CANTRCVMORE) == 0) {
                   1608:                sbunlock(&so->so_rcv);
                   1609:                goto restart;
                   1610:        }
1.108     perry    1611:
1.144     dyoung   1612:        if (flagsp != NULL)
1.1       cgd      1613:                *flagsp |= flags;
1.54      lukem    1614:  release:
1.1       cgd      1615:        sbunlock(&so->so_rcv);
1.160     ad       1616:        sounlock(so);
1.1       cgd      1617:        splx(s);
1.144     dyoung   1618:        return error;
1.1       cgd      1619: }
                   1620:
1.14      mycroft  1621: int
1.54      lukem    1622: soshutdown(struct socket *so, int how)
1.1       cgd      1623: {
1.99      matt     1624:        const struct protosw    *pr;
1.160     ad       1625:        int     error;
                   1626:
                   1627:        KASSERT(solocked(so));
1.34      kleink   1628:
1.54      lukem    1629:        pr = so->so_proto;
1.34      kleink   1630:        if (!(how == SHUT_RD || how == SHUT_WR || how == SHUT_RDWR))
                   1631:                return (EINVAL);
1.1       cgd      1632:
1.160     ad       1633:        if (how == SHUT_RD || how == SHUT_RDWR) {
1.1       cgd      1634:                sorflush(so);
1.160     ad       1635:                error = 0;
                   1636:        }
1.34      kleink   1637:        if (how == SHUT_WR || how == SHUT_RDWR)
1.160     ad       1638:                error = (*pr->pr_usrreq)(so, PRU_SHUTDOWN, NULL,
1.140     dyoung   1639:                    NULL, NULL, NULL);
1.160     ad       1640:
                   1641:        return error;
1.1       cgd      1642: }
                   1643:
1.195     dsl      1644: void
1.196     dsl      1645: sorestart(struct socket *so)
1.188     ad       1646: {
1.196     dsl      1647:        /*
                   1648:         * An application has called close() on an fd on which another
                   1649:         * of its threads has called a socket system call.
                   1650:         * Mark this and wake everyone up, and code that would block again
                   1651:         * instead returns ERESTART.
                   1652:         * On system call re-entry the fd is validated and EBADF returned.
                   1653:         * Any other fd will block again on the 2nd syscall.
                   1654:         */
1.188     ad       1655:        solock(so);
1.196     dsl      1656:        so->so_state |= SS_RESTARTSYS;
1.188     ad       1657:        cv_broadcast(&so->so_cv);
1.196     dsl      1658:        cv_broadcast(&so->so_snd.sb_cv);
                   1659:        cv_broadcast(&so->so_rcv.sb_cv);
1.188     ad       1660:        sounlock(so);
                   1661: }
                   1662:
1.14      mycroft  1663: void
1.54      lukem    1664: sorflush(struct socket *so)
1.1       cgd      1665: {
1.54      lukem    1666:        struct sockbuf  *sb, asb;
1.99      matt     1667:        const struct protosw    *pr;
1.160     ad       1668:
                   1669:        KASSERT(solocked(so));
1.1       cgd      1670:
1.54      lukem    1671:        sb = &so->so_rcv;
                   1672:        pr = so->so_proto;
1.160     ad       1673:        socantrcvmore(so);
1.1       cgd      1674:        sb->sb_flags |= SB_NOINTR;
1.160     ad       1675:        (void )sblock(sb, M_WAITOK);
1.1       cgd      1676:        sbunlock(sb);
                   1677:        asb = *sb;
1.86      wrstuden 1678:        /*
                   1679:         * Clear most of the sockbuf structure, but leave some of the
                   1680:         * fields valid.
