Re: [SRM] proposed krb5 upload for stable

2014-09-20 Thread Julien Cristau
On Thu, Apr 25, 2013 at 15:53:40 -0400, Benjamin Kaduk wrote:

 Hi all,
 
 I've been working with hartmans to get some updates to krb5 into Debian (he
 is pretty busy these days, and I'm at upstream).  A couple of issues are
 serious enough that I think they merit being fixed in stable:
 
 #704775 - KDC crash by authenticated user with unmodified client tools
 #704647 - use of reverse DNS cannot be disabled, which breaks many
   configurations
 
 CVE-2013-1415 (KDC null pointer dereference with PKINIT) was not assigned a
 Debian bug number because we imported the fix right away.  It is less
 serious than #704775, but the patch is very small and it seemed worth
 including if an upload was to happen.
 
 These patches are all taken from upstream, and are already in wheezy.
 
 Per the quoted mail below (from #704775), Sam is prepared to upload the
 attached debdiff given SRM approval.
 
Hi Benjamin,

I'm really sorry we didn't reply to this mail in anywhere close to a
reasonable time.  As squeeze is now EOL I guess this is now moot.  In
the future, please open bugs against the release.debian.org
pseudo-package for proposed stable or oldstable updates, there's less
chance of something slipping through the cracks that way.

Thanks for understanding,
Julien


signature.asc
Description: Digital signature


Re: [SRM] proposed krb5 upload for stable

2014-09-20 Thread Michael Gilbert
On Sat, Sep 20, 2014 at 12:24 PM, Julien Cristau wrote:
 On Thu, Apr 25, 2013 at 15:53:40 -0400, Benjamin Kaduk wrote:

 Hi all,

 I've been working with hartmans to get some updates to krb5 into Debian (he
 is pretty busy these days, and I'm at upstream).  A couple of issues are
 serious enough that I think they merit being fixed in stable:

 #704775 - KDC crash by authenticated user with unmodified client tools
 #704647 - use of reverse DNS cannot be disabled, which breaks many
   configurations

 CVE-2013-1415 (KDC null pointer dereference with PKINIT) was not assigned a
 Debian bug number because we imported the fix right away.  It is less
 serious than #704775, but the patch is very small and it seemed worth
 including if an upload was to happen.

 These patches are all taken from upstream, and are already in wheezy.

 Per the quoted mail below (from #704775), Sam is prepared to upload the
 attached debdiff given SRM approval.

 Hi Benjamin,

 I'm really sorry we didn't reply to this mail in anywhere close to a
 reasonable time.  As squeeze is now EOL I guess this is now moot.  In
 the future, please open bugs against the release.debian.org
 pseudo-package for proposed stable or oldstable updates, there's less
 chance of something slipping through the cracks that way.

Hi Benjamin,

You can do a squeeze LTS upload on your own now.  Please see:
https://wiki.debian.org/LTS/Development

Best wishes,
Mike


-- 
To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/CANTw=motkzfmjb8wwwt1ldahgswx4kujs-uay_duckraimy...@mail.gmail.com



[SRM] proposed krb5 upload for stable

2013-04-25 Thread Benjamin Kaduk

Hi all,

I've been working with hartmans to get some updates to krb5 into Debian 
(he is pretty busy these days, and I'm at upstream).  A couple of 
issues are serious enough that I think they merit being fixed in stable:


#704775 - KDC crash by authenticated user with unmodified client tools
#704647 - use of reverse DNS cannot be disabled, which breaks many
  configurations

CVE-2013-1415 (KDC null pointer dereference with PKINIT) was not assigned 
a Debian bug number because we imported the fix right away.  It is less 
serious than #704775, but the patch is very small and it seemed worth 
including if an upload was to happen.


These patches are all taken from upstream, and are already in wheezy.

Per the quoted mail below (from #704775), Sam is prepared to upload the 
attached debdiff given SRM approval.


