mirror/freebsd
1
0
Fork 0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-15 10:17:20 +00:00
Code Issues Releases Activity
37bde6c0a3
freebsd/sys/kern/uipc_usrreq.c
Robert Watson 6ea48a903c Introduce support for Mandatory Access Control and extensible 
kernel access control.

Authorize the creation of UNIX domain sockets in the file system
namespace via an appropriate invocation a MAC framework entry
point.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, NAI Labs
2002-08-01 01:18:42 +00:00

1517 lines
35 KiB
C
Raw Blame History

/*
* Copyright (c) 1982, 1986, 1989, 1991, 1993
* The Regents of the University of California. All rights reserved.
*
* 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 University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University 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 REGENTS 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 REGENTS 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.
*
* From: @(#)uipc_usrreq.c 8.3 (Berkeley) 1/4/94
* $FreeBSD$
*/
#include "opt_mac.h"
#include <sys/param.h>
#include <sys/domain.h>
#include <sys/fcntl.h>
#include <sys/malloc.h> /* XXX must be before <sys/file.h> */
#include <sys/file.h>
#include <sys/filedesc.h>
#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/mac.h>
#include <sys/mbuf.h>
#include <sys/mutex.h>
#include <sys/namei.h>
#include <sys/proc.h>
#include <sys/protosw.h>
#include <sys/resourcevar.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/signalvar.h>
#include <sys/stat.h>
#include <sys/sx.h>
#include <sys/sysctl.h>
#include <sys/systm.h>
#include <sys/un.h>
#include <sys/unpcb.h>
#include <sys/vnode.h>
#include <vm/uma.h>
static uma_zone_t unp_zone;
static unp_gen_t unp_gencnt;
static u_int unp_count;
static struct unp_head unp_shead, unp_dhead;
/*
* Unix communications domain.
*
* TODO:
* SEQPACKET, RDM
* rethink name space problems
* need a proper out-of-band
* lock pushdown
*/
static struct sockaddr sun_noname = { sizeof(sun_noname), AF_LOCAL };
static ino_t unp_ino; /* prototype for fake inode numbers */
static int unp_attach(struct socket *);
static void unp_detach(struct unpcb *);
static int unp_bind(struct unpcb *,struct sockaddr *, struct thread *);
static int unp_connect(struct socket *,struct sockaddr *, struct thread *);
static void unp_disconnect(struct unpcb *);
static void unp_shutdown(struct unpcb *);
static void unp_drop(struct unpcb *, int);
static void unp_gc(void);
static void unp_scan(struct mbuf *, void (*)(struct file *));
static void unp_mark(struct file *);
static void unp_discard(struct file *);
static void unp_freerights(struct file **, int);
static int unp_internalize(struct mbuf **, struct thread *);
static int unp_listen(struct unpcb *, struct thread *);
static int
uipc_abort(struct socket *so)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
unp_drop(unp, ECONNABORTED);
unp_detach(unp);
sotryfree(so);
return 0;
}
static int
uipc_accept(struct socket *so, struct sockaddr **nam)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
/*
* Pass back name of connected socket,
* if it was bound and we are still connected
* (our peer may have closed already!).
*/
if (unp->unp_conn && unp->unp_conn->unp_addr) {
*nam = dup_sockaddr((struct sockaddr *)unp->unp_conn->unp_addr,
1);
} else {
*nam = dup_sockaddr((struct sockaddr *)&sun_noname, 1);
}
return 0;
}
static int
uipc_attach(struct socket *so, int proto, struct thread *td)
{
struct unpcb *unp = sotounpcb(so);
if (unp != 0)
return EISCONN;
return unp_attach(so);
}
static int
uipc_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
return unp_bind(unp, nam, td);
}
static int
uipc_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
return unp_connect(so, nam, curthread);
}
static int
uipc_connect2(struct socket *so1, struct socket *so2)
{
struct unpcb *unp = sotounpcb(so1);
if (unp == 0)
return EINVAL;
return unp_connect2(so1, so2);
}
/* control is EOPNOTSUPP */
static int
uipc_detach(struct socket *so)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
unp_detach(unp);
return 0;
}
static int
uipc_disconnect(struct socket *so)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
unp_disconnect(unp);
return 0;
}
static int
uipc_listen(struct socket *so, struct thread *td)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0 || unp->unp_vnode == 0)
return EINVAL;
return unp_listen(unp, td);
}
static int
uipc_peeraddr(struct socket *so, struct sockaddr **nam)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
if (unp->unp_conn && unp->unp_conn->unp_addr)
*nam = dup_sockaddr((struct sockaddr *)unp->unp_conn->unp_addr,
1);
return 0;
}
static int
uipc_rcvd(struct socket *so, int flags)
{
struct unpcb *unp = sotounpcb(so);
struct socket *so2;
u_long newhiwat;
if (unp == 0)
return EINVAL;
switch (so->so_type) {
case SOCK_DGRAM:
panic("uipc_rcvd DGRAM?");
/*NOTREACHED*/
case SOCK_STREAM:
if (unp->unp_conn == 0)
break;
so2 = unp->unp_conn->unp_socket;
/*
* Adjust backpressure on sender
* and wakeup any waiting to write.