                   1681:         */
                   1682:        memset(&sb->sb_startzero, 0,
                   1683:            sizeof(*sb) - offsetof(struct sockbuf, sb_startzero));
1.160     ad       1684:        if (pr->pr_flags & PR_RIGHTS && pr->pr_domain->dom_dispose) {
                   1685:                sounlock(so);
1.1       cgd      1686:                (*pr->pr_domain->dom_dispose)(asb.sb_mb);
1.160     ad       1687:                solock(so);
                   1688:        }
1.98      christos 1689:        sbrelease(&asb, so);
1.1       cgd      1690: }
                   1691:
1.171     plunky   1692: /*
                   1693:  * internal set SOL_SOCKET options
                   1694:  */
1.142     dyoung   1695: static int
1.171     plunky   1696: sosetopt1(struct socket *so, const struct sockopt *sopt)
1.1       cgd      1697: {
1.182     christos 1698:        int error = EINVAL, optval, opt;
1.171     plunky   1699:        struct linger l;
                   1700:        struct timeval tv;
1.142     dyoung   1701:
1.179     christos 1702:        switch ((opt = sopt->sopt_name)) {
1.142     dyoung   1703:
1.170     tls      1704:        case SO_ACCEPTFILTER:
1.177     ad       1705:                error = accept_filt_setopt(so, sopt);
                   1706:                KASSERT(solocked(so));
1.170     tls      1707:                break;
                   1708:
1.171     plunky   1709:        case SO_LINGER:
                   1710:                error = sockopt_get(sopt, &l, sizeof(l));
1.177     ad       1711:                solock(so);
1.171     plunky   1712:                if (error)
1.177     ad       1713:                        break;
1.171     plunky   1714:                if (l.l_linger < 0 || l.l_linger > USHRT_MAX ||
1.177     ad       1715:                    l.l_linger > (INT_MAX / hz)) {
                   1716:                        error = EDOM;
                   1717:                        break;
                   1718:                }
1.171     plunky   1719:                so->so_linger = l.l_linger;
                   1720:                if (l.l_onoff)
                   1721:                        so->so_options |= SO_LINGER;
                   1722:                else
                   1723:                        so->so_options &= ~SO_LINGER;
1.177     ad       1724:                break;
1.1       cgd      1725:
1.142     dyoung   1726:        case SO_DEBUG:
                   1727:        case SO_KEEPALIVE:
                   1728:        case SO_DONTROUTE:
                   1729:        case SO_USELOOPBACK:
                   1730:        case SO_BROADCAST:
                   1731:        case SO_REUSEADDR:
                   1732:        case SO_REUSEPORT:
                   1733:        case SO_OOBINLINE:
                   1734:        case SO_TIMESTAMP:
1.184     christos 1735: #ifdef SO_OTIMESTAMP
                   1736:        case SO_OTIMESTAMP:
                   1737: #endif
1.171     plunky   1738:                error = sockopt_getint(sopt, &optval);
1.177     ad       1739:                solock(so);
1.171     plunky   1740:                if (error)
1.177     ad       1741:                        break;
1.171     plunky   1742:                if (optval)
1.179     christos 1743:                        so->so_options |= opt;
1.142     dyoung   1744:                else
1.179     christos 1745:                        so->so_options &= ~opt;
1.142     dyoung   1746:                break;
                   1747:
                   1748:        case SO_SNDBUF:
                   1749:        case SO_RCVBUF:
                   1750:        case SO_SNDLOWAT:
                   1751:        case SO_RCVLOWAT:
1.171     plunky   1752:                error = sockopt_getint(sopt, &optval);
1.177     ad       1753:                solock(so);
1.171     plunky   1754:                if (error)
1.177     ad       1755:                        break;
1.1       cgd      1756:
1.142     dyoung   1757:                /*
                   1758:                 * Values < 1 make no sense for any of these
                   1759:                 * options, so disallow them.
                   1760:                 */
1.177     ad       1761:                if (optval < 1) {
                   1762:                        error = EINVAL;
                   1763:                        break;
                   1764:                }
1.1       cgd      1765:
1.179     christos 1766:                switch (opt) {
1.171     plunky   1767:                case SO_SNDBUF:
1.177     ad       1768:                        if (sbreserve(&so->so_snd, (u_long)optval, so) == 0) {
                   1769:                                error = ENOBUFS;
                   1770:                                break;
                   1771:                        }
1.171     plunky   1772:                        so->so_snd.sb_flags &= ~SB_AUTOSIZE;
                   1773:                        break;
1.1       cgd      1774:
                   1775:                case SO_RCVBUF:
1.177     ad       1776:                        if (sbreserve(&so->so_rcv, (u_long)optval, so) == 0) {
                   1777:                                error = ENOBUFS;
                   1778:                                break;
                   1779:                        }
1.171     plunky   1780:                        so->so_rcv.sb_flags &= ~SB_AUTOSIZE;
1.142     dyoung   1781:                        break;
                   1782:
                   1783:                /*
                   1784:                 * Make sure the low-water is never greater than
                   1785:                 * the high-water.
