d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
6ff44ffc53
freebsd-dev/usr.sbin/rpcbind/util.c

377 lines
10 KiB
C
Raw Normal View History

Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
/*
* $NetBSD: util.c,v 1.4 2000/08/03 00:04:30 fvdl Exp $
* $FreeBSD$
*/
/*-
* Copyright (c) 2000 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Frank van der Linden.
*
* 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.
*
* 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.
*/
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/queue.h>
#include <net/if.h>
#include <netinet/in.h>
#include <ifaddrs.h>
#include <sys/poll.h>
#include <rpc/rpc.h>
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <netdb.h>
#include <netconfig.h>
#include <stdio.h>
#include <arpa/inet.h>
#include "rpcbind.h"
static struct sockaddr_in *local_in4;
#ifdef INET6
static struct sockaddr_in6 *local_in6;
#endif
Cleanup of userland __P use
2007-11-07 10:53:41 +00:00
static int bitmaskcmp(void *, void *, void *, int);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
#ifdef INET6
Cleanup of userland __P use
2007-11-07 10:53:41 +00:00
static void in6_fillscopeid(struct sockaddr_in6 *);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
#endif
/*
* For all bits set in "mask", compare the corresponding bits in
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
* "dst" and "src", and see if they match. Returns 0 if the addresses
* match.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
static int
bitmaskcmp(void *dst, void *src, void *mask, int bytelen)
{
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
int i;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
u_int8_t *p1 = dst, *p2 = src, *netmask = mask;
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
for (i = 0; i < bytelen; i++)
if ((p1[i] & netmask[i]) != (p2[i] & netmask[i]))
return (1);
return (0);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
}
/*
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
* Similar to code in ifconfig.c. Fill in the scope ID for link-local
* addresses returned by getifaddrs().
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
#ifdef INET6
static void
in6_fillscopeid(struct sockaddr_in6 *sin6)
{
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
u_int16_t ifindex;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
if (IN6_IS_ADDR_LINKLOCAL(&sin6->sin6_addr)) {
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
ifindex = ntohs(*(u_int16_t *)&sin6->sin6_addr.s6_addr[2]);
if (sin6->sin6_scope_id == 0 && ifindex != 0) {
sin6->sin6_scope_id = ifindex;
*(u_int16_t *)&sin6->sin6_addr.s6_addr[2] = 0;
}
}
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
}
#endif
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
/*
* Find a server address that can be used by `caller' to contact
* the local service specified by `serv_uaddr'. If `clnt_uaddr' is
* non-NULL, it is used instead of `caller' as a hint suggesting
* the best address (e.g. the `r_addr' field of an rpc, which
* contains the rpcbind server address that the caller used).
*
* Returns the best server address as a malloc'd "universal address"
* string which should be freed by the caller. On error, returns NULL.
*/
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
char *
addrmerge(struct netbuf *caller, char *serv_uaddr, char *clnt_uaddr,
char *netid)
{
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
struct ifaddrs *ifap, *ifp = NULL, *bestif;
struct netbuf *serv_nbp = NULL, *hint_nbp = NULL, tbuf;
struct sockaddr *caller_sa, *hint_sa, *ifsa, *ifmasksa, *serv_sa;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
struct sockaddr_storage ss;
struct netconfig *nconf;
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
char *caller_uaddr = NULL, *hint_uaddr = NULL;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
char *ret = NULL;
#ifdef ND_DEBUG
if (debugging)
fprintf(stderr, "addrmerge(caller, %s, %s, %s\n", serv_uaddr,
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
clnt_uaddr == NULL ? "NULL" : clnt_uaddr, netid);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
#endif
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
caller_sa = caller->buf;
if ((nconf = rpcbind_get_conf(netid)) == NULL)
goto freeit;
if ((caller_uaddr = taddr2uaddr(nconf, caller)) == NULL)
goto freeit;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
/*
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
* Use `clnt_uaddr' as the hint if non-NULL, but ignore it if its
* address family is different from that of the caller.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
hint_sa = NULL;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
if (clnt_uaddr != NULL) {
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
hint_uaddr = clnt_uaddr;
if ((hint_nbp = uaddr2taddr(nconf, clnt_uaddr)) == NULL)
goto freeit;
hint_sa = hint_nbp->buf;
}
if (hint_sa == NULL || hint_sa->sa_family != caller_sa->sa_family) {
hint_uaddr = caller_uaddr;
hint_sa = caller->buf;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
}
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
#ifdef ND_DEBUG
if (debugging)
fprintf(stderr, "addrmerge: hint %s\n", hint_uaddr);
#endif
/* Local caller, just return the server address. */
if (strncmp(caller_uaddr, "0.0.0.0.", 8) == 0 ||
strncmp(caller_uaddr, "::.", 3) == 0 || caller_uaddr[0] == '/') {
ret = strdup(serv_uaddr);
