| version 1.56.2.7, 2002/09/06 08:48:16 |
version 1.57, 2001/09/17 18:59:29 |
|
|
| /* $NetBSD$ */ |
/* $NetBSD$ */ |
| |
|
| /*- |
|
| * Copyright (c) 2002 The NetBSD Foundation, Inc. |
|
| * All rights reserved. |
|
| * |
|
| * This code is derived from software contributed to The NetBSD Foundation |
|
| * by Jason R. Thorpe of Wasabi Systems, Inc. |
|
| * |
|
| * Redistribution and use in source and binary forms, with or without |
|
| * modification, are permitted provided that the following conditions |
|
| * are met: |
|
| * 1. Redistributions of source code must retain the above copyright |
|
| * notice, this list of conditions and the following disclaimer. |
|
| * 2. Redistributions in binary form must reproduce the above copyright |
|
| * notice, this list of conditions and the following disclaimer in the |
|
| * documentation and/or other materials provided with the distribution. |
|
| * 3. All advertising materials mentioning features or use of this software |
|
| * must display the following acknowledgement: |
|
| * This product includes software developed by the NetBSD |
|
| * Foundation, Inc. and its contributors. |
|
| * 4. Neither the name of The NetBSD Foundation nor the names of its |
|
| * contributors may be used to endorse or promote products derived |
|
| * from this software without specific prior written permission. |
|
| * |
|
| * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS |
|
| * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED |
|
| * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR |
|
| * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS |
|
| * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
|
| * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
|
| * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
|
| * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN |
|
| * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
|
| * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
|
| * POSSIBILITY OF SUCH DAMAGE. |
|
| */ |
|
| |
|
| /* |
/* |
| * Copyright (c) 1982, 1986, 1988, 1990, 1993 |
* Copyright (c) 1982, 1986, 1988, 1990, 1993 |
| * The Regents of the University of California. All rights reserved. |
* The Regents of the University of California. All rights reserved. |
|
|
| * @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95 |
* @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95 |
| */ |
*/ |
| |
|
| #include <sys/cdefs.h> |
#include "opt_compat_sunos.h" |
| __KERNEL_RCSID(0, "$NetBSD$"); |
|
| |
|
| #include "opt_sock_counters.h" |
|
| #include "opt_sosend_loan.h" |
|
| |
|
| #include <sys/param.h> |
#include <sys/param.h> |
| #include <sys/systm.h> |
#include <sys/systm.h> |
| Line 91 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| Line 51 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| #include <sys/signalvar.h> |
#include <sys/signalvar.h> |
| #include <sys/resourcevar.h> |
#include <sys/resourcevar.h> |
| #include <sys/pool.h> |
#include <sys/pool.h> |
| #include <sys/event.h> |
|
| |
|
| #include <uvm/uvm.h> |
|
| |
|
| static void filt_sordetach(struct knote *kn); |
|
| static int filt_soread(struct knote *kn, long hint); |
|
| static void filt_sowdetach(struct knote *kn); |
|
| static int filt_sowrite(struct knote *kn, long hint); |
|
