mountd --no-tcp -p bugs?

mountd --no-tcp -p bugs?

[Ben Leslie]

Hi all,

I'm running into a couple of problems with mountd.

I'm using nfs-kernel-server version 1:1.0-2 package 
provided in Debian stable. However looking at the source
seems to indicate the same problems would exist in current
cvs version.

Problem 1/

The -n flag (short for --no-tcp) is ignored.

To reproduce: rpc.mountd -n

The problem seems to be in utils/mountd/mountd.c.

while ((c = getopt_long(argc, argv, "o:Fd:f:p:P:hN:V:v", longopts, NULL)) != EOF)

Probably should be:

while ((c = getopt_long(argc, argv, "o:Fd:f:p:P:hN:V:vn", longopts, NULL)) != EOF)


Problem 2/

The --no-tcp and -p <portnum> options don't play well together. 
Specifically, if you specify --no-tcp the -p option is ignored. I couldn't
find anything to indicate that this was the desired behaviour.

The problem here seems to be in the rpc_init() function (support/nfs/rpcmisc.c).

Specifically this bit:

if ((_rpcfdtype == 0) || (_rpcfdtype == SOCK_DGRAM)) {
	static SVCXPRT *last_transp = NULL;
 
	if (_rpcfdtype == 0) {
		if (last_transp
		    && (!defport || defport == last_transp->xp_port)) {
			transp = last_transp;
			goto udp_transport;
		}
		if (defport == 0)
			sock = RPC_ANYSOCK;
		else if ((sock = makesock(defport, IPPROTO_UDP)) < 0) {
			xlog(L_FATAL, "%s: cannot make a UDP socket\n",
					name);
		}
	}
	if (sock == RPC_ANYSOCK)
		sock = svcudp_socket (prog, 1);
	transp = svcudp_create(sock);


defport is only looked at if the _rpcfgtype is set to zero, which 
seems to me to be incorrect behaviour. However, since the rpc_init()
function is probably used by things other than mountd, chaning this
behaviour may have other unintended side affects. (Or I could be wrong
and this is indeed the correct behaviour, in which case the docs should
probably explain this.)

Cheers,

Benno



-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: mountd --no-tcp -p bugs?

[Ben Leslie]

On Tue, 11 Feb 2003, Benno wrote:

> Hi all,
> 
> I'm running into a couple of problems with mountd.

Hi again,

I haven't had any feedback on this email from a few days
ago. Is this not a bug and I'm missing something, or just 
no one cares about bugs in client programs, or am I just being
impatient?

Thanks again,

Benno


-------------------------------------------------------
This SF.NET email is sponsored by: FREE  SSL Guide from Thawte
are you planning your Web Server Security? Click here to get a FREE
Thawte SSL guide and find the answers to all your  SSL security issues.
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: mountd --no-tcp -p bugs?

[Neil Brown]

On Friday February 14, benno@sesgroup.net wrote:
> On Tue, 11 Feb 2003, Benno wrote:
> 
> > Hi all,
> > 
> > I'm running into a couple of problems with mountd.
> 
> Hi again,
> 
> I haven't had any feedback on this email from a few days
> ago. Is this not a bug and I'm missing something, or just 
> no one cares about bugs in client programs, or am I just being
> impatient?

It is a bug.
You aren't missing much.
People do care, but are also busy.
You were reasonably patient, and it does help to ask again (and
again...) if you don't get a response.  It gives an indication of how
important it is, and helps people with short memory and long todo
lists.

I have just committed the following patch to the nfs-utils CVS.  It
should address your issues.

NeilBrown

Index: ChangeLog
===================================================================
RCS file: /cvsroot/nfs/nfs-utils/ChangeLog,v
retrieving revision 1.187
diff -u -r1.187 ChangeLog
--- ChangeLog	13 Feb 2003 02:29:42 -0000	1.187
+++ ChangeLog	19 Feb 2003 04:23:27 -0000
@@ -1,3 +1,16 @@
+2003-02-19 NeilBrown <neilb@cse.unsw.edu.au>	
+	Ben Leslie <benno@sesgroup.net>
+
+	* support/nfs/rpcmisc.c(rpc_init):  change test for ignoring
+	"defport" from (__rpcfdtype == 0) to (__rpcpmstart==0).  The
+	former would incorectly ignore defport if the application 
+	explicitly set __rpcfdtype.  The later is sufficient to avoid the
+	special handling in the case where stdin is an internet socket.
+
+	* utils/mountd/mountd.c(main): Add 'n' to option list in
+	getopt_long as this is a short form of --no-tcp which want being
+	checked.
+	
 2003-02-12  Chip Salzenberg  <chip@pobox.com>
 
 	* debian/changelog: Version 1.0.2-2.
Index: support/nfs/rpcmisc.c
===================================================================
RCS file: /cvsroot/nfs/nfs-utils/support/nfs/rpcmisc.c,v
retrieving revision 1.13
diff -u -r1.13 rpcmisc.c
--- support/nfs/rpcmisc.c	16 Sep 2002 02:27:45 -0000	1.13
+++ support/nfs/rpcmisc.c	19 Feb 2003 04:23:28 -0000
@@ -66,7 +66,7 @@
 	if ((_rpcfdtype == 0) || (_rpcfdtype == SOCK_DGRAM)) {
 		static SVCXPRT *last_transp = NULL;
  
-		if (_rpcfdtype == 0) {
+		if (_rpcpmstart == 0) {
 			if (last_transp
 			    && (!defport || defport == last_transp->xp_port)) {
 				transp = last_transp;
@@ -96,7 +96,7 @@
 	if ((_rpcfdtype == 0) || (_rpcfdtype == SOCK_STREAM)) {
 		static SVCXPRT *last_transp = NULL;
 
-		if (_rpcfdtype == 0) {
+		if (_rpcpmstart == 0) {
 			if (last_transp
 			    && (!defport || defport == last_transp->xp_port)) {
 				transp = last_transp;
Index: utils/mountd/mountd.c
===================================================================
RCS file: /cvsroot/nfs/nfs-utils/utils/mountd/mountd.c,v
retrieving revision 1.13
diff -u -r1.13 mountd.c
--- utils/mountd/mountd.c	8 Apr 2002 21:43:25 -0000	1.13
+++ utils/mountd/mountd.c	19 Feb 2003 04:23:28 -0000
@@ -399,7 +399,7 @@
 
 	/* Parse the command line options and arguments. */
 	opterr = 0;
-	while ((c = getopt_long(argc, argv, "o:Fd:f:p:P:hN:V:v", longopts, NULL)) != EOF)
+	while ((c = getopt_long(argc, argv, "on:Fd:f:p:P:hN:V:v", longopts, NULL)) != EOF)
 		switch (c) {
 		case 'o':
 			descriptors = atoi(optarg);


-------------------------------------------------------
This SF.net email is sponsored by: SlickEdit Inc. Develop an edge.
The most comprehensive and flexible code editor you can use.
Code faster. C/C++, C#, Java, HTML, XML, many more. FREE 30-Day Trial.
www.slickedit.com/sourceforge
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs