d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
372481e5b3
freebsd-dev/usr.sbin/ypserv/yp_dnslookup.c

543 lines
13 KiB
C
Raw Normal View History

Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
/*
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
* Copyright (c) 1995, 1996
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
* Bill Paul <wpaul@ctr.columbia.edu>. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by Bill Paul.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL Bill Paul OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
#ifndef lint
static const char rcsid[] =
$Id$ -> $FreeBSD$
1999-08-28 01:35:59 +00:00
"$FreeBSD$";
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
#endif /* not lint */
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
/*
* Do standard and reverse DNS lookups using the resolver library.
* Take care of all the dirty work here so the main program only has to
* pass us a pointer to an array of characters.
*
* We have to use direct resolver calls here otherwise the YP server
* could end up looping by calling itself over and over again until
* it disappeared up its own belly button.
*/
#include <sys/param.h>
#include <sys/socket.h>
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#include <sys/time.h>
#include <sys/fcntl.h>
#include <sys/queue.h>
#include <netinet/in.h>
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
#include <arpa/inet.h>
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#include <arpa/nameser.h>
#include <ctype.h>
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
#include <errno.h>
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#include <netdb.h>
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
#include <stdio.h>
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#include <stdlib.h>
#include <string.h>
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
#include <resolv.h>
#include <unistd.h>
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#include <rpcsvc/yp.h>
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
#include "yp_extern.h"
static char *parse(hp)
struct hostent *hp;
{
static char result[MAXHOSTNAMELEN * 2];
int len,i;
struct in_addr addr;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (hp == NULL)
return(NULL);
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
len = 16 + strlen(hp->h_name);
for (i = 0; hp->h_aliases[i]; i++)
len += strlen(hp->h_aliases[i]) + 1;
If we get back too much data to fit in result, return NULL. This avoids a buffer overflow, but might negatively impact those hosts who have enough aliases to fill MAXHOSTNAMELEN * 2 characters in them. Good candidate for merging back into -stable. Lightly tested by me, but it came from OpenBSD a while ago. Obtained from: OpenBSD
1998-06-09 05:06:27 +00:00
len++;
if (len > sizeof(result))
return(NULL);
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
bzero(result, sizeof(result));
bcopy(hp->h_addr, &addr, sizeof(struct in_addr));
snprintf(result, sizeof(result), "%s %s", inet_ntoa(addr), hp->h_name);
for (i = 0; hp->h_aliases[i]; i++) {
strcat(result, " ");
strcat(result, hp->h_aliases[i]);
}
return ((char *)&result);
}
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
#define MAXPACKET 1024
#define DEF_TTL 50
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
#define BY_DNS_ID 1
#define BY_RPC_XID 2
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
extern struct hostent *__dns_getanswer __P((char *, int, char *, int));
static CIRCLEQ_HEAD(dns_qhead, circleq_dnsentry) qhead;
struct circleq_dnsentry {
SVCXPRT *xprt;
unsigned long xid;
struct sockaddr_in client_addr;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
unsigned long ypvers;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
unsigned long id;
unsigned long ttl;
unsigned long type;
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
unsigned short prot_type;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
char **domain;
char *name;
struct in_addr addr;
CIRCLEQ_ENTRY(circleq_dnsentry) links;
};
static int pending = 0;
int yp_init_resolver()
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
{
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
CIRCLEQ_INIT(&qhead);
if (!(_res.options & RES_INIT) && res_init() == -1) {
yp_error("res_init failed");
return(1);
}
if ((resfd = socket(AF_INET, SOCK_DGRAM, 0)) == -1) {
yp_error("couldn't create socket");
return(1);
}
if (fcntl(resfd, F_SETFL, O_NONBLOCK) == -1) {
yp_error("couldn't make resolver socket non-blocking");
return(1);
}
return(0);
}
static struct circleq_dnsentry *yp_malloc_dnsent()
{
register struct circleq_dnsentry *q;
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q = (struct circleq_dnsentry *)malloc(sizeof(struct circleq_dnsentry));
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (q == NULL) {
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("failed to malloc() circleq dns entry");
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
return(NULL);
}
return(q);
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
}
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
/*
* Transmit a query.