                   1786:                 */
1.1       cgd      1787:                case SO_SNDLOWAT:
1.171     plunky   1788:                        if (optval > so->so_snd.sb_hiwat)
                   1789:                                optval = so->so_snd.sb_hiwat;
                   1790:
                   1791:                        so->so_snd.sb_lowat = optval;
1.142     dyoung   1792:                        break;
1.171     plunky   1793:
1.1       cgd      1794:                case SO_RCVLOWAT:
1.171     plunky   1795:                        if (optval > so->so_rcv.sb_hiwat)
                   1796:                                optval = so->so_rcv.sb_hiwat;
                   1797:
                   1798:                        so->so_rcv.sb_lowat = optval;
1.142     dyoung   1799:                        break;
                   1800:                }
                   1801:                break;
1.28      thorpej  1802:
1.179     christos 1803: #ifdef COMPAT_50
                   1804:        case SO_OSNDTIMEO:
                   1805:        case SO_ORCVTIMEO: {
                   1806:                struct timeval50 otv;
                   1807:                error = sockopt_get(sopt, &otv, sizeof(otv));
1.186     pooka    1808:                if (error) {
                   1809:                        solock(so);
1.183     christos 1810:                        break;
1.186     pooka    1811:                }
1.179     christos 1812:                timeval50_to_timeval(&otv, &tv);
                   1813:                opt = opt == SO_OSNDTIMEO ? SO_SNDTIMEO : SO_RCVTIMEO;
1.182     christos 1814:                error = 0;
1.179     christos 1815:                /*FALLTHROUGH*/
                   1816:        }
                   1817: #endif /* COMPAT_50 */
                   1818:
1.142     dyoung   1819:        case SO_SNDTIMEO:
                   1820:        case SO_RCVTIMEO:
1.182     christos 1821:                if (error)
1.179     christos 1822:                        error = sockopt_get(sopt, &tv, sizeof(tv));
1.177     ad       1823:                solock(so);
1.171     plunky   1824:                if (error)
1.177     ad       1825:                        break;
1.171     plunky   1826:
1.177     ad       1827:                if (tv.tv_sec > (INT_MAX - tv.tv_usec / tick) / hz) {
                   1828:                        error = EDOM;
                   1829:                        break;
                   1830:                }
1.28      thorpej  1831:
1.171     plunky   1832:                optval = tv.tv_sec * hz + tv.tv_usec / tick;
                   1833:                if (optval == 0 && tv.tv_usec != 0)
                   1834:                        optval = 1;
1.28      thorpej  1835:
1.179     christos 1836:                switch (opt) {
1.142     dyoung   1837:                case SO_SNDTIMEO:
1.171     plunky   1838:                        so->so_snd.sb_timeo = optval;
1.1       cgd      1839:                        break;
                   1840:                case SO_RCVTIMEO:
1.171     plunky   1841:                        so->so_rcv.sb_timeo = optval;
1.142     dyoung   1842:                        break;
                   1843:                }
                   1844:                break;
1.1       cgd      1845:
1.142     dyoung   1846:        default:
1.177     ad       1847:                solock(so);
                   1848:                error = ENOPROTOOPT;
                   1849:                break;
1.142     dyoung   1850:        }
1.177     ad       1851:        KASSERT(solocked(so));
                   1852:        return error;
1.142     dyoung   1853: }
1.1       cgd      1854:
1.142     dyoung   1855: int
1.171     plunky   1856: sosetopt(struct socket *so, struct sockopt *sopt)
1.142     dyoung   1857: {
                   1858:        int error, prerr;
1.1       cgd      1859:
1.177     ad       1860:        if (sopt->sopt_level == SOL_SOCKET) {
1.171     plunky   1861:                error = sosetopt1(so, sopt);
1.177     ad       1862:                KASSERT(solocked(so));
                   1863:        } else {
1.142     dyoung   1864:                error = ENOPROTOOPT;
1.177     ad       1865:                solock(so);
                   1866:        }
1.1       cgd      1867:
1.142     dyoung   1868:        if ((error == 0 || error == ENOPROTOOPT) &&
                   1869:            so->so_proto != NULL && so->so_proto->pr_ctloutput != NULL) {
                   1870:                /* give the protocol stack a shot */
1.171     plunky   1871:                prerr = (*so->so_proto->pr_ctloutput)(PRCO_SETOPT, so, sopt);
1.142     dyoung   1872:                if (prerr == 0)
                   1873:                        error = 0;
                   1874:                else if (prerr != ENOPROTOOPT)
                   1875:                        error = prerr;
1.171     plunky   1876:        }
1.160     ad       1877:        sounlock(so);
1.142     dyoung   1878:        return error;
1.1       cgd      1879: }
                   1880:
1.171     plunky   1881: /*
                   1882:  * so_setsockopt() is a wrapper providing a sockopt structure for sosetopt()
                   1883:  */
                   1884: int
                   1885: so_setsockopt(struct lwp *l, struct socket *so, int level, int name,
                   1886:     const void *val, size_t valsize)
                   1887: {
                   1888:        struct sockopt sopt;
                   1889:        int error;
                   1890:
                   1891:        KASSERT(valsize == 0 || val != NULL);
                   1892:
                   1893:        sockopt_init(&sopt, level, name, valsize);
                   1894:        sockopt_set(&sopt, val, valsize);
