d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
c89f86ce8f
freebsd-nq/lib/libc/yp/yplib.c

1034 lines
23 KiB
C
Raw Normal View History

Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
/*
* Copyright (c) 1992/3 Theo de Raadt <deraadt@fsa.ca>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
*/
#ifndef LINT
Don't even think about processing bogus domain names here.
1997-04-10 20:26:04 +00:00
static char *rcsid = "$Id: yplib.c,v 1.28 1997/02/22 15:05:02 peter Exp $";
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#endif
#include <sys/param.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/file.h>
#include <sys/uio.h>
#include <errno.h>
#include <stdio.h>
#include <string.h>
More of the same: we also need to handle multiple domains properly even if /var/yp/binding/DOMAIN.VERS doesn't exist yet.
1995-04-09 21:52:31 +00:00
#include <stdlib.h>
#include <unistd.h>
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#include <rpc/rpc.h>
#include <rpc/xdr.h>
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
#include <rpcsvc/yp.h>
/*
* We have to define these here due to clashes between yp_prot.h and
* yp.h.
*/
struct dom_binding {
struct dom_binding *dom_pnext;
char dom_domain[YPMAXDOMAIN + 1];
struct sockaddr_in dom_server_addr;
u_short dom_server_port;
int dom_socket;
CLIENT *dom_client;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
u_short dom_local_port; /* now I finally know what this is for. */
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
long dom_vers;
};
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#include <rpcsvc/ypclnt.h>
#ifndef BINDINGDIR
#define BINDINGDIR "/var/yp/binding"
#endif
#define YPMATCHCACHE
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
#define MAX_RETRIES 20
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
extern bool_t xdr_domainname(), xdr_ypbind_resp();
extern bool_t xdr_ypreq_key(), xdr_ypresp_val();
extern bool_t xdr_ypreq_nokey(), xdr_ypresp_key_val();
extern bool_t xdr_ypresp_all(), xdr_ypresp_all_seq();
extern bool_t xdr_ypresp_master();
int (*ypresp_allfn)();
void *ypresp_data;
Fix for bug noticed by Christoph Kukulies. _yp_dobind() checks to see if a fork() happens (by checking PIDs) and invalidates all bindings if it finds itself in a newly created child process. (This avoids sharing RPC client handles and socket descriptors with the parent, which would be bad.) Unfortunately, it summarily calls clnt_destroy() on the handles, which may result in the destruction of a descriptor that isn't really a socket. This is fixed by replacing the explicit call to clnt_destroy() with a call to _yp_unbind(), which deals with potentially hosed socket descriptors an a safe manner. This is basically a one-liner. Once I confirm that it fixes Christoph's problem, I'd like permission to put it in the 2.2-RELENG branch.
1996-12-20 19:41:15 +00:00
static void _yp_unbind __P(( struct dom_binding * ));
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
struct dom_binding *_ypbindlist;
static char _yp_domain[MAXHOSTNAMELEN];
int _yplib_timeout = 10;
#ifdef YPMATCHCACHE
int _yplib_cache = 5;
static struct ypmatch_ent {
struct ypmatch_ent *next;
char *map, *key, *val;
int keylen, vallen;
time_t expire_t;
} *ypmc;
static void
ypmatch_add(map, key, keylen, val, vallen)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *map;
char *key;
int keylen;
char *val;
int vallen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct ypmatch_ent *ep;
time_t t;
time(&t);
for(ep=ypmc; ep; ep=ep->next)
if(ep->expire_t < t)
break;
if(ep==NULL) {
ep = (struct ypmatch_ent *)malloc(sizeof *ep);
bzero((char *)ep, sizeof *ep);
if(ypmc)
ep->next = ypmc;
ypmc = ep;
}
if(ep->key)
free(ep->key);
if(ep->val)
free(ep->val);
ep->key = NULL;
ep->val = NULL;
ep->key = (char *)malloc(keylen);
if(ep->key==NULL)
return;
ep->val = (char *)malloc(vallen);
if(ep->key==NULL) {
free(ep->key);
ep->key = NULL;
return;
}
ep->keylen = keylen;
ep->vallen = vallen;
bcopy(key, ep->key, ep->keylen);
bcopy(val, ep->val, ep->vallen);
if(ep->map) {
if( strcmp(ep->map, map) ) {
free(ep->map);
ep->map = strdup(map);
}
} else {
ep->map = strdup(map);
}
ep->expire_t = t + _yplib_cache;
}
static bool_t
ypmatch_find(map, key, keylen, val, vallen)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *map;
char *key;
int keylen;
char **val;
int *vallen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct ypmatch_ent *ep;
time_t t;
if(ypmc==NULL)
return 0;
time(&t);
for(ep=ypmc; ep; ep=ep->next) {
if(ep->keylen != keylen)
continue;
if(strcmp(ep->map, map))
continue;
if(bcmp(ep->key, key, keylen))
continue;
if(t > ep->expire_t)
continue;
*val = ep->val;
*vallen = ep->vallen;
return 1;
}
return 0;
}
#endif
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
char *
ypbinderr_string(incode)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
int incode;
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
{
static char err[80];
switch(incode) {
case 0:
return "Success";
Cosmetic changes and paranoia checks: ypbind.c: Make fewer assumtions about the state of the dom_alive and dom_broadcasting flags in roc_received(). If select() fails, use syslog() to report the error rather than perror(). Check that all our malloc()s succeed. Report malloc() failure in ypbindproc_setdom_2() to callers. yplib.c: Use #defined constants in ypbinderr_string() rather than hard-coded values.
1995-05-03 18:33:10 +00:00
case YPBIND_ERR_ERR:
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
return "Internal ypbind error";
Cosmetic changes and paranoia checks: ypbind.c: Make fewer assumtions about the state of the dom_alive and dom_broadcasting flags in roc_received(). If select() fails, use syslog() to report the error rather than perror(). Check that all our malloc()s succeed. Report malloc() failure in ypbindproc_setdom_2() to callers. yplib.c: Use #defined constants in ypbinderr_string() rather than hard-coded values.
1995-05-03 18:33:10 +00:00
case YPBIND_ERR_NOSERV:
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
return "Domain not bound";
Cosmetic changes and paranoia checks: ypbind.c: Make fewer assumtions about the state of the dom_alive and dom_broadcasting flags in roc_received(). If select() fails, use syslog() to report the error rather than perror(). Check that all our malloc()s succeed. Report malloc() failure in ypbindproc_setdom_2() to callers. yplib.c: Use #defined constants in ypbinderr_string() rather than hard-coded values.