| static int filt_solisten(struct knote *kn, long hint); |
|
| |
|
| static const struct filterops solisten_filtops = |
|
| { 1, NULL, filt_sordetach, filt_solisten }; |
|
| const struct filterops soread_filtops = |
|
| { 1, NULL, filt_sordetach, filt_soread }; |
|
| const struct filterops sowrite_filtops = |
|
| { 1, NULL, filt_sowdetach, filt_sowrite }; |
|
| |
|
| struct pool socket_pool; |
struct pool socket_pool; |
| |
|
| extern int somaxconn; /* patchable (XXX sysctl) */ |
extern int somaxconn; /* patchable (XXX sysctl) */ |
| int somaxconn = SOMAXCONN; |
int somaxconn = SOMAXCONN; |
| |
|
| #ifdef SOSEND_COUNTERS |
|
| #include <sys/device.h> |
|
| |
|
| struct evcnt sosend_loan_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, |
|
| NULL, "sosend", "loan big"); |
|
| struct evcnt sosend_copy_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, |
|
| NULL, "sosend", "copy big"); |
|
| struct evcnt sosend_copy_small = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, |
|
| NULL, "sosend", "copy small"); |
|
| struct evcnt sosend_kvalimit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, |
|
| NULL, "sosend", "kva limit"); |
|
| |
|
| #define SOSEND_COUNTER_INCR(ev) (ev)->ev_count++ |
|
| |
|
| #else |
|
| |
|
| #define SOSEND_COUNTER_INCR(ev) /* nothing */ |
|
| |
|
| #endif /* SOSEND_COUNTERS */ |
|
| |
|
| void |
void |
| soinit(void) |
soinit(void) |
| { |
{ |
| |
|
| pool_init(&socket_pool, sizeof(struct socket), 0, 0, 0, |
pool_init(&socket_pool, sizeof(struct socket), 0, 0, 0, |
| "sockpl", NULL); |
"sockpl", 0, NULL, NULL, M_SOCKET); |
| |
|
| #ifdef SOSEND_COUNTERS |
|
| evcnt_attach_static(&sosend_loan_big); |
|
| evcnt_attach_static(&sosend_copy_big); |
|
| evcnt_attach_static(&sosend_copy_small); |
|
| evcnt_attach_static(&sosend_kvalimit); |
|
| #endif /* SOSEND_COUNTERS */ |
|
| } |
|
| |
|
| #ifdef SOSEND_NO_LOAN |
|
| int use_sosend_loan = 0; |
|
| #else |
|
| int use_sosend_loan = 1; |
|
| #endif |
|
| |
|
| struct mbuf *so_pendfree; |
|
| |
|
| int somaxkva = 16 * 1024 * 1024; |
|
| int socurkva; |
|
| int sokvawaiters; |
|
| |
|
| #define SOCK_LOAN_THRESH 4096 |
|
| #define SOCK_LOAN_CHUNK 65536 |
|
| |
|
| static void |
|
| sodoloanfree(caddr_t buf, u_int size) |
|
| { |
|
| struct vm_page **pgs; |
|
| vaddr_t va, sva, eva; |
|
| vsize_t len; |
|
| paddr_t pa; |
|
| int i, npgs; |
|
| |
|
| eva = round_page((vaddr_t) buf + size); |
|
| sva = trunc_page((vaddr_t) buf); |
|
| len = eva - sva; |
|
| npgs = len >> PAGE_SHIFT; |
|
| |
|
| pgs = alloca(npgs * sizeof(*pgs)); |
|
| |
|
| for (i = 0, va = sva; va < eva; i++, va += PAGE_SIZE) { |
|
| if (pmap_extract(pmap_kernel(), va, &pa) == FALSE) |
|
| panic("sodoloanfree: va 0x%lx not mapped", va); |
|
| pgs[i] = PHYS_TO_VM_PAGE(pa); |
|
| } |
|
| |
|
| pmap_kremove(sva, len); |
|
| pmap_update(pmap_kernel()); |
|
| uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE); |
|
| uvm_km_free(kernel_map, sva, len); |
|
| socurkva -= len; |
|
| if (sokvawaiters) |
|
| wakeup(&socurkva); |
|
| } |
|
| |
|
| static size_t |
|
| sodopendfree(struct socket *so) |
|
| { |
|
| struct mbuf *m; |
|
| size_t rv = 0; |
|
| int s; |
|
| |
|
| s = splvm(); |
|
| |
|
| for (;;) { |
|
| m = so_pendfree; |
|
| if (m == NULL) |
|
| break; |
|
| so_pendfree = m->m_next; |
|