                   1895:
                   1896:        error = sosetopt(so, &sopt);
                   1897:
                   1898:        sockopt_destroy(&sopt);
                   1899:
                   1900:        return error;
                   1901: }
                   1902:
                   1903: /*
                   1904:  * internal get SOL_SOCKET options
                   1905:  */
                   1906: static int
                   1907: sogetopt1(struct socket *so, struct sockopt *sopt)
                   1908: {
1.179     christos 1909:        int error, optval, opt;
1.171     plunky   1910:        struct linger l;
                   1911:        struct timeval tv;
                   1912:
1.179     christos 1913:        switch ((opt = sopt->sopt_name)) {
1.171     plunky   1914:
                   1915:        case SO_ACCEPTFILTER:
1.177     ad       1916:                error = accept_filt_getopt(so, sopt);
1.171     plunky   1917:                break;
                   1918:
                   1919:        case SO_LINGER:
                   1920:                l.l_onoff = (so->so_options & SO_LINGER) ? 1 : 0;
                   1921:                l.l_linger = so->so_linger;
                   1922:
                   1923:                error = sockopt_set(sopt, &l, sizeof(l));
                   1924:                break;
                   1925:
                   1926:        case SO_USELOOPBACK:
                   1927:        case SO_DONTROUTE:
                   1928:        case SO_DEBUG:
                   1929:        case SO_KEEPALIVE:
                   1930:        case SO_REUSEADDR:
                   1931:        case SO_REUSEPORT:
                   1932:        case SO_BROADCAST:
                   1933:        case SO_OOBINLINE:
                   1934:        case SO_TIMESTAMP:
1.184     christos 1935: #ifdef SO_OTIMESTAMP
                   1936:        case SO_OTIMESTAMP:
                   1937: #endif
1.179     christos 1938:                error = sockopt_setint(sopt, (so->so_options & opt) ? 1 : 0);
1.171     plunky   1939:                break;
                   1940:
                   1941:        case SO_TYPE:
                   1942:                error = sockopt_setint(sopt, so->so_type);
                   1943:                break;
                   1944:
                   1945:        case SO_ERROR:
                   1946:                error = sockopt_setint(sopt, so->so_error);
                   1947:                so->so_error = 0;
                   1948:                break;
                   1949:
                   1950:        case SO_SNDBUF:
                   1951:                error = sockopt_setint(sopt, so->so_snd.sb_hiwat);
                   1952:                break;
                   1953:
                   1954:        case SO_RCVBUF:
                   1955:                error = sockopt_setint(sopt, so->so_rcv.sb_hiwat);
                   1956:                break;
                   1957:
                   1958:        case SO_SNDLOWAT:
                   1959:                error = sockopt_setint(sopt, so->so_snd.sb_lowat);
                   1960:                break;
                   1961:
                   1962:        case SO_RCVLOWAT:
                   1963:                error = sockopt_setint(sopt, so->so_rcv.sb_lowat);
                   1964:                break;
                   1965:
1.179     christos 1966: #ifdef COMPAT_50
                   1967:        case SO_OSNDTIMEO:
                   1968:        case SO_ORCVTIMEO: {
                   1969:                struct timeval50 otv;
                   1970:
                   1971:                optval = (opt == SO_OSNDTIMEO ?
                   1972:                     so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
                   1973:
                   1974:                otv.tv_sec = optval / hz;
                   1975:                otv.tv_usec = (optval % hz) * tick;
                   1976:
                   1977:                error = sockopt_set(sopt, &otv, sizeof(otv));
                   1978:                break;
                   1979:        }
                   1980: #endif /* COMPAT_50 */
                   1981:
1.171     plunky   1982:        case SO_SNDTIMEO:
                   1983:        case SO_RCVTIMEO:
1.179     christos 1984:                optval = (opt == SO_SNDTIMEO ?
1.171     plunky   1985:                     so->so_snd.sb_timeo : so->so_rcv.sb_timeo);
                   1986:
                   1987:                tv.tv_sec = optval / hz;
                   1988:                tv.tv_usec = (optval % hz) * tick;
                   1989:
                   1990:                error = sockopt_set(sopt, &tv, sizeof(tv));
                   1991:                break;
                   1992:
                   1993:        case SO_OVERFLOWED:
                   1994:                error = sockopt_setint(sopt, so->so_rcv.sb_overflowed);
                   1995:                break;
                   1996:
                   1997:        default:
                   1998:                error = ENOPROTOOPT;
                   1999:                break;
                   2000:        }
                   2001:
                   2002:        return (error);
                   2003: }
                   2004:
1.14      mycroft  2005: int
1.171     plunky   2006: sogetopt(struct socket *so, struct sockopt *sopt)
1.1       cgd      2007: {
1.160     ad       2008:        int             error;
1.1       cgd      2009:
1.160     ad       2010:        solock(so);
1.171     plunky   2011:        if (sopt->sopt_level != SOL_SOCKET) {
1.1       cgd      2012:                if (so->so_proto && so->so_proto->pr_ctloutput) {