1995-05-03 18:33:10 +00:00
case YPBIND_ERR_RESC:
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
return "System resource allocation failure";
}
Cosmetic changes and paranoia checks: ypbind.c: Make fewer assumtions about the state of the dom_alive and dom_broadcasting flags in roc_received(). If select() fails, use syslog() to report the error rather than perror(). Check that all our malloc()s succeed. Report malloc() failure in ypbindproc_setdom_2() to callers. yplib.c: Use #defined constants in ypbinderr_string() rather than hard-coded values.
1995-05-03 18:33:10 +00:00
sprintf(err, "Unknown ypbind error: #%d\n", incode);
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
return err;
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
int
_yp_dobind(dom, ypdb)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *dom;
struct dom_binding **ypdb;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
Code clean up: Changed type of pid from int to pid_t.
1996-06-03 13:19:10 +00:00
static pid_t pid = -1;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
char path[MAXPATHLEN];
struct dom_binding *ysd, *ysd2;
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
struct ypbind_resp ypbr;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
struct timeval tv;
struct sockaddr_in clnt_sin;
General -Wall warning cleanup, part I. Submitted-By: Kent Vander Velden <graphix@iastate.edu>
1996-07-12 18:57:58 +00:00
int clnt_sock, fd;
Code clean up: Changed type of pid from int to pid_t. (Missed one!)
1996-06-04 17:35:15 +00:00
pid_t gpid;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
CLIENT *client;
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
int new = 0, r;
int retries = 0;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
struct sockaddr_in check;
int checklen = sizeof(struct sockaddr_in);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
Don't even think about processing bogus domain names here.
1997-04-10 20:26:04 +00:00
/* Not allowed; bad doggie. Bad. */
if (strchr(dom, '/') != NULL)
return(YPERR_BADARGS);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
gpid = getpid();
if( !(pid==-1 || pid==gpid) ) {
ysd = _ypbindlist;
while(ysd) {
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
if(ysd->dom_client != NULL)
Fix for bug noticed by Christoph Kukulies. _yp_dobind() checks to see if a fork() happens (by checking PIDs) and invalidates all bindings if it finds itself in a newly created child process. (This avoids sharing RPC client handles and socket descriptors with the parent, which would be bad.) Unfortunately, it summarily calls clnt_destroy() on the handles, which may result in the destruction of a descriptor that isn't really a socket. This is fixed by replacing the explicit call to clnt_destroy() with a call to _yp_unbind(), which deals with potentially hosed socket descriptors an a safe manner. This is basically a one-liner. Once I confirm that it fixes Christoph's problem, I'd like permission to put it in the 2.2-RELENG branch.
1996-12-20 19:41:15 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ysd2 = ysd->dom_pnext;
free(ysd);
ysd = ysd2;
}
_ypbindlist = NULL;
}
pid = gpid;
if(ypdb!=NULL)
*ypdb = NULL;
if(dom==NULL || strlen(dom)==0)
return YPERR_BADARGS;
for(ysd = _ypbindlist; ysd; ysd = ysd->dom_pnext)
if( strcmp(dom, ysd->dom_domain) == 0)
break;
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if(ysd==NULL) {
ysd = (struct dom_binding *)malloc(sizeof *ysd);
bzero((char *)ysd, sizeof *ysd);
ysd->dom_socket = -1;
ysd->dom_vers = 0;
new = 1;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
} else {
/* Check the socket -- may have been hosed by the caller. */
if (getsockname(ysd->dom_socket, (struct sockaddr *)&check,
&checklen) == -1 || check.sin_family != AF_INET ||
check.sin_port != ysd->dom_local_port) {
/* Socket became bogus somehow... need to rebind. */
int save, sock;
sock = ysd->dom_socket;
save = dup(ysd->dom_socket);
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
if (ysd->dom_client != NULL)
clnt_destroy(ysd->dom_client);
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
ysd->dom_vers = 0;
ysd->dom_client = NULL;
sock = dup2(save, sock);
close(save);
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
- Make _do_ypbind() check for /var/run/ypbind.lock and attempt to flock() it before before trying to establish a binding. If /var/run/ypbind.lock doesn't exist, or if it exists and isn't locked, then ypbind isn't running, which means NIS is either turned off or hosed. - Have _yp_check() call yp_unbind() after it sucessfully calls yp_bind() to make sure it frees resources correctly. (I don't think there's really a memory leak here, but it seems somehow wrong to call yp_bind() without making a corresponding call to yp_unbind() afterwards.) This makes the NIS code behave a little better in cases where libc makes calls to NIS, but it isn't running correctly (i.e. there's no ypbind). This cleans up some strange libc behavior that manifests itself if you have the system domain name set, but aren't actually running NIS. In this event, the getrpcent(3) code could try to call into NIS and cause several inexplicable "clnttcp_create error: RPC program not registered" messages to appear. This happens because _yp_check() checks if the system domain name is set and, if it is, proceeds to call yp_bind() to attempt to establish a binding. Since there is no binding file (remember: ypbind isn't running, so /var/yp/binding will be empty), _yp_dobind() will attempt to contact ypbind to prod it into binding the domain. And because ypbind isn't running, the code generates the 'clnttcp_create' error. Ultimately the _yp_check() fails and the getrpcent(3) code rolls over to the /etc/rpc file, but the error messages are annoying, and the code should be smart enough to forgo the binding attempt when NIS is turned off.
1995-11-05 05:39:04 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
again:
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
retries++;
if (retries > MAX_RETRIES) {
if (new)
free(ysd);
return(YPERR_YPBIND);
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#ifdef BINDINGDIR
if(ysd->dom_vers==0) {
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
/*
* We're trying to make a new binding: zorch the
* existing handle now (if any).
*/
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
if(ysd->dom_client != NULL) {
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
clnt_destroy(ysd->dom_client);
ysd->dom_client = NULL;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
ysd->dom_socket = -1;
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
sprintf(path, "%s/%s.%d", BINDINGDIR, dom, 2);
if( (fd=open(path, O_RDONLY)) == -1) {
/* no binding file, YP is dead. */
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
/* Try to bring it back to life. */
More of the same: we also need to handle multiple domains properly even if /var/yp/binding/DOMAIN.VERS doesn't exist yet.
