Upcoming nfs-utils 1.1.3 release.

FYI...

I'm thinking about making a 1.1.3 nfs-utils update release in the relatively near future. I've made about 20 or so commits since the last release so I was thinking that was a nice round number to cut a release...

So if your sitting on some patches that you want include please post them to the list. Also if you don't see something in the git tree that should be there (i.e. I missed the patch posting) please bring it to my attention...

steved.

Hello,

atively

so

them to

be

ion...=20

have there been any changes to rpc.gssd since 1.1.2? Searching for a pr= oper
kerberos ticket was broken. I have patches to fix it and also adding an=
extra
option to set preffered realm of the tickets (it comes handy in cross-r= ealm
environment). I would like to see these patches included.

--=20
Luk=E1=B9 Hejtm=E1nek

I just took quick look and no there does not appear to be any rpc.gssd changes. I think the last ones to go in were the "[PATCH 0/4] nfs-utils gssd" patch set that Kevin posted.

Sorry I must have missed the posting... would you mind sending me a pointer to those posting?

tia,

steved.

er

I did not send it yet. Tomorrow, I will pull git of nfs utils and rediff my patches against it and send you the patch set tomorrow.

(can you send me a pointer to the git repository?)

-- =

Luk=E1=B9 Hejtm=E1nek

SSBkb24ndCByZWNhbGwgYW55IHJlcG9ydHMgb2YgdGhpbmdzIGJlaW5nIGJyb2tlbi4gIEknbGwg d2FpdCBhbmQKY2hlY2sgb3V0IHRoZSBwYXRjaCB0b21vcnJvdy4KCk1lYW53aGlsZSwgdGhlIGdp dCB0cmVlIGlzIGF0CmdpdDovL2dpdC5saW51eC1uZnMub3JnL3Byb2plY3RzL3N0ZXZlZC9uZnMt dXRpbHMuZ2l0CgoyMDA4LzYvMjMgTHVrYXMgSGVqdG1hbmVrIDx4aGVqdG1hbkBpY3MubXVuaS5j ej46Cj4gT24gTW9uLCBKdW4gMjMsIDIwMDggYXQgMDU6MzM6MzhQTSAtMDQwMCwgU3RldmUgRGlj a3NvbiB3cm90ZToKPj4gU29ycnkgSSBtdXN0IGhhdmUgbWlzc2VkIHRoZSBwb3N0aW5nLi4uIHdv dWxkIHlvdSBtaW5kIHNlbmRpbmcgbWUgYSBwb2ludGVyCj4+IHRvIHRob3NlIHBvc3Rpbmc/Cj4K PiBJIGRpZCBub3Qgc2VuZCBpdCB5ZXQuIFRvbW9ycm93LCBJIHdpbGwgcHVsbCBnaXQgb2YgbmZz IHV0aWxzIGFuZCByZWRpZmYgbXkKPiBwYXRjaGVzIGFnYWluc3QgaXQgYW5kIHNlbmQgeW91IHRo ZSBwYXRjaCBzZXQgdG9tb3Jyb3cuCj4KPiAoY2FuIHlvdSBzZW5kIG1lIGEgcG9pbnRlciB0byB0 aGUgZ2l0IHJlcG9zaXRvcnk/KQo+Cj4gLS0KPiBMdWvhuSBIZWp0beFuZWsKPgo+Cg==

as promissed, my patch is attached. hope you like it :)

Lukáš Hejtmánek

index e8612a5..1ad14f7 100644
@@ -60,6 +60,8 @@ char ccachedir[PATH_MAX] = GSSD_DEFAULT_CRED_DIR;
 char *ccachesearch[GSSD_MAX_CCACHE_SEARCH + 1];
 int  use_memcache = 0;
 int  root_uses_machine_creds = 1;
+char *preferred_realm = NULL;
+int ccache_timeout = 0;
 
 void
 sig_die(int signal)

@@ -82,7 +84,7 @@ sig_hup(int signal)
 static void
 usage(char *progname)
 {
-	fprintf(stderr, "usage: %s [-f] [-M] [-n] [-v] [-r] [-p pipefsdir] [-k keytab] [-d ccachedir]\n",
+	fprintf(stderr, "usage: %s [-f] [-M] [-n] [-v] [-r] [-p pipefsdir] [-k keytab] [-d ccachedir][-R preferred realm] [-t ticketstimeout]\n",
 		progname);
 	exit(1);
 }

@@ -99,7 +101,7 @@ main(int argc, char *argv[])
 	char *progname;
 
 	memset(ccachesearch, 0, sizeof(ccachesearch));
-	while ((opt = getopt(argc, argv, "fvrmnMp:k:d:")) != -1) {
+	while ((opt = getopt(argc, argv, "fvrmnMp:k:d:R:t:")) != -1) {
 		switch (opt) {
 			case 'f':
 				fg = 1;