1.160     ad       2013:                        error = ((*so->so_proto->pr_ctloutput)
1.171     plunky   2014:                            (PRCO_GETOPT, so, sopt));
1.1       cgd      2015:                } else
1.160     ad       2016:                        error = (ENOPROTOOPT);
1.1       cgd      2017:        } else {
1.171     plunky   2018:                error = sogetopt1(so, sopt);
                   2019:        }
                   2020:        sounlock(so);
                   2021:        return (error);
                   2022: }
                   2023:
                   2024: /*
                   2025:  * alloc sockopt data buffer buffer
                   2026:  *     - will be released at destroy
                   2027:  */
1.176     plunky   2028: static int
                   2029: sockopt_alloc(struct sockopt *sopt, size_t len, km_flag_t kmflag)
1.171     plunky   2030: {
                   2031:
                   2032:        KASSERT(sopt->sopt_size == 0);
                   2033:
1.176     plunky   2034:        if (len > sizeof(sopt->sopt_buf)) {
                   2035:                sopt->sopt_data = kmem_zalloc(len, kmflag);
                   2036:                if (sopt->sopt_data == NULL)
                   2037:                        return ENOMEM;
                   2038:        } else
1.171     plunky   2039:                sopt->sopt_data = sopt->sopt_buf;
                   2040:
                   2041:        sopt->sopt_size = len;
1.176     plunky   2042:        return 0;
1.171     plunky   2043: }
                   2044:
                   2045: /*
                   2046:  * initialise sockopt storage
1.176     plunky   2047:  *     - MAY sleep during allocation
1.171     plunky   2048:  */
                   2049: void
                   2050: sockopt_init(struct sockopt *sopt, int level, int name, size_t size)
                   2051: {
1.1       cgd      2052:
1.171     plunky   2053:        memset(sopt, 0, sizeof(*sopt));
1.1       cgd      2054:
1.171     plunky   2055:        sopt->sopt_level = level;
                   2056:        sopt->sopt_name = name;
1.176     plunky   2057:        (void)sockopt_alloc(sopt, size, KM_SLEEP);
1.171     plunky   2058: }
                   2059:
                   2060: /*
                   2061:  * destroy sockopt storage
                   2062:  *     - will release any held memory references
                   2063:  */
                   2064: void
                   2065: sockopt_destroy(struct sockopt *sopt)
                   2066: {
                   2067:
                   2068:        if (sopt->sopt_data != sopt->sopt_buf)
1.173     plunky   2069:                kmem_free(sopt->sopt_data, sopt->sopt_size);
1.171     plunky   2070:
                   2071:        memset(sopt, 0, sizeof(*sopt));
                   2072: }
                   2073:
                   2074: /*
                   2075:  * set sockopt value
                   2076:  *     - value is copied into sockopt
1.176     plunky   2077:  *     - memory is allocated when necessary, will not sleep
1.171     plunky   2078:  */
                   2079: int
                   2080: sockopt_set(struct sockopt *sopt, const void *buf, size_t len)
                   2081: {
1.176     plunky   2082:        int error;
1.171     plunky   2083:
1.176     plunky   2084:        if (sopt->sopt_size == 0) {
                   2085:                error = sockopt_alloc(sopt, len, KM_NOSLEEP);
                   2086:                if (error)
                   2087:                        return error;
                   2088:        }
1.171     plunky   2089:
                   2090:        KASSERT(sopt->sopt_size == len);
                   2091:        memcpy(sopt->sopt_data, buf, len);
                   2092:        return 0;
                   2093: }
                   2094:
                   2095: /*
                   2096:  * common case of set sockopt integer value
                   2097:  */
                   2098: int
                   2099: sockopt_setint(struct sockopt *sopt, int val)
                   2100: {
                   2101:
                   2102:        return sockopt_set(sopt, &val, sizeof(int));
                   2103: }
                   2104:
                   2105: /*
                   2106:  * get sockopt value
                   2107:  *     - correct size must be given
                   2108:  */
                   2109: int
                   2110: sockopt_get(const struct sockopt *sopt, void *buf, size_t len)
                   2111: {
1.170     tls      2112:
1.171     plunky   2113:        if (sopt->sopt_size != len)
                   2114:                return EINVAL;
1.1       cgd      2115:
1.171     plunky   2116:        memcpy(buf, sopt->sopt_data, len);
                   2117:        return 0;
                   2118: }
1.1       cgd      2119:
1.171     plunky   2120: /*
                   2121:  * common case of get sockopt integer value
                   2122:  */
                   2123: int
                   2124: sockopt_getint(const struct sockopt *sopt, int *valp)
                   2125: {
1.1       cgd      2126:
1.171     plunky   2127:        return sockopt_get(sopt, valp, sizeof(int));
                   2128: }
1.1       cgd      2129:
1.171     plunky   2130: /*
                   2131:  * set sockopt value from mbuf
                   2132:  *     - ONLY for legacy code
                   2133:  *     - mbuf is released by sockopt
1.176     plunky   2134:  *     - will not sleep
1.171     plunky   2135:  */
                   2136: int
                   2137: sockopt_setmbuf(struct sockopt *sopt, struct mbuf *m)
                   2138: {
                   2139:        size_t len;
1.176     plunky   2140:        int error;