1995-04-09 21:52:31 +00:00
close(fd);
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
goto skipit;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
if( flock(fd, LOCK_EX|LOCK_NB) == -1 && errno==EWOULDBLOCK) {
struct iovec iov[2];
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
struct ypbind_resp ybr;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
u_short ypb_port;
iov[0].iov_base = (caddr_t)&ypb_port;
iov[0].iov_len = sizeof ypb_port;
iov[1].iov_base = (caddr_t)&ybr;
iov[1].iov_len = sizeof ybr;
r = readv(fd, iov, 2);
if(r != iov[0].iov_len + iov[1].iov_len) {
close(fd);
ysd->dom_vers = -1;
goto again;
}
bzero(&ysd->dom_server_addr, sizeof ysd->dom_server_addr);
ysd->dom_server_addr.sin_family = AF_INET;
ysd->dom_server_addr.sin_len = sizeof(struct sockaddr_in);
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
ysd->dom_server_addr.sin_addr.s_addr =
*(u_long *)&ybr.ypbind_resp_u.ypbind_bindinfo.ypbind_binding_addr;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ysd->dom_server_addr.sin_port =
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
*(u_short *)&ybr.ypbind_resp_u.ypbind_bindinfo.ypbind_binding_port;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ysd->dom_server_port = ysd->dom_server_addr.sin_port;
close(fd);
goto gotit;
} else {
/* no lock on binding file, YP is dead. */
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
/* Try to bring it back to life. */
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
close(fd);
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
goto skipit;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
}
More of the same: we also need to handle multiple domains properly even if /var/yp/binding/DOMAIN.VERS doesn't exist yet.
1995-04-09 21:52:31 +00:00
skipit:
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#endif
if(ysd->dom_vers==-1 || ysd->dom_vers==0) {
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
/*
* We're trying to make a new binding: zorch the
* existing handle now (if any).
*/
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
if(ysd->dom_client != NULL) {
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
clnt_destroy(ysd->dom_client);
ysd->dom_client = NULL;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
ysd->dom_socket = -1;
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
bzero((char *)&clnt_sin, sizeof clnt_sin);
clnt_sin.sin_family = AF_INET;
clnt_sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
Remove trailing whitespace.
1995-05-30 05:51:47 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
clnt_sock = RPC_ANYSOCK;
client = clnttcp_create(&clnt_sin, YPBINDPROG, YPBINDVERS, &clnt_sock,
0, 0);
if(client==NULL) {
Don't bother trying to flock() /var/run/ypbind.lock; this breaks when /var/run resides on an NFS filesystem (flock() always returns 0 in this case, so we falsely assume that ypbind is dead and bail out). Settle instead for better failure checking when using clnttcp_create() and clnt_call() to interact with ypbind. We still try to flock() /var/yp/binding/$DOMAINNAME.2, but if this doesn't work, we drop into the code that retrieves the binding information from ypbind directly. If that also fails, then we're toast. On NFS filesystems, this means we'll be ignoring the binding file for no reason and always talking to ypbind even though we don't have to, but at least things will work. (I could just replace the flock(/var/run/ypbind.lock) check with an RPC call to ypbind's NULLPROC procedure, but if the flock() of the binding file doesn't pan out we're going to try to talk to ypbind later anyway. *sigh* Is NFS file locking ever going to work?)
1996-03-23 22:48:19 +00:00
/*
* These conditions indicate ypbind just isn't
* alive -- we probably don't want to shoot our
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
* mouth off in this case; instead generate error
Don't bother trying to flock() /var/run/ypbind.lock; this breaks when /var/run resides on an NFS filesystem (flock() always returns 0 in this case, so we falsely assume that ypbind is dead and bail out). Settle instead for better failure checking when using clnttcp_create() and clnt_call() to interact with ypbind. We still try to flock() /var/yp/binding/$DOMAINNAME.2, but if this doesn't work, we drop into the code that retrieves the binding information from ypbind directly. If that also fails, then we're toast. On NFS filesystems, this means we'll be ignoring the binding file for no reason and always talking to ypbind even though we don't have to, but at least things will work. (I could just replace the flock(/var/run/ypbind.lock) check with an RPC call to ypbind's NULLPROC procedure, but if the flock() of the binding file doesn't pan out we're going to try to talk to ypbind later anyway. *sigh* Is NFS file locking ever going to work?)
1996-03-23 22:48:19 +00:00
* messages only for really exotic problems.
*/
if (rpc_createerr.cf_stat != RPC_PROGNOTREGISTERED &&
(rpc_createerr.cf_stat != RPC_SYSTEMERROR &&
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
rpc_createerr.cf_error.re_errno == ECONNREFUSED))
clnt_pcreateerror("clnttcp_create");
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if(new)
free(ysd);
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
return (YPERR_YPBIND);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
In _yp_dobind(), if we find ourselves required to contact the local ypbind directly in order to obtain binding information, check that the local ypbind is using a reserved port and return YPERR_YPBIND if it isn't. We should not trust any ypbind running on a port >= IPPORT_RESERVED; it may have been started by a malicious user hoping to trick us into talking to a bogus ypserv. Note that we do not check the ypserv port returned to us from ypbind. It is assumed that ypbind has already done a reserved port test (or not, depending on whether or not it was started with -s); if we trust the authenticity of the local ypbind, we should also trust its judgement. Obtained from: OpenBSD
1996-07-13 20:23:13 +00:00
/*
* Check the port number -- should be < IPPORT_RESERVED.
* If not, it's possible someone has registered a bogus
* ypbind with the portmapper and is trying to trick us.
*/
if (ntohs(clnt_sin.sin_port) >= IPPORT_RESERVED) {
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
if (client != NULL)
clnt_destroy(client);
In _yp_dobind(), if we find ourselves required to contact the local ypbind directly in order to obtain binding information, check that the local ypbind is using a reserved port and return YPERR_YPBIND if it isn't. We should not trust any ypbind running on a port >= IPPORT_RESERVED; it may have been started by a malicious user hoping to trick us into talking to a bogus ypserv. Note that we do not check the ypserv port returned to us from ypbind. It is assumed that ypbind has already done a reserved port test (or not, depending on whether or not it was started with -s); if we trust the authenticity of the local ypbind, we should also trust its judgement. Obtained from: OpenBSD
1996-07-13 20:23:13 +00:00
if (new)
free(ysd);
return(YPERR_YPBIND);
}
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
tv.tv_sec = _yplib_timeout/2;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
tv.tv_usec = 0;
r = clnt_call(client, YPBINDPROC_DOMAIN,
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
xdr_domainname, (char *)&dom, xdr_ypbind_resp, &ypbr, tv);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if(r != RPC_SUCCESS) {
clnt_destroy(client);
ysd->dom_vers = -1;
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
if (r == RPC_PROGUNAVAIL || r == RPC_PROCUNAVAIL) {
if (new)
free(ysd);
Don't bother trying to flock() /var/run/ypbind.lock; this breaks when /var/run resides on an NFS filesystem (flock() always returns 0 in this case, so we falsely assume that ypbind is dead and bail out). Settle instead for better failure checking when using clnttcp_create() and clnt_call() to interact with ypbind. We still try to flock() /var/yp/binding/$DOMAINNAME.2, but if this doesn't work, we drop into the code that retrieves the binding information from ypbind directly. If that also fails, then we're toast. On NFS filesystems, this means we'll be ignoring the binding file for no reason and always talking to ypbind even though we don't have to, but at least things will work. (I could just replace the flock(/var/run/ypbind.lock) check with an RPC call to ypbind's NULLPROC procedure, but if the flock() of the binding file doesn't pan out we're going to try to talk to ypbind later anyway. *sigh* Is NFS file locking ever going to work?)