*/
static unsigned long yp_send_dns_query(name, type)
char *name;
int type;
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
{
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
char buf[MAXPACKET];
int n;
HEADER *hptr;
int ns;
int rval;
unsigned long id;
bzero(buf, sizeof(buf));
n = res_mkquery(QUERY,name,C_IN,type,NULL,0,NULL,buf,sizeof(buf));
if (n <= 0) {
yp_error("res_mkquery failed");
return(0);
}
hptr = (HEADER *)&buf;
id = ntohs(hptr->id);
for (ns = 0; ns < _res.nscount; ns++) {
rval = sendto(resfd, buf, n, 0,
(struct sockaddr *)&_res.nsaddr_list[ns],
sizeof(struct sockaddr));
if (rval == -1) {
yp_error("sendto failed");
return(0);
}
}
return(id);
}
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
static struct circleq_dnsentry *yp_find_dnsqent(id, type)
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
unsigned long id;
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
int type;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
{
register struct circleq_dnsentry *q;
for (q = qhead.cqh_first; q != (void *)&qhead; q = q->links.cqe_next) {
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
switch(type) {
case BY_RPC_XID:
if (id == q->xid)
return(q);
break;
case BY_DNS_ID:
default:
if (id == q->id)
return(q);
break;
}
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
}
return (NULL);
}
static void yp_send_dns_reply(q, buf)
struct circleq_dnsentry *q;
char *buf;
{
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
ypresponse result_v1;
ypresp_val result_v2;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
unsigned long xid;
struct sockaddr_in client_addr;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
xdrproc_t xdrfunc;
char *result;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Set up correct reply struct and
* XDR filter depending on ypvers.
*/
switch(q->ypvers) {
case YPVERS:
bzero((char *)&result_v2, sizeof(result_v2));
if (buf == NULL)
result_v2.stat = YP_NOKEY;
else {
result_v2.val.valdat_len = strlen(buf);
result_v2.val.valdat_val = buf;
result_v2.stat = YP_TRUE;
}
result = (char *)&result_v2;
xdrfunc = (xdrproc_t)xdr_ypresp_val;
break;
case YPOLDVERS:
/*
* The odds are we will _never_ execute this
* particular code, but we include it anyway
* for the sake of completeness.
*/
bzero((char *)&result_v1, sizeof(result_v1));
result_v1.yp_resptype = YPRESP_VAL;
# define YPVAL ypresponse_u.yp_resp_valtype
if (buf == NULL)
result_v1.YPVAL.stat = YP_NOKEY;
else {
result_v1.YPVAL.val.valdat_len = strlen(buf);
result_v1.YPVAL.val.valdat_val = buf;
result_v1.YPVAL.stat = YP_TRUE;
}
result = (char *)&result_v1;
xdrfunc = (xdrproc_t)xdr_ypresponse;
break;
default:
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("bad YP program version (%lu)!", q->ypvers);
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
return;
break;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
}
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("sending dns reply to %s (%lu)",
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
inet_ntoa(q->client_addr.sin_addr), q->id);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
/*
* XXX This is disgusting. There's basically one transport
* handle for UDP, but we're holding off on replying to a
* client until we're ready, by which time we may have received
* several other queries from other clients with different
* transaction IDs. So to make the delayed response thing work,
* we have to save the transaction ID and client address of
* each request, then jam them into the transport handle when
* we're ready to send a reply. Then after we've send the reply,
* we put the old transaction ID and remote address back the
* way we found 'em. This is _INCREDIBLY_ non-portable; it's
* not even supported by the RPC library.
*/
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
/*
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
* XXX Don't frob the transaction ID for TCP handles.
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
*/
if (q->prot_type == SOCK_DGRAM)
xid = svcudp_set_xid(q->xprt, q->xid);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
client_addr = q->xprt->xp_raddr;
q->xprt->xp_raddr = q->client_addr;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
if (!svc_sendreply(q->xprt, xdrfunc, result))
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
yp_error("svc_sendreply failed");
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Now that we sent the reply,
* put the handle back the way it was.
*/
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
if (q->prot_type == SOCK_DGRAM)
svcudp_set_xid(q->xprt, xid);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->xprt->xp_raddr = client_addr;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
return;
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Decrement TTL on all queue entries, possibly nuking
* any that have been around too long without being serviced.
*/
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
void yp_prune_dnsq()
{
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
register struct circleq_dnsentry *q, *n;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
q = qhead.cqh_first;
while(q != (void *)&qhead) {
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->ttl--;
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
n = q->links.cqe_next;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (!q->ttl) {
CIRCLEQ_REMOVE(&qhead, q, links);
free(q->name);
free(q);
pending--;
}
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
q = n;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
}
if (pending < 0)
pending = 0;
return;
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Data is pending on the DNS socket; check for valid replies
* to our queries and dispatch them to waiting clients.