1.1       cgd      2141:
1.171     plunky   2142:        len = m_length(m);
1.1       cgd      2143:
1.176     plunky   2144:        if (sopt->sopt_size == 0) {
                   2145:                error = sockopt_alloc(sopt, len, KM_NOSLEEP);
                   2146:                if (error)
                   2147:                        return error;
                   2148:        }
1.1       cgd      2149:
1.171     plunky   2150:        KASSERT(sopt->sopt_size == len);
                   2151:        m_copydata(m, 0, len, sopt->sopt_data);
                   2152:        m_freem(m);
1.1       cgd      2153:
1.171     plunky   2154:        return 0;
                   2155: }
1.1       cgd      2156:
1.171     plunky   2157: /*
                   2158:  * get sockopt value into mbuf
                   2159:  *     - ONLY for legacy code
                   2160:  *     - mbuf to be released by the caller
1.176     plunky   2161:  *     - will not sleep
1.171     plunky   2162:  */
                   2163: struct mbuf *
                   2164: sockopt_getmbuf(const struct sockopt *sopt)
                   2165: {
                   2166:        struct mbuf *m;
1.107     darrenr  2167:
1.176     plunky   2168:        if (sopt->sopt_size > MCLBYTES)
                   2169:                return NULL;
                   2170:
                   2171:        m = m_get(M_DONTWAIT, MT_SOOPTS);
1.171     plunky   2172:        if (m == NULL)
                   2173:                return NULL;
                   2174:
1.176     plunky   2175:        if (sopt->sopt_size > MLEN) {
                   2176:                MCLGET(m, M_DONTWAIT);
                   2177:                if ((m->m_flags & M_EXT) == 0) {
                   2178:                        m_free(m);
                   2179:                        return NULL;
                   2180:                }
1.1       cgd      2181:        }
1.176     plunky   2182:
                   2183:        memcpy(mtod(m, void *), sopt->sopt_data, sopt->sopt_size);
                   2184:        m->m_len = sopt->sopt_size;
1.160     ad       2185:
1.171     plunky   2186:        return m;
1.1       cgd      2187: }
                   2188:
1.14      mycroft  2189: void
1.54      lukem    2190: sohasoutofband(struct socket *so)
1.1       cgd      2191: {
1.153     rmind    2192:
1.90      christos 2193:        fownsignal(so->so_pgid, SIGURG, POLL_PRI, POLLPRI|POLLRDBAND, so);
1.189     ad       2194:        selnotify(&so->so_rcv.sb_sel, POLLPRI | POLLRDBAND, NOTE_SUBMIT);
1.1       cgd      2195: }
1.72      jdolecek 2196:
                   2197: static void
                   2198: filt_sordetach(struct knote *kn)
                   2199: {
                   2200:        struct socket   *so;
                   2201:
1.155     ad       2202:        so = ((file_t *)kn->kn_obj)->f_data;
1.160     ad       2203:        solock(so);
1.73      christos 2204:        SLIST_REMOVE(&so->so_rcv.sb_sel.sel_klist, kn, knote, kn_selnext);
                   2205:        if (SLIST_EMPTY(&so->so_rcv.sb_sel.sel_klist))
1.72      jdolecek 2206:                so->so_rcv.sb_flags &= ~SB_KNOTE;
1.160     ad       2207:        sounlock(so);
1.72      jdolecek 2208: }
                   2209:
                   2210: /*ARGSUSED*/
                   2211: static int
1.129     yamt     2212: filt_soread(struct knote *kn, long hint)
1.72      jdolecek 2213: {
                   2214:        struct socket   *so;
1.160     ad       2215:        int rv;
1.72      jdolecek 2216:
1.155     ad       2217:        so = ((file_t *)kn->kn_obj)->f_data;
1.160     ad       2218:        if (hint != NOTE_SUBMIT)
                   2219:                solock(so);
1.72      jdolecek 2220:        kn->kn_data = so->so_rcv.sb_cc;
                   2221:        if (so->so_state & SS_CANTRCVMORE) {
1.108     perry    2222:                kn->kn_flags |= EV_EOF;
1.72      jdolecek 2223:                kn->kn_fflags = so->so_error;
1.160     ad       2224:                rv = 1;
                   2225:        } else if (so->so_error)        /* temporary udp error */
                   2226:                rv = 1;
                   2227:        else if (kn->kn_sfflags & NOTE_LOWAT)
                   2228:                rv = (kn->kn_data >= kn->kn_sdata);
                   2229:        else
                   2230:                rv = (kn->kn_data >= so->so_rcv.sb_lowat);
                   2231:        if (hint != NOTE_SUBMIT)
                   2232:                sounlock(so);
                   2233:        return rv;
1.72      jdolecek 2234: }
                   2235:
                   2236: static void
                   2237: filt_sowdetach(struct knote *kn)
                   2238: {
                   2239:        struct socket   *so;
                   2240:
1.155     ad       2241:        so = ((file_t *)kn->kn_obj)->f_data;
1.160     ad       2242:        solock(so);
1.73      christos 2243:        SLIST_REMOVE(&so->so_snd.sb_sel.sel_klist, kn, knote, kn_selnext);
                   2244:        if (SLIST_EMPTY(&so->so_snd.sb_sel.sel_klist))
1.72      jdolecek 2245:                so->so_snd.sb_flags &= ~SB_KNOTE;
1.160     ad       2246:        sounlock(so);
1.72      jdolecek 2247: }
                   2248:
                   2249: /*ARGSUSED*/
                   2250: static int
1.129     yamt     2251: filt_sowrite(struct knote *kn, long hint)
1.72      jdolecek 2252: {
                   2253:        struct socket   *so;
1.160     ad       2254:        int rv;
1.72      jdolecek 2255:
1.155     ad       2256:        so = ((file_t *)kn->kn_obj)->f_data;
1.160     ad       2257:        if (hint != NOTE_SUBMIT)
                   2258:                solock(so);
1.72      jdolecek 2259:        kn->kn_data = sbspace(&so->so_snd);