1996-03-23 22:48:19 +00:00
return(YPERR_YPBIND);
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
}
Don't bother trying to flock() /var/run/ypbind.lock; this breaks when /var/run resides on an NFS filesystem (flock() always returns 0 in this case, so we falsely assume that ypbind is dead and bail out). Settle instead for better failure checking when using clnttcp_create() and clnt_call() to interact with ypbind. We still try to flock() /var/yp/binding/$DOMAINNAME.2, but if this doesn't work, we drop into the code that retrieves the binding information from ypbind directly. If that also fails, then we're toast. On NFS filesystems, this means we'll be ignoring the binding file for no reason and always talking to ypbind even though we don't have to, but at least things will work. (I could just replace the flock(/var/run/ypbind.lock) check with an RPC call to ypbind's NULLPROC procedure, but if the flock() of the binding file doesn't pan out we're going to try to talk to ypbind later anyway. *sigh* Is NFS file locking ever going to work?)
1996-03-23 22:48:19 +00:00
fprintf(stderr,
"YP: server for domain %s not responding, retrying\n", dom);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
} else {
if (ypbr.ypbind_status != YPBIND_SUCC_VAL) {
ypbind.c: Major overhaul. - Moved to a more client-driven model. We aggressively attempt to keep the default domain bound (as before) but we give up on non-default domains if we lose contact with a server and fail to get a response after one round of broadcasting. This helps drastically reduce the amount of network bandwitdh that ypbind consumes: if a client references the secondary domain at some later point, this will prod ypbind into establishing a new binding anyway, so continuously broadcasting without need is pointless. Note that we still actively seek out a binding for our default domain even if no client program has queried us yet. I'm not exactly sure if this matches SunOS's behavior or not, but I decided to do it this way since we can get into all sorts of trouble if our default domain comes unbound. Even so, we're still much quieter than we used to be. - Removed a bunch of no-longer pertinent comments and a couple of chunks of #ifdef 0'ed code that no longer fit in to the new layout. - Theo deRaadt must have become frustrated with the callback mechanism in clnt_broadcast(), because he shamelessly stole the clnt_broadcast() code right out of the RPC library and hacked it up to suit his needs. (Comments and all! :) I can understand why: clnt_broadcast() blocks while awaiting replies. Changing this behavior requires surgery. However, you can work around this: fork the broadcast into a child process and relay the results back to the parent via a pipe. (Careful obervation has shown that the SunOS ypbind forks children for broadcasting too, though I can only guess what sort of interprocess communication it uses. pipe() seems to do the job well enough.) This may seem like the long way around, but it's not really that hard to implement, and I'd prefer to use documented RPC library functions wherever possible. We're careful to limit the number of simultaneous broadcasters to avoid swamping the system (the current limit is 5). Each clnt_broadcast() call only sends out a small number of packets at increasing intervals. We're also careful not to spawn more than one bradcaster for a given domain. - Used clntudp_bufcreate() and clnt_call() to implement a ping() function for directly querying a particular server so that we can check if it's still alive. This lets me completely remove the old bradcasting code and use actual RPC library calls instead, at the cost of more than a few handfulls of torn-out hair. (Make no mistake folks: I *HATE* RPC.) Currently, the ping interval is one minute. - Fixed another potential 'nfds too big for select()' bug: use _rpc_dtablesize() instead of getdtablesize(). - Quieted gcc -Wall a bit. - Probably a bunch of other stuff that I've forgotten. ypbind.8: - Updated man page to reflect modifications. ypwhich.c: - Small mind-o fix from last time: decode error results from ypbind correctly (*groan*) yplib.c: - same as above - Change behavior of _yp_dobind() a little: if we get back a 'Domain not bound' error for a given domain, retry a few times before giving up and passing the error back to the caller. We have to sleep for a few seconds between tries since the 'Domain not bound' error comes back immediately (by repeatedly looping, we end up pounding on ypbind). We retry at most 20 times at 5 second intervals. This gives us a full minute to get a response. This seems to deviate a bit from SunOS behavior -- it appears to wait forever -- but I don't like the idea of perpetually hanging inside a library call. Note that this should fix the problems some people have with bindings not being established fast enough at boot time; sometimes amd is started in /etc/rc after ypbind has run but before it gets a binding set up. The automounter gets annoyed at this and tends to exit. By pausing ther YP calls until a binding is ready, we avoid this situation. - Another _yp_dobind() change: if we determine that our binding files are unlocked or nonexistent, jump directly to code that pokes ypbind into restablishing the binding. Again, if it fails, we'll time out eventually and return.
1995-04-26 19:03:16 +00:00
clnt_destroy(client);
ysd->dom_vers = -1;
sleep(_yplib_timeout/2);
goto again;
small NIS binding fixes: ypbind.c: if a client program asks ypbind for the name of the server for a particular domain, and there isn't a binding for that domain available yet, ypbind needs to supply a status value along with its failure message. Set yprespbody.ypbind_error before returning from a ypbindproc_domain request. yplib.c: properly handle the error status messages ypbind now has the ability to send us. Add a ypbinderr_string() function to decode the error values. ypwhich.c: handle ypbind errors correctly: yperr_string() can't handle ypbind_status messages -- use ypbinderr_string instead.
1995-04-21 18:04:36 +00:00
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
clnt_destroy(client);
bzero((char *)&ysd->dom_server_addr, sizeof ysd->dom_server_addr);
ysd->dom_server_addr.sin_family = AF_INET;
ysd->dom_server_addr.sin_port =
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
*(u_short *)&ypbr.ypbind_resp_u.ypbind_bindinfo.ypbind_binding_port;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ysd->dom_server_addr.sin_addr.s_addr =
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
*(u_long *)&ypbr.ypbind_resp_u.ypbind_bindinfo.ypbind_binding_addr;
In _yp_dobind(), if we find ourselves required to contact the local ypbind directly in order to obtain binding information, check that the local ypbind is using a reserved port and return YPERR_YPBIND if it isn't. We should not trust any ypbind running on a port >= IPPORT_RESERVED; it may have been started by a malicious user hoping to trick us into talking to a bogus ypserv. Note that we do not check the ypserv port returned to us from ypbind. It is assumed that ypbind has already done a reserved port test (or not, depending on whether or not it was started with -s); if we trust the authenticity of the local ypbind, we should also trust its judgement. Obtained from: OpenBSD
1996-07-13 20:23:13 +00:00
/*
* We could do a reserved port check here too, but this
* could pose compatibility problems. The local ypbind is
* supposed to decide whether or not to trust yp servers
* on insecure ports. For now, we trust its judgement.