*/
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
void yp_run_dnsq()
{
register struct circleq_dnsentry *q;
char buf[sizeof(HEADER) + MAXPACKET];
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
char retrybuf[MAXHOSTNAMELEN];
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
struct sockaddr_in sin;
int rval;
int len;
HEADER *hptr;
struct hostent *hent;
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("running dns queue");
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
bzero(buf, sizeof(buf));
len = sizeof(struct sockaddr_in);
rval = recvfrom(resfd, buf, sizeof(buf), 0,
(struct sockaddr *)&sin, &len);
if (rval == -1) {
yp_error("recvfrom failed: %s", strerror(errno));
return;
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* We may have data left in the socket that represents
* replies to earlier queries that we don't care about
* anymore. If there are no lookups pending or the packet
* ID doesn't match any of the queue IDs, just drop it
* on the floor.
*/
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
hptr = (HEADER *)&buf;
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
if (!pending ||
(q = yp_find_dnsqent(ntohs(hptr->id), BY_DNS_ID)) == NULL) {
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/* ignore */
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
return;
}
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("got dns reply from %s", inet_ntoa(sin.sin_addr));
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
hent = __dns_getanswer(buf, rval, q->name, q->type);
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* If the lookup failed, try appending one of the domains
* from resolv.conf. If we have no domains to test, the
* query has failed.
*/
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (hent == NULL) {
Fix for PR #4147: apparently BIND now returns a status of NO_RECOVERY at the end of gethostanswer()/getanswer()/whatever where it used to return TRY_AGAIN. This breaks the domain list traversal in ypserv's async DNS lookup module: it would only retry using the domain(s) from the 'domain' or 'search' lines in /etc/resolv.conf if __dns_getanswer() returned TRY_AGAIN. Changed the test so that either TRY_AGAIN or NO_RECOVERY will work. This seemed to me the best solution in the event somebody tries to compile this code on an older system with a different version of BIND. (You shouldn't do that of course, but then there's a lot of things in the world that you shouldn't do and people do them anyway.)
1997-07-28 14:13:57 +00:00
if ((h_errno == TRY_AGAIN || h_errno == NO_RECOVERY)
&& q->domain && *q->domain) {
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
snprintf(retrybuf, sizeof(retrybuf), "%s.%s",
q->name, *q->domain);
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("retrying with: %s", retrybuf);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->id = yp_send_dns_query(retrybuf, q->type);
q->ttl = DEF_TTL;
q->domain++;
return;
}
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
} else {
if (q->type == T_PTR) {
hent->h_addr = (char *)&q->addr.s_addr;
hent->h_length = sizeof(struct in_addr);
}
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/* Got an answer ready for a client -- send it off. */
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
yp_send_dns_reply(q, parse(hent));
pending--;
CIRCLEQ_REMOVE(&qhead, q, links);
free(q->name);
free(q);
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/* Decrement TTLs on other entries while we're here. */
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
yp_prune_dnsq();
return;
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Queue and transmit an asynchronous DNS hostname lookup.
*/
ypstat yp_async_lookup_name(rqstp, name)
struct svc_req *rqstp;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
char *name;
{
register struct circleq_dnsentry *q;
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
int type, len;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
/* Check for SOCK_DGRAM or SOCK_STREAM -- we need to know later */
Ack. When I moved the getsockopt() calls, I forgot to move the lines that initialize the getsockopt() args (type and len) too.
1997-01-12 08:18:17 +00:00
type = -1; len = sizeof(type);
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
if (getsockopt(rqstp->rq_xprt->xp_sock, SOL_SOCKET,
SO_TYPE, &type, &len) == -1) {
yp_error("getsockopt failed: %s", strerror(errno));
return(YP_YPERR);
}
/* Avoid transmitting dupe requests. */
if (type == SOCK_DGRAM &&
yp_find_dnsqent(svcudp_get_xid(rqstp->rq_xprt),BY_RPC_XID) != NULL)
return(YP_TRUE);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if ((q = yp_malloc_dnsent()) == NULL)
return(YP_YPERR);
q->type = T_A;
q->ttl = DEF_TTL;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
q->xprt = rqstp->rq_xprt;
q->ypvers = rqstp->rq_vers;
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
q->prot_type = type;
if (q->prot_type == SOCK_DGRAM)
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
q->xid = svcudp_get_xid(q->xprt);
q->client_addr = q->xprt->xp_raddr;
DNS lookups for partially qualified hosts didn't work right. I was trying to be clever by avoiding the 'check all domains in the search list' cycle in certain cases, but this would lose if handed a name like "foo.ctr" which refers to an FQDN of "foo.ctr.columbia.edu". If "columbia.edu" is in the search list in /etc/resolv.conf then the DNS lookup code should resolve it, but it didn't.