*/
so2->so_snd.sb_mbmax += unp->unp_mbcnt - so->so_rcv.sb_mbcnt;
unp->unp_mbcnt = so->so_rcv.sb_mbcnt;
newhiwat = so2->so_snd.sb_hiwat + unp->unp_cc -
so->so_rcv.sb_cc;
(void)chgsbsize(so2->so_cred->cr_uidinfo, &so2->so_snd.sb_hiwat,
newhiwat, RLIM_INFINITY);
unp->unp_cc = so->so_rcv.sb_cc;
sowwakeup(so2);
break;
default:
panic("uipc_rcvd unknown socktype");
}
return 0;
}
/* pru_rcvoob is EOPNOTSUPP */
static int
uipc_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
struct mbuf *control, struct thread *td)
{
int error = 0;
struct unpcb *unp = sotounpcb(so);
struct socket *so2;
u_long newhiwat;
if (unp == 0) {
error = EINVAL;
goto release;
}
if (flags & PRUS_OOB) {
error = EOPNOTSUPP;
goto release;
}
if (control && (error = unp_internalize(&control, td)))
goto release;
switch (so->so_type) {
case SOCK_DGRAM:
{
struct sockaddr *from;
if (nam) {
if (unp->unp_conn) {
error = EISCONN;
break;
}
error = unp_connect(so, nam, td);
if (error)
break;
} else {
if (unp->unp_conn == 0) {
error = ENOTCONN;
break;
}
}
so2 = unp->unp_conn->unp_socket;
if (unp->unp_addr)
from = (struct sockaddr *)unp->unp_addr;
else
from = &sun_noname;
if (sbappendaddr(&so2->so_rcv, from, m, control)) {
sorwakeup(so2);
m = 0;
control = 0;
} else
error = ENOBUFS;
if (nam)
unp_disconnect(unp);
break;
}
case SOCK_STREAM:
/* Connect if not connected yet. */
/*
* Note: A better implementation would complain
* if not equal to the peer's address.
*/
if ((so->so_state & SS_ISCONNECTED) == 0) {
if (nam) {
error = unp_connect(so, nam, td);
if (error)
break; /* XXX */
} else {
error = ENOTCONN;
break;
}
}
if (so->so_state & SS_CANTSENDMORE) {
error = EPIPE;
break;
}
if (unp->unp_conn == 0)
panic("uipc_send connected but no connection?");
so2 = unp->unp_conn->unp_socket;
/*
* Send to paired receive port, and then reduce
* send buffer hiwater marks to maintain backpressure.
* Wake up readers.
*/
if (control) {
if (sbappendcontrol(&so2->so_rcv, m, control))
control = 0;
} else
sbappend(&so2->so_rcv, m);
so->so_snd.sb_mbmax -=
so2->so_rcv.sb_mbcnt - unp->unp_conn->unp_mbcnt;
unp->unp_conn->unp_mbcnt = so2->so_rcv.sb_mbcnt;
newhiwat = so->so_snd.sb_hiwat -
(so2->so_rcv.sb_cc - unp->unp_conn->unp_cc);
(void)chgsbsize(so->so_cred->cr_uidinfo, &so->so_snd.sb_hiwat,
newhiwat, RLIM_INFINITY);
unp->unp_conn->unp_cc = so2->so_rcv.sb_cc;
sorwakeup(so2);
m = 0;
break;
default:
panic("uipc_send unknown socktype");
}
/*
* SEND_EOF is equivalent to a SEND followed by
* a SHUTDOWN.
*/
if (flags & PRUS_EOF) {
socantsendmore(so);
unp_shutdown(unp);
}
if (control && error != 0)
unp_dispose(control);
release:
if (control)
m_freem(control);
if (m)
m_freem(m);
return error;
}
static int
uipc_sense(struct socket *so, struct stat *sb)
{
struct unpcb *unp = sotounpcb(so);
struct socket *so2;
if (unp == 0)
return EINVAL;
sb->st_blksize = so->so_snd.sb_hiwat;
if (so->so_type == SOCK_STREAM && unp->unp_conn != 0) {
so2 = unp->unp_conn->unp_socket;
sb->st_blksize += so2->so_rcv.sb_cc;
}
sb->st_dev = NOUDEV;
if (unp->unp_ino == 0)
unp->unp_ino = unp_ino++;
sb->st_ino = unp->unp_ino;
return (0);
}
static int
uipc_shutdown(struct socket *so)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
socantsendmore(so);
unp_shutdown(unp);
return 0;
}
static int
uipc_sockaddr(struct socket *so, struct sockaddr **nam)
{
struct unpcb *unp = sotounpcb(so);
if (unp == 0)
return EINVAL;
if (unp->unp_addr)
*nam = dup_sockaddr((struct sockaddr *)unp->unp_addr, 1);
else
*nam = dup_sockaddr((struct sockaddr *)&sun_noname, 1);
return 0;
}
struct pr_usrreqs uipc_usrreqs = {
uipc_abort, uipc_accept, uipc_attach, uipc_bind, uipc_connect,
uipc_connect2, pru_control_notsupp, uipc_detach, uipc_disconnect,
uipc_listen, uipc_peeraddr, uipc_rcvd, pru_rcvoob_notsupp,
uipc_send, uipc_sense, uipc_shutdown, uipc_sockaddr,
sosend, soreceive, sopoll
};
int
uipc_ctloutput(so, sopt)
struct socket *so;
struct sockopt *sopt;
{
struct unpcb *unp = sotounpcb(so);
int error;
switch (sopt->sopt_dir) {
case SOPT_GET:
switch (sopt->sopt_name) {
case LOCAL_PEERCRED:
if (unp->unp_flags & UNP_HAVEPC)
error = sooptcopyout(sopt, &unp->unp_peercred,
sizeof(unp->unp_peercred));
else {
if (so->so_type == SOCK_STREAM)
error = ENOTCONN;
else
error = EINVAL;
}
break;
default:
error = EOPNOTSUPP;
break;
}
break;
case SOPT_SET:
default:
error = EOPNOTSUPP;
break;
}
return (error);
}
/*
* Both send and receive buffers are allocated PIPSIZ bytes of buffering
* for stream sockets, although the total for sender and receiver is
* actually only PIPSIZ.