*/
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ysd->dom_server_port =
Clean up a few things left dangling after the last commit. The new XDR routines auto-generated by rpcgen don't quite match the format of the original ones even though tey have the same names (that was one of the things wrong with the old XDR routines).
1995-12-15 03:26:40 +00:00
*(u_short *)&ypbr.ypbind_resp_u.ypbind_bindinfo.ypbind_binding_port;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
gotit:
ysd->dom_vers = YPVERS;
strcpy(ysd->dom_domain, dom);
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
/* Don't rebuild the connection to the server unless we have to. */
if (ysd->dom_client == NULL) {
tv.tv_sec = _yplib_timeout/2;
tv.tv_usec = 0;
ysd->dom_socket = RPC_ANYSOCK;
ysd->dom_client = clntudp_create(&ysd->dom_server_addr,
YPPROG, YPVERS, tv, &ysd->dom_socket);
if(ysd->dom_client==NULL) {
clnt_pcreateerror("clntudp_create");
ysd->dom_vers = -1;
goto again;
}
if( fcntl(ysd->dom_socket, F_SETFD, 1) == -1)
perror("fcntl: F_SETFD");
Make _yp_dobind() a litle smarter: Now that we preserve RPC handles instead of rebuilding them each time a ypcln function is called, we have to be careful about keeping our sockets in a sane state. It's possible that the caller may call a ypclnt function, and then decide to close all its file descriptors. This would also close the socket descriptor held by the yplib code. Worse, it could re-open the same descriptor number for its own use. If it then calls another ypclnt function, the subsequent RPC will fail because the socket will either be gone or replaced with Something Completely Different. The yplib code will recover by rebinding, but it doing so it may wreck the descriptor which now belongs to the caller. To fix this, _yp_dobind() needs to label the descriptor somehow so that it can test it later to make sure it hasn't been altered between ypclnt calls. It does this by binding the socket, thus associating a port number with it. It then saves this port number in the dom_local_port member of the dom_binding structure for the given domain. When _yp_dobind() is called again (which it is at the start of each ypclnt function), it checks to see if the domain is already bound, and if it is, it does a getsockname() on the socket and compares the port number to the one it saved. If the getsockname() fails, or the port number doesn't match, it abandons the socket and sets up a new client handle. This still incurs some syscall overhead, which is what I was trying to avoid, but it's still not as bad as before.
1996-06-01 05:08:31 +00:00
/*
* We want a port number associated with this socket
* so that we can check its authenticity later.
*/
checklen = sizeof(struct sockaddr_in);
bzero((char *)&check, checklen);
bind(ysd->dom_socket, (struct sockaddr *)&check, checklen);
check.sin_family = AF_INET;
if (!getsockname(ysd->dom_socket,
(struct sockaddr *)&check, &checklen)) {
ysd->dom_local_port = check.sin_port;
} else {
clnt_destroy(ysd->dom_client);
if (new)
free(ysd);
return(YPERR_YPBIND);
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
if(new) {
ysd->dom_pnext = _ypbindlist;
_ypbindlist = ysd;
}
if(ypdb!=NULL)
*ypdb = ysd;
return 0;
}
static void
_yp_unbind(ypb)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
struct dom_binding *ypb;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
struct sockaddr_in check;
int checklen = sizeof(struct sockaddr_in);
if (ypb->dom_client != NULL) {
/* Check the socket -- may have been hosed by the caller. */
if (getsockname(ypb->dom_socket, (struct sockaddr *)&check,
&checklen) == -1 || check.sin_family != AF_INET ||
check.sin_port != ypb->dom_local_port) {
int save, sock;
sock = ypb->dom_socket;
save = dup(ypb->dom_socket);
clnt_destroy(ypb->dom_client);
sock = dup2(save, sock);
close(save);
} else
clnt_destroy(ypb->dom_client);
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
ypb->dom_client = NULL;
ypb->dom_socket = -1;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
ypb->dom_vers = -1;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
int
yp_bind(dom)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *dom;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
return _yp_dobind(dom, NULL);
}
void
yp_unbind(dom)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *dom;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct dom_binding *ypb, *ypbp;
ypbp = NULL;
for(ypb=_ypbindlist; ypb; ypb=ypb->dom_pnext) {
if( strcmp(dom, ypb->dom_domain) == 0) {
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ypb);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if(ypbp)
ypbp->dom_pnext = ypb->dom_pnext;
else
_ypbindlist = ypb->dom_pnext;
free(ypb);
return;
}
ypbp = ypb;
}
return;
}
int
yp_match(indomain, inmap, inkey, inkeylen, outval, outvallen)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
const char *inkey;
int inkeylen;
char **outval;
int *outvallen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct dom_binding *ysd;
struct ypresp_val yprv;
struct timeval tv;
struct ypreq_key yprk;
int r;
*outval = NULL;
*outvallen = 0;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
Don't let yp_match() or yp_next() operate on null or empty keys: asking ypserv to do a yp_match() with an a null or empty key causes much havok. (Note that this could be construed as a denial of service attack if used maliciously.)
1995-03-23 22:23:15 +00:00
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
if (inkey == NULL || !strlen(inkey) || inkeylen <= 0 ||
inmap == NULL || !strlen(inmap) ||
indomain == NULL || !strlen(indomain))
return YPERR_BADARGS;
Don't let yp_match() or yp_next() operate on null or empty keys: asking ypserv to do a yp_match() with an a null or empty key causes much havok. (Note that this could be construed as a denial of service attack if used maliciously.)
1995-03-23 22:23:15 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
#ifdef YPMATCHCACHE
if( !strcmp(_yp_domain, indomain) && ypmatch_find(inmap, inkey,
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
inkeylen, &yprv.val.valdat_val, &yprv.val.valdat_len)) {
*outvallen = yprv.val.valdat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outval = (char *)malloc(*outvallen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprv.val.valdat_val, *outval, *outvallen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outval)[*outvallen] = '\0';
return 0;
}
#endif
Clear up a minor bogosity in yp_match(): we have YPMATCHCACHE turned on, which is fine, except that _yp_dobind() is called before we check the cache. The means we can return from the cache check (if we have a hit) without calling _yp_unbind(). We should do the cache check first and _then_ drop into the section that binds the server and does the yp_match query.