                   2260:        if (so->so_state & SS_CANTSENDMORE) {
1.108     perry    2261:                kn->kn_flags |= EV_EOF;
1.72      jdolecek 2262:                kn->kn_fflags = so->so_error;
1.160     ad       2263:                rv = 1;
                   2264:        } else if (so->so_error)        /* temporary udp error */
                   2265:                rv = 1;
                   2266:        else if (((so->so_state & SS_ISCONNECTED) == 0) &&
1.72      jdolecek 2267:            (so->so_proto->pr_flags & PR_CONNREQUIRED))
1.160     ad       2268:                rv = 0;
                   2269:        else if (kn->kn_sfflags & NOTE_LOWAT)
                   2270:                rv = (kn->kn_data >= kn->kn_sdata);
                   2271:        else
                   2272:                rv = (kn->kn_data >= so->so_snd.sb_lowat);
                   2273:        if (hint != NOTE_SUBMIT)
                   2274:                sounlock(so);
                   2275:        return rv;
1.72      jdolecek 2276: }
                   2277:
                   2278: /*ARGSUSED*/
                   2279: static int
1.129     yamt     2280: filt_solisten(struct knote *kn, long hint)
1.72      jdolecek 2281: {
                   2282:        struct socket   *so;
1.160     ad       2283:        int rv;
1.72      jdolecek 2284:
1.155     ad       2285:        so = ((file_t *)kn->kn_obj)->f_data;
1.72      jdolecek 2286:
                   2287:        /*
                   2288:         * Set kn_data to number of incoming connections, not
                   2289:         * counting partial (incomplete) connections.
1.108     perry    2290:         */
1.160     ad       2291:        if (hint != NOTE_SUBMIT)
                   2292:                solock(so);
1.72      jdolecek 2293:        kn->kn_data = so->so_qlen;
1.160     ad       2294:        rv = (kn->kn_data > 0);
                   2295:        if (hint != NOTE_SUBMIT)
                   2296:                sounlock(so);
                   2297:        return rv;
1.72      jdolecek 2298: }
                   2299:
                   2300: static const struct filterops solisten_filtops =
                   2301:        { 1, NULL, filt_sordetach, filt_solisten };
                   2302: static const struct filterops soread_filtops =
                   2303:        { 1, NULL, filt_sordetach, filt_soread };
                   2304: static const struct filterops sowrite_filtops =
                   2305:        { 1, NULL, filt_sowdetach, filt_sowrite };
                   2306:
                   2307: int
1.129     yamt     2308: soo_kqfilter(struct file *fp, struct knote *kn)
1.72      jdolecek 2309: {
                   2310:        struct socket   *so;
                   2311:        struct sockbuf  *sb;
                   2312:
1.155     ad       2313:        so = ((file_t *)kn->kn_obj)->f_data;
1.160     ad       2314:        solock(so);
1.72      jdolecek 2315:        switch (kn->kn_filter) {
                   2316:        case EVFILT_READ:
                   2317:                if (so->so_options & SO_ACCEPTCONN)
                   2318:                        kn->kn_fop = &solisten_filtops;
                   2319:                else
                   2320:                        kn->kn_fop = &soread_filtops;
                   2321:                sb = &so->so_rcv;
                   2322:                break;
                   2323:        case EVFILT_WRITE:
                   2324:                kn->kn_fop = &sowrite_filtops;
                   2325:                sb = &so->so_snd;
                   2326:                break;
                   2327:        default:
1.160     ad       2328:                sounlock(so);
1.149     pooka    2329:                return (EINVAL);
1.72      jdolecek 2330:        }
1.73      christos 2331:        SLIST_INSERT_HEAD(&sb->sb_sel.sel_klist, kn, kn_selnext);
1.72      jdolecek 2332:        sb->sb_flags |= SB_KNOTE;
1.160     ad       2333:        sounlock(so);
1.72      jdolecek 2334:        return (0);
                   2335: }
                   2336:
1.154     ad       2337: static int
                   2338: sodopoll(struct socket *so, int events)
                   2339: {
                   2340:        int revents;
                   2341:
                   2342:        revents = 0;
                   2343:
                   2344:        if (events & (POLLIN | POLLRDNORM))
                   2345:                if (soreadable(so))
                   2346:                        revents |= events & (POLLIN | POLLRDNORM);
                   2347:
                   2348:        if (events & (POLLOUT | POLLWRNORM))
                   2349:                if (sowritable(so))
                   2350:                        revents |= events & (POLLOUT | POLLWRNORM);
                   2351:
                   2352:        if (events & (POLLPRI | POLLRDBAND))
                   2353:                if (so->so_oobmark || (so->so_state & SS_RCVATMARK))
                   2354:                        revents |= events & (POLLPRI | POLLRDBAND);
                   2355:
                   2356:        return revents;
                   2357: }
                   2358:
                   2359: int
                   2360: sopoll(struct socket *so, int events)
                   2361: {
                   2362:        int revents = 0;
                   2363:
1.160     ad       2364: #ifndef DIAGNOSTIC
                   2365:        /*
                   2366:         * Do a quick, unlocked check in expectation that the socket
                   2367:         * will be ready for I/O.  Don't do this check if DIAGNOSTIC,
                   2368:         * as the solocked() assertions will fail.