goto freeit;
Bring in some bugfixes from NetBSD. I'm going to make a more extensive pass through the rpcbind code soon, but I might as well bring these in now. - (NetBSD util.c r1.5) Move the initialisation of `tbuf' to avoid a case where it could end up containing junk from the stack. This should address the issue in PR bin/26806. - (NetBSD util.c r1.6) Don't `merge' AF_LOCAL addresses, fix a few memory leaks. PR: bin/26806 Submitted by: Martin Blapp <mb@imp.ch> Obtained from: NetBSD
2001-04-26 17:05:55 +00:00
}
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
if (getifaddrs(&ifp) < 0)
goto freeit;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
/*
When you have multiple addresses on the same network on different interfaces (such as when you are part of a carp pool), and you run rpcbind -h to restrict which interfaces have rpc services, rpcbind can none-the-less return addresses that aren't in the -h list. This patch enforces the rule that when you specify -h on the command line, then services returned from rpcbind must be to one of the addresses listed in -h, or be a loopback address (since localhost is implicit when running -h). The root cause of this is the assumption in addrmerge that there can be only one interface that matches a given network IP address. This turns out not to be the case. To retain historical behavior, I didn't try to fix the routine to prefer the address that the request came into, since I didn't know the side effects that might cause in the normal case. My quick analysis suggests that it wouldn't be a problem, but since this code is tricky I opted for the more conservative patch of only restricting the reply when -h is in effect. Hence, this change will have no effect when you are running rpcbind without -h. Reviewed by: alfred@ Sponsored by: iX Systems MFC after: 2 weeks
2010-02-09 18:10:56 +00:00
* Loop through all interfaces. For each interface, see if it
* is either the loopback interface (which we always listen
* on) or is one of the addresses the program bound to (the
* wildcard by default, or a subset if -h is specified) and
* the network portion of its address is equal to that of the
* client. If so, we have found the interface that we want to
* use.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
bestif = NULL;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
for (ifap = ifp; ifap != NULL; ifap = ifap->ifa_next) {
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
ifsa = ifap->ifa_addr;
ifmasksa = ifap->ifa_netmask;
if (ifsa == NULL || ifsa->sa_family != hint_sa->sa_family ||
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
!(ifap->ifa_flags & IFF_UP))
continue;
When you have multiple addresses on the same network on different interfaces (such as when you are part of a carp pool), and you run rpcbind -h to restrict which interfaces have rpc services, rpcbind can none-the-less return addresses that aren't in the -h list. This patch enforces the rule that when you specify -h on the command line, then services returned from rpcbind must be to one of the addresses listed in -h, or be a loopback address (since localhost is implicit when running -h). The root cause of this is the assumption in addrmerge that there can be only one interface that matches a given network IP address. This turns out not to be the case. To retain historical behavior, I didn't try to fix the routine to prefer the address that the request came into, since I didn't know the side effects that might cause in the normal case. My quick analysis suggests that it wouldn't be a problem, but since this code is tricky I opted for the more conservative patch of only restricting the reply when -h is in effect. Hence, this change will have no effect when you are running rpcbind without -h. Reviewed by: alfred@ Sponsored by: iX Systems MFC after: 2 weeks
2010-02-09 18:10:56 +00:00
if (!(ifap->ifa_flags & IFF_LOOPBACK) && !listen_addr(ifsa))
continue;
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
switch (hint_sa->sa_family) {
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
case AF_INET:
/*
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
* If the hint address matches this interface
* address/netmask, then we're done.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
if (!bitmaskcmp(&SA2SINADDR(ifsa),
&SA2SINADDR(hint_sa), &SA2SINADDR(ifmasksa),
sizeof(struct in_addr))) {
bestif = ifap;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
goto found;
}
break;
#ifdef INET6
case AF_INET6:
/*
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
* For v6 link local addresses, if the caller is on
* a link-local address then use the scope id to see
* which one.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
in6_fillscopeid(SA2SIN6(ifsa));
if (IN6_IS_ADDR_LINKLOCAL(&SA2SIN6ADDR(ifsa)) &&
IN6_IS_ADDR_LINKLOCAL(&SA2SIN6ADDR(caller_sa)) &&
IN6_IS_ADDR_LINKLOCAL(&SA2SIN6ADDR(hint_sa))) {
if (SA2SIN6(ifsa)->sin6_scope_id ==
SA2SIN6(caller_sa)->sin6_scope_id) {
bestif = ifap;
goto found;
}
} else if (!bitmaskcmp(&SA2SIN6ADDR(ifsa),
&SA2SIN6ADDR(hint_sa), &SA2SIN6ADDR(ifmasksa),
sizeof(struct in6_addr))) {
bestif = ifap;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
goto found;
}
break;
#endif
default:
continue;
}
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
/*
* Remember the first possibly useful interface, preferring
* "normal" to point-to-point and loopback ones.