* Datagram sockets really use the sendspace as the maximum datagram size,
* and don't really want to reserve the sendspace. Their recvspace should
* be large enough for at least one max-size datagram plus address.
*/
#ifndef PIPSIZ
#define PIPSIZ 8192
#endif
static u_long unpst_sendspace = PIPSIZ;
static u_long unpst_recvspace = PIPSIZ;
static u_long unpdg_sendspace = 2*1024; /* really max datagram size */
static u_long unpdg_recvspace = 4*1024;
static int unp_rights; /* file descriptors in flight */
SYSCTL_DECL(_net_local_stream);
SYSCTL_INT(_net_local_stream, OID_AUTO, sendspace, CTLFLAG_RW,
&unpst_sendspace, 0, "");
SYSCTL_INT(_net_local_stream, OID_AUTO, recvspace, CTLFLAG_RW,
&unpst_recvspace, 0, "");
SYSCTL_DECL(_net_local_dgram);
SYSCTL_INT(_net_local_dgram, OID_AUTO, maxdgram, CTLFLAG_RW,
&unpdg_sendspace, 0, "");
SYSCTL_INT(_net_local_dgram, OID_AUTO, recvspace, CTLFLAG_RW,
&unpdg_recvspace, 0, "");
SYSCTL_DECL(_net_local);
SYSCTL_INT(_net_local, OID_AUTO, inflight, CTLFLAG_RD, &unp_rights, 0, "");
static int
unp_attach(so)
struct socket *so;
{
register struct unpcb *unp;
int error;
if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
switch (so->so_type) {
case SOCK_STREAM:
error = soreserve(so, unpst_sendspace, unpst_recvspace);
break;
case SOCK_DGRAM:
error = soreserve(so, unpdg_sendspace, unpdg_recvspace);
break;
default:
panic("unp_attach");
}
if (error)
return (error);
}
unp = uma_zalloc(unp_zone, M_WAITOK);
if (unp == NULL)
return (ENOBUFS);
bzero(unp, sizeof *unp);
unp->unp_gencnt = ++unp_gencnt;
unp_count++;
LIST_INIT(&unp->unp_refs);
unp->unp_socket = so;
FILEDESC_LOCK(curproc->p_fd);
unp->unp_rvnode = curthread->td_proc->p_fd->fd_rdir;
FILEDESC_UNLOCK(curproc->p_fd);
LIST_INSERT_HEAD(so->so_type == SOCK_DGRAM ? &unp_dhead
: &unp_shead, unp, unp_link);
so->so_pcb = unp;
return (0);
}
static void
unp_detach(unp)
register struct unpcb *unp;
{
LIST_REMOVE(unp, unp_link);
unp->unp_gencnt = ++unp_gencnt;
--unp_count;
if (unp->unp_vnode) {
unp->unp_vnode->v_socket = 0;
vrele(unp->unp_vnode);
unp->unp_vnode = 0;
}
if (unp->unp_conn)
unp_disconnect(unp);
while (!LIST_EMPTY(&unp->unp_refs))
unp_drop(LIST_FIRST(&unp->unp_refs), ECONNRESET);
soisdisconnected(unp->unp_socket);
unp->unp_socket->so_pcb = 0;
if (unp_rights) {
/*
* Normally the receive buffer is flushed later,
* in sofree, but if our receive buffer holds references
* to descriptors that are now garbage, we will dispose
* of those descriptor references after the garbage collector
* gets them (resulting in a "panic: closef: count < 0").