| splx(s); |
|
| |
|
| rv += m->m_ext.ext_size; |
|
| sodoloanfree(m->m_ext.ext_buf, m->m_ext.ext_size); |
|
| s = splvm(); |
|
| pool_cache_put(&mbpool_cache, m); |
|
| } |
|
| |
|
| for (;;) { |
|
| m = so->so_pendfree; |
|
| if (m == NULL) |
|
| break; |
|
| so->so_pendfree = m->m_next; |
|
| splx(s); |
|
| |
|
| rv += m->m_ext.ext_size; |
|
| sodoloanfree(m->m_ext.ext_buf, m->m_ext.ext_size); |
|
| s = splvm(); |
|
| pool_cache_put(&mbpool_cache, m); |
|
| } |
|
| |
|
| splx(s); |
|
| return (rv); |
|
| } |
|
| |
|
| static void |
|
| soloanfree(struct mbuf *m, caddr_t buf, u_int size, void *arg) |
|
| { |
|
| struct socket *so = arg; |
|
| int s; |
|
| |
|
| if (m == NULL) { |
|
| sodoloanfree(buf, size); |
|
| return; |
|
| } |
|
| |
|
| s = splvm(); |
|
| m->m_next = so->so_pendfree; |
|
| so->so_pendfree = m; |
|
| splx(s); |
|
| if (sokvawaiters) |
|
| wakeup(&socurkva); |
|
| } |
|
| |
|
| static long |
|
| sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space) |
|
| { |
|
| struct iovec *iov = uio->uio_iov; |
|
| vaddr_t sva, eva; |
|
| vsize_t len; |
|
| struct vm_page **pgs; |
|
| vaddr_t lva, va; |
|
| int npgs, s, i, error; |
|
| |
|
| if (uio->uio_segflg != UIO_USERSPACE) |
|
| return (0); |
|
| |
|
| if (iov->iov_len < (size_t) space) |
|
| space = iov->iov_len; |
|
| if (space > SOCK_LOAN_CHUNK) |
|
| space = SOCK_LOAN_CHUNK; |
|
| |
|
| eva = round_page((vaddr_t) iov->iov_base + space); |
|
| sva = trunc_page((vaddr_t) iov->iov_base); |
|
| len = eva - sva; |
|
| npgs = len >> PAGE_SHIFT; |
|
| |
|
| while (socurkva + len > somaxkva) { |
|
| if (sodopendfree(so)) |
|
| continue; |
|
| SOSEND_COUNTER_INCR(&sosend_kvalimit); |
|
| s = splvm(); |
|
| sokvawaiters++; |
|
| (void) tsleep(&socurkva, PVM, "sokva", 0); |
|
| sokvawaiters--; |
|
| splx(s); |
|
| } |
|
| |
|
| lva = uvm_km_valloc_wait(kernel_map, len); |
|
| if (lva == 0) |
|
| return (0); |
|
| socurkva += len; |
|
| |
|
| pgs = alloca(npgs * sizeof(*pgs)); |
|
| |
|
| error = uvm_loan(&uio->uio_procp->p_vmspace->vm_map, sva, len, |
|
| pgs, UVM_LOAN_TOPAGE); |
|
| if (error) { |
|
| uvm_km_free(kernel_map, lva, len); |
|
| socurkva -= len; |
|
| return (0); |
|
| } |
|
| |
|
| for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE) |
|
| pmap_kenter_pa(va, VM_PAGE_TO_PHYS(pgs[i]), VM_PROT_READ); |
|
| pmap_update(pmap_kernel()); |
|
| |
|
| lva += (vaddr_t) iov->iov_base & PAGE_MASK; |
|
| |
|
| MEXTADD(m, (caddr_t) lva, space, M_MBUF, soloanfree, so); |
|
| |
|
| uio->uio_resid -= space; |
|
| /* uio_offset not updated, not set/used for write(2) */ |
|
| uio->uio_iov->iov_base = (caddr_t) uio->uio_iov->iov_base + space; |
|
| uio->uio_iov->iov_len -= space; |
|
| if (uio->uio_iov->iov_len == 0) { |
|
| uio->uio_iov++; |
|
| uio->uio_iovcnt--; |
|
| } |
|
| |
|
| return (space); |
|
| } |
} |
| |
|
| /* |
/* |
| Line 365 socreate(int dom, struct socket **aso, i |
|
| Line 109 socreate(int dom, struct socket **aso, i |
|
| splx(s); |
splx(s); |
| return (error); |
return (error); |
| } |
} |
| |
#ifdef COMPAT_SUNOS |
| |
{ |
| |
extern struct emul emul_sunos; |
| |
if (p->p_emul == &emul_sunos && type == SOCK_DGRAM) |
| |
so->so_options |= SO_BROADCAST; |
| |
} |
| |
#endif |
| splx(s); |
splx(s); |
| *aso = so; |
*aso = so; |
| return (0); |
return (0); |