*/
if (bestif == NULL ||
(!(ifap->ifa_flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) &&
(bestif->ifa_flags & (IFF_LOOPBACK | IFF_POINTOPOINT))))
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
bestif = ifap;
}
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
if (bestif == NULL)
goto freeit;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
found:
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
/*
Fix typos - remove duplicate "the". PR: bin/154928 Submitted by: Eitan Adler <lists at eitanadler.com> MFC after: 3 days
2011-02-21 09:01:34 +00:00
* Construct the new address using the address from
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
* `bestif', and the port number from `serv_uaddr'.
*/
serv_nbp = uaddr2taddr(nconf, serv_uaddr);
if (serv_nbp == NULL)
goto freeit;
serv_sa = serv_nbp->buf;
memcpy(&ss, bestif->ifa_addr, bestif->ifa_addr->sa_len);
switch (ss.ss_family) {
Bring in some bugfixes from NetBSD. I'm going to make a more extensive pass through the rpcbind code soon, but I might as well bring these in now. - (NetBSD util.c r1.5) Move the initialisation of `tbuf' to avoid a case where it could end up containing junk from the stack. This should address the issue in PR bin/26806. - (NetBSD util.c r1.6) Don't `merge' AF_LOCAL addresses, fix a few memory leaks. PR: bin/26806 Submitted by: Martin Blapp <mb@imp.ch> Obtained from: NetBSD
2001-04-26 17:05:55 +00:00
case AF_INET:
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
SA2SIN(&ss)->sin_port = SA2SIN(serv_sa)->sin_port;
break;
Bring in some bugfixes from NetBSD. I'm going to make a more extensive pass through the rpcbind code soon, but I might as well bring these in now. - (NetBSD util.c r1.5) Move the initialisation of `tbuf' to avoid a case where it could end up containing junk from the stack. This should address the issue in PR bin/26806. - (NetBSD util.c r1.6) Don't `merge' AF_LOCAL addresses, fix a few memory leaks. PR: bin/26806 Submitted by: Martin Blapp <mb@imp.ch> Obtained from: NetBSD
2001-04-26 17:05:55 +00:00
#ifdef INET6
case AF_INET6:
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
SA2SIN6(&ss)->sin6_port = SA2SIN6(serv_sa)->sin6_port;
Bring in some bugfixes from NetBSD. I'm going to make a more extensive pass through the rpcbind code soon, but I might as well bring these in now. - (NetBSD util.c r1.5) Move the initialisation of `tbuf' to avoid a case where it could end up containing junk from the stack. This should address the issue in PR bin/26806. - (NetBSD util.c r1.6) Don't `merge' AF_LOCAL addresses, fix a few memory leaks. PR: bin/26806 Submitted by: Martin Blapp <mb@imp.ch> Obtained from: NetBSD
2001-04-26 17:05:55 +00:00
break;
#endif
}
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
tbuf.len = ss.ss_len;
tbuf.maxlen = sizeof(ss);
tbuf.buf = &ss;
ret = taddr2uaddr(nconf, &tbuf);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
freeit:
Clean up the addrmerge() function, which was over-complicated and contained a number of memory leaks. The changes include: - Add a comment describing what addrmerge() does. - Deal with 0.0.0.0./::. or AF_LOCAL callers correctly. - Use rpcbind_get_conf() instead of getnetconfigent() so we don't have to remember to free the returned netconfig struct. - Make just one pass through the ifaddrs list; we can pick up a fallback interface address in the same pass as the netmask comparison. - Define and use SA2SIN* macros to avoid the need for loads of protocol-specific local variables. - Use mostly protocol-independent code for building the netbuf version of the address to be returned. - Use the common cleanup code for virtually all error and non-error cases, fixing a number of memory leaks.
2001-06-24 15:03:06 +00:00
if (caller_uaddr != NULL)
free(caller_uaddr);
if (hint_nbp != NULL) {
free(hint_nbp->buf);
free(hint_nbp);
}
if (serv_nbp != NULL) {
free(serv_nbp->buf);
free(serv_nbp);
}
if (ifp != NULL)
freeifaddrs(ifp);
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
#ifdef ND_DEBUG
if (debugging)
fprintf(stderr, "addrmerge: returning %s\n", ret);
#endif
return ret;
}
void
Use prototype. While I'm there, add a pair of parenthesis to mark an if statment's border. MFC after: 1 month
2011-07-14 07:28:49 +00:00
network_init(void)
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
{
#ifdef INET6
struct ifaddrs *ifap, *ifp;
struct ipv6_mreq mreq6;
WARNS=3 safety (mostly), use __unused for unused params and unsigned where needed to avoid warnings about comparing signed and unsigned values.