*/
sorflush(unp->unp_socket);
unp_gc();
}
if (unp->unp_addr)
FREE(unp->unp_addr, M_SONAME);
uma_zfree(unp_zone, unp);
}
static int
unp_bind(unp, nam, td)
struct unpcb *unp;
struct sockaddr *nam;
struct thread *td;
{
struct sockaddr_un *soun = (struct sockaddr_un *)nam;
struct vnode *vp;
struct mount *mp;
struct vattr vattr;
int error, namelen;
struct nameidata nd;
char *buf;
if (unp->unp_vnode != NULL)
return (EINVAL);
namelen = soun->sun_len - offsetof(struct sockaddr_un, sun_path);
if (namelen <= 0)
return EINVAL;
buf = malloc(SOCK_MAXADDRLEN, M_TEMP, M_WAITOK);
strncpy(buf, soun->sun_path, namelen);
buf[namelen] = 0; /* null-terminate the string */
restart:
NDINIT(&nd, CREATE, NOFOLLOW | LOCKPARENT | SAVENAME, UIO_SYSSPACE,
buf, td);
/* SHOULD BE ABLE TO ADOPT EXISTING AND wakeup() ALA FIFO's */
error = namei(&nd);
if (error) {
free(buf, M_TEMP);
return (error);
}
vp = nd.ni_vp;
if (vp != NULL || vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
NDFREE(&nd, NDF_ONLY_PNBUF);
if (nd.ni_dvp == vp)
vrele(nd.ni_dvp);
else
vput(nd.ni_dvp);
if (vp != NULL) {
vrele(vp);
free(buf, M_TEMP);
return (EADDRINUSE);
}
error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH);
if (error) {
free(buf, M_TEMP);
return (error);
}
goto restart;
}
VATTR_NULL(&vattr);
vattr.va_type = VSOCK;
FILEDESC_LOCK(td->td_proc->p_fd);
vattr.va_mode = (ACCESSPERMS & ~td->td_proc->p_fd->fd_cmask);
FILEDESC_UNLOCK(td->td_proc->p_fd);
#ifdef MAC
error = mac_check_vnode_create(td->td_ucred, nd.ni_dvp, &nd.ni_cnd,
&vattr);
#endif /* MAC */
if (error == 0) {
VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
error = VOP_CREATE(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr);
}
NDFREE(&nd, NDF_ONLY_PNBUF);
vput(nd.ni_dvp);
if (error) {
free(buf, M_TEMP);
return (error);
}
vp = nd.ni_vp;
vp->v_socket = unp->unp_socket;
unp->unp_vnode = vp;
unp->unp_addr = (struct sockaddr_un *)dup_sockaddr(nam, 1);
VOP_UNLOCK(vp, 0, td);
vn_finished_write(mp);
free(buf, M_TEMP);
return (0);
}
static int
unp_connect(so, nam, td)
struct socket *so;
struct sockaddr *nam;
struct thread *td;
{
register struct sockaddr_un *soun = (struct sockaddr_un *)nam;
register struct vnode *vp;
register struct socket *so2, *so3;
struct unpcb *unp, *unp2, *unp3;
int error, len;
struct nameidata nd;
char buf[SOCK_MAXADDRLEN];
len = nam->sa_len - offsetof(struct sockaddr_un, sun_path);
if (len <= 0)
return EINVAL;
strncpy(buf, soun->sun_path, len);
buf[len] = 0;
NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, buf, td);
error = namei(&nd);
if (error)
return (error);
vp = nd.ni_vp;
NDFREE(&nd, NDF_ONLY_PNBUF);
if (vp->v_type != VSOCK) {
error = ENOTSOCK;
goto bad;
}
error = VOP_ACCESS(vp, VWRITE, td->td_ucred, td);
if (error)
goto bad;
so2 = vp->v_socket;
if (so2 == 0) {
error = ECONNREFUSED;
goto bad;
}
if (so->so_type != so2->so_type) {
error = EPROTOTYPE;
goto bad;
}
if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
if ((so2->so_options & SO_ACCEPTCONN) == 0 ||
(so3 = sonewconn(so2, 0)) == 0) {
error = ECONNREFUSED;
goto bad;
}
unp = sotounpcb(so);
unp2 = sotounpcb(so2);
unp3 = sotounpcb(so3);
if (unp2->unp_addr)
unp3->unp_addr = (struct sockaddr_un *)
dup_sockaddr((struct sockaddr *)
unp2->unp_addr, 1);
/*
* unp_peercred management:
*
* The connecter's (client's) credentials are copied
* from its process structure at the time of connect()
* (which is now).
*/
cru2x(td->td_ucred, &unp3->unp_peercred);
unp3->unp_flags |= UNP_HAVEPC;
/*
* The receiver's (server's) credentials are copied
* from the unp_peercred member of socket on which the
* former called listen(); unp_listen() cached that
* process's credentials at that time so we can use
* them now.