@@ -134,6 +136,12 @@ main(int argc, char *argv[])
 				if (ccachedir[sizeof(ccachedir)-1] != '\0')
 					errx(1, "ccachedir path name too long");
 				break;
+			case 'R':
+				preferred_realm = strdup(optarg);
+				break;
+			case 't':
+				ccache_timeout = atoi(optarg);
+				break;
 			default:
 				usage(argv[0]);
 				break;

index 0f9f428..b3d6c82 100644
@@ -65,6 +65,8 @@ extern char			keytabfile[PATH_MAX];
 extern char			*ccachesearch[];
 extern int			use_memcache;
 extern int			root_uses_machine_creds;
+extern char			*preferred_realm;
+extern int			ccache_timeout;
 
 TAILQ_HEAD(clnt_list_head, clnt_info) clnt_list;
 

index 8fa4f4a..2b5ff27 100644
@@ -86,6 +86,15 @@ Increases the verbosity of the output (can be specified multiple times).
 .B -r
 If the rpcsec_gss library supports setting debug level,
 increases the verbosity of the output (can be specified multiple times).
+.TP
+.B -R realm
+Tickets from this realm will be preffered when scaning ccache dir.
+By default no realm is preferred.
+.TP
+.B -t timeout
+Timeout for kernel tickets cache. This is workaround in case you want to 
+change kerberos tickets and identities frequently. Default value is 
+0 - no timeout.
 .SH SEE ALSO
 .BR rpc.svcgssd(8)
 .SH AUTHORS

index be6f440..81e3bb8 100644
@@ -427,7 +427,7 @@ do_downcall(int k5_fd, uid_t uid, struct authgss_private_data *pd,
 	    gss_buffer_desc *context_token)
 {
 	char    *buf = NULL, *p = NULL, *end = NULL;
-	unsigned int timeout = 0; /* XXX decide on a reasonable value */
+	unsigned int timeout = ccache_timeout;
 	unsigned int buf_size = 0;
 
 	printerr(1, "doing downcall\n");

@@ -703,11 +703,10 @@ handle_krb5_upcall(struct clnt_info *clp)
 	if (uid != 0 || (uid == 0 && root_uses_machine_creds == 0)) {
 		/* Tell krb5 gss which credentials cache to use */
 		for (dirname = ccachesearch; *dirname != NULL; dirname++) {
-			gssd_setup_krb5_user_gss_ccache(uid, clp->servername, *dirname);
-
-			create_resp = create_auth_rpc_client(clp, &rpc_clnt, &auth, uid,
+			if(!gssd_setup_krb5_user_gss_ccache(uid, clp->servername, *dirname))
+				create_resp = create_auth_rpc_client(clp, &rpc_clnt, &auth, uid,
 							     AUTHTYPE_KRB5);
-			if (create_resp == 0)
+			if (create_resp == 0) 
 				break;
 		}
 	}

@@ -729,7 +728,7 @@ handle_krb5_upcall(struct clnt_info *clp)
 			}
 			for (ccname = credlist; ccname && *ccname; ccname++) {
 				gssd_setup_krb5_machine_gss_ccache(*ccname);
-				if ((create_auth_rpc_client(clp, &rpc_clnt,
+				if((create_auth_rpc_client(clp, &rpc_clnt,
 							    &auth, uid,
 							    AUTHTYPE_KRB5)) == 0) {
 					/* Success! */

index 431fdaf..27bd692 100644
@@ -17,7 +17,7 @@ struct gssd_k5_kt_princ {
 };
 
 
-void gssd_setup_krb5_user_gss_ccache(uid_t uid, char *servername,
+int  gssd_setup_krb5_user_gss_ccache(uid_t uid, char *servername,
 				     char *dirname);
 int  gssd_get_krb5_machine_cred_list(char ***list);
 void gssd_free_krb5_machine_cred_list(char **list);

_______________________________________________
NFSv4 mailing list
[email protected]
http://linux-nfs.org/cgi-bin/mailman/listinfo/nfsv4