2002-10-07 02:56:59 +00:00
unsigned int ifindex;
int s;
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
#endif
int ecode;
struct addrinfo hints, *res;
memset(&hints, 0, sizeof hints);
hints.ai_family = AF_INET;
if ((ecode = getaddrinfo(NULL, "sunrpc", &hints, &res))) {
if (debugging)
fprintf(stderr, "can't get local ip4 address: %s\n",
gai_strerror(ecode));
} else {
local_in4 = (struct sockaddr_in *)malloc(sizeof *local_in4);
if (local_in4 == NULL) {
if (debugging)
fprintf(stderr, "can't alloc local ip4 addr\n");
}
memcpy(local_in4, res->ai_addr, sizeof *local_in4);
}
#ifdef INET6
hints.ai_family = AF_INET6;
if ((ecode = getaddrinfo(NULL, "sunrpc", &hints, &res))) {
if (debugging)
fprintf(stderr, "can't get local ip6 address: %s\n",
gai_strerror(ecode));
} else {
local_in6 = (struct sockaddr_in6 *)malloc(sizeof *local_in6);
if (local_in6 == NULL) {
if (debugging)
fprintf(stderr, "can't alloc local ip6 addr\n");
}
memcpy(local_in6, res->ai_addr, sizeof *local_in6);
}
/*
* Now join the RPC ipv6 multicast group on all interfaces.
*/
if (getifaddrs(&ifp) < 0)
return;
mreq6.ipv6mr_interface = 0;
inet_pton(AF_INET6, RPCB_MULTICAST_ADDR, &mreq6.ipv6mr_multiaddr);
s = socket(AF_INET6, SOCK_DGRAM, IPPROTO_UDP);
/*
Simplify to bitmaskcmp() to use the obvious approach instead of comparing bit by bit. Make the logic in in6_fillscopeid() match that in our ifconfig(8): only set the scope ID if there is one in the address and none in sin6_scope_id. Correct a comment in network_init() that didn't make sense; it was probably never updated after it was pasted from similar code in addrmerge().
2001-07-14 17:04:26 +00:00
* Loop through all interfaces. For each IPv6 multicast-capable
* interface, join the RPC multicast group on that interface.
Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and associated changes that had to happen to make this possible as well as bugs fixed along the way. Bring in required TLI library routines to support this. Since we don't support TLI we've essentially copied what NetBSD has done, adding a thin layer to emulate direct the TLI calls into BSD socket calls. This is mostly from Sun's tirpc release that was made in 1994, however some fixes were backported from the 1999 release (supposedly only made available after this porting effort was underway). The submitter has agreed to continue on and bring us up to the 1999 release. Several key features are introduced with this update: Client calls are thread safe. (1999 code has server side thread safe) Updated, a more modern interface. Many userland updates were done to bring the code up to par with the recent RPC API. There is an update to the pthreads library, a function pthread_main_np() was added to emulate a function of Sun's threads library. While we're at it, bring in NetBSD's lockd, it's been far too long of a wait. New rpcbind(8) replaces portmap(8) (supporting communication over an authenticated Unix-domain socket, and by default only allowing set and unset requests over that channel). It's much more secure than the old portmapper. Umount(8), mountd(8), mount_nfs(8), nfsd(8) have also been upgraded to support TI-RPC and to support IPV6. Umount(8) is also fixed to unmount pathnames longer than 80 chars, which are currently truncated by the Kernel statfs structure. Submitted by: Martin Blapp <mb@imp.ch> Manpage review: ru Secure RPC implemented by: wpaul
2001-03-19 12:50:13 +00:00
*/
for (ifap = ifp; ifap != NULL; ifap = ifap->ifa_next) {
if (ifap->ifa_addr->sa_family != AF_INET6 ||
!(ifap->ifa_flags & IFF_MULTICAST))
continue;
ifindex = if_nametoindex(ifap->ifa_name);
if (ifindex == mreq6.ipv6mr_interface)
/*
* Already did this one.
*/
continue;
mreq6.ipv6mr_interface = ifindex;
if (setsockopt(s, IPPROTO_IPV6, IPV6_JOIN_GROUP, &mreq6,
sizeof mreq6) < 0)
if (debugging)
perror("setsockopt v6 multicast");
}
#endif
/* close(s); */
}
struct sockaddr *
local_sa(int af)
{
switch (af) {
case AF_INET:
return (struct sockaddr *)local_in4;
#ifdef INET6
case AF_INET6:
return (struct sockaddr *)local_in6;
#endif
default:
return NULL;
}
}
Reference in New Issue Copy Permalink