*/
KASSERT(unp2->unp_flags & UNP_HAVEPCCACHED,
("unp_connect: listener without cached peercred"));
memcpy(&unp->unp_peercred, &unp2->unp_peercred,
sizeof(unp->unp_peercred));
unp->unp_flags |= UNP_HAVEPC;
#ifdef MAC
mac_set_socket_peer_from_socket(so, so3);
mac_set_socket_peer_from_socket(so3, so);
#endif
so2 = so3;
}
error = unp_connect2(so, so2);
bad:
vput(vp);
return (error);
}
int
unp_connect2(so, so2)
register struct socket *so;
register struct socket *so2;
{
register struct unpcb *unp = sotounpcb(so);
register struct unpcb *unp2;
if (so2->so_type != so->so_type)
return (EPROTOTYPE);
unp2 = sotounpcb(so2);
unp->unp_conn = unp2;
switch (so->so_type) {
case SOCK_DGRAM:
LIST_INSERT_HEAD(&unp2->unp_refs, unp, unp_reflink);
soisconnected(so);
break;
case SOCK_STREAM:
unp2->unp_conn = unp;
soisconnected(so);
soisconnected(so2);
break;
default:
panic("unp_connect2");
}
return (0);
}
static void
unp_disconnect(unp)
struct unpcb *unp;
{
register struct unpcb *unp2 = unp->unp_conn;
if (unp2 == 0)
return;
unp->unp_conn = 0;
switch (unp->unp_socket->so_type) {
case SOCK_DGRAM:
LIST_REMOVE(unp, unp_reflink);
unp->unp_socket->so_state &= ~SS_ISCONNECTED;
break;
case SOCK_STREAM:
soisdisconnected(unp->unp_socket);
unp2->unp_conn = 0;
soisdisconnected(unp2->unp_socket);
break;
}
}
#ifdef notdef
void
unp_abort(unp)
struct unpcb *unp;
{
unp_detach(unp);
}
#endif
static int
unp_pcblist(SYSCTL_HANDLER_ARGS)
{
int error, i, n;
struct unpcb *unp, **unp_list;
unp_gen_t gencnt;
struct xunpgen *xug;
struct unp_head *head;
struct xunpcb *xu;
head = ((intptr_t)arg1 == SOCK_DGRAM ? &unp_dhead : &unp_shead);
/*
* The process of preparing the PCB list is too time-consuming and
* resource-intensive to repeat twice on every request.
*/
if (req->oldptr == 0) {
n = unp_count;
req->oldidx = 2 * (sizeof *xug)
+ (n + n/8) * sizeof(struct xunpcb);
return 0;
}
if (req->newptr != 0)
return EPERM;
/*
* OK, now we're committed to doing something.
*/
xug = malloc(sizeof(*xug), M_TEMP, M_WAITOK);
gencnt = unp_gencnt;
n = unp_count;
xug->xug_len = sizeof *xug;
xug->xug_count = n;
xug->xug_gen = gencnt;
xug->xug_sogen = so_gencnt;
error = SYSCTL_OUT(req, xug, sizeof *xug);
if (error) {
free(xug, M_TEMP);
return error;
}
unp_list = malloc(n * sizeof *unp_list, M_TEMP, M_WAITOK);
for (unp = LIST_FIRST(head), i = 0; unp && i < n;
unp = LIST_NEXT(unp, unp_link)) {
if (unp->unp_gencnt <= gencnt) {
if (cr_cansee(req->td->td_ucred,
unp->unp_socket->so_cred))
continue;
unp_list[i++] = unp;
}
}
n = i; /* in case we lost some during malloc */
error = 0;
xu = malloc(sizeof(*xu), M_TEMP, M_WAITOK);
for (i = 0; i < n; i++) {
unp = unp_list[i];
if (unp->unp_gencnt <= gencnt) {
xu->xu_len = sizeof *xu;
xu->xu_unpp = unp;
/*
* XXX - need more locking here to protect against
* connect/disconnect races for SMP.
*/
if (unp->unp_addr)
bcopy(unp->unp_addr, &xu->xu_addr,
unp->unp_addr->sun_len);
if (unp->unp_conn && unp->unp_conn->unp_addr)
bcopy(unp->unp_conn->unp_addr,
&xu->xu_caddr,
unp->unp_conn->unp_addr->sun_len);
bcopy(unp, &xu->xu_unp, sizeof *unp);
sotoxsocket(unp->unp_socket, &xu->xu_socket);
error = SYSCTL_OUT(req, xu, sizeof *xu);
}
}
free(xu, M_TEMP);
if (!error) {
/*
* Give the user an updated idea of our state.
* If the generation differs from what we told
* her before, she knows that something happened
* while we were processing this request, and it
* might be necessary to retry.
*/
xug->xug_gen = unp_gencnt;
xug->xug_sogen = so_gencnt;
xug->xug_count = unp_count;
error = SYSCTL_OUT(req, xug, sizeof *xug);
}
free(unp_list, M_TEMP);
free(xug, M_TEMP);
return error;
}
SYSCTL_PROC(_net_local_dgram, OID_AUTO, pcblist, CTLFLAG_RD,
(caddr_t)(long)SOCK_DGRAM, 0, unp_pcblist, "S,xunpcb",
"List of active local datagram sockets");
SYSCTL_PROC(_net_local_stream, OID_AUTO, pcblist, CTLFLAG_RD,
(caddr_t)(long)SOCK_STREAM, 0, unp_pcblist, "S,xunpcb",
"List of active local stream sockets");
static void
unp_shutdown(unp)
struct unpcb *unp;
{
struct socket *so;
if (unp->unp_socket->so_type == SOCK_STREAM && unp->unp_conn &&
(so = unp->unp_conn->unp_socket))
socantrcvmore(so);
}
static void
unp_drop(unp, errno)
struct unpcb *unp;
int errno;
{
struct socket *so = unp->unp_socket;
so->so_error = errno;
unp_disconnect(unp);
}
#ifdef notdef
void
unp_drain()
{
}
#endif
static void
unp_freerights(rp, fdcount)
struct file **rp;
int fdcount;
{
int i;
struct file *fp;
for (i = 0; i < fdcount; i++) {
fp = *rp;
/*
* zero the pointer before calling
* unp_discard since it may end up
* in unp_gc()..