| Line 394 solisten(struct socket *so, int backlog) |
|
| Line 145 solisten(struct socket *so, int backlog) |
|
| splx(s); |
splx(s); |
| return (error); |
return (error); |
| } |
} |
| if (TAILQ_EMPTY(&so->so_q)) |
if (so->so_q.tqh_first == NULL) |
| so->so_options |= SO_ACCEPTCONN; |
so->so_options |= SO_ACCEPTCONN; |
| if (backlog < 0) |
if (backlog < 0) |
| backlog = 0; |
backlog = 0; |
| Line 406 solisten(struct socket *so, int backlog) |
|
| Line 157 solisten(struct socket *so, int backlog) |
|
| void |
void |
| sofree(struct socket *so) |
sofree(struct socket *so) |
| { |
{ |
| struct mbuf *m; |
|
| |
|
| if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0) |
if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0) |
| return; |
return; |
| Line 421 sofree(struct socket *so) |
|
| Line 171 sofree(struct socket *so) |
|
| } |
} |
| sbrelease(&so->so_snd); |
sbrelease(&so->so_snd); |
| sorflush(so); |
sorflush(so); |
| while ((m = so->so_pendfree) != NULL) { |
|
| so->so_pendfree = m->m_next; |
|
| m->m_next = so_pendfree; |
|
| so_pendfree = m; |
|
| } |
|
| pool_put(&socket_pool, so); |
pool_put(&socket_pool, so); |
| } |
} |
| |
|
| Line 443 soclose(struct socket *so) |
|
| Line 188 soclose(struct socket *so) |
|
| error = 0; |
error = 0; |
| s = splsoftnet(); /* conservative */ |
s = splsoftnet(); /* conservative */ |
| if (so->so_options & SO_ACCEPTCONN) { |
if (so->so_options & SO_ACCEPTCONN) { |
| while ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) { |
while ((so2 = so->so_q0.tqh_first) != 0) { |
| (void) soqremque(so2, 0); |
(void) soqremque(so2, 0); |
| (void) soabort(so2); |
(void) soabort(so2); |
| } |
} |
| while ((so2 = TAILQ_FIRST(&so->so_q)) != 0) { |
while ((so2 = so->so_q.tqh_first) != 0) { |
| (void) soqremque(so2, 1); |
(void) soqremque(so2, 1); |
| (void) soabort(so2); |
(void) soabort(so2); |
| } |
} |
| Line 581 sodisconnect(struct socket *so) |
|
| Line 326 sodisconnect(struct socket *so) |
|
| (struct proc *)0); |
(struct proc *)0); |
| bad: |
bad: |
| splx(s); |
splx(s); |
| sodopendfree(so); |
|
| return (error); |
return (error); |
| } |
} |
| |
|
| Line 609 sosend(struct socket *so, struct mbuf *a |
|
| Line 353 sosend(struct socket *so, struct mbuf *a |
|
| { |
{ |
| struct proc *p; |
struct proc *p; |
| struct mbuf **mp, *m; |
struct mbuf **mp, *m; |
| long space, len, resid, clen, mlen; |
long space, len, resid; |
| int error, s, dontroute, atomic; |
int clen, error, s, dontroute, mlen, atomic; |
| |
|
| sodopendfree(so); |
|
| |
|
| p = curproc; /* XXX */ |
p = curproc; /* XXX */ |
| clen = 0; |
clen = 0; |
| Line 699 sosend(struct socket *so, struct mbuf *a |
|
| Line 441 sosend(struct socket *so, struct mbuf *a |
|
| MGET(m, M_WAIT, MT_DATA); |
MGET(m, M_WAIT, MT_DATA); |
| mlen = MLEN; |
mlen = MLEN; |
| } |
} |
| if (use_sosend_loan && |
|
| uio->uio_iov->iov_len >= SOCK_LOAN_THRESH && |
|
| space >= SOCK_LOAN_THRESH && |
|
| (len = sosend_loan(so, uio, m, |
|
| space)) != 0) { |
|
| SOSEND_COUNTER_INCR(&sosend_loan_big); |
|
| space -= len; |
|
| goto have_data; |
|
| } |
|
| if (resid >= MINCLSIZE && space >= MCLBYTES) { |
if (resid >= MINCLSIZE && space >= MCLBYTES) { |
| SOSEND_COUNTER_INCR(&sosend_copy_big); |
|
| MCLGET(m, M_WAIT); |
MCLGET(m, M_WAIT); |
| if ((m->m_flags & M_EXT) == 0) |