Thanks,

Ben Kaduk



On Thu, 25 Apr 2013, Sam Hartman wrote:


OK.
Why don't you run that patch by debian-release@lists.debian.org with a
SRM tag in the subject.
If you get an ack, then I'm happy to  to sign and upload.

diff -u krb5-1.8.3+dfsg/src/lib/krb5/os/sn2princ.c 
krb5-1.8.3+dfsg/src/lib/krb5/os/sn2princ.c
--- krb5-1.8.3+dfsg/src/lib/krb5/os/sn2princ.c
+++ krb5-1.8.3+dfsg/src/lib/krb5/os/sn2princ.c
@@ -111,19 +111,12 @@
hostnames associated.  */
 
 memset(hints, 0, sizeof(hints));
-hints.ai_family = AF_INET;
 hints.ai_flags = AI_CANONNAME;
-try_getaddrinfo_again:
 err = getaddrinfo(hostname, 0, hints, ai);
 if (err) {
 #ifdef DEBUG_REFERRALS
 printf(sname_to_princ: probably punting due to bad hostname 
of %s\n,hostname);
 #endif
-if (hints.ai_family == AF_INET) {
-/* Just in case it's an IPv6-only name.  */
-hints.ai_family = 0;
-goto try_getaddrinfo_again;
-}
 return KRB5_ERR_BAD_HOSTNAME;
 }
 remote_host = strdup(ai-ai_canonname ? ai-ai_canonname : 
hostname);
diff -u krb5-1.8.3+dfsg/debian/changelog krb5-1.8.3+dfsg/debian/changelog
--- krb5-1.8.3+dfsg/debian/changelog
+++ krb5-1.8.3+dfsg/debian/changelog
@@ -1,3 +1,11 @@
+krb5 (1.8.3+dfsg-4squeeze7) stable-proposed-updates; urgency=high
+
+  * CVE-2013-1416 TGS-REQ null pointer dereference in KDC, Closes: #704775
+  * CVE-2013-1415 KDC null pointer dereference with PKINIT
+  * Import upstream's workaround for a getaddrinfo bug, Closes: #704647
+
+ -- Benjamin Kaduk ka...@mit.edu  Sun, 21 Apr 2013 15:49:14 -0400
+
 krb5 (1.8.3+dfsg-4squeeze6) stable-security; urgency=high
 
   * MITKRB5-SA-2012-001 CVE-2012-1015: KDC frees uninitialized pointer
only in patch2:
unchanged:
--- krb5-1.8.3+dfsg.orig/src/plugins/preauth/pkinit/pkinit_crypto_openssl.c
+++ krb5-1.8.3+dfsg/src/plugins/preauth/pkinit/pkinit_crypto_openssl.c
@@ -2879,7 +2879,7 @@
 pkiDebug(found kdcPkId in AS REQ\n);
 is = d2i_PKCS7_ISSUER_AND_SERIAL(NULL, p, (int)pkid_len);
 if (is == NULL)
-goto cleanup;
+return retval;
 
 status = X509_NAME_cmp(X509_get_issuer_name(kdc_cert), is-issuer);
 if (!status) {
@@ -2889,7 +2889,6 @@
 }
 
 retval = 0;
-cleanup:
 X509_NAME_free(is-issuer);
 ASN1_INTEGER_free(is-serial);
 free(is);
only in patch2:
unchanged:
--- krb5-1.8.3+dfsg.orig/src/kdc/do_tgs_req.c
+++ krb5-1.8.3+dfsg/src/kdc/do_tgs_req.c
@@ -1216,7 +1216,8 @@
 retval = ENOMEM;
 goto cleanup;
 }
-strlcpy(comp1_str,comp1-data,comp1-length+1);
+if (comp1-data != NULL)
+memcpy(comp1_str, comp1-data, comp1-length);
 
 if ((krb5_princ_type(kdc_context, request-server) == KRB5_NT_SRV_HST 
||
  krb5_princ_type(kdc_context, request-server) == KRB5_NT_SRV_INST 
||
@@ -1239,7 +1240,8 @@
 retval = ENOMEM;
 goto cleanup;
 }
-strlcpy(temp_buf, comp2-data,comp2-length+1);
+if (comp2-data != NULL)
+memcpy(temp_buf, comp2-data, comp2-length);
 retval = krb5int_get_domain_realm_mapping(kdc_context, temp_buf, 
realms);
 free(temp_buf);
 if (retval) {