1995-09-02 04:16:21 +00:00
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
yprk.domain = indomain;
yprk.map = inmap;
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
yprk.key.keydat_val = (char *)inkey;
yprk.key.keydat_len = inkeylen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
bzero((char *)&yprv, sizeof yprv);
r = clnt_call(ysd->dom_client, YPPROC_MATCH,
xdr_ypreq_key, &yprk, xdr_ypresp_val, &yprv, tv);
if(r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_match: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
if( !(r=ypprot_err(yprv.stat)) ) {
*outvallen = yprv.val.valdat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outval = (char *)malloc(*outvallen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprv.val.valdat_val, *outval, *outvallen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outval)[*outvallen] = '\0';
#ifdef YPMATCHCACHE
if( strcmp(_yp_domain, indomain)==0 )
ypmatch_add(inmap, inkey, inkeylen, *outval, *outvallen);
#endif
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_val, (char *)&yprv);
return r;
}
int
yp_get_default_domain(domp)
char **domp;
{
*domp = NULL;
if(_yp_domain[0] == '\0')
if( getdomainname(_yp_domain, sizeof _yp_domain))
return YPERR_NODOM;
*domp = _yp_domain;
return 0;
}
int
yp_first(indomain, inmap, outkey, outkeylen, outval, outvallen)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
char **outkey;
int *outkeylen;
char **outval;
int *outvallen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct ypresp_key_val yprkv;
struct ypreq_nokey yprnk;
struct dom_binding *ysd;
struct timeval tv;
int r;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
if (indomain == NULL || !strlen(indomain) ||
inmap == NULL || !strlen(inmap))
return YPERR_BADARGS;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outkey = *outval = NULL;
*outkeylen = *outvallen = 0;
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
yprnk.domain = indomain;
yprnk.map = inmap;
bzero((char *)&yprkv, sizeof yprkv);
r = clnt_call(ysd->dom_client, YPPROC_FIRST,
xdr_ypreq_nokey, &yprnk, xdr_ypresp_key_val, &yprkv, tv);
if(r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_first: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
if( !(r=ypprot_err(yprkv.stat)) ) {
*outkeylen = yprkv.key.keydat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outkey = (char *)malloc(*outkeylen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprkv.key.keydat_val, *outkey, *outkeylen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outkey)[*outkeylen] = '\0';
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
*outvallen = yprkv.val.valdat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outval = (char *)malloc(*outvallen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprkv.val.valdat_val, *outval, *outvallen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outval)[*outvallen] = '\0';
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_key_val, (char *)&yprkv);
return r;
}
int
yp_next(indomain, inmap, inkey, inkeylen, outkey, outkeylen, outval, outvallen)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
char *inkey;
int inkeylen;
char **outkey;
int *outkeylen;
char **outval;
int *outvallen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct ypresp_key_val yprkv;
struct ypreq_key yprk;
struct dom_binding *ysd;
struct timeval tv;
int r;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
if (inkey == NULL || !strlen(inkey) || inkeylen <= 0 ||
inmap == NULL || !strlen(inmap) ||
indomain == NULL || !strlen(indomain))
return YPERR_BADARGS;
Don't let yp_match() or yp_next() operate on null or empty keys: asking ypserv to do a yp_match() with an a null or empty key causes much havok. (Note that this could be construed as a denial of service attack if used maliciously.)
1995-03-23 22:23:15 +00:00
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
*outkey = *outval = NULL;
*outkeylen = *outvallen = 0;
Don't let yp_match() or yp_next() operate on null or empty keys: asking ypserv to do a yp_match() with an a null or empty key causes much havok. (Note that this could be construed as a denial of service attack if used maliciously.)
1995-03-23 22:23:15 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
yprk.domain = indomain;
yprk.map = inmap;
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
yprk.key.keydat_val = inkey;
yprk.key.keydat_len = inkeylen;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
bzero((char *)&yprkv, sizeof yprkv);
r = clnt_call(ysd->dom_client, YPPROC_NEXT,
xdr_ypreq_key, &yprk, xdr_ypresp_key_val, &yprkv, tv);
if(r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_next: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
if( !(r=ypprot_err(yprkv.stat)) ) {
*outkeylen = yprkv.key.keydat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outkey = (char *)malloc(*outkeylen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprkv.key.keydat_val, *outkey, *outkeylen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outkey)[*outkeylen] = '\0';
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
*outvallen = yprkv.val.valdat_len;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
*outval = (char *)malloc(*outvallen+1);
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
bcopy(yprkv.val.valdat_val, *outval, *outvallen);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
(*outval)[*outvallen] = '\0';
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_key_val, (char *)&yprkv);
return r;
}
int
yp_all(indomain, inmap, incallback)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
struct ypall_callback *incallback;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct ypreq_nokey yprnk;
struct dom_binding *ysd;
struct timeval tv;
struct sockaddr_in clnt_sin;
CLIENT *clnt;
Fix yp_all() so that it doesn't bogusly return YP_NOMORE at the end of a successful map retrieval. (This has to do with a previous change to xdr_ypresp_all_seq() and ypxfr_get_map(); originally, yp_all() would look for a return value of YP_FALSE to signal success, but now it should be looking for YP_NOMORE. It should not be passing YP_NOMORE back up to the caller though.) Noticed by: <aagero@aage.priv.no> There is also another small bug here, which is that the call to xdr_free() that happens immediately after the clnt_call() in yp_all() clobbers the return status value. I've worked around this for now, but I think the xdr_free() is actually bogus and should be removed. I want to check some more before I do that though.
1996-03-19 19:27:03 +00:00
u_long status, savstat;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
int clnt_sock;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
if (indomain == NULL || !strlen(indomain) ||
inmap == NULL || !strlen(inmap))
return YPERR_BADARGS;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
again:
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
/* YPPROC_ALL manufactures its own channel to ypserv using TCP */
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
clnt_sock = RPC_ANYSOCK;
clnt_sin = ysd->dom_server_addr;
clnt_sin.sin_port = 0;
clnt = clnttcp_create(&clnt_sin, YPPROG, YPVERS, &clnt_sock, 0, 0);
if(clnt==NULL) {
printf("clnttcp_create failed\n");
return YPERR_PMAP;
}
yprnk.domain = indomain;
yprnk.map = inmap;
ypresp_allfn = incallback->foreach;
ypresp_data = (void *)incallback->data;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
if (clnt_call(clnt, YPPROC_ALL,
xdr_ypreq_nokey, &yprnk,
xdr_ypresp_all_seq, &status, tv) != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_next: clnt_call");
clnt_destroy(clnt);
_yp_unbind(ysd);
goto again;
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
clnt_destroy(clnt);
Fix yp_all() so that it doesn't bogusly return YP_NOMORE at the end of a successful map retrieval. (This has to do with a previous change to xdr_ypresp_all_seq() and ypxfr_get_map(); originally, yp_all() would look for a return value of YP_FALSE to signal success, but now it should be looking for YP_NOMORE. It should not be passing YP_NOMORE back up to the caller though.) Noticed by: <aagero@aage.priv.no> There is also another small bug here, which is that the call to xdr_free() that happens immediately after the clnt_call() in yp_all() clobbers the return status value. I've worked around this for now, but I think the xdr_free() is actually bogus and should be removed. I want to check some more before I do that though.