if ((m->m_flags & M_EXT) == 0) |
| goto nopages; |
goto nopages; |
| mlen = MCLBYTES; |
mlen = MCLBYTES; |
| |
#ifdef MAPPED_MBUFS |
| |
len = min(MCLBYTES, resid); |
| |
#else |
| if (atomic && top == 0) { |
if (atomic && top == 0) { |
| len = lmin(MCLBYTES - max_hdr, |
len = min(MCLBYTES - max_hdr, |
| resid); |
resid); |
| m->m_data += max_hdr; |
m->m_data += max_hdr; |
| } else |
} else |
| len = lmin(MCLBYTES, resid); |
len = min(MCLBYTES, resid); |
| |
#endif |
| space -= len; |
space -= len; |
| } else { |
} else { |
| nopages: |
nopages: |
| SOSEND_COUNTER_INCR(&sosend_copy_small); |
len = min(min(mlen, resid), space); |
| len = lmin(lmin(mlen, resid), space); |
|
| space -= len; |
space -= len; |
| /* |
/* |
| * For datagram protocols, leave room |
* For datagram protocols, leave room |
| Line 735 sosend(struct socket *so, struct mbuf *a |
|
| Line 470 sosend(struct socket *so, struct mbuf *a |
|
| } |
} |
| error = uiomove(mtod(m, caddr_t), (int)len, |
error = uiomove(mtod(m, caddr_t), (int)len, |
| uio); |
uio); |
| have_data: |
|
| resid = uio->uio_resid; |
resid = uio->uio_resid; |
| m->m_len = len; |
m->m_len = len; |
| *mp = m; |
*mp = m; |
| Line 811 soreceive(struct socket *so, struct mbuf |
|
| Line 545 soreceive(struct socket *so, struct mbuf |
|
| int flags, len, error, s, offset, moff, type, orig_resid; |
int flags, len, error, s, offset, moff, type, orig_resid; |
| struct protosw *pr; |
struct protosw *pr; |
| struct mbuf *nextrecord; |
struct mbuf *nextrecord; |
| int mbuf_removed = 0; |
|
| |
|
| pr = so->so_proto; |
pr = so->so_proto; |
| mp = mp0; |
mp = mp0; |
| Line 825 soreceive(struct socket *so, struct mbuf |
|
| Line 558 soreceive(struct socket *so, struct mbuf |
|
| flags = *flagsp &~ MSG_EOR; |
flags = *flagsp &~ MSG_EOR; |
| else |
else |
| flags = 0; |
flags = 0; |
| |
|
| if ((flags & MSG_DONTWAIT) == 0) |
|
| sodopendfree(so); |
|
| |
|
| if (flags & MSG_OOB) { |
if (flags & MSG_OOB) { |
| m = m_get(M_WAIT, MT_DATA); |
m = m_get(M_WAIT, MT_DATA); |
| error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m, |
error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m, |
| Line 908 soreceive(struct socket *so, struct mbuf |
|
| Line 637 soreceive(struct socket *so, struct mbuf |
|
| error = EWOULDBLOCK; |
error = EWOULDBLOCK; |
| goto release; |
goto release; |
| } |
} |
| SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1"); |
|
| sbunlock(&so->so_rcv); |
sbunlock(&so->so_rcv); |
| error = sbwait(&so->so_rcv); |
error = sbwait(&so->so_rcv); |
| splx(s); |
splx(s); |
| Line 918 soreceive(struct socket *so, struct mbuf |
|
| Line 645 soreceive(struct socket *so, struct mbuf |
|
| goto restart; |
goto restart; |
| } |
} |
| dontblock: |
dontblock: |
| /* |
|
| * On entry here, m points to the first record of the socket buffer. |
|
| * While we process the initial mbufs containing address and control |
|
| * info, we save a copy of m->m_nextpkt into nextrecord. |
|
| */ |
|
| #ifdef notyet /* XXXX */ |
#ifdef notyet /* XXXX */ |
| if (uio->uio_procp) |
if (uio->uio_procp) |
| uio->uio_procp->p_stats->p_ru.ru_msgrcv++; |
uio->uio_procp->p_stats->p_ru.ru_msgrcv++; |
| #endif |
#endif |
| KASSERT(m == so->so_rcv.sb_mb); |
|
| SBLASTRECORDCHK(&so->so_rcv, "soreceive 1"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive 1"); |