1998-08-10 19:43:33 +00:00
q->domain = _res.dnsrch;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->id = yp_send_dns_query(name, q->type);
if (q->id == 0) {
yp_error("DNS query failed");
free(q);
return(YP_YPERR);
}
q->name = strdup(name);
CIRCLEQ_INSERT_HEAD(&qhead, q, links);
pending++;
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("queueing async DNS name lookup (%d)", q->id);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
yp_prune_dnsq();
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
return(YP_TRUE);
}
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
/*
* Queue and transmit an asynchronous DNS IP address lookup.
*/
ypstat yp_async_lookup_addr(rqstp, addr)
struct svc_req *rqstp;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
char *addr;
{
register struct circleq_dnsentry *q;
char buf[MAXHOSTNAMELEN];
int a, b, c, d;
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
int type, len;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
/* Check for SOCK_DGRAM or SOCK_STREAM -- we need to know later */
Ack. When I moved the getsockopt() calls, I forgot to move the lines that initialize the getsockopt() args (type and len) too.
1997-01-12 08:18:17 +00:00
type = -1; len = sizeof(type);
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
if (getsockopt(rqstp->rq_xprt->xp_sock, SOL_SOCKET,
SO_TYPE, &type, &len) == -1) {
yp_error("getsockopt failed: %s", strerror(errno));
return(YP_YPERR);
}
/* Avoid transmitting dupe requests. */
if (type == SOCK_DGRAM &&
yp_find_dnsqent(svcudp_get_xid(rqstp->rq_xprt),BY_RPC_XID) != NULL)
return(YP_TRUE);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if ((q = yp_malloc_dnsent()) == NULL)
return(YP_YPERR);
if (sscanf(addr, "%d.%d.%d.%d", &a, &b, &c, &d) != 4)
return(YP_NOKEY);
yp_server.c: - Fail YPPROC_ALL requests when we hit the child process limit. This is a little harsh, but it helps prevent the parent from blocking and causing other requests to time out. yp_dnslookup.c: - Check for duplicate RPC transaction IDs that indicate duplicate requests sent due to RPC retransmissions. We don't want to send a second DNS request for the same data while an existing request is in progress. - Fix small formatting bogon in snprintf() in yp_async_lookup_addr().
1997-01-07 06:07:21 +00:00
snprintf(buf, sizeof(buf), "%d.%d.%d.%d.in-addr.arpa", d, c, b, a);
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (debug)
yp_error("DNS address is: %s", buf);
q->type = T_PTR;
q->ttl = DEF_TTL;
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
q->xprt = rqstp->rq_xprt;
q->ypvers = rqstp->rq_vers;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->domain = NULL;
Fix some bugs: - Don't dereference a NULL hostent pointer (if T_PTR lookup fails). - Today I asked myself: "Self, you wrote this nifty async resolver that does a great job handling delayed replies to clients using the UDP transport, and the yplib code in libc always uses UDP (except for yp_all()). But what if some dork makes a DNS lookup using TCP?" Being the only dork on hand at the time, I tried it and was enlightened. As I suspected, my transaction ID frobbing hacks cause fireworks if called on a TCP transport handle (duh: the structures are different). Fix: check the type of socket in xprt->xp_sock using getsockopt() and don't use svcudp_get_xid() and svcudp_set_xid() for anything except SOCK_DGRAM sockets. (Since accept() gives you a new socket for each connection, the transaction ID munging isn't needed for TCP anyway.)
1996-12-24 02:44:52 +00:00
q->prot_type = type;
if (q->prot_type == SOCK_DGRAM)
More async resolver refinements: - yp_main.c: Always add the resolver socket to the set of fds monitored by select(). It can happen that pending == 0 but we still have some data in the socket buffer from an old query. This way, the data will be flushed in a timely manner. - yp_extern.h: remove proto for yp_dns_pending() since we don't need it anynmore. - yp_server.c: call yp_async_lookup_name()/yp_async_lookup_addr() functions with the svc_req pointer as an arg instead of the xprt. (The svc_req struct includes a pointer to the transport handle, and it also has the service version number which the async DNS code will need. (see below)) - yp_dnslookup.c: o Nuke yp_dns_pending() since we don't need it anymore. o In yp_run_dnsq(), swallow up and ignore replies if no requests are pending or the ID doesn't match any of the IDs in the queue. o In yp_send_dns_reply(), we assume that we will always be replying to an NIS v2 client. While this will probably always be the case, we do support the v1 'match' procedure, and it has a different result struct than v2. For completeness, support replying to both NIS v1 and v2 clients. o Update the queue entry structure to include a member to keep track of the NIS version number. o Have yp_async_lookup_name/addr() extract the version number from the svc_req structure and save it with the queue entry for yp_send_dns_reply() to inspect later. o Add some comments.
