Re: sasl2 and ipv6
On Mon, Nov 03, 2003 at 11:52:17AM -0500, brendan@xxxxxxxxxx wrote:
...
> I'm currently connected over IPv6 and SASL2 on debian unstable. What
> OS/version
> are you using?
Which begs the question, how big is a sockaddr in Debian?
The ipv6 bug is squashed by the following patch.
Cheers,
Patrick
Index: mutt_sasl.c
===================================================================
RCS file: /home/roessler/cvs/mutt/mutt_sasl.c,v
retrieving revision 3.4
diff -u -r3.4 mutt_sasl.c
--- mutt_sasl.c 11 Sep 2003 16:59:54 -0000 3.4
+++ mutt_sasl.c 5 Nov 2003 19:15:03 -0000
@@ -32,6 +32,48 @@
#include <sys/socket.h>
#include <netinet/in.h>
+static int getnameinfo_err(int ret)
+{
+ int err;
+ dprint (1, (debugfile, "getnameinfo: "));
+ switch(ret)
+ {
+ case EAI_AGAIN:
+ dprint (1, (debugfile, "The name could not be resolved at this time.
Future attempts may succeed.\n"));
+ err=SASL_TRYAGAIN;
+ break;
+ case EAI_BADFLAGS:
+ dprint (1, (debugfile, "The flags had an invalid value.\n"));
+ err=SASL_BADPARAM;
+ break;
+ case EAI_FAIL:
+ dprint (1, (debugfile, "A non-recoverable error occurred.\n"));
+ err=SASL_FAIL;
+ break;
+ case EAI_FAMILY:
+ dprint (1, (debugfile, "The address family was not recognized or the
address length was invalid for the specified family.\n"));
+ err=SASL_BADPROT;
+ break;
+ case EAI_MEMORY:
+ dprint (1, (debugfile, "There was a memory allocation failure.\n"));
+ err=SASL_NOMEM;
+ break;
+ case EAI_NONAME:
+ dprint (1, (debugfile, "The name does not resolve for the supplied
parameters. NI_NAMEREQD is set and the host's name cannot be located, or both
nodename and servname were null.\n"));
+ err=SASL_FAIL; /* no real equivalent */
+ break;
+ case EAI_SYSTEM:
+ dprint (1, (debugfile, "A system error occurred. The error code can be
found in errno(%d,%s)).\n",errno,strerror(errno)));
+ err=SASL_FAIL; /* no real equivalent */
+ break;
+ default:
+ dprint (1, (debugfile, "Unknown error %d\n",ret));
+ err=SASL_FAIL; /* no real equivalent */
+ break;
+ }
+ return err;
+}
+
/* arbitrary. SASL will probably use a smaller buffer anyway. OTOH it's
* been a while since I've had access to an SASL server which negotiated
* a protection buffer. */
@@ -64,15 +106,18 @@
static int iptostring(const struct sockaddr *addr, socklen_t addrlen,
char *out, unsigned outlen) {
char hbuf[NI_MAXHOST], pbuf[NI_MAXSERV];
+ int ret;
if(!addr || !out) return SASL_BADPARAM;
- getnameinfo(addr, addrlen, hbuf, sizeof(hbuf), pbuf, sizeof(pbuf),
- NI_NUMERICHOST |
+ ret=getnameinfo(addr, addrlen, hbuf, sizeof(hbuf), pbuf, sizeof(pbuf),
+ NI_NUMERICHOST |
#ifdef NI_WITHSCOPEID
- NI_WITHSCOPEID |
+ NI_WITHSCOPEID |
#endif
- NI_NUMERICSERV);
+ NI_NUMERICSERV);
+ if(ret)
+ return getnameinfo_err(ret);
if(outlen < strlen(hbuf) + strlen(pbuf) + 2)
return SASL_BUFOVER;
@@ -124,7 +169,7 @@
{
sasl_security_properties_t secprops;
#ifdef USE_SASL2
- struct sockaddr local, remote;
+ struct sockaddr_storage local, remote;
socklen_t size;
char iplocalport[IP_PORT_BUFLEN], ipremoteport[IP_PORT_BUFLEN];
#else
@@ -151,23 +196,23 @@
#ifdef USE_SASL2
size = sizeof (local);
- if (getsockname (conn->fd, &local, &size)){
+ if (getsockname (conn->fd, (struct sockaddr *)&local, &size)){
dprint (1, (debugfile, "mutt_sasl_client_new: getsockname for local
failed\n"));
return -1;
}
else
- if (iptostring(&local, size, iplocalport, IP_PORT_BUFLEN) != SASL_OK){
+ if (iptostring((struct sockaddr *)&local, local.ss_len, iplocalport,
IP_PORT_BUFLEN) != SASL_OK){
dprint (1, (debugfile, "mutt_sasl_client_new: iptostring for local
failed\n"));
return -1;
}
size = sizeof (remote);
- if (getpeername (conn->fd, &remote, &size)){
+ if (getpeername (conn->fd, (struct sockaddr *)&remote, &size)){
dprint (1, (debugfile, "mutt_sasl_client_new: getsockname for remote
failed\n"));
return -1;
}
else
- if (iptostring(&remote, size, ipremoteport, IP_PORT_BUFLEN) != SASL_OK){
+ if (iptostring((struct sockaddr *)&remote, remote.ss_len, ipremoteport,
IP_PORT_BUFLEN) != SASL_OK){
dprint (1, (debugfile, "mutt_sasl_client_new: iptostring for remote
failed\n"));
return -1;
}