|
| nextrecord = m->m_nextpkt; |
nextrecord = m->m_nextpkt; |
| if (pr->pr_flags & PR_ADDR) { |
if (pr->pr_flags & PR_ADDR) { |
| #ifdef DIAGNOSTIC |
#ifdef DIAGNOSTIC |
| Line 943 soreceive(struct socket *so, struct mbuf |
|
| Line 662 soreceive(struct socket *so, struct mbuf |
|
| m = m->m_next; |
m = m->m_next; |
| } else { |
} else { |
| sbfree(&so->so_rcv, m); |
sbfree(&so->so_rcv, m); |
| mbuf_removed = 1; |
|
| if (paddr) { |
if (paddr) { |
| *paddr = m; |
*paddr = m; |
| so->so_rcv.sb_mb = m->m_next; |
so->so_rcv.sb_mb = m->m_next; |
| Line 962 soreceive(struct socket *so, struct mbuf |
|
| Line 680 soreceive(struct socket *so, struct mbuf |
|
| m = m->m_next; |
m = m->m_next; |
| } else { |
} else { |
| sbfree(&so->so_rcv, m); |
sbfree(&so->so_rcv, m); |
| mbuf_removed = 1; |
|
| if (controlp) { |
if (controlp) { |
| if (pr->pr_domain->dom_externalize && |
if (pr->pr_domain->dom_externalize && |
| mtod(m, struct cmsghdr *)->cmsg_type == |
mtod(m, struct cmsghdr *)->cmsg_type == |
| Line 982 soreceive(struct socket *so, struct mbuf |
|
| Line 699 soreceive(struct socket *so, struct mbuf |
|
| controlp = &(*controlp)->m_next; |
controlp = &(*controlp)->m_next; |
| } |
} |
| } |
} |
| |
|
| /* |
|
| * If m is non-NULL, we have some data to read. From now on, |
|
| * make sure to keep sb_lastrecord consistent when working on |
|
| * the last packet on the chain (nextrecord == NULL) and we |
|
| * change m->m_nextpkt. |
|
| */ |
|
| if (m) { |
if (m) { |
| if ((flags & MSG_PEEK) == 0) { |
if ((flags & MSG_PEEK) == 0) |
| m->m_nextpkt = nextrecord; |
m->m_nextpkt = nextrecord; |
| /* |
|
| * If nextrecord == NULL (this is a single chain), |
|
| * then sb_lastrecord may not be valid here if m |
|
| * was changed earlier. |
|
| */ |
|
| if (nextrecord == NULL) { |
|
| KASSERT(so->so_rcv.sb_mb == m); |
|
| so->so_rcv.sb_lastrecord = m; |
|
| } |
|
| } |
|
| type = m->m_type; |
type = m->m_type; |
| if (type == MT_OOBDATA) |
if (type == MT_OOBDATA) |
| flags |= MSG_OOB; |
flags |= MSG_OOB; |
| } else { |
|
| if ((flags & MSG_PEEK) == 0) { |
|
| KASSERT(so->so_rcv.sb_mb == m); |
|
| so->so_rcv.sb_mb = nextrecord; |
|
| SB_EMPTY_FIXUP(&so->so_rcv); |
|
| } |
|
| } |
} |
| SBLASTRECORDCHK(&so->so_rcv, "soreceive 2"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive 2"); |
|
| |
|
| moff = 0; |
moff = 0; |
| offset = 0; |
offset = 0; |
| while (m && uio->uio_resid > 0 && error == 0) { |
while (m && uio->uio_resid > 0 && error == 0) { |
| Line 1042 soreceive(struct socket *so, struct mbuf |
|
| Line 733 soreceive(struct socket *so, struct mbuf |
|
| * block interrupts again. |
* block interrupts again. |
| */ |
*/ |
| if (mp == 0) { |
if (mp == 0) { |
| SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove"); |
|
| splx(s); |
splx(s); |
| error = uiomove(mtod(m, caddr_t) + moff, (int)len, uio); |
error = uiomove(mtod(m, caddr_t) + moff, (int)len, uio); |
| s = splsoftnet(); |
s = splsoftnet(); |
| if (error) { |
if (error) |
| /* |
|
| * If any part of the record has been removed |
|
| * (such as the MT_SONAME mbuf, which will |
|
| * happen when PR_ADDR, and thus also |
|
| * PR_ATOMIC, is set), then drop the entire |
|
| * record to maintain the atomicity of the |
|
| * receive operation. |
|