                   2369:         */
1.154     ad       2370:        if ((revents = sodopoll(so, events)) != 0)
                   2371:                return revents;
1.160     ad       2372: #endif
1.154     ad       2373:
1.160     ad       2374:        solock(so);
1.154     ad       2375:        if ((revents = sodopoll(so, events)) == 0) {
                   2376:                if (events & (POLLIN | POLLPRI | POLLRDNORM | POLLRDBAND)) {
                   2377:                        selrecord(curlwp, &so->so_rcv.sb_sel);
1.160     ad       2378:                        so->so_rcv.sb_flags |= SB_NOTIFY;
1.154     ad       2379:                }
                   2380:
                   2381:                if (events & (POLLOUT | POLLWRNORM)) {
                   2382:                        selrecord(curlwp, &so->so_snd.sb_sel);
1.160     ad       2383:                        so->so_snd.sb_flags |= SB_NOTIFY;
1.154     ad       2384:                }
                   2385:        }
1.160     ad       2386:        sounlock(so);
1.154     ad       2387:
                   2388:        return revents;
                   2389: }
                   2390:
                   2391:
1.94      yamt     2392: #include <sys/sysctl.h>
                   2393:
                   2394: static int sysctl_kern_somaxkva(SYSCTLFN_PROTO);
                   2395:
                   2396: /*
                   2397:  * sysctl helper routine for kern.somaxkva.  ensures that the given
                   2398:  * value is not too small.
                   2399:  * (XXX should we maybe make sure it's not too large as well?)
                   2400:  */
                   2401: static int
                   2402: sysctl_kern_somaxkva(SYSCTLFN_ARGS)
                   2403: {
                   2404:        int error, new_somaxkva;
                   2405:        struct sysctlnode node;
                   2406:
                   2407:        new_somaxkva = somaxkva;
                   2408:        node = *rnode;
                   2409:        node.sysctl_data = &new_somaxkva;
                   2410:        error = sysctl_lookup(SYSCTLFN_CALL(&node));
                   2411:        if (error || newp == NULL)
                   2412:                return (error);
                   2413:
                   2414:        if (new_somaxkva < (16 * 1024 * 1024)) /* sanity */
                   2415:                return (EINVAL);
                   2416:
1.136     ad       2417:        mutex_enter(&so_pendfree_lock);
1.94      yamt     2418:        somaxkva = new_somaxkva;
1.136     ad       2419:        cv_broadcast(&socurkva_cv);
                   2420:        mutex_exit(&so_pendfree_lock);
1.94      yamt     2421:
                   2422:        return (error);
                   2423: }
                   2424:
1.178     pooka    2425: static void
1.187     cegger   2426: sysctl_kern_somaxkva_setup(void)
1.94      yamt     2427: {
                   2428:
1.178     pooka    2429:        KASSERT(socket_sysctllog == NULL);
                   2430:        sysctl_createv(&socket_sysctllog, 0, NULL, NULL,
1.97      atatat   2431:                       CTLFLAG_PERMANENT,
                   2432:                       CTLTYPE_NODE, "kern", NULL,
                   2433:                       NULL, 0, NULL, 0,
                   2434:                       CTL_KERN, CTL_EOL);
                   2435:
1.178     pooka    2436:        sysctl_createv(&socket_sysctllog, 0, NULL, NULL,
1.97      atatat   2437:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.103     atatat   2438:                       CTLTYPE_INT, "somaxkva",
                   2439:                       SYSCTL_DESCR("Maximum amount of kernel memory to be "
                   2440:                                    "used for socket buffers"),
1.94      yamt     2441:                       sysctl_kern_somaxkva, 0, NULL, 0,
                   2442:                       CTL_KERN, KERN_SOMAXKVA, CTL_EOL);
                   2443: }

CVSweb <webmaster@jp.NetBSD.org>