1996-03-19 19:27:03 +00:00
savstat = status;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_all_seq, (char *)&status); /* not really needed... */
Fix yp_all() so that it doesn't bogusly return YP_NOMORE at the end of a successful map retrieval. (This has to do with a previous change to xdr_ypresp_all_seq() and ypxfr_get_map(); originally, yp_all() would look for a return value of YP_FALSE to signal success, but now it should be looking for YP_NOMORE. It should not be passing YP_NOMORE back up to the caller though.) Noticed by: <aagero@aage.priv.no> There is also another small bug here, which is that the call to xdr_free() that happens immediately after the clnt_call() in yp_all() clobbers the return status value. I've worked around this for now, but I think the xdr_free() is actually bogus and should be removed. I want to check some more before I do that though.
1996-03-19 19:27:03 +00:00
if(savstat != YP_NOMORE)
return ypprot_err(savstat);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
return 0;
}
int
yp_order(indomain, inmap, outorder)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
int *outorder;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct dom_binding *ysd;
struct ypresp_order ypro;
struct ypreq_nokey yprnk;
struct timeval tv;
int r;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
if (indomain == NULL || !strlen(indomain) ||
inmap == NULL || !strlen(inmap))
return YPERR_BADARGS;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
yprnk.domain = indomain;
yprnk.map = inmap;
bzero((char *)(char *)&ypro, sizeof ypro);
r = clnt_call(ysd->dom_client, YPPROC_ORDER,
xdr_ypreq_nokey, &yprnk, xdr_ypresp_order, &ypro, tv);
- Avoid possible SEGVs: never call clnt_destroy() without checking for NULL RPC client handles. This should hopefully fix the problems Satoshi reported on -current. - Add socket descriptor sanity checks to _yp_unbind(). - Fix yp_order() so that it handles the RPC_PROCUNAVAIL error gracefully. NIS+ in YP compat mode doesn't support the YPPROC_ORDER procedure. This is a 2.2 candidate with bells on.
1996-11-08 01:42:02 +00:00
/*
* NIS+ in YP compat mode doesn't support the YPPROC_ORDER
* procedure.
*/
if (r == RPC_PROCUNAVAIL) {
return(YPERR_YPERR);
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if(r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_order: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
if( !(r=ypprot_err(ypro.stat)) ) {
*outorder = ypro.ordernum;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_order, (char *)&ypro);
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
return (r);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
int
yp_master(indomain, inmap, outname)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
char *inmap;
char **outname;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct dom_binding *ysd;
struct ypresp_master yprm;
struct ypreq_nokey yprnk;
struct timeval tv;
int r;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
if (indomain == NULL || !strlen(indomain) ||
inmap == NULL || !strlen(inmap))
return YPERR_BADARGS;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
yprnk.domain = indomain;
yprnk.map = inmap;
bzero((char *)&yprm, sizeof yprm);
r = clnt_call(ysd->dom_client, YPPROC_MASTER,
xdr_ypreq_nokey, &yprnk, xdr_ypresp_master, &yprm, tv);
if(r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_master: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
Now that rpcgen is squared away, arrange to have all the NIS XDR routines rpcgen-erated on the fly (just like librpcsvc). Makefile: Add rule for generating yp_xdr.c and yp.h. xdryp.c: gut everything except the special ypresp_all XDR function needed to to handle yp_all() (this one can't be created on the fly), and xdr_datum(), which isn't used internally by libc, but which as documented as being there in yp_prot.h, so what the hell. We now get everything else from yp_xdr.c. yplib.c: change a few structure member names to match those found in yp.h instead of those declared in yp_prot.h.
1995-12-14 05:16:45 +00:00
if( !(r=ypprot_err(yprm.stat)) ) {
*outname = (char *)strdup(yprm.peer);
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
xdr_free(xdr_ypresp_master, (char *)&yprm);
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
return (r);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
More of the same: we also need to handle multiple domains properly even if /var/yp/binding/DOMAIN.VERS doesn't exist yet.
1995-04-09 21:52:31 +00:00
int
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
yp_maplist(indomain, outmaplist)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char *indomain;
struct ypmaplist **outmaplist;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
struct dom_binding *ysd;
struct ypresp_maplist ypml;
struct timeval tv;
int r;
Add more sanity checks. *Lots* of sanity checks. Huge tracts of sanity checks. Make sure all arguments to the yp_*() functions are valid before sending them off to the server. This is somewhat distressing: once again my FreeBSD box brought down my entire network because of NIS bogosities. I *think* the poor argument checking in this module is the cause, but I still haven't been able to reproduce the exact series of events that lead to the ypserv crashes. For now I've resorted to sticking my FreeBSD box in a seprate domain. Hopefully a weekend of heavy testing will uncover the problem.
1995-03-24 21:21:37 +00:00
/* Sanity check */
if (indomain == NULL || !strlen(indomain))
return YPERR_BADARGS;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
again:
if( _yp_dobind(indomain, &ysd) != 0)
return YPERR_DOMAIN;
tv.tv_sec = _yplib_timeout;
tv.tv_usec = 0;
bzero((char *)&ypml, sizeof ypml);
r = clnt_call(ysd->dom_client, YPPROC_MAPLIST,
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
xdr_domainname,(char *)&indomain,xdr_ypresp_maplist,&ypml,tv);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
if (r != RPC_SUCCESS) {
clnt_perror(ysd->dom_client, "yp_maplist: clnt_call");
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
_yp_unbind(ysd);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
goto again;
}
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
if( !(r=ypprot_err(ypml.stat)) ) {
*outmaplist = ypml.maps;
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
/* NO: xdr_free(xdr_ypresp_maplist, &ypml);*/
- Patch around amd core dump problem: don't allow yp_unbind() or _yp_unbind() to call clnt_destroy() on a potentially NULL RPC handle. Somebody should bang on this a bit to make sure the problem is really gone; I seem to have difficulty reproducing it. Patch provided by Peter Wemm and slightly tweaked by me. - Don't call _yp_unbind() in individual ypclnt functions unless we encounter an RPC error while making a clnt_call().