| * |
|
| * This avoids a later panic("receive 1a") |
|
| * when compiled with DIAGNOSTIC. |
|
| */ |
|
| if (m && mbuf_removed |
|
| && (pr->pr_flags & PR_ATOMIC)) |
|
| (void) sbdroprecord(&so->so_rcv); |
|
| |
|
| goto release; |
goto release; |
| } |
|
| } else |
} else |
| uio->uio_resid -= len; |
uio->uio_resid -= len; |
| if (len == m->m_len - moff) { |
if (len == m->m_len - moff) { |
| Line 1085 soreceive(struct socket *so, struct mbuf |
|
| Line 758 soreceive(struct socket *so, struct mbuf |
|
| MFREE(m, so->so_rcv.sb_mb); |
MFREE(m, so->so_rcv.sb_mb); |
| m = so->so_rcv.sb_mb; |
m = so->so_rcv.sb_mb; |
| } |
} |
| /* |
if (m) |
| * If m != NULL, we also know that |
|
| * so->so_rcv.sb_mb != NULL. |
|
| */ |
|
| KASSERT(so->so_rcv.sb_mb == m); |
|
| if (m) { |
|
| m->m_nextpkt = nextrecord; |
m->m_nextpkt = nextrecord; |
| if (nextrecord == NULL) |
|
| so->so_rcv.sb_lastrecord = m; |
|
| } else { |
|
| so->so_rcv.sb_mb = nextrecord; |
|
| SB_EMPTY_FIXUP(&so->so_rcv); |
|
| } |
|
| SBLASTRECORDCHK(&so->so_rcv, "soreceive 3"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive 3"); |
|
| } |
} |
| } else { |
} else { |
| if (flags & MSG_PEEK) |
if (flags & MSG_PEEK) |
| Line 1138 soreceive(struct socket *so, struct mbuf |
|
| Line 798 soreceive(struct socket *so, struct mbuf |
|
| !sosendallatonce(so) && !nextrecord) { |
!sosendallatonce(so) && !nextrecord) { |
| if (so->so_error || so->so_state & SS_CANTRCVMORE) |
if (so->so_error || so->so_state & SS_CANTRCVMORE) |
| break; |
break; |
| /* |
|
| * If we are peeking and the socket receive buffer is |
|
| * full, stop since we can't get more data to peek at. |
|
| */ |
|
| if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0) |
|
| break; |
|
| /* |
|
| * If we've drained the socket buffer, tell the |
|
| * protocol in case it needs to do something to |
|
| * get it filled again. |
|
| */ |
|
| if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb) |
|
| (*pr->pr_usrreq)(so, PRU_RCVD, |
|
| (struct mbuf *)0, |
|
| (struct mbuf *)(long)flags, |
|
| (struct mbuf *)0, |
|
| (struct proc *)0); |
|
| SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2"); |
|
| error = sbwait(&so->so_rcv); |
error = sbwait(&so->so_rcv); |
| if (error) { |
if (error) { |
| sbunlock(&so->so_rcv); |
sbunlock(&so->so_rcv); |
| Line 1174 soreceive(struct socket *so, struct mbuf |
|
| Line 815 soreceive(struct socket *so, struct mbuf |
|
| (void) sbdroprecord(&so->so_rcv); |
(void) sbdroprecord(&so->so_rcv); |
| } |
} |
| if ((flags & MSG_PEEK) == 0) { |
if ((flags & MSG_PEEK) == 0) { |
| if (m == 0) { |
if (m == 0) |
| /* |
|
| * First part is an inline SB_EMPTY_FIXUP(). Second |
|
| * part makes sure sb_lastrecord is up-to-date if |
|
| * there is still data in the socket buffer. |
|
| */ |
|
| so->so_rcv.sb_mb = nextrecord; |
so->so_rcv.sb_mb = nextrecord; |
| if (so->so_rcv.sb_mb == NULL) { |
|
| so->so_rcv.sb_mbtail = NULL; |
|
| so->so_rcv.sb_lastrecord = NULL; |
|
| } else if (nextrecord->m_nextpkt == NULL) |
|
| so->so_rcv.sb_lastrecord = nextrecord; |
|
| } |
|
| SBLASTRECORDCHK(&so->so_rcv, "soreceive 4"); |
|
| SBLASTMBUFCHK(&so->so_rcv, "soreceive 4"); |
|
| if (pr->pr_flags & PR_WANTRCVD && so->so_pcb) |
if (pr->pr_flags & PR_WANTRCVD && so->so_pcb) |
| (*pr->pr_usrreq)(so, PRU_RCVD, (struct mbuf *)0, |