1996-12-25 18:10:35 +00:00
q->xid = svcudp_get_xid(q->xprt);
q->client_addr = q->xprt->xp_raddr;
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
q->id = yp_send_dns_query(buf, q->type);
if (q->id == 0) {
yp_error("DNS query failed");
free(q);
return(YP_YPERR);
}
inet_aton(addr, &q->addr);
q->name = strdup(buf);
CIRCLEQ_INSERT_HEAD(&qhead, q, links);
pending++;
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
if (debug)
Typos in man page. Cosmetics in error strings.
1997-10-29 07:25:05 +00:00
yp_error("queueing async DNS address lookup (%d)", q->id);
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
Fix a bug in the async DNS resolver that can crash ypserv. yp_prune_dnsq() is not sane: if the TTL on a pending but unanswered query hits 0 and the circular queue entry is removed and free()d, the for() loop may still try to use the entry pointer (which now points at no longer valid memory). usually, deleting only the last entry off the end of the queue worked, but if more than one was deleted, the server would crash. I changed things a bit so this shouldn't happen anymore. Also arranged to call the prune routine a bit more often.
1997-07-27 03:41:53 +00:00
yp_prune_dnsq();
Big round o changes: - yp_dblookup.c: Create non-DB specific database access functions. Using these allows access to the underlying database functions without needing explicit knowledge of Berkeley DB. (These are used only when DB_CACHE is #defined. Other programs that use the non-caching functions (yp_mkdb, ypxfr, yppush, rpc.yppasswdd) shouldn't notice the difference.) - yp_dnslookup: Implement async DNS lookups. We send our own DNS requests using UDP and put the request in a queue. When the response arrives, we use the ID in the header to find the corresponsing queue entry and then send the response to the client. We can go about our business and handle other YP requests in the meantime. This way, we can deal with time consuming DNS requests without blocking and without forking. - yp_server.c: Convert to using new non-DB-specific database access functions. This simplifies the code a bit and removes the need for this module to know anything about Berkeley DB. Also convert the ypproc_match_2_svc() function to use the async DNS lookup routines. - yp_main.c: tweak yp_svc_run() to add the resolver socket to the set of descriptors monitored in the select() loop. Also add a timeout to select(); we may get stale DNS requests stuck in the queue which we want to invalidate after a while. If the timeout hits, we decrement the ttl on all pending DNS requests and nuke those requests that aren't handled before ttl hits zero. - yp_extern.h: Add prototypes for new stuff. - yp_svc_udp.c (new file): The async resolver code needs to be able to rummage around inside the RPC UDP transport handle in order to work correcty. There's basically one transport handle, and each time a request comes in, the transaction ID in the handle is changed. This means that if we queue a DNS request, then we handle some other unrelated requests, we will be unable to send the DNS response because the transaction ID and remote address of the client that made the DNS request will have been lost. What we need to do is save the client address and transaction ID in the queue entry for the DNS request, then put the transaction ID and address back in the transport handle when we're ready to reply. (And then we have to undo the change so as not to confuse any other part of the server.) The trouble is that the transaction ID is hidden in an opaque part of the transport handle, and only the code in the svc_udp module in the RPC library knows how to handle it. This file contains a couple of functions that let us read and set the transaction ID in spite of this. This is really a dirty trick and I should be taken out and shot for even thinking about it, but there's no other way to get this stuff to work. - Makefile: add yp_svc_udp.c to SRCS.
1996-12-22 22:30:58 +00:00
return(YP_TRUE);
Import the new, non-GPL ypserv, written by yours truly. Functionally equivalent to the old ypserv, except that it doesn't support the -p [port] option to force the server to use a particular port. The server stubs and yp.h header file are auto-generated from the yp.x protocol definition file. The auto-generated XDR routines in libc/yp are also used. The database access code has been broken out into a seperate module so that other NIS utilities (ypxfr in particular) can use it. Note that the old mknetid script is being temporarily moved here; it will be replaced by an mknetid program which will eventually have a home under /usr/src/libexec. (The existing script is actually somewhat broken -- it doesn't handle hosts -- but this isn't a big deal at this point since the netid.byname map is really only useful fopr Secure RPC, which we don't have yet.)
1995-12-16 20:54:17 +00:00
}
Reference in New Issue Copy Permalink