SGkgTHVr4bksCkNvdWxkIHlvdSBwb2ludCBvdXQgd2hhdCBpcyAiYnJva2VuIiB3aXRoIHRoZSBl eGlzdGluZyBjb2RlLCBhbmQgd2hlcmUKaW4gdGhpcyBwYXRjaCB0aGF0IGlzIGZpeGVkPwoKSSBj YW4gc2VlIHRoZSB2YWx1ZSBvZiBhIHByZWZlcnJlZCByZWFsbS4gIFdvdWxkIGl0IGFsc28gYmUg YSBnb29kCmlkZWEgdG8gdXNlIHRoZSBtYWNoaW5lJ3MgZGVmYXVsdCByZWFsbSBpZiBubyBwcmVm ZXJyZWQgcmVhbG0gaXMKc3BlY2lmaWVkPwoKSSdkIGxpa2UgdG8gZ2V0IG90aGVyJ3Mgb3Bpbmlv bnMgb24gdGhlIGFkZGl0aW9uIG9mIHRoZSB0aW1lb3V0IHBhcmFtZXRlci4KCkl0IHNlZW1zIHRo aXMgc2hvdWxkIHJlYWxseSBiZSB0aHJlZSBzZXBhcmF0ZSBwYXRjaGVzOiAgMSkgdGhlIGZpeCBm b3IKdGhlICJicm9rZW4iIGNvZGUsIDIpIGFkZGluZyB0aGUgcHJlZmVycmVkIHJlYWxtLCAzKSBh ZGRpbmcgdGhlCmNvbnRleHQgdGltZW91dC4KCkkgZG9uJ3QgaGF2ZSB0aW1lIHRvIGxvb2sgYW55 IGNsb3NlciB0b2RheSwgaG9wZWZ1bGx5IHRvbW9ycm93LgoKSy5DLgoKMjAwOC82LzI0IEx1a2Fz IEhlanRtYW5layA8eGhlanRtYW5AaWNzLm11bmkuY3o+Ogo+IE9uIE1vbiwgSnVuIDIzLCAyMDA4 IGF0IDA1OjMzOjM4UE0gLTA0MDAsIFN0ZXZlIERpY2tzb24gd3JvdGU6Cj4+IFNvcnJ5IEkgbXVz dCBoYXZlIG1pc3NlZCB0aGUgcG9zdGluZy4uLiB3b3VsZCB5b3UgbWluZCBzZW5kaW5nIG1lIGEg cG9pbnRlcgo+PiB0byB0aG9zZSBwb3N0aW5nPwo+Cj4gYXMgcHJvbWlzc2VkLCBteSBwYXRjaCBp cyBhdHRhY2hlZC4gaG9wZSB5b3UgbGlrZSBpdCA6KQo+Cj4gLS0KPiBMdWvhuSBIZWp0beFuZWsK Pgo=

Btw, the current git head cannot be compiled as I don't have strnstr functi= on.
Configure does not check for it and select_krb5_ccache unconditionally uses it.

-- =

Luk=E1=B9 Hejtm=E1nek

2008/6/24 Lukas Hejtmanek <[email protected]>:

Near as I can tell, the use of strnstr() is from your patch? select_krb5_ccache() uses strstr().

I guess no. I believe it came to me from git head.

-- =

Luk=E1=B9 Hejtm=E1nek

Hi Lukas,

Again, not to be a pain but, we have a standard way of posting patches. Here are some guidelines: http://kernel.org/pub/linux/docs/lkml/#s1-10

But in general you need to describe at the top of the patch what problem the patch addresses and how that patch fixes that problem as well as having a "Signed-Off-By:" line.

So if you could please repost the patch following the above guidelines, it would be appreciated.

steved.

I tried my best. I sent two patches. The last one (that fixes "broken" gssd= ) =

I will send tomorrow.

-- =

Luk=E1=B9 Hejtm=E1nek

Hi,

Could you confirm there will be the Chuck Lever's patches about the mount ipv6 patchs serie?
Will there be also something about ipv6 server side commands?

Thanks

Steve Dickson a =E9crit :

tively near

was thinking

hem

ld be there (i.e.

n

--=20
-----------------------------------------------------------------
Company : Bull, Architect of an Open World TM (www.bull.com)
Name : Aime Le Rouzic=20
Mail : Bull - BP 208 - 38432 Echirolles Cedex - France
E-Mail : [email protected]
Phone : 33 (4) 76.29.75.51
=46ax : 33 (4) 76.29.75.18
-----------------------------------------------------------------=20

I'm not speaking for Steve, but I think my nfs-utils IPv6 patches are a little too much churn for a small dot release. And besides, I still consider them a prototype.

There is a large amount of dust yet to settle for the text-based mount interface. I hope nfs-utils 1.1.3 and the 2.6.27 kernel will address most of these issues.

Out of curiosity, who is working on server-side user space IPv6 support?

-

-- =

I am certain that these presidents will understand the cry of the people of Bolivia, of the people of Latin America and the whole world, which wants to have more food and not more cars. First food, then if something's left over, more cars, more automobiles. I think that life has to come first.
-- Evo Morales