*/
*rp++ = 0;
unp_discard(fp);
}
}
int
unp_externalize(control, controlp)
struct mbuf *control, **controlp;
{
struct thread *td = curthread; /* XXX */
struct cmsghdr *cm = mtod(control, struct cmsghdr *);
int i;
int *fdp;
struct file **rp;
struct file *fp;
void *data;
socklen_t clen = control->m_len, datalen;
int error, newfds;
int f;
u_int newlen;
error = 0;
if (controlp != NULL) /* controlp == NULL => free control messages */
*controlp = NULL;
while (cm != NULL) {
if (sizeof(*cm) > clen || cm->cmsg_len > clen) {
error = EINVAL;
break;
}
data = CMSG_DATA(cm);
datalen = (caddr_t)cm + cm->cmsg_len - (caddr_t)data;
if (cm->cmsg_level == SOL_SOCKET
&& cm->cmsg_type == SCM_RIGHTS) {
newfds = datalen / sizeof(struct file *);
rp = data;
/* If we're not outputting the discriptors free them. */
if (error || controlp == NULL) {
unp_freerights(rp, newfds);
goto next;
}
FILEDESC_LOCK(td->td_proc->p_fd);
/* if the new FD's will not fit free them. */
if (!fdavail(td, newfds)) {
FILEDESC_UNLOCK(td->td_proc->p_fd);
error = EMSGSIZE;
unp_freerights(rp, newfds);
goto next;
}
/*
* now change each pointer to an fd in the global
* table to an integer that is the index to the
* local fd table entry that we set up to point
* to the global one we are transferring.
*/
newlen = newfds * sizeof(int);
*controlp = sbcreatecontrol(NULL, newlen,
SCM_RIGHTS, SOL_SOCKET);
if (*controlp == NULL) {
FILEDESC_UNLOCK(td->td_proc->p_fd);
error = E2BIG;
unp_freerights(rp, newfds);
goto next;
}
fdp = (int *)
CMSG_DATA(mtod(*controlp, struct cmsghdr *));
for (i = 0; i < newfds; i++) {
if (fdalloc(td, 0, &f))
panic("unp_externalize fdalloc failed");
fp = *rp++;
td->td_proc->p_fd->fd_ofiles[f] = fp;
FILE_LOCK(fp);
fp->f_msgcount--;
FILE_UNLOCK(fp);
unp_rights--;
*fdp++ = f;
}
FILEDESC_UNLOCK(td->td_proc->p_fd);
} else { /* We can just copy anything else across */
if (error || controlp == NULL)
goto next;
*controlp = sbcreatecontrol(NULL, datalen,
cm->cmsg_type, cm->cmsg_level);
if (*controlp == NULL) {
error = ENOBUFS;
goto next;
}
bcopy(data,
CMSG_DATA(mtod(*controlp, struct cmsghdr *)),
datalen);
}
controlp = &(*controlp)->m_next;
next:
if (CMSG_SPACE(datalen) < clen) {
clen -= CMSG_SPACE(datalen);
cm = (struct cmsghdr *)
((caddr_t)cm + CMSG_SPACE(datalen));
} else {
clen = 0;
cm = NULL;
}
}
m_freem(control);
return (error);
}
void
unp_init(void)
{
unp_zone = uma_zcreate("unpcb", sizeof(struct unpcb), NULL, NULL,
NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
uma_zone_set_max(unp_zone, nmbclusters);
if (unp_zone == 0)
panic("unp_init");
LIST_INIT(&unp_dhead);
LIST_INIT(&unp_shead);
}
#ifndef MIN
#define MIN(a,b) (((a)<(b))?(a):(b))
#endif
static int
unp_internalize(controlp, td)
struct mbuf **controlp;
struct thread *td;
{
struct mbuf *control = *controlp;
struct proc *p = td->td_proc;
struct filedesc *fdescp = p->p_fd;
struct cmsghdr *cm = mtod(control, struct cmsghdr *);
struct cmsgcred *cmcred;
struct file **rp;
struct file *fp;
struct timeval *tv;
int i, fd, *fdp;
void *data;
socklen_t clen = control->m_len, datalen;
int error, oldfds;
u_int newlen;
error = 0;
*controlp = NULL;
while (cm != NULL) {
if (sizeof(*cm) > clen || cm->cmsg_level != SOL_SOCKET
|| cm->cmsg_len > clen) {
error = EINVAL;
goto out;
}
data = CMSG_DATA(cm);
datalen = (caddr_t)cm + cm->cmsg_len - (caddr_t)data;
switch (cm->cmsg_type) {
/*
* Fill in credential information.
*/
case SCM_CREDS:
*controlp = sbcreatecontrol(NULL, sizeof(*cmcred),
SCM_CREDS, SOL_SOCKET);
if (*controlp == NULL) {
error = ENOBUFS;
goto out;
}
cmcred = (struct cmsgcred *)
CMSG_DATA(mtod(*controlp, struct cmsghdr *));
cmcred->cmcred_pid = p->p_pid;
cmcred->cmcred_uid = td->td_ucred->cr_ruid;
cmcred->cmcred_gid = td->td_ucred->cr_rgid;
cmcred->cmcred_euid = td->td_ucred->cr_uid;
cmcred->cmcred_ngroups = MIN(td->td_ucred->cr_ngroups,
CMGROUP_MAX);
for (i = 0; i < cmcred->cmcred_ngroups; i++)
cmcred->cmcred_groups[i] =
td->td_ucred->cr_groups[i];
break;
case SCM_RIGHTS:
oldfds = datalen / sizeof (int);
/*
* check that all the FDs passed in refer to legal files
* If not, reject the entire operation.
*/
fdp = data;
FILEDESC_LOCK(fdescp);
for (i = 0; i < oldfds; i++) {
fd = *fdp++;
if ((unsigned)fd >= fdescp->fd_nfiles ||
fdescp->fd_ofiles[fd] == NULL) {
FILEDESC_UNLOCK(fdescp);
error = EBADF;
goto out;
}
}
/*
* Now replace the integer FDs with pointers to
* the associated global file table entry..
*/
newlen = oldfds * sizeof(struct file *);
*controlp = sbcreatecontrol(NULL, newlen,
SCM_RIGHTS, SOL_SOCKET);
if (*controlp == NULL) {
FILEDESC_UNLOCK(fdescp);
error = E2BIG;
goto out;
}
fdp = data;
rp = (struct file **)
CMSG_DATA(mtod(*controlp, struct cmsghdr *));
for (i = 0; i < oldfds; i++) {
fp = fdescp->fd_ofiles[*fdp++];
*rp++ = fp;
FILE_LOCK(fp);
fp->f_count++;
fp->f_msgcount++;
FILE_UNLOCK(fp);
unp_rights++;
}
FILEDESC_UNLOCK(fdescp);
break;
case SCM_TIMESTAMP:
*controlp = sbcreatecontrol(NULL, sizeof(*tv),
SCM_TIMESTAMP, SOL_SOCKET);
if (*controlp == NULL) {
error = ENOBUFS;
goto out;
}
tv = (struct timeval *)
CMSG_DATA(mtod(*controlp, struct cmsghdr *));
microtime(tv);
break;
default:
error = EINVAL;
goto out;
}
controlp = &(*controlp)->m_next;
if (CMSG_SPACE(datalen) < clen) {
clen -= CMSG_SPACE(datalen);
cm = (struct cmsghdr *)
((caddr_t)cm + CMSG_SPACE(datalen));
} else {
clen = 0;
cm = NULL;
}
}
out:
m_freem(control);
return (error);
}
static int unp_defer, unp_gcing;
static void
unp_gc()
{
register struct file *fp, *nextfp;
register struct socket *so;
struct file **extra_ref, **fpp;
int nunref, i;
if (unp_gcing)
return;
unp_gcing = 1;
unp_defer = 0;
/*
* before going through all this, set all FDs to
* be NOT defered and NOT externally accessible
*/
sx_slock(&filelist_lock);
LIST_FOREACH(fp, &filehead, f_list)
fp->f_gcflag &= ~(FMARK|FDEFER);
do {
LIST_FOREACH(fp, &filehead, f_list) {
FILE_LOCK(fp);
/*
* If the file is not open, skip it
*/
if (fp->f_count == 0) {
FILE_UNLOCK(fp);
continue;
}
/*
* If we already marked it as 'defer' in a
* previous pass, then try process it this time
* and un-mark it
*/
if (fp->f_gcflag & FDEFER) {
fp->f_gcflag &= ~FDEFER;
unp_defer--;
} else {
/*
* if it's not defered, then check if it's
* already marked.. if so skip it
*/
if (fp->f_gcflag & FMARK) {
FILE_UNLOCK(fp);
continue;
}
/*
* If all references are from messages
* in transit, then skip it. it's not
* externally accessible.
*/
if (fp->f_count == fp->f_msgcount) {
FILE_UNLOCK(fp);
continue;
}
/*
* If it got this far then it must be
* externally accessible.
*/
fp->f_gcflag |= FMARK;
}
/*
* either it was defered, or it is externally
* accessible and not already marked so.
* Now check if it is possibly one of OUR sockets.
*/
if (fp->f_type != DTYPE_SOCKET ||
(so = (struct socket *)fp->f_data) == 0) {
FILE_UNLOCK(fp);
continue;
}
FILE_UNLOCK(fp);
if (so->so_proto->pr_domain != &localdomain ||
(so->so_proto->pr_flags&PR_RIGHTS) == 0)
continue;
#ifdef notdef
if (so->so_rcv.sb_flags & SB_LOCK) {
/*
* This is problematical; it's not clear
* we need to wait for the sockbuf to be
* unlocked (on a uniprocessor, at least),
* and it's also not clear what to do
* if sbwait returns an error due to receipt
* of a signal. If sbwait does return
* an error, we'll go into an infinite
* loop. Delete all of this for now.
*/
(void) sbwait(&so->so_rcv);
goto restart;
}
#endif
/*
* So, Ok, it's one of our sockets and it IS externally
* accessible (or was defered). Now we look
* to see if we hold any file descriptors in its
* message buffers. Follow those links and mark them
* as accessible too.
*/
unp_scan(so->so_rcv.sb_mb, unp_mark);
}
} while (unp_defer);
sx_sunlock(&filelist_lock);
/*
* We grab an extra reference to each of the file table entries
* that are not otherwise accessible and then free the rights
* that are stored in messages on them.
*
* The bug in the orginal code is a little tricky, so I'll describe
* what's wrong with it here.