1996-05-16 18:01:17 +00:00
return (r);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
}
char *
yperr_string(incode)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
int incode;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
static char err[80];
switch(incode) {
case 0:
return "Success";
case YPERR_BADARGS:
return "Request arguments bad";
case YPERR_RPC:
return "RPC failure";
case YPERR_DOMAIN:
return "Can't bind to server which serves this domain";
case YPERR_MAP:
return "No such map in server's domain";
case YPERR_KEY:
return "No such key in map";
case YPERR_YPERR:
return "YP server error";
case YPERR_RESRC:
return "Local resource allocation failure";
case YPERR_NOMORE:
return "No more records in map database";
case YPERR_PMAP:
return "Can't communicate with portmapper";
case YPERR_YPBIND:
return "Can't communicate with ypbind";
case YPERR_YPSERV:
return "Can't communicate with ypserv";
case YPERR_NODOM:
return "Local domain name not set";
case YPERR_BADDB:
return "Server data base is bad";
case YPERR_VERS:
return "YP server version mismatch - server can't supply service.";
case YPERR_ACCESS:
return "Access violation";
case YPERR_BUSY:
return "Database is busy";
}
sprintf(err, "YP unknown error %d\n", incode);
return err;
}
int
ypprot_err(incode)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
unsigned int incode;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
switch(incode) {
case YP_TRUE:
return 0;
case YP_FALSE:
return YPERR_YPBIND;
case YP_NOMORE:
return YPERR_NOMORE;
case YP_NOMAP:
return YPERR_MAP;
case YP_NODOM:
The ypprot_err() function incorrectly maps YP_NODOM to YPERR_NODOM. Strange as it sounds, it should map to YPERR_DOMAIN instead. The YP_NODOM protocol error code is generally returned by ypserv when you ask it for data from a domain that it doesn't support. By contrast, the YPERR_NODOM error code means 'local domain name not set.' Consequently, this incorrect mapping leads to yperr_string() generating a very confusing error message. YPERR_DOMAIN says 'couldn't bind to a server which serves this domain' which is much closer to the truth.
1995-07-05 06:04:20 +00:00
return YPERR_DOMAIN;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
case YP_NOKEY:
return YPERR_KEY;
case YP_BADOP:
return YPERR_YPERR;
case YP_BADDB:
return YPERR_BADDB;
case YP_YPERR:
return YPERR_YPERR;
case YP_BADARGS:
return YPERR_BADARGS;
case YP_VERS:
return YPERR_VERS;
}
return YPERR_YPERR;
}
int
_yp_check(dom)
NIS client-side performance tweak: Each of the ypclnt functions does a _yp_dobind() when it starts and then a _yp_unbind() when it finishes. This is not strictly necessary and it wastes cycles: it means we do a new clnt_create() and clnt_destroy() for each yp_whatever() call. In fact, you can do multiple clnt_call()s using a single RPC client handle returned by clnt_create(). Ideally we only have to create a handle to ypserv once (the first time we call a ypclnt function) and then destroy it and rebind only if a call to ypserv fails. - Modify _yp_dobind() so that it only creates a new RPC client handle when establishing a new binding or when one of the ypclnt calls invalidates an existing binding and calls _yp_dobind() to establish a new one. - Modify the various ypclnt functions to only call _yp_unbind() if a call to ypserv fails.
1996-05-02 15:44:53 +00:00
char **dom;
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
{
char *unused;
if( _yp_domain[0]=='\0' )
if( yp_get_default_domain(&unused) )
return 0;
if(dom)
*dom = _yp_domain;
- Make _do_ypbind() check for /var/run/ypbind.lock and attempt to flock() it before before trying to establish a binding. If /var/run/ypbind.lock doesn't exist, or if it exists and isn't locked, then ypbind isn't running, which means NIS is either turned off or hosed. - Have _yp_check() call yp_unbind() after it sucessfully calls yp_bind() to make sure it frees resources correctly. (I don't think there's really a memory leak here, but it seems somehow wrong to call yp_bind() without making a corresponding call to yp_unbind() afterwards.) This makes the NIS code behave a little better in cases where libc makes calls to NIS, but it isn't running correctly (i.e. there's no ypbind). This cleans up some strange libc behavior that manifests itself if you have the system domain name set, but aren't actually running NIS. In this event, the getrpcent(3) code could try to call into NIS and cause several inexplicable "clnttcp_create error: RPC program not registered" messages to appear. This happens because _yp_check() checks if the system domain name is set and, if it is, proceeds to call yp_bind() to attempt to establish a binding. Since there is no binding file (remember: ypbind isn't running, so /var/yp/binding will be empty), _yp_dobind() will attempt to contact ypbind to prod it into binding the domain. And because ypbind isn't running, the code generates the 'clnttcp_create' error. Ultimately the _yp_check() fails and the getrpcent(3) code rolls over to the /etc/rpc file, but the error messages are annoying, and the code should be smart enough to forgo the binding attempt when NIS is turned off.
1995-11-05 05:39:04 +00:00
if( yp_bind(_yp_domain)==0 ) {
yp_unbind(_yp_domain);
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
return 1;
- Make _do_ypbind() check for /var/run/ypbind.lock and attempt to flock() it before before trying to establish a binding. If /var/run/ypbind.lock doesn't exist, or if it exists and isn't locked, then ypbind isn't running, which means NIS is either turned off or hosed. - Have _yp_check() call yp_unbind() after it sucessfully calls yp_bind() to make sure it frees resources correctly. (I don't think there's really a memory leak here, but it seems somehow wrong to call yp_bind() without making a corresponding call to yp_unbind() afterwards.) This makes the NIS code behave a little better in cases where libc makes calls to NIS, but it isn't running correctly (i.e. there's no ypbind). This cleans up some strange libc behavior that manifests itself if you have the system domain name set, but aren't actually running NIS. In this event, the getrpcent(3) code could try to call into NIS and cause several inexplicable "clnttcp_create error: RPC program not registered" messages to appear. This happens because _yp_check() checks if the system domain name is set and, if it is, proceeds to call yp_bind() to attempt to establish a binding. Since there is no binding file (remember: ypbind isn't running, so /var/yp/binding will be empty), _yp_dobind() will attempt to contact ypbind to prod it into binding the domain. And because ypbind isn't running, the code generates the 'clnttcp_create' error. Ultimately the _yp_check() fails and the getrpcent(3) code rolls over to the /etc/rpc file, but the error messages are annoying, and the code should be smart enough to forgo the binding attempt when NIS is turned off.
1995-11-05 05:39:04 +00:00
}
Add back in the YP code from 1.1.5. (This attribution brought to you by Theo de Raadt.) Added a new make flag variable, NO_YP_LIBC, which disables YP entirely. User-land programs to come later.
1994-08-07 23:04:55 +00:00
return 0;
}
Reference in New Issue Copy Permalink