(*pr->pr_usrreq)(so, PRU_RCVD, (struct mbuf *)0, |
| (struct mbuf *)(long)flags, (struct mbuf *)0, |
(struct mbuf *)(long)flags, (struct mbuf *)0, |
| Line 1482 sohasoutofband(struct socket *so) |
|
| Line 1110 sohasoutofband(struct socket *so) |
|
| psignal(p, SIGURG); |
psignal(p, SIGURG); |
| selwakeup(&so->so_rcv.sb_sel); |
selwakeup(&so->so_rcv.sb_sel); |
| } |
} |
| |
|
| |
|
| int |
|
| soo_kqfilter(struct file *fp, struct knote *kn) |
|
| { |
|
| struct socket *so; |
|
| struct sockbuf *sb; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| switch (kn->kn_filter) { |
|
| case EVFILT_READ: |
|
| if (so->so_options & SO_ACCEPTCONN) |
|
| kn->kn_fop = &solisten_filtops; |
|
| else |
|
| kn->kn_fop = &soread_filtops; |
|
| sb = &so->so_rcv; |
|
| break; |
|
| case EVFILT_WRITE: |
|
| kn->kn_fop = &sowrite_filtops; |
|
| sb = &so->so_snd; |
|
| break; |
|
| default: |
|
| return (1); |
|
| } |
|
| SLIST_INSERT_HEAD(&sb->sb_sel.si_klist, kn, kn_selnext); |
|
| sb->sb_flags |= SB_KNOTE; |
|
| return (0); |
|
| } |
|
| |
|
| static void |
|
| filt_sordetach(struct knote *kn) |
|
| { |
|
| struct socket *so; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| SLIST_REMOVE(&so->so_rcv.sb_sel.si_klist, kn, knote, kn_selnext); |
|
| if (SLIST_EMPTY(&so->so_rcv.sb_sel.si_klist)) |
|
| so->so_rcv.sb_flags &= ~SB_KNOTE; |
|
| } |
|
| |
|
| /*ARGSUSED*/ |
|
| static int |
|
| filt_soread(struct knote *kn, long hint) |
|
| { |
|
| struct socket *so; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| kn->kn_data = so->so_rcv.sb_cc; |
|
| if (so->so_state & SS_CANTRCVMORE) { |
|
| kn->kn_flags |= EV_EOF; |
|
| kn->kn_fflags = so->so_error; |
|
| return (1); |
|
| } |
|
| if (so->so_error) /* temporary udp error */ |
|
| return (1); |
|
| if (kn->kn_sfflags & NOTE_LOWAT) |
|
| return (kn->kn_data >= kn->kn_sdata); |
|
| return (kn->kn_data >= so->so_rcv.sb_lowat); |
|
| } |
|
| |
|
| static void |
|
| filt_sowdetach(struct knote *kn) |
|
| { |
|
| struct socket *so; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| SLIST_REMOVE(&so->so_snd.sb_sel.si_klist, kn, knote, kn_selnext); |
|
| if (SLIST_EMPTY(&so->so_snd.sb_sel.si_klist)) |
|
| so->so_snd.sb_flags &= ~SB_KNOTE; |
|
| } |
|
| |
|
| /*ARGSUSED*/ |
|
| static int |
|
| filt_sowrite(struct knote *kn, long hint) |
|
| { |
|
| struct socket *so; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| kn->kn_data = sbspace(&so->so_snd); |
|
| if (so->so_state & SS_CANTSENDMORE) { |
|
| kn->kn_flags |= EV_EOF; |
|
| kn->kn_fflags = so->so_error; |
|
| return (1); |
|
| } |
|
| if (so->so_error) /* temporary udp error */ |
|
| return (1); |
|
| if (((so->so_state & SS_ISCONNECTED) == 0) && |
|
| (so->so_proto->pr_flags & PR_CONNREQUIRED)) |
|
| return (0); |
|
| if (kn->kn_sfflags & NOTE_LOWAT) |
|
| return (kn->kn_data >= kn->kn_sdata); |
|
| return (kn->kn_data >= so->so_snd.sb_lowat); |
|
| } |
|
| |
|
| /*ARGSUSED*/ |
|
| static int |
|
| filt_solisten(struct knote *kn, long hint) |
|
| { |
|
| struct socket *so; |
|
| |
|
| so = (struct socket *)kn->kn_fp->f_data; |
|
| |
|
| /* |
|
| * Set kn_data to number of incoming connections, not |
|
| * counting partial (incomplete) connections. |
|
| */ |
|
| kn->kn_data = so->so_qlen; |
|
| return (kn->kn_data > 0); |
|
| } |
|