*
* It is incorrect to simply unp_discard each entry for f_msgcount
* times -- consider the case of sockets A and B that contain
* references to each other. On a last close of some other socket,
* we trigger a gc since the number of outstanding rights (unp_rights)
* is non-zero. If during the sweep phase the gc code un_discards,
* we end up doing a (full) closef on the descriptor. A closef on A
* results in the following chain. Closef calls soo_close, which
* calls soclose. Soclose calls first (through the switch
* uipc_usrreq) unp_detach, which re-invokes unp_gc. Unp_gc simply
* returns because the previous instance had set unp_gcing, and
* we return all the way back to soclose, which marks the socket
* with SS_NOFDREF, and then calls sofree. Sofree calls sorflush
* to free up the rights that are queued in messages on the socket A,
* i.e., the reference on B. The sorflush calls via the dom_dispose
* switch unp_dispose, which unp_scans with unp_discard. This second
* instance of unp_discard just calls closef on B.
*
* Well, a similar chain occurs on B, resulting in a sorflush on B,
* which results in another closef on A. Unfortunately, A is already
* being closed, and the descriptor has already been marked with
* SS_NOFDREF, and soclose panics at this point.
*
* Here, we first take an extra reference to each inaccessible
* descriptor. Then, we call sorflush ourself, since we know
* it is a Unix domain socket anyhow. After we destroy all the
* rights carried in messages, we do a last closef to get rid
* of our extra reference. This is the last close, and the
* unp_detach etc will shut down the socket.
*
* 91/09/19, bsy@cs.cmu.edu
*/
extra_ref = malloc(nfiles * sizeof(struct file *), M_TEMP, M_WAITOK);
sx_slock(&filelist_lock);
for (nunref = 0, fp = LIST_FIRST(&filehead), fpp = extra_ref; fp != 0;
fp = nextfp) {
nextfp = LIST_NEXT(fp, f_list);
FILE_LOCK(fp);
/*
* If it's not open, skip it
*/
if (fp->f_count == 0) {
FILE_UNLOCK(fp);
continue;
}
/*
* If all refs are from msgs, and it's not marked accessible
* then it must be referenced from some unreachable cycle
* of (shut-down) FDs, so include it in our
* list of FDs to remove
*/
if (fp->f_count == fp->f_msgcount && !(fp->f_gcflag & FMARK)) {
*fpp++ = fp;
nunref++;
fp->f_count++;
}
FILE_UNLOCK(fp);
}
sx_sunlock(&filelist_lock);
/*
* for each FD on our hit list, do the following two things
*/
for (i = nunref, fpp = extra_ref; --i >= 0; ++fpp) {
struct file *tfp = *fpp;
FILE_LOCK(tfp);
if (tfp->f_type == DTYPE_SOCKET && tfp->f_data != NULL) {
FILE_UNLOCK(tfp);
sorflush((struct socket *)(tfp->f_data));
} else
FILE_UNLOCK(tfp);
}
for (i = nunref, fpp = extra_ref; --i >= 0; ++fpp)
closef(*fpp, (struct thread *) NULL);
free(extra_ref, M_TEMP);
unp_gcing = 0;
}
void
unp_dispose(m)
struct mbuf *m;
{
if (m)
unp_scan(m, unp_discard);
}
static int
unp_listen(unp, td)
struct unpcb *unp;
struct thread *td;
{
cru2x(td->td_ucred, &unp->unp_peercred);
unp->unp_flags |= UNP_HAVEPCCACHED;
return (0);
}
static void
unp_scan(m0, op)
register struct mbuf *m0;
void (*op)(struct file *);
{
struct mbuf *m;
struct file **rp;
struct cmsghdr *cm;
void *data;
int i;
socklen_t clen, datalen;
int qfds;
while (m0) {
for (m = m0; m; m = m->m_next) {
if (m->m_type != MT_CONTROL)
continue;
cm = mtod(m, struct cmsghdr *);
clen = m->m_len;
while (cm != NULL) {
if (sizeof(*cm) > clen || cm->cmsg_len > clen)
break;
data = CMSG_DATA(cm);
datalen = (caddr_t)cm + cm->cmsg_len
- (caddr_t)data;
if (cm->cmsg_level == SOL_SOCKET &&
cm->cmsg_type == SCM_RIGHTS) {
qfds = datalen / sizeof (struct file *);
rp = data;
for (i = 0; i < qfds; i++)
(*op)(*rp++);
}
if (CMSG_SPACE(datalen) < clen) {
clen -= CMSG_SPACE(datalen);
cm = (struct cmsghdr *)
((caddr_t)cm + CMSG_SPACE(datalen));
} else {
clen = 0;
cm = NULL;
}
}
}
m0 = m0->m_act;
}
}
static void
unp_mark(fp)
struct file *fp;
{
if (fp->f_gcflag & FMARK)
return;
unp_defer++;
fp->f_gcflag |= (FMARK|FDEFER);
}
static void
unp_discard(fp)
struct file *fp;
{
FILE_LOCK(fp);
fp->f_msgcount--;
unp_rights--;
FILE_UNLOCK(fp);
(void) closef(fp, (struct thread *)NULL);
}
Reference in New Issue View Git Blame Copy Permalink